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                 wfProfileIn( __METHOD__ );
  99                 $searches = $this->search( $search, $limit, $namespaces, $offset );
 100
 101                 // if the content language has variants, try to retrieve fallback results
 102                 $fallbackLimit = $limit - count( $searches );
 103                 if ( $fallbackLimit > 0 ) {
 104                         global $wgContLang;
 105
 106                         $fallbackSearches = $wgContLang->autoConvertToAllVariants( $search );
 107                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), array( $search ) );
 108
 109                         foreach ( $fallbackSearches as $fbs ) {
 110                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
 111                                 $searches = array_merge( $searches, $fallbackSearchResult );
 112                                 $fallbackLimit -= count( $fallbackSearchResult );
 113
 114                                 if ( $fallbackLimit == 0 ) {
 115                                         break;
 116                                 }
 117                         }
 118                 }
 119                 wfProfileOut( __METHOD__ );
 120                 return $searches;
 121         }
 122
 123         /**
 124          * When implemented in a descendant class, receives an array of Title objects and returns
 125          * either an unmodified array or an array of strings corresponding to titles passed to it.
 126          *
 127          * @param array $titles
 128          * @return array
 129          */
 130         abstract protected function titles( array $titles );
 131
 132         /**
 133          * When implemented in a descendant class, receives an array of titles as strings and returns
 134          * either an unmodified array or an array of Title objects corresponding to strings received.
 135          *
 136          * @param array $strings
 137          *
 138          * @return array
 139          */
 140         abstract protected function strings( array $strings );
 141
 142         /**
 143          * Do a prefix search of titles and return a list of matching page names.
 144          * @param array $namespaces
 145          * @param string $search
 146          * @param int $limit
 147          * @param int $offset How many results to offset from the beginning
 148          * @return array Array of strings
 149          */
 150         protected function searchBackend( $namespaces, $search, $limit, $offset ) {
 151                 if ( count( $namespaces ) == 1 ) {
 152                         $ns = $namespaces[0];
 153                         if ( $ns == NS_MEDIA ) {
 154                                 $namespaces = array( NS_FILE );
 155                         } elseif ( $ns == NS_SPECIAL ) {
 156                                 return $this->titles( $this->specialSearch( $search, $limit, $offset ) );
 157                         }
 158                 }
 159                 $srchres = array();
 160                 if ( Hooks::run( 'PrefixSearchBackend', array( $namespaces, $search, $limit, &$srchres, $offset ) ) ) {
 161                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) );
 162                 }
 163                 return $this->strings( $this->handleResultFromHook( $srchres, $namespaces, $search, $limit ) );
 164         }
 165
 166         /**
 167          * Default search backend does proper prefix searching, but custom backends
 168          * may sort based on other algorythms that may cause the exact title match
 169          * to not be in the results or be lower down the list.
 170          * @param array $srchres results from the hook
 171          * @return array munged results from the hook
 172          */
 173         private function handleResultFromHook( $srchres, $namespaces, $search, $limit ) {
 174                 // Pick namespace (based on PrefixSearch::defaultSearchBackend)
 175                 $ns = in_array( NS_MAIN, $namespaces ) ? NS_MAIN : $namespaces[0];
 176                 $t = Title::newFromText( $search, $ns );
 177                 if ( !$t || !$t->exists() ) {
 178                         // No exact match so just return the search results
 179                         return $srchres;
 180                 }
 181                 $string = $t->getPrefixedText();
 182                 $key = array_search( $string, $srchres );
 183                 if ( $key !== false ) {
 184                         // Exact match was in the results so just move it to the front
 185                         return $this->pullFront( $key, $srchres );
 186                 }
 187                 // Exact match not in the search results so check for some redirect handling cases
 188                 if ( $t->isRedirect() ) {
 189                         $target = $this->getRedirectTarget( $t );
 190                         $key = array_search( $target, $srchres );
 191                         if ( $key !== false ) {
 192                                 // Exact match is a redirect to one of the returned matches so pull the
 193                                 // returned match to the front.  This might look odd but the alternative
 194                                 // is to put the redirect in front and drop the match.  The name of the
 195                                 // found match is often more descriptive/better formed than the name of
 196                                 // the redirect AND by definition they share a prefix.  Hopefully this
 197                                 // choice is less confusing and more helpful.  But it might not be.  But
 198                                 // it is the choice we're going with for now.
