includes/cache/BacklinkCache.php

   1 <?php
   2 /**
   3  * Class for fetching backlink lists, approximate backlink counts and
   4  * partitions.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License as published by
   8  * the Free Software Foundation; either version 2 of the License, or
   9  * (at your option) any later version.
  10  *
  11  * This program is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14  * GNU General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU General Public License along
  17  * with this program; if not, write to the Free Software Foundation, Inc.,
  18  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  19  * http://www.gnu.org/copyleft/gpl.html
  20  *
  21  * @file
  22  * @author Tim Starling
  23  * @author Aaron Schulz
  24  * @copyright © 2009, Tim Starling, Domas Mituzas
  25  * @copyright © 2010, Max Sem
  26  * @copyright © 2011, Antoine Musso
  27  */
  28
  29 /**
  30  * Class for fetching backlink lists, approximate backlink counts and
  31  * partitions. This is a shared cache.
  32  *
  33  * Instances of this class should typically be fetched with the method
  34  * $title->getBacklinkCache().
  35  *
  36  * Ideally you should only get your backlinks from here when you think
  37  * there is some advantage in caching them. Otherwise it's just a waste
  38  * of memory.
  39  *
  40  * Introduced by r47317
  41  */
  42 class BacklinkCache {
  43         /** @var ProcessCacheLRU */
  44         protected static $cache;
  45
  46         /**
  47          * Multi dimensions array representing batches. Keys are:
  48          *  > (string) links table name
  49          *   > (int) batch size
  50          *    > 'numRows' : Number of rows for this link table
  51          *    > 'batches' : array( $start, $end )
  52          *
  53          * @see BacklinkCache::partitionResult()
  54          *
  55          * Cleared with BacklinkCache::clear()
  56          */
  57         protected $partitionCache = array();
  58
  59         /**
  60          * Contains the whole links from a database result.
  61          * This is raw data that will be partitioned in $partitionCache
  62          *
  63          * Initialized with BacklinkCache::getLinks()
  64          * Cleared with BacklinkCache::clear()
  65          */
  66         protected $fullResultCache = array();
  67
  68         /**
  69          * Local copy of a database object.
  70          *
  71          * Accessor: BacklinkCache::getDB()
  72          * Mutator : BacklinkCache::setDB()
  73          * Cleared with BacklinkCache::clear()
  74          */
  75         protected $db;
  76
  77         /**
  78          * Local copy of a Title object
  79          */
  80         protected $title;
  81
  82         const CACHE_EXPIRY = 3600;
  83
  84         /**
  85          * Create a new BacklinkCache
  86          *
  87          * @param Title $title : Title object to create a backlink cache for
  88          */
  89         public function __construct( Title $title ) {
  90                 $this->title = $title;
  91         }
  92
  93         /**
  94          * Create a new BacklinkCache or reuse any existing one.
  95          * Currently, only one cache instance can exist; callers that
  96          * need multiple backlink cache objects should keep them in scope.
  97          *
  98          * @param Title $title Title object to get a backlink cache for
  99          * @return BacklinkCache
 100          */
 101         public static function get( Title $title ) {
 102                 if ( !self::$cache ) { // init cache
 103                         self::$cache = new ProcessCacheLRU( 1 );
 104                 }
 105                 $dbKey = $title->getPrefixedDBkey();
 106                 if ( !self::$cache->has( $dbKey, 'obj', 3600 ) ) {
 107                         self::$cache->set( $dbKey, 'obj', new self( $title ) );
 108                 }
 109
 110                 return self::$cache->get( $dbKey, 'obj' );
 111         }
 112
 113         /**
 114          * Serialization handler, diasallows to serialize the database to prevent
 115          * failures after this class is deserialized from cache with dead DB
 116          * connection.
 117          *
 118          * @return array
 119          */
 120         function __sleep() {
 121                 return array( 'partitionCache', 'fullResultCache', 'title' );
 122         }
 123
 124         /**
 125          * Clear locally stored data and database object.
 126          */
 127         public function clear() {
 128                 $this->partitionCache = array();
 129                 $this->fullResultCache = array();
 130                 unset( $this->db );
 131         }
 132
 133         /**
 134          * Set the Database object to use
 135          *
 136          * @param DatabaseBase $db
 137          */
 138         public function setDB( $db ) {
 139                 $this->db = $db;
 140         }
 141
 142         /**
 143          * Get the slave connection to the database
 144          * When non existing, will initialize the connection.
