16 releases

0.3.23 Mar 9, 2024
0.3.22 Mar 8, 2024
0.3.13 Feb 28, 2024
0.3.10 May 23, 2023

#223 in Algorithms

Download history 77/week @ 2023-12-06 98/week @ 2023-12-13 186/week @ 2023-12-20 14/week @ 2023-12-27 36/week @ 2024-01-03 94/week @ 2024-01-10 104/week @ 2024-01-17 21/week @ 2024-01-24 6/week @ 2024-01-31 151/week @ 2024-02-07 36/week @ 2024-02-14 216/week @ 2024-02-21 802/week @ 2024-02-28 1074/week @ 2024-03-06 288/week @ 2024-03-13 53/week @ 2024-03-20

2,326 downloads per month
Used in 2 crates (via two_percent)

MIT license

80KB
2K SLoC

Crates.io

Fuzzy Matcher

Fuzzy matching algorithm(s) in Rust!

Usage

In your Cargo.toml add the following:

[dependencies]
fuzzy-matcher = "*"

Here are some code example:

use fuzzy_matcher::FuzzyMatcher;
use fuzzy_matcher::skim::SkimMatcherV2;

let matcher = SkimMatcherV2::default();
assert_eq!(None, matcher.fuzzy_match("abc", "abx"));
assert!(matcher.fuzzy_match("axbycz", "abc").is_some());
assert!(matcher.fuzzy_match("axbycz", "xyz").is_some());

let (score, indices) = matcher.fuzzy_indices("axbycz", "abc").unwrap();
assert_eq!(indices, [0, 2, 4]);
  • fuzzy_match only return scores while fuzzy_indices returns the matching indices as well.
  • Both function return None if the pattern won't match.
  • The score is the higher the better.

More example

echo "axbycz" | cargo run --example fz "abc" and check what happens.

About the Algorithm

Skim

The skim is currently used by skim, a fuzzy finder.

Skim V2

  • Just like fzf v2, the algorithm is based on Smith-Waterman algorithm which is normally used in DNA sequence alignment
  • Also checkout https://www.cs.cmu.edu/~ckingsf/bioinfo-lectures/gaps.pdf for more details
  • The time complexity is O(mn) where m, n are the length of the pattern and input line.
  • Space complexity is O(mn) for fuzzy_indices and O(2n) for fuzzy_match which will compress the table for dynamic programming.
  • V2 matcher has an option to set the max element of the score matrix, if m*n exceeded the limit, it will fallback to a linear search.

Skim V1

Clangd

  • The algorithm is based on clangd's FuzzyMatch.cpp.
  • Also checkout https://github.com/lewang/flx/issues/98 for some variants.
  • The algorithm is O(mn) where m, n are the length of the pattern and input line.
  • Space complexity is O(mn) for fuzzy_indices and O(2n) for fuzzy_match which will compress the table for dynamic programming.

Dependencies

~89KB