Browse Source

language detection library

pull/297/head
friendica 9 years ago
parent
commit
62727012d3
11 changed files with 4713 additions and 0 deletions
  1. +1708
    -0
      library/langdet/Text/LanguageDetect.php
  2. +57
    -0
      library/langdet/Text/LanguageDetect/Exception.php
  3. +341
    -0
      library/langdet/Text/LanguageDetect/ISO639.php
  4. +349
    -0
      library/langdet/Text/LanguageDetect/Parser.php
  5. +1
    -0
      library/langdet/data/lang.dat
  6. +1
    -0
      library/langdet/data/unicode_blocks.dat
  7. +35
    -0
      library/langdet/docs/example_clui.php
  8. +72
    -0
      library/langdet/docs/example_web.php
  9. +21
    -0
      library/langdet/docs/iso.php
  10. +2056
    -0
      library/langdet/tests/Text_LanguageDetectTest.php
  11. +72
    -0
      library/langdet/tests/Text_LanguageDetect_ISO639Test.php

+ 1708
- 0
library/langdet/Text/LanguageDetect.php
File diff suppressed because it is too large
View File


+ 57
- 0
library/langdet/Text/LanguageDetect/Exception.php View File

@ -0,0 +1,57 @@
<?php
class Text_LanguageDetect_Exception extends Exception
{
/**
* Database file could not be found
*/
const DB_NOT_FOUND = 10;
/**
* Database file found, but not readable
*/
const DB_NOT_READABLE = 11;
/**
* Database file is empty
*/
const DB_EMPTY = 12;
/**
* Database contents is not a PHP array
*/
const DB_NOT_ARRAY = 13;
/**
* Magic quotes are activated
*/
const MAGIC_QUOTES = 14;
/**
* Parameter of invalid type passed to method
*/
const PARAM_TYPE = 20;
/**
* Character in parameter is invalid
*/
const INVALID_CHAR = 21;
/**
* Language is not in the database
*/
const UNKNOWN_LANGUAGE = 30;
/**
* Error during block detection
*/
const BLOCK_DETECTION = 40;
/**
* Error while clustering languages
*/
const NO_HIGHEST_KEY = 50;
}

