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