Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
81.82% |
90 / 110 |
|
33.33% |
2 / 6 |
CRAP | |
0.00% |
0 / 1 |
| PrefixSearch | |
81.82% |
90 / 110 |
|
33.33% |
2 / 6 |
45.23 | |
0.00% |
0 / 1 |
| search | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
3 | |||
| searchWithVariants | |
0.00% |
0 / 15 |
|
0.00% |
0 / 1 |
20 | |||
| titles | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| strings | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| searchBackend | |
91.67% |
11 / 12 |
|
0.00% |
0 / 1 |
5.01 | |||
| handleResultFromHook | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| specialSearch | |
94.59% |
35 / 37 |
|
0.00% |
0 / 1 |
14.03 | |||
| defaultSearchBackend | |
94.29% |
33 / 35 |
|
0.00% |
0 / 1 |
9.02 | |||
| 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\HookContainer\HookRunner; |
| 24 | use MediaWiki\MediaWikiServices; |
| 25 | use MediaWiki\Title\MediaWikiTitleCodec; |
| 26 | use MediaWiki\Title\Title; |
| 27 | use Wikimedia\Rdbms\IExpression; |
| 28 | use Wikimedia\Rdbms\LikeValue; |
| 29 | use Wikimedia\Rdbms\OrExpressionGroup; |
| 30 | |
| 31 | /** |
| 32 | * Handles searching prefixes of titles and finding any page |
| 33 | * names that match. Used largely by the OpenSearch implementation. |
| 34 | * @deprecated Since 1.27, Use SearchEngine::defaultPrefixSearch or SearchEngine::completionSearch |
| 35 | * |
| 36 | * @stable to extend |
| 37 | * @ingroup Search |
| 38 | */ |
| 39 | abstract class PrefixSearch { |
| 40 | /** |
| 41 | * Do a prefix search of titles and return a list of matching page names. |
| 42 | * |
| 43 | * @param string $search |
| 44 | * @param int $limit |
| 45 | * @param array $namespaces Used if query is not explicitly prefixed |
| 46 | * @param int $offset How many results to offset from the beginning |
| 47 | * @return (Title|string)[] |
| 48 | */ |
| 49 | public function search( $search, $limit, $namespaces = [], $offset = 0 ) { |
| 50 | $search = trim( $search ); |
| 51 | if ( $search == '' ) { |
| 52 | return []; // Return empty result |
| 53 | } |
| 54 | |
| 55 | $hasNamespace = SearchEngine::parseNamespacePrefixes( $search, false, true ); |
| 56 | if ( $hasNamespace !== false ) { |
| 57 | [ $search, $namespaces ] = $hasNamespace; |
| 58 | } |
| 59 | |
| 60 | return $this->searchBackend( $namespaces, $search, $limit, $offset ); |
| 61 | } |
| 62 | |
| 63 | /** |
| 64 | * Do a prefix search for all possible variants of the prefix |
| 65 | * @param string $search |
| 66 | * @param int $limit |
| 67 | * @param array $namespaces |
| 68 | * @param int $offset How many results to offset from the beginning |
| 69 | * |
| 70 | * @return (Title|string)[] |
| 71 | */ |
| 72 | public function searchWithVariants( $search, $limit, array $namespaces, $offset = 0 ) { |
| 73 | $searches = $this->search( $search, $limit, $namespaces, $offset ); |
| 74 | |
| 75 | // if the content language has variants, try to retrieve fallback results |
| 76 | $fallbackLimit = $limit - count( $searches ); |
| 77 | if ( $fallbackLimit > 0 ) { |
| 78 | $services = MediaWikiServices::getInstance(); |
| 79 | $fallbackSearches = $services->getLanguageConverterFactory() |
| 80 | ->getLanguageConverter( $services->getContentLanguage() ) |
