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          * @return array Array of strings
  38          */
  39         public static function titleSearch( $search, $limit, $namespaces = array() ) {
  40                 $prefixSearch = new StringPrefixSearch;
  41                 return $prefixSearch->search( $search, $limit, $namespaces );
  42         }
  43
  44         /**
  45          * Do a prefix search of titles and return a list of matching page names.
  46          *
  47          * @param string $search
  48          * @param int $limit
  49          * @param array $namespaces Used if query is not explicitly prefixed
  50          * @return array Array of strings or Title objects
  51          */
  52         public function search( $search, $limit, $namespaces = array() ) {
  53                 $search = trim( $search );
  54                 if ( $search == '' ) {
  55                         return array(); // Return empty result
  56                 }
  57                 $namespaces = $this->validateNamespaces( $namespaces );
  58
  59                 // Find a Title which is not an interwiki and is in NS_MAIN
  60                 $title = Title::newFromText( $search );
  61                 if ( $title && !$title->isExternal() ) {
  62                         $ns = array( $title->getNamespace() );
  63                         $search = $title->getText();
  64                         if ( $ns[0] == NS_MAIN ) {
  65                                 $ns = $namespaces; // no explicit prefix, use default namespaces
  66                                 wfRunHooks( 'PrefixSearchExtractNamespace', array( &$ns, &$search ) );
  67                         }
  68                         return $this->searchBackend( $ns, $search, $limit );
  69                 }
  70
  71                 // Is this a namespace prefix?
  72                 $title = Title::newFromText( $search . 'Dummy' );
  73                 if ( $title && $title->getText() == 'Dummy'
  74                         && $title->getNamespace() != NS_MAIN
  75                         && !$title->isExternal() )
  76                 {
  77                         $namespaces = array( $title->getNamespace() );
  78                         $search = '';
  79                 } else {
  80                         wfRunHooks( 'PrefixSearchExtractNamespace', array( &$namespaces, &$search ) );
  81                 }
  82
  83                 return $this->searchBackend( $namespaces, $search, $limit );
  84         }
  85
  86         /**
  87          * Do a prefix search for all possible variants of the prefix
  88          * @param string $search
  89          * @param int $limit
  90          * @param array $namespaces
  91          *
  92          * @return array
  93          */
  94         public function searchWithVariants( $search, $limit, array $namespaces ) {
  95                 wfProfileIn( __METHOD__ );
  96                 $searches = $this->search( $search, $limit, $namespaces );
  97
  98                 // if the content language has variants, try to retrieve fallback results
  99                 $fallbackLimit = $limit - count( $searches );
 100                 if ( $fallbackLimit > 0 ) {
 101                         global $wgContLang;
 102
 103                         $fallbackSearches = $wgContLang->autoConvertToAllVariants( $search );
 104                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), array( $search ) );
 105
 106                         foreach ( $fallbackSearches as $fbs ) {
 107                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
 108                                 $searches = array_merge( $searches, $fallbackSearchResult );
 109                                 $fallbackLimit -= count( $fallbackSearchResult );
 110
 111                                 if ( $fallbackLimit == 0 ) {
 112                                         break;
 113                                 }
 114                         }
 115                 }
 116                 wfProfileOut( __METHOD__ );
 117                 return $searches;
 118         }
 119
 120         /**
 121          * When implemented in a descendant class, receives an array of Title objects and returns
 122          * either an unmodified array or an array of strings corresponding to titles passed to it.
 123          *
 124          * @param array $titles
 125          * @return array
 126          */
 127         abstract protected function titles( array $titles );
 128
 129         /**
 130          * When implemented in a descendant class, receives an array of titles as strings and returns
 131          * either an unmodified array or an array of Title objects corresponding to strings received.
 132          *
 133          * @param array $strings
 134          *
 135          * @return array
 136          */
 137         abstract protected function strings( array $strings );
 138
 139         /**
 140          * Do a prefix search of titles and return a list of matching page names.
