nitotm/efficient-language-detector

Fast and accurate natural language detection. Detector written in PHP. Nito-ELD, ELD.

v2.1.0 2023-10-01 16:14 UTC

This package is not auto-updated.

Last update: 2024-04-07 19:17:49 UTC


README

supported PHP versions license supported languages

Efficient language detector (Nito-ELD or ELD) is a fast and accurate natural language detection software, written in PHP, with a speed comparable to existent fast C++ compiled detectors, and accuracy within the range of the heaviest and slowest detectors.

It has no dependencies, 100% PHP, easy installation, all it's needed is PHP with the mb extension.
ELD is also available in Javascript and Python.

  1. Installation
  2. How to use
  3. Benchmarks
  4. Testing
  5. Languages

Installation

$ composer require nitotm/efficient-language-detector

Alternatively, download / clone the files will work just fine.

How to use?

detect() expects a UTF-8 string, returns an object, with a value (ISO 639-1 code or null) named language

$eld = new Nitotm\Eld\LanguageDetector();
// require_once 'src/LanguageDetector.php'; To load ELD without composer/autoload. Update path.

var_dump($eld->detect('Hola, cómo te llamas?'));
// object( language => 'es', scores => ['es' => 0.5, 'et' => 0.2], isReliable() => true )
// object( language => null|string, scores => null|array, isReliable() => bool )

print $eld->detect('Hola, cómo te llamas?')->language;
// 'es'

// if cleanText(True), detect() removes Urls, .com domains, emails, alphanumerical & numbers
$eld->cleanText(true); // Default is false
  • To reduce the languages to be detected, there are 3 different options, they only need to be executed once. (Check available languages below)
$langSubset = ['en','es','fr','it','nl','de'];

/* Option #1
 With dynamicLangSubset() the detector executes normally, and then filters excluded languages
*/
$eld->dynamicLangSubset($langSubset);
// Returns an object with an ?array property named 'languages', with the validated languages

/* Option #2
 langSubset($langs, save: true) removes excluded languages from the Ngrams database
 For a single detection is slower than dynamicLangSubset(), but for several will be faster
 If $save option is true (default), the new Ngrams subset will be stored, and loaded next call
*/
var_dump($eld->langSubset($langSubset)); // returns the subset file name if saved
// Object ( success => bool, languages => null|[], error => null|string, file => null|string )

// To remove either dynamicLangSubset() or langSubset(), call the methods without any argument
$eld->langSubset(); 

/* Option #3
 Finally the optimal way to regularly use a language subset: we create the instance with a file
 The file in the argument can be a subset by langSubset() or another database like ngramsL60.php
*/
$eld_L = new Nitotm\Eld\LanguageDetector('ngramsL60.php');

If needed, we can get the current status of eld: languages, database type and subset

var_dump($eld->info());

Benchmarks

I compared ELD with a different variety of detectors, since the interesting part is the algorithm.

URL Version Language
https://github.com/nitotm/efficient-language-detector/ 1.0.0 PHP
https://github.com/pemistahl/lingua-py 1.3.2 Python
https://github.com/CLD2Owners/cld2 Aug 21, 2015 C++
https://github.com/google/cld3 Aug 28, 2020 C++
https://github.com/wooorm/franc 6.1.0 Javascript
https://github.com/patrickschur/language-detection 5.2.0 PHP

Benchmarks: Tweets: 760KB, short sentences of 140 chars max.; Big test: 10MB, sentences in all 60 languages supported; Sentences: 8MB, this is the Lingua sentences test, minus unsupported languages.
Short sentences is what ELD and most detectors focus on, as very short text is unreliable, but I included the Lingua Word pairs 1.5MB, and Single words 880KB tests to see how they all compare beyond their reliable limits.

These are the results, first, execution time and then accuracy.

time table

accuracy table

1. Lingua could have a small advantage as it participates with 54 languages, 6 less.
2. CLD2 and CLD3, return a list of languages, the ones not included in this test where discarded, but usually they return one language, I believe they have a disadvantage. Also, I confirm the results of CLD2 for short text are correct, contrary to the test on the Lingua page, they did not use the parameter "bestEffort = True", their benchmark for CLD2 is unfair.

Lingua is the average accuracy winner, but at what cost, the same test that in ELD or CLD2 lasts 2 seconds, in Lingua takes more than 5 hours! It acts like a brute-force software. Also, its lead comes from single and pair words, which are unreliable regardless.

I added ELD-L for comparison, which has a 2.3x bigger database, but only increases execution time marginally, a testament to the efficiency of the algorithm. ELD-L is not the main database as it does not improve language detection in sentences.

Here is the average, per benchmark, of Tweets, Big test & Sentences.

Sentences tests average

Testing

  • To make sure everything works on your setup, you can execute the following file:
$ php efficient-language-detector/tests/tests.php # Update path
  • Also, for composer "autoload-dev", the following line will also execute the tests
new Nitotm\Eld\Tests\TestsAutoload();
  • To run the accuracy benchmarks run the benchmark/bench.php file .

Languages

These are the ISO 639-1 codes of the 60 supported languages for Nito-ELD v1

am, ar, az, be, bg, bn, ca, cs, da, de, el, en, es, et, eu, fa, fi, fr, gu, he, hi, hr, hu, hy, is, it, ja, ka, kn, ko, ku, lo, lt, lv, ml, mr, ms, nl, no, or, pa, pl, pt, ro, ru, sk, sl, sq, sr, sv, ta, te, th, tl, tr, uk, ur, vi, yo, zh

Full name languages:

Amharic, Arabic, Azerbaijani (Latin), Belarusian, Bulgarian, Bengali, Catalan, Czech, Danish, German, Greek, English, Spanish, Estonian, Basque, Persian, Finnish, French, Gujarati, Hebrew, Hindi, Croatian, Hungarian, Armenian, Icelandic, Italian, Japanese, Georgian, Kannada, Korean, Kurdish (Arabic), Lao, Lithuanian, Latvian, Malayalam, Marathi, Malay (Latin), Dutch, Norwegian, Oriya, Punjabi, Polish, Portuguese, Romanian, Russian, Slovak, Slovene, Albanian, Serbian (Cyrillic), Swedish, Tamil, Telugu, Thai, Tagalog, Turkish, Ukrainian, Urdu, Vietnamese, Yoruba, Chinese

Future improvements

  • Train from bigger datasets, and more languages.
  • The tokenizer could separate characters from languages that have their own alphabet, potentially improving accuracy and reducing the N-grams database. Retraining and testing is needed.

Donate / Hire
If you wish to Donate for open source improvements, Hire me for private modifications / upgrades, or to Contact me, use the following link: https://linktr.ee/nitotm