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  * @deprecated Since 1.27, Use SearchEngine::prefixSearchSubpages or SearchEngine::completionSearch
  27  *
  28  * @ingroup Search
  29  */
  30 abstract class PrefixSearch {
  31         /**
  32          * Do a prefix search of titles and return a list of matching page names.
  33          * @deprecated Since 1.23, use TitlePrefixSearch or StringPrefixSearch classes
  34          *
  35          * @param string $search
  36          * @param int $limit
  37          * @param array $namespaces Used if query is not explicitly prefixed
  38          * @param int $offset How many results to offset from the beginning
  39          * @return array Array of strings
  40          */
  41         public static function titleSearch( $search, $limit, $namespaces = [], $offset = 0 ) {
  42                 $prefixSearch = new StringPrefixSearch;
  43                 return $prefixSearch->search( $search, $limit, $namespaces, $offset );
  44         }
  45
  46         /**
  47          * Do a prefix search of titles and return a list of matching page names.
  48          *
  49          * @param string $search
  50          * @param int $limit
  51          * @param array $namespaces Used if query is not explicitly prefixed
  52          * @param int $offset How many results to offset from the beginning
  53          * @return array Array of strings or Title objects
  54          */
  55         public function search( $search, $limit, $namespaces = [], $offset = 0 ) {
  56                 $search = trim( $search );
  57                 if ( $search == '' ) {
  58                         return []; // Return empty result
  59                 }
  60
  61                 $hasNamespace = $this->extractNamespace( $search );
  62                 if ( $hasNamespace ) {
  63                         list( $namespace, $search ) = $hasNamespace;
  64                         $namespaces = [ $namespace ];
  65                 } else {
  66                         $namespaces = $this->validateNamespaces( $namespaces );
  67                         Hooks::run( 'PrefixSearchExtractNamespace', [ &$namespaces, &$search ] );
  68                 }
  69
  70                 return $this->searchBackend( $namespaces, $search, $limit, $offset );
  71         }
  72
  73         /**
  74          * Figure out if given input contains an explicit namespace.
  75          *
  76          * @param string $input
  77          * @return false|array Array of namespace and remaining text, or false if no namespace given.
  78          */
  79         protected function extractNamespace( $input ) {
  80                 if ( strpos( $input, ':' ) === false ) {
  81                         return false;
  82                 }
  83
  84                 // Namespace prefix only
  85                 $title = Title::newFromText( $input . 'Dummy' );
  86                 if (
  87                         $title &&
  88                         $title->getText() === 'Dummy' &&
  89                         !$title->inNamespace( NS_MAIN ) &&
  90                         !$title->isExternal()
  91                 ) {
  92                         return [ $title->getNamespace(), '' ];
  93                 }
  94
  95                 // Namespace prefix with additional input
  96                 $title = Title::newFromText( $input );
  97                 if (
  98                         $title &&
  99                         !$title->inNamespace( NS_MAIN ) &&
 100                         !$title->isExternal()
 101                 ) {
 102                         // getText provides correct capitalization
 103                         return [ $title->getNamespace(), $title->getText() ];
 104                 }
 105
 106                 return false;
 107         }
 108
 109         /**
 110          * Do a prefix search for all possible variants of the prefix
 111          * @param string $search
 112          * @param int $limit
 113          * @param array $namespaces
 114          * @param int $offset How many results to offset from the beginning
 115          *
 116          * @return array
 117          */
 118         public function searchWithVariants( $search, $limit, array $namespaces, $offset = 0 ) {
 119                 $searches = $this->search( $search, $limit, $namespaces, $offset );
 120
 121                 // if the content language has variants, try to retrieve fallback results
 122                 $fallbackLimit = $limit - count( $searches );
 123                 if ( $fallbackLimit > 0 ) {
 124                         global $wgContLang;
 125
 126                         $fallbackSearches = $wgContLang->autoConvertToAllVariants( $search );
 127                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), [ $search ] );
 128
 129                         foreach ( $fallbackSearches as $fbs ) {
 130                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
 131                                 $searches = array_merge( $searches, $fallbackSearchResult );
 132                                 $fallbackLimit -= count( $fallbackSearchResult );
 133
 134                                 if ( $fallbackLimit == 0 ) {
 135                                         break;
 136                                 }
 137                         }
 138                 }
 139                 return $searches;
 140         }
 141
 142         /**
 143          * When implemented in a descendant class, receives an array of Title objects and returns
 144          * either an unmodified array or an array of strings corresponding to titles passed to it.
 145          *
 146          * @param array $titles
 147          * @return array
 148          */
 149         abstract protected function titles( array $titles );
 150
 151         /**
 152          * When implemented in a descendant class, receives an array of titles as strings and returns
 153          * either an unmodified array or an array of Title objects corresponding to strings received.