+ 341
- 0
library/langdet/Text/LanguageDetect/ISO639.php View File

@ -0,0 +1,341 @@
<?php
/**
* Part of Text_LanguageDetect
*
* PHP version 5
*
* @category Text
* @package Text_LanguageDetect
* @author Christian Weiske <cweiske@php.net>
* @copyright 2011 Christian Weiske <cweiske@php.net>
* @license http://www.debian.org/misc/bsd.license BSD
* @version SVN: $Id$
* @link http://pear.php.net/package/Text_LanguageDetect/
*/
/**
* Provides a mapping between the languages from lang.dat and the
* ISO 639-1 and ISO-639-2 codes.
*
* Note that this class contains only languages that exist in lang.dat.
*
* @category Text
* @package Text_LanguageDetect
* @author Christian Weiske <cweiske@php.net>
* @copyright 2011 Christian Weiske <cweiske@php.net>
* @license http://www.debian.org/misc/bsd.license BSD
* @link http://www.loc.gov/standards/iso639-2/php/code_list.php
*/
class Text_LanguageDetect_ISO639
{
/**
* Maps all language names from the language database to the
* ISO 639-1 2-letter language code.
*
* NULL indicates that there is no 2-letter code.
*
* @var array
*/
public static $nameToCode2 = array(
'albanian' => 'sq',
'arabic' => 'ar',
'azeri' => 'az',
'bengali' => 'bn',
'bulgarian' => 'bg',
'cebuano' => null,
'croatian' => 'hr',
'czech' => 'cs',
'danish' => 'da',
'dutch' => 'nl',
'english' => 'en',
'estonian' => 'et',
'farsi' => 'fa',
'finnish' => 'fi',
'french' => 'fr',
'german' => 'de',
'hausa' => 'ha',
'hawaiian' => null,
'hindi' => 'hi',
'hungarian' => 'hu',
'icelandic' => 'is',
'indonesian' => 'id',
'italian' => 'it',
'kazakh' => 'kk',
'kyrgyz' => 'ky',
'latin' => 'la',
'latvian' => 'lv',
'lithuanian' => 'lt',
'macedonian' => 'mk',
'mongolian' => 'mn',
'nepali' => 'ne',
'norwegian' => 'no',
'pashto' => 'ps',
'pidgin' => null,
'polish' => 'pl',
'portuguese' => 'pt',
'romanian' => 'ro',
'russian' => 'ru',
'serbian' => 'sr',
'slovak' => 'sk',
'slovene' => 'sl',
'somali' => 'so',
'spanish' => 'es',
'swahili' => 'sw',
'swedish' => 'sv',
'tagalog' => 'tl',
'turkish' => 'tr',
'ukrainian' => 'uk',
'urdu' => 'ur',
'uzbek' => 'uz',
'vietnamese' => 'vi',
'welsh' => 'cy',
);
/**
* Maps all language names from the language database to the
* ISO 639-2 3-letter language code.
*
* @var array
*/
public static $nameToCode3 = array(
'albanian' => 'sqi',
'arabic' => 'ara',
'azeri' => 'aze',
'bengali' => 'ben',
'bulgarian' => 'bul',
'cebuano' => 'ceb',
'croatian' => 'hrv',
'czech' => 'ces',
'danish' => 'dan',
'dutch' => 'nld',
'english' => 'eng',
'estonian' => 'est',
'farsi' => 'fas',
'finnish' => 'fin',
'french' => 'fra',
'german' => 'deu',
'hausa' => 'hau',
'hawaiian' => 'haw',
'hindi' => 'hin',
'hungarian' => 'hun',
'icelandic' => 'isl',
'indonesian' => 'ind',
'italian' => 'ita',
'kazakh' => 'kaz',
'kyrgyz' => 'kir',
'latin' => 'lat',
'latvian' => 'lav',
'lithuanian' => 'lit',
'macedonian' => 'mkd',
'mongolian' => 'mon',
'nepali' => 'nep',
'norwegian' => 'nor',
'pashto' => 'pus',
'pidgin' => 'crp',
'polish' => 'pol',
'portuguese' => 'por',
'romanian' => 'ron',
'russian' => 'rus',
'serbian' => 'srp',
'slovak' => 'slk',
'slovene' => 'slv',
'somali' => 'som',
'spanish' => 'spa',
'swahili' => 'swa',
'swedish' => 'swe',
'tagalog' => 'tgl',
'turkish' => 'tur',
'ukrainian' => 'ukr',
'urdu' => 'urd',
'uzbek' => 'uzb',
'vietnamese' => 'vie',
'welsh' => 'cym',
);
/**
* Maps ISO 639-1 2-letter language codes to the language names
* in the language database
*
* Not all languages have a 2 letter code, so some are missing
*
* @var array
*/
public static $code2ToName = array(
'ar' => 'arabic',
'az' => 'azeri',
