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  * @internal documentation reviewed on 18 Mar 2011 by hashar
  43  */
  44 class BacklinkCache {
  45         /** @var ProcessCacheLRU */
  46         protected static $cache;
  47
  48         /**
  49          * Multi dimensions array representing batches. Keys are:
  50          *  > (string) links table name
  51          *   > (int) batch size
  52          *    > 'numRows' : Number of rows for this link table
  53          *    > 'batches' : array( $start, $end )
  54          *
  55          * @see BacklinkCache::partitionResult()
  56          *
  57          * Cleared with BacklinkCache::clear()
  58          */
  59         protected $partitionCache = array();
  60
  61         /**
  62          * Contains the whole links from a database result.
  63          * This is raw data that will be partitioned in $partitionCache
  64          *
  65          * Initialized with BacklinkCache::getLinks()
  66          * Cleared with BacklinkCache::clear()
  67          */
  68         protected $fullResultCache = array();
  69
  70         /**
  71          * Local copy of a database object.
  72          *
  73          * Accessor: BacklinkCache::getDB()
  74          * Mutator : BacklinkCache::setDB()
  75          * Cleared with BacklinkCache::clear()
  76          */
  77         protected $db;
  78
  79         /**
  80          * Local copy of a Title object
  81          */
  82         protected $title;
  83
  84         const CACHE_EXPIRY = 3600;
  85
  86         /**
  87          * Create a new BacklinkCache
  88          *
  89          * @param Title $title : Title object to create a backlink cache for
  90          */
  91         public function __construct( Title $title ) {
  92                 $this->title = $title;
  93         }
  94
  95         /**
  96          * Create a new BacklinkCache or reuse any existing one.
  97          * Currently, only one cache instance can exist; callers that
  98          * need multiple backlink cache objects should keep them in scope.
  99          *
 100          * @param Title $title : Title object to get a backlink cache for
 101          * @return BacklinkCache
 102          */
 103         public static function get( Title $title ) {
 104                 if ( !self::$cache ) { // init cache
 105                         self::$cache = new ProcessCacheLRU( 1 );
 106                 }
 107                 $dbKey = $title->getPrefixedDBkey();
 108                 if ( !self::$cache->has( $dbKey, 'obj', 3600 ) ) {
 109                         self::$cache->set( $dbKey, 'obj', new self( $title ) );
 110                 }
 111                 return self::$cache->get( $dbKey, 'obj' );
 112         }
 113
 114         /**
 115          * Serialization handler, diasallows to serialize the database to prevent
 116          * failures after this class is deserialized from cache with dead DB
 117          * connection.
 118          *
 119          * @return array
 120          */
 121         function __sleep() {
 122                 return array( 'partitionCache', 'fullResultCache', 'title' );
 123         }
 124
 125         /**
 126          * Clear locally stored data and database object.
 127          */
 128         public function clear() {
 129                 $this->partitionCache = array();
 130                 $this->fullResultCache = array();
 131                 unset( $this->db );
 132         }
 133
 134         /**
 135          * Set the Database object to use
 136          *
 137          * @param $db DatabaseBase
 138          */
 139         public function setDB( $db ) {
 140                 $this->db = $db;
 141         }
 142
 143         /**
 144          * Get the slave connection to the database
 145          * When non existing, will initialize the connection.
 146          * @return DatabaseBase object
 147          */
 148         protected function getDB() {
 149                 if ( !isset( $this->db ) ) {
 150                         $this->db = wfGetDB( DB_SLAVE );
 151                 }
 152                 return $this->db;
 153         }
 154
 155         /**
 156          * Get the backlinks for a given table. Cached in process memory only.
 157          * @param $table String
 158          * @param $startId Integer|false
 159          * @param $endId Integer|false
 160          * @param $max Integer|INF
 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 $table String
 170          * @param $startId Integer|false
 171          * @param $endId Integer|false
 172          * @param $max Integer|INF
 173          * @param $select string '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' ), $options )
 221                                 );
 222                         }
 223
 224                         if ( !$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                 return $res;
 234         }
 235
 236         /**
 237          * Get the field name prefix for a given table
 238          * @param $table String
 239          * @throws MWException
 240          * @return null|string
 241          */
 242         protected function getPrefix( $table ) {
 243                 static $prefixes = array(
 244                         'pagelinks' => 'pl',
 245                         'imagelinks' => 'il',
 246                         'categorylinks' => 'cl',
 247                         'templatelinks' => 'tl',
 248                         'redirect' => 'rd',
 249                 );
 250
 251                 if ( isset( $prefixes[$table] ) ) {
 252                         return $prefixes[$table];
 253                 } else {
 254                         $prefix = null;
 255                         wfRunHooks( 'BacklinkCacheGetPrefix', array( $table, &$prefix ) );
 256                         if ( $prefix ) {
 257                                 return $prefix;
 258                         } else {
 259                                 throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 260                         }
 261                 }
 262         }
 263
 264         /**
 265          * Get the SQL condition array for selecting backlinks, with a join
 266          * on the page table.
