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 use MediaWiki\MediaWikiServices;
  24
  25 /**
  26  * Handles searching prefixes of titles and finding any page
  27  * names that match. Used largely by the OpenSearch implementation.
  28  * @deprecated Since 1.27, Use SearchEngine::defaultPrefixSearch or SearchEngine::completionSearch
  29  *
  30  * @ingroup Search
  31  */
  32 abstract class PrefixSearch {
  33         /**
  34          * Do a prefix search of titles and return a list of matching page names.
  35          * @deprecated Since 1.23, use TitlePrefixSearch or StringPrefixSearch classes
  36          *
  37          * @param string $search
  38          * @param int $limit
  39          * @param array $namespaces Used if query is not explicitly prefixed
  40          * @param int $offset How many results to offset from the beginning
  41          * @return array Array of strings
  42          */
  43         public static function titleSearch( $search, $limit, $namespaces = [], $offset = 0 ) {
  44                 $prefixSearch = new StringPrefixSearch;
  45                 return $prefixSearch->search( $search, $limit, $namespaces, $offset );
  46         }
  47
  48         /**
  49          * Do a prefix search of titles and return a list of matching page names.
  50          *
  51          * @param string $search
  52          * @param int $limit
  53          * @param array $namespaces Used if query is not explicitly prefixed
  54          * @param int $offset How many results to offset from the beginning
  55          * @return array Array of strings or Title objects
  56          */
  57         public function search( $search, $limit, $namespaces = [], $offset = 0 ) {
  58                 $search = trim( $search );
  59                 if ( $search == '' ) {
  60                         return []; // Return empty result
  61                 }
  62
  63                 $hasNamespace = SearchEngine::parseNamespacePrefixes( $search, false, true );
  64                 if ( $hasNamespace !== false ) {
  65                         list( $search, $namespaces ) = $hasNamespace;
  66                 }
  67
  68                 return $this->searchBackend( $namespaces, $search, $limit, $offset );
  69         }
  70
  71         /**
  72          * Do a prefix search for all possible variants of the prefix
  73          * @param string $search
  74          * @param int $limit
  75          * @param array $namespaces
  76          * @param int $offset How many results to offset from the beginning
  77          *
  78          * @return array
  79          */
  80         public function searchWithVariants( $search, $limit, array $namespaces, $offset = 0 ) {
  81                 $searches = $this->search( $search, $limit, $namespaces, $offset );
  82
  83                 // if the content language has variants, try to retrieve fallback results
  84                 $fallbackLimit = $limit - count( $searches );
  85                 if ( $fallbackLimit > 0 ) {
  86                         $fallbackSearches = MediaWikiServices::getInstance()->getContentLanguage()->
  87                                 autoConvertToAllVariants( $search );
  88                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), [ $search ] );
  89
  90                         foreach ( $fallbackSearches as $fbs ) {
  91                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
  92                                 $searches = array_merge( $searches, $fallbackSearchResult );
  93                                 $fallbackLimit -= count( $fallbackSearchResult );
  94
  95                                 if ( $fallbackLimit == 0 ) {
  96                                         break;
  97                                 }
  98                         }
  99                 }
 100                 return $searches;
 101         }
 102
 103         /**
 104          * When implemented in a descendant class, receives an array of Title objects and returns
 105          * either an unmodified array or an array of strings corresponding to titles passed to it.
 106          *
 107          * @param array $titles
 108          * @return array
 109          */
 110         abstract protected function titles( array $titles );
 111
 112         /**
 113          * When implemented in a descendant class, receives an array of titles as strings and returns
 114          * either an unmodified array or an array of Title objects corresponding to strings received.
 115          *
 116          * @param array $strings
 117          *
 118          * @return array
 119          */
 120         abstract protected function strings( array $strings );
 121
 122         /**
 123          * Do a prefix search of titles and return a list of matching page names.