'bg' => 'bulgarian',
'bn' => 'bengali',
'cs' => 'czech',
'cy' => 'welsh',
'da' => 'danish',
'de' => 'german',
'en' => 'english',
'es' => 'spanish',
'et' => 'estonian',
'fa' => 'farsi',
'fi' => 'finnish',
'fr' => 'french',
'ha' => 'hausa',
'hi' => 'hindi',
'hr' => 'croatian',
'hu' => 'hungarian',
'id' => 'indonesian',
'is' => 'icelandic',
'it' => 'italian',
'kk' => 'kazakh',
'ky' => 'kyrgyz',
'la' => 'latin',
'lt' => 'lithuanian',
'lv' => 'latvian',
'mk' => 'macedonian',
'mn' => 'mongolian',
'ne' => 'nepali',
'nl' => 'dutch',
'no' => 'norwegian',
'pl' => 'polish',
'ps' => 'pashto',
'pt' => 'portuguese',
'ro' => 'romanian',
'ru' => 'russian',
'sk' => 'slovak',
'sl' => 'slovene',
'so' => 'somali',
'sq' => 'albanian',
'sr' => 'serbian',
'sv' => 'swedish',
'sw' => 'swahili',
'tl' => 'tagalog',
'tr' => 'turkish',
'uk' => 'ukrainian',
'ur' => 'urdu',
'uz' => 'uzbek',
'vi' => 'vietnamese',
);
/**
* Maps ISO 639-2 3-letter language codes to the language names
* in the language database.
*
* @var array
*/
public static $code3ToName = array(
'ara' => 'arabic',
'aze' => 'azeri',
'ben' => 'bengali',
'bul' => 'bulgarian',
'ceb' => 'cebuano',
'ces' => 'czech',
'crp' => 'pidgin',
'cym' => 'welsh',
'dan' => 'danish',
'deu' => 'german',
'eng' => 'english',
'est' => 'estonian',
'fas' => 'farsi',
'fin' => 'finnish',
'fra' => 'french',
'hau' => 'hausa',
'haw' => 'hawaiian',
'hin' => 'hindi',
'hrv' => 'croatian',
'hun' => 'hungarian',
'ind' => 'indonesian',
'isl' => 'icelandic',
'ita' => 'italian',
'kaz' => 'kazakh',
'kir' => 'kyrgyz',
'lat' => 'latin',
'lav' => 'latvian',
'lit' => 'lithuanian',
'mkd' => 'macedonian',
'mon' => 'mongolian',
'nep' => 'nepali',
'nld' => 'dutch',
'nor' => 'norwegian',
'pol' => 'polish',
'por' => 'portuguese',
'pus' => 'pashto',
'rom' => 'romanian',
'rus' => 'russian',
'slk' => 'slovak',
'slv' => 'slovene',
'som' => 'somali',
'spa' => 'spanish',
'sqi' => 'albanian',
'srp' => 'serbian',
'swa' => 'swahili',
'swe' => 'swedish',
'tgl' => 'tagalog',
'tur' => 'turkish',
'ukr' => 'ukrainian',
'urd' => 'urdu',
'uzb' => 'uzbek',
'vie' => 'vietnamese',
);
/**
* Returns the 2-letter ISO 639-1 code for the given language name.
*
* @param string $lang English language name like "swedish"
*
* @return string Two-letter language code (e.g. "sv") or NULL if not found
*/
public static function nameToCode2($lang)
{
$lang = strtolower($lang);
if (!isset(self::$nameToCode2[$lang])) {
return null;
}
return self::$nameToCode2[$lang];
}
/**
* Returns the 3-letter ISO 639-2 code for the given language name.
*
* @param string $lang English language name like "swedish"
*
* @return string Three-letter language code (e.g. "swe") or NULL if not found
*/
public static function nameToCode3($lang)
{
$lang = strtolower($lang);
if (!isset(self::$nameToCode3[$lang])) {
return null;
}
return self::$nameToCode3[$lang];
}
/**
* Returns the language name for the given 2-letter ISO 639-1 code.
*
* @param string $code Two-letter language code (e.g. "sv")
*
* @return string English language name like "swedish"
*/
public static function code2ToName($code)
{
$lang = strtolower($code);
if (!isset(self::$code2ToName[$code])) {
return null;
}
return self::$code2ToName[$code];
}
/**
* Returns the language name for the given 3-letter ISO 639-2 code.
*
* @param string $code Three-letter language code (e.g. "swe")
*
* @return string English language name like "swedish"
*/
public static function code3ToName($code)
{
$lang = strtolower($code);
if (!isset(self::$code3ToName[$code])) {
return null;
}
return self::$code3ToName[$code];
}
}
?>

