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