| 81 | ->autoConvertToAllVariants( $search ); |
| 82 | $fallbackSearches = array_diff( array_unique( $fallbackSearches ), [ $search ] ); |
| 83 | |
| 84 | foreach ( $fallbackSearches as $fbs ) { |
| 85 | $fallbackSearchResult = $this->search( $fbs, $fallbackLimit, $namespaces ); |
| 86 | $searches = array_merge( $searches, $fallbackSearchResult ); |
| 87 | $fallbackLimit -= count( $fallbackSearchResult ); |
| 88 | |
| 89 | if ( $fallbackLimit == 0 ) { |
| 90 | break; |
| 91 | } |
| 92 | } |
| 93 | } |
| 94 | return $searches; |
| 95 | } |
| 96 | |
| 97 | /** |
| 98 | * When implemented in a descendant class, receives an array of Title objects and returns |
| 99 | * either an unmodified array or an array of strings corresponding to titles passed to it. |
| 100 | * |
| 101 | * @param Title[] $titles |
| 102 | * @return (Title|string)[] |
| 103 | */ |
| 104 | abstract protected function titles( array $titles ); |
| 105 | |
| 106 | /** |
| 107 | * When implemented in a descendant class, receives an array of titles as strings and returns |
| 108 | * either an unmodified array or an array of Title objects corresponding to strings received. |
| 109 | * |
| 110 | * @param string[] $strings |
| 111 | * @return (Title|string)[] |
| 112 | */ |
| 113 | abstract protected function strings( array $strings ); |
| 114 | |
| 115 | /** |
| 116 | * Do a prefix search of titles and return a list of matching page names. |
| 117 | * @param int[] $namespaces |
| 118 | * @param string $search |
| 119 | * @param int $limit |
| 120 | * @param int $offset How many results to offset from the beginning |
| 121 | * @return (Title|string)[] |
| 122 | */ |
| 123 | protected function searchBackend( $namespaces, $search, $limit, $offset ) { |
| 124 | if ( count( $namespaces ) == 1 ) { |
| 125 | $ns = $namespaces[0]; |
| 126 | if ( $ns == NS_MEDIA ) { |
| 127 | $namespaces = [ NS_FILE ]; |
| 128 | } elseif ( $ns == NS_SPECIAL ) { |
| 129 | return $this->titles( $this->specialSearch( $search, $limit, $offset ) ); |
| 130 | } |
| 131 | } |
| 132 | $srchres = []; |
| 133 | if ( ( new HookRunner( MediaWikiServices::getInstance()->getHookContainer() ) )->onPrefixSearchBackend( |
| 134 | $namespaces, $search, $limit, $srchres, $offset ) |
| 135 | ) { |
| 136 | return $this->titles( $this->defaultSearchBackend( $namespaces, $search, $limit, $offset ) ); |
| 137 | } |
| 138 | return $this->strings( |
| 139 | $this->handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) ); |
| 140 | } |
| 141 | |
| 142 | private function handleResultFromHook( $srchres, $namespaces, $search, $limit, $offset ) { |
| 143 | if ( $offset === 0 ) { |
| 144 | // Only perform exact db match if offset === 0 |
| 145 | // This is still far from perfect but at least we avoid returning the |
| 146 | // same title again and again when the user is scrolling with a query |
| 147 | // that matches a title in the db. |
| 148 | $rescorer = new SearchExactMatchRescorer(); |
| 149 | $srchres = $rescorer->rescore( $search, $namespaces, $srchres, $limit ); |
| 150 | } |
| 151 | return $srchres; |
| 152 | } |
| 153 | |
| 154 | /** |
| 155 | * Prefix search special-case for Special: namespace. |
| 156 | * |