 154          *
 155          * @param array $strings
 156          *
 157          * @return array
 158          */
 159         abstract protected function strings( array $strings );
 160
 161         /**
 162          * Do a prefix search of titles and return a list of matching page names.
 163          * @param array $namespaces
 164          * @param string $search
 165          * @param int $limit
 166          * @param int $offset How many results to offset from the beginning
 167          * @return array Array of strings
 168          */
 169         protected function searchBackend( $namespaces, $search, $limit, $offset ) {
 170                 if ( count( $namespaces ) == 1 ) {
 171                         $ns = $namespaces[0];
 172                         if ( $ns == NS_MEDIA ) {
 173                                 $namespaces = [ NS_FILE ];
 174                         } elseif ( $ns == NS_SPECIAL ) {
 175                                 return $this->titles( $this->specialSearch( $search, $limit, $offset ) );
 176                         }
 177                 }
 178                 $srchres = [];
 179                 if ( Hooks::run(
 180                         'PrefixSearchBackend',
 181                         [ $namespaces, $search, $limit, &$srchres, $offset ]
 182                 ) ) {
 183                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) );
 184                 }
 185                 return $this->strings(
 186                         $this->handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) );
 187         }
 188
 189         private function handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) {
 190                 if ( $offset === 0 ) {
 191                         // Only perform exact db match if offset === 0
 192                         // This is still far from perfect but at least we avoid returning the
 193                         // same title afain and again when the user is scrolling with a query
 194                         // that matches a title in the db.
 195                         $rescorer = new SearchExactMatchRescorer();
 196                         $srchres = $rescorer->rescore( $search, $namespaces, $srchres, $limit );
 197                 }
 198                 return $srchres;
 199         }
 200
 201         /**
 202          * Prefix search special-case for Special: namespace.
 203          *
 204          * @param string $search Term
 205          * @param int $limit Max number of items to return
 206          * @param int $offset Number of items to offset
 207          * @return array
 208          */
 209         protected function specialSearch( $search, $limit, $offset ) {
 210                 global $wgContLang;
 211
 212                 $searchParts = explode( '/', $search, 2 );
 213                 $searchKey = $searchParts[0];
 214                 $subpageSearch = isset( $searchParts[1] ) ? $searchParts[1] : null;
 215
 216                 // Handle subpage search separately.
 217                 if ( $subpageSearch !== null ) {
 218                         // Try matching the full search string as a page name
 219                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 220                         if ( !$specialTitle ) {
 221                                 return [];
 222                         }
 223                         $special = SpecialPageFactory::getPage( $specialTitle->getText() );
 224                         if ( $special ) {
 225                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset );
 226                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 227                                         return $specialTitle->getSubpage( $sub );
 228                                 }, $subpages );
 229                         } else {
 230                                 return [];
 231                         }
 232                 }
 233
 234                 # normalize searchKey, so aliases with spaces can be found - bug 25675
 235                 $searchKey = str_replace( ' ', '_', $searchKey );
 236                 $searchKey = $wgContLang->caseFold( $searchKey );
 237
 238                 // Unlike SpecialPage itself, we want the canonical forms of both
 239                 // canonical and alias title forms...
 240                 $keys = [];
 241                 foreach ( SpecialPageFactory::getNames() as $page ) {
 242                         $keys[$wgContLang->caseFold( $page )] = [ 'page' => $page, 'rank' => 0 ];
 243                 }
 244
 245                 foreach ( $wgContLang->getSpecialPageAliases() as $page => $aliases ) {
 246                         if ( !in_array( $page, SpecialPageFactory::getNames() ) ) {# bug 20885
 247                                 continue;
 248                         }
 249
 250                         foreach ( $aliases as $key => $alias ) {
 251                                 $keys[$wgContLang->caseFold( $alias )] = [ 'page' => $alias, 'rank' => $key ];
 252                         }
 253                 }
 254                 ksort( $keys );
 255
 256                 $matches = [];
 257                 foreach ( $keys as $pageKey => $page ) {
 258                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 259                                 // bug 27671: Don't use SpecialPage::getTitleFor() here because it
 260                                 // localizes its input leading to searches for e.g. Special:All
 261                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 262                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 263                                 $matches[$page['rank']][] = Title::makeTitleSafe( NS_SPECIAL, $page['page'] );
 264
 265                                 if ( isset( $matches[0] ) && count( $matches[0] ) >= $limit + $offset ) {
 266                                         // We have enough items in primary rank, no use to continue
 267                                         break;
 268                                 }
 269                         }
 270
 271                 }
 272
 273                 // Ensure keys are in order
 274                 ksort( $matches );
 275                 // Flatten the array
 276                 $matches = array_reduce( $matches, 'array_merge', [] );
 277
 278                 return array_slice( $matches, $offset, $limit );
 279         }
 280
 281         /**
 282          * Unless overridden by PrefixSearchBackend hook...
