includes/cache/LinkCache.php

   1 <?php
   2 /**
   3  * Page existence cache.
   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  * @ingroup Cache
  22  */
  23
  24 use Wikimedia\Rdbms\Database;
  25 use Wikimedia\Rdbms\IDatabase;
  26 use MediaWiki\Linker\LinkTarget;
  27 use MediaWiki\MediaWikiServices;
  28
  29 /**
  30  * Cache for article titles (prefixed DB keys) and ids linked from one source
  31  *
  32  * @ingroup Cache
  33  */
  34 class LinkCache {
  35         /** @var MapCacheLRU */
  36         private $goodLinks;
  37         /** @var MapCacheLRU */
  38         private $badLinks;
  39         /** @var WANObjectCache */
  40         private $wanCache;
  41
  42         /** @var bool */
  43         private $mForUpdate = false;
  44
  45         /** @var TitleFormatter */
  46         private $titleFormatter;
  47
  48         /** @var NamespaceInfo */
  49         private $nsInfo;
  50
  51         /**
  52          * How many Titles to store. There are two caches, so the amount actually
  53          * stored in memory can be up to twice this.
  54          */
  55         const MAX_SIZE = 10000;
  56
  57         public function __construct(
  58                 TitleFormatter $titleFormatter,
  59                 WANObjectCache $cache,
  60                 NamespaceInfo $nsInfo = null
  61         ) {
  62                 if ( !$nsInfo ) {
  63                         wfDeprecated( __METHOD__ . ' with no NamespaceInfo argument', '1.34' );
  64                         $nsInfo = MediaWikiServices::getInstance()->getNamespaceInfo();
  65                 }
  66                 $this->goodLinks = new MapCacheLRU( self::MAX_SIZE );
  67                 $this->badLinks = new MapCacheLRU( self::MAX_SIZE );
  68                 $this->wanCache = $cache;
  69                 $this->titleFormatter = $titleFormatter;
  70                 $this->nsInfo = $nsInfo;
  71         }
  72
  73         /**
  74          * Get an instance of this class.
  75          *
  76          * @return LinkCache
  77          * @deprecated since 1.28, use MediaWikiServices instead
  78          */
  79         public static function singleton() {
  80                 return MediaWikiServices::getInstance()->getLinkCache();
  81         }
  82
  83         /**
  84          * General accessor to get/set whether the master DB should be used
  85          *
  86          * This used to also set the FOR UPDATE option (locking the rows read
  87          * in order to avoid link table inconsistency), which was later removed
  88          * for performance on wikis with a high edit rate.
  89          *
  90          * @param bool|null $update
  91          * @return bool
  92          */
  93         public function forUpdate( $update = null ) {
  94                 return wfSetVar( $this->mForUpdate, $update );
  95         }
  96
  97         /**
  98          * @param string $title Prefixed DB key
  99          * @return int Page ID or zero
 100          */
 101         public function getGoodLinkID( $title ) {
 102                 $info = $this->goodLinks->get( $title );
 103                 if ( !$info ) {
 104                         return 0;
 105                 }
 106                 return $info['id'];
 107         }
 108
 109         /**
 110          * Get a field of a title object from cache.
 111          * If this link is not a cached good title, it will return NULL.
 112          * @param LinkTarget $target
 113          * @param string $field ('length','redirect','revision','model')
 114          * @return string|int|null
 115          */
 116         public function getGoodLinkFieldObj( LinkTarget $target, $field ) {
 117                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 118                 $info = $this->goodLinks->get( $dbkey );
 119                 if ( !$info ) {
 120                         return null;
 121                 }
 122                 return $info[$field];
 123         }
 124
 125         /**
 126          * @param string $title Prefixed DB key
 127          * @return bool
 128          */
 129         public function isBadLink( $title ) {
 130                 // Use get() to ensure it records as used for LRU.
