This crate implements Fuzzy Searching with trigrams
Fuzzy searching allows to compare strings by similarity rather than by equality:\
Similar strings will get a high score (close to 1.0f32) while dissimilar strings will get a lower score (closer to 0.0f32).
Fuzzy searching tolerates changes in word order:\
ex. "John Dep" and "Dep John" will get a high score.
The Algorithm used is taken from : https://dev.to/kaleman15/fuzzy-searching-with-postgresql-97o
Basic idea:
From both strings extracts all groups of 3 adjacent letters.\
("House" becomes [' H', ' Ho', 'Hou', 'ous', 'use', 'se ']).\
Note the 2 spaces added to the head of the string and the one on the tail, used to make the algorithm work on zero length words.
Then counts the number of trigrams of the first words that are also present on the second word and divide by the number of trigrams of the first word.
Example: Comparing 2 strings
rust
fn test () {
use rust_fuzzy_search::fuzzy_compare;
let score : f32 = fuzzy_compare("kolbasobulko", "kolbasobulko");
println!("score = {:?}", score);
}
Example: Comparing a string with a list of strings and retrieving only the best matches
rust
fn test() {
use rust_fuzzy_search::fuzzy_search_best_n;
let s = "bulko";
let list : Vec<&str> = vec![
"kolbasobulko",
"sandviĉo",
"ŝatas",
"domo",
"emuo",
"fabo",
"fazano"
];
let n : usize = 3;
let res : Vec<(&str, f32)> = fuzzy_search_best_n(s,&list, n);
for (_word, score) in res {
println!("{:?}",score)
}
}
Example: if you have a Vec of Strings you need to convert it to a list of &str
rust
fn works_with_strings() {
use rust_fuzzy_search::fuzzy_search;
let s = String::from("varma");
let list: Vec<String> = vec![String::from("varma vetero"), String::from("varma ĉokolado")];
fuzzy_search(&s, &list.iter().map(String::as_ref).collect::<Vec<&str>>());
}
The crate exposes 5 main functions:
- fuzzycompare will take 2 strings and return a score representing how similar those strings are.
- fuzzysearch applies fuzzycompare to a list of strings and returns a list of tuples: (word, score).
- fuzzysearchsorted is similar to fuzzysearch but orders the output in descending order.
- fuzzysearchthreshold will take an additional f32 as input and returns only tuples with score greater than the threshold.
- fuzzysearchbest_n will take an additional usize arguments and returns the first n tuples.
To use this crate, first add this to your Cargo.toml:
toml
[dependencies]
rust-fuzzy-search = { git = "https://gitlab.com/EnricoCh/rust-fuzzy-search"}
Next, add this to your crate:
```rust extern crate rustfuzzysearch;
use rustfuzzysearch::*;
fn main() { // ... } ```
To build the library use cargo build
To test the library use cargo test
Licensed under either of
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.