 267          * @param $table String
 268          * @throws MWException
 269          * @return array|null
 270          */
 271         protected function getConditions( $table ) {
 272                 $prefix = $this->getPrefix( $table );
 273
 274                 // @todo FIXME: imagelinks and categorylinks do not rely on getNamespace,
 275                 // they could be moved up for nicer case statements
 276                 switch ( $table ) {
 277                         case 'pagelinks':
 278                         case 'templatelinks':
 279                                 $conds = array(
 280                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 281                                         "{$prefix}_title" => $this->title->getDBkey(),
 282                                         "page_id={$prefix}_from"
 283                                 );
 284                                 break;
 285                         case 'redirect':
 286                                 $conds = array(
 287                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 288                                         "{$prefix}_title" => $this->title->getDBkey(),
 289                                         $this->getDb()->makeList( array(
 290                                                 "{$prefix}_interwiki" => '',
 291                                                 "{$prefix}_interwiki IS NULL",
 292                                         ), LIST_OR ),
 293                                         "page_id={$prefix}_from"
 294                                 );
 295                                 break;
 296                         case 'imagelinks':
 297                                 $conds = array(
 298                                         'il_to' => $this->title->getDBkey(),
 299                                         "page_id={$prefix}_from"
 300                                 );
 301                                 break;
 302                         case 'categorylinks':
 303                                 $conds = array(
 304                                         'cl_to' => $this->title->getDBkey(),
 305                                         "page_id={$prefix}_from"
 306                                 );
 307                                 break;
 308                         default:
 309                                 $conds = null;
 310                                 wfRunHooks( 'BacklinkCacheGetConditions', array( $table, $this->title, &$conds ) );
 311                                 if ( !$conds ) {
 312                                         throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 313                                 }
 314                 }
 315
 316                 return $conds;
 317         }
 318
 319         /**
 320          * Check if there are any backlinks
 321          * @param $table String
 322          * @return bool
 323          */
 324         public function hasLinks( $table ) {
 325                 return ( $this->getNumLinks( $table, 1 ) > 0 );
 326         }
 327
 328         /**
 329          * Get the approximate number of backlinks
 330          * @param $table String
 331          * @param $max integer|INF Only count up to this many backlinks
 332          * @return integer
 333          */
 334         public function getNumLinks( $table, $max = INF ) {
 335                 global $wgMemc, $wgUpdateRowsPerJob;
 336
 337                 // 1) try partition cache ...
 338                 if ( isset( $this->partitionCache[$table] ) ) {
 339                         $entry = reset( $this->partitionCache[$table] );
 340                         return min( $max, $entry['numRows'] );
 341                 }
 342
 343                 // 2) ... then try full result cache ...
 344                 if ( isset( $this->fullResultCache[$table] ) ) {
 345                         return min( $max, $this->fullResultCache[$table]->numRows() );
 346                 }
 347
 348                 $memcKey = wfMemcKey( 'numbacklinks', md5( $this->title->getPrefixedDBkey() ), $table );
 349
 350                 // 3) ... fallback to memcached ...
 351                 $count = $wgMemc->get( $memcKey );
 352                 if ( $count ) {
 353                         return min( $max, $count );
 354                 }
 355
 356                 // 4) fetch from the database ...
 357                 if ( is_infinite( $max ) ) { // no limit at all
 358                         // Use partition() since it will batch the query and skip the JOIN.
 359                         // Use $wgUpdateRowsPerJob just to encourage cache reuse for jobs.
 360                         $this->partition( $table, $wgUpdateRowsPerJob ); // updates $this->partitionCache
 361                         return $this->partitionCache[$table][$wgUpdateRowsPerJob]['numRows'];
 362                 } else { // probably some sane limit
 363                         // Fetch the full title info, since the caller will likely need it next
 364                         $count = $this->getLinks( $table, false, false, $max )->count();
 365                         if ( $count < $max ) { // full count
 366                                 $wgMemc->set( $memcKey, $count, self::CACHE_EXPIRY );
 367                         }
 368                 }
 369
 370                 return min( $max, $count );
 371         }
 372
 373         /**
 374          * Partition the backlinks into batches.
 375          * Returns an array giving the start and end of each range. The first
 376          * batch has a start of false, and the last batch has an end of false.
 377          *
 378          * @param string $table the links table name
 379          * @param $batchSize Integer
 380          * @return Array
 381          */
 382         public function partition( $table, $batchSize ) {
 383                 global $wgMemc;
 384
 385                 // 1) try partition cache ...
