includes/PrefixSearch.php

   1 <?php
   2 /**
   3  * Prefix search of page names.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Handles searching prefixes of titles and finding any page
  25  * names that match. Used largely by the OpenSearch implementation.
  26  *
  27  * @ingroup Search
  28  */
  29 abstract class PrefixSearch {
  30         /**
  31          * Do a prefix search of titles and return a list of matching page names.
  32          * @deprecated Since 1.23, use TitlePrefixSearch or StringPrefixSearch classes
  33          *
  34          * @param string $search
  35          * @param int $limit
  36          * @param array $namespaces Used if query is not explicitly prefixed
  37          * @param int $offset How many results to offset from the beginning
  38          * @return array Array of strings
  39          */
  40         public static function titleSearch( $search, $limit, $namespaces = array(), $offset = 0 ) {
  41                 $prefixSearch = new StringPrefixSearch;
  42                 return $prefixSearch->search( $search, $limit, $namespaces, $offset );
  43         }
  44
  45         /**
  46          * Do a prefix search of titles and return a list of matching page names.
  47          *
  48          * @param string $search
  49          * @param int $limit
  50          * @param array $namespaces Used if query is not explicitly prefixed
  51          * @param int $offset How many results to offset from the beginning
  52          * @return array Array of strings or Title objects
  53          */
  54         public function search( $search, $limit, $namespaces = array(), $offset = 0 ) {
  55                 $search = trim( $search );
  56                 if ( $search == '' ) {
  57                         return array(); // Return empty result
  58                 }
  59                 $namespaces = $this->validateNamespaces( $namespaces );
  60
  61                 // Find a Title which is not an interwiki and is in NS_MAIN
  62                 $title = Title::newFromText( $search );
  63                 if ( $title && !$title->isExternal() ) {
  64                         $ns = array( $title->getNamespace() );
  65                         $search = $title->getText();
  66                         if ( $ns[0] == NS_MAIN ) {
  67                                 $ns = $namespaces; // no explicit prefix, use default namespaces
  68                                 Hooks::run( 'PrefixSearchExtractNamespace', array( &$ns, &$search ) );
  69                         }
  70                         return $this->searchBackend( $ns, $search, $limit, $offset );
  71                 }
  72
  73                 // Is this a namespace prefix?
  74                 $title = Title::newFromText( $search . 'Dummy' );
  75                 if ( $title && $title->getText() == 'Dummy'
  76                         && $title->getNamespace() != NS_MAIN
  77                         && !$title->isExternal() )
  78                 {
  79                         $namespaces = array( $title->getNamespace() );
  80                         $search = '';
  81                 } else {
  82                         Hooks::run( 'PrefixSearchExtractNamespace', array( &$namespaces, &$search ) );
  83                 }
  84
  85                 return $this->searchBackend( $namespaces, $search, $limit, $offset );
  86         }
  87
  88         /**
  89          * Do a prefix search for all possible variants of the prefix
  90          * @param string $search
  91          * @param int $limit
  92          * @param array $namespaces
  93          * @param int $offset How many results to offset from the beginning
  94          *
  95          * @return array
  96          */
  97         public function searchWithVariants( $search, $limit, array $namespaces, $offset = 0 ) {
  98                 $searches = $this->search( $search, $limit, $namespaces, $offset );
  99
 100                 // if the content language has variants, try to retrieve fallback results
 101                 $fallbackLimit = $limit - count( $searches );
 102                 if ( $fallbackLimit > 0 ) {
 103                         global $wgContLang;
 104
 105                         $fallbackSearches = $wgContLang->autoConvertToAllVariants( $search );
 106                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), array( $search ) );
 107
 108                         foreach ( $fallbackSearches as $fbs ) {
 109                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
 110                                 $searches = array_merge( $searches, $fallbackSearchResult );
 111                                 $fallbackLimit -= count( $fallbackSearchResult );
 112
 113                                 if ( $fallbackLimit == 0 ) {
 114                                         break;
 115                                 }
 116                         }
 117                 }
 118                 return $searches;
 119         }
 120
 121         /**
 122          * When implemented in a descendant class, receives an array of Title objects and returns
 123          * either an unmodified array or an array of strings corresponding to titles passed to it.
