Spaces:
Runtime error
Runtime error
File size: 14,019 Bytes
63deadc |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 |
"""String distance evaluators based on the RapidFuzz library."""
from enum import Enum
from typing import Any, Callable, Dict, List, Optional
from langchain_core.callbacks.manager import (
AsyncCallbackManagerForChainRun,
CallbackManagerForChainRun,
Callbacks,
)
from langchain_core.pydantic_v1 import Field, root_validator
from langchain.chains.base import Chain
from langchain.evaluation.schema import PairwiseStringEvaluator, StringEvaluator
from langchain.schema import RUN_KEY
def _load_rapidfuzz() -> Any:
"""
Load the RapidFuzz library.
Raises:
ImportError: If the rapidfuzz library is not installed.
Returns:
Any: The rapidfuzz.distance module.
"""
try:
import rapidfuzz
except ImportError:
raise ImportError(
"Please install the rapidfuzz library to use the FuzzyMatchStringEvaluator."
"Please install it with `pip install rapidfuzz`."
)
return rapidfuzz.distance
class StringDistance(str, Enum):
"""Distance metric to use.
Attributes:
DAMERAU_LEVENSHTEIN: The Damerau-Levenshtein distance.
LEVENSHTEIN: The Levenshtein distance.
JARO: The Jaro distance.
JARO_WINKLER: The Jaro-Winkler distance.
HAMMING: The Hamming distance.
INDEL: The Indel distance.
"""
DAMERAU_LEVENSHTEIN = "damerau_levenshtein"
LEVENSHTEIN = "levenshtein"
JARO = "jaro"
JARO_WINKLER = "jaro_winkler"
HAMMING = "hamming"
INDEL = "indel"
class _RapidFuzzChainMixin(Chain):
"""Shared methods for the rapidfuzz string distance evaluators."""
distance: StringDistance = Field(default=StringDistance.JARO_WINKLER)
normalize_score: bool = Field(default=True)
"""Whether to normalize the score to a value between 0 and 1.
Applies only to the Levenshtein and Damerau-Levenshtein distances."""
@root_validator
def validate_dependencies(cls, values: Dict[str, Any]) -> Dict[str, Any]:
"""
Validate that the rapidfuzz library is installed.
Args:
values (Dict[str, Any]): The input values.
Returns:
Dict[str, Any]: The validated values.
"""
_load_rapidfuzz()
return values
@property
def output_keys(self) -> List[str]:
"""
Get the output keys.
Returns:
List[str]: The output keys.
"""
return ["score"]
def _prepare_output(self, result: Dict[str, Any]) -> Dict[str, Any]:
"""
Prepare the output dictionary.
Args:
result (Dict[str, Any]): The evaluation results.
Returns:
Dict[str, Any]: The prepared output dictionary.
"""
result = {"score": result["score"]}
if RUN_KEY in result:
result[RUN_KEY] = result[RUN_KEY].dict()
return result
@staticmethod
def _get_metric(distance: str, normalize_score: bool = False) -> Callable:
"""
Get the distance metric function based on the distance type.
Args:
distance (str): The distance type.
Returns:
Callable: The distance metric function.
Raises:
ValueError: If the distance metric is invalid.
"""
from rapidfuzz import distance as rf_distance
module_map: Dict[str, Any] = {
StringDistance.DAMERAU_LEVENSHTEIN: rf_distance.DamerauLevenshtein,
StringDistance.LEVENSHTEIN: rf_distance.Levenshtein,
StringDistance.JARO: rf_distance.Jaro,
StringDistance.JARO_WINKLER: rf_distance.JaroWinkler,
StringDistance.HAMMING: rf_distance.Hamming,
StringDistance.INDEL: rf_distance.Indel,
}
if distance not in module_map:
raise ValueError(
f"Invalid distance metric: {distance}"
f"\nMust be one of: {list(StringDistance)}"
)
module = module_map[distance]
if normalize_score:
return module.normalized_distance
else:
return module.distance
@property
def metric(self) -> Callable:
"""
Get the distance metric function.