 386                 if ( isset( $this->partitionCache[$table][$batchSize] ) ) {
 387                         wfDebug( __METHOD__ . ": got from partition cache\n" );
 388                         return $this->partitionCache[$table][$batchSize]['batches'];
 389                 }
 390
 391                 $this->partitionCache[$table][$batchSize] = false;
 392                 $cacheEntry =& $this->partitionCache[$table][$batchSize];
 393
 394                 // 2) ... then try full result cache ...
 395                 if ( isset( $this->fullResultCache[$table] ) ) {
 396                         $cacheEntry = $this->partitionResult( $this->fullResultCache[$table], $batchSize );
 397                         wfDebug( __METHOD__ . ": got from full result cache\n" );
 398                         return $cacheEntry['batches'];
 399                 }
 400
 401                 $memcKey = wfMemcKey(
 402                         'backlinks',
 403                         md5( $this->title->getPrefixedDBkey() ),
 404                         $table,
 405                         $batchSize
 406                 );
 407
 408                 // 3) ... fallback to memcached ...
 409                 $memcValue = $wgMemc->get( $memcKey );
 410                 if ( is_array( $memcValue ) ) {
 411                         $cacheEntry = $memcValue;
 412                         wfDebug( __METHOD__ . ": got from memcached $memcKey\n" );
 413                         return $cacheEntry['batches'];
 414                 }
 415
 416                 // 4) ... finally fetch from the slow database :(
 417                 $cacheEntry = array( 'numRows' => 0, 'batches' => array() ); // final result
 418                 // Do the selects in batches to avoid client-side OOMs (bug 43452).
 419                 // Use a LIMIT that plays well with $batchSize to keep equal sized partitions.
 420                 $selectSize = max( $batchSize, 200000 - ( 200000 % $batchSize ) );
 421                 $start = false;
 422                 do {
 423                         $res = $this->queryLinks( $table, $start, false, $selectSize, 'ids' );
 424                         $partitions = $this->partitionResult( $res, $batchSize, false );
 425                         // Merge the link count and range partitions for this chunk
 426                         $cacheEntry['numRows'] += $partitions['numRows'];
 427                         $cacheEntry['batches'] = array_merge( $cacheEntry['batches'], $partitions['batches'] );
 428                         if ( count( $partitions['batches'] ) ) {
 429                                 list( $lStart, $lEnd ) = end( $partitions['batches'] );
 430                                 $start = $lEnd + 1; // pick up after this inclusive range
 431                         }
 432                 } while ( $partitions['numRows'] >= $selectSize );
 433                 // Make sure the first range has start=false and the last one has end=false
 434                 if ( count( $cacheEntry['batches'] ) ) {
 435                         $cacheEntry['batches'][0][0] = false;
 436                         $cacheEntry['batches'][count( $cacheEntry['batches'] ) - 1][1] = false;
 437                 }
 438
 439                 // Save partitions to memcached
 440                 $wgMemc->set( $memcKey, $cacheEntry, self::CACHE_EXPIRY );
 441
 442                 // Save backlink count to memcached
 443                 $memcKey = wfMemcKey( 'numbacklinks', md5( $this->title->getPrefixedDBkey() ), $table );
 444                 $wgMemc->set( $memcKey, $cacheEntry['numRows'], self::CACHE_EXPIRY );
 445
 446                 wfDebug( __METHOD__ . ": got from database\n" );
 447                 return $cacheEntry['batches'];
 448         }
 449
 450         /**
 451          * Partition a DB result with backlinks in it into batches
 452          * @param $res ResultWrapper database result
 453          * @param $batchSize integer
 454          * @param $isComplete bool Whether $res includes all the backlinks
 455          * @throws MWException
 456          * @return array
 457          */
 458         protected function partitionResult( $res, $batchSize, $isComplete = true ) {
 459                 $batches = array();
 460                 $numRows = $res->numRows();
 461                 $numBatches = ceil( $numRows / $batchSize );
 462
 463                 for ( $i = 0; $i < $numBatches; $i++ ) {
 464                         if ( $i == 0 && $isComplete ) {
 465                                 $start = false;
 466                         } else {
 467                                 $rowNum = $i * $batchSize;
 468                                 $res->seek( $rowNum );
 469                                 $row = $res->fetchObject();
 470                                 $start = (int)$row->page_id;
 471                         }
 472
 473                         if ( $i == ( $numBatches - 1 ) && $isComplete ) {
 474                                 $end = false;
 475                         } else {
 476                                 $rowNum = min( $numRows - 1, ( $i + 1 ) * $batchSize - 1 );
 477                                 $res->seek( $rowNum );
 478                                 $row = $res->fetchObject();
 479                                 $end = (int)$row->page_id;
 480                         }
 481
 482                         # Sanity check order
 483                         if ( $start && $end && $start > $end ) {
 484                                 throw new MWException( __METHOD__ . ': Internal error: query result out of order' );
 485                         }
 486
 487                         $batches[] = array( $start, $end );
 488                 }
 489
 490                 return array( 'numRows' => $numRows, 'batches' => $batches );
 491         }
 492 }