File size: 18,130 Bytes
5ef0f8d 2c86f91 5f1cdfa 2ba1434 035141c 5ef0f8d 8ac47d4 5ef0f8d 5f1cdfa 5ef0f8d 46800f4 5ef0f8d 2c86f91 5ef0f8d f3bf993 035141c f6bffda 5ef0f8d d2dc29e 5ef0f8d 2c86f91 6607a5c 2c86f91 5ef0f8d eb8cbe5 6607a5c 5f1cdfa 6607a5c 5f1cdfa 6607a5c 2c86f91 6607a5c 5f1cdfa 6607a5c 5f1cdfa 6607a5c 2c86f91 5f1cdfa 2c86f91 f3bf993 eb8cbe5 6607a5c 2c86f91 5f1cdfa 2c86f91 5f1cdfa 2c86f91 5ef0f8d 2c86f91 eb8cbe5 2c86f91 5ef0f8d 2c86f91 5ef0f8d 46800f4 eb8cbe5 46800f4 5ef0f8d 46800f4 5ef0f8d 46800f4 5ef0f8d 46800f4 5ef0f8d 46800f4 5ef0f8d 46800f4 5ef0f8d 5f1cdfa 5ef0f8d 2c86f91 6607a5c eb8cbe5 5ef0f8d eb8cbe5 2c86f91 eb8cbe5 2c86f91 eb8cbe5 2c86f91 eb8cbe5 2c86f91 eb8cbe5 2c86f91 eb8cbe5 5f1cdfa 2c86f91 5ef0f8d f3bf993 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d f3bf993 5ef0f8d f3bf993 8ac47d4 595613d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 595613d f3bf993 5ef0f8d 595613d 5ef0f8d f3bf993 5ef0f8d a83bff5 8ac47d4 a83bff5 6607a5c a83bff5 5ef0f8d 6607a5c a83bff5 6607a5c 5ef0f8d a83bff5 eb8cbe5 a83bff5 6607a5c 5ef0f8d 6607a5c eb8cbe5 6607a5c eb8cbe5 6607a5c 5ef0f8d a83bff5 5ef0f8d a83bff5 5ef0f8d f3bf993 8ac47d4 5ef0f8d d2dc29e 5ef0f8d 6607a5c 5ef0f8d 5f1cdfa 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d 8ac47d4 5ef0f8d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 |
import asyncio
from aiolimiter import AsyncLimiter
from pathlib import Path
import traceback
from typing import Literal, Tuple
from fastapi.routing import APIRouter
import logging
import io
import zipfile
import os
from httpx import AsyncClient
from pydantic import BaseModel
import subprocess
import pandas as pd
import re
import tempfile
from lxml import etree
from bs4 import BeautifulSoup
from fastapi import Depends, HTTPException
from dependencies import get_http_client, get_llm_router
from fastapi.responses import StreamingResponse
from litellm.router import Router
from kreuzberg import ExtractionConfig, extract_bytes
from schemas import GetMeetingDocsRequest, GetMeetingDocsResponse, DocRequirements, DownloadDocsRequest, GetMeetingsRequest, GetMeetingsResponse, ExtractRequirementsRequest, ExtractRequirementsResponse
# API router for requirement extraction from docs / doc list retrieval / download
router = APIRouter(tags=["document extraction"])
# ==================================================== Utilities =================================================================
NSMAP = {
'w': 'http://schemas.openxmlformats.org/wordprocessingml/2006/main',
'v': 'urn:schemas-microsoft-com:vml'
}
# ================================== Converting of files to .txt ====================================
KREUZBERG_CONFIG: ExtractionConfig = ExtractionConfig(
force_ocr=False, ocr_backend=None)
# Unfortunately needs to be kept to 1, as libreoffice isn't built to support parallel instances
LO_CONVERSION_MUTEX = asyncio.Semaphore(1)
async def convert_file(contents: io.BytesIO, filename: str, input_ext: str, output_ext: str, filter: str = None) -> io.BytesIO:
"""
Converts the given file bytes using Libreoffice headless to the specified file type.
This is an asynchronous version.
Args:
contents: File contents
filename: File base name WITHOUT THE EXTENSION
input_ext: Input extension (WITHOUT THE DOT)
output_ext: Output extension (WITHOUT THE DOT)
filter: The conversion filter to use.
