includes/SiteStats.php

   1 <?php
   2 /**
   3  * Accessors and mutators for the site-wide statistics.
   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 Wikimedia\Rdbms\Database;
  24 use Wikimedia\Rdbms\IDatabase;
  25 use MediaWiki\MediaWikiServices;
  26
  27 /**
  28  * Static accessor class for site_stats and related things
  29  */
  30 class SiteStats {
  31         /** @var bool|stdClass */
  32         private static $row;
  33
  34         /** @var bool */
  35         private static $loaded = false;
  36         /** @var int[] */
  37         private static $pageCount = [];
  38
  39         static function unload() {
  40                 self::$loaded = false;
  41         }
  42
  43         static function recache() {
  44                 self::load( true );
  45         }
  46
  47         /**
  48          * @param bool $recache
  49          */
  50         static function load( $recache = false ) {
  51                 if ( self::$loaded && !$recache ) {
  52                         return;
  53                 }
  54
  55                 self::$row = self::loadAndLazyInit();
  56
  57                 self::$loaded = true;
  58         }
  59
  60         /**
  61          * @return bool|stdClass
  62          */
  63         static function loadAndLazyInit() {
  64                 global $wgMiserMode;
  65
  66                 wfDebug( __METHOD__ . ": reading site_stats from replica DB\n" );
  67                 $row = self::doLoad( wfGetDB( DB_REPLICA ) );
  68
  69                 if ( !self::isSane( $row ) ) {
  70                         $lb = MediaWikiServices::getInstance()->getDBLoadBalancer();
  71                         if ( $lb->hasOrMadeRecentMasterChanges() ) {
  72                                 // Might have just been initialized during this request? Underflow?
  73                                 wfDebug( __METHOD__ . ": site_stats damaged or missing on replica DB\n" );
  74                                 $row = self::doLoad( wfGetDB( DB_MASTER ) );
  75                         }
  76                 }
  77
  78                 if ( !self::isSane( $row ) ) {
  79                         if ( $wgMiserMode ) {
  80                                 // Start off with all zeroes, assuming that this is a new wiki or any
  81                                 // repopulations where done manually via script.
  82                                 SiteStatsInit::doPlaceholderInit();
  83                         } else {
  84                                 // Normally the site_stats table is initialized at install time.
  85                                 // Some manual construction scenarios may leave the table empty or
  86                                 // broken, however, for instance when importing from a dump into a
  87                                 // clean schema with mwdumper.
  88                                 wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  89                                 SiteStatsInit::doAllAndCommit( wfGetDB( DB_REPLICA ) );
  90                         }
  91
  92                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  93                 }
  94
  95                 if ( !self::isSane( $row ) ) {
  96                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  97                         $row = (object)array_fill_keys( self::selectFields(), 0 );
  98                 }
  99
 100                 return $row;
 101         }
 102
 103         /**
 104          * @param IDatabase $db
 105          * @return bool|stdClass
 106          */
 107         static function doLoad( $db ) {
 108                 return $db->selectRow(
 109                         'site_stats',
 110                         self::selectFields(),
 111                         [ 'ss_row_id' => 1 ],
 112                         __METHOD__
 113                 );
 114         }
 115
 116         /**
 117          * Return the total number of page views. Except we don't track those anymore.
 118          * Stop calling this function, it will be removed some time in the future. It's
 119          * kept here simply to prevent fatal errors.
 120          *
 121          * @deprecated since 1.25
 122          * @return int
 123          */
 124         static function views() {
 125                 wfDeprecated( __METHOD__, '1.25' );
 126                 return 0;
 127         }
 128
 129         /**
 130          * @return int
 131          */
 132         static function edits() {
 133                 self::load();
 134                 return self::$row->ss_total_edits;
 135         }
 136
 137         /**
 138          * @return int
 139          */
 140         static function articles() {
 141                 self::load();
 142                 return self::$row->ss_good_articles;
 143         }
 144
 145         /**
 146          * @return int
 147          */
 148         static function pages() {
 149                 self::load();
 150                 return self::$row->ss_total_pages;
 151         }
 152
 153         /**
 154          * @return int
 155          */
 156         static function users() {
 157                 self::load();
 158                 return self::$row->ss_users;
 159         }
 160
 161         /**
 162          * @return int
 163          */
 164         static function activeUsers() {
 165                 self::load();
 166                 return self::$row->ss_active_users;
 167         }
 168
 169         /**
 170          * @return int
 171          */
 172         static function images() {
 173                 self::load();
 174                 return self::$row->ss_images;
 175         }
 176
 177         /**
 178          * Find the number of users in a given user group.