 131                 return $this->badLinks->has( $title );
 132         }
 133
 134         /**
 135          * Add a link for the title to the link cache
 136          *
 137          * @param int $id Page's ID
 138          * @param LinkTarget $target
 139          * @param int $len Text's length
 140          * @param int|null $redir Whether the page is a redirect
 141          * @param int $revision Latest revision's ID
 142          * @param string|null $model Latest revision's content model ID
 143          * @param string|null $lang Language code of the page, if not the content language
 144          */
 145         public function addGoodLinkObj( $id, LinkTarget $target, $len = -1, $redir = null,
 146                 $revision = 0, $model = null, $lang = null
 147         ) {
 148                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 149                 $this->goodLinks->set( $dbkey, [
 150                         'id' => (int)$id,
 151                         'length' => (int)$len,
 152                         'redirect' => (int)$redir,
 153                         'revision' => (int)$revision,
 154                         'model' => $model ? (string)$model : null,
 155                         'lang' => $lang ? (string)$lang : null,
 156                         'restrictions' => null
 157                 ] );
 158         }
 159
 160         /**
 161          * Same as above with better interface.
 162          * @since 1.19
 163          * @param LinkTarget $target
 164          * @param stdClass $row Object which has the fields page_id, page_is_redirect,
 165          *  page_latest and page_content_model
 166          */
 167         public function addGoodLinkObjFromRow( LinkTarget $target, $row ) {
 168                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 169                 $this->goodLinks->set( $dbkey, [
 170                         'id' => intval( $row->page_id ),
 171                         'length' => intval( $row->page_len ),
 172                         'redirect' => intval( $row->page_is_redirect ),
 173                         'revision' => intval( $row->page_latest ),
 174                         'model' => !empty( $row->page_content_model )
 175                                 ? strval( $row->page_content_model )
 176                                 : null,
 177                         'lang' => !empty( $row->page_lang )
 178                                 ? strval( $row->page_lang )
 179                                 : null,
 180                         'restrictions' => !empty( $row->page_restrictions )
 181                                 ? strval( $row->page_restrictions )
 182                                 : null
 183                 ] );
 184         }
 185
 186         /**
 187          * @param LinkTarget $target
 188          */
 189         public function addBadLinkObj( LinkTarget $target ) {
 190                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 191                 if ( !$this->isBadLink( $dbkey ) ) {
 192                         $this->badLinks->set( $dbkey, 1 );
 193                 }
 194         }
 195
 196         /**
 197          * @param string $title Prefixed DB key
 198          */
 199         public function clearBadLink( $title ) {
 200                 $this->badLinks->clear( $title );
 201         }
 202
 203         /**
 204          * @param LinkTarget $target
 205          */
 206         public function clearLink( LinkTarget $target ) {
 207                 $dbkey = $this->titleFormatter->getPrefixedDBkey( $target );
 208                 $this->badLinks->clear( $dbkey );
 209                 $this->goodLinks->clear( $dbkey );
 210         }
 211
 212         /**
 213          * Fields that LinkCache needs to select
 214          *
 215          * @since 1.28
 216          * @return array
 217          */
 218         public static function getSelectFields() {
 219                 global $wgContentHandlerUseDB, $wgPageLanguageUseDB;
 220
 221                 $fields = [
 222                         'page_id',
 223                         'page_len',
 224                         'page_is_redirect',
 225                         'page_latest',
 226                         'page_restrictions'
 227                 ];
 228                 if ( $wgContentHandlerUseDB ) {
 229                         $fields[] = 'page_content_model';
 230                 }
 231                 if ( $wgPageLanguageUseDB ) {
 232                         $fields[] = 'page_lang';
 233                 }
 234
 235                 return $fields;
 236         }
 237
 238         /**
 239          * Add a title to the link cache, return the page_id or zero if non-existent
 240          *
 241          * @param LinkTarget $nt LinkTarget object to add
 242          * @return int Page ID or zero
 243          */
 244         public function addLinkObj( LinkTarget $nt ) {
