instruction
stringclasses
100 values
code
stringlengths
78
193k
response
stringlengths
259
170k
file
stringlengths
59
203
Provide docstrings following PEP 257
from __future__ import annotations from typing import TYPE_CHECKING from scrapy import Request, Spider, signals from scrapy.utils.decorators import _warn_spider_arg from scrapy.utils.deprecate import warn_on_deprecated_spider_attribute if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.http import Response class UserAgentMiddleware: def __init__(self, user_agent: str = "Scrapy"): self.user_agent = user_agent @classmethod def from_crawler(cls, crawler: Crawler) -> Self: o = cls(crawler.settings["USER_AGENT"]) crawler.signals.connect(o.spider_opened, signal=signals.spider_opened) return o def spider_opened(self, spider: Spider) -> None: if hasattr(spider, "user_agent"): # pragma: no cover warn_on_deprecated_spider_attribute("user_agent", "USER_AGENT") self.user_agent = getattr(spider, "user_agent", self.user_agent) @_warn_spider_arg def process_request( self, request: Request, spider: Spider | None = None ) -> Request | Response | None: if self.user_agent: request.headers.setdefault(b"User-Agent", self.user_agent) return None
--- +++ @@ -1,3 +1,4 @@+"""Set User-Agent header per spider or use a default value from settings""" from __future__ import annotations @@ -16,6 +17,7 @@ class UserAgentMiddleware: + """This middleware allows spiders to override the user_agent""" def __init__(self, user_agent: str = "Scrapy"): self.user_agent = user_agent @@ -38,4 +40,4 @@ ) -> Request | Response | None: if self.user_agent: request.headers.setdefault(b"User-Agent", self.user_agent) - return None+ return None
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/downloadermiddlewares/useragent.py
Add professional docstrings to my codebase
from __future__ import annotations from logging import Logger, getLogger from typing import TYPE_CHECKING from scrapy.exceptions import NotConfigured from scrapy.utils.decorators import _warn_spider_arg from scrapy.utils.misc import load_object from scrapy.utils.python import global_object_name from scrapy.utils.response import response_status_message if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.http import Response from scrapy.http.request import Request from scrapy.settings import BaseSettings from scrapy.spiders import Spider retry_logger = getLogger(__name__) def get_retry_request( request: Request, *, spider: Spider, reason: str | Exception | type[Exception] = "unspecified", max_retry_times: int | None = None, priority_adjust: int | None = None, logger: Logger = retry_logger, stats_base_key: str = "retry", ) -> Request | None: settings = spider.crawler.settings assert spider.crawler.stats stats = spider.crawler.stats retry_times = request.meta.get("retry_times", 0) + 1 if max_retry_times is None: max_retry_times = request.meta.get("max_retry_times") if max_retry_times is None: max_retry_times = settings.getint("RETRY_TIMES") if retry_times <= max_retry_times: logger.debug( "Retrying %(request)s (failed %(retry_times)d times): %(reason)s", {"request": request, "retry_times": retry_times, "reason": reason}, extra={"spider": spider}, ) new_request: Request = request.copy() new_request.meta["retry_times"] = retry_times new_request.dont_filter = True if priority_adjust is None: priority_adjust = settings.getint("RETRY_PRIORITY_ADJUST") new_request.priority = request.priority + priority_adjust if callable(reason): reason = reason() if isinstance(reason, Exception): reason = global_object_name(reason.__class__) stats.inc_value(f"{stats_base_key}/count") stats.inc_value(f"{stats_base_key}/reason_count/{reason}") return new_request stats.inc_value(f"{stats_base_key}/max_reached") logger.error( "Gave up retrying %(request)s (failed %(retry_times)d times): %(reason)s", {"request": request, "retry_times": retry_times, "reason": reason}, extra={"spider": spider}, ) return None class RetryMiddleware: crawler: Crawler def __init__(self, settings: BaseSettings): if not settings.getbool("RETRY_ENABLED"): raise NotConfigured self.max_retry_times = settings.getint("RETRY_TIMES") self.retry_http_codes = {int(x) for x in settings.getlist("RETRY_HTTP_CODES")} self.priority_adjust = settings.getint("RETRY_PRIORITY_ADJUST") self.exceptions_to_retry = tuple( load_object(x) if isinstance(x, str) else x for x in settings.getlist("RETRY_EXCEPTIONS") ) @classmethod def from_crawler(cls, crawler: Crawler) -> Self: o = cls(crawler.settings) o.crawler = crawler return o @_warn_spider_arg def process_response( self, request: Request, response: Response, spider: Spider | None = None ) -> Request | Response: if request.meta.get("dont_retry", False): return response if response.status in self.retry_http_codes: reason = response_status_message(response.status) return self._retry(request, reason) or response return response @_warn_spider_arg def process_exception( self, request: Request, exception: Exception, spider: Spider | None = None ) -> Request | Response | None: if isinstance(exception, self.exceptions_to_retry) and not request.meta.get( "dont_retry", False ): return self._retry(request, exception) return None def _retry( self, request: Request, reason: str | Exception | type[Exception] ) -> Request | None: max_retry_times = request.meta.get("max_retry_times", self.max_retry_times) priority_adjust = request.meta.get("priority_adjust", self.priority_adjust) assert self.crawler.spider return get_retry_request( request, reason=reason, spider=self.crawler.spider, max_retry_times=max_retry_times, priority_adjust=priority_adjust, )
--- +++ @@ -1,3 +1,14 @@+""" +An extension to retry failed requests that are potentially caused by temporary +problems such as a connection timeout or HTTP 500 error. + +You can change the behaviour of this middleware by modifying the scraping settings: +RETRY_TIMES - how many times to retry a failed page +RETRY_HTTP_CODES - which HTTP response codes to retry + +Failed pages are collected on the scraping process and rescheduled at the end, +once the spider has finished crawling all regular (non-failed) pages. +""" from __future__ import annotations @@ -34,6 +45,46 @@ logger: Logger = retry_logger, stats_base_key: str = "retry", ) -> Request | None: + """ + Returns a new :class:`~scrapy.Request` object to retry the specified + request, or ``None`` if retries of the specified request have been + exhausted. + + For example, in a :class:`~scrapy.Spider` callback, you could use it as + follows:: + + def parse(self, response): + if not response.text: + new_request_or_none = get_retry_request( + response.request, + spider=self, + reason='empty', + ) + return new_request_or_none + + *spider* is the :class:`~scrapy.Spider` instance which is asking for the + retry request. It is used to access the :ref:`settings <topics-settings>` + and :ref:`stats <topics-stats>`, and to provide extra logging context (see + :func:`logging.debug`). + + *reason* is a string or an :class:`Exception` object that indicates the + reason why the request needs to be retried. It is used to name retry stats. + + *max_retry_times* is a number that determines the maximum number of times + that *request* can be retried. If not specified or ``None``, the number is + read from the :reqmeta:`max_retry_times` meta key of the request. If the + :reqmeta:`max_retry_times` meta key is not defined or ``None``, the number + is read from the :setting:`RETRY_TIMES` setting. + + *priority_adjust* is a number that determines how the priority of the new + request changes in relation to *request*. If not specified, the number is + read from the :setting:`RETRY_PRIORITY_ADJUST` setting. + + *logger* is the logging.Logger object to be used when logging messages + + *stats_base_key* is a string to be used as the base key for the + retry-related job stats + """ settings = spider.crawler.settings assert spider.crawler.stats stats = spider.crawler.stats @@ -125,4 +176,4 @@ spider=self.crawler.spider, max_retry_times=max_retry_times, priority_adjust=priority_adjust, - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/downloadermiddlewares/retry.py
Create docstrings for each class method
from __future__ import annotations import re import time from http.cookiejar import Cookie, CookiePolicy, DefaultCookiePolicy from http.cookiejar import CookieJar as _CookieJar from typing import TYPE_CHECKING, Any, cast from scrapy.utils.httpobj import urlparse_cached from scrapy.utils.python import to_unicode if TYPE_CHECKING: from collections.abc import Iterator, Sequence # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Request from scrapy.http import Response # Defined in the http.cookiejar module, but undocumented: # https://github.com/python/cpython/blob/v3.9.0/Lib/http/cookiejar.py#L527 IPV4_RE = re.compile(r"\.\d+$", re.ASCII) class CookieJar: def __init__( self, policy: CookiePolicy | None = None, check_expired_frequency: int = 10000, ): self.policy: CookiePolicy = policy or DefaultCookiePolicy() self.jar: _CookieJar = _CookieJar(self.policy) self.jar._cookies_lock = _DummyLock() # type: ignore[attr-defined] self.check_expired_frequency: int = check_expired_frequency self.processed: int = 0 def extract_cookies(self, response: Response, request: Request) -> None: wreq = WrappedRequest(request) wrsp = WrappedResponse(response) self.jar.extract_cookies(wrsp, wreq) # type: ignore[arg-type] def add_cookie_header(self, request: Request) -> None: wreq = WrappedRequest(request) self.policy._now = self.jar._now = int(time.time()) # type: ignore[attr-defined] # the cookiejar implementation iterates through all domains # instead we restrict to potential matches on the domain req_host = urlparse_cached(request).hostname if not req_host: return if not IPV4_RE.search(req_host): hosts = potential_domain_matches(req_host) if "." not in req_host: hosts += [req_host + ".local"] else: hosts = [req_host] cookies = [] for host in hosts: if host in self.jar._cookies: # type: ignore[attr-defined] cookies += self.jar._cookies_for_domain(host, wreq) # type: ignore[attr-defined] attrs = self.jar._cookie_attrs(cookies) # type: ignore[attr-defined] if attrs and not wreq.has_header("Cookie"): wreq.add_unredirected_header("Cookie", "; ".join(attrs)) self.processed += 1 if self.processed % self.check_expired_frequency == 0: # This is still quite inefficient for large number of cookies self.jar.clear_expired_cookies() @property def _cookies(self) -> dict[str, dict[str, dict[str, Cookie]]]: return self.jar._cookies # type: ignore[attr-defined] def clear_session_cookies(self) -> None: return self.jar.clear_session_cookies() def clear( self, domain: str | None = None, path: str | None = None, name: str | None = None, ) -> None: self.jar.clear(domain, path, name) def __iter__(self) -> Iterator[Cookie]: return iter(self.jar) def __len__(self) -> int: return len(self.jar) def set_policy(self, pol: CookiePolicy) -> None: self.jar.set_policy(pol) def make_cookies(self, response: Response, request: Request) -> Sequence[Cookie]: wreq = WrappedRequest(request) wrsp = WrappedResponse(response) return self.jar.make_cookies(wrsp, wreq) # type: ignore[arg-type] def set_cookie(self, cookie: Cookie) -> None: self.jar.set_cookie(cookie) def set_cookie_if_ok(self, cookie: Cookie, request: Request) -> None: self.jar.set_cookie_if_ok(cookie, WrappedRequest(request)) # type: ignore[arg-type] def potential_domain_matches(domain: str) -> list[str]: matches = [domain] try: start = domain.index(".") + 1 end = domain.rindex(".") while start < end: matches.append(domain[start:]) start = domain.index(".", start) + 1 except ValueError: pass return matches + ["." + d for d in matches] class _DummyLock: def acquire(self) -> None: pass def release(self) -> None: pass class WrappedRequest: def __init__(self, request: Request): self.request = request def get_full_url(self) -> str: return self.request.url def get_host(self) -> str: return urlparse_cached(self.request).netloc def get_type(self) -> str: return urlparse_cached(self.request).scheme def is_unverifiable(self) -> bool: return cast("bool", self.request.meta.get("is_unverifiable", False)) @property def full_url(self) -> str: return self.get_full_url() @property def host(self) -> str: return self.get_host() @property def type(self) -> str: return self.get_type() @property def unverifiable(self) -> bool: return self.is_unverifiable() @property def origin_req_host(self) -> str: return cast("str", urlparse_cached(self.request).hostname) def has_header(self, name: str) -> bool: return name in self.request.headers def get_header(self, name: str, default: str | None = None) -> str | None: value = self.request.headers.get(name, default) return to_unicode(value, errors="replace") if value is not None else None def header_items(self) -> list[tuple[str, list[str]]]: return [ ( to_unicode(k, errors="replace"), [to_unicode(x, errors="replace") for x in v], ) for k, v in self.request.headers.items() ] def add_unredirected_header(self, name: str, value: str) -> None: self.request.headers.appendlist(name, value) class WrappedResponse: def __init__(self, response: Response): self.response = response def info(self) -> Self: return self def get_all(self, name: str, default: Any = None) -> list[str]: return [ to_unicode(v, errors="replace") for v in self.response.headers.getlist(name) ]
--- +++ @@ -109,6 +109,12 @@ def potential_domain_matches(domain: str) -> list[str]: + """Potential domain matches for a cookie + + >>> potential_domain_matches('www.example.com') + ['www.example.com', 'example.com', '.www.example.com', '.example.com'] + + """ matches = [domain] try: start = domain.index(".") + 1 @@ -130,6 +136,10 @@ class WrappedRequest: + """Wraps a scrapy Request class with methods defined by urllib2.Request class to interact with CookieJar class + + see http://docs.python.org/library/urllib2.html#urllib2.Request + """ def __init__(self, request: Request): self.request = request @@ -144,6 +154,13 @@ return urlparse_cached(self.request).scheme def is_unverifiable(self) -> bool: + """Unverifiable should indicate whether the request is unverifiable, as defined by RFC 2965. + + It defaults to False. An unverifiable request is one whose URL the user did not have the + option to approve. For example, if the request is for an image in an + HTML document, and the user had no option to approve the automatic + fetching of the image, this should be true. + """ return cast("bool", self.request.meta.get("is_unverifiable", False)) @property @@ -196,4 +213,4 @@ def get_all(self, name: str, default: Any = None) -> list[str]: return [ to_unicode(v, errors="replace") for v in self.response.headers.getlist(name) - ]+ ]
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/http/cookies.py
Add docstrings with type hints explained
from __future__ import annotations import logging from typing import TYPE_CHECKING, Any, cast from urllib.parse import urljoin, urlparse from w3lib.url import safe_url_string from scrapy import signals from scrapy.exceptions import IgnoreRequest, NotConfigured from scrapy.http import HtmlResponse, Response from scrapy.spidermiddlewares.referer import RefererMiddleware from scrapy.utils.decorators import _warn_spider_arg from scrapy.utils.httpobj import urlparse_cached from scrapy.utils.python import global_object_name from scrapy.utils.response import get_meta_refresh if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Request, Spider from scrapy.crawler import Crawler from scrapy.settings import BaseSettings logger = logging.getLogger(__name__) class BaseRedirectMiddleware: crawler: Crawler enabled_setting: str = "REDIRECT_ENABLED" def __init__(self, settings: BaseSettings): if not settings.getbool(self.enabled_setting): raise NotConfigured self.max_redirect_times: int = settings.getint("REDIRECT_MAX_TIMES") self.priority_adjust: int = settings.getint("REDIRECT_PRIORITY_ADJUST") self._referer_spider_middleware: RefererMiddleware | None = None @classmethod def from_crawler(cls, crawler: Crawler) -> Self: o = cls(crawler.settings) o.crawler = crawler crawler.signals.connect(o._engine_started, signal=signals.engine_started) return o def handle_referer(self, request: Request, response: Response) -> None: request.headers.pop("Referer", None) if not self._referer_spider_middleware: return self._referer_spider_middleware.get_processed_request(request, response) def _engine_started(self) -> None: self._referer_spider_middleware = self.crawler.get_spider_middleware( RefererMiddleware ) if self._referer_spider_middleware: return redirect_cls = global_object_name(self.__class__) referer_cls = global_object_name(RefererMiddleware) if self.__class__ in (RedirectMiddleware, MetaRefreshMiddleware): replacement = ( f"replace {redirect_cls} with a subclass that overrides the " f"handle_referer() method" ) else: replacement = ( f"or edit {redirect_cls} (if defined in your code base) to " f"override the handle_referer() method, or replace " f"{redirect_cls} with a subclass that overrides the " f"handle_referer() method." ) logger.warning( f"{redirect_cls} found no {referer_cls} instance to handle " f"Referer header handling, so the Referer header will be removed " f"on redirects. To set a Referer header on redirects, enable " f"{referer_cls} (or a subclass), or {replacement}.", ) def _redirect(self, redirected: Request, request: Request, reason: Any) -> Request: ttl = request.meta.setdefault("redirect_ttl", self.max_redirect_times) redirects = request.meta.get("redirect_times", 0) + 1 if ttl and redirects <= self.max_redirect_times: redirected.meta["redirect_times"] = redirects redirected.meta["redirect_ttl"] = ttl - 1 redirected.meta["redirect_urls"] = [ *request.meta.get("redirect_urls", []), request.url, ] redirected.meta["redirect_reasons"] = [ *request.meta.get("redirect_reasons", []), reason, ] redirected.dont_filter = request.dont_filter redirected.priority = request.priority + self.priority_adjust logger.debug( "Redirecting (%(reason)s) to %(redirected)s from %(request)s", {"reason": reason, "redirected": redirected, "request": request}, extra={"spider": self.crawler.spider}, ) return redirected logger.debug( "Discarding %(request)s: max redirections reached", {"request": request}, extra={"spider": self.crawler.spider}, ) raise IgnoreRequest("max redirections reached") def _build_redirect_request( self, source_request: Request, response: Response, *, url: str, **kwargs: Any ) -> Request: redirect_request = source_request.replace( url=url, **kwargs, cls=None, cookies=None, ) if "_scheme_proxy" in redirect_request.meta: source_request_scheme = urlparse_cached(source_request).scheme redirect_request_scheme = urlparse_cached(redirect_request).scheme if source_request_scheme != redirect_request_scheme: redirect_request.meta.pop("_scheme_proxy") redirect_request.meta.pop("proxy", None) redirect_request.meta.pop("_auth_proxy", None) redirect_request.headers.pop(b"Proxy-Authorization", None) has_cookie_header = "Cookie" in redirect_request.headers has_authorization_header = "Authorization" in redirect_request.headers if has_cookie_header or has_authorization_header: default_ports = {"http": 80, "https": 443} parsed_source_request = urlparse_cached(source_request) source_scheme, source_host, source_port = ( parsed_source_request.scheme, parsed_source_request.hostname, parsed_source_request.port or default_ports.get(parsed_source_request.scheme), ) parsed_redirect_request = urlparse_cached(redirect_request) redirect_scheme, redirect_host, redirect_port = ( parsed_redirect_request.scheme, parsed_redirect_request.hostname, parsed_redirect_request.port or default_ports.get(parsed_redirect_request.scheme), ) if has_cookie_header and ( redirect_scheme not in {source_scheme, "https"} or source_host != redirect_host ): del redirect_request.headers["Cookie"] # https://fetch.spec.whatwg.org/#ref-for-cors-non-wildcard-request-header-name if has_authorization_header and ( source_scheme != redirect_scheme or source_host != redirect_host or source_port != redirect_port ): del redirect_request.headers["Authorization"] self.handle_referer(redirect_request, response) return redirect_request def _redirect_request_using_get( self, request: Request, response: Response, redirect_url: str ) -> Request: redirect_request = self._build_redirect_request( request, response, url=redirect_url, method="GET", body="", ) redirect_request.headers.pop("Content-Type", None) redirect_request.headers.pop("Content-Length", None) redirect_request.headers.pop("Content-Encoding", None) redirect_request.headers.pop("Content-Language", None) redirect_request.headers.pop("Content-Location", None) return redirect_request class RedirectMiddleware(BaseRedirectMiddleware): @_warn_spider_arg def process_response( self, request: Request, response: Response, spider: Spider | None = None ) -> Request | Response: if ( request.meta.get("dont_redirect", False) or response.status in getattr(self.crawler.spider, "handle_httpstatus_list", []) or response.status in request.meta.get("handle_httpstatus_list", []) or request.meta.get("handle_httpstatus_all", False) ): return response allowed_status = (301, 302, 303, 307, 308) if "Location" not in response.headers or response.status not in allowed_status: return response assert response.headers["Location"] is not None location = safe_url_string(response.headers["Location"]) if response.headers["Location"].startswith(b"//"): request_scheme = urlparse_cached(request).scheme location = request_scheme + "://" + location.lstrip("/") redirected_url = urljoin(request.url, location) if not urlparse(redirected_url).fragment: fragment = urlparse_cached(request).fragment if fragment: redirected_url = urljoin(redirected_url, f"#{fragment}") redirected = self._build_redirect_request(request, response, url=redirected_url) if urlparse_cached(redirected).scheme not in {"http", "https"}: return response if (response.status in (301, 302) and request.method == "POST") or ( response.status == 303 and request.method not in ("GET", "HEAD") ): redirected = self._redirect_request_using_get( request, response, redirected_url ) return self._redirect(redirected, request, response.status) class MetaRefreshMiddleware(BaseRedirectMiddleware): enabled_setting = "METAREFRESH_ENABLED" def __init__(self, settings: BaseSettings): super().__init__(settings) self._ignore_tags: list[str] = settings.getlist("METAREFRESH_IGNORE_TAGS") self._maxdelay: int = settings.getint("METAREFRESH_MAXDELAY") @_warn_spider_arg def process_response( self, request: Request, response: Response, spider: Spider | None = None ) -> Request | Response: if ( request.meta.get("dont_redirect", False) or request.method == "HEAD" or not isinstance(response, HtmlResponse) or urlparse_cached(request).scheme not in {"http", "https"} ): return response interval, url = get_meta_refresh(response, ignore_tags=self._ignore_tags) if not url: return response redirected = self._redirect_request_using_get(request, response, url) if urlparse_cached(redirected).scheme not in {"http", "https"}: return response if cast("float", interval) < self._maxdelay: return self._redirect(redirected, request, "meta refresh") return response
--- +++ @@ -47,6 +47,17 @@ return o def handle_referer(self, request: Request, response: Response) -> None: + """Remove, modify or keep the Referer header of *request* based on the + *response* that triggered *request*. + + By default, this method finds a run-time instance of + scrapy.spidermiddlewares.referer.RefererMiddleware (or of a subclass) + and uses it to set the right Referer header. + + Override this method if you use a different Scrapy component to handle + Referer headers, of if you want to use a custom logic to set the + Referer header on redirects. + """ request.headers.pop("Referer", None) if not self._referer_spider_middleware: return @@ -185,6 +196,10 @@ class RedirectMiddleware(BaseRedirectMiddleware): + """ + Handle redirection of requests based on response status + and meta-refresh html tag. + """ @_warn_spider_arg def process_response( @@ -258,4 +273,4 @@ return response if cast("float", interval) < self._maxdelay: return self._redirect(redirected, request, "meta refresh") - return response+ return response
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/downloadermiddlewares/redirect.py
Insert docstrings into my code
from __future__ import annotations import asyncio import contextlib import logging import re import sys import warnings from abc import ABC, abstractmethod from collections.abc import Callable, Coroutine from datetime import datetime, timezone from pathlib import Path, PureWindowsPath from tempfile import NamedTemporaryFile from typing import IO, TYPE_CHECKING, Any, Protocol, TypeAlias, cast from urllib.parse import unquote, urlparse from twisted.internet.defer import Deferred, DeferredList from twisted.internet.threads import deferToThread from w3lib.url import file_uri_to_path from zope.interface import Interface, implementer from scrapy import Spider, signals from scrapy.exceptions import NotConfigured, ScrapyDeprecationWarning from scrapy.extensions.postprocessing import PostProcessingManager from scrapy.utils.asyncio import is_asyncio_available from scrapy.utils.conf import feed_complete_default_values_from_settings from scrapy.utils.defer import deferred_from_coro, ensure_awaitable from scrapy.utils.ftp import ftp_store_file from scrapy.utils.misc import build_from_crawler, load_object from scrapy.utils.python import without_none_values if TYPE_CHECKING: from _typeshed import OpenBinaryMode # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.exporters import BaseItemExporter from scrapy.settings import BaseSettings, Settings logger = logging.getLogger(__name__) UriParamsCallableT: TypeAlias = Callable[ [dict[str, Any], Spider], dict[str, Any] | None ] class ItemFilter: feed_options: dict[str, Any] | None item_classes: tuple[type, ...] def __init__(self, feed_options: dict[str, Any] | None) -> None: self.feed_options = feed_options if feed_options is not None: self.item_classes = tuple( load_object(item_class) for item_class in feed_options.get("item_classes") or () ) else: self.item_classes = () def accepts(self, item: Any) -> bool: if self.item_classes: return isinstance(item, self.item_classes) return True # accept all items by default class IFeedStorage(Interface): # type: ignore[misc] # pylint: disable=no-self-argument def __init__(uri, *, feed_options=None): # type: ignore[no-untyped-def] # pylint: disable=super-init-not-called def open(spider): # type: ignore[no-untyped-def] def store(file): # type: ignore[no-untyped-def] class FeedStorageProtocol(Protocol): def __init__(self, uri: str, *, feed_options: dict[str, Any] | None = None): def open(self, spider: Spider) -> IO[bytes]: def store(self, file: IO[bytes]) -> Deferred[None] | None: @implementer(IFeedStorage) class BlockingFeedStorage(ABC): def open(self, spider: Spider) -> IO[bytes]: path = spider.crawler.settings["FEED_TEMPDIR"] if path and not Path(path).is_dir(): raise OSError("Not a Directory: " + str(path)) return NamedTemporaryFile(prefix="feed-", dir=path) def store(self, file: IO[bytes]) -> Deferred[None] | None: return deferToThread(self._store_in_thread, file) @abstractmethod def _store_in_thread(self, file: IO[bytes]) -> None: raise NotImplementedError @implementer(IFeedStorage) class StdoutFeedStorage: def __init__( self, uri: str, _stdout: IO[bytes] | None = None, *, feed_options: dict[str, Any] | None = None, ): if not _stdout: _stdout = sys.stdout.buffer self._stdout: IO[bytes] = _stdout if feed_options and feed_options.get("overwrite", False) is True: logger.warning( "Standard output (stdout) storage does not support " "overwriting. To suppress this warning, remove the " "overwrite option from your FEEDS setting, or set " "it to False." ) def open(self, spider: Spider) -> IO[bytes]: return self._stdout def store(self, file: IO[bytes]) -> Deferred[None] | None: pass @implementer(IFeedStorage) class FileFeedStorage: def __init__(self, uri: str, *, feed_options: dict[str, Any] | None = None): self.path: str = file_uri_to_path(uri) if uri.startswith("file://") else uri feed_options = feed_options or {} self.write_mode: OpenBinaryMode = ( "wb" if feed_options.get("overwrite", False) else "ab" ) def open(self, spider: Spider) -> IO[bytes]: dirname = Path(self.path).parent if dirname and not dirname.exists(): dirname.mkdir(parents=True) return Path(self.path).open(self.write_mode) def store(self, file: IO[bytes]) -> Deferred[None] | None: file.close() return None class S3FeedStorage(BlockingFeedStorage): def __init__( self, uri: str, access_key: str | None = None, secret_key: str | None = None, acl: str | None = None, endpoint_url: str | None = None, *, feed_options: dict[str, Any] | None = None, session_token: str | None = None, region_name: str | None = None, ): try: import boto3.session # noqa: PLC0415 except ImportError: raise NotConfigured("missing boto3 library") u = urlparse(uri) assert u.hostname self.bucketname: str = u.hostname self.access_key: str | None = u.username or access_key self.secret_key: str | None = u.password or secret_key self.session_token: str | None = session_token self.keyname: str = u.path[1:] # remove first "/" self.acl: str | None = acl self.endpoint_url: str | None = endpoint_url self.region_name: str | None = region_name boto3_session = boto3.session.Session() self.s3_client = boto3_session.client( "s3", aws_access_key_id=self.access_key, aws_secret_access_key=self.secret_key, aws_session_token=self.session_token, endpoint_url=self.endpoint_url, region_name=self.region_name, ) if feed_options and feed_options.get("overwrite", True) is False: logger.warning( "S3 does not support appending to files. To " "suppress this warning, remove the overwrite " "option from your FEEDS setting or set it to True." ) @classmethod def from_crawler( cls, crawler: Crawler, uri: str, *, feed_options: dict[str, Any] | None = None, ) -> Self: return cls( uri, access_key=crawler.settings["AWS_ACCESS_KEY_ID"], secret_key=crawler.settings["AWS_SECRET_ACCESS_KEY"], session_token=crawler.settings["AWS_SESSION_TOKEN"], acl=crawler.settings["FEED_STORAGE_S3_ACL"] or None, endpoint_url=crawler.settings["AWS_ENDPOINT_URL"] or None, region_name=crawler.settings["AWS_REGION_NAME"] or None, feed_options=feed_options, ) def _store_in_thread(self, file: IO[bytes]) -> None: file.seek(0) kwargs: dict[str, Any] = {"ExtraArgs": {"ACL": self.acl}} if self.acl else {} self.s3_client.upload_fileobj( Bucket=self.bucketname, Key=self.keyname, Fileobj=file, **kwargs ) file.close() class GCSFeedStorage(BlockingFeedStorage): def __init__( self, uri: str, project_id: str | None, acl: str | None, *, feed_options: dict[str, Any] | None = None, ): self.project_id: str | None = project_id self.acl: str | None = acl u = urlparse(uri) assert u.hostname self.bucket_name: str = u.hostname self.blob_name: str = u.path[1:] # remove first "/" if feed_options and feed_options.get("overwrite", True) is False: logger.warning( "GCS does not support appending to files. To " "suppress this warning, remove the overwrite " "option from your FEEDS setting or set it to True." ) @classmethod def from_crawler( cls, crawler: Crawler, uri: str, *, feed_options: dict[str, Any] | None = None, ) -> Self: return cls( uri, crawler.settings["GCS_PROJECT_ID"], crawler.settings["FEED_STORAGE_GCS_ACL"] or None, feed_options=feed_options, ) def _store_in_thread(self, file: IO[bytes]) -> None: file.seek(0) from google.cloud.storage import Client # noqa: PLC0415 client = Client(project=self.project_id) bucket = client.get_bucket(self.bucket_name) blob = bucket.blob(self.blob_name) blob.upload_from_file(file, predefined_acl=self.acl) class FTPFeedStorage(BlockingFeedStorage): def __init__( self, uri: str, use_active_mode: bool = False, *, feed_options: dict[str, Any] | None = None, ): u = urlparse(uri) if not u.hostname: raise ValueError(f"Got a storage URI without a hostname: {uri}") self.host: str = u.hostname self.port: int = int(u.port or "21") self.username: str = u.username or "" self.password: str = unquote(u.password or "") self.path: str = u.path self.use_active_mode: bool = use_active_mode self.overwrite: bool = not feed_options or feed_options.get("overwrite", True) @classmethod def from_crawler( cls, crawler: Crawler, uri: str, *, feed_options: dict[str, Any] | None = None, ) -> Self: return cls( uri, use_active_mode=crawler.settings.getbool("FEED_STORAGE_FTP_ACTIVE"), feed_options=feed_options, ) def _store_in_thread(self, file: IO[bytes]) -> None: ftp_store_file( path=self.path, file=file, host=self.host, port=self.port, username=self.username, password=self.password, use_active_mode=self.use_active_mode, overwrite=self.overwrite, ) class FeedSlot: def __init__( self, storage: FeedStorageProtocol, uri: str, format: str, # noqa: A002 store_empty: bool, batch_id: int, uri_template: str, filter: ItemFilter, # noqa: A002 feed_options: dict[str, Any], spider: Spider, exporters: dict[str, type[BaseItemExporter]], settings: BaseSettings, crawler: Crawler, ): self.file: IO[bytes] | None = None self.exporter: BaseItemExporter | None = None self.storage: FeedStorageProtocol = storage # feed params self.batch_id: int = batch_id self.format: str = format self.store_empty: bool = store_empty self.uri_template: str = uri_template self.uri: str = uri self.filter: ItemFilter = filter # exporter params self.feed_options: dict[str, Any] = feed_options self.spider: Spider = spider self.exporters: dict[str, type[BaseItemExporter]] = exporters self.settings: BaseSettings = settings self.crawler: Crawler = crawler # flags self.itemcount: int = 0 self._exporting: bool = False self._fileloaded: bool = False def start_exporting(self) -> None: if not self._fileloaded: self.file = self.storage.open(self.spider) if "postprocessing" in self.feed_options: self.file = cast( "IO[bytes]", PostProcessingManager( self.feed_options["postprocessing"], self.file, self.feed_options, ), ) self.exporter = self._get_exporter( file=self.file, format_=self.feed_options["format"], fields_to_export=self.feed_options["fields"], encoding=self.feed_options["encoding"], indent=self.feed_options["indent"], **self.feed_options["item_export_kwargs"], ) self._fileloaded = True if not self._exporting: assert self.exporter self.exporter.start_exporting() self._exporting = True def _get_exporter( self, file: IO[bytes], format_: str, *args: Any, **kwargs: Any ) -> BaseItemExporter: return build_from_crawler( self.exporters[format_], self.crawler, file, *args, **kwargs ) def finish_exporting(self) -> None: if self._exporting: assert self.exporter self.exporter.finish_exporting() self._exporting = False class FeedExporter: @classmethod def from_crawler(cls, crawler: Crawler) -> Self: exporter = cls(crawler) crawler.signals.connect(exporter.open_spider, signals.spider_opened) crawler.signals.connect(exporter.close_spider, signals.spider_closed) crawler.signals.connect(exporter.item_scraped, signals.item_scraped) return exporter def __init__(self, crawler: Crawler): self.crawler: Crawler = crawler self.settings: Settings = crawler.settings self.feeds = {} self.slots: list[FeedSlot] = [] self.filters: dict[str, ItemFilter] = {} self._pending_close_coros: list[Coroutine[Any, Any, None]] = [] if not self.settings["FEEDS"] and not self.settings["FEED_URI"]: raise NotConfigured # Begin: Backward compatibility for FEED_URI and FEED_FORMAT settings if self.settings["FEED_URI"]: warnings.warn( "The `FEED_URI` and `FEED_FORMAT` settings have been deprecated in favor of " "the `FEEDS` setting. Please see the `FEEDS` setting docs for more details", category=ScrapyDeprecationWarning, stacklevel=2, ) uri = self.settings["FEED_URI"] # handle pathlib.Path objects uri = str(uri) if not isinstance(uri, Path) else uri.absolute().as_uri() feed_options = {"format": self.settings["FEED_FORMAT"]} self.feeds[uri] = feed_complete_default_values_from_settings( feed_options, self.settings ) self.filters[uri] = self._load_filter(feed_options) # End: Backward compatibility for FEED_URI and FEED_FORMAT settings # 'FEEDS' setting takes precedence over 'FEED_URI' for uri, feed_options in self.settings.getdict("FEEDS").items(): # handle pathlib.Path objects uri = str(uri) if not isinstance(uri, Path) else uri.absolute().as_uri() self.feeds[uri] = feed_complete_default_values_from_settings( feed_options, self.settings ) self.filters[uri] = self._load_filter(feed_options) self.storages: dict[str, type[FeedStorageProtocol]] = self._load_components( "FEED_STORAGES" ) self.exporters: dict[str, type[BaseItemExporter]] = self._load_components( "FEED_EXPORTERS" ) for uri, feed_options in self.feeds.items(): if not self._storage_supported(uri, feed_options): raise NotConfigured if not self._settings_are_valid(): raise NotConfigured if not self._exporter_supported(feed_options["format"]): raise NotConfigured def open_spider(self, spider: Spider) -> None: for uri, feed_options in self.feeds.items(): uri_params = self._get_uri_params(spider, feed_options["uri_params"]) self.slots.append( self._start_new_batch( batch_id=1, uri=uri % uri_params, feed_options=feed_options, spider=spider, uri_template=uri, ) ) async def close_spider(self, spider: Spider) -> None: self._pending_close_coros.extend( self._close_slot(slot, spider) for slot in self.slots ) if self._pending_close_coros: if is_asyncio_available(): await asyncio.wait( [asyncio.create_task(coro) for coro in self._pending_close_coros] ) else: await DeferredList( deferred_from_coro(coro) for coro in self._pending_close_coros ) # Send FEED_EXPORTER_CLOSED signal await self.crawler.signals.send_catch_log_async(signals.feed_exporter_closed) async def _close_slot(self, slot: FeedSlot, spider: Spider) -> None: def get_file(slot_: FeedSlot) -> IO[bytes]: assert slot_.file if isinstance(slot_.file, PostProcessingManager): slot_.file.close() return slot_.file.file return slot_.file if slot.itemcount: # Normal case slot.finish_exporting() elif slot.store_empty and slot.batch_id == 1: # Need to store the empty file slot.start_exporting() slot.finish_exporting() else: # In this case, the file is not stored, so no processing is required. return logmsg = f"{slot.format} feed ({slot.itemcount} items) in: {slot.uri}" slot_type = type(slot.storage).__name__ assert self.crawler.stats try: await ensure_awaitable(slot.storage.store(get_file(slot))) except Exception: logger.error( "Error storing %s", logmsg, exc_info=True, extra={"spider": spider}, ) self.crawler.stats.inc_value(f"feedexport/failed_count/{slot_type}") else: logger.info("Stored %s", logmsg, extra={"spider": spider}) self.crawler.stats.inc_value(f"feedexport/success_count/{slot_type}") await self.crawler.signals.send_catch_log_async( signals.feed_slot_closed, slot=slot ) def _start_new_batch( self, batch_id: int, uri: str, feed_options: dict[str, Any], spider: Spider, uri_template: str, ) -> FeedSlot: storage = self._get_storage(uri, feed_options) return FeedSlot( storage=storage, uri=uri, format=feed_options["format"], store_empty=feed_options["store_empty"], batch_id=batch_id, uri_template=uri_template, filter=self.filters[uri_template], feed_options=feed_options, spider=spider, exporters=self.exporters, settings=self.settings, crawler=self.crawler, ) def item_scraped(self, item: Any, spider: Spider) -> None: slots = [] for slot in self.slots: if not slot.filter.accepts(item): slots.append( slot ) # if slot doesn't accept item, continue with next slot continue slot.start_exporting() assert slot.exporter slot.exporter.export_item(item) slot.itemcount += 1 # create new slot for each slot with itemcount == FEED_EXPORT_BATCH_ITEM_COUNT and close the old one if ( self.feeds[slot.uri_template]["batch_item_count"] and slot.itemcount >= self.feeds[slot.uri_template]["batch_item_count"] ): uri_params = self._get_uri_params( spider, self.feeds[slot.uri_template]["uri_params"], slot ) self._pending_close_coros.append(self._close_slot(slot, spider)) slots.append( self._start_new_batch( batch_id=slot.batch_id + 1, uri=slot.uri_template % uri_params, feed_options=self.feeds[slot.uri_template], spider=spider, uri_template=slot.uri_template, ) ) else: slots.append(slot) self.slots = slots def _load_components(self, setting_prefix: str) -> dict[str, Any]: conf = without_none_values( cast("dict[str, str]", self.settings.getwithbase(setting_prefix)) ) d = {} for k, v in conf.items(): with contextlib.suppress(NotConfigured): d[k] = load_object(v) return d def _exporter_supported(self, format_: str) -> bool: if format_ in self.exporters: return True logger.error("Unknown feed format: %(format)s", {"format": format_}) return False def _settings_are_valid(self) -> bool: for uri_template, values in self.feeds.items(): if values["batch_item_count"] and not re.search( r"%\(batch_time\)s|%\(batch_id\)", uri_template ): logger.error( "%%(batch_time)s or %%(batch_id)d must be in the feed URI (%s) if FEED_EXPORT_BATCH_ITEM_COUNT " "setting or FEEDS.batch_item_count is specified and greater than 0. For more info see: " "https://docs.scrapy.org/en/latest/topics/feed-exports.html#feed-export-batch-item-count", uri_template, ) return False return True def _storage_supported(self, uri: str, feed_options: dict[str, Any]) -> bool: scheme = urlparse(uri).scheme if scheme in self.storages or PureWindowsPath(uri).drive: try: self._get_storage(uri, feed_options) return True except NotConfigured as e: logger.error( "Disabled feed storage scheme: %(scheme)s. Reason: %(reason)s", {"scheme": scheme, "reason": str(e)}, ) else: logger.error("Unknown feed storage scheme: %(scheme)s", {"scheme": scheme}) return False def _get_storage( self, uri: str, feed_options: dict[str, Any] ) -> FeedStorageProtocol: cls = self.storages.get(urlparse(uri).scheme, self.storages["file"]) return build_from_crawler(cls, self.crawler, uri, feed_options=feed_options) def _get_uri_params( self, spider: Spider, uri_params_function: str | UriParamsCallableT | None, slot: FeedSlot | None = None, ) -> dict[str, Any]: params = {} for k in dir(spider): params[k] = getattr(spider, k) utc_now = datetime.now(tz=timezone.utc) params["time"] = utc_now.replace(microsecond=0).isoformat().replace(":", "-") params["batch_time"] = utc_now.isoformat().replace(":", "-") params["batch_id"] = slot.batch_id + 1 if slot is not None else 1 uripar_function: UriParamsCallableT = ( load_object(uri_params_function) if uri_params_function else lambda params, _: params ) new_params = uripar_function(params, spider) return new_params if new_params is not None else params def _load_filter(self, feed_options: dict[str, Any]) -> ItemFilter: # load the item filter if declared else load the default filter class item_filter_class: type[ItemFilter] = load_object( feed_options.get("item_filter", ItemFilter) ) return item_filter_class(feed_options)
--- +++ @@ -1,3 +1,8 @@+""" +Feed Exports extension + +See documentation in docs/topics/feed-exports.rst +""" from __future__ import annotations @@ -49,6 +54,13 @@ class ItemFilter: + """ + This will be used by FeedExporter to decide if an item should be allowed + to be exported to a particular feed. + + :param feed_options: feed specific options passed from FeedExporter + :type feed_options: dict + """ feed_options: dict[str, Any] | None item_classes: tuple[type, ...] @@ -64,29 +76,49 @@ self.item_classes = () def accepts(self, item: Any) -> bool: + """ + Return ``True`` if `item` should be exported or ``False`` otherwise. + + :param item: scraped item which user wants to check if is acceptable + :type item: :ref:`Scrapy items <topics-items>` + :return: `True` if accepted, `False` otherwise + :rtype: bool + """ if self.item_classes: return isinstance(item, self.item_classes) return True # accept all items by default class IFeedStorage(Interface): # type: ignore[misc] + """Interface that all Feed Storages must implement""" # pylint: disable=no-self-argument def __init__(uri, *, feed_options=None): # type: ignore[no-untyped-def] # pylint: disable=super-init-not-called + """Initialize the storage with the parameters given in the URI and the + feed-specific options (see :setting:`FEEDS`)""" def open(spider): # type: ignore[no-untyped-def] + """Open the storage for the given spider. It must return a file-like + object that will be used for the exporters""" def store(file): # type: ignore[no-untyped-def] + """Store the given file stream""" class FeedStorageProtocol(Protocol): + """Reimplementation of ``IFeedStorage`` that can be used in type hints.""" def __init__(self, uri: str, *, feed_options: dict[str, Any] | None = None): + """Initialize the storage with the parameters given in the URI and the + feed-specific options (see :setting:`FEEDS`)""" def open(self, spider: Spider) -> IO[bytes]: + """Open the storage for the given spider. It must return a file-like + object that will be used for the exporters""" def store(self, file: IO[bytes]) -> Deferred[None] | None: + """Store the given file stream""" @implementer(IFeedStorage) @@ -538,6 +570,15 @@ spider: Spider, uri_template: str, ) -> FeedSlot: + """ + Redirect the output data stream to a new file. + Execute multiple times if FEED_EXPORT_BATCH_ITEM_COUNT setting or FEEDS.batch_item_count is specified + :param batch_id: sequence number of current batch + :param uri: uri of the new batch to start + :param feed_options: dict with parameters of feed + :param spider: user spider + :param uri_template: template of uri which contains %(batch_time)s or %(batch_id)d to create new uri + """ storage = self._get_storage(uri, feed_options) return FeedSlot( storage=storage, @@ -606,6 +647,10 @@ return False def _settings_are_valid(self) -> bool: + """ + If FEED_EXPORT_BATCH_ITEM_COUNT setting or FEEDS.batch_item_count is specified uri has to contain + %(batch_time)s or %(batch_id)d to distinguish different files of partial output + """ for uri_template, values in self.feeds.items(): if values["batch_item_count"] and not re.search( r"%\(batch_time\)s|%\(batch_id\)", uri_template @@ -637,6 +682,8 @@ def _get_storage( self, uri: str, feed_options: dict[str, Any] ) -> FeedStorageProtocol: + """Build a storage object for the specified *uri* with the specified + *feed_options*.""" cls = self.storages.get(urlparse(uri).scheme, self.storages["file"]) return build_from_crawler(cls, self.crawler, uri, feed_options=feed_options) @@ -666,4 +713,4 @@ item_filter_class: type[ItemFilter] = load_object( feed_options.get("item_filter", ItemFilter) ) - return item_filter_class(feed_options)+ return item_filter_class(feed_options)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/extensions/feedexport.py
Write docstrings describing each step
from bz2 import BZ2File from gzip import GzipFile from io import IOBase from lzma import LZMAFile from typing import IO, Any, BinaryIO, cast from scrapy.utils.misc import load_object class GzipPlugin: def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file self.feed_options = feed_options compress_level = self.feed_options.get("gzip_compresslevel", 9) mtime = self.feed_options.get("gzip_mtime") filename = self.feed_options.get("gzip_filename") self.gzipfile = GzipFile( fileobj=self.file, mode="wb", compresslevel=compress_level, mtime=mtime, filename=filename, ) def write(self, data: bytes) -> int: return self.gzipfile.write(data) def close(self) -> None: self.gzipfile.close() class Bz2Plugin: def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file self.feed_options = feed_options compress_level = self.feed_options.get("bz2_compresslevel", 9) self.bz2file = BZ2File( filename=self.file, mode="wb", compresslevel=compress_level ) def write(self, data: bytes) -> int: return self.bz2file.write(data) def close(self) -> None: self.bz2file.close() class LZMAPlugin: def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file self.feed_options = feed_options format_ = self.feed_options.get("lzma_format") check = self.feed_options.get("lzma_check", -1) preset = self.feed_options.get("lzma_preset") filters = self.feed_options.get("lzma_filters") self.lzmafile = LZMAFile( filename=self.file, mode="wb", format=format_, check=check, preset=preset, filters=filters, ) def write(self, data: bytes) -> int: return self.lzmafile.write(data) def close(self) -> None: self.lzmafile.close() # io.IOBase is subclassed here, so that exporters can use the PostProcessingManager # instance as a file like writable object. This could be needed by some exporters # such as CsvItemExporter which wraps the feed storage with io.TextIOWrapper. class PostProcessingManager(IOBase): def __init__( self, plugins: list[Any], file: IO[bytes], feed_options: dict[str, Any] ) -> None: self.plugins = self._load_plugins(plugins) self.file = file self.feed_options = feed_options self.head_plugin = self._get_head_plugin() def write(self, data: bytes) -> int: return cast("int", self.head_plugin.write(data)) def tell(self) -> int: return self.file.tell() def close(self) -> None: self.head_plugin.close() def writable(self) -> bool: return True def _load_plugins(self, plugins: list[Any]) -> list[Any]: return [load_object(plugin) for plugin in plugins] def _get_head_plugin(self) -> Any: prev = self.file for plugin in self.plugins[::-1]: prev = plugin(prev, self.feed_options) return prev
--- +++ @@ -1,3 +1,6 @@+""" +Extension for processing data before they are exported to feeds. +""" from bz2 import BZ2File from gzip import GzipFile @@ -9,6 +12,17 @@ class GzipPlugin: + """ + Compresses received data using `gzip <https://en.wikipedia.org/wiki/Gzip>`_. + + Accepted ``feed_options`` parameters: + + - `gzip_compresslevel` + - `gzip_mtime` + - `gzip_filename` + + See :py:class:`gzip.GzipFile` for more info about parameters. + """ def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file @@ -32,6 +46,15 @@ class Bz2Plugin: + """ + Compresses received data using `bz2 <https://en.wikipedia.org/wiki/Bzip2>`_. + + Accepted ``feed_options`` parameters: + + - `bz2_compresslevel` + + See :py:class:`bz2.BZ2File` for more info about parameters. + """ def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file @@ -49,6 +72,21 @@ class LZMAPlugin: + """ + Compresses received data using `lzma <https://en.wikipedia.org/wiki/Lempel–Ziv–Markov_chain_algorithm>`_. + + Accepted ``feed_options`` parameters: + + - `lzma_format` + - `lzma_check` + - `lzma_preset` + - `lzma_filters` + + .. note:: + ``lzma_filters`` cannot be used in pypy version 7.3.1 and older. + + See :py:class:`lzma.LZMAFile` for more info about parameters. + """ def __init__(self, file: BinaryIO, feed_options: dict[str, Any]) -> None: self.file = file @@ -78,6 +116,14 @@ # instance as a file like writable object. This could be needed by some exporters # such as CsvItemExporter which wraps the feed storage with io.TextIOWrapper. class PostProcessingManager(IOBase): + """ + This will manage and use declared plugins to process data in a + pipeline-ish way. + :param plugins: all the declared plugins for the feed + :type plugins: list + :param file: final target file where the processed data will be written + :type file: file like object + """ def __init__( self, plugins: list[Any], file: IO[bytes], feed_options: dict[str, Any] @@ -88,12 +134,23 @@ self.head_plugin = self._get_head_plugin() def write(self, data: bytes) -> int: + """ + Uses all the declared plugins to process data first, then writes + the processed data to target file. + :param data: data passed to be written to target file + :type data: bytes + :return: returns number of bytes written + :rtype: int + """ return cast("int", self.head_plugin.write(data)) def tell(self) -> int: return self.file.tell() def close(self) -> None: + """ + Close the target file along with all the plugins. + """ self.head_plugin.close() def writable(self) -> bool: @@ -106,4 +163,4 @@ prev = self.file for plugin in self.plugins[::-1]: prev = plugin(prev, self.feed_options) - return prev+ return prev
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/extensions/postprocessing.py
Write docstrings for algorithm functions
from __future__ import annotations from io import StringIO from mimetypes import MimeTypes from pkgutil import get_data from typing import TYPE_CHECKING from scrapy.http import Response from scrapy.utils.misc import load_object from scrapy.utils.python import binary_is_text, to_bytes, to_unicode if TYPE_CHECKING: from collections.abc import Mapping class ResponseTypes: CLASSES = { "text/html": "scrapy.http.HtmlResponse", "application/atom+xml": "scrapy.http.XmlResponse", "application/rdf+xml": "scrapy.http.XmlResponse", "application/rss+xml": "scrapy.http.XmlResponse", "application/xhtml+xml": "scrapy.http.HtmlResponse", "application/vnd.wap.xhtml+xml": "scrapy.http.HtmlResponse", "application/xml": "scrapy.http.XmlResponse", "application/json": "scrapy.http.JsonResponse", "application/x-json": "scrapy.http.JsonResponse", "application/json-amazonui-streaming": "scrapy.http.JsonResponse", "application/javascript": "scrapy.http.TextResponse", "application/x-javascript": "scrapy.http.TextResponse", "text/xml": "scrapy.http.XmlResponse", "text/*": "scrapy.http.TextResponse", } def __init__(self) -> None: self.classes: dict[str, type[Response]] = {} self.mimetypes: MimeTypes = MimeTypes() mimedata = get_data("scrapy", "mime.types") if not mimedata: raise ValueError( "The mime.types file is not found in the Scrapy installation" ) self.mimetypes.readfp(StringIO(mimedata.decode("utf8"))) for mimetype, cls in self.CLASSES.items(): self.classes[mimetype] = load_object(cls) def from_mimetype(self, mimetype: str) -> type[Response]: if mimetype is None: return Response if mimetype in self.classes: return self.classes[mimetype] basetype = f"{mimetype.split('/', maxsplit=1)[0]}/*" return self.classes.get(basetype, Response) def from_content_type( self, content_type: str | bytes, content_encoding: bytes | None = None ) -> type[Response]: if content_encoding: return Response mimetype = ( to_unicode(content_type, encoding="latin-1").split(";")[0].strip().lower() ) return self.from_mimetype(mimetype) def from_content_disposition( self, content_disposition: str | bytes ) -> type[Response]: try: filename = ( to_unicode(content_disposition, encoding="latin-1", errors="replace") .split(";")[1] .split("=")[1] .strip("\"'") ) return self.from_filename(filename) except IndexError: return Response def from_headers(self, headers: Mapping[bytes, bytes]) -> type[Response]: cls = Response if b"Content-Type" in headers: cls = self.from_content_type( content_type=headers[b"Content-Type"], content_encoding=headers.get(b"Content-Encoding"), ) if cls is Response and b"Content-Disposition" in headers: cls = self.from_content_disposition(headers[b"Content-Disposition"]) return cls def from_filename(self, filename: str) -> type[Response]: mimetype, encoding = self.mimetypes.guess_type(filename) if mimetype and not encoding: return self.from_mimetype(mimetype) return Response def from_body(self, body: bytes) -> type[Response]: chunk = body[:5000] chunk = to_bytes(chunk) if not binary_is_text(chunk): return self.from_mimetype("application/octet-stream") lowercase_chunk = chunk.lower() if b"<html>" in lowercase_chunk: return self.from_mimetype("text/html") if b"<?xml" in lowercase_chunk: return self.from_mimetype("text/xml") if b"<!doctype html>" in lowercase_chunk: return self.from_mimetype("text/html") return self.from_mimetype("text") def from_args( self, headers: Mapping[bytes, bytes] | None = None, url: str | None = None, filename: str | None = None, body: bytes | None = None, ) -> type[Response]: cls = Response if headers is not None: cls = self.from_headers(headers) if cls is Response and url is not None: cls = self.from_filename(url) if cls is Response and filename is not None: cls = self.from_filename(filename) if cls is Response and body is not None: cls = self.from_body(body) return cls responsetypes = ResponseTypes()
--- +++ @@ -1,3 +1,7 @@+""" +This module implements a class which returns the appropriate Response class +based on different criteria. +""" from __future__ import annotations @@ -45,6 +49,7 @@ self.classes[mimetype] = load_object(cls) def from_mimetype(self, mimetype: str) -> type[Response]: + """Return the most appropriate Response class for the given mimetype""" if mimetype is None: return Response if mimetype in self.classes: @@ -55,6 +60,8 @@ def from_content_type( self, content_type: str | bytes, content_encoding: bytes | None = None ) -> type[Response]: + """Return the most appropriate Response class from an HTTP Content-Type + header""" if content_encoding: return Response mimetype = ( @@ -77,6 +84,8 @@ return Response def from_headers(self, headers: Mapping[bytes, bytes]) -> type[Response]: + """Return the most appropriate Response class by looking at the HTTP + headers""" cls = Response if b"Content-Type" in headers: cls = self.from_content_type( @@ -88,12 +97,17 @@ return cls def from_filename(self, filename: str) -> type[Response]: + """Return the most appropriate Response class from a file name""" mimetype, encoding = self.mimetypes.guess_type(filename) if mimetype and not encoding: return self.from_mimetype(mimetype) return Response def from_body(self, body: bytes) -> type[Response]: + """Try to guess the appropriate response based on the body content. + This method is a bit magic and could be improved in the future, but + it's not meant to be used except for special cases where response types + cannot be guess using more straightforward methods.""" chunk = body[:5000] chunk = to_bytes(chunk) if not binary_is_text(chunk): @@ -114,6 +128,8 @@ filename: str | None = None, body: bytes | None = None, ) -> type[Response]: + """Guess the most appropriate Response class based on + the given arguments.""" cls = Response if headers is not None: cls = self.from_headers(headers) @@ -126,4 +142,4 @@ return cls -responsetypes = ResponseTypes()+responsetypes = ResponseTypes()
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/responsetypes.py
Document functions with detailed explanations
from __future__ import annotations import inspect from typing import ( TYPE_CHECKING, Any, AnyStr, Concatenate, NoReturn, TypeAlias, TypedDict, TypeVar, overload, ) from w3lib.url import safe_url_string # a workaround for the docs "more than one target found" problem import scrapy # noqa: TC001 from scrapy.http.headers import Headers from scrapy.utils.curl import curl_to_request_kwargs from scrapy.utils.python import to_bytes from scrapy.utils.trackref import object_ref if TYPE_CHECKING: from collections.abc import Callable, Iterable, Mapping from twisted.python.failure import Failure # typing.NotRequired and typing.Self require Python 3.11 from typing_extensions import NotRequired, Self # circular import from scrapy.http import Response CallbackT: TypeAlias = Callable[Concatenate[Response, ...], Any] class VerboseCookie(TypedDict): name: str | bytes value: str | bytes | bool | float | int domain: NotRequired[str | bytes] path: NotRequired[str | bytes] secure: NotRequired[bool] CookiesT: TypeAlias = dict[str, str] | list[VerboseCookie] RequestTypeVar = TypeVar("RequestTypeVar", bound="Request") def NO_CALLBACK(*args: Any, **kwargs: Any) -> NoReturn: raise RuntimeError( "The NO_CALLBACK callback has been called. This is a special callback " "value intended for requests whose callback is never meant to be " "called." ) class Request(object_ref): __attrs_and_slots = ("callback", "dont_filter", "errback", "method", "priority") attributes: tuple[str, ...] = ( "url", "headers", "body", "cookies", "meta", "encoding", "flags", "cb_kwargs", *__attrs_and_slots, ) """A tuple of :class:`str` objects containing the name of all public attributes of the class that are also keyword parameters of the ``__init__()`` method. Currently used by :meth:`.Request.replace`, :meth:`.Request.to_dict` and :func:`~scrapy.utils.request.request_from_dict`. """ __slots__ = ( "__weakref__", "_body", "_cb_kwargs", "_cookies", "_encoding", "_flags", "_headers", "_meta", "_url", *__attrs_and_slots, ) del __attrs_and_slots def __init__( self, url: str, callback: CallbackT | None = None, method: str = "GET", headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes | str | None = None, cookies: CookiesT | None = None, meta: dict[str, Any] | None = None, encoding: str = "utf-8", priority: int = 0, dont_filter: bool = False, errback: Callable[[Failure], Any] | None = None, flags: list[str] | None = None, cb_kwargs: dict[str, Any] | None = None, ) -> None: self._encoding: str = encoding # this one has to be set first self.method: str = str(method).upper() self._set_url(url) self._set_body(body) if not isinstance(priority, int): raise TypeError(f"Request priority not an integer: {priority!r}") #: Default: ``0`` #: #: Value that the :ref:`scheduler <topics-scheduler>` may use for #: request prioritization. #: #: Built-in schedulers prioritize requests with a higher priority #: value. #: #: Negative values are allowed. self.priority: int = priority if not (callable(callback) or callback is None): raise TypeError( f"callback must be a callable, got {type(callback).__name__}" ) if not (callable(errback) or errback is None): raise TypeError(f"errback must be a callable, got {type(errback).__name__}") #: :class:`~collections.abc.Callable` to parse the #: :class:`~scrapy.http.Response` to this request once received. #: #: The callable must expect the response as its first parameter, and #: support any additional keyword arguments set through #: :attr:`cb_kwargs`. #: #: In addition to an arbitrary callable, the following values are also #: supported: #: #: - ``None`` (default), which indicates that the #: :meth:`~scrapy.Spider.parse` method of the spider must be used. #: #: - :func:`~scrapy.http.request.NO_CALLBACK`. #: #: If an unhandled exception is raised during request or response #: processing, i.e. by a :ref:`spider middleware #: <topics-spider-middleware>`, :ref:`downloader middleware #: <topics-downloader-middleware>` or download handler #: (:setting:`DOWNLOAD_HANDLERS`), :attr:`errback` is called instead. #: #: .. tip:: #: :class:`~scrapy.spidermiddlewares.httperror.HttpErrorMiddleware` #: raises exceptions for non-2xx responses by default, sending them #: to the :attr:`errback` instead. #: #: .. seealso:: #: :ref:`topics-request-response-ref-request-callback-arguments` self.callback: CallbackT | None = callback #: :class:`~collections.abc.Callable` to handle exceptions raised #: during request or response processing. #: #: The callable must expect a :exc:`~twisted.python.failure.Failure` as #: its first parameter. #: #: .. seealso:: :ref:`topics-request-response-ref-errbacks` self.errback: Callable[[Failure], Any] | None = errback self._cookies: CookiesT | None = cookies or None self._headers: Headers | None = ( Headers(headers, encoding=encoding) if headers else None ) #: Whether this request may be filtered out by :ref:`components #: <topics-components>` that support filtering out requests (``False``, #: default), or those components should not filter out this request #: (``True``). #: #: The following built-in components check this attribute: #: #: - The :ref:`scheduler <topics-scheduler>` uses it to skip #: duplicate request filtering (see #: :setting:`DUPEFILTER_CLASS`). When set to ``True``, the #: request is not checked against the duplicate filter, #: allowing requests that would otherwise be considered duplicates #: to be scheduled multiple times. #: - :class:`~scrapy.downloadermiddlewares.offsite.OffsiteMiddleware` #: uses it to allow requests to domains not in #: :attr:`~scrapy.Spider.allowed_domains`. To skip only the offsite #: filter without affecting other components, consider using the #: :reqmeta:`allow_offsite` request meta key instead. #: #: Third-party components may also use this attribute to decide whether #: to filter out a request. #: #: When defining the start URLs of a spider through #: :attr:`~scrapy.Spider.start_urls`, this attribute is enabled by #: default. See :meth:`~scrapy.Spider.start`. self.dont_filter: bool = dont_filter self._meta: dict[str, Any] | None = dict(meta) if meta else None self._cb_kwargs: dict[str, Any] | None = dict(cb_kwargs) if cb_kwargs else None self._flags: list[str] | None = list(flags) if flags else None @property def cb_kwargs(self) -> dict[str, Any]: if self._cb_kwargs is None: self._cb_kwargs = {} return self._cb_kwargs @cb_kwargs.setter def cb_kwargs(self, value: dict[str, Any] | None) -> None: self._cb_kwargs = value or None @property def meta(self) -> dict[str, Any]: if self._meta is None: self._meta = {} return self._meta @meta.setter def meta(self, value: dict[str, Any] | None) -> None: self._meta = value or None @property def url(self) -> str: return self._url def _set_url(self, url: str) -> None: if not isinstance(url, str): raise TypeError(f"Request url must be str, got {type(url).__name__}") self._url = safe_url_string(url, self.encoding) if ( "://" not in self._url and not self._url.startswith("about:") and not self._url.startswith("data:") ): raise ValueError(f"Missing scheme in request url: {self._url}") @property def body(self) -> bytes: return self._body def _set_body(self, body: str | bytes | None) -> None: self._body = b"" if not body else to_bytes(body, self.encoding) @property def encoding(self) -> str: return self._encoding @property def flags(self) -> list[str]: if self._flags is None: self._flags = [] return self._flags @flags.setter def flags(self, value: list[str] | None) -> None: self._flags = value or None @property def cookies(self) -> CookiesT: if self._cookies is None: self._cookies = {} return self._cookies @cookies.setter def cookies(self, value: CookiesT | None) -> None: self._cookies = value or None @property def headers(self) -> Headers: if self._headers is None: self._headers = Headers(encoding=self.encoding) return self._headers @headers.setter def headers( self, value: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None ) -> None: if isinstance(value, Headers): self._headers = value else: self._headers = Headers(value, encoding=self.encoding) if value else None def __repr__(self) -> str: return f"<{self.method} {self.url}>" def copy(self) -> Self: return self.replace() @overload def replace( self, *args: Any, cls: type[RequestTypeVar], **kwargs: Any ) -> RequestTypeVar: ... @overload def replace(self, *args: Any, cls: None = None, **kwargs: Any) -> Self: ... def replace( self, *args: Any, cls: type[Request] | None = None, **kwargs: Any ) -> Request: for x in self.attributes: kwargs.setdefault(x, getattr(self, x)) if cls is None: cls = self.__class__ return cls(*args, **kwargs) @classmethod def from_curl( cls, curl_command: str, ignore_unknown_options: bool = True, **kwargs: Any, ) -> Self: request_kwargs = curl_to_request_kwargs(curl_command, ignore_unknown_options) request_kwargs.update(kwargs) return cls(**request_kwargs) def to_dict(self, *, spider: scrapy.Spider | None = None) -> dict[str, Any]: d = { "url": self.url, # urls are safe (safe_string_url) "callback": ( _find_method(spider, self.callback) if callable(self.callback) else self.callback ), "errback": ( _find_method(spider, self.errback) if callable(self.errback) else self.errback ), "headers": dict(self.headers), } for attr in self.attributes: d.setdefault(attr, getattr(self, attr)) if type(self) is not Request: # pylint: disable=unidiomatic-typecheck d["_class"] = self.__module__ + "." + self.__class__.__name__ return d def _find_method(obj: Any, func: Callable[..., Any]) -> str: # Only instance methods contain ``__func__`` if obj and hasattr(func, "__func__"): members = inspect.getmembers(obj, predicate=inspect.ismethod) for name, obj_func in members: # We need to use __func__ to access the original function object because instance # method objects are generated each time attribute is retrieved from instance. # # Reference: The standard type hierarchy # https://docs.python.org/3/reference/datamodel.html if obj_func.__func__ is func.__func__: return name raise ValueError(f"Function {func} is not an instance method in: {obj}")
--- +++ @@ -1,3 +1,9 @@+""" +This module implements the Request class which is used to represent HTTP +requests in Scrapy. + +See documentation in docs/topics/request-response.rst +""" from __future__ import annotations @@ -52,6 +58,22 @@ def NO_CALLBACK(*args: Any, **kwargs: Any) -> NoReturn: + """When assigned to the ``callback`` parameter of + :class:`~scrapy.Request`, it indicates that the request is not meant + to have a spider callback at all. + + For example: + + .. code-block:: python + + Request("https://example.com", callback=NO_CALLBACK) + + This value should be used by :ref:`components <topics-components>` that + create and handle their own requests, e.g. through + :meth:`scrapy.core.engine.ExecutionEngine.download`, so that downloader + middlewares handling such requests can treat them differently from requests + intended for the :meth:`~scrapy.Spider.parse` callback. + """ raise RuntimeError( "The NO_CALLBACK callback has been called. This is a special callback " "value intended for requests whose callback is never meant to be " @@ -60,6 +82,9 @@ class Request(object_ref): + """Represents an HTTP request, which is usually generated in a Spider and + executed by the Downloader, thus generating a :class:`~scrapy.http.Response`. + """ __attrs_and_slots = ("callback", "dont_filter", "errback", "method", "priority") attributes: tuple[str, ...] = ( @@ -311,6 +336,7 @@ def replace( self, *args: Any, cls: type[Request] | None = None, **kwargs: Any ) -> Request: + """Create a new Request with the same attributes except for those given new values""" for x in self.attributes: kwargs.setdefault(x, getattr(self, x)) if cls is None: @@ -324,11 +350,45 @@ ignore_unknown_options: bool = True, **kwargs: Any, ) -> Self: + """Create a Request object from a string containing a `cURL + <https://curl.se/>`_ command. It populates the HTTP method, the + URL, the headers, the cookies and the body. It accepts the same + arguments as the :class:`Request` class, taking preference and + overriding the values of the same arguments contained in the cURL + command. + + Unrecognized options are ignored by default. To raise an error when + finding unknown options call this method by passing + ``ignore_unknown_options=False``. + + .. caution:: Using :meth:`from_curl` from :class:`~scrapy.Request` + subclasses, such as :class:`~scrapy.http.JsonRequest`, or + :class:`~scrapy.http.XmlRpcRequest`, as well as having + :ref:`downloader middlewares <topics-downloader-middleware>` + and + :ref:`spider middlewares <topics-spider-middleware>` + enabled, such as + :class:`~scrapy.downloadermiddlewares.defaultheaders.DefaultHeadersMiddleware`, + :class:`~scrapy.downloadermiddlewares.useragent.UserAgentMiddleware`, + or + :class:`~scrapy.downloadermiddlewares.httpcompression.HttpCompressionMiddleware`, + may modify the :class:`~scrapy.Request` object. + + To translate a cURL command into a Scrapy request, + you may use `curl2scrapy <https://michael-shub.github.io/curl2scrapy/>`_. + """ request_kwargs = curl_to_request_kwargs(curl_command, ignore_unknown_options) request_kwargs.update(kwargs) return cls(**request_kwargs) def to_dict(self, *, spider: scrapy.Spider | None = None) -> dict[str, Any]: + """Return a dictionary containing the Request's data. + + Use :func:`~scrapy.utils.request.request_from_dict` to convert back into a :class:`~scrapy.Request` object. + + If a spider is given, this method will try to find out the name of the spider methods used as callback + and errback and include them in the output dict, raising an exception if they cannot be found. + """ d = { "url": self.url, # urls are safe (safe_string_url) "callback": ( @@ -351,6 +411,7 @@ def _find_method(obj: Any, func: Callable[..., Any]) -> str: + """Helper function for Request.to_dict""" # Only instance methods contain ``__func__`` if obj and hasattr(func, "__func__"): members = inspect.getmembers(obj, predicate=inspect.ismethod) @@ -362,4 +423,4 @@ # https://docs.python.org/3/reference/datamodel.html if obj_func.__func__ is func.__func__: return name - raise ValueError(f"Function {func} is not an instance method in: {obj}")+ raise ValueError(f"Function {func} is not an instance method in: {obj}")
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/http/request/__init__.py
Improve documentation using docstrings
from __future__ import annotations import logging import socket import sys from importlib import import_module from pprint import pformat from typing import TYPE_CHECKING from scrapy import signals from scrapy.exceptions import NotConfigured from scrapy.mail import MailSender from scrapy.utils.asyncio import AsyncioLoopingCall, create_looping_call from scrapy.utils.defer import _schedule_coro from scrapy.utils.engine import get_engine_status if TYPE_CHECKING: from twisted.internet.task import LoopingCall # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler logger = logging.getLogger(__name__) class MemoryUsage: def __init__(self, crawler: Crawler): if not crawler.settings.getbool("MEMUSAGE_ENABLED"): raise NotConfigured try: # stdlib's resource module is only available on unix platforms. self.resource = import_module("resource") except ImportError: raise NotConfigured self.crawler: Crawler = crawler self.warned: bool = False self.notify_mails: list[str] = crawler.settings.getlist("MEMUSAGE_NOTIFY_MAIL") self.limit: int = crawler.settings.getint("MEMUSAGE_LIMIT_MB") * 1024 * 1024 self.warning: int = crawler.settings.getint("MEMUSAGE_WARNING_MB") * 1024 * 1024 self.check_interval: float = crawler.settings.getfloat( "MEMUSAGE_CHECK_INTERVAL_SECONDS" ) self.mail: MailSender = MailSender.from_crawler(crawler) crawler.signals.connect(self.engine_started, signal=signals.engine_started) crawler.signals.connect(self.engine_stopped, signal=signals.engine_stopped) @classmethod def from_crawler(cls, crawler: Crawler) -> Self: return cls(crawler) def get_virtual_size(self) -> int: size: int = self.resource.getrusage(self.resource.RUSAGE_SELF).ru_maxrss if sys.platform != "darwin": # on macOS ru_maxrss is in bytes, on Linux it is in KB size *= 1024 return size def engine_started(self) -> None: assert self.crawler.stats self.crawler.stats.set_value("memusage/startup", self.get_virtual_size()) self.tasks: list[AsyncioLoopingCall | LoopingCall] = [] tsk = create_looping_call(self.update) self.tasks.append(tsk) tsk.start(self.check_interval, now=True) if self.limit: tsk = create_looping_call(self._check_limit) self.tasks.append(tsk) tsk.start(self.check_interval, now=True) if self.warning: tsk = create_looping_call(self._check_warning) self.tasks.append(tsk) tsk.start(self.check_interval, now=True) def engine_stopped(self) -> None: for tsk in self.tasks: if tsk.running: tsk.stop() def update(self) -> None: assert self.crawler.stats self.crawler.stats.max_value("memusage/max", self.get_virtual_size()) def _check_limit(self) -> None: assert self.crawler.engine assert self.crawler.stats peak_mem_usage = self.get_virtual_size() if peak_mem_usage > self.limit: self.crawler.stats.set_value("memusage/limit_reached", 1) mem = self.limit / 1024 / 1024 logger.error( "Memory usage exceeded %(memusage)dMiB. Shutting down Scrapy...", {"memusage": mem}, extra={"crawler": self.crawler}, ) if self.notify_mails: subj = ( f"{self.crawler.settings['BOT_NAME']} terminated: " f"memory usage exceeded {mem}MiB at {socket.gethostname()}" ) self._send_report(self.notify_mails, subj) self.crawler.stats.set_value("memusage/limit_notified", 1) if self.crawler.engine.spider is not None: _schedule_coro( self.crawler.engine.close_spider_async(reason="memusage_exceeded") ) else: _schedule_coro(self.crawler.stop_async()) else: logger.info( "Peak memory usage is %(virtualsize)dMiB", {"virtualsize": peak_mem_usage / 1024 / 1024}, ) def _check_warning(self) -> None: if self.warned: # warn only once return assert self.crawler.stats if self.get_virtual_size() > self.warning: self.crawler.stats.set_value("memusage/warning_reached", 1) mem = self.warning / 1024 / 1024 logger.warning( "Memory usage reached %(memusage)dMiB", {"memusage": mem}, extra={"crawler": self.crawler}, ) if self.notify_mails: subj = ( f"{self.crawler.settings['BOT_NAME']} warning: " f"memory usage reached {mem}MiB at {socket.gethostname()}" ) self._send_report(self.notify_mails, subj) self.crawler.stats.set_value("memusage/warning_notified", 1) self.warned = True def _send_report(self, rcpts: list[str], subject: str) -> None: assert self.crawler.engine assert self.crawler.stats stats = self.crawler.stats s = f"Memory usage at engine startup : {stats.get_value('memusage/startup') / 1024 / 1024}M\r\n" s += f"Maximum memory usage : {stats.get_value('memusage/max') / 1024 / 1024}M\r\n" s += f"Current memory usage : {self.get_virtual_size() / 1024 / 1024}M\r\n" s += ( "ENGINE STATUS ------------------------------------------------------- \r\n" ) s += "\r\n" s += pformat(get_engine_status(self.crawler.engine)) s += "\r\n" self.mail.send(rcpts, subject, s)
--- +++ @@ -1,3 +1,8 @@+""" +MemoryUsage extension + +See documentation in docs/topics/extensions.rst +""" from __future__ import annotations @@ -139,6 +144,7 @@ self.warned = True def _send_report(self, rcpts: list[str], subject: str) -> None: + """send notification mail with some additional useful info""" assert self.crawler.engine assert self.crawler.stats stats = self.crawler.stats @@ -152,4 +158,4 @@ s += "\r\n" s += pformat(get_engine_status(self.crawler.engine)) s += "\r\n" - self.mail.send(rcpts, subject, s)+ self.mail.send(rcpts, subject, s)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/extensions/memusage.py
Add docstrings to make code maintainable
from __future__ import annotations import base64 import functools import hashlib import logging import mimetypes import time import warnings from collections import defaultdict from contextlib import suppress from ftplib import FTP from io import BytesIO from pathlib import Path from typing import IO, TYPE_CHECKING, Any, NoReturn, Protocol, TypedDict, cast from urllib.parse import urlparse from itemadapter import ItemAdapter from twisted.internet.defer import Deferred, maybeDeferred from twisted.internet.threads import deferToThread from scrapy.exceptions import IgnoreRequest, NotConfigured, ScrapyDeprecationWarning from scrapy.http import Request, Response from scrapy.http.request import NO_CALLBACK from scrapy.pipelines.media import FileInfo, FileInfoOrError, MediaPipeline from scrapy.utils.boto import is_botocore_available from scrapy.utils.datatypes import CaseInsensitiveDict from scrapy.utils.ftp import ftp_store_file from scrapy.utils.log import failure_to_exc_info from scrapy.utils.python import to_bytes from scrapy.utils.request import referer_str if TYPE_CHECKING: from os import PathLike from twisted.python.failure import Failure # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.settings import BaseSettings logger = logging.getLogger(__name__) def _to_string(path: str | PathLike[str]) -> str: return str(path) # convert a Path object to string def _md5sum(file: IO[bytes]) -> str: m = hashlib.md5() # noqa: S324 while True: d = file.read(8096) if not d: break m.update(d) return m.hexdigest() class FileException(Exception): class StatInfo(TypedDict, total=False): checksum: str last_modified: float class FilesStoreProtocol(Protocol): def __init__(self, basedir: str): ... def persist_file( self, path: str, buf: BytesIO, info: MediaPipeline.SpiderInfo, meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> Deferred[Any] | None: ... def stat_file( self, path: str, info: MediaPipeline.SpiderInfo ) -> StatInfo | Deferred[StatInfo]: ... class FSFilesStore: def __init__(self, basedir: str | PathLike[str]): basedir = _to_string(basedir) if "://" in basedir: basedir = basedir.split("://", 1)[1] self.basedir: str = basedir self._mkdir(Path(self.basedir)) self.created_directories: defaultdict[MediaPipeline.SpiderInfo, set[str]] = ( defaultdict(set) ) def persist_file( self, path: str | PathLike[str], buf: BytesIO, info: MediaPipeline.SpiderInfo, meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> None: absolute_path = self._get_filesystem_path(path) self._mkdir(absolute_path.parent, info) absolute_path.write_bytes(buf.getvalue()) def stat_file( self, path: str | PathLike[str], info: MediaPipeline.SpiderInfo ) -> StatInfo: absolute_path = self._get_filesystem_path(path) try: last_modified = absolute_path.stat().st_mtime except OSError: return {} with absolute_path.open("rb") as f: checksum = _md5sum(f) return {"last_modified": last_modified, "checksum": checksum} def _get_filesystem_path(self, path: str | PathLike[str]) -> Path: path_comps = _to_string(path).split("/") return Path(self.basedir, *path_comps) def _mkdir( self, dirname: Path, domain: MediaPipeline.SpiderInfo | None = None ) -> None: seen: set[str] = self.created_directories[domain] if domain else set() if str(dirname) not in seen: if not dirname.exists(): dirname.mkdir(parents=True) seen.add(str(dirname)) class S3FilesStore: AWS_ACCESS_KEY_ID = None AWS_SECRET_ACCESS_KEY = None AWS_SESSION_TOKEN = None AWS_ENDPOINT_URL = None AWS_REGION_NAME = None AWS_USE_SSL = None AWS_VERIFY = None POLICY = "private" # Overridden from settings.FILES_STORE_S3_ACL in FilesPipeline.from_crawler() HEADERS = { "Cache-Control": "max-age=172800", } def __init__(self, uri: str): if not is_botocore_available(): raise NotConfigured("missing botocore library") import botocore.session # noqa: PLC0415 session = botocore.session.get_session() self.s3_client = session.create_client( "s3", aws_access_key_id=self.AWS_ACCESS_KEY_ID, aws_secret_access_key=self.AWS_SECRET_ACCESS_KEY, aws_session_token=self.AWS_SESSION_TOKEN, endpoint_url=self.AWS_ENDPOINT_URL, region_name=self.AWS_REGION_NAME, use_ssl=self.AWS_USE_SSL, verify=self.AWS_VERIFY, ) if not uri.startswith("s3://"): raise ValueError(f"Incorrect URI scheme in {uri}, expected 's3'") self.bucket, self.prefix = uri[5:].split("/", 1) def stat_file( self, path: str, info: MediaPipeline.SpiderInfo ) -> Deferred[StatInfo]: def _onsuccess(boto_key: dict[str, Any]) -> StatInfo: checksum = boto_key["ETag"].strip('"') last_modified = boto_key["LastModified"] modified_stamp = time.mktime(last_modified.timetuple()) return {"checksum": checksum, "last_modified": modified_stamp} return self._get_boto_key(path).addCallback(_onsuccess) def _get_boto_key(self, path: str) -> Deferred[dict[str, Any]]: key_name = f"{self.prefix}{path}" return cast( "Deferred[dict[str, Any]]", deferToThread( self.s3_client.head_object, # type: ignore[attr-defined] Bucket=self.bucket, Key=key_name, ), ) def persist_file( self, path: str, buf: BytesIO, info: MediaPipeline.SpiderInfo, meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> Deferred[Any]: key_name = f"{self.prefix}{path}" buf.seek(0) extra = self._headers_to_botocore_kwargs(self.HEADERS) if headers: extra.update(self._headers_to_botocore_kwargs(headers)) return deferToThread( self.s3_client.put_object, # type: ignore[attr-defined] Bucket=self.bucket, Key=key_name, Body=buf, Metadata={k: str(v) for k, v in (meta or {}).items()}, ACL=self.POLICY, **extra, ) def _headers_to_botocore_kwargs(self, headers: dict[str, Any]) -> dict[str, Any]: # This is required while we need to support both boto and botocore. mapping = CaseInsensitiveDict( { "Content-Type": "ContentType", "Cache-Control": "CacheControl", "Content-Disposition": "ContentDisposition", "Content-Encoding": "ContentEncoding", "Content-Language": "ContentLanguage", "Content-Length": "ContentLength", "Content-MD5": "ContentMD5", "Expires": "Expires", "X-Amz-Grant-Full-Control": "GrantFullControl", "X-Amz-Grant-Read": "GrantRead", "X-Amz-Grant-Read-ACP": "GrantReadACP", "X-Amz-Grant-Write-ACP": "GrantWriteACP", "X-Amz-Object-Lock-Legal-Hold": "ObjectLockLegalHoldStatus", "X-Amz-Object-Lock-Mode": "ObjectLockMode", "X-Amz-Object-Lock-Retain-Until-Date": "ObjectLockRetainUntilDate", "X-Amz-Request-Payer": "RequestPayer", "X-Amz-Server-Side-Encryption": "ServerSideEncryption", "X-Amz-Server-Side-Encryption-Aws-Kms-Key-Id": "SSEKMSKeyId", "X-Amz-Server-Side-Encryption-Context": "SSEKMSEncryptionContext", "X-Amz-Server-Side-Encryption-Customer-Algorithm": "SSECustomerAlgorithm", "X-Amz-Server-Side-Encryption-Customer-Key": "SSECustomerKey", "X-Amz-Server-Side-Encryption-Customer-Key-Md5": "SSECustomerKeyMD5", "X-Amz-Storage-Class": "StorageClass", "X-Amz-Tagging": "Tagging", "X-Amz-Website-Redirect-Location": "WebsiteRedirectLocation", } ) extra: dict[str, Any] = {} for key, value in headers.items(): try: kwarg = mapping[key] except KeyError: raise TypeError(f'Header "{key}" is not supported by botocore') extra[kwarg] = value return extra class GCSFilesStore: GCS_PROJECT_ID = None CACHE_CONTROL = "max-age=172800" # The bucket's default object ACL will be applied to the object. # Overridden from settings.FILES_STORE_GCS_ACL in FilesPipeline.from_crawler(). POLICY = None def __init__(self, uri: str): from google.cloud import storage # noqa: PLC0415 client = storage.Client(project=self.GCS_PROJECT_ID) bucket, prefix = uri[5:].split("/", 1) self.bucket = client.bucket(bucket) self.prefix: str = prefix permissions = self.bucket.test_iam_permissions( ["storage.objects.get", "storage.objects.create"] ) if "storage.objects.get" not in permissions: logger.warning( "No 'storage.objects.get' permission for GSC bucket %(bucket)s. " "Checking if files are up to date will be impossible. Files will be downloaded every time.", {"bucket": bucket}, ) if "storage.objects.create" not in permissions: logger.error( "No 'storage.objects.create' permission for GSC bucket %(bucket)s. Saving files will be impossible!", {"bucket": bucket}, ) def stat_file( self, path: str, info: MediaPipeline.SpiderInfo ) -> Deferred[StatInfo]: def _onsuccess(blob: Any) -> StatInfo: if blob: checksum = base64.b64decode(blob.md5_hash).hex() last_modified = time.mktime(blob.updated.timetuple()) return {"checksum": checksum, "last_modified": last_modified} return {} blob_path = self._get_blob_path(path) return cast( "Deferred[StatInfo]", deferToThread(self.bucket.get_blob, blob_path).addCallback(_onsuccess), ) def _get_content_type(self, headers: dict[str, str] | None) -> str: if headers and "Content-Type" in headers: return headers["Content-Type"] return "application/octet-stream" def _get_blob_path(self, path: str) -> str: return self.prefix + path def persist_file( self, path: str, buf: BytesIO, info: MediaPipeline.SpiderInfo, meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> Deferred[Any]: blob_path = self._get_blob_path(path) blob = self.bucket.blob(blob_path) blob.cache_control = self.CACHE_CONTROL blob.metadata = {k: str(v) for k, v in (meta or {}).items()} return deferToThread( blob.upload_from_string, data=buf.getvalue(), content_type=self._get_content_type(headers), predefined_acl=self.POLICY, ) class FTPFilesStore: FTP_USERNAME: str | None = None FTP_PASSWORD: str | None = None USE_ACTIVE_MODE: bool | None = None def __init__(self, uri: str): if not uri.startswith("ftp://"): raise ValueError(f"Incorrect URI scheme in {uri}, expected 'ftp'") u = urlparse(uri) assert u.port assert u.hostname self.port: int = u.port self.host: str = u.hostname self.port = int(u.port or 21) assert self.FTP_USERNAME assert self.FTP_PASSWORD self.username: str = u.username or self.FTP_USERNAME self.password: str = u.password or self.FTP_PASSWORD self.basedir: str = u.path.rstrip("/") def persist_file( self, path: str, buf: BytesIO, info: MediaPipeline.SpiderInfo, meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> Deferred[Any]: path = f"{self.basedir}/{path}" return deferToThread( ftp_store_file, path=path, file=buf, host=self.host, port=self.port, username=self.username, password=self.password, use_active_mode=self.USE_ACTIVE_MODE, ) def stat_file( self, path: str, info: MediaPipeline.SpiderInfo ) -> Deferred[StatInfo]: def _stat_file(path: str) -> StatInfo: try: with FTP() as ftp: ftp.connect(self.host, self.port) ftp.login(self.username, self.password) if self.USE_ACTIVE_MODE: ftp.set_pasv(False) file_path = f"{self.basedir}/{path}" last_modified = float(ftp.voidcmd(f"MDTM {file_path}")[4:].strip()) m = hashlib.md5() # noqa: S324 ftp.retrbinary(f"RETR {file_path}", m.update) return {"last_modified": last_modified, "checksum": m.hexdigest()} # The file doesn't exist except Exception: return {} return cast("Deferred[StatInfo]", deferToThread(_stat_file, path)) class FilesPipeline(MediaPipeline): MEDIA_NAME: str = "file" EXPIRES: int = 90 STORE_SCHEMES: dict[str, type[FilesStoreProtocol]] = { "": FSFilesStore, "file": FSFilesStore, "s3": S3FilesStore, "gs": GCSFilesStore, "ftp": FTPFilesStore, } DEFAULT_FILES_URLS_FIELD: str = "file_urls" DEFAULT_FILES_RESULT_FIELD: str = "files" def __init__( self, store_uri: str | PathLike[str], download_func: None = None, *, crawler: Crawler, ): if download_func is not None: # pragma: no cover warnings.warn( "The download_func argument of FilesPipeline.__init__() is ignored" " and will be removed in a future Scrapy version.", category=ScrapyDeprecationWarning, stacklevel=2, ) if not (store_uri and (store_uri := _to_string(store_uri))): from scrapy.pipelines.images import ImagesPipeline # noqa: PLC0415 setting_name = ( "IMAGES_STORE" if isinstance(self, ImagesPipeline) else "FILES_STORE" ) raise NotConfigured( f"{setting_name} setting must be set to a valid path (not empty) " f"to enable {self.__class__.__name__}." ) settings = crawler.settings cls_name = "FilesPipeline" self.store: FilesStoreProtocol = self._get_store(store_uri) resolve = functools.partial( self._key_for_pipe, base_class_name=cls_name, settings=settings ) self.expires: int = settings.getint(resolve("FILES_EXPIRES"), self.EXPIRES) if not hasattr(self, "FILES_URLS_FIELD"): self.FILES_URLS_FIELD = self.DEFAULT_FILES_URLS_FIELD if not hasattr(self, "FILES_RESULT_FIELD"): self.FILES_RESULT_FIELD = self.DEFAULT_FILES_RESULT_FIELD self.files_urls_field: str = settings.get( resolve("FILES_URLS_FIELD"), self.FILES_URLS_FIELD ) self.files_result_field: str = settings.get( resolve("FILES_RESULT_FIELD"), self.FILES_RESULT_FIELD ) super().__init__(crawler=crawler) @classmethod def from_crawler(cls, crawler: Crawler) -> Self: settings = crawler.settings cls._update_stores(settings) store_uri = settings["FILES_STORE"] return cls(store_uri, crawler=crawler) @classmethod def _update_stores(cls, settings: BaseSettings) -> None: s3store: type[S3FilesStore] = cast( "type[S3FilesStore]", cls.STORE_SCHEMES["s3"] ) s3store.AWS_ACCESS_KEY_ID = settings["AWS_ACCESS_KEY_ID"] s3store.AWS_SECRET_ACCESS_KEY = settings["AWS_SECRET_ACCESS_KEY"] s3store.AWS_SESSION_TOKEN = settings["AWS_SESSION_TOKEN"] s3store.AWS_ENDPOINT_URL = settings["AWS_ENDPOINT_URL"] s3store.AWS_REGION_NAME = settings["AWS_REGION_NAME"] s3store.AWS_USE_SSL = settings["AWS_USE_SSL"] s3store.AWS_VERIFY = settings["AWS_VERIFY"] s3store.POLICY = settings["FILES_STORE_S3_ACL"] gcs_store: type[GCSFilesStore] = cast( "type[GCSFilesStore]", cls.STORE_SCHEMES["gs"] ) gcs_store.GCS_PROJECT_ID = settings["GCS_PROJECT_ID"] gcs_store.POLICY = settings["FILES_STORE_GCS_ACL"] or None ftp_store: type[FTPFilesStore] = cast( "type[FTPFilesStore]", cls.STORE_SCHEMES["ftp"] ) ftp_store.FTP_USERNAME = settings["FTP_USER"] ftp_store.FTP_PASSWORD = settings["FTP_PASSWORD"] ftp_store.USE_ACTIVE_MODE = settings.getbool("FEED_STORAGE_FTP_ACTIVE") def _get_store(self, uri: str) -> FilesStoreProtocol: # to support win32 paths like: C:\\some\dir scheme = "file" if Path(uri).is_absolute() else urlparse(uri).scheme store_cls = self.STORE_SCHEMES[scheme] return store_cls(uri) def media_to_download( self, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None ) -> Deferred[FileInfo | None] | None: def _onsuccess(result: StatInfo) -> FileInfo | None: if not result: return None # returning None force download last_modified = result.get("last_modified", None) if not last_modified: return None # returning None force download age_seconds = time.time() - last_modified age_days = age_seconds / 60 / 60 / 24 if age_days > self.expires: return None # returning None force download referer = referer_str(request) logger.debug( "File (uptodate): Downloaded %(medianame)s from %(request)s " "referred in <%(referer)s>", {"medianame": self.MEDIA_NAME, "request": request, "referer": referer}, extra={"spider": info.spider}, ) self.inc_stats("uptodate") checksum = result.get("checksum", None) return { "url": request.url, "path": path, "checksum": checksum, "status": "uptodate", } path = self.file_path(request, info=info, item=item) # maybeDeferred() overloads don't seem to support a Union[_T, Deferred[_T]] return type dfd: Deferred[StatInfo] = maybeDeferred(self.store.stat_file, path, info) # type: ignore[call-overload] dfd2: Deferred[FileInfo | None] = dfd.addCallback(_onsuccess) dfd2.addErrback(lambda _: None) dfd2.addErrback( lambda f: logger.error( self.__class__.__name__ + ".store.stat_file", exc_info=failure_to_exc_info(f), extra={"spider": info.spider}, ) ) return dfd2 def media_failed( self, failure: Failure, request: Request, info: MediaPipeline.SpiderInfo ) -> NoReturn: if not isinstance(failure.value, IgnoreRequest): referer = referer_str(request) logger.warning( "File (unknown-error): Error downloading %(medianame)s from " "%(request)s referred in <%(referer)s>: %(exception)s", { "medianame": self.MEDIA_NAME, "request": request, "referer": referer, "exception": failure.value, }, extra={"spider": info.spider}, ) raise FileException def media_downloaded( self, response: Response, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None, ) -> FileInfo: referer = referer_str(request) if response.status != 200: logger.warning( "File (code: %(status)s): Error downloading file from " "%(request)s referred in <%(referer)s>", {"status": response.status, "request": request, "referer": referer}, extra={"spider": info.spider}, ) raise FileException("download-error") if not response.body: logger.warning( "File (empty-content): Empty file from %(request)s referred " "in <%(referer)s>: no-content", {"request": request, "referer": referer}, extra={"spider": info.spider}, ) raise FileException("empty-content") status = "cached" if "cached" in response.flags else "downloaded" logger.debug( "File (%(status)s): Downloaded file from %(request)s referred in " "<%(referer)s>", {"status": status, "request": request, "referer": referer}, extra={"spider": info.spider}, ) self.inc_stats(status) try: path = self.file_path(request, response=response, info=info, item=item) checksum = self.file_downloaded(response, request, info, item=item) except FileException as exc: logger.warning( "File (error): Error processing file from %(request)s " "referred in <%(referer)s>: %(errormsg)s", {"request": request, "referer": referer, "errormsg": str(exc)}, extra={"spider": info.spider}, exc_info=True, ) raise except Exception as exc: logger.error( "File (unknown-error): Error processing file from %(request)s " "referred in <%(referer)s>", {"request": request, "referer": referer}, exc_info=True, extra={"spider": info.spider}, ) raise FileException(str(exc)) return { "url": request.url, "path": path, "checksum": checksum, "status": status, } def inc_stats(self, status: str) -> None: assert self.crawler.stats self.crawler.stats.inc_value("file_count") self.crawler.stats.inc_value(f"file_status_count/{status}") # Overridable Interface def get_media_requests( self, item: Any, info: MediaPipeline.SpiderInfo ) -> list[Request]: urls = ItemAdapter(item).get(self.files_urls_field, []) if not isinstance(urls, list): raise TypeError( f"{self.files_urls_field} must be a list of URLs, got {type(urls).__name__}. " ) return [Request(u, callback=NO_CALLBACK) for u in urls] def file_downloaded( self, response: Response, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None, ) -> str: path = self.file_path(request, response=response, info=info, item=item) buf = BytesIO(response.body) checksum = _md5sum(buf) buf.seek(0) self.store.persist_file(path, buf, info) return checksum def item_completed( self, results: list[FileInfoOrError], item: Any, info: MediaPipeline.SpiderInfo ) -> Any: with suppress(KeyError): ItemAdapter(item)[self.files_result_field] = [x for ok, x in results if ok] return item def file_path( self, request: Request, response: Response | None = None, info: MediaPipeline.SpiderInfo | None = None, *, item: Any = None, ) -> str: media_guid = hashlib.sha1(to_bytes(request.url)).hexdigest() # noqa: S324 media_ext = Path(request.url).suffix # Handles empty and wild extensions by trying to guess the # mime type then extension or default to empty string otherwise if media_ext not in mimetypes.types_map: media_ext = "" media_type = mimetypes.guess_type(request.url)[0] if media_type: media_ext = cast("str", mimetypes.guess_extension(media_type)) return f"full/{media_guid}{media_ext}"
--- +++ @@ -1,3 +1,8 @@+""" +Files Pipeline + +See documentation in topics/media-pipeline.rst +""" from __future__ import annotations @@ -51,6 +56,13 @@ def _md5sum(file: IO[bytes]) -> str: + """Calculate the md5 checksum of a file-like object without reading its + whole content in memory. + + >>> from io import BytesIO + >>> _md5sum(BytesIO(b'file content to hash')) + '784406af91dd5a54fbb9c84c2236595a' + """ m = hashlib.md5() # noqa: S324 while True: d = file.read(8096) @@ -61,6 +73,7 @@ class FileException(Exception): + """General media error exception""" class StatInfo(TypedDict, total=False): @@ -200,6 +213,7 @@ meta: dict[str, Any] | None = None, headers: dict[str, str] | None = None, ) -> Deferred[Any]: + """Upload file to S3 storage""" key_name = f"{self.prefix}{path}" buf.seek(0) extra = self._headers_to_botocore_kwargs(self.HEADERS) @@ -216,6 +230,7 @@ ) def _headers_to_botocore_kwargs(self, headers: dict[str, Any]) -> dict[str, Any]: + """Convert headers to botocore keyword arguments.""" # This is required while we need to support both boto and botocore. mapping = CaseInsensitiveDict( { @@ -394,6 +409,23 @@ class FilesPipeline(MediaPipeline): + """Abstract pipeline that implement the file downloading + + This pipeline tries to minimize network transfers and file processing, + doing stat of the files and determining if file is new, up-to-date or + expired. + + ``new`` files are those that pipeline never processed and needs to be + downloaded from supplier site the first time. + + ``uptodate`` files are the ones that the pipeline processed and are still + valid files. + + ``expired`` files are those that pipeline already processed but the last + modification was made long time ago, so a reprocessing is recommended to + refresh it in case of change. + + """ MEDIA_NAME: str = "file" EXPIRES: int = 90 @@ -680,4 +712,4 @@ media_type = mimetypes.guess_type(request.url)[0] if media_type: media_ext = cast("str", mimetypes.guess_extension(media_type)) - return f"full/{media_guid}{media_ext}"+ return f"full/{media_guid}{media_ext}"
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/pipelines/files.py
Add docstrings for better understanding
from __future__ import annotations import logging import sys from abc import ABCMeta, abstractmethod from typing import TYPE_CHECKING from urllib.robotparser import RobotFileParser from protego import Protego from scrapy.utils.python import to_unicode if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Spider from scrapy.crawler import Crawler logger = logging.getLogger(__name__) def decode_robotstxt( robotstxt_body: bytes, spider: Spider | None, to_native_str_type: bool = False ) -> str: try: if to_native_str_type: body_decoded = to_unicode(robotstxt_body) else: body_decoded = robotstxt_body.decode("utf-8-sig", errors="ignore") except UnicodeDecodeError: # If we found garbage or robots.txt in an encoding other than UTF-8, disregard it. # Switch to 'allow all' state. logger.warning( "Failure while parsing robots.txt. File either contains garbage or " "is in an encoding other than UTF-8, treating it as an empty file.", exc_info=sys.exc_info(), extra={"spider": spider}, ) body_decoded = "" return body_decoded class RobotParser(metaclass=ABCMeta): @classmethod @abstractmethod def from_crawler(cls, crawler: Crawler, robotstxt_body: bytes) -> Self: @abstractmethod def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: class PythonRobotParser(RobotParser): def __init__(self, robotstxt_body: bytes, spider: Spider | None): self.spider: Spider | None = spider body_decoded = decode_robotstxt(robotstxt_body, spider, to_native_str_type=True) self.rp: RobotFileParser = RobotFileParser() self.rp.parse(body_decoded.splitlines()) @classmethod def from_crawler(cls, crawler: Crawler, robotstxt_body: bytes) -> Self: spider = None if not crawler else crawler.spider return cls(robotstxt_body, spider) def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: user_agent = to_unicode(user_agent) url = to_unicode(url) return self.rp.can_fetch(user_agent, url) class RerpRobotParser(RobotParser): def __init__(self, robotstxt_body: bytes, spider: Spider | None): from robotexclusionrulesparser import RobotExclusionRulesParser # noqa: PLC0415 self.spider: Spider | None = spider self.rp: RobotExclusionRulesParser = RobotExclusionRulesParser() body_decoded = decode_robotstxt(robotstxt_body, spider) self.rp.parse(body_decoded) @classmethod def from_crawler(cls, crawler: Crawler, robotstxt_body: bytes) -> Self: spider = None if not crawler else crawler.spider return cls(robotstxt_body, spider) def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: user_agent = to_unicode(user_agent) url = to_unicode(url) return self.rp.is_allowed(user_agent, url) class ProtegoRobotParser(RobotParser): def __init__(self, robotstxt_body: bytes, spider: Spider | None): self.spider: Spider | None = spider body_decoded = decode_robotstxt(robotstxt_body, spider) self.rp = Protego.parse(body_decoded) @classmethod def from_crawler(cls, crawler: Crawler, robotstxt_body: bytes) -> Self: spider = None if not crawler else crawler.spider return cls(robotstxt_body, spider) def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: user_agent = to_unicode(user_agent) url = to_unicode(url) return self.rp.can_fetch(url, user_agent)
--- +++ @@ -46,9 +46,26 @@ @classmethod @abstractmethod def from_crawler(cls, crawler: Crawler, robotstxt_body: bytes) -> Self: + """Parse the content of a robots.txt_ file as bytes. This must be a class method. + It must return a new instance of the parser backend. + + :param crawler: crawler which made the request + :type crawler: :class:`~scrapy.crawler.Crawler` instance + + :param robotstxt_body: content of a robots.txt_ file. + :type robotstxt_body: bytes + """ @abstractmethod def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: + """Return ``True`` if ``user_agent`` is allowed to crawl ``url``, otherwise return ``False``. + + :param url: Absolute URL + :type url: str or bytes + + :param user_agent: User agent + :type user_agent: str or bytes + """ class PythonRobotParser(RobotParser): @@ -103,4 +120,4 @@ def allowed(self, url: str | bytes, user_agent: str | bytes) -> bool: user_agent = to_unicode(user_agent) url = to_unicode(url) - return self.rp.can_fetch(url, user_agent)+ return self.rp.can_fetch(url, user_agent)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/robotstxt.py
Replace inline comments with docstrings
from __future__ import annotations import functools import hashlib import warnings from contextlib import suppress from io import BytesIO from typing import TYPE_CHECKING, Any from itemadapter import ItemAdapter from scrapy.exceptions import NotConfigured, ScrapyDeprecationWarning from scrapy.http import Request, Response from scrapy.http.request import NO_CALLBACK from scrapy.pipelines.files import FileException, FilesPipeline, _md5sum from scrapy.utils.python import to_bytes if TYPE_CHECKING: from collections.abc import Iterable from os import PathLike from PIL import Image # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.pipelines.media import FileInfoOrError, MediaPipeline class ImageException(FileException): class ImagesPipeline(FilesPipeline): MEDIA_NAME: str = "image" # Uppercase attributes kept for backward compatibility with code that subclasses # ImagesPipeline. They may be overridden by settings. MIN_WIDTH: int = 0 MIN_HEIGHT: int = 0 EXPIRES: int = 90 THUMBS: dict[str, tuple[int, int]] = {} DEFAULT_IMAGES_URLS_FIELD = "image_urls" DEFAULT_IMAGES_RESULT_FIELD = "images" def __init__( self, store_uri: str | PathLike[str], download_func: None = None, *, crawler: Crawler, ): if download_func is not None: # pragma: no cover warnings.warn( "The download_func argument of ImagesPipeline.__init__() is ignored" " and will be removed in a future Scrapy version.", category=ScrapyDeprecationWarning, stacklevel=2, ) try: from PIL import Image, ImageOps # noqa: PLC0415 self._Image = Image self._ImageOps = ImageOps except ImportError: raise NotConfigured( "ImagesPipeline requires installing Pillow 8.3.2 or later" ) super().__init__(store_uri, crawler=crawler) settings = crawler.settings resolve = functools.partial( self._key_for_pipe, base_class_name="ImagesPipeline", settings=settings, ) self.expires: int = settings.getint(resolve("IMAGES_EXPIRES"), self.EXPIRES) if not hasattr(self, "IMAGES_RESULT_FIELD"): self.IMAGES_RESULT_FIELD: str = self.DEFAULT_IMAGES_RESULT_FIELD if not hasattr(self, "IMAGES_URLS_FIELD"): self.IMAGES_URLS_FIELD: str = self.DEFAULT_IMAGES_URLS_FIELD self.images_urls_field: str = settings.get( resolve("IMAGES_URLS_FIELD"), self.IMAGES_URLS_FIELD ) self.images_result_field: str = settings.get( resolve("IMAGES_RESULT_FIELD"), self.IMAGES_RESULT_FIELD ) self.min_width: int = settings.getint( resolve("IMAGES_MIN_WIDTH"), self.MIN_WIDTH ) self.min_height: int = settings.getint( resolve("IMAGES_MIN_HEIGHT"), self.MIN_HEIGHT ) self.thumbs: dict[str, tuple[int, int]] = settings.get( resolve("IMAGES_THUMBS"), self.THUMBS ) @classmethod def from_crawler(cls, crawler: Crawler) -> Self: settings = crawler.settings cls._update_stores(settings) store_uri = settings["IMAGES_STORE"] return cls(store_uri, crawler=crawler) def file_downloaded( self, response: Response, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None, ) -> str: return self.image_downloaded(response, request, info, item=item) def image_downloaded( self, response: Response, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None, ) -> str: checksum: str | None = None for path, image, buf in self.get_images(response, request, info, item=item): if checksum is None: buf.seek(0) checksum = _md5sum(buf) width, height = image.size self.store.persist_file( path, buf, info, meta={"width": width, "height": height}, headers={"Content-Type": "image/jpeg"}, ) assert checksum is not None return checksum def get_images( self, response: Response, request: Request, info: MediaPipeline.SpiderInfo, *, item: Any = None, ) -> Iterable[tuple[str, Image.Image, BytesIO]]: path = self.file_path(request, response=response, info=info, item=item) orig_image = self._Image.open(BytesIO(response.body)) transposed_image = self._ImageOps.exif_transpose(orig_image) width, height = transposed_image.size if width < self.min_width or height < self.min_height: raise ImageException( "Image too small " f"({width}x{height} < " f"{self.min_width}x{self.min_height})" ) image, buf = self.convert_image( transposed_image, response_body=BytesIO(response.body) ) yield path, image, buf for thumb_id, size in self.thumbs.items(): thumb_path = self.thumb_path( request, thumb_id, response=response, info=info, item=item ) thumb_image, thumb_buf = self.convert_image(image, size, response_body=buf) yield thumb_path, thumb_image, thumb_buf def convert_image( self, image: Image.Image, size: tuple[int, int] | None = None, *, response_body: BytesIO, ) -> tuple[Image.Image, BytesIO]: if image.format in ("PNG", "WEBP") and image.mode == "RGBA": background = self._Image.new("RGBA", image.size, (255, 255, 255)) background.paste(image, image) image = background.convert("RGB") elif image.mode == "P": image = image.convert("RGBA") background = self._Image.new("RGBA", image.size, (255, 255, 255)) background.paste(image, image) image = background.convert("RGB") elif image.mode != "RGB": image = image.convert("RGB") if size: image = image.copy() try: # Image.Resampling.LANCZOS was added in Pillow 9.1.0 # remove this try except block, # when updating the minimum requirements for Pillow. resampling_filter = self._Image.Resampling.LANCZOS except AttributeError: resampling_filter = self._Image.ANTIALIAS # type: ignore[attr-defined] image.thumbnail(size, resampling_filter) elif image.format == "JPEG": return image, response_body buf = BytesIO() image.save(buf, "JPEG") return image, buf def get_media_requests( self, item: Any, info: MediaPipeline.SpiderInfo ) -> list[Request]: urls = ItemAdapter(item).get(self.images_urls_field, []) if not isinstance(urls, list): raise TypeError( f"{self.images_urls_field} must be a list of URLs, got {type(urls).__name__}. " ) return [Request(u, callback=NO_CALLBACK) for u in urls] def item_completed( self, results: list[FileInfoOrError], item: Any, info: MediaPipeline.SpiderInfo ) -> Any: with suppress(KeyError): ItemAdapter(item)[self.images_result_field] = [x for ok, x in results if ok] return item def file_path( self, request: Request, response: Response | None = None, info: MediaPipeline.SpiderInfo | None = None, *, item: Any = None, ) -> str: image_guid = hashlib.sha1(to_bytes(request.url)).hexdigest() # noqa: S324 return f"full/{image_guid}.jpg" def thumb_path( self, request: Request, thumb_id: str, response: Response | None = None, info: MediaPipeline.SpiderInfo | None = None, *, item: Any = None, ) -> str: thumb_guid = hashlib.sha1(to_bytes(request.url)).hexdigest() # noqa: S324 return f"thumbs/{thumb_id}/{thumb_guid}.jpg"
--- +++ @@ -1,3 +1,8 @@+""" +Images Pipeline + +See documentation in topics/media-pipeline.rst +""" from __future__ import annotations @@ -30,9 +35,11 @@ class ImageException(FileException): + """General image error exception""" class ImagesPipeline(FilesPipeline): + """Abstract pipeline that implement the image thumbnail generation logic""" MEDIA_NAME: str = "image" @@ -248,4 +255,4 @@ item: Any = None, ) -> str: thumb_guid = hashlib.sha1(to_bytes(request.url)).hexdigest() # noqa: S324 - return f"thumbs/{thumb_id}/{thumb_guid}.jpg"+ return f"thumbs/{thumb_id}/{thumb_guid}.jpg"
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/pipelines/images.py
Document classes and their methods
from __future__ import annotations from collections.abc import Iterable from typing import TYPE_CHECKING, Any, TypeAlias, cast from urllib.parse import urlencode, urljoin, urlsplit, urlunsplit from parsel.csstranslator import HTMLTranslator from w3lib.html import strip_html5_whitespace from scrapy.http.request import Request from scrapy.utils.python import is_listlike, to_bytes if TYPE_CHECKING: # typing.Self requires Python 3.11 from lxml.html import ( FormElement, InputElement, MultipleSelectOptions, SelectElement, TextareaElement, ) from typing_extensions import Self from scrapy.http.response.text import TextResponse FormdataVType: TypeAlias = str | Iterable[str] FormdataKVType: TypeAlias = tuple[str, FormdataVType] FormdataType: TypeAlias = dict[str, FormdataVType] | list[FormdataKVType] | None class FormRequest(Request): __slots__ = () valid_form_methods = ["GET", "POST"] def __init__( self, *args: Any, formdata: FormdataType = None, **kwargs: Any ) -> None: if formdata and kwargs.get("method") is None: kwargs["method"] = "POST" super().__init__(*args, **kwargs) if formdata: items = formdata.items() if isinstance(formdata, dict) else formdata form_query_str = _urlencode(items, self.encoding) if self.method == "POST": self.headers.setdefault( b"Content-Type", b"application/x-www-form-urlencoded" ) self._set_body(form_query_str) else: self._set_url( urlunsplit(urlsplit(self.url)._replace(query=form_query_str)) ) @classmethod def from_response( cls, response: TextResponse, formname: str | None = None, formid: str | None = None, formnumber: int = 0, formdata: FormdataType = None, clickdata: dict[str, str | int] | None = None, dont_click: bool = False, formxpath: str | None = None, formcss: str | None = None, **kwargs: Any, ) -> Self: kwargs.setdefault("encoding", response.encoding) if formcss is not None: formxpath = HTMLTranslator().css_to_xpath(formcss) form = _get_form(response, formname, formid, formnumber, formxpath) formdata = _get_inputs(form, formdata, dont_click, clickdata) url = _get_form_url(form, kwargs.pop("url", None)) method = kwargs.pop("method", form.method) if method is not None: method = method.upper() if method not in cls.valid_form_methods: method = "GET" return cls(url=url, method=method, formdata=formdata, **kwargs) def _get_form_url(form: FormElement, url: str | None) -> str: assert form.base_url is not None # typing if url is None: action = form.get("action") if action is None: return form.base_url return urljoin(form.base_url, strip_html5_whitespace(action)) return urljoin(form.base_url, url) def _urlencode(seq: Iterable[FormdataKVType], enc: str) -> str: values = [ (to_bytes(k, enc), to_bytes(v, enc)) for k, vs in seq for v in (vs if is_listlike(vs) else [cast("str", vs)]) ] return urlencode(values, doseq=True) def _get_form( response: TextResponse, formname: str | None, formid: str | None, formnumber: int, formxpath: str | None, ) -> FormElement: root = response.selector.root forms = root.xpath("//form") if not forms: raise ValueError(f"No <form> element found in {response}") if formname is not None: f = root.xpath(f'//form[@name="{formname}"]') if f: return cast("FormElement", f[0]) if formid is not None: f = root.xpath(f'//form[@id="{formid}"]') if f: return cast("FormElement", f[0]) # Get form element from xpath, if not found, go up if formxpath is not None: nodes = root.xpath(formxpath) if nodes: el = nodes[0] while True: if el.tag == "form": return cast("FormElement", el) el = el.getparent() if el is None: break raise ValueError(f"No <form> element found with {formxpath}") # If we get here, it means that either formname was None or invalid try: form = forms[formnumber] except IndexError: raise IndexError(f"Form number {formnumber} not found in {response}") return cast("FormElement", form) def _get_inputs( form: FormElement, formdata: FormdataType, dont_click: bool, clickdata: dict[str, str | int] | None, ) -> list[FormdataKVType]: try: formdata_keys = dict(formdata or ()).keys() except (ValueError, TypeError): raise ValueError("formdata should be a dict or iterable of tuples") if not formdata: formdata = [] inputs = form.xpath( "descendant::textarea" "|descendant::select" "|descendant::input[not(@type) or @type[" ' not(re:test(., "^(?:submit|image|reset)$", "i"))' " and (../@checked or" ' not(re:test(., "^(?:checkbox|radio)$", "i")))]]', namespaces={"re": "http://exslt.org/regular-expressions"}, ) values: list[FormdataKVType] = [ (k, "" if v is None else v) for k, v in (_value(e) for e in inputs) if k and k not in formdata_keys ] if not dont_click: clickable = _get_clickable(clickdata, form) if clickable and clickable[0] not in formdata and clickable[0] is not None: values.append(clickable) formdata_items = formdata.items() if isinstance(formdata, dict) else formdata values.extend((k, v) for k, v in formdata_items if v is not None) return values def _value( ele: InputElement | SelectElement | TextareaElement, ) -> tuple[str | None, str | MultipleSelectOptions | None]: n = ele.name v = ele.value if ele.tag == "select": return _select_value(cast("SelectElement", ele), n, v) return n, v def _select_value( ele: SelectElement, n: str | None, v: str | MultipleSelectOptions | None ) -> tuple[str | None, str | MultipleSelectOptions | None]: multiple = ele.multiple if v is None and not multiple: # Match browser behaviour on simple select tag without options selected # And for select tags without options o = ele.value_options return (n, o[0]) if o else (None, None) return n, v def _get_clickable( clickdata: dict[str, str | int] | None, form: FormElement ) -> tuple[str, str] | None: clickables = list( form.xpath( 'descendant::input[re:test(@type, "^(submit|image)$", "i")]' '|descendant::button[not(@type) or re:test(@type, "^submit$", "i")]', namespaces={"re": "http://exslt.org/regular-expressions"}, ) ) if not clickables: return None # If we don't have clickdata, we just use the first clickable element if clickdata is None: el = clickables[0] return (el.get("name"), el.get("value") or "") # If clickdata is given, we compare it to the clickable elements to find a # match. We first look to see if the number is specified in clickdata, # because that uniquely identifies the element nr = clickdata.get("nr", None) if nr is not None: assert isinstance(nr, int) try: el = list(form.inputs)[nr] except IndexError: pass else: return (cast("str", el.get("name")), el.get("value") or "") # We didn't find it, so now we build an XPath expression out of the other # arguments, because they can be used as such xpath = ".//*" + "".join(f'[@{k}="{v}"]' for k, v in clickdata.items()) el = form.xpath(xpath) if len(el) == 1: return (el[0].get("name"), el[0].get("value") or "") if len(el) > 1: raise ValueError( f"Multiple elements found ({el!r}) matching the " f"criteria in clickdata: {clickdata!r}" ) raise ValueError(f"No clickable element matching clickdata: {clickdata!r}")