 199                                 return $this->pullFront( $key, $srchres );
 200                         }
 201                         $redirectTargetsToRedirect = $this->redirectTargetsToRedirect( $srchres );
 202                         if ( isset( $redirectTargetsToRedirect[$target] ) ) {
 203                                 // The exact match and something in the results list are both redirects
 204                                 // to the same thing!  In this case we'll pull the returned match to the
 205                                 // top following the same logic above.  Again, it might not be a perfect
 206                                 // choice but it'll do.
 207                                 return $this->pullFront( $redirectTargetsToRedirect[$target], $srchres );
 208                         }
 209                 } else {
 210                         $redirectTargetsToRedirect = $this->redirectTargetsToRedirect( $srchres );
 211                         if ( isset( $redirectTargetsToRedirect[$string] ) ) {
 212                                 // The exact match is the target of a redirect already in the results list so remove
 213                                 // the redirect from the results list and push the exact match to the front
 214                                 array_splice( $srchres, $redirectTargetsToRedirect[$string], 1 );
 215                                 array_unshift( $srchres, $string );
 216                                 return $srchres;
 217                         }
 218                 }
 219
 220                 // Exact match is totally unique from the other results so just add it to the front
 221                 array_unshift( $srchres, $string );
 222                 // And roll one off the end if the results are too long
 223                 if ( count( $srchres ) > $limit ) {
 224                         array_pop( $srchres );
 225                 }
 226                 return $srchres;
 227         }
 228
 229         /**
 230          * @param Array(string) $titles as strings
 231          * @return Array(string => int) redirect target prefixedText to index of title in titles
 232          *   that is a redirect to it.
 233          */
 234         private function redirectTargetsToRedirect( $titles ) {
 235                 $result = array();
 236                 foreach ( $titles as $key => $titleText ) {
 237                         $title = Title::newFromText( $titleText );
 238                         if ( !$title || !$title->isRedirect() ) {
 239                                 continue;
 240                         }
 241                         $target = $this->getRedirectTarget( $title );
 242                         if ( !$target ) {
 243                                 continue;
 244                         }
 245                         $result[$target] = $key;
 246                 }
 247                 return $result;
 248         }
 249
 250         /**
 251          * @param int $key key to pull to the front
 252          * @return array $array with the item at $key pulled to the front
 253          */
 254         private function pullFront( $key, $array ) {
 255                 $cut = array_splice( $array, $key, 1 );
 256                 array_unshift( $array, $cut[0] );
 257                 return $array;
 258         }
 259
 260         private function getRedirectTarget( $title ) {
 261                 $page = WikiPage::factory( $title );
 262                 if ( !$page->exists() ) {
 263                         return null;
 264                 }
 265                 return $page->getRedirectTarget()->getPrefixedText();
 266         }
 267
 268         /**
 269          * Prefix search special-case for Special: namespace.
 270          *
 271          * @param string $search Term
 272          * @param int $limit Max number of items to return
 273          * @param int $offset Number of items to offset
 274          * @return array
 275          */
 276         protected function specialSearch( $search, $limit, $offset ) {
 277                 global $wgContLang;
 278
 279                 $searchParts = explode( '/', $search, 2 );
 280                 $searchKey = $searchParts[0];
 281                 $subpageSearch = isset( $searchParts[1] ) ? $searchParts[1] : null;
 282
 283                 // Handle subpage search separately.
 284                 if ( $subpageSearch !== null ) {
 285                         // Try matching the full search string as a page name
 286                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 287                         if ( !$specialTitle ) {
 288                                 return array();
 289                         }
 290                         $special = SpecialPageFactory::getPage( $specialTitle->getText() );
 291                         if ( $special ) {
 292                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset );
 293                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 294                                         return $specialTitle->getSubpage( $sub );
 295                                 }, $subpages );
 296                         } else {
 297                                 return array();
 298                         }
 299                 }
 300
 301                 # normalize searchKey, so aliases with spaces can be found - bug 25675
 302                 $searchKey = str_replace( ' ', '_', $searchKey );
 303                 $searchKey = $wgContLang->caseFold( $searchKey );
 304
 305                 // Unlike SpecialPage itself, we want the canonical forms of both
 306                 // canonical and alias title forms...
