starfuzz/search

A module providing fuzzy search collections matching capabilities. Collections of custom candidates can be searched in-memory with custom selectors, similarity scoring functions, score thresholds, and limits.

Types

Match

</>

A matched result from a fuzzy search query.

pub type Match(candidate) {
  Match(candidate: candidate, score: Float, index: Int)
}

Constructors

```
Match(candidate: candidate, score: Float, index: Int)
```
Arguments

candidate

The original candidate item from the list.

score

The similarity score calculated for this match (between 0.0 and 1.0).

index

The original index of this candidate in the input list (used for stable tie-breaking).

SearchBuilder

</>

An immutable search configuration builder.

pub type SearchBuilder(candidate) {
  SearchBuilder(
    candidates: List(candidate),
    selector: fn(candidate) -> String,
    scorer: fn(String, String) -> Float,
    normalizer: option.Option(normalize.Normalizer),
    min_score: Float,
    limit: option.Option(Int),
  )
}

Constructors

SearchBuilder(
  candidates: List(candidate),
  selector: fn(candidate) -> String,
  scorer: fn(String, String) -> Float,
  normalizer: option.Option(normalize.Normalizer),
  min_score: Float,
  limit: option.Option(Int),
)

Values

by

</>

pub fn by(
  builder: SearchBuilder(candidate),
  selector: fn(candidate) -> String,
) -> SearchBuilder(candidate)

Configures a custom selector function to retrieve the string value to match from candidates.

new

</>

pub fn new(
  candidates: List(candidate),
  selector: fn(candidate) -> String,
) -> SearchBuilder(candidate)

Creates a new SearchBuilder with the list of candidates and a string selector function. Defaults to using Levenshtein similarity, no normalizer, minimum score of 0.0, and no limit.

run

</>

pub fn run(
  builder: SearchBuilder(candidate),
  query: String,
) -> List(Match(candidate))

Runs the fuzzy search query against the candidates list. Returns matched candidates sorted descending by score, with original index-based tie-breaking.

strings

</>

pub fn strings(
  query: String,
  candidates: List(String),
) -> List(Match(String))

Quick direct search over a list of strings using default Levenshtein similarity.

using

</>

pub fn using(
  builder: SearchBuilder(candidate),
  scorer: fn(String, String) -> Float,
) -> SearchBuilder(candidate)

Configures the similarity scoring function to use for comparison (e.g. similarity.jaro_winkler).

with_limit

</>

pub fn with_limit(
  builder: SearchBuilder(candidate),
  limit: Int,
) -> SearchBuilder(candidate)

Configures a limit on the number of matches returned.

with_minimum_score

</>

pub fn with_minimum_score(
  builder: SearchBuilder(candidate),
  min_score: Float,
) -> SearchBuilder(candidate)

Configures the minimum score threshold. Matches below this score are discarded.

with_normalizer

</>

pub fn with_normalizer(
  builder: SearchBuilder(candidate),
  normalizer: normalize.Normalizer,
) -> SearchBuilder(candidate)

Configures the normalizer to clean up both query and candidate strings before scoring.

Constructors

Arguments

Constructors