includes/deferred/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         /** @var int Page id being updated */
  33         private $id = 0;
  34
  35         /** @var Title Title we're updating */
  36         private $title;
  37
  38         /** @var Content|bool Content of the page (not text) */
  39         private $content;
  40
  41         /**
  42          * Constructor
  43          *
  44          * @param int $id Page id to update
  45          * @param Title|string $title Title of page to update
  46          * @param Content|string|bool $c Content of the page to update. Default: false.
  47          *  If a Content object, text will be gotten from it. String is for back-compat.
  48          *  Passing false tells the backend to just update the title, not the content
  49          */
  50         public function __construct( $id, $title, $c = false ) {
  51                 if ( is_string( $title ) ) {
  52                         $nt = Title::newFromText( $title );
  53                 } else {
  54                         $nt = $title;
  55                 }
  56
  57                 if ( $nt ) {
  58                         $this->id = $id;
  59                         // is_string() check is back-compat for ApprovedRevs
  60                         if ( is_string( $c ) ) {
  61                                 $this->content = new TextContent( $c );
  62                         } else {
  63                                 $this->content = $c ?: false;
  64                         }
  65                         $this->title = $nt;
  66                 } else {
  67                         wfDebug( "SearchUpdate object created with invalid title '$title'\n" );
  68                 }
  69         }
  70
  71         /**
  72          * Perform actual update for the entry
  73          */
  74         public function doUpdate() {
  75                 global $wgDisableSearchUpdate;
  76
  77                 if ( $wgDisableSearchUpdate || !$this->id ) {
  78                         return;
  79                 }
  80
  81                 $page = WikiPage::newFromID( $this->id, WikiPage::READ_LATEST );
  82
  83                 foreach ( SearchEngine::getSearchTypes() as $type ) {
  84                         $search = SearchEngine::create( $type );
  85                         $indexTitle = $this->indexTitle( $search );
  86                         if ( !$search->supports( 'search-update' ) ) {
  87                                 continue;
  88                         }
  89
  90                         $normalTitle = $search->normalizeText( $indexTitle );
  91
  92                         if ( $page === null ) {
  93                                 $search->delete( $this->id, $normalTitle );
  94                                 continue;
  95                         } elseif ( $this->content === false ) {
  96                                 $search->updateTitle( $this->id, $normalTitle );
  97                                 continue;
  98                         }
  99
 100                         $text = $search->getTextFromContent( $this->title, $this->content );
 101                         if ( !$search->textAlreadyUpdatedForIndex() ) {
 102                                 $text = self::updateText( $text );
 103                         }
 104
 105                         # Perform the actual update
 106                         $search->update( $this->id, $normalTitle, $search->normalizeText( $text ) );
 107                 }
 108
 109         }
 110
 111         /**
 112          * Clean text for indexing. Only really suitable for indexing in databases.
 113          * If you're using a real search engine, you'll probably want to override
 114          * this behavior and do something nicer with the original wikitext.
 115          * @param string $text
 116          * @return string
 117          */
 118         public static function updateText( $text ) {
 119                 global $wgContLang;
 120
 121                 # Language-specific strip/conversion
 122                 $text = $wgContLang->normalizeForSearch( $text );
 123                 $lc = SearchEngine::legalSearchChars() . '&#;';
 124
 125                 $text = preg_replace( "/<\\/?\\s*[A-Za-z][^>]*?>/",
 126                         ' ', $wgContLang->lc( " " . $text . " " ) ); # Strip HTML markup
 127                 $text = preg_replace( "/(^|\\n)==\\s*([^\\n]+)\\s*==(\\s)/sD",
 128                         "\\1\\2 \\2 \\2\\3", $text ); # Emphasize headings
 129
 130                 # Strip external URLs
 131                 $uc = "A-Za-z0-9_\\/:.,~%\\-+&;#?!=()@\\x80-\\xFF";
 132                 $protos = "http|https|ftp|mailto|news|gopher";
 133                 $pat = "/(^|[^\\[])({$protos}):[{$uc}]+([^{$uc}]|$)/";
 134                 $text = preg_replace( $pat, "\\1 \\3", $text );
 135
 136                 $p1 = "/([^\\[])\\[({$protos}):[{$uc}]+]/";
 137                 $p2 = "/([^\\[])\\[({$protos}):[{$uc}]+\\s+([^\\]]+)]/";
 138                 $text = preg_replace( $p1, "\\1 ", $text );
 139                 $text = preg_replace( $p2, "\\1 \\3 ", $text );
 140
 141                 # Internal image links
 142                 $pat2 = "/\\[\\[image:([{$uc}]+)\\.(gif|png|jpg|jpeg)([^{$uc}])/i";
 143                 $text = preg_replace( $pat2, " \\1 \\3", $text );
 144
 145                 $text = preg_replace( "/([^{$lc}])([{$lc}]+)]]([a-z]+)/",
 146                         "\\1\\2 \\2\\3", $text ); # Handle [[game]]s
 147
 148                 # Strip all remaining non-search characters
 149                 $text = preg_replace( "/[^{$lc}]+/", " ", $text );
 150
 151                 # Handle 's, s'
 152                 #
 153                 #   $text = preg_replace( "/([{$lc}]+)'s /", "\\1 \\1's ", $text );
 154                 #   $text = preg_replace( "/([{$lc}]+)s' /", "\\1s ", $text );
 155                 #
 156                 # These tail-anchored regexps are insanely slow. The worst case comes
 157                 # when Japanese or Chinese text (ie, no word spacing) is written on
 158                 # a wiki configured for Western UTF-8 mode. The Unicode characters are
 159                 # expanded to hex codes and the "words" are very long paragraph-length
 160                 # monstrosities. On a large page the above regexps may take over 20
 161                 # seconds *each* on a 1GHz-level processor.
 162                 #
 163                 # Following are reversed versions which are consistently fast
 164                 # (about 3 milliseconds on 1GHz-level processor).
 165                 #
 166                 $text = strrev( preg_replace( "/ s'([{$lc}]+)/", " s'\\1 \\1", strrev( $text ) ) );
 167                 $text = strrev( preg_replace( "/ 's([{$lc}]+)/", " s\\1", strrev( $text ) ) );
 168
 169                 # Strip wiki '' and '''
 170                 $text = preg_replace( "/''[']*/", " ", $text );
 171
 172                 return $text;
 173         }
 174
 175         /**
 176          * Get a string representation of a title suitable for
 177          * including in a search index
 178          *
 179          * @param SearchEngine $search
 180          * @return string A stripped-down title string ready for the search index
 181          */
 182         private function indexTitle( SearchEngine $search ) {
 183                 global $wgContLang;
 184
 185                 $ns = $this->title->getNamespace();
 186                 $title = $this->title->getText();
 187
 188                 $lc = $search->legalSearchChars() . '&#;';
 189                 $t = $wgContLang->normalizeForSearch( $title );
 190                 $t = preg_replace( "/[^{$lc}]+/", ' ', $t );
 191                 $t = $wgContLang->lc( $t );
 192
 193                 # Handle 's, s'
 194                 $t = preg_replace( "/([{$lc}]+)'s( |$)/", "\\1 \\1's ", $t );
 195                 $t = preg_replace( "/([{$lc}]+)s'( |$)/", "\\1s ", $t );
 196
 197                 $t = preg_replace( "/\\s+/", ' ', $t );
 198
 199                 if ( $ns == NS_FILE ) {
 200                         $t = preg_replace( "/ (png|gif|jpg|jpeg|ogg)$/", "", $t );
 201                 }
 202                 return trim( $t );
 203         }
 204 }