 141          * @param array $namespaces
 142          * @param string $search
 143          * @param int $limit
 144          * @return array Array of strings
 145          */
 146         protected function searchBackend( $namespaces, $search, $limit ) {
 147                 if ( count( $namespaces ) == 1 ) {
 148                         $ns = $namespaces[0];
 149                         if ( $ns == NS_MEDIA ) {
 150                                 $namespaces = array( NS_FILE );
 151                         } elseif ( $ns == NS_SPECIAL ) {
 152                                 return $this->titles( $this->specialSearch( $search, $limit ) );
 153                         }
 154                 }
 155                 $srchres = array();
 156                 if ( wfRunHooks( 'PrefixSearchBackend', array( $namespaces, $search, $limit, &$srchres ) ) ) {
 157                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit ) );
 158                 } else {
 159                         // Default search backend does proper prefix searching, but custom backends
 160                         // may sort based on other algorythms that may cause the exact title match
 161                         // to not be in the results or be lower down the list.
 162
 163                         // Pick namespace (based on PrefixSearch::defaultSearchBackend)
 164                         $ns = in_array( NS_MAIN, $namespaces ) ? NS_MAIN : $namespaces[0];
 165                         $t = Title::newFromText( $search, $ns );
 166                         if ( $t ) {
 167                                 // If text is a valid title and is in the search results
 168                                 $string = $t->getPrefixedText();
 169                                 $key = array_search( $string, $srchres );
 170                                 if ( $key !== false ) {
 171                                         // Move it to the front
 172                                         $cut = array_splice( $srchres, $key, 1 );
 173                                         array_unshift( $srchres, $cut[0] );
 174                                 } elseif ( $t->exists() ) {
 175                                         // Add it in front
 176                                         array_unshift( $srchres, $string );
 177
 178                                         if ( count( $srchres ) > $limit ) {
 179                                                 array_pop( $srchres );
 180                                         }
 181                                 }
 182                         }
 183                 }
 184
 185                 return $this->strings( $srchres );
 186         }
 187
 188         /**
 189          * Prefix search special-case for Special: namespace.
 190          *
 191          * @param string $search Term
 192          * @param int $limit Max number of items to return
 193          * @return array
 194          */
 195         protected function specialSearch( $search, $limit ) {
 196                 global $wgContLang;
 197
 198                 $searchParts = explode( '/', $search, 2 );
 199                 $searchKey = $searchParts[0];
 200                 $subpageSearch = isset( $searchParts[1] ) ? $searchParts[1] : null;
 201
 202                 // Handle subpage search separately.
 203                 if ( $subpageSearch !== null ) {
 204                         // Try matching the full search string as a page name
 205                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 206                         if ( !$specialTitle ) {
 207                                 return array();
 208                         }
 209                         $special = SpecialPageFactory::getPage( $specialTitle->getText() );
 210                         if ( $special ) {
 211                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit );
 212                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 213                                         return $specialTitle->getSubpage( $sub );
 214                                 }, $subpages );
 215                         } else {
 216                                 return array();
 217                         }
 218                 }
 219
 220                 # normalize searchKey, so aliases with spaces can be found - bug 25675
 221                 $searchKey = str_replace( ' ', '_', $searchKey );
 222                 $searchKey = $wgContLang->caseFold( $searchKey );
 223
 224                 // Unlike SpecialPage itself, we want the canonical forms of both
 225                 // canonical and alias title forms...