 124          *
 125          * @param array $titles
 126          * @return array
 127          */
 128         abstract protected function titles( array $titles );
 129
 130         /**
 131          * When implemented in a descendant class, receives an array of titles as strings and returns
 132          * either an unmodified array or an array of Title objects corresponding to strings received.
 133          *
 134          * @param array $strings
 135          *
 136          * @return array
 137          */
 138         abstract protected function strings( array $strings );
 139
 140         /**
 141          * Do a prefix search of titles and return a list of matching page names.
 142          * @param array $namespaces
 143          * @param string $search
 144          * @param int $limit
 145          * @param int $offset How many results to offset from the beginning
 146          * @return array Array of strings
 147          */
 148         protected function searchBackend( $namespaces, $search, $limit, $offset ) {
 149                 if ( count( $namespaces ) == 1 ) {
 150                         $ns = $namespaces[0];
 151                         if ( $ns == NS_MEDIA ) {
 152                                 $namespaces = array( NS_FILE );
 153                         } elseif ( $ns == NS_SPECIAL ) {
 154                                 return $this->titles( $this->specialSearch( $search, $limit, $offset ) );
 155                         }
 156                 }
 157                 $srchres = array();
 158                 if ( Hooks::run(
 159                         'PrefixSearchBackend',
 160                         array( $namespaces, $search, $limit, &$srchres, $offset )
 161                 ) ) {
 162                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) );
 163                 }
 164                 return $this->strings( $this->handleResultFromHook( $srchres, $namespaces, $search, $limit ) );
 165         }
 166
 167         private function handleResultFromHook( $srchres, $namespaces, $search, $limit ) {
 168                 $rescorer = new SearchExactMatchRescorer();
 169                 return $rescorer->rescore( $search, $namespaces, $srchres, $limit );
 170         }
 171
 172         /**
 173          * Prefix search special-case for Special: namespace.
 174          *
 175          * @param string $search Term
 176          * @param int $limit Max number of items to return
 177          * @param int $offset Number of items to offset
 178          * @return array
 179          */
 180         protected function specialSearch( $search, $limit, $offset ) {
 181                 global $wgContLang;
 182
 183                 $searchParts = explode( '/', $search, 2 );
 184                 $searchKey = $searchParts[0];
 185                 $subpageSearch = isset( $searchParts[1] ) ? $searchParts[1] : null;
 186
 187                 // Handle subpage search separately.
 188                 if ( $subpageSearch !== null ) {
 189                         // Try matching the full search string as a page name
 190                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 191                         if ( !$specialTitle ) {
 192                                 return array();
 193                         }
 194                         $special = SpecialPageFactory::getPage( $specialTitle->getText() );
 195                         if ( $special ) {
 196                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset );
 197                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 198                                         return $specialTitle->getSubpage( $sub );
 199                                 }, $subpages );
 200                         } else {
 201                                 return array();
 202                         }
 203                 }
 204
 205                 # normalize searchKey, so aliases with spaces can be found - bug 25675
 206                 $searchKey = str_replace( ' ', '_', $searchKey );
 207                 $searchKey = $wgContLang->caseFold( $searchKey );
 208
 209                 // Unlike SpecialPage itself, we want the canonical forms of both
 210                 // canonical and alias title forms...
 211                 $keys = array();
 212                 foreach ( SpecialPageFactory::getNames() as $page ) {
 213                         $keys[$wgContLang->caseFold( $page )] = $page;
 214                 }
 215
 216                 foreach ( $wgContLang->getSpecialPageAliases() as $page => $aliases ) {
 217                         if ( !in_array( $page, SpecialPageFactory::getNames() ) ) {# bug 20885
 218                                 continue;
 219                         }
 220
 221                         foreach ( $aliases as $alias ) {
 222                                 $keys[$wgContLang->caseFold( $alias )] = $alias;
 223                         }
 224                 }
 225                 ksort( $keys );
 226
 227                 $srchres = array();
 228                 $skipped = 0;
 229                 foreach ( $keys as $pageKey => $page ) {
 230                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 231                                 // bug 27671: Don't use SpecialPage::getTitleFor() here because it
 232                                 // localizes its input leading to searches for e.g. Special:All
 233                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 234                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 235                                 if ( $offset > 0 && $skipped < $offset ) {
 236                                         $skipped++;
 237                                         continue;
 238                                 }
 239                                 $srchres[] = Title::makeTitleSafe( NS_SPECIAL, $page );
 240                         }
 241
 242                         if ( count( $srchres ) >= $limit ) {
 243                                 break;
 244                         }
 245                 }
 246
 247                 return $srchres;
 248         }
 249
 250         /**
 251          * Unless overridden by PrefixSearchBackend hook...
