zectbynmo/franc

Natural language detection

Detect the language of text.

What’s so cool about franc?

1. franc supports more languages^(†) than any other library, or Google;
2. franc is easily forked to support 339 languages;
3. franc is just as fast as the competition.

† - If humans write in the language, on the web, and the language has more than one million speakers, franc detects it.

Installation

npm:

npm install franc

franc is also available pre-built as an AMD, CommonJS, and globals module, supporting 75, 176, and 339 languages.

Usage

var franc = require('franc');

franc('Alle menslike wesens word vry'); // "afr"
franc('এটি একটি ভাষা একক IBM স্ক্রিপ্ট'); // "ben"
franc('Alle mennesker er født frie og'); // "nno"
franc(''); // "und"

franc.all('O Brasil caiu 26 posições');
/*
 * [
 *  [ 'por', 1 ],
 *  [ 'src', 0.8948665297741273 ],
 *  [ 'glg', 0.8862422997946612 ],
 *  [ 'snn', 0.8804928131416838 ],
 *  [ 'bos', 0.8394250513347022 ],
 *  [ 'hrv', 0.8336755646817249 ],
 *  [ 'lav', 0.833264887063655 ],
 *  [ 'cat', 0.8303901437371664 ],
 *  [ 'spa', 0.8242299794661191 ],
 *  [ 'bam', 0.8242299794661191 ],
 *  [ 'sco', 0.8069815195071869 ],
 *  [ 'rmy', 0.7839835728952772 ],
 *   ...
 * ]
 */

/* "und" is returned for too-short input: */
franc('the'); // 'und'

/* You can change what’s too short (default: 10): */
franc('the', {'minLength': 3}); // 'sco'

/* Provide a whitelist: */
franc.all('O Brasil caiu 26 posições', {
    'whitelist' : ['por', 'src', 'glg', 'spa']
});
/*
 * [
 *   [ 'por', 1 ],
 *   [ 'src', 0.8948665297741273 ],
 *   [ 'glg', 0.8862422997946612 ],
 *   [ 'spa', 0.8242299794661191 ]
 * ]
*/

/* Provide a blacklist: */
franc.all('O Brasil caiu 26 posições', {
    'blacklist' : ['src', 'glg', 'lav']
});
/*
 * [
 *   [ 'por', 1 ],
 *   [ 'snn', 0.8804928131416838 ],
 *   [ 'bos', 0.8394250513347022 ],
 *   [ 'hrv', 0.8336755646817249 ],
 *   [ 'cat', 0.8303901437371664 ],
 *   [ 'spa', 0.8242299794661191 ],
 *   [ 'bam', 0.8242299794661191 ],
 *   [ 'sco', 0.8069815195071869 ],
 *   [ 'rmy', 0.7839835728952772 ],
 *   ...
 * ]
 */

CLI

Install:

npm install --global franc

Use:

Usage: franc [options] <string>

Detect the language of text

Options:

  -h, --help                    output usage information
  -v, --version                 output version number
  -m, --min-length <number>     minimum length to accept
  -w, --whitelist <string>      allow languages
  -b, --blacklist <string>      disallow languages

Usage:

# output language
$ franc "Alle menslike wesens word vry"
# afr

# output language from stdin (expects utf8)
$ echo "এটি একটি ভাষা একক IBM স্ক্রিপ্ট" | franc
# ben

# blacklist certain languages
$ franc --blacklist por,glg "O Brasil caiu 26 posições"
# src

# output language from stdin with whitelist
$ echo "Alle mennesker er født frie og" | franc --whitelist nob,dan
# nob

Supported languages

franc supports 176 “languages”, by default. For a complete list, check out supported-languages.md.

Supporting more or less languages

Supporting more or less languages is easy: fork the project and run the following:

npm install # Install development dependencies.
export THRESHOLD=100000 # Set minimum speakers to a 100,000.
npm run build # Run the `build` script.

The above would create a version of franc with support for any language with 100,000 or more speakers. To support all languages, even dead ones like Latin, specify -1.

Browser

I’ve compiled three versions of franc for use in the browser. They’re UMD compliant: they work with AMD, CommonJS, and <script>s.

• franc.js — franc with support for languages with 8 million or more speakers (75 languages);

• franc-most.js — franc with support for languages with 1 million or more speakers (175 languages, the same as the npm version);

• franc-all.js — franc with support for all languages (339 languages, carful, huge!).

Derivation

Franc is a derivative work from guess-language (Python, LGPL), guesslanguage (C++, LGPL), and Language::Guess (Perl, GPL). Their creators granted me the rights to distribute franc under the MIT license: respectively, Maciej Ceglowski, Jacob R. Rideout, and Kent S. Johnson.

License

MIT © Titus Wormer