--- +++ @@ -1,3 +1,9 @@+""" +This module implements the FormRequest class which is a more convenient class +(than Request) to generate Requests based on form data. + +See documentation in docs/topics/request-response.rst +""" from __future__ import annotations @@ -114,6 +120,7 @@ formnumber: int, formxpath: str | None, ) -> FormElement: + """Find the wanted form element within the given response.""" root = response.selector.root forms = root.xpath("//form") if not forms: @@ -156,6 +163,7 @@ dont_click: bool, clickdata: dict[str, str | int] | None, ) -> list[FormdataKVType]: + """Return a list of key-value pairs for the inputs found in the given form.""" try: formdata_keys = dict(formdata or ()).keys() except (ValueError, TypeError): @@ -213,6 +221,11 @@ def _get_clickable( clickdata: dict[str, str | int] | None, form: FormElement ) -> tuple[str, str] | None: + """ + Returns the clickable element specified in clickdata, + if the latter is given. If not, it returns the first + clickable element found + """ clickables = list( form.xpath( 'descendant::input[re:test(@type, "^(submit|image)$", "i")]' @@ -252,4 +265,4 @@ f"Multiple elements found ({el!r}) matching the " f"criteria in clickdata: {clickdata!r}" ) - raise ValueError(f"No clickable element matching clickdata: {clickdata!r}")+ raise ValueError(f"No clickable element matching clickdata: {clickdata!r}")
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/http/request/form.py
Create structured documentation for my script
from __future__ import annotations from typing import TYPE_CHECKING, Any from scrapy import Request, Spider from scrapy.utils.decorators import _warn_spider_arg if TYPE_CHECKING: from collections.abc import AsyncIterator, Iterable # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.http import Response class BaseSpiderMiddleware: def __init__(self, crawler: Crawler): self.crawler: Crawler = crawler @classmethod def from_crawler(cls, crawler: Crawler) -> Self: return cls(crawler) def process_start_requests( self, start: Iterable[Any], spider: Spider ) -> Iterable[Any]: for o in start: if (o := self._get_processed(o, None)) is not None: yield o async def process_start(self, start: AsyncIterator[Any]) -> AsyncIterator[Any]: async for o in start: if (o := self._get_processed(o, None)) is not None: yield o @_warn_spider_arg def process_spider_output( self, response: Response, result: Iterable[Any], spider: Spider | None = None ) -> Iterable[Any]: for o in result: if (o := self._get_processed(o, response)) is not None: yield o @_warn_spider_arg async def process_spider_output_async( self, response: Response, result: AsyncIterator[Any], spider: Spider | None = None, ) -> AsyncIterator[Any]: async for o in result: if (o := self._get_processed(o, response)) is not None: yield o def _get_processed(self, o: Any, response: Response | None) -> Any: if isinstance(o, Request): return self.get_processed_request(o, response) return self.get_processed_item(o, response) def get_processed_request( self, request: Request, response: Response | None ) -> Request | None: return request def get_processed_item(self, item: Any, response: Response | None) -> Any: return item
--- +++ @@ -16,6 +16,23 @@ class BaseSpiderMiddleware: + """Optional base class for spider middlewares. + + .. versionadded:: 2.13 + + This class provides helper methods for asynchronous + ``process_spider_output()`` and ``process_start()`` methods. Middlewares + that don't have either of these methods don't need to use this class. + + You can override the + :meth:`~scrapy.spidermiddlewares.base.BaseSpiderMiddleware.get_processed_request` + method to add processing code for requests and the + :meth:`~scrapy.spidermiddlewares.base.BaseSpiderMiddleware.get_processed_item` + method to add processing code for items. These methods take a single + request or item from the spider output iterable and return a request or + item (the same or a new one), or ``None`` to remove this request or item + from the processing. + """ def __init__(self, crawler: Crawler): self.crawler: Crawler = crawler @@ -63,7 +80,37 @@ def get_processed_request( self, request: Request, response: Response | None ) -> Request | None: + """Return a processed request from the spider output. + + This method is called with a single request from the start seeds or the + spider output. It should return the same or a different request, or + ``None`` to ignore it. + + :param request: the input request + :type request: :class:`~scrapy.Request` object + + :param response: the response being processed + :type response: :class:`~scrapy.http.Response` object or ``None`` for + start seeds + + :return: the processed request or ``None`` + """ return request def get_processed_item(self, item: Any, response: Response | None) -> Any: - return item+ """Return a processed item from the spider output. + + This method is called with a single item from the start seeds or the + spider output. It should return the same or a different item, or + ``None`` to ignore it. + + :param item: the input item + :type item: item object + + :param response: the response being processed + :type response: :class:`~scrapy.http.Response` object or ``None`` for + start seeds + + :return: the processed item or ``None`` + """ + return item
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spidermiddlewares/base.py
Generate docstrings for this script
from __future__ import annotations from typing import Any from parsel import Selector as _ParselSelector from scrapy.http import HtmlResponse, TextResponse, XmlResponse from scrapy.utils.python import to_bytes from scrapy.utils.response import get_base_url from scrapy.utils.trackref import object_ref __all__ = ["Selector", "SelectorList"] _NOT_SET = object() def _st(response: TextResponse | None, st: str | None) -> str: if st is None: return "xml" if isinstance(response, XmlResponse) else "html" return st def _response_from_text(text: str | bytes, st: str | None) -> TextResponse: rt: type[TextResponse] = XmlResponse if st == "xml" else HtmlResponse return rt(url="about:blank", encoding="utf-8", body=to_bytes(text, "utf-8")) class SelectorList(_ParselSelector.selectorlist_cls, object_ref): class Selector(_ParselSelector, object_ref): __slots__ = ["response"] selectorlist_cls = SelectorList def __init__( self, response: TextResponse | None = None, text: str | None = None, type: str | None = None, # noqa: A002 root: Any | None = _NOT_SET, **kwargs: Any, ): if response is not None and text is not None: raise ValueError( f"{self.__class__.__name__}.__init__() received both response and text" ) st = _st(response, type) if text is not None: response = _response_from_text(text, st) if response is not None: text = response.text kwargs.setdefault("base_url", get_base_url(response)) self.response = response if root is not _NOT_SET: kwargs["root"] = root super().__init__(text=text, type=st, **kwargs)
--- +++ @@ -1,3 +1,6 @@+""" +XPath selectors based on lxml +""" from __future__ import annotations @@ -27,9 +30,43 @@ class SelectorList(_ParselSelector.selectorlist_cls, object_ref): + """ + The :class:`SelectorList` class is a subclass of the builtin ``list`` + class, which provides a few additional methods. + """ class Selector(_ParselSelector, object_ref): + """ + An instance of :class:`Selector` is a wrapper over response to select + certain parts of its content. + + ``response`` is an :class:`~scrapy.http.HtmlResponse` or an + :class:`~scrapy.http.XmlResponse` object that will be used for selecting + and extracting data. + + ``text`` is a unicode string or utf-8 encoded text for cases when a + ``response`` isn't available. Using ``text`` and ``response`` together is + undefined behavior. + + ``type`` defines the selector type, it can be ``"html"``, ``"xml"``, ``"json"`` + or ``None`` (default). + + If ``type`` is ``None``, the selector automatically chooses the best type + based on ``response`` type (see below), or defaults to ``"html"`` in case it + is used together with ``text``. + + If ``type`` is ``None`` and a ``response`` is passed, the selector type is + inferred from the response type as follows: + + * ``"html"`` for :class:`~scrapy.http.HtmlResponse` type + * ``"xml"`` for :class:`~scrapy.http.XmlResponse` type + * ``"json"`` for :class:`~scrapy.http.TextResponse` type + * ``"html"`` for anything else + + Otherwise, if ``type`` is set, the selector type will be forced and no + detection will occur. + """ __slots__ = ["response"] selectorlist_cls = SelectorList @@ -61,4 +98,4 @@ if root is not _NOT_SET: kwargs["root"] = root - super().__init__(text=text, type=st, **kwargs)+ super().__init__(text=text, type=st, **kwargs)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/selector/unified.py
Document functions with detailed explanations
from __future__ import annotations import traceback import warnings from collections import defaultdict from typing import TYPE_CHECKING, Protocol, cast from zope.interface import implementer from zope.interface.verify import verifyClass from scrapy.interfaces import ISpiderLoader from scrapy.utils.misc import load_object, walk_modules from scrapy.utils.spider import iter_spider_classes if TYPE_CHECKING: from types import ModuleType # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Request, Spider from scrapy.settings import BaseSettings def get_spider_loader(settings: BaseSettings) -> SpiderLoaderProtocol: cls_path = settings.get("SPIDER_LOADER_CLASS") loader_cls = load_object(cls_path) verifyClass(ISpiderLoader, loader_cls) return cast("SpiderLoaderProtocol", loader_cls.from_settings(settings.frozencopy())) class SpiderLoaderProtocol(Protocol): @classmethod def from_settings(cls, settings: BaseSettings) -> Self: def load(self, spider_name: str) -> type[Spider]: def list(self) -> list[str]: def find_by_request(self, request: Request) -> __builtins__.list[str]: @implementer(ISpiderLoader) class SpiderLoader: def __init__(self, settings: BaseSettings): self.spider_modules: list[str] = settings.getlist("SPIDER_MODULES") self.warn_only: bool = settings.getbool("SPIDER_LOADER_WARN_ONLY") self._spiders: dict[str, type[Spider]] = {} self._found: defaultdict[str, list[tuple[str, str]]] = defaultdict(list) self._load_all_spiders() def _check_name_duplicates(self) -> None: dupes = [] for name, locations in self._found.items(): dupes.extend( [ f" {cls} named {name!r} (in {mod})" for mod, cls in locations if len(locations) > 1 ] ) if dupes: dupes_string = "\n\n".join(dupes) warnings.warn( "There are several spiders with the same name:\n\n" f"{dupes_string}\n\n This can cause unexpected behavior.", category=UserWarning, ) def _load_spiders(self, module: ModuleType) -> None: for spcls in iter_spider_classes(module): self._found[spcls.name].append((module.__name__, spcls.__name__)) self._spiders[spcls.name] = spcls def _load_all_spiders(self) -> None: for name in self.spider_modules: try: for module in walk_modules(name): self._load_spiders(module) except (ImportError, SyntaxError): if self.warn_only: warnings.warn( f"\n{traceback.format_exc()}Could not load spiders " f"from module '{name}'. " "See above traceback for details.", category=RuntimeWarning, ) else: raise self._check_name_duplicates() @classmethod def from_settings(cls, settings: BaseSettings) -> Self: return cls(settings) def load(self, spider_name: str) -> type[Spider]: try: return self._spiders[spider_name] except KeyError: raise KeyError(f"Spider not found: {spider_name}") def find_by_request(self, request: Request) -> list[str]: return [ name for name, cls in self._spiders.items() if cls.handles_request(request) ] def list(self) -> list[str]: return list(self._spiders.keys()) @implementer(ISpiderLoader) class DummySpiderLoader: @classmethod def from_settings(cls, settings: BaseSettings) -> Self: return cls() def load(self, spider_name: str) -> type[Spider]: raise KeyError("DummySpiderLoader doesn't load any spiders") def list(self) -> list[str]: return [] def find_by_request(self, request: Request) -> __builtins__.list[str]: return []
--- +++ @@ -23,6 +23,7 @@ def get_spider_loader(settings: BaseSettings) -> SpiderLoaderProtocol: + """Get SpiderLoader instance from settings""" cls_path = settings.get("SPIDER_LOADER_CLASS") loader_cls = load_object(cls_path) verifyClass(ISpiderLoader, loader_cls) @@ -32,16 +33,26 @@ class SpiderLoaderProtocol(Protocol): @classmethod def from_settings(cls, settings: BaseSettings) -> Self: + """Return an instance of the class for the given settings""" def load(self, spider_name: str) -> type[Spider]: + """Return the Spider class for the given spider name. If the spider + name is not found, it must raise a KeyError.""" def list(self) -> list[str]: + """Return a list with the names of all spiders available in the + project""" def find_by_request(self, request: Request) -> __builtins__.list[str]: + """Return the list of spiders names that can handle the given request""" @implementer(ISpiderLoader) class SpiderLoader: + """ + SpiderLoader is a class which locates and loads spiders + in a Scrapy project. + """ def __init__(self, settings: BaseSettings): self.spider_modules: list[str] = settings.getlist("SPIDER_MODULES") @@ -96,22 +107,33 @@ return cls(settings) def load(self, spider_name: str) -> type[Spider]: + """ + Return the Spider class for the given spider name. If the spider + name is not found, raise a KeyError. + """ try: return self._spiders[spider_name] except KeyError: raise KeyError(f"Spider not found: {spider_name}") def find_by_request(self, request: Request) -> list[str]: + """ + Return the list of spider names that can handle the given request. + """ return [ name for name, cls in self._spiders.items() if cls.handles_request(request) ] def list(self) -> list[str]: + """ + Return a list with the names of all spiders available in the project. + """ return list(self._spiders.keys()) @implementer(ISpiderLoader) class DummySpiderLoader: + """A dummy spider loader that does not load any spiders.""" @classmethod def from_settings(cls, settings: BaseSettings) -> Self: @@ -124,4 +146,4 @@ return [] def find_by_request(self, request: Request) -> __builtins__.list[str]: - return []+ return []
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spiderloader.py
Write documentation strings for class attributes
class Link: __slots__ = ["fragment", "nofollow", "text", "url"] def __init__( self, url: str, text: str = "", fragment: str = "", nofollow: bool = False ): if not isinstance(url, str): got = url.__class__.__name__ raise TypeError(f"Link urls must be str objects, got {got}") self.url: str = url self.text: str = text self.fragment: str = fragment self.nofollow: bool = nofollow def __eq__(self, other: object) -> bool: if not isinstance(other, Link): raise NotImplementedError return ( self.url == other.url and self.text == other.text and self.fragment == other.fragment and self.nofollow == other.nofollow ) def __hash__(self) -> int: return ( hash(self.url) ^ hash(self.text) ^ hash(self.fragment) ^ hash(self.nofollow) ) def __repr__(self) -> str: return ( f"Link(url={self.url!r}, text={self.text!r}, " f"fragment={self.fragment!r}, nofollow={self.nofollow!r})" )
--- +++ @@ -1,6 +1,28 @@+""" +This module defines the Link object used in Link extractors. + +For actual link extractors implementation see scrapy.linkextractors, or +its documentation in: docs/topics/link-extractors.rst +""" class Link: + """Link objects represent an extracted link by the LinkExtractor. + + Using the anchor tag sample below to illustrate the parameters:: + + <a href="https://example.com/nofollow.html#foo" rel="nofollow">Dont follow this one</a> + + :param url: the absolute url being linked to in the anchor tag. + From the sample, this is ``https://example.com/nofollow.html``. + + :param text: the text in the anchor tag. From the sample, this is ``Dont follow this one``. + + :param fragment: the part of the url after the hash symbol. From the sample, this is ``foo``. + + :param nofollow: an indication of the presence or absence of a nofollow value in the ``rel`` attribute + of the anchor tag. + """ __slots__ = ["fragment", "nofollow", "text", "url"] @@ -34,4 +56,4 @@ return ( f"Link(url={self.url!r}, text={self.text!r}, " f"fragment={self.fragment!r}, nofollow={self.nofollow!r})" - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/link.py
Write docstrings for backend logic
# pylint: disable=no-method-argument,no-self-argument from zope.interface import Interface class ISpiderLoader(Interface): def from_settings(settings): def load(spider_name): def list(): def find_by_request(request):
--- +++ @@ -5,9 +5,15 @@ class ISpiderLoader(Interface): def from_settings(settings): + """Return an instance of the class for the given settings""" def load(spider_name): + """Return the Spider class for the given spider name. If the spider + name is not found, it must raise a KeyError.""" def list(): + """Return a list with the names of all spiders available in the + project""" - def find_by_request(request):+ def find_by_request(request): + """Return the list of spiders names that can handle the given request"""
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/interfaces.py
Add docstrings including usage examples
from __future__ import annotations from abc import ABCMeta from collections.abc import MutableMapping from copy import deepcopy from pprint import pformat from typing import TYPE_CHECKING, Any, NoReturn from scrapy.utils.trackref import object_ref if TYPE_CHECKING: from collections.abc import Iterator, KeysView # typing.Self requires Python 3.11 from typing_extensions import Self class Field(dict[str, Any]): class ItemMeta(ABCMeta): def __new__( mcs, class_name: str, bases: tuple[type, ...], attrs: dict[str, Any] ) -> ItemMeta: classcell = attrs.pop("__classcell__", None) new_bases = tuple(base._class for base in bases if hasattr(base, "_class")) _class = super().__new__(mcs, "x_" + class_name, new_bases, attrs) fields = getattr(_class, "fields", {}) new_attrs = {} for n in dir(_class): v = getattr(_class, n) if isinstance(v, Field): fields[n] = v elif n in attrs: new_attrs[n] = attrs[n] new_attrs["fields"] = fields new_attrs["_class"] = _class if classcell is not None: new_attrs["__classcell__"] = classcell return super().__new__(mcs, class_name, bases, new_attrs) class Item(MutableMapping[str, Any], object_ref, metaclass=ItemMeta): #: A dictionary containing *all declared fields* for this Item, not only #: those populated. The keys are the field names and the values are the #: :class:`Field` objects used in the :ref:`Item declaration #: <topics-items-declaring>`. fields: dict[str, Field] def __init__(self, *args: Any, **kwargs: Any): self._values: dict[str, Any] = {} if args or kwargs: # avoid creating dict for most common case for k, v in dict(*args, **kwargs).items(): self[k] = v def __getitem__(self, key: str) -> Any: return self._values[key] def __setitem__(self, key: str, value: Any) -> None: if key in self.fields: self._values[key] = value else: raise KeyError(f"{self.__class__.__name__} does not support field: {key}") def __delitem__(self, key: str) -> None: del self._values[key] def __getattr__(self, name: str) -> NoReturn: if name in self.fields: raise AttributeError(f"Use item[{name!r}] to get field value") raise AttributeError(name) def __setattr__(self, name: str, value: Any) -> None: if not name.startswith("_"): raise AttributeError(f"Use item[{name!r}] = {value!r} to set field value") super().__setattr__(name, value) def __len__(self) -> int: return len(self._values) def __iter__(self) -> Iterator[str]: return iter(self._values) __hash__ = object_ref.__hash__ def keys(self) -> KeysView[str]: return self._values.keys() def __repr__(self) -> str: return pformat(dict(self)) def copy(self) -> Self: return self.__class__(self) def deepcopy(self) -> Self: return deepcopy(self)
--- +++ @@ -1,3 +1,8 @@+""" +Scrapy Item + +See documentation in docs/topics/item.rst +""" from __future__ import annotations @@ -17,9 +22,14 @@ class Field(dict[str, Any]): + """Container of field metadata""" class ItemMeta(ABCMeta): + """Metaclass_ of :class:`Item` that handles field definitions. + + .. _metaclass: https://realpython.com/python-metaclasses + """ def __new__( mcs, class_name: str, bases: tuple[type, ...], attrs: dict[str, Any] @@ -45,6 +55,26 @@ class Item(MutableMapping[str, Any], object_ref, metaclass=ItemMeta): + """Base class for scraped items. + + In Scrapy, an object is considered an ``item`` if it's supported by the + `itemadapter`_ library. For example, when the output of a spider callback + is evaluated, only such objects are passed to :ref:`item pipelines + <topics-item-pipeline>`. :class:`Item` is one of the classes supported by + `itemadapter`_ by default. + + Items must declare :class:`Field` attributes, which are processed and stored + in the ``fields`` attribute. This restricts the set of allowed field names + and prevents typos, raising ``KeyError`` when referring to undefined fields. + Additionally, fields can be used to define metadata and control the way + data is processed internally. Please refer to the :ref:`documentation + about fields <topics-items-fields>` for additional information. + + Unlike instances of :class:`dict`, instances of :class:`Item` may be + :ref:`tracked <topics-leaks-trackrefs>` to debug memory leaks. + + .. _itemadapter: https://github.com/scrapy/itemadapter + """ #: A dictionary containing *all declared fields* for this Item, not only #: those populated. The keys are the field names and the values are the @@ -98,4 +128,5 @@ return self.__class__(self) def deepcopy(self) -> Self: - return deepcopy(self)+ """Return a :func:`~copy.deepcopy` of this item.""" + return deepcopy(self)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/item.py
Replace inline comments with docstrings
from __future__ import annotations from typing import TYPE_CHECKING, Any, AnyStr, TypeVar, overload from urllib.parse import urljoin from scrapy.exceptions import NotSupported from scrapy.http.headers import Headers from scrapy.http.request import Request from scrapy.link import Link from scrapy.utils.trackref import object_ref if TYPE_CHECKING: from collections.abc import Callable, Iterable, Mapping from ipaddress import IPv4Address, IPv6Address from twisted.internet.ssl import Certificate from twisted.python.failure import Failure # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.http.request import CallbackT, CookiesT from scrapy.selector import SelectorList ResponseTypeVar = TypeVar("ResponseTypeVar", bound="Response") class Response(object_ref): attributes: tuple[str, ...] = ( "url", "status", "headers", "body", "flags", "request", "certificate", "ip_address", "protocol", ) """A tuple of :class:`str` objects containing the name of all public attributes of the class that are also keyword parameters of the ``__init__()`` method. Currently used by :meth:`Response.replace`. """ def __init__( self, url: str, status: int = 200, headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes = b"", flags: list[str] | None = None, request: Request | None = None, certificate: Certificate | None = None, ip_address: IPv4Address | IPv6Address | None = None, protocol: str | None = None, ): self.headers: Headers = Headers(headers or {}) self.status: int = int(status) self._set_body(body) self._set_url(url) self.request: Request | None = request self.flags: list[str] = [] if flags is None else list(flags) self.certificate: Certificate | None = certificate self.ip_address: IPv4Address | IPv6Address | None = ip_address self.protocol: str | None = protocol @property def cb_kwargs(self) -> dict[str, Any]: try: return self.request.cb_kwargs # type: ignore[union-attr] except AttributeError: raise AttributeError( "Response.cb_kwargs not available, this response " "is not tied to any request" ) @property def meta(self) -> dict[str, Any]: try: return self.request.meta # type: ignore[union-attr] except AttributeError: raise AttributeError( "Response.meta not available, this response is not tied to any request" ) @property def url(self) -> str: return self._url def _set_url(self, url: str) -> None: if isinstance(url, str): self._url: str = url else: raise TypeError( f"{type(self).__name__} url must be str, got {type(url).__name__}" ) @property def body(self) -> bytes: return self._body def _set_body(self, body: bytes | None) -> None: if body is None: self._body = b"" elif not isinstance(body, bytes): raise TypeError( "Response body must be bytes. " "If you want to pass unicode body use TextResponse " "or HtmlResponse." ) else: self._body = body def __repr__(self) -> str: return f"<{self.status} {self.url}>" def copy(self) -> Self: return self.replace() @overload def replace( self, *args: Any, cls: type[ResponseTypeVar], **kwargs: Any ) -> ResponseTypeVar: ... @overload def replace(self, *args: Any, cls: None = None, **kwargs: Any) -> Self: ... def replace( self, *args: Any, cls: type[Response] | None = None, **kwargs: Any ) -> Response: for x in self.attributes: kwargs.setdefault(x, getattr(self, x)) if cls is None: cls = self.__class__ return cls(*args, **kwargs) def urljoin(self, url: str) -> str: return urljoin(self.url, url) @property def text(self) -> str: raise AttributeError("Response content isn't text") def css(self, *a: Any, **kw: Any) -> SelectorList: raise NotSupported("Response content isn't text") def jmespath(self, *a: Any, **kw: Any) -> SelectorList: raise NotSupported("Response content isn't text") def xpath(self, *a: Any, **kw: Any) -> SelectorList: raise NotSupported("Response content isn't text") def follow( self, url: str | Link, callback: CallbackT | None = None, method: str = "GET", headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes | str | None = None, cookies: CookiesT | None = None, meta: dict[str, Any] | None = None, encoding: str | None = "utf-8", priority: int = 0, dont_filter: bool = False, errback: Callable[[Failure], Any] | None = None, cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Request: if encoding is None: raise ValueError("encoding can't be None") if isinstance(url, Link): url = url.url elif url is None: raise ValueError("url can't be None") url = self.urljoin(url) return Request( url=url, callback=callback, method=method, headers=headers, body=body, cookies=cookies, meta=meta, encoding=encoding, priority=priority, dont_filter=dont_filter, errback=errback, cb_kwargs=cb_kwargs, flags=flags, ) def follow_all( self, urls: Iterable[str | Link], callback: CallbackT | None = None, method: str = "GET", headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes | str | None = None, cookies: CookiesT | None = None, meta: dict[str, Any] | None = None, encoding: str | None = "utf-8", priority: int = 0, dont_filter: bool = False, errback: Callable[[Failure], Any] | None = None, cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Iterable[Request]: if not hasattr(urls, "__iter__"): raise TypeError("'urls' argument must be an iterable") return ( self.follow( url=url, callback=callback, method=method, headers=headers, body=body, cookies=cookies, meta=meta, encoding=encoding, priority=priority, dont_filter=dont_filter, errback=errback, cb_kwargs=cb_kwargs, flags=flags, ) for url in urls )
--- +++ @@ -1,3 +1,9 @@+""" +This module implements the Response class which is used to represent HTTP +responses in Scrapy. + +See documentation in docs/topics/request-response.rst +""" from __future__ import annotations @@ -28,6 +34,9 @@ class Response(object_ref): + """An object that represents an HTTP response, which is usually + downloaded (by the Downloader) and fed to the Spiders for processing. + """ attributes: tuple[str, ...] = ( "url", @@ -120,6 +129,7 @@ return f"<{self.status} {self.url}>" def copy(self) -> Self: + """Return a copy of this Response""" return self.replace() @overload @@ -133,6 +143,7 @@ def replace( self, *args: Any, cls: type[Response] | None = None, **kwargs: Any ) -> Response: + """Create a new Response with the same attributes except for those given new values""" for x in self.attributes: kwargs.setdefault(x, getattr(self, x)) if cls is None: @@ -140,19 +151,33 @@ return cls(*args, **kwargs) def urljoin(self, url: str) -> str: + """Join this Response's url with a possible relative url to form an + absolute interpretation of the latter.""" return urljoin(self.url, url) @property def text(self) -> str: + """For subclasses of TextResponse, this will return the body + as str + """ raise AttributeError("Response content isn't text") def css(self, *a: Any, **kw: Any) -> SelectorList: + """Shortcut method implemented only by responses whose content + is text (subclasses of TextResponse). + """ raise NotSupported("Response content isn't text") def jmespath(self, *a: Any, **kw: Any) -> SelectorList: + """Shortcut method implemented only by responses whose content + is text (subclasses of TextResponse). + """ raise NotSupported("Response content isn't text") def xpath(self, *a: Any, **kw: Any) -> SelectorList: + """Shortcut method implemented only by responses whose content + is text (subclasses of TextResponse). + """ raise NotSupported("Response content isn't text") def follow( @@ -171,6 +196,16 @@ cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Request: + """ + Return a :class:`~.Request` instance to follow a link ``url``. + It accepts the same arguments as ``Request.__init__()`` method, + but ``url`` can be a relative URL or a :class:`~scrapy.link.Link` object, + not only an absolute URL. + + :class:`~.TextResponse` provides a :meth:`~.TextResponse.follow` + method which supports selectors in addition to absolute/relative URLs + and Link objects. + """ if encoding is None: raise ValueError("encoding can't be None") if isinstance(url, Link): @@ -211,6 +246,16 @@ cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Iterable[Request]: + """ + Return an iterable of :class:`~.Request` instances to follow all links + in ``urls``. It accepts the same arguments as ``Request.__init__()`` method, + but elements of ``urls`` can be relative URLs or :class:`~scrapy.link.Link` objects, + not only absolute URLs. + + :class:`~.TextResponse` provides a :meth:`~.TextResponse.follow_all` + method which supports selectors in addition to absolute/relative URLs + and Link objects. + """ if not hasattr(urls, "__iter__"): raise TypeError("'urls' argument must be an iterable") return ( @@ -230,4 +275,4 @@ flags=flags, ) for url in urls - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/http/response/__init__.py
Generate helpful docstrings for debugging
from __future__ import annotations import logging import operator import re from collections.abc import Callable, Iterable from functools import partial from typing import TYPE_CHECKING, Any, TypeAlias, cast from urllib.parse import urljoin, urlparse from lxml import etree from parsel.csstranslator import HTMLTranslator from w3lib.html import strip_html5_whitespace from w3lib.url import canonicalize_url, safe_url_string from scrapy.link import Link from scrapy.linkextractors import IGNORED_EXTENSIONS, _is_valid_url, _matches from scrapy.utils.misc import arg_to_iter, rel_has_nofollow from scrapy.utils.python import unique as unique_list from scrapy.utils.response import get_base_url from scrapy.utils.url import url_has_any_extension, url_is_from_any_domain if TYPE_CHECKING: from lxml.html import HtmlElement from scrapy import Selector from scrapy.http import TextResponse logger = logging.getLogger(__name__) # from lxml/src/lxml/html/__init__.py XHTML_NAMESPACE = "http://www.w3.org/1999/xhtml" _collect_string_content = etree.XPath("string()") def _nons(tag: Any) -> Any: if ( isinstance(tag, str) and tag[0] == "{" and tag[1 : len(XHTML_NAMESPACE) + 1] == XHTML_NAMESPACE ): return tag.split("}")[-1] return tag def _identity(x: Any) -> Any: return x def _canonicalize_link_url(link: Link) -> str: return canonicalize_url(link.url, keep_fragments=True) class LxmlParserLinkExtractor: def __init__( self, tag: str | Callable[[str], bool] = "a", attr: str | Callable[[str], bool] = "href", process: Callable[[Any], Any] | None = None, unique: bool = False, strip: bool = True, canonicalized: bool = False, ): # mypy doesn't infer types for operator.* and also for partial() self.scan_tag: Callable[[str], bool] = ( tag if callable(tag) else cast("Callable[[str], bool]", partial(operator.eq, tag)) ) self.scan_attr: Callable[[str], bool] = ( attr if callable(attr) else cast("Callable[[str], bool]", partial(operator.eq, attr)) ) self.process_attr: Callable[[Any], Any] = ( process if callable(process) else _identity ) self.unique: bool = unique self.strip: bool = strip self.link_key: Callable[[Link], str] = ( cast("Callable[[Link], str]", operator.attrgetter("url")) if canonicalized else _canonicalize_link_url ) def _iter_links( self, document: HtmlElement ) -> Iterable[tuple[HtmlElement, str, str]]: for el in document.iter(etree.Element): if not self.scan_tag(_nons(el.tag)): continue attribs = el.attrib for attrib in attribs: if not self.scan_attr(attrib): continue yield el, attrib, attribs[attrib] def _extract_links( self, selector: Selector, response_url: str, response_encoding: str, base_url: str, ) -> list[Link]: links: list[Link] = [] # hacky way to get the underlying lxml parsed document for el, _, attr_val in self._iter_links(selector.root): # pseudo lxml.html.HtmlElement.make_links_absolute(base_url) try: if self.strip: attr_val = strip_html5_whitespace(attr_val) attr_val = urljoin(base_url, attr_val) except ValueError: continue # skipping bogus links else: url = self.process_attr(attr_val) if url is None: continue try: url = safe_url_string(url, encoding=response_encoding) except ValueError: logger.debug(f"Skipping extraction of link with bad URL {url!r}") continue # to fix relative links after process_value url = urljoin(response_url, url) link = Link( url, _collect_string_content(el) or "", nofollow=rel_has_nofollow(el.get("rel")), ) links.append(link) return self._deduplicate_if_needed(links) def extract_links(self, response: TextResponse) -> list[Link]: base_url = get_base_url(response) return self._extract_links( response.selector, response.url, response.encoding, base_url ) def _process_links(self, links: list[Link]) -> list[Link]: return self._deduplicate_if_needed(links) def _deduplicate_if_needed(self, links: list[Link]) -> list[Link]: if self.unique: return unique_list(links, key=self.link_key) return links _Regex: TypeAlias = str | re.Pattern[str] _RegexOrSeveral: TypeAlias = _Regex | Iterable[_Regex] class LxmlLinkExtractor: _csstranslator = HTMLTranslator() def __init__( self, allow: _RegexOrSeveral = (), deny: _RegexOrSeveral = (), allow_domains: str | Iterable[str] = (), deny_domains: str | Iterable[str] = (), restrict_xpaths: str | Iterable[str] = (), tags: str | Iterable[str] = ("a", "area"), attrs: str | Iterable[str] = ("href",), canonicalize: bool = False, unique: bool = True, process_value: Callable[[Any], Any] | None = None, deny_extensions: str | Iterable[str] | None = None, restrict_css: str | Iterable[str] = (), strip: bool = True, restrict_text: _RegexOrSeveral | None = None, ): tags, attrs = set(arg_to_iter(tags)), set(arg_to_iter(attrs)) self.link_extractor = LxmlParserLinkExtractor( tag=partial(operator.contains, tags), attr=partial(operator.contains, attrs), unique=unique, process=process_value, strip=strip, canonicalized=not canonicalize, ) self.allow_res: list[re.Pattern[str]] = self._compile_regexes(allow) self.deny_res: list[re.Pattern[str]] = self._compile_regexes(deny) self.allow_domains: set[str] = set(arg_to_iter(allow_domains)) self.deny_domains: set[str] = set(arg_to_iter(deny_domains)) self.restrict_xpaths: tuple[str, ...] = tuple(arg_to_iter(restrict_xpaths)) self.restrict_xpaths += tuple( map(self._csstranslator.css_to_xpath, arg_to_iter(restrict_css)) ) if deny_extensions is None: deny_extensions = IGNORED_EXTENSIONS self.canonicalize: bool = canonicalize self.deny_extensions: set[str] = {"." + e for e in arg_to_iter(deny_extensions)} self.restrict_text: list[re.Pattern[str]] = self._compile_regexes(restrict_text) @staticmethod def _compile_regexes(value: _RegexOrSeveral | None) -> list[re.Pattern[str]]: return [ x if isinstance(x, re.Pattern) else re.compile(x) for x in arg_to_iter(value) ] def _link_allowed(self, link: Link) -> bool: if not _is_valid_url(link.url): return False if self.allow_res and not _matches(link.url, self.allow_res): return False if self.deny_res and _matches(link.url, self.deny_res): return False parsed_url = urlparse(link.url) if self.allow_domains and not url_is_from_any_domain( parsed_url, self.allow_domains ): return False if self.deny_domains and url_is_from_any_domain(parsed_url, self.deny_domains): return False if self.deny_extensions and url_has_any_extension( parsed_url, self.deny_extensions ): return False return not self.restrict_text or _matches(link.text, self.restrict_text) def matches(self, url: str) -> bool: if self.allow_domains and not url_is_from_any_domain(url, self.allow_domains): return False if self.deny_domains and url_is_from_any_domain(url, self.deny_domains): return False allowed = ( (regex.search(url) for regex in self.allow_res) if self.allow_res else [True] ) denied = (regex.search(url) for regex in self.deny_res) if self.deny_res else [] return any(allowed) and not any(denied) def _process_links(self, links: list[Link]) -> list[Link]: links = [x for x in links if self._link_allowed(x)] if self.canonicalize: for link in links: link.url = canonicalize_url(link.url) return self.link_extractor._process_links(links) def _extract_links(self, *args: Any, **kwargs: Any) -> list[Link]: return self.link_extractor._extract_links(*args, **kwargs) def extract_links(self, response: TextResponse) -> list[Link]: base_url = get_base_url(response) if self.restrict_xpaths: docs = [ subdoc for x in self.restrict_xpaths for subdoc in response.xpath(x) ] else: docs = [response.selector] all_links = [] for doc in docs: links = self._extract_links(doc, response.url, response.encoding, base_url) all_links.extend(self._process_links(links)) if self.link_extractor.unique: return unique_list(all_links, key=self.link_extractor.link_key) return all_links
--- +++ @@ -1,3 +1,6 @@+""" +Link extractor based on lxml.html +""" from __future__ import annotations @@ -142,6 +145,10 @@ ) def _process_links(self, links: list[Link]) -> list[Link]: + """Normalize and filter extracted links + + The subclass should override it if necessary + """ return self._deduplicate_if_needed(links) def _deduplicate_if_needed(self, links: list[Link]) -> list[Link]: @@ -252,6 +259,15 @@ return self.link_extractor._extract_links(*args, **kwargs) def extract_links(self, response: TextResponse) -> list[Link]: + """Returns a list of :class:`~scrapy.link.Link` objects from the + specified :class:`response <scrapy.http.Response>`. + + Only links that match the settings passed to the ``__init__`` method of + the link extractor are returned. + + Duplicate links are omitted if the ``unique`` attribute is set to ``True``, + otherwise they are returned. + """ base_url = get_base_url(response) if self.restrict_xpaths: docs = [ @@ -265,4 +281,4 @@ all_links.extend(self._process_links(links)) if self.link_extractor.unique: return unique_list(all_links, key=self.link_extractor.link_key) - return all_links+ return all_links
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/linkextractors/lxmlhtml.py
Add detailed documentation for each class
from __future__ import annotations import json from contextlib import suppress from typing import TYPE_CHECKING, Any, AnyStr, cast from urllib.parse import urljoin import parsel from w3lib.encoding import ( html_body_declared_encoding, html_to_unicode, http_content_type_encoding, read_bom, resolve_encoding, ) from w3lib.html import strip_html5_whitespace from scrapy.http.response import Response from scrapy.utils.python import memoizemethod_noargs, to_unicode from scrapy.utils.response import get_base_url if TYPE_CHECKING: from collections.abc import Callable, Iterable, Mapping from twisted.python.failure import Failure from scrapy.http.request import CallbackT, CookiesT, Request from scrapy.link import Link from scrapy.selector import Selector, SelectorList _NONE = object() class TextResponse(Response): _DEFAULT_ENCODING = "ascii" _cached_decoded_json = _NONE attributes: tuple[str, ...] = (*Response.attributes, "encoding") def __init__(self, *args: Any, **kwargs: Any): self._encoding: str | None = kwargs.pop("encoding", None) self._cached_benc: str | None = None self._cached_ubody: str | None = None self._cached_selector: Selector | None = None super().__init__(*args, **kwargs) def _set_body(self, body: str | bytes | None) -> None: self._body: bytes = b"" # used by encoding detection if isinstance(body, str): if self._encoding is None: raise TypeError( "Cannot convert unicode body - " f"{type(self).__name__} has no encoding" ) self._body = body.encode(self._encoding) else: super()._set_body(body) @property def encoding(self) -> str: return self._declared_encoding() or self._body_inferred_encoding() def _declared_encoding(self) -> str | None: return ( self._encoding or self._bom_encoding() or self._headers_encoding() or self._body_declared_encoding() ) def json(self) -> Any: if self._cached_decoded_json is _NONE: self._cached_decoded_json = json.loads(self.body) return self._cached_decoded_json @property def text(self) -> str: # access self.encoding before _cached_ubody to make sure # _body_inferred_encoding is called benc = self.encoding if self._cached_ubody is None: charset = f"charset={benc}" self._cached_ubody = html_to_unicode(charset, self.body)[1] return self._cached_ubody def urljoin(self, url: str) -> str: return urljoin(get_base_url(self), url) @memoizemethod_noargs def _headers_encoding(self) -> str | None: content_type = cast("bytes", self.headers.get(b"Content-Type", b"")) return http_content_type_encoding(to_unicode(content_type, encoding="latin-1")) def _body_inferred_encoding(self) -> str: if self._cached_benc is None: content_type = to_unicode( cast("bytes", self.headers.get(b"Content-Type", b"")), encoding="latin-1", ) benc, ubody = html_to_unicode( content_type, self.body, auto_detect_fun=self._auto_detect_fun, default_encoding=self._DEFAULT_ENCODING, ) self._cached_benc = benc self._cached_ubody = ubody return self._cached_benc def _auto_detect_fun(self, text: bytes) -> str | None: for enc in (self._DEFAULT_ENCODING, "utf-8", "cp1252"): try: text.decode(enc) except UnicodeError: continue return resolve_encoding(enc) return None @memoizemethod_noargs def _body_declared_encoding(self) -> str | None: return html_body_declared_encoding(self.body) @memoizemethod_noargs def _bom_encoding(self) -> str | None: return read_bom(self.body)[0] @property def selector(self) -> Selector: # circular import from scrapy.selector import Selector # noqa: PLC0415 if self._cached_selector is None: self._cached_selector = Selector(self) return self._cached_selector def jmespath(self, query: str, **kwargs: Any) -> SelectorList: if not hasattr(self.selector, "jmespath"): raise AttributeError( "Please install parsel >= 1.8.1 to get jmespath support" ) return cast("SelectorList", self.selector.jmespath(query, **kwargs)) def xpath(self, query: str, **kwargs: Any) -> SelectorList: return cast("SelectorList", self.selector.xpath(query, **kwargs)) def css(self, query: str) -> SelectorList: return cast("SelectorList", self.selector.css(query)) def follow( self, url: str | Link | parsel.Selector, callback: CallbackT | None = None, method: str = "GET", headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes | str | None = None, cookies: CookiesT | None = None, meta: dict[str, Any] | None = None, encoding: str | None = None, priority: int = 0, dont_filter: bool = False, errback: Callable[[Failure], Any] | None = None, cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Request: if isinstance(url, parsel.Selector): url = _url_from_selector(url) elif isinstance(url, parsel.SelectorList): raise ValueError("SelectorList is not supported") encoding = self.encoding if encoding is None else encoding return super().follow( url=url, callback=callback, method=method, headers=headers, body=body, cookies=cookies, meta=meta, encoding=encoding, priority=priority, dont_filter=dont_filter, errback=errback, cb_kwargs=cb_kwargs, flags=flags, ) def follow_all( self, urls: Iterable[str | Link] | parsel.SelectorList | None = None, callback: CallbackT | None = None, method: str = "GET", headers: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, body: bytes | str | None = None, cookies: CookiesT | None = None, meta: dict[str, Any] | None = None, encoding: str | None = None, priority: int = 0, dont_filter: bool = False, errback: Callable[[Failure], Any] | None = None, cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, css: str | None = None, xpath: str | None = None, ) -> Iterable[Request]: arguments = [x for x in (urls, css, xpath) if x is not None] if len(arguments) != 1: raise ValueError( "Please supply exactly one of the following arguments: urls, css, xpath" ) if not urls: if css: urls = self.css(css) if xpath: urls = self.xpath(xpath) if isinstance(urls, parsel.SelectorList): selectors = urls urls = [] for sel in selectors: with suppress(_InvalidSelector): urls.append(_url_from_selector(sel)) return super().follow_all( urls=cast("Iterable[str | Link]", urls), callback=callback, method=method, headers=headers, body=body, cookies=cookies, meta=meta, encoding=encoding, priority=priority, dont_filter=dont_filter, errback=errback, cb_kwargs=cb_kwargs, flags=flags, ) class _InvalidSelector(ValueError): def _url_from_selector(sel: parsel.Selector) -> str: if isinstance(sel.root, str): # e.g. ::attr(href) result return strip_html5_whitespace(sel.root) if not hasattr(sel.root, "tag"): raise _InvalidSelector(f"Unsupported selector: {sel}") if sel.root.tag not in ("a", "link"): raise _InvalidSelector( f"Only <a> and <link> elements are supported; got <{sel.root.tag}>" ) href = sel.root.get("href") if href is None: raise _InvalidSelector(f"<{sel.root.tag}> element has no href attribute: {sel}") return strip_html5_whitespace(href)
--- +++ @@ -1,3 +1,9 @@+""" +This module implements the TextResponse class which adds encoding handling and +discovering (through HTTP headers) to base Response class. + +See documentation in docs/topics/request-response.rst +""" from __future__ import annotations @@ -71,12 +77,14 @@ ) def json(self) -> Any: + """Deserialize a JSON document to a Python object.""" if self._cached_decoded_json is _NONE: self._cached_decoded_json = json.loads(self.body) return self._cached_decoded_json @property def text(self) -> str: + """Body as unicode""" # access self.encoding before _cached_ubody to make sure # _body_inferred_encoding is called benc = self.encoding @@ -86,6 +94,8 @@ return self._cached_ubody def urljoin(self, url: str) -> str: + """Join this Response's url with a possible relative url to form an + absolute interpretation of the latter.""" return urljoin(get_base_url(self), url) @memoizemethod_noargs @@ -164,6 +174,22 @@ cb_kwargs: dict[str, Any] | None = None, flags: list[str] | None = None, ) -> Request: + """ + Return a :class:`~.Request` instance to follow a link ``url``. + It accepts the same arguments as ``Request.__init__()`` method, + but ``url`` can be not only an absolute URL, but also + + * a relative URL + * a :class:`~scrapy.link.Link` object, e.g. the result of + :ref:`topics-link-extractors` + * a :class:`~scrapy.Selector` object for a ``<link>`` or ``<a>`` element, e.g. + ``response.css('a.my_link')[0]`` + * an attribute :class:`~scrapy.Selector` (not SelectorList), e.g. + ``response.css('a::attr(href)')[0]`` or + ``response.xpath('//img/@src')[0]`` + + See :ref:`response-follow-example` for usage examples. + """ if isinstance(url, parsel.Selector): url = _url_from_selector(url) elif isinstance(url, parsel.SelectorList): @@ -203,6 +229,29 @@ css: str | None = None, xpath: str | None = None, ) -> Iterable[Request]: + """ + A generator that produces :class:`~.Request` instances to follow all + links in ``urls``. It accepts the same arguments as the :class:`~.Request`'s + ``__init__()`` method, except that each ``urls`` element does not need to be + an absolute URL, it can be any of the following: + + * a relative URL + * a :class:`~scrapy.link.Link` object, e.g. the result of + :ref:`topics-link-extractors` + * a :class:`~scrapy.Selector` object for a ``<link>`` or ``<a>`` element, e.g. + ``response.css('a.my_link')[0]`` + * an attribute :class:`~scrapy.Selector` (not SelectorList), e.g. + ``response.css('a::attr(href)')[0]`` or + ``response.xpath('//img/@src')[0]`` + + In addition, ``css`` and ``xpath`` arguments are accepted to perform the link extraction + within the ``follow_all()`` method (only one of ``urls``, ``css`` and ``xpath`` is accepted). + + Note that when passing a ``SelectorList`` as argument for the ``urls`` parameter or + using the ``css`` or ``xpath`` parameters, this method will not produce requests for + selectors from which links cannot be obtained (for instance, anchor tags without an + ``href`` attribute) + """ arguments = [x for x in (urls, css, xpath) if x is not None] if len(arguments) != 1: raise ValueError( @@ -237,6 +286,9 @@ class _InvalidSelector(ValueError): + """ + Raised when a URL cannot be obtained from a Selector + """ def _url_from_selector(sel: parsel.Selector) -> str: @@ -252,4 +304,4 @@ href = sel.root.get("href") if href is None: raise _InvalidSelector(f"<{sel.root.tag}> element has no href attribute: {sel}") - return strip_html5_whitespace(href)+ return strip_html5_whitespace(href)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/http/response/text.py
Add docstrings to improve code quality
from __future__ import annotations import logging import os from typing import TYPE_CHECKING, Any, TypedDict from twisted.python.failure import Failure # working around https://github.com/sphinx-doc/sphinx/issues/10400 from scrapy import Request, Spider # noqa: TC001 from scrapy.http import Response # noqa: TC001 from scrapy.utils.python import global_object_name from scrapy.utils.request import referer_str if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler SCRAPEDMSG = "Scraped from %(src)s" + os.linesep + "%(item)s" DROPPEDMSG = "Dropped: %(exception)s" + os.linesep + "%(item)s" CRAWLEDMSG = "Crawled (%(status)s) %(request)s%(request_flags)s (referer: %(referer)s)%(response_flags)s" ITEMERRORMSG = "Error processing %(item)s" SPIDERERRORMSG = "Spider error processing %(request)s (referer: %(referer)s)" DOWNLOADERRORMSG_SHORT = "Error downloading %(request)s" DOWNLOADERRORMSG_LONG = "Error downloading %(request)s: %(errmsg)s" class LogFormatterResult(TypedDict): level: int msg: str args: dict[str, Any] | tuple[Any, ...] class LogFormatter: def crawled( self, request: Request, response: Response, spider: Spider ) -> LogFormatterResult: request_flags = f" {request.flags!s}" if request.flags else "" response_flags = f" {response.flags!s}" if response.flags else "" return { "level": logging.DEBUG, "msg": CRAWLEDMSG, "args": { "status": response.status, "request": request, "request_flags": request_flags, "referer": referer_str(request), "response_flags": response_flags, # backward compatibility with Scrapy logformatter below 1.4 version "flags": response_flags, }, } def scraped( self, item: Any, response: Response | Failure | None, spider: Spider ) -> LogFormatterResult: src: Any if response is None: src = f"{global_object_name(spider.__class__)}.start" elif isinstance(response, Failure): src = response.getErrorMessage() else: src = response return { "level": logging.DEBUG, "msg": SCRAPEDMSG, "args": { "src": src, "item": item, }, } def dropped( self, item: Any, exception: BaseException, response: Response | Failure | None, spider: Spider, ) -> LogFormatterResult: if (level := getattr(exception, "log_level", None)) is None: level = spider.crawler.settings["DEFAULT_DROPITEM_LOG_LEVEL"] if isinstance(level, str): level = getattr(logging, level) return { "level": level, "msg": DROPPEDMSG, "args": { "exception": exception, "item": item, }, } def item_error( self, item: Any, exception: BaseException, response: Response | Failure | None, spider: Spider, ) -> LogFormatterResult: return { "level": logging.ERROR, "msg": ITEMERRORMSG, "args": { "item": item, }, } def spider_error( self, failure: Failure, request: Request, response: Response | Failure, spider: Spider, ) -> LogFormatterResult: return { "level": logging.ERROR, "msg": SPIDERERRORMSG, "args": { "request": request, "referer": referer_str(request), }, } def download_error( self, failure: Failure, request: Request, spider: Spider, errmsg: str | None = None, ) -> LogFormatterResult: args: dict[str, Any] = {"request": request} if errmsg: msg = DOWNLOADERRORMSG_LONG args["errmsg"] = errmsg else: msg = DOWNLOADERRORMSG_SHORT return { "level": logging.ERROR, "msg": msg, "args": args, } @classmethod def from_crawler(cls, crawler: Crawler) -> Self: return cls()
--- +++ @@ -35,10 +35,47 @@ class LogFormatter: + """Class for generating log messages for different actions. + + All methods must return a dictionary listing the parameters ``level``, ``msg`` + and ``args`` which are going to be used for constructing the log message when + calling ``logging.log``. + + Dictionary keys for the method outputs: + + * ``level`` is the log level for that action, you can use those from the + `python logging library <https://docs.python.org/3/library/logging.html>`_ : + ``logging.DEBUG``, ``logging.INFO``, ``logging.WARNING``, ``logging.ERROR`` + and ``logging.CRITICAL``. + * ``msg`` should be a string that can contain different formatting placeholders. + This string, formatted with the provided ``args``, is going to be the long message + for that action. + * ``args`` should be a tuple or dict with the formatting placeholders for ``msg``. + The final log message is computed as ``msg % args``. + + Users can define their own ``LogFormatter`` class if they want to customize how + each action is logged or if they want to omit it entirely. In order to omit + logging an action the method must return ``None``. + + Here is an example on how to create a custom log formatter to lower the severity level of + the log message when an item is dropped from the pipeline:: + + class PoliteLogFormatter(logformatter.LogFormatter): + def dropped(self, item, exception, response, spider): + return { + 'level': logging.INFO, # lowering the level from logging.WARNING + 'msg': "Dropped: %(exception)s" + os.linesep + "%(item)s", + 'args': { + 'exception': exception, + 'item': item, + } + } + """ def crawled( self, request: Request, response: Response, spider: Spider ) -> LogFormatterResult: + """Logs a message when the crawler finds a webpage.""" request_flags = f" {request.flags!s}" if request.flags else "" response_flags = f" {response.flags!s}" if response.flags else "" return { @@ -58,6 +95,7 @@ def scraped( self, item: Any, response: Response | Failure | None, spider: Spider ) -> LogFormatterResult: + """Logs a message when an item is scraped by a spider.""" src: Any if response is None: src = f"{global_object_name(spider.__class__)}.start" @@ -81,6 +119,7 @@ response: Response | Failure | None, spider: Spider, ) -> LogFormatterResult: + """Logs a message when an item is dropped while it is passing through the item pipeline.""" if (level := getattr(exception, "log_level", None)) is None: level = spider.crawler.settings["DEFAULT_DROPITEM_LOG_LEVEL"] if isinstance(level, str): @@ -101,6 +140,9 @@ response: Response | Failure | None, spider: Spider, ) -> LogFormatterResult: + """Logs a message when an item causes an error while it is passing + through the item pipeline. + """ return { "level": logging.ERROR, "msg": ITEMERRORMSG, @@ -116,6 +158,7 @@ response: Response | Failure, spider: Spider, ) -> LogFormatterResult: + """Logs an error message from a spider.""" return { "level": logging.ERROR, "msg": SPIDERERRORMSG, @@ -132,6 +175,9 @@ spider: Spider, errmsg: str | None = None, ) -> LogFormatterResult: + """Logs a download error message from a spider (typically coming from + the engine). + """ args: dict[str, Any] = {"request": request} if errmsg: msg = DOWNLOADERRORMSG_LONG @@ -146,4 +192,4 @@ @classmethod def from_crawler(cls, crawler: Crawler) -> Self: - return cls()+ return cls()
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/logformatter.py
Generate docstrings for each module
from __future__ import annotations import logging from typing import TYPE_CHECKING, Any from scrapy.exceptions import IgnoreRequest from scrapy.utils.decorators import _warn_spider_arg if TYPE_CHECKING: from collections.abc import Iterable # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Spider from scrapy.crawler import Crawler from scrapy.http import Response from scrapy.settings import BaseSettings logger = logging.getLogger(__name__) class HttpError(IgnoreRequest): def __init__(self, response: Response, *args: Any, **kwargs: Any): self.response = response super().__init__(*args, **kwargs) class HttpErrorMiddleware: crawler: Crawler def __init__(self, settings: BaseSettings): self.handle_httpstatus_all: bool = settings.getbool("HTTPERROR_ALLOW_ALL") self.handle_httpstatus_list: list[int] = settings.getlist( "HTTPERROR_ALLOWED_CODES" ) @classmethod def from_crawler(cls, crawler: Crawler) -> Self: o = cls(crawler.settings) o.crawler = crawler return o @_warn_spider_arg def process_spider_input( self, response: Response, spider: Spider | None = None ) -> None: if 200 <= response.status < 300: # common case return meta = response.meta if meta.get("handle_httpstatus_all", False): return if "handle_httpstatus_list" in meta: allowed_statuses = meta["handle_httpstatus_list"] elif self.handle_httpstatus_all: return else: allowed_statuses = getattr( self.crawler.spider, "handle_httpstatus_list", self.handle_httpstatus_list, ) if response.status in allowed_statuses: return raise HttpError(response, "Ignoring non-200 response") @_warn_spider_arg def process_spider_exception( self, response: Response, exception: Exception, spider: Spider | None = None ) -> Iterable[Any] | None: if isinstance(exception, HttpError): assert self.crawler.stats self.crawler.stats.inc_value("httperror/response_ignored_count") self.crawler.stats.inc_value( f"httperror/response_ignored_status_count/{response.status}" ) logger.info( "Ignoring response %(response)r: HTTP status code is not handled or not allowed", {"response": response}, extra={"spider": self.crawler.spider}, ) return [] return None
--- +++ @@ -1,3 +1,8 @@+""" +HttpError Spider Middleware + +See documentation in docs/topics/spider-middleware.rst +""" from __future__ import annotations @@ -23,6 +28,7 @@ class HttpError(IgnoreRequest): + """A non-200 response was filtered""" def __init__(self, response: Response, *args: Any, **kwargs: Any): self.response = response @@ -83,4 +89,4 @@ extra={"spider": self.crawler.spider}, ) return [] - return None+ return None
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spidermiddlewares/httperror.py
Write Python docstrings for this snippet
from __future__ import annotations from typing import TYPE_CHECKING, Any from twisted.internet import defer from twisted.internet.base import ReactorBase, ThreadedResolver from twisted.internet.interfaces import ( IAddress, IHostnameResolver, IHostResolution, IResolutionReceiver, IResolverSimple, ) from zope.interface.declarations import implementer, provider from scrapy.utils.datatypes import LocalCache if TYPE_CHECKING: from collections.abc import Sequence from twisted.internet.defer import Deferred # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler # TODO: cache misses dnscache: LocalCache[str, Any] = LocalCache(10000) @implementer(IResolverSimple) class CachingThreadedResolver(ThreadedResolver): def __init__(self, reactor: ReactorBase, cache_size: int, timeout: float): super().__init__(reactor) dnscache.limit = cache_size self.timeout = timeout @classmethod def from_crawler(cls, crawler: Crawler, reactor: ReactorBase) -> Self: if crawler.settings.getbool("DNSCACHE_ENABLED"): cache_size = crawler.settings.getint("DNSCACHE_SIZE") else: cache_size = 0 return cls(reactor, cache_size, crawler.settings.getfloat("DNS_TIMEOUT")) def install_on_reactor(self) -> None: self.reactor.installResolver(self) def getHostByName(self, name: str, timeout: Sequence[int] = ()) -> Deferred[str]: if name in dnscache: return defer.succeed(dnscache[name]) # in Twisted<=16.6, getHostByName() is always called with # a default timeout of 60s (actually passed as (1, 3, 11, 45) tuple), # so the input argument above is simply overridden # to enforce Scrapy's DNS_TIMEOUT setting's value # The timeout arg is typed as Sequence[int] but supports floats. timeout = (self.timeout,) # type: ignore[assignment] d = super().getHostByName(name, timeout) if dnscache.limit: d.addCallback(self._cache_result, name) return d def _cache_result(self, result: Any, name: str) -> Any: dnscache[name] = result return result @implementer(IHostResolution) class HostResolution: def __init__(self, name: str): self.name: str = name def cancel(self) -> None: raise NotImplementedError @provider(IResolutionReceiver) class _CachingResolutionReceiver: def __init__(self, resolutionReceiver: IResolutionReceiver, hostName: str): self.resolutionReceiver: IResolutionReceiver = resolutionReceiver self.hostName: str = hostName self.addresses: list[IAddress] = [] def resolutionBegan(self, resolution: IHostResolution) -> None: self.resolutionReceiver.resolutionBegan(resolution) self.resolution = resolution def addressResolved(self, address: IAddress) -> None: self.resolutionReceiver.addressResolved(address) self.addresses.append(address) def resolutionComplete(self) -> None: self.resolutionReceiver.resolutionComplete() if self.addresses: dnscache[self.hostName] = self.addresses @implementer(IHostnameResolver) class CachingHostnameResolver: def __init__(self, reactor: ReactorBase, cache_size: int): self.reactor: ReactorBase = reactor self.original_resolver: IHostnameResolver = reactor.nameResolver dnscache.limit = cache_size @classmethod def from_crawler(cls, crawler: Crawler, reactor: ReactorBase) -> Self: if crawler.settings.getbool("DNSCACHE_ENABLED"): cache_size = crawler.settings.getint("DNSCACHE_SIZE") else: cache_size = 0 return cls(reactor, cache_size) def install_on_reactor(self) -> None: self.reactor.installNameResolver(self) def resolveHostName( self, resolutionReceiver: IResolutionReceiver, hostName: str, portNumber: int = 0, addressTypes: Sequence[type[IAddress]] | None = None, transportSemantics: str = "TCP", ) -> IHostResolution: try: addresses = dnscache[hostName] except KeyError: return self.original_resolver.resolveHostName( _CachingResolutionReceiver(resolutionReceiver, hostName), hostName, portNumber, addressTypes, transportSemantics, ) resolutionReceiver.resolutionBegan(HostResolution(hostName)) for addr in addresses: resolutionReceiver.addressResolved(addr) resolutionReceiver.resolutionComplete() return resolutionReceiver
--- +++ @@ -31,6 +31,9 @@ @implementer(IResolverSimple) class CachingThreadedResolver(ThreadedResolver): + """ + Default caching resolver. IPv4 only, supports setting a timeout value for DNS requests. + """ def __init__(self, reactor: ReactorBase, cache_size: int, timeout: float): super().__init__(reactor) @@ -99,6 +102,10 @@ @implementer(IHostnameResolver) class CachingHostnameResolver: + """ + Experimental caching resolver. Resolves IPv4 and IPv6 addresses, + does not support setting a timeout value for DNS requests. + """ def __init__(self, reactor: ReactorBase, cache_size: int): self.reactor: ReactorBase = reactor @@ -138,4 +145,4 @@ for addr in addresses: resolutionReceiver.addressResolved(addr) resolutionReceiver.resolutionComplete() - return resolutionReceiver+ return resolutionReceiver
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/resolver.py
Add clean documentation to messy code
from __future__ import annotations import warnings from abc import ABC, abstractmethod from typing import TYPE_CHECKING, cast from urllib.parse import urlparse from warnings import warn from scrapy.exceptions import NotConfigured from scrapy.http import Request, Response from scrapy.spidermiddlewares.base import BaseSpiderMiddleware from scrapy.utils.misc import load_object from scrapy.utils.python import _looks_like_import_path, to_unicode from scrapy.utils.url import strip_url if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self, TypedDict, Unpack from scrapy.crawler import Crawler from scrapy.settings import BaseSettings class _PolicyKwargs(TypedDict, total=False): resp_or_url: Response | str LOCAL_SCHEMES: tuple[str, ...] = ( "about", "blob", "data", "filesystem", ) POLICY_NO_REFERRER = "no-referrer" POLICY_NO_REFERRER_WHEN_DOWNGRADE = "no-referrer-when-downgrade" POLICY_SAME_ORIGIN = "same-origin" POLICY_ORIGIN = "origin" POLICY_STRICT_ORIGIN = "strict-origin" POLICY_ORIGIN_WHEN_CROSS_ORIGIN = "origin-when-cross-origin" POLICY_STRICT_ORIGIN_WHEN_CROSS_ORIGIN = "strict-origin-when-cross-origin" POLICY_UNSAFE_URL = "unsafe-url" POLICY_SCRAPY_DEFAULT = "scrapy-default" class ReferrerPolicy(ABC): NOREFERRER_SCHEMES: tuple[str, ...] = LOCAL_SCHEMES name: str @abstractmethod def referrer(self, response_url: str, request_url: str) -> str | None: raise NotImplementedError def stripped_referrer(self, url: str) -> str | None: if urlparse(url).scheme not in self.NOREFERRER_SCHEMES: return self.strip_url(url) return None def origin_referrer(self, url: str) -> str | None: if urlparse(url).scheme not in self.NOREFERRER_SCHEMES: return self.origin(url) return None def strip_url(self, url: str, origin_only: bool = False) -> str | None: if not url: return None return strip_url( url, strip_credentials=True, strip_fragment=True, strip_default_port=True, origin_only=origin_only, ) def origin(self, url: str) -> str | None: return self.strip_url(url, origin_only=True) def potentially_trustworthy(self, url: str) -> bool: # Note: this does not follow https://w3c.github.io/webappsec-secure-contexts/#is-url-trustworthy parsed_url = urlparse(url) if parsed_url.scheme == "data": return False return self.tls_protected(url) def tls_protected(self, url: str) -> bool: return urlparse(url).scheme in ("https", "ftps") class NoReferrerPolicy(ReferrerPolicy): name: str = POLICY_NO_REFERRER def referrer(self, response_url: str, request_url: str) -> str | None: return None class NoReferrerWhenDowngradePolicy(ReferrerPolicy): name: str = POLICY_NO_REFERRER_WHEN_DOWNGRADE def referrer(self, response_url: str, request_url: str) -> str | None: if not self.tls_protected(response_url) or self.tls_protected(request_url): return self.stripped_referrer(response_url) return None class SameOriginPolicy(ReferrerPolicy): name: str = POLICY_SAME_ORIGIN def referrer(self, response_url: str, request_url: str) -> str | None: if self.origin(response_url) == self.origin(request_url): return self.stripped_referrer(response_url) return None class OriginPolicy(ReferrerPolicy): name: str = POLICY_ORIGIN def referrer(self, response_url: str, request_url: str) -> str | None: return self.origin_referrer(response_url) class StrictOriginPolicy(ReferrerPolicy): name: str = POLICY_STRICT_ORIGIN def referrer(self, response_url: str, request_url: str) -> str | None: if ( self.tls_protected(response_url) and self.potentially_trustworthy(request_url) ) or not self.tls_protected(response_url): return self.origin_referrer(response_url) return None class OriginWhenCrossOriginPolicy(ReferrerPolicy): name: str = POLICY_ORIGIN_WHEN_CROSS_ORIGIN def referrer(self, response_url: str, request_url: str) -> str | None: origin = self.origin(response_url) if origin == self.origin(request_url): return self.stripped_referrer(response_url) return origin class StrictOriginWhenCrossOriginPolicy(ReferrerPolicy): name: str = POLICY_STRICT_ORIGIN_WHEN_CROSS_ORIGIN def referrer(self, response_url: str, request_url: str) -> str | None: origin = self.origin(response_url) if origin == self.origin(request_url): return self.stripped_referrer(response_url) if ( self.tls_protected(response_url) and self.potentially_trustworthy(request_url) ) or not self.tls_protected(response_url): return self.origin_referrer(response_url) return None class UnsafeUrlPolicy(ReferrerPolicy): name: str = POLICY_UNSAFE_URL def referrer(self, response_url: str, request_url: str) -> str | None: return self.stripped_referrer(response_url) class DefaultReferrerPolicy(NoReferrerWhenDowngradePolicy): NOREFERRER_SCHEMES: tuple[str, ...] = (*LOCAL_SCHEMES, "file", "s3") name: str = POLICY_SCRAPY_DEFAULT class RefererMiddleware(BaseSpiderMiddleware): def __init__(self, settings: BaseSettings | None = None): # pylint: disable=super-init-not-called self.default_policy: type[ReferrerPolicy] = DefaultReferrerPolicy self.policies: dict[str, type[ReferrerPolicy]] = { p.name: p for p in ( NoReferrerPolicy, NoReferrerWhenDowngradePolicy, SameOriginPolicy, OriginPolicy, StrictOriginPolicy, OriginWhenCrossOriginPolicy, StrictOriginWhenCrossOriginPolicy, UnsafeUrlPolicy, DefaultReferrerPolicy, ) } # Reference: https://www.w3.org/TR/referrer-policy/#referrer-policy-empty-string self.policies[""] = NoReferrerWhenDowngradePolicy if settings is None: return setting_policies = settings.getdict("REFERRER_POLICIES") for policy_name, policy_class_import_path in setting_policies.items(): if policy_class_import_path is None: del self.policies[policy_name] else: self.policies[policy_name] = load_object(policy_class_import_path) settings_policy = self._load_policy_class( settings.get("REFERRER_POLICY"), allow_import_path=True ) assert settings_policy self.default_policy = settings_policy @classmethod def from_crawler(cls, crawler: Crawler) -> Self: if not crawler.settings.getbool("REFERER_ENABLED"): raise NotConfigured return cls(crawler.settings) def policy( self, response: Response | str | None = None, request: Request | None = None, **kwargs: Unpack[_PolicyKwargs], ) -> ReferrerPolicy: if "resp_or_url" in kwargs: if response is not None: raise TypeError("Cannot pass both 'response' and 'resp_or_url'") response = kwargs.pop("resp_or_url") warn( "Passing 'resp_or_url' is deprecated, use 'response' instead.", DeprecationWarning, stacklevel=2, ) if response is None: raise TypeError("Missing required argument: 'response'") if request is None: raise TypeError("Missing required argument: 'request'") if isinstance(response, str): warn( "Passing a response URL to RefererMiddleware.policy() instead " "of a Response object is deprecated.", DeprecationWarning, stacklevel=2, ) allow_import_path = True policy_name = request.meta.get("referrer_policy") if policy_name is None and isinstance(response, Response): policy_header = response.headers.get("Referrer-Policy") if policy_header is not None: policy_name = to_unicode(policy_header.decode("latin1")) allow_import_path = False if policy_name is None: return self.default_policy() cls = self._load_policy_class( policy_name, warning_only=True, allow_import_path=allow_import_path ) return cls() if cls else self.default_policy() def _load_policy_class( self, policy: str, warning_only: bool = False, *, allow_import_path: bool = False, ) -> type[ReferrerPolicy] | None: if allow_import_path: try: return cast("type[ReferrerPolicy]", load_object(policy)) except ValueError: pass policy_names = [ policy_name.strip() for policy_name in policy.lower().split(",") ] # https://www.w3.org/TR/referrer-policy/#parse-referrer-policy-from-header for policy_name in policy_names[::-1]: if policy_name in self.policies: return self.policies[policy_name] msg = f"Could not load referrer policy {policy!r}" if not allow_import_path and _looks_like_import_path(policy): msg += " (import paths from the response Referrer-Policy header are not allowed)" if not warning_only: raise RuntimeError(msg) warnings.warn(msg, RuntimeWarning) return None def get_processed_request( self, request: Request, response: Response | None ) -> Request | None: if response is None: # start requests return request referrer = self.policy(response, request).referrer(response.url, request.url) if referrer is not None: request.headers.setdefault("Referer", referrer) return request
--- +++ @@ -1,3 +1,7 @@+""" +RefererMiddleware: populates Request referer field, based on the Response which +originated it. +""" from __future__ import annotations @@ -44,6 +48,7 @@ class ReferrerPolicy(ABC): + """Abstract base class for referrer policies.""" NOREFERRER_SCHEMES: tuple[str, ...] = LOCAL_SCHEMES name: str @@ -63,6 +68,19 @@ return None def strip_url(self, url: str, origin_only: bool = False) -> str | None: + """ + https://www.w3.org/TR/referrer-policy/#strip-url + + If url is null, return no referrer. + If url's scheme is a local scheme, then return no referrer. + Set url's username to the empty string. + Set url's password to null. + Set url's fragment to null. + If the origin-only flag is true, then: + Set url's path to null. + Set url's query to null. + Return url. + """ if not url: return None return strip_url( @@ -74,6 +92,7 @@ ) def origin(self, url: str) -> str | None: + """Return serialized origin (scheme, host, path) for a request or response URL.""" return self.strip_url(url, origin_only=True) def potentially_trustworthy(self, url: str) -> bool: @@ -88,6 +107,13 @@ class NoReferrerPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-no-referrer + + The simplest policy is "no-referrer", which specifies that no referrer information + is to be sent along with requests made from a particular request client to any origin. + The header will be omitted entirely. + """ name: str = POLICY_NO_REFERRER @@ -96,6 +122,19 @@ class NoReferrerWhenDowngradePolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-no-referrer-when-downgrade + + The "no-referrer-when-downgrade" policy sends a full URL along with requests + from a TLS-protected environment settings object to a potentially trustworthy URL, + and requests from clients which are not TLS-protected to any origin. + + Requests from TLS-protected clients to non-potentially trustworthy URLs, + on the other hand, will contain no referrer information. + A Referer HTTP header will not be sent. + + This is a user agent's default behavior, if no policy is otherwise specified. + """ name: str = POLICY_NO_REFERRER_WHEN_DOWNGRADE @@ -106,6 +145,15 @@ class SameOriginPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-same-origin + + The "same-origin" policy specifies that a full URL, stripped for use as a referrer, + is sent as referrer information when making same-origin requests from a particular request client. + + Cross-origin requests, on the other hand, will contain no referrer information. + A Referer HTTP header will not be sent. + """ name: str = POLICY_SAME_ORIGIN @@ -116,6 +164,14 @@ class OriginPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-origin + + The "origin" policy specifies that only the ASCII serialization + of the origin of the request client is sent as referrer information + when making both same-origin requests and cross-origin requests + from a particular request client. + """ name: str = POLICY_ORIGIN @@ -124,6 +180,18 @@ class StrictOriginPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-strict-origin + + The "strict-origin" policy sends the ASCII serialization + of the origin of the request client when making requests: + - from a TLS-protected environment settings object to a potentially trustworthy URL, and + - from non-TLS-protected environment settings objects to any origin. + + Requests from TLS-protected request clients to non- potentially trustworthy URLs, + on the other hand, will contain no referrer information. + A Referer HTTP header will not be sent. + """ name: str = POLICY_STRICT_ORIGIN @@ -137,6 +205,16 @@ class OriginWhenCrossOriginPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-origin-when-cross-origin + + The "origin-when-cross-origin" policy specifies that a full URL, + stripped for use as a referrer, is sent as referrer information + when making same-origin requests from a particular request client, + and only the ASCII serialization of the origin of the request client + is sent as referrer information when making cross-origin requests + from a particular request client. + """ name: str = POLICY_ORIGIN_WHEN_CROSS_ORIGIN @@ -148,6 +226,22 @@ class StrictOriginWhenCrossOriginPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-strict-origin-when-cross-origin + + The "strict-origin-when-cross-origin" policy specifies that a full URL, + stripped for use as a referrer, is sent as referrer information + when making same-origin requests from a particular request client, + and only the ASCII serialization of the origin of the request client + when making cross-origin requests: + + - from a TLS-protected environment settings object to a potentially trustworthy URL, and + - from non-TLS-protected environment settings objects to any origin. + + Requests from TLS-protected clients to non- potentially trustworthy URLs, + on the other hand, will contain no referrer information. + A Referer HTTP header will not be sent. + """ name: str = POLICY_STRICT_ORIGIN_WHEN_CROSS_ORIGIN @@ -164,6 +258,18 @@ class UnsafeUrlPolicy(ReferrerPolicy): + """ + https://www.w3.org/TR/referrer-policy/#referrer-policy-unsafe-url + + The "unsafe-url" policy specifies that a full URL, stripped for use as a referrer, + is sent along with both cross-origin requests + and same-origin requests made from a particular request client. + + Note: The policy's name doesn't lie; it is unsafe. + This policy will leak origins and paths from TLS-protected resources + to insecure origins. + Carefully consider the impact of setting such a policy for potentially sensitive documents. + """ name: str = POLICY_UNSAFE_URL @@ -172,6 +278,11 @@ class DefaultReferrerPolicy(NoReferrerWhenDowngradePolicy): + """ + A variant of "no-referrer-when-downgrade", + with the addition that "Referer" is not sent if the parent request was + using ``file://`` or ``s3://`` scheme. + """ NOREFERRER_SCHEMES: tuple[str, ...] = (*LOCAL_SCHEMES, "file", "s3") name: str = POLICY_SCRAPY_DEFAULT @@ -222,6 +333,16 @@ request: Request | None = None, **kwargs: Unpack[_PolicyKwargs], ) -> ReferrerPolicy: + """Return the referrer policy to use for *request* based on *request* + meta, *response* and settings. + + - if a valid policy is set in Request meta, it is used. + - if the policy is set in meta but is wrong (e.g. a typo error), the + policy from settings is used + - if the policy is not set in Request meta, but there is a + Referrer-Policy header in the parent response, it is used if valid + - otherwise, the policy from settings is used. + """ if "resp_or_url" in kwargs: if response is not None: raise TypeError("Cannot pass both 'response' and 'resp_or_url'") @@ -263,6 +384,29 @@ *, allow_import_path: bool = False, ) -> type[ReferrerPolicy] | None: + """Load the :class:`ReferrerPolicy` class to use for *policy*. + + *policy* may be any of the following: + + - A standard policy name, e.g. ``"no-referrer"``, + ``"origin-when-cross-origin"``, etc. + + - The special ``"scrapy-default"`` policy. + + - The import path of a :class:`ReferrerPolicy` subclass, e.g. + ``"scrapy.spidermiddlewares.referer.NoReferrerPolicy"`` or + ``"myproject.policies.CustomReferrerPolicy"``. + + If *warning_only* is ``False`` (default) and *policy* cannot be turned + into a :class:`ReferrerPolicy` subclass, a :exc:`RuntimeError` is + raised. If *warning_only* is ``True``, a warning is logged and ``None`` + is returned instead. + + If *allow_import_path* is ``False`` (default), import paths are not + allowed, resulting in :exc:`RuntimeError` or ``None``. If ``True``, + they are allowed. Use ``True`` only if you trust the source of the + *policy* value. + """ if allow_import_path: try: return cast("type[ReferrerPolicy]", load_object(policy)) @@ -292,4 +436,4 @@ referrer = self.policy(response, request).referrer(response.url, request.url) if referrer is not None: request.headers.setdefault("Referer", referrer) - return request+ return request
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spidermiddlewares/referer.py
Add docstrings including usage examples
from __future__ import annotations from typing import TYPE_CHECKING, Any import itemloaders from scrapy.item import Item from scrapy.selector import Selector if TYPE_CHECKING: from scrapy.http import TextResponse class ItemLoader(itemloaders.ItemLoader): default_item_class: type = Item default_selector_class = Selector def __init__( self, item: Any = None, selector: Selector | None = None, response: TextResponse | None = None, parent: itemloaders.ItemLoader | None = None, **context: Any, ): if selector is None and response is not None: try: selector = self.default_selector_class(response) except AttributeError: selector = None context.update(response=response) super().__init__(item=item, selector=selector, parent=parent, **context)
--- +++ @@ -1,3 +1,8 @@+""" +Item Loader + +See documentation in docs/topics/loaders.rst +""" from __future__ import annotations @@ -13,6 +18,73 @@ class ItemLoader(itemloaders.ItemLoader): + """ + A user-friendly abstraction to populate an :ref:`item <topics-items>` with data + by applying :ref:`field processors <topics-loaders-processors>` to scraped data. + When instantiated with a ``selector`` or a ``response`` it supports + data extraction from web pages using :ref:`selectors <topics-selectors>`. + + :param item: The item instance to populate using subsequent calls to + :meth:`~ItemLoader.add_xpath`, :meth:`~ItemLoader.add_css`, + or :meth:`~ItemLoader.add_value`. + :type item: scrapy.item.Item + + :param selector: The selector to extract data from, when using the + :meth:`add_xpath`, :meth:`add_css`, :meth:`replace_xpath`, or + :meth:`replace_css` method. + :type selector: :class:`~scrapy.Selector` object + + :param response: The response used to construct the selector using the + :attr:`default_selector_class`, unless the selector argument is given, + in which case this argument is ignored. + :type response: :class:`~scrapy.http.Response` object + + If no item is given, one is instantiated automatically using the class in + :attr:`default_item_class`. + + The item, selector, response and remaining keyword arguments are + assigned to the Loader context (accessible through the :attr:`context` attribute). + + .. attribute:: item + + The item object being parsed by this Item Loader. + This is mostly used as a property so, when attempting to override this + value, you may want to check out :attr:`default_item_class` first. + + .. attribute:: context + + The currently active :ref:`Context <loaders-context>` of this Item Loader. + + .. attribute:: default_item_class + + An :ref:`item <topics-items>` class (or factory), used to instantiate + items when not given in the ``__init__`` method. + + .. attribute:: default_input_processor + + The default input processor to use for those fields which don't specify + one. + + .. attribute:: default_output_processor + + The default output processor to use for those fields which don't specify + one. + + .. attribute:: default_selector_class + + The class used to construct the :attr:`selector` of this + :class:`ItemLoader`, if only a response is given in the ``__init__`` method. + If a selector is given in the ``__init__`` method this attribute is ignored. + This attribute is sometimes overridden in subclasses. + + .. attribute:: selector + + The :class:`~scrapy.Selector` object to extract data from. + It's either the selector given in the ``__init__`` method or one created from + the response given in the ``__init__`` method using the + :attr:`default_selector_class`. This attribute is meant to be + read-only. + """ default_item_class: type = Item default_selector_class = Selector @@ -31,4 +103,4 @@ except AttributeError: selector = None context.update(response=response) - super().__init__(item=item, selector=selector, parent=parent, **context)+ super().__init__(item=item, selector=selector, parent=parent, **context)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/loader/__init__.py
Write docstrings for backend logic
from __future__ import annotations import copy import json import warnings from collections.abc import Iterable, Iterator, Mapping, MutableMapping from importlib import import_module from logging import getLogger from pprint import pformat from typing import TYPE_CHECKING, Any, TypeAlias, cast from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.settings import default_settings from scrapy.utils.misc import load_object from scrapy.utils.python import global_object_name logger = getLogger(__name__) # The key types are restricted in BaseSettings._get_key() to ones supported by JSON, # see https://github.com/scrapy/scrapy/issues/5383. _SettingsKey: TypeAlias = bool | float | int | str | None if TYPE_CHECKING: from types import ModuleType # https://github.com/python/typing/issues/445#issuecomment-1131458824 from _typeshed import SupportsItems # typing.Self requires Python 3.11 from typing_extensions import Self _SettingsInput: TypeAlias = SupportsItems[_SettingsKey, Any] | str | None SETTINGS_PRIORITIES: dict[str, int] = { "default": 0, "command": 10, "addon": 15, "project": 20, "spider": 30, "cmdline": 40, } def get_settings_priority(priority: int | str) -> int: if isinstance(priority, str): return SETTINGS_PRIORITIES[priority] return priority class SettingsAttribute: def __init__(self, value: Any, priority: int): self.value: Any = value self.priority: int if isinstance(self.value, BaseSettings): self.priority = max(self.value.maxpriority(), priority) else: self.priority = priority def set(self, value: Any, priority: int) -> None: if priority >= self.priority: if isinstance(self.value, BaseSettings): value = BaseSettings(value, priority=priority) self.value = value self.priority = priority def __repr__(self) -> str: return f"<SettingsAttribute value={self.value!r} priority={self.priority}>" class BaseSettings(MutableMapping[_SettingsKey, Any]): __default = object() def __init__(self, values: _SettingsInput = None, priority: int | str = "project"): self.frozen: bool = False self.attributes: dict[_SettingsKey, SettingsAttribute] = {} if values: self.update(values, priority) def __getitem__(self, opt_name: _SettingsKey) -> Any: if opt_name not in self: return None return self.attributes[opt_name].value def __contains__(self, name: Any) -> bool: return name in self.attributes def add_to_list(self, name: _SettingsKey, item: Any) -> None: value: list[str] = self.getlist(name) if item not in value: self.set(name, [*value, item], self.getpriority(name) or 0) def remove_from_list(self, name: _SettingsKey, item: Any) -> None: value: list[str] = self.getlist(name) if item not in value: raise ValueError(f"{item!r} not found in the {name} setting ({value!r}).") self.set(name, [v for v in value if v != item], self.getpriority(name) or 0) def get(self, name: _SettingsKey, default: Any = None) -> Any: if name == "CONCURRENT_REQUESTS_PER_IP" and ( isinstance(self[name], int) and self[name] != 0 ): warnings.warn( "The CONCURRENT_REQUESTS_PER_IP setting is deprecated, use CONCURRENT_REQUESTS_PER_DOMAIN instead.", ScrapyDeprecationWarning, stacklevel=2, ) return self[name] if self[name] is not None else default def getbool(self, name: _SettingsKey, default: bool = False) -> bool: got = self.get(name, default) try: return bool(int(got)) except ValueError: if got in ("True", "true"): return True if got in ("False", "false"): return False raise ValueError( "Supported values for boolean settings " "are 0/1, True/False, '0'/'1', " "'True'/'False' and 'true'/'false'" ) def getint(self, name: _SettingsKey, default: int = 0) -> int: return int(self.get(name, default)) def getfloat(self, name: _SettingsKey, default: float = 0.0) -> float: return float(self.get(name, default)) def getlist( self, name: _SettingsKey, default: list[Any] | None = None ) -> list[Any]: value = self.get(name, default or []) if not value: return [] if isinstance(value, str): value = value.split(",") return list(value) def getdict( self, name: _SettingsKey, default: dict[Any, Any] | None = None ) -> dict[Any, Any]: value = self.get(name, default or {}) if isinstance(value, str): value = json.loads(value) return dict(value) def getdictorlist( self, name: _SettingsKey, default: dict[Any, Any] | list[Any] | tuple[Any] | None = None, ) -> dict[Any, Any] | list[Any]: value = self.get(name, default) if value is None: return {} if isinstance(value, str): try: value_loaded = json.loads(value) if not isinstance(value_loaded, (dict, list)): raise ValueError( f"JSON string for setting '{name}' must evaluate to a dict or list, " f"got {type(value_loaded).__name__}: {value_loaded!r}" ) return value_loaded except ValueError: return value.split(",") if isinstance(value, tuple): return list(value) if not isinstance(value, (dict, list)): raise ValueError( f"Setting '{name}' must be a dict, list, tuple, or string, " f"got {type(value).__name__}: {value!r}" ) return copy.deepcopy(value) def getwithbase(self, name: _SettingsKey) -> BaseSettings: if not isinstance(name, str): raise ValueError(f"Base setting key must be a string, got {name}") normalized_keys = {} obj_keys = set() def track_loaded_key(k: Any) -> None: if k not in obj_keys: obj_keys.add(k) return logger.warning( f"Setting {name} contains multiple keys that refer to the " f"same object: {global_object_name(k)}. Only the last one will " f"be kept." ) def normalize_key(key: Any) -> str: try: loaded_key = load_object(key) except (AttributeError, TypeError, ValueError): loaded_key = key else: import_path = global_object_name(loaded_key) normalized_keys[import_path] = key key = import_path track_loaded_key(loaded_key) return key def restore_key(k: str) -> Any: return normalized_keys.get(k, k) result = dict(self[name + "_BASE"] or {}) override = {normalize_key(k): v for k, v in (self[name] or {}).items()} result.update(override) return BaseSettings( {restore_key(k): v for k, v in result.items() if v is not None} ) def getpriority(self, name: _SettingsKey) -> int | None: if name not in self: return None return self.attributes[name].priority def maxpriority(self) -> int: if len(self) > 0: return max(cast("int", self.getpriority(name)) for name in self) return get_settings_priority("default") def replace_in_component_priority_dict( self, name: _SettingsKey, old_cls: type, new_cls: type, priority: int | None = None, ) -> None: component_priority_dict = self.getdict(name) old_priority = None for cls_or_path in tuple(component_priority_dict): if load_object(cls_or_path) != old_cls: continue if (old_priority := component_priority_dict.pop(cls_or_path)) is None: break if old_priority is None: raise KeyError( f"{old_cls} not found in the {name} setting ({component_priority_dict!r})." ) component_priority_dict[new_cls] = ( old_priority if priority is None else priority ) self.set(name, component_priority_dict, priority=self.getpriority(name) or 0) def __setitem__(self, name: _SettingsKey, value: Any) -> None: self.set(name, value) def set( self, name: _SettingsKey, value: Any, priority: int | str = "project" ) -> None: self._assert_mutability() priority = get_settings_priority(priority) if name not in self: if isinstance(value, SettingsAttribute): self.attributes[name] = value else: self.attributes[name] = SettingsAttribute(value, priority) else: self.attributes[name].set(value, priority) def set_in_component_priority_dict( self, name: _SettingsKey, cls: type, priority: int | None ) -> None: component_priority_dict = self.getdict(name) for cls_or_path in tuple(component_priority_dict): if not isinstance(cls_or_path, str): continue _cls = load_object(cls_or_path) if _cls == cls: del component_priority_dict[cls_or_path] component_priority_dict[cls] = priority self.set(name, component_priority_dict, self.getpriority(name) or 0) def setdefault( self, name: _SettingsKey, default: Any = None, priority: int | str = "project", ) -> Any: if name not in self: self.set(name, default, priority) return default return self.attributes[name].value def setdefault_in_component_priority_dict( self, name: _SettingsKey, cls: type, priority: int | None ) -> None: component_priority_dict = self.getdict(name) for cls_or_path in tuple(component_priority_dict): if load_object(cls_or_path) == cls: return component_priority_dict[cls] = priority self.set(name, component_priority_dict, self.getpriority(name) or 0) def setdict(self, values: _SettingsInput, priority: int | str = "project") -> None: self.update(values, priority) def setmodule( self, module: ModuleType | str, priority: int | str = "project" ) -> None: self._assert_mutability() if isinstance(module, str): module = import_module(module) for key in dir(module): if key.isupper(): self.set(key, getattr(module, key), priority) # BaseSettings.update() doesn't support all inputs that MutableMapping.update() supports def update(self, values: _SettingsInput, priority: int | str = "project") -> None: # type: ignore[override] self._assert_mutability() if isinstance(values, str): values = cast("dict[_SettingsKey, Any]", json.loads(values)) if values is not None: if isinstance(values, BaseSettings): for name, value in values.items(): self.set(name, value, cast("int", values.getpriority(name))) else: for name, value in values.items(): self.set(name, value, priority) def delete(self, name: _SettingsKey, priority: int | str = "project") -> None: if name not in self: raise KeyError(name) self._assert_mutability() priority = get_settings_priority(priority) if priority >= cast("int", self.getpriority(name)): del self.attributes[name] def __delitem__(self, name: _SettingsKey) -> None: self._assert_mutability() del self.attributes[name] def _assert_mutability(self) -> None: if self.frozen: raise TypeError("Trying to modify an immutable Settings object") def copy(self) -> Self: return copy.deepcopy(self) def freeze(self) -> None: self.frozen = True def frozencopy(self) -> Self: copy = self.copy() copy.freeze() return copy def __iter__(self) -> Iterator[_SettingsKey]: return iter(self.attributes) def __len__(self) -> int: return len(self.attributes) def _to_dict(self) -> dict[_SettingsKey, Any]: return { self._get_key(k): (v._to_dict() if isinstance(v, BaseSettings) else v) for k, v in self.items() } def _get_key(self, key_value: Any) -> _SettingsKey: return ( key_value if isinstance(key_value, (bool, float, int, str, type(None))) else str(key_value) ) def copy_to_dict(self) -> dict[_SettingsKey, Any]: settings = self.copy() return settings._to_dict() # https://ipython.readthedocs.io/en/stable/config/integrating.html#pretty-printing def _repr_pretty_(self, p: Any, cycle: bool) -> None: if cycle: p.text(repr(self)) else: p.text(pformat(self.copy_to_dict())) def pop(self, name: _SettingsKey, default: Any = __default) -> Any: try: value = self.attributes[name].value except KeyError: if default is self.__default: raise return default self.__delitem__(name) return value class Settings(BaseSettings): def __init__(self, values: _SettingsInput = None, priority: int | str = "project"): # Do not pass kwarg values here. We don't want to promote user-defined # dicts, and we want to update, not replace, default dicts with the # values given by the user super().__init__() self.setmodule(default_settings, "default") # Promote default dictionaries to BaseSettings instances for per-key # priorities for name, val in self.items(): if isinstance(val, dict): self.set(name, BaseSettings(val, "default"), "default") self.update(values, priority) def iter_default_settings() -> Iterable[tuple[str, Any]]: for name in dir(default_settings): if name.isupper(): yield name, getattr(default_settings, name) def overridden_settings( settings: Mapping[_SettingsKey, Any], ) -> Iterable[tuple[str, Any]]: for name, defvalue in iter_default_settings(): value = settings[name] if not isinstance(defvalue, dict) and value != defvalue: yield name, value
--- +++ @@ -43,12 +43,22 @@ def get_settings_priority(priority: int | str) -> int: + """ + Small helper function that looks up a given string priority in the + :attr:`~scrapy.settings.SETTINGS_PRIORITIES` dictionary and returns its + numerical value, or directly returns a given numerical priority. + """ if isinstance(priority, str): return SETTINGS_PRIORITIES[priority] return priority class SettingsAttribute: + """Class for storing data related to settings attributes. + + This class is intended for internal usage, you should try Settings class + for settings configuration, not this one. + """ def __init__(self, value: Any, priority: int): self.value: Any = value @@ -59,6 +69,7 @@ self.priority = priority def set(self, value: Any, priority: int) -> None: + """Sets value if priority is higher or equal than current priority.""" if priority >= self.priority: if isinstance(self.value, BaseSettings): value = BaseSettings(value, priority=priority) @@ -70,6 +81,26 @@ class BaseSettings(MutableMapping[_SettingsKey, Any]): + """ + Instances of this class behave like dictionaries, but store priorities + along with their ``(key, value)`` pairs, and can be frozen (i.e. marked + immutable). + + Key-value entries can be passed on initialization with the ``values`` + argument, and they would take the ``priority`` level (unless ``values`` is + already an instance of :class:`~scrapy.settings.BaseSettings`, in which + case the existing priority levels will be kept). If the ``priority`` + argument is a string, the priority name will be looked up in + :attr:`~scrapy.settings.SETTINGS_PRIORITIES`. Otherwise, a specific integer + should be provided. + + Once the object is created, new settings can be loaded or updated with the + :meth:`~scrapy.settings.BaseSettings.set` method, and can be accessed with + the square bracket notation of dictionaries, or with the + :meth:`~scrapy.settings.BaseSettings.get` method of the instance and its + value conversion variants. When requesting a stored key, the value with the + highest priority will be retrieved. + """ __default = object() @@ -88,17 +119,40 @@ return name in self.attributes def add_to_list(self, name: _SettingsKey, item: Any) -> None: + """Append *item* to the :class:`list` setting with the specified *name* + if *item* is not already in that list. + + This change is applied regardless of the priority of the *name* + setting. The setting priority is not affected by this change either. + """ value: list[str] = self.getlist(name) if item not in value: self.set(name, [*value, item], self.getpriority(name) or 0) def remove_from_list(self, name: _SettingsKey, item: Any) -> None: + """Remove *item* from the :class:`list` setting with the specified + *name*. + + If *item* is missing, raise :exc:`ValueError`. + + This change is applied regardless of the priority of the *name* + setting. The setting priority is not affected by this change either. + """ value: list[str] = self.getlist(name) if item not in value: raise ValueError(f"{item!r} not found in the {name} setting ({value!r}).") self.set(name, [v for v in value if v != item], self.getpriority(name) or 0) def get(self, name: _SettingsKey, default: Any = None) -> Any: + """ + Get a setting value without affecting its original type. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ if name == "CONCURRENT_REQUESTS_PER_IP" and ( isinstance(self[name], int) and self[name] != 0 ): @@ -111,6 +165,21 @@ return self[name] if self[name] is not None else default def getbool(self, name: _SettingsKey, default: bool = False) -> bool: + """ + Get a setting value as a boolean. + + ``1``, ``'1'``, `True`` and ``'True'`` return ``True``, + while ``0``, ``'0'``, ``False``, ``'False'`` and ``None`` return ``False``. + + For example, settings populated through environment variables set to + ``'0'`` will return ``False`` when using this method. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ got = self.get(name, default) try: return bool(int(got)) @@ -126,14 +195,46 @@ ) def getint(self, name: _SettingsKey, default: int = 0) -> int: + """ + Get a setting value as an int. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ return int(self.get(name, default)) def getfloat(self, name: _SettingsKey, default: float = 0.0) -> float: + """ + Get a setting value as a float. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ return float(self.get(name, default)) def getlist( self, name: _SettingsKey, default: list[Any] | None = None ) -> list[Any]: + """ + Get a setting value as a list. If the setting original type is a list, + a copy of it will be returned. If it's a string it will be split by + ",". If it is an empty string, an empty list will be returned. + + For example, settings populated through environment variables set to + ``'one,two'`` will return a list ['one', 'two'] when using this method. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ value = self.get(name, default or []) if not value: return [] @@ -144,6 +245,21 @@ def getdict( self, name: _SettingsKey, default: dict[Any, Any] | None = None ) -> dict[Any, Any]: + """ + Get a setting value as a dictionary. If the setting original type is a + dictionary, a copy of it will be returned. If it is a string it will be + evaluated as a JSON dictionary. In the case that it is a + :class:`~scrapy.settings.BaseSettings` instance itself, it will be + converted to a dictionary, containing all its current settings values + as they would be returned by :meth:`~scrapy.settings.BaseSettings.get`, + and losing all information about priority and mutability. + + :param name: the setting name + :type name: str + + :param default: the value to return if no setting is found + :type default: object + """ value = self.get(name, default or {}) if isinstance(value, str): value = json.loads(value) @@ -154,6 +270,27 @@ name: _SettingsKey, default: dict[Any, Any] | list[Any] | tuple[Any] | None = None, ) -> dict[Any, Any] | list[Any]: + """Get a setting value as either a :class:`dict` or a :class:`list`. + + If the setting is already a dict or a list, a copy of it will be + returned. + + If it is a string it will be evaluated as JSON, or as a comma-separated + list of strings as a fallback. + + For example, settings populated from the command line will return: + + - ``{'key1': 'value1', 'key2': 'value2'}`` if set to + ``'{"key1": "value1", "key2": "value2"}'`` + + - ``['one', 'two']`` if set to ``'["one", "two"]'`` or ``'one,two'`` + + :param name: the setting name + :type name: string + + :param default: the value to return if no setting is found + :type default: any + """ value = self.get(name, default) if value is None: return {} @@ -178,6 +315,12 @@ return copy.deepcopy(value) def getwithbase(self, name: _SettingsKey) -> BaseSettings: + """Get a composition of a dictionary-like setting and its `_BASE` + counterpart. + + :param name: name of the dictionary-like setting + :type name: str + """ if not isinstance(name, str): raise ValueError(f"Base setting key must be a string, got {name}") @@ -217,11 +360,24 @@ ) def getpriority(self, name: _SettingsKey) -> int | None: + """ + Return the current numerical priority value of a setting, or ``None`` if + the given ``name`` does not exist. + + :param name: the setting name + :type name: str + """ if name not in self: return None return self.attributes[name].priority def maxpriority(self) -> int: + """ + Return the numerical value of the highest priority present throughout + all settings, or the numerical value for ``default`` from + :attr:`~scrapy.settings.SETTINGS_PRIORITIES` if there are no settings + stored. + """ if len(self) > 0: return max(cast("int", self.getpriority(name)) for name in self) return get_settings_priority("default") @@ -233,6 +389,24 @@ new_cls: type, priority: int | None = None, ) -> None: + """Replace *old_cls* with *new_cls* in the *name* :ref:`component + priority dictionary <component-priority-dictionaries>`. + + If *old_cls* is missing, or has :data:`None` as value, :exc:`KeyError` + is raised. + + If *old_cls* was present as an import string, even more than once, + those keys are dropped and replaced by *new_cls*. + + If *priority* is specified, that is the value assigned to *new_cls* in + the component priority dictionary. Otherwise, the value of *old_cls* is + used. If *old_cls* was present multiple times (possible with import + strings) with different values, the value assigned to *new_cls* is one + of them, with no guarantee about which one it is. + + This change is applied regardless of the priority of the *name* + setting. The setting priority is not affected by this change either. + """ component_priority_dict = self.getdict(name) old_priority = None for cls_or_path in tuple(component_priority_dict): @@ -255,6 +429,23 @@ def set( self, name: _SettingsKey, value: Any, priority: int | str = "project" ) -> None: + """ + Store a key/value attribute with a given priority. + + Settings should be populated *before* configuring the Crawler object + (through the :meth:`~scrapy.crawler.Crawler.configure` method), + otherwise they won't have any effect. + + :param name: the setting name + :type name: str + + :param value: the value to associate with the setting + :type value: object + + :param priority: the priority of the setting. Should be a key of + :attr:`~scrapy.settings.SETTINGS_PRIORITIES` or an integer + :type priority: str or int + """ self._assert_mutability() priority = get_settings_priority(priority) if name not in self: @@ -268,6 +459,17 @@ def set_in_component_priority_dict( self, name: _SettingsKey, cls: type, priority: int | None ) -> None: + """Set the *cls* component in the *name* :ref:`component priority + dictionary <component-priority-dictionaries>` setting with *priority*. + + If *cls* already exists, its value is updated. + + If *cls* was present as an import string, even more than once, those + keys are dropped and replaced by *cls*. + + This change is applied regardless of the priority of the *name* + setting. The setting priority is not affected by this change either. + """ component_priority_dict = self.getdict(name) for cls_or_path in tuple(component_priority_dict): if not isinstance(cls_or_path, str): @@ -293,6 +495,14 @@ def setdefault_in_component_priority_dict( self, name: _SettingsKey, cls: type, priority: int | None ) -> None: + """Set the *cls* component in the *name* :ref:`component priority + dictionary <component-priority-dictionaries>` setting with *priority* + if not already defined (even as an import string). + + If *cls* is not already defined, it is set regardless of the priority + of the *name* setting. The setting priority is not affected by this + change either. + """ component_priority_dict = self.getdict(name) for cls_or_path in tuple(component_priority_dict): if load_object(cls_or_path) == cls: @@ -306,6 +516,20 @@ def setmodule( self, module: ModuleType | str, priority: int | str = "project" ) -> None: + """ + Store settings from a module with a given priority. + + This is a helper function that calls + :meth:`~scrapy.settings.BaseSettings.set` for every globally declared + uppercase variable of ``module`` with the provided ``priority``. + + :param module: the module or the path of the module + :type module: types.ModuleType or str + + :param priority: the priority of the settings. Should be a key of + :attr:`~scrapy.settings.SETTINGS_PRIORITIES` or an integer + :type priority: str or int + """ self._assert_mutability() if isinstance(module, str): module = import_module(module) @@ -315,6 +539,27 @@ # BaseSettings.update() doesn't support all inputs that MutableMapping.update() supports def update(self, values: _SettingsInput, priority: int | str = "project") -> None: # type: ignore[override] + """ + Store key/value pairs with a given priority. + + This is a helper function that calls + :meth:`~scrapy.settings.BaseSettings.set` for every item of ``values`` + with the provided ``priority``. + + If ``values`` is a string, it is assumed to be JSON-encoded and parsed + into a dict with ``json.loads()`` first. If it is a + :class:`~scrapy.settings.BaseSettings` instance, the per-key priorities + will be used and the ``priority`` parameter ignored. This allows + inserting/updating settings with different priorities with a single + command. + + :param values: the settings names and values + :type values: dict or string or :class:`~scrapy.settings.BaseSettings` + + :param priority: the priority of the settings. Should be a key of + :attr:`~scrapy.settings.SETTINGS_PRIORITIES` or an integer + :type priority: str or int + """ self._assert_mutability() if isinstance(values, str): values = cast("dict[_SettingsKey, Any]", json.loads(values)) @@ -343,12 +588,33 @@ raise TypeError("Trying to modify an immutable Settings object") def copy(self) -> Self: + """ + Make a deep copy of current settings. + + This method returns a new instance of the :class:`Settings` class, + populated with the same values and their priorities. + + Modifications to the new object won't be reflected on the original + settings. + """ return copy.deepcopy(self) def freeze(self) -> None: + """ + Disable further changes to the current settings. + + After calling this method, the present state of the settings will become + immutable. Trying to change values through the :meth:`~set` method and + its variants won't be possible and will be alerted. + """ self.frozen = True def frozencopy(self) -> Self: + """ + Return an immutable copy of the current settings. + + Alias for a :meth:`~freeze` call in the object returned by :meth:`copy`. + """ copy = self.copy() copy.freeze() return copy @@ -373,6 +639,18 @@ ) def copy_to_dict(self) -> dict[_SettingsKey, Any]: + """ + Make a copy of current settings and convert to a dict. + + This method returns a new dict populated with the same values + and their priorities as the current settings. + + Modifications to the returned dict won't be reflected on the original + settings. + + This method can be useful for example for printing settings + in Scrapy shell. + """ settings = self.copy() return settings._to_dict() @@ -395,6 +673,15 @@ class Settings(BaseSettings): + """ + This object stores Scrapy settings for the configuration of internal + components, and can be used for any further customization. + + It is a direct subclass and supports all methods of + :class:`~scrapy.settings.BaseSettings`. Additionally, after instantiation + of this class, the new object will have the global default settings + described on :ref:`topics-settings-ref` already populated. + """ def __init__(self, values: _SettingsInput = None, priority: int | str = "project"): # Do not pass kwarg values here. We don't want to promote user-defined @@ -411,6 +698,7 @@ def iter_default_settings() -> Iterable[tuple[str, Any]]: + """Return the default settings as an iterator of (name, value) tuples""" for name in dir(default_settings): if name.isupper(): yield name, getattr(default_settings, name) @@ -419,7 +707,8 @@ def overridden_settings( settings: Mapping[_SettingsKey, Any], ) -> Iterable[tuple[str, Any]]: + """Return an iterable of the settings that have been overridden""" for name, defvalue in iter_default_settings(): value = settings[name] if not isinstance(defvalue, dict) and value != defvalue: - yield name, value+ yield name, value
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/settings/__init__.py
Add docstrings to existing functions