 283          * This is case-sensitive (First character may
 284          * be automatically capitalized by Title::secureAndSpit()
 285          * later on depending on $wgCapitalLinks)
 286          *
 287          * @param array|null $namespaces Namespaces to search in
 288          * @param string $search Term
 289          * @param int $limit Max number of items to return
 290          * @param int $offset Number of items to skip
 291          * @return Title[] Array of Title objects
 292          */
 293         public function defaultSearchBackend( $namespaces, $search, $limit, $offset ) {
 294                 // Backwards compatability with old code. Default to NS_MAIN if no namespaces provided.
 295                 if ( $namespaces === null ) {
 296                         $namespaces = [];
 297                 }
 298                 if ( !$namespaces ) {
 299                         $namespaces[] = NS_MAIN;
 300                 }
 301
 302                 // Construct suitable prefix for each namespace. They differ in cases where
 303                 // some namespaces always capitalize and some don't.
 304                 $prefixes = [];
 305                 foreach ( $namespaces as $namespace ) {
 306                         // For now, if special is included, ignore the other namespaces
 307                         if ( $namespace == NS_SPECIAL ) {
 308                                 return $this->specialSearch( $search, $limit, $offset );
 309                         }
 310
 311                         $title = Title::makeTitleSafe( $namespace, $search );
 312                         // Why does the prefix default to empty?
 313                         $prefix = $title ? $title->getDBkey() : '';
 314                         $prefixes[$prefix][] = $namespace;
 315                 }
 316
 317                 $dbr = wfGetDB( DB_REPLICA );
 318                 // Often there is only one prefix that applies to all requested namespaces,
 319                 // but sometimes there are two if some namespaces do not always capitalize.
 320                 $conds = [];
 321                 foreach ( $prefixes as $prefix => $namespaces ) {
 322                         $condition = [
 323                                 'page_namespace' => $namespaces,
 324                                 'page_title' . $dbr->buildLike( $prefix, $dbr->anyString() ),
 325                         ];
 326                         $conds[] = $dbr->makeList( $condition, LIST_AND );
 327                 }
 328
 329                 $table = 'page';
 330                 $fields = [ 'page_id', 'page_namespace', 'page_title' ];
 331                 $conds = $dbr->makeList( $conds, LIST_OR );
 332                 $options = [
 333                         'LIMIT' => $limit,
 334                         'ORDER BY' => [ 'page_title', 'page_namespace' ],
 335                         'OFFSET' => $offset
 336                 ];
 337
 338                 $res = $dbr->select( $table, $fields, $conds, __METHOD__, $options );
 339
 340                 return iterator_to_array( TitleArray::newFromResult( $res ) );
 341         }
 342
 343         /**
 344          * Validate an array of numerical namespace indexes
 345          *
 346          * @param array $namespaces
 347          * @return array (default: contains only NS_MAIN)
 348          */
 349         protected function validateNamespaces( $namespaces ) {
 350                 global $wgContLang;
 351
 352                 // We will look at each given namespace against wgContLang namespaces
 353                 $validNamespaces = $wgContLang->getNamespaces();
 354                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 355                         $valid = [];
 356                         foreach ( $namespaces as $ns ) {
 357                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 358                                         $valid[] = $ns;
 359                                 }
 360                         }
 361                         if ( count( $valid ) > 0 ) {
 362                                 return $valid;
 363                         }
 364                 }
 365
 366                 return [ NS_MAIN ];
 367         }
 368 }
 369
 370 /**
 371  * Performs prefix search, returning Title objects
 372  * @deprecated Since 1.27, Use SearchEngine::prefixSearchSubpages or SearchEngine::completionSearch
 373  * @ingroup Search
 374  */
 375 class TitlePrefixSearch extends PrefixSearch {
 376
 377         protected function titles( array $titles ) {
 378                 return $titles;
 379         }
 380
 381         protected function strings( array $strings ) {
 382                 $titles = array_map( 'Title::newFromText', $strings );
 383                 $lb = new LinkBatch( $titles );
 384                 $lb->setCaller( __METHOD__ );
 385                 $lb->execute();
 386                 return $titles;
 387         }
 388 }
 389
 390 /**
 391  * Performs prefix search, returning strings
 392  * @deprecated Since 1.27, Use SearchEngine::prefixSearchSubpages or SearchEngine::completionSearch
 393  * @ingroup Search
 394  */
 395 class StringPrefixSearch extends PrefixSearch {
 396
 397         protected function titles( array $titles ) {
 398                 return array_map( function ( Title $t ) {
 399                         return $t->getPrefixedText();
 400                 }, $titles );
 401         }
 402
 403         protected function strings( array $strings ) {
 404                 return $strings;
 405         }
 406 }