includes/search/SearchUpdate.php

   1 <?php
   2 /**
   3  * Search index updater
   4  *
   5  * See deferred.txt
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @file
  23  * @ingroup Search
  24  */
  25
  26 /**
  27  * Database independant search index updater
  28  *
  29  * @ingroup Search
  30  */
  31 class SearchUpdate implements DeferrableUpdate {
  32
  33         private $mId = 0, $mNamespace, $mTitle, $mText;
  34         private $mTitleWords;
  35
  36         /**
  37          * Constructor
  38          *
  39          * @param int $id Page id to update
  40          * @param Title|string $title Title of page to update
  41          * @param Content|string|false $content Content of the page to update.
  42          *  If a Content object, text will be gotten from it. String is for back-compat.
  43          *  Passing false tells the backend to just update the title, not the content
  44          */
  45         public function __construct( $id, $title, $content = false ) {
  46                 if ( is_string( $title ) ) {
  47                         $nt = Title::newFromText( $title );
  48                 } else {
  49                         $nt = $title;
  50                 }
  51
  52                 if ( $nt ) {
  53                         $this->mId = $id;
  54                         // @todo This isn't ideal, we'd really like to have content-specific
  55                         // handling here. See similar content in SearchEngine::initText().
  56                         if( is_string( $content ) ) {
  57                                 // b/c for ApprovedRevs
  58                                 $this->mText = $content;
  59                         } else {
  60                                 $this->mText = $content ? $content->getTextForSearchIndex() : false;
  61                         }
  62
  63                         $this->mNamespace = $nt->getNamespace();
  64                         $this->mTitle = $nt->getText(); # Discard namespace
  65
  66                         $this->mTitleWords = $this->mTextWords = array();
  67                 } else {
  68                         wfDebug( "SearchUpdate object created with invalid title '$title'\n" );
  69                 }
  70         }
  71
  72         public function doUpdate() {
  73                 global $wgDisableSearchUpdate;
  74
  75                 if ( $wgDisableSearchUpdate || !$this->mId ) {
  76                         return;
  77                 }
  78
  79                 wfProfileIn( __METHOD__ );
  80
  81                 $search = SearchEngine::create();
  82                 $normalTitle = $search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) );
  83
  84                 if ( WikiPage::newFromId( $this->mId ) === null ) {
  85                         $search->delete( $this->mId, $normalTitle );
  86                         wfProfileOut( __METHOD__ );
  87                         return;
  88                 } elseif ( $this->mText === false ) {
  89                         $search->updateTitle( $this->mId, $normalTitle );
  90                         wfProfileOut( __METHOD__ );
  91                         return;
  92                 }
  93
  94                 $text = self::updateText( $this->mText );
  95
  96                 wfRunHooks( 'SearchUpdate', array( $this->mId, $this->mNamespace, $this->mTitle, &$text ) );
  97
  98                 # Perform the actual update
  99                 $search->update( $this->mId, $normalTitle, $search->normalizeText( $text ) );
 100
 101                 wfProfileOut( __METHOD__ );
 102         }
 103
 104         public static function updateText( $text ) {
 105                 global $wgContLang;
 106
 107                 # Language-specific strip/conversion
 108                 $text = $wgContLang->normalizeForSearch( $text );
 109                 $lc = SearchEngine::legalSearchChars() . '&#;';
 110
 111                 wfProfileIn( __METHOD__ . '-regexps' );
 112                 $text = preg_replace( "/<\\/?\\s*[A-Za-z][^>]*?>/",
 113                         ' ', $wgContLang->lc( " " . $text . " " ) ); # Strip HTML markup
 114                 $text = preg_replace( "/(^|\\n)==\\s*([^\\n]+)\\s*==(\\s)/sD",
 115                         "\\1\\2 \\2 \\2\\3", $text ); # Emphasize headings
 116
 117                 # Strip external URLs
 118                 $uc = "A-Za-z0-9_\\/:.,~%\\-+&;#?!=()@\\x80-\\xFF";
 119                 $protos = "http|https|ftp|mailto|news|gopher";
 120                 $pat = "/(^|[^\\[])({$protos}):[{$uc}]+([^{$uc}]|$)/";
 121                 $text = preg_replace( $pat, "\\1 \\3", $text );
 122
 123                 $p1 = "/([^\\[])\\[({$protos}):[{$uc}]+]/";
 124                 $p2 = "/([^\\[])\\[({$protos}):[{$uc}]+\\s+([^\\]]+)]/";
 125                 $text = preg_replace( $p1, "\\1 ", $text );
 126                 $text = preg_replace( $p2, "\\1 \\3 ", $text );
 127
 128                 # Internal image links
 129                 $pat2 = "/\\[\\[image:([{$uc}]+)\\.(gif|png|jpg|jpeg)([^{$uc}])/i";
 130                 $text = preg_replace( $pat2, " \\1 \\3", $text );
 131
 132                 $text = preg_replace( "/([^{$lc}])([{$lc}]+)]]([a-z]+)/",
 133                         "\\1\\2 \\2\\3", $text ); # Handle [[game]]s
 134
 135                 # Strip all remaining non-search characters
 136                 $text = preg_replace( "/[^{$lc}]+/", " ", $text );
 137
 138                 # Handle 's, s'
 139                 #
 140                 #   $text = preg_replace( "/([{$lc}]+)'s /", "\\1 \\1's ", $text );
 141                 #   $text = preg_replace( "/([{$lc}]+)s' /", "\\1s ", $text );
 142                 #
 143                 # These tail-anchored regexps are insanely slow. The worst case comes
 144                 # when Japanese or Chinese text (ie, no word spacing) is written on
 145                 # a wiki configured for Western UTF-8 mode. The Unicode characters are
 146                 # expanded to hex codes and the "words" are very long paragraph-length
 147                 # monstrosities. On a large page the above regexps may take over 20
 148                 # seconds *each* on a 1GHz-level processor.
 149                 #
 150                 # Following are reversed versions which are consistently fast
 151                 # (about 3 milliseconds on 1GHz-level processor).
 152                 #
 153                 $text = strrev( preg_replace( "/ s'([{$lc}]+)/", " s'\\1 \\1", strrev( $text ) ) );
 154                 $text = strrev( preg_replace( "/ 's([{$lc}]+)/", " s\\1", strrev( $text ) ) );
 155
 156                 # Strip wiki '' and '''
 157                 $text = preg_replace( "/''[']*/", " ", $text );
 158                 wfProfileOut( __METHOD__ . '-regexps' );
 159                 return $text;
 160         }
 161 }