from __future__ import annotations import hashlib import logging from typing import TYPE_CHECKING, Protocol, cast from scrapy.utils.misc import build_from_crawler if TYPE_CHECKING: from collections.abc import Iterable # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Request from scrapy.core.downloader import Downloader from scrapy.crawler import Crawler logger = logging.getLogger(__name__) def _path_safe(text: str) -> str: pathable_slot = "".join([c if c.isalnum() or c in "-._" else "_" for c in text]) # as we replace some letters we can get collision for different slots # add we add unique part unique_slot = hashlib.md5(text.encode("utf8")).hexdigest() # noqa: S324 return f"{pathable_slot}-{unique_slot}" class QueueProtocol(Protocol): def push(self, request: Request) -> None: ... def pop(self) -> Request | None: ... def close(self) -> None: ... def __len__(self) -> int: ... class ScrapyPriorityQueue: @classmethod def from_crawler( cls, crawler: Crawler, downstream_queue_cls: type[QueueProtocol], key: str, startprios: Iterable[int] = (), *, start_queue_cls: type[QueueProtocol] | None = None, ) -> Self: return cls( crawler, downstream_queue_cls, key, startprios, start_queue_cls=start_queue_cls, ) def __init__( self, crawler: Crawler, downstream_queue_cls: type[QueueProtocol], key: str, startprios: Iterable[int] = (), *, start_queue_cls: type[QueueProtocol] | None = None, ): self.crawler: Crawler = crawler self.downstream_queue_cls: type[QueueProtocol] = downstream_queue_cls self._start_queue_cls: type[QueueProtocol] | None = start_queue_cls self.key: str = key self.queues: dict[int, QueueProtocol] = {} self._start_queues: dict[int, QueueProtocol] = {} self.curprio: int | None = None self.init_prios(startprios) def init_prios(self, startprios: Iterable[int]) -> None: if not startprios: return for priority in startprios: q = self.qfactory(priority) if q: self.queues[priority] = q if self._start_queue_cls: q = self._sqfactory(priority) if q: self._start_queues[priority] = q self.curprio = min(startprios) def qfactory(self, key: int) -> QueueProtocol: return build_from_crawler( self.downstream_queue_cls, self.crawler, self.key + "/" + str(key), ) def _sqfactory(self, key: int) -> QueueProtocol: assert self._start_queue_cls is not None return build_from_crawler( self._start_queue_cls, self.crawler, f"{self.key}/{key}s", ) def priority(self, request: Request) -> int: return -request.priority def push(self, request: Request) -> None: priority = self.priority(request) is_start_request = request.meta.get("is_start_request", False) if is_start_request and self._start_queue_cls: if priority not in self._start_queues: self._start_queues[priority] = self._sqfactory(priority) q = self._start_queues[priority] else: if priority not in self.queues: self.queues[priority] = self.qfactory(priority) q = self.queues[priority] q.push(request) # this may fail (eg. serialization error) if self.curprio is None or priority < self.curprio: self.curprio = priority def pop(self) -> Request | None: while self.curprio is not None: try: q = self.queues[self.curprio] except KeyError: pass else: m = q.pop() if not q: del self.queues[self.curprio] q.close() if not self._start_queues: self._update_curprio() return m if self._start_queues: try: q = self._start_queues[self.curprio] except KeyError: self._update_curprio() else: m = q.pop() if not q: del self._start_queues[self.curprio] q.close() self._update_curprio() return m else: self._update_curprio() return None def _update_curprio(self) -> None: prios = { p for queues in (self.queues, self._start_queues) for p, q in queues.items() if q } self.curprio = min(prios) if prios else None def peek(self) -> Request | None: if self.curprio is None: return None try: queue = self._start_queues[self.curprio] except KeyError: queue = self.queues[self.curprio] # Protocols can't declare optional members return cast("Request", queue.peek()) # type: ignore[attr-defined] def close(self) -> list[int]: active: set[int] = set() for queues in (self.queues, self._start_queues): for p, q in queues.items(): active.add(p) q.close() return list(active) def __len__(self) -> int: return ( sum( len(x) for queues in (self.queues, self._start_queues) for x in queues.values() ) if self.queues or self._start_queues else 0 ) class DownloaderInterface: def __init__(self, crawler: Crawler): assert crawler.engine self.downloader: Downloader = crawler.engine.downloader def stats(self, possible_slots: Iterable[str]) -> list[tuple[int, str]]: return [(self._active_downloads(slot), slot) for slot in possible_slots] def get_slot_key(self, request: Request) -> str: return self.downloader.get_slot_key(request) def _active_downloads(self, slot: str) -> int: if slot not in self.downloader.slots: return 0 return len(self.downloader.slots[slot].active) class DownloaderAwarePriorityQueue: @classmethod def from_crawler( cls, crawler: Crawler, downstream_queue_cls: type[QueueProtocol], key: str, startprios: dict[str, Iterable[int]] | None = None, *, start_queue_cls: type[QueueProtocol] | None = None, ) -> Self: return cls( crawler, downstream_queue_cls, key, startprios, start_queue_cls=start_queue_cls, ) def __init__( self, crawler: Crawler, downstream_queue_cls: type[QueueProtocol], key: str, slot_startprios: dict[str, Iterable[int]] | None = None, *, start_queue_cls: type[QueueProtocol] | None = None, ): if crawler.settings.getint("CONCURRENT_REQUESTS_PER_IP") != 0: raise ValueError( f'"{self.__class__}" does not support CONCURRENT_REQUESTS_PER_IP' ) if slot_startprios and not isinstance(slot_startprios, dict): raise ValueError( "DownloaderAwarePriorityQueue accepts " "``slot_startprios`` as a dict; " f"{slot_startprios.__class__!r} instance " "is passed. Most likely, it means the state is " "created by an incompatible priority queue. " "Only a crawl started with the same priority " "queue class can be resumed." ) self._downloader_interface: DownloaderInterface = DownloaderInterface(crawler) self.downstream_queue_cls: type[QueueProtocol] = downstream_queue_cls self._start_queue_cls: type[QueueProtocol] | None = start_queue_cls self.key: str = key self.crawler: Crawler = crawler self.pqueues: dict[str, ScrapyPriorityQueue] = {} # slot -> priority queue for slot, startprios in (slot_startprios or {}).items(): self.pqueues[slot] = self.pqfactory(slot, startprios) def pqfactory( self, slot: str, startprios: Iterable[int] = () ) -> ScrapyPriorityQueue: return ScrapyPriorityQueue( self.crawler, self.downstream_queue_cls, self.key + "/" + _path_safe(slot), startprios, start_queue_cls=self._start_queue_cls, ) def pop(self) -> Request | None: stats = self._downloader_interface.stats(self.pqueues) if not stats: return None slot = min(stats)[1] queue = self.pqueues[slot] request = queue.pop() if len(queue) == 0: del self.pqueues[slot] return request def push(self, request: Request) -> None: slot = self._downloader_interface.get_slot_key(request) if slot not in self.pqueues: self.pqueues[slot] = self.pqfactory(slot) queue = self.pqueues[slot] queue.push(request) def peek(self) -> Request | None: stats = self._downloader_interface.stats(self.pqueues) if not stats: return None slot = min(stats)[1] queue = self.pqueues[slot] return queue.peek() def close(self) -> dict[str, list[int]]: active = {slot: queue.close() for slot, queue in self.pqueues.items()} self.pqueues.clear() return active def __len__(self) -> int: return sum(len(x) for x in self.pqueues.values()) if self.pqueues else 0 def __contains__(self, slot: str) -> bool: return slot in self.pqueues
--- +++ @@ -20,6 +20,16 @@ def _path_safe(text: str) -> str: + """ + Return a filesystem-safe version of a string ``text`` + + >>> _path_safe('simple.org').startswith('simple.org') + True + >>> _path_safe('dash-underscore_.org').startswith('dash-underscore_.org') + True + >>> _path_safe('some@symbol?').startswith('some_symbol_') + True + """ pathable_slot = "".join([c if c.isalnum() or c in "-._" else "_" for c in text]) # as we replace some letters we can get collision for different slots # add we add unique part @@ -28,6 +38,7 @@ class QueueProtocol(Protocol): + """Protocol for downstream queues of ``ScrapyPriorityQueue``.""" def push(self, request: Request) -> None: ... @@ -39,6 +50,52 @@ class ScrapyPriorityQueue: + """A priority queue implemented using multiple internal queues (typically, + FIFO queues). It uses one internal queue for each priority value. The + internal queue must implement the following methods: + + * push(obj) + * pop() + * close() + * __len__() + + Optionally, the queue could provide a ``peek`` method, that should return + the next object to be returned by ``pop``, but without removing it from the + queue. + + ``__init__`` method of ScrapyPriorityQueue receives a downstream_queue_cls + argument, which is a class used to instantiate a new (internal) queue when + a new priority is allocated. + + Only integer priorities should be used. Lower numbers are higher + priorities. + + startprios is a sequence of priorities to start with. If the queue was + previously closed leaving some priority buckets non-empty, those priorities + should be passed in startprios. + + Disk persistence + ================ + + .. warning:: The files that this class generates on disk are an + implementation detail, and may change without a warning in a future + version of Scrapy. Do not rely on the following information for + anything other than debugging purposes. + + When a component instantiates this class with a non-empty *key* argument, + *key* is used as a persistence directory. + + For every request enqueued, this class checks: + + - Whether the request is a :ref:`start request <start-requests>` or not. + + - The :data:`~scrapy.Request.priority` of the request. + + For each combination of the above seen, this class creates an instance of + *downstream_queue_cls* with *key* set to a subdirectory of the persistence + directory, named as the request priority (e.g. ``1``), with an ``s`` suffix + in case of a start request (e.g. ``1s``). + """ @classmethod def from_crawler( @@ -164,6 +221,12 @@ self.curprio = min(prios) if prios else None def peek(self) -> Request | None: + """Returns the next object to be returned by :meth:`pop`, + but without removing it from the queue. + + Raises :exc:`NotImplementedError` if the underlying queue class does + not implement a ``peek`` method, which is optional for queues. + """ if self.curprio is None: return None try: @@ -205,12 +268,37 @@ return self.downloader.get_slot_key(request) def _active_downloads(self, slot: str) -> int: + """Return a number of requests in a Downloader for a given slot""" if slot not in self.downloader.slots: return 0 return len(self.downloader.slots[slot].active) class DownloaderAwarePriorityQueue: + """PriorityQueue which takes Downloader activity into account: + domains (slots) with the least amount of active downloads are dequeued + first. + + Disk persistence + ================ + + .. warning:: The files that this class generates on disk are an + implementation detail, and may change without a warning in a future + version of Scrapy. Do not rely on the following information for + anything other than debugging purposes. + + When a component instantiates this class with a non-empty *key* argument, + *key* is used as a persistence directory, and inside that directory this + class creates a subdirectory per download slot (domain). + + Those subdirectories are named after the corresponding download slot, with + path-unsafe characters replaced by underscores and an MD5 hash suffix to + avoid collisions. + + For each download slot, this class creates an instance of + :class:`ScrapyPriorityQueue` with the download slot subdirectory as *key* + and its own *downstream_queue_cls*. + """ @classmethod def from_crawler( @@ -297,6 +385,12 @@ queue.push(request) def peek(self) -> Request | None: + """Returns the next object to be returned by :meth:`pop`, + but without removing it from the queue. + + Raises :exc:`NotImplementedError` if the underlying queue class does + not implement a ``peek`` method, which is optional for queues. + """ stats = self._downloader_interface.stats(self.pqueues) if not stats: return None @@ -313,4 +407,4 @@ return sum(len(x) for x in self.pqueues.values()) if self.pqueues else 0 def __contains__(self, slot: str) -> bool: - return slot in self.pqueues+ return slot in self.pqueues
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/pqueues.py
Create documentation for each function signature
from __future__ import annotations import asyncio import functools import logging import warnings from abc import ABC, abstractmethod from collections import defaultdict from typing import TYPE_CHECKING, Any, Literal, TypeAlias, TypedDict, cast from twisted import version as twisted_version from twisted.internet.defer import Deferred, DeferredList from twisted.python.failure import Failure from twisted.python.versions import Version from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.http.request import NO_CALLBACK, Request from scrapy.utils.asyncio import call_later, is_asyncio_available from scrapy.utils.datatypes import SequenceExclude from scrapy.utils.decorators import _warn_spider_arg from scrapy.utils.defer import ( _DEFER_DELAY, _defer_sleep_async, deferred_from_coro, ensure_awaitable, maybe_deferred_to_future, ) from scrapy.utils.log import failure_to_exc_info from scrapy.utils.misc import arg_to_iter from scrapy.utils.python import global_object_name if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Spider from scrapy.crawler import Crawler from scrapy.http import Response from scrapy.settings import Settings from scrapy.utils.request import RequestFingerprinterProtocol class FileInfo(TypedDict): url: str path: str checksum: str | None status: str FileInfoOrError: TypeAlias = ( tuple[Literal[True], FileInfo] | tuple[Literal[False], Failure] ) logger = logging.getLogger(__name__) class MediaPipeline(ABC): LOG_FAILED_RESULTS: bool = True class SpiderInfo: def __init__(self, spider: Spider): self.spider: Spider = spider self.downloading: set[bytes] = set() self.downloaded: dict[bytes, FileInfo | Failure] = {} self.waiting: defaultdict[bytes, list[Deferred[FileInfo]]] = defaultdict( list ) def __init__( self, download_func: None = None, *, crawler: Crawler, ): if download_func is not None: # pragma: no cover warnings.warn( "The download_func argument of MediaPipeline.__init__() is ignored" " and will be removed in a future Scrapy version.", category=ScrapyDeprecationWarning, stacklevel=2, ) self.crawler: Crawler = crawler assert crawler.request_fingerprinter self._fingerprinter: RequestFingerprinterProtocol = ( crawler.request_fingerprinter ) settings = crawler.settings resolve = functools.partial( self._key_for_pipe, base_class_name="MediaPipeline", settings=settings ) self.allow_redirects: bool = settings.getbool( resolve("MEDIA_ALLOW_REDIRECTS"), False ) self._handle_statuses(self.allow_redirects) def _handle_statuses(self, allow_redirects: bool) -> None: self.handle_httpstatus_list = None if allow_redirects: self.handle_httpstatus_list = SequenceExclude(range(300, 400)) def _key_for_pipe( self, key: str, base_class_name: str | None = None, settings: Settings | None = None, ) -> str: class_name = self.__class__.__name__ formatted_key = f"{class_name.upper()}_{key}" if ( not base_class_name or class_name == base_class_name or (settings and not settings.get(formatted_key)) ): return key return formatted_key @classmethod def from_crawler(cls, crawler: Crawler) -> Self: return cls(crawler=crawler) @_warn_spider_arg def open_spider(self, spider: Spider | None = None) -> None: assert self.crawler.spider self.spiderinfo = self.SpiderInfo(self.crawler.spider) @_warn_spider_arg async def process_item(self, item: Any, spider: Spider | None = None) -> Any: info = self.spiderinfo requests = arg_to_iter(self.get_media_requests(item, info)) coros = [self._process_request(r, info, item) for r in requests] results: list[FileInfoOrError] = [] if coros: if is_asyncio_available(): results_asyncio = await asyncio.gather(*coros, return_exceptions=True) for res in results_asyncio: if isinstance(res, BaseException): results.append((False, Failure(res))) else: results.append((True, res)) else: results = await cast( "Deferred[list[FileInfoOrError]]", DeferredList( (deferred_from_coro(coro) for coro in coros), consumeErrors=True ), ) return self.item_completed(results, item, info) async def _process_request( self, request: Request, info: SpiderInfo, item: Any ) -> FileInfo: fp = self._fingerprinter.fingerprint(request) eb = request.errback request.callback = NO_CALLBACK request.errback = None # Return cached result if request was already seen if fp in info.downloaded: await _defer_sleep_async() cached_result = info.downloaded[fp] if isinstance(cached_result, Failure): if eb: return eb(cached_result) cached_result.raiseException() return cached_result # Otherwise, wait for result wad: Deferred[FileInfo] = Deferred() if eb: wad.addErrback(eb) info.waiting[fp].append(wad) # Check if request is downloading right now to avoid doing it twice if fp in info.downloading: return await maybe_deferred_to_future(wad) # Download request checking media_to_download hook output first info.downloading.add(fp) await _defer_sleep_async() result: FileInfo | Failure try: file_info: FileInfo | None = await ensure_awaitable( self.media_to_download(request, info, item=item) ) if file_info: # got a result without downloading result = file_info else: # download the result result = await self._check_media_to_download(request, info, item=item) except Exception: result = Failure() logger.exception(result) self._cache_result_and_execute_waiters(result, fp, info) return await maybe_deferred_to_future(wad) # it must return wad at last def _modify_media_request(self, request: Request) -> None: if self.handle_httpstatus_list: request.meta["handle_httpstatus_list"] = self.handle_httpstatus_list else: request.meta["handle_httpstatus_all"] = True async def _check_media_to_download( self, request: Request, info: SpiderInfo, item: Any ) -> FileInfo: try: self._modify_media_request(request) assert self.crawler.engine response = await self.crawler.engine.download_async(request) return self.media_downloaded(response, request, info, item=item) except Exception: failure = self.media_failed(Failure(), request, info) if isinstance(failure, Failure): warnings.warn( f"{global_object_name(self.media_failed)} returned a Failure instance." f" This is deprecated, please raise an exception instead, e.g. via failure.raiseException().", category=ScrapyDeprecationWarning, stacklevel=2, ) failure.raiseException() def _cache_result_and_execute_waiters( self, result: FileInfo | Failure, fp: bytes, info: SpiderInfo ) -> None: if isinstance(result, Failure): # minimize cached information for failure result.cleanFailure() result.frames = [] if twisted_version < Version("twisted", 24, 10, 0): result.stack = [] # type: ignore[method-assign] # This code fixes a memory leak by avoiding to keep references to # the Request and Response objects on the Media Pipeline cache. # # What happens when the media_downloaded callback raises an # exception, for example a FileException('download-error') when # the Response status code is not 200 OK, is that the original # StopIteration exception (which in turn contains the failed # Response and by extension, the original Request) gets encapsulated # within the FileException context. # # Originally, Scrapy was using twisted.internet.defer.returnValue # inside functions decorated with twisted.internet.defer.inlineCallbacks, # encapsulating the returned Response in a _DefGen_Return exception # instead of a StopIteration. # # To avoid keeping references to the Response and therefore Request # objects on the Media Pipeline cache, we should wipe the context of # the encapsulated exception when it is a StopIteration instance context = getattr(result.value, "__context__", None) if isinstance(context, StopIteration): assert result.value is not None result.value.__context__ = None info.downloading.remove(fp) info.downloaded[fp] = result # cache result for wad in info.waiting.pop(fp): if isinstance(result, Failure): call_later(_DEFER_DELAY, wad.errback, result) else: call_later(_DEFER_DELAY, wad.callback, result) # Overridable Interface @abstractmethod def media_to_download( self, request: Request, info: SpiderInfo, *, item: Any = None ) -> Deferred[FileInfo | None] | None: raise NotImplementedError @abstractmethod def get_media_requests(self, item: Any, info: SpiderInfo) -> list[Request]: raise NotImplementedError @abstractmethod def media_downloaded( self, response: Response, request: Request, info: SpiderInfo, *, item: Any = None, ) -> FileInfo: raise NotImplementedError @abstractmethod def media_failed( self, failure: Failure, request: Request, info: SpiderInfo ) -> Failure: raise NotImplementedError def item_completed( self, results: list[FileInfoOrError], item: Any, info: SpiderInfo ) -> Any: if self.LOG_FAILED_RESULTS: for ok, value in results: if not ok: assert isinstance(value, Failure) logger.error( "%(class)s found errors processing %(item)s", {"class": self.__class__.__name__, "item": item}, exc_info=failure_to_exc_info(value), extra={"spider": info.spider}, ) return item @abstractmethod def file_path( self, request: Request, response: Response | None = None, info: SpiderInfo | None = None, *, item: Any = None, ) -> str: raise NotImplementedError
--- +++ @@ -266,10 +266,12 @@ def media_to_download( self, request: Request, info: SpiderInfo, *, item: Any = None ) -> Deferred[FileInfo | None] | None: + """Check request before starting download""" raise NotImplementedError @abstractmethod def get_media_requests(self, item: Any, info: SpiderInfo) -> list[Request]: + """Returns the media requests to download""" raise NotImplementedError @abstractmethod @@ -281,17 +283,20 @@ *, item: Any = None, ) -> FileInfo: + """Handler for success downloads""" raise NotImplementedError @abstractmethod def media_failed( self, failure: Failure, request: Request, info: SpiderInfo ) -> Failure: + """Handler for failed downloads""" raise NotImplementedError def item_completed( self, results: list[FileInfoOrError], item: Any, info: SpiderInfo ) -> Any: + """Called per item when all media requests has been processed""" if self.LOG_FAILED_RESULTS: for ok, value in results: if not ok: @@ -313,4 +318,5 @@ *, item: Any = None, ) -> str: - raise NotImplementedError+ """Returns the path where downloaded media should be stored""" + raise NotImplementedError
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/pipelines/media.py
Help me write clear docstrings
from __future__ import annotations import inspect import warnings from functools import wraps from typing import TYPE_CHECKING, Any, ParamSpec, TypeVar, overload from twisted.internet.defer import Deferred, maybeDeferred from twisted.internet.threads import deferToThread from scrapy.exceptions import ScrapyDeprecationWarning if TYPE_CHECKING: from collections.abc import AsyncGenerator, Callable, Coroutine _T = TypeVar("_T") _P = ParamSpec("_P") def deprecated( use_instead: Any = None, ) -> Callable[[Callable[_P, _T]], Callable[_P, _T]]: def deco(func: Callable[_P, _T]) -> Callable[_P, _T]: @wraps(func) def wrapped(*args: _P.args, **kwargs: _P.kwargs) -> _T: message = f"Call to deprecated function {func.__name__}." if use_instead: message += f" Use {use_instead} instead." warnings.warn(message, category=ScrapyDeprecationWarning, stacklevel=2) return func(*args, **kwargs) return wrapped if callable(use_instead): deco = deco(use_instead) use_instead = None return deco def defers(func: Callable[_P, _T]) -> Callable[_P, Deferred[_T]]: # pragma: no cover warnings.warn( "@defers is deprecated, you can use maybeDeferred() directly if needed.", category=ScrapyDeprecationWarning, stacklevel=2, ) @wraps(func) def wrapped(*a: _P.args, **kw: _P.kwargs) -> Deferred[_T]: return maybeDeferred(func, *a, **kw) return wrapped def inthread(func: Callable[_P, _T]) -> Callable[_P, Deferred[_T]]: @wraps(func) def wrapped(*a: _P.args, **kw: _P.kwargs) -> Deferred[_T]: return deferToThread(func, *a, **kw) return wrapped @overload def _warn_spider_arg( func: Callable[_P, Coroutine[Any, Any, _T]], ) -> Callable[_P, Coroutine[Any, Any, _T]]: ... @overload def _warn_spider_arg( func: Callable[_P, AsyncGenerator[_T]], ) -> Callable[_P, AsyncGenerator[_T]]: ... @overload def _warn_spider_arg(func: Callable[_P, _T]) -> Callable[_P, _T]: ... def _warn_spider_arg( func: Callable[_P, _T], ) -> ( Callable[_P, _T] | Callable[_P, Coroutine[Any, Any, _T]] | Callable[_P, AsyncGenerator[_T]] ): sig = inspect.signature(func) def check_args(*args: _P.args, **kwargs: _P.kwargs) -> None: bound = sig.bind(*args, **kwargs) if "spider" in bound.arguments: warnings.warn( f"Passing a 'spider' argument to {func.__qualname__}() is deprecated and " "the argument will be removed in a future Scrapy version.", category=ScrapyDeprecationWarning, stacklevel=3, ) if inspect.iscoroutinefunction(func): @wraps(func) async def async_inner(*args: _P.args, **kwargs: _P.kwargs) -> _T: check_args(*args, **kwargs) return await func(*args, **kwargs) return async_inner if inspect.isasyncgenfunction(func): @wraps(func) async def asyncgen_inner( *args: _P.args, **kwargs: _P.kwargs ) -> AsyncGenerator[_T]: check_args(*args, **kwargs) async for item in func(*args, **kwargs): yield item return asyncgen_inner @wraps(func) def sync_inner(*args: _P.args, **kwargs: _P.kwargs) -> _T: check_args(*args, **kwargs) return func(*args, **kwargs) return sync_inner
--- +++ @@ -21,6 +21,9 @@ def deprecated( use_instead: Any = None, ) -> Callable[[Callable[_P, _T]], Callable[_P, _T]]: + """This is a decorator which can be used to mark functions + as deprecated. It will result in a warning being emitted + when the function is used.""" def deco(func: Callable[_P, _T]) -> Callable[_P, _T]: @wraps(func) @@ -40,6 +43,7 @@ def defers(func: Callable[_P, _T]) -> Callable[_P, Deferred[_T]]: # pragma: no cover + """Decorator to make sure a function always returns a deferred""" warnings.warn( "@defers is deprecated, you can use maybeDeferred() directly if needed.", category=ScrapyDeprecationWarning, @@ -54,6 +58,9 @@ def inthread(func: Callable[_P, _T]) -> Callable[_P, Deferred[_T]]: + """Decorator to call a function in a thread and return a deferred with the + result + """ @wraps(func) def wrapped(*a: _P.args, **kw: _P.kwargs) -> Deferred[_T]: @@ -85,6 +92,7 @@ | Callable[_P, Coroutine[Any, Any, _T]] | Callable[_P, AsyncGenerator[_T]] ): + """Decorator to warn if a ``spider`` argument is passed to a function.""" sig = inspect.signature(func) @@ -124,4 +132,4 @@ check_args(*args, **kwargs) return func(*args, **kwargs) - return sync_inner+ return sync_inner
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/decorators.py
Add docstrings to existing functions
from __future__ import annotations import asyncio import logging import time from collections.abc import AsyncIterator, Callable, Coroutine, Iterable from typing import TYPE_CHECKING, Any, Concatenate, ParamSpec, TypeVar from twisted.internet.defer import Deferred from twisted.internet.task import LoopingCall from scrapy.utils.asyncgen import as_async_generator from scrapy.utils.reactor import is_asyncio_reactor_installed, is_reactor_installed if TYPE_CHECKING: from twisted.internet.base import DelayedCall # typing.Self, typing.TypeVarTuple and typing.Unpack require Python 3.11 from typing_extensions import Self, TypeVarTuple, Unpack _Ts = TypeVarTuple("_Ts") _T = TypeVar("_T") _P = ParamSpec("_P") logger = logging.getLogger(__name__) def is_asyncio_available() -> bool: # Check if there is a running asyncio loop. # Can't easily check for an installed but not running one, and if we # checked that there could be false positives due to some 3rd-party code # installing it as a side effect (e.g. by calling get_event_loop()). try: asyncio.get_running_loop() except RuntimeError: pass else: return True # Check if there is an installed asyncio reactor (it doesn't need to be # running). if not is_reactor_installed(): raise RuntimeError( "is_asyncio_available() called without an installed reactor" " or running asyncio loop." ) return is_asyncio_reactor_installed() async def _parallel_asyncio( iterable: Iterable[_T] | AsyncIterator[_T], count: int, callable_: Callable[Concatenate[_T, _P], Coroutine[Any, Any, None]], *args: _P.args, **kwargs: _P.kwargs, ) -> None: queue: asyncio.Queue[_T | None] = asyncio.Queue(count * 2) async def worker() -> None: while True: item = await queue.get() if item is None: break try: await callable_(item, *args, **kwargs) finally: queue.task_done() async def fill_queue() -> None: async for item in as_async_generator(iterable): await queue.put(item) for _ in range(count): await queue.put(None) fill_task = asyncio.create_task(fill_queue()) work_tasks = [asyncio.create_task(worker()) for _ in range(count)] await asyncio.wait([fill_task, *work_tasks]) class AsyncioLoopingCall: def __init__(self, func: Callable[_P, _T], *args: _P.args, **kwargs: _P.kwargs): self._func: Callable[_P, _T] = func self._args: tuple[Any, ...] = args self._kwargs: dict[str, Any] = kwargs self._task: asyncio.Task | None = None self.interval: float | None = None self._start_time: float | None = None @property def running(self) -> bool: return self._start_time is not None def start(self, interval: float, now: bool = True) -> None: if self.running: raise RuntimeError("AsyncioLoopingCall already running") if interval <= 0: raise ValueError("Interval must be greater than 0") self.interval = interval self._start_time = time.time() if now: self._call() loop = asyncio.get_event_loop() self._task = loop.create_task(self._loop()) def _to_sleep(self) -> float: assert self.interval is not None assert self._start_time is not None now = time.time() running_for = now - self._start_time return self.interval - (running_for % self.interval) async def _loop(self) -> None: while self.running: await asyncio.sleep(self._to_sleep()) self._call() def stop(self) -> None: self.interval = self._start_time = None if self._task is not None: self._task.cancel() self._task = None def _call(self) -> None: try: result = self._func(*self._args, **self._kwargs) except Exception: logger.exception("Error calling the AsyncioLoopingCall function") self.stop() else: if isinstance(result, (Coroutine, Deferred)): self.stop() raise TypeError( "The AsyncioLoopingCall function must not return a coroutine or a Deferred" ) def create_looping_call( func: Callable[_P, _T], *args: _P.args, **kwargs: _P.kwargs ) -> AsyncioLoopingCall | LoopingCall: if is_asyncio_available(): return AsyncioLoopingCall(func, *args, **kwargs) return LoopingCall(func, *args, **kwargs) def call_later( delay: float, func: Callable[[Unpack[_Ts]], object], *args: Unpack[_Ts] ) -> CallLaterResult: if is_asyncio_available(): loop = asyncio.get_event_loop() return CallLaterResult.from_asyncio(loop.call_later(delay, func, *args)) from twisted.internet import reactor return CallLaterResult.from_twisted(reactor.callLater(delay, func, *args)) class CallLaterResult: _timer_handle: asyncio.TimerHandle | None = None _delayed_call: DelayedCall | None = None @classmethod def from_asyncio(cls, timer_handle: asyncio.TimerHandle) -> Self: o = cls() o._timer_handle = timer_handle return o @classmethod def from_twisted(cls, delayed_call: DelayedCall) -> Self: o = cls() o._delayed_call = delayed_call return o def cancel(self) -> None: if self._timer_handle: self._timer_handle.cancel() self._timer_handle = None elif self._delayed_call and self._delayed_call.active(): self._delayed_call.cancel() self._delayed_call = None
--- +++ @@ -1,3 +1,4 @@+"""Utilities related to asyncio and its support in Scrapy.""" from __future__ import annotations @@ -30,6 +31,43 @@ def is_asyncio_available() -> bool: + """Check if it's possible to call asyncio code that relies on the asyncio event loop. + + .. versionadded:: 2.14 + + This function returns ``True`` if there is a running asyncio event loop. If + there is no such loop, it returns ``True`` if the Twisted reactor that is + installed is + :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor`, returns + ``False`` if a different reactor is installed, and raises a + :exc:`RuntimeError` if no reactor is installed. + + Code that doesn't directly require a Twisted reactor should use this + function while code that requires + :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor` should use + :func:`~scrapy.utils.reactor.is_asyncio_reactor_installed`. + + When this returns ``True``, an asyncio loop is installed and used by + Scrapy. It's possible to call functions that require it, such as + :func:`asyncio.sleep`, and await on :class:`asyncio.Future` objects in + Scrapy-related code. + + When this returns ``False``, a non-asyncio Twisted reactor is installed. + It's not possible to use asyncio features that require an asyncio event + loop or await on :class:`asyncio.Future` objects in Scrapy-related code, + but it's possible to await on :class:`~twisted.internet.defer.Deferred` + objects. + + .. note:: As this function uses :func:`asyncio.get_running_loop()`, it will + only detect the event loop if called in the same thread and from the + code that runs inside that loop (this shouldn't be a problem when + calling it from code such as spiders and Scrapy components, if Scrapy + is run using one of the supported ways). + + .. versionchanged:: VERSION + This function now also returns ``True`` if there is a running asyncio + loop, even if no Twisted reactor is installed. + """ # Check if there is a running asyncio loop. # Can't easily check for an installed but not running one, and if we @@ -60,6 +98,14 @@ *args: _P.args, **kwargs: _P.kwargs, ) -> None: + """Execute a callable over the objects in the given iterable, in parallel, + using no more than ``count`` concurrent calls. + + This function is only used in + :meth:`scrapy.core.scraper.Scraper.handle_spider_output_async` and so it + assumes that neither *callable* nor iterating *iterable* will raise an + exception. + """ queue: asyncio.Queue[_T | None] = asyncio.Queue(count * 2) async def worker() -> None: @@ -84,6 +130,15 @@ class AsyncioLoopingCall: + """A simple implementation of a periodic call using asyncio, keeping + some API and behavior compatibility with the Twisted ``LoopingCall``. + + The function is called every *interval* seconds, independent of the finish + time of the previous call. If the function is still running when it's time + to call it again, calls are skipped until the function finishes. + + The function must not return a coroutine or a ``Deferred``. + """ def __init__(self, func: Callable[_P, _T], *args: _P.args, **kwargs: _P.kwargs): self._func: Callable[_P, _T] = func @@ -98,6 +153,14 @@ return self._start_time is not None def start(self, interval: float, now: bool = True) -> None: + """Start calling the function every *interval* seconds. + + :param interval: The interval in seconds between calls. + :type interval: float + + :param now: If ``True``, also call the function immediately. + :type now: bool + """ if self.running: raise RuntimeError("AsyncioLoopingCall already running") @@ -112,6 +175,7 @@ self._task = loop.create_task(self._loop()) def _to_sleep(self) -> float: + """Return the time to sleep until the next call.""" assert self.interval is not None assert self._start_time is not None now = time.time() @@ -119,17 +183,20 @@ return self.interval - (running_for % self.interval) async def _loop(self) -> None: + """Run an infinite loop that calls the function periodically.""" while self.running: await asyncio.sleep(self._to_sleep()) self._call() def stop(self) -> None: + """Stop the periodic calls.""" self.interval = self._start_time = None if self._task is not None: self._task.cancel() self._task = None def _call(self) -> None: + """Execute the function.""" try: result = self._func(*self._args, **self._kwargs) except Exception: @@ -146,6 +213,11 @@ def create_looping_call( func: Callable[_P, _T], *args: _P.args, **kwargs: _P.kwargs ) -> AsyncioLoopingCall | LoopingCall: + """Create an instance of a looping call class. + + This creates an instance of :class:`AsyncioLoopingCall` or + :class:`LoopingCall`, depending on whether asyncio support is available. + """ if is_asyncio_available(): return AsyncioLoopingCall(func, *args, **kwargs) return LoopingCall(func, *args, **kwargs) @@ -154,6 +226,11 @@ def call_later( delay: float, func: Callable[[Unpack[_Ts]], object], *args: Unpack[_Ts] ) -> CallLaterResult: + """Schedule a function to be called after a delay. + + This uses either ``loop.call_later()`` or ``reactor.callLater()``, depending + on whether asyncio support is available. + """ if is_asyncio_available(): loop = asyncio.get_event_loop() return CallLaterResult.from_asyncio(loop.call_later(delay, func, *args)) @@ -164,26 +241,40 @@ class CallLaterResult: + """An universal result for :func:`call_later`, wrapping either + :class:`asyncio.TimerHandle` or :class:`twisted.internet.base.DelayedCall`. + + The provided API is close to the :class:`asyncio.TimerHandle` one: there is + no ``active()`` (as there is no such public API in + :class:`asyncio.TimerHandle`) but ``cancel()`` can be called on already + called or cancelled instances. + """ _timer_handle: asyncio.TimerHandle | None = None _delayed_call: DelayedCall | None = None @classmethod def from_asyncio(cls, timer_handle: asyncio.TimerHandle) -> Self: + """Create a CallLaterResult from an asyncio TimerHandle.""" o = cls() o._timer_handle = timer_handle return o @classmethod def from_twisted(cls, delayed_call: DelayedCall) -> Self: + """Create a CallLaterResult from a Twisted DelayedCall.""" o = cls() o._delayed_call = delayed_call return o def cancel(self) -> None: + """Cancel the underlying delayed call. + + Does nothing if the delayed call was already called or cancelled. + """ if self._timer_handle: self._timer_handle.cancel() self._timer_handle = None elif self._delayed_call and self._delayed_call.active(): self._delayed_call.cancel() - self._delayed_call = None+ self._delayed_call = None
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/asyncio.py
Generate docstrings for this script
from __future__ import annotations import logging import warnings from typing import TYPE_CHECKING, Any, cast from scrapy import signals from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.http import Request, Response from scrapy.utils.trackref import object_ref from scrapy.utils.url import url_is_from_spider if TYPE_CHECKING: from collections.abc import AsyncIterator, Iterable from twisted.internet.defer import Deferred # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.http.request import CallbackT from scrapy.settings import BaseSettings, _SettingsKey from scrapy.utils.log import SpiderLoggerAdapter class Spider(object_ref): name: str custom_settings: dict[_SettingsKey, Any] | None = None #: Start URLs. See :meth:`start`. start_urls: list[str] def __init__(self, name: str | None = None, **kwargs: Any): if name is not None: self.name: str = name elif not getattr(self, "name", None): raise ValueError(f"{type(self).__name__} must have a name") self.__dict__.update(kwargs) if not hasattr(self, "start_urls"): self.start_urls: list[str] = [] @property def logger(self) -> SpiderLoggerAdapter: # circular import from scrapy.utils.log import SpiderLoggerAdapter # noqa: PLC0415 logger = logging.getLogger(self.name) return SpiderLoggerAdapter(logger, {"spider": self}) def log(self, message: Any, level: int = logging.DEBUG, **kw: Any) -> None: self.logger.log(level, message, **kw) @classmethod def from_crawler(cls, crawler: Crawler, *args: Any, **kwargs: Any) -> Self: spider = cls(*args, **kwargs) spider._set_crawler(crawler) return spider def _set_crawler(self, crawler: Crawler) -> None: self.crawler: Crawler = crawler self.settings: BaseSettings = crawler.settings crawler.signals.connect(self.close, signals.spider_closed) async def start(self) -> AsyncIterator[Any]: with warnings.catch_warnings(): warnings.filterwarnings( "ignore", category=ScrapyDeprecationWarning, module=r"^scrapy\.spiders$" ) for item_or_request in self.start_requests(): yield item_or_request def start_requests(self) -> Iterable[Any]: warnings.warn( ( "The Spider.start_requests() method is deprecated, use " "Spider.start() instead. If you are calling " "super().start_requests() from a Spider.start() override, " "iterate super().start() instead." ), ScrapyDeprecationWarning, stacklevel=2, ) if not self.start_urls and hasattr(self, "start_url"): raise AttributeError( "Crawling could not start: 'start_urls' not found " "or empty (but found 'start_url' attribute instead, " "did you miss an 's'?)" ) for url in self.start_urls: yield Request(url, dont_filter=True) def _parse(self, response: Response, **kwargs: Any) -> Any: return self.parse(response, **kwargs) if TYPE_CHECKING: parse: CallbackT else: def parse(self, response: Response, **kwargs: Any) -> Any: raise NotImplementedError( f"{self.__class__.__name__}.parse callback is not defined" ) @classmethod def update_settings(cls, settings: BaseSettings) -> None: settings.setdict(cls.custom_settings or {}, priority="spider") @classmethod def handles_request(cls, request: Request) -> bool: return url_is_from_spider(request.url, cls) @staticmethod def close(spider: Spider, reason: str) -> Deferred[None] | None: closed = getattr(spider, "closed", None) if callable(closed): return cast("Deferred[None] | None", closed(reason)) return None def __repr__(self) -> str: return f"<{type(self).__name__} {self.name!r} at 0x{id(self):0x}>" # Top-level imports from scrapy.spiders.crawl import CrawlSpider, Rule from scrapy.spiders.feed import CSVFeedSpider, XMLFeedSpider from scrapy.spiders.sitemap import SitemapSpider __all__ = [ "CSVFeedSpider", "CrawlSpider", "Rule", "SitemapSpider", "Spider", "XMLFeedSpider", ]
--- +++ @@ -1,3 +1,8 @@+""" +Base class for Scrapy spiders + +See documentation in docs/topics/spiders.rst +""" from __future__ import annotations @@ -26,6 +31,12 @@ class Spider(object_ref): + """Base class that any spider must subclass. + + It provides a default :meth:`start` implementation that sends + requests based on the :attr:`start_urls` class attribute and calls the + :meth:`parse` method for each response. + """ name: str custom_settings: dict[_SettingsKey, Any] | None = None @@ -51,6 +62,12 @@ return SpiderLoggerAdapter(logger, {"spider": self}) def log(self, message: Any, level: int = logging.DEBUG, **kw: Any) -> None: + """Log the given message at the given log level + + This helper wraps a log call to the logger within the spider, but you + can use it directly (e.g. Spider.logger.info('msg')) or use any other + Python logger too. + """ self.logger.log(level, message, **kw) @classmethod @@ -65,6 +82,51 @@ crawler.signals.connect(self.close, signals.spider_closed) async def start(self) -> AsyncIterator[Any]: + """Yield the initial :class:`~scrapy.Request` objects to send. + + .. versionadded:: 2.13 + + For example: + + .. code-block:: python + + from scrapy import Request, Spider + + + class MySpider(Spider): + name = "myspider" + + async def start(self): + yield Request("https://toscrape.com/") + + The default implementation reads URLs from :attr:`start_urls` and + yields a request for each with :attr:`~scrapy.Request.dont_filter` + enabled. It is functionally equivalent to: + + .. code-block:: python + + async def start(self): + for url in self.start_urls: + yield Request(url, dont_filter=True) + + You can also yield :ref:`items <topics-items>`. For example: + + .. code-block:: python + + async def start(self): + yield {"foo": "bar"} + + To write spiders that work on Scrapy versions lower than 2.13, + define also a synchronous ``start_requests()`` method that returns an + iterable. For example: + + .. code-block:: python + + def start_requests(self): + yield Request("https://toscrape.com/") + + .. seealso:: :ref:`start-requests` + """ with warnings.catch_warnings(): warnings.filterwarnings( "ignore", category=ScrapyDeprecationWarning, module=r"^scrapy\.spiders$" @@ -135,4 +197,4 @@ "SitemapSpider", "Spider", "XMLFeedSpider", -]+]
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spiders/__init__.py
Generate consistent docstrings
from __future__ import annotations import os import warnings from importlib import import_module from pathlib import Path from scrapy.exceptions import NotConfigured from scrapy.settings import Settings from scrapy.utils.conf import closest_scrapy_cfg, get_config, init_env ENVVAR = "SCRAPY_SETTINGS_MODULE" DATADIR_CFG_SECTION = "datadir" def inside_project() -> bool: scrapy_module = os.environ.get(ENVVAR) if scrapy_module: try: import_module(scrapy_module) except ImportError as exc: warnings.warn( f"Cannot import scrapy settings module {scrapy_module}: {exc}" ) else: return True return bool(closest_scrapy_cfg()) def project_data_dir(project: str = "default") -> str: if not inside_project(): raise NotConfigured("Not inside a project") cfg = get_config() if cfg.has_option(DATADIR_CFG_SECTION, project): d = Path(cfg.get(DATADIR_CFG_SECTION, project)) else: scrapy_cfg = closest_scrapy_cfg() if not scrapy_cfg: raise NotConfigured( "Unable to find scrapy.cfg file to infer project data dir" ) d = (Path(scrapy_cfg).parent / ".scrapy").resolve() if not d.exists(): d.mkdir(parents=True) return str(d) def data_path(path: str | os.PathLike[str], createdir: bool = False) -> str: path_obj = Path(path) if not path_obj.is_absolute(): if inside_project(): path_obj = Path(project_data_dir(), path) else: path_obj = Path(".scrapy", path) if createdir and not path_obj.exists(): path_obj.mkdir(parents=True) return str(path_obj) def get_project_settings() -> Settings: if ENVVAR not in os.environ: project = os.environ.get("SCRAPY_PROJECT", "default") init_env(project) settings = Settings() settings_module_path = os.environ.get(ENVVAR) if settings_module_path: settings.setmodule(settings_module_path, priority="project") valid_envvars = { "CHECK", "PROJECT", "PYTHON_SHELL", "SETTINGS_MODULE", } scrapy_envvars = { k[7:]: v for k, v in os.environ.items() if k.startswith("SCRAPY_") and k.replace("SCRAPY_", "") in valid_envvars } settings.setdict(scrapy_envvars, priority="project") return settings
--- +++ @@ -28,6 +28,7 @@ def project_data_dir(project: str = "default") -> str: + """Return the current project data dir, creating it if it doesn't exist""" if not inside_project(): raise NotConfigured("Not inside a project") cfg = get_config() @@ -46,6 +47,10 @@ def data_path(path: str | os.PathLike[str], createdir: bool = False) -> str: + """ + Return the given path joined with the .scrapy data directory. + If given an absolute path, return it unmodified. + """ path_obj = Path(path) if not path_obj.is_absolute(): if inside_project(): @@ -82,4 +87,4 @@ settings.setdict(scrapy_envvars, priority="project") - return settings+ return settings
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/project.py
Insert docstrings into my code
from __future__ import annotations import marshal import pickle from pathlib import Path from typing import TYPE_CHECKING, Any from queuelib import queue from scrapy.utils.request import request_from_dict if TYPE_CHECKING: from collections.abc import Callable from os import PathLike # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy import Request from scrapy.crawler import Crawler def _with_mkdir(queue_class: type[queue.BaseQueue]) -> type[queue.BaseQueue]: class DirectoriesCreated(queue_class): # type: ignore[valid-type,misc] def __init__(self, path: str | PathLike, *args: Any, **kwargs: Any): dirname = Path(path).parent if not dirname.exists(): dirname.mkdir(parents=True, exist_ok=True) super().__init__(path, *args, **kwargs) return DirectoriesCreated def _serializable_queue( queue_class: type[queue.BaseQueue], serialize: Callable[[Any], bytes], deserialize: Callable[[bytes], Any], ) -> type[queue.BaseQueue]: class SerializableQueue(queue_class): # type: ignore[valid-type,misc] def push(self, obj: Any) -> None: s = serialize(obj) super().push(s) def pop(self) -> Any | None: s = super().pop() if s: return deserialize(s) return None def peek(self) -> Any | None: try: s = super().peek() except AttributeError as ex: raise NotImplementedError( "The underlying queue class does not implement 'peek'" ) from ex if s: return deserialize(s) return None return SerializableQueue def _scrapy_serialization_queue( queue_class: type[queue.BaseQueue], ) -> type[queue.BaseQueue]: class ScrapyRequestQueue(queue_class): # type: ignore[valid-type,misc] def __init__(self, crawler: Crawler, key: str): self.spider = crawler.spider super().__init__(key) @classmethod def from_crawler( cls, crawler: Crawler, key: str, *args: Any, **kwargs: Any ) -> Self: return cls(crawler, key) def push(self, request: Request) -> None: request_dict = request.to_dict(spider=self.spider) super().push(request_dict) def pop(self) -> Request | None: request = super().pop() if not request: return None return request_from_dict(request, spider=self.spider) def peek(self) -> Request | None: request = super().peek() if not request: return None return request_from_dict(request, spider=self.spider) return ScrapyRequestQueue def _scrapy_non_serialization_queue( queue_class: type[queue.BaseQueue], ) -> type[queue.BaseQueue]: class ScrapyRequestQueue(queue_class): # type: ignore[valid-type,misc] @classmethod def from_crawler(cls, crawler: Crawler, *args: Any, **kwargs: Any) -> Self: return cls() def peek(self) -> Any | None: try: s = super().peek() except AttributeError as ex: raise NotImplementedError( "The underlying queue class does not implement 'peek'" ) from ex return s return ScrapyRequestQueue def _pickle_serialize(obj: Any) -> bytes: try: return pickle.dumps(obj, protocol=4) # Both pickle.PicklingError and AttributeError can be raised by pickle.dump(s) # TypeError is raised from parsel.Selector except (pickle.PicklingError, AttributeError, TypeError) as e: raise ValueError(str(e)) from e # queue.*Queue aren't subclasses of queue.BaseQueue _PickleFifoSerializationDiskQueue = _serializable_queue( _with_mkdir(queue.FifoDiskQueue), # type: ignore[arg-type] _pickle_serialize, pickle.loads, ) _PickleLifoSerializationDiskQueue = _serializable_queue( _with_mkdir(queue.LifoDiskQueue), # type: ignore[arg-type] _pickle_serialize, pickle.loads, ) _MarshalFifoSerializationDiskQueue = _serializable_queue( _with_mkdir(queue.FifoDiskQueue), # type: ignore[arg-type] marshal.dumps, marshal.loads, ) _MarshalLifoSerializationDiskQueue = _serializable_queue( _with_mkdir(queue.LifoDiskQueue), # type: ignore[arg-type] marshal.dumps, marshal.loads, ) # public queue classes PickleFifoDiskQueue = _scrapy_serialization_queue(_PickleFifoSerializationDiskQueue) PickleLifoDiskQueue = _scrapy_serialization_queue(_PickleLifoSerializationDiskQueue) MarshalFifoDiskQueue = _scrapy_serialization_queue(_MarshalFifoSerializationDiskQueue) MarshalLifoDiskQueue = _scrapy_serialization_queue(_MarshalLifoSerializationDiskQueue) FifoMemoryQueue = _scrapy_non_serialization_queue(queue.FifoMemoryQueue) # type: ignore[arg-type] LifoMemoryQueue = _scrapy_non_serialization_queue(queue.LifoMemoryQueue) # type: ignore[arg-type]
--- +++ @@ -1,3 +1,6 @@+""" +Scheduler queues +""" from __future__ import annotations @@ -49,6 +52,12 @@ return None def peek(self) -> Any | None: + """Returns the next object to be returned by :meth:`pop`, + but without removing it from the queue. + + Raises :exc:`NotImplementedError` if the underlying queue class does + not implement a ``peek`` method, which is optional for queues. + """ try: s = super().peek() except AttributeError as ex: @@ -87,6 +96,12 @@ return request_from_dict(request, spider=self.spider) def peek(self) -> Request | None: + """Returns the next object to be returned by :meth:`pop`, + but without removing it from the queue. + + Raises :exc:`NotImplementedError` if the underlying queue class does + not implement a ``peek`` method, which is optional for queues. + """ request = super().peek() if not request: return None @@ -104,6 +119,12 @@ return cls() def peek(self) -> Any | None: + """Returns the next object to be returned by :meth:`pop`, + but without removing it from the queue. + + Raises :exc:`NotImplementedError` if the underlying queue class does + not implement a ``peek`` method, which is optional for queues. + """ try: s = super().peek() except AttributeError as ex: @@ -152,4 +173,4 @@ MarshalFifoDiskQueue = _scrapy_serialization_queue(_MarshalFifoSerializationDiskQueue) MarshalLifoDiskQueue = _scrapy_serialization_queue(_MarshalLifoSerializationDiskQueue) FifoMemoryQueue = _scrapy_non_serialization_queue(queue.FifoMemoryQueue) # type: ignore[arg-type] -LifoMemoryQueue = _scrapy_non_serialization_queue(queue.LifoMemoryQueue) # type: ignore[arg-type]+LifoMemoryQueue = _scrapy_non_serialization_queue(queue.LifoMemoryQueue) # type: ignore[arg-type]
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/squeues.py
Write docstrings describing each step
from __future__ import annotations import logging import pprint import sys from collections.abc import MutableMapping from logging.config import dictConfig from typing import TYPE_CHECKING, Any, cast from twisted.internet import asyncioreactor from twisted.python import log as twisted_log from twisted.python.failure import Failure import scrapy from scrapy.settings import Settings, _SettingsKey from scrapy.utils.versions import get_versions if TYPE_CHECKING: from types import TracebackType from scrapy.crawler import Crawler from scrapy.logformatter import LogFormatterResult logger = logging.getLogger(__name__) def failure_to_exc_info( failure: Failure, ) -> tuple[type[BaseException], BaseException, TracebackType | None] | None: if isinstance(failure, Failure): assert failure.type assert failure.value return ( failure.type, failure.value, cast("TracebackType | None", failure.getTracebackObject()), ) return None class TopLevelFormatter(logging.Filter): def __init__(self, loggers: list[str] | None = None): super().__init__() self.loggers: list[str] = loggers or [] def filter(self, record: logging.LogRecord) -> bool: if any(record.name.startswith(logger + ".") for logger in self.loggers): record.name = record.name.split(".", 1)[0] return True DEFAULT_LOGGING = { "version": 1, "disable_existing_loggers": False, "loggers": { "filelock": { "level": "ERROR", }, "hpack": { "level": "ERROR", }, "httpcore": { "level": "ERROR", }, "httpx": { "level": "WARNING", }, "scrapy": { "level": "DEBUG", }, "twisted": { "level": "ERROR", }, }, } def configure_logging( settings: Settings | dict[_SettingsKey, Any] | None = None, install_root_handler: bool = True, ) -> None: if not sys.warnoptions: # Route warnings through python logging logging.captureWarnings(True) observer = twisted_log.PythonLoggingObserver("twisted") observer.start() dictConfig(DEFAULT_LOGGING) if isinstance(settings, dict) or settings is None: settings = Settings(settings) if settings.getbool("LOG_STDOUT"): sys.stdout = StreamLogger(logging.getLogger("stdout")) if install_root_handler: install_scrapy_root_handler(settings) _scrapy_root_handler: logging.Handler | None = None def install_scrapy_root_handler(settings: Settings) -> None: global _scrapy_root_handler # noqa: PLW0603 _uninstall_scrapy_root_handler() logging.root.setLevel(logging.NOTSET) _scrapy_root_handler = _get_handler(settings) logging.root.addHandler(_scrapy_root_handler) def _uninstall_scrapy_root_handler() -> None: global _scrapy_root_handler # noqa: PLW0603 if ( _scrapy_root_handler is not None and _scrapy_root_handler in logging.root.handlers ): logging.root.removeHandler(_scrapy_root_handler) _scrapy_root_handler = None def get_scrapy_root_handler() -> logging.Handler | None: return _scrapy_root_handler def _get_handler(settings: Settings) -> logging.Handler: filename = settings.get("LOG_FILE") handler: logging.Handler if filename: mode = "a" if settings.getbool("LOG_FILE_APPEND") else "w" encoding = settings.get("LOG_ENCODING") handler = logging.FileHandler(filename, mode=mode, encoding=encoding) elif settings.getbool("LOG_ENABLED"): handler = logging.StreamHandler() else: handler = logging.NullHandler() formatter = logging.Formatter( fmt=settings.get("LOG_FORMAT"), datefmt=settings.get("LOG_DATEFORMAT") ) handler.setFormatter(formatter) handler.setLevel(settings.get("LOG_LEVEL")) if settings.getbool("LOG_SHORT_NAMES"): handler.addFilter(TopLevelFormatter(["scrapy"])) return handler def log_scrapy_info(settings: Settings) -> None: logger.info( "Scrapy %(version)s started (bot: %(bot)s)", {"version": scrapy.__version__, "bot": settings["BOT_NAME"]}, ) software: list[str] = settings.getlist("LOG_VERSIONS") if not software: return versions = pprint.pformat(dict(get_versions(software)), sort_dicts=False) logger.info(f"Versions:\n{versions}") def log_reactor_info() -> None: from twisted.internet import reactor logger.debug("Using reactor: %s.%s", reactor.__module__, reactor.__class__.__name__) if isinstance(reactor, asyncioreactor.AsyncioSelectorReactor): logger.debug( "Using asyncio event loop: %s.%s", reactor._asyncioEventloop.__module__, reactor._asyncioEventloop.__class__.__name__, ) class StreamLogger: def __init__(self, logger: logging.Logger, log_level: int = logging.INFO): self.logger: logging.Logger = logger self.log_level: int = log_level self.linebuf: str = "" def write(self, buf: str) -> None: for line in buf.rstrip().splitlines(): self.logger.log(self.log_level, line.rstrip()) def flush(self) -> None: for h in self.logger.handlers: h.flush() class LogCounterHandler(logging.Handler): def __init__(self, crawler: Crawler, *args: Any, **kwargs: Any): super().__init__(*args, **kwargs) self.crawler: Crawler = crawler def emit(self, record: logging.LogRecord) -> None: sname = f"log_count/{record.levelname}" assert self.crawler.stats self.crawler.stats.inc_value(sname) def logformatter_adapter( logkws: LogFormatterResult, ) -> tuple[int, str, dict[str, Any] | tuple[Any, ...]]: level = logkws.get("level", logging.INFO) message = logkws.get("msg") or "" # NOTE: This also handles 'args' being an empty dict, that case doesn't # play well in logger.log calls args = cast("dict[str, Any]", logkws) if not logkws.get("args") else logkws["args"] return (level, message, args) class SpiderLoggerAdapter(logging.LoggerAdapter): def process( self, msg: str, kwargs: MutableMapping[str, Any] ) -> tuple[str, MutableMapping[str, Any]]: if isinstance(kwargs.get("extra"), MutableMapping): kwargs["extra"].update(self.extra) else: kwargs["extra"] = self.extra return msg, kwargs
--- +++ @@ -28,6 +28,7 @@ def failure_to_exc_info( failure: Failure, ) -> tuple[type[BaseException], BaseException, TracebackType | None] | None: + """Extract exc_info from Failure instances""" if isinstance(failure, Failure): assert failure.type assert failure.value @@ -40,6 +41,16 @@ class TopLevelFormatter(logging.Filter): + """Keep only top level loggers' name (direct children from root) from + records. + + This filter will replace Scrapy loggers' names with 'scrapy'. This mimics + the old Scrapy log behaviour and helps shortening long names. + + Since it can't be set for just one logger (it won't propagate for its + children), it's going to be set in the root handler, with a parametrized + ``loggers`` list where it should act. + """ def __init__(self, loggers: list[str] | None = None): super().__init__() @@ -81,6 +92,29 @@ settings: Settings | dict[_SettingsKey, Any] | None = None, install_root_handler: bool = True, ) -> None: + """ + Initialize logging defaults for Scrapy. + + :param settings: settings used to create and configure a handler for the + root logger (default: None). + :type settings: dict, :class:`~scrapy.settings.Settings` object or ``None`` + + :param install_root_handler: whether to install root logging handler + (default: True) + :type install_root_handler: bool + + This function does: + + - Route warnings and twisted logging through Python standard logging + - Assign DEBUG and ERROR level to Scrapy and Twisted loggers respectively + - Route stdout to log if LOG_STDOUT setting is True + + When ``install_root_handler`` is True (default), this function also + creates a handler for the root logger according to given settings + (see :ref:`topics-logging-settings`). You can override default options + using ``settings`` argument. When ``settings`` is empty or None, defaults + are used. + """ if not sys.warnoptions: # Route warnings through python logging logging.captureWarnings(True) @@ -128,6 +162,7 @@ def _get_handler(settings: Settings) -> logging.Handler: + """Return a log handler object according to settings""" filename = settings.get("LOG_FILE") handler: logging.Handler if filename: @@ -174,6 +209,11 @@ class StreamLogger: + """Fake file-like stream object that redirects writes to a logger instance + + Taken from: + https://www.electricmonk.nl/log/2011/08/14/redirect-stdout-and-stderr-to-a-logger-in-python/ + """ def __init__(self, logger: logging.Logger, log_level: int = logging.INFO): self.logger: logging.Logger = logger @@ -190,6 +230,7 @@ class LogCounterHandler(logging.Handler): + """Record log levels count into a crawler stats""" def __init__(self, crawler: Crawler, *args: Any, **kwargs: Any): super().__init__(*args, **kwargs) @@ -204,6 +245,11 @@ def logformatter_adapter( logkws: LogFormatterResult, ) -> tuple[int, str, dict[str, Any] | tuple[Any, ...]]: + """ + Helper that takes the dictionary output from the methods in LogFormatter + and adapts it into a tuple of positional arguments for logger.log calls, + handling backward compatibility as well. + """ level = logkws.get("level", logging.INFO) message = logkws.get("msg") or "" @@ -218,9 +264,10 @@ def process( self, msg: str, kwargs: MutableMapping[str, Any] ) -> tuple[str, MutableMapping[str, Any]]: + """Method that augments logging with additional 'extra' data""" if isinstance(kwargs.get("extra"), MutableMapping): kwargs["extra"].update(self.extra) else: kwargs["extra"] = self.extra - return msg, kwargs+ return msg, kwargs
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/log.py
Add docstrings to clarify complex logic
from __future__ import annotations # used in global tests code from time import time # noqa: F401 from typing import TYPE_CHECKING, Any if TYPE_CHECKING: from scrapy.core.engine import ExecutionEngine def get_engine_status(engine: ExecutionEngine) -> list[tuple[str, Any]]: tests = [ "time()-engine.start_time", "len(engine.downloader.active)", "engine.scraper.is_idle()", "engine.spider.name", "engine.spider_is_idle()", "engine._slot.closing", "len(engine._slot.inprogress)", "len(engine._slot.scheduler.dqs or [])", "len(engine._slot.scheduler.mqs)", "len(engine.scraper.slot.queue)", "len(engine.scraper.slot.active)", "engine.scraper.slot.active_size", "engine.scraper.slot.itemproc_size", "engine.scraper.slot.needs_backout()", ] checks: list[tuple[str, Any]] = [] for test in tests: try: checks += [(test, eval(test))] # noqa: S307 except Exception as e: checks += [(test, f"{type(e).__name__} (exception)")] return checks def format_engine_status(engine: ExecutionEngine) -> str: checks = get_engine_status(engine) s = "Execution engine status\n\n" for test, result in checks: s += f"{test:<47} : {result}\n" s += "\n" return s def print_engine_status(engine: ExecutionEngine) -> None: print(format_engine_status(engine))
--- +++ @@ -1,3 +1,4 @@+"""Some debugging functions for working with the Scrapy engine""" from __future__ import annotations @@ -10,6 +11,7 @@ def get_engine_status(engine: ExecutionEngine) -> list[tuple[str, Any]]: + """Return a report of the current engine status""" tests = [ "time()-engine.start_time", "len(engine.downloader.active)", @@ -48,4 +50,4 @@ def print_engine_status(engine: ExecutionEngine) -> None: - print(format_engine_status(engine))+ print(format_engine_status(engine))
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/engine.py
Create docstrings for all classes and functions
from __future__ import annotations from abc import ABC from contextlib import contextmanager from typing import TYPE_CHECKING, Any from twisted.internet.defer import CancelledError from twisted.internet.error import ConnectionRefusedError as TxConnectionRefusedError from twisted.internet.error import DNSLookupError from twisted.internet.error import TimeoutError as TxTimeoutError from twisted.python.failure import Failure from twisted.web.client import ResponseFailed from twisted.web.error import SchemeNotSupported from scrapy import responsetypes from scrapy.core.downloader.handlers.base import BaseDownloadHandler from scrapy.exceptions import ( CannotResolveHostError, DownloadCancelledError, DownloadConnectionRefusedError, DownloadFailedError, DownloadTimeoutError, StopDownload, UnsupportedURLSchemeError, ) from scrapy.utils.log import logger if TYPE_CHECKING: from collections.abc import Iterator from ipaddress import IPv4Address, IPv6Address from twisted.internet.ssl import Certificate from scrapy import Request from scrapy.crawler import Crawler from scrapy.http import Headers, Response class BaseHttpDownloadHandler(BaseDownloadHandler, ABC): def __init__(self, crawler: Crawler): super().__init__(crawler) self._default_maxsize: int = crawler.settings.getint("DOWNLOAD_MAXSIZE") self._default_warnsize: int = crawler.settings.getint("DOWNLOAD_WARNSIZE") self._fail_on_dataloss: bool = crawler.settings.getbool( "DOWNLOAD_FAIL_ON_DATALOSS" ) self._fail_on_dataloss_warned: bool = False @contextmanager def wrap_twisted_exceptions() -> Iterator[None]: try: yield except SchemeNotSupported as e: raise UnsupportedURLSchemeError(str(e)) from e except CancelledError as e: raise DownloadCancelledError(str(e)) from e except TxConnectionRefusedError as e: raise DownloadConnectionRefusedError(str(e)) from e except DNSLookupError as e: raise CannotResolveHostError(str(e)) from e except ResponseFailed as e: raise DownloadFailedError(str(e)) from e except TxTimeoutError as e: raise DownloadTimeoutError(str(e)) from e def check_stop_download( signal: object, crawler: Crawler, request: Request, **kwargs: Any ) -> StopDownload | None: signal_result = crawler.signals.send_catch_log( signal=signal, request=request, spider=crawler.spider, **kwargs, ) for handler, result in signal_result: if isinstance(result, Failure) and isinstance(result.value, StopDownload): logger.debug( f"Download stopped for {request} from signal handler {handler.__qualname__}" ) return result.value return None def make_response( url: str, status: int, headers: Headers, body: bytes = b"", flags: list[str] | None = None, certificate: Certificate | None = None, ip_address: IPv4Address | IPv6Address | None = None, protocol: str | None = None, stop_download: StopDownload | None = None, ) -> Response: respcls = responsetypes.responsetypes.from_args(headers=headers, url=url, body=body) response = respcls( url=url, status=status, headers=headers, body=body, flags=flags, certificate=certificate, ip_address=ip_address, protocol=protocol, ) if stop_download: response.flags.append("download_stopped") if stop_download.fail: stop_download.response = response raise stop_download return response def get_maxsize_msg(size: int, limit: int, request: Request, *, expected: bool) -> str: prefix = "Expected to receive" if expected else "Received" return ( f"{prefix} {size} bytes which is larger than download " f"max size ({limit}) in request {request}." ) def get_warnsize_msg(size: int, limit: int, request: Request, *, expected: bool) -> str: prefix = "Expected to receive" if expected else "Received" return ( f"{prefix} {size} bytes which is larger than download " f"warn size ({limit}) in request {request}." ) def get_dataloss_msg(url: str) -> str: return ( f"Got data loss in {url}. If you want to process broken " f"responses set the setting DOWNLOAD_FAIL_ON_DATALOSS = False" f" -- This message won't be shown in further requests" )
--- +++ @@ -1,3 +1,4 @@+"""Utils for built-in HTTP download handlers.""" from __future__ import annotations @@ -38,6 +39,7 @@ class BaseHttpDownloadHandler(BaseDownloadHandler, ABC): + """Base class for built-in HTTP download handlers.""" def __init__(self, crawler: Crawler): super().__init__(crawler) @@ -51,6 +53,7 @@ @contextmanager def wrap_twisted_exceptions() -> Iterator[None]: + """Context manager that wraps Twisted exceptions into Scrapy exceptions.""" try: yield except SchemeNotSupported as e: @@ -70,6 +73,11 @@ def check_stop_download( signal: object, crawler: Crawler, request: Request, **kwargs: Any ) -> StopDownload | None: + """Send the given signal and check if any of its handlers raised + :exc:`~scrapy.exceptions.StopDownload`. + + Return the raised exception or ``None``. + """ signal_result = crawler.signals.send_catch_log( signal=signal, request=request, @@ -137,4 +145,4 @@ f"Got data loss in {url}. If you want to process broken " f"responses set the setting DOWNLOAD_FAIL_ON_DATALOSS = False" f" -- This message won't be shown in further requests" - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/_download_handlers.py
Help me write clear docstrings
from __future__ import annotations import asyncio import sys from contextlib import suppress from typing import TYPE_CHECKING, Any, Generic, ParamSpec, TypeVar from warnings import catch_warnings, filterwarnings from twisted.internet import asyncioreactor, error from twisted.internet.defer import Deferred from scrapy.utils.misc import load_object from scrapy.utils.python import global_object_name if TYPE_CHECKING: from asyncio import AbstractEventLoop from collections.abc import Callable from twisted.internet.protocol import ServerFactory from twisted.internet.tcp import Port from scrapy.utils.asyncio import CallLaterResult _T = TypeVar("_T") _P = ParamSpec("_P") def listen_tcp(portrange: list[int], host: str, factory: ServerFactory) -> Port: # type: ignore[return] # noqa: RET503 from twisted.internet import reactor if len(portrange) > 2: raise ValueError(f"invalid portrange: {portrange}") if not portrange: return reactor.listenTCP(0, factory, interface=host) if len(portrange) == 1: return reactor.listenTCP(portrange[0], factory, interface=host) for x in range(portrange[0], portrange[1] + 1): try: return reactor.listenTCP(x, factory, interface=host) except error.CannotListenError: if x == portrange[1]: raise class CallLaterOnce(Generic[_T]): def __init__(self, func: Callable[_P, _T], *a: _P.args, **kw: _P.kwargs): self._func: Callable[_P, _T] = func self._a: tuple[Any, ...] = a self._kw: dict[str, Any] = kw self._call: CallLaterResult | None = None self._deferreds: list[Deferred] = [] def schedule(self, delay: float = 0) -> None: # circular import from scrapy.utils.asyncio import call_later # noqa: PLC0415 if self._call is None: self._call = call_later(delay, self) def cancel(self) -> None: if self._call: self._call.cancel() def __call__(self) -> _T: # circular import from scrapy.utils.asyncio import call_later # noqa: PLC0415 self._call = None result = self._func(*self._a, **self._kw) for d in self._deferreds: call_later(0, d.callback, None) self._deferreds = [] return result async def wait(self) -> None: # circular import from scrapy.utils.defer import maybe_deferred_to_future # noqa: PLC0415 d: Deferred[None] = Deferred() self._deferreds.append(d) await maybe_deferred_to_future(d) _asyncio_reactor_path = "twisted.internet.asyncioreactor.AsyncioSelectorReactor" def set_asyncio_event_loop_policy() -> None: policy = asyncio.get_event_loop_policy() if sys.platform == "win32" and not isinstance( policy, asyncio.WindowsSelectorEventLoopPolicy ): policy = asyncio.WindowsSelectorEventLoopPolicy() asyncio.set_event_loop_policy(policy) def install_reactor(reactor_path: str, event_loop_path: str | None = None) -> None: reactor_class = load_object(reactor_path) if reactor_class is asyncioreactor.AsyncioSelectorReactor: set_asyncio_event_loop_policy() with suppress(error.ReactorAlreadyInstalledError): event_loop = set_asyncio_event_loop(event_loop_path) asyncioreactor.install(eventloop=event_loop) else: *module, _ = reactor_path.split(".") installer_path = [*module, "install"] installer = load_object(".".join(installer_path)) with suppress(error.ReactorAlreadyInstalledError): installer() def _get_asyncio_event_loop() -> AbstractEventLoop: return set_asyncio_event_loop(None) def set_asyncio_event_loop(event_loop_path: str | None) -> AbstractEventLoop: if event_loop_path is not None: event_loop_class: type[AbstractEventLoop] = load_object(event_loop_path) event_loop = _get_asyncio_event_loop() if not isinstance(event_loop, event_loop_class): event_loop = event_loop_class() asyncio.set_event_loop(event_loop) else: try: with catch_warnings(): # In Python 3.10.9, 3.11.1, 3.12 and 3.13, a DeprecationWarning # is emitted about the lack of a current event loop, because in # Python 3.14 and later `get_event_loop` will raise a # RuntimeError in that event. Because our code is already # prepared for that future behavior, we ignore the deprecation # warning. filterwarnings( "ignore", message="There is no current event loop", category=DeprecationWarning, ) event_loop = asyncio.get_event_loop() except RuntimeError: # `get_event_loop` raises RuntimeError when called with no asyncio # event loop yet installed in the following scenarios: # - Previsibly on Python 3.14 and later. # https://github.com/python/cpython/issues/100160#issuecomment-1345581902 event_loop = asyncio.new_event_loop() asyncio.set_event_loop(event_loop) return event_loop def verify_installed_reactor(reactor_path: str) -> None: if not is_reactor_installed(): raise RuntimeError( "verify_installed_reactor() called without an installed reactor." ) from twisted.internet import reactor expected_reactor_type = load_object(reactor_path) reactor_type = type(reactor) if not reactor_type == expected_reactor_type: raise RuntimeError( f"The installed reactor ({global_object_name(reactor_type)}) " f"does not match the requested one ({reactor_path})" ) def verify_installed_asyncio_event_loop(loop_path: str) -> None: if not is_reactor_installed(): raise RuntimeError( "verify_installed_asyncio_event_loop() called without an installed reactor." ) from twisted.internet import reactor loop_class = load_object(loop_path) if isinstance(reactor._asyncioEventloop, loop_class): return installed = ( f"{reactor._asyncioEventloop.__class__.__module__}" f".{reactor._asyncioEventloop.__class__.__qualname__}" ) raise RuntimeError( "Scrapy found an asyncio Twisted reactor already " f"installed, and its event loop class ({installed}) does " "not match the one specified in the ASYNCIO_EVENT_LOOP " f"setting ({global_object_name(loop_class)})" ) def is_reactor_installed() -> bool: return "twisted.internet.reactor" in sys.modules def is_asyncio_reactor_installed() -> bool: if not is_reactor_installed(): raise RuntimeError( "is_asyncio_reactor_installed() called without an installed reactor." ) from twisted.internet import reactor return isinstance(reactor, asyncioreactor.AsyncioSelectorReactor)
--- +++ @@ -27,6 +27,7 @@ def listen_tcp(portrange: list[int], host: str, factory: ServerFactory) -> Port: # type: ignore[return] # noqa: RET503 + """Like reactor.listenTCP but tries different ports in a range.""" from twisted.internet import reactor if len(portrange) > 2: @@ -44,6 +45,9 @@ class CallLaterOnce(Generic[_T]): + """Schedule a function to be called in the next reactor loop, but only if + it hasn't been already scheduled since the last time it ran. + """ def __init__(self, func: Callable[_P, _T], *a: _P.args, **kw: _P.kwargs): self._func: Callable[_P, _T] = func @@ -89,6 +93,10 @@ def set_asyncio_event_loop_policy() -> None: + """The policy functions from asyncio often behave unexpectedly, + so we restrict their use to the absolutely essential case. + This should only be used to install the reactor. + """ policy = asyncio.get_event_loop_policy() if sys.platform == "win32" and not isinstance( policy, asyncio.WindowsSelectorEventLoopPolicy @@ -98,6 +106,9 @@ def install_reactor(reactor_path: str, event_loop_path: str | None = None) -> None: + """Installs the :mod:`~twisted.internet.reactor` with the specified + import path. Also installs the asyncio event loop with the specified import + path if the asyncio reactor is enabled""" reactor_class = load_object(reactor_path) if reactor_class is asyncioreactor.AsyncioSelectorReactor: set_asyncio_event_loop_policy() @@ -117,6 +128,7 @@ def set_asyncio_event_loop(event_loop_path: str | None) -> AbstractEventLoop: + """Sets and returns the event loop with specified import path.""" if event_loop_path is not None: event_loop_class: type[AbstractEventLoop] = load_object(event_loop_path) event_loop = _get_asyncio_event_loop() @@ -149,6 +161,9 @@ def verify_installed_reactor(reactor_path: str) -> None: + """Raise :exc:`RuntimeError` if the installed + :mod:`~twisted.internet.reactor` does not match the specified import + path or if no reactor is installed.""" if not is_reactor_installed(): raise RuntimeError( "verify_installed_reactor() called without an installed reactor." @@ -166,6 +181,9 @@ def verify_installed_asyncio_event_loop(loop_path: str) -> None: + """Raise :exc:`RuntimeError` if the even loop of the installed + :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor` + does not match the specified import path or if no reactor is installed.""" if not is_reactor_installed(): raise RuntimeError( "verify_installed_asyncio_event_loop() called without an installed reactor." @@ -189,10 +207,26 @@ def is_reactor_installed() -> bool: + """Check whether a :mod:`~twisted.internet.reactor` is installed.""" return "twisted.internet.reactor" in sys.modules def is_asyncio_reactor_installed() -> bool: + """Check whether the installed reactor is :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor`. + + Raise a :exc:`RuntimeError` if no reactor is installed. + + In a future Scrapy version, when Scrapy supports running without a Twisted + reactor, this function won't be useful for checking if it's possible to use + asyncio features, so the code that that doesn't directly require a Twisted + reactor should use :func:`scrapy.utils.asyncio.is_asyncio_available` + instead of this function. + + .. versionchanged:: 2.13 + In earlier Scrapy versions this function silently installed the default + reactor if there was no reactor installed. Now it raises an exception to + prevent silent problems in this case. + """ if not is_reactor_installed(): raise RuntimeError( "is_asyncio_reactor_installed() called without an installed reactor." @@ -200,4 +234,4 @@ from twisted.internet import reactor - return isinstance(reactor, asyncioreactor.AsyncioSelectorReactor)+ return isinstance(reactor, asyncioreactor.AsyncioSelectorReactor)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/reactor.py
Create simple docstrings for beginners
from __future__ import annotations import sys from importlib.abc import MetaPathFinder from typing import TYPE_CHECKING from scrapy.utils.asyncio import is_asyncio_available from scrapy.utils.reactor import is_reactor_installed if TYPE_CHECKING: from collections.abc import Sequence from importlib.machinery import ModuleSpec from types import ModuleType def is_reactorless() -> bool: return is_asyncio_available() and not is_reactor_installed() class ReactorImportHook(MetaPathFinder): def find_spec( self, fullname: str, path: Sequence[str] | None, target: ModuleType | None = None, ) -> ModuleSpec | None: if fullname == "twisted.internet.reactor": raise ImportError( f"Import of {fullname} is forbidden when running without a Twisted reactor," f" as importing it installs the reactor, which can lead to unexpected behavior." ) return None def install_reactor_import_hook() -> None: sys.meta_path.insert(0, ReactorImportHook())
--- +++ @@ -14,10 +14,24 @@ def is_reactorless() -> bool: + """Check if we are running in the reactorless mode, i.e. with + :setting:`TWISTED_ENABLED` set to ``False``. + + As this checks the runtime state and not the setting itself, it can be + wrong when executed very early, before the reactor and/or the asyncio event + loop are initialized. + + .. note:: As this function uses + :func:`scrapy.utils.asyncio.is_asyncio_available()`, it has the same + limitations for detecting a running asyncio event loop as that one. + + .. versionadded:: VERSION + """ return is_asyncio_available() and not is_reactor_installed() class ReactorImportHook(MetaPathFinder): + """Hook that prevents importing :mod:`twisted.internet.reactor`.""" def find_spec( self, @@ -34,5 +48,6 @@ def install_reactor_import_hook() -> None: + """Prevent importing :mod:`twisted.internet.reactor`.""" - sys.meta_path.insert(0, ReactorImportHook())+ sys.meta_path.insert(0, ReactorImportHook())
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/reactorless.py
Add docstrings for production code
from __future__ import annotations import logging import re # Iterable is needed at the run time for the SitemapSpider._parse_sitemap() annotation from collections.abc import AsyncIterator, Iterable, Sequence # noqa: TC003 from typing import TYPE_CHECKING, Any, cast from scrapy.http import Request, Response, XmlResponse from scrapy.spiders import Spider from scrapy.utils._compression import _DecompressionMaxSizeExceeded from scrapy.utils.gz import gunzip, gzip_magic_number from scrapy.utils.sitemap import Sitemap, sitemap_urls_from_robots if TYPE_CHECKING: # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler from scrapy.http.request import CallbackT logger = logging.getLogger(__name__) class SitemapSpider(Spider): sitemap_urls: Sequence[str] = () sitemap_rules: Sequence[tuple[re.Pattern[str] | str, str | CallbackT]] = [ ("", "parse") ] sitemap_follow: Sequence[re.Pattern[str] | str] = [""] sitemap_alternate_links: bool = False _max_size: int _warn_size: int @classmethod def from_crawler(cls, crawler: Crawler, *args: Any, **kwargs: Any) -> Self: spider = super().from_crawler(crawler, *args, **kwargs) spider._max_size = getattr( spider, "download_maxsize", spider.settings.getint("DOWNLOAD_MAXSIZE") ) spider._warn_size = getattr( spider, "download_warnsize", spider.settings.getint("DOWNLOAD_WARNSIZE") ) return spider def __init__(self, *a: Any, **kw: Any): super().__init__(*a, **kw) self._cbs: list[tuple[re.Pattern[str], CallbackT]] = [] for r, c in self.sitemap_rules: if isinstance(c, str): c = cast("CallbackT", getattr(self, c)) self._cbs.append((regex(r), c)) self._follow: list[re.Pattern[str]] = [regex(x) for x in self.sitemap_follow] async def start(self) -> AsyncIterator[Any]: for item_or_request in self.start_requests(): yield item_or_request def start_requests(self) -> Iterable[Request]: for url in self.sitemap_urls: yield Request(url, self._parse_sitemap) def sitemap_filter( self, entries: Iterable[dict[str, Any]] ) -> Iterable[dict[str, Any]]: yield from entries def _parse_sitemap(self, response: Response) -> Iterable[Request]: if response.url.endswith("/robots.txt"): for url in sitemap_urls_from_robots(response.text, base_url=response.url): yield Request(url, callback=self._parse_sitemap) else: body = self._get_sitemap_body(response) if body is None: logger.warning( "Ignoring invalid sitemap: %(response)s", {"response": response}, extra={"spider": self}, ) return s = Sitemap(body) it = self.sitemap_filter(s) if s.type == "sitemapindex": for loc in iterloc(it, self.sitemap_alternate_links): if any(x.search(loc) for x in self._follow): yield Request(loc, callback=self._parse_sitemap) elif s.type == "urlset": for loc in iterloc(it, self.sitemap_alternate_links): for r, c in self._cbs: if r.search(loc): yield Request(loc, callback=c) break def _get_sitemap_body(self, response: Response) -> bytes | None: if isinstance(response, XmlResponse): return response.body if gzip_magic_number(response): uncompressed_size = len(response.body) max_size = response.meta.get("download_maxsize", self._max_size) warn_size = response.meta.get("download_warnsize", self._warn_size) try: body = gunzip(response.body, max_size=max_size) except _DecompressionMaxSizeExceeded: return None if uncompressed_size < warn_size <= len(body): logger.warning( f"{response} body size after decompression ({len(body)} B) " f"is larger than the download warning size ({warn_size} B)." ) return body # actual gzipped sitemap files are decompressed above ; # if we are here (response body is not gzipped) # and have a response for .xml.gz, # it usually means that it was already gunzipped # by HttpCompression middleware, # the HTTP response being sent with "Content-Encoding: gzip" # without actually being a .xml.gz file in the first place, # merely XML gzip-compressed on the fly, # in other word, here, we have plain XML if response.url.endswith(".xml") or response.url.endswith(".xml.gz"): return response.body return None def regex(x: re.Pattern[str] | str) -> re.Pattern[str]: if isinstance(x, str): return re.compile(x) return x def iterloc(it: Iterable[dict[str, Any]], alt: bool = False) -> Iterable[str]: for d in it: yield d["loc"] # Also consider alternate URLs (xhtml:link rel="alternate") if alt and "alternate" in d: yield from d["alternate"]
--- +++ @@ -64,6 +64,10 @@ def sitemap_filter( self, entries: Iterable[dict[str, Any]] ) -> Iterable[dict[str, Any]]: + """This method can be used to filter sitemap entries by their + attributes, for example, you can filter locs with lastmod greater + than a given date (see docs). + """ yield from entries def _parse_sitemap(self, response: Response) -> Iterable[Request]: @@ -95,6 +99,9 @@ break def _get_sitemap_body(self, response: Response) -> bytes | None: + """Return the sitemap body contained in the given response, + or None if the response is not a sitemap. + """ if isinstance(response, XmlResponse): return response.body if gzip_magic_number(response): @@ -137,4 +144,4 @@ # Also consider alternate URLs (xhtml:link rel="alternate") if alt and "alternate" in d: - yield from d["alternate"]+ yield from d["alternate"]
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spiders/sitemap.py
Add docstrings to incomplete code
import posixpath from ftplib import FTP, error_perm from posixpath import dirname from typing import IO def ftp_makedirs_cwd(ftp: FTP, path: str, first_call: bool = True) -> None: try: ftp.cwd(path) except error_perm: ftp_makedirs_cwd(ftp, dirname(path), False) ftp.mkd(path) if first_call: ftp.cwd(path) def ftp_store_file( *, path: str, file: IO[bytes], host: str, port: int, username: str, password: str, use_active_mode: bool = False, overwrite: bool = True, ) -> None: with FTP() as ftp: ftp.connect(host, port) ftp.login(username, password) if use_active_mode: ftp.set_pasv(False) file.seek(0) dirname, filename = posixpath.split(path) ftp_makedirs_cwd(ftp, dirname) command = "STOR" if overwrite else "APPE" ftp.storbinary(f"{command} {filename}", file) file.close()
--- +++ @@ -5,6 +5,10 @@ def ftp_makedirs_cwd(ftp: FTP, path: str, first_call: bool = True) -> None: + """Set the current directory of the FTP connection given in the ``ftp`` + argument (as a ftplib.FTP object), creating all parent directories if they + don't exist. The ftplib.FTP object must be already connected and logged in. + """ try: ftp.cwd(path) except error_perm: @@ -25,6 +29,9 @@ use_active_mode: bool = False, overwrite: bool = True, ) -> None: + """Opens a FTP connection with passed credentials,sets current directory + to the directory extracted from given path, then uploads the file to server + """ with FTP() as ftp: ftp.connect(host, port) ftp.login(username, password) @@ -35,4 +42,4 @@ ftp_makedirs_cwd(ftp, dirname) command = "STOR" if overwrite else "APPE" ftp.storbinary(f"{command} {filename}", file) - file.close()+ file.close()
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/ftp.py
Write proper docstrings for these functions
from __future__ import annotations import csv import logging import re from io import StringIO from typing import TYPE_CHECKING, Any, Literal, cast, overload from warnings import warn from lxml import etree from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.http import Response, TextResponse from scrapy.selector import Selector from scrapy.utils.python import re_rsearch if TYPE_CHECKING: from collections.abc import Callable, Iterator logger = logging.getLogger(__name__) def xmliter(obj: Response | str | bytes, nodename: str) -> Iterator[Selector]: warn( ( "xmliter is deprecated and its use strongly discouraged because " "it is vulnerable to ReDoS attacks. Use xmliter_lxml instead. See " "https://github.com/scrapy/scrapy/security/advisories/GHSA-cc65-xxvf-f7r9" ), ScrapyDeprecationWarning, stacklevel=2, ) nodename_patt = re.escape(nodename) DOCUMENT_HEADER_RE = re.compile(r"<\?xml[^>]+>\s*", re.DOTALL) HEADER_END_RE = re.compile(rf"<\s*/{nodename_patt}\s*>", re.DOTALL) END_TAG_RE = re.compile(r"<\s*/([^\s>]+)\s*>", re.DOTALL) NAMESPACE_RE = re.compile(r"((xmlns[:A-Za-z]*)=[^>\s]+)", re.DOTALL) text = _body_or_str(obj) document_header_match = re.search(DOCUMENT_HEADER_RE, text) document_header = ( document_header_match.group().strip() if document_header_match else "" ) header_end_idx = re_rsearch(HEADER_END_RE, text) header_end = text[header_end_idx[1] :].strip() if header_end_idx else "" namespaces: dict[str, str] = {} if header_end: for tagname in reversed(re.findall(END_TAG_RE, header_end)): assert header_end_idx tag = re.search( rf"<\s*{tagname}.*?xmlns[:=][^>]*>", text[: header_end_idx[1]], re.DOTALL, ) if tag: for x in re.findall(NAMESPACE_RE, tag.group()): namespaces[x[1]] = x[0] r = re.compile(rf"<{nodename_patt}[\s>].*?</{nodename_patt}>", re.DOTALL) for match in r.finditer(text): nodetext = ( document_header + match.group().replace( nodename, f"{nodename} {' '.join(namespaces.values())}", 1 ) + header_end ) yield Selector(text=nodetext, type="xml") def xmliter_lxml( obj: Response | str | bytes, nodename: str, namespace: str | None = None, prefix: str = "x", ) -> Iterator[Selector]: reader = _StreamReader(obj) tag = f"{{{namespace}}}{nodename}" if namespace else nodename iterable = etree.iterparse( reader, encoding=reader.encoding, events=("end", "start-ns"), resolve_entities=False, huge_tree=True, ) selxpath = "//" + (f"{prefix}:{nodename}" if namespace else nodename) needs_namespace_resolution = not namespace and ":" in nodename if needs_namespace_resolution: prefix, nodename = nodename.split(":", maxsplit=1) for event, data in iterable: if event == "start-ns": assert isinstance(data, tuple) if needs_namespace_resolution: _prefix, _namespace = data if _prefix != prefix: continue namespace = _namespace needs_namespace_resolution = False selxpath = f"//{prefix}:{nodename}" tag = f"{{{namespace}}}{nodename}" continue assert isinstance(data, etree._Element) node = data if node.tag != tag: continue nodetext = etree.tostring(node, encoding="unicode") node.clear() xs = Selector(text=nodetext, type="xml") if namespace: xs.register_namespace(prefix, namespace) yield xs.xpath(selxpath)[0] class _StreamReader: def __init__(self, obj: Response | str | bytes): self._ptr: int = 0 self._text: str | bytes if isinstance(obj, TextResponse): self._text, self.encoding = obj.body, obj.encoding elif isinstance(obj, Response): self._text, self.encoding = obj.body, "utf-8" else: self._text, self.encoding = obj, "utf-8" self._is_unicode: bool = isinstance(self._text, str) self._is_first_read: bool = True def read(self, n: int = 65535) -> bytes: method: Callable[[int], bytes] = ( self._read_unicode if self._is_unicode else self._read_string ) result = method(n) if self._is_first_read: self._is_first_read = False result = result.lstrip() return result def _read_string(self, n: int = 65535) -> bytes: s, e = self._ptr, self._ptr + n self._ptr = e return cast("bytes", self._text)[s:e] def _read_unicode(self, n: int = 65535) -> bytes: s, e = self._ptr, self._ptr + n self._ptr = e return cast("str", self._text)[s:e].encode("utf-8") def csviter( obj: Response | str | bytes, delimiter: str | None = None, headers: list[str] | None = None, encoding: str | None = None, quotechar: str | None = None, ) -> Iterator[dict[str, str]]: if encoding is not None: # pragma: no cover warn( "The encoding argument of csviter() is ignored and will be removed" " in a future Scrapy version.", category=ScrapyDeprecationWarning, stacklevel=2, ) lines = StringIO(_body_or_str(obj, unicode=True)) kwargs: dict[str, Any] = {} if delimiter: kwargs["delimiter"] = delimiter if quotechar: kwargs["quotechar"] = quotechar csv_r = csv.reader(lines, **kwargs) if not headers: try: headers = next(csv_r) except StopIteration: return for row in csv_r: if len(row) != len(headers): logger.warning( "ignoring row %(csvlnum)d (length: %(csvrow)d, " "should be: %(csvheader)d)", { "csvlnum": csv_r.line_num, "csvrow": len(row), "csvheader": len(headers), }, ) continue yield dict(zip(headers, row, strict=False)) @overload def _body_or_str(obj: Response | str | bytes) -> str: ... @overload def _body_or_str(obj: Response | str | bytes, unicode: Literal[True]) -> str: ... @overload def _body_or_str(obj: Response | str | bytes, unicode: Literal[False]) -> bytes: ... def _body_or_str(obj: Response | str | bytes, unicode: bool = True) -> str | bytes: expected_types = (Response, str, bytes) if not isinstance(obj, expected_types): expected_types_str = " or ".join(t.__name__ for t in expected_types) raise TypeError( f"Object {obj!r} must be {expected_types_str}, not {type(obj).__name__}" ) if isinstance(obj, Response): if not unicode: return obj.body if isinstance(obj, TextResponse): return obj.text return obj.body.decode("utf-8") if isinstance(obj, str): return obj if unicode else obj.encode("utf-8") return obj.decode("utf-8") if unicode else obj
--- +++ @@ -21,6 +21,14 @@ def xmliter(obj: Response | str | bytes, nodename: str) -> Iterator[Selector]: + """Return a iterator of Selector's over all nodes of a XML document, + given the name of the node to iterate. Useful for parsing XML feeds. + + obj can be: + - a Response object + - a unicode string + - a string encoded as utf-8 + """ warn( ( "xmliter is deprecated and its use strongly discouraged because " @@ -154,6 +162,20 @@ encoding: str | None = None, quotechar: str | None = None, ) -> Iterator[dict[str, str]]: + """Returns an iterator of dictionaries from the given csv object + + obj can be: + - a Response object + - a unicode string + - a string encoded as utf-8 + + delimiter is the character used to separate fields on the given obj. + + headers is an iterable that when provided offers the keys + for the returned dictionaries, if not the first row is used. + + quotechar is the character used to enclosure fields on the given obj. + """ if encoding is not None: # pragma: no cover warn( @@ -220,4 +242,4 @@ return obj.body.decode("utf-8") if isinstance(obj, str): return obj if unicode else obj.encode("utf-8") - return obj.decode("utf-8") if unicode else obj+ return obj.decode("utf-8") if unicode else obj
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/iterators.py
Add return value explanations in docstrings
from __future__ import annotations import ast import hashlib import inspect import os import re import warnings from collections import deque from contextlib import contextmanager from functools import partial from importlib import import_module from pkgutil import iter_modules from typing import IO, TYPE_CHECKING, Any, TypeVar, cast from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.item import Item from scrapy.utils.datatypes import LocalWeakReferencedCache if TYPE_CHECKING: from collections.abc import Callable, Iterable, Iterator from types import ModuleType from scrapy import Spider from scrapy.crawler import Crawler _ITERABLE_SINGLE_VALUES = dict, Item, str, bytes T = TypeVar("T") def arg_to_iter(arg: Any) -> Iterable[Any]: if arg is None: return [] if not isinstance(arg, _ITERABLE_SINGLE_VALUES) and hasattr(arg, "__iter__"): return cast("Iterable[Any]", arg) return [arg] def load_object(path: str | Callable[..., Any]) -> Any: if not isinstance(path, str): if callable(path): return path raise TypeError( f"Unexpected argument type, expected string or object, got: {type(path)}" ) try: dot = path.rindex(".") except ValueError: raise ValueError(f"Error loading object '{path}': not a full path") module, name = path[:dot], path[dot + 1 :] mod = import_module(module) try: obj = getattr(mod, name) except AttributeError: raise NameError(f"Module '{module}' doesn't define any object named '{name}'") return obj def walk_modules(path: str) -> list[ModuleType]: mods: list[ModuleType] = [] mod = import_module(path) mods.append(mod) if hasattr(mod, "__path__"): for _, subpath, ispkg in iter_modules(mod.__path__): fullpath = path + "." + subpath if ispkg: mods += walk_modules(fullpath) else: submod = import_module(fullpath) mods.append(submod) return mods def md5sum(file: IO[bytes]) -> str: warnings.warn( ( "The scrapy.utils.misc.md5sum function is deprecated and will be " "removed in a future version of Scrapy." ), ScrapyDeprecationWarning, stacklevel=2, ) m = hashlib.md5() # noqa: S324 while True: d = file.read(8096) if not d: break m.update(d) return m.hexdigest() def rel_has_nofollow(rel: str | None) -> bool: return rel is not None and "nofollow" in rel.replace(",", " ").split() def build_from_crawler( objcls: type[T], crawler: Crawler, /, *args: Any, **kwargs: Any ) -> T: if hasattr(objcls, "from_crawler"): instance = objcls.from_crawler(crawler, *args, **kwargs) # type: ignore[attr-defined] method_name = "from_crawler" else: instance = objcls(*args, **kwargs) method_name = "__new__" if instance is None: raise TypeError(f"{objcls.__qualname__}.{method_name} returned None") return cast("T", instance) @contextmanager def set_environ(**kwargs: str) -> Iterator[None]: original_env = {k: os.environ.get(k) for k in kwargs} os.environ.update(kwargs) try: yield finally: for k, v in original_env.items(): if v is None: del os.environ[k] else: os.environ[k] = v def walk_callable(node: ast.AST) -> Iterable[ast.AST]: todo: deque[ast.AST] = deque([node]) walked_func_def = False while todo: node = todo.popleft() if isinstance(node, ast.FunctionDef): if walked_func_def: continue walked_func_def = True todo.extend(ast.iter_child_nodes(node)) yield node _generator_callbacks_cache = LocalWeakReferencedCache(limit=128) def is_generator_with_return_value(callable: Callable[..., Any]) -> bool: # noqa: A002 if callable in _generator_callbacks_cache: return bool(_generator_callbacks_cache[callable]) def returns_none(return_node: ast.Return) -> bool: value = return_node.value return value is None or ( isinstance(value, ast.Constant) and value.value is None ) if inspect.isgeneratorfunction(callable): func = callable while isinstance(func, partial): func = func.func src = inspect.getsource(func) pattern = re.compile(r"(^[\t ]+)") code = pattern.sub("", src) match = pattern.match(src) # finds indentation if match: code = re.sub(f"\n{match.group(0)}", "\n", code) # remove indentation tree = ast.parse(code) for node in walk_callable(tree): if isinstance(node, ast.Return) and not returns_none(node): _generator_callbacks_cache[callable] = True return bool(_generator_callbacks_cache[callable]) _generator_callbacks_cache[callable] = False return bool(_generator_callbacks_cache[callable]) def warn_on_generator_with_return_value( spider: Spider, callable: Callable[..., Any], # noqa: A002 ) -> None: if not spider.settings.getbool("WARN_ON_GENERATOR_RETURN_VALUE"): return try: if is_generator_with_return_value(callable): warnings.warn( f'The "{spider.__class__.__name__}.{callable.__name__}" method is ' 'a generator and includes a "return" statement with a value ' "different than None. This could lead to unexpected behaviour. Please see " "https://docs.python.org/3/reference/simple_stmts.html#the-return-statement " 'for details about the semantics of the "return" statement within generators', stacklevel=2, ) except IndentationError: callable_name = spider.__class__.__name__ + "." + callable.__name__ warnings.warn( f'Unable to determine whether or not "{callable_name}" is a generator with a return value. ' "This will not prevent your code from working, but it prevents Scrapy from detecting " f'potential issues in your implementation of "{callable_name}". Please, report this in the ' "Scrapy issue tracker (https://github.com/scrapy/scrapy/issues), " f'including the code of "{callable_name}"', stacklevel=2, )
--- +++ @@ -1,3 +1,4 @@+"""Helper functions which don't fit anywhere else""" from __future__ import annotations @@ -31,6 +32,11 @@ def arg_to_iter(arg: Any) -> Iterable[Any]: + """Convert an argument to an iterable. The argument can be a None, single + value, or an iterable. + + Exception: if arg is a dict, [arg] will be returned + """ if arg is None: return [] if not isinstance(arg, _ITERABLE_SINGLE_VALUES) and hasattr(arg, "__iter__"): @@ -39,6 +45,14 @@ def load_object(path: str | Callable[..., Any]) -> Any: + """Load an object given its absolute object path, and return it. + + The object can be the import path of a class, function, variable or an + instance, e.g. 'scrapy.downloadermiddlewares.redirect.RedirectMiddleware'. + + If ``path`` is not a string, but is a callable object, such as a class or + a function, then return it as is. + """ if not isinstance(path, str): if callable(path): @@ -64,6 +78,12 @@ def walk_modules(path: str) -> list[ModuleType]: + """Loads a module and all its submodules from the given module path and + returns them. If *any* module throws an exception while importing, that + exception is thrown back. + + For example: walk_modules('scrapy.utils') + """ mods: list[ModuleType] = [] mod = import_module(path) @@ -80,6 +100,13 @@ def md5sum(file: IO[bytes]) -> str: + """Calculate the md5 checksum of a file-like object without reading its + whole content in memory. + + >>> from io import BytesIO + >>> md5sum(BytesIO(b'file content to hash')) + '784406af91dd5a54fbb9c84c2236595a' + """ warnings.warn( ( "The scrapy.utils.misc.md5sum function is deprecated and will be " @@ -98,12 +125,21 @@ def rel_has_nofollow(rel: str | None) -> bool: + """Return True if link rel attribute has nofollow type""" return rel is not None and "nofollow" in rel.replace(",", " ").split() def build_from_crawler( objcls: type[T], crawler: Crawler, /, *args: Any, **kwargs: Any ) -> T: + """Construct a class instance using its ``from_crawler()`` or ``__init__()`` constructor. + + .. versionadded:: 2.12 + + ``*args`` and ``**kwargs`` are forwarded to the constructor. + + Raises ``TypeError`` if the resulting instance is ``None``. + """ if hasattr(objcls, "from_crawler"): instance = objcls.from_crawler(crawler, *args, **kwargs) # type: ignore[attr-defined] method_name = "from_crawler" @@ -117,6 +153,9 @@ @contextmanager def set_environ(**kwargs: str) -> Iterator[None]: + """Temporarily set environment variables inside the context manager and + fully restore previous environment afterwards + """ original_env = {k: os.environ.get(k) for k in kwargs} os.environ.update(kwargs) @@ -131,6 +170,9 @@ def walk_callable(node: ast.AST) -> Iterable[ast.AST]: + """Similar to ``ast.walk``, but walks only function body and skips nested + functions defined within the node. + """ todo: deque[ast.AST] = deque([node]) walked_func_def = False while todo: @@ -147,6 +189,10 @@ def is_generator_with_return_value(callable: Callable[..., Any]) -> bool: # noqa: A002 + """ + Returns True if a callable is a generator function which includes a + 'return' statement with a value different than None, False otherwise + """ if callable in _generator_callbacks_cache: return bool(_generator_callbacks_cache[callable]) @@ -183,6 +229,10 @@ spider: Spider, callable: Callable[..., Any], # noqa: A002 ) -> None: + """ + Logs a warning if a callable is a generator function and includes + a 'return' statement with a value different than None + """ if not spider.settings.getbool("WARN_ON_GENERATOR_RETURN_VALUE"): return try: @@ -204,4 +254,4 @@ "Scrapy issue tracker (https://github.com/scrapy/scrapy/issues), " f'including the code of "{callable_name}"', stacklevel=2, - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/misc.py
Fully document this Python code with docstrings
from __future__ import annotations import numbers import os import sys from configparser import ConfigParser from operator import itemgetter from pathlib import Path from typing import TYPE_CHECKING, Any, cast from scrapy.exceptions import UsageError from scrapy.settings import BaseSettings from scrapy.utils.deprecate import update_classpath from scrapy.utils.python import without_none_values if TYPE_CHECKING: from collections.abc import Callable, Collection, Iterable, Mapping, MutableMapping def build_component_list( compdict: MutableMapping[Any, Any], *, convert: Callable[[Any], Any] = update_classpath, ) -> list[Any]: def _check_components(complist: Collection[Any]) -> None: if len({convert(c) for c in complist}) != len(complist): raise ValueError( f"Some paths in {complist!r} convert to the same object, " "please update your settings" ) def _map_keys(compdict: Mapping[Any, Any]) -> BaseSettings | dict[Any, Any]: if isinstance(compdict, BaseSettings): compbs = BaseSettings() for k, v in compdict.items(): prio = compdict.getpriority(k) assert prio is not None if compbs.getpriority(convert(k)) == prio: raise ValueError( f"Some paths in {list(compdict.keys())!r} " "convert to the same " "object, please update your settings" ) compbs.set(convert(k), v, priority=prio) return compbs _check_components(compdict) return {convert(k): v for k, v in compdict.items()} def _validate_values(compdict: Mapping[Any, Any]) -> None: for name, value in compdict.items(): if value is not None and not isinstance(value, numbers.Real): raise ValueError( f"Invalid value {value} for component {name}, " "please provide a real number or None instead" ) _validate_values(compdict) compdict = without_none_values(_map_keys(compdict)) return [k for k, v in sorted(compdict.items(), key=itemgetter(1))] def arglist_to_dict(arglist: list[str]) -> dict[str, str]: return dict(x.split("=", 1) for x in arglist) def closest_scrapy_cfg( path: str | os.PathLike = ".", prevpath: str | os.PathLike | None = None, ) -> str: if prevpath is not None and str(path) == str(prevpath): return "" path = Path(path).resolve() cfgfile = path / "scrapy.cfg" if cfgfile.exists(): return str(cfgfile) return closest_scrapy_cfg(path.parent, path) def init_env(project: str = "default", set_syspath: bool = True) -> None: cfg = get_config() if cfg.has_option("settings", project): os.environ["SCRAPY_SETTINGS_MODULE"] = cfg.get("settings", project) closest = closest_scrapy_cfg() if closest: projdir = str(Path(closest).parent) if set_syspath and projdir not in sys.path: sys.path.append(projdir) def get_config(use_closest: bool = True) -> ConfigParser: sources = get_sources(use_closest) cfg = ConfigParser() cfg.read(sources) return cfg def get_sources(use_closest: bool = True) -> list[str]: xdg_config_home = ( os.environ.get("XDG_CONFIG_HOME") or Path("~/.config").expanduser() ) sources = [ "/etc/scrapy.cfg", r"c:\scrapy\scrapy.cfg", str(Path(xdg_config_home) / "scrapy.cfg"), str(Path("~/.scrapy.cfg").expanduser()), ] if use_closest: sources.append(closest_scrapy_cfg()) return sources def feed_complete_default_values_from_settings( feed: dict[str, Any], settings: BaseSettings ) -> dict[str, Any]: out = feed.copy() out.setdefault("batch_item_count", settings.getint("FEED_EXPORT_BATCH_ITEM_COUNT")) out.setdefault("encoding", settings["FEED_EXPORT_ENCODING"]) out.setdefault("fields", settings.getdictorlist("FEED_EXPORT_FIELDS") or None) out.setdefault("store_empty", settings.getbool("FEED_STORE_EMPTY")) out.setdefault("uri_params", settings["FEED_URI_PARAMS"]) out.setdefault("item_export_kwargs", {}) if settings["FEED_EXPORT_INDENT"] is None: out.setdefault("indent", None) else: out.setdefault("indent", settings.getint("FEED_EXPORT_INDENT")) return out def feed_process_params_from_cli( settings: BaseSettings, output: list[str], *, overwrite_output: list[str] | None = None, ) -> dict[str, dict[str, Any]]: valid_output_formats: Iterable[str] = without_none_values( cast("dict[str, str]", settings.getwithbase("FEED_EXPORTERS")) ).keys() def check_valid_format(output_format: str) -> None: if output_format not in valid_output_formats: raise UsageError( f"Unrecognized output format '{output_format}'. " f"Set a supported one ({tuple(valid_output_formats)}) " "after a colon at the end of the output URI (i.e. -o/-O " "<URI>:<FORMAT>) or as a file extension." ) overwrite = False if overwrite_output: if output: raise UsageError( "Please use only one of -o/--output and -O/--overwrite-output" ) output = overwrite_output overwrite = True result: dict[str, dict[str, Any]] = {} for element in output: try: feed_uri, feed_format = element.rsplit(":", 1) check_valid_format(feed_format) except (ValueError, UsageError): feed_uri = element feed_format = Path(element).suffix.replace(".", "") else: if feed_uri == "-": feed_uri = "stdout:" check_valid_format(feed_format) result[feed_uri] = {"format": feed_format} if overwrite: result[feed_uri]["overwrite"] = True # FEEDS setting should take precedence over the matching CLI options result.update(settings.getdict("FEEDS")) return result
--- +++ @@ -22,6 +22,8 @@ *, convert: Callable[[Any], Any] = update_classpath, ) -> list[Any]: + """Compose a component list from a :ref:`component priority dictionary + <component-priority-dictionaries>`.""" def _check_components(complist: Collection[Any]) -> None: if len({convert(c) for c in complist}) != len(complist): @@ -48,6 +50,7 @@ return {convert(k): v for k, v in compdict.items()} def _validate_values(compdict: Mapping[Any, Any]) -> None: + """Fail if a value in the components dict is not a real number or None.""" for name, value in compdict.items(): if value is not None and not isinstance(value, numbers.Real): raise ValueError( @@ -61,6 +64,9 @@ def arglist_to_dict(arglist: list[str]) -> dict[str, str]: + """Convert a list of arguments like ['arg1=val1', 'arg2=val2', ...] to a + dict + """ return dict(x.split("=", 1) for x in arglist) @@ -68,6 +74,9 @@ path: str | os.PathLike = ".", prevpath: str | os.PathLike | None = None, ) -> str: + """Return the path to the closest scrapy.cfg file by traversing the current + directory and its parents + """ if prevpath is not None and str(path) == str(prevpath): return "" path = Path(path).resolve() @@ -78,6 +87,10 @@ def init_env(project: str = "default", set_syspath: bool = True) -> None: + """Initialize environment to use command-line tool from inside a project + dir. This sets the Scrapy settings module and modifies the Python path to + be able to locate the project module. + """ cfg = get_config() if cfg.has_option("settings", project): os.environ["SCRAPY_SETTINGS_MODULE"] = cfg.get("settings", project) @@ -89,6 +102,7 @@ def get_config(use_closest: bool = True) -> ConfigParser: + """Get Scrapy config file as a ConfigParser""" sources = get_sources(use_closest) cfg = ConfigParser() cfg.read(sources) @@ -133,6 +147,11 @@ *, overwrite_output: list[str] | None = None, ) -> dict[str, dict[str, Any]]: + """ + Receives feed export params (from the 'crawl' or 'runspider' commands), + checks for inconsistencies in their quantities and returns a dictionary + suitable to be used as the FEEDS setting. + """ valid_output_formats: Iterable[str] = without_none_values( cast("dict[str, str]", settings.getwithbase("FEED_EXPORTERS")) ).keys() @@ -174,4 +193,4 @@ # FEEDS setting should take precedence over the matching CLI options result.update(settings.getdict("FEEDS")) - return result+ return result
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/conf.py
Document classes and their methods
from __future__ import annotations import hashlib import json from typing import TYPE_CHECKING, Any, Protocol from urllib.parse import urlunparse from weakref import WeakKeyDictionary from w3lib.url import canonicalize_url from scrapy import Request, Spider from scrapy.utils.httpobj import urlparse_cached from scrapy.utils.misc import load_object from scrapy.utils.python import to_bytes, to_unicode if TYPE_CHECKING: from collections.abc import Iterable # typing.Self requires Python 3.11 from typing_extensions import Self from scrapy.crawler import Crawler _fingerprint_cache: WeakKeyDictionary[ Request, dict[tuple[tuple[bytes, ...] | None, bool], bytes] ] = WeakKeyDictionary() def fingerprint( request: Request, *, include_headers: Iterable[bytes | str] | None = None, keep_fragments: bool = False, ) -> bytes: processed_include_headers: tuple[bytes, ...] | None = None if include_headers: processed_include_headers = tuple( to_bytes(h.lower()) for h in sorted(include_headers) ) cache = _fingerprint_cache.setdefault(request, {}) cache_key = (processed_include_headers, keep_fragments) if cache_key not in cache: # To decode bytes reliably (JSON does not support bytes), regardless of # character encoding, we use bytes.hex() headers: dict[str, list[str]] = {} if processed_include_headers: for header in processed_include_headers: if header in request.headers: headers[header.hex()] = [ header_value.hex() for header_value in request.headers.getlist(header) ] fingerprint_data = { "method": to_unicode(request.method), "url": canonicalize_url(request.url, keep_fragments=keep_fragments), "body": (request.body or b"").hex(), "headers": headers, } fingerprint_json = json.dumps(fingerprint_data, sort_keys=True) cache[cache_key] = hashlib.sha1( # noqa: S324 fingerprint_json.encode() ).digest() return cache[cache_key] class RequestFingerprinterProtocol(Protocol): def fingerprint(self, request: Request) -> bytes: ... class RequestFingerprinter: @classmethod def from_crawler(cls, crawler: Crawler) -> Self: return cls(crawler) def __init__(self, crawler: Crawler | None = None): self._fingerprint = fingerprint def fingerprint(self, request: Request) -> bytes: return self._fingerprint(request) def request_httprepr(request: Request) -> bytes: parsed = urlparse_cached(request) path = urlunparse(("", "", parsed.path or "/", parsed.params, parsed.query, "")) s = to_bytes(request.method) + b" " + to_bytes(path) + b" HTTP/1.1\r\n" s += b"Host: " + to_bytes(parsed.hostname or b"") + b"\r\n" if request.headers: s += request.headers.to_string() + b"\r\n" s += b"\r\n" s += request.body return s def referer_str(request: Request) -> str | None: referrer = request.headers.get("Referer") if referrer is None: return referrer return to_unicode(referrer, errors="replace") def request_from_dict(d: dict[str, Any], *, spider: Spider | None = None) -> Request: request_cls: type[Request] = load_object(d["_class"]) if "_class" in d else Request kwargs = {key: value for key, value in d.items() if key in request_cls.attributes} if d.get("callback") and spider: kwargs["callback"] = _get_method(spider, d["callback"]) if d.get("errback") and spider: kwargs["errback"] = _get_method(spider, d["errback"]) return request_cls(**kwargs) def _get_method(obj: Any, name: Any) -> Any: name = str(name) try: return getattr(obj, name) except AttributeError: raise ValueError(f"Method {name!r} not found in: {obj}") def request_to_curl(request: Request) -> str: method = request.method data = f"--data-raw '{request.body.decode('utf-8')}'" if request.body else "" headers = " ".join( f"-H '{k.decode()}: {v[0].decode()}'" for k, v in request.headers.items() ) url = request.url cookies = "" if request.cookies: if isinstance(request.cookies, dict): cookie = "; ".join(f"{k}={v}" for k, v in request.cookies.items()) cookies = f"--cookie '{cookie}'" elif isinstance(request.cookies, list): cookie = "; ".join( f"{next(iter(c.keys()))}={next(iter(c.values()))}" for c in request.cookies ) cookies = f"--cookie '{cookie}'" curl_cmd = f"curl -X {method} {url} {data} {headers} {cookies}".strip() return " ".join(curl_cmd.split())
--- +++ @@ -1,3 +1,7 @@+""" +This module provides some useful functions for working with +scrapy.Request objects +""" from __future__ import annotations @@ -34,6 +38,34 @@ include_headers: Iterable[bytes | str] | None = None, keep_fragments: bool = False, ) -> bytes: + """ + Return the request fingerprint. + + The request fingerprint is a hash that uniquely identifies the resource the + request points to. For example, take the following two urls: + ``http://www.example.com/query?id=111&cat=222``, + ``http://www.example.com/query?cat=222&id=111``. + + Even though those are two different URLs both point to the same resource + and are equivalent (i.e. they should return the same response). + + Another example are cookies used to store session ids. Suppose the + following page is only accessible to authenticated users: + ``http://www.example.com/members/offers.html``. + + Lots of sites use a cookie to store the session id, which adds a random + component to the HTTP Request and thus should be ignored when calculating + the fingerprint. + + For this reason, request headers are ignored by default when calculating + the fingerprint. If you want to include specific headers use the + include_headers argument, which is a list of Request headers to include. + + Also, servers usually ignore fragments in urls when handling requests, + so they are also ignored by default when calculating the fingerprint. + If you want to include them, set the keep_fragments argument to True + (for instance when handling requests with a headless browser). + """ processed_include_headers: tuple[bytes, ...] | None = None if include_headers: processed_include_headers = tuple( @@ -70,6 +102,15 @@ class RequestFingerprinter: + """Default fingerprinter. + + It takes into account a canonical version + (:func:`w3lib.url.canonicalize_url`) of :attr:`request.url + <scrapy.Request.url>` and the values of :attr:`request.method + <scrapy.Request.method>` and :attr:`request.body + <scrapy.Request.body>`. It then generates an `SHA1 + <https://en.wikipedia.org/wiki/SHA-1>`_ hash. + """ @classmethod def from_crawler(cls, crawler: Crawler) -> Self: @@ -83,6 +124,11 @@ def request_httprepr(request: Request) -> bytes: + """Return the raw HTTP representation (as bytes) of the given request. + This is provided only for reference since it's not the actual stream of + bytes that will be send when performing the request (that's controlled + by Twisted). + """ parsed = urlparse_cached(request) path = urlunparse(("", "", parsed.path or "/", parsed.params, parsed.query, "")) s = to_bytes(request.method) + b" " + to_bytes(path) + b" HTTP/1.1\r\n" @@ -95,6 +141,7 @@ def referer_str(request: Request) -> str | None: + """Return Referer HTTP header suitable for logging.""" referrer = request.headers.get("Referer") if referrer is None: return referrer @@ -102,6 +149,11 @@ def request_from_dict(d: dict[str, Any], *, spider: Spider | None = None) -> Request: + """Create a :class:`~scrapy.Request` object from a dict. + + If a spider is given, it will try to resolve the callbacks looking at the + spider for methods with the same name. + """ request_cls: type[Request] = load_object(d["_class"]) if "_class" in d else Request kwargs = {key: value for key, value in d.items() if key in request_cls.attributes} if d.get("callback") and spider: @@ -112,6 +164,7 @@ def _get_method(obj: Any, name: Any) -> Any: + """Helper function for request_from_dict""" name = str(name) try: return getattr(obj, name) @@ -120,6 +173,12 @@ def request_to_curl(request: Request) -> str: + """ + Converts a :class:`~scrapy.Request` object to a curl command. + + :param :class:`~scrapy.Request`: Request object to be converted + :return: string containing the curl command + """ method = request.method data = f"--data-raw '{request.body.decode('utf-8')}'" if request.body else "" @@ -142,4 +201,4 @@ cookies = f"--cookie '{cookie}'" curl_cmd = f"curl -X {method} {url} {data} {headers} {cookies}".strip() - return " ".join(curl_cmd.split())+ return " ".join(curl_cmd.split())