 307                 $keys = array();
 308                 foreach ( SpecialPageFactory::getNames() as $page ) {
 309                         $keys[$wgContLang->caseFold( $page )] = $page;
 310                 }
 311
 312                 foreach ( $wgContLang->getSpecialPageAliases() as $page => $aliases ) {
 313                         if ( !in_array( $page, SpecialPageFactory::getNames() ) ) {# bug 20885
 314                                 continue;
 315                         }
 316
 317                         foreach ( $aliases as $alias ) {
 318                                 $keys[$wgContLang->caseFold( $alias )] = $alias;
 319                         }
 320                 }
 321                 ksort( $keys );
 322
 323                 $srchres = array();
 324                 $skipped = 0;
 325                 foreach ( $keys as $pageKey => $page ) {
 326                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 327                                 // bug 27671: Don't use SpecialPage::getTitleFor() here because it
 328                                 // localizes its input leading to searches for e.g. Special:All
 329                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 330                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 331                                 if ( $offset > 0 && $skipped < $offset ) {
 332                                         $skipped++;
 333                                         continue;
 334                                 }
 335                                 $srchres[] = Title::makeTitleSafe( NS_SPECIAL, $page );
 336                         }
 337
 338                         if ( count( $srchres ) >= $limit ) {
 339                                 break;
 340                         }
 341                 }
 342
 343                 return $srchres;
 344         }
 345
 346         /**
 347          * Unless overridden by PrefixSearchBackend hook...
 348          * This is case-sensitive (First character may
 349          * be automatically capitalized by Title::secureAndSpit()
 350          * later on depending on $wgCapitalLinks)
 351          *
 352          * @param array $namespaces Namespaces to search in
 353          * @param string $search Term
 354          * @param int $limit Max number of items to return
 355          * @param int $offset Number of items to skip
 356          * @return array Array of Title objects
 357          */
 358         protected function defaultSearchBackend( $namespaces, $search, $limit, $offset ) {
 359                 $ns = array_shift( $namespaces ); // support only one namespace
 360                 if ( in_array( NS_MAIN, $namespaces ) ) {
 361                         $ns = NS_MAIN; // if searching on many always default to main
 362                 }
 363
 364                 $t = Title::newFromText( $search, $ns );
 365                 $prefix = $t ? $t->getDBkey() : '';
 366                 $dbr = wfGetDB( DB_SLAVE );
 367                 $res = $dbr->select( 'page',
 368                         array( 'page_id', 'page_namespace', 'page_title' ),
 369                         array(
 370                                 'page_namespace' => $ns,
 371                                 'page_title ' . $dbr->buildLike( $prefix, $dbr->anyString() )
 372                         ),
 373                         __METHOD__,
 374                         array(
 375                                 'LIMIT' => $limit,
 376                                 'ORDER BY' => 'page_title',
 377                                 'OFFSET' => $offset
 378                         )
 379                 );
 380                 $srchres = array();
 381                 foreach ( $res as $row ) {
 382                         $srchres[] = Title::newFromRow( $row );
 383                 }
 384                 return $srchres;
 385         }
 386
 387         /**
 388          * Validate an array of numerical namespace indexes
 389          *
 390          * @param array $namespaces
 391          * @return array (default: contains only NS_MAIN)
 392          */
 393         protected function validateNamespaces( $namespaces ) {
 394                 global $wgContLang;
 395
 396                 // We will look at each given namespace against wgContLang namespaces
 397                 $validNamespaces = $wgContLang->getNamespaces();
 398                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 399                         $valid = array();
 400                         foreach ( $namespaces as $ns ) {
 401                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 402                                         $valid[] = $ns;
 403                                 }
 404                         }
 405                         if ( count( $valid ) > 0 ) {
 406                                 return $valid;
 407                         }
 408                 }
 409
 410                 return array( NS_MAIN );
 411         }
 412 }
 413
 414 /**
 415  * Performs prefix search, returning Title objects
 416  * @ingroup Search
 417  */
 418 class TitlePrefixSearch extends PrefixSearch {
 419
 420         protected function titles( array $titles ) {
 421                 return $titles;
 422         }
 423
 424         protected function strings( array $strings ) {
 425                 $titles = array_map( 'Title::newFromText', $strings );
 426                 $lb = new LinkBatch( $titles );
 427                 $lb->setCaller( __METHOD__ );
 428                 $lb->execute();
 429                 return $titles;
 430         }
 431 }
 432
 433 /**
 434  * Performs prefix search, returning strings
 435  * @ingroup Search
 436  */
 437 class StringPrefixSearch extends PrefixSearch {
 438
 439         protected function titles( array $titles ) {
 440                 return array_map( function ( Title $t ) {
 441                         return $t->getPrefixedText();
 442                 }, $titles );
 443         }
 444
 445         protected function strings( array $strings ) {
 446                 return $strings;
 447         }
 448 }