includes/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  * @copyright © 2009, Tim Starling, Domas Mituzas
  24  * @copyright © 2010, Max Sem
  25  * @copyright © 2011, Antoine Musso
  26  */
  27
  28 /**
  29  * Class for fetching backlink lists, approximate backlink counts and
  30  * partitions. This is a shared cache.
  31  *
  32  * Instances of this class should typically be fetched with the method
  33  * $title->getBacklinkCache().
  34  *
  35  * Ideally you should only get your backlinks from here when you think
  36  * there is some advantage in caching them. Otherwise it's just a waste
  37  * of memory.
  38  *
  39  * Introduced by r47317
  40  *
  41  * @internal documentation reviewed on 18 Mar 2011 by hashar
  42  */
  43 class BacklinkCache {
  44         /** @var ProcessCacheLRU */
  45         protected static $cache;
  46
  47         /**
  48          * Multi dimensions array representing batches. Keys are:
  49          *  > (string) links table name
  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          *
  96          * @param Title $title : Title object to get a backlink cache for
  97          * @return BacklinkCache
  98          */
  99         public static function get( Title $title ) {
 100                 if ( !self::$cache ) { // init cache
 101                         self::$cache = new ProcessCacheLRU( 2 );
 102                 }
 103                 $dbKey = $title->getPrefixedDBkey();
 104                 if ( !self::$cache->has( $dbKey, 'obj' ) ) {
 105                         self::$cache->set( $dbKey, 'obj', new self( $title ) );
 106                 }
 107                 return self::$cache->get( $dbKey, 'obj' );
 108         }
 109
 110         /**
 111          * Serialization handler, diasallows to serialize the database to prevent
 112          * failures after this class is deserialized from cache with dead DB
 113          * connection.
 114          *
 115          * @return array
 116          */
 117         function __sleep() {
 118                 return array( 'partitionCache', 'fullResultCache', 'title' );
 119         }
 120
 121         /**
 122          * Clear locally stored data and database object.
 123          */
 124         public function clear() {
 125                 $this->partitionCache = array();
 126                 $this->fullResultCache = array();
 127                 unset( $this->db );
 128         }
 129
 130         /**
 131          * Set the Database object to use
 132          *
 133          * @param $db DatabaseBase
 134          */
 135         public function setDB( $db ) {
 136                 $this->db = $db;
 137         }
 138
 139         /**
 140          * Get the slave connection to the database
 141          * When non existing, will initialize the connection.
 142          * @return DatabaseBase object
 143          */
 144         protected function getDB() {
 145                 if ( !isset( $this->db ) ) {
 146                         $this->db = wfGetDB( DB_SLAVE );
 147                 }
 148
 149                 return $this->db;
 150         }
 151
 152         /**
 153          * Get the backlinks for a given table. Cached in process memory only.
 154          * @param $table String
 155          * @param $startId Integer or false
 156          * @param $endId Integer or false
 157          * @return TitleArrayFromResult
 158          */
 159         public function getLinks( $table, $startId = false, $endId = false ) {
 160                 wfProfileIn( __METHOD__ );
 161
 162                 $fromField = $this->getPrefix( $table ) . '_from';
 163
 164                 if ( $startId || $endId ) {
 165                         // Partial range, not cached
 166                         wfDebug( __METHOD__ . ": from DB (uncacheable range)\n" );
 167                         $conds = $this->getConditions( $table );
 168
 169                         // Use the from field in the condition rather than the joined page_id,
 170                         // because databases are stupid and don't necessarily propagate indexes.
 171                         if ( $startId ) {
 172                                 $conds[] = "$fromField >= " . intval( $startId );
 173                         }
 174
 175                         if ( $endId ) {
 176                                 $conds[] = "$fromField <= " . intval( $endId );
 177                         }
 178
 179                         $res = $this->getDB()->select(
 180                                 array( $table, 'page' ),
 181                                 array( 'page_namespace', 'page_title', 'page_id' ),
 182                                 $conds,
 183                                 __METHOD__,
 184                                 array(
 185                                         'STRAIGHT_JOIN',
 186                                         'ORDER BY' => $fromField
 187                                 ) );
 188                         $ta = TitleArray::newFromResult( $res );
 189
 190                         wfProfileOut( __METHOD__ );
 191                         return $ta;
 192                 }
 193
 194                 // @todo FIXME: Make this a function?