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/request.py
Write reusable docstrings
from __future__ import annotations import asyncio import inspect import warnings from asyncio import Future from collections.abc import Awaitable, Coroutine, Iterable, Iterator from functools import wraps from typing import ( TYPE_CHECKING, Any, Concatenate, Generic, ParamSpec, TypeVar, cast, overload, ) from twisted.internet.defer import Deferred, DeferredList, FirstError, fail, succeed from twisted.internet.task import Cooperator from twisted.python import failure from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.utils.asyncio import is_asyncio_available from scrapy.utils.python import global_object_name if TYPE_CHECKING: from collections.abc import AsyncIterator, Callable from twisted.python.failure import Failure _T = TypeVar("_T") _T2 = TypeVar("_T2") _P = ParamSpec("_P") _DEFER_DELAY = 0.1 def defer_fail(_failure: Failure) -> Deferred[Any]: # pragma: no cover warnings.warn( "scrapy.utils.defer.defer_fail() is deprecated, use" " twisted.internet.defer.fail(), plus an explicit sleep if needed.", category=ScrapyDeprecationWarning, stacklevel=2, ) from twisted.internet import reactor d: Deferred[Any] = Deferred() reactor.callLater(_DEFER_DELAY, d.errback, _failure) return d def defer_succeed(result: _T) -> Deferred[_T]: # pragma: no cover warnings.warn( "scrapy.utils.defer.defer_succeed() is deprecated, use" " twisted.internet.defer.succeed(), plus an explicit sleep if needed.", category=ScrapyDeprecationWarning, stacklevel=2, ) from twisted.internet import reactor d: Deferred[_T] = Deferred() reactor.callLater(_DEFER_DELAY, d.callback, result) return d async def _defer_sleep_async() -> None: if is_asyncio_available(): await asyncio.sleep(_DEFER_DELAY) else: from twisted.internet import reactor d: Deferred[None] = Deferred() reactor.callLater(_DEFER_DELAY, d.callback, None) await d def defer_result(result: Any) -> Deferred[Any]: # pragma: no cover warnings.warn( "scrapy.utils.defer.defer_result() is deprecated, use" " twisted.internet.defer.success() and twisted.internet.defer.fail()," " plus an explicit sleep if needed, or explicit reactor.callLater().", category=ScrapyDeprecationWarning, stacklevel=2, ) if isinstance(result, Deferred): return result from twisted.internet import reactor d: Deferred[Any] = Deferred() if isinstance(result, failure.Failure): reactor.callLater(_DEFER_DELAY, d.errback, result) else: reactor.callLater(_DEFER_DELAY, d.callback, result) return d @overload def mustbe_deferred( f: Callable[_P, Deferred[_T]], *args: _P.args, **kw: _P.kwargs ) -> Deferred[_T]: ... @overload def mustbe_deferred( f: Callable[_P, _T], *args: _P.args, **kw: _P.kwargs ) -> Deferred[_T]: ... def mustbe_deferred( f: Callable[_P, Deferred[_T] | _T], *args: _P.args, **kw: _P.kwargs, ) -> Deferred[_T]: # pragma: no cover warnings.warn( "scrapy.utils.defer.mustbe_deferred() is deprecated, use" " twisted.internet.defer.maybeDeferred(), with an explicit sleep if needed.", category=ScrapyDeprecationWarning, stacklevel=2, ) result: _T | Deferred[_T] | Failure try: result = f(*args, **kw) except Exception: result = failure.Failure() return defer_result(result) def parallel( iterable: Iterable[_T], count: int, callable: Callable[Concatenate[_T, _P], _T2], # noqa: A002 *args: _P.args, **named: _P.kwargs, ) -> Deferred[list[tuple[bool, Iterator[_T2]]]]: coop = Cooperator() work: Iterator[_T2] = (callable(elem, *args, **named) for elem in iterable) return DeferredList([coop.coiterate(work) for _ in range(count)]) class _AsyncCooperatorAdapter(Iterator, Generic[_T]): def __init__( self, aiterable: AsyncIterator[_T], callable_: Callable[Concatenate[_T, _P], Deferred[Any] | None], *callable_args: _P.args, **callable_kwargs: _P.kwargs, ): self.aiterator: AsyncIterator[_T] = aiterable.__aiter__() self.callable: Callable[Concatenate[_T, _P], Deferred[Any] | None] = callable_ self.callable_args: tuple[Any, ...] = callable_args self.callable_kwargs: dict[str, Any] = callable_kwargs self.finished: bool = False self.waiting_deferreds: list[Deferred[Any]] = [] self.anext_deferred: Deferred[_T] | None = None def _callback(self, result: _T) -> None: # This gets called when the result from aiterator.__anext__() is available. # It calls the callable on it and sends the result to the oldest waiting Deferred # (by chaining if the result is a Deferred too or by firing if not). self.anext_deferred = None callable_result = self.callable( result, *self.callable_args, **self.callable_kwargs ) d = self.waiting_deferreds.pop(0) if isinstance(callable_result, Deferred): callable_result.chainDeferred(d) else: d.callback(None) if self.waiting_deferreds: self._call_anext() def _errback(self, failure: Failure) -> None: # This gets called on any exceptions in aiterator.__anext__(). # It handles StopAsyncIteration by stopping the iteration and reraises all others. self.anext_deferred = None failure.trap(StopAsyncIteration) self.finished = True for d in self.waiting_deferreds: d.callback(None) def _call_anext(self) -> None: # This starts waiting for the next result from aiterator. # If aiterator is exhausted, _errback will be called. self.anext_deferred = deferred_from_coro(self.aiterator.__anext__()) self.anext_deferred.addCallbacks(self._callback, self._errback) def __next__(self) -> Deferred[Any]: # This puts a new Deferred into self.waiting_deferreds and returns it. # It also calls __anext__() if needed. if self.finished: raise StopIteration d: Deferred[Any] = Deferred() self.waiting_deferreds.append(d) if not self.anext_deferred: self._call_anext() return d def parallel_async( async_iterable: AsyncIterator[_T], count: int, callable: Callable[Concatenate[_T, _P], Deferred[Any] | None], # noqa: A002 *args: _P.args, **named: _P.kwargs, ) -> Deferred[list[tuple[bool, Iterator[Deferred[Any]]]]]: coop = Cooperator() work: Iterator[Deferred[Any]] = _AsyncCooperatorAdapter( async_iterable, callable, *args, **named ) dl: Deferred[list[tuple[bool, Iterator[Deferred[Any]]]]] = DeferredList( [coop.coiterate(work) for _ in range(count)] ) return dl def process_chain( callbacks: Iterable[Callable[Concatenate[_T, _P], _T]], input: _T, # noqa: A002 *a: _P.args, **kw: _P.kwargs, ) -> Deferred[_T]: # pragma: no cover warnings.warn( "process_chain() is deprecated.", category=ScrapyDeprecationWarning, stacklevel=2, ) d: Deferred[_T] = Deferred() for x in callbacks: d.addCallback(x, *a, **kw) d.callback(input) return d def process_parallel( callbacks: Iterable[Callable[Concatenate[_T, _P], _T2]], input: _T, # noqa: A002 *a: _P.args, **kw: _P.kwargs, ) -> Deferred[list[_T2]]: # pragma: no cover warnings.warn( "process_parallel() is deprecated.", category=ScrapyDeprecationWarning, stacklevel=2, ) dfds = [succeed(input).addCallback(x, *a, **kw) for x in callbacks] d: Deferred[list[tuple[bool, _T2]]] = DeferredList( dfds, fireOnOneErrback=True, consumeErrors=True ) d2: Deferred[list[_T2]] = d.addCallback(lambda r: [x[1] for x in r]) def eb(failure: Failure) -> Failure: assert isinstance(failure.value, FirstError) return failure.value.subFailure d2.addErrback(eb) return d2 def iter_errback( iterable: Iterable[_T], errback: Callable[Concatenate[Failure, _P], Any], *a: _P.args, **kw: _P.kwargs, ) -> Iterable[_T]: it = iter(iterable) while True: try: yield next(it) except StopIteration: break except Exception: errback(failure.Failure(), *a, **kw) async def aiter_errback( aiterable: AsyncIterator[_T], errback: Callable[Concatenate[Failure, _P], Any], *a: _P.args, **kw: _P.kwargs, ) -> AsyncIterator[_T]: it = aiterable.__aiter__() while True: try: yield await it.__anext__() except StopAsyncIteration: break except Exception: errback(failure.Failure(), *a, **kw) @overload def deferred_from_coro(o: Awaitable[_T]) -> Deferred[_T]: ... @overload def deferred_from_coro(o: _T2) -> _T2: ... def deferred_from_coro(o: Awaitable[_T] | _T2) -> Deferred[_T] | _T2: if isinstance(o, Deferred): return o if inspect.isawaitable(o): if not is_asyncio_available(): # wrapping the coroutine directly into a Deferred, this doesn't work correctly with coroutines # that use asyncio, e.g. "await asyncio.sleep(1)" return Deferred.fromCoroutine(cast("Coroutine[Deferred[Any], Any, _T]", o)) # wrapping the coroutine into a Future and then into a Deferred, this requires AsyncioSelectorReactor return Deferred.fromFuture(asyncio.ensure_future(o)) return o def deferred_f_from_coro_f( coro_f: Callable[_P, Awaitable[_T]], ) -> Callable[_P, Deferred[_T]]: @wraps(coro_f) def f(*coro_args: _P.args, **coro_kwargs: _P.kwargs) -> Deferred[_T]: return deferred_from_coro(coro_f(*coro_args, **coro_kwargs)) return f def maybeDeferred_coro( f: Callable[_P, Any], *args: _P.args, **kw: _P.kwargs ) -> Deferred[Any]: # pragma: no cover warnings.warn( "maybeDeferred_coro() is deprecated and will be removed in a future Scrapy version.", ScrapyDeprecationWarning, stacklevel=2, ) return _maybeDeferred_coro(f, False, *args, **kw) def _maybeDeferred_coro( f: Callable[_P, Any], warn: bool, *args: _P.args, **kw: _P.kwargs ) -> Deferred[Any]: try: result = f(*args, **kw) except: # noqa: E722 return fail(failure.Failure(captureVars=Deferred.debug)) # when the deprecation period has ended we need to make sure the behavior # of the public maybeDeferred_coro() function isn't changed, or drop it in # the same release if isinstance(result, Deferred): if warn: warnings.warn( f"{global_object_name(f)} returned a Deferred, this is deprecated." f" Please refactor this function to return a coroutine.", ScrapyDeprecationWarning, stacklevel=2, ) return result if asyncio.isfuture(result) or inspect.isawaitable(result): return deferred_from_coro(result) if isinstance(result, failure.Failure): # pragma: no cover if warn: warnings.warn( f"{global_object_name(f)} returned a Failure, this is deprecated." f" Please refactor this function to return a coroutine.", ScrapyDeprecationWarning, stacklevel=2, ) return fail(result) return succeed(result) def deferred_to_future(d: Deferred[_T]) -> Future[_T]: if not is_asyncio_available(): raise RuntimeError("deferred_to_future() requires AsyncioSelectorReactor.") return d.asFuture(asyncio.get_event_loop()) def maybe_deferred_to_future(d: Deferred[_T]) -> Deferred[_T] | Future[_T]: if not is_asyncio_available(): return d return deferred_to_future(d) def _schedule_coro(coro: Coroutine[Any, Any, Any]) -> None: if not is_asyncio_available(): Deferred.fromCoroutine(coro) return loop = asyncio.get_event_loop() loop.create_task(coro) # noqa: RUF006 @overload def ensure_awaitable(o: Awaitable[_T], _warn: str | None = None) -> Awaitable[_T]: ... @overload def ensure_awaitable(o: _T, _warn: str | None = None) -> Awaitable[_T]: ... def ensure_awaitable(o: _T | Awaitable[_T], _warn: str | None = None) -> Awaitable[_T]: if isinstance(o, Deferred): if _warn: warnings.warn( f"{_warn} returned a Deferred, this is deprecated." f" Please refactor this function to return a coroutine.", ScrapyDeprecationWarning, stacklevel=2, ) return maybe_deferred_to_future(o) if inspect.isawaitable(o): return o async def coro() -> _T: return o return coro()
--- +++ @@ -1,3 +1,6 @@+""" +Helper functions for dealing with Twisted deferreds +""" from __future__ import annotations @@ -41,6 +44,12 @@ def defer_fail(_failure: Failure) -> Deferred[Any]: # pragma: no cover + """Same as twisted.internet.defer.fail but delay calling errback until + next reactor loop + + It delays by 100ms so reactor has a chance to go through readers and writers + before attending pending delayed calls, so do not set delay to zero. + """ warnings.warn( "scrapy.utils.defer.defer_fail() is deprecated, use" " twisted.internet.defer.fail(), plus an explicit sleep if needed.", @@ -56,6 +65,12 @@ def defer_succeed(result: _T) -> Deferred[_T]: # pragma: no cover + """Same as twisted.internet.defer.succeed but delay calling callback until + next reactor loop + + It delays by 100ms so reactor has a chance to go through readers and writers + before attending pending delayed calls, so do not set delay to zero. + """ warnings.warn( "scrapy.utils.defer.defer_succeed() is deprecated, use" " twisted.internet.defer.succeed(), plus an explicit sleep if needed.", @@ -71,6 +86,9 @@ async def _defer_sleep_async() -> None: + """Delay by _DEFER_DELAY so reactor has a chance to go through readers and writers + before attending pending delayed calls, so do not set delay to zero. + """ if is_asyncio_available(): await asyncio.sleep(_DEFER_DELAY) else: @@ -120,6 +138,9 @@ *args: _P.args, **kw: _P.kwargs, ) -> Deferred[_T]: # pragma: no cover + """Same as twisted.internet.defer.maybeDeferred, but delay calling + callback/errback to next reactor loop + """ warnings.warn( "scrapy.utils.defer.mustbe_deferred() is deprecated, use" " twisted.internet.defer.maybeDeferred(), with an explicit sleep if needed.", @@ -141,12 +162,61 @@ *args: _P.args, **named: _P.kwargs, ) -> Deferred[list[tuple[bool, Iterator[_T2]]]]: + """Execute a callable over the objects in the given iterable, in parallel, + using no more than ``count`` concurrent calls. + + Taken from: https://jcalderone.livejournal.com/24285.html + """ coop = Cooperator() work: Iterator[_T2] = (callable(elem, *args, **named) for elem in iterable) return DeferredList([coop.coiterate(work) for _ in range(count)]) class _AsyncCooperatorAdapter(Iterator, Generic[_T]): + """A class that wraps an async iterable into a normal iterator suitable + for using in Cooperator.coiterate(). As it's only needed for parallel_async(), + it calls the callable directly in the callback, instead of providing a more + generic interface. + + On the outside, this class behaves as an iterator that yields Deferreds. + Each Deferred is fired with the result of the callable which was called on + the next result from aiterator. It raises StopIteration when aiterator is + exhausted, as expected. + + Cooperator calls __next__() multiple times and waits on the Deferreds + returned from it. As async generators (since Python 3.8) don't support + awaiting on __anext__() several times in parallel, we need to serialize + this. It's done by storing the Deferreds returned from __next__() and + firing the oldest one when a result from __anext__() is available. + + The workflow: + 1. When __next__() is called for the first time, it creates a Deferred, stores it + in self.waiting_deferreds and returns it. It also makes a Deferred that will wait + for self.aiterator.__anext__() and puts it into self.anext_deferred. + 2. If __next__() is called again before self.anext_deferred fires, more Deferreds + are added to self.waiting_deferreds. + 3. When self.anext_deferred fires, it either calls _callback() or _errback(). Both + clear self.anext_deferred. + 3.1. _callback() calls the callable passing the result value that it takes, pops a + Deferred from self.waiting_deferreds, and if the callable result was a Deferred, it + chains those Deferreds so that the waiting Deferred will fire when the result + Deferred does, otherwise it fires it directly. This causes one awaiting task to + receive a result. If self.waiting_deferreds is still not empty, new __anext__() is + called and self.anext_deferred is populated. + 3.2. _errback() checks the exception class. If it's StopAsyncIteration it means + self.aiterator is exhausted and so it sets self.finished and fires all + self.waiting_deferreds. Other exceptions are propagated. + 4. If __next__() is called after __anext__() was handled, then if self.finished is + True, it raises StopIteration, otherwise it acts like in step 2, but if + self.anext_deferred is now empty is also populates it with a new __anext__(). + + Note that CooperativeTask ignores the value returned from the Deferred that it waits + for, so we fire them with None when needed. + + It may be possible to write an async iterator-aware replacement for + Cooperator/CooperativeTask and use it instead of this adapter to achieve the same + goal. + """ def __init__( self, @@ -213,6 +283,7 @@ *args: _P.args, **named: _P.kwargs, ) -> Deferred[list[tuple[bool, Iterator[Deferred[Any]]]]]: + """Like ``parallel`` but for async iterators""" coop = Cooperator() work: Iterator[Deferred[Any]] = _AsyncCooperatorAdapter( async_iterable, callable, *args, **named @@ -229,6 +300,7 @@ *a: _P.args, **kw: _P.kwargs, ) -> Deferred[_T]: # pragma: no cover + """Return a Deferred built by chaining the given callbacks""" warnings.warn( "process_chain() is deprecated.", category=ScrapyDeprecationWarning, @@ -247,6 +319,9 @@ *a: _P.args, **kw: _P.kwargs, ) -> Deferred[list[_T2]]: # pragma: no cover + """Return a Deferred with the output of all successful calls to the given + callbacks + """ warnings.warn( "process_parallel() is deprecated.", category=ScrapyDeprecationWarning, @@ -272,6 +347,9 @@ *a: _P.args, **kw: _P.kwargs, ) -> Iterable[_T]: + """Wrap an iterable calling an errback if an error is caught while + iterating it. + """ it = iter(iterable) while True: try: @@ -288,6 +366,9 @@ *a: _P.args, **kw: _P.kwargs, ) -> AsyncIterator[_T]: + """Wrap an async iterable calling an errback if an error is caught while + iterating it. Similar to :func:`scrapy.utils.defer.iter_errback`. + """ it = aiterable.__aiter__() while True: try: @@ -307,6 +388,8 @@ def deferred_from_coro(o: Awaitable[_T] | _T2) -> Deferred[_T] | _T2: + """Convert a coroutine or other awaitable object into a Deferred, + or return the object as is if it isn't a coroutine.""" if isinstance(o, Deferred): return o if inspect.isawaitable(o): @@ -322,6 +405,11 @@ def deferred_f_from_coro_f( coro_f: Callable[_P, Awaitable[_T]], ) -> Callable[_P, Deferred[_T]]: + """Convert a coroutine function into a function that returns a Deferred. + + The coroutine function will be called at the time when the wrapper is called. Wrapper args will be passed to it. + This is useful for callback chains, as callback functions are called with the previous callback result. + """ @wraps(coro_f) def f(*coro_args: _P.args, **coro_kwargs: _P.kwargs) -> Deferred[_T]: @@ -333,6 +421,7 @@ def maybeDeferred_coro( f: Callable[_P, Any], *args: _P.args, **kw: _P.kwargs ) -> Deferred[Any]: # pragma: no cover + """Copy of defer.maybeDeferred that also converts coroutines to Deferreds.""" warnings.warn( "maybeDeferred_coro() is deprecated and will be removed in a future Scrapy version.", ScrapyDeprecationWarning, @@ -344,6 +433,7 @@ def _maybeDeferred_coro( f: Callable[_P, Any], warn: bool, *args: _P.args, **kw: _P.kwargs ) -> Deferred[Any]: + """Copy of defer.maybeDeferred that also converts coroutines to Deferreds.""" try: result = f(*args, **kw) except: # noqa: E722 @@ -376,18 +466,71 @@ def deferred_to_future(d: Deferred[_T]) -> Future[_T]: + """Return an :class:`asyncio.Future` object that wraps *d*. + + This function requires + :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor` to be + installed. + + When :ref:`using the asyncio reactor <install-asyncio>`, you cannot await + on :class:`~twisted.internet.defer.Deferred` objects from :ref:`Scrapy + callables defined as coroutines <coroutine-support>`, you can only await on + ``Future`` objects. Wrapping ``Deferred`` objects into ``Future`` objects + allows you to wait on them:: + + class MySpider(Spider): + ... + async def parse(self, response): + additional_request = scrapy.Request('https://example.org/price') + deferred = self.crawler.engine.download(additional_request) + additional_response = await deferred_to_future(deferred) + + .. versionchanged:: 2.14 + This function no longer installs an asyncio loop if called before the + Twisted asyncio reactor is installed. A :exc:`RuntimeError` is raised + in this case. + """ if not is_asyncio_available(): raise RuntimeError("deferred_to_future() requires AsyncioSelectorReactor.") return d.asFuture(asyncio.get_event_loop()) def maybe_deferred_to_future(d: Deferred[_T]) -> Deferred[_T] | Future[_T]: + """Return *d* as an object that can be awaited from a :ref:`Scrapy callable + defined as a coroutine <coroutine-support>`. + + What you can await in Scrapy callables defined as coroutines depends on the + value of :setting:`TWISTED_REACTOR`: + + - When :ref:`using the asyncio reactor <install-asyncio>`, you can only + await on :class:`asyncio.Future` objects. + + - When not using the asyncio reactor, you can only await on + :class:`~twisted.internet.defer.Deferred` objects. + + If you want to write code that uses ``Deferred`` objects but works with any + reactor, use this function on all ``Deferred`` objects:: + + class MySpider(Spider): + ... + async def parse(self, response): + additional_request = scrapy.Request('https://example.org/price') + deferred = self.crawler.engine.download(additional_request) + additional_response = await maybe_deferred_to_future(deferred) + """ if not is_asyncio_available(): return d return deferred_to_future(d) def _schedule_coro(coro: Coroutine[Any, Any, Any]) -> None: + """Schedule the coroutine as a task or a Deferred. + + This doesn't store the reference to the task/Deferred, so a better + alternative is calling :func:`scrapy.utils.defer.deferred_from_coro`, + keeping the result, and adding proper exception handling (e.g. errbacks) to + it. + """ if not is_asyncio_available(): Deferred.fromCoroutine(coro) return @@ -404,6 +547,15 @@ def ensure_awaitable(o: _T | Awaitable[_T], _warn: str | None = None) -> Awaitable[_T]: + """Convert any value to an awaitable object. + + For a :class:`~twisted.internet.defer.Deferred` object, use + :func:`maybe_deferred_to_future` to wrap it into a suitable object. For an + awaitable object of a different type, return it as is. For any other + value, return a coroutine that completes with that value. + + .. versionadded:: 2.14 + """ if isinstance(o, Deferred): if _warn: warnings.warn( @@ -419,4 +571,4 @@ async def coro() -> _T: return o - return coro()+ return coro()
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/defer.py
Generate helpful docstrings for debugging
from __future__ import annotations import os import re import tempfile import webbrowser from typing import TYPE_CHECKING, Any from weakref import WeakKeyDictionary from twisted.web import http from w3lib import html from scrapy.utils.python import to_bytes, to_unicode if TYPE_CHECKING: from collections.abc import Callable, Iterable from scrapy.http import Response, TextResponse _baseurl_cache: WeakKeyDictionary[Response, str] = WeakKeyDictionary() def get_base_url(response: TextResponse) -> str: if response not in _baseurl_cache: text = response.text[0:4096] _baseurl_cache[response] = html.get_base_url( text, response.url, response.encoding ) return _baseurl_cache[response] _metaref_cache: WeakKeyDictionary[Response, tuple[None, None] | tuple[float, str]] = ( WeakKeyDictionary() ) def get_meta_refresh( response: TextResponse, ignore_tags: Iterable[str] = ("script", "noscript"), ) -> tuple[None, None] | tuple[float, str]: if response not in _metaref_cache: text = response.text[0:4096] _metaref_cache[response] = html.get_meta_refresh( text, get_base_url(response), response.encoding, ignore_tags=ignore_tags ) return _metaref_cache[response] def response_status_message(status: bytes | float | str) -> str: status_int = int(status) message = http.RESPONSES.get(status_int, "Unknown Status") return f"{status_int} {to_unicode(message)}" def _remove_html_comments(body: bytes) -> bytes: start = body.find(b"<!--") while start != -1: end = body.find(b"-->", start + 1) if end == -1: return body[:start] body = body[:start] + body[end + 3 :] start = body.find(b"<!--") return body def open_in_browser( response: TextResponse, _openfunc: Callable[[str], Any] = webbrowser.open, ) -> Any: # circular imports from scrapy.http import HtmlResponse, TextResponse # noqa: PLC0415 # XXX: this implementation is a bit dirty and could be improved body = response.body if isinstance(response, HtmlResponse): if b"<base" not in body: _remove_html_comments(body) repl = rf'\0<base href="{response.url}">' body = re.sub(rb"<head(?:[^<>]*?>)", to_bytes(repl), body, count=1) ext = ".html" elif isinstance(response, TextResponse): ext = ".txt" else: raise TypeError(f"Unsupported response type: {response.__class__.__name__}") fd, fname = tempfile.mkstemp(ext) os.write(fd, body) os.close(fd) return _openfunc(f"file://{fname}")
--- +++ @@ -1,3 +1,7 @@+""" +This module provides some useful functions for working with +scrapy.http.Response objects +""" from __future__ import annotations @@ -22,6 +26,7 @@ def get_base_url(response: TextResponse) -> str: + """Return the base url of the given response, joined with the response url""" if response not in _baseurl_cache: text = response.text[0:4096] _baseurl_cache[response] = html.get_base_url( @@ -39,6 +44,7 @@ response: TextResponse, ignore_tags: Iterable[str] = ("script", "noscript"), ) -> tuple[None, None] | tuple[float, str]: + """Parse the http-equiv refresh parameter from the given response""" if response not in _metaref_cache: text = response.text[0:4096] _metaref_cache[response] = html.get_meta_refresh( @@ -48,6 +54,7 @@ def response_status_message(status: bytes | float | str) -> str: + """Return status code plus status text descriptive message""" status_int = int(status) message = http.RESPONSES.get(status_int, "Unknown Status") return f"{status_int} {to_unicode(message)}" @@ -68,6 +75,22 @@ response: TextResponse, _openfunc: Callable[[str], Any] = webbrowser.open, ) -> Any: + """Open *response* in a local web browser, adjusting the `base tag`_ for + external links to work, e.g. so that images and styles are displayed. + + .. _base tag: https://www.w3schools.com/tags/tag_base.asp + + For example: + + .. code-block:: python + + from scrapy.utils.response import open_in_browser + + + def parse_details(self, response): + if "item name" not in response.body: + open_in_browser(response) + """ # circular imports from scrapy.http import HtmlResponse, TextResponse # noqa: PLC0415 @@ -86,4 +109,4 @@ fd, fname = tempfile.mkstemp(ext) os.write(fd, body) os.close(fd) - return _openfunc(f"file://{fname}")+ return _openfunc(f"file://{fname}")
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/response.py
Add clean documentation to messy code
from __future__ import annotations import collections import contextlib import warnings import weakref from collections import OrderedDict from collections.abc import Mapping from typing import TYPE_CHECKING, Any, AnyStr, TypeVar from scrapy.exceptions import ScrapyDeprecationWarning if TYPE_CHECKING: from collections.abc import Iterable, Sequence # typing.Self requires Python 3.11 from typing_extensions import Self _KT = TypeVar("_KT") _VT = TypeVar("_VT") class CaselessDict(dict): __slots__ = () def __new__(cls, *args: Any, **kwargs: Any) -> Self: # circular import from scrapy.http.headers import Headers # noqa: PLC0415 if issubclass(cls, CaselessDict) and not issubclass(cls, Headers): warnings.warn( "scrapy.utils.datatypes.CaselessDict is deprecated," " please use scrapy.utils.datatypes.CaseInsensitiveDict instead", category=ScrapyDeprecationWarning, stacklevel=2, ) return super().__new__(cls, *args, **kwargs) def __init__( self, seq: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]] | None = None, ): super().__init__() if seq: self.update(seq) def __getitem__(self, key: AnyStr) -> Any: return dict.__getitem__(self, self.normkey(key)) def __setitem__(self, key: AnyStr, value: Any) -> None: dict.__setitem__(self, self.normkey(key), self.normvalue(value)) def __delitem__(self, key: AnyStr) -> None: dict.__delitem__(self, self.normkey(key)) def __contains__(self, key: AnyStr) -> bool: # type: ignore[override] return dict.__contains__(self, self.normkey(key)) has_key = __contains__ def __copy__(self) -> Self: return self.__class__(self) copy = __copy__ def normkey(self, key: AnyStr) -> AnyStr: return key.lower() def normvalue(self, value: Any) -> Any: return value def get(self, key: AnyStr, def_val: Any = None) -> Any: return dict.get(self, self.normkey(key), self.normvalue(def_val)) def setdefault(self, key: AnyStr, def_val: Any = None) -> Any: return dict.setdefault(self, self.normkey(key), self.normvalue(def_val)) # doesn't fully implement MutableMapping.update() def update(self, seq: Mapping[AnyStr, Any] | Iterable[tuple[AnyStr, Any]]) -> None: # type: ignore[override] seq = seq.items() if isinstance(seq, Mapping) else seq iseq = ((self.normkey(k), self.normvalue(v)) for k, v in seq) super().update(iseq) @classmethod def fromkeys(cls, keys: Iterable[AnyStr], value: Any = None) -> Self: # type: ignore[override] return cls((k, value) for k in keys) # type: ignore[misc] def pop(self, key: AnyStr, *args: Any) -> Any: return dict.pop(self, self.normkey(key), *args) class CaseInsensitiveDict(collections.UserDict): def __init__(self, *args: Any, **kwargs: Any) -> None: self._keys: dict = {} super().__init__(*args, **kwargs) def __getitem__(self, key: AnyStr) -> Any: normalized_key = self._normkey(key) return super().__getitem__(self._keys[normalized_key.lower()]) def __setitem__(self, key: AnyStr, value: Any) -> None: normalized_key = self._normkey(key) try: lower_key = self._keys[normalized_key.lower()] del self[lower_key] except KeyError: pass super().__setitem__(normalized_key, self._normvalue(value)) self._keys[normalized_key.lower()] = normalized_key def __delitem__(self, key: AnyStr) -> None: normalized_key = self._normkey(key) stored_key = self._keys.pop(normalized_key.lower()) super().__delitem__(stored_key) def __contains__(self, key: AnyStr) -> bool: # type: ignore[override] normalized_key = self._normkey(key) return normalized_key.lower() in self._keys def __repr__(self) -> str: return f"<{self.__class__.__name__}: {super().__repr__()}>" def _normkey(self, key: AnyStr) -> AnyStr: return key def _normvalue(self, value: Any) -> Any: return value class LocalCache(OrderedDict[_KT, _VT]): def __init__(self, limit: int | None = None): super().__init__() self.limit: int | None = limit def __setitem__(self, key: _KT, value: _VT) -> None: if self.limit: while len(self) >= self.limit: self.popitem(last=False) super().__setitem__(key, value) class LocalWeakReferencedCache(weakref.WeakKeyDictionary): def __init__(self, limit: int | None = None): super().__init__() self.data: LocalCache = LocalCache(limit=limit) def __setitem__(self, key: _KT, value: _VT) -> None: # if raised, key is not weak-referenceable, skip caching with contextlib.suppress(TypeError): super().__setitem__(key, value) def __getitem__(self, key: _KT) -> _VT | None: try: return super().__getitem__(key) except (TypeError, KeyError): return None # key is either not weak-referenceable or not cached class SequenceExclude: def __init__(self, seq: Sequence[Any]): self.seq: Sequence[Any] = seq def __contains__(self, item: Any) -> bool: return item not in self.seq
--- +++ @@ -1,3 +1,9 @@+""" +This module contains data types used by Scrapy which are not included in the +Python Standard Library. + +This module must not depend on any module outside the Standard Library. +""" from __future__ import annotations @@ -66,9 +72,11 @@ copy = __copy__ def normkey(self, key: AnyStr) -> AnyStr: + """Method to normalize dictionary key access""" return key.lower() def normvalue(self, value: Any) -> Any: + """Method to normalize values prior to be set""" return value def get(self, key: AnyStr, def_val: Any = None) -> Any: @@ -92,6 +100,9 @@ class CaseInsensitiveDict(collections.UserDict): + """A dict-like structure that accepts strings or bytes + as keys and allows case-insensitive lookups. + """ def __init__(self, *args: Any, **kwargs: Any) -> None: self._keys: dict = {} @@ -131,6 +142,10 @@ class LocalCache(OrderedDict[_KT, _VT]): + """Dictionary with a finite number of keys. + + Older items expires first. + """ def __init__(self, limit: int | None = None): super().__init__() @@ -144,6 +159,16 @@ class LocalWeakReferencedCache(weakref.WeakKeyDictionary): + """ + A weakref.WeakKeyDictionary implementation that uses LocalCache as its + underlying data structure, making it ordered and capable of being size-limited. + + Useful for memoization, while avoiding keeping received + arguments in memory only because of the cached references. + + Note: like LocalCache and unlike weakref.WeakKeyDictionary, + it cannot be instantiated with an initial dictionary. + """ def __init__(self, limit: int | None = None): super().__init__() @@ -162,9 +187,10 @@ class SequenceExclude: + """Object to test if an item is NOT within some sequence.""" def __init__(self, seq: Sequence[Any]): self.seq: Sequence[Any] = seq def __contains__(self, item: Any) -> bool: - return item not in self.seq+ return item not in self.seq
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/datatypes.py
Write Python docstrings for this snippet
from __future__ import annotations import contextlib import os import signal from typing import TYPE_CHECKING, Any from itemadapter import is_item from twisted.internet import defer, threads from twisted.python import threadable from w3lib.url import any_to_uri import scrapy from scrapy.crawler import Crawler from scrapy.exceptions import IgnoreRequest from scrapy.http import Request, Response from scrapy.settings import Settings from scrapy.spiders import Spider from scrapy.utils.conf import get_config from scrapy.utils.console import DEFAULT_PYTHON_SHELLS, start_python_console from scrapy.utils.datatypes import SequenceExclude from scrapy.utils.defer import _schedule_coro, deferred_f_from_coro_f from scrapy.utils.misc import load_object from scrapy.utils.reactor import is_asyncio_reactor_installed, set_asyncio_event_loop from scrapy.utils.response import open_in_browser if TYPE_CHECKING: from collections.abc import Callable class Shell: relevant_classes: tuple[type, ...] = (Crawler, Spider, Request, Response, Settings) def __init__( self, crawler: Crawler, update_vars: Callable[[dict[str, Any]], None] | None = None, code: str | None = None, ): self.crawler: Crawler = crawler self.update_vars: Callable[[dict[str, Any]], None] = update_vars or ( lambda x: None ) self.item_class: type = load_object(crawler.settings["DEFAULT_ITEM_CLASS"]) self.spider: Spider | None = None self.inthread: bool = not threadable.isInIOThread() self.code: str | None = code self.vars: dict[str, Any] = {} def start( self, url: str | None = None, request: Request | None = None, response: Response | None = None, spider: Spider | None = None, redirect: bool = True, ) -> None: # disable accidental Ctrl-C key press from shutting down the engine signal.signal(signal.SIGINT, signal.SIG_IGN) if url: self.fetch(url, spider, redirect=redirect) elif request: self.fetch(request, spider) elif response: request = response.request self.populate_vars(response, request, spider) else: self.populate_vars() if self.code: print(eval(self.code, globals(), self.vars)) # noqa: S307 else: # Detect interactive shell setting in scrapy.cfg # e.g.: ~/.config/scrapy.cfg or ~/.scrapy.cfg # [settings] # # shell can be one of ipython, bpython or python; # # to be used as the interactive python console, if available. # # (default is ipython, fallbacks in the order listed above) # shell = python cfg = get_config() section, option = "settings", "shell" env = os.environ.get("SCRAPY_PYTHON_SHELL") shells = [] if env: shells += env.strip().lower().split(",") elif cfg.has_option(section, option): shells += [cfg.get(section, option).strip().lower()] else: # try all by default shells += DEFAULT_PYTHON_SHELLS.keys() # always add standard shell as fallback shells += ["python"] start_python_console( self.vars, shells=shells, banner=self.vars.pop("banner", "") ) def _schedule(self, request: Request, spider: Spider | None) -> defer.Deferred[Any]: if is_asyncio_reactor_installed(): # set the asyncio event loop for the current thread event_loop_path = self.crawler.settings["ASYNCIO_EVENT_LOOP"] set_asyncio_event_loop(event_loop_path) def crawl_request(_: None) -> None: assert self.crawler.engine is not None self.crawler.engine.crawl(request) d2 = self._open_spider(request, spider) d2.addCallback(crawl_request) d = _request_deferred(request) d.addCallback(lambda x: (x, spider)) return d @deferred_f_from_coro_f async def _open_spider(self, request: Request, spider: Spider | None) -> None: if self.spider: return if spider is None: spider = self.crawler.spider or self.crawler._create_spider() self.crawler.spider = spider assert self.crawler.engine await self.crawler.engine.open_spider_async(close_if_idle=False) _schedule_coro(self.crawler.engine._start_request_processing()) self.spider = spider def fetch( self, request_or_url: Request | str, spider: Spider | None = None, redirect: bool = True, **kwargs: Any, ) -> None: from twisted.internet import reactor if isinstance(request_or_url, Request): request = request_or_url else: url = any_to_uri(request_or_url) request = Request(url, dont_filter=True, **kwargs) if redirect: request.meta["handle_httpstatus_list"] = SequenceExclude( range(300, 400) ) else: request.meta["handle_httpstatus_all"] = True response = None with contextlib.suppress(IgnoreRequest): response, spider = threads.blockingCallFromThread( reactor, self._schedule, request, spider ) self.populate_vars(response, request, spider) def populate_vars( self, response: Response | None = None, request: Request | None = None, spider: Spider | None = None, ) -> None: self.vars["scrapy"] = scrapy self.vars["crawler"] = self.crawler self.vars["item"] = self.item_class() self.vars["settings"] = self.crawler.settings self.vars["spider"] = spider self.vars["request"] = request self.vars["response"] = response if self.inthread: self.vars["fetch"] = self.fetch self.vars["view"] = open_in_browser self.vars["shelp"] = self.print_help self.update_vars(self.vars) if not self.code: self.vars["banner"] = self.get_help() def print_help(self) -> None: print(self.get_help()) def get_help(self) -> str: b = [] b.append("Available Scrapy objects:") b.append( " scrapy scrapy module (contains scrapy.Request, scrapy.Selector, etc)" ) for k, v in sorted(self.vars.items()): if self._is_relevant(v): b.append(f" {k:<10} {v}") b.append("Useful shortcuts:") if self.inthread: b.append( " fetch(url[, redirect=True]) " "Fetch URL and update local objects (by default, redirects are followed)" ) b.append( " fetch(req) " "Fetch a scrapy.Request and update local objects " ) b.append(" shelp() Shell help (print this help)") b.append(" view(response) View response in a browser") return "\n".join(f"[s] {line}" for line in b) + "\n" def _is_relevant(self, value: Any) -> bool: return isinstance(value, self.relevant_classes) or is_item(value) def inspect_response(response: Response, spider: Spider) -> None: # Shell.start removes the SIGINT handler, so save it and re-add it after # the shell has closed sigint_handler = signal.getsignal(signal.SIGINT) Shell(spider.crawler).start(response=response, spider=spider) signal.signal(signal.SIGINT, sigint_handler) def _request_deferred(request: Request) -> defer.Deferred[Any]: request_callback = request.callback request_errback = request.errback def _restore_callbacks(result: Any) -> Any: request.callback = request_callback request.errback = request_errback return result d: defer.Deferred[Any] = defer.Deferred() d.addBoth(_restore_callbacks) if request.callback: d.addCallback(request.callback) if request.errback: d.addErrback(request.errback) request.callback, request.errback = d.callback, d.errback return d
--- +++ @@ -1,3 +1,8 @@+"""Scrapy Shell + +See documentation in docs/topics/shell.rst + +""" from __future__ import annotations @@ -204,6 +209,7 @@ def inspect_response(response: Response, spider: Spider) -> None: + """Open a shell to inspect the given response""" # Shell.start removes the SIGINT handler, so save it and re-add it after # the shell has closed sigint_handler = signal.getsignal(signal.SIGINT) @@ -212,6 +218,16 @@ def _request_deferred(request: Request) -> defer.Deferred[Any]: + """Wrap a request inside a Deferred. + + This function is harmful, do not use it until you know what you are doing. + + This returns a Deferred whose first pair of callbacks are the request + callback and errback. The Deferred also triggers when the request + callback/errback is executed (i.e. when the request is downloaded) + + WARNING: Do not call request.replace() until after the deferred is called. + """ request_callback = request.callback request_errback = request.errback @@ -228,4 +244,4 @@ d.addErrback(request.errback) request.callback, request.errback = d.callback, d.errback - return d+ return d
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/shell.py
Add docstrings to clarify complex logic
from __future__ import annotations import inspect import warnings from typing import TYPE_CHECKING, Any, overload from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.utils.python import get_func_args_dict if TYPE_CHECKING: from collections.abc import Callable def attribute(obj: Any, oldattr: str, newattr: str, version: str = "0.12") -> None: cname = obj.__class__.__name__ warnings.warn( f"{cname}.{oldattr} attribute is deprecated and will be no longer supported " f"in Scrapy {version}, use {cname}.{newattr} attribute instead", ScrapyDeprecationWarning, stacklevel=3, ) def create_deprecated_class( name: str, new_class: type, clsdict: dict[str, Any] | None = None, warn_category: type[Warning] = ScrapyDeprecationWarning, warn_once: bool = True, old_class_path: str | None = None, new_class_path: str | None = None, subclass_warn_message: str = "{cls} inherits from deprecated class {old}, please inherit from {new}.", instance_warn_message: str = "{cls} is deprecated, instantiate {new} instead.", ) -> type: # https://github.com/python/mypy/issues/4177 class DeprecatedClass(new_class.__class__): # type: ignore[misc,name-defined] # pylint: disable=no-self-argument deprecated_class: type | None = None warned_on_subclass: bool = False def __new__( # pylint: disable=bad-classmethod-argument metacls, name: str, bases: tuple[type, ...], clsdict_: dict[str, Any] ) -> type: cls = super().__new__(metacls, name, bases, clsdict_) if metacls.deprecated_class is None: metacls.deprecated_class = cls return cls def __init__(cls, name: str, bases: tuple[type, ...], clsdict_: dict[str, Any]): meta = cls.__class__ old = meta.deprecated_class if old in bases and not (warn_once and meta.warned_on_subclass): meta.warned_on_subclass = True msg = subclass_warn_message.format( cls=_clspath(cls), old=_clspath(old, old_class_path), new=_clspath(new_class, new_class_path), ) if warn_once: msg += " (warning only on first subclass, there may be others)" warnings.warn(msg, warn_category, stacklevel=2) super().__init__(name, bases, clsdict_) # see https://www.python.org/dev/peps/pep-3119/#overloading-isinstance-and-issubclass # and https://docs.python.org/reference/datamodel.html#customizing-instance-and-subclass-checks # for implementation details def __instancecheck__(cls, inst: Any) -> bool: return any(cls.__subclasscheck__(c) for c in (type(inst), inst.__class__)) def __subclasscheck__(cls, sub: type) -> bool: if cls is not DeprecatedClass.deprecated_class: # we should do the magic only if second `issubclass` argument # is the deprecated class itself - subclasses of the # deprecated class should not use custom `__subclasscheck__` # method. return super().__subclasscheck__(sub) if not inspect.isclass(sub): raise TypeError("issubclass() arg 1 must be a class") mro = getattr(sub, "__mro__", ()) return any(c in {cls, new_class} for c in mro) def __call__(cls, *args: Any, **kwargs: Any) -> Any: old = DeprecatedClass.deprecated_class if cls is old: msg = instance_warn_message.format( cls=_clspath(cls, old_class_path), new=_clspath(new_class, new_class_path), ) warnings.warn(msg, warn_category, stacklevel=2) return super().__call__(*args, **kwargs) deprecated_cls = DeprecatedClass(name, (new_class,), clsdict or {}) try: frm = inspect.stack()[1] parent_module = inspect.getmodule(frm[0]) if parent_module is not None: deprecated_cls.__module__ = parent_module.__name__ except Exception as e: # Sometimes inspect.stack() fails (e.g. when the first import of # deprecated class is in jinja2 template). __module__ attribute is not # important enough to raise an exception as users may be unable # to fix inspect.stack() errors. warnings.warn(f"Error detecting parent module: {e!r}") return deprecated_cls def _clspath(cls: type, forced: str | None = None) -> str: if forced is not None: return forced return f"{cls.__module__}.{cls.__name__}" DEPRECATION_RULES: list[tuple[str, str]] = [] @overload def update_classpath(path: str) -> str: ... @overload def update_classpath(path: Any) -> Any: ... def update_classpath(path: Any) -> Any: for prefix, replacement in DEPRECATION_RULES: if isinstance(path, str) and path.startswith(prefix): new_path = path.replace(prefix, replacement, 1) warnings.warn( f"`{path}` class is deprecated, use `{new_path}` instead", ScrapyDeprecationWarning, ) return new_path return path def method_is_overridden(subclass: type, base_class: type, method_name: str) -> bool: base_method = getattr(base_class, method_name) sub_method = getattr(subclass, method_name) return base_method.__code__ is not sub_method.__code__ def argument_is_required(func: Callable[..., Any], arg_name: str) -> bool: args = get_func_args_dict(func) param = args.get(arg_name) return param is not None and param.default is inspect.Parameter.empty def warn_on_deprecated_spider_attribute(attribute_name: str, setting_name: str) -> None: warnings.warn( f"The '{attribute_name}' spider attribute is deprecated. " "Use Spider.custom_settings or Spider.update_settings() instead. " f"The corresponding setting name is '{setting_name}'.", category=ScrapyDeprecationWarning, stacklevel=2, )
--- +++ @@ -1,3 +1,4 @@+"""Some helpers for deprecation messages""" from __future__ import annotations @@ -33,6 +34,30 @@ subclass_warn_message: str = "{cls} inherits from deprecated class {old}, please inherit from {new}.", instance_warn_message: str = "{cls} is deprecated, instantiate {new} instead.", ) -> type: + """ + Return a "deprecated" class that causes its subclasses to issue a warning. + Subclasses of ``new_class`` are considered subclasses of this class. + It also warns when the deprecated class is instantiated, but do not when + its subclasses are instantiated. + + It can be used to rename a base class in a library. For example, if we + have + + class OldName(SomeClass): + # ... + + and we want to rename it to NewName, we can do the following:: + + class NewName(SomeClass): + # ... + + OldName = create_deprecated_class('OldName', NewName) + + Then, if user class inherits from OldName, warning is issued. Also, if + some code uses ``issubclass(sub, OldName)`` or ``isinstance(sub(), OldName)`` + checks they'll still return True if sub is a subclass of NewName instead of + OldName. + """ # https://github.com/python/mypy/issues/4177 class DeprecatedClass(new_class.__class__): # type: ignore[misc,name-defined] @@ -128,6 +153,7 @@ def update_classpath(path: Any) -> Any: + """Update a deprecated path from an object with its new location""" for prefix, replacement in DEPRECATION_RULES: if isinstance(path, str) and path.startswith(prefix): new_path = path.replace(prefix, replacement, 1) @@ -140,12 +166,56 @@ def method_is_overridden(subclass: type, base_class: type, method_name: str) -> bool: + """ + Return True if a method named ``method_name`` of a ``base_class`` + is overridden in a ``subclass``. + + >>> class Base: + ... def foo(self): + ... pass + >>> class Sub1(Base): + ... pass + >>> class Sub2(Base): + ... def foo(self): + ... pass + >>> class Sub3(Sub1): + ... def foo(self): + ... pass + >>> class Sub4(Sub2): + ... pass + >>> method_is_overridden(Base, Base, 'foo') + False + >>> method_is_overridden(Sub1, Base, 'foo') + False + >>> method_is_overridden(Sub2, Base, 'foo') + True + >>> method_is_overridden(Sub3, Base, 'foo') + True + >>> method_is_overridden(Sub4, Base, 'foo') + True + """ base_method = getattr(base_class, method_name) sub_method = getattr(subclass, method_name) return base_method.__code__ is not sub_method.__code__ def argument_is_required(func: Callable[..., Any], arg_name: str) -> bool: + """ + Check if a function argument is required (exists and doesn't have a default value). + + .. versionadded:: 2.14 + + >>> def func(a, b=1, c=None): + ... pass + >>> argument_is_required(func, 'a') + True + >>> argument_is_required(func, 'b') + False + >>> argument_is_required(func, 'c') + False + >>> argument_is_required(func, 'd') + False + """ args = get_func_args_dict(func) param = args.get(arg_name) return param is not None and param.default is inspect.Parameter.empty @@ -158,4 +228,4 @@ f"The corresponding setting name is '{setting_name}'.", category=ScrapyDeprecationWarning, stacklevel=2, - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/deprecate.py
Write docstrings for backend logic
from __future__ import annotations import gc import inspect import re import sys import weakref from collections.abc import AsyncIterator, Iterable, Mapping from functools import partial, wraps from itertools import chain from typing import TYPE_CHECKING, Any, Concatenate, ParamSpec, TypeVar, overload from scrapy.utils.asyncgen import as_async_generator if TYPE_CHECKING: from collections.abc import Callable, Iterator from re import Pattern # typing.Self requires Python 3.11 from typing_extensions import Self _T = TypeVar("_T") _KT = TypeVar("_KT") _VT = TypeVar("_VT") _P = ParamSpec("_P") def is_listlike(x: Any) -> bool: return hasattr(x, "__iter__") and not isinstance(x, (str, bytes)) def unique(list_: Iterable[_T], key: Callable[[_T], Any] = lambda x: x) -> list[_T]: seen = set() result: list[_T] = [] for item in list_: seenkey = key(item) if seenkey in seen: continue seen.add(seenkey) result.append(item) return result def to_unicode( text: str | bytes, encoding: str | None = None, errors: str = "strict" ) -> str: if isinstance(text, str): return text if not isinstance(text, (bytes, str)): raise TypeError( f"to_unicode must receive a bytes or str object, got {type(text).__name__}" ) if encoding is None: encoding = "utf-8" return text.decode(encoding, errors) def to_bytes( text: str | bytes, encoding: str | None = None, errors: str = "strict" ) -> bytes: if isinstance(text, bytes): return text if not isinstance(text, str): raise TypeError( f"to_bytes must receive a str or bytes object, got {type(text).__name__}" ) if encoding is None: encoding = "utf-8" return text.encode(encoding, errors) def re_rsearch( pattern: str | Pattern[str], text: str, chunk_size: int = 1024 ) -> tuple[int, int] | None: def _chunk_iter() -> Iterable[tuple[str, int]]: offset = len(text) while True: offset -= chunk_size * 1024 if offset <= 0: break yield (text[offset:], offset) yield (text, 0) if isinstance(pattern, str): pattern = re.compile(pattern) for chunk, offset in _chunk_iter(): matches = list(pattern.finditer(chunk)) if matches: start, end = matches[-1].span() return offset + start, offset + end return None _SelfT = TypeVar("_SelfT") def memoizemethod_noargs( method: Callable[Concatenate[_SelfT, _P], _T], ) -> Callable[Concatenate[_SelfT, _P], _T]: cache: weakref.WeakKeyDictionary[_SelfT, _T] = weakref.WeakKeyDictionary() @wraps(method) def new_method(self: _SelfT, *args: _P.args, **kwargs: _P.kwargs) -> _T: if self not in cache: cache[self] = method(self, *args, **kwargs) return cache[self] return new_method _BINARYCHARS = { i for i in range(32) if to_bytes(chr(i)) not in {b"\0", b"\t", b"\n", b"\r"} } def binary_is_text(data: bytes) -> bool: if not isinstance(data, bytes): raise TypeError(f"data must be bytes, got '{type(data).__name__}'") return all(c not in _BINARYCHARS for c in data) def get_func_args_dict( func: Callable[..., Any], stripself: bool = False ) -> Mapping[str, inspect.Parameter]: if not callable(func): raise TypeError(f"func must be callable, got '{type(func).__name__}'") args: Mapping[str, inspect.Parameter] try: sig = inspect.signature(func) except ValueError: return {} if isinstance(func, partial): partial_args = func.args partial_kw = func.keywords args = {} for name, param in sig.parameters.items(): if name in partial_args: continue if partial_kw and name in partial_kw: continue args[name] = param else: args = sig.parameters if stripself and args and "self" in args: args = {k: v for k, v in args.items() if k != "self"} return args def get_func_args(func: Callable[..., Any], stripself: bool = False) -> list[str]: return list(get_func_args_dict(func, stripself=stripself)) def get_spec(func: Callable[..., Any]) -> tuple[list[str], dict[str, Any]]: if inspect.isfunction(func) or inspect.ismethod(func): spec = inspect.getfullargspec(func) elif hasattr(func, "__call__"): # noqa: B004 spec = inspect.getfullargspec(func.__call__) else: raise TypeError(f"{type(func)} is not callable") defaults: tuple[Any, ...] = spec.defaults or () firstdefault = len(spec.args) - len(defaults) args = spec.args[:firstdefault] kwargs = dict(zip(spec.args[firstdefault:], defaults, strict=False)) return args, kwargs @overload def without_none_values(iterable: Mapping[_KT, _VT]) -> dict[_KT, _VT]: ... @overload def without_none_values(iterable: Iterable[_KT]) -> Iterable[_KT]: ... def without_none_values( iterable: Mapping[_KT, _VT] | Iterable[_KT], ) -> dict[_KT, _VT] | Iterable[_KT]: if isinstance(iterable, Mapping): return {k: v for k, v in iterable.items() if v is not None} # the iterable __init__ must take another iterable return type(iterable)(v for v in iterable if v is not None) # type: ignore[call-arg] def global_object_name(obj: Any) -> str: return f"{obj.__module__}.{obj.__qualname__}" if hasattr(sys, "pypy_version_info"): def garbage_collect() -> None: # Collecting weakreferences can take two collections on PyPy. gc.collect() gc.collect() else: def garbage_collect() -> None: gc.collect() class MutableChain(Iterable[_T]): def __init__(self, *args: Iterable[_T]): self.data: Iterator[_T] = chain.from_iterable(args) def extend(self, *iterables: Iterable[_T]) -> None: self.data = chain(self.data, chain.from_iterable(iterables)) def __iter__(self) -> Iterator[_T]: return self def __next__(self) -> _T: return next(self.data) async def _async_chain( *iterables: Iterable[_T] | AsyncIterator[_T], ) -> AsyncIterator[_T]: for it in iterables: async for o in as_async_generator(it): yield o class MutableAsyncChain(AsyncIterator[_T]): def __init__(self, *args: Iterable[_T] | AsyncIterator[_T]): self.data: AsyncIterator[_T] = _async_chain(*args) def extend(self, *iterables: Iterable[_T] | AsyncIterator[_T]) -> None: self.data = _async_chain(self.data, _async_chain(*iterables)) def __aiter__(self) -> Self: return self async def __anext__(self) -> _T: return await self.data.__anext__() def _looks_like_import_path(value: str) -> bool: if not value: return False if any(c.isspace() for c in value): return False allowed_chars = set( "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_." ) if any(c not in allowed_chars for c in value): return False if value[0] == "." or value[-1] == ".": return False parts = value.split(".") if any(part == "" for part in parts): return False return all(part.isidentifier() for part in parts)
--- +++ @@ -1,3 +1,6 @@+""" +This module contains essential stuff that should've come with Python itself ;) +""" from __future__ import annotations @@ -28,10 +31,31 @@ def is_listlike(x: Any) -> bool: + """ + >>> is_listlike("foo") + False + >>> is_listlike(5) + False + >>> is_listlike(b"foo") + False + >>> is_listlike([b"foo"]) + True + >>> is_listlike((b"foo",)) + True + >>> is_listlike({}) + True + >>> is_listlike(set()) + True + >>> is_listlike((x for x in range(3))) + True + >>> is_listlike(range(5)) + True + """ return hasattr(x, "__iter__") and not isinstance(x, (str, bytes)) def unique(list_: Iterable[_T], key: Callable[[_T], Any] = lambda x: x) -> list[_T]: + """efficient function to uniquify a list preserving item order""" seen = set() result: list[_T] = [] for item in list_: @@ -46,6 +70,8 @@ def to_unicode( text: str | bytes, encoding: str | None = None, errors: str = "strict" ) -> str: + """Return the unicode representation of a bytes object ``text``. If + ``text`` is already an unicode object, return it as-is.""" if isinstance(text, str): return text if not isinstance(text, (bytes, str)): @@ -60,6 +86,8 @@ def to_bytes( text: str | bytes, encoding: str | None = None, errors: str = "strict" ) -> bytes: + """Return the binary representation of ``text``. If ``text`` + is already a bytes object, return it as-is.""" if isinstance(text, bytes): return text if not isinstance(text, str): @@ -74,6 +102,18 @@ def re_rsearch( pattern: str | Pattern[str], text: str, chunk_size: int = 1024 ) -> tuple[int, int] | None: + """ + This function does a reverse search in a text using a regular expression + given in the attribute 'pattern'. + Since the re module does not provide this functionality, we have to find for + the expression into chunks of text extracted from the end (for the sake of efficiency). + At first, a chunk of 'chunk_size' kilobytes is extracted from the end, and searched for + the pattern. If the pattern is not found, another chunk is extracted, and another + search is performed. + This process continues until a match is found, or until the whole file is read. + In case the pattern wasn't found, None is returned, otherwise it returns a tuple containing + the start position of the match, and the ending (regarding the entire text). + """ def _chunk_iter() -> Iterable[tuple[str, int]]: offset = len(text) @@ -101,6 +141,9 @@ def memoizemethod_noargs( method: Callable[Concatenate[_SelfT, _P], _T], ) -> Callable[Concatenate[_SelfT, _P], _T]: + """Decorator to cache the result of a method (without arguments) using a + weak reference to its object + """ cache: weakref.WeakKeyDictionary[_SelfT, _T] = weakref.WeakKeyDictionary() @wraps(method) @@ -118,6 +161,9 @@ def binary_is_text(data: bytes) -> bool: + """Returns ``True`` if the given ``data`` argument (a ``bytes`` object) + does not contain unprintable control characters. + """ if not isinstance(data, bytes): raise TypeError(f"data must be bytes, got '{type(data).__name__}'") return all(c not in _BINARYCHARS for c in data) @@ -126,6 +172,10 @@ def get_func_args_dict( func: Callable[..., Any], stripself: bool = False ) -> Mapping[str, inspect.Parameter]: + """Return the argument dict of a callable object. + + .. versionadded:: 2.14 + """ if not callable(func): raise TypeError(f"func must be callable, got '{type(func).__name__}'") @@ -155,10 +205,31 @@ def get_func_args(func: Callable[..., Any], stripself: bool = False) -> list[str]: + """Return the argument name list of a callable object""" return list(get_func_args_dict(func, stripself=stripself)) def get_spec(func: Callable[..., Any]) -> tuple[list[str], dict[str, Any]]: + """Returns (args, kwargs) tuple for a function + >>> import re + >>> get_spec(re.match) + (['pattern', 'string'], {'flags': 0}) + + >>> class Test: + ... def __call__(self, val): + ... pass + ... def method(self, val, flags=0): + ... pass + + >>> get_spec(Test) + (['self', 'val'], {}) + + >>> get_spec(Test.method) + (['self', 'val'], {'flags': 0}) + + >>> get_spec(Test().method) + (['self', 'val'], {'flags': 0}) + """ if inspect.isfunction(func) or inspect.ismethod(func): spec = inspect.getfullargspec(func) @@ -186,6 +257,11 @@ def without_none_values( iterable: Mapping[_KT, _VT] | Iterable[_KT], ) -> dict[_KT, _VT] | Iterable[_KT]: + """Return a copy of ``iterable`` with all ``None`` entries removed. + + If ``iterable`` is a mapping, return a dictionary where all pairs that have + value ``None`` have been removed. + """ if isinstance(iterable, Mapping): return {k: v for k, v in iterable.items() if v is not None} # the iterable __init__ must take another iterable @@ -193,6 +269,14 @@ def global_object_name(obj: Any) -> str: + """Return the full import path of the given object. + + >>> from scrapy import Request + >>> global_object_name(Request) + 'scrapy.http.request.Request' + >>> global_object_name(Request.replace) + 'scrapy.http.request.Request.replace' + """ return f"{obj.__module__}.{obj.__qualname__}" @@ -210,6 +294,9 @@ class MutableChain(Iterable[_T]): + """ + Thin wrapper around itertools.chain, allowing to add iterables "in-place" + """ def __init__(self, *args: Iterable[_T]): self.data: Iterator[_T] = chain.from_iterable(args) @@ -233,6 +320,9 @@ class MutableAsyncChain(AsyncIterator[_T]): + """ + Similar to MutableChain but for async iterables + """ def __init__(self, *args: Iterable[_T] | AsyncIterator[_T]): self.data: AsyncIterator[_T] = _async_chain(*args) @@ -248,6 +338,8 @@ def _looks_like_import_path(value: str) -> bool: + """Return True if **value** looks like a valid Python import path or False + otherwise.""" if not value: return False if any(c.isspace() for c in value): @@ -262,4 +354,4 @@ parts = value.split(".") if any(part == "" for part in parts): return False - return all(part.isidentifier() for part in parts)+ return all(part.isidentifier() for part in parts)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/python.py
Auto-generate documentation strings for this file
from __future__ import annotations from typing import TYPE_CHECKING, Any from scrapy.exceptions import NotConfigured, NotSupported from scrapy.http import Response, TextResponse from scrapy.selector import Selector from scrapy.spiders import Spider from scrapy.utils.iterators import csviter, xmliter_lxml from scrapy.utils.spider import iterate_spider_output if TYPE_CHECKING: from collections.abc import Iterable, Sequence class XMLFeedSpider(Spider): iterator: str = "iternodes" itertag: str = "item" namespaces: Sequence[tuple[str, str]] = () def process_results( self, response: Response, results: Iterable[Any] ) -> Iterable[Any]: return results def adapt_response(self, response: Response) -> Response: return response def parse_node(self, response: Response, selector: Selector) -> Any: if hasattr(self, "parse_item"): # backward compatibility return self.parse_item(response, selector) raise NotImplementedError def parse_nodes(self, response: Response, nodes: Iterable[Selector]) -> Any: for selector in nodes: ret = iterate_spider_output(self.parse_node(response, selector)) yield from self.process_results(response, ret) def _parse(self, response: Response, **kwargs: Any) -> Any: if not hasattr(self, "parse_node"): raise NotConfigured( "You must define parse_node method in order to scrape this XML feed" ) response = self.adapt_response(response) nodes: Iterable[Selector] if self.iterator == "iternodes": nodes = self._iternodes(response) elif self.iterator == "xml": if not isinstance(response, TextResponse): raise ValueError("Response content isn't text") selector = Selector(response, type="xml") self._register_namespaces(selector) nodes = selector.xpath(f"//{self.itertag}") elif self.iterator == "html": if not isinstance(response, TextResponse): raise ValueError("Response content isn't text") selector = Selector(response, type="html") self._register_namespaces(selector) nodes = selector.xpath(f"//{self.itertag}") else: raise NotSupported("Unsupported node iterator") return self.parse_nodes(response, nodes) def _iternodes(self, response: Response) -> Iterable[Selector]: for node in xmliter_lxml(response, self.itertag): self._register_namespaces(node) yield node def _register_namespaces(self, selector: Selector) -> None: for prefix, uri in self.namespaces: selector.register_namespace(prefix, uri) class CSVFeedSpider(Spider): delimiter: str | None = ( None # When this is None, python's csv module's default delimiter is used ) quotechar: str | None = ( None # When this is None, python's csv module's default quotechar is used ) headers: list[str] | None = None def process_results( self, response: Response, results: Iterable[Any] ) -> Iterable[Any]: return results def adapt_response(self, response: Response) -> Response: return response def parse_row(self, response: Response, row: dict[str, str]) -> Any: raise NotImplementedError def parse_rows(self, response: Response) -> Any: for row in csviter( response, self.delimiter, self.headers, quotechar=self.quotechar ): ret = iterate_spider_output(self.parse_row(response, row)) yield from self.process_results(response, ret) def _parse(self, response: Response, **kwargs: Any) -> Any: if not hasattr(self, "parse_row"): raise NotConfigured( "You must define parse_row method in order to scrape this CSV feed" ) response = self.adapt_response(response) return self.parse_rows(response)
--- +++ @@ -1,3 +1,9 @@+""" +This module implements the XMLFeedSpider which is the recommended spider to use +for scraping from an XML feed. + +See documentation in docs/topics/spiders.rst +""" from __future__ import annotations @@ -15,6 +21,14 @@ class XMLFeedSpider(Spider): + """ + This class intends to be the base class for spiders that scrape + from XML feeds. + + You can choose whether to parse the file using the 'iternodes' iterator, an + 'xml' selector, or an 'html' selector. In most cases, it's convenient to + use iternodes, since it's a faster and cleaner. + """ iterator: str = "iternodes" itertag: str = "item" @@ -23,17 +37,35 @@ def process_results( self, response: Response, results: Iterable[Any] ) -> Iterable[Any]: + """This overridable method is called for each result (item or request) + returned by the spider, and it's intended to perform any last time + processing required before returning the results to the framework core, + for example setting the item GUIDs. It receives a list of results and + the response which originated that results. It must return a list of + results (items or requests). + """ return results def adapt_response(self, response: Response) -> Response: + """You can override this function in order to make any changes you want + to into the feed before parsing it. This function must return a + response. + """ return response def parse_node(self, response: Response, selector: Selector) -> Any: + """This method must be overridden with your custom spider functionality""" if hasattr(self, "parse_item"): # backward compatibility return self.parse_item(response, selector) raise NotImplementedError def parse_nodes(self, response: Response, nodes: Iterable[Selector]) -> Any: + """This method is called for the nodes matching the provided tag name + (itertag). Receives the response and an Selector for each node. + Overriding this method is mandatory. Otherwise, you spider won't work. + This method must return either an item, a request, or a list + containing any of them. + """ for selector in nodes: ret = iterate_spider_output(self.parse_node(response, selector)) @@ -77,6 +109,13 @@ class CSVFeedSpider(Spider): + """Spider for parsing CSV feeds. + It receives a CSV file in a response; iterates through each of its rows, + and calls parse_row with a dict containing each field's data. + + You can set some options regarding the CSV file, such as the delimiter, quotechar + and the file's headers. + """ delimiter: str | None = ( None # When this is None, python's csv module's default delimiter is used @@ -89,15 +128,23 @@ def process_results( self, response: Response, results: Iterable[Any] ) -> Iterable[Any]: + """This method has the same purpose as the one in XMLFeedSpider""" return results def adapt_response(self, response: Response) -> Response: + """This method has the same purpose as the one in XMLFeedSpider""" return response def parse_row(self, response: Response, row: dict[str, str]) -> Any: + """This method must be overridden with your custom spider functionality""" raise NotImplementedError def parse_rows(self, response: Response) -> Any: + """Receives a response and a dict (representing each row) with a key for + each provided (or detected) header of the CSV file. This spider also + gives the opportunity to override adapt_response and + process_results methods for pre and post-processing purposes. + """ for row in csviter( response, self.delimiter, self.headers, quotechar=self.quotechar @@ -111,4 +158,4 @@ "You must define parse_row method in order to scrape this CSV feed" ) response = self.adapt_response(response) - return self.parse_rows(response)+ return self.parse_rows(response)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spiders/feed.py
Add detailed docstrings explaining each function
from __future__ import annotations import warnings from typing import TYPE_CHECKING, Any, cast from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.spiders import Spider from scrapy.utils.spider import iterate_spider_output if TYPE_CHECKING: from collections.abc import AsyncIterator, Iterable from scrapy import Request from scrapy.http import Response class InitSpider(Spider): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) warnings.warn( "InitSpider is deprecated. Copy its code from Scrapy's source if needed. " "Will be removed in a future version.", ScrapyDeprecationWarning, stacklevel=2, ) async def start(self) -> AsyncIterator[Any]: with warnings.catch_warnings(): warnings.filterwarnings( "ignore", category=ScrapyDeprecationWarning, module=r"^scrapy\.spiders$" ) for item_or_request in self.start_requests(): yield item_or_request def start_requests(self) -> Iterable[Request]: self._postinit_reqs: Iterable[Request] = super().start_requests() return cast("Iterable[Request]", iterate_spider_output(self.init_request())) def initialized(self, response: Response | None = None) -> Any: return self.__dict__.pop("_postinit_reqs") def init_request(self) -> Any: return self.initialized()
--- +++ @@ -15,6 +15,11 @@ class InitSpider(Spider): + """Base Spider with initialization facilities + + .. warning:: This class is deprecated. Copy its code into your project if needed. + It will be removed in a future Scrapy version. + """ def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) @@ -38,7 +43,22 @@ return cast("Iterable[Request]", iterate_spider_output(self.init_request())) def initialized(self, response: Response | None = None) -> Any: + """This method must be set as the callback of your last initialization + request. See self.init_request() docstring for more info. + """ return self.__dict__.pop("_postinit_reqs") def init_request(self) -> Any: - return self.initialized()+ """This function should return one initialization request, with the + self.initialized method as callback. When the self.initialized method + is called this spider is considered initialized. If you need to perform + several requests for initializing your spider, you can do so by using + different callbacks. The only requirement is that the final callback + (of the last initialization request) must be self.initialized. + + The default implementation calls self.initialized immediately, and + means that no initialization is needed. This method should be + overridden only when you need to perform requests to initialize your + spider + """ + return self.initialized()
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/spiders/init.py
Write docstrings including parameters and return values
from __future__ import annotations from typing import TYPE_CHECKING, Any from urllib.parse import urljoin import lxml.etree if TYPE_CHECKING: from collections.abc import Iterable, Iterator class Sitemap: def __init__(self, xmltext: str | bytes): xmlp = lxml.etree.XMLParser( recover=True, remove_comments=True, resolve_entities=False ) self._root = lxml.etree.fromstring(xmltext, parser=xmlp) rt = self._root.tag assert isinstance(rt, str) self.type = rt.split("}", 1)[1] if "}" in rt else rt def __iter__(self) -> Iterator[dict[str, Any]]: for elem in self._root.getchildren(): d: dict[str, Any] = {} for el in elem.getchildren(): tag = el.tag assert isinstance(tag, str) name = tag.split("}", 1)[1] if "}" in tag else tag if name == "link": if "href" in el.attrib: d.setdefault("alternate", []).append(el.get("href")) else: d[name] = el.text.strip() if el.text else "" if "loc" in d: yield d def sitemap_urls_from_robots( robots_text: str, base_url: str | None = None ) -> Iterable[str]: for line in robots_text.splitlines(): if line.lstrip().lower().startswith("sitemap:"): url = line.split(":", 1)[1].strip() yield urljoin(base_url or "", url)
--- +++ @@ -1,3 +1,9 @@+""" +Module for processing Sitemaps. + +Note: The main purpose of this module is to provide support for the +SitemapSpider, its API is subject to change without notice. +""" from __future__ import annotations @@ -11,6 +17,8 @@ class Sitemap: + """Class to parse Sitemap (type=urlset) and Sitemap Index + (type=sitemapindex) files""" def __init__(self, xmltext: str | bytes): xmlp = lxml.etree.XMLParser( @@ -42,7 +50,10 @@ def sitemap_urls_from_robots( robots_text: str, base_url: str | None = None ) -> Iterable[str]: + """Return an iterator over all sitemap urls contained in the given + robots.txt file + """ for line in robots_text.splitlines(): if line.lstrip().lower().startswith("sitemap:"): url = line.split(":", 1)[1].strip() - yield urljoin(base_url or "", url)+ yield urljoin(base_url or "", url)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/sitemap.py
Provide docstrings following PEP 257
from __future__ import annotations import code from collections.abc import Callable from functools import wraps from typing import TYPE_CHECKING, Any if TYPE_CHECKING: from collections.abc import Iterable EmbedFuncT = Callable[..., None] KnownShellsT = dict[str, Callable[..., EmbedFuncT]] def _embed_ipython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: try: from IPython.terminal.embed import InteractiveShellEmbed # noqa: T100,PLC0415 from IPython.terminal.ipapp import load_default_config # noqa: PLC0415 except ImportError: from IPython.frontend.terminal.embed import ( # type: ignore[import-not-found,no-redef] # noqa: T100,PLC0415 InteractiveShellEmbed, ) from IPython.frontend.terminal.ipapp import ( # type: ignore[import-not-found,no-redef] # noqa: PLC0415 load_default_config, ) @wraps(_embed_ipython_shell) def wrapper(namespace: dict[str, Any] = namespace, banner: str = "") -> None: config = load_default_config() # Always use .instance() to ensure _instance propagation to all parents # this is needed for <TAB> completion works well for new imports # and clear the instance to always have the fresh env # on repeated breaks like with inspect_response() InteractiveShellEmbed.clear_instance() shell = InteractiveShellEmbed.instance( banner1=banner, user_ns=namespace, config=config ) shell() return wrapper def _embed_bpython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: import bpython # noqa: PLC0415 @wraps(_embed_bpython_shell) def wrapper(namespace: dict[str, Any] = namespace, banner: str = "") -> None: bpython.embed(locals_=namespace, banner=banner) return wrapper def _embed_ptpython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: import ptpython.repl # noqa: PLC0415 # pylint: disable=import-error @wraps(_embed_ptpython_shell) def wrapper(namespace: dict[str, Any] = namespace, banner: str = "") -> None: print(banner) ptpython.repl.embed(locals=namespace) return wrapper def _embed_standard_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: try: # readline module is only available on unix systems import readline # noqa: PLC0415 except ImportError: pass else: import rlcompleter # noqa: F401,PLC0415 readline.parse_and_bind("tab:complete") # type: ignore[attr-defined,unused-ignore] @wraps(_embed_standard_shell) def wrapper(namespace: dict[str, Any] = namespace, banner: str = "") -> None: code.interact(banner=banner, local=namespace) return wrapper DEFAULT_PYTHON_SHELLS: KnownShellsT = { "ptpython": _embed_ptpython_shell, "ipython": _embed_ipython_shell, "bpython": _embed_bpython_shell, "python": _embed_standard_shell, } def get_shell_embed_func( shells: Iterable[str] | None = None, known_shells: KnownShellsT | None = None ) -> EmbedFuncT | None: if shells is None: # list, preference order of shells shells = DEFAULT_PYTHON_SHELLS.keys() if known_shells is None: # available embeddable shells known_shells = DEFAULT_PYTHON_SHELLS.copy() for shell in shells: if shell in known_shells: try: # function test: run all setup code (imports), # but dont fall into the shell return known_shells[shell]() except ImportError: continue return None def start_python_console( namespace: dict[str, Any] | None = None, banner: str = "", shells: Iterable[str] | None = None, ) -> None: if namespace is None: namespace = {} try: shell = get_shell_embed_func(shells) if shell is not None: shell(namespace=namespace, banner=banner) except SystemExit: # raised when using exit() in python code.interact pass
--- +++ @@ -15,6 +15,7 @@ def _embed_ipython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: + """Start an IPython Shell""" try: from IPython.terminal.embed import InteractiveShellEmbed # noqa: T100,PLC0415 from IPython.terminal.ipapp import load_default_config # noqa: PLC0415 @@ -45,6 +46,7 @@ def _embed_bpython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: + """Start a bpython shell""" import bpython # noqa: PLC0415 @wraps(_embed_bpython_shell) @@ -57,6 +59,7 @@ def _embed_ptpython_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: + """Start a ptpython shell""" import ptpython.repl # noqa: PLC0415 # pylint: disable=import-error @wraps(_embed_ptpython_shell) @@ -70,6 +73,7 @@ def _embed_standard_shell( namespace: dict[str, Any] = {}, banner: str = "" ) -> EmbedFuncT: + """Start a standard python shell""" try: # readline module is only available on unix systems import readline # noqa: PLC0415 except ImportError: @@ -97,6 +101,9 @@ def get_shell_embed_func( shells: Iterable[str] | None = None, known_shells: KnownShellsT | None = None ) -> EmbedFuncT | None: + """Return the first acceptable shell-embed function + from a given list of shell names. + """ if shells is None: # list, preference order of shells shells = DEFAULT_PYTHON_SHELLS.keys() if known_shells is None: # available embeddable shells @@ -117,6 +124,9 @@ banner: str = "", shells: Iterable[str] | None = None, ) -> None: + """Start Python console bound to the given namespace. + Readline support and tab completion will be used on Unix, if available. + """ if namespace is None: namespace = {} @@ -125,4 +135,4 @@ if shell is not None: shell(namespace=namespace, banner=banner) except SystemExit: # raised when using exit() in python code.interact - pass+ pass
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/console.py
Generate helpful docstrings for debugging
from __future__ import annotations import asyncio import logging import warnings from collections.abc import Awaitable, Callable, Generator, Sequence from typing import Any as TypingAny from pydispatch.dispatcher import ( Anonymous, Any, disconnect, getAllReceivers, liveReceivers, ) from pydispatch.robustapply import robustApply from twisted.internet.defer import Deferred, DeferredList, inlineCallbacks from twisted.python.failure import Failure from scrapy.exceptions import ScrapyDeprecationWarning, StopDownload from scrapy.utils.asyncio import is_asyncio_available from scrapy.utils.defer import ( _maybeDeferred_coro, ensure_awaitable, maybe_deferred_to_future, ) from scrapy.utils.log import failure_to_exc_info from scrapy.utils.python import global_object_name logger = logging.getLogger(__name__) def send_catch_log( signal: TypingAny = Any, sender: TypingAny = Anonymous, *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: dont_log = named.pop("dont_log", ()) dont_log = tuple(dont_log) if isinstance(dont_log, Sequence) else (dont_log,) dont_log += (StopDownload,) spider = named.get("spider") responses: list[tuple[TypingAny, TypingAny]] = [] for receiver in liveReceivers(getAllReceivers(sender, signal)): result: TypingAny try: response = robustApply( receiver, *arguments, signal=signal, sender=sender, **named ) if isinstance(response, Deferred): logger.error( "Cannot return deferreds from signal handler: %(receiver)s", {"receiver": receiver}, extra={"spider": spider}, ) except dont_log: result = Failure() except Exception: result = Failure() logger.error( "Error caught on signal handler: %(receiver)s", {"receiver": receiver}, exc_info=True, extra={"spider": spider}, ) else: result = response responses.append((receiver, result)) return responses def send_catch_log_deferred( signal: TypingAny = Any, sender: TypingAny = Anonymous, *arguments: TypingAny, **named: TypingAny, ) -> Deferred[list[tuple[TypingAny, TypingAny]]]: warnings.warn( "send_catch_log_deferred() is deprecated, use send_catch_log_async() instead", ScrapyDeprecationWarning, stacklevel=2, ) return _send_catch_log_deferred(signal, sender, *arguments, **named) @inlineCallbacks def _send_catch_log_deferred( signal: TypingAny, sender: TypingAny, *arguments: TypingAny, **named: TypingAny, ) -> Generator[Deferred[TypingAny], TypingAny, list[tuple[TypingAny, TypingAny]]]: def logerror(failure: Failure, recv: TypingAny) -> Failure: if dont_log is None or not isinstance(failure.value, dont_log): logger.error( "Error caught on signal handler: %(receiver)s", {"receiver": recv}, exc_info=failure_to_exc_info(failure), extra={"spider": spider}, ) return failure dont_log = named.pop("dont_log", None) spider = named.get("spider") dfds: list[Deferred[tuple[TypingAny, TypingAny]]] = [] for receiver in liveReceivers(getAllReceivers(sender, signal)): d: Deferred[TypingAny] = _maybeDeferred_coro( robustApply, True, receiver, *arguments, signal=signal, sender=sender, **named, ) d.addErrback(logerror, receiver) # TODO https://pylint.readthedocs.io/en/latest/user_guide/messages/warning/cell-var-from-loop.html d2: Deferred[tuple[TypingAny, TypingAny]] = d.addBoth( lambda result: ( receiver, # pylint: disable=cell-var-from-loop # noqa: B023 result, ) ) dfds.append(d2) results = yield DeferredList(dfds) return [result[1] for result in results] async def send_catch_log_async( signal: TypingAny = Any, sender: TypingAny = Anonymous, *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: # note that this returns exceptions instead of Failures in the second tuple member if is_asyncio_available(): return await _send_catch_log_asyncio(signal, sender, *arguments, **named) results = await maybe_deferred_to_future( _send_catch_log_deferred(signal, sender, *arguments, **named) ) return [ (receiver, result.value if isinstance(result, Failure) else result) for receiver, result in results ] async def _send_catch_log_asyncio( signal: TypingAny = Any, sender: TypingAny = Anonymous, *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: dont_log = named.pop("dont_log", ()) dont_log = tuple(dont_log) if isinstance(dont_log, Sequence) else (dont_log,) spider = named.get("spider") handlers: list[Awaitable[TypingAny]] = [] for receiver in liveReceivers(getAllReceivers(sender, signal)): async def handler(receiver: Callable) -> TypingAny: result: TypingAny try: result = await ensure_awaitable( robustApply( receiver, *arguments, signal=signal, sender=sender, **named ), _warn=global_object_name(receiver), ) except dont_log as ex: # pylint: disable=catching-non-exception result = ex except Exception as ex: logger.error( "Error caught on signal handler: %(receiver)s", {"receiver": receiver}, exc_info=True, extra={"spider": spider}, ) result = ex return (receiver, result) handlers.append(handler(receiver)) return await asyncio.gather(*handlers, return_exceptions=True) def disconnect_all(signal: TypingAny = Any, sender: TypingAny = Any) -> None: for receiver in liveReceivers(getAllReceivers(sender, signal)): disconnect(receiver, signal=signal, sender=sender)
--- +++ @@ -1,3 +1,4 @@+"""Helper functions for working with signals""" from __future__ import annotations @@ -37,6 +38,9 @@ *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: + """Like ``pydispatcher.robust.sendRobust()`` but it also logs errors and returns + Failures instead of exceptions. + """ dont_log = named.pop("dont_log", ()) dont_log = tuple(dont_log) if isinstance(dont_log, Sequence) else (dont_log,) dont_log += (StopDownload,) @@ -76,6 +80,11 @@ *arguments: TypingAny, **named: TypingAny, ) -> Deferred[list[tuple[TypingAny, TypingAny]]]: + """Like :func:`send_catch_log` but supports :ref:`asynchronous signal handlers + <signal-deferred>`. + + Returns a deferred that gets fired once all signal handlers have finished. + """ warnings.warn( "send_catch_log_deferred() is deprecated, use send_catch_log_async() instead", ScrapyDeprecationWarning, @@ -134,6 +143,13 @@ *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: + """Like :func:`send_catch_log` but supports :ref:`asynchronous signal handlers + <signal-deferred>`. + + Returns a coroutine that completes once all signal handlers have finished. + + .. versionadded:: 2.14 + """ # note that this returns exceptions instead of Failures in the second tuple member if is_asyncio_available(): return await _send_catch_log_asyncio(signal, sender, *arguments, **named) @@ -152,6 +168,17 @@ *arguments: TypingAny, **named: TypingAny, ) -> list[tuple[TypingAny, TypingAny]]: + """Like :func:`send_catch_log` but supports :ref:`asynchronous signal handlers + <signal-deferred>`. + + Returns a coroutine that completes once all signal handlers have finished. + + This function requires + :class:`~twisted.internet.asyncioreactor.AsyncioSelectorReactor` to be + installed. + + .. versionadded:: 2.14 + """ dont_log = named.pop("dont_log", ()) dont_log = tuple(dont_log) if isinstance(dont_log, Sequence) else (dont_log,) spider = named.get("spider") @@ -185,5 +212,8 @@ def disconnect_all(signal: TypingAny = Any, sender: TypingAny = Any) -> None: - for receiver in liveReceivers(getAllReceivers(sender, signal)): - disconnect(receiver, signal=signal, sender=sender)+ """Disconnect all signal handlers. Useful for cleaning up after running + tests. + """ + for receiver in liveReceivers(getAllReceivers(sender, signal)): + disconnect(receiver, signal=signal, sender=sender)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/signal.py
Add standardized docstrings across the file
from __future__ import annotations from typing import TYPE_CHECKING from urllib.parse import ParseResult, urlparse from weakref import WeakKeyDictionary if TYPE_CHECKING: from scrapy.http import Request, Response _urlparse_cache: WeakKeyDictionary[Request | Response, ParseResult] = ( WeakKeyDictionary() ) def urlparse_cached(request_or_response: Request | Response) -> ParseResult: if request_or_response not in _urlparse_cache: _urlparse_cache[request_or_response] = urlparse(request_or_response.url) return _urlparse_cache[request_or_response]
--- +++ @@ -1,3 +1,4 @@+"""Helper functions for scrapy.http objects (Request, Response)""" from __future__ import annotations @@ -15,6 +16,9 @@ def urlparse_cached(request_or_response: Request | Response) -> ParseResult: + """Return urlparse.urlparse caching the result, where the argument can be a + Request or Response object + """ if request_or_response not in _urlparse_cache: _urlparse_cache[request_or_response] = urlparse(request_or_response.url) - return _urlparse_cache[request_or_response]+ return _urlparse_cache[request_or_response]
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/httpobj.py
Write docstrings for backend logic
from __future__ import annotations from collections import defaultdict from operator import itemgetter from time import time from typing import TYPE_CHECKING, Any from weakref import WeakKeyDictionary if TYPE_CHECKING: from collections.abc import Iterable # typing.Self requires Python 3.11 from typing_extensions import Self NoneType = type(None) live_refs: defaultdict[type, WeakKeyDictionary] = defaultdict(WeakKeyDictionary) class object_ref: __slots__ = () def __new__(cls, *args: Any, **kwargs: Any) -> Self: obj = object.__new__(cls) live_refs[cls][obj] = time() return obj # using Any as it's hard to type type(None) def format_live_refs(ignore: Any = NoneType) -> str: s = "Live References\n\n" now = time() for cls, wdict in sorted(live_refs.items(), key=lambda x: x[0].__name__): if not wdict: continue if issubclass(cls, ignore): continue oldest = min(wdict.values()) s += f"{cls.__name__:<30} {len(wdict):6} oldest: {int(now - oldest)}s ago\n" return s def print_live_refs(*a: Any, **kw: Any) -> None: print(format_live_refs(*a, **kw)) def get_oldest(class_name: str) -> Any: for cls, wdict in live_refs.items(): if cls.__name__ == class_name: if not wdict: break return min(wdict.items(), key=itemgetter(1))[0] return None def iter_all(class_name: str) -> Iterable[Any]: for cls, wdict in live_refs.items(): if cls.__name__ == class_name: return wdict.keys() return []
--- +++ @@ -1,3 +1,13 @@+"""This module provides some functions and classes to record and report +references to live object instances. + +If you want live objects for a particular class to be tracked, you only have to +subclass from object_ref (instead of object). + +About performance: This library has a minimal performance impact when enabled, +and no performance penalty at all when disabled (as object_ref becomes just an +alias to object in that case). +""" from __future__ import annotations @@ -19,6 +29,7 @@ class object_ref: + """Inherit from this class to a keep a record of live instances""" __slots__ = () @@ -30,6 +41,7 @@ # using Any as it's hard to type type(None) def format_live_refs(ignore: Any = NoneType) -> str: + """Return a tabular representation of tracked objects""" s = "Live References\n\n" now = time() for cls, wdict in sorted(live_refs.items(), key=lambda x: x[0].__name__): @@ -43,10 +55,12 @@ def print_live_refs(*a: Any, **kw: Any) -> None: + """Print tracked objects""" print(format_live_refs(*a, **kw)) def get_oldest(class_name: str) -> Any: + """Get the oldest object for a specific class name""" for cls, wdict in live_refs.items(): if cls.__name__ == class_name: if not wdict: @@ -56,7 +70,8 @@ def iter_all(class_name: str) -> Iterable[Any]: + """Iterate over all objects of the same class by its class name""" for cls, wdict in live_refs.items(): if cls.__name__ == class_name: return wdict.keys() - return []+ return []
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/trackref.py
Provide clean and structured docstrings
from __future__ import annotations import inspect import logging from typing import TYPE_CHECKING, Any, TypeVar, overload from scrapy.spiders import Spider from scrapy.utils.defer import deferred_from_coro from scrapy.utils.misc import arg_to_iter if TYPE_CHECKING: from collections.abc import AsyncGenerator, Iterable from types import CoroutineType, ModuleType from twisted.internet.defer import Deferred from scrapy import Request from scrapy.spiderloader import SpiderLoaderProtocol logger = logging.getLogger(__name__) _T = TypeVar("_T") # https://stackoverflow.com/questions/60222982 @overload def iterate_spider_output(result: AsyncGenerator[_T]) -> AsyncGenerator[_T]: ... # type: ignore[overload-overlap] @overload def iterate_spider_output(result: CoroutineType[Any, Any, _T]) -> Deferred[_T]: ... @overload def iterate_spider_output(result: _T) -> Iterable[Any]: ... def iterate_spider_output( result: Any, ) -> Iterable[Any] | AsyncGenerator[_T] | Deferred[_T]: if inspect.isasyncgen(result): return result if inspect.iscoroutine(result): d = deferred_from_coro(result) d.addCallback(iterate_spider_output) return d return arg_to_iter(deferred_from_coro(result)) def iter_spider_classes(module: ModuleType) -> Iterable[type[Spider]]: for obj in vars(module).values(): if ( inspect.isclass(obj) and issubclass(obj, Spider) and obj.__module__ == module.__name__ and getattr(obj, "name", None) ): yield obj @overload def spidercls_for_request( spider_loader: SpiderLoaderProtocol, request: Request, default_spidercls: type[Spider], log_none: bool = ..., log_multiple: bool = ..., ) -> type[Spider]: ... @overload def spidercls_for_request( spider_loader: SpiderLoaderProtocol, request: Request, default_spidercls: None, log_none: bool = ..., log_multiple: bool = ..., ) -> type[Spider] | None: ... @overload def spidercls_for_request( spider_loader: SpiderLoaderProtocol, request: Request, *, log_none: bool = ..., log_multiple: bool = ..., ) -> type[Spider] | None: ... def spidercls_for_request( spider_loader: SpiderLoaderProtocol, request: Request, default_spidercls: type[Spider] | None = None, log_none: bool = False, log_multiple: bool = False, ) -> type[Spider] | None: snames = spider_loader.find_by_request(request) if len(snames) == 1: return spider_loader.load(snames[0]) if len(snames) > 1 and log_multiple: logger.error( "More than one spider can handle: %(request)s - %(snames)s", {"request": request, "snames": ", ".join(snames)}, ) if len(snames) == 0 and log_none: logger.error( "Unable to find spider that handles: %(request)s", {"request": request} ) return default_spidercls class DefaultSpider(Spider): name = "default"
--- +++ @@ -49,6 +49,9 @@ def iter_spider_classes(module: ModuleType) -> Iterable[type[Spider]]: + """Return an iterator over all spider classes defined in the given module + that can be instantiated (i.e. which have name) + """ for obj in vars(module).values(): if ( inspect.isclass(obj) @@ -96,6 +99,16 @@ log_none: bool = False, log_multiple: bool = False, ) -> type[Spider] | None: + """Return a spider class that handles the given Request. + + This will look for the spiders that can handle the given request (using + the spider loader) and return a Spider class if (and only if) there is + only one Spider able to handle the Request. + + If multiple spiders (or no spider) are found, it will return the + default_spidercls passed. It can optionally log if multiple or no spiders + are found. + """ snames = spider_loader.find_by_request(request) if len(snames) == 1: return spider_loader.load(snames[0]) @@ -115,4 +128,4 @@ class DefaultSpider(Spider): - name = "default"+ name = "default"
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/spider.py
Please document this code using docstrings
from __future__ import annotations import re import warnings from importlib import import_module from typing import TYPE_CHECKING, Any, TypeAlias from urllib.parse import ParseResult, urldefrag, urlparse, urlunparse from warnings import warn from w3lib.url import __all__ as _public_w3lib_objects from w3lib.url import add_or_replace_parameter as _add_or_replace_parameter from w3lib.url import any_to_uri as _any_to_uri from w3lib.url import parse_url as _parse_url from scrapy.exceptions import ScrapyDeprecationWarning def __getattr__(name: str) -> Any: if name in ("_unquotepath", "_safe_chars", "parse_url", *_public_w3lib_objects): obj_type = "attribute" if name == "_safe_chars" else "function" warnings.warn( f"The scrapy.utils.url.{name} {obj_type} is deprecated, use w3lib.url.{name} instead.", ScrapyDeprecationWarning, ) return getattr(import_module("w3lib.url"), name) raise AttributeError if TYPE_CHECKING: from collections.abc import Iterable from scrapy import Spider UrlT: TypeAlias = str | bytes | ParseResult def url_is_from_any_domain(url: UrlT, domains: Iterable[str]) -> bool: host = _parse_url(url).netloc.lower() if not host: return False domains = [d.lower() for d in domains] return any((host == d) or (host.endswith(f".{d}")) for d in domains) def url_is_from_spider(url: UrlT, spider: type[Spider]) -> bool: return url_is_from_any_domain( url, [spider.name, *getattr(spider, "allowed_domains", [])] ) def url_has_any_extension(url: UrlT, extensions: Iterable[str]) -> bool: lowercase_path = _parse_url(url).path.lower() return any(lowercase_path.endswith(ext) for ext in extensions) def escape_ajax(url: str) -> str: warn( "escape_ajax() is deprecated and will be removed in a future Scrapy version.", ScrapyDeprecationWarning, stacklevel=2, ) defrag, frag = urldefrag(url) if not frag.startswith("!"): return url return _add_or_replace_parameter(defrag, "_escaped_fragment_", frag[1:]) def add_http_if_no_scheme(url: str) -> str: match = re.match(r"^\w+://", url, flags=re.IGNORECASE) if not match: parts = urlparse(url) scheme = "http:" if parts.netloc else "http://" url = scheme + url return url def _is_posix_path(string: str) -> bool: return bool( re.match( r""" ^ # start with... ( \. # ...a single dot, ( \. | [^/\.]+ # optionally followed by )? # either a second dot or some characters | ~ # $HOME )? # optional match of ".", ".." or ".blabla" / # at least one "/" for a file path, . # and something after the "/" """, string, flags=re.VERBOSE, ) ) def _is_windows_path(string: str) -> bool: return bool( re.match( r""" ^ ( [a-z]:\\ | \\\\ ) """, string, flags=re.IGNORECASE | re.VERBOSE, ) ) def _is_filesystem_path(string: str) -> bool: return _is_posix_path(string) or _is_windows_path(string) def guess_scheme(url: str) -> str: if _is_filesystem_path(url): return _any_to_uri(url) return add_http_if_no_scheme(url) def strip_url( url: str, strip_credentials: bool = True, strip_default_port: bool = True, origin_only: bool = False, strip_fragment: bool = True, ) -> str: parsed_url = urlparse(url) netloc = parsed_url.netloc if (strip_credentials or origin_only) and ( parsed_url.username or parsed_url.password ): netloc = netloc.split("@")[-1] if ( strip_default_port and parsed_url.port and (parsed_url.scheme, parsed_url.port) in ( ("http", 80), ("https", 443), ("ftp", 21), ) ): netloc = netloc.replace(f":{parsed_url.port}", "") return urlunparse( ( parsed_url.scheme, netloc, "/" if origin_only else parsed_url.path, "" if origin_only else parsed_url.params, "" if origin_only else parsed_url.query, "" if strip_fragment else parsed_url.fragment, ) )
--- +++ @@ -1,3 +1,7 @@+""" +This module contains general purpose URL functions not found in the standard +library. +""" from __future__ import annotations @@ -37,6 +41,7 @@ def url_is_from_any_domain(url: UrlT, domains: Iterable[str]) -> bool: + """Return True if the url belongs to any of the given domains""" host = _parse_url(url).netloc.lower() if not host: return False @@ -45,17 +50,40 @@ def url_is_from_spider(url: UrlT, spider: type[Spider]) -> bool: + """Return True if the url belongs to the given spider""" return url_is_from_any_domain( url, [spider.name, *getattr(spider, "allowed_domains", [])] ) def url_has_any_extension(url: UrlT, extensions: Iterable[str]) -> bool: + """Return True if the url ends with one of the extensions provided""" lowercase_path = _parse_url(url).path.lower() return any(lowercase_path.endswith(ext) for ext in extensions) def escape_ajax(url: str) -> str: + """ + Return the crawlable url + + >>> escape_ajax("www.example.com/ajax.html#!key=value") + 'www.example.com/ajax.html?_escaped_fragment_=key%3Dvalue' + >>> escape_ajax("www.example.com/ajax.html?k1=v1&k2=v2#!key=value") + 'www.example.com/ajax.html?k1=v1&k2=v2&_escaped_fragment_=key%3Dvalue' + >>> escape_ajax("www.example.com/ajax.html?#!key=value") + 'www.example.com/ajax.html?_escaped_fragment_=key%3Dvalue' + >>> escape_ajax("www.example.com/ajax.html#!") + 'www.example.com/ajax.html?_escaped_fragment_=' + + URLs that are not "AJAX crawlable" (according to Google) returned as-is: + + >>> escape_ajax("www.example.com/ajax.html#key=value") + 'www.example.com/ajax.html#key=value' + >>> escape_ajax("www.example.com/ajax.html#") + 'www.example.com/ajax.html#' + >>> escape_ajax("www.example.com/ajax.html") + 'www.example.com/ajax.html' + """ warn( "escape_ajax() is deprecated and will be removed in a future Scrapy version.", ScrapyDeprecationWarning, @@ -68,6 +96,7 @@ def add_http_if_no_scheme(url: str) -> str: + """Add http as the default scheme if it is missing from the url.""" match = re.match(r"^\w+://", url, flags=re.IGNORECASE) if not match: parts = urlparse(url) @@ -120,6 +149,8 @@ def guess_scheme(url: str) -> str: + """Add an URL scheme if missing: file:// for filepath-like input or + http:// otherwise.""" if _is_filesystem_path(url): return _any_to_uri(url) return add_http_if_no_scheme(url) @@ -132,6 +163,15 @@ origin_only: bool = False, strip_fragment: bool = True, ) -> str: + """Strip URL string from some of its components: + + - ``strip_credentials`` removes "user:password@" + - ``strip_default_port`` removes ":80" (resp. ":443", ":21") + from http:// (resp. https://, ftp://) URLs + - ``origin_only`` replaces path component with "/", also dropping + query and fragment components ; it also strips credentials + - ``strip_fragment`` drops any #fragment component + """ parsed_url = urlparse(url) netloc = parsed_url.netloc @@ -161,4 +201,4 @@ "" if origin_only else parsed_url.query, "" if strip_fragment else parsed_url.fragment, ) - )+ )
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/url.py
Add docstrings to existing functions
from __future__ import annotations import warnings from typing import Any from pydispatch import dispatcher from twisted.internet.defer import Deferred from scrapy.exceptions import ScrapyDeprecationWarning from scrapy.utils import signal as _signal from scrapy.utils.defer import maybe_deferred_to_future class SignalManager: def __init__(self, sender: Any = dispatcher.Anonymous): self.sender: Any = sender def connect(self, receiver: Any, signal: Any, **kwargs: Any) -> None: kwargs.setdefault("sender", self.sender) dispatcher.connect(receiver, signal, **kwargs) def disconnect(self, receiver: Any, signal: Any, **kwargs: Any) -> None: kwargs.setdefault("sender", self.sender) dispatcher.disconnect(receiver, signal, **kwargs) def send_catch_log(self, signal: Any, **kwargs: Any) -> list[tuple[Any, Any]]: kwargs.setdefault("sender", self.sender) return _signal.send_catch_log(signal, **kwargs) def send_catch_log_deferred( self, signal: Any, **kwargs: Any ) -> Deferred[list[tuple[Any, Any]]]: # pragma: no cover kwargs.setdefault("sender", self.sender) warnings.warn( "send_catch_log_deferred() is deprecated, use send_catch_log_async() instead", ScrapyDeprecationWarning, stacklevel=2, ) return _signal._send_catch_log_deferred(signal, **kwargs) async def send_catch_log_async( self, signal: Any, **kwargs: Any ) -> list[tuple[Any, Any]]: # note that this returns exceptions instead of Failures in the second tuple member kwargs.setdefault("sender", self.sender) return await _signal.send_catch_log_async(signal, **kwargs) def disconnect_all(self, signal: Any, **kwargs: Any) -> None: kwargs.setdefault("sender", self.sender) _signal.disconnect_all(signal, **kwargs) async def wait_for(self, signal: Any) -> None: d: Deferred[None] = Deferred() def handle() -> None: self.disconnect(handle, signal) d.callback(None) self.connect(handle, signal) await maybe_deferred_to_future(d)
--- +++ @@ -16,20 +16,54 @@ self.sender: Any = sender def connect(self, receiver: Any, signal: Any, **kwargs: Any) -> None: + """ + Connect a receiver function to a signal. + + The signal can be any object, although Scrapy comes with some + predefined signals that are documented in the :ref:`topics-signals` + section. + + :param receiver: the function to be connected + :type receiver: collections.abc.Callable + + :param signal: the signal to connect to + :type signal: object + """ kwargs.setdefault("sender", self.sender) dispatcher.connect(receiver, signal, **kwargs) def disconnect(self, receiver: Any, signal: Any, **kwargs: Any) -> None: + """ + Disconnect a receiver function from a signal. This has the + opposite effect of the :meth:`connect` method, and the arguments + are the same. + """ kwargs.setdefault("sender", self.sender) dispatcher.disconnect(receiver, signal, **kwargs) def send_catch_log(self, signal: Any, **kwargs: Any) -> list[tuple[Any, Any]]: + """ + Send a signal, catch exceptions and log them. + + The keyword arguments are passed to the signal handlers (connected + through the :meth:`connect` method). + """ kwargs.setdefault("sender", self.sender) return _signal.send_catch_log(signal, **kwargs) def send_catch_log_deferred( self, signal: Any, **kwargs: Any ) -> Deferred[list[tuple[Any, Any]]]: # pragma: no cover + """ + Like :meth:`send_catch_log` but supports :ref:`asynchronous signal + handlers <signal-deferred>`. + + Returns a Deferred that gets fired once all signal handlers + have finished. Send a signal, catch exceptions and log them. + + The keyword arguments are passed to the signal handlers (connected + through the :meth:`connect` method). + """ kwargs.setdefault("sender", self.sender) warnings.warn( "send_catch_log_deferred() is deprecated, use send_catch_log_async() instead", @@ -41,15 +75,37 @@ async def send_catch_log_async( self, signal: Any, **kwargs: Any ) -> list[tuple[Any, Any]]: + """ + Like :meth:`send_catch_log` but supports :ref:`asynchronous signal + handlers <signal-deferred>`. + + Returns a coroutine that completes once all signal handlers + have finished. Send a signal, catch exceptions and log them. + + The keyword arguments are passed to the signal handlers (connected + through the :meth:`connect` method). + + .. versionadded:: 2.14 + """ # note that this returns exceptions instead of Failures in the second tuple member kwargs.setdefault("sender", self.sender) return await _signal.send_catch_log_async(signal, **kwargs) def disconnect_all(self, signal: Any, **kwargs: Any) -> None: + """ + Disconnect all receivers from the given signal. + + :param signal: the signal to disconnect from + :type signal: object + """ kwargs.setdefault("sender", self.sender) _signal.disconnect_all(signal, **kwargs) async def wait_for(self, signal: Any) -> None: + """Await the next *signal*. + + See :ref:`start-requests-lazy` for an example. + """ d: Deferred[None] = Deferred() def handle() -> None: @@ -57,4 +113,4 @@ d.callback(None) self.connect(handle, signal) - await maybe_deferred_to_future(d)+ await maybe_deferred_to_future(d)
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/signalmanager.py
Create docstrings for all classes and functions
from __future__ import annotations import re import string from pathlib import Path from typing import TYPE_CHECKING, Any if TYPE_CHECKING: from os import PathLike def render_templatefile(path: str | PathLike, **kwargs: Any) -> None: path_obj = Path(path) raw = path_obj.read_text("utf8") content = string.Template(raw).substitute(**kwargs) render_path = path_obj.with_suffix("") if path_obj.suffix == ".tmpl" else path_obj if path_obj.suffix == ".tmpl": path_obj.rename(render_path) render_path.write_text(content, "utf8") CAMELCASE_INVALID_CHARS = re.compile(r"[^a-zA-Z\d]") def string_camelcase(string: str) -> str: return CAMELCASE_INVALID_CHARS.sub("", string.title())
--- +++ @@ -1,3 +1,4 @@+"""Helper functions for working with templates""" from __future__ import annotations @@ -28,4 +29,13 @@ def string_camelcase(string: str) -> str: - return CAMELCASE_INVALID_CHARS.sub("", string.title())+ """Convert a word to its CamelCase version and remove invalid chars + + >>> string_camelcase('lost-pound') + 'LostPound' + + >>> string_camelcase('missing_images') + 'MissingImages' + + """ + return CAMELCASE_INVALID_CHARS.sub("", string.title())
https://raw.githubusercontent.com/scrapy/scrapy/HEAD/scrapy/utils/template.py
Please document this code using docstrings
# Copyright (c) Meta Platforms, Inc. and affiliates. # This software may be used and distributed according to the terms of the Llama 2 Community License Agreement. import math from dataclasses import dataclass from typing import Optional, Tuple import fairscale.nn.model_parallel.initialize as fs_init import torch import torch.nn.functional as F from fairscale.nn.model_parallel.layers import ( ColumnParallelLinear, ParallelEmbedding, RowParallelLinear, ) from torch import nn @dataclass class ModelArgs: dim: int = 4096 n_layers: int = 32 n_heads: int = 32 n_kv_heads: Optional[int] = None vocab_size: int = -1 # defined later by tokenizer multiple_of: int = 256 # make SwiGLU hidden layer size multiple of large power of 2 ffn_dim_multiplier: Optional[float] = None norm_eps: float = 1e-5 max_batch_size: int = 32 max_seq_len: int = 2048 class RMSNorm(torch.nn.Module): def __init__(self, dim: int, eps: float = 1e-6): super().__init__() self.eps = eps self.weight = nn.Parameter(torch.ones(dim)) def _norm(self, x): return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps) def forward(self, x): output = self._norm(x.float()).type_as(x) return output * self.weight def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0): freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim)) t = torch.arange(end, device=freqs.device) # type: ignore freqs = torch.outer(t, freqs).float() # type: ignore freqs_cis = torch.polar(torch.ones_like(freqs), freqs) # complex64 return freqs_cis def reshape_for_broadcast(freqs_cis: torch.Tensor, x: torch.Tensor): ndim = x.ndim assert 0 <= 1 < ndim assert freqs_cis.shape == (x.shape[1], x.shape[-1]) shape = [d if i == 1 or i == ndim - 1 else 1 for i, d in enumerate(x.shape)] return freqs_cis.view(*shape) def apply_rotary_emb( xq: torch.Tensor, xk: torch.Tensor, freqs_cis: torch.Tensor, ) -> Tuple[torch.Tensor, torch.Tensor]: xq_ = torch.view_as_complex(xq.float().reshape(*xq.shape[:-1], -1, 2)) xk_ = torch.view_as_complex(xk.float().reshape(*xk.shape[:-1], -1, 2)) freqs_cis = reshape_for_broadcast(freqs_cis, xq_) xq_out = torch.view_as_real(xq_ * freqs_cis).flatten(3) xk_out = torch.view_as_real(xk_ * freqs_cis).flatten(3) return xq_out.type_as(xq), xk_out.type_as(xk) def repeat_kv(x: torch.Tensor, n_rep: int) -> torch.Tensor: bs, slen, n_kv_heads, head_dim = x.shape if n_rep == 1: return x return ( x[:, :, :, None, :] .expand(bs, slen, n_kv_heads, n_rep, head_dim) .reshape(bs, slen, n_kv_heads * n_rep, head_dim) ) class Attention(nn.Module): def __init__(self, args: ModelArgs): super().__init__() self.n_kv_heads = args.n_heads if args.n_kv_heads is None else args.n_kv_heads model_parallel_size = fs_init.get_model_parallel_world_size() self.n_local_heads = args.n_heads // model_parallel_size self.n_local_kv_heads = self.n_kv_heads // model_parallel_size self.n_rep = self.n_local_heads // self.n_local_kv_heads self.head_dim = args.dim // args.n_heads self.wq = ColumnParallelLinear( args.dim, args.n_heads * self.head_dim, bias=False, gather_output=False, init_method=lambda x: x, ) self.wk = ColumnParallelLinear( args.dim, self.n_kv_heads * self.head_dim, bias=False, gather_output=False, init_method=lambda x: x, ) self.wv = ColumnParallelLinear( args.dim, self.n_kv_heads * self.head_dim, bias=False, gather_output=False, init_method=lambda x: x, ) self.wo = RowParallelLinear( args.n_heads * self.head_dim, args.dim, bias=False, input_is_parallel=True, init_method=lambda x: x, ) self.cache_k = torch.zeros( ( args.max_batch_size, args.max_seq_len, self.n_local_kv_heads, self.head_dim, ) ).cuda() self.cache_v = torch.zeros( ( args.max_batch_size, args.max_seq_len, self.n_local_kv_heads, self.head_dim, ) ).cuda() def forward( self, x: torch.Tensor, start_pos: int, freqs_cis: torch.Tensor, mask: Optional[torch.Tensor], ): bsz, seqlen, _ = x.shape xq, xk, xv = self.wq(x), self.wk(x), self.wv(x) xq = xq.view(bsz, seqlen, self.n_local_heads, self.head_dim) xk = xk.view(bsz, seqlen, self.n_local_kv_heads, self.head_dim) xv = xv.view(bsz, seqlen, self.n_local_kv_heads, self.head_dim) xq, xk = apply_rotary_emb(xq, xk, freqs_cis=freqs_cis) self.cache_k = self.cache_k.to(xq) self.cache_v = self.cache_v.to(xq) self.cache_k[:bsz, start_pos : start_pos + seqlen] = xk self.cache_v[:bsz, start_pos : start_pos + seqlen] = xv keys = self.cache_k[:bsz, : start_pos + seqlen] values = self.cache_v[:bsz, : start_pos + seqlen] # repeat k/v heads if n_kv_heads < n_heads keys = repeat_kv(keys, self.n_rep) # (bs, cache_len + seqlen, n_local_heads, head_dim) values = repeat_kv(values, self.n_rep) # (bs, cache_len + seqlen, n_local_heads, head_dim) xq = xq.transpose(1, 2) # (bs, n_local_heads, seqlen, head_dim) keys = keys.transpose(1, 2) # (bs, n_local_heads, cache_len + seqlen, head_dim) values = values.transpose(1, 2) # (bs, n_local_heads, cache_len + seqlen, head_dim) scores = torch.matmul(xq, keys.transpose(2, 3)) / math.sqrt(self.head_dim) if mask is not None: scores = scores + mask # (bs, n_local_heads, seqlen, cache_len + seqlen) scores = F.softmax(scores.float(), dim=-1).type_as(xq) output = torch.matmul(scores, values) # (bs, n_local_heads, seqlen, head_dim) output = output.transpose(1, 2).contiguous().view(bsz, seqlen, -1) return self.wo(output) class FeedForward(nn.Module): def __init__( self, dim: int, hidden_dim: int, multiple_of: int, ffn_dim_multiplier: Optional[float], ): super().__init__() hidden_dim = int(2 * hidden_dim / 3) # custom dim factor multiplier if ffn_dim_multiplier is not None: hidden_dim = int(ffn_dim_multiplier * hidden_dim) hidden_dim = multiple_of * ((hidden_dim + multiple_of - 1) // multiple_of) self.w1 = ColumnParallelLinear( dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x ) self.w2 = RowParallelLinear( hidden_dim, dim, bias=False, input_is_parallel=True, init_method=lambda x: x ) self.w3 = ColumnParallelLinear( dim, hidden_dim, bias=False, gather_output=False, init_method=lambda x: x ) def forward(self, x): return self.w2(F.silu(self.w1(x)) * self.w3(x)) class TransformerBlock(nn.Module): def __init__(self, layer_id: int, args: ModelArgs): super().__init__() self.n_heads = args.n_heads self.dim = args.dim self.head_dim = args.dim // args.n_heads self.attention = Attention(args) self.feed_forward = FeedForward( dim=args.dim, hidden_dim=4 * args.dim, multiple_of=args.multiple_of, ffn_dim_multiplier=args.ffn_dim_multiplier, ) self.layer_id = layer_id self.attention_norm = RMSNorm(args.dim, eps=args.norm_eps) self.ffn_norm = RMSNorm(args.dim, eps=args.norm_eps) def forward( self, x: torch.Tensor, start_pos: int, freqs_cis: torch.Tensor, mask: Optional[torch.Tensor], ): h = x + self.attention( self.attention_norm(x), start_pos, freqs_cis, mask ) out = h + self.feed_forward(self.ffn_norm(h)) return out class Transformer(nn.Module): def __init__(self, params: ModelArgs): super().__init__() self.params = params self.vocab_size = params.vocab_size self.n_layers = params.n_layers self.tok_embeddings = ParallelEmbedding( params.vocab_size, params.dim, init_method=lambda x: x ) self.layers = torch.nn.ModuleList() for layer_id in range(params.n_layers): self.layers.append(TransformerBlock(layer_id, params)) self.norm = RMSNorm(params.dim, eps=params.norm_eps) self.output = ColumnParallelLinear( params.dim, params.vocab_size, bias=False, init_method=lambda x: x ) self.freqs_cis = precompute_freqs_cis( # Note that self.params.max_seq_len is multiplied by 2 because the token limit for the Llama 2 generation of models is 4096. # Adding this multiplier instead of using 4096 directly allows for dynamism of token lengths while training or fine-tuning. self.params.dim // self.params.n_heads, self.params.max_seq_len * 2 ) @torch.inference_mode() def forward(self, tokens: torch.Tensor, start_pos: int): _bsz, seqlen = tokens.shape h = self.tok_embeddings(tokens) self.freqs_cis = self.freqs_cis.to(h.device) freqs_cis = self.freqs_cis[start_pos : start_pos + seqlen] mask = None if seqlen > 1: mask = torch.full( (seqlen, seqlen), float("-inf"), device=tokens.device ) mask = torch.triu(mask, diagonal=1) # When performing key-value caching, we compute the attention scores # only for the new sequence. Thus, the matrix of scores is of size # (seqlen, cache_len + seqlen), and the only masked entries are (i, j) for # j > cache_len + i, since row i corresponds to token cache_len + i. mask = torch.hstack([ torch.zeros((seqlen, start_pos), device=tokens.device), mask ]).type_as(h) for layer in self.layers: h = layer(h, start_pos, freqs_cis, mask) h = self.norm(h) output = self.output(h).float() return output
--- +++ @@ -33,19 +33,70 @@ class RMSNorm(torch.nn.Module): def __init__(self, dim: int, eps: float = 1e-6): + """ + Initialize the RMSNorm normalization layer. + + Args: + dim (int): The dimension of the input tensor. + eps (float, optional): A small value added to the denominator for numerical stability. Default is 1e-6. + + Attributes: + eps (float): A small value added to the denominator for numerical stability. + weight (nn.Parameter): Learnable scaling parameter. + + """ super().__init__() self.eps = eps self.weight = nn.Parameter(torch.ones(dim)) def _norm(self, x): + """ + Apply the RMSNorm normalization to the input tensor. + + Args: + x (torch.Tensor): The input tensor. + + Returns: + torch.Tensor: The normalized tensor. + + """ return x * torch.rsqrt(x.pow(2).mean(-1, keepdim=True) + self.eps) def forward(self, x): + """ + Forward pass through the RMSNorm layer. + + Args: + x (torch.Tensor): The input tensor. + + Returns: + torch.Tensor: The output tensor after applying RMSNorm. + + """ output = self._norm(x.float()).type_as(x) return output * self.weight def precompute_freqs_cis(dim: int, end: int, theta: float = 10000.0): + """ + Precompute the frequency tensor for complex exponentials (cis) with given dimensions. + + This function calculates a frequency tensor with complex exponentials using the given dimension 'dim' + and the end index 'end'. The 'theta' parameter scales the frequencies. + The returned tensor contains complex values in complex64 data type. + + Args: + dim (int): Dimension of the frequency tensor. + end (int): End index for precomputing frequencies. + theta (float, optional): Scaling factor for frequency computation. Defaults to 10000.0. + + Returns: + torch.Tensor: Precomputed frequency tensor with complex exponentials. + + + + + """ freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim)) t = torch.arange(end, device=freqs.device) # type: ignore freqs = torch.outer(t, freqs).float() # type: ignore @@ -54,6 +105,23 @@ def reshape_for_broadcast(freqs_cis: torch.Tensor, x: torch.Tensor): + """ + Reshape frequency tensor for broadcasting it with another tensor. + + This function reshapes the frequency tensor to have the same shape as the target tensor 'x' + for the purpose of broadcasting the frequency tensor during element-wise operations. + + Args: + freqs_cis (torch.Tensor): Frequency tensor to be reshaped. + x (torch.Tensor): Target tensor for broadcasting compatibility. + + Returns: + torch.Tensor: Reshaped frequency tensor. + + Raises: + AssertionError: If the frequency tensor doesn't match the expected shape. + AssertionError: If the target tensor 'x' doesn't have the expected number of dimensions. + """ ndim = x.ndim assert 0 <= 1 < ndim assert freqs_cis.shape == (x.shape[1], x.shape[-1]) @@ -66,6 +134,25 @@ xk: torch.Tensor, freqs_cis: torch.Tensor, ) -> Tuple[torch.Tensor, torch.Tensor]: + """ + Apply rotary embeddings to input tensors using the given frequency tensor. + + This function applies rotary embeddings to the given query 'xq' and key 'xk' tensors using the provided + frequency tensor 'freqs_cis'. The input tensors are reshaped as complex numbers, and the frequency tensor + is reshaped for broadcasting compatibility. The resulting tensors contain rotary embeddings and are + returned as real tensors. + + Args: + xq (torch.Tensor): Query tensor to apply rotary embeddings. + xk (torch.Tensor): Key tensor to apply rotary embeddings. + freqs_cis (torch.Tensor): Precomputed frequency tensor for complex exponentials. + + Returns: + Tuple[torch.Tensor, torch.Tensor]: Tuple of modified query tensor and key tensor with rotary embeddings. + + + + """ xq_ = torch.view_as_complex(xq.float().reshape(*xq.shape[:-1], -1, 2)) xk_ = torch.view_as_complex(xk.float().reshape(*xk.shape[:-1], -1, 2)) freqs_cis = reshape_for_broadcast(freqs_cis, xq_) @@ -75,6 +162,7 @@ def repeat_kv(x: torch.Tensor, n_rep: int) -> torch.Tensor: + """torch.repeat_interleave(x, dim=2, repeats=n_rep)""" bs, slen, n_kv_heads, head_dim = x.shape if n_rep == 1: return x @@ -86,7 +174,28 @@ class Attention(nn.Module): + """Multi-head attention module.""" def __init__(self, args: ModelArgs): + """ + Initialize the Attention module. + + Args: + args (ModelArgs): Model configuration parameters. + + Attributes: + n_kv_heads (int): Number of key and value heads. + n_local_heads (int): Number of local query heads. + n_local_kv_heads (int): Number of local key and value heads. + n_rep (int): Number of repetitions for local heads. + head_dim (int): Dimension size of each attention head. + wq (ColumnParallelLinear): Linear transformation for queries. + wk (ColumnParallelLinear): Linear transformation for keys. + wv (ColumnParallelLinear): Linear transformation for values. + wo (RowParallelLinear): Linear transformation for output. + cache_k (torch.Tensor): Cached keys for attention. + cache_v (torch.Tensor): Cached values for attention. + + """ super().__init__() self.n_kv_heads = args.n_heads if args.n_kv_heads is None else args.n_kv_heads model_parallel_size = fs_init.get_model_parallel_world_size() @@ -148,6 +257,19 @@ freqs_cis: torch.Tensor, mask: Optional[torch.Tensor], ): + """ + Forward pass of the attention module. + + Args: + x (torch.Tensor): Input tensor. + start_pos (int): Starting position for caching. + freqs_cis (torch.Tensor): Precomputed frequency tensor. + mask (torch.Tensor, optional): Attention mask tensor. + + Returns: + torch.Tensor: Output tensor after attention. + + """ bsz, seqlen, _ = x.shape xq, xk, xv = self.wq(x), self.wk(x), self.wv(x) @@ -190,6 +312,21 @@ multiple_of: int, ffn_dim_multiplier: Optional[float], ): + """ + Initialize the FeedForward module. + + Args: + dim (int): Input dimension. + hidden_dim (int): Hidden dimension of the feedforward layer. + multiple_of (int): Value to ensure hidden dimension is a multiple of this value. + ffn_dim_multiplier (float, optional): Custom multiplier for hidden dimension. Defaults to None. + + Attributes: + w1 (ColumnParallelLinear): Linear transformation for the first layer. + w2 (RowParallelLinear): Linear transformation for the second layer. + w3 (ColumnParallelLinear): Linear transformation for the third layer. + + """ super().__init__() hidden_dim = int(2 * hidden_dim / 3) # custom dim factor multiplier @@ -213,6 +350,24 @@ class TransformerBlock(nn.Module): def __init__(self, layer_id: int, args: ModelArgs): + """ + Initialize a TransformerBlock. + + Args: + layer_id (int): Identifier for the layer. + args (ModelArgs): Model configuration parameters. + + Attributes: + n_heads (int): Number of attention heads. + dim (int): Dimension size of the model. + head_dim (int): Dimension size of each attention head. + attention (Attention): Attention module. + feed_forward (FeedForward): FeedForward module. + layer_id (int): Identifier for the layer. + attention_norm (RMSNorm): Layer normalization for attention output. + ffn_norm (RMSNorm): Layer normalization for feedforward output. + + """ super().__init__() self.n_heads = args.n_heads self.dim = args.dim @@ -235,6 +390,19 @@ freqs_cis: torch.Tensor, mask: Optional[torch.Tensor], ): + """ + Perform a forward pass through the TransformerBlock. + + Args: + x (torch.Tensor): Input tensor. + start_pos (int): Starting position for attention caching. + freqs_cis (torch.Tensor): Precomputed cosine and sine frequencies. + mask (torch.Tensor, optional): Masking tensor for attention. Defaults to None. + + Returns: + torch.Tensor: Output tensor after applying attention and feedforward layers. + + """ h = x + self.attention( self.attention_norm(x), start_pos, freqs_cis, mask ) @@ -244,6 +412,23 @@ class Transformer(nn.Module): def __init__(self, params: ModelArgs): + """ + Initialize a Transformer model. + + Args: + params (ModelArgs): Model configuration parameters. + + Attributes: + params (ModelArgs): Model configuration parameters. + vocab_size (int): Vocabulary size. + n_layers (int): Number of layers in the model. + tok_embeddings (ParallelEmbedding): Token embeddings. + layers (torch.nn.ModuleList): List of Transformer blocks. + norm (RMSNorm): Layer normalization for the model output. + output (ColumnParallelLinear): Linear layer for final output. + freqs_cis (torch.Tensor): Precomputed cosine and sine frequencies. + + """ super().__init__() self.params = params self.vocab_size = params.vocab_size @@ -270,6 +455,17 @@ @torch.inference_mode() def forward(self, tokens: torch.Tensor, start_pos: int): + """ + Perform a forward pass through the Transformer model. + + Args: + tokens (torch.Tensor): Input token indices. + start_pos (int): Starting position for attention caching. + + Returns: + torch.Tensor: Output logits after applying the Transformer model. + + """ _bsz, seqlen = tokens.shape h = self.tok_embeddings(tokens) self.freqs_cis = self.freqs_cis.to(h.device) @@ -296,4 +492,4 @@ h = layer(h, start_pos, freqs_cis, mask) h = self.norm(h) output = self.output(h).float() - return output+ return output
