includes/search/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                 wfDeprecated( __METHOD__, '1.34' );
  45                 $prefixSearch = new StringPrefixSearch;
  46                 return $prefixSearch->search( $search, $limit, $namespaces, $offset );
  47         }
  48
  49         /**
  50          * Do a prefix search of titles and return a list of matching page names.
  51          *
  52          * @param string $search
  53          * @param int $limit
  54          * @param array $namespaces Used if query is not explicitly prefixed
  55          * @param int $offset How many results to offset from the beginning
  56          * @return array Array of strings or Title objects
  57          */
  58         public function search( $search, $limit, $namespaces = [], $offset = 0 ) {
  59                 $search = trim( $search );
  60                 if ( $search == '' ) {
  61                         return []; // Return empty result
  62                 }
  63
  64                 $hasNamespace = SearchEngine::parseNamespacePrefixes( $search, false, true );
  65                 if ( $hasNamespace !== false ) {
  66                         list( $search, $namespaces ) = $hasNamespace;
  67                 }
  68
  69                 return $this->searchBackend( $namespaces, $search, $limit, $offset );
  70         }
  71
  72         /**
  73          * Do a prefix search for all possible variants of the prefix
  74          * @param string $search
  75          * @param int $limit
  76          * @param array $namespaces
  77          * @param int $offset How many results to offset from the beginning
  78          *
  79          * @return array
  80          */
  81         public function searchWithVariants( $search, $limit, array $namespaces, $offset = 0 ) {
  82                 $searches = $this->search( $search, $limit, $namespaces, $offset );
  83
  84                 // if the content language has variants, try to retrieve fallback results
  85                 $fallbackLimit = $limit - count( $searches );
  86                 if ( $fallbackLimit > 0 ) {
  87                         $fallbackSearches = MediaWikiServices::getInstance()->getContentLanguage()->
  88                                 autoConvertToAllVariants( $search );
  89                         $fallbackSearches = array_diff( array_unique( $fallbackSearches ), [ $search ] );
  90
  91                         foreach ( $fallbackSearches as $fbs ) {
  92                                 $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces );
  93                                 $searches = array_merge( $searches, $fallbackSearchResult );
  94                                 $fallbackLimit -= count( $fallbackSearchResult );
  95
  96                                 if ( $fallbackLimit == 0 ) {
  97                                         break;
  98                                 }
  99                         }
 100                 }
 101                 return $searches;
 102         }
 103
 104         /**
 105          * When implemented in a descendant class, receives an array of Title objects and returns
 106          * either an unmodified array or an array of strings corresponding to titles passed to it.
 107          *
 108          * @param array $titles
 109          * @return array
 110          */
 111         abstract protected function titles( array $titles );
 112
 113         /**
 114          * When implemented in a descendant class, receives an array of titles as strings and returns
 115          * either an unmodified array or an array of Title objects corresponding to strings received.
 116          *
 117          * @param array $strings
 118          *
 119          * @return array
 120          */
 121         abstract protected function strings( array $strings );
 122
 123         /**
 124          * Do a prefix search of titles and return a list of matching page names.
 125          * @param array $namespaces
 126          * @param string $search
 127          * @param int $limit
 128          * @param int $offset How many results to offset from the beginning
 129          * @return array Array of strings
 130          */
 131         protected function searchBackend( $namespaces, $search, $limit, $offset ) {
 132                 if ( count( $namespaces ) == 1 ) {
 133                         $ns = $namespaces[0];
 134                         if ( $ns == NS_MEDIA ) {
 135                                 $namespaces = [ NS_FILE ];
 136                         } elseif ( $ns == NS_SPECIAL ) {
 137                                 return $this->titles( $this->specialSearch( $search, $limit, $offset ) );
 138                         }
 139                 }
 140                 $srchres = [];
 141                 if ( Hooks::run(
 142                         'PrefixSearchBackend',
 143                         [ $namespaces, $search, $limit, &$srchres, $offset ]
 144                 ) ) {
 145                         return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) );
 146                 }
 147                 return $this->strings(
 148                         $this->handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) );
 149         }
 150
 151         private function handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) {
 152                 if ( $offset === 0 ) {
 153                         // Only perform exact db match if offset === 0
 154                         // This is still far from perfect but at least we avoid returning the
 155                         // same title afain and again when the user is scrolling with a query
 156                         // that matches a title in the db.
 157                         $rescorer = new SearchExactMatchRescorer();
 158                         $srchres = $rescorer->rescore( $search, $namespaces, $srchres, $limit );
 159                 }
 160                 return $srchres;
 161         }
 162
 163         /**
 164          * Prefix search special-case for Special: namespace.
 165          *
 166          * @param string $search Term
 167          * @param int $limit Max number of items to return
 168          * @param int $offset Number of items to offset
 169          * @return array
 170          */
 171         protected function specialSearch( $search, $limit, $offset ) {
 172                 $searchParts = explode( '/', $search, 2 );
 173                 $searchKey = $searchParts[0];
 174                 $subpageSearch = $searchParts[1] ?? null;
 175
 176                 // Handle subpage search separately.
 177                 $spFactory = MediaWikiServices::getInstance()->getSpecialPageFactory();
 178                 if ( $subpageSearch !== null ) {
 179                         // Try matching the full search string as a page name
 180                         $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey );
 181                         if ( !$specialTitle ) {
 182                                 return [];
 183                         }
 184                         $special = $spFactory->getPage( $specialTitle->getText() );
 185                         if ( $special ) {
 186                                 $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset );
 187                                 return array_map( function ( $sub ) use ( $specialTitle ) {
 188                                         return $specialTitle->getSubpage( $sub );
 189                                 }, $subpages );
 190                         } else {
 191                                 return [];
 192                         }
 193                 }
 194
 195                 # normalize searchKey, so aliases with spaces can be found - T27675
 196                 $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 197                 $searchKey = str_replace( ' ', '_', $searchKey );
 198                 $searchKey = $contLang->caseFold( $searchKey );
 199
 200                 // Unlike SpecialPage itself, we want the canonical forms of both
 201                 // canonical and alias title forms...