 179          * @param string $group Name of group
 180          * @return int
 181          */
 182         static function numberingroup( $group ) {
 183                 $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 184                 return $cache->getWithSetCallback(
 185                         $cache->makeKey( 'SiteStats', 'groupcounts', $group ),
 186                         $cache::TTL_HOUR,
 187                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 188                                 $dbr = wfGetDB( DB_REPLICA );
 189
 190                                 $setOpts += Database::getCacheSetOptions( $dbr );
 191
 192                                 return $dbr->selectField(
 193                                         'user_groups',
 194                                         'COUNT(*)',
 195                                         [
 196                                                 'ug_group' => $group,
 197                                                 'ug_expiry IS NULL OR ug_expiry >= ' . $dbr->addQuotes( $dbr->timestamp() )
 198                                         ],
 199                                         __METHOD__
 200                                 );
 201                         },
 202                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 203                 );
 204         }
 205
 206         /**
 207          * Total number of jobs in the job queue.
 208          * @return int
 209          */
 210         static function jobs() {
 211                 $cache = MediaWikiServices::getInstance()->getMainWANObjectCache();
 212                 return $cache->getWithSetCallback(
 213                         $cache->makeKey( 'SiteStats', 'jobscount' ),
 214                         $cache::TTL_MINUTE,
 215                         function ( $oldValue, &$ttl, array &$setOpts ) {
 216                                 try{
 217                                         $jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 218                                 } catch ( JobQueueError $e ) {
 219                                         $jobs = 0;
 220                                 }
 221                                 return $jobs;
 222                         },
 223                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 224                 );
 225         }
 226
 227         /**
 228          * @param int $ns
 229          *
 230          * @return int
 231          */
 232         static function pagesInNs( $ns ) {
 233                 if ( !isset( self::$pageCount[$ns] ) ) {
 234                         $dbr = wfGetDB( DB_REPLICA );
 235                         self::$pageCount[$ns] = (int)$dbr->selectField(
 236                                 'page',
 237                                 'COUNT(*)',
 238                                 [ 'page_namespace' => $ns ],
 239                                 __METHOD__
 240                         );
 241                 }
 242                 return self::$pageCount[$ns];
 243         }
 244
 245         /**
 246          * @return array
 247          */
 248         public static function selectFields() {
 249                 return [
 250                         'ss_total_edits',
 251                         'ss_good_articles',
 252                         'ss_total_pages',
 253                         'ss_users',
 254                         'ss_active_users',
 255                         'ss_images',
 256                 ];
 257         }
 258
 259         /**
 260          * Is the provided row of site stats sane, or should it be regenerated?
 261          *
 262          * Checks only fields which are filled by SiteStatsInit::refresh.
 263          *
 264          * @param bool|object $row
 265          * @return bool
 266          */
 267         private static function isSane( $row ) {
 268                 if ( $row === false
 269                         || $row->ss_total_pages < $row->ss_good_articles
 270                         || $row->ss_total_edits < $row->ss_total_pages
 271                 ) {
 272                         return false;
 273                 }
 274                 // Now check for underflow/overflow
 275                 foreach ( [
 276                         'ss_total_edits',
 277                         'ss_good_articles',
 278                         'ss_total_pages',
 279                         'ss_users',
 280                         'ss_images',
 281                 ] as $member ) {
 282                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 283                                 return false;
 284                         }
 285                 }
 286                 return true;
 287         }
 288 }
 289
 290 /**
 291  * Class designed for counting of stats.