 145          * @return DatabaseBase
 146          */
 147         protected function getDB() {
 148                 if ( !isset( $this->db ) ) {
 149                         $this->db = wfGetDB( DB_SLAVE );
 150                 }
 151
 152                 return $this->db;
 153         }
 154
 155         /**
 156          * Get the backlinks for a given table. Cached in process memory only.
 157          * @param string $table
 158          * @param int|bool $startId
 159          * @param int|bool $endId
 160          * @param int|INF $max
 161          * @return TitleArrayFromResult
 162          */
 163         public function getLinks( $table, $startId = false, $endId = false, $max = INF ) {
 164                 return TitleArray::newFromResult( $this->queryLinks( $table, $startId, $endId, $max ) );
 165         }
 166
 167         /**
 168          * Get the backlinks for a given table. Cached in process memory only.
 169          * @param string $table
 170          * @param int|bool $startId
 171          * @param int|bool $endId
 172          * @param int|INF $max
 173          * @param string $select 'all' or 'ids'
 174          * @return ResultWrapper
 175          */
 176         protected function queryLinks( $table, $startId, $endId, $max, $select = 'all' ) {
 177                 wfProfileIn( __METHOD__ );
 178
 179                 $fromField = $this->getPrefix( $table ) . '_from';
 180
 181                 if ( !$startId && !$endId && is_infinite( $max )
 182                         && isset( $this->fullResultCache[$table] )
 183                 ) {
 184                         wfDebug( __METHOD__ . ": got results from cache\n" );
 185                         $res = $this->fullResultCache[$table];
 186                 } else {
 187                         wfDebug( __METHOD__ . ": got results from DB\n" );
 188                         $conds = $this->getConditions( $table );
 189                         // Use the from field in the condition rather than the joined page_id,
 190                         // because databases are stupid and don't necessarily propagate indexes.
 191                         if ( $startId ) {
 192                                 $conds[] = "$fromField >= " . intval( $startId );
 193                         }
 194                         if ( $endId ) {
 195                                 $conds[] = "$fromField <= " . intval( $endId );
 196                         }
 197                         $options = array( 'ORDER BY' => $fromField );
 198                         if ( is_finite( $max ) && $max > 0 ) {
 199                                 $options['LIMIT'] = $max;
 200                         }
 201
 202                         if ( $select === 'ids' ) {
 203                                 // Just select from the backlink table and ignore the page JOIN
 204                                 $res = $this->getDB()->select(
 205                                         $table,
 206                                         array( $this->getPrefix( $table ) . '_from AS page_id' ),
 207                                         array_filter( $conds, function ( $clause ) { // kind of janky
 208                                                 return !preg_match( '/(\b|=)page_id(\b|=)/', $clause );
 209                                         } ),
 210                                         __METHOD__,
 211                                         $options
 212                                 );
 213                         } else {
 214                                 // Select from the backlink table and JOIN with page title information
 215                                 $res = $this->getDB()->select(
 216                                         array( $table, 'page' ),
 217                                         array( 'page_namespace', 'page_title', 'page_id' ),
 218                                         $conds,
 219                                         __METHOD__,
 220                                         array_merge( array( 'STRAIGHT_JOIN' ), $options )
 221                                 );
 222                         }
 223
 224                         if ( $select === 'all' && !$startId && !$endId && $res->numRows() < $max ) {
 225                                 // The full results fit within the limit, so cache them
 226                                 $this->fullResultCache[$table] = $res;
 227                         } else {
 228                                 wfDebug( __METHOD__ . ": results from DB were uncacheable\n" );
 229                         }
 230                 }
 231
 232                 wfProfileOut( __METHOD__ );
 233
 234                 return $res;
 235         }
 236
 237         /**
 238          * Get the field name prefix for a given table
 239          * @param string $table
 240          * @throws MWException
 241          * @return null|string
 242          */
 243         protected function getPrefix( $table ) {
 244                 static $prefixes = array(
 245                         'pagelinks' => 'pl',
 246                         'imagelinks' => 'il',
 247                         'categorylinks' => 'cl',
 248                         'templatelinks' => 'tl',
 249                         'redirect' => 'rd',
 250                 );
 251
 252                 if ( isset( $prefixes[$table] ) ) {
 253                         return $prefixes[$table];
 254                 } else {
 255                         $prefix = null;
 256                         Hooks::run( 'BacklinkCacheGetPrefix', array( $table, &$prefix ) );
 257                         if ( $prefix ) {
 258                                 return $prefix;
 259                         } else {
 260                                 throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 261                         }
 262                 }
 263         }
 264
 265         /**
 266          * Get the SQL condition array for selecting backlinks, with a join
 267          * on the page table.