"""
await LO_CONVERSION_MUTEX.acquire()
with tempfile.TemporaryDirectory() as tmpdir:
dir_path = Path(tmpdir)
input_file_path = dir_path / f"{filename}.{input_ext}"
output_file_path = dir_path / f"{filename}.{output_ext}"
# write the memory contents to the input file
with open(input_file_path, "wb") as in_file:
in_file.write(contents.read())
out_bytes = io.BytesIO()
# construct the command
command = [
"libreoffice",
"--headless",
"--convert-to", f"{output_ext}:{filter}" if filter else output_ext,
"--outdir", tmpdir,
str(input_file_path) # Ensure path is a string for subprocess
]
# convert using libreoffice asynchronously
process = await asyncio.create_subprocess_exec(
*command,
stdout=asyncio.subprocess.PIPE,
stderr=asyncio.subprocess.PIPE
)
stdout, stderr = await process.communicate()
exit_code = await process.wait()
if exit_code != 0 and not output_file_path.exists():
raise subprocess.CalledProcessError(
exit_code,
command,
output=stdout,
stderr=stderr
)
LO_CONVERSION_MUTEX.release()
with open(output_file_path, mode="rb") as out:
out_bytes.write(out.read())
out_bytes.seek(0)
return out_bytes
# Rate limit of FTP downloads per minute
FTP_DOWNLOAD_RATE_LIMITER = AsyncLimiter(max_rate=120, time_period=60)
async def get_doc_archive(url: str, client: AsyncClient) -> tuple[str, str, io.BytesIO]:
"""Récupère le docx depuis l'URL et le retourne un tuple (nom, extension, contenu)"""
async with FTP_DOWNLOAD_RATE_LIMITER:
if not url.endswith("zip"):
raise ValueError("URL doit pointer vers un fichier ZIP")
doc_id = os.path.splitext(os.path.basename(url))[0]
resp = await client.get(url, headers={
"User-Agent": 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36'
})
resp.raise_for_status()
with zipfile.ZipFile(io.BytesIO(resp.content)) as zf:
# there should be a single file per file
for entry in zf.infolist():
if entry.is_dir():
continue
file_name = entry.filename
root, ext = os.path.splitext(file_name)
doc_bytes = zf.read(file_name)
return (root, ext.lower(), io.BytesIO(doc_bytes))
raise ValueError("Aucun fichier trouvé dans l'archive")
def apply_docx_revisions(docx_zip: zipfile.ZipFile) -> io.BytesIO:
"""
Applique les révisions des .docx avant de retourner le contenu.
Args:
docx_zip: Le document word sous forme de zip
"""
try:
xml_bytes = docx_zip.read('word/document.xml')
except KeyError:
raise FileNotFoundError(
"word/document.xml not found in the DOCX archive.")
parser = etree.XMLParser(remove_blank_text=True)
root = etree.fromstring(xml_bytes, parser=parser)
# Suppression des balises <w:del> et leur contenu
for del_elem in root.xpath('//w:del', namespaces=NSMAP):
parent = del_elem.getparent()
if parent is not None:
parent.remove(del_elem)
# Désencapsulation des balises <w:ins>
for ins_elem in root.xpath('//w:ins', namespaces=NSMAP):
parent = ins_elem.getparent()
if parent is not None:
index = parent.index(ins_elem)
for child in ins_elem.iterchildren():
parent.insert(index, child)
index += 1
parent.remove(ins_elem)
# Nettoyage des commentaires
for tag in ['w:commentRangeStart', 'w:commentRangeEnd', 'w:commentReference']:
for elem in root.xpath(f'//{tag}', namespaces=NSMAP):
parent = elem.getparent()
if parent is not None:
parent.remove(elem)
# 3. Create a new docx with the modified XML
output = io.BytesIO()
with zipfile.ZipFile(output, 'w', compression=zipfile.ZIP_DEFLATED) as new_zip:
# Copier tous les fichiers non modifiés
for file_info in docx_zip.infolist():
if file_info.filename != 'word/document.xml':
new_zip.writestr(file_info, docx_zip.read(file_info.filename))
# Ajouter le document.xml modifié
xml_str = etree.tostring(
root,
xml_declaration=True,
encoding='UTF-8',
pretty_print=True
)
new_zip.writestr('word/document.xml', xml_str)
output.seek(0)
return output
FORMAT_MIME_TYPES = {
".docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
".pdf": "application/pdf",
".pptx": "application/vnd.openxmlformats-officedocument.presentationml.presentation"
}
async def doc_to_txt(doc_id: str, url: str, client: AsyncClient) -> str:
"""
Télécharge le TDoc spécifié et le convertit en texte.