 226                 $keys = array();
 227                 foreach ( SpecialPageFactory::getNames() as $page ) {
 228                         $keys[$wgContLang->caseFold( $page )] = $page;
 229                 }
 230
 231                 foreach ( $wgContLang->getSpecialPageAliases() as $page => $aliases ) {
 232                         if ( !in_array( $page, SpecialPageFactory::getNames() ) ) {# bug 20885
 233                                 continue;
 234                         }
 235
 236                         foreach ( $aliases as $alias ) {
 237                                 $keys[$wgContLang->caseFold( $alias )] = $alias;
 238                         }
 239                 }
 240                 ksort( $keys );
 241
 242                 $srchres = array();
 243                 foreach ( $keys as $pageKey => $page ) {
 244                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 245                                 // bug 27671: Don't use SpecialPage::getTitleFor() here because it
 246                                 // localizes its input leading to searches for e.g. Special:All
 247                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 248                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 249                                 $srchres[] = Title::makeTitleSafe( NS_SPECIAL, $page );
 250                         }
 251
 252                         if ( count( $srchres ) >= $limit ) {
 253                                 break;
 254                         }
 255                 }
 256
 257                 return $srchres;
 258         }
 259
 260         /**
 261          * Unless overridden by PrefixSearchBackend hook...
 262          * This is case-sensitive (First character may
 263          * be automatically capitalized by Title::secureAndSpit()
 264          * later on depending on $wgCapitalLinks)
 265          *
 266          * @param array $namespaces Namespaces to search in
 267          * @param string $search Term
 268          * @param int $limit Max number of items to return
 269          * @return array Array of Title objects
 270          */
 271         protected function defaultSearchBackend( $namespaces, $search, $limit ) {
 272                 $ns = array_shift( $namespaces ); // support only one namespace
 273                 if ( in_array( NS_MAIN, $namespaces ) ) {
 274                         $ns = NS_MAIN; // if searching on many always default to main
 275                 }
 276
 277                 $t = Title::newFromText( $search, $ns );
 278                 $prefix = $t ? $t->getDBkey() : '';
 279                 $dbr = wfGetDB( DB_SLAVE );
 280                 $res = $dbr->select( 'page',
 281                         array( 'page_id', 'page_namespace', 'page_title' ),
 282                         array(
 283                                 'page_namespace' => $ns,
 284                                 'page_title ' . $dbr->buildLike( $prefix, $dbr->anyString() )
 285                         ),
 286                         __METHOD__,
 287                         array( 'LIMIT' => $limit, 'ORDER BY' => 'page_title' )
 288                 );
 289                 $srchres = array();
 290                 foreach ( $res as $row ) {
 291                         $srchres[] = Title::newFromRow( $row );
 292                 }
 293                 return $srchres;
 294         }
 295
 296         /**
 297          * Validate an array of numerical namespace indexes
 298          *
 299          * @param array $namespaces
 300          * @return array (default: contains only NS_MAIN)
 301          */
 302         protected function validateNamespaces( $namespaces ) {
 303                 global $wgContLang;
 304
 305                 // We will look at each given namespace against wgContLang namespaces
 306                 $validNamespaces = $wgContLang->getNamespaces();
 307                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 308                         $valid = array();
 309                         foreach ( $namespaces as $ns ) {
 310                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 311                                         $valid[] = $ns;
 312                                 }
 313                         }
 314                         if ( count( $valid ) > 0 ) {
 315                                 return $valid;
 316                         }
 317                 }
 318
 319                 return array( NS_MAIN );
 320         }
 321 }
 322
 323 /**
 324  * Performs prefix search, returning Title objects
 325  * @ingroup Search
 326  */
 327 class TitlePrefixSearch extends PrefixSearch {
 328
 329         protected function titles( array $titles ) {
 330                 return $titles;
 331         }
 332
 333         protected function strings( array $strings ) {
 334                 $titles = array_map( 'Title::newFromText', $strings );
 335                 $lb = new LinkBatch( $titles );
 336                 $lb->setCaller( __METHOD__ );
 337                 $lb->execute();
 338                 return $titles;
 339         }
 340 }
 341
 342 /**
 343  * Performs prefix search, returning strings
 344  * @ingroup Search
 345  */
 346 class StringPrefixSearch extends PrefixSearch {
 347
 348         protected function titles( array $titles ) {
 349                 return array_map( function ( Title $t ) {
 350                         return $t->getPrefixedText();
 351                 }, $titles );
 352         }
 353
 354         protected function strings( array $strings ) {
 355                 return $strings;
 356         }
 357 }