foodshare-search

High-performance fuzzy search and text matching utilities.

Features

• Relevance Scoring - Multi-level scoring (exact, starts-with, contains, fuzzy)
• Fuzzy Matching - Find matches even with typos
• Levenshtein Distance - Calculate edit distance between strings
• Unicode Support - Proper handling of international characters
• WASM Support - Compile to WebAssembly for browser usage

Installation

Add to your Cargo.toml:

[dependencies]
foodshare-search = "1.3"

Feature Flags

Usage

Relevance Scoring

Calculate how well a query matches text:

use foodshare_search::calculate_relevance;

// Returns score from 0-50 (higher is better)
let score = calculate_relevance("Hello World", "hello");
// score = 40 (StartsWith)

let score = calculate_relevance("Say Hello", "hello");
// score = 30 (WordBoundary)

let score = calculate_relevance("Hello", "hello");
// score = 50 (Exact match)

Score Levels

Fuzzy Matching

Check if all query characters appear in text in order:

use foodshare_search::fuzzy_match;

assert!(fuzzy_match("Hello World", "hwo"));  // h...w...o
assert!(fuzzy_match("Hello", "hel"));        // true
assert!(!fuzzy_match("Hello", "leh"));       // wrong order

Levenshtein Distance

Calculate the number of edits to transform one string into another:

use foodshare_search::levenshtein_distance;

let dist = levenshtein_distance("hello", "hallo");  // 1 (substitute)
let dist = levenshtein_distance("hello", "helo");   // 1 (delete)
let dist = levenshtein_distance("cat", "dog");      // 3 (all different)

Search and Rank Results

use foodshare_search::calculate_relevance;

struct Product {
    id: String,
    name: String,
}

fn search_products(query: &str, products: &[Product]) -> Vec<(&Product, u32)> {
    let mut results: Vec<_> = products
        .iter()
        .map(|p| (p, calculate_relevance(&p.name, query)))
        .filter(|(_, score)| *score > 0)
        .collect();
    
    results.sort_by(|a, b| b.1.cmp(&a.1));
    results
}

WASM Usage

For browser usage, see @foodshare/search-wasm.

import init, { relevance_score, search_items } from '@foodshare/search-wasm';

await init();

// Single score
const score = relevance_score("hello", "Hello World");

// Batch search
const items = JSON.stringify([
  { id: "1", text: "Fresh Apples" },
  { id: "2", text: "Apple Pie" },
]);
const results = JSON.parse(search_items("apple", items, 10));

Performance

• ~10x faster than JavaScript implementations
• Efficient memory usage with optimized algorithms
• Optional parallel processing with rayon

License

MIT License - see LICENSE for details.