didyoumean3

Features

• Built-in fastest levenshtein algorithm
• Support custom return results
• Typescript
• Super fast
• More flexible configuration
• Super small (production.min.js < 2kb) and tree shaking! more info

Usage

install

npm i didyoumean3

// or
yarn add didyoumean3

use case

• base use

import { didyoumean3 } from 'didyoumean3'
// or
const { default: didyoumean3 } = require('didyoumean3');

let input = 'insargrm'
let list = [
  'facebook', 'INSTAgram', ' in stagram', 'baidu', 'twitter', 'wechat', 'instagram', 'linkedin'
]

console.log(didyoumean3(input, list));
// will print:
// {
//   winner: 'instagram', // 🔥 This is the best match, our winner！
//   matched: [
//     {
//       score: 8,
//       target: 'facebook',
//     },
//     {
//       score: 3,
//       target: 'instagram',
//     },
//     {
//       score: 7,
//       target: 'linkedin',
//     },
//     // ...
//   ],
// }

• optional configuration

didyoumea3 has some built-in string formatting configuration items：

• ignore: Ignore case when comparing, default false.

• trim: Use string.trim format the string, default true.

• trimAll: Use regexp /\s+/g formate the string, default false.

• diacritics: Use normalize Api, e.g. 'café' -> 'café'.normalize(), default false

• normalize: Customize the formatting function by yourself.

🔥 If these parameters don't meet your requirements, you can customize the formatting function through normalize.

🔥 When using the custom normalize function, the above string formatting configurations will be ignored.

didyoumean3(input, target, { normalize: (s: string) => s.trim() } );

result: Customize the structure of the results you want to return

// default result may be null or {winner: xx, matched: []}
type Res = null | { matched: any[], winner: string }

// you can custom your own result style!!
const result = (res: Res) => {
  if (!res) return 'nothing matched!'
  else return res
}

didyoumean3(input, target, { result } );

filter: You can filter the results you want, such as those with a score greater than 5

let i2 = 'insargrm';
let l2 = ['facebook', 'instagram', 'linkedin'];
expect(
  didyoumean3(i2, l2, { filter: (score: number, item: any) => score >= 7 })
    ?.matched.length
).toBe(2);

benchmark

didyoumean x 159,428 ops/sec ±1.63% (85 runs sampled)
didyoumean2 x 227,343 ops/sec ±1.26% (90 runs sampled)
didyoumean3 x 400,683 ops/sec ±0.59% (89 runs sampled)
Fastest is didyoumean3

contributors

Both issure and pr are welcome!

license

MIT

changelog

v-1.2.5

1. update the docs
2. val property is no longer be used

v-1.2.4

1. upgrade some depencies
2. update the docs

v-1.2.0

1. delete dice-coefficient algorithm, more small more faster.

v-1.1.4

1. delete opt compartor
2. add opt filter

v-1.1.2

1. update the doc

v-1.1.1

1. remove test file in dist

v-1.1.0

1. improve the performance of the dice-coefficient algorithm
2. add CI/CD
3. coverage test

v-1.0.0

1. refactor the beta version, and we can custom our algorithm
2. we can custom our result now
3. we can custom our normalize string function now
4. builtin dice-coefficient or levenshtein algorithm