"""
# Grab the document archive
filename, ext, bytes = await get_doc_archive(url, client)
final_text: str = None
if ext == ".doc":
logging.debug(f"Converting {filename} .doc --> .docx")
docx_bytes = await convert_file(bytes, doc_id, "doc", "docx")
extracted_data = await extract_bytes(docx_bytes.read(), FORMAT_MIME_TYPES[".docx"], config=KREUZBERG_CONFIG)
final_text = extracted_data.content
elif ext == ".docx":
# Applying doc revisions to docx files (especially for pCR / draftCR files)
logging.debug(f"Updating .docx revisions for {doc_id}.")
applied_revision = apply_docx_revisions(zipfile.ZipFile(bytes))
extracted_data = await extract_bytes(applied_revision.read(), FORMAT_MIME_TYPES[".docx"], config=KREUZBERG_CONFIG)
final_text = extracted_data.content
else:
if ext in FORMAT_MIME_TYPES: # file extension is supported
extracted_data = await extract_bytes(bytes.read(), FORMAT_MIME_TYPES[ext], config=KREUZBERG_CONFIG)
final_text = extracted_data.content
else:
raise Exception(
f"Unsupported file type: {ext}, filename: {filename}")
txt_data = [line.strip()
for line in final_text.splitlines() if line.strip()]
return txt_data
# ============================================= Doc routes =========================================================
@router.post("/get_meetings", response_model=GetMeetingsResponse)
async def get_meetings(req: GetMeetingsRequest, http_client: AsyncClient = Depends(get_http_client)):
"""
Retrieves the list of meetings for the given working group.
"""
# Extracting WG
working_group = req.working_group
tsg = re.sub(r"\d+", "", working_group)
wg_number = re.search(r"\d", working_group).group(0)
# building corresponding FTP url
logging.debug(tsg, wg_number)
url = "https://www.3gpp.org/ftp/tsg_" + tsg
logging.debug(url)
ftp_request = await http_client.get(url)
soup = BeautifulSoup(ftp_request.text, "html.parser")
meeting_folders = []
all_meetings = []
wg_folders = [item.get_text() for item in soup.select("tr td a")]
selected_folder = None
# sanity check to ensure the requested workgroup is present in the ftp directories
for folder in wg_folders:
if "wg" + str(wg_number) in folder.lower():
selected_folder = folder
break
url += "/" + selected_folder
logging.debug(url)
if selected_folder:
resp = await http_client.get(url)
soup = BeautifulSoup(resp.text, "html.parser")
meeting_folders = [item.get_text() for item in soup.select("tr td a") if item.get_text(
).startswith("TSG") or (item.get_text().startswith("CT") and "-" in item.get_text())]
all_meetings = [working_group + "#" + meeting.split("_", 1)[1].replace("_", " ").replace(
"-", " ") if meeting.startswith('TSG') else meeting.replace("-", "#") for meeting in meeting_folders]
return GetMeetingsResponse(meetings=dict(zip(all_meetings, meeting_folders)))
# ============================================================================================================================================
@router.post("/get_meeting_docs", response_model=GetMeetingDocsResponse)
async def get_meeting_docs(req: GetMeetingDocsRequest, http_client: AsyncClient = Depends(get_http_client)) -> GetMeetingDocsResponse:
"""
Downloads the document list dataframe for a given meeting
"""
# FIXME: extract the document URLS from the hyperlinks in the excelsheet using openpyxl?
# Extracting WG
working_group = req.working_group
tsg = re.sub(r"\d+", "", working_group)
wg_number = re.search(r"\d", working_group).group(0)
url = "https://www.3gpp.org/ftp/tsg_" + tsg
logging.info("Fetching TDocs dataframe")
resp = await http_client.get(url)
soup = BeautifulSoup(resp.text, "html.parser")
wg_folders = [item.get_text() for item in soup.select("tr td a")]
selected_folder = None
for folder in wg_folders:
if "wg" + str(wg_number) in folder.lower():
selected_folder = folder
break
url += "/" + selected_folder + "/" + req.meeting + "/docs"
resp = await http_client.get(url)
soup = BeautifulSoup(resp.text, "html.parser")
files = [item.get_text() for item in soup.select("tr td a")
if item.get_text().endswith(".xlsx")]
if files == []:
raise HTTPException(status_code=404, detail="No XLSX has been found")
df = pd.read_excel(str(url + "/" + files[0]).replace("#", "%23"))
filtered_df = df[~(
df["Uploaded"].isna())][["TDoc", "Title", "CR category", "For", "Source", "Type", "Agenda item", "Agenda item description", "TDoc Status"]]
filtered_df["URL"] = filtered_df["TDoc"].apply(
lambda tdoc: f"{url}/{tdoc}.zip")
df = filtered_df.fillna("")
return GetMeetingDocsResponse(data=df[["TDoc", "Title", "Type", "For", "TDoc Status", "Agenda item description", "URL"]].to_dict(orient="records"))
# ==================================================================================================================================
@router.post("/download_docs")
async def download_docs(req: DownloadDocsRequest, http_client: AsyncClient = Depends(get_http_client)) -> StreamingResponse:
"""Download the specified TDocs and zips them in a single archive"""
# Document IDs to download
document_ids = [doc.document for doc in req.documents]
logging.info(f"Downloading TDocs: {document_ids}")
async def _process_single_document(doc_id: str, doc_url: str) -> Tuple[bool, bytes]:
"""Attempts to convert a document to text and returns success status and content."""
