<?php
+/**
+ * @package MediaWiki
+ * @subpackage Language
+ */
+require_once( "LanguageConverter.php" );
require_once( "LanguageZh_cn.php");
require_once( "LanguageZh_tw.php");
+require_once( "LanguageZh_sg.php");
+require_once( "LanguageZh_hk.php");
-/* caching the conversion tables */
-$zh2TW = $wgMemc->get($key1 = "$wgDBname:zhConvert:tw");
-$zh2CN = $wgMemc->get($key2 = "$wgDBname:zhConvert:cn");
-$zh2SG = $wgMemc->get($key3 = "$wgDBname:zhConvert:sg");
-$zh2HK = $wgMemc->get($key4 = "$wgDBname:zhConvert:hk");
-if(empty($zhSimp2Trad) || empty($zhTrad2Simp)) {
- require_once("includes/ZhConversion.php");
- $wgMemc->set($key1, $zh2TW);
- $wgMemc->set($key2, $zh2CN);
- $wgMemc->set($key3, $zh2SG);
- $wgMemc->set($key4, $zh2HK);
+class ZhConverter extends LanguageConverter {
+ function loadDefaultTables() {
+ require( "includes/ZhConversion.php" );
+ $this->mTables = array();
+ $this->mTables['zh-cn'] = $zh2CN;
+ $this->mTables['zh-tw'] = $zh2TW;
+ $this->mTables['zh-sg'] = array_merge($zh2CN, $zh2SG);
+ $this->mTables['zh-hk'] = array_merge($zh2TW, $zh2HK);
+ $this->mTables['zh'] = array();
+ }
+
+ function postLoadTables() {
+ $this->mTables['zh-sg'] = array_merge($this->mTables['zh-cn'], $this->mTables['zh-sg']);
+ $this->mTables['zh-hk'] = array_merge($this->mTables['zh-tw'], $this->mTables['zh-hk']);
+ }
+
+ /* there shouldn't be any latin text in Chinese conversion, so no need
+ to mark anything
+ */
+ function markNoConversion($text) {
+ return $text;
+ }
+
+ function convertCategoryKey( $key ) {
+ return $this->autoConvert( $key, 'zh-cn' );
+ }
}
+
/* class that handles both Traditional and Simplified Chinese
- right now it only distinguish zh_cn and zh_tw (actuall, zh_cn and
- non-zh_cn), will add support for zh_sg, zh_hk, etc, later.
+ right now it only distinguish zh_cn, zh_tw, zh_sg and zh_hk.
*/
class LanguageZh extends LanguageZh_cn {
- var $mZhLanguageCode=false;
-
function LanguageZh() {
- $this->mZhLanguageCode = $this->getPreferredVariant();
+ global $wgHooks;
+ $this->mConverter = new ZhConverter($this, 'zh',
+ array('zh', 'zh-cn', 'zh-tw', 'zh-sg', 'zh-hk'),
+ array('zh'=>'zh-cn',
+ 'zh-cn'=>'zh-sg',
+ 'zh-sg'=>'zh-cn',
+ 'zh-tw'=>'zh-hk',
+ 'zh-hk'=>'zh-tw'));
+ $wgHooks['ArticleSaveComplete'][] = $this->mConverter;
}
+
- /*
- get preferred language variants. eventually this will check the
- user's preference setting as well, once the language option in
- the setting pages is finalized.
- */
- function getPreferredVariant() {
- global $wgUser;
-
- if($this->mZhLanguageCode)
- return $this->mZhLanguageCode;
-
- // get language variant preference for logged in users
- if($wgUser->getID()!=0) {
- $this->mZhLanguageCode = $wgUser->getOption('variant');
- }
- else {
- // see if some zh- variant is set in the http header,
- $this->mZhLanguageCode="zh-cn";
- $header = str_replace( '_', '-', strtolower($_SERVER["HTTP_ACCEPT_LANGUAGE"]));
- $zh = strstr($header, 'zh-');
- if($zh) {
- $this->mZhLanguageCode = substr($zh,0,5);
- }
- }
- return $this->mZhLanguageCode;
+ # this should give much better diff info
+ function segmentForDiff( $text ) {
+ return preg_replace(
+ "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
+ "' ' .\"$1\"", $text);
}
-
-
- /* the Simplified/Traditional conversion stuff */
- function zh2tw($text) {
- global $zh2TW;
- return strtr($text, $zh2TW);
+ function unsegmentForDiff( $text ) {
+ return preg_replace(
+ "/ ([\\xc0-\\xff][\\x80-\\xbf]*)/e",
+ "\"$1\"", $text);
}
- function zh2cn($text) {
- global $zh2CN;
- return strtr($text, $zh2CN);
- }
+ // word segmentation
+ function stripForSearch( $string ) {
+ $fname="LanguageZh::stripForSearch";
+ wfProfileIn( $fname );
- function zh2sg($text) {
- global $zh2SG, $zh2CN;
- return strtr(strtr($text, $zh2CN), $zh2SG);
- }
+ // eventually this should be a word segmentation
+ // for now just treat each character as a word
+ $t = preg_replace(
+ "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
+ "' ' .\"$1\"", $string);
- function zh2hk($text) {
- global $zh2HK, $zh2TW;
- return strtr(strtr($text, $zh2TW), $zh2HK);
- }
-
- function autoConvert($text, $toVariant=false) {
- if(!$toVariant)
- $toVariant = $this->getPreferredVariant();
- $fname="zhconvert";
- wfProfileIn( $fname );
- $t = $text;
- switch($toVariant) {
- case 'zh-cn':
- $t = $this->zh2cn($text);
- break;
- case 'zh-tw':
- $t = $this->zh2tw($text);
- break;
- case 'zh-sg':
- $t = $this->zh2sg($text);
- break;
- case 'zh-hk':
- $t = $this->zh2hk($text);
- break;
- }
+ //always convert to zh-cn before indexing. it should be
+ //better to use zh-cn for search, since conversion from
+ //Traditional to Simplified is less ambiguous than the
+ //other way around
+
+ $t = $this->mConverter->autoConvert($t, 'zh-cn');
+ $t = LanguageUtf8::stripForSearch( $t );
wfProfileOut( $fname );
return $t;
+
}
- function getVariants() {
- return array("zh-cn", "zh-tw", "zh-sg", "zh-hk");
+ function convertForSearchResult( $termsArray ) {
+ $terms = implode( '|', $termsArray );
+ $terms = implode( '|', $this->mConverter->autoConvertToAllVariants( $terms ) );
+ $ret = array_unique( explode('|', $terms) );
+ return $ret;
}
+
}
?>