+ 349
- 0
library/langdet/Text/LanguageDetect/Parser.php View File

@ -0,0 +1,349 @@
<?php
/**
* This class represents a text sample to be parsed.
*
* @category Text
* @package Text_LanguageDetect
* @author Nicholas Pisarro
* @copyright 2006
* @license BSD
* @version CVS: $Id: Parser.php 322327 2012-01-15 17:55:59Z cweiske $
* @link http://pear.php.net/package/Text_LanguageDetect/
* @link http://langdetect.blogspot.com/
*/
/**
* This class represents a text sample to be parsed.
*
* This separates the analysis of a text sample from the primary LanguageDetect
* class. After a new profile has been built, the data can be retrieved using
* the accessor functions.
*
* This class is intended to be used by the Text_LanguageDetect class, not
* end-users.
*
* @category Text
* @package Text_LanguageDetect
* @author Nicholas Pisarro
* @copyright 2006
* @license BSD
* @version release: 0.3.0
*/
class Text_LanguageDetect_Parser extends Text_LanguageDetect
{
/**
* the piece of text being parsed
*
* @access private
* @var string
*/
var $_string;
/**
* stores the trigram frequencies of the sample
*
* @access private
* @var string
*/
var $_trigrams = array();
/**
* stores the trigram ranks of the sample
*
* @access private
* @var array
*/
var $_trigram_ranks = array();
/**
* stores the unicode blocks of the sample
*
* @access private
* @var array
*/
var $_unicode_blocks = array();
/**
* Whether the parser should compile the unicode ranges
*
* @access private
* @var bool
*/
var $_compile_unicode = false;
/**
* Whether the parser should compile trigrams
*
* @access private
* @var bool
*/
var $_compile_trigram = false;
/**
* Whether the trigram parser should pad the beginning of the string
*
* @access private
* @var bool
*/
var $_trigram_pad_start = false;
/**
* Whether the unicode parser should skip non-alphabetical ascii chars
*
* @access private
* @var bool
*/
var $_unicode_skip_symbols = true;
/**
* Constructor
*
* @access private
* @param string $string string to be parsed
*/
function Text_LanguageDetect_Parser($string) {
$this->_string = $string;
}
/**
* Returns true if a string is suitable for parsing
*
* @param string $str input string to test
* @return bool true if acceptable, false if not
*/
public static function validateString($str) {
if (!empty($str) && strlen($str) > 3 && preg_match('/\S/', $str)) {
return true;
} else {
return false;
}
}
/**
* turn on/off trigram counting
*
* @access public
* @param bool $bool true for on, false for off
*/
function prepareTrigram($bool = true)
{
$this->_compile_trigram = $bool;
}
/**
* turn on/off unicode block counting
*
* @access public
* @param bool $bool true for on, false for off
*/
function prepareUnicode($bool = true)
{
$this->_compile_unicode = $bool;
}
/**
* turn on/off padding the beginning of the sample string
*
* @access public
* @param bool $bool true for on, false for off
*/
function setPadStart($bool = true)
{
$this->_trigram_pad_start = $bool;
}
/**
* Should the unicode block counter skip non-alphabetical ascii chars?
*
* @access public
* @param bool $bool true for on, false for off
*/
function setUnicodeSkipSymbols($bool = true)
{
$this->_unicode_skip_symbols = $bool;
}
/**
* Returns the trigram ranks for the text sample
*
* @access public
* @return array trigram ranks in the text sample
*/
function &getTrigramRanks()
{
return $this->_trigram_ranks;
}
/**
* Return the trigram freqency table
*
