lexical_identity#

scikitplot.rank_bm25.lexical_identity(recipe, analyzer, parameters)[source]#

Return the identity of a lexical build.

Parameters:

recipeRecipe

The scoring formula used.

analyzerAnalyzer or None

How the text was tokenised. None records that the caller supplied tokens directly, which is a different build from any analysed one.

parametersmapping

Scoring parameters, e.g. {"k1": 1.5, "b": 0.75}.

Returns:

str

A 64-character hex digest covering what produced the numbers.

Parameters:

• recipe (Recipe)

• analyzer (Analyzer | None)

• parameters (Mapping[str, Any])

Return type:

str

Notes

Developer. Parameters are sorted by name, so identity does not depend on the order a caller happened to pass them. The corpus content is deliberately not part of this: it answers “how was this built”, and the documents answer “what was built from”. Mixing them would mean a re-tokenisation and a new document looked like the same kind of change.