 245                 $key = $this->titleFormatter->getPrefixedDBkey( $nt );
 246                 if ( $this->isBadLink( $key ) || $nt->isExternal() || $nt->getNamespace() < 0 ) {
 247                         return 0;
 248                 }
 249                 $id = $this->getGoodLinkID( $key );
 250                 if ( $id != 0 ) {
 251                         return $id;
 252                 }
 253
 254                 if ( $key === '' ) {
 255                         return 0;
 256                 }
 257
 258                 // Cache template/file pages as they are less often viewed but heavily used
 259                 if ( $this->mForUpdate ) {
 260                         $row = $this->fetchPageRow( wfGetDB( DB_MASTER ), $nt );
 261                 } elseif ( $this->isCacheable( $nt ) ) {
 262                         // These pages are often transcluded heavily, so cache them
 263                         $cache = $this->wanCache;
 264                         $row = $cache->getWithSetCallback(
 265                                 $cache->makeKey( 'page', $nt->getNamespace(), sha1( $nt->getDBkey() ) ),
 266                                 $cache::TTL_DAY,
 267                                 function ( $curValue, &$ttl, array &$setOpts ) use ( $cache, $nt ) {
 268                                         $dbr = wfGetDB( DB_REPLICA );
 269                                         $setOpts += Database::getCacheSetOptions( $dbr );
 270
 271                                         $row = $this->fetchPageRow( $dbr, $nt );
 272                                         $mtime = $row ? wfTimestamp( TS_UNIX, $row->page_touched ) : false;
 273                                         $ttl = $cache->adaptiveTTL( $mtime, $ttl );
 274
 275                                         return $row;
 276                                 }
 277                         );
 278                 } else {
 279                         $row = $this->fetchPageRow( wfGetDB( DB_REPLICA ), $nt );
 280                 }
 281
 282                 if ( $row ) {
 283                         $this->addGoodLinkObjFromRow( $nt, $row );
 284                         $id = intval( $row->page_id );
 285                 } else {
 286                         $this->addBadLinkObj( $nt );
 287                         $id = 0;
 288                 }
 289
 290                 return $id;
 291         }
 292
 293         /**
 294          * @param WANObjectCache $cache
 295          * @param LinkTarget $t
 296          * @return string[]
 297          * @since 1.28
 298          */
 299         public function getMutableCacheKeys( WANObjectCache $cache, LinkTarget $t ) {
 300                 if ( $this->isCacheable( $t ) ) {
 301                         return [ $cache->makeKey( 'page', $t->getNamespace(), sha1( $t->getDBkey() ) ) ];
 302                 }
 303
 304                 return [];
 305         }
 306
 307         private function isCacheable( LinkTarget $title ) {
 308                 $ns = $title->getNamespace();
 309                 if ( in_array( $ns, [ NS_TEMPLATE, NS_FILE, NS_CATEGORY, NS_MEDIAWIKI ] ) ) {
 310                         return true;
 311                 }
 312                 // Focus on transcluded pages more than the main content
 313                 if ( $this->nsInfo->isContent( $ns ) ) {
 314                         return false;
 315                 }
 316                 // Non-talk extension namespaces (e.g. NS_MODULE)
 317                 return ( $ns >= 100 && $this->nsInfo->isSubject( $ns ) );
 318         }
 319
 320         private function fetchPageRow( IDatabase $db, LinkTarget $nt ) {
 321                 $fields = self::getSelectFields();
 322                 if ( $this->isCacheable( $nt ) ) {
 323                         $fields[] = 'page_touched';
 324                 }
 325
 326                 return $db->selectRow(
 327                         'page',
 328                         $fields,
 329                         [ 'page_namespace' => $nt->getNamespace(), 'page_title' => $nt->getDBkey() ],
 330                         __METHOD__
 331                 );
 332         }
 333
 334         /**
 335          * Purge the link cache for a title
 336          *
 337          * @param LinkTarget $title
 338          * @since 1.28
 339          */
 340         public function invalidateTitle( LinkTarget $title ) {
 341                 if ( $this->isCacheable( $title ) ) {
 342                         $cache = $this->wanCache;
 343                         $cache->delete(
 344                                 $cache->makeKey( 'page', $title->getNamespace(), sha1( $title->getDBkey() ) )
 345                         );
 346                 }
 347         }
 348
 349         /**
 350          * Clears cache
 351          */
 352         public function clear() {
 353                 $this->goodLinks->clear();
 354                 $this->badLinks->clear();
 355         }
 356 }