* only used in testing to make sure the parser is working
*
* @access public
* @return array trigram freqencies in the text sample
*/
function &getTrigramFreqs()
{
return $this->_trigram;
}
/**
* returns the array of unicode blocks
*
* @access public
* @return array unicode blocks in the text sample
*/
function &getUnicodeBlocks()
{
return $this->_unicode_blocks;
}
/**
* Executes the parsing operation
*
* Be sure to call the set*() functions to set options and the
* prepare*() functions first to tell it what kind of data to compute
*
* Afterwards the get*() functions can be used to access the compiled
* information.
*
* @access public
*/
function analyze()
{
$len = strlen($this->_string);
$byte_counter = 0;
// unicode startup
if ($this->_compile_unicode) {
$blocks = $this->_read_unicode_block_db();
$block_count = count($blocks);
$skipped_count = 0;
$unicode_chars = array();
}
// trigram startup
if ($this->_compile_trigram) {
// initialize them as blank so the parser will skip the first two
// (since it skips trigrams with more than 2 contiguous spaces)
$a = ' ';
$b = ' ';
// kludge
// if it finds a valid trigram to start and the start pad option is
// off, then set a variable that will be used to reduce this
// trigram after parsing has finished
if (!$this->_trigram_pad_start) {
$a = $this->_next_char($this->_string, $byte_counter, true);
if ($a != ' ') {
$b = $this->_next_char($this->_string, $byte_counter, true);
$dropone = " $a$b";
}
$byte_counter = 0;
$a = ' ';
$b = ' ';
}
}
while ($byte_counter < $len) {
$char = $this->_next_char($this->_string, $byte_counter, true);
// language trigram detection
if ($this->_compile_trigram) {
if (!($b == ' ' && ($a == ' ' || $char == ' '))) {
if (!isset($this->_trigram[$a . $b . $char])) {
$this->_trigram[$a . $b . $char] = 1;
} else {
$this->_trigram[$a . $b . $char]++;
}
}
$a = $b;
$b = $char;
}
// unicode block detection
if ($this->_compile_unicode) {
if ($this->_unicode_skip_symbols
&& strlen($char) == 1
&& ($char < 'A' || $char > 'z'
|| ($char > 'Z' && $char < 'a'))
&& $char != "'") { // does not skip the apostrophe
// since it's included in the language
// models
$skipped_count++;
continue;
}
// build an array of all the characters
if (isset($unicode_chars[$char])) {
$unicode_chars[$char]++;
} else {
$unicode_chars[$char] = 1;
}
}
// todo: add byte detection here
}
// unicode cleanup
if ($this->_compile_unicode) {
foreach ($unicode_chars as $utf8_char => $count) {
$search_result = $this->_unicode_block_name(
$this->_utf8char2unicode($utf8_char), $blocks, $block_count);
if ($search_result != -1) {
$block_name = $search_result[2];
} else {
$block_name = '[Malformatted]';
}
if (isset($this->_unicode_blocks[$block_name])) {
$this->_unicode_blocks[$block_name] += $count;
} else {
$this->_unicode_blocks[$block_name] = $count;
}
}
}
// trigram cleanup
if ($this->_compile_trigram) {
// pad the end
if ($b != ' ') {
if (!isset($this->_trigram["$a$b "])) {
$this->_trigram["$a$b "] = 1;
} else {
$this->_trigram["$a$b "]++;
}
}
// perl compatibility; Language::Guess does not pad the beginning
// kludge
if (isset($dropone)) {
if ($this->_trigram[$dropone] == 1) {
unset($this->_trigram[$dropone]);
} else {
$this->_trigram[$dropone]--;
}
}
if (!empty($this->_trigram)) {
$this->_trigram_ranks = $this->_arr_rank($this->_trigram);
} else {
$this->_trigram_ranks = array();
}
}
}
}
/* vim: set expandtab tabstop=4 shiftwidth=4 softtabstop=4: */
?>

