string-metric
Version:
Get string similarity in JavaScript or TypeScript
172 lines (120 loc) • 4.62 kB
Markdown
# String Metric
[](https://travis-ci.com/hellojayjay/string-metric) [](https://coveralls.io/github/hellojayjay/string-metric?branch=master)
A library implementing different string similarity and distance measures, and Implement by **TypeScript**. Also, you can use in **JavaScript**.
Algorithm reference [java-string-similarity](https://github.com/tdebatty/java-string-similarity)
## Install
`npm install string-metric`
## Progress
| Algorithm | Complete? |
| ------------------------------------------------- | --------- |
| [Jaro-Winkler](#Jaro-Winkler) | Yes |
| [Levenshtein](#Levenshtein) | Yes |
| [Normalized-Levenshtein](#Normalized-Levenshtein) | Yes |
| [Weighted-Levenshtein](#Weighted-Levenshtein) | Yes |
| [Damerau](#Damerau) | Yes |
| [Optimal-String-Alignment](#Optimal-String-Alignment) |Yes|
| [Longest-Common-Subsequence](#Longest-Common-Subsequence) | Yes |
| [Metric-Longest-Common-Subsequence](#Metric-Longest-Common-Subsequence) | Yes |
| [N-Gram](#N-Gram) | Yes |
| Q-Gram | No |
| Shingle (n-gram) based algorithms | No |
| Cosine similarity | No |
| Jaccard index | No |
| Sorensen-Dice coefficient | No |
| Ratcliff-Obershelp | No |
## Jaro-Winkler
For more specs, please go to `tests/JaroWinkler.spec.ts` in the repository.
```typescript
const instance = new JaroWinkler();
const s1 = 'My string';
const s2 = 'My string';
instance.similarity(s1, s2); // 1
const s1 = 'My string';
const s2 = 'My tsring';
instance.similarity(s1, s2); // 0.974074
const s1 = 'My string';
const s2 = 'My ntrisg';
instance.similarity(s1, s2); // 0.896296
```
## Levenshtein
For more specs, please go to `tests/Levenshtein.spec.ts` in the repository.
```typescript
const instance = new Levenshtein();
const s1 = 'My string';
const s2 = 'My string';
instance.distance(s1, s2); // 0
const s1 = 'My string';
const s2 = 'My tring';
instance.distance(s1, s2); // 1
const s1 = 'My string';
const s2 = 'M string2';
instance.distance(s1, s2); // 2
```
## Normalized-Levenshtein
For more specs, please go to `tests/NormalizedLevenshtein.spec.ts` in the repository.
```typescript
const instance = new NormalizedLevenshtein();
```
## Weighted-Levenshtein
For more specs, please go to `tests/WeightedLevenshtein.spec.ts` in the repository.
```typescript
const instance = new WeightedLevenshtein();
```
## Damerau
For more specs, please go to `tests/Damerau.spec.ts` in the repository.
```typescript
const instance = new Damerau();
const s1 = 'ABCDEF';
const s2 = 'ABDCEF';
instance.distance(s1, s2); // 1
const s1 = 'ABCDEF';
const s2 = 'BACDFE';
instance.distance(s1, s2); // 2
const s1 = 'ABCDEF';
const s2 = 'ABCDE';
instance.distance(s1, s2); // 1
```
## Optimal-String-Alignment
For more specs, please go to `tests/OptimalStringAlignment.spec.ts` in the repository.
```typescript
const instance = new OptimalStringAlignment();
const s1 = 'ABDCEF';
const s2 = 'ADCEF';
instance.distance(s1, s2); // 1
const s1 = 'BAC';
const s2 = 'CAB';
instance.distance(s1, s2); // 2
const s1 = 'CA';
const s2 = 'ABC';
instance.distance(s1, s2); // 3
```
## Longest-Common-Subsequence
For more specs, please go to `tests/LongestCommonSubsequence.spec.ts` in the repository.
```typescript
const instance = new LongestCommonSubsequence();
const s1 = 'AGCAT';
const s2 = 'GAC';
instance.distance(s1, s2); // 4
const s1 = 'AGCAT';
const s2 = 'AGCT';
instance.distance(s1, s2); // 1
```
## Metric-Longest-Common-Subsequence
For more specs, please go to `tests/MetricLCS.spec.ts` in the repository.
```typescript
const instance = new MetricLCS();
```
## N-Gram
For more specs, please go to `tests/NGram.spec.ts` in the repository.
```typescript
const instance = new NGram();
const s1 = 'SIJK';
const s2 = 'SIJK';
instance.distance(s1, s2); // 0
const s0 = 'ABABABAB';
const s1 = 'ABCABCABCABC';
const s2 = 'POIULKJH';
instance.distance(s0, s1) < instance.distance(s0, s2); // true
```