| 157 | * @param string $search Term |
| 158 | * @param int $limit Max number of items to return |
| 159 | * @param int $offset Number of items to offset |
| 160 | * @return array |
| 161 | */ |
| 162 | protected function specialSearch( $search, $limit, $offset ) { |
| 163 | $searchParts = explode( '/', $search, 2 ); |
| 164 | $searchKey = $searchParts[0]; |
| 165 | $subpageSearch = $searchParts[1] ?? null; |
| 166 | |
| 167 | // Handle subpage search separately. |
| 168 | $spFactory = MediaWikiServices::getInstance()->getSpecialPageFactory(); |
| 169 | if ( $subpageSearch !== null ) { |
| 170 | // Try matching the full search string as a page name |
| 171 | $specialTitle = Title::makeTitleSafe( NS_SPECIAL, $searchKey ); |
| 172 | if ( !$specialTitle ) { |
| 173 | return []; |
| 174 | } |
| 175 | $special = $spFactory->getPage( $specialTitle->getText() ); |
| 176 | if ( $special ) { |
| 177 | $subpages = $special->prefixSearchSubpages( $subpageSearch, $limit, $offset ); |
| 178 | return array_map( [ $specialTitle, 'getSubpage' ], $subpages ); |
| 179 | } else { |
| 180 | return []; |
| 181 | } |
| 182 | } |
| 183 | |
| 184 | # normalize searchKey, so aliases with spaces can be found - T27675 |
| 185 | $contLang = MediaWikiServices::getInstance()->getContentLanguage(); |
| 186 | $searchKey = str_replace( ' ', '_', $searchKey ); |
| 187 | $searchKey = $contLang->caseFold( $searchKey ); |
| 188 | |
| 189 | // Unlike SpecialPage itself, we want the canonical forms of both |
| 190 | // canonical and alias title forms... |
| 191 | $keys = []; |
| 192 | $listedPages = $spFactory->getListedPages(); |
| 193 | foreach ( $listedPages as $page => $_obj ) { |
| 194 | $keys[$contLang->caseFold( $page )] = [ 'page' => $page, 'rank' => 0 ]; |
| 195 | } |
| 196 | |
| 197 | foreach ( $contLang->getSpecialPageAliases() as $page => $aliases ) { |
| 198 | // Exclude localisation aliases for pages that are not defined (T22885), |
| 199 | // e.g. if an extension registers a page based on site configuration. |
| 200 | if ( !in_array( $page, $spFactory->getNames() ) ) { |
| 201 | continue; |
| 202 | } |
| 203 | // Exclude aliases for unlisted pages |
| 204 | if ( !isset( $listedPages[ $page ] ) ) { |
| 205 | continue; |
| 206 | } |
| 207 | |
| 208 | foreach ( $aliases as $key => $alias ) { |
| 209 | $keys[$contLang->caseFold( $alias )] = [ 'page' => $alias, 'rank' => $key ]; |
| 210 | } |
| 211 | } |
| 212 | ksort( $keys ); |
| 213 | |
| 214 | $matches = []; |
| 215 | foreach ( $keys as $pageKey => $page ) { |
| 216 | if ( $searchKey === '' || strpos( $pageKey, $searchKey ) === 0 ) { |
| 217 | // T29671: Don't use SpecialPage::getTitleFor() here because it |
| 218 | // localizes its input leading to searches for e.g. Special:All |
| 219 | // returning Spezial:MediaWiki-Systemnachrichten and returning |
| 220 | // Spezial:Alle_Seiten twice when $wgLanguageCode == 'de' |
| 221 | $matches[$page['rank']][] = Title::makeTitleSafe( NS_SPECIAL, $page['page'] ); |
| 222 | |
| 223 | if ( isset( $matches[0] ) && count( $matches[0] ) >= $limit + $offset ) { |
| 224 | // We have enough items in primary rank, no use to continue |
| 225 | break; |
| 226 | } |
| 227 | } |
| 228 | |
| 229 | } |
| 230 | |