https://raw.githubusercontent.com/meta-llama/llama/HEAD/llama/model.py
Write docstrings for utility functions
# Copyright (c) Meta Platforms, Inc. and affiliates. # This software may be used and distributed according to the terms of the Llama 2 Community License Agreement. import json import os import sys import time from pathlib import Path from typing import List, Literal, Optional, Tuple, TypedDict import torch import torch.nn.functional as F from fairscale.nn.model_parallel.initialize import ( get_model_parallel_rank, initialize_model_parallel, model_parallel_is_initialized, ) from llama.model import ModelArgs, Transformer from llama.tokenizer import Tokenizer Role = Literal["system", "user", "assistant"] class Message(TypedDict): role: Role content: str class CompletionPrediction(TypedDict, total=False): generation: str tokens: List[str] # not required logprobs: List[float] # not required class ChatPrediction(TypedDict, total=False): generation: Message tokens: List[str] # not required logprobs: List[float] # not required Dialog = List[Message] B_INST, E_INST = "[INST]", "[/INST]" B_SYS, E_SYS = "<<SYS>>\n", "\n<</SYS>>\n\n" SPECIAL_TAGS = [B_INST, E_INST, "<<SYS>>", "<</SYS>>"] UNSAFE_ERROR = "Error: special tags are not allowed as part of the prompt." class Llama: @staticmethod def build( ckpt_dir: str, tokenizer_path: str, max_seq_len: int, max_batch_size: int, model_parallel_size: Optional[int] = None, seed: int = 1, ) -> "Llama": if not torch.distributed.is_initialized(): torch.distributed.init_process_group("nccl") if not model_parallel_is_initialized(): if model_parallel_size is None: model_parallel_size = int(os.environ.get("WORLD_SIZE", 1)) initialize_model_parallel(model_parallel_size) local_rank = int(os.environ.get("LOCAL_RANK", 0)) torch.cuda.set_device(local_rank) # seed must be the same in all processes torch.manual_seed(seed) if local_rank > 0: sys.stdout = open(os.devnull, "w") start_time = time.time() checkpoints = sorted(Path(ckpt_dir).glob("*.pth")) assert len(checkpoints) > 0, f"no checkpoint files found in {ckpt_dir}" assert model_parallel_size == len( checkpoints ), f"Loading a checkpoint for MP={len(checkpoints)} but world size is {model_parallel_size}" ckpt_path = checkpoints[get_model_parallel_rank()] checkpoint = torch.load(ckpt_path, map_location="cpu") with open(Path(ckpt_dir) / "params.json", "r") as f: params = json.loads(f.read()) model_args: ModelArgs = ModelArgs( max_seq_len=max_seq_len, max_batch_size=max_batch_size, **params, ) tokenizer = Tokenizer(model_path=tokenizer_path) model_args.vocab_size = tokenizer.n_words torch.set_default_tensor_type(torch.cuda.HalfTensor) model = Transformer(model_args) model.load_state_dict(checkpoint, strict=False) print(f"Loaded in {time.time() - start_time:.2f} seconds") return Llama(model, tokenizer) def __init__(self, model: Transformer, tokenizer: Tokenizer): self.model = model self.tokenizer = tokenizer @torch.inference_mode() def generate( self, prompt_tokens: List[List[int]], max_gen_len: int, temperature: float = 0.6, top_p: float = 0.9, logprobs: bool = False, echo: bool = False, ) -> Tuple[List[List[int]], Optional[List[List[float]]]]: params = self.model.params bsz = len(prompt_tokens) assert bsz <= params.max_batch_size, (bsz, params.max_batch_size) min_prompt_len = min(len(t) for t in prompt_tokens) max_prompt_len = max(len(t) for t in prompt_tokens) assert max_prompt_len <= params.max_seq_len total_len = min(params.max_seq_len, max_gen_len + max_prompt_len) pad_id = self.tokenizer.pad_id tokens = torch.full((bsz, total_len), pad_id, dtype=torch.long, device="cuda") for k, t in enumerate(prompt_tokens): tokens[k, : len(t)] = torch.tensor(t, dtype=torch.long, device="cuda") if logprobs: token_logprobs = torch.zeros_like(tokens, dtype=torch.float) prev_pos = 0 eos_reached = torch.tensor([False] * bsz, device="cuda") input_text_mask = tokens != pad_id if min_prompt_len == total_len: logits = self.model.forward(tokens, prev_pos) token_logprobs = -F.cross_entropy( input=logits.transpose(1, 2), target=tokens, reduction="none", ignore_index=pad_id, ) for cur_pos in range(min_prompt_len, total_len): logits = self.model.forward(tokens[:, prev_pos:cur_pos], prev_pos) if temperature > 0: probs = torch.softmax(logits[:, -1] / temperature, dim=-1) next_token = sample_top_p(probs, top_p) else: next_token = torch.argmax(logits[:, -1], dim=-1) next_token = next_token.reshape(-1) # only replace token if prompt has already been generated next_token = torch.where( input_text_mask[:, cur_pos], tokens[:, cur_pos], next_token ) tokens[:, cur_pos] = next_token if logprobs: token_logprobs[:, prev_pos + 1 : cur_pos + 1] = -F.cross_entropy( input=logits.transpose(1, 2), target=tokens[:, prev_pos + 1 : cur_pos + 1], reduction="none", ignore_index=pad_id, ) eos_reached |= (~input_text_mask[:, cur_pos]) & ( next_token == self.tokenizer.eos_id ) prev_pos = cur_pos if all(eos_reached): break if logprobs: token_logprobs = token_logprobs.tolist() out_tokens, out_logprobs = [], [] for i, toks in enumerate(tokens.tolist()): # cut to max gen len start = 0 if echo else len(prompt_tokens[i]) toks = toks[start : len(prompt_tokens[i]) + max_gen_len] probs = None if logprobs: probs = token_logprobs[i][start : len(prompt_tokens[i]) + max_gen_len] # cut to eos tok if any if self.tokenizer.eos_id in toks: eos_idx = toks.index(self.tokenizer.eos_id) toks = toks[:eos_idx] probs = probs[:eos_idx] if logprobs else None out_tokens.append(toks) out_logprobs.append(probs) return (out_tokens, out_logprobs if logprobs else None) def text_completion( self, prompts: List[str], temperature: float = 0.6, top_p: float = 0.9, max_gen_len: Optional[int] = None, logprobs: bool = False, echo: bool = False, ) -> List[CompletionPrediction]: if max_gen_len is None: max_gen_len = self.model.params.max_seq_len - 1 prompt_tokens = [self.tokenizer.encode(x, bos=True, eos=False) for x in prompts] generation_tokens, generation_logprobs = self.generate( prompt_tokens=prompt_tokens, max_gen_len=max_gen_len, temperature=temperature, top_p=top_p, logprobs=logprobs, echo=echo, ) if logprobs: return [ { "generation": self.tokenizer.decode(t), "tokens": [self.tokenizer.decode(x) for x in t], "logprobs": logprobs_i, } for t, logprobs_i in zip(generation_tokens, generation_logprobs) ] return [{"generation": self.tokenizer.decode(t)} for t in generation_tokens] def chat_completion( self, dialogs: List[Dialog], temperature: float = 0.6, top_p: float = 0.9, max_gen_len: Optional[int] = None, logprobs: bool = False, ) -> List[ChatPrediction]: if max_gen_len is None: max_gen_len = self.model.params.max_seq_len - 1 prompt_tokens = [] unsafe_requests = [] for dialog in dialogs: unsafe_requests.append( any([tag in msg["content"] for tag in SPECIAL_TAGS for msg in dialog]) ) if dialog[0]["role"] == "system": dialog = [ { "role": dialog[1]["role"], "content": B_SYS + dialog[0]["content"] + E_SYS + dialog[1]["content"], } ] + dialog[2:] assert all([msg["role"] == "user" for msg in dialog[::2]]) and all( [msg["role"] == "assistant" for msg in dialog[1::2]] ), ( "model only supports 'system', 'user' and 'assistant' roles, " "starting with 'system', then 'user' and alternating (u/a/u/a/u...)" ) dialog_tokens: List[int] = sum( [ self.tokenizer.encode( f"{B_INST} {(prompt['content']).strip()} {E_INST} {(answer['content']).strip()} ", bos=True, eos=True, ) for prompt, answer in zip( dialog[::2], dialog[1::2], ) ], [], ) assert ( dialog[-1]["role"] == "user" ), f"Last message must be from user, got {dialog[-1]['role']}" dialog_tokens += self.tokenizer.encode( f"{B_INST} {(dialog[-1]['content']).strip()} {E_INST}", bos=True, eos=False, ) prompt_tokens.append(dialog_tokens) generation_tokens, generation_logprobs = self.generate( prompt_tokens=prompt_tokens, max_gen_len=max_gen_len, temperature=temperature, top_p=top_p, logprobs=logprobs, ) if logprobs: return [ { "generation": { "role": "assistant", "content": self.tokenizer.decode(t) if not unsafe else UNSAFE_ERROR, }, "tokens": [self.tokenizer.decode(x) for x in t], "logprobs": logprobs_i, } for t, logprobs_i, unsafe in zip( generation_tokens, generation_logprobs, unsafe_requests ) ] return [ { "generation": { "role": "assistant", "content": self.tokenizer.decode(t) if not unsafe else UNSAFE_ERROR, } } for t, unsafe in zip(generation_tokens, unsafe_requests) ] def sample_top_p(probs, p): probs_sort, probs_idx = torch.sort(probs, dim=-1, descending=True) probs_sum = torch.cumsum(probs_sort, dim=-1) mask = probs_sum - probs_sort > p probs_sort[mask] = 0.0 probs_sort.div_(probs_sort.sum(dim=-1, keepdim=True)) next_token = torch.multinomial(probs_sort, num_samples=1) next_token = torch.gather(probs_idx, -1, next_token) return next_token
--- +++ @@ -58,6 +58,29 @@ model_parallel_size: Optional[int] = None, seed: int = 1, ) -> "Llama": + """ + Build a Llama instance by initializing and loading a pre-trained model. + + Args: + ckpt_dir (str): Path to the directory containing checkpoint files. + tokenizer_path (str): Path to the tokenizer file. + max_seq_len (int): Maximum sequence length for input text. + max_batch_size (int): Maximum batch size for inference. + model_parallel_size (Optional[int], optional): Number of model parallel processes. + If not provided, it's determined from the environment. Defaults to None. + + Returns: + Llama: An instance of the Llama class with the loaded model and tokenizer. + + Raises: + AssertionError: If there are no checkpoint files in the specified directory, + or if the model parallel size does not match the number of checkpoint files. + + Note: + This method initializes the distributed process group, sets the device to CUDA, + and loads the pre-trained model and tokenizer. + + """ if not torch.distributed.is_initialized(): torch.distributed.init_process_group("nccl") if not model_parallel_is_initialized(): @@ -113,6 +136,25 @@ logprobs: bool = False, echo: bool = False, ) -> Tuple[List[List[int]], Optional[List[List[float]]]]: + """ + Generate text sequences based on provided prompts using the language generation model. + + Args: + prompt_tokens (List[List[int]]): List of tokenized prompts, where each prompt is represented as a list of integers. + max_gen_len (int): Maximum length of the generated text sequence. + temperature (float, optional): Temperature value for controlling randomness in sampling. Defaults to 0.6. + top_p (float, optional): Top-p probability threshold for nucleus sampling. Defaults to 0.9. + logprobs (bool, optional): Flag indicating whether to compute token log probabilities. Defaults to False. + echo (bool, optional): Flag indicating whether to include prompt tokens in the generated output. Defaults to False. + + Returns: + Tuple[List[List[int]], Optional[List[List[float]]]]: A tuple containing generated token sequences and, if logprobs is True, corresponding token log probabilities. + + Note: + This method uses the provided prompts as a basis for generating text. It employs nucleus sampling to produce text with controlled randomness. + If logprobs is True, token log probabilities are computed for each generated token. + + """ params = self.model.params bsz = len(prompt_tokens) assert bsz <= params.max_batch_size, (bsz, params.max_batch_size) @@ -197,6 +239,26 @@ logprobs: bool = False, echo: bool = False, ) -> List[CompletionPrediction]: + """ + Perform text completion for a list of prompts using the language generation model. + + Args: + prompts (List[str]): List of text prompts for completion. + temperature (float, optional): Temperature value for controlling randomness in sampling. Defaults to 0.6. + top_p (float, optional): Top-p probability threshold for nucleus sampling. Defaults to 0.9. + max_gen_len (Optional[int], optional): Maximum length of the generated completion sequence. + If not provided, it's set to the model's maximum sequence length minus 1. + logprobs (bool, optional): Flag indicating whether to compute token log probabilities. Defaults to False. + echo (bool, optional): Flag indicating whether to include prompt tokens in the generated output. Defaults to False. + + Returns: + List[CompletionPrediction]: List of completion predictions, each containing the generated text completion. + + Note: + This method generates text completions for the provided prompts, employing nucleus sampling to introduce controlled randomness. + If logprobs is True, token log probabilities are computed for each generated token. + + """ if max_gen_len is None: max_gen_len = self.model.params.max_seq_len - 1 prompt_tokens = [self.tokenizer.encode(x, bos=True, eos=False) for x in prompts] @@ -227,6 +289,30 @@ max_gen_len: Optional[int] = None, logprobs: bool = False, ) -> List[ChatPrediction]: + """ + Generate assistant responses for a list of conversational dialogs using the language generation model. + + Args: + dialogs (List[Dialog]): List of conversational dialogs, where each dialog is a list of messages. + temperature (float, optional): Temperature value for controlling randomness in sampling. Defaults to 0.6. + top_p (float, optional): Top-p probability threshold for nucleus sampling. Defaults to 0.9. + max_gen_len (Optional[int], optional): Maximum length of the generated response sequence. + If not provided, it's set to the model's maximum sequence length minus 1. + logprobs (bool, optional): Flag indicating whether to compute token log probabilities. Defaults to False. + + Returns: + List[ChatPrediction]: List of chat predictions, each containing the assistant's generated response. + + Raises: + AssertionError: If the last message in a dialog is not from the user. + AssertionError: If the dialog roles are not in the required 'user', 'assistant', and optional 'system' order. + + Note: + This method generates assistant responses for the provided conversational dialogs. + It employs nucleus sampling to introduce controlled randomness in text generation. + If logprobs is True, token log probabilities are computed for each generated token. + + """ if max_gen_len is None: max_gen_len = self.model.params.max_seq_len - 1 prompt_tokens = [] @@ -310,6 +396,21 @@ def sample_top_p(probs, p): + """ + Perform top-p (nucleus) sampling on a probability distribution. + + Args: + probs (torch.Tensor): Probability distribution tensor. + p (float): Probability threshold for top-p sampling. + + Returns: + torch.Tensor: Sampled token indices. + + Note: + Top-p sampling selects the smallest set of tokens whose cumulative probability mass + exceeds the threshold p. The distribution is renormalized based on the selected tokens. + + """ probs_sort, probs_idx = torch.sort(probs, dim=-1, descending=True) probs_sum = torch.cumsum(probs_sort, dim=-1) mask = probs_sum - probs_sort > p @@ -317,4 +418,4 @@ probs_sort.div_(probs_sort.sum(dim=-1, keepdim=True)) next_token = torch.multinomial(probs_sort, num_samples=1) next_token = torch.gather(probs_idx, -1, next_token) - return next_token+ return next_token
https://raw.githubusercontent.com/meta-llama/llama/HEAD/llama/generation.py
Generate documentation strings for clarity
# Copyright (c) Meta Platforms, Inc. and affiliates. # This software may be used and distributed according to the terms of the Llama 2 Community License Agreement. import os from logging import getLogger from typing import List from sentencepiece import SentencePieceProcessor logger = getLogger() class Tokenizer: def __init__(self, model_path: str): # reload tokenizer assert os.path.isfile(model_path), model_path self.sp_model = SentencePieceProcessor(model_file=model_path) logger.info(f"Reloaded SentencePiece model from {model_path}") # BOS / EOS token IDs self.n_words: int = self.sp_model.vocab_size() self.bos_id: int = self.sp_model.bos_id() self.eos_id: int = self.sp_model.eos_id() self.pad_id: int = self.sp_model.pad_id() logger.info( f"#words: {self.n_words} - BOS ID: {self.bos_id} - EOS ID: {self.eos_id}" ) assert self.sp_model.vocab_size() == self.sp_model.get_piece_size() def encode(self, s: str, bos: bool, eos: bool) -> List[int]: assert type(s) is str t = self.sp_model.encode(s) if bos: t = [self.bos_id] + t if eos: t = t + [self.eos_id] return t def decode(self, t: List[int]) -> str: return self.sp_model.decode(t)
--- +++ @@ -12,7 +12,14 @@ class Tokenizer: + """tokenizing and encoding/decoding text using SentencePiece.""" def __init__(self, model_path: str): + """ + Initializes the Tokenizer with a SentencePiece model. + + Args: + model_path (str): The path to the SentencePiece model file. + """ # reload tokenizer assert os.path.isfile(model_path), model_path self.sp_model = SentencePieceProcessor(model_file=model_path) @@ -29,6 +36,17 @@ assert self.sp_model.vocab_size() == self.sp_model.get_piece_size() def encode(self, s: str, bos: bool, eos: bool) -> List[int]: + """ + Encodes a string into a list of token IDs. + + Args: + s (str): The input string to be encoded. + bos (bool): Whether to prepend the beginning-of-sequence token. + eos (bool): Whether to append the end-of-sequence token. + + Returns: + List[int]: A list of token IDs. + """ assert type(s) is str t = self.sp_model.encode(s) if bos: @@ -38,4 +56,13 @@ return t def decode(self, t: List[int]) -> str: - return self.sp_model.decode(t)+ """ + Decodes a list of token IDs into a string. + + Args: + t (List[int]): The list of token IDs to be decoded. + + Returns: + str: The decoded string. + """ + return self.sp_model.decode(t)
https://raw.githubusercontent.com/meta-llama/llama/HEAD/llama/tokenizer.py
Expand my code with proper documentation strings
import os import re import typing from typing import Any, TextIO from yaml import SafeLoader _env_replace_matcher = re.compile(r"\$\{(\w|_)+:?.*}") @typing.no_type_check # pyaml does not have good hints, everything is Any def load_yaml_with_envvars( stream: TextIO, environ: dict[str, Any] = os.environ ) -> dict[str, Any]: loader = SafeLoader(stream) def load_env_var(_, node) -> str: value = str(node.value).removeprefix("${").removesuffix("}") split = value.split(":", 1) env_var = split[0] value = environ.get(env_var) default = None if len(split) == 1 else split[1] if value is None and default is None: raise ValueError( f"Environment variable {env_var} is not set and not default was provided" ) return value or default loader.add_implicit_resolver("env_var_replacer", _env_replace_matcher, None) loader.add_constructor("env_var_replacer", load_env_var) try: return loader.get_single_data() finally: loader.dispose()
--- +++ @@ -12,9 +12,15 @@ def load_yaml_with_envvars( stream: TextIO, environ: dict[str, Any] = os.environ ) -> dict[str, Any]: + """Load yaml file with environment variable expansion. + + The pattern ${VAR} or ${VAR:default} will be replaced with + the value of the environment variable. + """ loader = SafeLoader(stream) def load_env_var(_, node) -> str: + """Extract the matched value, expand env variable, and replace the match.""" value = str(node.value).removeprefix("${").removesuffix("}") split = value.split(":", 1) env_var = split[0] @@ -32,4 +38,4 @@ try: return loader.get_single_data() finally: - loader.dispose()+ loader.dispose()
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/settings/yaml.py
Replace inline comments with docstrings
import base64 import logging import time from collections.abc import Iterable from enum import Enum from pathlib import Path from typing import Any import gradio as gr # type: ignore from fastapi import FastAPI from gradio.themes.utils.colors import slate # type: ignore from injector import inject, singleton from llama_index.core.llms import ChatMessage, ChatResponse, MessageRole from llama_index.core.types import TokenGen from pydantic import BaseModel from private_gpt.constants import PROJECT_ROOT_PATH from private_gpt.di import global_injector from private_gpt.open_ai.extensions.context_filter import ContextFilter from private_gpt.server.chat.chat_service import ChatService, CompletionGen from private_gpt.server.chunks.chunks_service import Chunk, ChunksService from private_gpt.server.ingest.ingest_service import IngestService from private_gpt.server.recipes.summarize.summarize_service import SummarizeService from private_gpt.settings.settings import settings from private_gpt.ui.images import logo_svg logger = logging.getLogger(__name__) THIS_DIRECTORY_RELATIVE = Path(__file__).parent.relative_to(PROJECT_ROOT_PATH) # Should be "private_gpt/ui/avatar-bot.ico" AVATAR_BOT = THIS_DIRECTORY_RELATIVE / "avatar-bot.ico" UI_TAB_TITLE = "My Private GPT" SOURCES_SEPARATOR = "<hr>Sources: \n" class Modes(str, Enum): RAG_MODE = "RAG" SEARCH_MODE = "Search" BASIC_CHAT_MODE = "Basic" SUMMARIZE_MODE = "Summarize" MODES: list[Modes] = [ Modes.RAG_MODE, Modes.SEARCH_MODE, Modes.BASIC_CHAT_MODE, Modes.SUMMARIZE_MODE, ] class Source(BaseModel): file: str page: str text: str class Config: frozen = True @staticmethod def curate_sources(sources: list[Chunk]) -> list["Source"]: curated_sources = [] for chunk in sources: doc_metadata = chunk.document.doc_metadata file_name = doc_metadata.get("file_name", "-") if doc_metadata else "-" page_label = doc_metadata.get("page_label", "-") if doc_metadata else "-" source = Source(file=file_name, page=page_label, text=chunk.text) curated_sources.append(source) curated_sources = list( dict.fromkeys(curated_sources).keys() ) # Unique sources only return curated_sources @singleton class PrivateGptUi: @inject def __init__( self, ingest_service: IngestService, chat_service: ChatService, chunks_service: ChunksService, summarizeService: SummarizeService, ) -> None: self._ingest_service = ingest_service self._chat_service = chat_service self._chunks_service = chunks_service self._summarize_service = summarizeService # Cache the UI blocks self._ui_block = None self._selected_filename = None # Initialize system prompt based on default mode default_mode_map = {mode.value: mode for mode in Modes} self._default_mode = default_mode_map.get( settings().ui.default_mode, Modes.RAG_MODE ) self._system_prompt = self._get_default_system_prompt(self._default_mode) def _chat( self, message: str, history: list[list[str]], mode: Modes, *_: Any ) -> Any: def yield_deltas(completion_gen: CompletionGen) -> Iterable[str]: full_response: str = "" stream = completion_gen.response for delta in stream: if isinstance(delta, str): full_response += str(delta) elif isinstance(delta, ChatResponse): full_response += delta.delta or "" yield full_response time.sleep(0.02) if completion_gen.sources: full_response += SOURCES_SEPARATOR cur_sources = Source.curate_sources(completion_gen.sources) sources_text = "\n\n\n" used_files = set() for index, source in enumerate(cur_sources, start=1): if f"{source.file}-{source.page}" not in used_files: sources_text = ( sources_text + f"{index}. {source.file} (page {source.page}) \n\n" ) used_files.add(f"{source.file}-{source.page}") sources_text += "<hr>\n\n" full_response += sources_text yield full_response def yield_tokens(token_gen: TokenGen) -> Iterable[str]: full_response: str = "" for token in token_gen: full_response += str(token) yield full_response def build_history() -> list[ChatMessage]: history_messages: list[ChatMessage] = [] for interaction in history: history_messages.append( ChatMessage(content=interaction[0], role=MessageRole.USER) ) if len(interaction) > 1 and interaction[1] is not None: history_messages.append( ChatMessage( # Remove from history content the Sources information content=interaction[1].split(SOURCES_SEPARATOR)[0], role=MessageRole.ASSISTANT, ) ) # max 20 messages to try to avoid context overflow return history_messages[:20] new_message = ChatMessage(content=message, role=MessageRole.USER) all_messages = [*build_history(), new_message] # If a system prompt is set, add it as a system message if self._system_prompt: all_messages.insert( 0, ChatMessage( content=self._system_prompt, role=MessageRole.SYSTEM, ), ) match mode: case Modes.RAG_MODE: # Use only the selected file for the query context_filter = None if self._selected_filename is not None: docs_ids = [] for ingested_document in self._ingest_service.list_ingested(): if ( ingested_document.doc_metadata["file_name"] == self._selected_filename ): docs_ids.append(ingested_document.doc_id) context_filter = ContextFilter(docs_ids=docs_ids) query_stream = self._chat_service.stream_chat( messages=all_messages, use_context=True, context_filter=context_filter, ) yield from yield_deltas(query_stream) case Modes.BASIC_CHAT_MODE: llm_stream = self._chat_service.stream_chat( messages=all_messages, use_context=False, ) yield from yield_deltas(llm_stream) case Modes.SEARCH_MODE: response = self._chunks_service.retrieve_relevant( text=message, limit=4, prev_next_chunks=0 ) sources = Source.curate_sources(response) yield "\n\n\n".join( f"{index}. **{source.file} " f"(page {source.page})**\n " f"{source.text}" for index, source in enumerate(sources, start=1) ) case Modes.SUMMARIZE_MODE: # Summarize the given message, optionally using selected files context_filter = None if self._selected_filename: docs_ids = [] for ingested_document in self._ingest_service.list_ingested(): if ( ingested_document.doc_metadata["file_name"] == self._selected_filename ): docs_ids.append(ingested_document.doc_id) context_filter = ContextFilter(docs_ids=docs_ids) summary_stream = self._summarize_service.stream_summarize( use_context=True, context_filter=context_filter, instructions=message, ) yield from yield_tokens(summary_stream) # On initialization and on mode change, this function set the system prompt # to the default prompt based on the mode (and user settings). @staticmethod def _get_default_system_prompt(mode: Modes) -> str: p = "" match mode: # For query chat mode, obtain default system prompt from settings case Modes.RAG_MODE: p = settings().ui.default_query_system_prompt # For chat mode, obtain default system prompt from settings case Modes.BASIC_CHAT_MODE: p = settings().ui.default_chat_system_prompt # For summarization mode, obtain default system prompt from settings case Modes.SUMMARIZE_MODE: p = settings().ui.default_summarization_system_prompt # For any other mode, clear the system prompt case _: p = "" return p @staticmethod def _get_default_mode_explanation(mode: Modes) -> str: match mode: case Modes.RAG_MODE: return "Get contextualized answers from selected files." case Modes.SEARCH_MODE: return "Find relevant chunks of text in selected files." case Modes.BASIC_CHAT_MODE: return "Chat with the LLM using its training data. Files are ignored." case Modes.SUMMARIZE_MODE: return "Generate a summary of the selected files. Prompt to customize the result." case _: return "" def _set_system_prompt(self, system_prompt_input: str) -> None: logger.info(f"Setting system prompt to: {system_prompt_input}") self._system_prompt = system_prompt_input def _set_explanatation_mode(self, explanation_mode: str) -> None: self._explanation_mode = explanation_mode def _set_current_mode(self, mode: Modes) -> Any: self.mode = mode self._set_system_prompt(self._get_default_system_prompt(mode)) self._set_explanatation_mode(self._get_default_mode_explanation(mode)) interactive = self._system_prompt is not None return [ gr.update(placeholder=self._system_prompt, interactive=interactive), gr.update(value=self._explanation_mode), ] def _list_ingested_files(self) -> list[list[str]]: files = set() for ingested_document in self._ingest_service.list_ingested(): if ingested_document.doc_metadata is None: # Skipping documents without metadata continue file_name = ingested_document.doc_metadata.get( "file_name", "[FILE NAME MISSING]" ) files.add(file_name) return [[row] for row in files] def _upload_file(self, files: list[str]) -> None: logger.debug("Loading count=%s files", len(files)) paths = [Path(file) for file in files] # remove all existing Documents with name identical to a new file upload: file_names = [path.name for path in paths] doc_ids_to_delete = [] for ingested_document in self._ingest_service.list_ingested(): if ( ingested_document.doc_metadata and ingested_document.doc_metadata["file_name"] in file_names ): doc_ids_to_delete.append(ingested_document.doc_id) if len(doc_ids_to_delete) > 0: logger.info( "Uploading file(s) which were already ingested: %s document(s) will be replaced.", len(doc_ids_to_delete), ) for doc_id in doc_ids_to_delete: self._ingest_service.delete(doc_id) self._ingest_service.bulk_ingest([(str(path.name), path) for path in paths]) def _delete_all_files(self) -> Any: ingested_files = self._ingest_service.list_ingested() logger.debug("Deleting count=%s files", len(ingested_files)) for ingested_document in ingested_files: self._ingest_service.delete(ingested_document.doc_id) return [ gr.List(self._list_ingested_files()), gr.components.Button(interactive=False), gr.components.Button(interactive=False), gr.components.Textbox("All files"), ] def _delete_selected_file(self) -> Any: logger.debug("Deleting selected %s", self._selected_filename) # Note: keep looping for pdf's (each page became a Document) for ingested_document in self._ingest_service.list_ingested(): if ( ingested_document.doc_metadata and ingested_document.doc_metadata["file_name"] == self._selected_filename ): self._ingest_service.delete(ingested_document.doc_id) return [ gr.List(self._list_ingested_files()), gr.components.Button(interactive=False), gr.components.Button(interactive=False), gr.components.Textbox("All files"), ] def _deselect_selected_file(self) -> Any: self._selected_filename = None return [ gr.components.Button(interactive=False), gr.components.Button(interactive=False), gr.components.Textbox("All files"), ] def _selected_a_file(self, select_data: gr.SelectData) -> Any: self._selected_filename = select_data.value return [ gr.components.Button(interactive=True), gr.components.Button(interactive=True), gr.components.Textbox(self._selected_filename), ] def _build_ui_blocks(self) -> gr.Blocks: logger.debug("Creating the UI blocks") with gr.Blocks( title=UI_TAB_TITLE, theme=gr.themes.Soft(primary_hue=slate), css=".logo { " "display:flex;" "background-color: #C7BAFF;" "height: 80px;" "border-radius: 8px;" "align-content: center;" "justify-content: center;" "align-items: center;" "}" ".logo img { height: 25% }" ".contain { display: flex !important; flex-direction: column !important; }" "#component-0, #component-3, #component-10, #component-8 { height: 100% !important; }" "#chatbot { flex-grow: 1 !important; overflow: auto !important;}" "#col { height: calc(100vh - 112px - 16px) !important; }" "hr { margin-top: 1em; margin-bottom: 1em; border: 0; border-top: 1px solid #FFF; }" ".avatar-image { background-color: antiquewhite; border-radius: 2px; }" ".footer { text-align: center; margin-top: 20px; font-size: 14px; display: flex; align-items: center; justify-content: center; }" ".footer-zylon-link { display:flex; margin-left: 5px; text-decoration: auto; color: var(--body-text-color); }" ".footer-zylon-link:hover { color: #C7BAFF; }" ".footer-zylon-ico { height: 20px; margin-left: 5px; background-color: antiquewhite; border-radius: 2px; }", ) as blocks: with gr.Row(): gr.HTML(f"<div class='logo'/><img src={logo_svg} alt=PrivateGPT></div") with gr.Row(equal_height=False): with gr.Column(scale=3): default_mode = self._default_mode mode = gr.Radio( [mode.value for mode in MODES], label="Mode", value=default_mode, ) explanation_mode = gr.Textbox( placeholder=self._get_default_mode_explanation(default_mode), show_label=False, max_lines=3, interactive=False, ) upload_button = gr.components.UploadButton( "Upload File(s)", type="filepath", file_count="multiple", size="sm", ) ingested_dataset = gr.List( self._list_ingested_files, headers=["File name"], label="Ingested Files", height=235, interactive=False, render=False, # Rendered under the button ) upload_button.upload( self._upload_file, inputs=upload_button, outputs=ingested_dataset, ) ingested_dataset.change( self._list_ingested_files, outputs=ingested_dataset, ) ingested_dataset.render() deselect_file_button = gr.components.Button( "De-select selected file", size="sm", interactive=False ) selected_text = gr.components.Textbox( "All files", label="Selected for Query or Deletion", max_lines=1 ) delete_file_button = gr.components.Button( "🗑️ Delete selected file", size="sm", visible=settings().ui.delete_file_button_enabled, interactive=False, ) delete_files_button = gr.components.Button( "⚠️ Delete ALL files", size="sm", visible=settings().ui.delete_all_files_button_enabled, ) deselect_file_button.click( self._deselect_selected_file, outputs=[ delete_file_button, deselect_file_button, selected_text, ], ) ingested_dataset.select( fn=self._selected_a_file, outputs=[ delete_file_button, deselect_file_button, selected_text, ], ) delete_file_button.click( self._delete_selected_file, outputs=[ ingested_dataset, delete_file_button, deselect_file_button, selected_text, ], ) delete_files_button.click( self._delete_all_files, outputs=[ ingested_dataset, delete_file_button, deselect_file_button, selected_text, ], ) system_prompt_input = gr.Textbox( placeholder=self._system_prompt, label="System Prompt", lines=2, interactive=True, render=False, ) # When mode changes, set default system prompt, and other stuffs mode.change( self._set_current_mode, inputs=mode, outputs=[system_prompt_input, explanation_mode], ) # On blur, set system prompt to use in queries system_prompt_input.blur( self._set_system_prompt, inputs=system_prompt_input, ) def get_model_label() -> str | None: # Get model label from llm mode setting YAML # Labels: local, openai, openailike, sagemaker, mock, ollama config_settings = settings() if config_settings is None: raise ValueError("Settings are not configured.") # Get llm_mode from settings llm_mode = config_settings.llm.mode # Mapping of 'llm_mode' to corresponding model labels model_mapping = { "llamacpp": config_settings.llamacpp.llm_hf_model_file, "openai": config_settings.openai.model, "openailike": config_settings.openai.model, "azopenai": config_settings.azopenai.llm_model, "sagemaker": config_settings.sagemaker.llm_endpoint_name, "mock": llm_mode, "ollama": config_settings.ollama.llm_model, "gemini": config_settings.gemini.model, } if llm_mode not in model_mapping: print(f"Invalid 'llm mode': {llm_mode}") return None return model_mapping[llm_mode] with gr.Column(scale=7, elem_id="col"): # Determine the model label based on the value of PGPT_PROFILES model_label = get_model_label() if model_label is not None: label_text = ( f"LLM: {settings().llm.mode} | Model: {model_label}" ) else: label_text = f"LLM: {settings().llm.mode}" _ = gr.ChatInterface( self._chat, chatbot=gr.Chatbot( label=label_text, show_copy_button=True, elem_id="chatbot", render=False, avatar_images=( None, AVATAR_BOT, ), ), additional_inputs=[mode, upload_button, system_prompt_input], ) with gr.Row(): avatar_byte = AVATAR_BOT.read_bytes() f_base64 = f"data:image/png;base64,{base64.b64encode(avatar_byte).decode('utf-8')}" gr.HTML( f"<div class='footer'><a class='footer-zylon-link' href='https://zylon.ai/'>Maintained by Zylon <img class='footer-zylon-ico' src='{f_base64}' alt=Zylon></a></div>" ) return blocks def get_ui_blocks(self) -> gr.Blocks: if self._ui_block is None: self._ui_block = self._build_ui_blocks() return self._ui_block def mount_in_app(self, app: FastAPI, path: str) -> None: blocks = self.get_ui_blocks() blocks.queue() logger.info("Mounting the gradio UI, at path=%s", path) gr.mount_gradio_app(app, blocks, path=path, favicon_path=AVATAR_BOT) if __name__ == "__main__": ui = global_injector.get(PrivateGptUi) _blocks = ui.get_ui_blocks() _blocks.queue() _blocks.launch(debug=False, show_api=False)
--- +++ @@ -1,3 +1,4 @@+"""This file should be imported if and only if you want to run the UI locally.""" import base64 import logging @@ -500,6 +501,14 @@ ) def get_model_label() -> str | None: + """Get model label from llm mode setting YAML. + + Raises: + ValueError: If an invalid 'llm_mode' is encountered. + + Returns: + str: The corresponding model label. + """ # Get model label from llm mode setting YAML # Labels: local, openai, openailike, sagemaker, mock, ollama config_settings = settings() @@ -577,4 +586,4 @@ ui = global_injector.get(PrivateGptUi) _blocks = ui.get_ui_blocks() _blocks.queue() - _blocks.launch(debug=False, show_api=False)+ _blocks.launch(debug=False, show_api=False)
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/ui/ui.py
Add docstrings that explain inputs and outputs
# mypy: ignore-errors # Disabled mypy error: All conditional function variants must have identical signatures # We are changing the implementation of the authenticated method, based on # the config. If the auth is not enabled, we are not defining the complex method # with its dependencies. import logging import secrets from typing import Annotated from fastapi import Depends, Header, HTTPException from private_gpt.settings.settings import settings # 401 signify that the request requires authentication. # 403 signify that the authenticated user is not authorized to perform the operation. NOT_AUTHENTICATED = HTTPException( status_code=401, detail="Not authenticated", headers={"WWW-Authenticate": 'Basic realm="All the API", charset="UTF-8"'}, ) logger = logging.getLogger(__name__) def _simple_authentication(authorization: Annotated[str, Header()] = "") -> bool: if not secrets.compare_digest(authorization, settings().server.auth.secret): # If the "Authorization" header is not the expected one, raise an exception. raise NOT_AUTHENTICATED return True if not settings().server.auth.enabled: logger.debug( "Defining a dummy authentication mechanism for fastapi, always authenticating requests" ) # Define a dummy authentication method that always returns True. def authenticated() -> bool: return True else: logger.info("Defining the given authentication mechanism for the API") # Method to be used as a dependency to check if the request is authenticated. def authenticated( _simple_authentication: Annotated[bool, Depends(_simple_authentication)] ) -> bool: assert settings().server.auth.enabled if not _simple_authentication: raise NOT_AUTHENTICATED return True
--- +++ @@ -1,3 +1,17 @@+"""Authentication mechanism for the API. + +Define a simple mechanism to authenticate requests. +More complex authentication mechanisms can be defined here, and be placed in the +`authenticated` method (being a 'bean' injected in fastapi routers). + +Authorization can also be made after the authentication, and depends on +the authentication. Authorization should not be implemented in this file. + +Authorization can be done by following fastapi's guides: +* https://fastapi.tiangolo.com/advanced/security/oauth2-scopes/ +* https://fastapi.tiangolo.com/tutorial/security/ +* https://fastapi.tiangolo.com/tutorial/dependencies/dependencies-in-path-operation-decorators/ +""" # mypy: ignore-errors # Disabled mypy error: All conditional function variants must have identical signatures @@ -24,6 +38,7 @@ def _simple_authentication(authorization: Annotated[str, Header()] = "") -> bool: + """Check if the request is authenticated.""" if not secrets.compare_digest(authorization, settings().server.auth.secret): # If the "Authorization" header is not the expected one, raise an exception. raise NOT_AUTHENTICATED @@ -37,6 +52,7 @@ # Define a dummy authentication method that always returns True. def authenticated() -> bool: + """Check if the request is authenticated.""" return True else: @@ -46,7 +62,8 @@ def authenticated( _simple_authentication: Annotated[bool, Depends(_simple_authentication)] ) -> bool: + """Check if the request is authenticated.""" assert settings().server.auth.enabled if not _simple_authentication: raise NOT_AUTHENTICATED - return True+ return True
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/server/utils/auth.py
Add structured docstrings to improve clarity
import abc import itertools import logging import multiprocessing import multiprocessing.pool import os import threading from pathlib import Path from queue import Queue from typing import Any from llama_index.core.data_structs import IndexDict from llama_index.core.embeddings.utils import EmbedType from llama_index.core.indices import VectorStoreIndex, load_index_from_storage from llama_index.core.indices.base import BaseIndex from llama_index.core.ingestion import run_transformations from llama_index.core.schema import BaseNode, Document, TransformComponent from llama_index.core.storage import StorageContext from private_gpt.components.ingest.ingest_helper import IngestionHelper from private_gpt.paths import local_data_path from private_gpt.settings.settings import Settings from private_gpt.utils.eta import eta logger = logging.getLogger(__name__) class BaseIngestComponent(abc.ABC): def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], *args: Any, **kwargs: Any, ) -> None: logger.debug("Initializing base ingest component type=%s", type(self).__name__) self.storage_context = storage_context self.embed_model = embed_model self.transformations = transformations @abc.abstractmethod def ingest(self, file_name: str, file_data: Path) -> list[Document]: pass @abc.abstractmethod def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[Document]: pass @abc.abstractmethod def delete(self, doc_id: str) -> None: pass class BaseIngestComponentWithIndex(BaseIngestComponent, abc.ABC): def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], *args: Any, **kwargs: Any, ) -> None: super().__init__(storage_context, embed_model, transformations, *args, **kwargs) self.show_progress = True self._index_thread_lock = ( threading.Lock() ) # Thread lock! Not Multiprocessing lock self._index = self._initialize_index() def _initialize_index(self) -> BaseIndex[IndexDict]: try: # Load the index with store_nodes_override=True to be able to delete them index = load_index_from_storage( storage_context=self.storage_context, store_nodes_override=True, # Force store nodes in index and document stores show_progress=self.show_progress, embed_model=self.embed_model, transformations=self.transformations, ) except ValueError: # There are no index in the storage context, creating a new one logger.info("Creating a new vector store index") index = VectorStoreIndex.from_documents( [], storage_context=self.storage_context, store_nodes_override=True, # Force store nodes in index and document stores show_progress=self.show_progress, embed_model=self.embed_model, transformations=self.transformations, ) index.storage_context.persist(persist_dir=local_data_path) return index def _save_index(self) -> None: self._index.storage_context.persist(persist_dir=local_data_path) def delete(self, doc_id: str) -> None: with self._index_thread_lock: # Delete the document from the index self._index.delete_ref_doc(doc_id, delete_from_docstore=True) # Save the index self._save_index() class SimpleIngestComponent(BaseIngestComponentWithIndex): def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], *args: Any, **kwargs: Any, ) -> None: super().__init__(storage_context, embed_model, transformations, *args, **kwargs) def ingest(self, file_name: str, file_data: Path) -> list[Document]: logger.info("Ingesting file_name=%s", file_name) documents = IngestionHelper.transform_file_into_documents(file_name, file_data) logger.info( "Transformed file=%s into count=%s documents", file_name, len(documents) ) logger.debug("Saving the documents in the index and doc store") return self._save_docs(documents) def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[Document]: saved_documents = [] for file_name, file_data in files: documents = IngestionHelper.transform_file_into_documents( file_name, file_data ) saved_documents.extend(self._save_docs(documents)) return saved_documents def _save_docs(self, documents: list[Document]) -> list[Document]: logger.debug("Transforming count=%s documents into nodes", len(documents)) with self._index_thread_lock: for document in documents: self._index.insert(document, show_progress=True) logger.debug("Persisting the index and nodes") # persist the index and nodes self._save_index() logger.debug("Persisted the index and nodes") return documents class BatchIngestComponent(BaseIngestComponentWithIndex): def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], count_workers: int, *args: Any, **kwargs: Any, ) -> None: super().__init__(storage_context, embed_model, transformations, *args, **kwargs) # Make an efficient use of the CPU and GPU, the embedding # must be in the transformations assert ( len(self.transformations) >= 2 ), "Embeddings must be in the transformations" assert count_workers > 0, "count_workers must be > 0" self.count_workers = count_workers self._file_to_documents_work_pool = multiprocessing.Pool( processes=self.count_workers ) def ingest(self, file_name: str, file_data: Path) -> list[Document]: logger.info("Ingesting file_name=%s", file_name) documents = IngestionHelper.transform_file_into_documents(file_name, file_data) logger.info( "Transformed file=%s into count=%s documents", file_name, len(documents) ) logger.debug("Saving the documents in the index and doc store") return self._save_docs(documents) def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[Document]: documents = list( itertools.chain.from_iterable( self._file_to_documents_work_pool.starmap( IngestionHelper.transform_file_into_documents, files ) ) ) logger.info( "Transformed count=%s files into count=%s documents", len(files), len(documents), ) return self._save_docs(documents) def _save_docs(self, documents: list[Document]) -> list[Document]: logger.debug("Transforming count=%s documents into nodes", len(documents)) nodes = run_transformations( documents, # type: ignore[arg-type] self.transformations, show_progress=self.show_progress, ) # Locking the index to avoid concurrent writes with self._index_thread_lock: logger.info("Inserting count=%s nodes in the index", len(nodes)) self._index.insert_nodes(nodes, show_progress=True) for document in documents: self._index.docstore.set_document_hash( document.get_doc_id(), document.hash ) logger.debug("Persisting the index and nodes") # persist the index and nodes self._save_index() logger.debug("Persisted the index and nodes") return documents class ParallelizedIngestComponent(BaseIngestComponentWithIndex): def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], count_workers: int, *args: Any, **kwargs: Any, ) -> None: super().__init__(storage_context, embed_model, transformations, *args, **kwargs) # To make an efficient use of the CPU and GPU, the embeddings # must be in the transformations (to be computed in batches) assert ( len(self.transformations) >= 2 ), "Embeddings must be in the transformations" assert count_workers > 0, "count_workers must be > 0" self.count_workers = count_workers # We are doing our own multiprocessing # To do not collide with the multiprocessing of huggingface, we disable it os.environ["TOKENIZERS_PARALLELISM"] = "false" self._ingest_work_pool = multiprocessing.pool.ThreadPool( processes=self.count_workers ) self._file_to_documents_work_pool = multiprocessing.Pool( processes=self.count_workers ) def ingest(self, file_name: str, file_data: Path) -> list[Document]: logger.info("Ingesting file_name=%s", file_name) # Running in a single (1) process to release the current # thread, and take a dedicated CPU core for computation documents = self._file_to_documents_work_pool.apply( IngestionHelper.transform_file_into_documents, (file_name, file_data) ) logger.info( "Transformed file=%s into count=%s documents", file_name, len(documents) ) logger.debug("Saving the documents in the index and doc store") return self._save_docs(documents) def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[Document]: # Lightweight threads, used for parallelize the # underlying IO calls made in the ingestion documents = list( itertools.chain.from_iterable( self._ingest_work_pool.starmap(self.ingest, files) ) ) return documents def _save_docs(self, documents: list[Document]) -> list[Document]: logger.debug("Transforming count=%s documents into nodes", len(documents)) nodes = run_transformations( documents, # type: ignore[arg-type] self.transformations, show_progress=self.show_progress, ) # Locking the index to avoid concurrent writes with self._index_thread_lock: logger.info("Inserting count=%s nodes in the index", len(nodes)) self._index.insert_nodes(nodes, show_progress=True) for document in documents: self._index.docstore.set_document_hash( document.get_doc_id(), document.hash ) logger.debug("Persisting the index and nodes") # persist the index and nodes self._save_index() logger.debug("Persisted the index and nodes") return documents def __del__(self) -> None: # We need to do the appropriate cleanup of the multiprocessing pools # when the object is deleted. Using root logger to avoid # the logger to be deleted before the pool logging.debug("Closing the ingest work pool") self._ingest_work_pool.close() self._ingest_work_pool.join() self._ingest_work_pool.terminate() logging.debug("Closing the file to documents work pool") self._file_to_documents_work_pool.close() self._file_to_documents_work_pool.join() self._file_to_documents_work_pool.terminate() class PipelineIngestComponent(BaseIngestComponentWithIndex): NODE_FLUSH_COUNT = 5000 # Save the index every # nodes. def __init__( self, storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], count_workers: int, *args: Any, **kwargs: Any, ) -> None: super().__init__(storage_context, embed_model, transformations, *args, **kwargs) self.count_workers = count_workers assert ( len(self.transformations) >= 2 ), "Embeddings must be in the transformations" assert count_workers > 0, "count_workers must be > 0" self.count_workers = count_workers # We are doing our own multiprocessing # To do not collide with the multiprocessing of huggingface, we disable it os.environ["TOKENIZERS_PARALLELISM"] = "false" # doc_q stores parsed files as Document chunks. # Using a shallow queue causes the filesystem parser to block # when it reaches capacity. This ensures it doesn't outpace the # computationally intensive embeddings phase, avoiding unnecessary # memory consumption. The semaphore is used to bound the async worker # embedding computations to cause the doc Q to fill and block. self.doc_semaphore = multiprocessing.Semaphore( self.count_workers ) # limit the doc queue to # items. self.doc_q: Queue[tuple[str, str | None, list[Document] | None]] = Queue(20) # node_q stores documents parsed into nodes (embeddings). # Larger queue size so we don't block the embedding workers during a slow # index update. self.node_q: Queue[ tuple[str, str | None, list[Document] | None, list[BaseNode] | None] ] = Queue(40) threading.Thread(target=self._doc_to_node, daemon=True).start() threading.Thread(target=self._write_nodes, daemon=True).start() def _doc_to_node(self) -> None: # Parse documents into nodes with multiprocessing.pool.ThreadPool(processes=self.count_workers) as pool: while True: try: cmd, file_name, documents = self.doc_q.get( block=True ) # Documents for a file if cmd == "process": # Push CPU/GPU embedding work to the worker pool # Acquire semaphore to control access to worker pool self.doc_semaphore.acquire() pool.apply_async( self._doc_to_node_worker, (file_name, documents) ) elif cmd == "quit": break finally: if cmd != "process": self.doc_q.task_done() # unblock Q joins def _doc_to_node_worker(self, file_name: str, documents: list[Document]) -> None: # CPU/GPU intensive work in its own process try: nodes = run_transformations( documents, # type: ignore[arg-type] self.transformations, show_progress=self.show_progress, ) self.node_q.put(("process", file_name, documents, list(nodes))) finally: self.doc_semaphore.release() self.doc_q.task_done() # unblock Q joins def _save_docs( self, files: list[str], documents: list[Document], nodes: list[BaseNode] ) -> None: try: logger.info( f"Saving {len(files)} files ({len(documents)} documents / {len(nodes)} nodes)" ) self._index.insert_nodes(nodes) for document in documents: self._index.docstore.set_document_hash( document.get_doc_id(), document.hash ) self._save_index() except Exception: # Tell the user so they can investigate these files logger.exception(f"Processing files {files}") finally: # Clearing work, even on exception, maintains a clean state. nodes.clear() documents.clear() files.clear() def _write_nodes(self) -> None: # Save nodes to index. I/O intensive. node_stack: list[BaseNode] = [] doc_stack: list[Document] = [] file_stack: list[str] = [] while True: try: cmd, file_name, documents, nodes = self.node_q.get(block=True) if cmd in ("flush", "quit"): if file_stack: self._save_docs(file_stack, doc_stack, node_stack) if cmd == "quit": break elif cmd == "process": node_stack.extend(nodes) # type: ignore[arg-type] doc_stack.extend(documents) # type: ignore[arg-type] file_stack.append(file_name) # type: ignore[arg-type] # Constant saving is heavy on I/O - accumulate to a threshold if len(node_stack) >= self.NODE_FLUSH_COUNT: self._save_docs(file_stack, doc_stack, node_stack) finally: self.node_q.task_done() def _flush(self) -> None: self.doc_q.put(("flush", None, None)) self.doc_q.join() self.node_q.put(("flush", None, None, None)) self.node_q.join() def ingest(self, file_name: str, file_data: Path) -> list[Document]: documents = IngestionHelper.transform_file_into_documents(file_name, file_data) self.doc_q.put(("process", file_name, documents)) self._flush() return documents def bulk_ingest(self, files: list[tuple[str, Path]]) -> list[Document]: docs = [] for file_name, file_data in eta(files): try: documents = IngestionHelper.transform_file_into_documents( file_name, file_data ) self.doc_q.put(("process", file_name, documents)) docs.extend(documents) except Exception: logger.exception(f"Skipping {file_data.name}") self._flush() return docs def get_ingestion_component( storage_context: StorageContext, embed_model: EmbedType, transformations: list[TransformComponent], settings: Settings, ) -> BaseIngestComponent: ingest_mode = settings.embedding.ingest_mode if ingest_mode == "batch": return BatchIngestComponent( storage_context=storage_context, embed_model=embed_model, transformations=transformations, count_workers=settings.embedding.count_workers, ) elif ingest_mode == "parallel": return ParallelizedIngestComponent( storage_context=storage_context, embed_model=embed_model, transformations=transformations, count_workers=settings.embedding.count_workers, ) elif ingest_mode == "pipeline": return PipelineIngestComponent( storage_context=storage_context, embed_model=embed_model, transformations=transformations, count_workers=settings.embedding.count_workers, ) else: return SimpleIngestComponent( storage_context=storage_context, embed_model=embed_model, transformations=transformations, )
--- +++ @@ -70,6 +70,7 @@ self._index = self._initialize_index() def _initialize_index(self) -> BaseIndex[IndexDict]: + """Initialize the index from the storage context.""" try: # Load the index with store_nodes_override=True to be able to delete them index = load_index_from_storage( @@ -147,6 +148,10 @@ class BatchIngestComponent(BaseIngestComponentWithIndex): + """Parallelize the file reading and parsing on multiple CPU core. + + This also makes the embeddings to be computed in batches (on GPU or CPU). + """ def __init__( self, @@ -217,6 +222,11 @@ class ParallelizedIngestComponent(BaseIngestComponentWithIndex): + """Parallelize the file ingestion (file reading, embeddings, and index insertion). + + This use the CPU and GPU in parallel (both running at the same time), and + reduce the memory pressure by not loading all the files in memory at the same time. + """ def __init__( self, @@ -307,6 +317,21 @@ class PipelineIngestComponent(BaseIngestComponentWithIndex): + """Pipeline ingestion - keeping the embedding worker pool as busy as possible. + + This class implements a threaded ingestion pipeline, which comprises two threads + and two queues. The primary thread is responsible for reading and parsing files + into documents. These documents are then placed into a queue, which is + distributed to a pool of worker processes for embedding computation. After + embedding, the documents are transferred to another queue where they are + accumulated until a threshold is reached. Upon reaching this threshold, the + accumulated documents are flushed to the document store, index, and vector + store. + + Exception handling ensures robustness against erroneous files. However, in the + pipelined design, one error can lead to the discarding of multiple files. Any + discarded files will be reported. + """ NODE_FLUSH_COUNT = 5000 # Save the index every # nodes. @@ -461,6 +486,7 @@ transformations: list[TransformComponent], settings: Settings, ) -> BaseIngestComponent: + """Get the ingestion component for the given configuration.""" ingest_mode = settings.embedding.ingest_mode if ingest_mode == "batch": return BatchIngestComponent( @@ -488,4 +514,4 @@ storage_context=storage_context, embed_model=embed_model, transformations=transformations, - )+ )
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/components/ingest/ingest_component.py
Document this script properly
from collections.abc import Generator, Sequence from typing import TYPE_CHECKING, Any from llama_index.core.schema import BaseNode, MetadataMode from llama_index.core.vector_stores.utils import node_to_metadata_dict from llama_index.vector_stores.chroma import ChromaVectorStore # type: ignore if TYPE_CHECKING: from collections.abc import Mapping def chunk_list( lst: Sequence[BaseNode], max_chunk_size: int ) -> Generator[Sequence[BaseNode], None, None]: for i in range(0, len(lst), max_chunk_size): yield lst[i : i + max_chunk_size] class BatchedChromaVectorStore(ChromaVectorStore): # type: ignore chroma_client: Any | None def __init__( self, chroma_client: Any, chroma_collection: Any, host: str | None = None, port: str | None = None, ssl: bool = False, headers: dict[str, str] | None = None, collection_kwargs: dict[Any, Any] | None = None, ) -> None: super().__init__( chroma_collection=chroma_collection, host=host, port=port, ssl=ssl, headers=headers, collection_kwargs=collection_kwargs or {}, ) self.chroma_client = chroma_client def add(self, nodes: Sequence[BaseNode], **add_kwargs: Any) -> list[str]: if not self.chroma_client: raise ValueError("Client not initialized") if not self._collection: raise ValueError("Collection not initialized") max_chunk_size = self.chroma_client.max_batch_size node_chunks = chunk_list(nodes, max_chunk_size) all_ids = [] for node_chunk in node_chunks: embeddings: list[Sequence[float]] = [] metadatas: list[Mapping[str, Any]] = [] ids = [] documents = [] for node in node_chunk: embeddings.append(node.get_embedding()) metadatas.append( node_to_metadata_dict( node, remove_text=True, flat_metadata=self.flat_metadata ) ) ids.append(node.node_id) documents.append(node.get_content(metadata_mode=MetadataMode.NONE)) self._collection.add( embeddings=embeddings, ids=ids, metadatas=metadatas, documents=documents, ) all_ids.extend(ids) return all_ids
--- +++ @@ -12,11 +12,34 @@ def chunk_list( lst: Sequence[BaseNode], max_chunk_size: int ) -> Generator[Sequence[BaseNode], None, None]: + """Yield successive max_chunk_size-sized chunks from lst. + + Args: + lst (List[BaseNode]): list of nodes with embeddings + max_chunk_size (int): max chunk size + + Yields: + Generator[List[BaseNode], None, None]: list of nodes with embeddings + """ for i in range(0, len(lst), max_chunk_size): yield lst[i : i + max_chunk_size] class BatchedChromaVectorStore(ChromaVectorStore): # type: ignore + """Chroma vector store, batching additions to avoid reaching the max batch limit. + + In this vector store, embeddings are stored within a ChromaDB collection. + + During query time, the index uses ChromaDB to query for the top + k most similar nodes. + + Args: + chroma_client (from chromadb.api.API): + API instance + chroma_collection (chromadb.api.models.Collection.Collection): + ChromaDB collection instance + + """ chroma_client: Any | None @@ -41,6 +64,12 @@ self.chroma_client = chroma_client def add(self, nodes: Sequence[BaseNode], **add_kwargs: Any) -> list[str]: + """Add nodes to index, batching the insertion to avoid issues. + + Args: + nodes: List[BaseNode]: list of nodes with embeddings + add_kwargs: _ + """ if not self.chroma_client: raise ValueError("Client not initialized") @@ -74,4 +103,4 @@ ) all_ids.extend(ids) - return all_ids+ return all_ids
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/components/vector_store/batched_chroma.py
Turn comments into proper docstrings
import abc import logging from collections.abc import Sequence from typing import Any, Literal from llama_index.core.llms import ChatMessage, MessageRole logger = logging.getLogger(__name__) class AbstractPromptStyle(abc.ABC): def __init__(self, *args: Any, **kwargs: Any) -> None: logger.debug("Initializing prompt_style=%s", self.__class__.__name__) @abc.abstractmethod def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: pass @abc.abstractmethod def _completion_to_prompt(self, completion: str) -> str: pass def messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: prompt = self._messages_to_prompt(messages) logger.debug("Got for messages='%s' the prompt='%s'", messages, prompt) return prompt def completion_to_prompt(self, prompt: str) -> str: completion = prompt # Fix: Llama-index parameter has to be named as prompt prompt = self._completion_to_prompt(completion) logger.debug("Got for completion='%s' the prompt='%s'", completion, prompt) return prompt class DefaultPromptStyle(AbstractPromptStyle): def __init__(self, *args: Any, **kwargs: Any) -> None: super().__init__(*args, **kwargs) # Hacky way to override the functions # Override the functions to be None, and pass None to the LLM. self.messages_to_prompt = None # type: ignore[method-assign, assignment] self.completion_to_prompt = None # type: ignore[method-assign, assignment] def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: return "" def _completion_to_prompt(self, completion: str) -> str: return "" class Llama2PromptStyle(AbstractPromptStyle): BOS, EOS = "<s>", "</s>" B_INST, E_INST = "[INST]", "[/INST]" B_SYS, E_SYS = "<<SYS>>\n", "\n<</SYS>>\n\n" DEFAULT_SYSTEM_PROMPT = """\ You are a helpful, respectful and honest assistant. \ Always answer as helpfully as possible and follow ALL given instructions. \ Do not speculate or make up information. \ Do not reference any given instructions or context. \ """ def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: string_messages: list[str] = [] if messages[0].role == MessageRole.SYSTEM: # pull out the system message (if it exists in messages) system_message_str = messages[0].content or "" messages = messages[1:] else: system_message_str = self.DEFAULT_SYSTEM_PROMPT system_message_str = f"{self.B_SYS} {system_message_str.strip()} {self.E_SYS}" for i in range(0, len(messages), 2): # first message should always be a user user_message = messages[i] assert user_message.role == MessageRole.USER if i == 0: # make sure system prompt is included at the start str_message = f"{self.BOS} {self.B_INST} {system_message_str} " else: # end previous user-assistant interaction string_messages[-1] += f" {self.EOS}" # no need to include system prompt str_message = f"{self.BOS} {self.B_INST} " # include user message content str_message += f"{user_message.content} {self.E_INST}" if len(messages) > (i + 1): # if assistant message exists, add to str_message assistant_message = messages[i + 1] assert assistant_message.role == MessageRole.ASSISTANT str_message += f" {assistant_message.content}" string_messages.append(str_message) return "".join(string_messages) def _completion_to_prompt(self, completion: str) -> str: system_prompt_str = self.DEFAULT_SYSTEM_PROMPT return ( f"{self.BOS} {self.B_INST} {self.B_SYS} {system_prompt_str.strip()} {self.E_SYS} " f"{completion.strip()} {self.E_INST}" ) class Llama3PromptStyle(AbstractPromptStyle): BOS, EOS = "<|begin_of_text|>", "<|end_of_text|>" B_INST, E_INST = "<|start_header_id|>", "<|end_header_id|>" EOT = "<|eot_id|>" B_SYS, E_SYS = "<|start_header_id|>system<|end_header_id|>", "<|eot_id|>" ASSISTANT_INST = "<|start_header_id|>assistant<|end_header_id|>" DEFAULT_SYSTEM_PROMPT = """\ You are a helpful, respectful and honest assistant. \ Always answer as helpfully as possible and follow ALL given instructions. \ Do not speculate or make up information. \ Do not reference any given instructions or context. \ """ def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: prompt = "" has_system_message = False for i, message in enumerate(messages): if not message or message.content is None: continue if message.role == MessageRole.SYSTEM: prompt += f"{self.B_SYS}\n\n{message.content.strip()}{self.E_SYS}" has_system_message = True else: role_header = f"{self.B_INST}{message.role.value}{self.E_INST}" prompt += f"{role_header}\n\n{message.content.strip()}{self.EOT}" # Add assistant header if the last message is not from the assistant if i == len(messages) - 1 and message.role != MessageRole.ASSISTANT: prompt += f"{self.ASSISTANT_INST}\n\n" # Add default system prompt if no system message was provided if not has_system_message: prompt = ( f"{self.B_SYS}\n\n{self.DEFAULT_SYSTEM_PROMPT}{self.E_SYS}" + prompt ) # TODO: Implement tool handling logic return prompt def _completion_to_prompt(self, completion: str) -> str: return ( f"{self.B_SYS}\n\n{self.DEFAULT_SYSTEM_PROMPT}{self.E_SYS}" f"{self.B_INST}user{self.E_INST}\n\n{completion.strip()}{self.EOT}" f"{self.ASSISTANT_INST}\n\n" ) class TagPromptStyle(AbstractPromptStyle): def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: prompt = "" for message in messages: role = message.role content = message.content or "" message_from_user = f"<|{role.lower()}|>: {content.strip()}" message_from_user += "\n" prompt += message_from_user # we are missing the last <|assistant|> tag that will trigger a completion prompt += "<|assistant|>: " return prompt def _completion_to_prompt(self, completion: str) -> str: return self._messages_to_prompt( [ChatMessage(content=completion, role=MessageRole.USER)] ) class MistralPromptStyle(AbstractPromptStyle): def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: inst_buffer = [] text = "" for message in messages: if message.role == MessageRole.SYSTEM or message.role == MessageRole.USER: inst_buffer.append(str(message.content).strip()) elif message.role == MessageRole.ASSISTANT: text += "<s>[INST] " + "\n".join(inst_buffer) + " [/INST]" text += " " + str(message.content).strip() + "</s>" inst_buffer.clear() else: raise ValueError(f"Unknown message role {message.role}") if len(inst_buffer) > 0: text += "<s>[INST] " + "\n".join(inst_buffer) + " [/INST]" return text def _completion_to_prompt(self, completion: str) -> str: return self._messages_to_prompt( [ChatMessage(content=completion, role=MessageRole.USER)] ) class ChatMLPromptStyle(AbstractPromptStyle): def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: prompt = "<|im_start|>system\n" for message in messages: role = message.role content = message.content or "" if role.lower() == "system": message_from_user = f"{content.strip()}" prompt += message_from_user elif role.lower() == "user": prompt += "<|im_end|>\n<|im_start|>user\n" message_from_user = f"{content.strip()}<|im_end|>\n" prompt += message_from_user prompt += "<|im_start|>assistant\n" return prompt def _completion_to_prompt(self, completion: str) -> str: return self._messages_to_prompt( [ChatMessage(content=completion, role=MessageRole.USER)] ) def get_prompt_style( prompt_style: ( Literal["default", "llama2", "llama3", "tag", "mistral", "chatml"] | None ) ) -> AbstractPromptStyle: if prompt_style is None or prompt_style == "default": return DefaultPromptStyle() elif prompt_style == "llama2": return Llama2PromptStyle() elif prompt_style == "llama3": return Llama3PromptStyle() elif prompt_style == "tag": return TagPromptStyle() elif prompt_style == "mistral": return MistralPromptStyle() elif prompt_style == "chatml": return ChatMLPromptStyle() raise ValueError(f"Unknown prompt_style='{prompt_style}'")
--- +++ @@ -9,6 +9,20 @@ class AbstractPromptStyle(abc.ABC): + """Abstract class for prompt styles. + + This class is used to format a series of messages into a prompt that can be + understood by the models. A series of messages represents the interaction(s) + between a user and an assistant. This series of messages can be considered as a + session between a user X and an assistant Y.This session holds, through the + messages, the state of the conversation. This session, to be understood by the + model, needs to be formatted into a prompt (i.e. a string that the models + can understand). Prompts can be formatted in different ways, + depending on the model. + + The implementations of this class represent the different ways to format a + series of messages into a prompt. + """ def __init__(self, *args: Any, **kwargs: Any) -> None: logger.debug("Initializing prompt_style=%s", self.__class__.__name__) @@ -34,6 +48,11 @@ class DefaultPromptStyle(AbstractPromptStyle): + """Default prompt style that uses the defaults from llama_utils. + + It basically passes None to the LLM, indicating it should use + the default functions. + """ def __init__(self, *args: Any, **kwargs: Any) -> None: super().__init__(*args, **kwargs) @@ -51,6 +70,17 @@ class Llama2PromptStyle(AbstractPromptStyle): + """Simple prompt style that uses llama 2 prompt style. + + Inspired by llama_index/legacy/llms/llama_utils.py + + It transforms the sequence of messages into a prompt that should look like: + ```text + <s> [INST] <<SYS>> your system prompt here. <</SYS>> + + user message here [/INST] assistant (model) response here </s> + ``` + """ BOS, EOS = "<s>", "</s>" B_INST, E_INST = "[INST]", "[/INST]" @@ -110,6 +140,22 @@ class Llama3PromptStyle(AbstractPromptStyle): + r"""Template for Meta's Llama 3.1. + + The format follows this structure: + <|begin_of_text|> + <|start_header_id|>system<|end_header_id|> + + [System message content]<|eot_id|> + <|start_header_id|>user<|end_header_id|> + + [User message content]<|eot_id|> + <|start_header_id|>assistant<|end_header_id|> + + [Assistant message content]<|eot_id|> + ... + (Repeat for each message, including possible 'ipython' role) + """ BOS, EOS = "<|begin_of_text|>", "<|end_of_text|>" B_INST, E_INST = "<|start_header_id|>", "<|end_header_id|>" @@ -160,8 +206,21 @@ class TagPromptStyle(AbstractPromptStyle): - - def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: + """Tag prompt style (used by Vigogne) that uses the prompt style `<|ROLE|>`. + + It transforms the sequence of messages into a prompt that should look like: + ```text + <|system|>: your system prompt here. + <|user|>: user message here + (possibly with context and question) + <|assistant|>: assistant (model) response here. + ``` + + FIXME: should we add surrounding `<s>` and `</s>` tags, like in llama2? + """ + + def _messages_to_prompt(self, messages: Sequence[ChatMessage]) -> str: + """Format message to prompt with `<|ROLE|>: MSG` style.""" prompt = "" for message in messages: role = message.role @@ -231,6 +290,11 @@ Literal["default", "llama2", "llama3", "tag", "mistral", "chatml"] | None ) ) -> AbstractPromptStyle: + """Get the prompt style to use from the given string. + + :param prompt_style: The prompt style to use. + :return: The prompt style to use. + """ if prompt_style is None or prompt_style == "default": return DefaultPromptStyle() elif prompt_style == "llama2": @@ -243,4 +307,4 @@ return MistralPromptStyle() elif prompt_style == "chatml": return ChatMLPromptStyle() - raise ValueError(f"Unknown prompt_style='{prompt_style}'")+ raise ValueError(f"Unknown prompt_style='{prompt_style}'")
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/components/llm/prompt_helper.py
Add structured docstrings to improve clarity
import time import uuid from collections.abc import Iterator from typing import Literal from llama_index.core.llms import ChatResponse, CompletionResponse from pydantic import BaseModel, Field from private_gpt.server.chunks.chunks_service import Chunk class OpenAIDelta(BaseModel): content: str | None class OpenAIMessage(BaseModel): role: Literal["assistant", "system", "user"] = Field(default="user") content: str | None class OpenAIChoice(BaseModel): finish_reason: str | None = Field(examples=["stop"]) delta: OpenAIDelta | None = None message: OpenAIMessage | None = None sources: list[Chunk] | None = None index: int = 0 class OpenAICompletion(BaseModel): id: str object: Literal["completion", "completion.chunk"] = Field(default="completion") created: int = Field(..., examples=[1623340000]) model: Literal["private-gpt"] choices: list[OpenAIChoice] @classmethod def from_text( cls, text: str | None, finish_reason: str | None = None, sources: list[Chunk] | None = None, ) -> "OpenAICompletion": return OpenAICompletion( id=str(uuid.uuid4()), object="completion", created=int(time.time()), model="private-gpt", choices=[ OpenAIChoice( message=OpenAIMessage(role="assistant", content=text), finish_reason=finish_reason, sources=sources, ) ], ) @classmethod def json_from_delta( cls, *, text: str | None, finish_reason: str | None = None, sources: list[Chunk] | None = None, ) -> str: chunk = OpenAICompletion( id=str(uuid.uuid4()), object="completion.chunk", created=int(time.time()), model="private-gpt", choices=[ OpenAIChoice( delta=OpenAIDelta(content=text), finish_reason=finish_reason, sources=sources, ) ], ) return chunk.model_dump_json() def to_openai_response( response: str | ChatResponse, sources: list[Chunk] | None = None ) -> OpenAICompletion: if isinstance(response, ChatResponse): return OpenAICompletion.from_text(response.delta, finish_reason="stop") else: return OpenAICompletion.from_text( response, finish_reason="stop", sources=sources ) def to_openai_sse_stream( response_generator: Iterator[str | CompletionResponse | ChatResponse], sources: list[Chunk] | None = None, ) -> Iterator[str]: for response in response_generator: if isinstance(response, CompletionResponse | ChatResponse): yield f"data: {OpenAICompletion.json_from_delta(text=response.delta)}\n\n" else: yield f"data: {OpenAICompletion.json_from_delta(text=response, sources=sources)}\n\n" yield f"data: {OpenAICompletion.json_from_delta(text='', finish_reason='stop')}\n\n" yield "data: [DONE]\n\n"
--- +++ @@ -10,17 +10,28 @@ class OpenAIDelta(BaseModel): + """A piece of completion that needs to be concatenated to get the full message.""" content: str | None class OpenAIMessage(BaseModel): + """Inference result, with the source of the message. + + Role could be the assistant or system + (providing a default response, not AI generated). + """ role: Literal["assistant", "system", "user"] = Field(default="user") content: str | None class OpenAIChoice(BaseModel): + """Response from AI. + + Either the delta or the message will be present, but never both. + Sources used will be returned in case context retrieval was enabled. + """ finish_reason: str | None = Field(examples=["stop"]) delta: OpenAIDelta | None = None @@ -30,6 +41,10 @@ class OpenAICompletion(BaseModel): + """Clone of OpenAI Completion model. + + For more information see: https://platform.openai.com/docs/api-reference/chat/object + """ id: str object: Literal["completion", "completion.chunk"] = Field(default="completion") @@ -104,4 +119,4 @@ else: yield f"data: {OpenAICompletion.json_from_delta(text=response, sources=sources)}\n\n" yield f"data: {OpenAICompletion.json_from_delta(text='', finish_reason='stop')}\n\n" - yield "data: [DONE]\n\n"+ yield "data: [DONE]\n\n"
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/open_ai/openai_models.py
Add verbose docstrings with examples
from typing import Any, Literal from pydantic import BaseModel, Field from private_gpt.settings.settings_loader import load_active_settings class CorsSettings(BaseModel): enabled: bool = Field( description="Flag indicating if CORS headers are set or not." "If set to True, the CORS headers will be set to allow all origins, methods and headers.", default=False, ) allow_credentials: bool = Field( description="Indicate that cookies should be supported for cross-origin requests", default=False, ) allow_origins: list[str] = Field( description="A list of origins that should be permitted to make cross-origin requests.", default=[], ) allow_origin_regex: list[str] = Field( description="A regex string to match against origins that should be permitted to make cross-origin requests.", default=None, ) allow_methods: list[str] = Field( description="A list of HTTP methods that should be allowed for cross-origin requests.", default=[ "GET", ], ) allow_headers: list[str] = Field( description="A list of HTTP request headers that should be supported for cross-origin requests.", default=[], ) class AuthSettings(BaseModel): enabled: bool = Field( description="Flag indicating if authentication is enabled or not.", default=False, ) secret: str = Field( description="The secret to be used for authentication. " "It can be any non-blank string. For HTTP basic authentication, " "this value should be the whole 'Authorization' header that is expected" ) class IngestionSettings(BaseModel): enabled: bool = Field( description="Flag indicating if local ingestion is enabled or not.", default=False, ) allow_ingest_from: list[str] = Field( description="A list of folders that should be permitted to make ingest requests.", default=[], ) class ServerSettings(BaseModel): env_name: str = Field( description="Name of the environment (prod, staging, local...)" ) port: int = Field(description="Port of PrivateGPT FastAPI server, defaults to 8001") cors: CorsSettings = Field( description="CORS configuration", default=CorsSettings(enabled=False) ) auth: AuthSettings = Field( description="Authentication configuration", default_factory=lambda: AuthSettings(enabled=False, secret="secret-key"), ) class DataSettings(BaseModel): local_ingestion: IngestionSettings = Field( description="Ingestion configuration", default_factory=lambda: IngestionSettings(allow_ingest_from=["*"]), ) local_data_folder: str = Field( description="Path to local storage." "It will be treated as an absolute path if it starts with /" ) class LLMSettings(BaseModel): mode: Literal[ "llamacpp", "openai", "openailike", "azopenai", "sagemaker", "mock", "ollama", "gemini", ] max_new_tokens: int = Field( 256, description="The maximum number of token that the LLM is authorized to generate in one completion.", ) context_window: int = Field( 3900, description="The maximum number of context tokens for the model.", ) tokenizer: str = Field( None, description="The model id of a predefined tokenizer hosted inside a model repo on " "huggingface.co. Valid model ids can be located at the root-level, like " "`bert-base-uncased`, or namespaced under a user or organization name, " "like `HuggingFaceH4/zephyr-7b-beta`. If not set, will load a tokenizer matching " "gpt-3.5-turbo LLM.", ) temperature: float = Field( 0.1, description="The temperature of the model. Increasing the temperature will make the model answer more creatively. A value of 0.1 would be more factual.", ) prompt_style: Literal["default", "llama2", "llama3", "tag", "mistral", "chatml"] = ( Field( "llama2", description=( "The prompt style to use for the chat engine. " "If `default` - use the default prompt style from the llama_index. It should look like `role: message`.\n" "If `llama2` - use the llama2 prompt style from the llama_index. Based on `<s>`, `[INST]` and `<<SYS>>`.\n" "If `llama3` - use the llama3 prompt style from the llama_index." "If `tag` - use the `tag` prompt style. It should look like `<|role|>: message`. \n" "If `mistral` - use the `mistral prompt style. It shoudl look like <s>[INST] {System Prompt} [/INST]</s>[INST] { UserInstructions } [/INST]" "`llama2` is the historic behaviour. `default` might work better with your custom models." ), ) ) class VectorstoreSettings(BaseModel): database: Literal["chroma", "qdrant", "postgres", "clickhouse", "milvus"] class NodeStoreSettings(BaseModel): database: Literal["simple", "postgres"] class LlamaCPPSettings(BaseModel): llm_hf_repo_id: str llm_hf_model_file: str tfs_z: float = Field( 1.0, description="Tail free sampling is used to reduce the impact of less probable tokens from the output. A higher value (e.g., 2.0) will reduce the impact more, while a value of 1.0 disables this setting.", ) top_k: int = Field( 40, description="Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative. (Default: 40)", ) top_p: float = Field( 0.9, description="Works together with top-k. A higher value (e.g., 0.95) will lead to more diverse text, while a lower value (e.g., 0.5) will generate more focused and conservative text. (Default: 0.9)", ) repeat_penalty: float = Field( 1.1, description="Sets how strongly to penalize repetitions. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. (Default: 1.1)", ) class HuggingFaceSettings(BaseModel): embedding_hf_model_name: str = Field( description="Name of the HuggingFace model to use for embeddings" ) access_token: str = Field( None, description="Huggingface access token, required to download some models", ) trust_remote_code: bool = Field( False, description="If set to True, the code from the remote model will be trusted and executed.", ) class EmbeddingSettings(BaseModel): mode: Literal[ "huggingface", "openai", "azopenai", "sagemaker", "ollama", "mock", "gemini", "mistralai", ] ingest_mode: Literal["simple", "batch", "parallel", "pipeline"] = Field( "simple", description=( "The ingest mode to use for the embedding engine:\n" "If `simple` - ingest files sequentially and one by one. It is the historic behaviour.\n" "If `batch` - if multiple files, parse all the files in parallel, " "and send them in batch to the embedding model.\n" "In `pipeline` - The Embedding engine is kept as busy as possible\n" "If `parallel` - parse the files in parallel using multiple cores, and embedd them in parallel.\n" "`parallel` is the fastest mode for local setup, as it parallelize IO RW in the index.\n" "For modes that leverage parallelization, you can specify the number of " "workers to use with `count_workers`.\n" ), ) count_workers: int = Field( 2, description=( "The number of workers to use for file ingestion.\n" "In `batch` mode, this is the number of workers used to parse the files.\n" "In `parallel` mode, this is the number of workers used to parse the files and embed them.\n" "In `pipeline` mode, this is the number of workers that can perform embeddings.\n" "This is only used if `ingest_mode` is not `simple`.\n" "Do not go too high with this number, as it might cause memory issues. (especially in `parallel` mode)\n" "Do not set it higher than your number of threads of your CPU." ), ) embed_dim: int = Field( 384, description="The dimension of the embeddings stored in the Postgres database", ) class SagemakerSettings(BaseModel): llm_endpoint_name: str embedding_endpoint_name: str class OpenAISettings(BaseModel): api_base: str = Field( None, description="Base URL of OpenAI API. Example: 'https://api.openai.com/v1'.", ) api_key: str model: str = Field( "gpt-3.5-turbo", description="OpenAI Model to use. Example: 'gpt-4'.", ) request_timeout: float = Field( 120.0, description="Time elapsed until openailike server times out the request. Default is 120s. Format is float. ", ) embedding_api_base: str = Field( None, description="Base URL of OpenAI API. Example: 'https://api.openai.com/v1'.", ) embedding_api_key: str embedding_model: str = Field( "text-embedding-ada-002", description="OpenAI embedding Model to use. Example: 'text-embedding-3-large'.", ) class GeminiSettings(BaseModel): api_key: str model: str = Field( "models/gemini-pro", description="Google Model to use. Example: 'models/gemini-pro'.", ) embedding_model: str = Field( "models/embedding-001", description="Google Embedding Model to use. Example: 'models/embedding-001'.", ) class OllamaSettings(BaseModel): api_base: str = Field( "http://localhost:11434", description="Base URL of Ollama API. Example: 'https://localhost:11434'.", ) embedding_api_base: str = Field( "http://localhost:11434", description="Base URL of Ollama embedding API. Example: 'https://localhost:11434'.", ) llm_model: str = Field( None, description="Model to use. Example: 'llama2-uncensored'.", ) embedding_model: str = Field( None, description="Model to use. Example: 'nomic-embed-text'.", ) keep_alive: str = Field( "5m", description="Time the model will stay loaded in memory after a request. examples: 5m, 5h, '-1' ", ) tfs_z: float = Field( 1.0, description="Tail free sampling is used to reduce the impact of less probable tokens from the output. A higher value (e.g., 2.0) will reduce the impact more, while a value of 1.0 disables this setting.", ) num_predict: int = Field( None, description="Maximum number of tokens to predict when generating text. (Default: 128, -1 = infinite generation, -2 = fill context)", ) top_k: int = Field( 40, description="Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative. (Default: 40)", ) top_p: float = Field( 0.9, description="Works together with top-k. A higher value (e.g., 0.95) will lead to more diverse text, while a lower value (e.g., 0.5) will generate more focused and conservative text. (Default: 0.9)", ) repeat_last_n: int = Field( 64, description="Sets how far back for the model to look back to prevent repetition. (Default: 64, 0 = disabled, -1 = num_ctx)", ) repeat_penalty: float = Field( 1.1, description="Sets how strongly to penalize repetitions. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. (Default: 1.1)", ) request_timeout: float = Field( 120.0, description="Time elapsed until ollama times out the request. Default is 120s. Format is float. ", ) autopull_models: bool = Field( False, description="If set to True, the Ollama will automatically pull the models from the API base.", ) class AzureOpenAISettings(BaseModel): api_key: str azure_endpoint: str api_version: str = Field( "2023_05_15", description="The API version to use for this operation. This follows the YYYY-MM-DD format.", ) embedding_deployment_name: str embedding_model: str = Field( "text-embedding-ada-002", description="OpenAI Model to use. Example: 'text-embedding-ada-002'.", ) llm_deployment_name: str llm_model: str = Field( "gpt-35-turbo", description="OpenAI Model to use. Example: 'gpt-4'.", ) class UISettings(BaseModel): enabled: bool path: str default_mode: Literal["RAG", "Search", "Basic", "Summarize"] = Field( "RAG", description="The default mode.", ) default_chat_system_prompt: str = Field( None, description="The default system prompt to use for the chat mode.", ) default_query_system_prompt: str = Field( None, description="The default system prompt to use for the query mode." ) default_summarization_system_prompt: str = Field( None, description="The default system prompt to use for the summarization mode.", ) delete_file_button_enabled: bool = Field( True, description="If the button to delete a file is enabled or not." ) delete_all_files_button_enabled: bool = Field( False, description="If the button to delete all files is enabled or not." ) class RerankSettings(BaseModel): enabled: bool = Field( False, description="This value controls whether a reranker should be included in the RAG pipeline.", ) model: str = Field( "cross-encoder/ms-marco-MiniLM-L-2-v2", description="Rerank model to use. Limited to SentenceTransformer cross-encoder models.", ) top_n: int = Field( 2, description="This value controls the number of documents returned by the RAG pipeline.", ) class RagSettings(BaseModel): similarity_top_k: int = Field( 2, description="This value controls the number of documents returned by the RAG pipeline or considered for reranking if enabled.", ) similarity_value: float = Field( None, description="If set, any documents retrieved from the RAG must meet a certain match score. Acceptable values are between 0 and 1.", ) rerank: RerankSettings class SummarizeSettings(BaseModel): use_async: bool = Field( True, description="If set to True, the summarization will be done asynchronously.", ) class ClickHouseSettings(BaseModel): host: str = Field( "localhost", description="The server hosting the ClickHouse database", ) port: int = Field( 8443, description="The port on which the ClickHouse database is accessible", ) username: str = Field( "default", description="The username to use to connect to the ClickHouse database", ) password: str = Field( "", description="The password to use to connect to the ClickHouse database", ) database: str = Field( "__default__", description="The default database to use for connections", ) secure: bool | str = Field( False, description="Use https/TLS for secure connection to the server", ) interface: str | None = Field( None, description="Must be either 'http' or 'https'. Determines the protocol to use for the connection", ) settings: dict[str, Any] | None = Field( None, description="Specific ClickHouse server settings to be used with the session", ) connect_timeout: int | None = Field( None, description="Timeout in seconds for establishing a connection", ) send_receive_timeout: int | None = Field( None, description="Read timeout in seconds for http connection", ) verify: bool | None = Field( None, description="Verify the server certificate in secure/https mode", ) ca_cert: str | None = Field( None, description="Path to Certificate Authority root certificate (.pem format)", ) client_cert: str | None = Field( None, description="Path to TLS Client certificate (.pem format)", ) client_cert_key: str | None = Field( None, description="Path to the private key for the TLS Client certificate", ) http_proxy: str | None = Field( None, description="HTTP proxy address", ) https_proxy: str | None = Field( None, description="HTTPS proxy address", ) server_host_name: str | None = Field( None, description="Server host name to be checked against the TLS certificate", ) class PostgresSettings(BaseModel): host: str = Field( "localhost", description="The server hosting the Postgres database", ) port: int = Field( 5432, description="The port on which the Postgres database is accessible", ) user: str = Field( "postgres", description="The user to use to connect to the Postgres database", ) password: str = Field( "postgres", description="The password to use to connect to the Postgres database", ) database: str = Field( "postgres", description="The database to use to connect to the Postgres database", ) schema_name: str = Field( "public", description="The name of the schema in the Postgres database to use", ) class QdrantSettings(BaseModel): location: str | None = Field( None, description=( "If `:memory:` - use in-memory Qdrant instance.\n" "If `str` - use it as a `url` parameter.\n" ), ) url: str | None = Field( None, description=( "Either host or str of 'Optional[scheme], host, Optional[port], Optional[prefix]'." ), ) port: int | None = Field(6333, description="Port of the REST API interface.") grpc_port: int | None = Field(6334, description="Port of the gRPC interface.") prefer_grpc: bool | None = Field( False, description="If `true` - use gRPC interface whenever possible in custom methods.", ) https: bool | None = Field( None, description="If `true` - use HTTPS(SSL) protocol.", ) api_key: str | None = Field( None, description="API key for authentication in Qdrant Cloud.", ) prefix: str | None = Field( None, description=( "Prefix to add to the REST URL path." "Example: `service/v1` will result in " "'http://localhost:6333/service/v1/{qdrant-endpoint}' for REST API." ), ) timeout: float | None = Field( None, description="Timeout for REST and gRPC API requests.", ) host: str | None = Field( None, description="Host name of Qdrant service. If url and host are None, set to 'localhost'.", ) path: str | None = Field(None, description="Persistence path for QdrantLocal.") force_disable_check_same_thread: bool | None = Field( True, description=( "For QdrantLocal, force disable check_same_thread. Default: `True`" "Only use this if you can guarantee that you can resolve the thread safety outside QdrantClient." ), ) class MilvusSettings(BaseModel): uri: str = Field( "local_data/private_gpt/milvus/milvus_local.db", description="The URI of the Milvus instance. For example: 'local_data/private_gpt/milvus/milvus_local.db' for Milvus Lite.", ) token: str = Field( "", description=( "A valid access token to access the specified Milvus instance. " "This can be used as a recommended alternative to setting user and password separately. " ), ) collection_name: str = Field( "make_this_parameterizable_per_api_call", description="The name of the collection in Milvus. Default is 'make_this_parameterizable_per_api_call'.", ) overwrite: bool = Field( True, description="Overwrite the previous collection schema if it exists." ) class Settings(BaseModel): server: ServerSettings data: DataSettings ui: UISettings llm: LLMSettings embedding: EmbeddingSettings llamacpp: LlamaCPPSettings huggingface: HuggingFaceSettings sagemaker: SagemakerSettings openai: OpenAISettings gemini: GeminiSettings ollama: OllamaSettings azopenai: AzureOpenAISettings vectorstore: VectorstoreSettings nodestore: NodeStoreSettings rag: RagSettings summarize: SummarizeSettings qdrant: QdrantSettings | None = None postgres: PostgresSettings | None = None clickhouse: ClickHouseSettings | None = None milvus: MilvusSettings | None = None """ This is visible just for DI or testing purposes. Use dependency injection or `settings()` method instead. """ unsafe_settings = load_active_settings() """ This is visible just for DI or testing purposes. Use dependency injection or `settings()` method instead. """ unsafe_typed_settings = Settings(**unsafe_settings) def settings() -> Settings: from private_gpt.di import global_injector return global_injector.get(Settings)
--- +++ @@ -6,6 +6,12 @@ class CorsSettings(BaseModel): + """CORS configuration. + + For more details on the CORS configuration, see: + # * https://fastapi.tiangolo.com/tutorial/cors/ + # * https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS + """ enabled: bool = Field( description="Flag indicating if CORS headers are set or not." @@ -37,6 +43,10 @@ class AuthSettings(BaseModel): + """Authentication configuration. + + The implementation of the authentication strategy must + """ enabled: bool = Field( description="Flag indicating if authentication is enabled or not.", @@ -50,6 +60,15 @@ class IngestionSettings(BaseModel): + """Ingestion configuration. + + This configuration is used to control the ingestion of data into the system + using non-server methods. This is useful for local development and testing; + or to ingest in bulk from a folder. + + Please note that this configuration is not secure and should be used in + a controlled environment only (setting right permissions, etc.). + """ enabled: bool = Field( description="Flag indicating if local ingestion is enabled or not.", @@ -607,6 +626,13 @@ def settings() -> Settings: + """Get the current loaded settings from the DI container. + + This method exists to keep compatibility with the existing code, + that require global access to the settings. + + For regular components use dependency injection instead. + """ from private_gpt.di import global_injector - return global_injector.get(Settings)+ return global_injector.get(Settings)
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/settings/settings.py
Add detailed documentation for each class
# mypy: ignore-errors from __future__ import annotations import io import json import logging from typing import TYPE_CHECKING, Any import boto3 # type: ignore from llama_index.core.base.llms.generic_utils import ( completion_response_to_chat_response, stream_completion_response_to_chat_response, ) from llama_index.core.bridge.pydantic import Field from llama_index.core.llms import ( CompletionResponse, CustomLLM, LLMMetadata, ) from llama_index.core.llms.callbacks import ( llm_chat_callback, llm_completion_callback, ) if TYPE_CHECKING: from collections.abc import Sequence from llama_index.callbacks import CallbackManager from llama_index.llms import ( ChatMessage, ChatResponse, ChatResponseGen, CompletionResponseGen, ) logger = logging.getLogger(__name__) class LineIterator: def __init__(self, stream: Any) -> None: self.byte_iterator = iter(stream) self.buffer = io.BytesIO() self.read_pos = 0 def __iter__(self) -> Any: return self def __next__(self) -> Any: while True: self.buffer.seek(self.read_pos) line = self.buffer.readline() if line and line[-1] == ord("\n"): self.read_pos += len(line) return line[:-1] try: chunk = next(self.byte_iterator) except StopIteration: if self.read_pos < self.buffer.getbuffer().nbytes: continue raise if "PayloadPart" not in chunk: logger.warning("Unknown event type=%s", chunk) continue self.buffer.seek(0, io.SEEK_END) self.buffer.write(chunk["PayloadPart"]["Bytes"]) class SagemakerLLM(CustomLLM): endpoint_name: str = Field(description="") temperature: float = Field(description="The temperature to use for sampling.") max_new_tokens: int = Field(description="The maximum number of tokens to generate.") context_window: int = Field( description="The maximum number of context tokens for the model." ) messages_to_prompt: Any = Field( description="The function to convert messages to a prompt.", exclude=True ) completion_to_prompt: Any = Field( description="The function to convert a completion to a prompt.", exclude=True ) generate_kwargs: dict[str, Any] = Field( default_factory=dict, description="Kwargs used for generation." ) model_kwargs: dict[str, Any] = Field( default_factory=dict, description="Kwargs used for model initialization." ) verbose: bool = Field(description="Whether to print verbose output.") _boto_client: Any = boto3.client( "sagemaker-runtime", ) # TODO make it an optional field def __init__( self, endpoint_name: str | None = "", temperature: float = 0.1, max_new_tokens: int = 512, # to review defaults context_window: int = 2048, # to review defaults messages_to_prompt: Any = None, completion_to_prompt: Any = None, callback_manager: CallbackManager | None = None, generate_kwargs: dict[str, Any] | None = None, model_kwargs: dict[str, Any] | None = None, verbose: bool = True, ) -> None: model_kwargs = model_kwargs or {} model_kwargs.update({"n_ctx": context_window, "verbose": verbose}) messages_to_prompt = messages_to_prompt or {} completion_to_prompt = completion_to_prompt or {} generate_kwargs = generate_kwargs or {} generate_kwargs.update( {"temperature": temperature, "max_tokens": max_new_tokens} ) super().__init__( endpoint_name=endpoint_name, temperature=temperature, context_window=context_window, max_new_tokens=max_new_tokens, messages_to_prompt=messages_to_prompt, completion_to_prompt=completion_to_prompt, callback_manager=callback_manager, generate_kwargs=generate_kwargs, model_kwargs=model_kwargs, verbose=verbose, ) @property def inference_params(self): # TODO expose the rest of params return { "do_sample": True, "top_p": 0.7, "temperature": self.temperature, "top_k": 50, "max_new_tokens": self.max_new_tokens, } @property def metadata(self) -> LLMMetadata: return LLMMetadata( context_window=self.context_window, num_output=self.max_new_tokens, model_name="Sagemaker LLama 2", ) @llm_completion_callback() def complete(self, prompt: str, **kwargs: Any) -> CompletionResponse: self.generate_kwargs.update({"stream": False}) is_formatted = kwargs.pop("formatted", False) if not is_formatted: prompt = self.completion_to_prompt(prompt) request_params = { "inputs": prompt, "stream": False, "parameters": self.inference_params, } resp = self._boto_client.invoke_endpoint( EndpointName=self.endpoint_name, Body=json.dumps(request_params), ContentType="application/json", ) response_body = resp["Body"] response_str = response_body.read().decode("utf-8") response_dict = json.loads(response_str) return CompletionResponse( text=response_dict[0]["generated_text"][len(prompt) :], raw=resp ) @llm_completion_callback() def stream_complete(self, prompt: str, **kwargs: Any) -> CompletionResponseGen: def get_stream(): text = "" request_params = { "inputs": prompt, "stream": True, "parameters": self.inference_params, } resp = self._boto_client.invoke_endpoint_with_response_stream( EndpointName=self.endpoint_name, Body=json.dumps(request_params), ContentType="application/json", ) event_stream = resp["Body"] start_json = b"{" stop_token = "<|endoftext|>" first_token = True for line in LineIterator(event_stream): if line != b"" and start_json in line: data = json.loads(line[line.find(start_json) :].decode("utf-8")) special = data["token"]["special"] stop = data["token"]["text"] == stop_token if not special and not stop: delta = data["token"]["text"] # trim the leading space for the first token if present if first_token: delta = delta.lstrip() first_token = False text += delta yield CompletionResponse(delta=delta, text=text, raw=data) return get_stream() @llm_chat_callback() def chat(self, messages: Sequence[ChatMessage], **kwargs: Any) -> ChatResponse: prompt = self.messages_to_prompt(messages) completion_response = self.complete(prompt, formatted=True, **kwargs) return completion_response_to_chat_response(completion_response) @llm_chat_callback() def stream_chat( self, messages: Sequence[ChatMessage], **kwargs: Any ) -> ChatResponseGen: prompt = self.messages_to_prompt(messages) completion_response = self.stream_complete(prompt, formatted=True, **kwargs) return stream_completion_response_to_chat_response(completion_response)
--- +++ @@ -37,16 +37,46 @@ class LineIterator: + r"""A helper class for parsing the byte stream input from TGI container. + + The output of the model will be in the following format: + ``` + b'data:{"token": {"text": " a"}}\n\n' + b'data:{"token": {"text": " challenging"}}\n\n' + b'data:{"token": {"text": " problem" + b'}}' + ... + ``` + + While usually each PayloadPart event from the event stream will contain a byte array + with a full json, this is not guaranteed and some of the json objects may be split + across PayloadPart events. For example: + ``` + {'PayloadPart': {'Bytes': b'{"outputs": '}} + {'PayloadPart': {'Bytes': b'[" problem"]}\n'}} + ``` + + + This class accounts for this by concatenating bytes written via the 'write' function + and then exposing a method which will return lines (ending with a '\n' character) + within the buffer via the 'scan_lines' function. It maintains the position of the + last read position to ensure that previous bytes are not exposed again. It will + also save any pending lines that doe not end with a '\n' to make sure truncations + are concatinated + """ def __init__(self, stream: Any) -> None: + """Line iterator initializer.""" self.byte_iterator = iter(stream) self.buffer = io.BytesIO() self.read_pos = 0 def __iter__(self) -> Any: + """Self iterator.""" return self def __next__(self) -> Any: + """Next element from iterator.""" while True: self.buffer.seek(self.read_pos) line = self.buffer.readline() @@ -67,6 +97,22 @@ class SagemakerLLM(CustomLLM): + """Sagemaker Inference Endpoint models. + + To use, you must supply the endpoint name from your deployed + Sagemaker model & the region where it is deployed. + + To authenticate, the AWS client uses the following methods to + automatically load credentials: + https://boto3.amazonaws.com/v1/documentation/api/latest/guide/credentials.html + + If a specific credential profile should be used, you must pass + the name of the profile from the ~/.aws/credentials file that is to be used. + + Make sure the credentials / roles used have the required policies to + access the Sagemaker endpoint. + See: https://docs.aws.amazon.com/IAM/latest/UserGuide/access_policies.html + """ endpoint_name: str = Field(description="") temperature: float = Field(description="The temperature to use for sampling.") @@ -105,6 +151,7 @@ model_kwargs: dict[str, Any] | None = None, verbose: bool = True, ) -> None: + """SagemakerLLM initializer.""" model_kwargs = model_kwargs or {} model_kwargs.update({"n_ctx": context_window, "verbose": verbose}) @@ -142,6 +189,7 @@ @property def metadata(self) -> LLMMetadata: + """Get LLM metadata.""" return LLMMetadata( context_window=self.context_window, num_output=self.max_new_tokens, @@ -225,4 +273,4 @@ ) -> ChatResponseGen: prompt = self.messages_to_prompt(messages) completion_response = self.stream_complete(prompt, formatted=True, **kwargs) - return stream_completion_response_to_chat_response(completion_response)+ return stream_completion_response_to_chat_response(completion_response)
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/components/llm/custom/sagemaker.py
Add standardized docstrings across the file
from typing import Literal from fastapi import APIRouter, Depends, HTTPException, Request, UploadFile from pydantic import BaseModel, Field from private_gpt.server.ingest.ingest_service import IngestService from private_gpt.server.ingest.model import IngestedDoc from private_gpt.server.utils.auth import authenticated ingest_router = APIRouter(prefix="/v1", dependencies=[Depends(authenticated)]) class IngestTextBody(BaseModel): file_name: str = Field(examples=["Avatar: The Last Airbender"]) text: str = Field( examples=[ "Avatar is set in an Asian and Arctic-inspired world in which some " "people can telekinetically manipulate one of the four elements—water, " "earth, fire or air—through practices known as 'bending', inspired by " "Chinese martial arts." ] ) class IngestResponse(BaseModel): object: Literal["list"] model: Literal["private-gpt"] data: list[IngestedDoc] @ingest_router.post("/ingest", tags=["Ingestion"], deprecated=True) def ingest(request: Request, file: UploadFile) -> IngestResponse: return ingest_file(request, file) @ingest_router.post("/ingest/file", tags=["Ingestion"]) def ingest_file(request: Request, file: UploadFile) -> IngestResponse: service = request.state.injector.get(IngestService) if file.filename is None: raise HTTPException(400, "No file name provided") ingested_documents = service.ingest_bin_data(file.filename, file.file) return IngestResponse(object="list", model="private-gpt", data=ingested_documents) @ingest_router.post("/ingest/text", tags=["Ingestion"]) def ingest_text(request: Request, body: IngestTextBody) -> IngestResponse: service = request.state.injector.get(IngestService) if len(body.file_name) == 0: raise HTTPException(400, "No file name provided") ingested_documents = service.ingest_text(body.file_name, body.text) return IngestResponse(object="list", model="private-gpt", data=ingested_documents) @ingest_router.get("/ingest/list", tags=["Ingestion"]) def list_ingested(request: Request) -> IngestResponse: service = request.state.injector.get(IngestService) ingested_documents = service.list_ingested() return IngestResponse(object="list", model="private-gpt", data=ingested_documents) @ingest_router.delete("/ingest/{doc_id}", tags=["Ingestion"]) def delete_ingested(request: Request, doc_id: str) -> None: service = request.state.injector.get(IngestService) service.delete(doc_id)