 292  */
 293 class SiteStatsInit {
 294
 295         // Database connection
 296         private $db;
 297
 298         // Various stats
 299         private $mEdits = null, $mArticles = null, $mPages = null;
 300         private $mUsers = null, $mFiles = null;
 301
 302         /**
 303          * @param bool|IDatabase $database
 304          * - bool: Whether to use the master DB
 305          * - IDatabase: Database connection to use
 306          */
 307         public function __construct( $database = false ) {
 308                 if ( $database instanceof IDatabase ) {
 309                         $this->db = $database;
 310                 } elseif ( $database ) {
 311                         $this->db = wfGetDB( DB_MASTER );
 312                 } else {
 313                         $this->db = wfGetDB( DB_REPLICA, 'vslow' );
 314                 }
 315         }
 316
 317         /**
 318          * Count the total number of edits
 319          * @return int
 320          */
 321         public function edits() {
 322                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 323                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 324                 return $this->mEdits;
 325         }
 326
 327         /**
 328          * Count pages in article space(s)
 329          * @return int
 330          */
 331         public function articles() {
 332                 global $wgArticleCountMethod;
 333
 334                 $tables = [ 'page' ];
 335                 $conds = [
 336                         'page_namespace' => MWNamespace::getContentNamespaces(),
 337                         'page_is_redirect' => 0,
 338                 ];
 339
 340                 if ( $wgArticleCountMethod == 'link' ) {
 341                         $tables[] = 'pagelinks';
 342                         $conds[] = 'pl_from=page_id';
 343                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 344                         // To make a correct check for this, we would need, for each page,
 345                         // to load the text, maybe uncompress it, maybe decode it and then
 346                         // check if there's one comma.
 347                         // But one thing we are sure is that if the page is empty, it can't
 348                         // contain a comma :)
 349                         $conds[] = 'page_len > 0';
 350                 }
 351
 352                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 353                         $conds, __METHOD__ );
 354                 return $this->mArticles;
 355         }
 356
 357         /**
 358          * Count total pages
 359          * @return int
 360          */
 361         public function pages() {
 362                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 363                 return $this->mPages;
 364         }
 365
 366         /**
 367          * Count total users
 368          * @return int
 369          */
 370         public function users() {
 371                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 372                 return $this->mUsers;
 373         }
 374
 375         /**
 376          * Count total files
 377          * @return int
 378          */
 379         public function files() {
 380                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 381                 return $this->mFiles;
 382         }
 383
 384         /**
 385          * Do all updates and commit them. More or less a replacement
 386          * for the original initStats, but without output.
 387          *
 388          * @param IDatabase|bool $database
 389          * - bool: Whether to use the master DB
 390          * - IDatabase: Database connection to use
 391          * @param array $options Array of options, may contain the following values
 392          * - activeUsers bool: Whether to update the number of active users (default: false)
 393          */
 394         public static function doAllAndCommit( $database, array $options = [] ) {
 395                 $options += [ 'update' => false, 'activeUsers' => false ];
 396
 397                 // Grab the object and count everything
 398                 $counter = new SiteStatsInit( $database );
 399
 400                 $counter->edits();
 401                 $counter->articles();
 402                 $counter->pages();
 403                 $counter->users();
 404                 $counter->files();
 405
 406                 $counter->refresh();
 407
 408                 // Count active users if need be
 409                 if ( $options['activeUsers'] ) {
 410                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 411                 }
 412         }
 413
 414         /**
 415          * Insert a dummy row with all zeroes if no row is present
 416          */
 417         public static function doPlaceholderInit() {
 418                 $dbw = wfGetDB( DB_MASTER );
 419                 $exists = $dbw->selectField( 'site_stats', '1', [ 'ss_row_id' => 1 ],  __METHOD__ );
 420                 if ( $exists === false ) {
 421                         $dbw->insert(
 422                                 'site_stats',
 423                                 [ 'ss_row_id' => 1 ] + array_fill_keys( SiteStats::selectFields(), 0 ),
 424                                 __METHOD__,
 425                                 [ 'IGNORE' ]
 426                         );
 427                 }
 428         }
 429
 430         /**
 431          * Refresh site_stats
 432          */
 433         public function refresh() {
 434                 $values = [
 435                         'ss_row_id' => 1,
 436                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 437                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 438                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 439                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 440                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 441                 ];
 442
 443                 $dbw = wfGetDB( DB_MASTER );
 444                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 445         }
 446 }