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 use MediaWiki\MediaWikiServices;
  27
  28 /**
  29  * Database independant search index updater
  30  *
  31  * @ingroup Search
  32  */
  33 class SearchUpdate implements DeferrableUpdate {
  34         /** @var int Page id being updated */
  35         private $id = 0;
  36
  37         /** @var Title Title we're updating */
  38         private $title;
  39
  40         /** @var Content|bool Content of the page (not text) */
  41         private $content;
  42
  43         /** @var WikiPage **/
  44         private $page;
  45
  46         /**
  47          * @param int $id Page id to update
  48          * @param Title|string $title Title of page to update
  49          * @param Content|string|bool $c Content of the page to update. Default: false.
  50          *  If a Content object, text will be gotten from it. String is for back-compat.
  51          *  Passing false tells the backend to just update the title, not the content
  52          */
  53         public function __construct( $id, $title, $c = false ) {
  54                 if ( is_string( $title ) ) {
  55                         $nt = Title::newFromText( $title );
  56                 } else {
  57                         $nt = $title;
  58                 }
  59
  60                 if ( $nt ) {
  61                         $this->id = $id;
  62                         // is_string() check is back-compat for ApprovedRevs
  63                         if ( is_string( $c ) ) {
  64                                 $this->content = new TextContent( $c );
  65                         } else {
  66                                 $this->content = $c ?: false;
  67                         }
  68                         $this->title = $nt;
  69                 } else {
  70                         wfDebug( "SearchUpdate object created with invalid title '$title'\n" );
  71                 }
  72         }
  73
  74         /**
  75          * Perform actual update for the entry
  76          */
  77         public function doUpdate() {
  78                 $config = MediaWikiServices::getInstance()->getSearchEngineConfig();
  79
  80                 if ( $config->getConfig()->get( 'DisableSearchUpdate' ) || !$this->id ) {
  81                         return;
  82                 }
  83
  84                 $seFactory = MediaWikiServices::getInstance()->getSearchEngineFactory();
  85                 foreach ( $config->getSearchTypes() as $type ) {
  86                         $search = $seFactory->create( $type );
  87                         if ( !$search->supports( 'search-update' ) ) {
  88                                 continue;
  89                         }
  90
  91                         $normalTitle = $this->getNormalizedTitle( $search );
  92
  93                         if ( $this->getLatestPage() === null ) {
  94                                 $search->delete( $this->id, $normalTitle );
  95                                 continue;
  96                         } elseif ( $this->content === false ) {
  97                                 $search->updateTitle( $this->id, $normalTitle );
  98                                 continue;
  99                         }
 100
 101                         $text = $search->getTextFromContent( $this->title, $this->content );
 102                         if ( !$search->textAlreadyUpdatedForIndex() ) {
 103                                 $text = $this->updateText( $text, $search );
 104                         }
 105
 106                         # Perform the actual update
 107                         $search->update( $this->id, $normalTitle, $search->normalizeText( $text ) );
 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          * @param SearchEngine|null $se Search engine
 117          * @return string
 118          */
 119         public function updateText( $text, SearchEngine $se = null ) {
 120                 # Language-specific strip/conversion
 121                 $text = MediaWikiServices::getInstance()->getContentLanguage()->normalizeForSearch( $text );
 122                 $se = $se ?: MediaWikiServices::getInstance()->newSearchEngine();
 123                 $lc = $se->legalSearchChars() . '&#;';
 124
 125                 # Strip HTML markup
 126                 $text = preg_replace( "/<\\/?\\s*[A-Za-z][^>]*?>/",
 127                         ' ', MediaWikiServices::getInstance()->getContentLanguage()->lc( " " . $text . " " ) );
 128                 $text = preg_replace( "/(^|\\n)==\\s*([^\\n]+)\\s*==(\\s)/sD",