 124          * @param array $namespaces
 125          * @param string $search
 126          * @param int $limit
 127          * @param int $offset How many results to offset from the beginning
 128          * @return array Array of strings
 129          */
 130         protected function searchBackend( $namespaces, $search, $limit, $offset ) {
 131                 if ( count( $namespaces ) == 1 ) {
 132                         $ns = $namespaces[0];
 133                         if ( $ns == NS_MEDIA ) {
 134                                 $namespaces = [ NS_FILE ];
 135                         } elseif ( $ns == NS_SPECIAL ) {
 136                                 return $this->titles( $this->specialSearch( $search, $limit, $offset ) );
 137                         }
 138                 }
 139                 $srchres = [];
 140                 if ( Hooks::run(
 141                         'PrefixSearchBackend',
 142                         [ $namespaces, $search, $limit, &$srchres, $offset ]
 143                 ) ) {
 144                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) );
 145                 }
 146                 return $this->strings(
 147                         $this->handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) );
 148         }
 149
 150         private function handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) {
 151                 if ( $offset === 0 ) {
 152                         // Only perform exact db match if offset === 0
 153                         // This is still far from perfect but at least we avoid returning the
 154                         // same title afain and again when the user is scrolling with a query
 155                         // that matches a title in the db.
 156                         $rescorer = new SearchExactMatchRescorer();
 157                         $srchres = $rescorer->rescore( $search, $namespaces, $srchres, $limit );
 158                 }
 159                 return $srchres;
 160         }
 161
 162         /**
 163          * Prefix search special-case for Special: namespace.
 164          *
 165          * @param string $search Term
 166          * @param int $limit Max number of items to return
 167          * @param int $offset Number of items to offset
 168          * @return array
 169          */
 170         protected function specialSearch( $search, $limit, $offset ) {
 171                 $searchParts = explode( '/', $search, 2 );
 172                 $searchKey = $searchParts[0];
 173                 $subpageSearch = $searchParts[1] ?? null;
 174
 175                 // Handle subpage search separately.
 176                 if ( $subpageSearch !== null ) {
 177                         // Try matching the full search string as a page name
 178                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 179                         if ( !$specialTitle ) {
 180                                 return [];
 181                         }
 182                         $special = SpecialPageFactory::getPage( $specialTitle->getText() );
 183                         if ( $special ) {
 184                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset );
 185                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 186                                         return $specialTitle->getSubpage( $sub );
 187                                 }, $subpages );
 188                         } else {
 189                                 return [];
 190                         }
 191                 }
 192
 193                 # normalize searchKey, so aliases with spaces can be found - T27675
 194                 $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 195                 $searchKey = str_replace( ' ', '_', $searchKey );
 196                 $searchKey = $contLang->caseFold( $searchKey );
 197
 198                 // Unlike SpecialPage itself, we want the canonical forms of both
 199                 // canonical and alias title forms...
 200                 $keys = [];
 201                 foreach ( SpecialPageFactory::getNames() as $page ) {
 202                         $keys[$contLang->caseFold( $page )] = [ 'page' => $page, 'rank' => 0 ];
 203                 }
 204
 205                 foreach ( $contLang->getSpecialPageAliases() as $page => $aliases ) {
 206                         if ( !in_array( $page, SpecialPageFactory::getNames() ) ) {# T22885
 207                                 continue;
 208                         }
 209
 210                         foreach ( $aliases as $key => $alias ) {
 211                                 $keys[$contLang->caseFold( $alias )] = [ 'page' => $alias, 'rank' => $key ];
 212                         }
 213                 }
 214                 ksort( $keys );
 215
 216                 $matches = [];
 217                 foreach ( $keys as $pageKey => $page ) {
 218                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 219                                 // T29671: Don't use SpecialPage::getTitleFor() here because it
 220                                 // localizes its input leading to searches for e.g. Special:All
 221                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 222                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 223                                 $matches[$page['rank']][] = Title::makeTitleSafe( NS_SPECIAL, $page['page'] );
 224
 225                                 if ( isset( $matches[0] ) && count( $matches[0] ) >= $limit + $offset ) {
 226                                         // We have enough items in primary rank, no use to continue
 227                                         break;
 228                                 }
 229                         }
 230
 231                 }
 232
 233                 // Ensure keys are in order
 234                 ksort( $matches );
 235                 // Flatten the array
 236                 $matches = array_reduce( $matches, 'array_merge', [] );
 237
 238                 return array_slice( $matches, $offset, $limit );
 239         }
 240
 241         /**
 242          * Unless overridden by PrefixSearchBackend hook...
 243          * This is case-sensitive (First character may
 244          * be automatically capitalized by Title::secureAndSpit()
 245          * later on depending on $wgCapitalLinks)
 246          *
 247          * @param array|null $namespaces Namespaces to search in
 248          * @param string $search Term
 249          * @param int $limit Max number of items to return
 250          * @param int $offset Number of items to skip
 251          * @return Title[] Array of Title objects
 252          */
 253         public function defaultSearchBackend( $namespaces, $search, $limit, $offset ) {
 254                 // Backwards compatability with old code. Default to NS_MAIN if no namespaces provided.
