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 = ObjectCache::getMainWANInstance();
 193                 return $cache->getWithSetCallback(
 194                         wfMemcKey( 'SiteStats', 'groupcounts', $group ),
 195                         $cache::TTL_HOUR,
 196                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 197                                 global $wgDisableUserGroupExpiry;
 198                                 $dbr = wfGetDB( DB_REPLICA );
 199
 200                                 $setOpts += Database::getCacheSetOptions( $dbr );
 201
 202                                 return $dbr->selectField(
 203                                         'user_groups',
 204                                         'COUNT(*)',
 205                                         [
 206                                                 'ug_group' => $group,
 207                                                 $wgDisableUserGroupExpiry ?
 208                                                         '1' :
 209                                                         'ug_expiry IS NULL OR ug_expiry >= ' . $dbr->addQuotes( $dbr->timestamp() )
 210                                         ],
 211                                         __METHOD__
 212                                 );
 213                         },
 214                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 215                 );
 216         }
 217
 218         /**
 219          * @return int
 220          */
 221         static function jobs() {
 222                 if ( !isset( self::$jobs ) ) {
 223                         try{
 224                                 self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 225                         } catch ( JobQueueError $e ) {
 226                                 self::$jobs = 0;
 227                         }
 228                         /**
 229                          * Zero rows still do single row read for row that doesn't exist,
 230                          * but people are annoyed by that
 231                          */
 232                         if ( self::$jobs == 1 ) {
 233                                 self::$jobs = 0;
 234                         }
 235                 }
 236                 return self::$jobs;
 237         }
 238
 239         /**
 240          * @param int $ns
 241          *
 242          * @return int
 243          */
 244         static function pagesInNs( $ns ) {
 245                 if ( !isset( self::$pageCount[$ns] ) ) {
 246                         $dbr = wfGetDB( DB_REPLICA );
 247                         self::$pageCount[$ns] = (int)$dbr->selectField(
 248                                 'page',
 249                                 'COUNT(*)',
 250                                 [ 'page_namespace' => $ns ],
 251                                 __METHOD__
 252                         );
 253                 }
 254                 return self::$pageCount[$ns];
 255         }
 256
 257         /**
 258          * Is the provided row of site stats sane, or should it be regenerated?
 259          *
 260          * Checks only fields which are filled by SiteStatsInit::refresh.
 261          *
 262          * @param bool|object $row
 263          *
 264          * @return bool
 265          */
 266         private static function isSane( $row ) {
 267                 if ( $row === false
 268                         || $row->ss_total_pages < $row->ss_good_articles
 269                         || $row->ss_total_edits < $row->ss_total_pages
 270                 ) {
 271                         return false;
 272                 }
 273                 // Now check for underflow/overflow
 274                 foreach ( [
 275                         'ss_total_edits',
 276                         'ss_good_articles',
 277                         'ss_total_pages',
 278                         'ss_users',
 279                         'ss_images',
 280                 ] as $member ) {
 281                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 282                                 return false;
 283                         }
 284                 }
 285                 return true;
 286         }
 287 }
 288
 289 /**
 290  * Class designed for counting of stats.
 291  */
 292 class SiteStatsInit {
 293
 294         // Database connection
 295         private $db;
 296
 297         // Various stats
 298         private $mEdits = null, $mArticles = null, $mPages = null;
 299         private $mUsers = null, $mFiles = null;
 300
 301         /**
 302          * Constructor
 303          * @param bool|IDatabase $database
 304          * - boolean: 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          * - boolean: 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 boolean: 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          * Refresh site_stats
 416          */
 417         public function refresh() {
 418                 $values = [
 419                         'ss_row_id' => 1,
 420                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 421                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 422                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 423                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 424                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 425                 ];
 426
 427                 $dbw = wfGetDB( DB_MASTER );
 428                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 429         }
 430 }