 252          * This is case-sensitive (First character may
 253          * be automatically capitalized by Title::secureAndSpit()
 254          * later on depending on $wgCapitalLinks)
 255          *
 256          * @param array $namespaces Namespaces to search in
 257          * @param string $search Term
 258          * @param int $limit Max number of items to return
 259          * @param int $offset Number of items to skip
 260          * @return array Array of Title objects
 261          */
 262         protected function defaultSearchBackend( $namespaces, $search, $limit, $offset ) {
 263                 $ns = array_shift( $namespaces ); // support only one namespace
 264                 if ( in_array( NS_MAIN, $namespaces ) ) {
 265                         $ns = NS_MAIN; // if searching on many always default to main
 266                 }
 267
 268                 $t = Title::newFromText( $search, $ns );
 269
 270                 $prefix = $t ? $t->getDBkey() : '';
 271                 $dbr = wfGetDB( DB_SLAVE );
 272                 $res = $dbr->select( 'page',
 273                         array( 'page_id', 'page_namespace', 'page_title' ),
 274                         array(
 275                                 'page_namespace' => $ns,
 276                                 'page_title ' . $dbr->buildLike( $prefix, $dbr->anyString() )
 277                         ),
 278                         __METHOD__,
 279                         array(
 280                                 'LIMIT' => $limit,
 281                                 'ORDER BY' => 'page_title',
 282                                 'OFFSET' => $offset
 283                         )
 284                 );
 285                 $srchres = array();
 286                 foreach ( $res as $row ) {
 287                         $srchres[] = Title::newFromRow( $row );
 288                 }
 289                 return $srchres;
 290         }
 291
 292         /**
 293          * Validate an array of numerical namespace indexes
 294          *
 295          * @param array $namespaces
 296          * @return array (default: contains only NS_MAIN)
 297          */
 298         protected function validateNamespaces( $namespaces ) {
 299                 global $wgContLang;
 300
 301                 // We will look at each given namespace against wgContLang namespaces
 302                 $validNamespaces = $wgContLang->getNamespaces();
 303                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 304                         $valid = array();
 305                         foreach ( $namespaces as $ns ) {
 306                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 307                                         $valid[] = $ns;
 308                                 }
 309                         }
 310                         if ( count( $valid ) > 0 ) {
 311                                 return $valid;
 312                         }
 313                 }
 314
 315                 return array( NS_MAIN );
 316         }
 317 }
 318
 319 /**
 320  * Performs prefix search, returning Title objects
 321  * @ingroup Search
 322  */
 323 class TitlePrefixSearch extends PrefixSearch {
 324
 325         protected function titles( array $titles ) {
 326                 return $titles;
 327         }
 328
 329         protected function strings( array $strings ) {
 330                 $titles = array_map( 'Title::newFromText', $strings );
 331                 $lb = new LinkBatch( $titles );
 332                 $lb->setCaller( __METHOD__ );
 333                 $lb->execute();
 334                 return $titles;
 335         }
 336 }
 337
 338 /**
 339  * Performs prefix search, returning strings
 340  * @ingroup Search
 341  */
 342 class StringPrefixSearch extends PrefixSearch {
 343
 344         protected function titles( array $titles ) {
 345                 return array_map( function ( Title $t ) {
 346                         return $t->getPrefixedText();
 347                 }, $titles );
 348         }
 349
 350         protected function strings( array $strings ) {
 351                 return $strings;
 352         }
 353 }