add text_languagedetect to guess article language for better hyphenation

(bump schema)
This commit is contained in:
Andrew Dolgov 2013-07-31 10:30:14 +04:00
parent f035e6dc82
commit 6b4617970f
13 changed files with 2498 additions and 6 deletions

View file

@ -621,7 +621,7 @@ class Feeds extends Handler_Protected {
}
$reply['content'] .= "</div>";
$reply['content'] .= "<div class=\"cdmContentInner\" lang=\"en\">";
$reply['content'] .= "<div class=\"cdmContentInner\" lang=\"".$line['lang']."\">";
if ($line["orig_feed_id"]) {

View file

@ -1,6 +1,6 @@
<?php
define('EXPECTED_CONFIG_VERSION', 26);
define('SCHEMA_VERSION', 121);
define('SCHEMA_VERSION', 122);
define('LABEL_BASE_INDEX', -1024);
define('PLUGIN_FEED_BASE_INDEX', -128);
@ -87,6 +87,7 @@
require_once "lib/accept-to-gettext.php";
require_once "lib/gettext/gettext.inc";
require_once "lib/languagedetect/LanguageDetect.php";
function startup_gettext() {
@ -2650,6 +2651,7 @@
comments,
int_id,
uuid,
lang,
hide_images,
unread,feed_id,marked,published,link,last_read,orig_feed_id,
last_marked, last_published,
@ -2692,6 +2694,7 @@
"tag_cache," .
"label_cache," .
"link," .
"lang," .
"uuid," .
"last_read," .
"(SELECT hide_images FROM ttrss_feeds WHERE id = feed_id) AS hide_images," .
@ -3118,7 +3121,7 @@
ccache_update($feed_id, $owner_uid);
}
$result = db_query("SELECT id,title,link,content,feed_id,comments,int_id,
$result = db_query("SELECT id,title,link,content,feed_id,comments,int_id,lang,
".SUBSTRING_FOR_DATE."(updated,1,16) as updated,
(SELECT site_url FROM ttrss_feeds WHERE id = feed_id) as site_url,
(SELECT hide_images FROM ttrss_feeds WHERE id = feed_id) as hide_images,
@ -3290,7 +3293,7 @@
}
$rv['content'] .= "</div>";
$rv['content'] .= "<div class=\"postContent\" lang=\"en\">";
$rv['content'] .= "<div class=\"postContent\" lang=\"".$line['lang']."\">";
$rv['content'] .= $line["content"];
$rv['content'] .= format_article_enclosures($id,

View file

@ -354,6 +354,11 @@
$rss->init();
}
require_once "lib/languagedetect/LanguageDetect.php";
$lang = new Text_LanguageDetect();
$lang->setNameMode(2);
// print_r($rss);
$feed = db_escape_string($feed);
@ -565,6 +570,15 @@
print "\n";
}
$entry_language = $lang->detect($entry_content, 1);
if (count($entry_language) > 0) {
$entry_language = array_keys($entry_language);
$entry_language = db_escape_string($entry_language[0]);
_debug("detected language: $entry_language", $debug_enabled);
}
$entry_comments = $item->get_comments_url();
$entry_author = $item->get_author();
@ -677,6 +691,7 @@
comments,
num_comments,
plugin_data,
lang,
author)
VALUES
('$entry_title',
@ -691,6 +706,7 @@
'$entry_comments',
'$num_comments',
'$entry_plugin_data',
'$entry_language',
'$entry_author')");
$article_labels = array();

File diff suppressed because it is too large Load diff

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;
}

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];
}
}
?>

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: */
?>

File diff suppressed because one or more lines are too long

File diff suppressed because one or more lines are too long

View file

@ -163,6 +163,7 @@ create table ttrss_entries (id integer not null primary key auto_increment,
date_updated datetime not null,
num_comments integer not null default 0,
plugin_data longtext,
lang varchar(2),
comments varchar(250) not null default '',
author varchar(250) not null default '') ENGINE=InnoDB DEFAULT CHARSET=UTF8;
@ -301,7 +302,7 @@ create table ttrss_tags (id integer primary key auto_increment,
create table ttrss_version (schema_version int not null) ENGINE=InnoDB DEFAULT CHARSET=UTF8;
insert into ttrss_version values (121);
insert into ttrss_version values (122);
create table ttrss_enclosures (id integer primary key auto_increment,
content_url text not null,

View file

@ -144,6 +144,7 @@ create table ttrss_entries (id serial not null primary key,
num_comments integer not null default 0,
comments varchar(250) not null default '',
plugin_data text,
lang varchar(2),
author varchar(250) not null default '');
create index ttrss_entries_guid_index on ttrss_entries(guid);
@ -259,7 +260,7 @@ create index ttrss_tags_post_int_id_idx on ttrss_tags(post_int_id);
create table ttrss_version (schema_version int not null);
insert into ttrss_version values (121);
insert into ttrss_version values (122);
create table ttrss_enclosures (id serial not null primary key,
content_url text not null,

View file

@ -0,0 +1,7 @@
begin;
alter table ttrss_entries add column lang varchar(2);
update ttrss_version set schema_version = 122;
commit;

View file

@ -0,0 +1,7 @@
begin;
alter table ttrss_entries add column lang varchar(2);
update ttrss_version set schema_version = 122;
commit;