 268          * @param string $table
 269          * @throws MWException
 270          * @return array|null
 271          */
 272         protected function getConditions( $table ) {
 273                 $prefix = $this->getPrefix( $table );
 274
 275                 switch ( $table ) {
 276                         case 'pagelinks':
 277                         case 'templatelinks':
 278                                 $conds = array(
 279                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 280                                         "{$prefix}_title" => $this->title->getDBkey(),
 281                                         "page_id={$prefix}_from"
 282                                 );
 283                                 break;
 284                         case 'redirect':
 285                                 $conds = array(
 286                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 287                                         "{$prefix}_title" => $this->title->getDBkey(),
 288                                         $this->getDb()->makeList( array(
 289                                                 "{$prefix}_interwiki" => '',
 290                                                 "{$prefix}_interwiki IS NULL",
 291                                         ), LIST_OR ),
 292                                         "page_id={$prefix}_from"
 293                                 );
 294                                 break;
 295                         case 'imagelinks':
 296                         case 'categorylinks':
 297                                 $conds = array(
 298                                         "{$prefix}_to" => $this->title->getDBkey(),
 299                                         "page_id={$prefix}_from"
 300                                 );
 301                                 break;
 302                         default:
 303                                 $conds = null;
 304                                 Hooks::run( 'BacklinkCacheGetConditions', array( $table, $this->title, &$conds ) );
 305                                 if ( !$conds ) {
 306                                         throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 307                                 }
 308                 }
 309
 310                 return $conds;
 311         }
 312
 313         /**
 314          * Check if there are any backlinks
 315          * @param string $table
 316          * @return bool
 317          */
 318         public function hasLinks( $table ) {
 319                 return ( $this->getNumLinks( $table, 1 ) > 0 );
 320         }
 321
 322         /**
 323          * Get the approximate number of backlinks
 324          * @param string $table
 325          * @param int|INF $max Only count up to this many backlinks
 326          * @return int
 327          */
 328         public function getNumLinks( $table, $max = INF ) {
 329                 global $wgMemc, $wgUpdateRowsPerJob;
 330
 331                 // 1) try partition cache ...
 332                 if ( isset( $this->partitionCache[$table] ) ) {
 333                         $entry = reset( $this->partitionCache[$table] );
 334
 335                         return min( $max, $entry['numRows'] );
 336                 }
 337
 338                 // 2) ... then try full result cache ...
 339                 if ( isset( $this->fullResultCache[$table] ) ) {
 340                         return min( $max, $this->fullResultCache[$table]->numRows() );
 341                 }
 342
 343                 $memcKey = wfMemcKey( 'numbacklinks', md5( $this->title->getPrefixedDBkey() ), $table );
 344
 345                 // 3) ... fallback to memcached ...
 346                 $count = $wgMemc->get( $memcKey );
 347                 if ( $count ) {
 348                         return min( $max, $count );
 349                 }
 350
 351                 // 4) fetch from the database ...
 352                 if ( is_infinite( $max ) ) { // no limit at all
 353                         // Use partition() since it will batch the query and skip the JOIN.
 354                         // Use $wgUpdateRowsPerJob just to encourage cache reuse for jobs.
 355                         $this->partition( $table, $wgUpdateRowsPerJob ); // updates $this->partitionCache
 356                         return $this->partitionCache[$table][$wgUpdateRowsPerJob]['numRows'];
 357                 } else { // probably some sane limit
 358                         // Fetch the full title info, since the caller will likely need it next
 359                         $count = $this->getLinks( $table, false, false, $max )->count();
 360                         if ( $count < $max ) { // full count
 361                                 $wgMemc->set( $memcKey, $count, self::CACHE_EXPIRY );
 362                         }
 363                 }
 364
 365                 return min( $max, $count );
 366         }
 367
 368         /**
 369          * Partition the backlinks into batches.
 370          * Returns an array giving the start and end of each range. The first
 371          * batch has a start of false, and the last batch has an end of false.
 372          *
 373          * @param string $table The links table name
 374          * @param int $batchSize
 375          * @return array
 376          */
 377         public function partition( $table, $batchSize ) {
 378                 global $wgMemc;
 379
 380                 // 1) try partition cache ...
 381                 if ( isset( $this->partitionCache[$table][$batchSize] ) ) {
 382                         wfDebug( __METHOD__ . ": got from partition cache\n" );
 383
 384                         return $this->partitionCache[$table][$batchSize]['batches'];
 385                 }
 386
 387                 $this->partitionCache[$table][$batchSize] = false;
 388                 $cacheEntry =& $this->partitionCache[$table][$batchSize];
 389
 390                 // 2) ... then try full result cache ...
