scikitplot.levenshtein#

Levenshtein distance facade with safe optional acceleration.

scikitplot.levenshtein is always importable. The default backend chain is bundled Cython -> RapidFuzz -> dependency-free Python. The GPL-licensed Levenshtein package is available only when explicitly requested.

User guide. See the Levenshtein distance and fuzzy matching section for further details.

Backend-neutral Levenshtein edit distance helpers.#

`BackendInfo`	Read-only description of one implementation backend.
`Match`	One ranked fuzzy-match result with actual backend provenance.
`available_backends`	Return backend capability records in preferred order.
`backend_info`	Return capability information for one backend.
`closest`	Return the closest choice, or `None` for an empty iterable.
`distance`	Return the Levenshtein edit distance between two sequences.
`make_corpus_scorer`	Return a scorer compatible with `corpus.CustomRetrievalIndex`.
`normalized_distance`	Return distance normalised to `[0, 1]` by maximum sequence length.
`normalized_similarity`	Return normalised Levenshtein similarity in `[0, 1]`.
`rank`	Rank choices from most to least similar to query.
`similarity`	Return `max(len(a), len(b)) - distance(a, b)`.