langchain.evaluation.parsing.json_distance.JsonEditDistanceEvaluator

class langchain.evaluation.parsing.json_distance.JsonEditDistanceEvaluator(string_distance: Optional[Callable[[str, str], float]] = None, canonicalize: Optional[Callable[[Any], Any]] = None, **kwargs: Any)[source]

An evaluator that calculates the edit distance between JSON strings.

This evaluator computes a normalized Damerau-Levenshtein distance between two JSON strings after parsing them and converting them to a canonical format (i.e., whitespace and key order are normalized). It can be customized with alternative distance and canonicalization functions.

Parameters
  • string_distance (Optional[Callable[[str, str], float]]) – A callable that computes the distance between two strings. If not provided, a Damerau-Levenshtein distance from the rapidfuzz package will be used.

  • canonicalize (Optional[Callable[[Any], Any]]) – A callable that converts a parsed JSON object into its canonical string form. If not provided, the default behavior is to serialize the JSON with sorted keys and no extra whitespace.

  • **kwargs (Any) – Additional keyword arguments.

_string_distance

The internal distance computation function.

Type

Callable[[str, str], float]

_canonicalize

The internal canonicalization function.

Type

Callable[[Any], Any]

Examples

>>> evaluator = JsonEditDistanceEvaluator()
>>> result = evaluator.evaluate_strings(prediction='{"a": 1, "b": 2}', reference='{"a": 1, "b": 3}')
>>> assert result["score"] is not None
Raises

ImportError – If rapidfuzz is not installed and no alternative string_distance function is provided.

Attributes

evaluation_name

The name of the evaluation.

requires_input

Whether this evaluator requires an input string.

requires_reference

Whether this evaluator requires a reference label.

Methods

__init__([string_distance, canonicalize])

aevaluate_strings(*, prediction[, ...])

Asynchronously evaluate Chain or LLM output, based on optional input and label.

evaluate_strings(*, prediction[, reference, ...])

Evaluate Chain or LLM output, based on optional input and label.

__init__(string_distance: Optional[Callable[[str, str], float]] = None, canonicalize: Optional[Callable[[Any], Any]] = None, **kwargs: Any) None[source]
async aevaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) dict

Asynchronously evaluate Chain or LLM output, based on optional input and label.

Parameters
  • prediction (str) – The LLM or chain prediction to evaluate.

  • reference (Optional[str], optional) – The reference label to evaluate against.

  • input (Optional[str], optional) – The input to consider during evaluation.

  • **kwargs – Additional keyword arguments, including callbacks, tags, etc.

Returns

The evaluation results containing the score or value.

Return type

dict

evaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) dict

Evaluate Chain or LLM output, based on optional input and label.

Parameters
  • prediction (str) – The LLM or chain prediction to evaluate.

  • reference (Optional[str], optional) – The reference label to evaluate against.

  • input (Optional[str], optional) – The input to consider during evaluation.

  • **kwargs – Additional keyword arguments, including callbacks, tags, etc.

Returns

The evaluation results containing the score or value.

Return type

dict