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 /**
  24  * Static accessor class for site_stats and related things
  25  */
  26 class SiteStats {
  27         /** @var bool|stdClass */
  28         private static $row;
  29
  30         /** @var bool */
  31         private static $loaded = false;
  32
  33         /** @var int */
  34         private static $jobs;
  35
  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                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  58                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  59                         # Update schema
  60                         $u = new SiteStatsUpdate( 0, 0, 0 );
  61                         $u->doUpdate();
  62                         self::$row = self::doLoad( wfGetDB( DB_REPLICA ) );
  63                 }
  64
  65                 self::$loaded = true;
  66         }
  67
  68         /**
  69          * @return bool|stdClass
  70          */
  71         static function loadAndLazyInit() {
  72                 global $wgMiserMode;
  73
  74                 wfDebug( __METHOD__ . ": reading site_stats from replica DB\n" );
  75                 $row = self::doLoad( wfGetDB( DB_REPLICA ) );
  76
  77                 if ( !self::isSane( $row ) ) {
  78                         // Might have just been initialized during this request? Underflow?
  79                         wfDebug( __METHOD__ . ": site_stats damaged or missing on replica DB\n" );
  80                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  81                 }
  82
  83                 if ( !$wgMiserMode && !self::isSane( $row ) ) {
  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
  90                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_REPLICA ) );
  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                 return $row;
  99         }
 100
 101         /**
 102          * @param IDatabase $db
 103          * @return bool|stdClass
 104          */
 105         static function doLoad( $db ) {
 106                 return $db->selectRow( 'site_stats', [
 107                                 'ss_row_id',
 108                                 'ss_total_edits',
 109                                 'ss_good_articles',
 110                                 'ss_total_pages',
 111                                 'ss_users',
 112                                 'ss_active_users',
 113                                 'ss_images',
 114                         ], [], __METHOD__ );
 115         }
 116
 117         /**
 118          * Return the total number of page views. Except we don't track those anymore.
 119          * Stop calling this function, it will be removed some time in the future. It's
 120          * kept here simply to prevent fatal errors.
 121          *
 122          * @deprecated since 1.25
 123          * @return int
 124          */
 125         static function views() {
 126                 wfDeprecated( __METHOD__, '1.25' );
 127                 return 0;
 128         }
 129
 130         /**
 131          * @return int
 132          */
 133         static function edits() {
 134                 self::load();
 135                 return self::$row->ss_total_edits;
 136         }
 137
 138         /**
 139          * @return int
 140          */
 141         static function articles() {
 142                 self::load();
 143                 return self::$row->ss_good_articles;
 144         }
 145
 146         /**
 147          * @return int
 148          */
 149         static function pages() {
 150                 self::load();
 151                 return self::$row->ss_total_pages;
 152         }
 153
 154         /**
 155          * @return int
 156          */
 157         static function users() {
 158                 self::load();
 159                 return self::$row->ss_users;
 160         }
 161
 162         /**
 163          * @return int
 164          */
 165         static function activeUsers() {
 166                 self::load();
 167                 return self::$row->ss_active_users;
 168         }
 169
 170         /**
 171          * @return int
 172          */
 173         static function images() {
 174                 self::load();
 175                 return self::$row->ss_images;
 176         }
 177
 178         /**
 179          * Find the number of users in a given user group.
 180          * @param string $group Name of group
 181          * @return int
 182          */
 183         static function numberingroup( $group ) {
 184                 $cache = ObjectCache::getMainWANInstance();
 185                 return $cache->getWithSetCallback(
 186                         wfMemcKey( 'SiteStats', 'groupcounts', $group ),
 187                         $cache::TTL_HOUR,
 188                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 189                                 global $wgDisableUserGroupExpiry;
 190                                 $dbr = wfGetDB( DB_REPLICA );
 191
 192                                 $setOpts += Database::getCacheSetOptions( $dbr );
 193
 194                                 return $dbr->selectField(
 195                                         'user_groups',
 196                                         'COUNT(*)',
 197                                         [
 198                                                 'ug_group' => $group,
 199                                                 $wgDisableUserGroupExpiry ?