 391                 if ( isset( $this->fullResultCache[$table] ) ) {
 392                         $cacheEntry = $this->partitionResult( $this->fullResultCache[$table], $batchSize );
 393                         wfDebug( __METHOD__ . ": got from full result cache\n" );
 394
 395                         return $cacheEntry['batches'];
 396                 }
 397
 398                 $memcKey = wfMemcKey(
 399                         'backlinks',
 400                         md5( $this->title->getPrefixedDBkey() ),
 401                         $table,
 402                         $batchSize
 403                 );
 404
 405                 // 3) ... fallback to memcached ...
 406                 $memcValue = $wgMemc->get( $memcKey );
 407                 if ( is_array( $memcValue ) ) {
 408                         $cacheEntry = $memcValue;
 409                         wfDebug( __METHOD__ . ": got from memcached $memcKey\n" );
 410
 411                         return $cacheEntry['batches'];
 412                 }
 413
 414                 // 4) ... finally fetch from the slow database :(
 415                 $cacheEntry = array( 'numRows' => 0, 'batches' => array() ); // final result
 416                 // Do the selects in batches to avoid client-side OOMs (bug 43452).
 417                 // Use a LIMIT that plays well with $batchSize to keep equal sized partitions.
 418                 $selectSize = max( $batchSize, 200000 - ( 200000 % $batchSize ) );
 419                 $start = false;
 420                 do {
 421                         $res = $this->queryLinks( $table, $start, false, $selectSize, 'ids' );
 422                         $partitions = $this->partitionResult( $res, $batchSize, false );
 423                         // Merge the link count and range partitions for this chunk
 424                         $cacheEntry['numRows'] += $partitions['numRows'];
 425                         $cacheEntry['batches'] = array_merge( $cacheEntry['batches'], $partitions['batches'] );
 426                         if ( count( $partitions['batches'] ) ) {
 427                                 list( , $lEnd ) = end( $partitions['batches'] );
 428                                 $start = $lEnd + 1; // pick up after this inclusive range
 429                         }
 430                 } while ( $partitions['numRows'] >= $selectSize );
 431                 // Make sure the first range has start=false and the last one has end=false
 432                 if ( count( $cacheEntry['batches'] ) ) {
 433                         $cacheEntry['batches'][0][0] = false;
 434                         $cacheEntry['batches'][count( $cacheEntry['batches'] ) - 1][1] = false;
 435                 }
 436
 437                 // Save partitions to memcached
 438                 $wgMemc->set( $memcKey, $cacheEntry, self::CACHE_EXPIRY );
 439
 440                 // Save backlink count to memcached
 441                 $memcKey = wfMemcKey( 'numbacklinks', md5( $this->title->getPrefixedDBkey() ), $table );
 442                 $wgMemc->set( $memcKey, $cacheEntry['numRows'], self::CACHE_EXPIRY );
 443
 444                 wfDebug( __METHOD__ . ": got from database\n" );
 445
 446                 return $cacheEntry['batches'];
 447         }
 448
 449         /**
 450          * Partition a DB result with backlinks in it into batches
 451          * @param ResultWrapper $res Database result
 452          * @param int $batchSize
 453          * @param bool $isComplete Whether $res includes all the backlinks
 454          * @throws MWException
 455          * @return array
 456          */
 457         protected function partitionResult( $res, $batchSize, $isComplete = true ) {
 458                 $batches = array();
 459                 $numRows = $res->numRows();
 460                 $numBatches = ceil( $numRows / $batchSize );
 461
 462                 for ( $i = 0; $i < $numBatches; $i++ ) {
 463                         if ( $i == 0 && $isComplete ) {
 464                                 $start = false;
 465                         } else {
 466                                 $rowNum = $i * $batchSize;
 467                                 $res->seek( $rowNum );
 468                                 $row = $res->fetchObject();
 469                                 $start = (int)$row->page_id;
 470                         }
 471
 472                         if ( $i == ( $numBatches - 1 ) && $isComplete ) {
 473                                 $end = false;
 474                         } else {
 475                                 $rowNum = min( $numRows - 1, ( $i + 1 ) * $batchSize - 1 );
 476                                 $res->seek( $rowNum );
 477                                 $row = $res->fetchObject();
 478                                 $end = (int)$row->page_id;
 479                         }
 480
 481                         # Sanity check order
 482                         if ( $start && $end && $start > $end ) {
 483                                 throw new MWException( __METHOD__ . ': Internal error: query result out of order' );
 484                         }
 485
 486                         $batches[] = array( $start, $end );
 487                 }
 488
 489                 return array( 'numRows' => $numRows, 'batches' => $batches );
 490         }
 491 }