+ 1
- 0
library/langdet/data/lang.dat
File diff suppressed because it is too large
View File


+ 1
- 0
library/langdet/data/unicode_blocks.dat
File diff suppressed because it is too large
View File


+ 35
- 0
library/langdet/docs/example_clui.php View File

@ -0,0 +1,35 @@
<?php
/**
* example usage (CLI)
*
* @package Text_LanguageDetect
* @version CVS: $Id: example_clui.php 322305 2012-01-15 00:04:17Z clockwerx $
*/
require_once 'Text/LanguageDetect.php';
$l = new Text_LanguageDetect;
$stdin = fopen('php://stdin', 'r');
echo "Supported languages:\n";
$langs = $l->getLanguages();
sort($langs);
echo join(', ', $langs);
echo "\ntotal ", count($langs), "\n\n";
while ($line = fgets($stdin)) {
$result = $l->detect($line, 4);
print_r($result);
$blocks = $l->detectUnicodeBlocks($line, true);
print_r($blocks);
}
fclose($stdin);
unset($l);
/* vim: set expandtab tabstop=4 shiftwidth=4 softtabstop=4: */
?>

+ 72
- 0
library/langdet/docs/example_web.php View File

@ -0,0 +1,72 @@
<?php
/**
* example usage (web)
*
* @package Text_LanguageDetect
* @version CVS: $Id: example_web.php 205493 2006-01-18 00:26:57Z taak $
*/
// browsers will encode multi-byte characters wrong unless they think the page is utf8-encoded
header('Content-type: text/html; charset=utf-8', true);
require_once 'Text/LanguageDetect.php';
$l = new Text_LanguageDetect;
if (isset($_REQUEST['q'])) {
$q = stripslashes($_REQUEST['q']);
}
?>
<html>
<head>
<title>Text_LanguageDetect demonstration</title>
</head>
<body>
<h2>Text_LanguageDetect</h2>
<?
echo "<small>Supported languages:\n";
$langs = $l->getLanguages();
sort($langs);
foreach ($langs as $lang) {
echo ucfirst($lang), ', ';
$i++;
}
echo "<br />total $i</small><br /><br />";
?>
<form method="post">
Enter text to identify language (at least a couple of sentences):<br />
<textarea name="q" wrap="virtual" cols="80" rows="8"><?= $q ?></textarea>
<br />
<input type="submit" value="Submit" />
</form>
<?
if (isset($q) && strlen($q)) {
$len = $l->utf8strlen($q);
if ($len < 20) { // this value picked somewhat arbitrarily
echo "Warning: string not very long ($len chars)<br />\n";
}
$result = $l->detectConfidence($q);
if ($result == null) {
echo "Text_LanguageDetect cannot identify this piece of text. <br /><br />\n";
} else {
echo "Text_LanguageDetect thinks this text is written in <b>{$result['language']}</b> ({$result['similarity']}, {$result['confidence']})<br /><br />\n";
}
$result = $l->detectUnicodeBlocks($q, false);
if (!empty($result)) {
arsort($result);
echo "Unicode blocks present: ", join(', ', array_keys($result)), "\n<br /><br />";
}
}
unset($l);
/* vim: set expandtab tabstop=4 shiftwidth=4 softtabstop=4: */
?>
</body></html>

+ 21
- 0
library/langdet/docs/iso.php View File

@ -0,0 +1,21 @@
<?php
/**
* Demonstrates how to use ISO language codes.
*
* The "name mode" changes the way languages are accepted and returned.
*/
require_once 'Text/LanguageDetect.php';
$l = new Text_LanguageDetect();
//will output the ISO 639-1 two-letter language code
// "de"
$l->setNameMode(2);
echo $l->detectSimple('Das ist ein kleiner Text') . "\n";
//will output the ISO 639-2 three-letter language code
// "deu"
$l->setNameMode(3);
echo $l->detectSimple('Das ist ein kleiner Text') . "\n";
?>

+ 2056
- 0
library/langdet/tests/Text_LanguageDetectTest.php
File diff suppressed because it is too large
View File


+ 72
- 0
library/langdet/tests/Text_LanguageDetect_ISO639Test.php View File

@ -0,0 +1,72 @@
<?php
set_include_path(
__DIR__ . '/../' . PATH_SEPARATOR . get_include_path()
);
require_once 'Text/LanguageDetect/ISO639.php';
class Text_LanguageDetect_ISO639Test extends PHPUnit_Framework_TestCase
{
public function testNameToCode2()
{
$this->assertEquals(
'de',
Text_LanguageDetect_ISO639::nameToCode2('german')
);
}
public function testNameToCode2Fail()
{
$this->assertNull(
Text_LanguageDetect_ISO639::nameToCode2('doesnotexist')
);
}
public function testNameToCode3()
{
$this->assertEquals(
'fra',
Text_LanguageDetect_ISO639::nameToCode3('french')
);
}
public function testNameToCode3Fail()
{
$this->assertNull(
Text_LanguageDetect_ISO639::nameToCode3('doesnotexist')
);
}
public function testCode2ToName()
{
$this->assertEquals(
'english',
Text_LanguageDetect_ISO639::code2ToName('en')
);
}
public function testCode2ToNameFail()
{
$this->assertNull(
Text_LanguageDetect_ISO639::code2ToName('nx')
);
}
public function testCode3ToName()
{
$this->assertEquals(
'romanian',
Text_LanguageDetect_ISO639::code3ToName('rom')
);
}
public function testCode3ToNameFail()
{
$this->assertNull(
Text_LanguageDetect_ISO639::code3ToName('nxx')
);
}
}
?>

Loading…
Cancel
Save