 129                         "\\1\\2 \\2 \\2\\3", $text ); # Emphasize headings
 130
 131                 # Strip external URLs
 132                 $uc = "A-Za-z0-9_\\/:.,~%\\-+&;#?!=()@\\x80-\\xFF";
 133                 $protos = "http|https|ftp|mailto|news|gopher";
 134                 $pat = "/(^|[^\\[])({$protos}):[{$uc}]+([^{$uc}]|$)/";
 135                 $text = preg_replace( $pat, "\\1 \\3", $text );
 136
 137                 $p1 = "/([^\\[])\\[({$protos}):[{$uc}]+]/";
 138                 $p2 = "/([^\\[])\\[({$protos}):[{$uc}]+\\s+([^\\]]+)]/";
 139                 $text = preg_replace( $p1, "\\1 ", $text );
 140                 $text = preg_replace( $p2, "\\1 \\3 ", $text );
 141
 142                 # Internal image links
 143                 $pat2 = "/\\[\\[image:([{$uc}]+)\\.(gif|png|jpg|jpeg)([^{$uc}])/i";
 144                 $text = preg_replace( $pat2, " \\1 \\3", $text );
 145
 146                 $text = preg_replace( "/([^{$lc}])([{$lc}]+)]]([a-z]+)/",
 147                         "\\1\\2 \\2\\3", $text ); # Handle [[game]]s
 148
 149                 # Strip all remaining non-search characters
 150                 $text = preg_replace( "/[^{$lc}]+/", " ", $text );
 151
 152                 /**
 153                  * Handle 's, s'
 154                  *
 155                  *   $text = preg_replace( "/([{$lc}]+)'s /", "\\1 \\1's ", $text );
 156                  *   $text = preg_replace( "/([{$lc}]+)s' /", "\\1s ", $text );
 157                  *
 158                  * These tail-anchored regexps are insanely slow. The worst case comes
 159                  * when Japanese or Chinese text (ie, no word spacing) is written on
 160                  * a wiki configured for Western UTF-8 mode. The Unicode characters are
 161                  * expanded to hex codes and the "words" are very long paragraph-length
 162                  * monstrosities. On a large page the above regexps may take over 20
 163                  * seconds *each* on a 1GHz-level processor.
 164                  *
 165                  * Following are reversed versions which are consistently fast
 166                  * (about 3 milliseconds on 1GHz-level processor).
 167                  */
 168                 $text = strrev( preg_replace( "/ s'([{$lc}]+)/", " s'\\1 \\1", strrev( $text ) ) );
 169                 $text = strrev( preg_replace( "/ 's([{$lc}]+)/", " s\\1", strrev( $text ) ) );
 170
 171                 # Strip wiki '' and '''
 172                 $text = preg_replace( "/''[']*/", " ", $text );
 173
 174                 return $text;
 175         }
 176
 177         /**
 178          * Get WikiPage for the SearchUpdate $id using WikiPage::READ_LATEST
 179          * and ensure using the same WikiPage object if there are multiple
 180          * SearchEngine types.
 181          *
 182          * Returns null if a page has been deleted or is not found.
 183          *
 184          * @return WikiPage|null
 185          */
 186         private function getLatestPage() {
 187                 if ( !isset( $this->page ) ) {
 188                         $this->page = WikiPage::newFromID( $this->id, WikiPage::READ_LATEST );
 189                 }
 190
 191                 return $this->page;
 192         }
 193
 194         /**
 195          * Get a normalized string representation of a title suitable for
 196          * including in a search index
 197          *
 198          * @param SearchEngine $search
 199          * @return string A stripped-down title string ready for the search index
 200          */
 201         private function getNormalizedTitle( SearchEngine $search ) {
 202                 $ns = $this->title->getNamespace();
 203                 $title = $this->title->getText();
 204
 205                 $lc = $search->legalSearchChars() . '&#;';
 206                 $t = MediaWikiServices::getInstance()->getContentLanguage()->normalizeForSearch( $title );
 207                 $t = preg_replace( "/[^{$lc}]+/", ' ', $t );
 208                 $t = MediaWikiServices::getInstance()->getContentLanguage()->lc( $t );
 209
 210                 # Handle 's, s'
 211                 $t = preg_replace( "/([{$lc}]+)'s( |$)/", "\\1 \\1's ", $t );
 212                 $t = preg_replace( "/([{$lc}]+)s'( |$)/", "\\1s ", $t );
 213
 214                 $t = preg_replace( "/\\s+/", ' ', $t );
 215
 216                 if ( $ns == NS_FILE ) {
 217                         $t = preg_replace( "/ (png|gif|jpg|jpeg|ogg)$/", "", $t );
 218                 }
 219
 220                 return $search->normalizeText( trim( $t ) );
 221         }
 222 }