Scoring Functions

Scoring functions are the core of the fuzzy matching logic in FuzzyFinder. They are responsible for calculating the score for each candidate item in curses_fzf.FuzzyFinder.all_items based on the query the user has entered.

You can use the built-in scoring functions or implement your own and pass them to FuzzyFinder’s score() function parameter.

All scoring functions take the same parameters as ScoringResult’s constructor, query and candidate. The scoring functions create and populate a ScoringResult internally and return it after score calculation. A higher score indicates a better match, while a score of 0 means the candidate was not matched by the query, which will filter it out of the resulting curses_fzf.FuzzyFinder.filtered list.

Learn more about ScoringResult if you plan to implement your own scoring functions.

Built-in Scoring Functions

curses_fzf.scoring_fzf(query: str, candidate: str) → ScoringResult

A fzf-like fuzzy scoring.

This is the default scoring function used by FuzzyFinder if no other scoring function is provided.

The query characters are matched as a subsequence against the candidate (characters must appear in order). There are bonuses for consecutive matches, matches on boundaries and matches early in the candidate.

curses_fzf.scoring_full_words(query: str, candidate: str) → ScoringResult

The query and the candidate string both get lowercased and split on whitespaces (see query_words_with_index & candidate_words_with_index).

Each query word is supposed to find a unique match among the words of the candidate. The closer a match is to a full word the higher the score.

An additional bonus is given, if the match is at the beginning of a word. The query words may appear in the candidate in any order, however if the original order is found a small bonus will be granted.