--- +++ @@ -30,11 +30,30 @@ @ingest_router.post("/ingest", tags=["Ingestion"], deprecated=True) def ingest(request: Request, file: UploadFile) -> IngestResponse: + """Ingests and processes a file. + + Deprecated. Use ingest/file instead. + """ return ingest_file(request, file) @ingest_router.post("/ingest/file", tags=["Ingestion"]) def ingest_file(request: Request, file: UploadFile) -> IngestResponse: + """Ingests and processes a file, storing its chunks to be used as context. + + The context obtained from files is later used in + `/chat/completions`, `/completions`, and `/chunks` APIs. + + Most common document + formats are supported, but you may be prompted to install an extra dependency to + manage a specific file type. + + A file can generate different Documents (for example a PDF generates one Document + per page). All Documents IDs are returned in the response, together with the + extracted Metadata (which is later used to improve context retrieval). Those IDs + can be used to filter the context used to create responses in + `/chat/completions`, `/completions`, and `/chunks` APIs. + """ service = request.state.injector.get(IngestService) if file.filename is None: raise HTTPException(400, "No file name provided") @@ -44,6 +63,17 @@ @ingest_router.post("/ingest/text", tags=["Ingestion"]) def ingest_text(request: Request, body: IngestTextBody) -> IngestResponse: + """Ingests and processes a text, storing its chunks to be used as context. + + The context obtained from files is later used in + `/chat/completions`, `/completions`, and `/chunks` APIs. + + A Document will be generated with the given text. The Document + ID is returned in the response, together with the + extracted Metadata (which is later used to improve context retrieval). That ID + can be used to filter the context used to create responses in + `/chat/completions`, `/completions`, and `/chunks` APIs. + """ service = request.state.injector.get(IngestService) if len(body.file_name) == 0: raise HTTPException(400, "No file name provided") @@ -53,6 +83,11 @@ @ingest_router.get("/ingest/list", tags=["Ingestion"]) def list_ingested(request: Request) -> IngestResponse: + """Lists already ingested Documents including their Document ID and metadata. + + Those IDs can be used to filter the context used to create responses + in `/chat/completions`, `/completions`, and `/chunks` APIs. + """ service = request.state.injector.get(IngestService) ingested_documents = service.list_ingested() return IngestResponse(object="list", model="private-gpt", data=ingested_documents) @@ -60,5 +95,10 @@ @ingest_router.delete("/ingest/{doc_id}", tags=["Ingestion"]) def delete_ingested(request: Request, doc_id: str) -> None: + """Delete the specified ingested Document. + + The `doc_id` can be obtained from the `GET /ingest/list` endpoint. + The document will be effectively deleted from your storage context. + """ service = request.state.injector.get(IngestService) - service.delete(doc_id)+ service.delete(doc_id)
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/server/ingest/ingest_router.py
Add well-formatted docstrings
import datetime import logging import math import time from collections import deque from typing import Any logger = logging.getLogger(__name__) def human_time(*args: Any, **kwargs: Any) -> str: def timedelta_total_seconds(timedelta: datetime.timedelta) -> float: return ( timedelta.microseconds + 0.0 + (timedelta.seconds + timedelta.days * 24 * 3600) * 10**6 ) / 10**6 secs = float(timedelta_total_seconds(datetime.timedelta(*args, **kwargs))) # We want (ms) precision below 2 seconds if secs < 2: return f"{secs * 1000}ms" units = [("y", 86400 * 365), ("d", 86400), ("h", 3600), ("m", 60), ("s", 1)] parts = [] for unit, mul in units: if secs / mul >= 1 or mul == 1: if mul > 1: n = int(math.floor(secs / mul)) secs -= n * mul else: # >2s we drop the (ms) component. n = int(secs) if n: parts.append(f"{n}{unit}") return " ".join(parts) def eta(iterator: list[Any]) -> Any: total = len(iterator) _eta = ETA(total) _eta.needReport(30) for processed, data in enumerate(iterator, start=1): yield data _eta.update(processed) if _eta.needReport(60): logger.info(f"{processed}/{total} - ETA {_eta.human_time()}") class ETA: def __init__(self, total: int): self.total: int = total # Total expected records. self.rate: float = 0.0 # per second self._timing_data: deque[tuple[float, int]] = deque(maxlen=100) self.secondsLeft: float = 0.0 self.nexttime: float = 0.0 def human_time(self) -> str: if self._calc(): return f"{human_time(seconds=self.secondsLeft)} @ {int(self.rate * 60)}/min" return "(computing)" def update(self, count: int) -> None: # count should be in the range 0 to self.total assert count > 0 assert count <= self.total self._timing_data.append((time.time(), count)) # (X,Y) for pearson def needReport(self, whenSecs: int) -> bool: now = time.time() if now > self.nexttime: self.nexttime = now + whenSecs return True return False def _calc(self) -> bool: # A sample before a prediction. Need two points to compute slope! if len(self._timing_data) < 3: return False # http://en.wikipedia.org/wiki/Pearson_product-moment_correlation_coefficient # Calculate means and standard deviations. samples = len(self._timing_data) # column wise sum of the timing tuples to compute their mean. mean_x, mean_y = ( sum(i) / samples for i in zip(*self._timing_data, strict=False) ) std_x = math.sqrt( sum(pow(i[0] - mean_x, 2) for i in self._timing_data) / (samples - 1) ) std_y = math.sqrt( sum(pow(i[1] - mean_y, 2) for i in self._timing_data) / (samples - 1) ) # Calculate coefficient. sum_xy, sum_sq_v_x, sum_sq_v_y = 0.0, 0.0, 0 for x, y in self._timing_data: x -= mean_x y -= mean_y sum_xy += x * y sum_sq_v_x += pow(x, 2) sum_sq_v_y += pow(y, 2) pearson_r = sum_xy / math.sqrt(sum_sq_v_x * sum_sq_v_y) # Calculate regression line. # y = mx + b where m is the slope and b is the y-intercept. m = self.rate = pearson_r * (std_y / std_x) y = self.total b = mean_y - m * mean_x x = (y - b) / m # Calculate fitted line (transformed/shifted regression line horizontally). fitted_b = self._timing_data[-1][1] - (m * self._timing_data[-1][0]) fitted_x = (y - fitted_b) / m _, count = self._timing_data[-1] # adjust last data point progress count adjusted_x = ((fitted_x - x) * (count / self.total)) + x eta_epoch = adjusted_x self.secondsLeft = max([eta_epoch - time.time(), 0]) return True
--- +++ @@ -36,6 +36,7 @@ def eta(iterator: list[Any]) -> Any: + """Report an ETA after 30s and every 60s thereafter.""" total = len(iterator) _eta = ETA(total) _eta.needReport(30) @@ -47,6 +48,7 @@ class ETA: + """Predict how long something will take to complete.""" def __init__(self, total: int): self.total: int = total # Total expected records. @@ -117,4 +119,4 @@ eta_epoch = adjusted_x self.secondsLeft = max([eta_epoch - time.time(), 0]) - return True+ return True
https://raw.githubusercontent.com/zylon-ai/private-gpt/HEAD/private_gpt/utils/eta.py
Create documentation for each function signature
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import io import torch from flask import Flask, request from PIL import Image app = Flask(__name__) models = {} DETECTION_URL = "/v1/object-detection/<model>" @app.route(DETECTION_URL, methods=["POST"]) def predict(model): if request.method != "POST": return if request.files.get("image"): # Method 1 # with request.files["image"] as f: # im = Image.open(io.BytesIO(f.read())) # Method 2 im_file = request.files["image"] im_bytes = im_file.read() im = Image.open(io.BytesIO(im_bytes)) if model in models: results = models[model](im, size=640) # reduce size=320 for faster inference return results.pandas().xyxy[0].to_json(orient="records") if __name__ == "__main__": parser = argparse.ArgumentParser(description="Flask API exposing YOLOv5 model") parser.add_argument("--port", default=5000, type=int, help="port number") parser.add_argument("--model", nargs="+", default=["yolov5s"], help="model(s) to run, i.e. --model yolov5n yolov5s") opt = parser.parse_args() for m in opt.model: models[m] = torch.hub.load("ultralytics/yolov5", m, force_reload=True, skip_validation=True) app.run(host="0.0.0.0", port=opt.port) # debug=True causes Restarting with stat
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Run a Flask REST API exposing one or more YOLOv5s models.""" import argparse import io @@ -15,6 +16,9 @@ @app.route(DETECTION_URL, methods=["POST"]) def predict(model): + """Predict and return object detections in JSON format given an image and model name via a Flask REST API POST + request. + """ if request.method != "POST": return @@ -42,4 +46,4 @@ for m in opt.model: models[m] = torch.hub.load("ultralytics/yolov5", m, force_reload=True, skip_validation=True) - app.run(host="0.0.0.0", port=opt.port) # debug=True causes Restarting with stat+ app.run(host="0.0.0.0", port=opt.port) # debug=True causes Restarting with stat
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/flask_rest_api/restapi.py
Add standardized docstrings across the file
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import math import random import cv2 import numpy as np from ..augmentations import box_candidates from ..general import resample_segments, segment2box def mixup(im, labels, segments, im2, labels2, segments2): r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0 im = (im * r + im2 * (1 - r)).astype(np.uint8) labels = np.concatenate((labels, labels2), 0) segments = np.concatenate((segments, segments2), 0) return im, labels, segments def random_perspective( im, targets=(), segments=(), degrees=10, translate=0.1, scale=0.1, shear=10, perspective=0.0, border=(0, 0) ): # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(.1, .1), scale=(.9, 1.1), shear=(-10, 10)) # targets = [cls, xyxy] height = im.shape[0] + border[0] * 2 # shape(h,w,c) width = im.shape[1] + border[1] * 2 # Center C = np.eye(3) C[0, 2] = -im.shape[1] / 2 # x translation (pixels) C[1, 2] = -im.shape[0] / 2 # y translation (pixels) # Perspective P = np.eye(3) P[2, 0] = random.uniform(-perspective, perspective) # x perspective (about y) P[2, 1] = random.uniform(-perspective, perspective) # y perspective (about x) # Rotation and Scale R = np.eye(3) a = random.uniform(-degrees, degrees) # a += random.choice([-180, -90, 0, 90]) # add 90deg rotations to small rotations s = random.uniform(1 - scale, 1 + scale) # s = 2 ** random.uniform(-scale, scale) R[:2] = cv2.getRotationMatrix2D(angle=a, center=(0, 0), scale=s) # Shear S = np.eye(3) S[0, 1] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # x shear (deg) S[1, 0] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # y shear (deg) # Translation T = np.eye(3) T[0, 2] = random.uniform(0.5 - translate, 0.5 + translate) * width # x translation (pixels) T[1, 2] = random.uniform(0.5 - translate, 0.5 + translate) * height # y translation (pixels) # Combined rotation matrix M = T @ S @ R @ P @ C # order of operations (right to left) is IMPORTANT if (border[0] != 0) or (border[1] != 0) or (M != np.eye(3)).any(): # image changed if perspective: im = cv2.warpPerspective(im, M, dsize=(width, height), borderValue=(114, 114, 114)) else: # affine im = cv2.warpAffine(im, M[:2], dsize=(width, height), borderValue=(114, 114, 114)) new_segments = [] if n := len(targets): new = np.zeros((n, 4)) segments = resample_segments(segments) # upsample for i, segment in enumerate(segments): xy = np.ones((len(segment), 3)) xy[:, :2] = segment xy = xy @ M.T # transform xy = xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2] # perspective rescale or affine # clip new[i] = segment2box(xy, width, height) new_segments.append(xy) # filter candidates i = box_candidates(box1=targets[:, 1:5].T * s, box2=new.T, area_thr=0.01) targets = targets[i] targets[:, 1:5] = new[i] new_segments = np.array(new_segments)[i] return im, targets, new_segments
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Image augmentation functions.""" import math import random @@ -11,6 +12,10 @@ def mixup(im, labels, segments, im2, labels2, segments2): + """Applies MixUp augmentation blending two images, labels, and segments with a random ratio. + + See https://arxiv.org/pdf/1710.09412.pdf + """ r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0 im = (im * r + im2 * (1 - r)).astype(np.uint8) labels = np.concatenate((labels, labels2), 0) @@ -23,6 +28,7 @@ ): # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(.1, .1), scale=(.9, 1.1), shear=(-10, 10)) # targets = [cls, xyxy] + """Applies random perspective, rotation, scale, shear, and translation augmentations to an image and targets.""" height = im.shape[0] + border[0] * 2 # shape(h,w,c) width = im.shape[1] + border[1] * 2 @@ -82,4 +88,4 @@ targets[:, 1:5] = new[i] new_segments = np.array(new_segments)[i] - return im, targets, new_segments+ return im, targets, new_segments
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/segment/augmentations.py
Add docstrings to clarify complex logic
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license from copy import deepcopy import numpy as np import torch from utils.general import LOGGER, colorstr from utils.torch_utils import profile def check_train_batch_size(model, imgsz=640, amp=True): with torch.cuda.amp.autocast(amp): return autobatch(deepcopy(model).train(), imgsz) # compute optimal batch size def autobatch(model, imgsz=640, fraction=0.8, batch_size=16): # Usage: # import torch # from utils.autobatch import autobatch # model = torch.hub.load('ultralytics/yolov5', 'yolov5s', autoshape=False) # print(autobatch(model)) # Check device prefix = colorstr("AutoBatch: ") LOGGER.info(f"{prefix}Computing optimal batch size for --imgsz {imgsz}") device = next(model.parameters()).device # get model device if device.type == "cpu": LOGGER.info(f"{prefix}CUDA not detected, using default CPU batch-size {batch_size}") return batch_size if torch.backends.cudnn.benchmark: LOGGER.info(f"{prefix} ⚠️ Requires torch.backends.cudnn.benchmark=False, using default batch-size {batch_size}") return batch_size # Inspect CUDA memory gb = 1 << 30 # bytes to GiB (1024 ** 3) d = str(device).upper() # 'CUDA:0' properties = torch.cuda.get_device_properties(device) # device properties t = properties.total_memory / gb # GiB total r = torch.cuda.memory_reserved(device) / gb # GiB reserved a = torch.cuda.memory_allocated(device) / gb # GiB allocated f = t - (r + a) # GiB free LOGGER.info(f"{prefix}{d} ({properties.name}) {t:.2f}G total, {r:.2f}G reserved, {a:.2f}G allocated, {f:.2f}G free") # Profile batch sizes batch_sizes = [1, 2, 4, 8, 16] try: img = [torch.empty(b, 3, imgsz, imgsz) for b in batch_sizes] results = profile(img, model, n=3, device=device) except Exception as e: LOGGER.warning(f"{prefix}{e}") # Fit a solution y = [x[2] for x in results if x] # memory [2] p = np.polyfit(batch_sizes[: len(y)], y, deg=1) # first degree polynomial fit b = int((f * fraction - p[1]) / p[0]) # y intercept (optimal batch size) if None in results: # some sizes failed i = results.index(None) # first fail index if b >= batch_sizes[i]: # y intercept above failure point b = batch_sizes[max(i - 1, 0)] # select prior safe point if b < 1 or b > 1024: # b outside of safe range b = batch_size LOGGER.warning(f"{prefix}WARNING ⚠️ CUDA anomaly detected, recommend restart environment and retry command.") fraction = (np.polyval(p, b) + r + a) / t # actual fraction predicted LOGGER.info(f"{prefix}Using batch-size {b} for {d} {t * fraction:.2f}G/{t:.2f}G ({fraction * 100:.0f}%) ✅") return b
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Auto-batch utils.""" from copy import deepcopy @@ -10,11 +11,13 @@ def check_train_batch_size(model, imgsz=640, amp=True): + """Checks and computes optimal training batch size for YOLOv5 model, given image size and AMP setting.""" with torch.cuda.amp.autocast(amp): return autobatch(deepcopy(model).train(), imgsz) # compute optimal batch size def autobatch(model, imgsz=640, fraction=0.8, batch_size=16): + """Estimates optimal YOLOv5 batch size using `fraction` of CUDA memory.""" # Usage: # import torch # from utils.autobatch import autobatch @@ -64,4 +67,4 @@ fraction = (np.polyval(p, b) + r + a) / t # actual fraction predicted LOGGER.info(f"{prefix}Using batch-size {b} for {d} {t * fraction:.2f}G/{t:.2f}G ({fraction * 100:.0f}%) ✅") - return b+ return b
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/autobatch.py
Write docstrings that follow conventions
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import glob import re from pathlib import Path import matplotlib.image as mpimg import matplotlib.pyplot as plt import numpy as np import yaml from ultralytics.utils.plotting import Annotator, colors try: import clearml from clearml import Dataset, Task assert hasattr(clearml, "__version__") # verify package import not local dir except (ImportError, AssertionError): clearml = None def construct_dataset(clearml_info_string): dataset_id = clearml_info_string.replace("clearml://", "") dataset = Dataset.get(dataset_id=dataset_id) dataset_root_path = Path(dataset.get_local_copy()) # We'll search for the yaml file definition in the dataset yaml_filenames = list(glob.glob(str(dataset_root_path / "*.yaml")) + glob.glob(str(dataset_root_path / "*.yml"))) if len(yaml_filenames) > 1: raise ValueError( "More than one yaml file was found in the dataset root, cannot determine which one contains " "the dataset definition this way." ) elif not yaml_filenames: raise ValueError( "No yaml definition found in dataset root path, check that there is a correct yaml file " "inside the dataset root path." ) with open(yaml_filenames[0]) as f: dataset_definition = yaml.safe_load(f) assert set(dataset_definition.keys()).issuperset({"train", "test", "val", "nc", "names"}), ( "The right keys were not found in the yaml file, make sure it at least has the following keys: ('train', 'test', 'val', 'nc', 'names')" ) data_dict = { "train": ( str((dataset_root_path / dataset_definition["train"]).resolve()) if dataset_definition["train"] else None ) } data_dict["test"] = ( str((dataset_root_path / dataset_definition["test"]).resolve()) if dataset_definition["test"] else None ) data_dict["val"] = ( str((dataset_root_path / dataset_definition["val"]).resolve()) if dataset_definition["val"] else None ) data_dict["nc"] = dataset_definition["nc"] data_dict["names"] = dataset_definition["names"] return data_dict class ClearmlLogger: def __init__(self, opt, hyp): self.current_epoch = 0 # Keep tracked of amount of logged images to enforce a limit self.current_epoch_logged_images = set() # Maximum number of images to log to clearML per epoch self.max_imgs_to_log_per_epoch = 16 # Get the interval of epochs when bounding box images should be logged # Only for detection task though! if "bbox_interval" in opt: self.bbox_interval = opt.bbox_interval self.clearml = clearml self.task = None self.data_dict = None if self.clearml: self.task = Task.init( project_name="YOLOv5" if str(opt.project).startswith("runs/") else opt.project, task_name=opt.name if opt.name != "exp" else "Training", tags=["YOLOv5"], output_uri=True, reuse_last_task_id=opt.exist_ok, auto_connect_frameworks={"pytorch": False, "matplotlib": False}, # We disconnect pytorch auto-detection, because we added manual model save points in the code ) # ClearML's hooks will already grab all general parameters # Only the hyperparameters coming from the yaml config file # will have to be added manually! self.task.connect(hyp, name="Hyperparameters") self.task.connect(opt, name="Args") # Make sure the code is easily remotely runnable by setting the docker image to use by the remote agent self.task.set_base_docker( "ultralytics/yolov5:latest", docker_arguments=':ipc=host -e="CLEARML_AGENT_SKIP_PYTHON_ENV_INSTALL=1"', docker_setup_bash_script="pip install clearml", ) # Get ClearML Dataset Version if requested if opt.data.startswith("clearml://"): # data_dict should have the following keys: # names, nc (number of classes), test, train, val (all three relative paths to ../datasets) self.data_dict = construct_dataset(opt.data) # Set data to data_dict because wandb will crash without this information and opt is the best way # to give it to them opt.data = self.data_dict def log_scalars(self, metrics, epoch): for k, v in metrics.items(): title, series = k.split("/") self.task.get_logger().report_scalar(title, series, v, epoch) def log_model(self, model_path, model_name, epoch=0): self.task.update_output_model( model_path=str(model_path), name=model_name, iteration=epoch, auto_delete_file=False ) def log_summary(self, metrics): for k, v in metrics.items(): self.task.get_logger().report_single_value(k, v) def log_plot(self, title, plot_path): img = mpimg.imread(plot_path) fig = plt.figure() ax = fig.add_axes([0, 0, 1, 1], frameon=False, aspect="auto", xticks=[], yticks=[]) # no ticks ax.imshow(img) self.task.get_logger().report_matplotlib_figure(title, "", figure=fig, report_interactive=False) def log_debug_samples(self, files, title="Debug Samples"): for f in files: if f.exists(): it = re.search(r"_batch(\d+)", f.name) iteration = int(it.groups()[0]) if it else 0 self.task.get_logger().report_image( title=title, series=f.name.replace(f"_batch{iteration}", ""), local_path=str(f), iteration=iteration ) def log_image_with_boxes(self, image_path, boxes, class_names, image, conf_threshold=0.25): if ( len(self.current_epoch_logged_images) < self.max_imgs_to_log_per_epoch and self.current_epoch >= 0 and (self.current_epoch % self.bbox_interval == 0 and image_path not in self.current_epoch_logged_images) ): im = np.ascontiguousarray(np.moveaxis(image.mul(255).clamp(0, 255).byte().cpu().numpy(), 0, 2)) annotator = Annotator(im=im, pil=True) for i, (conf, class_nr, box) in enumerate(zip(boxes[:, 4], boxes[:, 5], boxes[:, :4])): color = colors(i) class_name = class_names[int(class_nr)] confidence_percentage = round(float(conf) * 100, 2) label = f"{class_name}: {confidence_percentage}%" if conf > conf_threshold: annotator.rectangle(box.cpu().numpy(), outline=color) annotator.box_label(box.cpu().numpy(), label=label, color=color) annotated_image = annotator.result() self.task.get_logger().report_image( title="Bounding Boxes", series=image_path.name, iteration=self.current_epoch, image=annotated_image ) self.current_epoch_logged_images.add(image_path)
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Main Logger class for ClearML experiment tracking.""" import glob import re @@ -20,6 +21,7 @@ def construct_dataset(clearml_info_string): + """Load in a clearml dataset and fill the internal data_dict with its contents.""" dataset_id = clearml_info_string.replace("clearml://", "") dataset = Dataset.get(dataset_id=dataset_id) dataset_root_path = Path(dataset.get_local_copy()) @@ -61,8 +63,23 @@ class ClearmlLogger: + """Log training runs, datasets, models, and predictions to ClearML. + + This logger sends information to ClearML at app.clear.ml or to your own hosted server. By default, this information + includes hyperparameters, system configuration and metrics, model metrics, code information and basic data metrics + and analyses. + + By providing additional command line arguments to train.py, datasets, models and predictions can also be logged. + """ def __init__(self, opt, hyp): + """- Initialize ClearML Task, this object will capture the experiment - Upload dataset version to ClearML Data + if opt.upload_dataset is True. + + Args: + opt (namespace): Commandline arguments for this run + hyp (dict): Hyperparameters for this run + """ self.current_epoch = 0 # Keep tracked of amount of logged images to enforce a limit self.current_epoch_logged_images = set() @@ -108,20 +125,44 @@ opt.data = self.data_dict def log_scalars(self, metrics, epoch): + """Log scalars/metrics to ClearML. + + Args: + metrics (dict): Metrics in dict format: {"metrics/mAP": 0.8, ...} + epoch (int): iteration number for the current set of metrics + """ for k, v in metrics.items(): title, series = k.split("/") self.task.get_logger().report_scalar(title, series, v, epoch) def log_model(self, model_path, model_name, epoch=0): + """Log model weights to ClearML. + + Args: + model_path (PosixPath or str): Path to the model weights + model_name (str): Name of the model visible in ClearML + epoch (int): Iteration / epoch of the model weights + """ self.task.update_output_model( model_path=str(model_path), name=model_name, iteration=epoch, auto_delete_file=False ) def log_summary(self, metrics): + """Log final metrics to a summary table. + + Args: + metrics (dict): Metrics in dict format: {"metrics/mAP": 0.8, ...} + """ for k, v in metrics.items(): self.task.get_logger().report_single_value(k, v) def log_plot(self, title, plot_path): + """Log image as plot in the plot section of ClearML. + + Args: + title (str): Title of the plot + plot_path (PosixPath or str): Path to the saved image file + """ img = mpimg.imread(plot_path) fig = plt.figure() ax = fig.add_axes([0, 0, 1, 1], frameon=False, aspect="auto", xticks=[], yticks=[]) # no ticks @@ -130,6 +171,12 @@ self.task.get_logger().report_matplotlib_figure(title, "", figure=fig, report_interactive=False) def log_debug_samples(self, files, title="Debug Samples"): + """Log files (images) as debug samples in the ClearML task. + + Args: + files (List(PosixPath)): a list of file paths in PosixPath format + title (str): A title that groups together images with the same values + """ for f in files: if f.exists(): it = re.search(r"_batch(\d+)", f.name) @@ -139,6 +186,14 @@ ) def log_image_with_boxes(self, image_path, boxes, class_names, image, conf_threshold=0.25): + """Draw the bounding boxes on a single image and report the result as a ClearML debug sample. + + Args: + image_path (PosixPath) the path the original image file + boxes (list): list of scaled predictions in the format - [xmin, ymin, xmax, ymax, confidence, class] + class_names (dict): dict containing mapping of class int to class name + image (Tensor): A torch tensor containing the actual image data + """ if ( len(self.current_epoch_logged_images) < self.max_imgs_to_log_per_epoch and self.current_epoch >= 0 @@ -161,4 +216,4 @@ self.task.get_logger().report_image( title="Bounding Boxes", series=image_path.name, iteration=self.current_epoch, image=annotated_image ) - self.current_epoch_logged_images.add(image_path)+ self.current_epoch_logged_images.add(image_path)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/loggers/clearml/clearml_utils.py
Document this code for team use
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import json import logging import os import sys from pathlib import Path import comet_ml logger = logging.getLogger(__name__) FILE = Path(__file__).resolve() ROOT = FILE.parents[3] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH from train import train from utils.callbacks import Callbacks from utils.general import increment_path from utils.torch_utils import select_device # Project Configuration config = comet_ml.config.get_config() COMET_PROJECT_NAME = config.get_string(os.getenv("COMET_PROJECT_NAME"), "comet.project_name", default="yolov5") def get_args(known=False): parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s.pt", help="initial weights path") parser.add_argument("--cfg", type=str, default="", help="model.yaml path") parser.add_argument("--data", type=str, default=ROOT / "data/coco128.yaml", help="dataset.yaml path") parser.add_argument("--hyp", type=str, default=ROOT / "data/hyps/hyp.scratch-low.yaml", help="hyperparameters path") parser.add_argument("--epochs", type=int, default=300, help="total training epochs") parser.add_argument("--batch-size", type=int, default=16, help="total batch size for all GPUs, -1 for autobatch") parser.add_argument("--imgsz", "--img", "--img-size", type=int, default=640, help="train, val image size (pixels)") parser.add_argument("--rect", action="store_true", help="rectangular training") parser.add_argument("--resume", nargs="?", const=True, default=False, help="resume most recent training") parser.add_argument("--nosave", action="store_true", help="only save final checkpoint") parser.add_argument("--noval", action="store_true", help="only validate final epoch") parser.add_argument("--noautoanchor", action="store_true", help="disable AutoAnchor") parser.add_argument("--noplots", action="store_true", help="save no plot files") parser.add_argument("--evolve", type=int, nargs="?", const=300, help="evolve hyperparameters for x generations") parser.add_argument("--bucket", type=str, default="", help="gsutil bucket") parser.add_argument("--cache", type=str, nargs="?", const="ram", help='--cache images in "ram" (default) or "disk"') parser.add_argument("--image-weights", action="store_true", help="use weighted image selection for training") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--multi-scale", action="store_true", help="vary img-size +/- 50%%") parser.add_argument("--single-cls", action="store_true", help="train multi-class data as single-class") parser.add_argument("--optimizer", type=str, choices=["SGD", "Adam", "AdamW"], default="SGD", help="optimizer") parser.add_argument("--sync-bn", action="store_true", help="use SyncBatchNorm, only available in DDP mode") parser.add_argument("--workers", type=int, default=8, help="max dataloader workers (per RANK in DDP mode)") parser.add_argument("--project", default=ROOT / "runs/train", help="save to project/name") parser.add_argument("--name", default="exp", help="save to project/name") parser.add_argument("--exist-ok", action="store_true", help="existing project/name ok, do not increment") parser.add_argument("--quad", action="store_true", help="quad dataloader") parser.add_argument("--cos-lr", action="store_true", help="cosine LR scheduler") parser.add_argument("--label-smoothing", type=float, default=0.0, help="Label smoothing epsilon") parser.add_argument("--patience", type=int, default=100, help="EarlyStopping patience (epochs without improvement)") parser.add_argument("--freeze", nargs="+", type=int, default=[0], help="Freeze layers: backbone=10, first3=0 1 2") parser.add_argument("--save-period", type=int, default=-1, help="Save checkpoint every x epochs (disabled if < 1)") parser.add_argument("--seed", type=int, default=0, help="Global training seed") parser.add_argument("--local_rank", type=int, default=-1, help="Automatic DDP Multi-GPU argument, do not modify") # Weights & Biases arguments parser.add_argument("--entity", default=None, help="W&B: Entity") parser.add_argument("--upload_dataset", nargs="?", const=True, default=False, help='W&B: Upload data, "val" option') parser.add_argument("--bbox_interval", type=int, default=-1, help="W&B: Set bounding-box image logging interval") parser.add_argument("--artifact_alias", type=str, default="latest", help="W&B: Version of dataset artifact to use") # Comet Arguments parser.add_argument("--comet_optimizer_config", type=str, help="Comet: Path to a Comet Optimizer Config File.") parser.add_argument("--comet_optimizer_id", type=str, help="Comet: ID of the Comet Optimizer sweep.") parser.add_argument("--comet_optimizer_objective", type=str, help="Comet: Set to 'minimize' or 'maximize'.") parser.add_argument("--comet_optimizer_metric", type=str, help="Comet: Metric to Optimize.") parser.add_argument( "--comet_optimizer_workers", type=int, default=1, help="Comet: Number of Parallel Workers to use with the Comet Optimizer.", ) return parser.parse_known_args()[0] if known else parser.parse_args() def run(parameters, opt): hyp_dict = {k: v for k, v in parameters.items() if k not in ["epochs", "batch_size"]} opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok or opt.evolve)) opt.batch_size = parameters.get("batch_size") opt.epochs = parameters.get("epochs") device = select_device(opt.device, batch_size=opt.batch_size) train(hyp_dict, opt, device, callbacks=Callbacks()) if __name__ == "__main__": opt = get_args(known=True) opt.weights = str(opt.weights) opt.cfg = str(opt.cfg) opt.data = str(opt.data) opt.project = str(opt.project) optimizer_id = os.getenv("COMET_OPTIMIZER_ID") if optimizer_id is None: with open(opt.comet_optimizer_config) as f: optimizer_config = json.load(f) optimizer = comet_ml.Optimizer(optimizer_config) else: optimizer = comet_ml.Optimizer(optimizer_id) opt.comet_optimizer_id = optimizer.id status = optimizer.status() opt.comet_optimizer_objective = status["spec"]["objective"] opt.comet_optimizer_metric = status["spec"]["metric"] logger.info("COMET INFO: Starting Hyperparameter Sweep") for parameter in optimizer.get_parameters(): run(parameter["parameters"], opt)
--- +++ @@ -27,6 +27,9 @@ def get_args(known=False): + """Parses command-line arguments for YOLOv5 training, supporting configuration of weights, data paths, + hyperparameters, and more. + """ parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s.pt", help="initial weights path") parser.add_argument("--cfg", type=str, default="", help="model.yaml path") @@ -85,6 +88,7 @@ def run(parameters, opt): + """Executes YOLOv5 training with given hyperparameters and options, setting up device and training directories.""" hyp_dict = {k: v for k, v in parameters.items() if k not in ["epochs", "batch_size"]} opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok or opt.evolve)) @@ -119,4 +123,4 @@ logger.info("COMET INFO: Starting Hyperparameter Sweep") for parameter in optimizer.get_parameters(): - run(parameter["parameters"], opt)+ run(parameter["parameters"], opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/loggers/comet/hpo.py
Add detailed documentation for each class
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import sys from copy import deepcopy from pathlib import Path FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH # ROOT = ROOT.relative_to(Path.cwd()) # relative import numpy as np import tensorflow as tf import torch import torch.nn as nn from tensorflow import keras from models.common import ( C3, SPP, SPPF, Bottleneck, BottleneckCSP, C3x, Concat, Conv, CrossConv, DWConv, DWConvTranspose2d, Focus, autopad, ) from models.experimental import MixConv2d, attempt_load from models.yolo import Detect, Segment from utils.activations import SiLU from utils.general import LOGGER, make_divisible, print_args class TFBN(keras.layers.Layer): def __init__(self, w=None): super().__init__() self.bn = keras.layers.BatchNormalization( beta_initializer=keras.initializers.Constant(w.bias.numpy()), gamma_initializer=keras.initializers.Constant(w.weight.numpy()), moving_mean_initializer=keras.initializers.Constant(w.running_mean.numpy()), moving_variance_initializer=keras.initializers.Constant(w.running_var.numpy()), epsilon=w.eps, ) def call(self, inputs): return self.bn(inputs) class TFPad(keras.layers.Layer): def __init__(self, pad): super().__init__() if isinstance(pad, int): self.pad = tf.constant([[0, 0], [pad, pad], [pad, pad], [0, 0]]) else: # tuple/list self.pad = tf.constant([[0, 0], [pad[0], pad[0]], [pad[1], pad[1]], [0, 0]]) def call(self, inputs): return tf.pad(inputs, self.pad, mode="constant", constant_values=0) class TFConv(keras.layers.Layer): def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None): super().__init__() assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument" # TensorFlow convolution padding is inconsistent with PyTorch (e.g. k=3 s=2 'SAME' padding) # see https://stackoverflow.com/questions/52975843/comparing-conv2d-with-padding-between-tensorflow-and-pytorch conv = keras.layers.Conv2D( filters=c2, kernel_size=k, strides=s, padding="SAME" if s == 1 else "VALID", use_bias=not hasattr(w, "bn"), kernel_initializer=keras.initializers.Constant(w.conv.weight.permute(2, 3, 1, 0).numpy()), bias_initializer="zeros" if hasattr(w, "bn") else keras.initializers.Constant(w.conv.bias.numpy()), ) self.conv = conv if s == 1 else keras.Sequential([TFPad(autopad(k, p)), conv]) self.bn = TFBN(w.bn) if hasattr(w, "bn") else tf.identity self.act = activations(w.act) if act else tf.identity def call(self, inputs): return self.act(self.bn(self.conv(inputs))) class TFDWConv(keras.layers.Layer): def __init__(self, c1, c2, k=1, s=1, p=None, act=True, w=None): super().__init__() assert c2 % c1 == 0, f"TFDWConv() output={c2} must be a multiple of input={c1} channels" conv = keras.layers.DepthwiseConv2D( kernel_size=k, depth_multiplier=c2 // c1, strides=s, padding="SAME" if s == 1 else "VALID", use_bias=not hasattr(w, "bn"), depthwise_initializer=keras.initializers.Constant(w.conv.weight.permute(2, 3, 1, 0).numpy()), bias_initializer="zeros" if hasattr(w, "bn") else keras.initializers.Constant(w.conv.bias.numpy()), ) self.conv = conv if s == 1 else keras.Sequential([TFPad(autopad(k, p)), conv]) self.bn = TFBN(w.bn) if hasattr(w, "bn") else tf.identity self.act = activations(w.act) if act else tf.identity def call(self, inputs): return self.act(self.bn(self.conv(inputs))) class TFDWConvTranspose2d(keras.layers.Layer): def __init__(self, c1, c2, k=1, s=1, p1=0, p2=0, w=None): super().__init__() assert c1 == c2, f"TFDWConv() output={c2} must be equal to input={c1} channels" assert k == 4 and p1 == 1, "TFDWConv() only valid for k=4 and p1=1" weight, bias = w.weight.permute(2, 3, 1, 0).numpy(), w.bias.numpy() self.c1 = c1 self.conv = [ keras.layers.Conv2DTranspose( filters=1, kernel_size=k, strides=s, padding="VALID", output_padding=p2, use_bias=True, kernel_initializer=keras.initializers.Constant(weight[..., i : i + 1]), bias_initializer=keras.initializers.Constant(bias[i]), ) for i in range(c1) ] def call(self, inputs): return tf.concat([m(x) for m, x in zip(self.conv, tf.split(inputs, self.c1, 3))], 3)[:, 1:-1, 1:-1] class TFFocus(keras.layers.Layer): def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None): super().__init__() self.conv = TFConv(c1 * 4, c2, k, s, p, g, act, w.conv) def call(self, inputs): inputs = [inputs[:, ::2, ::2, :], inputs[:, 1::2, ::2, :], inputs[:, ::2, 1::2, :], inputs[:, 1::2, 1::2, :]] return self.conv(tf.concat(inputs, 3)) class TFBottleneck(keras.layers.Layer): def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None): super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv(c_, c2, 3, 1, g=g, w=w.cv2) self.add = shortcut and c1 == c2 def call(self, inputs): return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs)) class TFCrossConv(keras.layers.Layer): def __init__(self, c1, c2, k=3, s=1, g=1, e=1.0, shortcut=False, w=None): super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, (1, k), (1, s), w=w.cv1) self.cv2 = TFConv(c_, c2, (k, 1), (s, 1), g=g, w=w.cv2) self.add = shortcut and c1 == c2 def call(self, inputs): return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs)) class TFConv2d(keras.layers.Layer): def __init__(self, c1, c2, k, s=1, g=1, bias=True, w=None): super().__init__() assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument" self.conv = keras.layers.Conv2D( filters=c2, kernel_size=k, strides=s, padding="VALID", use_bias=bias, kernel_initializer=keras.initializers.Constant(w.weight.permute(2, 3, 1, 0).numpy()), bias_initializer=keras.initializers.Constant(w.bias.numpy()) if bias else None, ) def call(self, inputs): return self.conv(inputs) class TFBottleneckCSP(keras.layers.Layer): def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv2d(c1, c_, 1, 1, bias=False, w=w.cv2) self.cv3 = TFConv2d(c_, c_, 1, 1, bias=False, w=w.cv3) self.cv4 = TFConv(2 * c_, c2, 1, 1, w=w.cv4) self.bn = TFBN(w.bn) self.act = lambda x: keras.activations.swish(x) self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)]) def call(self, inputs): y1 = self.cv3(self.m(self.cv1(inputs))) y2 = self.cv2(inputs) return self.cv4(self.act(self.bn(tf.concat((y1, y2), axis=3)))) class TFC3(keras.layers.Layer): def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv(c1, c_, 1, 1, w=w.cv2) self.cv3 = TFConv(2 * c_, c2, 1, 1, w=w.cv3) self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)]) def call(self, inputs): return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3)) class TFC3x(keras.layers.Layer): def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv(c1, c_, 1, 1, w=w.cv2) self.cv3 = TFConv(2 * c_, c2, 1, 1, w=w.cv3) self.m = keras.Sequential( [TFCrossConv(c_, c_, k=3, s=1, g=g, e=1.0, shortcut=shortcut, w=w.m[j]) for j in range(n)] ) def call(self, inputs): return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3)) class TFSPP(keras.layers.Layer): def __init__(self, c1, c2, k=(5, 9, 13), w=None): super().__init__() c_ = c1 // 2 # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv(c_ * (len(k) + 1), c2, 1, 1, w=w.cv2) self.m = [keras.layers.MaxPool2D(pool_size=x, strides=1, padding="SAME") for x in k] def call(self, inputs): x = self.cv1(inputs) return self.cv2(tf.concat([x] + [m(x) for m in self.m], 3)) class TFSPPF(keras.layers.Layer): def __init__(self, c1, c2, k=5, w=None): super().__init__() c_ = c1 // 2 # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) self.cv2 = TFConv(c_ * 4, c2, 1, 1, w=w.cv2) self.m = keras.layers.MaxPool2D(pool_size=k, strides=1, padding="SAME") def call(self, inputs): x = self.cv1(inputs) y1 = self.m(x) y2 = self.m(y1) return self.cv2(tf.concat([x, y1, y2, self.m(y2)], 3)) class TFDetect(keras.layers.Layer): def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None): super().__init__() self.stride = tf.convert_to_tensor(w.stride.numpy(), dtype=tf.float32) self.nc = nc # number of classes self.no = nc + 5 # number of outputs per anchor self.nl = len(anchors) # number of detection layers self.na = len(anchors[0]) // 2 # number of anchors self.grid = [tf.zeros(1)] * self.nl # init grid self.anchors = tf.convert_to_tensor(w.anchors.numpy(), dtype=tf.float32) self.anchor_grid = tf.reshape(self.anchors * tf.reshape(self.stride, [self.nl, 1, 1]), [self.nl, 1, -1, 1, 2]) self.m = [TFConv2d(x, self.no * self.na, 1, w=w.m[i]) for i, x in enumerate(ch)] self.training = False # set to False after building model self.imgsz = imgsz for i in range(self.nl): ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i] self.grid[i] = self._make_grid(nx, ny) def call(self, inputs): z = [] # inference output x = [] for i in range(self.nl): x.append(self.m[i](inputs[i])) # x(bs,20,20,255) to x(bs,3,20,20,85) ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i] x[i] = tf.reshape(x[i], [-1, ny * nx, self.na, self.no]) if not self.training: # inference y = x[i] grid = tf.transpose(self.grid[i], [0, 2, 1, 3]) - 0.5 anchor_grid = tf.transpose(self.anchor_grid[i], [0, 2, 1, 3]) * 4 xy = (tf.sigmoid(y[..., 0:2]) * 2 + grid) * self.stride[i] # xy wh = tf.sigmoid(y[..., 2:4]) ** 2 * anchor_grid # Normalize xywh to 0-1 to reduce calibration error xy /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32) wh /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32) y = tf.concat([xy, wh, tf.sigmoid(y[..., 4 : 5 + self.nc]), y[..., 5 + self.nc :]], -1) z.append(tf.reshape(y, [-1, self.na * ny * nx, self.no])) return tf.transpose(x, [0, 2, 1, 3]) if self.training else (tf.concat(z, 1),) @staticmethod def _make_grid(nx=20, ny=20): # return torch.stack((xv, yv), 2).view((1, 1, ny, nx, 2)).float() xv, yv = tf.meshgrid(tf.range(nx), tf.range(ny)) return tf.cast(tf.reshape(tf.stack([xv, yv], 2), [1, 1, ny * nx, 2]), dtype=tf.float32) class TFSegment(TFDetect): def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), imgsz=(640, 640), w=None): super().__init__(nc, anchors, ch, imgsz, w) self.nm = nm # number of masks self.npr = npr # number of protos self.no = 5 + nc + self.nm # number of outputs per anchor self.m = [TFConv2d(x, self.no * self.na, 1, w=w.m[i]) for i, x in enumerate(ch)] # output conv self.proto = TFProto(ch[0], self.npr, self.nm, w=w.proto) # protos self.detect = TFDetect.call def call(self, x): p = self.proto(x[0]) # p = TFUpsample(None, scale_factor=4, mode='nearest')(self.proto(x[0])) # (optional) full-size protos p = tf.transpose(p, [0, 3, 1, 2]) # from shape(1,160,160,32) to shape(1,32,160,160) x = self.detect(self, x) return (x, p) if self.training else (x[0], p) class TFProto(keras.layers.Layer): def __init__(self, c1, c_=256, c2=32, w=None): super().__init__() self.cv1 = TFConv(c1, c_, k=3, w=w.cv1) self.upsample = TFUpsample(None, scale_factor=2, mode="nearest") self.cv2 = TFConv(c_, c_, k=3, w=w.cv2) self.cv3 = TFConv(c_, c2, w=w.cv3) def call(self, inputs): return self.cv3(self.cv2(self.upsample(self.cv1(inputs)))) class TFUpsample(keras.layers.Layer): def __init__(self, size, scale_factor, mode, w=None): super().__init__() assert scale_factor % 2 == 0, "scale_factor must be multiple of 2" self.upsample = lambda x: tf.image.resize(x, (x.shape[1] * scale_factor, x.shape[2] * scale_factor), mode) # self.upsample = keras.layers.UpSampling2D(size=scale_factor, interpolation=mode) # with default arguments: align_corners=False, half_pixel_centers=False # self.upsample = lambda x: tf.raw_ops.ResizeNearestNeighbor(images=x, # size=(x.shape[1] * 2, x.shape[2] * 2)) def call(self, inputs): return self.upsample(inputs) class TFConcat(keras.layers.Layer): def __init__(self, dimension=1, w=None): super().__init__() assert dimension == 1, "convert only NCHW to NHWC concat" self.d = 3 def call(self, inputs): return tf.concat(inputs, self.d) def parse_model(d, ch, model, imgsz): LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}") anchors, nc, gd, gw, ch_mul = ( d["anchors"], d["nc"], d["depth_multiple"], d["width_multiple"], d.get("channel_multiple"), ) na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors no = na * (nc + 5) # number of outputs = anchors * (classes + 5) if not ch_mul: ch_mul = 8 layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out for i, (f, n, m, args) in enumerate(d["backbone"] + d["head"]): # from, number, module, args m_str = m m = eval(m) if isinstance(m, str) else m # eval strings for j, a in enumerate(args): try: args[j] = eval(a) if isinstance(a, str) else a # eval strings except NameError: pass n = max(round(n * gd), 1) if n > 1 else n # depth gain if m in [ nn.Conv2d, Conv, DWConv, DWConvTranspose2d, Bottleneck, SPP, SPPF, MixConv2d, Focus, CrossConv, BottleneckCSP, C3, C3x, ]: c1, c2 = ch[f], args[0] c2 = make_divisible(c2 * gw, ch_mul) if c2 != no else c2 args = [c1, c2, *args[1:]] if m in [BottleneckCSP, C3, C3x]: args.insert(2, n) n = 1 elif m is nn.BatchNorm2d: args = [ch[f]] elif m is Concat: c2 = sum(ch[-1 if x == -1 else x + 1] for x in f) elif m in [Detect, Segment]: args.append([ch[x + 1] for x in f]) if isinstance(args[1], int): # number of anchors args[1] = [list(range(args[1] * 2))] * len(f) if m is Segment: args[3] = make_divisible(args[3] * gw, ch_mul) args.append(imgsz) else: c2 = ch[f] tf_m = eval("TF" + m_str.replace("nn.", "")) m_ = ( keras.Sequential([tf_m(*args, w=model.model[i][j]) for j in range(n)]) if n > 1 else tf_m(*args, w=model.model[i]) ) # module torch_m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module t = str(m)[8:-2].replace("__main__.", "") # module type np = sum(x.numel() for x in torch_m_.parameters()) # number params m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params LOGGER.info(f"{i:>3}{f!s:>18}{n!s:>3}{np:>10} {t:<40}{args!s:<30}") # print save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist layers.append(m_) ch.append(c2) return keras.Sequential(layers), sorted(save) class TFModel: def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, model=None, imgsz=(640, 640)): super().__init__() if isinstance(cfg, dict): self.yaml = cfg # model dict else: # is *.yaml import yaml # for torch hub self.yaml_file = Path(cfg).name with open(cfg) as f: self.yaml = yaml.load(f, Loader=yaml.FullLoader) # model dict # Define model if nc and nc != self.yaml["nc"]: LOGGER.info(f"Overriding {cfg} nc={self.yaml['nc']} with nc={nc}") self.yaml["nc"] = nc # override yaml value self.model, self.savelist = parse_model(deepcopy(self.yaml), ch=[ch], model=model, imgsz=imgsz) def predict( self, inputs, tf_nms=False, agnostic_nms=False, topk_per_class=100, topk_all=100, iou_thres=0.45, conf_thres=0.25, ): y = [] # outputs x = inputs for m in self.model.layers: if m.f != -1: # if not from previous layer x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers x = m(x) # run y.append(x if m.i in self.savelist else None) # save output # Add TensorFlow NMS if tf_nms: boxes = self._xywh2xyxy(x[0][..., :4]) probs = x[0][:, :, 4:5] classes = x[0][:, :, 5:] scores = probs * classes if agnostic_nms: nms = AgnosticNMS()((boxes, classes, scores), topk_all, iou_thres, conf_thres) else: boxes = tf.expand_dims(boxes, 2) nms = tf.image.combined_non_max_suppression( boxes, scores, topk_per_class, topk_all, iou_thres, conf_thres, clip_boxes=False ) return (nms,) return x # output [1,6300,85] = [xywh, conf, class0, class1, ...] # x = x[0] # [x(1,6300,85), ...] to x(6300,85) # xywh = x[..., :4] # x(6300,4) boxes # conf = x[..., 4:5] # x(6300,1) confidences # cls = tf.reshape(tf.cast(tf.argmax(x[..., 5:], axis=1), tf.float32), (-1, 1)) # x(6300,1) classes # return tf.concat([conf, cls, xywh], 1) @staticmethod def _xywh2xyxy(xywh): x, y, w, h = tf.split(xywh, num_or_size_splits=4, axis=-1) return tf.concat([x - w / 2, y - h / 2, x + w / 2, y + h / 2], axis=-1) class AgnosticNMS(keras.layers.Layer): def call(self, input, topk_all, iou_thres, conf_thres): return tf.map_fn( lambda x: self._nms(x, topk_all, iou_thres, conf_thres), input, fn_output_signature=(tf.float32, tf.float32, tf.float32, tf.int32), name="agnostic_nms", ) @staticmethod def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25): boxes, classes, scores = x class_inds = tf.cast(tf.argmax(classes, axis=-1), tf.float32) scores_inp = tf.reduce_max(scores, -1) selected_inds = tf.image.non_max_suppression( boxes, scores_inp, max_output_size=topk_all, iou_threshold=iou_thres, score_threshold=conf_thres ) selected_boxes = tf.gather(boxes, selected_inds) padded_boxes = tf.pad( selected_boxes, paddings=[[0, topk_all - tf.shape(selected_boxes)[0]], [0, 0]], mode="CONSTANT", constant_values=0.0, ) selected_scores = tf.gather(scores_inp, selected_inds) padded_scores = tf.pad( selected_scores, paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]], mode="CONSTANT", constant_values=-1.0, ) selected_classes = tf.gather(class_inds, selected_inds) padded_classes = tf.pad( selected_classes, paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]], mode="CONSTANT", constant_values=-1.0, ) valid_detections = tf.shape(selected_inds)[0] return padded_boxes, padded_scores, padded_classes, valid_detections def activations(act=nn.SiLU): if isinstance(act, nn.LeakyReLU): return lambda x: keras.activations.relu(x, alpha=0.1) elif isinstance(act, nn.Hardswish): return lambda x: x * tf.nn.relu6(x + 3) * 0.166666667 elif isinstance(act, (nn.SiLU, SiLU)): return lambda x: keras.activations.swish(x) else: raise Exception(f"no matching TensorFlow activation found for PyTorch activation {act}") def representative_dataset_gen(dataset, ncalib=100): for n, (path, img, im0s, vid_cap, string) in enumerate(dataset): im = np.transpose(img, [1, 2, 0]) im = np.expand_dims(im, axis=0).astype(np.float32) im /= 255 yield [im] if n >= ncalib: break def run( weights=ROOT / "yolov5s.pt", # weights path imgsz=(640, 640), # inference size h,w batch_size=1, # batch size dynamic=False, # dynamic batch size ): # PyTorch model im = torch.zeros((batch_size, 3, *imgsz)) # BCHW image model = attempt_load(weights, device=torch.device("cpu"), inplace=True, fuse=False) _ = model(im) # inference model.info() # TensorFlow model im = tf.zeros((batch_size, *imgsz, 3)) # BHWC image tf_model = TFModel(cfg=model.yaml, model=model, nc=model.nc, imgsz=imgsz) _ = tf_model.predict(im) # inference # Keras model im = keras.Input(shape=(*imgsz, 3), batch_size=None if dynamic else batch_size) keras_model = keras.Model(inputs=im, outputs=tf_model.predict(im)) keras_model.summary() LOGGER.info("PyTorch, TensorFlow and Keras models successfully verified.\nUse export.py for TF model export.") def parse_opt(): parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s.pt", help="weights path") parser.add_argument("--imgsz", "--img", "--img-size", nargs="+", type=int, default=[640], help="inference size h,w") parser.add_argument("--batch-size", type=int, default=1, help="batch size") parser.add_argument("--dynamic", action="store_true", help="dynamic batch size") opt = parser.parse_args() opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1 # expand print_args(vars(opt)) return opt def main(opt): run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() main(opt)
--- +++ @@ -1,4 +1,14 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +TensorFlow, Keras and TFLite versions of YOLOv5 +Authored by https://github.com/zldrobit in PR https://github.com/ultralytics/yolov5/pull/1127. + +Usage: + $ python models/tf.py --weights yolov5s.pt + +Export: + $ python export.py --weights yolov5s.pt --include saved_model pb tflite tfjs +""" import argparse import sys @@ -39,8 +49,10 @@ class TFBN(keras.layers.Layer): + """TensorFlow BatchNormalization wrapper for initializing with optional pretrained weights.""" def __init__(self, w=None): + """Initializes a TensorFlow BatchNormalization layer with optional pretrained weights.""" super().__init__() self.bn = keras.layers.BatchNormalization( beta_initializer=keras.initializers.Constant(w.bias.numpy()), @@ -51,12 +63,19 @@ ) def call(self, inputs): + """Applies batch normalization to the inputs.""" return self.bn(inputs) class TFPad(keras.layers.Layer): + """Pads input tensors in spatial dimensions 1 and 2 with specified integer or tuple padding values.""" def __init__(self, pad): + """Initializes a padding layer for spatial dimensions 1 and 2 with specified padding, supporting both int and + tuple inputs. + + Inputs are + """ super().__init__() if isinstance(pad, int): self.pad = tf.constant([[0, 0], [pad, pad], [pad, pad], [0, 0]]) @@ -64,12 +83,19 @@ self.pad = tf.constant([[0, 0], [pad[0], pad[0]], [pad[1], pad[1]], [0, 0]]) def call(self, inputs): + """Pads input tensor with zeros using specified padding, suitable for int and tuple pad dimensions.""" return tf.pad(inputs, self.pad, mode="constant", constant_values=0) class TFConv(keras.layers.Layer): + """Implements a standard convolutional layer with optional batch normalization and activation for TensorFlow.""" def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None): + """Initializes a standard convolution layer with optional batch normalization and activation; supports only + group=1. + + Inputs are ch_in, ch_out, weights, kernel, stride, padding, groups. + """ super().__init__() assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument" # TensorFlow convolution padding is inconsistent with PyTorch (e.g. k=3 s=2 'SAME' padding) @@ -88,12 +114,19 @@ self.act = activations(w.act) if act else tf.identity def call(self, inputs): + """Applies convolution, batch normalization, and activation function to input tensors.""" return self.act(self.bn(self.conv(inputs))) class TFDWConv(keras.layers.Layer): + """Initializes a depthwise convolution layer with optional batch normalization and activation for TensorFlow.""" def __init__(self, c1, c2, k=1, s=1, p=None, act=True, w=None): + """Initializes a depthwise convolution layer with optional batch normalization and activation for TensorFlow + models. + + Input are ch_in, ch_out, weights, kernel, stride, padding, groups. + """ super().__init__() assert c2 % c1 == 0, f"TFDWConv() output={c2} must be a multiple of input={c1} channels" conv = keras.layers.DepthwiseConv2D( @@ -110,12 +143,18 @@ self.act = activations(w.act) if act else tf.identity def call(self, inputs): + """Applies convolution, batch normalization, and activation function to input tensors.""" return self.act(self.bn(self.conv(inputs))) class TFDWConvTranspose2d(keras.layers.Layer): + """Implements a depthwise ConvTranspose2D layer for TensorFlow with specific settings.""" def __init__(self, c1, c2, k=1, s=1, p1=0, p2=0, w=None): + """Initializes depthwise ConvTranspose2D layer with specific channel, kernel, stride, and padding settings. + + Inputs are ch_in, ch_out, weights, kernel, stride, padding, groups. + """ super().__init__() assert c1 == c2, f"TFDWConv() output={c2} must be equal to input={c1} channels" assert k == 4 and p1 == 1, "TFDWConv() only valid for k=4 and p1=1" @@ -136,23 +175,40 @@ ] def call(self, inputs): + """Processes input through parallel convolutions and concatenates results, trimming border pixels.""" return tf.concat([m(x) for m, x in zip(self.conv, tf.split(inputs, self.c1, 3))], 3)[:, 1:-1, 1:-1] class TFFocus(keras.layers.Layer): + """Focuses spatial information into channel space using pixel shuffling and convolution for TensorFlow models.""" def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None): + """Initializes TFFocus layer to focus width and height information into channel space with custom convolution + parameters. + + Inputs are ch_in, ch_out, kernel, stride, padding, groups. + """ super().__init__() self.conv = TFConv(c1 * 4, c2, k, s, p, g, act, w.conv) def call(self, inputs): + """Performs pixel shuffling and convolution on input tensor, downsampling by 2 and expanding channels by 4. + + Example x(b,w,h,c) -> y(b,w/2,h/2,4c). + """ inputs = [inputs[:, ::2, ::2, :], inputs[:, 1::2, ::2, :], inputs[:, ::2, 1::2, :], inputs[:, 1::2, 1::2, :]] return self.conv(tf.concat(inputs, 3)) class TFBottleneck(keras.layers.Layer): + """Implements a TensorFlow bottleneck layer with optional shortcut connections for efficient feature extraction.""" def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None): + """Initializes a standard bottleneck layer for TensorFlow models, expanding and contracting channels with + optional shortcut. + + Arguments are ch_in, ch_out, shortcut, groups, expansion. + """ super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -160,12 +216,17 @@ self.add = shortcut and c1 == c2 def call(self, inputs): + """Performs forward pass; if shortcut is True & input/output channels match, adds input to the convolution + result. + """ return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs)) class TFCrossConv(keras.layers.Layer): + """Implements a cross convolutional layer with optional expansion, grouping, and shortcut for TensorFlow.""" def __init__(self, c1, c2, k=3, s=1, g=1, e=1.0, shortcut=False, w=None): + """Initializes cross convolution layer with optional expansion, grouping, and shortcut addition capabilities.""" super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, (1, k), (1, s), w=w.cv1) @@ -173,12 +234,17 @@ self.add = shortcut and c1 == c2 def call(self, inputs): + """Passes input through two convolutions optionally adding the input if channel dimensions match.""" return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs)) class TFConv2d(keras.layers.Layer): + """Implements a TensorFlow 2D convolution layer, mimicking PyTorch's nn.Conv2D for specified filters and stride.""" def __init__(self, c1, c2, k, s=1, g=1, bias=True, w=None): + """Initializes a TensorFlow 2D convolution layer, mimicking PyTorch's nn.Conv2D functionality for given filter + sizes and stride. + """ super().__init__() assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument" self.conv = keras.layers.Conv2D( @@ -192,12 +258,19 @@ ) def call(self, inputs): + """Applies a convolution operation to the inputs and returns the result.""" return self.conv(inputs) class TFBottleneckCSP(keras.layers.Layer): + """Implements a CSP bottleneck layer for TensorFlow models to enhance gradient flow and efficiency.""" def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): + """Initializes CSP bottleneck layer with specified channel sizes, count, shortcut option, groups, and expansion + ratio. + + Inputs are ch_in, ch_out, number, shortcut, groups, expansion. + """ super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -209,14 +282,22 @@ self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)]) def call(self, inputs): + """Processes input through the model layers, concatenates, normalizes, activates, and reduces the output + dimensions. + """ y1 = self.cv3(self.m(self.cv1(inputs))) y2 = self.cv2(inputs) return self.cv4(self.act(self.bn(tf.concat((y1, y2), axis=3)))) class TFC3(keras.layers.Layer): + """CSP bottleneck layer with 3 convolutions for TensorFlow, supporting optional shortcuts and group convolutions.""" def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): + """Initializes CSP Bottleneck with 3 convolutions, supporting optional shortcuts and group convolutions. + + Inputs are ch_in, ch_out, number, shortcut, groups, expansion. + """ super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -225,12 +306,21 @@ self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)]) def call(self, inputs): + """Processes input through a sequence of transformations for object detection (YOLOv5). + + See https://github.com/ultralytics/yolov5. + """ return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3)) class TFC3x(keras.layers.Layer): + """A TensorFlow layer for enhanced feature extraction using cross-convolutions in object detection models.""" def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None): + """Initializes layer with cross-convolutions for enhanced feature extraction in object detection models. + + Inputs are ch_in, ch_out, number, shortcut, groups, expansion. + """ super().__init__() c_ = int(c2 * e) # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -241,12 +331,15 @@ ) def call(self, inputs): + """Processes input through cascaded convolutions and merges features, returning the final tensor output.""" return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3)) class TFSPP(keras.layers.Layer): + """Implements spatial pyramid pooling for YOLOv3-SPP with specific channels and kernel sizes.""" def __init__(self, c1, c2, k=(5, 9, 13), w=None): + """Initializes a YOLOv3-SPP layer with specific input/output channels and kernel sizes for pooling.""" super().__init__() c_ = c1 // 2 # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -254,13 +347,16 @@ self.m = [keras.layers.MaxPool2D(pool_size=x, strides=1, padding="SAME") for x in k] def call(self, inputs): + """Processes input through two TFConv layers and concatenates with max-pooled outputs at intermediate stage.""" x = self.cv1(inputs) return self.cv2(tf.concat([x] + [m(x) for m in self.m], 3)) class TFSPPF(keras.layers.Layer): + """Implements a fast spatial pyramid pooling layer for TensorFlow with optimized feature extraction.""" def __init__(self, c1, c2, k=5, w=None): + """Initialize a fast spatial pyramid pooling layer with customizable channels, kernel size, and weights.""" super().__init__() c_ = c1 // 2 # hidden channels self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1) @@ -268,6 +364,9 @@ self.m = keras.layers.MaxPool2D(pool_size=k, strides=1, padding="SAME") def call(self, inputs): + """Executes the model's forward pass, concatenating input features with three max-pooled versions before final + convolution. + """ x = self.cv1(inputs) y1 = self.m(x) y2 = self.m(y1) @@ -275,8 +374,12 @@ class TFDetect(keras.layers.Layer): + """Implements YOLOv5 object detection layer in TensorFlow for predicting bounding boxes and class probabilities.""" def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None): + """Initializes YOLOv5 detection layer for TensorFlow with configurable classes, anchors, channels, and image + size. + """ super().__init__() self.stride = tf.convert_to_tensor(w.stride.numpy(), dtype=tf.float32) self.nc = nc # number of classes @@ -294,6 +397,7 @@ self.grid[i] = self._make_grid(nx, ny) def call(self, inputs): + """Performs forward pass through the model layers to predict object bounding boxes and classifications.""" z = [] # inference output x = [] for i in range(self.nl): @@ -318,14 +422,19 @@ @staticmethod def _make_grid(nx=20, ny=20): + """Generates a 2D grid of coordinates in (x, y) format with shape [1, 1, ny*nx, 2].""" # return torch.stack((xv, yv), 2).view((1, 1, ny, nx, 2)).float() xv, yv = tf.meshgrid(tf.range(nx), tf.range(ny)) return tf.cast(tf.reshape(tf.stack([xv, yv], 2), [1, 1, ny * nx, 2]), dtype=tf.float32) class TFSegment(TFDetect): + """YOLOv5 segmentation head for TensorFlow, combining detection and segmentation.""" def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), imgsz=(640, 640), w=None): + """Initializes YOLOv5 Segment head with specified channel depths, anchors, and input size for segmentation + models. + """ super().__init__(nc, anchors, ch, imgsz, w) self.nm = nm # number of masks self.npr = npr # number of protos @@ -335,6 +444,7 @@ self.detect = TFDetect.call def call(self, x): + """Applies detection and proto layers on input, returning detections and optionally protos if training.""" p = self.proto(x[0]) # p = TFUpsample(None, scale_factor=4, mode='nearest')(self.proto(x[0])) # (optional) full-size protos p = tf.transpose(p, [0, 3, 1, 2]) # from shape(1,160,160,32) to shape(1,32,160,160) @@ -343,8 +453,10 @@ class TFProto(keras.layers.Layer): + """Implements convolutional and upsampling layers for feature extraction in YOLOv5 segmentation.""" def __init__(self, c1, c_=256, c2=32, w=None): + """Initialize TFProto layer with convolutional and upsampling for feature extraction and transformation.""" super().__init__() self.cv1 = TFConv(c1, c_, k=3, w=w.cv1) self.upsample = TFUpsample(None, scale_factor=2, mode="nearest") @@ -352,12 +464,19 @@ self.cv3 = TFConv(c_, c2, w=w.cv3) def call(self, inputs): + """Performs forward pass through the model, applying convolutions and upscaling on input tensor.""" return self.cv3(self.cv2(self.upsample(self.cv1(inputs)))) class TFUpsample(keras.layers.Layer): + """Implements a TensorFlow upsampling layer with specified size, scale factor, and interpolation mode.""" def __init__(self, size, scale_factor, mode, w=None): + """Initializes a TensorFlow upsampling layer with specified size, scale_factor, and mode, ensuring scale_factor + is even. + + Warning: all arguments needed including 'w' + """ super().__init__() assert scale_factor % 2 == 0, "scale_factor must be multiple of 2" self.upsample = lambda x: tf.image.resize(x, (x.shape[1] * scale_factor, x.shape[2] * scale_factor), mode) @@ -367,21 +486,26 @@ # size=(x.shape[1] * 2, x.shape[2] * 2)) def call(self, inputs): + """Applies upsample operation to inputs using nearest neighbor interpolation.""" return self.upsample(inputs) class TFConcat(keras.layers.Layer): + """Implements TensorFlow's version of torch.concat() for concatenating tensors along the last dimension.""" def __init__(self, dimension=1, w=None): + """Initializes a TensorFlow layer for NCHW to NHWC concatenation, requiring dimension=1.""" super().__init__() assert dimension == 1, "convert only NCHW to NHWC concat" self.d = 3 def call(self, inputs): + """Concatenates a list of tensors along the last dimension, used for NCHW to NHWC conversion.""" return tf.concat(inputs, self.d) def parse_model(d, ch, model, imgsz): + """Parses a model definition dict `d` to create YOLOv5 model layers, including dynamic channel adjustments.""" LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}") anchors, nc, gd, gw, ch_mul = ( d["anchors"], @@ -461,8 +585,10 @@ class TFModel: + """Implements YOLOv5 model in TensorFlow, supporting TensorFlow, Keras, and TFLite formats for object detection.""" def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, model=None, imgsz=(640, 640)): + """Initialize TF YOLOv5 model with specified channels, classes, model instance, and input size.""" super().__init__() if isinstance(cfg, dict): self.yaml = cfg # model dict @@ -489,6 +615,7 @@ iou_thres=0.45, conf_thres=0.25, ): + """Runs inference on input data, with an option for TensorFlow NMS.""" y = [] # outputs x = inputs for m in self.model.layers: @@ -521,13 +648,16 @@ @staticmethod def _xywh2xyxy(xywh): + """Convert box format from [x, y, w, h] to [x1, y1, x2, y2], where xy1=top-left and xy2=bottom- right.""" x, y, w, h = tf.split(xywh, num_or_size_splits=4, axis=-1) return tf.concat([x - w / 2, y - h / 2, x + w / 2, y + h / 2], axis=-1) class AgnosticNMS(keras.layers.Layer): + """Performs agnostic non-maximum suppression (NMS) on detected objects using IoU and confidence thresholds.""" def call(self, input, topk_all, iou_thres, conf_thres): + """Performs agnostic NMS on input tensors using given thresholds and top-K selection.""" return tf.map_fn( lambda x: self._nms(x, topk_all, iou_thres, conf_thres), input, @@ -537,6 +667,9 @@ @staticmethod def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25): + """Performs agnostic non-maximum suppression (NMS) on detected objects, filtering based on IoU and confidence + thresholds. + """ boxes, classes, scores = x class_inds = tf.cast(tf.argmax(classes, axis=-1), tf.float32) scores_inp = tf.reduce_max(scores, -1) @@ -569,6 +702,7 @@ def activations(act=nn.SiLU): + """Converts PyTorch activations to TensorFlow equivalents, supporting LeakyReLU, Hardswish, and SiLU/Swish.""" if isinstance(act, nn.LeakyReLU): return lambda x: keras.activations.relu(x, alpha=0.1) elif isinstance(act, nn.Hardswish): @@ -580,6 +714,7 @@ def representative_dataset_gen(dataset, ncalib=100): + """Generate representative dataset for calibration by yielding transformed numpy arrays from the input dataset.""" for n, (path, img, im0s, vid_cap, string) in enumerate(dataset): im = np.transpose(img, [1, 2, 0]) im = np.expand_dims(im, axis=0).astype(np.float32) @@ -596,6 +731,7 @@ dynamic=False, # dynamic batch size ): # PyTorch model + """Exports YOLOv5 model from PyTorch to TensorFlow and Keras formats, performing inference for validation.""" im = torch.zeros((batch_size, 3, *imgsz)) # BCHW image model = attempt_load(weights, device=torch.device("cpu"), inplace=True, fuse=False) _ = model(im) # inference @@ -615,6 +751,9 @@ def parse_opt(): + """Parses and returns command-line options for model inference, including weights path, image size, batch size, and + dynamic batching. + """ parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s.pt", help="weights path") parser.add_argument("--imgsz", "--img", "--img-size", nargs="+", type=int, default=[640], help="inference size h,w") @@ -627,9 +766,10 @@ def main(opt): + """Executes the YOLOv5 model run function with parsed command line options.""" run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() - main(opt)+ main(opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/models/tf.py
Generate docstrings with examples
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import contextlib import math import os import platform import sys from copy import deepcopy from pathlib import Path import torch import torch.nn as nn FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH if platform.system() != "Windows": ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative from models.common import ( C3, C3SPP, C3TR, SPP, SPPF, Bottleneck, BottleneckCSP, C3Ghost, C3x, Classify, Concat, Contract, Conv, CrossConv, DetectMultiBackend, DWConv, DWConvTranspose2d, Expand, Focus, GhostBottleneck, GhostConv, Proto, ) from models.experimental import MixConv2d from utils.autoanchor import check_anchor_order from utils.general import LOGGER, check_version, check_yaml, colorstr, make_divisible, print_args from utils.plots import feature_visualization from utils.torch_utils import ( fuse_conv_and_bn, initialize_weights, model_info, profile, scale_img, select_device, time_sync, ) try: import thop # for FLOPs computation except ImportError: thop = None class Detect(nn.Module): stride = None # strides computed during build dynamic = False # force grid reconstruction export = False # export mode def __init__(self, nc=80, anchors=(), ch=(), inplace=True): super().__init__() self.nc = nc # number of classes self.no = nc + 5 # number of outputs per anchor self.nl = len(anchors) # number of detection layers self.na = len(anchors[0]) // 2 # number of anchors self.grid = [torch.empty(0) for _ in range(self.nl)] # init grid self.anchor_grid = [torch.empty(0) for _ in range(self.nl)] # init anchor grid self.register_buffer("anchors", torch.tensor(anchors).float().view(self.nl, -1, 2)) # shape(nl,na,2) self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv self.inplace = inplace # use inplace ops (e.g. slice assignment) def forward(self, x): z = [] # inference output for i in range(self.nl): x[i] = self.m[i](x[i]) # conv bs, _, ny, nx = x[i].shape # x(bs,255,20,20) to x(bs,3,20,20,85) x[i] = x[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous() if not self.training: # inference if self.dynamic or self.grid[i].shape[2:4] != x[i].shape[2:4]: self.grid[i], self.anchor_grid[i] = self._make_grid(nx, ny, i) if isinstance(self, Segment): # (boxes + masks) xy, wh, conf, mask = x[i].split((2, 2, self.nc + 1, self.no - self.nc - 5), 4) xy = (xy.sigmoid() * 2 + self.grid[i]) * self.stride[i] # xy wh = (wh.sigmoid() * 2) ** 2 * self.anchor_grid[i] # wh y = torch.cat((xy, wh, conf.sigmoid(), mask), 4) else: # Detect (boxes only) xy, wh, conf = x[i].sigmoid().split((2, 2, self.nc + 1), 4) xy = (xy * 2 + self.grid[i]) * self.stride[i] # xy wh = (wh * 2) ** 2 * self.anchor_grid[i] # wh y = torch.cat((xy, wh, conf), 4) z.append(y.view(bs, self.na * nx * ny, self.no)) return x if self.training else (torch.cat(z, 1),) if self.export else (torch.cat(z, 1), x) def _make_grid(self, nx=20, ny=20, i=0, torch_1_10=check_version(torch.__version__, "1.10.0")): d = self.anchors[i].device t = self.anchors[i].dtype shape = 1, self.na, ny, nx, 2 # grid shape y, x = torch.arange(ny, device=d, dtype=t), torch.arange(nx, device=d, dtype=t) yv, xv = torch.meshgrid(y, x, indexing="ij") if torch_1_10 else torch.meshgrid(y, x) # torch>=0.7 compatibility grid = torch.stack((xv, yv), 2).expand(shape) - 0.5 # add grid offset, i.e. y = 2.0 * x - 0.5 anchor_grid = (self.anchors[i] * self.stride[i]).view((1, self.na, 1, 1, 2)).expand(shape) return grid, anchor_grid class Segment(Detect): def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), inplace=True): super().__init__(nc, anchors, ch, inplace) self.nm = nm # number of masks self.npr = npr # number of protos self.no = 5 + nc + self.nm # number of outputs per anchor self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv self.proto = Proto(ch[0], self.npr, self.nm) # protos self.detect = Detect.forward def forward(self, x): p = self.proto(x[0]) x = self.detect(self, x) return (x, p) if self.training else (x[0], p) if self.export else (x[0], p, x[1]) class BaseModel(nn.Module): def forward(self, x, profile=False, visualize=False): return self._forward_once(x, profile, visualize) # single-scale inference, train def _forward_once(self, x, profile=False, visualize=False): y, dt = [], [] # outputs for m in self.model: if m.f != -1: # if not from previous layer x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layers if profile: self._profile_one_layer(m, x, dt) x = m(x) # run y.append(x if m.i in self.save else None) # save output if visualize: feature_visualization(x, m.type, m.i, save_dir=visualize) return x def _profile_one_layer(self, m, x, dt): c = m == self.model[-1] # is final layer, copy input as inplace fix o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1e9 * 2 if thop else 0 # FLOPs t = time_sync() for _ in range(10): m(x.copy() if c else x) dt.append((time_sync() - t) * 100) if m == self.model[0]: LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s} module") LOGGER.info(f"{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f} {m.type}") if c: LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total") def fuse(self): LOGGER.info("Fusing layers... ") for m in self.model.modules(): if isinstance(m, (Conv, DWConv)) and hasattr(m, "bn"): m.conv = fuse_conv_and_bn(m.conv, m.bn) # update conv delattr(m, "bn") # remove batchnorm m.forward = m.forward_fuse # update forward self.info() return self def info(self, verbose=False, img_size=640): model_info(self, verbose, img_size) def _apply(self, fn): self = super()._apply(fn) m = self.model[-1] # Detect() if isinstance(m, (Detect, Segment)): m.stride = fn(m.stride) m.grid = list(map(fn, m.grid)) if isinstance(m.anchor_grid, list): m.anchor_grid = list(map(fn, m.anchor_grid)) return self class DetectionModel(BaseModel): def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, anchors=None): super().__init__() if isinstance(cfg, dict): self.yaml = cfg # model dict else: # is *.yaml import yaml # for torch hub self.yaml_file = Path(cfg).name with open(cfg, encoding="ascii", errors="ignore") as f: self.yaml = yaml.safe_load(f) # model dict # Define model ch = self.yaml["ch"] = self.yaml.get("ch", ch) # input channels if nc and nc != self.yaml["nc"]: LOGGER.info(f"Overriding model.yaml nc={self.yaml['nc']} with nc={nc}") self.yaml["nc"] = nc # override yaml value if anchors: LOGGER.info(f"Overriding model.yaml anchors with anchors={anchors}") self.yaml["anchors"] = round(anchors) # override yaml value self.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch]) # model, savelist self.names = [str(i) for i in range(self.yaml["nc"])] # default names self.inplace = self.yaml.get("inplace", True) # Build strides, anchors m = self.model[-1] # Detect() if isinstance(m, (Detect, Segment)): def _forward(x): return self.forward(x)[0] if isinstance(m, Segment) else self.forward(x) s = 256 # 2x min stride m.inplace = self.inplace m.stride = torch.tensor([s / x.shape[-2] for x in _forward(torch.zeros(1, ch, s, s))]) # forward check_anchor_order(m) m.anchors /= m.stride.view(-1, 1, 1) self.stride = m.stride self._initialize_biases() # only run once # Init weights, biases initialize_weights(self) self.info() LOGGER.info("") def forward(self, x, augment=False, profile=False, visualize=False): if augment: return self._forward_augment(x) # augmented inference, None return self._forward_once(x, profile, visualize) # single-scale inference, train def _forward_augment(self, x): img_size = x.shape[-2:] # height, width s = [1, 0.83, 0.67] # scales f = [None, 3, None] # flips (2-ud, 3-lr) y = [] # outputs for si, fi in zip(s, f): xi = scale_img(x.flip(fi) if fi else x, si, gs=int(self.stride.max())) yi = self._forward_once(xi)[0] # forward # cv2.imwrite(f'img_{si}.jpg', 255 * xi[0].cpu().numpy().transpose((1, 2, 0))[:, :, ::-1]) # save yi = self._descale_pred(yi, fi, si, img_size) y.append(yi) y = self._clip_augmented(y) # clip augmented tails return torch.cat(y, 1), None # augmented inference, train def _descale_pred(self, p, flips, scale, img_size): if self.inplace: p[..., :4] /= scale # de-scale if flips == 2: p[..., 1] = img_size[0] - p[..., 1] # de-flip ud elif flips == 3: p[..., 0] = img_size[1] - p[..., 0] # de-flip lr else: x, y, wh = p[..., 0:1] / scale, p[..., 1:2] / scale, p[..., 2:4] / scale # de-scale if flips == 2: y = img_size[0] - y # de-flip ud elif flips == 3: x = img_size[1] - x # de-flip lr p = torch.cat((x, y, wh, p[..., 4:]), -1) return p def _clip_augmented(self, y): nl = self.model[-1].nl # number of detection layers (P3-P5) g = sum(4**x for x in range(nl)) # grid points e = 1 # exclude layer count i = (y[0].shape[1] // g) * sum(4**x for x in range(e)) # indices y[0] = y[0][:, :-i] # large i = (y[-1].shape[1] // g) * sum(4 ** (nl - 1 - x) for x in range(e)) # indices y[-1] = y[-1][:, i:] # small return y def _initialize_biases(self, cf=None): # cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1. m = self.model[-1] # Detect() module for mi, s in zip(m.m, m.stride): # from b = mi.bias.view(m.na, -1) # conv.bias(255) to (3,85) b.data[:, 4] += math.log(8 / (640 / s) ** 2) # obj (8 objects per 640 image) b.data[:, 5 : 5 + m.nc] += ( math.log(0.6 / (m.nc - 0.99999)) if cf is None else torch.log(cf / cf.sum()) ) # cls mi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True) Model = DetectionModel # retain YOLOv5 'Model' class for backwards compatibility class SegmentationModel(DetectionModel): def __init__(self, cfg="yolov5s-seg.yaml", ch=3, nc=None, anchors=None): super().__init__(cfg, ch, nc, anchors) class ClassificationModel(BaseModel): def __init__(self, cfg=None, model=None, nc=1000, cutoff=10): super().__init__() self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg) def _from_detection_model(self, model, nc=1000, cutoff=10): if isinstance(model, DetectMultiBackend): model = model.model # unwrap DetectMultiBackend model.model = model.model[:cutoff] # backbone m = model.model[-1] # last layer ch = m.conv.in_channels if hasattr(m, "conv") else m.cv1.conv.in_channels # ch into module c = Classify(ch, nc) # Classify() c.i, c.f, c.type = m.i, m.f, "models.common.Classify" # index, from, type model.model[-1] = c # replace self.model = model.model self.stride = model.stride self.save = [] self.nc = nc def _from_yaml(self, cfg): self.model = None def parse_model(d, ch): LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}") anchors, nc, gd, gw, act, ch_mul = ( d["anchors"], d["nc"], d["depth_multiple"], d["width_multiple"], d.get("activation"), d.get("channel_multiple"), ) if act: Conv.default_act = eval(act) # redefine default activation, i.e. Conv.default_act = nn.SiLU() LOGGER.info(f"{colorstr('activation:')} {act}") # print if not ch_mul: ch_mul = 8 na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors no = na * (nc + 5) # number of outputs = anchors * (classes + 5) layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out for i, (f, n, m, args) in enumerate(d["backbone"] + d["head"]): # from, number, module, args m = eval(m) if isinstance(m, str) else m # eval strings for j, a in enumerate(args): with contextlib.suppress(NameError): args[j] = eval(a) if isinstance(a, str) else a # eval strings n = n_ = max(round(n * gd), 1) if n > 1 else n # depth gain if m in { Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv, BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d, DWConvTranspose2d, C3x, }: c1, c2 = ch[f], args[0] if c2 != no: # if not output c2 = make_divisible(c2 * gw, ch_mul) args = [c1, c2, *args[1:]] if m in {BottleneckCSP, C3, C3TR, C3Ghost, C3x}: args.insert(2, n) # number of repeats n = 1 elif m is nn.BatchNorm2d: args = [ch[f]] elif m is Concat: c2 = sum(ch[x] for x in f) # TODO: channel, gw, gd elif m in {Detect, Segment}: args.append([ch[x] for x in f]) if isinstance(args[1], int): # number of anchors args[1] = [list(range(args[1] * 2))] * len(f) if m is Segment: args[3] = make_divisible(args[3] * gw, ch_mul) elif m is Contract: c2 = ch[f] * args[0] ** 2 elif m is Expand: c2 = ch[f] // args[0] ** 2 else: c2 = ch[f] m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args) # module t = str(m)[8:-2].replace("__main__.", "") # module type np = sum(x.numel() for x in m_.parameters()) # number params m_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number params LOGGER.info(f"{i:>3}{f!s:>18}{n_:>3}{np:10.0f} {t:<40}{args!s:<30}") # print save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelist layers.append(m_) if i == 0: ch = [] ch.append(c2) return nn.Sequential(*layers), sorted(save) if __name__ == "__main__": parser = argparse.ArgumentParser() parser.add_argument("--cfg", type=str, default="yolov5s.yaml", help="model.yaml") parser.add_argument("--batch-size", type=int, default=1, help="total batch size for all GPUs") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--profile", action="store_true", help="profile model speed") parser.add_argument("--line-profile", action="store_true", help="profile model speed layer by layer") parser.add_argument("--test", action="store_true", help="test all yolo*.yaml") opt = parser.parse_args() opt.cfg = check_yaml(opt.cfg) # check YAML print_args(vars(opt)) device = select_device(opt.device) # Create model im = torch.rand(opt.batch_size, 3, 640, 640).to(device) model = Model(opt.cfg).to(device) # Options if opt.line_profile: # profile layer by layer model(im, profile=True) elif opt.profile: # profile forward-backward results = profile(input=im, ops=[model], n=3) elif opt.test: # test all models for cfg in Path(ROOT / "models").rglob("yolo*.yaml"): try: _ = Model(cfg) except Exception as e: print(f"Error in {cfg}: {e}") else: # report fused model summary model.fuse()
--- +++ @@ -1,4 +1,10 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +YOLO-specific modules. + +Usage: + $ python models/yolo.py --cfg yolov5s.yaml +""" import argparse import contextlib @@ -64,12 +70,14 @@ class Detect(nn.Module): + """YOLOv5 Detect head for processing input tensors and generating detection outputs in object detection models.""" stride = None # strides computed during build dynamic = False # force grid reconstruction export = False # export mode def __init__(self, nc=80, anchors=(), ch=(), inplace=True): + """Initializes YOLOv5 detection layer with specified classes, anchors, channels, and inplace operations.""" super().__init__() self.nc = nc # number of classes self.no = nc + 5 # number of outputs per anchor @@ -82,6 +90,7 @@ self.inplace = inplace # use inplace ops (e.g. slice assignment) def forward(self, x): + """Processes input through YOLOv5 layers, altering shape for detection: `x(bs, 3, ny, nx, 85)`.""" z = [] # inference output for i in range(self.nl): x[i] = self.m[i](x[i]) # conv @@ -107,6 +116,7 @@ return x if self.training else (torch.cat(z, 1),) if self.export else (torch.cat(z, 1), x) def _make_grid(self, nx=20, ny=20, i=0, torch_1_10=check_version(torch.__version__, "1.10.0")): + """Generates a mesh grid for anchor boxes with optional compatibility for torch versions < 1.10.""" d = self.anchors[i].device t = self.anchors[i].dtype shape = 1, self.na, ny, nx, 2 # grid shape @@ -118,8 +128,10 @@ class Segment(Detect): + """YOLOv5 Segment head for segmentation models, extending Detect with mask and prototype layers.""" def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), inplace=True): + """Initializes YOLOv5 Segment head with options for mask count, protos, and channel adjustments.""" super().__init__(nc, anchors, ch, inplace) self.nm = nm # number of masks self.npr = npr # number of protos @@ -129,17 +141,25 @@ self.detect = Detect.forward def forward(self, x): + """Processes input through the network, returning detections and prototypes; adjusts output based on + training/export mode. + """ p = self.proto(x[0]) x = self.detect(self, x) return (x, p) if self.training else (x[0], p) if self.export else (x[0], p, x[1]) class BaseModel(nn.Module): + """YOLOv5 base model.""" def forward(self, x, profile=False, visualize=False): + """Executes a single-scale inference or training pass on the YOLOv5 base model, with options for profiling and + visualization. + """ return self._forward_once(x, profile, visualize) # single-scale inference, train def _forward_once(self, x, profile=False, visualize=False): + """Performs a forward pass on the YOLOv5 model, enabling profiling and feature visualization options.""" y, dt = [], [] # outputs for m in self.model: if m.f != -1: # if not from previous layer @@ -153,6 +173,7 @@ return x def _profile_one_layer(self, m, x, dt): + """Profiles a single layer's performance by computing GFLOPs, execution time, and parameters.""" c = m == self.model[-1] # is final layer, copy input as inplace fix o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1e9 * 2 if thop else 0 # FLOPs t = time_sync() @@ -166,6 +187,7 @@ LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s} Total") def fuse(self): + """Fuses Conv2d() and BatchNorm2d() layers in the model to improve inference speed.""" LOGGER.info("Fusing layers... ") for m in self.model.modules(): if isinstance(m, (Conv, DWConv)) and hasattr(m, "bn"): @@ -176,9 +198,13 @@ return self def info(self, verbose=False, img_size=640): + """Prints model information given verbosity and image size, e.g., `info(verbose=True, img_size=640)`.""" model_info(self, verbose, img_size) def _apply(self, fn): + """Applies transformations like to(), cpu(), cuda(), half() to model tensors excluding parameters or registered + buffers. + """ self = super()._apply(fn) m = self.model[-1] # Detect() if isinstance(m, (Detect, Segment)): @@ -190,8 +216,10 @@ class DetectionModel(BaseModel): + """YOLOv5 detection model class for object detection tasks, supporting custom configurations and anchors.""" def __init__(self, cfg="yolov5s.yaml", ch=3, nc=None, anchors=None): + """Initializes YOLOv5 model with configuration file, input channels, number of classes, and custom anchors.""" super().__init__() if isinstance(cfg, dict): self.yaml = cfg # model dict @@ -219,6 +247,7 @@ if isinstance(m, (Detect, Segment)): def _forward(x): + """Passes the input 'x' through the model and returns the processed output.""" return self.forward(x)[0] if isinstance(m, Segment) else self.forward(x) s = 256 # 2x min stride @@ -235,11 +264,13 @@ LOGGER.info("") def forward(self, x, augment=False, profile=False, visualize=False): + """Performs single-scale or augmented inference and may include profiling or visualization.""" if augment: return self._forward_augment(x) # augmented inference, None return self._forward_once(x, profile, visualize) # single-scale inference, train def _forward_augment(self, x): + """Performs augmented inference across different scales and flips, returning combined detections.""" img_size = x.shape[-2:] # height, width s = [1, 0.83, 0.67] # scales f = [None, 3, None] # flips (2-ud, 3-lr) @@ -254,6 +285,7 @@ return torch.cat(y, 1), None # augmented inference, train def _descale_pred(self, p, flips, scale, img_size): + """De-scales predictions from augmented inference, adjusting for flips and image size.""" if self.inplace: p[..., :4] /= scale # de-scale if flips == 2: @@ -270,6 +302,9 @@ return p def _clip_augmented(self, y): + """Clips augmented inference tails for YOLOv5 models, affecting first and last tensors based on grid points and + layer counts. + """ nl = self.model[-1].nl # number of detection layers (P3-P5) g = sum(4**x for x in range(nl)) # grid points e = 1 # exclude layer count @@ -280,6 +315,10 @@ return y def _initialize_biases(self, cf=None): + """Initializes biases for YOLOv5's Detect() module, optionally using class frequencies (cf). + + For details see https://arxiv.org/abs/1708.02002 section 3.3. + """ # cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1. m = self.model[-1] # Detect() module for mi, s in zip(m.m, m.stride): # from @@ -295,18 +334,29 @@ class SegmentationModel(DetectionModel): + """YOLOv5 segmentation model for object detection and segmentation tasks with configurable parameters.""" def __init__(self, cfg="yolov5s-seg.yaml", ch=3, nc=None, anchors=None): + """Initializes a YOLOv5 segmentation model with configurable params: cfg (str) for configuration, ch (int) for + channels, nc (int) for num classes, anchors (list). + """ super().__init__(cfg, ch, nc, anchors) class ClassificationModel(BaseModel): + """YOLOv5 classification model for image classification tasks, initialized with a config file or detection model.""" def __init__(self, cfg=None, model=None, nc=1000, cutoff=10): + """Initializes YOLOv5 model with config file `cfg`, input channels `ch`, number of classes `nc`, and `cuttoff` + index. + """ super().__init__() self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg) def _from_detection_model(self, model, nc=1000, cutoff=10): + """Creates a classification model from a YOLOv5 detection model, slicing at `cutoff` and adding a classification + layer. + """ if isinstance(model, DetectMultiBackend): model = model.model # unwrap DetectMultiBackend model.model = model.model[:cutoff] # backbone @@ -321,10 +371,12 @@ self.nc = nc def _from_yaml(self, cfg): + """Creates a YOLOv5 classification model from a specified *.yaml configuration file.""" self.model = None def parse_model(d, ch): + """Parses a YOLOv5 model from a dict `d`, configuring layers based on input channels `ch` and model architecture.""" LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10} {'module':<40}{'arguments':<30}") anchors, nc, gd, gw, act, ch_mul = ( d["anchors"], @@ -441,4 +493,4 @@ print(f"Error in {cfg}: {e}") else: # report fused model summary - model.fuse()+ model.fuse()
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/models/yolo.py
Add docstrings to existing functions
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import math import os import random import subprocess import sys import time from copy import deepcopy from datetime import datetime from pathlib import Path import numpy as np import torch import torch.distributed as dist import torch.nn as nn import yaml from torch.optim import lr_scheduler from tqdm import tqdm FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative from ultralytics.utils.patches import torch_load import segment.val as validate # for end-of-epoch mAP from models.experimental import attempt_load from models.yolo import SegmentationModel from utils.autoanchor import check_anchors from utils.autobatch import check_train_batch_size from utils.callbacks import Callbacks from utils.downloads import attempt_download, is_url from utils.general import ( LOGGER, TQDM_BAR_FORMAT, check_amp, check_dataset, check_file, check_git_info, check_git_status, check_img_size, check_requirements, check_suffix, check_yaml, colorstr, get_latest_run, increment_path, init_seeds, intersect_dicts, labels_to_class_weights, labels_to_image_weights, one_cycle, print_args, print_mutation, strip_optimizer, yaml_save, ) from utils.loggers import GenericLogger from utils.plots import plot_evolve, plot_labels from utils.segment.dataloaders import create_dataloader from utils.segment.loss import ComputeLoss from utils.segment.metrics import KEYS, fitness from utils.segment.plots import plot_images_and_masks, plot_results_with_masks from utils.torch_utils import ( EarlyStopping, ModelEMA, de_parallel, select_device, smart_DDP, smart_optimizer, smart_resume, torch_distributed_zero_first, ) LOCAL_RANK = int(os.getenv("LOCAL_RANK", -1)) # https://pytorch.org/docs/stable/elastic/run.html RANK = int(os.getenv("RANK", -1)) WORLD_SIZE = int(os.getenv("WORLD_SIZE", 1)) GIT_INFO = check_git_info() def train(hyp, opt, device, callbacks): ( save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, noval, nosave, workers, freeze, mask_ratio, ) = ( Path(opt.save_dir), opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg, opt.resume, opt.noval, opt.nosave, opt.workers, opt.freeze, opt.mask_ratio, ) # callbacks.run('on_pretrain_routine_start') # Directories w = save_dir / "weights" # weights dir (w.parent if evolve else w).mkdir(parents=True, exist_ok=True) # make dir last, best = w / "last.pt", w / "best.pt" # Hyperparameters if isinstance(hyp, str): with open(hyp, errors="ignore") as f: hyp = yaml.safe_load(f) # load hyps dict LOGGER.info(colorstr("hyperparameters: ") + ", ".join(f"{k}={v}" for k, v in hyp.items())) opt.hyp = hyp.copy() # for saving hyps to checkpoints # Save run settings if not evolve: yaml_save(save_dir / "hyp.yaml", hyp) yaml_save(save_dir / "opt.yaml", vars(opt)) # Loggers data_dict = None if RANK in {-1, 0}: logger = GenericLogger(opt=opt, console_logger=LOGGER) # Config plots = not evolve and not opt.noplots # create plots overlap = not opt.no_overlap cuda = device.type != "cpu" init_seeds(opt.seed + 1 + RANK, deterministic=True) with torch_distributed_zero_first(LOCAL_RANK): data_dict = data_dict or check_dataset(data) # check if None train_path, val_path = data_dict["train"], data_dict["val"] nc = 1 if single_cls else int(data_dict["nc"]) # number of classes names = {0: "item"} if single_cls and len(data_dict["names"]) != 1 else data_dict["names"] # class names is_coco = isinstance(val_path, str) and val_path.endswith("coco/val2017.txt") # COCO dataset # Model check_suffix(weights, ".pt") # check weights pretrained = weights.endswith(".pt") if pretrained: with torch_distributed_zero_first(LOCAL_RANK): weights = attempt_download(weights) # download if not found locally ckpt = torch_load(weights, map_location="cpu") # load checkpoint to CPU to avoid CUDA memory leak model = SegmentationModel(cfg or ckpt["model"].yaml, ch=3, nc=nc, anchors=hyp.get("anchors")).to(device) exclude = ["anchor"] if (cfg or hyp.get("anchors")) and not resume else [] # exclude keys csd = ckpt["model"].float().state_dict() # checkpoint state_dict as FP32 csd = intersect_dicts(csd, model.state_dict(), exclude=exclude) # intersect model.load_state_dict(csd, strict=False) # load LOGGER.info(f"Transferred {len(csd)}/{len(model.state_dict())} items from {weights}") # report else: model = SegmentationModel(cfg, ch=3, nc=nc, anchors=hyp.get("anchors")).to(device) # create amp = check_amp(model) # check AMP # Freeze freeze = [f"model.{x}." for x in (freeze if len(freeze) > 1 else range(freeze[0]))] # layers to freeze for k, v in model.named_parameters(): v.requires_grad = True # train all layers # v.register_hook(lambda x: torch.nan_to_num(x)) # NaN to 0 (commented for erratic training results) if any(x in k for x in freeze): LOGGER.info(f"freezing {k}") v.requires_grad = False # Image size gs = max(int(model.stride.max()), 32) # grid size (max stride) imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2) # verify imgsz is gs-multiple # Batch size if RANK == -1 and batch_size == -1: # single-GPU only, estimate best batch size batch_size = check_train_batch_size(model, imgsz, amp) logger.update_params({"batch_size": batch_size}) # loggers.on_params_update({"batch_size": batch_size}) # Optimizer nbs = 64 # nominal batch size accumulate = max(round(nbs / batch_size), 1) # accumulate loss before optimizing hyp["weight_decay"] *= batch_size * accumulate / nbs # scale weight_decay optimizer = smart_optimizer(model, opt.optimizer, hyp["lr0"], hyp["momentum"], hyp["weight_decay"]) # Scheduler if opt.cos_lr: lf = one_cycle(1, hyp["lrf"], epochs) # cosine 1->hyp['lrf'] else: def lf(x): return (1 - x / epochs) * (1.0 - hyp["lrf"]) + hyp["lrf"] # linear scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs) # EMA ema = ModelEMA(model) if RANK in {-1, 0} else None # Resume best_fitness, start_epoch = 0.0, 0 if pretrained: if resume: best_fitness, start_epoch, epochs = smart_resume(ckpt, optimizer, ema, weights, epochs, resume) del ckpt, csd # DP mode if cuda and RANK == -1 and torch.cuda.device_count() > 1: LOGGER.warning( "WARNING ⚠️ DP not recommended, use torch.distributed.run for best DDP Multi-GPU results.\n" "See Multi-GPU Tutorial at https://docs.ultralytics.com/yolov5/tutorials/multi_gpu_training to get started." ) model = torch.nn.DataParallel(model) # SyncBatchNorm if opt.sync_bn and cuda and RANK != -1: model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device) LOGGER.info("Using SyncBatchNorm()") # Trainloader train_loader, dataset = create_dataloader( train_path, imgsz, batch_size // WORLD_SIZE, gs, single_cls, hyp=hyp, augment=True, cache=None if opt.cache == "val" else opt.cache, rect=opt.rect, rank=LOCAL_RANK, workers=workers, image_weights=opt.image_weights, quad=opt.quad, prefix=colorstr("train: "), shuffle=True, mask_downsample_ratio=mask_ratio, overlap_mask=overlap, ) labels = np.concatenate(dataset.labels, 0) mlc = int(labels[:, 0].max()) # max label class assert mlc < nc, f"Label class {mlc} exceeds nc={nc} in {data}. Possible class labels are 0-{nc - 1}" # Process 0 if RANK in {-1, 0}: val_loader = create_dataloader( val_path, imgsz, batch_size // WORLD_SIZE * 2, gs, single_cls, hyp=hyp, cache=None if noval else opt.cache, rect=True, rank=-1, workers=workers * 2, pad=0.5, mask_downsample_ratio=mask_ratio, overlap_mask=overlap, prefix=colorstr("val: "), )[0] if not resume: if not opt.noautoanchor: check_anchors(dataset, model=model, thr=hyp["anchor_t"], imgsz=imgsz) # run AutoAnchor model.half().float() # pre-reduce anchor precision if plots: plot_labels(labels, names, save_dir) # callbacks.run('on_pretrain_routine_end', labels, names) # DDP mode if cuda and RANK != -1: model = smart_DDP(model) # Model attributes nl = de_parallel(model).model[-1].nl # number of detection layers (to scale hyps) hyp["box"] *= 3 / nl # scale to layers hyp["cls"] *= nc / 80 * 3 / nl # scale to classes and layers hyp["obj"] *= (imgsz / 640) ** 2 * 3 / nl # scale to image size and layers hyp["label_smoothing"] = opt.label_smoothing model.nc = nc # attach number of classes to model model.hyp = hyp # attach hyperparameters to model model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device) * nc # attach class weights model.names = names # Start training t0 = time.time() nb = len(train_loader) # number of batches nw = max(round(hyp["warmup_epochs"] * nb), 100) # number of warmup iterations, max(3 epochs, 100 iterations) # nw = min(nw, (epochs - start_epoch) / 2 * nb) # limit warmup to < 1/2 of training last_opt_step = -1 maps = np.zeros(nc) # mAP per class results = (0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls) scheduler.last_epoch = start_epoch - 1 # do not move scaler = torch.cuda.amp.GradScaler(enabled=amp) stopper, stop = EarlyStopping(patience=opt.patience), False compute_loss = ComputeLoss(model, overlap=overlap) # init loss class # callbacks.run('on_train_start') LOGGER.info( f"Image sizes {imgsz} train, {imgsz} val\n" f"Using {train_loader.num_workers * WORLD_SIZE} dataloader workers\n" f"Logging results to {colorstr('bold', save_dir)}\n" f"Starting training for {epochs} epochs..." ) for epoch in range(start_epoch, epochs): # epoch ------------------------------------------------------------------ # callbacks.run('on_train_epoch_start') model.train() # Update image weights (optional, single-GPU only) if opt.image_weights: cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2 / nc # class weights iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw) # image weights dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n) # rand weighted idx # Update mosaic border (optional) # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs) # dataset.mosaic_border = [b - imgsz, -b] # height, width borders mloss = torch.zeros(4, device=device) # mean losses if RANK != -1: train_loader.sampler.set_epoch(epoch) pbar = enumerate(train_loader) LOGGER.info( ("\n" + "%11s" * 8) % ("Epoch", "GPU_mem", "box_loss", "seg_loss", "obj_loss", "cls_loss", "Instances", "Size") ) if RANK in {-1, 0}: pbar = tqdm(pbar, total=nb, bar_format=TQDM_BAR_FORMAT) # progress bar optimizer.zero_grad() for i, (imgs, targets, paths, _, masks) in pbar: # batch ------------------------------------------------------ # callbacks.run('on_train_batch_start') ni = i + nb * epoch # number integrated batches (since train start) imgs = imgs.to(device, non_blocking=True).float() / 255 # uint8 to float32, 0-255 to 0.0-1.0 # Warmup if ni <= nw: xi = [0, nw] # x interp # compute_loss.gr = np.interp(ni, xi, [0.0, 1.0]) # iou loss ratio (obj_loss = 1.0 or iou) accumulate = max(1, np.interp(ni, xi, [1, nbs / batch_size]).round()) for j, x in enumerate(optimizer.param_groups): # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0 x["lr"] = np.interp(ni, xi, [hyp["warmup_bias_lr"] if j == 0 else 0.0, x["initial_lr"] * lf(epoch)]) if "momentum" in x: x["momentum"] = np.interp(ni, xi, [hyp["warmup_momentum"], hyp["momentum"]]) # Multi-scale if opt.multi_scale: sz = random.randrange(int(imgsz * 0.5), int(imgsz * 1.5) + gs) // gs * gs # size sf = sz / max(imgs.shape[2:]) # scale factor if sf != 1: ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]] # new shape (stretched to gs-multiple) imgs = nn.functional.interpolate(imgs, size=ns, mode="bilinear", align_corners=False) # Forward with torch.cuda.amp.autocast(amp): pred = model(imgs) # forward loss, loss_items = compute_loss(pred, targets.to(device), masks=masks.to(device).float()) if RANK != -1: loss *= WORLD_SIZE # gradient averaged between devices in DDP mode if opt.quad: loss *= 4.0 # Backward scaler.scale(loss).backward() # Optimize - https://pytorch.org/docs/master/notes/amp_examples.html if ni - last_opt_step >= accumulate: scaler.unscale_(optimizer) # unscale gradients torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=10.0) # clip gradients scaler.step(optimizer) # optimizer.step scaler.update() optimizer.zero_grad() if ema: ema.update(model) last_opt_step = ni # Log if RANK in {-1, 0}: mloss = (mloss * i + loss_items) / (i + 1) # update mean losses mem = f"{torch.cuda.memory_reserved() / 1e9 if torch.cuda.is_available() else 0:.3g}G" # (GB) pbar.set_description( ("%11s" * 2 + "%11.4g" * 6) % (f"{epoch}/{epochs - 1}", mem, *mloss, targets.shape[0], imgs.shape[-1]) ) # callbacks.run('on_train_batch_end', model, ni, imgs, targets, paths) # if callbacks.stop_training: # return # Mosaic plots if plots: if ni < 3: plot_images_and_masks(imgs, targets, masks, paths, save_dir / f"train_batch{ni}.jpg") if ni == 10: files = sorted(save_dir.glob("train*.jpg")) logger.log_images(files, "Mosaics", epoch) # end batch ------------------------------------------------------------------------------------------------ # Scheduler lr = [x["lr"] for x in optimizer.param_groups] # for loggers scheduler.step() if RANK in {-1, 0}: # mAP # callbacks.run('on_train_epoch_end', epoch=epoch) ema.update_attr(model, include=["yaml", "nc", "hyp", "names", "stride", "class_weights"]) final_epoch = (epoch + 1 == epochs) or stopper.possible_stop if not noval or final_epoch: # Calculate mAP results, maps, _ = validate.run( data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz, half=amp, model=ema.ema, single_cls=single_cls, dataloader=val_loader, save_dir=save_dir, plots=False, callbacks=callbacks, compute_loss=compute_loss, mask_downsample_ratio=mask_ratio, overlap=overlap, ) # Update best mAP fi = fitness(np.array(results).reshape(1, -1)) # weighted combination of [P, R, mAP@.5, mAP@.5-.95] stop = stopper(epoch=epoch, fitness=fi) # early stop check if fi > best_fitness: best_fitness = fi log_vals = list(mloss) + list(results) + lr # callbacks.run('on_fit_epoch_end', log_vals, epoch, best_fitness, fi) # Log val metrics and media metrics_dict = dict(zip(KEYS, log_vals)) logger.log_metrics(metrics_dict, epoch) # Save model if (not nosave) or (final_epoch and not evolve): # if save ckpt = { "epoch": epoch, "best_fitness": best_fitness, "model": deepcopy(de_parallel(model)).half(), "ema": deepcopy(ema.ema).half(), "updates": ema.updates, "optimizer": optimizer.state_dict(), "opt": vars(opt), "git": GIT_INFO, # {remote, branch, commit} if a git repo "date": datetime.now().isoformat(), } # Save last, best and delete torch.save(ckpt, last) if best_fitness == fi: torch.save(ckpt, best) if opt.save_period > 0 and epoch % opt.save_period == 0: torch.save(ckpt, w / f"epoch{epoch}.pt") logger.log_model(w / f"epoch{epoch}.pt") del ckpt # callbacks.run('on_model_save', last, epoch, final_epoch, best_fitness, fi) # EarlyStopping if RANK != -1: # if DDP training broadcast_list = [stop if RANK == 0 else None] dist.broadcast_object_list(broadcast_list, 0) # broadcast 'stop' to all ranks if RANK != 0: stop = broadcast_list[0] if stop: break # must break all DDP ranks # end epoch ---------------------------------------------------------------------------------------------------- # end training ----------------------------------------------------------------------------------------------------- if RANK in {-1, 0}: LOGGER.info(f"\n{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.") for f in last, best: if f.exists(): strip_optimizer(f) # strip optimizers if f is best: LOGGER.info(f"\nValidating {f}...") results, _, _ = validate.run( data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz, model=attempt_load(f, device).half(), iou_thres=0.65 if is_coco else 0.60, # best pycocotools at iou 0.65 single_cls=single_cls, dataloader=val_loader, save_dir=save_dir, save_json=is_coco, verbose=True, plots=plots, callbacks=callbacks, compute_loss=compute_loss, mask_downsample_ratio=mask_ratio, overlap=overlap, ) # val best model with plots if is_coco: # callbacks.run('on_fit_epoch_end', list(mloss) + list(results) + lr, epoch, best_fitness, fi) metrics_dict = dict(zip(KEYS, list(mloss) + list(results) + lr)) logger.log_metrics(metrics_dict, epoch) # callbacks.run('on_train_end', last, best, epoch, results) # on train end callback using genericLogger logger.log_metrics(dict(zip(KEYS[4:16], results)), epochs) if not opt.evolve: logger.log_model(best, epoch) if plots: plot_results_with_masks(file=save_dir / "results.csv") # save results.png files = ["results.png", "confusion_matrix.png", *(f"{x}_curve.png" for x in ("F1", "PR", "P", "R"))] files = [(save_dir / f) for f in files if (save_dir / f).exists()] # filter LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}") logger.log_images(files, "Results", epoch + 1) logger.log_images(sorted(save_dir.glob("val*.jpg")), "Validation", epoch + 1) torch.cuda.empty_cache() return results def parse_opt(known=False): parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s-seg.pt", help="initial weights path") parser.add_argument("--cfg", type=str, default="", help="model.yaml path") parser.add_argument("--data", type=str, default=ROOT / "data/coco128-seg.yaml", help="dataset.yaml path") parser.add_argument("--hyp", type=str, default=ROOT / "data/hyps/hyp.scratch-low.yaml", help="hyperparameters path") parser.add_argument("--epochs", type=int, default=100, help="total training epochs") parser.add_argument("--batch-size", type=int, default=16, help="total batch size for all GPUs, -1 for autobatch") parser.add_argument("--imgsz", "--img", "--img-size", type=int, default=640, help="train, val image size (pixels)") parser.add_argument("--rect", action="store_true", help="rectangular training") parser.add_argument("--resume", nargs="?", const=True, default=False, help="resume most recent training") parser.add_argument("--nosave", action="store_true", help="only save final checkpoint") parser.add_argument("--noval", action="store_true", help="only validate final epoch") parser.add_argument("--noautoanchor", action="store_true", help="disable AutoAnchor") parser.add_argument("--noplots", action="store_true", help="save no plot files") parser.add_argument("--evolve", type=int, nargs="?", const=300, help="evolve hyperparameters for x generations") parser.add_argument("--bucket", type=str, default="", help="gsutil bucket") parser.add_argument("--cache", type=str, nargs="?", const="ram", help="image --cache ram/disk") parser.add_argument("--image-weights", action="store_true", help="use weighted image selection for training") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--multi-scale", action="store_true", help="vary img-size +/- 50%%") parser.add_argument("--single-cls", action="store_true", help="train multi-class data as single-class") parser.add_argument("--optimizer", type=str, choices=["SGD", "Adam", "AdamW"], default="SGD", help="optimizer") parser.add_argument("--sync-bn", action="store_true", help="use SyncBatchNorm, only available in DDP mode") parser.add_argument("--workers", type=int, default=8, help="max dataloader workers (per RANK in DDP mode)") parser.add_argument("--project", default=ROOT / "runs/train-seg", help="save to project/name") parser.add_argument("--name", default="exp", help="save to project/name") parser.add_argument("--exist-ok", action="store_true", help="existing project/name ok, do not increment") parser.add_argument("--quad", action="store_true", help="quad dataloader") parser.add_argument("--cos-lr", action="store_true", help="cosine LR scheduler") parser.add_argument("--label-smoothing", type=float, default=0.0, help="Label smoothing epsilon") parser.add_argument("--patience", type=int, default=100, help="EarlyStopping patience (epochs without improvement)") parser.add_argument("--freeze", nargs="+", type=int, default=[0], help="Freeze layers: backbone=10, first3=0 1 2") parser.add_argument("--save-period", type=int, default=-1, help="Save checkpoint every x epochs (disabled if < 1)") parser.add_argument("--seed", type=int, default=0, help="Global training seed") parser.add_argument("--local_rank", type=int, default=-1, help="Automatic DDP Multi-GPU argument, do not modify") # Instance Segmentation Args parser.add_argument("--mask-ratio", type=int, default=4, help="Downsample the truth masks to saving memory") parser.add_argument("--no-overlap", action="store_true", help="Overlap masks train faster at slightly less mAP") return parser.parse_known_args()[0] if known else parser.parse_args() def main(opt, callbacks=Callbacks()): if RANK in {-1, 0}: print_args(vars(opt)) check_git_status() check_requirements(ROOT / "requirements.txt") # Resume if opt.resume and not opt.evolve: # resume from specified or most recent last.pt last = Path(check_file(opt.resume) if isinstance(opt.resume, str) else get_latest_run()) opt_yaml = last.parent.parent / "opt.yaml" # train options yaml opt_data = opt.data # original dataset if opt_yaml.is_file(): with open(opt_yaml, errors="ignore") as f: d = yaml.safe_load(f) else: d = torch_load(last, map_location="cpu")["opt"] opt = argparse.Namespace(**d) # replace opt.cfg, opt.weights, opt.resume = "", str(last), True # reinstate if is_url(opt_data): opt.data = check_file(opt_data) # avoid HUB resume auth timeout else: opt.data, opt.cfg, opt.hyp, opt.weights, opt.project = ( check_file(opt.data), check_yaml(opt.cfg), check_yaml(opt.hyp), str(opt.weights), str(opt.project), ) # checks assert len(opt.cfg) or len(opt.weights), "either --cfg or --weights must be specified" if opt.evolve: if opt.project == str(ROOT / "runs/train-seg"): # if default project name, rename to runs/evolve-seg opt.project = str(ROOT / "runs/evolve-seg") opt.exist_ok, opt.resume = opt.resume, False # pass resume to exist_ok and disable resume if opt.name == "cfg": opt.name = Path(opt.cfg).stem # use model.yaml as name opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)) # DDP mode device = select_device(opt.device, batch_size=opt.batch_size) if LOCAL_RANK != -1: msg = "is not compatible with YOLOv5 Multi-GPU DDP training" assert not opt.image_weights, f"--image-weights {msg}" assert not opt.evolve, f"--evolve {msg}" assert opt.batch_size != -1, f"AutoBatch with --batch-size -1 {msg}, please pass a valid --batch-size" assert opt.batch_size % WORLD_SIZE == 0, f"--batch-size {opt.batch_size} must be multiple of WORLD_SIZE" assert torch.cuda.device_count() > LOCAL_RANK, "insufficient CUDA devices for DDP command" torch.cuda.set_device(LOCAL_RANK) device = torch.device("cuda", LOCAL_RANK) dist.init_process_group(backend="nccl" if dist.is_nccl_available() else "gloo") # Train if not opt.evolve: train(opt.hyp, opt, device, callbacks) # Evolve hyperparameters (optional) else: # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit) meta = { "lr0": (1, 1e-5, 1e-1), # initial learning rate (SGD=1E-2, Adam=1E-3) "lrf": (1, 0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf) "momentum": (0.3, 0.6, 0.98), # SGD momentum/Adam beta1 "weight_decay": (1, 0.0, 0.001), # optimizer weight decay "warmup_epochs": (1, 0.0, 5.0), # warmup epochs (fractions ok) "warmup_momentum": (1, 0.0, 0.95), # warmup initial momentum "warmup_bias_lr": (1, 0.0, 0.2), # warmup initial bias lr "box": (1, 0.02, 0.2), # box loss gain "cls": (1, 0.2, 4.0), # cls loss gain "cls_pw": (1, 0.5, 2.0), # cls BCELoss positive_weight "obj": (1, 0.2, 4.0), # obj loss gain (scale with pixels) "obj_pw": (1, 0.5, 2.0), # obj BCELoss positive_weight "iou_t": (0, 0.1, 0.7), # IoU training threshold "anchor_t": (1, 2.0, 8.0), # anchor-multiple threshold "anchors": (2, 2.0, 10.0), # anchors per output grid (0 to ignore) "fl_gamma": (0, 0.0, 2.0), # focal loss gamma (efficientDet default gamma=1.5) "hsv_h": (1, 0.0, 0.1), # image HSV-Hue augmentation (fraction) "hsv_s": (1, 0.0, 0.9), # image HSV-Saturation augmentation (fraction) "hsv_v": (1, 0.0, 0.9), # image HSV-Value augmentation (fraction) "degrees": (1, 0.0, 45.0), # image rotation (+/- deg) "translate": (1, 0.0, 0.9), # image translation (+/- fraction) "scale": (1, 0.0, 0.9), # image scale (+/- gain) "shear": (1, 0.0, 10.0), # image shear (+/- deg) "perspective": (0, 0.0, 0.001), # image perspective (+/- fraction), range 0-0.001 "flipud": (1, 0.0, 1.0), # image flip up-down (probability) "fliplr": (0, 0.0, 1.0), # image flip left-right (probability) "mosaic": (1, 0.0, 1.0), # image mixup (probability) "mixup": (1, 0.0, 1.0), # image mixup (probability) "copy_paste": (1, 0.0, 1.0), } # segment copy-paste (probability) with open(opt.hyp, errors="ignore") as f: hyp = yaml.safe_load(f) # load hyps dict if "anchors" not in hyp: # anchors commented in hyp.yaml hyp["anchors"] = 3 if opt.noautoanchor: del hyp["anchors"], meta["anchors"] opt.noval, opt.nosave, save_dir = True, True, Path(opt.save_dir) # only val/save final epoch # ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices evolve_yaml, evolve_csv = save_dir / "hyp_evolve.yaml", save_dir / "evolve.csv" if opt.bucket: # download evolve.csv if exists subprocess.run( [ "gsutil", "cp", f"gs://{opt.bucket}/evolve.csv", str(evolve_csv), ] ) for _ in range(opt.evolve): # generations to evolve if evolve_csv.exists(): # if evolve.csv exists: select best hyps and mutate # Select parent(s) parent = "single" # parent selection method: 'single' or 'weighted' x = np.loadtxt(evolve_csv, ndmin=2, delimiter=",", skiprows=1) n = min(5, len(x)) # number of previous results to consider x = x[np.argsort(-fitness(x))][:n] # top n mutations w = fitness(x) - fitness(x).min() + 1e-6 # weights (sum > 0) if parent == "single" or len(x) == 1: # x = x[random.randint(0, n - 1)] # random selection x = x[random.choices(range(n), weights=w)[0]] # weighted selection elif parent == "weighted": x = (x * w.reshape(n, 1)).sum(0) / w.sum() # weighted combination # Mutate mp, s = 0.8, 0.2 # mutation probability, sigma npr = np.random npr.seed(int(time.time())) g = np.array([meta[k][0] for k in hyp.keys()]) # gains 0-1 ng = len(meta) v = np.ones(ng) while all(v == 1): # mutate until a change occurs (prevent duplicates) v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0) for i, k in enumerate(hyp.keys()): # plt.hist(v.ravel(), 300) hyp[k] = float(x[i + 12] * v[i]) # mutate # Constrain to limits for k, v in meta.items(): hyp[k] = max(hyp[k], v[1]) # lower limit hyp[k] = min(hyp[k], v[2]) # upper limit hyp[k] = round(hyp[k], 5) # significant digits # Train mutation results = train(hyp.copy(), opt, device, callbacks) callbacks = Callbacks() # Write mutation results print_mutation(KEYS[4:16], results, hyp.copy(), save_dir, opt.bucket) # Plot results plot_evolve(evolve_csv) LOGGER.info( f"Hyperparameter evolution finished {opt.evolve} generations\n" f"Results saved to {colorstr('bold', save_dir)}\n" f"Usage example: $ python train.py --hyp {evolve_yaml}" ) def run(**kwargs): opt = parse_opt(True) for k, v in kwargs.items(): setattr(opt, k, v) main(opt) return opt if __name__ == "__main__": opt = parse_opt() main(opt)