 195                 if ( !isset( $this->fullResultCache[$table] ) ) {
 196                         wfDebug( __METHOD__ . ": from DB\n" );
 197                         $res = $this->getDB()->select(
 198                                 array( $table, 'page' ),
 199                                 array( 'page_namespace', 'page_title', 'page_id' ),
 200                                 $this->getConditions( $table ),
 201                                 __METHOD__,
 202                                 array(
 203                                         'STRAIGHT_JOIN',
 204                                         'ORDER BY' => $fromField,
 205                                 ) );
 206                         $this->fullResultCache[$table] = $res;
 207                 }
 208
 209                 $ta = TitleArray::newFromResult( $this->fullResultCache[$table] );
 210
 211                 wfProfileOut( __METHOD__ );
 212                 return $ta;
 213         }
 214
 215         /**
 216          * Get the field name prefix for a given table
 217          * @param $table String
 218          * @return null|string
 219          */
 220         protected function getPrefix( $table ) {
 221                 static $prefixes = array(
 222                         'pagelinks'     => 'pl',
 223                         'imagelinks'    => 'il',
 224                         'categorylinks' => 'cl',
 225                         'templatelinks' => 'tl',
 226                         'redirect'      => 'rd',
 227                 );
 228
 229                 if ( isset( $prefixes[$table] ) ) {
 230                         return $prefixes[$table];
 231                 } else {
 232                         $prefix = null;
 233                         wfRunHooks( 'BacklinkCacheGetPrefix', array( $table, &$prefix ) );
 234                         if( $prefix ) {
 235                                 return $prefix;
 236                         } else {
 237                                 throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 238                         }
 239                 }
 240         }
 241
 242         /**
 243          * Get the SQL condition array for selecting backlinks, with a join
 244          * on the page table.
 245          * @param $table String
 246          * @return array|null
 247          */
 248         protected function getConditions( $table ) {
 249                 $prefix = $this->getPrefix( $table );
 250
 251                 // @todo FIXME: imagelinks and categorylinks do not rely on getNamespace,
 252                 // they could be moved up for nicer case statements
 253                 switch ( $table ) {
 254                         case 'pagelinks':
 255                         case 'templatelinks':
 256                                 $conds = array(
 257                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 258                                         "{$prefix}_title"     => $this->title->getDBkey(),
 259                                         "page_id={$prefix}_from"
 260                                 );
 261                                 break;
 262                         case 'redirect':
 263                                 $conds = array(
 264                                         "{$prefix}_namespace" => $this->title->getNamespace(),
 265                                         "{$prefix}_title"     => $this->title->getDBkey(),
 266                                         $this->getDb()->makeList( array(
 267                                                 "{$prefix}_interwiki = ''",
 268                                                 "{$prefix}_interwiki is null",
 269                                         ), LIST_OR ),
 270                                         "page_id={$prefix}_from"
 271                                 );
 272                                 break;
 273                         case 'imagelinks':
 274                                 $conds = array(
 275                                         'il_to' => $this->title->getDBkey(),
 276                                         'page_id=il_from'
 277                                 );
 278                                 break;
 279                         case 'categorylinks':
 280                                 $conds = array(
 281                                         'cl_to' => $this->title->getDBkey(),
 282                                         'page_id=cl_from',
 283                                 );
 284                                 break;
 285                         default:
 286                                 $conds = null;
 287                                 wfRunHooks( 'BacklinkCacheGetConditions', array( $table, $this->title, &$conds ) );
 288                                 if( !$conds )
 289                                         throw new MWException( "Invalid table \"$table\" in " . __CLASS__ );
 290                 }
 291
 292                 return $conds;
 293         }
 294
 295         /**
 296          * Get the approximate number of backlinks
 297          * @param $table String
 298          * @return integer
 299          */
 300         public function getNumLinks( $table ) {
 301                 if ( isset( $this->fullResultCache[$table] ) ) {
 302                         return $this->fullResultCache[$table]->numRows();
 303                 }
 304
 305                 if ( isset( $this->partitionCache[$table] ) ) {
 306                         $entry = reset( $this->partitionCache[$table] );
 307                         return $entry['numRows'];
 308                 }
 309
 310                 $titleArray = $this->getLinks( $table );
 311
 312                 return $titleArray->count();
 313         }
 314
 315         /**
 316          * Partition the backlinks into batches.
