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\IDatabase;
  24
  25 /**
  26  * Static accessor class for site_stats and related things
  27  */
  28 class SiteStats {
  29         /** @var bool|stdClass */
  30         private static $row;
  31
  32         /** @var bool */
  33         private static $loaded = false;
  34
  35         /** @var int */
  36         private static $jobs;
  37
  38         /** @var int[] */
  39         private static $pageCount = [];
  40
  41         static function unload() {
  42                 self::$loaded = false;
  43         }
  44
  45         static function recache() {
  46                 self::load( true );
  47         }
  48
  49         /**
  50          * @param bool $recache
  51          */
  52         static function load( $recache = false ) {
  53                 if ( self::$loaded && !$recache ) {
  54                         return;
  55                 }
  56
  57                 self::$row = self::loadAndLazyInit();
  58
  59                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  60                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  61                         # Update schema
  62                         $u = new SiteStatsUpdate( 0, 0, 0 );
  63                         $u->doUpdate();
  64                         self::$row = self::doLoad( wfGetDB( DB_REPLICA ) );
  65                 }
  66
  67                 self::$loaded = true;
  68         }
  69
  70         /**
  71          * @return bool|stdClass
  72          */
  73         static function loadAndLazyInit() {
  74                 global $wgMiserMode;
  75
  76                 wfDebug( __METHOD__ . ": reading site_stats from replica DB\n" );
  77                 $row = self::doLoad( wfGetDB( DB_REPLICA ) );
  78
  79                 if ( !self::isSane( $row ) ) {
  80                         // Might have just been initialized during this request? Underflow?
  81                         wfDebug( __METHOD__ . ": site_stats damaged or missing on replica DB\n" );
  82                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  83                 }
  84
  85                 if ( !$wgMiserMode && !self::isSane( $row ) ) {
  86                         // Normally the site_stats table is initialized at install time.
  87                         // Some manual construction scenarios may leave the table empty or
  88                         // broken, however, for instance when importing from a dump into a
  89                         // clean schema with mwdumper.
  90                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  91
  92                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_REPLICA ) );
  93
  94                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  95                 }
  96
  97                 if ( !self::isSane( $row ) ) {
  98                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  99                 }
 100                 return $row;
 101         }
 102
 103         /**
 104          * @param IDatabase $db
 105          * @return bool|stdClass
 106          */
 107         static function doLoad( $db ) {
 108                 return $db->selectRow( 'site_stats', [
 109                                 'ss_row_id',
 110                                 'ss_total_edits',
 111                                 'ss_good_articles',
 112                                 'ss_total_pages',
 113                                 'ss_users',
 114                                 'ss_active_users',
 115                                 'ss_images',
 116                         ], [], __METHOD__ );
 117         }
 118
 119         /**
 120          * Return the total number of page views. Except we don't track those anymore.
 121          * Stop calling this function, it will be removed some time in the future. It's
 122          * kept here simply to prevent fatal errors.
 123          *
 124          * @deprecated since 1.25
 125          * @return int
 126          */
 127         static function views() {
 128                 wfDeprecated( __METHOD__, '1.25' );
 129                 return 0;
 130         }
 131
 132         /**
 133          * @return int
 134          */
 135         static function edits() {
 136                 self::load();
 137                 return self::$row->ss_total_edits;
 138         }
 139
 140         /**
 141          * @return int
 142          */
 143         static function articles() {
 144                 self::load();
 145                 return self::$row->ss_good_articles;
 146         }
 147
 148         /**
 149          * @return int
 150          */
 151         static function pages() {
 152                 self::load();
 153                 return self::$row->ss_total_pages;
 154         }
 155
 156         /**
 157          * @return int
 158          */
 159         static function users() {
 160                 self::load();
 161                 return self::$row->ss_users;
 162         }
 163
 164         /**
 165          * @return int
 166          */
 167         static function activeUsers() {
 168                 self::load();
 169                 return self::$row->ss_active_users;
 170         }
 171
 172         /**
 173          * @return int
 174          */
 175         static function images() {
 176                 self::load();
 177                 return self::$row->ss_images;
 178         }
 179
 180         /**
 181          * Find the number of users in a given user group.
 182          * @param string $group Name of group
 183          * @return int
 184          */
 185         static function numberingroup( $group ) {
 186                 $cache = ObjectCache::getMainWANInstance();
 187                 return $cache->getWithSetCallback(
 188                         wfMemcKey( 'SiteStats', 'groupcounts', $group ),
 189                         $cache::TTL_HOUR,
 190                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 191                                 global $wgDisableUserGroupExpiry;
 192                                 $dbr = wfGetDB( DB_REPLICA );
 193
 194                                 $setOpts += Database::getCacheSetOptions( $dbr );
 195
 196                                 return $dbr->selectField(
 197                                         'user_groups',
 198                                         'COUNT(*)',
 199                                         [
 200                                                 'ug_group' => $group,
 201                                                 $wgDisableUserGroupExpiry ?