--- +++ @@ -1,4 +1,19 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +Train a YOLOv5 segment model on a segment dataset Models and datasets download automatically from the latest YOLOv5 +release. + +Usage - Single-GPU training: + $ python segment/train.py --data coco128-seg.yaml --weights yolov5s-seg.pt --img 640 # from pretrained (recommended) + $ python segment/train.py --data coco128-seg.yaml --weights '' --cfg yolov5s-seg.yaml --img 640 # from scratch + +Usage - Multi-GPU DDP training: + $ python -m torch.distributed.run --nproc_per_node 4 --master_port 1 segment/train.py --data coco128-seg.yaml --weights yolov5s-seg.pt --img 640 --device 0,1,2,3 + +Models: https://github.com/ultralytics/yolov5/tree/master/models +Datasets: https://github.com/ultralytics/yolov5/tree/master/data +Tutorial: https://docs.ultralytics.com/yolov5/tutorials/train_custom_data +""" import argparse import math @@ -83,6 +98,10 @@ def train(hyp, opt, device, callbacks): + """Trains the YOLOv5 model on a dataset, managing hyperparameters, model optimization, logging, and validation. + + `hyp` is path/to/hyp.yaml or hyp dictionary. + """ ( save_dir, epochs, @@ -198,6 +217,7 @@ else: def lf(x): + """Linear learning rate scheduler decreasing from 1 to hyp['lrf'] over 'epochs'.""" return (1 - x / epochs) * (1.0 - hyp["lrf"]) + hyp["lrf"] # linear scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs) @@ -522,6 +542,10 @@ def parse_opt(known=False): + """Parses command line arguments for training configurations, returning parsed arguments. + + Supports both known and unknown args. + """ parser = argparse.ArgumentParser() parser.add_argument("--weights", type=str, default=ROOT / "yolov5s-seg.pt", help="initial weights path") parser.add_argument("--cfg", type=str, default="", help="model.yaml path") @@ -566,6 +590,7 @@ def main(opt, callbacks=Callbacks()): + """Initializes training or evolution of YOLOv5 models based on provided configuration and options.""" if RANK in {-1, 0}: print_args(vars(opt)) check_git_status() @@ -722,6 +747,10 @@ def run(**kwargs): + """Executes YOLOv5 training with given parameters, altering options programmatically; returns updated options. + + Example: import train; train.run(data='coco128.yaml', imgsz=320, weights='yolov5m.pt') + """ opt = parse_opt(True) for k, v in kwargs.items(): setattr(opt, k, v) @@ -731,4 +760,4 @@ if __name__ == "__main__": opt = parse_opt() - main(opt)+ main(opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/segment/train.py
Help me comply with documentation standards
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import torch import torch.nn as nn import torch.nn.functional as F from ..general import xywh2xyxy from ..loss import FocalLoss, smooth_BCE from ..metrics import bbox_iou from ..torch_utils import de_parallel from .general import crop_mask class ComputeLoss: def __init__(self, model, autobalance=False, overlap=False): self.sort_obj_iou = False self.overlap = overlap device = next(model.parameters()).device # get model device h = model.hyp # hyperparameters # Define criteria BCEcls = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h["cls_pw"]], device=device)) BCEobj = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h["obj_pw"]], device=device)) # Class label smoothing https://arxiv.org/pdf/1902.04103.pdf eqn 3 self.cp, self.cn = smooth_BCE(eps=h.get("label_smoothing", 0.0)) # positive, negative BCE targets # Focal loss g = h["fl_gamma"] # focal loss gamma if g > 0: BCEcls, BCEobj = FocalLoss(BCEcls, g), FocalLoss(BCEobj, g) m = de_parallel(model).model[-1] # Detect() module self.balance = {3: [4.0, 1.0, 0.4]}.get(m.nl, [4.0, 1.0, 0.25, 0.06, 0.02]) # P3-P7 self.ssi = list(m.stride).index(16) if autobalance else 0 # stride 16 index self.BCEcls, self.BCEobj, self.gr, self.hyp, self.autobalance = BCEcls, BCEobj, 1.0, h, autobalance self.na = m.na # number of anchors self.nc = m.nc # number of classes self.nl = m.nl # number of layers self.nm = m.nm # number of masks self.anchors = m.anchors self.device = device def __call__(self, preds, targets, masks): # predictions, targets, model p, proto = preds bs, nm, mask_h, mask_w = proto.shape # batch size, number of masks, mask height, mask width lcls = torch.zeros(1, device=self.device) lbox = torch.zeros(1, device=self.device) lobj = torch.zeros(1, device=self.device) lseg = torch.zeros(1, device=self.device) tcls, tbox, indices, anchors, tidxs, xywhn = self.build_targets(p, targets) # targets # Losses for i, pi in enumerate(p): # layer index, layer predictions b, a, gj, gi = indices[i] # image, anchor, gridy, gridx tobj = torch.zeros(pi.shape[:4], dtype=pi.dtype, device=self.device) # target obj if n := b.shape[0]: pxy, pwh, _, pcls, pmask = pi[b, a, gj, gi].split((2, 2, 1, self.nc, nm), 1) # subset of predictions # Box regression pxy = pxy.sigmoid() * 2 - 0.5 pwh = (pwh.sigmoid() * 2) ** 2 * anchors[i] pbox = torch.cat((pxy, pwh), 1) # predicted box iou = bbox_iou(pbox, tbox[i], CIoU=True).squeeze() # iou(prediction, target) lbox += (1.0 - iou).mean() # iou loss # Objectness iou = iou.detach().clamp(0).type(tobj.dtype) if self.sort_obj_iou: j = iou.argsort() b, a, gj, gi, iou = b[j], a[j], gj[j], gi[j], iou[j] if self.gr < 1: iou = (1.0 - self.gr) + self.gr * iou tobj[b, a, gj, gi] = iou # iou ratio # Classification if self.nc > 1: # cls loss (only if multiple classes) t = torch.full_like(pcls, self.cn, device=self.device) # targets t[range(n), tcls[i]] = self.cp lcls += self.BCEcls(pcls, t) # BCE # Mask regression if tuple(masks.shape[-2:]) != (mask_h, mask_w): # downsample masks = F.interpolate(masks[None], (mask_h, mask_w), mode="nearest")[0] marea = xywhn[i][:, 2:].prod(1) # mask width, height normalized mxyxy = xywh2xyxy(xywhn[i] * torch.tensor([mask_w, mask_h, mask_w, mask_h], device=self.device)) for bi in b.unique(): j = b == bi # matching index if self.overlap: mask_gti = torch.where(masks[bi][None] == tidxs[i][j].view(-1, 1, 1), 1.0, 0.0) else: mask_gti = masks[tidxs[i]][j] lseg += self.single_mask_loss(mask_gti, pmask[j], proto[bi], mxyxy[j], marea[j]) obji = self.BCEobj(pi[..., 4], tobj) lobj += obji * self.balance[i] # obj loss if self.autobalance: self.balance[i] = self.balance[i] * 0.9999 + 0.0001 / obji.detach().item() if self.autobalance: self.balance = [x / self.balance[self.ssi] for x in self.balance] lbox *= self.hyp["box"] lobj *= self.hyp["obj"] lcls *= self.hyp["cls"] lseg *= self.hyp["box"] / bs loss = lbox + lobj + lcls + lseg return loss * bs, torch.cat((lbox, lseg, lobj, lcls)).detach() def single_mask_loss(self, gt_mask, pred, proto, xyxy, area): pred_mask = (pred @ proto.view(self.nm, -1)).view(-1, *proto.shape[1:]) # (n,32) @ (32,80,80) -> (n,80,80) loss = F.binary_cross_entropy_with_logits(pred_mask, gt_mask, reduction="none") return (crop_mask(loss, xyxy).mean(dim=(1, 2)) / area).mean() def build_targets(self, p, targets): na, nt = self.na, targets.shape[0] # number of anchors, targets tcls, tbox, indices, anch, tidxs, xywhn = [], [], [], [], [], [] gain = torch.ones(8, device=self.device) # normalized to gridspace gain ai = torch.arange(na, device=self.device).float().view(na, 1).repeat(1, nt) # same as .repeat_interleave(nt) if self.overlap: batch = p[0].shape[0] ti = [] for i in range(batch): num = (targets[:, 0] == i).sum() # find number of targets of each image ti.append(torch.arange(num, device=self.device).float().view(1, num).repeat(na, 1) + 1) # (na, num) ti = torch.cat(ti, 1) # (na, nt) else: ti = torch.arange(nt, device=self.device).float().view(1, nt).repeat(na, 1) targets = torch.cat((targets.repeat(na, 1, 1), ai[..., None], ti[..., None]), 2) # append anchor indices g = 0.5 # bias off = ( torch.tensor( [ [0, 0], [1, 0], [0, 1], [-1, 0], [0, -1], # j,k,l,m # [1, 1], [1, -1], [-1, 1], [-1, -1], # jk,jm,lk,lm ], device=self.device, ).float() * g ) # offsets for i in range(self.nl): anchors, shape = self.anchors[i], p[i].shape gain[2:6] = torch.tensor(shape)[[3, 2, 3, 2]] # xyxy gain # Match targets to anchors t = targets * gain # shape(3,n,7) if nt: # Matches r = t[..., 4:6] / anchors[:, None] # wh ratio j = torch.max(r, 1 / r).max(2)[0] < self.hyp["anchor_t"] # compare # j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t'] # iou(3,n)=wh_iou(anchors(3,2), gwh(n,2)) t = t[j] # filter # Offsets gxy = t[:, 2:4] # grid xy gxi = gain[[2, 3]] - gxy # inverse j, k = ((gxy % 1 < g) & (gxy > 1)).T l, m = ((gxi % 1 < g) & (gxi > 1)).T j = torch.stack((torch.ones_like(j), j, k, l, m)) t = t.repeat((5, 1, 1))[j] offsets = (torch.zeros_like(gxy)[None] + off[:, None])[j] else: t = targets[0] offsets = 0 # Define bc, gxy, gwh, at = t.chunk(4, 1) # (image, class), grid xy, grid wh, anchors (a, tidx), (b, c) = at.long().T, bc.long().T # anchors, image, class gij = (gxy - offsets).long() gi, gj = gij.T # grid indices # Append indices.append((b, a, gj.clamp_(0, shape[2] - 1), gi.clamp_(0, shape[3] - 1))) # image, anchor, grid tbox.append(torch.cat((gxy - gij, gwh), 1)) # box anch.append(anchors[a]) # anchors tcls.append(c) # class tidxs.append(tidx) xywhn.append(torch.cat((gxy, gwh), 1) / gain[2:6]) # xywh normalized return tcls, tbox, indices, anch, tidxs, xywhn
--- +++ @@ -12,8 +12,10 @@ class ComputeLoss: + """Computes the YOLOv5 model's loss components including classification, objectness, box, and mask losses.""" def __init__(self, model, autobalance=False, overlap=False): + """Initialize compute loss function for YOLOv5 models with options for autobalancing and overlap handling.""" self.sort_obj_iou = False self.overlap = overlap device = next(model.parameters()).device # get model device @@ -43,6 +45,7 @@ self.device = device def __call__(self, preds, targets, masks): # predictions, targets, model + """Evaluates YOLOv5 model's loss for given predictions, targets, and masks; returns total loss components.""" p, proto = preds bs, nm, mask_h, mask_w = proto.shape # batch size, number of masks, mask height, mask width lcls = torch.zeros(1, device=self.device) @@ -110,11 +113,15 @@ return loss * bs, torch.cat((lbox, lseg, lobj, lcls)).detach() def single_mask_loss(self, gt_mask, pred, proto, xyxy, area): + """Calculates and normalizes single mask loss for YOLOv5 between predicted and ground truth masks.""" pred_mask = (pred @ proto.view(self.nm, -1)).view(-1, *proto.shape[1:]) # (n,32) @ (32,80,80) -> (n,80,80) loss = F.binary_cross_entropy_with_logits(pred_mask, gt_mask, reduction="none") return (crop_mask(loss, xyxy).mean(dim=(1, 2)) / area).mean() def build_targets(self, p, targets): + """Prepares YOLOv5 targets for loss computation; inputs targets (image, class, x, y, w, h), output target + classes/boxes. + """ na, nt = self.na, targets.shape[0] # number of anchors, targets tcls, tbox, indices, anch, tidxs, xywhn = [], [], [], [], [], [] gain = torch.ones(8, device=self.device) # normalized to gridspace gain @@ -185,4 +192,4 @@ tidxs.append(tidx) xywhn.append(torch.cat((gxy, gwh), 1) / gain[2:6]) # xywh normalized - return tcls, tbox, indices, anch, tidxs, xywhn+ return tcls, tbox, indices, anch, tidxs, xywhn
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/segment/loss.py
Document this code for team use
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import contextlib import glob import hashlib import json import math import os import random import shutil import time from itertools import repeat from multiprocessing.pool import Pool, ThreadPool from pathlib import Path from threading import Thread from urllib.parse import urlparse import numpy as np import psutil import torch import torch.nn.functional as F import torchvision import yaml from PIL import ExifTags, Image, ImageOps from torch.utils.data import DataLoader, Dataset, dataloader, distributed from tqdm import tqdm from utils.augmentations import ( Albumentations, augment_hsv, classify_albumentations, classify_transforms, copy_paste, letterbox, mixup, random_perspective, ) from utils.general import ( DATASETS_DIR, LOGGER, NUM_THREADS, TQDM_BAR_FORMAT, check_dataset, check_requirements, check_yaml, clean_str, cv2, is_colab, is_kaggle, segments2boxes, unzip_file, xyn2xy, xywh2xyxy, xywhn2xyxy, xyxy2xywhn, ) from utils.torch_utils import torch_distributed_zero_first # Parameters HELP_URL = "See https://docs.ultralytics.com/yolov5/tutorials/train_custom_data" IMG_FORMATS = "bmp", "dng", "jpeg", "jpg", "mpo", "png", "tif", "tiff", "webp", "pfm" # include image suffixes VID_FORMATS = "asf", "avi", "gif", "m4v", "mkv", "mov", "mp4", "mpeg", "mpg", "ts", "wmv" # include video suffixes LOCAL_RANK = int(os.getenv("LOCAL_RANK", -1)) # https://pytorch.org/docs/stable/elastic/run.html RANK = int(os.getenv("RANK", -1)) WORLD_SIZE = int(os.getenv("WORLD_SIZE", 1)) PIN_MEMORY = str(os.getenv("PIN_MEMORY", True)).lower() == "true" # global pin_memory for dataloaders # Get orientation exif tag for orientation in ExifTags.TAGS.keys(): if ExifTags.TAGS[orientation] == "Orientation": break def get_hash(paths): size = sum(os.path.getsize(p) for p in paths if os.path.exists(p)) # sizes h = hashlib.sha256(str(size).encode()) # hash sizes h.update("".join(paths).encode()) # hash paths return h.hexdigest() # return hash def exif_size(img): s = img.size # (width, height) with contextlib.suppress(Exception): rotation = dict(img._getexif().items())[orientation] if rotation in [6, 8]: # rotation 270 or 90 s = (s[1], s[0]) return s def exif_transpose(image): exif = image.getexif() orientation = exif.get(0x0112, 1) # default 1 if orientation > 1: method = { 2: Image.FLIP_LEFT_RIGHT, 3: Image.ROTATE_180, 4: Image.FLIP_TOP_BOTTOM, 5: Image.TRANSPOSE, 6: Image.ROTATE_270, 7: Image.TRANSVERSE, 8: Image.ROTATE_90, }.get(orientation) if method is not None: image = image.transpose(method) del exif[0x0112] image.info["exif"] = exif.tobytes() return image def seed_worker(worker_id): worker_seed = torch.initial_seed() % 2**32 np.random.seed(worker_seed) random.seed(worker_seed) # Inherit from DistributedSampler and override iterator # https://github.com/pytorch/pytorch/blob/master/torch/utils/data/distributed.py class SmartDistributedSampler(distributed.DistributedSampler): def __iter__(self): g = torch.Generator() g.manual_seed(self.seed + self.epoch) # determine the eventual size (n) of self.indices (DDP indices) n = int((len(self.dataset) - self.rank - 1) / self.num_replicas) + 1 # num_replicas == WORLD_SIZE idx = torch.randperm(n, generator=g) if not self.shuffle: idx = idx.sort()[0] idx = idx.tolist() if self.drop_last: idx = idx[: self.num_samples] else: padding_size = self.num_samples - len(idx) if padding_size <= len(idx): idx += idx[:padding_size] else: idx += (idx * math.ceil(padding_size / len(idx)))[:padding_size] return iter(idx) def create_dataloader( path, imgsz, batch_size, stride, single_cls=False, hyp=None, augment=False, cache=False, pad=0.0, rect=False, rank=-1, workers=8, image_weights=False, quad=False, prefix="", shuffle=False, seed=0, ): if rect and shuffle: LOGGER.warning("WARNING ⚠️ --rect is incompatible with DataLoader shuffle, setting shuffle=False") shuffle = False with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP dataset = LoadImagesAndLabels( path, imgsz, batch_size, augment=augment, # augmentation hyp=hyp, # hyperparameters rect=rect, # rectangular batches cache_images=cache, single_cls=single_cls, stride=int(stride), pad=pad, image_weights=image_weights, prefix=prefix, rank=rank, ) batch_size = min(batch_size, len(dataset)) nd = torch.cuda.device_count() # number of CUDA devices nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers]) # number of workers sampler = None if rank == -1 else SmartDistributedSampler(dataset, shuffle=shuffle) loader = DataLoader if image_weights else InfiniteDataLoader # only DataLoader allows for attribute updates generator = torch.Generator() generator.manual_seed(6148914691236517205 + seed + RANK) return loader( dataset, batch_size=batch_size, shuffle=shuffle and sampler is None, num_workers=nw, sampler=sampler, drop_last=quad, pin_memory=PIN_MEMORY, collate_fn=LoadImagesAndLabels.collate_fn4 if quad else LoadImagesAndLabels.collate_fn, worker_init_fn=seed_worker, generator=generator, ), dataset class InfiniteDataLoader(dataloader.DataLoader): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) object.__setattr__(self, "batch_sampler", _RepeatSampler(self.batch_sampler)) self.iterator = super().__iter__() def __len__(self): return len(self.batch_sampler.sampler) def __iter__(self): for _ in range(len(self)): yield next(self.iterator) class _RepeatSampler: def __init__(self, sampler): self.sampler = sampler def __iter__(self): while True: yield from iter(self.sampler) class LoadScreenshots: def __init__(self, source, img_size=640, stride=32, auto=True, transforms=None): check_requirements("mss") import mss source, *params = source.split() self.screen, left, top, width, height = 0, None, None, None, None # default to full screen 0 if len(params) == 1: self.screen = int(params[0]) elif len(params) == 4: left, top, width, height = (int(x) for x in params) elif len(params) == 5: self.screen, left, top, width, height = (int(x) for x in params) self.img_size = img_size self.stride = stride self.transforms = transforms self.auto = auto self.mode = "stream" self.frame = 0 self.sct = mss.mss() # Parse monitor shape monitor = self.sct.monitors[self.screen] self.top = monitor["top"] if top is None else (monitor["top"] + top) self.left = monitor["left"] if left is None else (monitor["left"] + left) self.width = width or monitor["width"] self.height = height or monitor["height"] self.monitor = {"left": self.left, "top": self.top, "width": self.width, "height": self.height} def __iter__(self): return self def __next__(self): im0 = np.array(self.sct.grab(self.monitor))[:, :, :3] # [:, :, :3] BGRA to BGR s = f"screen {self.screen} (LTWH): {self.left},{self.top},{self.width},{self.height}: " if self.transforms: im = self.transforms(im0) # transforms else: im = letterbox(im0, self.img_size, stride=self.stride, auto=self.auto)[0] # padded resize im = im.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB im = np.ascontiguousarray(im) # contiguous self.frame += 1 return str(self.screen), im, im0, None, s # screen, img, original img, im0s, s class LoadImages: def __init__(self, path, img_size=640, stride=32, auto=True, transforms=None, vid_stride=1): if isinstance(path, str) and Path(path).suffix == ".txt": # *.txt file with img/vid/dir on each line path = Path(path).read_text().rsplit() files = [] for p in sorted(path) if isinstance(path, (list, tuple)) else [path]: p = str(Path(p).resolve()) if "*" in p: files.extend(sorted(glob.glob(p, recursive=True))) # glob elif os.path.isdir(p): files.extend(sorted(glob.glob(os.path.join(p, "*.*")))) # dir elif os.path.isfile(p): files.append(p) # files else: raise FileNotFoundError(f"{p} does not exist") images = [x for x in files if x.split(".")[-1].lower() in IMG_FORMATS] videos = [x for x in files if x.split(".")[-1].lower() in VID_FORMATS] ni, nv = len(images), len(videos) self.img_size = img_size self.stride = stride self.files = images + videos self.nf = ni + nv # number of files self.video_flag = [False] * ni + [True] * nv self.mode = "image" self.auto = auto self.transforms = transforms # optional self.vid_stride = vid_stride # video frame-rate stride if any(videos): self._new_video(videos[0]) # new video else: self.cap = None assert self.nf > 0, ( f"No images or videos found in {p}. Supported formats are:\nimages: {IMG_FORMATS}\nvideos: {VID_FORMATS}" ) def __iter__(self): self.count = 0 return self def __next__(self): if self.count == self.nf: raise StopIteration path = self.files[self.count] if self.video_flag[self.count]: # Read video self.mode = "video" for _ in range(self.vid_stride): self.cap.grab() ret_val, im0 = self.cap.retrieve() while not ret_val: self.count += 1 self.cap.release() if self.count == self.nf: # last video raise StopIteration path = self.files[self.count] self._new_video(path) ret_val, im0 = self.cap.read() self.frame += 1 # im0 = self._cv2_rotate(im0) # for use if cv2 autorotation is False s = f"video {self.count + 1}/{self.nf} ({self.frame}/{self.frames}) {path}: " else: # Read image self.count += 1 im0 = cv2.imread(path) # BGR assert im0 is not None, f"Image Not Found {path}" s = f"image {self.count}/{self.nf} {path}: " if self.transforms: im = self.transforms(im0) # transforms else: im = letterbox(im0, self.img_size, stride=self.stride, auto=self.auto)[0] # padded resize im = im.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB im = np.ascontiguousarray(im) # contiguous return path, im, im0, self.cap, s def _new_video(self, path): self.frame = 0 self.cap = cv2.VideoCapture(path) self.frames = int(self.cap.get(cv2.CAP_PROP_FRAME_COUNT) / self.vid_stride) self.orientation = int(self.cap.get(cv2.CAP_PROP_ORIENTATION_META)) # rotation degrees # self.cap.set(cv2.CAP_PROP_ORIENTATION_AUTO, 0) # disable https://github.com/ultralytics/yolov5/issues/8493 def _cv2_rotate(self, im): if self.orientation == 0: return cv2.rotate(im, cv2.ROTATE_90_CLOCKWISE) elif self.orientation == 180: return cv2.rotate(im, cv2.ROTATE_90_COUNTERCLOCKWISE) elif self.orientation == 90: return cv2.rotate(im, cv2.ROTATE_180) return im def __len__(self): return self.nf # number of files class LoadStreams: def __init__(self, sources="file.streams", img_size=640, stride=32, auto=True, transforms=None, vid_stride=1): torch.backends.cudnn.benchmark = True # faster for fixed-size inference self.mode = "stream" self.img_size = img_size self.stride = stride self.vid_stride = vid_stride # video frame-rate stride sources = Path(sources).read_text().rsplit() if os.path.isfile(sources) else [sources] n = len(sources) self.sources = [clean_str(x) for x in sources] # clean source names for later self.imgs, self.fps, self.frames, self.threads = [None] * n, [0] * n, [0] * n, [None] * n for i, s in enumerate(sources): # index, source # Start thread to read frames from video stream st = f"{i + 1}/{n}: {s}... " if urlparse(s).hostname in ("www.youtube.com", "youtube.com", "youtu.be"): # if source is YouTube video # YouTube format i.e. 'https://www.youtube.com/watch?v=Zgi9g1ksQHc' or 'https://youtu.be/LNwODJXcvt4' check_requirements(("pafy", "youtube_dl==2020.12.2")) import pafy s = pafy.new(s).getbest(preftype="mp4").url # YouTube URL s = eval(s) if s.isnumeric() else s # i.e. s = '0' local webcam if s == 0: assert not is_colab(), "--source 0 webcam unsupported on Colab. Rerun command in a local environment." assert not is_kaggle(), "--source 0 webcam unsupported on Kaggle. Rerun command in a local environment." cap = cv2.VideoCapture(s) assert cap.isOpened(), f"{st}Failed to open {s}" w = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) h = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) fps = cap.get(cv2.CAP_PROP_FPS) # warning: may return 0 or nan self.frames[i] = max(int(cap.get(cv2.CAP_PROP_FRAME_COUNT)), 0) or float("inf") # infinite stream fallback self.fps[i] = max((fps if math.isfinite(fps) else 0) % 100, 0) or 30 # 30 FPS fallback _, self.imgs[i] = cap.read() # guarantee first frame self.threads[i] = Thread(target=self.update, args=([i, cap, s]), daemon=True) LOGGER.info(f"{st} Success ({self.frames[i]} frames {w}x{h} at {self.fps[i]:.2f} FPS)") self.threads[i].start() LOGGER.info("") # newline # check for common shapes s = np.stack([letterbox(x, img_size, stride=stride, auto=auto)[0].shape for x in self.imgs]) self.rect = np.unique(s, axis=0).shape[0] == 1 # rect inference if all shapes equal self.auto = auto and self.rect self.transforms = transforms # optional if not self.rect: LOGGER.warning("WARNING ⚠️ Stream shapes differ. For optimal performance supply similarly-shaped streams.") def update(self, i, cap, stream): n, f = 0, self.frames[i] # frame number, frame array while cap.isOpened() and n < f: n += 1 cap.grab() # .read() = .grab() followed by .retrieve() if n % self.vid_stride == 0: success, im = cap.retrieve() if success: self.imgs[i] = im else: LOGGER.warning("WARNING ⚠️ Video stream unresponsive, please check your IP camera connection.") self.imgs[i] = np.zeros_like(self.imgs[i]) cap.open(stream) # re-open stream if signal was lost time.sleep(0.0) # wait time def __iter__(self): self.count = -1 return self def __next__(self): self.count += 1 if not all(x.is_alive() for x in self.threads) or cv2.waitKey(1) == ord("q"): # q to quit cv2.destroyAllWindows() raise StopIteration im0 = self.imgs.copy() if self.transforms: im = np.stack([self.transforms(x) for x in im0]) # transforms else: im = np.stack([letterbox(x, self.img_size, stride=self.stride, auto=self.auto)[0] for x in im0]) # resize im = im[..., ::-1].transpose((0, 3, 1, 2)) # BGR to RGB, BHWC to BCHW im = np.ascontiguousarray(im) # contiguous return self.sources, im, im0, None, "" def __len__(self): return len(self.sources) # 1E12 frames = 32 streams at 30 FPS for 30 years def img2label_paths(img_paths): sa, sb = f"{os.sep}images{os.sep}", f"{os.sep}labels{os.sep}" # /images/, /labels/ substrings return [sb.join(x.rsplit(sa, 1)).rsplit(".", 1)[0] + ".txt" for x in img_paths] class LoadImagesAndLabels(Dataset): cache_version = 0.6 # dataset labels *.cache version rand_interp_methods = [cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4] def __init__( self, path, img_size=640, batch_size=16, augment=False, hyp=None, rect=False, image_weights=False, cache_images=False, single_cls=False, stride=32, pad=0.0, min_items=0, prefix="", rank=-1, seed=0, ): self.img_size = img_size self.augment = augment self.hyp = hyp self.image_weights = image_weights self.rect = False if image_weights else rect self.mosaic = self.augment and not self.rect # load 4 images at a time into a mosaic (only during training) self.mosaic_border = [-img_size // 2, -img_size // 2] self.stride = stride self.path = path self.albumentations = Albumentations(size=img_size) if augment else None try: f = [] # image files for p in path if isinstance(path, list) else [path]: p = Path(p) # os-agnostic if p.is_dir(): # dir f += glob.glob(str(p / "**" / "*.*"), recursive=True) # f = list(p.rglob('*.*')) # pathlib elif p.is_file(): # file with open(p) as t: t = t.read().strip().splitlines() parent = str(p.parent) + os.sep f += [x.replace("./", parent, 1) if x.startswith("./") else x for x in t] # to global path # f += [p.parent / x.lstrip(os.sep) for x in t] # to global path (pathlib) else: raise FileNotFoundError(f"{prefix}{p} does not exist") self.im_files = sorted(x.replace("/", os.sep) for x in f if x.split(".")[-1].lower() in IMG_FORMATS) # self.img_files = sorted([x for x in f if x.suffix[1:].lower() in IMG_FORMATS]) # pathlib assert self.im_files, f"{prefix}No images found" except Exception as e: raise Exception(f"{prefix}Error loading data from {path}: {e}\n{HELP_URL}") from e # Check cache self.label_files = img2label_paths(self.im_files) # labels cache_path = (p if p.is_file() else Path(self.label_files[0]).parent).with_suffix(".cache") try: cache, exists = np.load(cache_path, allow_pickle=True).item(), True # load dict assert cache["version"] == self.cache_version # matches current version assert cache["hash"] == get_hash(self.label_files + self.im_files) # identical hash except Exception: cache, exists = self.cache_labels(cache_path, prefix), False # run cache ops # Display cache nf, nm, ne, nc, n = cache.pop("results") # found, missing, empty, corrupt, total if exists and LOCAL_RANK in {-1, 0}: d = f"Scanning {cache_path}... {nf} images, {nm + ne} backgrounds, {nc} corrupt" tqdm(None, desc=prefix + d, total=n, initial=n, bar_format=TQDM_BAR_FORMAT) # display cache results if cache["msgs"]: LOGGER.info("\n".join(cache["msgs"])) # display warnings assert nf > 0 or not augment, f"{prefix}No labels found in {cache_path}, can not start training. {HELP_URL}" # Read cache [cache.pop(k) for k in ("hash", "version", "msgs")] # remove items labels, shapes, self.segments = zip(*cache.values()) nl = len(np.concatenate(labels, 0)) # number of labels assert nl > 0 or not augment, f"{prefix}All labels empty in {cache_path}, can not start training. {HELP_URL}" self.labels = list(labels) self.shapes = np.array(shapes) self.im_files = list(cache.keys()) # update self.label_files = img2label_paths(cache.keys()) # update # Filter images if min_items: include = np.array([len(x) >= min_items for x in self.labels]).nonzero()[0].astype(int) LOGGER.info(f"{prefix}{n - len(include)}/{n} images filtered from dataset") self.im_files = [self.im_files[i] for i in include] self.label_files = [self.label_files[i] for i in include] self.labels = [self.labels[i] for i in include] self.segments = [self.segments[i] for i in include] self.shapes = self.shapes[include] # wh # Create indices n = len(self.shapes) # number of images bi = np.floor(np.arange(n) / batch_size).astype(int) # batch index nb = bi[-1] + 1 # number of batches self.batch = bi # batch index of image self.n = n self.indices = np.arange(n) if rank > -1: # DDP indices (see: SmartDistributedSampler) # force each rank (i.e. GPU process) to sample the same subset of data on every epoch self.indices = self.indices[np.random.RandomState(seed=seed).permutation(n) % WORLD_SIZE == RANK] # Update labels include_class = [] # filter labels to include only these classes (optional) self.segments = list(self.segments) include_class_array = np.array(include_class).reshape(1, -1) for i, (label, segment) in enumerate(zip(self.labels, self.segments)): if include_class: j = (label[:, 0:1] == include_class_array).any(1) self.labels[i] = label[j] if segment: self.segments[i] = [segment[idx] for idx, elem in enumerate(j) if elem] if single_cls: # single-class training, merge all classes into 0 self.labels[i][:, 0] = 0 # Rectangular Training if self.rect: # Sort by aspect ratio s = self.shapes # wh ar = s[:, 1] / s[:, 0] # aspect ratio irect = ar.argsort() self.im_files = [self.im_files[i] for i in irect] self.label_files = [self.label_files[i] for i in irect] self.labels = [self.labels[i] for i in irect] self.segments = [self.segments[i] for i in irect] self.shapes = s[irect] # wh ar = ar[irect] # Set training image shapes shapes = [[1, 1]] * nb for i in range(nb): ari = ar[bi == i] mini, maxi = ari.min(), ari.max() if maxi < 1: shapes[i] = [maxi, 1] elif mini > 1: shapes[i] = [1, 1 / mini] self.batch_shapes = np.ceil(np.array(shapes) * img_size / stride + pad).astype(int) * stride # Cache images into RAM/disk for faster training if cache_images == "ram" and not self.check_cache_ram(prefix=prefix): cache_images = False self.ims = [None] * n self.npy_files = [Path(f).with_suffix(".npy") for f in self.im_files] if cache_images: b, gb = 0, 1 << 30 # bytes of cached images, bytes per gigabytes self.im_hw0, self.im_hw = [None] * n, [None] * n fcn = self.cache_images_to_disk if cache_images == "disk" else self.load_image with ThreadPool(NUM_THREADS) as pool: results = pool.imap(lambda i: (i, fcn(i)), self.indices) pbar = tqdm(results, total=len(self.indices), bar_format=TQDM_BAR_FORMAT, disable=LOCAL_RANK > 0) for i, x in pbar: if cache_images == "disk": b += self.npy_files[i].stat().st_size else: # 'ram' self.ims[i], self.im_hw0[i], self.im_hw[i] = x # im, hw_orig, hw_resized = load_image(self, i) b += self.ims[i].nbytes * WORLD_SIZE pbar.desc = f"{prefix}Caching images ({b / gb:.1f}GB {cache_images})" pbar.close() def check_cache_ram(self, safety_margin=0.1, prefix=""): b, gb = 0, 1 << 30 # bytes of cached images, bytes per gigabytes n = min(self.n, 30) # extrapolate from 30 random images for _ in range(n): im = cv2.imread(random.choice(self.im_files)) # sample image ratio = self.img_size / max(im.shape[0], im.shape[1]) # max(h, w) # ratio b += im.nbytes * ratio**2 mem_required = b * self.n / n # GB required to cache dataset into RAM mem = psutil.virtual_memory() cache = mem_required * (1 + safety_margin) < mem.available # to cache or not to cache, that is the question if not cache: LOGGER.info( f"{prefix}{mem_required / gb:.1f}GB RAM required, " f"{mem.available / gb:.1f}/{mem.total / gb:.1f}GB available, " f"{'caching images ✅' if cache else 'not caching images ⚠️'}" ) return cache def cache_labels(self, path=Path("./labels.cache"), prefix=""): x = {} # dict nm, nf, ne, nc, msgs = 0, 0, 0, 0, [] # number missing, found, empty, corrupt, messages desc = f"{prefix}Scanning {path.parent / path.stem}..." with Pool(NUM_THREADS) as pool: pbar = tqdm( pool.imap(verify_image_label, zip(self.im_files, self.label_files, repeat(prefix))), desc=desc, total=len(self.im_files), bar_format=TQDM_BAR_FORMAT, ) for im_file, lb, shape, segments, nm_f, nf_f, ne_f, nc_f, msg in pbar: nm += nm_f nf += nf_f ne += ne_f nc += nc_f if im_file: x[im_file] = [lb, shape, segments] if msg: msgs.append(msg) pbar.desc = f"{desc} {nf} images, {nm + ne} backgrounds, {nc} corrupt" pbar.close() if msgs: LOGGER.info("\n".join(msgs)) if nf == 0: LOGGER.warning(f"{prefix}WARNING ⚠️ No labels found in {path}. {HELP_URL}") x["hash"] = get_hash(self.label_files + self.im_files) x["results"] = nf, nm, ne, nc, len(self.im_files) x["msgs"] = msgs # warnings x["version"] = self.cache_version # cache version try: np.save(path, x) # save cache for next time path.with_suffix(".cache.npy").rename(path) # remove .npy suffix LOGGER.info(f"{prefix}New cache created: {path}") except Exception as e: LOGGER.warning(f"{prefix}WARNING ⚠️ Cache directory {path.parent} is not writeable: {e}") # not writeable return x def __len__(self): return len(self.im_files) # def __iter__(self): # self.count = -1 # print('ran dataset iter') # #self.shuffled_vector = np.random.permutation(self.nF) if self.augment else np.arange(self.nF) # return self def __getitem__(self, index): index = self.indices[index] # linear, shuffled, or image_weights hyp = self.hyp if mosaic := self.mosaic and random.random() < hyp["mosaic"]: # Load mosaic img, labels = self.load_mosaic(index) shapes = None # MixUp augmentation if random.random() < hyp["mixup"]: img, labels = mixup(img, labels, *self.load_mosaic(random.choice(self.indices))) else: # Load image img, (h0, w0), (h, w) = self.load_image(index) # Letterbox shape = self.batch_shapes[self.batch[index]] if self.rect else self.img_size # final letterboxed shape img, ratio, pad = letterbox(img, shape, auto=False, scaleup=self.augment) shapes = (h0, w0), ((h / h0, w / w0), pad) # for COCO mAP rescaling labels = self.labels[index].copy() if labels.size: # normalized xywh to pixel xyxy format labels[:, 1:] = xywhn2xyxy(labels[:, 1:], ratio[0] * w, ratio[1] * h, padw=pad[0], padh=pad[1]) if self.augment: img, labels = random_perspective( img, labels, degrees=hyp["degrees"], translate=hyp["translate"], scale=hyp["scale"], shear=hyp["shear"], perspective=hyp["perspective"], ) nl = len(labels) # number of labels if nl: labels[:, 1:5] = xyxy2xywhn(labels[:, 1:5], w=img.shape[1], h=img.shape[0], clip=True, eps=1e-3) if self.augment: # Albumentations img, labels = self.albumentations(img, labels) nl = len(labels) # update after albumentations # HSV color-space augment_hsv(img, hgain=hyp["hsv_h"], sgain=hyp["hsv_s"], vgain=hyp["hsv_v"]) # Flip up-down if random.random() < hyp["flipud"]: img = np.flipud(img) if nl: labels[:, 2] = 1 - labels[:, 2] # Flip left-right if random.random() < hyp["fliplr"]: img = np.fliplr(img) if nl: labels[:, 1] = 1 - labels[:, 1] # Cutouts # labels = cutout(img, labels, p=0.5) # nl = len(labels) # update after cutout labels_out = torch.zeros((nl, 6)) if nl: labels_out[:, 1:] = torch.from_numpy(labels) # Convert img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB img = np.ascontiguousarray(img) return torch.from_numpy(img), labels_out, self.im_files[index], shapes def load_image(self, i): im, f, fn = ( self.ims[i], self.im_files[i], self.npy_files[i], ) if im is None: # not cached in RAM if fn.exists(): # load npy im = np.load(fn) else: # read image im = cv2.imread(f) # BGR assert im is not None, f"Image Not Found {f}" h0, w0 = im.shape[:2] # orig hw r = self.img_size / max(h0, w0) # ratio if r != 1: # if sizes are not equal interp = cv2.INTER_LINEAR if (self.augment or r > 1) else cv2.INTER_AREA im = cv2.resize(im, (math.ceil(w0 * r), math.ceil(h0 * r)), interpolation=interp) return im, (h0, w0), im.shape[:2] # im, hw_original, hw_resized return self.ims[i], self.im_hw0[i], self.im_hw[i] # im, hw_original, hw_resized def cache_images_to_disk(self, i): f = self.npy_files[i] if not f.exists(): np.save(f.as_posix(), cv2.imread(self.im_files[i])) def load_mosaic(self, index): labels4, segments4 = [], [] s = self.img_size yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y indices = [index] + random.choices(self.indices, k=3) # 3 additional image indices random.shuffle(indices) for i, index in enumerate(indices): # Load image img, _, (h, w) = self.load_image(index) # place img in img4 if i == 0: # top left img4 = np.full((s * 2, s * 2, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles x1a, y1a, x2a, y2a = max(xc - w, 0), max(yc - h, 0), xc, yc # xmin, ymin, xmax, ymax (large image) x1b, y1b, x2b, y2b = w - (x2a - x1a), h - (y2a - y1a), w, h # xmin, ymin, xmax, ymax (small image) elif i == 1: # top right x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h elif i == 2: # bottom left x1a, y1a, x2a, y2a = max(xc - w, 0), yc, xc, min(s * 2, yc + h) x1b, y1b, x2b, y2b = w - (x2a - x1a), 0, w, min(y2a - y1a, h) elif i == 3: # bottom right x1a, y1a, x2a, y2a = xc, yc, min(xc + w, s * 2), min(s * 2, yc + h) x1b, y1b, x2b, y2b = 0, 0, min(w, x2a - x1a), min(y2a - y1a, h) img4[y1a:y2a, x1a:x2a] = img[y1b:y2b, x1b:x2b] # img4[ymin:ymax, xmin:xmax] padw = x1a - x1b padh = y1a - y1b # Labels labels, segments = self.labels[index].copy(), self.segments[index].copy() if labels.size: labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padw, padh) # normalized xywh to pixel xyxy format segments = [xyn2xy(x, w, h, padw, padh) for x in segments] labels4.append(labels) segments4.extend(segments) # Concat/clip labels labels4 = np.concatenate(labels4, 0) for x in (labels4[:, 1:], *segments4): np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective() # img4, labels4 = replicate(img4, labels4) # replicate # Augment img4, labels4, segments4 = copy_paste(img4, labels4, segments4, p=self.hyp["copy_paste"]) img4, labels4 = random_perspective( img4, labels4, segments4, degrees=self.hyp["degrees"], translate=self.hyp["translate"], scale=self.hyp["scale"], shear=self.hyp["shear"], perspective=self.hyp["perspective"], border=self.mosaic_border, ) # border to remove return img4, labels4 def load_mosaic9(self, index): labels9, segments9 = [], [] s = self.img_size indices = [index] + random.choices(self.indices, k=8) # 8 additional image indices random.shuffle(indices) hp, wp = -1, -1 # height, width previous for i, index in enumerate(indices): # Load image img, _, (h, w) = self.load_image(index) # place img in img9 if i == 0: # center img9 = np.full((s * 3, s * 3, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles h0, w0 = h, w c = s, s, s + w, s + h # xmin, ymin, xmax, ymax (base) coordinates elif i == 1: # top c = s, s - h, s + w, s elif i == 2: # top right c = s + wp, s - h, s + wp + w, s elif i == 3: # right c = s + w0, s, s + w0 + w, s + h elif i == 4: # bottom right c = s + w0, s + hp, s + w0 + w, s + hp + h elif i == 5: # bottom c = s + w0 - w, s + h0, s + w0, s + h0 + h elif i == 6: # bottom left c = s + w0 - wp - w, s + h0, s + w0 - wp, s + h0 + h elif i == 7: # left c = s - w, s + h0 - h, s, s + h0 elif i == 8: # top left c = s - w, s + h0 - hp - h, s, s + h0 - hp padx, pady = c[:2] x1, y1, x2, y2 = (max(x, 0) for x in c) # allocate coords # Labels labels, segments = self.labels[index].copy(), self.segments[index].copy() if labels.size: labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padx, pady) # normalized xywh to pixel xyxy format segments = [xyn2xy(x, w, h, padx, pady) for x in segments] labels9.append(labels) segments9.extend(segments) # Image img9[y1:y2, x1:x2] = img[y1 - pady :, x1 - padx :] # img9[ymin:ymax, xmin:xmax] hp, wp = h, w # height, width previous # Offset yc, xc = (int(random.uniform(0, s)) for _ in self.mosaic_border) # mosaic center x, y img9 = img9[yc : yc + 2 * s, xc : xc + 2 * s] # Concat/clip labels labels9 = np.concatenate(labels9, 0) labels9[:, [1, 3]] -= xc labels9[:, [2, 4]] -= yc c = np.array([xc, yc]) # centers segments9 = [x - c for x in segments9] for x in (labels9[:, 1:], *segments9): np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective() # img9, labels9 = replicate(img9, labels9) # replicate # Augment img9, labels9, segments9 = copy_paste(img9, labels9, segments9, p=self.hyp["copy_paste"]) img9, labels9 = random_perspective( img9, labels9, segments9, degrees=self.hyp["degrees"], translate=self.hyp["translate"], scale=self.hyp["scale"], shear=self.hyp["shear"], perspective=self.hyp["perspective"], border=self.mosaic_border, ) # border to remove return img9, labels9 @staticmethod def collate_fn(batch): im, label, path, shapes = zip(*batch) # transposed for i, lb in enumerate(label): lb[:, 0] = i # add target image index for build_targets() return torch.stack(im, 0), torch.cat(label, 0), path, shapes @staticmethod def collate_fn4(batch): im, label, path, shapes = zip(*batch) # transposed n = len(shapes) // 4 im4, label4, path4, shapes4 = [], [], path[:n], shapes[:n] ho = torch.tensor([[0.0, 0, 0, 1, 0, 0]]) wo = torch.tensor([[0.0, 0, 1, 0, 0, 0]]) s = torch.tensor([[1, 1, 0.5, 0.5, 0.5, 0.5]]) # scale for i in range(n): # zidane torch.zeros(16,3,720,1280) # BCHW i *= 4 if random.random() < 0.5: im1 = F.interpolate(im[i].unsqueeze(0).float(), scale_factor=2.0, mode="bilinear", align_corners=False)[ 0 ].type(im[i].type()) lb = label[i] else: im1 = torch.cat((torch.cat((im[i], im[i + 1]), 1), torch.cat((im[i + 2], im[i + 3]), 1)), 2) lb = torch.cat((label[i], label[i + 1] + ho, label[i + 2] + wo, label[i + 3] + ho + wo), 0) * s im4.append(im1) label4.append(lb) for i, lb in enumerate(label4): lb[:, 0] = i # add target image index for build_targets() return torch.stack(im4, 0), torch.cat(label4, 0), path4, shapes4 # Ancillary functions -------------------------------------------------------------------------------------------------- def flatten_recursive(path=DATASETS_DIR / "coco128"): new_path = Path(f"{path!s}_flat") if os.path.exists(new_path): shutil.rmtree(new_path) # delete output folder os.makedirs(new_path) # make new output folder for file in tqdm(glob.glob(f"{Path(path)!s}/**/*.*", recursive=True)): shutil.copyfile(file, new_path / Path(file).name) def extract_boxes(path=DATASETS_DIR / "coco128"): path = Path(path) # images dir shutil.rmtree(path / "classification") if (path / "classification").is_dir() else None # remove existing files = list(path.rglob("*.*")) n = len(files) # number of files for im_file in tqdm(files, total=n): if im_file.suffix[1:] in IMG_FORMATS: # image im = cv2.imread(str(im_file))[..., ::-1] # BGR to RGB h, w = im.shape[:2] # labels lb_file = Path(img2label_paths([str(im_file)])[0]) if Path(lb_file).exists(): with open(lb_file) as f: lb = np.array([x.split() for x in f.read().strip().splitlines()], dtype=np.float32) # labels for j, x in enumerate(lb): c = int(x[0]) # class f = (path / "classification") / f"{c}" / f"{path.stem}_{im_file.stem}_{j}.jpg" # new filename if not f.parent.is_dir(): f.parent.mkdir(parents=True) b = x[1:] * [w, h, w, h] # box # b[2:] = b[2:].max() # rectangle to square b[2:] = b[2:] * 1.2 + 3 # pad b = xywh2xyxy(b.reshape(-1, 4)).ravel().astype(int) b[[0, 2]] = np.clip(b[[0, 2]], 0, w) # clip boxes outside of image b[[1, 3]] = np.clip(b[[1, 3]], 0, h) assert cv2.imwrite(str(f), im[b[1] : b[3], b[0] : b[2]]), f"box failure in {f}" def autosplit(path=DATASETS_DIR / "coco128/images", weights=(0.9, 0.1, 0.0), annotated_only=False): path = Path(path) # images dir files = sorted(x for x in path.rglob("*.*") if x.suffix[1:].lower() in IMG_FORMATS) # image files only n = len(files) # number of files random.seed(0) # for reproducibility indices = random.choices([0, 1, 2], weights=weights, k=n) # assign each image to a split txt = ["autosplit_train.txt", "autosplit_val.txt", "autosplit_test.txt"] # 3 txt files for x in txt: if (path.parent / x).exists(): (path.parent / x).unlink() # remove existing print(f"Autosplitting images from {path}" + ", using *.txt labeled images only" * annotated_only) for i, img in tqdm(zip(indices, files), total=n): if not annotated_only or Path(img2label_paths([str(img)])[0]).exists(): # check label with open(path.parent / txt[i], "a") as f: f.write(f"./{img.relative_to(path.parent).as_posix()}" + "\n") # add image to txt file def verify_image_label(args): im_file, lb_file, prefix = args nm, nf, ne, nc, msg, segments = 0, 0, 0, 0, "", [] # number (missing, found, empty, corrupt), message, segments try: # verify images im = Image.open(im_file) im.verify() # PIL verify shape = exif_size(im) # image size assert (shape[0] > 9) & (shape[1] > 9), f"image size {shape} <10 pixels" assert im.format.lower() in IMG_FORMATS, f"invalid image format {im.format}" if im.format.lower() in ("jpg", "jpeg"): with open(im_file, "rb") as f: f.seek(-2, 2) if f.read() != b"\xff\xd9": # corrupt JPEG ImageOps.exif_transpose(Image.open(im_file)).save(im_file, "JPEG", subsampling=0, quality=100) msg = f"{prefix}WARNING ⚠️ {im_file}: corrupt JPEG restored and saved" # verify labels if os.path.isfile(lb_file): nf = 1 # label found with open(lb_file) as f: lb = [x.split() for x in f.read().strip().splitlines() if len(x)] if any(len(x) > 6 for x in lb): # is segment classes = np.array([x[0] for x in lb], dtype=np.float32) segments = [np.array(x[1:], dtype=np.float32).reshape(-1, 2) for x in lb] # (cls, xy1...) lb = np.concatenate((classes.reshape(-1, 1), segments2boxes(segments)), 1) # (cls, xywh) lb = np.array(lb, dtype=np.float32) if nl := len(lb): assert lb.shape[1] == 5, f"labels require 5 columns, {lb.shape[1]} columns detected" assert (lb >= 0).all(), f"negative label values {lb[lb < 0]}" assert (lb[:, 1:] <= 1).all(), f"non-normalized or out of bounds coordinates {lb[:, 1:][lb[:, 1:] > 1]}" _, i = np.unique(lb, axis=0, return_index=True) if len(i) < nl: # duplicate row check lb = lb[i] # remove duplicates if segments: segments = [segments[x] for x in i] msg = f"{prefix}WARNING ⚠️ {im_file}: {nl - len(i)} duplicate labels removed" else: ne = 1 # label empty lb = np.zeros((0, 5), dtype=np.float32) else: nm = 1 # label missing lb = np.zeros((0, 5), dtype=np.float32) return im_file, lb, shape, segments, nm, nf, ne, nc, msg except Exception as e: nc = 1 msg = f"{prefix}WARNING ⚠️ {im_file}: ignoring corrupt image/label: {e}" return [None, None, None, None, nm, nf, ne, nc, msg] class HUBDatasetStats: def __init__(self, path="coco128.yaml", autodownload=False): zipped, data_dir, yaml_path = self._unzip(Path(path)) try: with open(check_yaml(yaml_path), errors="ignore") as f: data = yaml.safe_load(f) # data dict if zipped: data["path"] = data_dir except Exception as e: raise Exception("error/HUB/dataset_stats/yaml_load") from e check_dataset(data, autodownload) # download dataset if missing self.hub_dir = Path(data["path"] + "-hub") self.im_dir = self.hub_dir / "images" self.im_dir.mkdir(parents=True, exist_ok=True) # makes /images self.stats = {"nc": data["nc"], "names": list(data["names"].values())} # statistics dictionary self.data = data @staticmethod def _find_yaml(dir): files = list(dir.glob("*.yaml")) or list(dir.rglob("*.yaml")) # try root level first and then recursive assert files, f"No *.yaml file found in {dir}" if len(files) > 1: files = [f for f in files if f.stem == dir.stem] # prefer *.yaml files that match dir name assert files, f"Multiple *.yaml files found in {dir}, only 1 *.yaml file allowed" assert len(files) == 1, f"Multiple *.yaml files found: {files}, only 1 *.yaml file allowed in {dir}" return files[0] def _unzip(self, path): if not str(path).endswith(".zip"): # path is data.yaml return False, None, path assert Path(path).is_file(), f"Error unzipping {path}, file not found" unzip_file(path, path=path.parent) dir = path.with_suffix("") # dataset directory == zip name assert dir.is_dir(), f"Error unzipping {path}, {dir} not found. path/to/abc.zip MUST unzip to path/to/abc/" return True, str(dir), self._find_yaml(dir) # zipped, data_dir, yaml_path def _hub_ops(self, f, max_dim=1920): f_new = self.im_dir / Path(f).name # dataset-hub image filename try: # use PIL im = Image.open(f) r = max_dim / max(im.height, im.width) # ratio if r < 1.0: # image too large im = im.resize((int(im.width * r), int(im.height * r))) im.save(f_new, "JPEG", quality=50, optimize=True) # save except Exception as e: # use OpenCV LOGGER.info(f"WARNING ⚠️ HUB ops PIL failure {f}: {e}") im = cv2.imread(f) im_height, im_width = im.shape[:2] r = max_dim / max(im_height, im_width) # ratio if r < 1.0: # image too large im = cv2.resize(im, (int(im_width * r), int(im_height * r)), interpolation=cv2.INTER_AREA) cv2.imwrite(str(f_new), im) def get_json(self, save=False, verbose=False): def _round(labels): return [[int(c), *(round(x, 4) for x in points)] for c, *points in labels] for split in "train", "val", "test": if self.data.get(split) is None: self.stats[split] = None # i.e. no test set continue dataset = LoadImagesAndLabels(self.data[split]) # load dataset x = np.array( [ np.bincount(label[:, 0].astype(int), minlength=self.data["nc"]) for label in tqdm(dataset.labels, total=dataset.n, desc="Statistics") ] ) # shape(128x80) self.stats[split] = { "instance_stats": {"total": int(x.sum()), "per_class": x.sum(0).tolist()}, "image_stats": { "total": dataset.n, "unlabelled": int(np.all(x == 0, 1).sum()), "per_class": (x > 0).sum(0).tolist(), }, "labels": [{str(Path(k).name): _round(v.tolist())} for k, v in zip(dataset.im_files, dataset.labels)], } # Save, print and return if save: stats_path = self.hub_dir / "stats.json" print(f"Saving {stats_path.resolve()}...") with open(stats_path, "w") as f: json.dump(self.stats, f) # save stats.json if verbose: print(json.dumps(self.stats, indent=2, sort_keys=False)) return self.stats def process_images(self): for split in "train", "val", "test": if self.data.get(split) is None: continue dataset = LoadImagesAndLabels(self.data[split]) # load dataset desc = f"{split} images" for _ in tqdm(ThreadPool(NUM_THREADS).imap(self._hub_ops, dataset.im_files), total=dataset.n, desc=desc): pass print(f"Done. All images saved to {self.im_dir}") return self.im_dir # Classification dataloaders ------------------------------------------------------------------------------------------- class ClassificationDataset(torchvision.datasets.ImageFolder): def __init__(self, root, augment, imgsz, cache=False): super().__init__(root=root) self.torch_transforms = classify_transforms(imgsz) self.album_transforms = classify_albumentations(augment, imgsz) if augment else None self.cache_ram = cache is True or cache == "ram" self.cache_disk = cache == "disk" self.samples = [[*list(x), Path(x[0]).with_suffix(".npy"), None] for x in self.samples] # file, index, npy, im def __getitem__(self, i): f, j, fn, im = self.samples[i] # filename, index, filename.with_suffix('.npy'), image if self.cache_ram and im is None: im = self.samples[i][3] = cv2.imread(f) elif self.cache_disk: if not fn.exists(): # load npy np.save(fn.as_posix(), cv2.imread(f)) im = np.load(fn) else: # read image im = cv2.imread(f) # BGR if self.album_transforms: sample = self.album_transforms(image=cv2.cvtColor(im, cv2.COLOR_BGR2RGB))["image"] else: sample = self.torch_transforms(im) return sample, j def create_classification_dataloader( path, imgsz=224, batch_size=16, augment=True, cache=False, rank=-1, workers=8, shuffle=True ): # Returns Dataloader object to be used with YOLOv5 Classifier with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP dataset = ClassificationDataset(root=path, imgsz=imgsz, augment=augment, cache=cache) batch_size = min(batch_size, len(dataset)) nd = torch.cuda.device_count() nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers]) sampler = None if rank == -1 else distributed.DistributedSampler(dataset, shuffle=shuffle) generator = torch.Generator() generator.manual_seed(6148914691236517205 + RANK) return InfiniteDataLoader( dataset, batch_size=batch_size, shuffle=shuffle and sampler is None, num_workers=nw, sampler=sampler, pin_memory=PIN_MEMORY, worker_init_fn=seed_worker, generator=generator, ) # or DataLoader(persistent_workers=True)
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Dataloaders and dataset utils.""" import contextlib import glob @@ -72,6 +73,7 @@ def get_hash(paths): + """Generates a single SHA256 hash for a list of file or directory paths by combining their sizes and paths.""" size = sum(os.path.getsize(p) for p in paths if os.path.exists(p)) # sizes h = hashlib.sha256(str(size).encode()) # hash sizes h.update("".join(paths).encode()) # hash paths @@ -79,6 +81,7 @@ def exif_size(img): + """Returns corrected PIL image size (width, height) considering EXIF orientation.""" s = img.size # (width, height) with contextlib.suppress(Exception): rotation = dict(img._getexif().items())[orientation] @@ -88,6 +91,13 @@ def exif_transpose(image): + """ + Transpose a PIL image accordingly if it has an EXIF Orientation tag. + Inplace version of https://github.com/python-pillow/Pillow/blob/master/src/PIL/ImageOps.py exif_transpose(). + + :param image: The image to transpose. + :return: An image. + """ exif = image.getexif() orientation = exif.get(0x0112, 1) # default 1 if orientation > 1: @@ -108,6 +118,10 @@ def seed_worker(worker_id): + """Sets the seed for a dataloader worker to ensure reproducibility, based on PyTorch's randomness notes. + + See https://pytorch.org/docs/stable/notes/randomness.html#dataloader. + """ worker_seed = torch.initial_seed() % 2**32 np.random.seed(worker_seed) random.seed(worker_seed) @@ -116,8 +130,10 @@ # Inherit from DistributedSampler and override iterator # https://github.com/pytorch/pytorch/blob/master/torch/utils/data/distributed.py class SmartDistributedSampler(distributed.DistributedSampler): + """A distributed sampler ensuring deterministic shuffling and balanced data distribution across GPUs.""" def __iter__(self): + """Yields indices for distributed data sampling, shuffled deterministically based on epoch and seed.""" g = torch.Generator() g.manual_seed(self.seed + self.epoch) @@ -159,6 +175,7 @@ shuffle=False, seed=0, ): + """Creates and returns a configured DataLoader instance for loading and processing image datasets.""" if rect and shuffle: LOGGER.warning("WARNING ⚠️ --rect is incompatible with DataLoader shuffle, setting shuffle=False") shuffle = False @@ -201,33 +218,55 @@ class InfiniteDataLoader(dataloader.DataLoader): + """Dataloader that reuses workers. + + Uses same syntax as vanilla DataLoader + """ def __init__(self, *args, **kwargs): + """Initializes an InfiniteDataLoader that reuses workers with standard DataLoader syntax, augmenting with a + repeating sampler. + """ super().__init__(*args, **kwargs) object.__setattr__(self, "batch_sampler", _RepeatSampler(self.batch_sampler)) self.iterator = super().__iter__() def __len__(self): + """Returns the length of the batch sampler's sampler in the InfiniteDataLoader.""" return len(self.batch_sampler.sampler) def __iter__(self): + """Yields batches of data indefinitely in a loop by resetting the sampler when exhausted.""" for _ in range(len(self)): yield next(self.iterator) class _RepeatSampler: + """Sampler that repeats forever. + + Args: + sampler (Sampler) + """ def __init__(self, sampler): + """Initializes a perpetual sampler wrapping a provided `Sampler` instance for endless data iteration.""" self.sampler = sampler def __iter__(self): + """Returns an infinite iterator over the dataset by repeatedly yielding from the given sampler.""" while True: yield from iter(self.sampler) class LoadScreenshots: + """Loads and processes screenshots for YOLOv5 detection from specified screen regions using mss.""" def __init__(self, source, img_size=640, stride=32, auto=True, transforms=None): + """Initializes a screenshot dataloader for YOLOv5 with specified source region, image size, stride, auto, and + transforms. + + Source = [screen_number left top width height] (pixels) + """ check_requirements("mss") import mss @@ -256,9 +295,13 @@ self.monitor = {"left": self.left, "top": self.top, "width": self.width, "height": self.height} def __iter__(self): + """Iterates over itself, enabling use in loops and iterable contexts.""" return self def __next__(self): + """Captures and returns the next screen frame as a BGR numpy array, cropping to only the first three channels + from BGRA. + """ im0 = np.array(self.sct.grab(self.monitor))[:, :, :3] # [:, :, :3] BGRA to BGR s = f"screen {self.screen} (LTWH): {self.left},{self.top},{self.width},{self.height}: " @@ -273,8 +316,10 @@ class LoadImages: + """YOLOv5 image/video dataloader, i.e. `python detect.py --source image.jpg/vid.mp4`.""" def __init__(self, path, img_size=640, stride=32, auto=True, transforms=None, vid_stride=1): + """Initializes YOLOv5 loader for images/videos, supporting glob patterns, directories, and lists of paths.""" if isinstance(path, str) and Path(path).suffix == ".txt": # *.txt file with img/vid/dir on each line path = Path(path).read_text().rsplit() files = [] @@ -311,10 +356,12 @@ ) def __iter__(self): + """Initializes iterator by resetting count and returns the iterator object itself.""" self.count = 0 return self def __next__(self): + """Advances to the next file in the dataset, raising StopIteration if at the end.""" if self.count == self.nf: raise StopIteration path = self.files[self.count] @@ -355,6 +402,7 @@ return path, im, im0, self.cap, s def _new_video(self, path): + """Initialize a new video capture object with path, frame count adjusted by stride, and orientation metadata.""" self.frame = 0 self.cap = cv2.VideoCapture(path) self.frames = int(self.cap.get(cv2.CAP_PROP_FRAME_COUNT) / self.vid_stride) @@ -362,6 +410,7 @@ # self.cap.set(cv2.CAP_PROP_ORIENTATION_AUTO, 0) # disable https://github.com/ultralytics/yolov5/issues/8493 def _cv2_rotate(self, im): + """Rotates a cv2 image based on its orientation; supports 0, 90, and 180 degrees rotations.""" if self.orientation == 0: return cv2.rotate(im, cv2.ROTATE_90_CLOCKWISE) elif self.orientation == 180: @@ -371,12 +420,17 @@ return im def __len__(self): + """Returns the number of files in the dataset.""" return self.nf # number of files class LoadStreams: + """Loads and processes video streams for YOLOv5, supporting various sources including YouTube and IP cameras.""" def __init__(self, sources="file.streams", img_size=640, stride=32, auto=True, transforms=None, vid_stride=1): + """Initializes a stream loader for processing video streams with YOLOv5, supporting various sources including + YouTube. + """ torch.backends.cudnn.benchmark = True # faster for fixed-size inference self.mode = "stream" self.img_size = img_size @@ -422,6 +476,7 @@ LOGGER.warning("WARNING ⚠️ Stream shapes differ. For optimal performance supply similarly-shaped streams.") def update(self, i, cap, stream): + """Reads frames from stream `i`, updating imgs array; handles stream reopening on signal loss.""" n, f = 0, self.frames[i] # frame number, frame array while cap.isOpened() and n < f: n += 1 @@ -437,10 +492,14 @@ time.sleep(0.0) # wait time def __iter__(self): + """Resets and returns the iterator for iterating over video frames or images in a dataset.""" self.count = -1 return self def __next__(self): + """Iterates over video frames or images, halting on thread stop or 'q' key press, raising `StopIteration` when + done. + """ self.count += 1 if not all(x.is_alive() for x in self.threads) or cv2.waitKey(1) == ord("q"): # q to quit cv2.destroyAllWindows() @@ -457,15 +516,20 @@ return self.sources, im, im0, None, "" def __len__(self): + """Returns the number of sources in the dataset, supporting up to 32 streams at 30 FPS over 30 years.""" return len(self.sources) # 1E12 frames = 32 streams at 30 FPS for 30 years def img2label_paths(img_paths): + """Generates label file paths from corresponding image file paths by replacing `/images/` with `/labels/` and + extension with `.txt`. + """ sa, sb = f"{os.sep}images{os.sep}", f"{os.sep}labels{os.sep}" # /images/, /labels/ substrings return [sb.join(x.rsplit(sa, 1)).rsplit(".", 1)[0] + ".txt" for x in img_paths] class LoadImagesAndLabels(Dataset): + """Loads images and their corresponding labels for training and validation in YOLOv5.""" cache_version = 0.6 # dataset labels *.cache version rand_interp_methods = [cv2.INTER_NEAREST, cv2.INTER_LINEAR, cv2.INTER_CUBIC, cv2.INTER_AREA, cv2.INTER_LANCZOS4] @@ -488,6 +552,7 @@ rank=-1, seed=0, ): + """Initializes the YOLOv5 dataset loader, handling images and their labels, caching, and preprocessing.""" self.img_size = img_size self.augment = augment self.hyp = hyp @@ -630,6 +695,7 @@ pbar.close() def check_cache_ram(self, safety_margin=0.1, prefix=""): + """Checks if available RAM is sufficient for caching images, adjusting for a safety margin.""" b, gb = 0, 1 << 30 # bytes of cached images, bytes per gigabytes n = min(self.n, 30) # extrapolate from 30 random images for _ in range(n): @@ -648,6 +714,7 @@ return cache def cache_labels(self, path=Path("./labels.cache"), prefix=""): + """Caches dataset labels, verifies images, reads shapes, and tracks dataset integrity.""" x = {} # dict nm, nf, ne, nc, msgs = 0, 0, 0, 0, [] # number missing, found, empty, corrupt, messages desc = f"{prefix}Scanning {path.parent / path.stem}..." @@ -687,6 +754,7 @@ return x def __len__(self): + """Returns the number of images in the dataset.""" return len(self.im_files) # def __iter__(self): @@ -696,6 +764,7 @@ # return self def __getitem__(self, index): + """Fetches the dataset item at the given index, considering linear, shuffled, or weighted sampling.""" index = self.indices[index] # linear, shuffled, or image_weights hyp = self.hyp @@ -771,6 +840,10 @@ return torch.from_numpy(img), labels_out, self.im_files[index], shapes def load_image(self, i): + """Loads an image by index, returning the image, its original dimensions, and resized dimensions. + + Returns (im, original hw, resized hw) + """ im, f, fn = ( self.ims[i], self.im_files[i], @@ -791,11 +864,13 @@ return self.ims[i], self.im_hw0[i], self.im_hw[i] # im, hw_original, hw_resized def cache_images_to_disk(self, i): + """Saves an image to disk as an *.npy file for quicker loading, identified by index `i`.""" f = self.npy_files[i] if not f.exists(): np.save(f.as_posix(), cv2.imread(self.im_files[i])) def load_mosaic(self, index): + """Loads a 4-image mosaic for YOLOv5, combining 1 selected and 3 random images, with labels and segments.""" labels4, segments4 = [], [] s = self.img_size yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y @@ -855,6 +930,9 @@ return img4, labels4 def load_mosaic9(self, index): + """Loads 1 image + 8 random images into a 9-image mosaic for augmented YOLOv5 training, returning labels and + segments. + """ labels9, segments9 = [], [] s = self.img_size indices = [index] + random.choices(self.indices, k=8) # 8 additional image indices @@ -934,6 +1012,7 @@ @staticmethod def collate_fn(batch): + """Batches images, labels, paths, and shapes, assigning unique indices to targets in merged label tensor.""" im, label, path, shapes = zip(*batch) # transposed for i, lb in enumerate(label): lb[:, 0] = i # add target image index for build_targets() @@ -941,6 +1020,7 @@ @staticmethod def collate_fn4(batch): + """Bundles a batch's data by quartering the number of shapes and paths, preparing it for model input.""" im, label, path, shapes = zip(*batch) # transposed n = len(shapes) // 4 im4, label4, path4, shapes4 = [], [], path[:n], shapes[:n] @@ -969,6 +1049,7 @@ # Ancillary functions -------------------------------------------------------------------------------------------------- def flatten_recursive(path=DATASETS_DIR / "coco128"): + """Flatten a directory by copying all files from subdirs to a new top-level directory, preserving filenames.""" new_path = Path(f"{path!s}_flat") if os.path.exists(new_path): shutil.rmtree(new_path) # delete output folder @@ -978,6 +1059,11 @@ def extract_boxes(path=DATASETS_DIR / "coco128"): + """Converts a detection dataset to a classification dataset, creating a directory for each class and extracting + bounding boxes. + + Example: from utils.dataloaders import *; extract_boxes() + """ path = Path(path) # images dir shutil.rmtree(path / "classification") if (path / "classification").is_dir() else None # remove existing files = list(path.rglob("*.*")) @@ -1011,6 +1097,14 @@ def autosplit(path=DATASETS_DIR / "coco128/images", weights=(0.9, 0.1, 0.0), annotated_only=False): + """Autosplit a dataset into train/val/test splits and save path/autosplit_*.txt files Usage: from utils.dataloaders + import *; autosplit(). + + Args: + path: Path to images directory + weights: Train, val, test weights (list, tuple) + annotated_only: Only use images with an annotated txt file + """ path = Path(path) # images dir files = sorted(x for x in path.rglob("*.*") if x.suffix[1:].lower() in IMG_FORMATS) # image files only n = len(files) # number of files @@ -1030,6 +1124,7 @@ def verify_image_label(args): + """Verifies a single image-label pair, ensuring image format, size, and legal label values.""" im_file, lb_file, prefix = args nm, nf, ne, nc, msg, segments = 0, 0, 0, 0, "", [] # number (missing, found, empty, corrupt), message, segments try: @@ -1080,8 +1175,24 @@ class HUBDatasetStats: + """Class for generating HUB dataset JSON and `-hub` dataset directory. + + Args: + path: Path to data.yaml or data.zip (with data.yaml inside data.zip) + autodownload: Attempt to download dataset if not found locally + + Usage + from utils.dataloaders import HUBDatasetStats + stats = HUBDatasetStats('coco128.yaml', autodownload=True) # usage 1 + stats = HUBDatasetStats('path/to/coco128.zip') # usage 2 + stats.get_json(save=False) + stats.process_images() + """ def __init__(self, path="coco128.yaml", autodownload=False): + """Initializes HUBDatasetStats with optional auto-download for datasets, given a path to dataset YAML or ZIP + file. + """ zipped, data_dir, yaml_path = self._unzip(Path(path)) try: with open(check_yaml(yaml_path), errors="ignore") as f: @@ -1100,6 +1211,9 @@ @staticmethod def _find_yaml(dir): + """Finds and returns the path to a single '.yaml' file in the specified directory, preferring files that match + the directory name. + """ files = list(dir.glob("*.yaml")) or list(dir.rglob("*.yaml")) # try root level first and then recursive assert files, f"No *.yaml file found in {dir}" if len(files) > 1: @@ -1109,6 +1223,7 @@ return files[0] def _unzip(self, path): + """Unzips a .zip file at 'path', returning success status, unzipped directory, and path to YAML file within.""" if not str(path).endswith(".zip"): # path is data.yaml return False, None, path assert Path(path).is_file(), f"Error unzipping {path}, file not found" @@ -1118,6 +1233,7 @@ return True, str(dir), self._find_yaml(dir) # zipped, data_dir, yaml_path def _hub_ops(self, f, max_dim=1920): + """Resizes and saves an image at reduced quality for web/app viewing, supporting both PIL and OpenCV.""" f_new = self.im_dir / Path(f).name # dataset-hub image filename try: # use PIL im = Image.open(f) @@ -1135,8 +1251,10 @@ cv2.imwrite(str(f_new), im) def get_json(self, save=False, verbose=False): + """Generates dataset JSON for Ultralytics Platform, optionally saves or prints it; save=bool, verbose=bool.""" def _round(labels): + """Rounds class labels to integers and coordinates to 4 decimal places for improved label accuracy.""" return [[int(c), *(round(x, 4) for x in points)] for c, *points in labels] for split in "train", "val", "test": @@ -1171,6 +1289,7 @@ return self.stats def process_images(self): + """Compress images across 'train', 'val', 'test' splits and saves to specified directory.""" for split in "train", "val", "test": if self.data.get(split) is None: continue @@ -1184,8 +1303,18 @@ # Classification dataloaders ------------------------------------------------------------------------------------------- class ClassificationDataset(torchvision.datasets.ImageFolder): + """YOLOv5 Classification Dataset. + + Args: + root: Dataset path + transform: torchvision transforms, used by default + album_transform: Albumentations transforms, used if installed + """ def __init__(self, root, augment, imgsz, cache=False): + """Initializes YOLOv5 Classification Dataset with optional caching, augmentations, and transforms for image + classification. + """ super().__init__(root=root) self.torch_transforms = classify_transforms(imgsz) self.album_transforms = classify_albumentations(augment, imgsz) if augment else None @@ -1194,6 +1323,7 @@ self.samples = [[*list(x), Path(x[0]).with_suffix(".npy"), None] for x in self.samples] # file, index, npy, im def __getitem__(self, i): + """Fetches and transforms an image sample by index, supporting RAM/disk caching and Augmentations.""" f, j, fn, im = self.samples[i] # filename, index, filename.with_suffix('.npy'), image if self.cache_ram and im is None: im = self.samples[i][3] = cv2.imread(f) @@ -1214,6 +1344,7 @@ path, imgsz=224, batch_size=16, augment=True, cache=False, rank=-1, workers=8, shuffle=True ): # Returns Dataloader object to be used with YOLOv5 Classifier + """Creates a DataLoader for image classification, supporting caching, augmentation, and distributed training.""" with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP dataset = ClassificationDataset(root=path, imgsz=imgsz, augment=augment, cache=cache) batch_size = min(batch_size, len(dataset)) @@ -1231,4 +1362,4 @@ pin_memory=PIN_MEMORY, worker_init_fn=seed_worker, generator=generator, - ) # or DataLoader(persistent_workers=True)+ ) # or DataLoader(persistent_workers=True)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/dataloaders.py
Generate missing documentation strings
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import math import random import cv2 import numpy as np import torch import torchvision.transforms as T import torchvision.transforms.functional as TF from utils.general import LOGGER, check_version, colorstr, resample_segments, segment2box, xywhn2xyxy from utils.metrics import bbox_ioa IMAGENET_MEAN = 0.485, 0.456, 0.406 # RGB mean IMAGENET_STD = 0.229, 0.224, 0.225 # RGB standard deviation class Albumentations: def __init__(self, size=640): self.transform = None prefix = colorstr("albumentations: ") try: import albumentations as A check_version(A.__version__, "1.0.3", hard=True) # version requirement T = [ A.RandomResizedCrop(height=size, width=size, scale=(0.8, 1.0), ratio=(0.9, 1.11), p=0.0), A.Blur(p=0.01), A.MedianBlur(p=0.01), A.ToGray(p=0.01), A.CLAHE(p=0.01), A.RandomBrightnessContrast(p=0.0), A.RandomGamma(p=0.0), A.ImageCompression(quality_lower=75, p=0.0), ] # transforms self.transform = A.Compose(T, bbox_params=A.BboxParams(format="yolo", label_fields=["class_labels"])) LOGGER.info(prefix + ", ".join(f"{x}".replace("always_apply=False, ", "") for x in T if x.p)) except ImportError: # package not installed, skip pass except Exception as e: LOGGER.info(f"{prefix}{e}") def __call__(self, im, labels, p=1.0): if self.transform and random.random() < p: new = self.transform(image=im, bboxes=labels[:, 1:], class_labels=labels[:, 0]) # transformed im, labels = new["image"], np.array([[c, *b] for c, b in zip(new["class_labels"], new["bboxes"])]) return im, labels def normalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD, inplace=False): return TF.normalize(x, mean, std, inplace=inplace) def denormalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD): for i in range(3): x[:, i] = x[:, i] * std[i] + mean[i] return x def augment_hsv(im, hgain=0.5, sgain=0.5, vgain=0.5): if hgain or sgain or vgain: r = np.random.uniform(-1, 1, 3) * [hgain, sgain, vgain] + 1 # random gains hue, sat, val = cv2.split(cv2.cvtColor(im, cv2.COLOR_BGR2HSV)) dtype = im.dtype # uint8 x = np.arange(0, 256, dtype=r.dtype) lut_hue = ((x * r[0]) % 180).astype(dtype) lut_sat = np.clip(x * r[1], 0, 255).astype(dtype) lut_val = np.clip(x * r[2], 0, 255).astype(dtype) im_hsv = cv2.merge((cv2.LUT(hue, lut_hue), cv2.LUT(sat, lut_sat), cv2.LUT(val, lut_val))) cv2.cvtColor(im_hsv, cv2.COLOR_HSV2BGR, dst=im) # no return needed def hist_equalize(im, clahe=True, bgr=False): yuv = cv2.cvtColor(im, cv2.COLOR_BGR2YUV if bgr else cv2.COLOR_RGB2YUV) if clahe: c = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8)) yuv[:, :, 0] = c.apply(yuv[:, :, 0]) else: yuv[:, :, 0] = cv2.equalizeHist(yuv[:, :, 0]) # equalize Y channel histogram return cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR if bgr else cv2.COLOR_YUV2RGB) # convert YUV image to RGB def replicate(im, labels): h, w = im.shape[:2] boxes = labels[:, 1:].astype(int) x1, y1, x2, y2 = boxes.T s = ((x2 - x1) + (y2 - y1)) / 2 # side length (pixels) for i in s.argsort()[: round(s.size * 0.5)]: # smallest indices x1b, y1b, x2b, y2b = boxes[i] bh, bw = y2b - y1b, x2b - x1b yc, xc = int(random.uniform(0, h - bh)), int(random.uniform(0, w - bw)) # offset x, y x1a, y1a, x2a, y2a = [xc, yc, xc + bw, yc + bh] im[y1a:y2a, x1a:x2a] = im[y1b:y2b, x1b:x2b] # im4[ymin:ymax, xmin:xmax] labels = np.append(labels, [[labels[i, 0], x1a, y1a, x2a, y2a]], axis=0) return im, labels def letterbox(im, new_shape=(640, 640), color=(114, 114, 114), auto=True, scaleFill=False, scaleup=True, stride=32): shape = im.shape[:2] # current shape [height, width] if isinstance(new_shape, int): new_shape = (new_shape, new_shape) # Scale ratio (new / old) r = min(new_shape[0] / shape[0], new_shape[1] / shape[1]) if not scaleup: # only scale down, do not scale up (for better val mAP) r = min(r, 1.0) # Compute padding ratio = r, r # width, height ratios new_unpad = round(shape[1] * r), round(shape[0] * r) dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1] # wh padding if auto: # minimum rectangle dw, dh = np.mod(dw, stride), np.mod(dh, stride) # wh padding elif scaleFill: # stretch dw, dh = 0.0, 0.0 new_unpad = (new_shape[1], new_shape[0]) ratio = new_shape[1] / shape[1], new_shape[0] / shape[0] # width, height ratios dw /= 2 # divide padding into 2 sides dh /= 2 if shape[::-1] != new_unpad: # resize im = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR) top, bottom = round(dh - 0.1), round(dh + 0.1) left, right = round(dw - 0.1), round(dw + 0.1) im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color) # add border return im, ratio, (dw, dh) def random_perspective( im, targets=(), segments=(), degrees=10, translate=0.1, scale=0.1, shear=10, perspective=0.0, border=(0, 0) ): # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(0.1, 0.1), scale=(0.9, 1.1), shear=(-10, 10)) # targets = [cls, xyxy] height = im.shape[0] + border[0] * 2 # shape(h,w,c) width = im.shape[1] + border[1] * 2 # Center C = np.eye(3) C[0, 2] = -im.shape[1] / 2 # x translation (pixels) C[1, 2] = -im.shape[0] / 2 # y translation (pixels) # Perspective P = np.eye(3) P[2, 0] = random.uniform(-perspective, perspective) # x perspective (about y) P[2, 1] = random.uniform(-perspective, perspective) # y perspective (about x) # Rotation and Scale R = np.eye(3) a = random.uniform(-degrees, degrees) # a += random.choice([-180, -90, 0, 90]) # add 90deg rotations to small rotations s = random.uniform(1 - scale, 1 + scale) # s = 2 ** random.uniform(-scale, scale) R[:2] = cv2.getRotationMatrix2D(angle=a, center=(0, 0), scale=s) # Shear S = np.eye(3) S[0, 1] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # x shear (deg) S[1, 0] = math.tan(random.uniform(-shear, shear) * math.pi / 180) # y shear (deg) # Translation T = np.eye(3) T[0, 2] = random.uniform(0.5 - translate, 0.5 + translate) * width # x translation (pixels) T[1, 2] = random.uniform(0.5 - translate, 0.5 + translate) * height # y translation (pixels) # Combined rotation matrix M = T @ S @ R @ P @ C # order of operations (right to left) is IMPORTANT if (border[0] != 0) or (border[1] != 0) or (M != np.eye(3)).any(): # image changed if perspective: im = cv2.warpPerspective(im, M, dsize=(width, height), borderValue=(114, 114, 114)) else: # affine im = cv2.warpAffine(im, M[:2], dsize=(width, height), borderValue=(114, 114, 114)) if n := len(targets): use_segments = any(x.any() for x in segments) and len(segments) == n new = np.zeros((n, 4)) if use_segments: # warp segments segments = resample_segments(segments) # upsample for i, segment in enumerate(segments): xy = np.ones((len(segment), 3)) xy[:, :2] = segment xy = xy @ M.T # transform xy = xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2] # perspective rescale or affine # clip new[i] = segment2box(xy, width, height) else: # warp boxes xy = np.ones((n * 4, 3)) xy[:, :2] = targets[:, [1, 2, 3, 4, 1, 4, 3, 2]].reshape(n * 4, 2) # x1y1, x2y2, x1y2, x2y1 xy = xy @ M.T # transform xy = (xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2]).reshape(n, 8) # perspective rescale or affine # create new boxes x = xy[:, [0, 2, 4, 6]] y = xy[:, [1, 3, 5, 7]] new = np.concatenate((x.min(1), y.min(1), x.max(1), y.max(1))).reshape(4, n).T # clip new[:, [0, 2]] = new[:, [0, 2]].clip(0, width) new[:, [1, 3]] = new[:, [1, 3]].clip(0, height) # filter candidates i = box_candidates(box1=targets[:, 1:5].T * s, box2=new.T, area_thr=0.01 if use_segments else 0.10) targets = targets[i] targets[:, 1:5] = new[i] return im, targets def copy_paste(im, labels, segments, p=0.5): n = len(segments) if p and n: _h, w, _c = im.shape # height, width, channels im_new = np.zeros(im.shape, np.uint8) for j in random.sample(range(n), k=round(p * n)): l, s = labels[j], segments[j] box = w - l[3], l[2], w - l[1], l[4] ioa = bbox_ioa(box, labels[:, 1:5]) # intersection over area if (ioa < 0.30).all(): # allow 30% obscuration of existing labels labels = np.concatenate((labels, [[l[0], *box]]), 0) segments.append(np.concatenate((w - s[:, 0:1], s[:, 1:2]), 1)) cv2.drawContours(im_new, [segments[j].astype(np.int32)], -1, (1, 1, 1), cv2.FILLED) result = cv2.flip(im, 1) # augment segments (flip left-right) i = cv2.flip(im_new, 1).astype(bool) im[i] = result[i] # cv2.imwrite('debug.jpg', im) # debug return im, labels, segments def cutout(im, labels, p=0.5): if random.random() < p: h, w = im.shape[:2] scales = [0.5] * 1 + [0.25] * 2 + [0.125] * 4 + [0.0625] * 8 + [0.03125] * 16 # image size fraction for s in scales: mask_h = random.randint(1, int(h * s)) # create random masks mask_w = random.randint(1, int(w * s)) # box xmin = max(0, random.randint(0, w) - mask_w // 2) ymin = max(0, random.randint(0, h) - mask_h // 2) xmax = min(w, xmin + mask_w) ymax = min(h, ymin + mask_h) # apply random color mask im[ymin:ymax, xmin:xmax] = [random.randint(64, 191) for _ in range(3)] # return unobscured labels if len(labels) and s > 0.03: box = np.array([xmin, ymin, xmax, ymax], dtype=np.float32) ioa = bbox_ioa(box, xywhn2xyxy(labels[:, 1:5], w, h)) # intersection over area labels = labels[ioa < 0.60] # remove >60% obscured labels return labels def mixup(im, labels, im2, labels2): r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0 im = (im * r + im2 * (1 - r)).astype(np.uint8) labels = np.concatenate((labels, labels2), 0) return im, labels def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16): w1, h1 = box1[2] - box1[0], box1[3] - box1[1] w2, h2 = box2[2] - box2[0], box2[3] - box2[1] ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps)) # aspect ratio return (w2 > wh_thr) & (h2 > wh_thr) & (w2 * h2 / (w1 * h1 + eps) > area_thr) & (ar < ar_thr) # candidates def classify_albumentations( augment=True, size=224, scale=(0.08, 1.0), ratio=(0.75, 1.0 / 0.75), # 0.75, 1.33 hflip=0.5, vflip=0.0, jitter=0.4, mean=IMAGENET_MEAN, std=IMAGENET_STD, auto_aug=False, ): # YOLOv5 classification Albumentations (optional, only used if package is installed) prefix = colorstr("albumentations: ") try: import albumentations as A from albumentations.pytorch import ToTensorV2 check_version(A.__version__, "1.0.3", hard=True) # version requirement if augment: # Resize and crop T = [A.RandomResizedCrop(height=size, width=size, scale=scale, ratio=ratio)] if auto_aug: # TODO: implement AugMix, AutoAug & RandAug in albumentation LOGGER.info(f"{prefix}auto augmentations are currently not supported") else: if hflip > 0: T += [A.HorizontalFlip(p=hflip)] if vflip > 0: T += [A.VerticalFlip(p=vflip)] if jitter > 0: color_jitter = (float(jitter),) * 3 # repeat value for brightness, contrast, saturation, 0 hue T += [A.ColorJitter(*color_jitter, 0)] else: # Use fixed crop for eval set (reproducibility) T = [A.SmallestMaxSize(max_size=size), A.CenterCrop(height=size, width=size)] T += [A.Normalize(mean=mean, std=std), ToTensorV2()] # Normalize and convert to Tensor LOGGER.info(prefix + ", ".join(f"{x}".replace("always_apply=False, ", "") for x in T if x.p)) return A.Compose(T) except ImportError: # package not installed, skip LOGGER.warning(f"{prefix}⚠️ not found, install with `pip install albumentations` (recommended)") except Exception as e: LOGGER.info(f"{prefix}{e}") def classify_transforms(size=224): assert isinstance(size, int), f"ERROR: classify_transforms size {size} must be integer, not (list, tuple)" # T.Compose([T.ToTensor(), T.Resize(size), T.CenterCrop(size), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)]) return T.Compose([CenterCrop(size), ToTensor(), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)]) class LetterBox: def __init__(self, size=(640, 640), auto=False, stride=32): super().__init__() self.h, self.w = (size, size) if isinstance(size, int) else size self.auto = auto # pass max size integer, automatically solve for short side using stride self.stride = stride # used with auto def __call__(self, im): imh, imw = im.shape[:2] r = min(self.h / imh, self.w / imw) # ratio of new/old h, w = round(imh * r), round(imw * r) # resized image hs, ws = (math.ceil(x / self.stride) * self.stride for x in (h, w)) if self.auto else self.h, self.w top, left = round((hs - h) / 2 - 0.1), round((ws - w) / 2 - 0.1) im_out = np.full((self.h, self.w, 3), 114, dtype=im.dtype) im_out[top : top + h, left : left + w] = cv2.resize(im, (w, h), interpolation=cv2.INTER_LINEAR) return im_out class CenterCrop: def __init__(self, size=640): super().__init__() self.h, self.w = (size, size) if isinstance(size, int) else size def __call__(self, im): imh, imw = im.shape[:2] m = min(imh, imw) # min dimension top, left = (imh - m) // 2, (imw - m) // 2 return cv2.resize(im[top : top + m, left : left + m], (self.w, self.h), interpolation=cv2.INTER_LINEAR) class ToTensor: def __init__(self, half=False): super().__init__() self.half = half def __call__(self, im): im = np.ascontiguousarray(im.transpose((2, 0, 1))[::-1]) # HWC to CHW -> BGR to RGB -> contiguous im = torch.from_numpy(im) # to torch im = im.half() if self.half else im.float() # uint8 to fp16/32 im /= 255.0 # 0-255 to 0.0-1.0 return im
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Image augmentation functions.""" import math import random @@ -17,8 +18,10 @@ class Albumentations: + """Provides optional data augmentation for YOLOv5 using Albumentations library if installed.""" def __init__(self, size=640): + """Initializes Albumentations class for optional data augmentation in YOLOv5 with specified input size.""" self.transform = None prefix = colorstr("albumentations: ") try: @@ -45,6 +48,7 @@ LOGGER.info(f"{prefix}{e}") def __call__(self, im, labels, p=1.0): + """Applies transformations to an image and labels with probability `p`, returning updated image and labels.""" if self.transform and random.random() < p: new = self.transform(image=im, bboxes=labels[:, 1:], class_labels=labels[:, 0]) # transformed im, labels = new["image"], np.array([[c, *b] for c, b in zip(new["class_labels"], new["bboxes"])]) @@ -52,16 +56,22 @@ def normalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD, inplace=False): + """Applies ImageNet normalization to RGB images in BCHW format, modifying them in-place if specified. + + Example: y = (x - mean) / std + """ return TF.normalize(x, mean, std, inplace=inplace) def denormalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD): + """Reverses ImageNet normalization for BCHW format RGB images by applying `x = x * std + mean`.""" for i in range(3): x[:, i] = x[:, i] * std[i] + mean[i] return x def augment_hsv(im, hgain=0.5, sgain=0.5, vgain=0.5): + """Applies HSV color-space augmentation to an image with random gains for hue, saturation, and value.""" if hgain or sgain or vgain: r = np.random.uniform(-1, 1, 3) * [hgain, sgain, vgain] + 1 # random gains hue, sat, val = cv2.split(cv2.cvtColor(im, cv2.COLOR_BGR2HSV)) @@ -77,6 +87,7 @@ def hist_equalize(im, clahe=True, bgr=False): + """Equalizes image histogram, with optional CLAHE, for BGR or RGB image with shape (n,m,3) and range 0-255.""" yuv = cv2.cvtColor(im, cv2.COLOR_BGR2YUV if bgr else cv2.COLOR_RGB2YUV) if clahe: c = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8)) @@ -87,6 +98,10 @@ def replicate(im, labels): + """Replicates half of the smallest object labels in an image for data augmentation. + + Returns augmented image and labels. + """ h, w = im.shape[:2] boxes = labels[:, 1:].astype(int) x1, y1, x2, y2 = boxes.T @@ -103,6 +118,7 @@ def letterbox(im, new_shape=(640, 640), color=(114, 114, 114), auto=True, scaleFill=False, scaleup=True, stride=32): + """Resizes and pads image to new_shape with stride-multiple constraints, returns resized image, ratio, padding.""" shape = im.shape[:2] # current shape [height, width] if isinstance(new_shape, int): new_shape = (new_shape, new_shape) @@ -139,6 +155,7 @@ ): # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(0.1, 0.1), scale=(0.9, 1.1), shear=(-10, 10)) # targets = [cls, xyxy] + """Applies random perspective transformation to an image, modifying the image and corresponding labels.""" height = im.shape[0] + border[0] * 2 # shape(h,w,c) width = im.shape[1] + border[1] * 2 @@ -216,6 +233,10 @@ def copy_paste(im, labels, segments, p=0.5): + """Applies Copy-Paste augmentation by flipping and merging segments and labels on an image. + + Details at https://arxiv.org/abs/2012.07177. + """ n = len(segments) if p and n: _h, w, _c = im.shape # height, width, channels @@ -237,6 +258,10 @@ def cutout(im, labels, p=0.5): + """Applies cutout augmentation to an image with optional label adjustment, using random masks of varying sizes. + + Details at https://arxiv.org/abs/1708.04552. + """ if random.random() < p: h, w = im.shape[:2] scales = [0.5] * 1 + [0.25] * 2 + [0.125] * 4 + [0.0625] * 8 + [0.03125] * 16 # image size fraction @@ -263,6 +288,10 @@ def mixup(im, labels, im2, labels2): + """Applies MixUp augmentation by blending images and labels. + + See https://arxiv.org/pdf/1710.09412.pdf for details. + """ r = np.random.beta(32.0, 32.0) # mixup ratio, alpha=beta=32.0 im = (im * r + im2 * (1 - r)).astype(np.uint8) labels = np.concatenate((labels, labels2), 0) @@ -270,6 +299,11 @@ def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16): + """Filters bounding box candidates by minimum width-height threshold `wh_thr` (pixels), aspect ratio threshold + `ar_thr`, and area ratio threshold `area_thr`. + + box1(4,n) is before augmentation, box2(4,n) is after augmentation. + """ w1, h1 = box1[2] - box1[0], box1[3] - box1[1] w2, h2 = box2[2] - box2[0], box2[3] - box2[1] ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps)) # aspect ratio @@ -289,6 +323,7 @@ auto_aug=False, ): # YOLOv5 classification Albumentations (optional, only used if package is installed) + """Sets up Albumentations transforms for YOLOv5 classification tasks depending on augmentation settings.""" prefix = colorstr("albumentations: ") try: import albumentations as A @@ -321,20 +356,29 @@ def classify_transforms(size=224): + """Applies a series of transformations including center crop, ToTensor, and normalization for classification.""" assert isinstance(size, int), f"ERROR: classify_transforms size {size} must be integer, not (list, tuple)" # T.Compose([T.ToTensor(), T.Resize(size), T.CenterCrop(size), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)]) return T.Compose([CenterCrop(size), ToTensor(), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)]) class LetterBox: + """Resizes and pads images to specified dimensions while maintaining aspect ratio for YOLOv5 preprocessing.""" def __init__(self, size=(640, 640), auto=False, stride=32): + """Initializes a LetterBox object for YOLOv5 image preprocessing with optional auto sizing and stride + adjustment. + """ super().__init__() self.h, self.w = (size, size) if isinstance(size, int) else size self.auto = auto # pass max size integer, automatically solve for short side using stride self.stride = stride # used with auto def __call__(self, im): + """Resizes and pads input image `im` (HWC format) to specified dimensions, maintaining aspect ratio. + + im = np.array HWC + """ imh, imw = im.shape[:2] r = min(self.h / imh, self.w / imw) # ratio of new/old h, w = round(imh * r), round(imw * r) # resized image @@ -346,12 +390,18 @@ class CenterCrop: + """Applies center crop to an image, resizing it to the specified size while maintaining aspect ratio.""" def __init__(self, size=640): + """Initializes CenterCrop for image preprocessing, accepting single int or tuple for size, defaults to 640.""" super().__init__() self.h, self.w = (size, size) if isinstance(size, int) else size def __call__(self, im): + """Applies center crop to the input image and resizes it to a specified size, maintaining aspect ratio. + + im = np.array HWC + """ imh, imw = im.shape[:2] m = min(imh, imw) # min dimension top, left = (imh - m) // 2, (imw - m) // 2 @@ -359,14 +409,21 @@ class ToTensor: + """Converts BGR np.array image from HWC to RGB CHW format, normalizes to [0, 1], and supports FP16 if half=True.""" def __init__(self, half=False): + """Initializes ToTensor for YOLOv5 image preprocessing, with optional half precision (half=True for FP16).""" super().__init__() self.half = half def __call__(self, im): + """Converts BGR np.array image from HWC to RGB CHW format, and normalizes to [0, 1], with support for FP16 if + `half=True`. + + im = np.array HWC in BGR order + """ im = np.ascontiguousarray(im.transpose((2, 0, 1))[::-1]) # HWC to CHW -> BGR to RGB -> contiguous im = torch.from_numpy(im) # to torch im = im.half() if self.half else im.float() # uint8 to fp16/32 im /= 255.0 # 0-255 to 0.0-1.0 - return im+ return im
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/augmentations.py
Add docstrings to improve code quality
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import logging import os from urllib.parse import urlparse try: import comet_ml except ImportError: comet_ml = None import yaml logger = logging.getLogger(__name__) COMET_PREFIX = "comet://" COMET_MODEL_NAME = os.getenv("COMET_MODEL_NAME", "yolov5") COMET_DEFAULT_CHECKPOINT_FILENAME = os.getenv("COMET_DEFAULT_CHECKPOINT_FILENAME", "last.pt") def download_model_checkpoint(opt, experiment): model_dir = f"{opt.project}/{experiment.name}" os.makedirs(model_dir, exist_ok=True) model_name = COMET_MODEL_NAME model_asset_list = experiment.get_model_asset_list(model_name) if len(model_asset_list) == 0: logger.error(f"COMET ERROR: No checkpoints found for model name : {model_name}") return model_asset_list = sorted( model_asset_list, key=lambda x: x["step"], reverse=True, ) logged_checkpoint_map = {asset["fileName"]: asset["assetId"] for asset in model_asset_list} resource_url = urlparse(opt.weights) checkpoint_filename = resource_url.query if checkpoint_filename: asset_id = logged_checkpoint_map.get(checkpoint_filename) else: asset_id = logged_checkpoint_map.get(COMET_DEFAULT_CHECKPOINT_FILENAME) checkpoint_filename = COMET_DEFAULT_CHECKPOINT_FILENAME if asset_id is None: logger.error(f"COMET ERROR: Checkpoint {checkpoint_filename} not found in the given Experiment") return try: logger.info(f"COMET INFO: Downloading checkpoint {checkpoint_filename}") asset_filename = checkpoint_filename model_binary = experiment.get_asset(asset_id, return_type="binary", stream=False) model_download_path = f"{model_dir}/{asset_filename}" with open(model_download_path, "wb") as f: f.write(model_binary) opt.weights = model_download_path except Exception as e: logger.warning("COMET WARNING: Unable to download checkpoint from Comet") logger.exception(e) def set_opt_parameters(opt, experiment): asset_list = experiment.get_asset_list() resume_string = opt.resume for asset in asset_list: if asset["fileName"] == "opt.yaml": asset_id = asset["assetId"] asset_binary = experiment.get_asset(asset_id, return_type="binary", stream=False) opt_dict = yaml.safe_load(asset_binary) for key, value in opt_dict.items(): setattr(opt, key, value) opt.resume = resume_string # Save hyperparameters to YAML file # Necessary to pass checks in training script save_dir = f"{opt.project}/{experiment.name}" os.makedirs(save_dir, exist_ok=True) hyp_yaml_path = f"{save_dir}/hyp.yaml" with open(hyp_yaml_path, "w") as f: yaml.dump(opt.hyp, f) opt.hyp = hyp_yaml_path def check_comet_weights(opt): if comet_ml is None: return if isinstance(opt.weights, str) and opt.weights.startswith(COMET_PREFIX): api = comet_ml.API() resource = urlparse(opt.weights) experiment_path = f"{resource.netloc}{resource.path}" experiment = api.get(experiment_path) download_model_checkpoint(opt, experiment) return True return None def check_comet_resume(opt): if comet_ml is None: return if isinstance(opt.resume, str) and opt.resume.startswith(COMET_PREFIX): api = comet_ml.API() resource = urlparse(opt.resume) experiment_path = f"{resource.netloc}{resource.path}" experiment = api.get(experiment_path) set_opt_parameters(opt, experiment) download_model_checkpoint(opt, experiment) return True return None
--- +++ @@ -19,6 +19,7 @@ def download_model_checkpoint(opt, experiment): + """Downloads YOLOv5 model checkpoint from Comet ML experiment, updating `opt.weights` with download path.""" model_dir = f"{opt.project}/{experiment.name}" os.makedirs(model_dir, exist_ok=True) @@ -66,6 +67,12 @@ def set_opt_parameters(opt, experiment): + """Update the opts Namespace with parameters from Comet's ExistingExperiment when resuming a run. + + Args: + opt (argparse.Namespace): Namespace of command line options + experiment (comet_ml.APIExperiment): Comet API Experiment object + """ asset_list = experiment.get_asset_list() resume_string = opt.resume @@ -90,6 +97,14 @@ def check_comet_weights(opt): + """Downloads model weights from Comet and updates the weights path to point to saved weights location. + + Args: + opt (argparse.Namespace): Command Line arguments passed to YOLOv5 training script + + Returns: + None/bool: Return True if weights are successfully downloaded else return None + """ if comet_ml is None: return @@ -105,6 +120,14 @@ def check_comet_resume(opt): + """Restores run parameters to its original state based on the model checkpoint and logged Experiment parameters. + + Args: + opt (argparse.Namespace): Command Line arguments passed to YOLOv5 training script + + Returns: + None/bool: Return True if the run is restored successfully else return None + """ if comet_ml is None: return @@ -118,4 +141,4 @@ return True - return None+ return None
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/loggers/comet/comet_utils.py
Document classes and their methods
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import contextlib import math import os from copy import copy from pathlib import Path import cv2 import matplotlib import matplotlib.pyplot as plt import numpy as np import pandas as pd import seaborn as sn import torch from PIL import Image, ImageDraw from scipy.ndimage.filters import gaussian_filter1d from ultralytics.utils.plotting import Annotator from utils import TryExcept, threaded from utils.general import LOGGER, clip_boxes, increment_path, xywh2xyxy, xyxy2xywh from utils.metrics import fitness # Settings RANK = int(os.getenv("RANK", -1)) matplotlib.rc("font", **{"size": 11}) matplotlib.use("Agg") # for writing to files only class Colors: def __init__(self): hexs = ( "FF3838", "FF9D97", "FF701F", "FFB21D", "CFD231", "48F90A", "92CC17", "3DDB86", "1A9334", "00D4BB", "2C99A8", "00C2FF", "344593", "6473FF", "0018EC", "8438FF", "520085", "CB38FF", "FF95C8", "FF37C7", ) self.palette = [self.hex2rgb(f"#{c}") for c in hexs] self.n = len(self.palette) def __call__(self, i, bgr=False): c = self.palette[int(i) % self.n] return (c[2], c[1], c[0]) if bgr else c @staticmethod def hex2rgb(h): return tuple(int(h[1 + i : 1 + i + 2], 16) for i in (0, 2, 4)) colors = Colors() # create instance for 'from utils.plots import colors' def feature_visualization(x, module_type, stage, n=32, save_dir=Path("runs/detect/exp")): if ("Detect" not in module_type) and ( "Segment" not in module_type ): # 'Detect' for Object Detect task,'Segment' for Segment task _batch, channels, height, width = x.shape # batch, channels, height, width if height > 1 and width > 1: f = save_dir / f"stage{stage}_{module_type.split('.')[-1]}_features.png" # filename blocks = torch.chunk(x[0].cpu(), channels, dim=0) # select batch index 0, block by channels n = min(n, channels) # number of plots _fig, ax = plt.subplots(math.ceil(n / 8), 8, tight_layout=True) ax = ax.ravel() plt.subplots_adjust(wspace=0.05, hspace=0.05) for i in range(n): ax[i].imshow(blocks[i].squeeze()) # cmap='gray' ax[i].axis("off") LOGGER.info(f"Saving {f}... ({n}/{channels})") plt.savefig(f, dpi=300, bbox_inches="tight") plt.close() np.save(str(f.with_suffix(".npy")), x[0].cpu().numpy()) # npy save def hist2d(x, y, n=100): xedges, yedges = np.linspace(x.min(), x.max(), n), np.linspace(y.min(), y.max(), n) hist, xedges, yedges = np.histogram2d(x, y, (xedges, yedges)) xidx = np.clip(np.digitize(x, xedges) - 1, 0, hist.shape[0] - 1) yidx = np.clip(np.digitize(y, yedges) - 1, 0, hist.shape[1] - 1) return np.log(hist[xidx, yidx]) def butter_lowpass_filtfilt(data, cutoff=1500, fs=50000, order=5): from scipy.signal import butter, filtfilt # https://stackoverflow.com/questions/28536191/how-to-filter-smooth-with-scipy-numpy def butter_lowpass(cutoff, fs, order): nyq = 0.5 * fs normal_cutoff = cutoff / nyq return butter(order, normal_cutoff, btype="low", analog=False) b, a = butter_lowpass(cutoff, fs, order=order) return filtfilt(b, a, data) # forward-backward filter def output_to_target(output, max_det=300): targets = [] for i, o in enumerate(output): box, conf, cls = o[:max_det, :6].cpu().split((4, 1, 1), 1) j = torch.full((conf.shape[0], 1), i) targets.append(torch.cat((j, cls, xyxy2xywh(box), conf), 1)) return torch.cat(targets, 0).numpy() @threaded def plot_images(images, targets, paths=None, fname="images.jpg", names=None): if isinstance(images, torch.Tensor): images = images.cpu().float().numpy() if isinstance(targets, torch.Tensor): targets = targets.cpu().numpy() max_size = 1920 # max image size max_subplots = 16 # max image subplots, i.e. 4x4 bs, _, h, w = images.shape # batch size, _, height, width bs = min(bs, max_subplots) # limit plot images ns = np.ceil(bs**0.5) # number of subplots (square) if np.max(images[0]) <= 1: images *= 255 # de-normalise (optional) # Build Image mosaic = np.full((int(ns * h), int(ns * w), 3), 255, dtype=np.uint8) # init for i, im in enumerate(images): if i == max_subplots: # if last batch has fewer images than we expect break x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin im = im.transpose(1, 2, 0) mosaic[y : y + h, x : x + w, :] = im # Resize (optional) scale = max_size / ns / max(h, w) if scale < 1: h = math.ceil(scale * h) w = math.ceil(scale * w) mosaic = cv2.resize(mosaic, tuple(int(x * ns) for x in (w, h))) # Annotate fs = int((h + w) * ns * 0.01) # font size annotator = Annotator(mosaic, line_width=round(fs / 10), font_size=fs, pil=True, example=names) for i in range(bs): x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin annotator.rectangle([x, y, x + w, y + h], None, (255, 255, 255), width=2) # borders if paths: annotator.text([x + 5, y + 5], text=Path(paths[i]).name[:40], txt_color=(220, 220, 220)) # filenames if len(targets) > 0: ti = targets[targets[:, 0] == i] # image targets boxes = xywh2xyxy(ti[:, 2:6]).T classes = ti[:, 1].astype("int") labels = ti.shape[1] == 6 # labels if no conf column conf = None if labels else ti[:, 6] # check for confidence presence (label vs pred) if boxes.shape[1]: if boxes.max() <= 1.01: # if normalized with tolerance 0.01 boxes[[0, 2]] *= w # scale to pixels boxes[[1, 3]] *= h elif scale < 1: # absolute coords need scale if image scales boxes *= scale boxes[[0, 2]] += x boxes[[1, 3]] += y for j, box in enumerate(boxes.T.tolist()): cls = classes[j] color = colors(cls) cls = names[cls] if names else cls if labels or conf[j] > 0.25: # 0.25 conf thresh label = f"{cls}" if labels else f"{cls} {conf[j]:.1f}" annotator.box_label(box, label, color=color) annotator.im.save(fname) # save def plot_lr_scheduler(optimizer, scheduler, epochs=300, save_dir=""): optimizer, scheduler = copy(optimizer), copy(scheduler) # do not modify originals y = [] for _ in range(epochs): scheduler.step() y.append(optimizer.param_groups[0]["lr"]) plt.plot(y, ".-", label="LR") plt.xlabel("epoch") plt.ylabel("LR") plt.grid() plt.xlim(0, epochs) plt.ylim(0) plt.savefig(Path(save_dir) / "LR.png", dpi=200) plt.close() def plot_val_txt(): x = np.loadtxt("val.txt", dtype=np.float32) box = xyxy2xywh(x[:, :4]) cx, cy = box[:, 0], box[:, 1] fig, ax = plt.subplots(1, 1, figsize=(6, 6), tight_layout=True) ax.hist2d(cx, cy, bins=600, cmax=10, cmin=0) ax.set_aspect("equal") plt.savefig("hist2d.png", dpi=300) _fig, ax = plt.subplots(1, 2, figsize=(12, 6), tight_layout=True) ax[0].hist(cx, bins=600) ax[1].hist(cy, bins=600) plt.savefig("hist1d.png", dpi=200) def plot_targets_txt(): x = np.loadtxt("targets.txt", dtype=np.float32).T s = ["x targets", "y targets", "width targets", "height targets"] _fig, ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True) ax = ax.ravel() for i in range(4): ax[i].hist(x[i], bins=100, label=f"{x[i].mean():.3g} +/- {x[i].std():.3g}") ax[i].legend() ax[i].set_title(s[i]) plt.savefig("targets.jpg", dpi=200) def plot_val_study(file="", dir="", x=None): save_dir = Path(file).parent if file else Path(dir) plot2 = False # plot additional results if plot2: ax = plt.subplots(2, 4, figsize=(10, 6), tight_layout=True)[1].ravel() _fig2, ax2 = plt.subplots(1, 1, figsize=(8, 4), tight_layout=True) # for f in [save_dir / f'study_coco_{x}.txt' for x in ['yolov5n6', 'yolov5s6', 'yolov5m6', 'yolov5l6', 'yolov5x6']]: for f in sorted(save_dir.glob("study*.txt")): y = np.loadtxt(f, dtype=np.float32, usecols=[0, 1, 2, 3, 7, 8, 9], ndmin=2).T x = np.arange(y.shape[1]) if x is None else np.array(x) if plot2: s = ["P", "R", "mAP@.5", "mAP@.5:.95", "t_preprocess (ms/img)", "t_inference (ms/img)", "t_NMS (ms/img)"] for i in range(7): ax[i].plot(x, y[i], ".-", linewidth=2, markersize=8) ax[i].set_title(s[i]) j = y[3].argmax() + 1 ax2.plot( y[5, 1:j], y[3, 1:j] * 1e2, ".-", linewidth=2, markersize=8, label=f.stem.replace("study_coco_", "").replace("yolo", "YOLO"), ) ax2.plot( 1e3 / np.array([209, 140, 97, 58, 35, 18]), [34.6, 40.5, 43.0, 47.5, 49.7, 51.5], "k.-", linewidth=2, markersize=8, alpha=0.25, label="EfficientDet", ) ax2.grid(alpha=0.2) ax2.set_yticks(np.arange(20, 60, 5)) ax2.set_xlim(0, 57) ax2.set_ylim(25, 55) ax2.set_xlabel("GPU Speed (ms/img)") ax2.set_ylabel("COCO AP val") ax2.legend(loc="lower right") f = save_dir / "study.png" print(f"Saving {f}...") plt.savefig(f, dpi=300) @TryExcept() # known issue https://github.com/ultralytics/yolov5/issues/5395 def plot_labels(labels, names=(), save_dir=Path("")): LOGGER.info(f"Plotting labels to {save_dir / 'labels.jpg'}... ") c, b = labels[:, 0], labels[:, 1:].transpose() # classes, boxes nc = int(c.max() + 1) # number of classes x = pd.DataFrame(b.transpose(), columns=["x", "y", "width", "height"]) # seaborn correlogram sn.pairplot(x, corner=True, diag_kind="auto", kind="hist", diag_kws=dict(bins=50), plot_kws=dict(pmax=0.9)) plt.savefig(save_dir / "labels_correlogram.jpg", dpi=200) plt.close() # matplotlib labels matplotlib.use("svg") # faster ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True)[1].ravel() y = ax[0].hist(c, bins=np.linspace(0, nc, nc + 1) - 0.5, rwidth=0.8) with contextlib.suppress(Exception): # color histogram bars by class [y[2].patches[i].set_color([x / 255 for x in colors(i)]) for i in range(nc)] # known issue #3195 ax[0].set_ylabel("instances") if 0 < len(names) < 30: ax[0].set_xticks(range(len(names))) ax[0].set_xticklabels(list(names.values()), rotation=90, fontsize=10) else: ax[0].set_xlabel("classes") sn.histplot(x, x="x", y="y", ax=ax[2], bins=50, pmax=0.9) sn.histplot(x, x="width", y="height", ax=ax[3], bins=50, pmax=0.9) # rectangles labels[:, 1:3] = 0.5 # center labels[:, 1:] = xywh2xyxy(labels[:, 1:]) * 2000 img = Image.fromarray(np.ones((2000, 2000, 3), dtype=np.uint8) * 255) for cls, *box in labels[:1000]: ImageDraw.Draw(img).rectangle(box, width=1, outline=colors(cls)) # plot ax[1].imshow(img) ax[1].axis("off") for a in [0, 1, 2, 3]: for s in ["top", "right", "left", "bottom"]: ax[a].spines[s].set_visible(False) plt.savefig(save_dir / "labels.jpg", dpi=200) matplotlib.use("Agg") plt.close() def imshow_cls(im, labels=None, pred=None, names=None, nmax=25, verbose=False, f=Path("images.jpg")): from utils.augmentations import denormalize names = names or [f"class{i}" for i in range(1000)] blocks = torch.chunk( denormalize(im.clone()).cpu().float(), len(im), dim=0 ) # select batch index 0, block by channels n = min(len(blocks), nmax) # number of plots m = min(8, round(n**0.5)) # 8 x 8 default _fig, ax = plt.subplots(math.ceil(n / m), m) # 8 rows x n/8 cols ax = ax.ravel() if m > 1 else [ax] # plt.subplots_adjust(wspace=0.05, hspace=0.05) for i in range(n): ax[i].imshow(blocks[i].squeeze().permute((1, 2, 0)).numpy().clip(0.0, 1.0)) ax[i].axis("off") if labels is not None: s = names[labels[i]] + (f"—{names[pred[i]]}" if pred is not None else "") ax[i].set_title(s, fontsize=8, verticalalignment="top") plt.savefig(f, dpi=300, bbox_inches="tight") plt.close() if verbose: LOGGER.info(f"Saving {f}") if labels is not None: LOGGER.info("True: " + " ".join(f"{names[i]:3s}" for i in labels[:nmax])) if pred is not None: LOGGER.info("Predicted:" + " ".join(f"{names[i]:3s}" for i in pred[:nmax])) return f def plot_evolve(evolve_csv="path/to/evolve.csv"): evolve_csv = Path(evolve_csv) data = pd.read_csv(evolve_csv) keys = [x.strip() for x in data.columns] x = data.values f = fitness(x) j = np.argmax(f) # max fitness index plt.figure(figsize=(10, 12), tight_layout=True) matplotlib.rc("font", **{"size": 8}) print(f"Best results from row {j} of {evolve_csv}:") for i, k in enumerate(keys[7:]): v = x[:, 7 + i] mu = v[j] # best single result plt.subplot(6, 5, i + 1) plt.scatter(v, f, c=hist2d(v, f, 20), cmap="viridis", alpha=0.8, edgecolors="none") plt.plot(mu, f.max(), "k+", markersize=15) plt.title(f"{k} = {mu:.3g}", fontdict={"size": 9}) # limit to 40 characters if i % 5 != 0: plt.yticks([]) print(f"{k:>15}: {mu:.3g}") f = evolve_csv.with_suffix(".png") # filename plt.savefig(f, dpi=200) plt.close() print(f"Saved {f}") def plot_results(file="path/to/results.csv", dir=""): save_dir = Path(file).parent if file else Path(dir) fig, ax = plt.subplots(2, 5, figsize=(12, 6), tight_layout=True) ax = ax.ravel() files = list(save_dir.glob("results*.csv")) assert len(files), f"No results.csv files found in {save_dir.resolve()}, nothing to plot." for f in files: try: data = pd.read_csv(f) s = [x.strip() for x in data.columns] x = data.values[:, 0] for i, j in enumerate([1, 2, 3, 4, 5, 8, 9, 10, 6, 7]): y = data.values[:, j].astype("float") # y[y == 0] = np.nan # don't show zero values ax[i].plot(x, y, marker=".", label=f.stem, linewidth=2, markersize=8) # actual results ax[i].plot(x, gaussian_filter1d(y, sigma=3), ":", label="smooth", linewidth=2) # smoothing line ax[i].set_title(s[j], fontsize=12) # if j in [8, 9, 10]: # share train and val loss y axes # ax[i].get_shared_y_axes().join(ax[i], ax[i - 5]) except Exception as e: LOGGER.info(f"Warning: Plotting error for {f}: {e}") ax[1].legend() fig.savefig(save_dir / "results.png", dpi=200) plt.close() def profile_idetection(start=0, stop=0, labels=(), save_dir=""): ax = plt.subplots(2, 4, figsize=(12, 6), tight_layout=True)[1].ravel() s = ["Images", "Free Storage (GB)", "RAM Usage (GB)", "Battery", "dt_raw (ms)", "dt_smooth (ms)", "real-world FPS"] files = list(Path(save_dir).glob("frames*.txt")) for fi, f in enumerate(files): try: results = np.loadtxt(f, ndmin=2).T[:, 90:-30] # clip first and last rows n = results.shape[1] # number of rows x = np.arange(start, min(stop, n) if stop else n) results = results[:, x] t = results[0] - results[0].min() # set t0=0s results[0] = x for i, a in enumerate(ax): if i < len(results): label = labels[fi] if len(labels) else f.stem.replace("frames_", "") a.plot(t, results[i], marker=".", label=label, linewidth=1, markersize=5) a.set_title(s[i]) a.set_xlabel("time (s)") # if fi == len(files) - 1: # a.set_ylim(bottom=0) for side in ["top", "right"]: a.spines[side].set_visible(False) else: a.remove() except Exception as e: print(f"Warning: Plotting error for {f}; {e}") ax[1].legend() plt.savefig(Path(save_dir) / "idetection_profile.png", dpi=200) def save_one_box(xyxy, im, file=Path("im.jpg"), gain=1.02, pad=10, square=False, BGR=False, save=True): xyxy = torch.tensor(xyxy).view(-1, 4) b = xyxy2xywh(xyxy) # boxes if square: b[:, 2:] = b[:, 2:].max(1)[0].unsqueeze(1) # attempt rectangle to square b[:, 2:] = b[:, 2:] * gain + pad # box wh * gain + pad xyxy = xywh2xyxy(b).long() clip_boxes(xyxy, im.shape) crop = im[int(xyxy[0, 1]) : int(xyxy[0, 3]), int(xyxy[0, 0]) : int(xyxy[0, 2]), :: (1 if BGR else -1)] if save: file.parent.mkdir(parents=True, exist_ok=True) # make directory f = str(increment_path(file).with_suffix(".jpg")) # cv2.imwrite(f, crop) # save BGR, https://github.com/ultralytics/yolov5/issues/7007 chroma subsampling issue Image.fromarray(crop[..., ::-1]).save(f, quality=95, subsampling=0) # save RGB return crop
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Plotting utils.""" import contextlib import math @@ -28,8 +29,14 @@ class Colors: + """Provides an RGB color palette derived from Ultralytics color scheme for visualization tasks.""" def __init__(self): + """Initializes the Colors class with a palette derived from Ultralytics color scheme, converting hex codes to + RGB. + + Colors derived from `hex = matplotlib.colors.TABLEAU_COLORS.values()`. + """ hexs = ( "FF3838", "FF9D97", @@ -56,11 +63,13 @@ self.n = len(self.palette) def __call__(self, i, bgr=False): + """Returns color from palette by index `i`, in BGR format if `bgr=True`, else RGB; `i` is an integer index.""" c = self.palette[int(i) % self.n] return (c[2], c[1], c[0]) if bgr else c @staticmethod def hex2rgb(h): + """Converts hexadecimal color `h` to an RGB tuple (PIL-compatible) with order (R, G, B).""" return tuple(int(h[1 + i : 1 + i + 2], 16) for i in (0, 2, 4)) @@ -68,6 +77,14 @@ def feature_visualization(x, module_type, stage, n=32, save_dir=Path("runs/detect/exp")): + """ + Args: + x: Features to be visualized + module_type: Module type + stage: Module stage within model + n: Maximum number of feature maps to plot + save_dir: Directory to save results. + """ if ("Detect" not in module_type) and ( "Segment" not in module_type ): # 'Detect' for Object Detect task,'Segment' for Segment task @@ -91,6 +108,10 @@ def hist2d(x, y, n=100): + """Generates a logarithmic 2D histogram, useful for visualizing label or evolution distributions. + + Used in used in labels.png and evolve.png. + """ xedges, yedges = np.linspace(x.min(), x.max(), n), np.linspace(y.min(), y.max(), n) hist, xedges, yedges = np.histogram2d(x, y, (xedges, yedges)) xidx = np.clip(np.digitize(x, xedges) - 1, 0, hist.shape[0] - 1) @@ -99,10 +120,14 @@ def butter_lowpass_filtfilt(data, cutoff=1500, fs=50000, order=5): + """Applies a low-pass Butterworth filter to `data` with specified `cutoff`, `fs`, and `order`.""" from scipy.signal import butter, filtfilt # https://stackoverflow.com/questions/28536191/how-to-filter-smooth-with-scipy-numpy def butter_lowpass(cutoff, fs, order): + """Applies a low-pass Butterworth filter to a signal with specified cutoff frequency, sample rate, and filter + order. + """ nyq = 0.5 * fs normal_cutoff = cutoff / nyq return butter(order, normal_cutoff, btype="low", analog=False) @@ -112,6 +137,9 @@ def output_to_target(output, max_det=300): + """Converts YOLOv5 model output to [batch_id, class_id, x, y, w, h, conf] format for plotting, limiting detections + to `max_det`. + """ targets = [] for i, o in enumerate(output): box, conf, cls = o[:max_det, :6].cpu().split((4, 1, 1), 1) @@ -122,6 +150,7 @@ @threaded def plot_images(images, targets, paths=None, fname="images.jpg", names=None): + """Plots an image grid with labels from YOLOv5 predictions or targets, saving to `fname`.""" if isinstance(images, torch.Tensor): images = images.cpu().float().numpy() if isinstance(targets, torch.Tensor): @@ -185,6 +214,7 @@ def plot_lr_scheduler(optimizer, scheduler, epochs=300, save_dir=""): + """Plots learning rate schedule for given optimizer and scheduler, saving plot to `save_dir`.""" optimizer, scheduler = copy(optimizer), copy(scheduler) # do not modify originals y = [] for _ in range(epochs): @@ -201,6 +231,11 @@ def plot_val_txt(): + """Plots 2D and 1D histograms of bounding box centers from 'val.txt' using matplotlib, saving as 'hist2d.png' and + 'hist1d.png'. + + Example: from utils.plots import *; plot_val() + """ x = np.loadtxt("val.txt", dtype=np.float32) box = xyxy2xywh(x[:, :4]) cx, cy = box[:, 0], box[:, 1] @@ -217,6 +252,10 @@ def plot_targets_txt(): + """Plots histograms of object detection targets from 'targets.txt', saving the figure as 'targets.jpg'. + + Example: from utils.plots import *; plot_targets_txt() + """ x = np.loadtxt("targets.txt", dtype=np.float32).T s = ["x targets", "y targets", "width targets", "height targets"] _fig, ax = plt.subplots(2, 2, figsize=(8, 8), tight_layout=True) @@ -229,6 +268,11 @@ def plot_val_study(file="", dir="", x=None): + """Plots validation study results from 'study*.txt' files in a directory or a specific file, comparing model + performance and speed. + + Example: from utils.plots import *; plot_val_study() + """ save_dir = Path(file).parent if file else Path(dir) plot2 = False # plot additional results if plot2: @@ -279,6 +323,7 @@ @TryExcept() # known issue https://github.com/ultralytics/yolov5/issues/5395 def plot_labels(labels, names=(), save_dir=Path("")): + """Plots dataset labels, saving correlogram and label images, handles classes, and visualizes bounding boxes.""" LOGGER.info(f"Plotting labels to {save_dir / 'labels.jpg'}... ") c, b = labels[:, 0], labels[:, 1:].transpose() # classes, boxes nc = int(c.max() + 1) # number of classes @@ -323,6 +368,7 @@ def imshow_cls(im, labels=None, pred=None, names=None, nmax=25, verbose=False, f=Path("images.jpg")): + """Displays a grid of images with optional labels and predictions, saving to a file.""" from utils.augmentations import denormalize names = names or [f"class{i}" for i in range(1000)] @@ -352,6 +398,10 @@ def plot_evolve(evolve_csv="path/to/evolve.csv"): + """Plots hyperparameter evolution results from a given CSV, saving the plot and displaying best results. + + Example: from utils.plots import *; plot_evolve() + """ evolve_csv = Path(evolve_csv) data = pd.read_csv(evolve_csv) keys = [x.strip() for x in data.columns] @@ -378,6 +428,10 @@ def plot_results(file="path/to/results.csv", dir=""): + """Plots training results from a 'results.csv' file; accepts file path and directory as arguments. + + Example: from utils.plots import *; plot_results('path/to/results.csv') + """ save_dir = Path(file).parent if file else Path(dir) fig, ax = plt.subplots(2, 5, figsize=(12, 6), tight_layout=True) ax = ax.ravel() @@ -404,6 +458,10 @@ def profile_idetection(start=0, stop=0, labels=(), save_dir=""): + """Plots per-image iDetection logs, comparing metrics like storage and performance over time. + + Example: from utils.plots import *; profile_idetection() + """ ax = plt.subplots(2, 4, figsize=(12, 6), tight_layout=True)[1].ravel() s = ["Images", "Free Storage (GB)", "RAM Usage (GB)", "Battery", "dt_raw (ms)", "dt_smooth (ms)", "real-world FPS"] files = list(Path(save_dir).glob("frames*.txt")) @@ -434,6 +492,9 @@ def save_one_box(xyxy, im, file=Path("im.jpg"), gain=1.02, pad=10, square=False, BGR=False, save=True): + """Crops and saves an image from bounding box `xyxy`, applied with `gain` and `pad`, optionally squares and adjusts + for BGR. + """ xyxy = torch.tensor(xyxy).view(-1, 4) b = xyxy2xywh(xyxy) # boxes if square: @@ -447,4 +508,4 @@ f = str(increment_path(file).with_suffix(".jpg")) # cv2.imwrite(f, crop) # save BGR, https://github.com/ultralytics/yolov5/issues/7007 chroma subsampling issue Image.fromarray(crop[..., ::-1]).save(f, quality=95, subsampling=0) # save RGB - return crop+ return crop
