Trigrams for 500+ languages.

• What is this?
• When should I use this?
• Install
• Use
• API
  • min()
  • top()
• Data
• Compatibility
• Contribute
• Security
• License

This package exposes all trigrams for natural languages. Based on the most translated copyright-free document on this planet: UDHR.

When you are dealing with natural language detection.

This package is ESM only. In Node.js (version 18+), install with npm:

npm install trigrams

In Deno with esm.sh:

import {min, top} from 'https://esm.sh/trigrams@6'

In browsers with esm.sh:

<script type="module"> import {min, top} from 'https://esm.sh/trigrams@6?bundle' </script>

import {min, top} from 'trigrams' console.log((await min()).nld) console.log((await top()).pam)

Yields:

[ // 300 top trigrams. ' ar', 'eer', 'tij', // … 'de ', 'an ', 'en ' // Most common trigram. ]

{ // 300 top trigrams. 'isa': 6, 'upa': 6, 'i k': 6, // … 'ang': 273, 'ing': 282, 'ng ': 572 // Most common trigram with how often it was found. }

This package exports the identifiers min and top. It exports no TypeScript types. There is no default export.

Get top trigrams.

Returns a promise resolving to arrays containing the top 300 trigrams sorted from least occurring to most occurring (Promise<Record<string, Array<string>>>).

Get top trigrams to occurrence counts.

Returns a promise resolving to an object mapping UDHR in Unicode codes to objects mapping the top 300 trigrams to occurrence counts (Promise<Record<string, Record<string, number>>>).

The trigrams are based on the unicode versions of the universal declaration of human rights.

The files are created from all paragraphs made available by wooorm/udhr and do not include headings and such.

Before creating trigrams,

• the unicode characters from \u0021 to \u0040 (both including) are removed
• one or more white space characters (\s+) are replaced with a single space
• alphabetic characters are lower cased ([A-Z])

Additionally, the input is padded with two spaces on both sides.

This package is at least compatible with all maintained versions of Node.js. As of now, that is Node.js 18+. It also works in Deno and modern browsers.

Yes please! See How to Contribute to Open Source.

This package is safe.

MIT © Titus Wormer