1. Add conditions to stripForSearch for LuceneSearch / MWSearch.
[lhc/web/wiklou.git] / languages / classes / LanguageJa.php
1 <?php
2
3 /**
4 * Japanese (日本語)
5 *
6 * @ingroup Language
7 */
8 class LanguageJa extends Language {
9 function stripForSearch( $string ) {
10 # MySQL fulltext index doesn't grok utf-8, so we
11 # need to fold cases and convert to hex
12 $s = $string;
13
14 # not for LuceneSearch, because LSearch
15 # will split the text to words itself
16 if ( $wgSearchType != 'LuceneSearch' ) {
17 # Strip known punctuation ?
18 #$s = preg_replace( '/\xe3\x80[\x80-\xbf]/', '', $s ); # U3000-303f
19
20 # Space strings of like hiragana/katakana/kanji
21 $hiragana = '(?:\xe3(?:\x81[\x80-\xbf]|\x82[\x80-\x9f]))'; # U3040-309f
22 $katakana = '(?:\xe3(?:\x82[\xa0-\xbf]|\x83[\x80-\xbf]))'; # U30a0-30ff
23 $kanji = '(?:\xe3[\x88-\xbf][\x80-\xbf]'
24 . '|[\xe4-\xe8][\x80-\xbf]{2}'
25 . '|\xe9[\x80-\xa5][\x80-\xbf]'
26 . '|\xe9\xa6[\x80-\x99])';
27 # U3200-9999 = \xe3\x88\x80-\xe9\xa6\x99
28 $s = preg_replace( "/({$hiragana}+|{$katakana}+|{$kanji}+)/", ' $1 ', $s );
29 }
30 # Double-width roman characters: ff00-ff5f ~= 0020-007f
31 $s = preg_replace( '/\xef\xbc([\x80-\xbf])/e', 'chr((ord("$1") & 0x3f) + 0x20)', $s );
32 $s = preg_replace( '/\xef\xbd([\x80-\x99])/e', 'chr((ord("$1") & 0x3f) + 0x60)', $s );
33
34 # Do general case folding and UTF-8 armoring
35 return parent::stripForSearch( $s );
36 }
37
38 # Italic is not appropriate for Japanese script
39 # Unfortunately most browsers do not recognise this, and render <em> as italic
40 function emphasize( $text ) {
41 return $text;
42 }
43 }