 202                                                         '1' :
 203                                                         'ug_expiry IS NULL OR ug_expiry >= ' . $dbr->addQuotes( $dbr->timestamp() )
 204                                         ],
 205                                         __METHOD__
 206                                 );
 207                         },
 208                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 209                 );
 210         }
 211
 212         /**
 213          * @return int
 214          */
 215         static function jobs() {
 216                 if ( !isset( self::$jobs ) ) {
 217                         try{
 218                                 self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 219                         } catch ( JobQueueError $e ) {
 220                                 self::$jobs = 0;
 221                         }
 222                         /**
 223                          * Zero rows still do single row read for row that doesn't exist,
 224                          * but people are annoyed by that
 225                          */
 226                         if ( self::$jobs == 1 ) {
 227                                 self::$jobs = 0;
 228                         }
 229                 }
 230                 return self::$jobs;
 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          * Constructor
 297          * @param bool|IDatabase $database
 298          * - boolean: Whether to use the master DB
 299          * - IDatabase: Database connection to use
 300          */
 301         public function __construct( $database = false ) {
 302                 if ( $database instanceof IDatabase ) {
 303                         $this->db = $database;
 304                 } elseif ( $database ) {
 305                         $this->db = wfGetDB( DB_MASTER );
 306                 } else {
 307                         $this->db = wfGetDB( DB_REPLICA, 'vslow' );
 308                 }
 309         }
 310
 311         /**
 312          * Count the total number of edits
 313          * @return int
 314          */
 315         public function edits() {
 316                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 317                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 318                 return $this->mEdits;
 319         }
 320
 321         /**
 322          * Count pages in article space(s)
 323          * @return int
 324          */
 325         public function articles() {
 326                 global $wgArticleCountMethod;
 327
 328                 $tables = [ 'page' ];
 329                 $conds = [
 330                         'page_namespace' => MWNamespace::getContentNamespaces(),
 331                         'page_is_redirect' => 0,
 332                 ];
 333
 334                 if ( $wgArticleCountMethod == 'link' ) {
 335                         $tables[] = 'pagelinks';
 336                         $conds[] = 'pl_from=page_id';
 337                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 338                         // To make a correct check for this, we would need, for each page,
 339                         // to load the text, maybe uncompress it, maybe decode it and then
 340                         // check if there's one comma.
 341                         // But one thing we are sure is that if the page is empty, it can't
 342                         // contain a comma :)
 343                         $conds[] = 'page_len > 0';
 344                 }
 345
 346                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 347                         $conds, __METHOD__ );
 348                 return $this->mArticles;
 349         }
 350
 351         /**
 352          * Count total pages
 353          * @return int
 354          */
 355         public function pages() {
 356                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 357                 return $this->mPages;
 358         }
 359
 360         /**
 361          * Count total users
 362          * @return int
 363          */
 364         public function users() {
 365                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 366                 return $this->mUsers;
 367         }
 368
 369         /**
 370          * Count total files
 371          * @return int
 372          */
 373         public function files() {
 374                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 375                 return $this->mFiles;
 376         }
 377
 378         /**
 379          * Do all updates and commit them. More or less a replacement
 380          * for the original initStats, but without output.
 381          *
 382          * @param IDatabase|bool $database
 383          * - boolean: Whether to use the master DB
 384          * - IDatabase: Database connection to use
 385          * @param array $options Array of options, may contain the following values
 386          * - activeUsers boolean: Whether to update the number of active users (default: false)
 387          */
 388         public static function doAllAndCommit( $database, array $options = [] ) {
 389                 $options += [ 'update' => false, 'activeUsers' => false ];
 390
 391                 // Grab the object and count everything
 392                 $counter = new SiteStatsInit( $database );
 393
 394                 $counter->edits();
 395                 $counter->articles();
 396                 $counter->pages();
 397                 $counter->users();
 398                 $counter->files();
 399
 400                 $counter->refresh();
 401
 402                 // Count active users if need be
 403                 if ( $options['activeUsers'] ) {
 404                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 405                 }
 406         }
 407
 408         /**
 409          * Refresh site_stats
 410          */
 411         public function refresh() {
 412                 $values = [
 413                         'ss_row_id' => 1,
 414                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 415                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 416                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 417                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 418                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 419                 ];
 420
 421                 $dbw = wfGetDB( DB_MASTER );
 422                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 423         }
 424 }