Skip to content

a library for matching human-quality input to a list of potential matches using the Levenshtein distance algorithm

License

Notifications You must be signed in to change notification settings

foray1010/didyoumean2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

didyoumean2

Build Status codecov.io

node npm npm

didyoumean2 is a library for matching human-quality input to a list of potential matches using the Levenshtein distance algorithm. It is inspired by didyoumean.js.

Why reinventing the wheel

  1. Based on fastest-levenshtein, the fastest JS implementation of the Levenshtein distance algorithm

  2. ~100% faster than didyoumean.js

  3. Well tested with 100% coverage

  4. Static type checking with TypeScript

  5. More control on what kind of matches you want to return

  6. Support matching object's path instead of just key

Installation

npm install didyoumean2
const didYouMean = require('didyoumean2').default
// or if you are using TypeScript or ES module
import didYouMean from 'didyoumean2'

// you can also access to Enums via:
const {
  default: didYouMean,
  ReturnTypeEnums,
  ThresholdTypeEnums,
} = require('didyoumean2')
// or
import didYouMean, { ReturnTypeEnums, ThresholdTypeEnums } from 'didyoumean2'

Development Setup

We are using corepack to manage the yarn version

corepack enable

Usage

didYouMean(input, matchList[, options])
  • input {string}: A string that you are not sure and want to match with matchList

  • matchList {Object[]|string[]}: A List for matching with input

  • options {Object}(optional): An options that allows you to modify the behavior

  • @return {Array|null|Object|string}: A list of or single matched result(s), return object if match is {Object[]}

Options

caseSensitive {boolean}

  • default: false

  • Perform case-sensitive matching

deburr {boolean}

matchPath {Array}

  • default: []

  • If your matchList is an array of object, you must use matchPath to point to the string that you want to match

  • Refer to ramda R.path for how to define the path, e.g. ['obj', 'array', 0, 'key']

returnType {string}

  • default: ReturnTypeEnums.FIRST_CLOSEST_MATCH
returnType Description
ReturnTypeEnums.ALL_CLOSEST_MATCHES Return all matches with the closest value to the input in array
ReturnTypeEnums.ALL_MATCHES Return all matches in array
ReturnTypeEnums.ALL_SORTED_MATCHES Return all matches in array, sorted from closest to furthest
ReturnTypeEnums.FIRST_CLOSEST_MATCH Return first match from ReturnTypeEnums.ALL_CLOSEST_MATCHES
ReturnTypeEnums.FIRST_MATCH Return first match (FASTEST)

threshold {integer|number}

  • depends on thresholdType

  • type: {number} (similarity) or {integer} (edit-distance)

  • default: 0.4 (similarity) or 20 (edit-distance)

  • If the result is larger (similarity) or smaller (edit-distance) than or equal to the threshold, that result is matched

thresholdType {string}

  • default: ThresholdTypeEnums.SIMILARITY
thresholdType Description
ThresholdTypeEnums.EDIT_DISTANCE Refer to Levenshtein distance algorithm, must be integer, lower value means more similar
ThresholdTypeEnums.SIMILARITY l = max(input.length, matchItem.length), similarity = (l - editDistance) / l, number from 0 to 1, higher value means more similar

trimSpaces {boolean}

  • default: true

  • Remove noises when matching

  • Trim all starting and ending spaces, and concatenate all continuous spaces to one space

Test

Before all:

npm install -g yarn
yarn install

Unit test and coverage:

yarn test

Linter:

yarn lint

About

a library for matching human-quality input to a list of potential matches using the Levenshtein distance algorithm

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published