 255                 if ( $namespaces === null ) {
 256                         $namespaces = [];
 257                 }
 258                 if ( !$namespaces ) {
 259                         $namespaces[] = NS_MAIN;
 260                 }
 261
 262                 // Construct suitable prefix for each namespace. They differ in cases where
 263                 // some namespaces always capitalize and some don't.
 264                 $prefixes = [];
 265                 foreach ( $namespaces as $namespace ) {
 266                         // For now, if special is included, ignore the other namespaces
 267                         if ( $namespace == NS_SPECIAL ) {
 268                                 return $this->specialSearch( $search, $limit, $offset );
 269                         }
 270
 271                         $title = Title::makeTitleSafe( $namespace, $search );
 272                         // Why does the prefix default to empty?
 273                         $prefix = $title ? $title->getDBkey() : '';
 274                         $prefixes[$prefix][] = $namespace;
 275                 }
 276
 277                 $dbr = wfGetDB( DB_REPLICA );
 278                 // Often there is only one prefix that applies to all requested namespaces,
 279                 // but sometimes there are two if some namespaces do not always capitalize.
 280                 $conds = [];
 281                 foreach ( $prefixes as $prefix => $namespaces ) {
 282                         $condition = [
 283                                 'page_namespace' => $namespaces,
 284                                 'page_title' . $dbr->buildLike( $prefix, $dbr->anyString() ),
 285                         ];
 286                         $conds[] = $dbr->makeList( $condition, LIST_AND );
 287                 }
 288
 289                 $table = 'page';
 290                 $fields = [ 'page_id', 'page_namespace', 'page_title' ];
 291                 $conds = $dbr->makeList( $conds, LIST_OR );
 292                 $options = [
 293                         'LIMIT' => $limit,
 294                         'ORDER BY' => [ 'page_title', 'page_namespace' ],
 295                         'OFFSET' => $offset
 296                 ];
 297
 298                 $res = $dbr->select( $table, $fields, $conds, __METHOD__, $options );
 299
 300                 return iterator_to_array( TitleArray::newFromResult( $res ) );
 301         }
 302
 303         /**
 304          * Validate an array of numerical namespace indexes
 305          *
 306          * @param array $namespaces
 307          * @return array (default: contains only NS_MAIN)
 308          */
 309         protected function validateNamespaces( $namespaces ) {
 310                 // We will look at each given namespace against content language namespaces
 311                 $validNamespaces = MediaWikiServices::getInstance()->getContentLanguage()->getNamespaces();
 312                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 313                         $valid = [];
 314                         foreach ( $namespaces as $ns ) {
 315                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 316                                         $valid[] = $ns;
 317                                 }
 318                         }
 319                         if ( count( $valid ) > 0 ) {
 320                                 return $valid;
 321                         }
 322                 }
 323
 324                 return [ NS_MAIN ];
 325         }
 326 }
 327
 328 /**
 329  * Performs prefix search, returning Title objects
 330  * @deprecated Since 1.27, Use SearchEngine::defaultPrefixSearch or SearchEngine::completionSearch
 331  * @ingroup Search
 332  */
 333 class TitlePrefixSearch extends PrefixSearch {
 334
 335         protected function titles( array $titles ) {
 336                 return $titles;
 337         }
 338
 339         protected function strings( array $strings ) {
 340                 $titles = array_map( 'Title::newFromText', $strings );
 341                 $lb = new LinkBatch( $titles );
 342                 $lb->setCaller( __METHOD__ );
 343                 $lb->execute();
 344                 return $titles;
 345         }
 346 }
 347
 348 /**
 349  * Performs prefix search, returning strings
 350  * @deprecated Since 1.27, Use SearchEngine::prefixSearchSubpages or SearchEngine::completionSearch
 351  * @ingroup Search
 352  */
 353 class StringPrefixSearch extends PrefixSearch {
 354
 355         protected function titles( array $titles ) {
 356                 return array_map( function ( Title $t ) {
 357                         return $t->getPrefixedText();
 358                 }, $titles );
 359         }
 360
 361         protected function strings( array $strings ) {
 362                 return $strings;
 363         }
 364 }