Returns:
Callable: The distance metric function.
"""
return _RapidFuzzChainMixin._get_metric(
self.distance, normalize_score=self.normalize_score
)
def compute_metric(self, a: str, b: str) -> float:
"""
Compute the distance between two strings.
Args:
a (str): The first string.
b (str): The second string.
Returns:
float: The distance between the two strings.
"""
return self.metric(a, b)
class StringDistanceEvalChain(StringEvaluator, _RapidFuzzChainMixin):
"""Compute string distances between the prediction and the reference.
Examples
----------
>>> from langchain.evaluation import StringDistanceEvalChain
>>> evaluator = StringDistanceEvalChain()
>>> evaluator.evaluate_strings(
prediction="Mindy is the CTO",
reference="Mindy is the CEO",
)
Using the `load_evaluator` function:
>>> from langchain.evaluation import load_evaluator
>>> evaluator = load_evaluator("string_distance")
>>> evaluator.evaluate_strings(
prediction="The answer is three",
reference="three",
)
"""
@property
def requires_input(self) -> bool:
"""
This evaluator does not require input.
"""
return False
@property
def requires_reference(self) -> bool:
"""
This evaluator does not require a reference.
"""
return True
@property
def input_keys(self) -> List[str]:
"""
Get the input keys.
Returns:
List[str]: The input keys.
"""
return ["reference", "prediction"]
@property
def evaluation_name(self) -> str:
"""
Get the evaluation name.
Returns:
str: The evaluation name.
"""
return f"{self.distance.value}_distance"
def _call(
self,
inputs: Dict[str, Any],
run_manager: Optional[CallbackManagerForChainRun] = None,
) -> Dict[str, Any]:
"""
Compute the string distance between the prediction and the reference.
Args:
inputs (Dict[str, Any]): The input values.
run_manager (Optional[CallbackManagerForChainRun]):
The callback manager.
Returns:
Dict[str, Any]: The evaluation results containing the score.
"""
return {"score": self.compute_metric(inputs["reference"], inputs["prediction"])}
async def _acall(
self,
inputs: Dict[str, Any],
run_manager: Optional[AsyncCallbackManagerForChainRun] = None,
) -> Dict[str, Any]:
"""
Asynchronously compute the string distance between the prediction
and the reference.
Args:
inputs (Dict[str, Any]): The input values.
run_manager (Optional[AsyncCallbackManagerForChainRun]:
The callback manager.
Returns:
Dict[str, Any]: The evaluation results containing the score.
"""
return {"score": self.compute_metric(inputs["reference"], inputs["prediction"])}
def _evaluate_strings(
self,
*,
prediction: str,
reference: Optional[str] = None,
input: Optional[str] = None,
callbacks: Callbacks = None,
tags: Optional[List[str]] = None,
metadata: Optional[Dict[str, Any]] = None,
include_run_info: bool = False,
**kwargs: Any,
) -> dict:
"""
Evaluate the string distance between the prediction and the reference.
Args:
prediction (str): The prediction string.
reference (Optional[str], optional): The reference string.
input (Optional[str], optional): The input string.
callbacks (Callbacks, optional): The callbacks to use.
**kwargs: Additional keyword arguments.
Returns:
dict: The evaluation results containing the score.
"""
result = self(
inputs={"prediction": prediction, "reference": reference},
callbacks=callbacks,
tags=tags,
metadata=metadata,
include_run_info=include_run_info,
)
return self._prepare_output(result)
async def _aevaluate_strings(
self,
*,
prediction: str,
reference: Optional[str] = None,
input: Optional[str] = None,
callbacks: Callbacks = None,
tags: Optional[List[str]] = None,
metadata: Optional[Dict[str, Any]] = None,
include_run_info: bool = False,
**kwargs: Any,
) -> dict:
"""
Asynchronously evaluate the string distance between the
prediction and the reference.