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/plots.py
Write beginner-friendly docstrings
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import os import sys from pathlib import Path import torch from tqdm import tqdm FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative from models.common import DetectMultiBackend from utils.dataloaders import create_classification_dataloader from utils.general import ( LOGGER, TQDM_BAR_FORMAT, Profile, check_img_size, check_requirements, colorstr, increment_path, print_args, ) from utils.torch_utils import select_device, smart_inference_mode @smart_inference_mode() def run( data=ROOT / "../datasets/mnist", # dataset dir weights=ROOT / "yolov5s-cls.pt", # model.pt path(s) batch_size=128, # batch size imgsz=224, # inference size (pixels) device="", # cuda device, i.e. 0 or 0,1,2,3 or cpu workers=8, # max dataloader workers (per RANK in DDP mode) verbose=False, # verbose output project=ROOT / "runs/val-cls", # save to project/name name="exp", # save to project/name exist_ok=False, # existing project/name ok, do not increment half=False, # use FP16 half-precision inference dnn=False, # use OpenCV DNN for ONNX inference model=None, dataloader=None, criterion=None, pbar=None, ): # Initialize/load model and set device training = model is not None if training: # called by train.py device, pt, jit, engine = next(model.parameters()).device, True, False, False # get model device, PyTorch model half &= device.type != "cpu" # half precision only supported on CUDA model.half() if half else model.float() else: # called directly device = select_device(device, batch_size=batch_size) # Directories save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run save_dir.mkdir(parents=True, exist_ok=True) # make dir # Load model model = DetectMultiBackend(weights, device=device, dnn=dnn, fp16=half) stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine imgsz = check_img_size(imgsz, s=stride) # check image size half = model.fp16 # FP16 supported on limited backends with CUDA if engine: batch_size = model.batch_size else: device = model.device if not (pt or jit): batch_size = 1 # export.py models default to batch-size 1 LOGGER.info(f"Forcing --batch-size 1 square inference (1,3,{imgsz},{imgsz}) for non-PyTorch models") # Dataloader data = Path(data) test_dir = data / "test" if (data / "test").exists() else data / "val" # data/test or data/val dataloader = create_classification_dataloader( path=test_dir, imgsz=imgsz, batch_size=batch_size, augment=False, rank=-1, workers=workers ) model.eval() pred, targets, loss, dt = [], [], 0, (Profile(device=device), Profile(device=device), Profile(device=device)) n = len(dataloader) # number of batches action = "validating" if dataloader.dataset.root.stem == "val" else "testing" desc = f"{pbar.desc[:-36]}{action:>36}" if pbar else f"{action}" bar = tqdm(dataloader, desc, n, not training, bar_format=TQDM_BAR_FORMAT, position=0) with torch.cuda.amp.autocast(enabled=device.type != "cpu"): for images, labels in bar: with dt[0]: images, labels = images.to(device, non_blocking=True), labels.to(device) with dt[1]: y = model(images) with dt[2]: pred.append(y.argsort(1, descending=True)[:, :5]) targets.append(labels) if criterion: loss += criterion(y, labels) loss /= n pred, targets = torch.cat(pred), torch.cat(targets) correct = (targets[:, None] == pred).float() acc = torch.stack((correct[:, 0], correct.max(1).values), dim=1) # (top1, top5) accuracy top1, top5 = acc.mean(0).tolist() if pbar: pbar.desc = f"{pbar.desc[:-36]}{loss:>12.3g}{top1:>12.3g}{top5:>12.3g}" if verbose: # all classes LOGGER.info(f"{'Class':>24}{'Images':>12}{'top1_acc':>12}{'top5_acc':>12}") LOGGER.info(f"{'all':>24}{targets.shape[0]:>12}{top1:>12.3g}{top5:>12.3g}") for i, c in model.names.items(): acc_i = acc[targets == i] top1i, top5i = acc_i.mean(0).tolist() LOGGER.info(f"{c:>24}{acc_i.shape[0]:>12}{top1i:>12.3g}{top5i:>12.3g}") # Print results t = tuple(x.t / len(dataloader.dataset.samples) * 1e3 for x in dt) # speeds per image shape = (1, 3, imgsz, imgsz) LOGGER.info(f"Speed: %.1fms pre-process, %.1fms inference, %.1fms post-process per image at shape {shape}" % t) LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}") return top1, top5, loss def parse_opt(): parser = argparse.ArgumentParser() parser.add_argument("--data", type=str, default=ROOT / "../datasets/mnist", help="dataset path") parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-cls.pt", help="model.pt path(s)") parser.add_argument("--batch-size", type=int, default=128, help="batch size") parser.add_argument("--imgsz", "--img", "--img-size", type=int, default=224, help="inference size (pixels)") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--workers", type=int, default=8, help="max dataloader workers (per RANK in DDP mode)") parser.add_argument("--verbose", nargs="?", const=True, default=True, help="verbose output") parser.add_argument("--project", default=ROOT / "runs/val-cls", help="save to project/name") parser.add_argument("--name", default="exp", help="save to project/name") parser.add_argument("--exist-ok", action="store_true", help="existing project/name ok, do not increment") parser.add_argument("--half", action="store_true", help="use FP16 half-precision inference") parser.add_argument("--dnn", action="store_true", help="use OpenCV DNN for ONNX inference") opt = parser.parse_args() print_args(vars(opt)) return opt def main(opt): check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() main(opt)
--- +++ @@ -1,4 +1,24 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +Validate a trained YOLOv5 classification model on a classification dataset. + +Usage: + $ bash data/scripts/get_imagenet.sh --val # download ImageNet val split (6.3G, 50000 images) + $ python classify/val.py --weights yolov5m-cls.pt --data ../datasets/imagenet --img 224 # validate ImageNet + +Usage - formats: + $ python classify/val.py --weights yolov5s-cls.pt # PyTorch + yolov5s-cls.torchscript # TorchScript + yolov5s-cls.onnx # ONNX Runtime or OpenCV DNN with --dnn + yolov5s-cls_openvino_model # OpenVINO + yolov5s-cls.engine # TensorRT + yolov5s-cls.mlmodel # CoreML (macOS-only) + yolov5s-cls_saved_model # TensorFlow SavedModel + yolov5s-cls.pb # TensorFlow GraphDef + yolov5s-cls.tflite # TensorFlow Lite + yolov5s-cls_edgetpu.tflite # TensorFlow Edge TPU + yolov5s-cls_paddle_model # PaddlePaddle +""" import argparse import os @@ -48,6 +68,7 @@ criterion=None, pbar=None, ): + """Validates a YOLOv5 classification model on a dataset, computing metrics like top1 and top5 accuracy.""" # Initialize/load model and set device training = model is not None if training: # called by train.py @@ -127,6 +148,7 @@ def parse_opt(): + """Parses and returns command line arguments for YOLOv5 model evaluation and inference settings.""" parser = argparse.ArgumentParser() parser.add_argument("--data", type=str, default=ROOT / "../datasets/mnist", help="dataset path") parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-cls.pt", help="model.pt path(s)") @@ -146,10 +168,11 @@ def main(opt): + """Executes the YOLOv5 model prediction workflow, handling argument parsing and requirement checks.""" check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() - main(opt)+ main(opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/classify/val.py
Add docstrings to my Python code
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import logging import subprocess import urllib from pathlib import Path import requests import torch def is_url(url, check=True): try: url = str(url) result = urllib.parse.urlparse(url) assert all([result.scheme, result.netloc]) # check if is url return (urllib.request.urlopen(url).getcode() == 200) if check else True # check if exists online except (AssertionError, urllib.request.HTTPError): return False def gsutil_getsize(url=""): output = subprocess.check_output(["gsutil", "du", url], shell=True, encoding="utf-8") return int(output.split()[0]) if output else 0 def url_getsize(url="https://ultralytics.com/images/bus.jpg"): response = requests.head(url, allow_redirects=True) return int(response.headers.get("content-length", -1)) def curl_download(url, filename, *, silent: bool = False) -> bool: silent_option = "sS" if silent else "" # silent proc = subprocess.run( [ "curl", "-#", f"-{silent_option}L", url, "--output", filename, "--retry", "9", "-C", "-", ] ) return proc.returncode == 0 def safe_download(file, url, url2=None, min_bytes=1e0, error_msg=""): from utils.general import LOGGER file = Path(file) assert_msg = f"Downloaded file '{file}' does not exist or size is < min_bytes={min_bytes}" try: # url1 LOGGER.info(f"Downloading {url} to {file}...") torch.hub.download_url_to_file(url, str(file), progress=LOGGER.level <= logging.INFO) assert file.exists() and file.stat().st_size > min_bytes, assert_msg # check except Exception as e: # url2 if file.exists(): file.unlink() # remove partial downloads LOGGER.info(f"ERROR: {e}\nRe-attempting {url2 or url} to {file}...") # curl download, retry and resume on fail curl_download(url2 or url, file) finally: if not file.exists() or file.stat().st_size < min_bytes: # check if file.exists(): file.unlink() # remove partial downloads LOGGER.info(f"ERROR: {assert_msg}\n{error_msg}") LOGGER.info("") def attempt_download(file, repo="ultralytics/yolov5", release="v7.0"): from utils.general import LOGGER def github_assets(repository, version="latest"): if version != "latest": version = f"tags/{version}" # i.e. tags/v7.0 response = requests.get(f"https://api.github.com/repos/{repository}/releases/{version}").json() # github api return response["tag_name"], [x["name"] for x in response["assets"]] # tag, assets file = Path(str(file).strip().replace("'", "")) if not file.exists(): # URL specified name = Path(urllib.parse.unquote(str(file))).name # decode '%2F' to '/' etc. if str(file).startswith(("http:/", "https:/")): # download url = str(file).replace(":/", "://") # Pathlib turns :// -> :/ file = name.split("?")[0] # parse authentication https://url.com/file.txt?auth... if Path(file).is_file(): LOGGER.info(f"Found {url} locally at {file}") # file already exists else: safe_download(file=file, url=url, min_bytes=1e5) return file # GitHub assets assets = [f"yolov5{size}{suffix}.pt" for size in "nsmlx" for suffix in ("", "6", "-cls", "-seg")] # default try: tag, assets = github_assets(repo, release) except Exception: try: tag, assets = github_assets(repo) # latest release except Exception: try: tag = subprocess.check_output("git tag", shell=True, stderr=subprocess.STDOUT).decode().split()[-1] except Exception: tag = release if name in assets: file.parent.mkdir(parents=True, exist_ok=True) # make parent dir (if required) safe_download( file, url=f"https://github.com/{repo}/releases/download/{tag}/{name}", min_bytes=1e5, error_msg=f"{file} missing, try downloading from https://github.com/{repo}/releases/{tag}", ) return str(file)
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Download utils.""" import logging import subprocess @@ -10,6 +11,7 @@ def is_url(url, check=True): + """Determines if a string is a URL and optionally checks its existence online, returning a boolean.""" try: url = str(url) result = urllib.parse.urlparse(url) @@ -20,16 +22,22 @@ def gsutil_getsize(url=""): + """Returns the size in bytes of a file at a Google Cloud Storage URL using `gsutil du`. + + Returns 0 if the command fails or output is empty. + """ output = subprocess.check_output(["gsutil", "du", url], shell=True, encoding="utf-8") return int(output.split()[0]) if output else 0 def url_getsize(url="https://ultralytics.com/images/bus.jpg"): + """Returns the size in bytes of a downloadable file at a given URL; defaults to -1 if not found.""" response = requests.head(url, allow_redirects=True) return int(response.headers.get("content-length", -1)) def curl_download(url, filename, *, silent: bool = False) -> bool: + """Download a file from a url to a filename using curl.""" silent_option = "sS" if silent else "" # silent proc = subprocess.run( [ @@ -49,6 +57,10 @@ def safe_download(file, url, url2=None, min_bytes=1e0, error_msg=""): + """Downloads a file from a URL (or alternate URL) to a specified path if file is above a minimum size. + + Removes incomplete downloads. + """ from utils.general import LOGGER file = Path(file) @@ -72,9 +84,11 @@ def attempt_download(file, repo="ultralytics/yolov5", release="v7.0"): + """Download a file from GitHub release assets or via direct URL if not found locally.""" from utils.general import LOGGER def github_assets(repository, version="latest"): + """Fetches GitHub repository release tag and asset names using the GitHub API.""" if version != "latest": version = f"tags/{version}" # i.e. tags/v7.0 response = requests.get(f"https://api.github.com/repos/{repository}/releases/{version}").json() # github api @@ -115,4 +129,4 @@ error_msg=f"{file} missing, try downloading from https://github.com/{repo}/releases/{tag}", ) - return str(file)+ return str(file)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/downloads.py
Add docstrings to incomplete code
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import random import numpy as np import torch import yaml from tqdm import tqdm from utils import TryExcept from utils.general import LOGGER, TQDM_BAR_FORMAT, colorstr PREFIX = colorstr("AutoAnchor: ") def check_anchor_order(m): a = m.anchors.prod(-1).mean(-1).view(-1) # mean anchor area per output layer da = a[-1] - a[0] # delta a ds = m.stride[-1] - m.stride[0] # delta s if da and (da.sign() != ds.sign()): # same order LOGGER.info(f"{PREFIX}Reversing anchor order") m.anchors[:] = m.anchors.flip(0) @TryExcept(f"{PREFIX}ERROR") def check_anchors(dataset, model, thr=4.0, imgsz=640): m = model.module.model[-1] if hasattr(model, "module") else model.model[-1] # Detect() shapes = imgsz * dataset.shapes / dataset.shapes.max(1, keepdims=True) scale = np.random.uniform(0.9, 1.1, size=(shapes.shape[0], 1)) # augment scale wh = torch.tensor(np.concatenate([l[:, 3:5] * s for s, l in zip(shapes * scale, dataset.labels)])).float() # wh def metric(k): # compute metric r = wh[:, None] / k[None] x = torch.min(r, 1 / r).min(2)[0] # ratio metric best = x.max(1)[0] # best_x aat = (x > 1 / thr).float().sum(1).mean() # anchors above threshold bpr = (best > 1 / thr).float().mean() # best possible recall return bpr, aat stride = m.stride.to(m.anchors.device).view(-1, 1, 1) # model strides anchors = m.anchors.clone() * stride # current anchors bpr, aat = metric(anchors.cpu().view(-1, 2)) s = f"\n{PREFIX}{aat:.2f} anchors/target, {bpr:.3f} Best Possible Recall (BPR). " if bpr > 0.98: # threshold to recompute LOGGER.info(f"{s}Current anchors are a good fit to dataset ✅") else: LOGGER.info(f"{s}Anchors are a poor fit to dataset ⚠️, attempting to improve...") na = m.anchors.numel() // 2 # number of anchors anchors = kmean_anchors(dataset, n=na, img_size=imgsz, thr=thr, gen=1000, verbose=False) new_bpr = metric(anchors)[0] if new_bpr > bpr: # replace anchors anchors = torch.tensor(anchors, device=m.anchors.device).type_as(m.anchors) m.anchors[:] = anchors.clone().view_as(m.anchors) check_anchor_order(m) # must be in pixel-space (not grid-space) m.anchors /= stride s = f"{PREFIX}Done ✅ (optional: update model *.yaml to use these anchors in the future)" else: s = f"{PREFIX}Done ⚠️ (original anchors better than new anchors, proceeding with original anchors)" LOGGER.info(s) def kmean_anchors(dataset="./data/coco128.yaml", n=9, img_size=640, thr=4.0, gen=1000, verbose=True): from scipy.cluster.vq import kmeans npr = np.random thr = 1 / thr def metric(k, wh): # compute metrics r = wh[:, None] / k[None] x = torch.min(r, 1 / r).min(2)[0] # ratio metric # x = wh_iou(wh, torch.tensor(k)) # iou metric return x, x.max(1)[0] # x, best_x def anchor_fitness(k): # mutation fitness _, best = metric(torch.tensor(k, dtype=torch.float32), wh) return (best * (best > thr).float()).mean() # fitness def print_results(k, verbose=True): k = k[np.argsort(k.prod(1))] # sort small to large x, best = metric(k, wh0) bpr, aat = (best > thr).float().mean(), (x > thr).float().mean() * n # best possible recall, anch > thr s = ( f"{PREFIX}thr={thr:.2f}: {bpr:.4f} best possible recall, {aat:.2f} anchors past thr\n" f"{PREFIX}n={n}, img_size={img_size}, metric_all={x.mean():.3f}/{best.mean():.3f}-mean/best, " f"past_thr={x[x > thr].mean():.3f}-mean: " ) for x in k: s += "%i,%i, " % (round(x[0]), round(x[1])) if verbose: LOGGER.info(s[:-2]) return k if isinstance(dataset, str): # *.yaml file with open(dataset, errors="ignore") as f: data_dict = yaml.safe_load(f) # model dict from utils.dataloaders import LoadImagesAndLabels dataset = LoadImagesAndLabels(data_dict["train"], augment=True, rect=True) # Get label wh shapes = img_size * dataset.shapes / dataset.shapes.max(1, keepdims=True) wh0 = np.concatenate([l[:, 3:5] * s for s, l in zip(shapes, dataset.labels)]) # wh # Filter i = (wh0 < 3.0).any(1).sum() if i: LOGGER.info(f"{PREFIX}WARNING ⚠️ Extremely small objects found: {i} of {len(wh0)} labels are <3 pixels in size") wh = wh0[(wh0 >= 2.0).any(1)].astype(np.float32) # filter > 2 pixels # wh = wh * (npr.rand(wh.shape[0], 1) * 0.9 + 0.1) # multiply by random scale 0-1 # Kmeans init try: LOGGER.info(f"{PREFIX}Running kmeans for {n} anchors on {len(wh)} points...") assert n <= len(wh) # apply overdetermined constraint s = wh.std(0) # sigmas for whitening k = kmeans(wh / s, n, iter=30)[0] * s # points assert n == len(k) # kmeans may return fewer points than requested if wh is insufficient or too similar except Exception: LOGGER.warning(f"{PREFIX}WARNING ⚠️ switching strategies from kmeans to random init") k = np.sort(npr.rand(n * 2)).reshape(n, 2) * img_size # random init wh, wh0 = (torch.tensor(x, dtype=torch.float32) for x in (wh, wh0)) k = print_results(k, verbose=False) # Plot # k, d = [None] * 20, [None] * 20 # for i in tqdm(range(1, 21)): # k[i-1], d[i-1] = kmeans(wh / s, i) # points, mean distance # fig, ax = plt.subplots(1, 2, figsize=(14, 7), tight_layout=True) # ax = ax.ravel() # ax[0].plot(np.arange(1, 21), np.array(d) ** 2, marker='.') # fig, ax = plt.subplots(1, 2, figsize=(14, 7)) # plot wh # ax[0].hist(wh[wh[:, 0]<100, 0],400) # ax[1].hist(wh[wh[:, 1]<100, 1],400) # fig.savefig('wh.png', dpi=200) # Evolve f, sh, mp, s = anchor_fitness(k), k.shape, 0.9, 0.1 # fitness, generations, mutation prob, sigma pbar = tqdm(range(gen), bar_format=TQDM_BAR_FORMAT) # progress bar for _ in pbar: v = np.ones(sh) while (v == 1).all(): # mutate until a change occurs (prevent duplicates) v = ((npr.random(sh) < mp) * random.random() * npr.randn(*sh) * s + 1).clip(0.3, 3.0) kg = (k.copy() * v).clip(min=2.0) fg = anchor_fitness(kg) if fg > f: f, k = fg, kg.copy() pbar.desc = f"{PREFIX}Evolving anchors with Genetic Algorithm: fitness = {f:.4f}" if verbose: print_results(k, verbose) return print_results(k).astype(np.float32)
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""AutoAnchor utils.""" import random @@ -14,6 +15,7 @@ def check_anchor_order(m): + """Checks and corrects anchor order against stride in YOLOv5 Detect() module if necessary.""" a = m.anchors.prod(-1).mean(-1).view(-1) # mean anchor area per output layer da = a[-1] - a[0] # delta a ds = m.stride[-1] - m.stride[0] # delta s @@ -24,12 +26,14 @@ @TryExcept(f"{PREFIX}ERROR") def check_anchors(dataset, model, thr=4.0, imgsz=640): + """Evaluates anchor fit to dataset and adjusts if necessary, supporting customizable threshold and image size.""" m = model.module.model[-1] if hasattr(model, "module") else model.model[-1] # Detect() shapes = imgsz * dataset.shapes / dataset.shapes.max(1, keepdims=True) scale = np.random.uniform(0.9, 1.1, size=(shapes.shape[0], 1)) # augment scale wh = torch.tensor(np.concatenate([l[:, 3:5] * s for s, l in zip(shapes * scale, dataset.labels)])).float() # wh def metric(k): # compute metric + """Computes ratio metric, anchors above threshold, and best possible recall for YOLOv5 anchor evaluation.""" r = wh[:, None] / k[None] x = torch.min(r, 1 / r).min(2)[0] # ratio metric best = x.max(1)[0] # best_x @@ -60,22 +64,41 @@ def kmean_anchors(dataset="./data/coco128.yaml", n=9, img_size=640, thr=4.0, gen=1000, verbose=True): + """Creates kmeans-evolved anchors from training dataset. + + Args: + dataset: path to data.yaml, or a loaded dataset + n: number of anchors + img_size: image size used for training + thr: anchor-label wh ratio threshold hyperparameter hyp['anchor_t'] used for training, default=4.0 + gen: generations to evolve anchors using genetic algorithm + verbose: print all results + + Returns: + k: kmeans evolved anchors + + Examples: + from utils.autoanchor import *; _ = kmean_anchors() + """ from scipy.cluster.vq import kmeans npr = np.random thr = 1 / thr def metric(k, wh): # compute metrics + """Computes ratio metric, anchors above threshold, and best possible recall for YOLOv5 anchor evaluation.""" r = wh[:, None] / k[None] x = torch.min(r, 1 / r).min(2)[0] # ratio metric # x = wh_iou(wh, torch.tensor(k)) # iou metric return x, x.max(1)[0] # x, best_x def anchor_fitness(k): # mutation fitness + """Evaluates fitness of YOLOv5 anchors by computing recall and ratio metrics for an anchor evolution process.""" _, best = metric(torch.tensor(k, dtype=torch.float32), wh) return (best * (best > thr).float()).mean() # fitness def print_results(k, verbose=True): + """Sorts and logs kmeans-evolved anchor metrics and best possible recall values for YOLOv5 anchor evaluation.""" k = k[np.argsort(k.prod(1))] # sort small to large x, best = metric(k, wh0) bpr, aat = (best > thr).float().mean(), (x > thr).float().mean() * n # best possible recall, anch > thr @@ -148,4 +171,4 @@ if verbose: print_results(k, verbose) - return print_results(k).astype(np.float32)+ return print_results(k).astype(np.float32)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/autoanchor.py
Document this script properly
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import os import platform import sys from pathlib import Path import torch import torch.nn.functional as F FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative from ultralytics.utils.plotting import Annotator from models.common import DetectMultiBackend from utils.augmentations import classify_transforms from utils.dataloaders import IMG_FORMATS, VID_FORMATS, LoadImages, LoadScreenshots, LoadStreams from utils.general import ( LOGGER, Profile, check_file, check_img_size, check_imshow, check_requirements, colorstr, cv2, increment_path, print_args, strip_optimizer, ) from utils.torch_utils import select_device, smart_inference_mode @smart_inference_mode() def run( weights=ROOT / "yolov5s-cls.pt", # model.pt path(s) source=ROOT / "data/images", # file/dir/URL/glob/screen/0(webcam) data=ROOT / "data/coco128.yaml", # dataset.yaml path imgsz=(224, 224), # inference size (height, width) device="", # cuda device, i.e. 0 or 0,1,2,3 or cpu view_img=False, # show results save_txt=False, # save results to *.txt nosave=False, # do not save images/videos augment=False, # augmented inference visualize=False, # visualize features update=False, # update all models project=ROOT / "runs/predict-cls", # save results to project/name name="exp", # save results to project/name exist_ok=False, # existing project/name ok, do not increment half=False, # use FP16 half-precision inference dnn=False, # use OpenCV DNN for ONNX inference vid_stride=1, # video frame-rate stride ): source = str(source) save_img = not nosave and not source.endswith(".txt") # save inference images is_file = Path(source).suffix[1:] in (IMG_FORMATS + VID_FORMATS) is_url = source.lower().startswith(("rtsp://", "rtmp://", "http://", "https://")) webcam = source.isnumeric() or source.endswith(".streams") or (is_url and not is_file) screenshot = source.lower().startswith("screen") if is_url and is_file: source = check_file(source) # download # Directories save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run (save_dir / "labels" if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir # Load model device = select_device(device) model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half) stride, names, pt = model.stride, model.names, model.pt imgsz = check_img_size(imgsz, s=stride) # check image size # Dataloader bs = 1 # batch_size if webcam: view_img = check_imshow(warn=True) dataset = LoadStreams(source, img_size=imgsz, transforms=classify_transforms(imgsz[0]), vid_stride=vid_stride) bs = len(dataset) elif screenshot: dataset = LoadScreenshots(source, img_size=imgsz, stride=stride, auto=pt) else: dataset = LoadImages(source, img_size=imgsz, transforms=classify_transforms(imgsz[0]), vid_stride=vid_stride) vid_path, vid_writer = [None] * bs, [None] * bs # Run inference model.warmup(imgsz=(1 if pt else bs, 3, *imgsz)) # warmup seen, windows, dt = 0, [], (Profile(device=device), Profile(device=device), Profile(device=device)) for path, im, im0s, vid_cap, s in dataset: with dt[0]: im = torch.Tensor(im).to(model.device) im = im.half() if model.fp16 else im.float() # uint8 to fp16/32 if len(im.shape) == 3: im = im[None] # expand for batch dim # Inference with dt[1]: results = model(im) # Post-process with dt[2]: pred = F.softmax(results, dim=1) # probabilities # Process predictions for i, prob in enumerate(pred): # per image seen += 1 if webcam: # batch_size >= 1 p, im0, frame = path[i], im0s[i].copy(), dataset.count s += f"{i}: " else: p, im0, frame = path, im0s.copy(), getattr(dataset, "frame", 0) p = Path(p) # to Path save_path = str(save_dir / p.name) # im.jpg txt_path = str(save_dir / "labels" / p.stem) + ("" if dataset.mode == "image" else f"_{frame}") # im.txt s += "{:g}x{:g} ".format(*im.shape[2:]) # print string annotator = Annotator(im0, example=str(names), pil=True) # Print results top5i = prob.argsort(0, descending=True)[:5].tolist() # top 5 indices s += f"{', '.join(f'{names[j]} {prob[j]:.2f}' for j in top5i)}, " # Write results text = "\n".join(f"{prob[j]:.2f} {names[j]}" for j in top5i) if save_img or view_img: # Add bbox to image annotator.text([32, 32], text, txt_color=(255, 255, 255)) if save_txt: # Write to file with open(f"{txt_path}.txt", "a") as f: f.write(text + "\n") # Stream results im0 = annotator.result() if view_img: if platform.system() == "Linux" and p not in windows: windows.append(p) cv2.namedWindow(str(p), cv2.WINDOW_NORMAL | cv2.WINDOW_KEEPRATIO) # allow window resize (Linux) cv2.resizeWindow(str(p), im0.shape[1], im0.shape[0]) cv2.imshow(str(p), im0) cv2.waitKey(1) # 1 millisecond # Save results (image with detections) if save_img: if dataset.mode == "image": cv2.imwrite(save_path, im0) else: # 'video' or 'stream' if vid_path[i] != save_path: # new video vid_path[i] = save_path if isinstance(vid_writer[i], cv2.VideoWriter): vid_writer[i].release() # release previous video writer if vid_cap: # video fps = vid_cap.get(cv2.CAP_PROP_FPS) w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH)) h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) else: # stream fps, w, h = 30, im0.shape[1], im0.shape[0] save_path = str(Path(save_path).with_suffix(".mp4")) # force *.mp4 suffix on results videos vid_writer[i] = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*"mp4v"), fps, (w, h)) vid_writer[i].write(im0) # Print time (inference-only) LOGGER.info(f"{s}{dt[1].dt * 1e3:.1f}ms") # Print results t = tuple(x.t / seen * 1e3 for x in dt) # speeds per image LOGGER.info(f"Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {(1, 3, *imgsz)}" % t) if save_txt or save_img: s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else "" LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}") if update: strip_optimizer(weights[0]) # update model (to fix SourceChangeWarning) def parse_opt(): parser = argparse.ArgumentParser() parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-cls.pt", help="model path(s)") parser.add_argument("--source", type=str, default=ROOT / "data/images", help="file/dir/URL/glob/screen/0(webcam)") parser.add_argument("--data", type=str, default=ROOT / "data/coco128.yaml", help="(optional) dataset.yaml path") parser.add_argument("--imgsz", "--img", "--img-size", nargs="+", type=int, default=[224], help="inference size h,w") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--view-img", action="store_true", help="show results") parser.add_argument("--save-txt", action="store_true", help="save results to *.txt") parser.add_argument("--nosave", action="store_true", help="do not save images/videos") parser.add_argument("--augment", action="store_true", help="augmented inference") parser.add_argument("--visualize", action="store_true", help="visualize features") parser.add_argument("--update", action="store_true", help="update all models") parser.add_argument("--project", default=ROOT / "runs/predict-cls", help="save results to project/name") parser.add_argument("--name", default="exp", help="save results to project/name") parser.add_argument("--exist-ok", action="store_true", help="existing project/name ok, do not increment") parser.add_argument("--half", action="store_true", help="use FP16 half-precision inference") parser.add_argument("--dnn", action="store_true", help="use OpenCV DNN for ONNX inference") parser.add_argument("--vid-stride", type=int, default=1, help="video frame-rate stride") opt = parser.parse_args() opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1 # expand print_args(vars(opt)) return opt def main(opt): check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() main(opt)
--- +++ @@ -1,4 +1,32 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +Run YOLOv5 classification inference on images, videos, directories, globs, YouTube, webcam, streams, etc. + +Usage - sources: + $ python classify/predict.py --weights yolov5s-cls.pt --source 0 # webcam + img.jpg # image + vid.mp4 # video + screen # screenshot + path/ # directory + list.txt # list of images + list.streams # list of streams + 'path/*.jpg' # glob + 'https://youtu.be/LNwODJXcvt4' # YouTube + 'rtsp://example.com/media.mp4' # RTSP, RTMP, HTTP stream + +Usage - formats: + $ python classify/predict.py --weights yolov5s-cls.pt # PyTorch + yolov5s-cls.torchscript # TorchScript + yolov5s-cls.onnx # ONNX Runtime or OpenCV DNN with --dnn + yolov5s-cls_openvino_model # OpenVINO + yolov5s-cls.engine # TensorRT + yolov5s-cls.mlmodel # CoreML (macOS-only) + yolov5s-cls_saved_model # TensorFlow SavedModel + yolov5s-cls.pb # TensorFlow GraphDef + yolov5s-cls.tflite # TensorFlow Lite + yolov5s-cls_edgetpu.tflite # TensorFlow Edge TPU + yolov5s-cls_paddle_model # PaddlePaddle +""" import argparse import os @@ -56,6 +84,7 @@ dnn=False, # use OpenCV DNN for ONNX inference vid_stride=1, # video frame-rate stride ): + """Conducts YOLOv5 classification inference on diverse input sources and saves results.""" source = str(source) save_img = not nosave and not source.endswith(".txt") # save inference images is_file = Path(source).suffix[1:] in (IMG_FORMATS + VID_FORMATS) @@ -176,6 +205,7 @@ def parse_opt(): + """Parses command line arguments for YOLOv5 inference settings including model, source, device, and image size.""" parser = argparse.ArgumentParser() parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-cls.pt", help="model path(s)") parser.add_argument("--source", type=str, default=ROOT / "data/images", help="file/dir/URL/glob/screen/0(webcam)") @@ -201,10 +231,11 @@ def main(opt): + """Executes YOLOv5 model inference with options for ONNX DNN and video frame-rate stride adjustments.""" check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) run(**vars(opt)) if __name__ == "__main__": opt = parse_opt() - main(opt)+ main(opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/classify/predict.py
Provide clean and structured docstrings
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import os import random import cv2 import numpy as np import torch from torch.utils.data import DataLoader from ..augmentations import augment_hsv, copy_paste, letterbox from ..dataloaders import InfiniteDataLoader, LoadImagesAndLabels, SmartDistributedSampler, seed_worker from ..general import LOGGER, xyn2xy, xywhn2xyxy, xyxy2xywhn from ..torch_utils import torch_distributed_zero_first from .augmentations import mixup, random_perspective RANK = int(os.getenv("RANK", -1)) def create_dataloader( path, imgsz, batch_size, stride, single_cls=False, hyp=None, augment=False, cache=False, pad=0.0, rect=False, rank=-1, workers=8, image_weights=False, quad=False, prefix="", shuffle=False, mask_downsample_ratio=1, overlap_mask=False, seed=0, ): if rect and shuffle: LOGGER.warning("WARNING ⚠️ --rect is incompatible with DataLoader shuffle, setting shuffle=False") shuffle = False with torch_distributed_zero_first(rank): # init dataset *.cache only once if DDP dataset = LoadImagesAndLabelsAndMasks( path, imgsz, batch_size, augment=augment, # augmentation hyp=hyp, # hyperparameters rect=rect, # rectangular batches cache_images=cache, single_cls=single_cls, stride=int(stride), pad=pad, image_weights=image_weights, prefix=prefix, downsample_ratio=mask_downsample_ratio, overlap=overlap_mask, rank=rank, ) batch_size = min(batch_size, len(dataset)) nd = torch.cuda.device_count() # number of CUDA devices nw = min([os.cpu_count() // max(nd, 1), batch_size if batch_size > 1 else 0, workers]) # number of workers sampler = None if rank == -1 else SmartDistributedSampler(dataset, shuffle=shuffle) loader = DataLoader if image_weights else InfiniteDataLoader # only DataLoader allows for attribute updates generator = torch.Generator() generator.manual_seed(6148914691236517205 + seed + RANK) return loader( dataset, batch_size=batch_size, shuffle=shuffle and sampler is None, num_workers=nw, sampler=sampler, drop_last=quad, pin_memory=True, collate_fn=LoadImagesAndLabelsAndMasks.collate_fn4 if quad else LoadImagesAndLabelsAndMasks.collate_fn, worker_init_fn=seed_worker, generator=generator, ), dataset class LoadImagesAndLabelsAndMasks(LoadImagesAndLabels): # for training/testing def __init__( self, path, img_size=640, batch_size=16, augment=False, hyp=None, rect=False, image_weights=False, cache_images=False, single_cls=False, stride=32, pad=0, min_items=0, prefix="", downsample_ratio=1, overlap=False, rank=-1, seed=0, ): super().__init__( path, img_size, batch_size, augment, hyp, rect, image_weights, cache_images, single_cls, stride, pad, min_items, prefix, rank, seed, ) self.downsample_ratio = downsample_ratio self.overlap = overlap def __getitem__(self, index): index = self.indices[index] # linear, shuffled, or image_weights hyp = self.hyp if mosaic := self.mosaic and random.random() < hyp["mosaic"]: # Load mosaic img, labels, segments = self.load_mosaic(index) shapes = None # MixUp augmentation if random.random() < hyp["mixup"]: img, labels, segments = mixup(img, labels, segments, *self.load_mosaic(random.randint(0, self.n - 1))) else: # Load image img, (h0, w0), (h, w) = self.load_image(index) # Letterbox shape = self.batch_shapes[self.batch[index]] if self.rect else self.img_size # final letterboxed shape img, ratio, pad = letterbox(img, shape, auto=False, scaleup=self.augment) shapes = (h0, w0), ((h / h0, w / w0), pad) # for COCO mAP rescaling labels = self.labels[index].copy() # [array, array, ....], array.shape=(num_points, 2), xyxyxyxy segments = self.segments[index].copy() if len(segments): for i_s in range(len(segments)): segments[i_s] = xyn2xy( segments[i_s], ratio[0] * w, ratio[1] * h, padw=pad[0], padh=pad[1], ) if labels.size: # normalized xywh to pixel xyxy format labels[:, 1:] = xywhn2xyxy(labels[:, 1:], ratio[0] * w, ratio[1] * h, padw=pad[0], padh=pad[1]) if self.augment: img, labels, segments = random_perspective( img, labels, segments=segments, degrees=hyp["degrees"], translate=hyp["translate"], scale=hyp["scale"], shear=hyp["shear"], perspective=hyp["perspective"], ) nl = len(labels) # number of labels masks = [] if nl: labels[:, 1:5] = xyxy2xywhn(labels[:, 1:5], w=img.shape[1], h=img.shape[0], clip=True, eps=1e-3) if self.overlap: masks, sorted_idx = polygons2masks_overlap( img.shape[:2], segments, downsample_ratio=self.downsample_ratio ) masks = masks[None] # (640, 640) -> (1, 640, 640) labels = labels[sorted_idx] else: masks = polygons2masks(img.shape[:2], segments, color=1, downsample_ratio=self.downsample_ratio) masks = ( torch.from_numpy(masks) if len(masks) else torch.zeros( 1 if self.overlap else nl, img.shape[0] // self.downsample_ratio, img.shape[1] // self.downsample_ratio ) ) # TODO: albumentations support if self.augment: # Albumentations # there are some augmentation that won't change boxes and masks, # so just be it for now. img, labels = self.albumentations(img, labels) nl = len(labels) # update after albumentations # HSV color-space augment_hsv(img, hgain=hyp["hsv_h"], sgain=hyp["hsv_s"], vgain=hyp["hsv_v"]) # Flip up-down if random.random() < hyp["flipud"]: img = np.flipud(img) if nl: labels[:, 2] = 1 - labels[:, 2] masks = torch.flip(masks, dims=[1]) # Flip left-right if random.random() < hyp["fliplr"]: img = np.fliplr(img) if nl: labels[:, 1] = 1 - labels[:, 1] masks = torch.flip(masks, dims=[2]) # Cutouts # labels = cutout(img, labels, p=0.5) labels_out = torch.zeros((nl, 6)) if nl: labels_out[:, 1:] = torch.from_numpy(labels) # Convert img = img.transpose((2, 0, 1))[::-1] # HWC to CHW, BGR to RGB img = np.ascontiguousarray(img) return (torch.from_numpy(img), labels_out, self.im_files[index], shapes, masks) def load_mosaic(self, index): labels4, segments4 = [], [] s = self.img_size yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y # 3 additional image indices indices = [index] + random.choices(self.indices, k=3) # 3 additional image indices for i, index in enumerate(indices): # Load image img, _, (h, w) = self.load_image(index) # place img in img4 if i == 0: # top left img4 = np.full((s * 2, s * 2, img.shape[2]), 114, dtype=np.uint8) # base image with 4 tiles x1a, y1a, x2a, y2a = max(xc - w, 0), max(yc - h, 0), xc, yc # xmin, ymin, xmax, ymax (large image) x1b, y1b, x2b, y2b = w - (x2a - x1a), h - (y2a - y1a), w, h # xmin, ymin, xmax, ymax (small image) elif i == 1: # top right x1a, y1a, x2a, y2a = xc, max(yc - h, 0), min(xc + w, s * 2), yc x1b, y1b, x2b, y2b = 0, h - (y2a - y1a), min(w, x2a - x1a), h elif i == 2: # bottom left x1a, y1a, x2a, y2a = max(xc - w, 0), yc, xc, min(s * 2, yc + h) x1b, y1b, x2b, y2b = w - (x2a - x1a), 0, w, min(y2a - y1a, h) elif i == 3: # bottom right x1a, y1a, x2a, y2a = xc, yc, min(xc + w, s * 2), min(s * 2, yc + h) x1b, y1b, x2b, y2b = 0, 0, min(w, x2a - x1a), min(y2a - y1a, h) img4[y1a:y2a, x1a:x2a] = img[y1b:y2b, x1b:x2b] # img4[ymin:ymax, xmin:xmax] padw = x1a - x1b padh = y1a - y1b labels, segments = self.labels[index].copy(), self.segments[index].copy() if labels.size: labels[:, 1:] = xywhn2xyxy(labels[:, 1:], w, h, padw, padh) # normalized xywh to pixel xyxy format segments = [xyn2xy(x, w, h, padw, padh) for x in segments] labels4.append(labels) segments4.extend(segments) # Concat/clip labels labels4 = np.concatenate(labels4, 0) for x in (labels4[:, 1:], *segments4): np.clip(x, 0, 2 * s, out=x) # clip when using random_perspective() # img4, labels4 = replicate(img4, labels4) # replicate # Augment img4, labels4, segments4 = copy_paste(img4, labels4, segments4, p=self.hyp["copy_paste"]) img4, labels4, segments4 = random_perspective( img4, labels4, segments4, degrees=self.hyp["degrees"], translate=self.hyp["translate"], scale=self.hyp["scale"], shear=self.hyp["shear"], perspective=self.hyp["perspective"], border=self.mosaic_border, ) # border to remove return img4, labels4, segments4 @staticmethod def collate_fn(batch): img, label, path, shapes, masks = zip(*batch) # transposed batched_masks = torch.cat(masks, 0) for i, l in enumerate(label): l[:, 0] = i # add target image index for build_targets() return torch.stack(img, 0), torch.cat(label, 0), path, shapes, batched_masks def polygon2mask(img_size, polygons, color=1, downsample_ratio=1): mask = np.zeros(img_size, dtype=np.uint8) polygons = np.asarray(polygons) polygons = polygons.astype(np.int32) shape = polygons.shape polygons = polygons.reshape(shape[0], -1, 2) cv2.fillPoly(mask, polygons, color=color) nh, nw = (img_size[0] // downsample_ratio, img_size[1] // downsample_ratio) # NOTE: fillPoly firstly then resize is trying the keep the same way # of loss calculation when mask-ratio=1. mask = cv2.resize(mask, (nw, nh)) return mask def polygons2masks(img_size, polygons, color, downsample_ratio=1): masks = [] for si in range(len(polygons)): mask = polygon2mask(img_size, [polygons[si].reshape(-1)], color, downsample_ratio) masks.append(mask) return np.array(masks) def polygons2masks_overlap(img_size, segments, downsample_ratio=1): masks = np.zeros( (img_size[0] // downsample_ratio, img_size[1] // downsample_ratio), dtype=np.int32 if len(segments) > 255 else np.uint8, ) areas = [] ms = [] for si in range(len(segments)): mask = polygon2mask( img_size, [segments[si].reshape(-1)], downsample_ratio=downsample_ratio, color=1, ) ms.append(mask) areas.append(mask.sum()) areas = np.asarray(areas) index = np.argsort(-areas) ms = np.array(ms)[index] for i in range(len(segments)): mask = ms[i] * (i + 1) masks = masks + mask masks = np.clip(masks, a_min=0, a_max=i + 1) return masks, index
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Dataloaders.""" import os import random @@ -38,6 +39,7 @@ overlap_mask=False, seed=0, ): + """Creates a dataloader for training, validating, or testing YOLO models with various dataset options.""" if rect and shuffle: LOGGER.warning("WARNING ⚠️ --rect is incompatible with DataLoader shuffle, setting shuffle=False") shuffle = False @@ -82,6 +84,7 @@ class LoadImagesAndLabelsAndMasks(LoadImagesAndLabels): # for training/testing + """Loads images, labels, and segmentation masks for training and testing YOLO models with augmentation support.""" def __init__( self, @@ -103,6 +106,7 @@ rank=-1, seed=0, ): + """Initializes the dataset with image, label, and mask loading capabilities for training/testing.""" super().__init__( path, img_size, @@ -124,6 +128,7 @@ self.overlap = overlap def __getitem__(self, index): + """Returns a transformed item from the dataset at the specified index, handling indexing and image weighting.""" index = self.indices[index] # linear, shuffled, or image_weights hyp = self.hyp @@ -230,6 +235,7 @@ return (torch.from_numpy(img), labels_out, self.im_files[index], shapes, masks) def load_mosaic(self, index): + """Loads 1 image + 3 random images into a 4-image YOLOv5 mosaic, adjusting labels and segments accordingly.""" labels4, segments4 = [], [] s = self.img_size yc, xc = (int(random.uniform(-x, 2 * s + x)) for x in self.mosaic_border) # mosaic center x, y @@ -290,6 +296,7 @@ @staticmethod def collate_fn(batch): + """Custom collation function for DataLoader, batches images, labels, paths, shapes, and segmentation masks.""" img, label, path, shapes, masks = zip(*batch) # transposed batched_masks = torch.cat(masks, 0) for i, l in enumerate(label): @@ -298,6 +305,11 @@ def polygon2mask(img_size, polygons, color=1, downsample_ratio=1): + """ + Args: + img_size (tuple): The image size. + polygons (np.ndarray): [N, M], N is the number of polygons, M is the number of points(Be divided by 2). + """ mask = np.zeros(img_size, dtype=np.uint8) polygons = np.asarray(polygons) polygons = polygons.astype(np.int32) @@ -312,6 +324,12 @@ def polygons2masks(img_size, polygons, color, downsample_ratio=1): + """ + Args: + img_size (tuple): The image size. + polygons (list[np.ndarray]): each polygon is [N, M], N is the number of polygons, M is the number of points(Be + divided by 2). + """ masks = [] for si in range(len(polygons)): mask = polygon2mask(img_size, [polygons[si].reshape(-1)], color, downsample_ratio) @@ -320,6 +338,7 @@ def polygons2masks_overlap(img_size, segments, downsample_ratio=1): + """Return a (640, 640) overlap mask.""" masks = np.zeros( (img_size[0] // downsample_ratio, img_size[1] // downsample_ratio), dtype=np.int32 if len(segments) > 255 else np.uint8, @@ -342,4 +361,4 @@ mask = ms[i] * (i + 1) masks = masks + mask masks = np.clip(masks, a_min=0, a_max=i + 1) - return masks, index+ return masks, index
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/segment/dataloaders.py
Document my Python code with docstrings
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import argparse import json import os import subprocess import sys from multiprocessing.pool import ThreadPool from pathlib import Path import numpy as np import torch from tqdm import tqdm FILE = Path(__file__).resolve() ROOT = FILE.parents[1] # YOLOv5 root directory if str(ROOT) not in sys.path: sys.path.append(str(ROOT)) # add ROOT to PATH ROOT = Path(os.path.relpath(ROOT, Path.cwd())) # relative import torch.nn.functional as F from models.common import DetectMultiBackend from models.yolo import SegmentationModel from utils.callbacks import Callbacks from utils.general import ( LOGGER, NUM_THREADS, TQDM_BAR_FORMAT, Profile, check_dataset, check_img_size, check_requirements, check_yaml, coco80_to_coco91_class, colorstr, increment_path, non_max_suppression, print_args, scale_boxes, xywh2xyxy, xyxy2xywh, ) from utils.metrics import ConfusionMatrix, box_iou from utils.plots import output_to_target, plot_val_study from utils.segment.dataloaders import create_dataloader from utils.segment.general import mask_iou, process_mask, process_mask_native, scale_image from utils.segment.metrics import Metrics, ap_per_class_box_and_mask from utils.segment.plots import plot_images_and_masks from utils.torch_utils import de_parallel, select_device, smart_inference_mode def save_one_txt(predn, save_conf, shape, file): gn = torch.tensor(shape)[[1, 0, 1, 0]] # normalization gain whwh for *xyxy, conf, cls in predn.tolist(): xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh line = (cls, *xywh, conf) if save_conf else (cls, *xywh) # label format with open(file, "a") as f: f.write(("%g " * len(line)).rstrip() % line + "\n") def save_one_json(predn, jdict, path, class_map, pred_masks): from pycocotools.mask import encode def single_encode(x): rle = encode(np.asarray(x[:, :, None], order="F", dtype="uint8"))[0] rle["counts"] = rle["counts"].decode("utf-8") return rle image_id = int(path.stem) if path.stem.isnumeric() else path.stem box = xyxy2xywh(predn[:, :4]) # xywh box[:, :2] -= box[:, 2:] / 2 # xy center to top-left corner pred_masks = np.transpose(pred_masks, (2, 0, 1)) with ThreadPool(NUM_THREADS) as pool: rles = pool.map(single_encode, pred_masks) for i, (p, b) in enumerate(zip(predn.tolist(), box.tolist())): jdict.append( { "image_id": image_id, "category_id": class_map[int(p[5])], "bbox": [round(x, 3) for x in b], "score": round(p[4], 5), "segmentation": rles[i], } ) def process_batch(detections, labels, iouv, pred_masks=None, gt_masks=None, overlap=False, masks=False): if masks: if overlap: nl = len(labels) index = torch.arange(nl, device=gt_masks.device).view(nl, 1, 1) + 1 gt_masks = gt_masks.repeat(nl, 1, 1) # shape(1,640,640) -> (n,640,640) gt_masks = torch.where(gt_masks == index, 1.0, 0.0) if gt_masks.shape[1:] != pred_masks.shape[1:]: gt_masks = F.interpolate(gt_masks[None], pred_masks.shape[1:], mode="bilinear", align_corners=False)[0] gt_masks = gt_masks.gt_(0.5) iou = mask_iou(gt_masks.view(gt_masks.shape[0], -1), pred_masks.view(pred_masks.shape[0], -1)) else: # boxes iou = box_iou(labels[:, 1:], detections[:, :4]) correct = np.zeros((detections.shape[0], iouv.shape[0])).astype(bool) correct_class = labels[:, 0:1] == detections[:, 5] for i in range(len(iouv)): x = torch.where((iou >= iouv[i]) & correct_class) # IoU > threshold and classes match if x[0].shape[0]: matches = torch.cat((torch.stack(x, 1), iou[x[0], x[1]][:, None]), 1).cpu().numpy() # [label, detect, iou] if x[0].shape[0] > 1: matches = matches[matches[:, 2].argsort()[::-1]] matches = matches[np.unique(matches[:, 1], return_index=True)[1]] # matches = matches[matches[:, 2].argsort()[::-1]] matches = matches[np.unique(matches[:, 0], return_index=True)[1]] correct[matches[:, 1].astype(int), i] = True return torch.tensor(correct, dtype=torch.bool, device=iouv.device) @smart_inference_mode() def run( data, weights=None, # model.pt path(s) batch_size=32, # batch size imgsz=640, # inference size (pixels) conf_thres=0.001, # confidence threshold iou_thres=0.6, # NMS IoU threshold max_det=300, # maximum detections per image task="val", # train, val, test, speed or study device="", # cuda device, i.e. 0 or 0,1,2,3 or cpu workers=8, # max dataloader workers (per RANK in DDP mode) single_cls=False, # treat as single-class dataset augment=False, # augmented inference verbose=False, # verbose output save_txt=False, # save results to *.txt save_hybrid=False, # save label+prediction hybrid results to *.txt save_conf=False, # save confidences in --save-txt labels save_json=False, # save a COCO-JSON results file project=ROOT / "runs/val-seg", # save to project/name name="exp", # save to project/name exist_ok=False, # existing project/name ok, do not increment half=True, # use FP16 half-precision inference dnn=False, # use OpenCV DNN for ONNX inference model=None, dataloader=None, save_dir=Path(""), plots=True, overlap=False, mask_downsample_ratio=1, compute_loss=None, callbacks=Callbacks(), ): if save_json: check_requirements("pycocotools>=2.0.6") process = process_mask_native # more accurate else: process = process_mask # faster # Initialize/load model and set device training = model is not None if training: # called by train.py device, pt, jit, engine = next(model.parameters()).device, True, False, False # get model device, PyTorch model half &= device.type != "cpu" # half precision only supported on CUDA model.half() if half else model.float() nm = de_parallel(model).model[-1].nm # number of masks else: # called directly device = select_device(device, batch_size=batch_size) # Directories save_dir = increment_path(Path(project) / name, exist_ok=exist_ok) # increment run (save_dir / "labels" if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir # Load model model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half) stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine imgsz = check_img_size(imgsz, s=stride) # check image size half = model.fp16 # FP16 supported on limited backends with CUDA nm = de_parallel(model).model.model[-1].nm if isinstance(model, SegmentationModel) else 32 # number of masks if engine: batch_size = model.batch_size else: device = model.device if not (pt or jit): batch_size = 1 # export.py models default to batch-size 1 LOGGER.info(f"Forcing --batch-size 1 square inference (1,3,{imgsz},{imgsz}) for non-PyTorch models") # Data data = check_dataset(data) # check # Configure model.eval() cuda = device.type != "cpu" is_coco = isinstance(data.get("val"), str) and data["val"].endswith(f"coco{os.sep}val2017.txt") # COCO dataset nc = 1 if single_cls else int(data["nc"]) # number of classes iouv = torch.linspace(0.5, 0.95, 10, device=device) # iou vector for mAP@0.5:0.95 niou = iouv.numel() # Dataloader if not training: if pt and not single_cls: # check --weights are trained on --data ncm = model.model.nc assert ncm == nc, ( f"{weights} ({ncm} classes) trained on different --data than what you passed ({nc} " f"classes). Pass correct combination of --weights and --data that are trained together." ) model.warmup(imgsz=(1 if pt else batch_size, 3, imgsz, imgsz)) # warmup pad, rect = (0.0, False) if task == "speed" else (0.5, pt) # square inference for benchmarks task = task if task in ("train", "val", "test") else "val" # path to train/val/test images dataloader = create_dataloader( data[task], imgsz, batch_size, stride, single_cls, pad=pad, rect=rect, workers=workers, prefix=colorstr(f"{task}: "), overlap_mask=overlap, mask_downsample_ratio=mask_downsample_ratio, )[0] seen = 0 confusion_matrix = ConfusionMatrix(nc=nc) names = model.names if hasattr(model, "names") else model.module.names # get class names if isinstance(names, (list, tuple)): # old format names = dict(enumerate(names)) class_map = coco80_to_coco91_class() if is_coco else list(range(1000)) s = ("%22s" + "%11s" * 10) % ( "Class", "Images", "Instances", "Box(P", "R", "mAP50", "mAP50-95)", "Mask(P", "R", "mAP50", "mAP50-95)", ) dt = Profile(device=device), Profile(device=device), Profile(device=device) metrics = Metrics() loss = torch.zeros(4, device=device) jdict, stats = [], [] # callbacks.run('on_val_start') pbar = tqdm(dataloader, desc=s, bar_format=TQDM_BAR_FORMAT) # progress bar for batch_i, (im, targets, paths, shapes, masks) in enumerate(pbar): # callbacks.run('on_val_batch_start') with dt[0]: if cuda: im = im.to(device, non_blocking=True) targets = targets.to(device) masks = masks.to(device) masks = masks.float() im = im.half() if half else im.float() # uint8 to fp16/32 im /= 255 # 0 - 255 to 0.0 - 1.0 nb, _, height, width = im.shape # batch size, channels, height, width # Inference with dt[1]: preds, protos, train_out = model(im) if compute_loss else (*model(im, augment=augment)[:2], None) # Loss if compute_loss: loss += compute_loss((train_out, protos), targets, masks)[1] # box, obj, cls # NMS targets[:, 2:] *= torch.tensor((width, height, width, height), device=device) # to pixels lb = [targets[targets[:, 0] == i, 1:] for i in range(nb)] if save_hybrid else [] # for autolabelling with dt[2]: preds = non_max_suppression( preds, conf_thres, iou_thres, labels=lb, multi_label=True, agnostic=single_cls, max_det=max_det, nm=nm ) # Metrics plot_masks = [] # masks for plotting for si, (pred, proto) in enumerate(zip(preds, protos)): labels = targets[targets[:, 0] == si, 1:] nl, npr = labels.shape[0], pred.shape[0] # number of labels, predictions path, shape = Path(paths[si]), shapes[si][0] correct_masks = torch.zeros(npr, niou, dtype=torch.bool, device=device) # init correct_bboxes = torch.zeros(npr, niou, dtype=torch.bool, device=device) # init seen += 1 if npr == 0: if nl: stats.append((correct_masks, correct_bboxes, *torch.zeros((2, 0), device=device), labels[:, 0])) if plots: confusion_matrix.process_batch(detections=None, labels=labels[:, 0]) continue # Masks midx = [si] if overlap else targets[:, 0] == si gt_masks = masks[midx] pred_masks = process(proto, pred[:, 6:], pred[:, :4], shape=im[si].shape[1:]) # Predictions if single_cls: pred[:, 5] = 0 predn = pred.clone() scale_boxes(im[si].shape[1:], predn[:, :4], shape, shapes[si][1]) # native-space pred # Evaluate if nl: tbox = xywh2xyxy(labels[:, 1:5]) # target boxes scale_boxes(im[si].shape[1:], tbox, shape, shapes[si][1]) # native-space labels labelsn = torch.cat((labels[:, 0:1], tbox), 1) # native-space labels correct_bboxes = process_batch(predn, labelsn, iouv) correct_masks = process_batch(predn, labelsn, iouv, pred_masks, gt_masks, overlap=overlap, masks=True) if plots: confusion_matrix.process_batch(predn, labelsn) stats.append((correct_masks, correct_bboxes, pred[:, 4], pred[:, 5], labels[:, 0])) # (conf, pcls, tcls) pred_masks = torch.as_tensor(pred_masks, dtype=torch.uint8) if plots and batch_i < 3: plot_masks.append(pred_masks[:15]) # filter top 15 to plot # Save/log if save_txt: save_one_txt(predn, save_conf, shape, file=save_dir / "labels" / f"{path.stem}.txt") if save_json: pred_masks = scale_image( im[si].shape[1:], pred_masks.permute(1, 2, 0).contiguous().cpu().numpy(), shape, shapes[si][1] ) save_one_json(predn, jdict, path, class_map, pred_masks) # append to COCO-JSON dictionary # callbacks.run('on_val_image_end', pred, predn, path, names, im[si]) # Plot images if plots and batch_i < 3: if len(plot_masks): plot_masks = torch.cat(plot_masks, dim=0) plot_images_and_masks(im, targets, masks, paths, save_dir / f"val_batch{batch_i}_labels.jpg", names) plot_images_and_masks( im, output_to_target(preds, max_det=15), plot_masks, paths, save_dir / f"val_batch{batch_i}_pred.jpg", names, ) # pred # callbacks.run('on_val_batch_end') # Compute metrics stats = [torch.cat(x, 0).cpu().numpy() for x in zip(*stats)] # to numpy if len(stats) and stats[0].any(): results = ap_per_class_box_and_mask(*stats, plot=plots, save_dir=save_dir, names=names) metrics.update(results) nt = np.bincount(stats[4].astype(int), minlength=nc) # number of targets per class # Print results pf = "%22s" + "%11i" * 2 + "%11.3g" * 8 # print format LOGGER.info(pf % ("all", seen, nt.sum(), *metrics.mean_results())) if nt.sum() == 0: LOGGER.warning(f"WARNING ⚠️ no labels found in {task} set, can not compute metrics without labels") # Print results per class if (verbose or (nc < 50 and not training)) and nc > 1 and len(stats): for i, c in enumerate(metrics.ap_class_index): LOGGER.info(pf % (names[c], seen, nt[c], *metrics.class_result(i))) # Print speeds t = tuple(x.t / seen * 1e3 for x in dt) # speeds per image if not training: shape = (batch_size, 3, imgsz, imgsz) LOGGER.info(f"Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {shape}" % t) # Plots if plots: confusion_matrix.plot(save_dir=save_dir, names=list(names.values())) # callbacks.run('on_val_end') mp_bbox, mr_bbox, map50_bbox, map_bbox, mp_mask, mr_mask, map50_mask, map_mask = metrics.mean_results() # Save JSON if save_json and len(jdict): w = Path(weights[0] if isinstance(weights, list) else weights).stem if weights is not None else "" # weights anno_json = str(Path("../datasets/coco/annotations/instances_val2017.json")) # annotations pred_json = str(save_dir / f"{w}_predictions.json") # predictions LOGGER.info(f"\nEvaluating pycocotools mAP... saving {pred_json}...") with open(pred_json, "w") as f: json.dump(jdict, f) try: # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb from pycocotools.coco import COCO from pycocotools.cocoeval import COCOeval anno = COCO(anno_json) # init annotations api pred = anno.loadRes(pred_json) # init predictions api results = [] for eval in COCOeval(anno, pred, "bbox"), COCOeval(anno, pred, "segm"): if is_coco: eval.params.imgIds = [int(Path(x).stem) for x in dataloader.dataset.im_files] # img ID to evaluate eval.evaluate() eval.accumulate() eval.summarize() results.extend(eval.stats[:2]) # update results (mAP@0.5:0.95, mAP@0.5) map_bbox, map50_bbox, map_mask, map50_mask = results except Exception as e: LOGGER.info(f"pycocotools unable to run: {e}") # Return results model.float() # for training if not training: s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else "" LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}") final_metric = mp_bbox, mr_bbox, map50_bbox, map_bbox, mp_mask, mr_mask, map50_mask, map_mask return (*final_metric, *(loss.cpu() / len(dataloader)).tolist()), metrics.get_maps(nc), t def parse_opt(): parser = argparse.ArgumentParser() parser.add_argument("--data", type=str, default=ROOT / "data/coco128-seg.yaml", help="dataset.yaml path") parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-seg.pt", help="model path(s)") parser.add_argument("--batch-size", type=int, default=32, help="batch size") parser.add_argument("--imgsz", "--img", "--img-size", type=int, default=640, help="inference size (pixels)") parser.add_argument("--conf-thres", type=float, default=0.001, help="confidence threshold") parser.add_argument("--iou-thres", type=float, default=0.6, help="NMS IoU threshold") parser.add_argument("--max-det", type=int, default=300, help="maximum detections per image") parser.add_argument("--task", default="val", help="train, val, test, speed or study") parser.add_argument("--device", default="", help="cuda device, i.e. 0 or 0,1,2,3 or cpu") parser.add_argument("--workers", type=int, default=8, help="max dataloader workers (per RANK in DDP mode)") parser.add_argument("--single-cls", action="store_true", help="treat as single-class dataset") parser.add_argument("--augment", action="store_true", help="augmented inference") parser.add_argument("--verbose", action="store_true", help="report mAP by class") parser.add_argument("--save-txt", action="store_true", help="save results to *.txt") parser.add_argument("--save-hybrid", action="store_true", help="save label+prediction hybrid results to *.txt") parser.add_argument("--save-conf", action="store_true", help="save confidences in --save-txt labels") parser.add_argument("--save-json", action="store_true", help="save a COCO-JSON results file") parser.add_argument("--project", default=ROOT / "runs/val-seg", help="save results to project/name") parser.add_argument("--name", default="exp", help="save to project/name") parser.add_argument("--exist-ok", action="store_true", help="existing project/name ok, do not increment") parser.add_argument("--half", action="store_true", help="use FP16 half-precision inference") parser.add_argument("--dnn", action="store_true", help="use OpenCV DNN for ONNX inference") opt = parser.parse_args() opt.data = check_yaml(opt.data) # check YAML # opt.save_json |= opt.data.endswith('coco.yaml') opt.save_txt |= opt.save_hybrid print_args(vars(opt)) return opt def main(opt): check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) if opt.task in ("train", "val", "test"): # run normally if opt.conf_thres > 0.001: # https://github.com/ultralytics/yolov5/issues/1466 LOGGER.warning(f"WARNING ⚠️ confidence threshold {opt.conf_thres} > 0.001 produces invalid results") if opt.save_hybrid: LOGGER.warning("WARNING ⚠️ --save-hybrid returns high mAP from hybrid labels, not from predictions alone") run(**vars(opt)) else: weights = opt.weights if isinstance(opt.weights, list) else [opt.weights] opt.half = torch.cuda.is_available() and opt.device != "cpu" # FP16 for fastest results if opt.task == "speed": # speed benchmarks # python val.py --task speed --data coco.yaml --batch 1 --weights yolov5n.pt yolov5s.pt... opt.conf_thres, opt.iou_thres, opt.save_json = 0.25, 0.45, False for opt.weights in weights: run(**vars(opt), plots=False) elif opt.task == "study": # speed vs mAP benchmarks # python val.py --task study --data coco.yaml --iou 0.7 --weights yolov5n.pt yolov5s.pt... for opt.weights in weights: f = f"study_{Path(opt.data).stem}_{Path(opt.weights).stem}.txt" # filename to save to x, y = list(range(256, 1536 + 128, 128)), [] # x axis (image sizes), y axis for opt.imgsz in x: # img-size LOGGER.info(f"\nRunning {f} --imgsz {opt.imgsz}...") r, _, t = run(**vars(opt), plots=False) y.append(r + t) # results and times np.savetxt(f, y, fmt="%10.4g") # save subprocess.run(["zip", "-r", "study.zip", "study_*.txt"]) plot_val_study(x=x) # plot else: raise NotImplementedError(f'--task {opt.task} not in ("train", "val", "test", "speed", "study")') if __name__ == "__main__": opt = parse_opt() main(opt)
--- +++ @@ -1,4 +1,24 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +""" +Validate a trained YOLOv5 segment model on a segment dataset. + +Usage: + $ bash data/scripts/get_coco.sh --val --segments # download COCO-segments val split (1G, 5000 images) + $ python segment/val.py --weights yolov5s-seg.pt --data coco.yaml --img 640 # validate COCO-segments + +Usage - formats: + $ python segment/val.py --weights yolov5s-seg.pt # PyTorch + yolov5s-seg.torchscript # TorchScript + yolov5s-seg.onnx # ONNX Runtime or OpenCV DNN with --dnn + yolov5s-seg_openvino_label # OpenVINO + yolov5s-seg.engine # TensorRT + yolov5s-seg.mlmodel # CoreML (macOS-only) + yolov5s-seg_saved_model # TensorFlow SavedModel + yolov5s-seg.pb # TensorFlow GraphDef + yolov5s-seg.tflite # TensorFlow Lite + yolov5s-seg_edgetpu.tflite # TensorFlow Edge TPU + yolov5s-seg_paddle_model # PaddlePaddle +""" import argparse import json @@ -51,6 +71,9 @@ def save_one_txt(predn, save_conf, shape, file): + """Saves detection results in txt format; includes class, xywh (normalized), optionally confidence if `save_conf` is + True. + """ gn = torch.tensor(shape)[[1, 0, 1, 0]] # normalization gain whwh for *xyxy, conf, cls in predn.tolist(): xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist() # normalized xywh @@ -60,9 +83,14 @@ def save_one_json(predn, jdict, path, class_map, pred_masks): + """Saves a JSON file with detection results including bounding boxes, category IDs, scores, and segmentation masks. + + Example JSON result: {"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}. + """ from pycocotools.mask import encode def single_encode(x): + """Encodes binary mask arrays into RLE (Run-Length Encoding) format for JSON serialization.""" rle = encode(np.asarray(x[:, :, None], order="F", dtype="uint8"))[0] rle["counts"] = rle["counts"].decode("utf-8") return rle @@ -86,6 +114,15 @@ def process_batch(detections, labels, iouv, pred_masks=None, gt_masks=None, overlap=False, masks=False): + """Return correct prediction matrix. + + Args: + detections (array[N, 6]): x1, y1, x2, y2, conf, class + labels (array[M, 5]): class, x1, y1, x2, y2 + + Returns: + correct (array[N, 10]), for 10 IoU levels. + """ if masks: if overlap: nl = len(labels) @@ -147,6 +184,7 @@ compute_loss=None, callbacks=Callbacks(), ): + """Validate a YOLOv5 segmentation model on specified dataset, producing metrics, plots, and optional JSON output.""" if save_json: check_requirements("pycocotools>=2.0.6") process = process_mask_native # more accurate @@ -407,6 +445,9 @@ def parse_opt(): + """Parses command line arguments for configuring YOLOv5 options like dataset path, weights, batch size, and + inference settings. + """ parser = argparse.ArgumentParser() parser.add_argument("--data", type=str, default=ROOT / "data/coco128-seg.yaml", help="dataset.yaml path") parser.add_argument("--weights", nargs="+", type=str, default=ROOT / "yolov5s-seg.pt", help="model path(s)") @@ -439,6 +480,7 @@ def main(opt): + """Executes YOLOv5 tasks including training, validation, testing, speed, and study with configurable options.""" check_requirements(ROOT / "requirements.txt", exclude=("tensorboard", "thop")) if opt.task in ("train", "val", "test"): # run normally @@ -475,4 +517,4 @@ if __name__ == "__main__": opt = parse_opt() - main(opt)+ main(opt)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/segment/val.py
Help me document legacy Python code
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import math import os import platform import subprocess import time import warnings from contextlib import contextmanager from copy import deepcopy from pathlib import Path import torch import torch.distributed as dist import torch.nn as nn import torch.nn.functional as F from torch.nn.parallel import DistributedDataParallel as DDP from utils.general import LOGGER, check_version, colorstr, file_date, git_describe LOCAL_RANK = int(os.getenv("LOCAL_RANK", -1)) # https://pytorch.org/docs/stable/elastic/run.html RANK = int(os.getenv("RANK", -1)) WORLD_SIZE = int(os.getenv("WORLD_SIZE", 1)) try: import thop # for FLOPs computation except ImportError: thop = None # Suppress PyTorch warnings warnings.filterwarnings("ignore", message="User provided device_type of 'cuda', but CUDA is not available. Disabling") warnings.filterwarnings("ignore", category=UserWarning) def smart_inference_mode(torch_1_9=check_version(torch.__version__, "1.9.0")): def decorate(fn): return (torch.inference_mode if torch_1_9 else torch.no_grad)()(fn) return decorate def smartCrossEntropyLoss(label_smoothing=0.0): if check_version(torch.__version__, "1.10.0"): return nn.CrossEntropyLoss(label_smoothing=label_smoothing) if label_smoothing > 0: LOGGER.warning(f"WARNING ⚠️ label smoothing {label_smoothing} requires torch>=1.10.0") return nn.CrossEntropyLoss() def smart_DDP(model): assert not check_version(torch.__version__, "1.12.0", pinned=True), ( "torch==1.12.0 torchvision==0.13.0 DDP training is not supported due to a known issue. " "Please upgrade or downgrade torch to use DDP. See https://github.com/ultralytics/yolov5/issues/8395" ) if check_version(torch.__version__, "1.11.0"): return DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK, static_graph=True) else: return DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK) def reshape_classifier_output(model, n=1000): from models.common import Classify name, m = list((model.model if hasattr(model, "model") else model).named_children())[-1] # last module if isinstance(m, Classify): # YOLOv5 Classify() head if m.linear.out_features != n: m.linear = nn.Linear(m.linear.in_features, n) elif isinstance(m, nn.Linear): # ResNet, EfficientNet if m.out_features != n: setattr(model, name, nn.Linear(m.in_features, n)) elif isinstance(m, nn.Sequential): types = [type(x) for x in m] if nn.Linear in types: i = len(types) - 1 - types[::-1].index(nn.Linear) # last nn.Linear index if m[i].out_features != n: m[i] = nn.Linear(m[i].in_features, n) elif nn.Conv2d in types: i = len(types) - 1 - types[::-1].index(nn.Conv2d) # last nn.Conv2d index if m[i].out_channels != n: m[i] = nn.Conv2d(m[i].in_channels, n, m[i].kernel_size, m[i].stride, bias=m[i].bias is not None) @contextmanager def torch_distributed_zero_first(local_rank: int): if local_rank not in [-1, 0]: dist.barrier(device_ids=[local_rank]) yield if local_rank == 0: dist.barrier(device_ids=[0]) def device_count(): assert platform.system() in ("Linux", "Windows"), "device_count() only supported on Linux or Windows" try: cmd = "nvidia-smi -L | wc -l" if platform.system() == "Linux" else 'nvidia-smi -L | find /c /v ""' # Windows return int(subprocess.run(cmd, shell=True, capture_output=True, check=True).stdout.decode().split()[-1]) except Exception: return 0 def select_device(device="", batch_size=0, newline=True): s = f"YOLOv5 🚀 {git_describe() or file_date()} Python-{platform.python_version()} torch-{torch.__version__} " device = str(device).strip().lower().replace("cuda:", "").replace("none", "") # to string, 'cuda:0' to '0' cpu = device == "cpu" mps = device == "mps" # Apple Metal Performance Shaders (MPS) if cpu or mps: os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # force torch.cuda.is_available() = False elif device: # non-cpu device requested os.environ["CUDA_VISIBLE_DEVICES"] = device # set environment variable - must be before assert is_available() assert torch.cuda.is_available() and torch.cuda.device_count() >= len(device.replace(",", "")), ( f"Invalid CUDA '--device {device}' requested, use '--device cpu' or pass valid CUDA device(s)" ) if not cpu and not mps and torch.cuda.is_available(): # prefer GPU if available devices = device.split(",") if device else "0" # range(torch.cuda.device_count()) # i.e. 0,1,6,7 n = len(devices) # device count if n > 1 and batch_size > 0: # check batch_size is divisible by device_count assert batch_size % n == 0, f"batch-size {batch_size} not multiple of GPU count {n}" space = " " * (len(s) + 1) for i, d in enumerate(devices): p = torch.cuda.get_device_properties(i) s += f"{'' if i == 0 else space}CUDA:{d} ({p.name}, {p.total_memory / (1 << 20):.0f}MiB)\n" # bytes to MB arg = "cuda:0" elif mps and getattr(torch, "has_mps", False) and torch.backends.mps.is_available(): # prefer MPS if available s += "MPS\n" arg = "mps" else: # revert to CPU s += "CPU\n" arg = "cpu" if not newline: s = s.rstrip() LOGGER.info(s) return torch.device(arg) def time_sync(): if torch.cuda.is_available(): torch.cuda.synchronize() return time.time() def profile(input, ops, n=10, device=None): results = [] if not isinstance(device, torch.device): device = select_device(device) print( f"{'Params':>12s}{'GFLOPs':>12s}{'GPU_mem (GB)':>14s}{'forward (ms)':>14s}{'backward (ms)':>14s}" f"{'input':>24s}{'output':>24s}" ) for x in input if isinstance(input, list) else [input]: x = x.to(device) x.requires_grad = True for m in ops if isinstance(ops, list) else [ops]: m = m.to(device) if hasattr(m, "to") else m # device m = m.half() if hasattr(m, "half") and isinstance(x, torch.Tensor) and x.dtype is torch.float16 else m tf, tb, t = 0, 0, [0, 0, 0] # dt forward, backward try: flops = thop.profile(m, inputs=(x,), verbose=False)[0] / 1e9 * 2 # GFLOPs except Exception: flops = 0 try: for _ in range(n): t[0] = time_sync() y = m(x) t[1] = time_sync() try: _ = (sum(yi.sum() for yi in y) if isinstance(y, list) else y).sum().backward() t[2] = time_sync() except Exception: # no backward method # print(e) # for debug t[2] = float("nan") tf += (t[1] - t[0]) * 1000 / n # ms per op forward tb += (t[2] - t[1]) * 1000 / n # ms per op backward mem = torch.cuda.memory_reserved() / 1e9 if torch.cuda.is_available() else 0 # (GB) s_in, s_out = (tuple(x.shape) if isinstance(x, torch.Tensor) else "list" for x in (x, y)) # shapes p = sum(x.numel() for x in m.parameters()) if isinstance(m, nn.Module) else 0 # parameters print(f"{p:12}{flops:12.4g}{mem:>14.3f}{tf:14.4g}{tb:14.4g}{s_in!s:>24s}{s_out!s:>24s}") results.append([p, flops, mem, tf, tb, s_in, s_out]) except Exception as e: print(e) results.append(None) torch.cuda.empty_cache() return results def is_parallel(model): return type(model) in (nn.parallel.DataParallel, nn.parallel.DistributedDataParallel) def de_parallel(model): return model.module if is_parallel(model) else model def initialize_weights(model): for m in model.modules(): t = type(m) if t is nn.Conv2d: pass # nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu') elif t is nn.BatchNorm2d: m.eps = 1e-3 m.momentum = 0.03 elif t in [nn.Hardswish, nn.LeakyReLU, nn.ReLU, nn.ReLU6, nn.SiLU]: m.inplace = True def find_modules(model, mclass=nn.Conv2d): return [i for i, m in enumerate(model.module_list) if isinstance(m, mclass)] def sparsity(model): a, b = 0, 0 for p in model.parameters(): a += p.numel() b += (p == 0).sum() return b / a def prune(model, amount=0.3): import torch.nn.utils.prune as prune for name, m in model.named_modules(): if isinstance(m, nn.Conv2d): prune.l1_unstructured(m, name="weight", amount=amount) # prune prune.remove(m, "weight") # make permanent LOGGER.info(f"Model pruned to {sparsity(model):.3g} global sparsity") def fuse_conv_and_bn(conv, bn): fusedconv = ( nn.Conv2d( conv.in_channels, conv.out_channels, kernel_size=conv.kernel_size, stride=conv.stride, padding=conv.padding, dilation=conv.dilation, groups=conv.groups, bias=True, ) .requires_grad_(False) .to(conv.weight.device) ) # Prepare filters w_conv = conv.weight.clone().view(conv.out_channels, -1) w_bn = torch.diag(bn.weight.div(torch.sqrt(bn.eps + bn.running_var))) fusedconv.weight.copy_(torch.mm(w_bn, w_conv).view(fusedconv.weight.shape)) # Prepare spatial bias b_conv = torch.zeros(conv.weight.size(0), device=conv.weight.device) if conv.bias is None else conv.bias b_bn = bn.bias - bn.weight.mul(bn.running_mean).div(torch.sqrt(bn.running_var + bn.eps)) fusedconv.bias.copy_(torch.mm(w_bn, b_conv.reshape(-1, 1)).reshape(-1) + b_bn) return fusedconv def model_info(model, verbose=False, imgsz=640): n_p = sum(x.numel() for x in model.parameters()) # number parameters n_g = sum(x.numel() for x in model.parameters() if x.requires_grad) # number gradients if verbose: print(f"{'layer':>5} {'name':>40} {'gradient':>9} {'parameters':>12} {'shape':>20} {'mu':>10} {'sigma':>10}") for i, (name, p) in enumerate(model.named_parameters()): name = name.replace("module_list.", "") print( "%5g %40s %9s %12g %20s %10.3g %10.3g" % (i, name, p.requires_grad, p.numel(), list(p.shape), p.mean(), p.std()) ) try: # FLOPs p = next(model.parameters()) stride = max(int(model.stride.max()), 32) if hasattr(model, "stride") else 32 # max stride im = torch.empty((1, p.shape[1], stride, stride), device=p.device) # input image in BCHW format flops = thop.profile(deepcopy(model), inputs=(im,), verbose=False)[0] / 1e9 * 2 # stride GFLOPs imgsz = imgsz if isinstance(imgsz, list) else [imgsz, imgsz] # expand if int/float fs = f", {flops * imgsz[0] / stride * imgsz[1] / stride:.1f} GFLOPs" # 640x640 GFLOPs except Exception: fs = "" name = Path(model.yaml_file).stem.replace("yolov5", "YOLOv5") if hasattr(model, "yaml_file") else "Model" LOGGER.info(f"{name} summary: {len(list(model.modules()))} layers, {n_p} parameters, {n_g} gradients{fs}") def scale_img(img, ratio=1.0, same_shape=False, gs=32): # img(16,3,256,416) if ratio == 1.0: return img h, w = img.shape[2:] s = (int(h * ratio), int(w * ratio)) # new size img = F.interpolate(img, size=s, mode="bilinear", align_corners=False) # resize if not same_shape: # pad/crop img h, w = (math.ceil(x * ratio / gs) * gs for x in (h, w)) return F.pad(img, [0, w - s[1], 0, h - s[0]], value=0.447) # value = imagenet mean def copy_attr(a, b, include=(), exclude=()): for k, v in b.__dict__.items(): if (len(include) and k not in include) or k.startswith("_") or k in exclude: continue else: setattr(a, k, v) def smart_optimizer(model, name="Adam", lr=0.001, momentum=0.9, decay=1e-5): g = [], [], [] # optimizer parameter groups bn = tuple(v for k, v in nn.__dict__.items() if "Norm" in k) # normalization layers, i.e. BatchNorm2d() for v in model.modules(): for p_name, p in v.named_parameters(recurse=0): if p_name == "bias": # bias (no decay) g[2].append(p) elif p_name == "weight" and isinstance(v, bn): # weight (no decay) g[1].append(p) else: g[0].append(p) # weight (with decay) if name == "Adam": optimizer = torch.optim.Adam(g[2], lr=lr, betas=(momentum, 0.999)) # adjust beta1 to momentum elif name == "AdamW": optimizer = torch.optim.AdamW(g[2], lr=lr, betas=(momentum, 0.999), weight_decay=0.0) elif name == "RMSProp": optimizer = torch.optim.RMSprop(g[2], lr=lr, momentum=momentum) elif name == "SGD": optimizer = torch.optim.SGD(g[2], lr=lr, momentum=momentum, nesterov=True) else: raise NotImplementedError(f"Optimizer {name} not implemented.") optimizer.add_param_group({"params": g[0], "weight_decay": decay}) # add g0 with weight_decay optimizer.add_param_group({"params": g[1], "weight_decay": 0.0}) # add g1 (BatchNorm2d weights) LOGGER.info( f"{colorstr('optimizer:')} {type(optimizer).__name__}(lr={lr}) with parameter groups " f"{len(g[1])} weight(decay=0.0), {len(g[0])} weight(decay={decay}), {len(g[2])} bias" ) return optimizer def smart_hub_load(repo="ultralytics/yolov5", model="yolov5s", **kwargs): if check_version(torch.__version__, "1.9.1"): kwargs["skip_validation"] = True # validation causes GitHub API rate limit errors if check_version(torch.__version__, "1.12.0"): kwargs["trust_repo"] = True # argument required starting in torch 0.12 try: return torch.hub.load(repo, model, **kwargs) except Exception: return torch.hub.load(repo, model, force_reload=True, **kwargs) def smart_resume(ckpt, optimizer, ema=None, weights="yolov5s.pt", epochs=300, resume=True): best_fitness = 0.0 start_epoch = ckpt["epoch"] + 1 if ckpt["optimizer"] is not None: optimizer.load_state_dict(ckpt["optimizer"]) # optimizer best_fitness = ckpt["best_fitness"] if ema and ckpt.get("ema"): ema.ema.load_state_dict(ckpt["ema"].float().state_dict()) # EMA ema.updates = ckpt["updates"] if resume: assert start_epoch > 0, ( f"{weights} training to {epochs} epochs is finished, nothing to resume.\n" f"Start a new training without --resume, i.e. 'python train.py --weights {weights}'" ) LOGGER.info(f"Resuming training from {weights} from epoch {start_epoch} to {epochs} total epochs") if epochs < start_epoch: LOGGER.info(f"{weights} has been trained for {ckpt['epoch']} epochs. Fine-tuning for {epochs} more epochs.") epochs += ckpt["epoch"] # finetune additional epochs return best_fitness, start_epoch, epochs class EarlyStopping: def __init__(self, patience=30): self.best_fitness = 0.0 # i.e. mAP self.best_epoch = 0 self.patience = patience or float("inf") # epochs to wait after fitness stops improving to stop self.possible_stop = False # possible stop may occur next epoch def __call__(self, epoch, fitness): if fitness >= self.best_fitness: # >= 0 to allow for early zero-fitness stage of training self.best_epoch = epoch self.best_fitness = fitness delta = epoch - self.best_epoch # epochs without improvement self.possible_stop = delta >= (self.patience - 1) # possible stop may occur next epoch stop = delta >= self.patience # stop training if patience exceeded if stop: LOGGER.info( f"Stopping training early as no improvement observed in last {self.patience} epochs. " f"Best results observed at epoch {self.best_epoch}, best model saved as best.pt.\n" f"To update EarlyStopping(patience={self.patience}) pass a new patience value, " f"i.e. `python train.py --patience 300` or use `--patience 0` to disable EarlyStopping." ) return stop class ModelEMA: def __init__(self, model, decay=0.9999, tau=2000, updates=0): self.ema = deepcopy(de_parallel(model)).eval() # FP32 EMA self.updates = updates # number of EMA updates self.decay = lambda x: decay * (1 - math.exp(-x / tau)) # decay exponential ramp (to help early epochs) for p in self.ema.parameters(): p.requires_grad_(False) def update(self, model): self.updates += 1 d = self.decay(self.updates) msd = de_parallel(model).state_dict() # model state_dict for k, v in self.ema.state_dict().items(): if v.dtype.is_floating_point: # true for FP16 and FP32 v *= d v += (1 - d) * msd[k].detach() # assert v.dtype == msd[k].dtype == torch.float32, f'{k}: EMA {v.dtype} and model {msd[k].dtype} must be FP32' def update_attr(self, model, include=(), exclude=("process_group", "reducer")): copy_attr(self.ema, model, include, exclude)
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""PyTorch utils.""" import math import os @@ -33,14 +34,17 @@ def smart_inference_mode(torch_1_9=check_version(torch.__version__, "1.9.0")): + """Applies torch.inference_mode() if torch>=1.9.0, else torch.no_grad() as a decorator for functions.""" def decorate(fn): + """Applies torch.inference_mode() if torch>=1.9.0, else torch.no_grad() to the decorated function.""" return (torch.inference_mode if torch_1_9 else torch.no_grad)()(fn) return decorate def smartCrossEntropyLoss(label_smoothing=0.0): + """Return CrossEntropyLoss with optional label smoothing for torch>=1.10.0; warns if smoothing on lower versions.""" if check_version(torch.__version__, "1.10.0"): return nn.CrossEntropyLoss(label_smoothing=label_smoothing) if label_smoothing > 0: @@ -49,6 +53,7 @@ def smart_DDP(model): + """Initializes DistributedDataParallel (DDP) for model training, respecting torch version constraints.""" assert not check_version(torch.__version__, "1.12.0", pinned=True), ( "torch==1.12.0 torchvision==0.13.0 DDP training is not supported due to a known issue. " "Please upgrade or downgrade torch to use DDP. See https://github.com/ultralytics/yolov5/issues/8395" @@ -60,6 +65,7 @@ def reshape_classifier_output(model, n=1000): + """Reshapes last layer of model to match class count 'n', supporting Classify, Linear, Sequential types.""" from models.common import Classify name, m = list((model.model if hasattr(model, "model") else model).named_children())[-1] # last module @@ -83,6 +89,9 @@ @contextmanager def torch_distributed_zero_first(local_rank: int): + """Context manager ensuring ordered operations in distributed training by making all processes wait for the leading + process. + """ if local_rank not in [-1, 0]: dist.barrier(device_ids=[local_rank]) yield @@ -91,6 +100,7 @@ def device_count(): + """Returns the number of available CUDA devices; works on Linux and Windows by invoking `nvidia-smi`.""" assert platform.system() in ("Linux", "Windows"), "device_count() only supported on Linux or Windows" try: cmd = "nvidia-smi -L | wc -l" if platform.system() == "Linux" else 'nvidia-smi -L | find /c /v ""' # Windows @@ -100,6 +110,7 @@ def select_device(device="", batch_size=0, newline=True): + """Selects computing device (CPU, CUDA GPU, MPS) for YOLOv5 model deployment, logging device info.""" s = f"YOLOv5 🚀 {git_describe() or file_date()} Python-{platform.python_version()} torch-{torch.__version__} " device = str(device).strip().lower().replace("cuda:", "").replace("none", "") # to string, 'cuda:0' to '0' cpu = device == "cpu" @@ -136,12 +147,21 @@ def time_sync(): + """Synchronizes PyTorch for accurate timing, leveraging CUDA if available, and returns the current time.""" if torch.cuda.is_available(): torch.cuda.synchronize() return time.time() def profile(input, ops, n=10, device=None): + """YOLOv5 speed/memory/FLOPs profiler. + + Examples: + >>> input = torch.randn(16, 3, 640, 640) + >>> m1 = lambda x: x * torch.sigmoid(x) + >>> m2 = nn.SiLU() + >>> profile(input, [m1, m2], n=100) # profile over 100 iterations. + """ results = [] if not isinstance(device, torch.device): device = select_device(device) @@ -188,14 +208,19 @@ def is_parallel(model): + """Checks if the model is using Data Parallelism (DP) or Distributed Data Parallelism (DDP).""" return type(model) in (nn.parallel.DataParallel, nn.parallel.DistributedDataParallel) def de_parallel(model): + """Returns a single-GPU model by removing Data Parallelism (DP) or Distributed Data Parallelism (DDP) if applied.""" return model.module if is_parallel(model) else model def initialize_weights(model): + """Initializes weights of Conv2d, BatchNorm2d, and activations (Hardswish, LeakyReLU, ReLU, ReLU6, SiLU) in the + model. + """ for m in model.modules(): t = type(m) if t is nn.Conv2d: @@ -208,10 +233,12 @@ def find_modules(model, mclass=nn.Conv2d): + """Finds and returns list of layer indices in `model.module_list` matching the specified `mclass`.""" return [i for i, m in enumerate(model.module_list) if isinstance(m, mclass)] def sparsity(model): + """Calculate global sparsity of a model as the ratio of zero-valued parameters to total parameters.""" a, b = 0, 0 for p in model.parameters(): a += p.numel() @@ -220,6 +247,7 @@ def prune(model, amount=0.3): + """Prunes Conv2d layers in a model to a specified sparsity using L1 unstructured pruning.""" import torch.nn.utils.prune as prune for name, m in model.named_modules(): @@ -230,6 +258,10 @@ def fuse_conv_and_bn(conv, bn): + """Fuses Conv2d and BatchNorm2d layers into a single Conv2d layer. + + See https://tehnokv.com/posts/fusing-batchnorm-and-conv/. + """ fusedconv = ( nn.Conv2d( conv.in_channels, @@ -259,6 +291,10 @@ def model_info(model, verbose=False, imgsz=640): + """Prints model summary including layers, parameters, gradients, and FLOPs; imgsz may be int or list. + + Example: img_size=640 or img_size=[640, 320] + """ n_p = sum(x.numel() for x in model.parameters()) # number parameters n_g = sum(x.numel() for x in model.parameters() if x.requires_grad) # number gradients if verbose: @@ -285,6 +321,9 @@ def scale_img(img, ratio=1.0, same_shape=False, gs=32): # img(16,3,256,416) + """Scales an image tensor `img` of shape (bs,3,y,x) by `ratio`, optionally maintaining the original shape, padded to + multiples of `gs`. + """ if ratio == 1.0: return img h, w = img.shape[2:] @@ -296,6 +335,7 @@ def copy_attr(a, b, include=(), exclude=()): + """Copies attributes from object b to a, optionally filtering with include and exclude lists.""" for k, v in b.__dict__.items(): if (len(include) and k not in include) or k.startswith("_") or k in exclude: continue @@ -304,6 +344,10 @@ def smart_optimizer(model, name="Adam", lr=0.001, momentum=0.9, decay=1e-5): + """Initializes YOLOv5 smart optimizer with 3 parameter groups for different decay configurations. + + Groups are 0) weights with decay, 1) weights no decay, 2) biases no decay. + """ g = [], [], [] # optimizer parameter groups bn = tuple(v for k, v in nn.__dict__.items() if "Norm" in k) # normalization layers, i.e. BatchNorm2d() for v in model.modules(): @@ -336,6 +380,7 @@ def smart_hub_load(repo="ultralytics/yolov5", model="yolov5s", **kwargs): + """YOLOv5 torch.hub.load() wrapper with smart error handling, adjusting torch arguments for compatibility.""" if check_version(torch.__version__, "1.9.1"): kwargs["skip_validation"] = True # validation causes GitHub API rate limit errors if check_version(torch.__version__, "1.12.0"): @@ -347,6 +392,7 @@ def smart_resume(ckpt, optimizer, ema=None, weights="yolov5s.pt", epochs=300, resume=True): + """Resumes training from a checkpoint, updating optimizer, ema, and epochs, with optional resume verification.""" best_fitness = 0.0 start_epoch = ckpt["epoch"] + 1 if ckpt["optimizer"] is not None: @@ -368,14 +414,17 @@ class EarlyStopping: + """Implements early stopping to halt training when no improvement is observed for a specified number of epochs.""" def __init__(self, patience=30): + """Initializes simple early stopping mechanism for YOLOv5, with adjustable patience for non-improving epochs.""" self.best_fitness = 0.0 # i.e. mAP self.best_epoch = 0 self.patience = patience or float("inf") # epochs to wait after fitness stops improving to stop self.possible_stop = False # possible stop may occur next epoch def __call__(self, epoch, fitness): + """Evaluates if training should stop based on fitness improvement and patience, returning a boolean.""" if fitness >= self.best_fitness: # >= 0 to allow for early zero-fitness stage of training self.best_epoch = epoch self.best_fitness = fitness @@ -393,8 +442,15 @@ class ModelEMA: + """Updated Exponential Moving Average (EMA) from https://github.com/rwightman/pytorch-image-models Keeps a moving + average of everything in the model state_dict (parameters and buffers) For EMA details + see https://www.tensorflow.org/api_docs/python/tf/train/ExponentialMovingAverage. + """ def __init__(self, model, decay=0.9999, tau=2000, updates=0): + """Initializes EMA with model parameters, decay rate, tau for decay adjustment, and update count; sets model to + evaluation mode. + """ self.ema = deepcopy(de_parallel(model)).eval() # FP32 EMA self.updates = updates # number of EMA updates self.decay = lambda x: decay * (1 - math.exp(-x / tau)) # decay exponential ramp (to help early epochs) @@ -402,6 +458,7 @@ p.requires_grad_(False) def update(self, model): + """Updates the Exponential Moving Average (EMA) parameters based on the current model's parameters.""" self.updates += 1 d = self.decay(self.updates) @@ -413,4 +470,7 @@ # assert v.dtype == msd[k].dtype == torch.float32, f'{k}: EMA {v.dtype} and model {msd[k].dtype} must be FP32' def update_attr(self, model, include=(), exclude=("process_group", "reducer")): - copy_attr(self.ema, model, include, exclude)+ """Updates EMA attributes by copying specified attributes from model to EMA, excluding certain attributes by + default. + """ + copy_attr(self.ema, model, include, exclude)
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/torch_utils.py
Generate descriptive docstrings automatically
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import torch import torch.nn as nn from utils.metrics import bbox_iou from utils.torch_utils import de_parallel def smooth_BCE(eps=0.1): return 1.0 - 0.5 * eps, 0.5 * eps class BCEBlurWithLogitsLoss(nn.Module): def __init__(self, alpha=0.05): super().__init__() self.loss_fcn = nn.BCEWithLogitsLoss(reduction="none") # must be nn.BCEWithLogitsLoss() self.alpha = alpha def forward(self, pred, true): loss = self.loss_fcn(pred, true) pred = torch.sigmoid(pred) # prob from logits dx = pred - true # reduce only missing label effects # dx = (pred - true).abs() # reduce missing label and false label effects alpha_factor = 1 - torch.exp((dx - 1) / (self.alpha + 1e-4)) loss *= alpha_factor return loss.mean() class FocalLoss(nn.Module): def __init__(self, loss_fcn, gamma=1.5, alpha=0.25): super().__init__() self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss() self.gamma = gamma self.alpha = alpha self.reduction = loss_fcn.reduction self.loss_fcn.reduction = "none" # required to apply FL to each element def forward(self, pred, true): loss = self.loss_fcn(pred, true) # p_t = torch.exp(-loss) # loss *= self.alpha * (1.000001 - p_t) ** self.gamma # non-zero power for gradient stability # TF implementation https://github.com/tensorflow/addons/blob/v0.7.1/tensorflow_addons/losses/focal_loss.py pred_prob = torch.sigmoid(pred) # prob from logits p_t = true * pred_prob + (1 - true) * (1 - pred_prob) alpha_factor = true * self.alpha + (1 - true) * (1 - self.alpha) modulating_factor = (1.0 - p_t) ** self.gamma loss *= alpha_factor * modulating_factor if self.reduction == "mean": return loss.mean() elif self.reduction == "sum": return loss.sum() else: # 'none' return loss class QFocalLoss(nn.Module): def __init__(self, loss_fcn, gamma=1.5, alpha=0.25): super().__init__() self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss() self.gamma = gamma self.alpha = alpha self.reduction = loss_fcn.reduction self.loss_fcn.reduction = "none" # required to apply FL to each element def forward(self, pred, true): loss = self.loss_fcn(pred, true) pred_prob = torch.sigmoid(pred) # prob from logits alpha_factor = true * self.alpha + (1 - true) * (1 - self.alpha) modulating_factor = torch.abs(true - pred_prob) ** self.gamma loss *= alpha_factor * modulating_factor if self.reduction == "mean": return loss.mean() elif self.reduction == "sum": return loss.sum() else: # 'none' return loss class ComputeLoss: sort_obj_iou = False # Compute losses def __init__(self, model, autobalance=False): device = next(model.parameters()).device # get model device h = model.hyp # hyperparameters # Define criteria BCEcls = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h["cls_pw"]], device=device)) BCEobj = nn.BCEWithLogitsLoss(pos_weight=torch.tensor([h["obj_pw"]], device=device)) # Class label smoothing https://arxiv.org/pdf/1902.04103.pdf eqn 3 self.cp, self.cn = smooth_BCE(eps=h.get("label_smoothing", 0.0)) # positive, negative BCE targets # Focal loss g = h["fl_gamma"] # focal loss gamma if g > 0: BCEcls, BCEobj = FocalLoss(BCEcls, g), FocalLoss(BCEobj, g) m = de_parallel(model).model[-1] # Detect() module self.balance = {3: [4.0, 1.0, 0.4]}.get(m.nl, [4.0, 1.0, 0.25, 0.06, 0.02]) # P3-P7 self.ssi = list(m.stride).index(16) if autobalance else 0 # stride 16 index self.BCEcls, self.BCEobj, self.gr, self.hyp, self.autobalance = BCEcls, BCEobj, 1.0, h, autobalance self.na = m.na # number of anchors self.nc = m.nc # number of classes self.nl = m.nl # number of layers self.anchors = m.anchors self.device = device def __call__(self, p, targets): # predictions, targets lcls = torch.zeros(1, device=self.device) # class loss lbox = torch.zeros(1, device=self.device) # box loss lobj = torch.zeros(1, device=self.device) # object loss tcls, tbox, indices, anchors = self.build_targets(p, targets) # targets # Losses for i, pi in enumerate(p): # layer index, layer predictions b, a, gj, gi = indices[i] # image, anchor, gridy, gridx tobj = torch.zeros(pi.shape[:4], dtype=pi.dtype, device=self.device) # target obj if n := b.shape[0]: # pxy, pwh, _, pcls = pi[b, a, gj, gi].tensor_split((2, 4, 5), dim=1) # faster, requires torch 1.8.0 pxy, pwh, _, pcls = pi[b, a, gj, gi].split((2, 2, 1, self.nc), 1) # target-subset of predictions # Regression pxy = pxy.sigmoid() * 2 - 0.5 pwh = (pwh.sigmoid() * 2) ** 2 * anchors[i] pbox = torch.cat((pxy, pwh), 1) # predicted box iou = bbox_iou(pbox, tbox[i], CIoU=True).squeeze() # iou(prediction, target) lbox += (1.0 - iou).mean() # iou loss # Objectness iou = iou.detach().clamp(0).type(tobj.dtype) if self.sort_obj_iou: j = iou.argsort() b, a, gj, gi, iou = b[j], a[j], gj[j], gi[j], iou[j] if self.gr < 1: iou = (1.0 - self.gr) + self.gr * iou tobj[b, a, gj, gi] = iou # iou ratio # Classification if self.nc > 1: # cls loss (only if multiple classes) t = torch.full_like(pcls, self.cn, device=self.device) # targets t[range(n), tcls[i]] = self.cp lcls += self.BCEcls(pcls, t) # BCE obji = self.BCEobj(pi[..., 4], tobj) lobj += obji * self.balance[i] # obj loss if self.autobalance: self.balance[i] = self.balance[i] * 0.9999 + 0.0001 / obji.detach().item() if self.autobalance: self.balance = [x / self.balance[self.ssi] for x in self.balance] lbox *= self.hyp["box"] lobj *= self.hyp["obj"] lcls *= self.hyp["cls"] bs = tobj.shape[0] # batch size return (lbox + lobj + lcls) * bs, torch.cat((lbox, lobj, lcls)).detach() def build_targets(self, p, targets): na, nt = self.na, targets.shape[0] # number of anchors, targets tcls, tbox, indices, anch = [], [], [], [] gain = torch.ones(7, device=self.device) # normalized to gridspace gain ai = torch.arange(na, device=self.device).float().view(na, 1).repeat(1, nt) # same as .repeat_interleave(nt) targets = torch.cat((targets.repeat(na, 1, 1), ai[..., None]), 2) # append anchor indices g = 0.5 # bias off = ( torch.tensor( [ [0, 0], [1, 0], [0, 1], [-1, 0], [0, -1], # j,k,l,m # [1, 1], [1, -1], [-1, 1], [-1, -1], # jk,jm,lk,lm ], device=self.device, ).float() * g ) # offsets for i in range(self.nl): anchors, shape = self.anchors[i], p[i].shape gain[2:6] = torch.tensor(shape)[[3, 2, 3, 2]] # xyxy gain # Match targets to anchors t = targets * gain # shape(3,n,7) if nt: # Matches r = t[..., 4:6] / anchors[:, None] # wh ratio j = torch.max(r, 1 / r).max(2)[0] < self.hyp["anchor_t"] # compare # j = wh_iou(anchors, t[:, 4:6]) > model.hyp['iou_t'] # iou(3,n)=wh_iou(anchors(3,2), gwh(n,2)) t = t[j] # filter # Offsets gxy = t[:, 2:4] # grid xy gxi = gain[[2, 3]] - gxy # inverse j, k = ((gxy % 1 < g) & (gxy > 1)).T l, m = ((gxi % 1 < g) & (gxi > 1)).T j = torch.stack((torch.ones_like(j), j, k, l, m)) t = t.repeat((5, 1, 1))[j] offsets = (torch.zeros_like(gxy)[None] + off[:, None])[j] else: t = targets[0] offsets = 0 # Define bc, gxy, gwh, a = t.chunk(4, 1) # (image, class), grid xy, grid wh, anchors a, (b, c) = a.long().view(-1), bc.long().T # anchors, image, class gij = (gxy - offsets).long() gi, gj = gij.T # grid indices # Append indices.append((b, a, gj.clamp_(0, shape[2] - 1), gi.clamp_(0, shape[3] - 1))) # image, anchor, grid tbox.append(torch.cat((gxy - gij, gwh), 1)) # box anch.append(anchors[a]) # anchors tcls.append(c) # class return tcls, tbox, indices, anch
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Loss functions.""" import torch import torch.nn as nn @@ -8,17 +9,27 @@ def smooth_BCE(eps=0.1): + """Returns label smoothing BCE targets for reducing overfitting; pos: `1.0 - 0.5*eps`, neg: `0.5*eps`. For details + see https://github.com/ultralytics/yolov3/issues/238#issuecomment-598028441. + """ return 1.0 - 0.5 * eps, 0.5 * eps class BCEBlurWithLogitsLoss(nn.Module): + """Modified BCEWithLogitsLoss to reduce missing label effects in YOLOv5 training with optional alpha smoothing.""" def __init__(self, alpha=0.05): + """Initializes a modified BCEWithLogitsLoss with reduced missing label effects, taking optional alpha smoothing + parameter. + """ super().__init__() self.loss_fcn = nn.BCEWithLogitsLoss(reduction="none") # must be nn.BCEWithLogitsLoss() self.alpha = alpha def forward(self, pred, true): + """Computes modified BCE loss for YOLOv5 with reduced missing label effects, taking pred and true tensors, + returns mean loss. + """ loss = self.loss_fcn(pred, true) pred = torch.sigmoid(pred) # prob from logits dx = pred - true # reduce only missing label effects @@ -29,8 +40,12 @@ class FocalLoss(nn.Module): + """Applies focal loss to address class imbalance by modifying BCEWithLogitsLoss with gamma and alpha parameters.""" def __init__(self, loss_fcn, gamma=1.5, alpha=0.25): + """Initializes FocalLoss with specified loss function, gamma, and alpha values; modifies loss reduction to + 'none'. + """ super().__init__() self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss() self.gamma = gamma @@ -39,6 +54,7 @@ self.loss_fcn.reduction = "none" # required to apply FL to each element def forward(self, pred, true): + """Calculates the focal loss between predicted and true labels using a modified BCEWithLogitsLoss.""" loss = self.loss_fcn(pred, true) # p_t = torch.exp(-loss) # loss *= self.alpha * (1.000001 - p_t) ** self.gamma # non-zero power for gradient stability @@ -59,8 +75,10 @@ class QFocalLoss(nn.Module): + """Implements Quality Focal Loss to address class imbalance by modulating loss based on prediction confidence.""" def __init__(self, loss_fcn, gamma=1.5, alpha=0.25): + """Initializes Quality Focal Loss with given loss function, gamma, alpha; modifies reduction to 'none'.""" super().__init__() self.loss_fcn = loss_fcn # must be nn.BCEWithLogitsLoss() self.gamma = gamma @@ -69,6 +87,9 @@ self.loss_fcn.reduction = "none" # required to apply FL to each element def forward(self, pred, true): + """Computes the focal loss between `pred` and `true` using BCEWithLogitsLoss, adjusting for imbalance with + `gamma` and `alpha`. + """ loss = self.loss_fcn(pred, true) pred_prob = torch.sigmoid(pred) # prob from logits @@ -85,11 +106,13 @@ class ComputeLoss: + """Computes the total loss for YOLOv5 model predictions, including classification, box, and objectness losses.""" sort_obj_iou = False # Compute losses def __init__(self, model, autobalance=False): + """Initializes ComputeLoss with model and autobalance option, autobalances losses if True.""" device = next(model.parameters()).device # get model device h = model.hyp # hyperparameters @@ -116,6 +139,7 @@ self.device = device def __call__(self, p, targets): # predictions, targets + """Performs forward pass, calculating class, box, and object loss for given predictions and targets.""" lcls = torch.zeros(1, device=self.device) # class loss lbox = torch.zeros(1, device=self.device) # box loss lobj = torch.zeros(1, device=self.device) # object loss @@ -167,6 +191,9 @@ return (lbox + lobj + lcls) * bs, torch.cat((lbox, lobj, lcls)).detach() def build_targets(self, p, targets): + """Prepares model targets from input targets (image,class,x,y,w,h) for loss computation, returning class, box, + indices, and anchors. + """ na, nt = self.na, targets.shape[0] # number of anchors, targets tcls, tbox, indices, anch = [], [], [], [] gain = torch.ones(7, device=self.device) # normalized to gridspace gain @@ -226,4 +253,4 @@ anch.append(anchors[a]) # anchors tcls.append(c) # class - return tcls, tbox, indices, anch+ return tcls, tbox, indices, anch
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/loss.py
Add docstrings including usage examples
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import contextlib import math from pathlib import Path import cv2 import matplotlib.pyplot as plt import numpy as np import pandas as pd import torch from .. import threaded from ..general import xywh2xyxy from ..plots import Annotator, colors @threaded def plot_images_and_masks(images, targets, masks, paths=None, fname="images.jpg", names=None): if isinstance(images, torch.Tensor): images = images.cpu().float().numpy() if isinstance(targets, torch.Tensor): targets = targets.cpu().numpy() if isinstance(masks, torch.Tensor): masks = masks.cpu().numpy().astype(int) max_size = 1920 # max image size max_subplots = 16 # max image subplots, i.e. 4x4 bs, _, h, w = images.shape # batch size, _, height, width bs = min(bs, max_subplots) # limit plot images ns = np.ceil(bs**0.5) # number of subplots (square) if np.max(images[0]) <= 1: images *= 255 # de-normalise (optional) # Build Image mosaic = np.full((int(ns * h), int(ns * w), 3), 255, dtype=np.uint8) # init for i, im in enumerate(images): if i == max_subplots: # if last batch has fewer images than we expect break x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin im = im.transpose(1, 2, 0) mosaic[y : y + h, x : x + w, :] = im # Resize (optional) scale = max_size / ns / max(h, w) if scale < 1: h = math.ceil(scale * h) w = math.ceil(scale * w) mosaic = cv2.resize(mosaic, tuple(int(x * ns) for x in (w, h))) # Annotate fs = int((h + w) * ns * 0.01) # font size annotator = Annotator(mosaic, line_width=round(fs / 10), font_size=fs, pil=True, example=names) for i in range(i + 1): x, y = int(w * (i // ns)), int(h * (i % ns)) # block origin annotator.rectangle([x, y, x + w, y + h], None, (255, 255, 255), width=2) # borders if paths: annotator.text([x + 5, y + 5], text=Path(paths[i]).name[:40], txt_color=(220, 220, 220)) # filenames if len(targets) > 0: idx = targets[:, 0] == i ti = targets[idx] # image targets boxes = xywh2xyxy(ti[:, 2:6]).T classes = ti[:, 1].astype("int") labels = ti.shape[1] == 6 # labels if no conf column conf = None if labels else ti[:, 6] # check for confidence presence (label vs pred) if boxes.shape[1]: if boxes.max() <= 1.01: # if normalized with tolerance 0.01 boxes[[0, 2]] *= w # scale to pixels boxes[[1, 3]] *= h elif scale < 1: # absolute coords need scale if image scales boxes *= scale boxes[[0, 2]] += x boxes[[1, 3]] += y for j, box in enumerate(boxes.T.tolist()): cls = classes[j] color = colors(cls) cls = names[cls] if names else cls if labels or conf[j] > 0.25: # 0.25 conf thresh label = f"{cls}" if labels else f"{cls} {conf[j]:.1f}" annotator.box_label(box, label, color=color) # Plot masks if len(masks): if masks.max() > 1.0: # mean that masks are overlap image_masks = masks[[i]] # (1, 640, 640) nl = len(ti) index = np.arange(nl).reshape(nl, 1, 1) + 1 image_masks = np.repeat(image_masks, nl, axis=0) image_masks = np.where(image_masks == index, 1.0, 0.0) else: image_masks = masks[idx] im = np.asarray(annotator.im).copy() for j, box in enumerate(boxes.T.tolist()): if labels or conf[j] > 0.25: # 0.25 conf thresh color = colors(classes[j]) mh, mw = image_masks[j].shape if mh != h or mw != w: mask = image_masks[j].astype(np.uint8) mask = cv2.resize(mask, (w, h)) mask = mask.astype(bool) else: mask = image_masks[j].astype(bool) with contextlib.suppress(Exception): im[y : y + h, x : x + w, :][mask] = ( im[y : y + h, x : x + w, :][mask] * 0.4 + np.array(color) * 0.6 ) annotator.fromarray(im) annotator.im.save(fname) # save def plot_results_with_masks(file="path/to/results.csv", dir="", best=True): save_dir = Path(file).parent if file else Path(dir) fig, ax = plt.subplots(2, 8, figsize=(18, 6), tight_layout=True) ax = ax.ravel() files = list(save_dir.glob("results*.csv")) assert len(files), f"No results.csv files found in {save_dir.resolve()}, nothing to plot." for f in files: try: data = pd.read_csv(f) index = np.argmax( 0.9 * data.values[:, 8] + 0.1 * data.values[:, 7] + 0.9 * data.values[:, 12] + 0.1 * data.values[:, 11] ) s = [x.strip() for x in data.columns] x = data.values[:, 0] for i, j in enumerate([1, 2, 3, 4, 5, 6, 9, 10, 13, 14, 15, 16, 7, 8, 11, 12]): y = data.values[:, j] # y[y == 0] = np.nan # don't show zero values ax[i].plot(x, y, marker=".", label=f.stem, linewidth=2, markersize=2) if best: # best ax[i].scatter(index, y[index], color="r", label=f"best:{index}", marker="*", linewidth=3) ax[i].set_title(s[j] + f"\n{round(y[index], 5)}") else: # last ax[i].scatter(x[-1], y[-1], color="r", label="last", marker="*", linewidth=3) ax[i].set_title(s[j] + f"\n{round(y[-1], 5)}") # if j in [8, 9, 10]: # share train and val loss y axes # ax[i].get_shared_y_axes().join(ax[i], ax[i - 5]) except Exception as e: print(f"Warning: Plotting error for {f}: {e}") ax[1].legend() fig.savefig(save_dir / "results.png", dpi=200) plt.close()
--- +++ @@ -17,6 +17,7 @@ @threaded def plot_images_and_masks(images, targets, masks, paths=None, fname="images.jpg", names=None): + """Plots a grid of images, their labels, and masks with optional resizing and annotations, saving to fname.""" if isinstance(images, torch.Tensor): images = images.cpu().float().numpy() if isinstance(targets, torch.Tensor): @@ -112,6 +113,10 @@ def plot_results_with_masks(file="path/to/results.csv", dir="", best=True): + """Plots training results from CSV files, plotting best or last result highlights based on `best` parameter. + + Example: from utils.plots import *; plot_results('path/to/results.csv') + """ save_dir = Path(file).parent if file else Path(dir) fig, ax = plt.subplots(2, 8, figsize=(18, 6), tight_layout=True) ax = ax.ravel() @@ -143,4 +148,4 @@ print(f"Warning: Plotting error for {f}: {e}") ax[1].legend() fig.savefig(save_dir / "results.png", dpi=200) - plt.close()+ plt.close()
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/utils/segment/plots.py
Provide clean and structured docstrings
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license import math import numpy as np import torch import torch.nn as nn from ultralytics.utils.patches import torch_load from utils.downloads import attempt_download class Sum(nn.Module): def __init__(self, n, weight=False): super().__init__() self.weight = weight # apply weights boolean self.iter = range(n - 1) # iter object if weight: self.w = nn.Parameter(-torch.arange(1.0, n) / 2, requires_grad=True) # layer weights def forward(self, x): y = x[0] # no weight if self.weight: w = torch.sigmoid(self.w) * 2 for i in self.iter: y = y + x[i + 1] * w[i] else: for i in self.iter: y = y + x[i + 1] return y class MixConv2d(nn.Module): def __init__(self, c1, c2, k=(1, 3), s=1, equal_ch=True): super().__init__() n = len(k) # number of convolutions if equal_ch: # equal c_ per group i = torch.linspace(0, n - 1e-6, c2).floor() # c2 indices c_ = [(i == g).sum() for g in range(n)] # intermediate channels else: # equal weight.numel() per group b = [c2] + [0] * n a = np.eye(n + 1, n, k=-1) a -= np.roll(a, 1, axis=1) a *= np.array(k) ** 2 a[0] = 1 c_ = np.linalg.lstsq(a, b, rcond=None)[0].round() # solve for equal weight indices, ax = b self.m = nn.ModuleList( [nn.Conv2d(c1, int(c_), k, s, k // 2, groups=math.gcd(c1, int(c_)), bias=False) for k, c_ in zip(k, c_)] ) self.bn = nn.BatchNorm2d(c2) self.act = nn.SiLU() def forward(self, x): return self.act(self.bn(torch.cat([m(x) for m in self.m], 1))) class Ensemble(nn.ModuleList): def __init__(self): super().__init__() def forward(self, x, augment=False, profile=False, visualize=False): y = [module(x, augment, profile, visualize)[0] for module in self] # y = torch.stack(y).max(0)[0] # max ensemble # y = torch.stack(y).mean(0) # mean ensemble y = torch.cat(y, 1) # nms ensemble return y, None # inference, train output def attempt_load(weights, device=None, inplace=True, fuse=True): from models.yolo import Detect, Model model = Ensemble() for w in weights if isinstance(weights, list) else [weights]: ckpt = torch_load(attempt_download(w), map_location="cpu") # load ckpt = (ckpt.get("ema") or ckpt["model"]).to(device).float() # FP32 model # Model compatibility updates if not hasattr(ckpt, "stride"): ckpt.stride = torch.tensor([32.0]) if hasattr(ckpt, "names") and isinstance(ckpt.names, (list, tuple)): ckpt.names = dict(enumerate(ckpt.names)) # convert to dict model.append(ckpt.fuse().eval() if fuse and hasattr(ckpt, "fuse") else ckpt.eval()) # model in eval mode # Module updates for m in model.modules(): t = type(m) if t in (nn.Hardswish, nn.LeakyReLU, nn.ReLU, nn.ReLU6, nn.SiLU, Detect, Model): m.inplace = inplace if t is Detect and not isinstance(m.anchor_grid, list): delattr(m, "anchor_grid") setattr(m, "anchor_grid", [torch.zeros(1)] * m.nl) elif t is nn.Upsample and not hasattr(m, "recompute_scale_factor"): m.recompute_scale_factor = None # torch 1.11.0 compatibility # Return model if len(model) == 1: return model[-1] # Return detection ensemble print(f"Ensemble created with {weights}\n") for k in "names", "nc", "yaml": setattr(model, k, getattr(model[0], k)) model.stride = model[torch.argmax(torch.tensor([m.stride.max() for m in model])).int()].stride # max stride assert all(model[0].nc == m.nc for m in model), f"Models have different class counts: {[m.nc for m in model]}" return model
--- +++ @@ -1,4 +1,5 @@ # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license +"""Experimental modules.""" import math @@ -11,8 +12,12 @@ class Sum(nn.Module): + """Weighted sum of 2 or more layers https://arxiv.org/abs/1911.09070.""" def __init__(self, n, weight=False): + """Initializes a module to sum outputs of layers with number of inputs `n` and optional weighting, supporting 2+ + inputs. + """ super().__init__() self.weight = weight # apply weights boolean self.iter = range(n - 1) # iter object @@ -20,6 +25,7 @@ self.w = nn.Parameter(-torch.arange(1.0, n) / 2, requires_grad=True) # layer weights def forward(self, x): + """Processes input through a customizable weighted sum of `n` inputs, optionally applying learned weights.""" y = x[0] # no weight if self.weight: w = torch.sigmoid(self.w) * 2 @@ -32,8 +38,12 @@ class MixConv2d(nn.Module): + """Mixed Depth-wise Conv https://arxiv.org/abs/1907.09595.""" def __init__(self, c1, c2, k=(1, 3), s=1, equal_ch=True): + """Initializes MixConv2d with mixed depth-wise convolutional layers, taking input and output channels (c1, c2), + kernel sizes (k), stride (s), and channel distribution strategy (equal_ch). + """ super().__init__() n = len(k) # number of convolutions if equal_ch: # equal c_ per group @@ -54,15 +64,21 @@ self.act = nn.SiLU() def forward(self, x): + """Performs forward pass by applying SiLU activation on batch-normalized concatenated convolutional layer + outputs. + """ return self.act(self.bn(torch.cat([m(x) for m in self.m], 1))) class Ensemble(nn.ModuleList): + """Ensemble of models.""" def __init__(self): + """Initializes an ensemble of models to be used for aggregated predictions.""" super().__init__() def forward(self, x, augment=False, profile=False, visualize=False): + """Performs forward pass aggregating outputs from an ensemble of models..""" y = [module(x, augment, profile, visualize)[0] for module in self] # y = torch.stack(y).max(0)[0] # max ensemble # y = torch.stack(y).mean(0) # mean ensemble @@ -71,6 +87,10 @@ def attempt_load(weights, device=None, inplace=True, fuse=True): + """Loads and fuses an ensemble or single YOLOv5 model from weights, handling device placement and model adjustments. + + Example inputs: weights=[a,b,c] or a single model weights=[a] or weights=a. + """ from models.yolo import Detect, Model model = Ensemble() @@ -107,4 +127,4 @@ setattr(model, k, getattr(model[0], k)) model.stride = model[torch.argmax(torch.tensor([m.stride.max() for m in model])).int()].stride # max stride assert all(model[0].nc == m.nc for m in model), f"Models have different class counts: {[m.nc for m in model]}" - return model+ return model
https://raw.githubusercontent.com/ultralytics/yolov5/HEAD/models/experimental.py