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