 317          * Returns an array giving the start and end of each range. The first
 318          * batch has a start of false, and the last batch has an end of false.
 319          *
 320          * @param $table String: the links table name
 321          * @param $batchSize Integer
 322          * @return Array
 323          */
 324         public function partition( $table, $batchSize ) {
 325
 326                 // 1) try partition cache ...
 327
 328                 if ( isset( $this->partitionCache[$table][$batchSize] ) ) {
 329                         wfDebug( __METHOD__ . ": got from partition cache\n" );
 330                         return $this->partitionCache[$table][$batchSize]['batches'];
 331                 }
 332
 333                 $this->partitionCache[$table][$batchSize] = false;
 334                 $cacheEntry =& $this->partitionCache[$table][$batchSize];
 335
 336                 // 2) ... then try full result cache ...
 337
 338                 if ( isset( $this->fullResultCache[$table] ) ) {
 339                         $cacheEntry = $this->partitionResult( $this->fullResultCache[$table], $batchSize );
 340                         wfDebug( __METHOD__ . ": got from full result cache\n" );
 341
 342                         return $cacheEntry['batches'];
 343                 }
 344
 345                 // 3) ... fallback to memcached ...
 346
 347                 global $wgMemc;
 348
 349                 $memcKey = wfMemcKey(
 350                         'backlinks',
 351                         md5( $this->title->getPrefixedDBkey() ),
 352                         $table,
 353                         $batchSize
 354                 );
 355
 356                 $memcValue = $wgMemc->get( $memcKey );
 357
 358                 if ( is_array( $memcValue ) ) {
 359                         $cacheEntry = $memcValue;
 360                         wfDebug( __METHOD__ . ": got from memcached $memcKey\n" );
 361
 362                         return $cacheEntry['batches'];
 363                 }
 364
 365
 366                 // 4) ... finally fetch from the slow database :(
 367
 368                 $this->getLinks( $table );
 369                 $cacheEntry = $this->partitionResult( $this->fullResultCache[$table], $batchSize );
 370                 // Save to memcached
 371                 $wgMemc->set( $memcKey, $cacheEntry, self::CACHE_EXPIRY );
 372
 373                 wfDebug( __METHOD__ . ": got from database\n" );
 374                 return $cacheEntry['batches'];
 375         }
 376
 377         /**
 378          * Partition a DB result with backlinks in it into batches
 379          * @param $res ResultWrapper database result
 380          * @param $batchSize integer
 381          * @return array @see
 382          */
 383         protected function partitionResult( $res, $batchSize ) {
 384                 $batches = array();
 385                 $numRows = $res->numRows();
 386                 $numBatches = ceil( $numRows / $batchSize );
 387
 388                 for ( $i = 0; $i < $numBatches; $i++ ) {
 389                         if ( $i == 0  ) {
 390                                 $start = false;
 391                         } else {
 392                                 $rowNum = intval( $numRows * $i / $numBatches );
 393                                 $res->seek( $rowNum );
 394                                 $row = $res->fetchObject();
 395                                 $start = $row->page_id;
 396                         }
 397
 398                         if ( $i == $numBatches - 1 ) {
 399                                 $end = false;
 400                         } else {
 401                                 $rowNum = intval( $numRows * ( $i + 1 ) / $numBatches );
 402                                 $res->seek( $rowNum );
 403                                 $row = $res->fetchObject();
 404                                 $end = $row->page_id - 1;
 405                         }
 406
 407                         # Sanity check order
 408                         if ( $start && $end && $start > $end ) {
 409                                 throw new MWException( __METHOD__ . ': Internal error: query result out of order' );
 410                         }
 411
 412                         $batches[] = array( $start, $end );
 413                 }
 414
 415                 return array( 'numRows' => $numRows, 'batches' => $batches );
 416         }
 417 }