 202                 $keys = [];
 203                 foreach ( $spFactory->getNames() as $page ) {
 204                         $keys[$contLang->caseFold( $page )] = [ 'page' => $page, 'rank' => 0 ];
 205                 }
 206
 207                 foreach ( $contLang->getSpecialPageAliases() as $page => $aliases ) {
 208                         if ( !in_array( $page, $spFactory->getNames() ) ) {# T22885
 209                                 continue;
 210                         }
 211
 212                         foreach ( $aliases as $key => $alias ) {
 213                                 $keys[$contLang->caseFold( $alias )] = [ 'page' => $alias, 'rank' => $key ];
 214                         }
 215                 }
 216                 ksort( $keys );
 217
 218                 $matches = [];
 219                 foreach ( $keys as $pageKey => $page ) {
 220                         if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) {
 221                                 // T29671: Don't use SpecialPage::getTitleFor() here because it
 222                                 // localizes its input leading to searches for e.g. Special:All
 223                                 // returning Spezial:MediaWiki-Systemnachrichten and returning
 224                                 // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de'
 225                                 $matches[$page['rank']][] = Title::makeTitleSafe( NS_SPECIAL, $page['page'] );
 226
 227                                 if ( isset( $matches[0] ) && count( $matches[0] ) >= $limit + $offset ) {
 228                                         // We have enough items in primary rank, no use to continue
 229                                         break;
 230                                 }
 231                         }
 232
 233                 }
 234
 235                 // Ensure keys are in order
 236                 ksort( $matches );
 237                 // Flatten the array
 238                 $matches = array_reduce( $matches, 'array_merge', [] );
 239
 240                 return array_slice( $matches, $offset, $limit );
 241         }
 242
 243         /**
 244          * Unless overridden by PrefixSearchBackend hook...
 245          * This is case-sensitive (First character may
 246          * be automatically capitalized by Title::secureAndSpit()
 247          * later on depending on $wgCapitalLinks)
 248          *
 249          * @param array|null $namespaces Namespaces to search in
 250          * @param string $search Term
 251          * @param int $limit Max number of items to return
 252          * @param int $offset Number of items to skip
 253          * @return Title[] Array of Title objects
 254          */
 255         public function defaultSearchBackend( $namespaces, $search, $limit, $offset ) {
 256                 // Backwards compatability with old code. Default to NS_MAIN if no namespaces provided.
 257                 if ( $namespaces === null ) {
 258                         $namespaces = [];
 259                 }
 260                 if ( !$namespaces ) {
 261                         $namespaces[] = NS_MAIN;
 262                 }
 263
 264                 // Construct suitable prefix for each namespace. They differ in cases where
 265                 // some namespaces always capitalize and some don't.
 266                 $prefixes = [];
 267                 foreach ( $namespaces as $namespace ) {
 268                         // For now, if special is included, ignore the other namespaces
 269                         if ( $namespace == NS_SPECIAL ) {
 270                                 return $this->specialSearch( $search, $limit, $offset );
 271                         }
 272
 273                         $title = Title::makeTitleSafe( $namespace, $search );
 274                         // Why does the prefix default to empty?
 275                         $prefix = $title ? $title->getDBkey() : '';
 276                         $prefixes[$prefix][] = $namespace;
 277                 }
 278
 279                 $dbr = wfGetDB( DB_REPLICA );
 280                 // Often there is only one prefix that applies to all requested namespaces,
 281                 // but sometimes there are two if some namespaces do not always capitalize.
 282                 $conds = [];
 283                 foreach ( $prefixes as $prefix => $namespaces ) {
 284                         $condition = [
 285                                 'page_namespace' => $namespaces,
 286                                 'page_title' . $dbr->buildLike( $prefix, $dbr->anyString() ),
 287                         ];
 288                         $conds[] = $dbr->makeList( $condition, LIST_AND );
 289                 }
 290
 291                 $table = 'page';
 292                 $fields = [ 'page_id', 'page_namespace', 'page_title' ];
 293                 $conds = $dbr->makeList( $conds, LIST_OR );
 294                 $options = [
 295                         'LIMIT' => $limit,
 296                         'ORDER BY' => [ 'page_title', 'page_namespace' ],
 297                         'OFFSET' => $offset
 298                 ];
 299
 300                 $res = $dbr->select( $table, $fields, $conds, __METHOD__, $options );
 301
 302                 return iterator_to_array( TitleArray::newFromResult( $res ) );
 303         }
 304
 305         /**
 306          * Validate an array of numerical namespace indexes
 307          *
 308          * @param array $namespaces
 309          * @return array (default: contains only NS_MAIN)
 310          */
 311         protected function validateNamespaces( $namespaces ) {
 312                 // We will look at each given namespace against content language namespaces
 313                 $validNamespaces = MediaWikiServices::getInstance()->getContentLanguage()->getNamespaces();
 314                 if ( is_array( $namespaces ) && count( $namespaces ) > 0 ) {
 315                         $valid = [];
 316                         foreach ( $namespaces as $ns ) {
 317                                 if ( is_numeric( $ns ) && array_key_exists( $ns, $validNamespaces ) ) {
 318                                         $valid[] = $ns;
 319                                 }
 320                         }
 321                         if ( count( $valid ) > 0 ) {
 322                                 return $valid;
 323                         }
 324                 }
 325
 326                 return [ NS_MAIN ];
 327         }
 328 }