Args:
prediction (str): The prediction string.
reference (Optional[str], optional): The reference string.
input (Optional[str], optional): The input string.
callbacks (Callbacks, optional): The callbacks to use.
**kwargs: Additional keyword arguments.
Returns:
dict: The evaluation results containing the score.
"""
result = await self.acall(
inputs={"prediction": prediction, "reference": reference},
callbacks=callbacks,
tags=tags,
metadata=metadata,
include_run_info=include_run_info,
)
return self._prepare_output(result)
class PairwiseStringDistanceEvalChain(PairwiseStringEvaluator, _RapidFuzzChainMixin):
"""Compute string edit distances between two predictions."""
@property
def input_keys(self) -> List[str]:
"""
Get the input keys.
Returns:
List[str]: The input keys.
"""
return ["prediction", "prediction_b"]
@property
def evaluation_name(self) -> str:
"""
Get the evaluation name.
Returns:
str: The evaluation name.
"""
return f"pairwise_{self.distance.value}_distance"
def _call(
self,
inputs: Dict[str, Any],
run_manager: Optional[CallbackManagerForChainRun] = None,
) -> Dict[str, Any]:
"""
Compute the string distance between two predictions.
Args:
inputs (Dict[str, Any]): The input values.
run_manager (CallbackManagerForChainRun , optional):
The callback manager.
Returns:
Dict[str, Any]: The evaluation results containing the score.
"""
return {
"score": self.compute_metric(inputs["prediction"], inputs["prediction_b"])
}
async def _acall(
self,
inputs: Dict[str, Any],
run_manager: Optional[AsyncCallbackManagerForChainRun] = None,
) -> Dict[str, Any]:
"""
Asynchronously compute the string distance between two predictions.
Args:
inputs (Dict[str, Any]): The input values.
run_manager (AsyncCallbackManagerForChainRun , optional):
The callback manager.
Returns:
Dict[str, Any]: The evaluation results containing the score.
"""
return {
"score": self.compute_metric(inputs["prediction"], inputs["prediction_b"])
}
def _evaluate_string_pairs(
self,
*,
prediction: str,
prediction_b: str,
callbacks: Callbacks = None,
tags: Optional[List[str]] = None,
metadata: Optional[Dict[str, Any]] = None,
include_run_info: bool = False,
**kwargs: Any,
) -> dict:
"""
Evaluate the string distance between two predictions.
Args:
prediction (str): The first prediction string.
prediction_b (str): The second prediction string.
callbacks (Callbacks, optional): The callbacks to use.
tags (List[str], optional): Tags to apply to traces.
metadata (Dict[str, Any], optional): Metadata to apply to traces.
**kwargs: Additional keyword arguments.
Returns:
dict: The evaluation results containing the score.
"""
result = self(
inputs={"prediction": prediction, "prediction_b": prediction_b},
callbacks=callbacks,
tags=tags,
metadata=metadata,
include_run_info=include_run_info,
)
return self._prepare_output(result)
async def _aevaluate_string_pairs(
self,
*,
prediction: str,
prediction_b: str,
callbacks: Callbacks = None,
tags: Optional[List[str]] = None,
metadata: Optional[Dict[str, Any]] = None,
include_run_info: bool = False,
**kwargs: Any,
) -> dict:
"""
Asynchronously evaluate the string distance between two predictions.
Args:
prediction (str): The first prediction string.
prediction_b (str): The second prediction string.
callbacks (Callbacks, optional): The callbacks to use.
tags (List[str], optional): Tags to apply to traces.
metadata (Dict[str, Any], optional): Metadata to apply to traces.
**kwargs: Additional keyword arguments.
Returns:
dict: The evaluation results containing the score.
"""
result = await self.acall(
inputs={"prediction": prediction, "prediction_b": prediction_b},
callbacks=callbacks,
tags=tags,
metadata=metadata,
include_run_info=include_run_info,
)
return self._prepare_output(result)
|