try:
text_lines = await doc_to_txt(doc_id, doc_url, http_client)
content_bytes = "\n".join(text_lines).encode("utf-8")
return {"doc_id": doc_id, "content": content_bytes}
except Exception as e:
logging.warning(
f"Failed to process document '{doc_id}' from URL '{doc_url}': {e}")
error_message = f"Document '{doc_id}' text extraction failed: {e}".encode(
"utf-8")
return {"doc_id": doc_id, "content": error_message, "failed": True}
convert_tasks = await asyncio.gather(*[_process_single_document(doc.document, doc.url) for doc in req.documents], return_exceptions=False)
zip_buffer = io.BytesIO()
with zipfile.ZipFile(zip_buffer, mode='w', compression=zipfile.ZIP_DEFLATED) as zip_file:
for task in convert_tasks:
failed = "failed" in task
doc_id = task["doc_id"]
safe_filename = f"failed_{doc_id}.txt" if failed else f"{doc_id}.txt"
zip_file.writestr(safe_filename, task["content"])
zip_buffer.seek(0)
return StreamingResponse(
zip_buffer,
media_type="application/zip",
headers={"Content-Disposition": "attachment; filename=tdocs.zip"}
)
# ======================================================================================================================================================================================
class ProgressUpdate(BaseModel):
"""Defines the structure of a single SSE message."""
status: Literal["progress", "complete"]
data: dict
total_docs: int
processed_docs: int
@router.post("/extract_requirements/sse")
async def extract_requirements_from_docs(req: ExtractRequirementsRequest, llm_router: Router = Depends(get_llm_router), http_client: AsyncClient = Depends(get_http_client)):
"""Extract requirements from the specified xxxxCR docs using a LLM and returns SSE events about the progress of ongoing operations"""
documents = req.documents
n_docs = len(documents)
logging.info(
"Generating requirements for documents: {}".format(req.documents))
# limit max concurrency of LLM requests to prevent a huge pile of errors because of small rate limits
concurrency_sema = asyncio.Semaphore(4)
def prompt(doc_id, full):
return f"Here's the document whose ID is {doc_id} : {full}\n\nExtract all requirements and group them by context, returning a list of objects where each object includes a document ID, a concise description of the context where the requirements apply (not a chapter title or copied text), and a list of associated requirements; always return the result as a list, even if only one context is found. Remove the errors"
async def _process_document(doc) -> list[DocRequirements]:
doc_id = doc.document
url = doc.url
# convert the docx to txt for use
try:
doc = await doc_to_txt(doc_id, url, http_client)
full = "\n".join(doc)
except Exception as e:
fmt = "".join(traceback.format_exception(e))
logging.error(f"Failed to process doc {doc_id} : {fmt}")
return [DocRequirements(document=doc_id, context="Failed to process document", requirements=[])]
try:
await concurrency_sema.acquire()
model_used = "gemini-v2"
resp_ai = await llm_router.acompletion(
model=model_used,
messages=[
{"role": "user", "content": prompt(doc_id, full)}],
response_format=ExtractRequirementsResponse
)
return ExtractRequirementsResponse.model_validate_json(resp_ai.choices[0].message.content).requirements
except Exception as e:
return [DocRequirements(document=doc_id, context="Error LLM", requirements=[])]
finally:
concurrency_sema.release()
# futures for all processed documents
process_futures = [_process_document(doc) for doc in documents]
# lambda to print progress
def progress_update(x): return f"data: {x.model_dump_json()}\n\n"
# async generator that generates the SSE events for progress
async def _stream_generator(docs: list[asyncio.Future]):
items = []
n_processed = 0
yield progress_update(ProgressUpdate(status="progress", data={}, total_docs=n_docs, processed_docs=0))
for doc in asyncio.as_completed(docs):
result = await doc
items.extend(result)
n_processed += 1
yield progress_update(ProgressUpdate(status="progress", data={}, total_docs=n_docs, processed_docs=n_processed))
final_response = ExtractRequirementsResponse(requirements=items)
yield progress_update(ProgressUpdate(status="complete", data=final_response.model_dump(), total_docs=n_docs, processed_docs=n_processed))
return StreamingResponse(_stream_generator(process_futures), media_type="text/event-stream")
|