| 231 | // Ensure keys are in order |
| 232 | ksort( $matches ); |
| 233 | // Flatten the array |
| 234 | $matches = array_reduce( $matches, 'array_merge', [] ); |
| 235 | |
| 236 | return array_slice( $matches, $offset, $limit ); |
| 237 | } |
| 238 | |
| 239 | /** |
| 240 | * Unless overridden by PrefixSearchBackend hook... |
| 241 | * This is case-sensitive (First character may |
| 242 | * be automatically capitalized by Title::secureAndSpit() |
| 243 | * later on depending on $wgCapitalLinks) |
| 244 | * |
| 245 | * @param int[]|null $namespaces Namespaces to search in |
| 246 | * @param string $search Term |
| 247 | * @param int $limit Max number of items to return |
| 248 | * @param int $offset Number of items to skip |
| 249 | * @return Title[] |
| 250 | */ |
| 251 | public function defaultSearchBackend( $namespaces, $search, $limit, $offset ) { |
| 252 | if ( !$namespaces ) { |
| 253 | $namespaces = [ NS_MAIN ]; |
| 254 | } |
| 255 | |
| 256 | if ( in_array( NS_SPECIAL, $namespaces ) ) { |
| 257 | // For now, if special is included, ignore the other namespaces |
| 258 | return $this->specialSearch( $search, $limit, $offset ); |
| 259 | } |
| 260 | |
| 261 | // Construct suitable prefix for each namespace. They differ in cases where |
| 262 | // some namespaces always capitalize and some don't. |
| 263 | $prefixes = []; |
| 264 | // Allow to do a prefix search for e.g. "Talk:" |
| 265 | if ( $search === '' ) { |
| 266 | $prefixes[$search] = $namespaces; |
| 267 | } else { |
| 268 | // Don't just ignore input like "[[Foo]]", but try to search for "Foo" |
| 269 | $search = preg_replace( MediaWikiTitleCodec::getTitleInvalidRegex(), '', $search ); |
| 270 | foreach ( $namespaces as $namespace ) { |
| 271 | $title = Title::makeTitleSafe( $namespace, $search ); |
| 272 | if ( $title ) { |
| 273 | $prefixes[ $title->getDBkey() ][] = $namespace; |
| 274 | } |
| 275 | } |
| 276 | } |
| 277 | if ( !$prefixes ) { |
| 278 | return []; |
| 279 | } |
| 280 | |
| 281 | $services = MediaWikiServices::getInstance(); |
| 282 | $dbr = $services->getConnectionProvider()->getReplicaDatabase(); |
| 283 | // Often there is only one prefix that applies to all requested namespaces, |
| 284 | // but sometimes there are two if some namespaces do not always capitalize. |
| 285 | $conds = []; |
| 286 | foreach ( $prefixes as $prefix => $namespaces ) { |
| 287 | $expr = $dbr->expr( 'page_namespace', '=', $namespaces ); |
| 288 | if ( $prefix !== '' ) { |
| 289 | $expr = $expr->and( |
| 290 | 'page_title', |
| 291 | IExpression::LIKE, |
| 292 | new LikeValue( $prefix, $dbr->anyString() ) |
| 293 | ); |
| 294 | } |
| 295 | $conds[] = $expr; |
| 296 | } |
| 297 | |
| 298 | $queryBuilder = $dbr->newSelectQueryBuilder() |
| 299 | ->select( [ 'page_id', 'page_namespace', 'page_title' ] ) |
| 300 | ->from( 'page' ) |
| 301 | ->where( new OrExpressionGroup( ...$conds ) ) |
| 302 | ->orderBy( [ 'page_title', 'page_namespace' ] ) |
| 303 | ->limit( $limit ) |
| 304 | ->offset( $offset ); |
| 305 | $res = $queryBuilder->caller( __METHOD__ )->fetchResultSet(); |
| 306 | |
| 307 | return iterator_to_array( $services->getTitleFactory()->newTitleArrayFromResult( $res ) ); |
| 308 | } |
| 309 | } |