 200                                                         '1' :
 201                                                         'ug_expiry IS NULL OR ug_expiry >= ' . $dbr->addQuotes( $dbr->timestamp() )
 202                                         ],
 203                                         __METHOD__
 204                                 );
 205                         },
 206                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 207                 );
 208         }
 209
 210         /**
 211          * @return int
 212          */
 213         static function jobs() {
 214                 if ( !isset( self::$jobs ) ) {
 215                         try{
 216                                 self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 217                         } catch ( JobQueueError $e ) {
 218                                 self::$jobs = 0;
 219                         }
 220                         /**
 221                          * Zero rows still do single row read for row that doesn't exist,
 222                          * but people are annoyed by that
 223                          */
 224                         if ( self::$jobs == 1 ) {
 225                                 self::$jobs = 0;
 226                         }
 227                 }
 228                 return self::$jobs;
 229         }
 230
 231         /**
 232          * @param int $ns
 233          *
 234          * @return int
 235          */
 236         static function pagesInNs( $ns ) {
 237                 if ( !isset( self::$pageCount[$ns] ) ) {
 238                         $dbr = wfGetDB( DB_REPLICA );
 239                         self::$pageCount[$ns] = (int)$dbr->selectField(
 240                                 'page',
 241                                 'COUNT(*)',
 242                                 [ 'page_namespace' => $ns ],
 243                                 __METHOD__
 244                         );
 245                 }
 246                 return self::$pageCount[$ns];
 247         }
 248
 249         /**
 250          * Is the provided row of site stats sane, or should it be regenerated?
 251          *
 252          * Checks only fields which are filled by SiteStatsInit::refresh.
 253          *
 254          * @param bool|object $row
 255          *
 256          * @return bool
 257          */
 258         private static function isSane( $row ) {
 259                 if ( $row === false
 260                         || $row->ss_total_pages < $row->ss_good_articles
 261                         || $row->ss_total_edits < $row->ss_total_pages
 262                 ) {
 263                         return false;
 264                 }
 265                 // Now check for underflow/overflow
 266                 foreach ( [
 267                         'ss_total_edits',
 268                         'ss_good_articles',
 269                         'ss_total_pages',
 270                         'ss_users',
 271                         'ss_images',
 272                 ] as $member ) {
 273                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 274                                 return false;
 275                         }
 276                 }
 277                 return true;
 278         }
 279 }
 280
 281 /**
 282  * Class designed for counting of stats.
 283  */
 284 class SiteStatsInit {
 285
 286         // Database connection
 287         private $db;
 288
 289         // Various stats
 290         private $mEdits = null, $mArticles = null, $mPages = null;
 291         private $mUsers = null, $mFiles = null;
 292
 293         /**
 294          * Constructor
 295          * @param bool|IDatabase $database
 296          * - boolean: Whether to use the master DB
 297          * - IDatabase: Database connection to use
 298          */
 299         public function __construct( $database = false ) {
 300                 if ( $database instanceof IDatabase ) {
 301                         $this->db = $database;
 302                 } elseif ( $database ) {
 303                         $this->db = wfGetDB( DB_MASTER );
 304                 } else {
 305                         $this->db = wfGetDB( DB_REPLICA, 'vslow' );
 306                 }
 307         }
 308
 309         /**
 310          * Count the total number of edits
 311          * @return int
 312          */
 313         public function edits() {
 314                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 315                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 316                 return $this->mEdits;
 317         }
 318
 319         /**
 320          * Count pages in article space(s)
 321          * @return int
 322          */
 323         public function articles() {
 324                 global $wgArticleCountMethod;
 325
 326                 $tables = [ 'page' ];
 327                 $conds = [
 328                         'page_namespace' => MWNamespace::getContentNamespaces(),
 329                         'page_is_redirect' => 0,
 330                 ];
 331
 332                 if ( $wgArticleCountMethod == 'link' ) {
 333                         $tables[] = 'pagelinks';
 334                         $conds[] = 'pl_from=page_id';
 335                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 336                         // To make a correct check for this, we would need, for each page,
 337                         // to load the text, maybe uncompress it, maybe decode it and then
 338                         // check if there's one comma.
 339                         // But one thing we are sure is that if the page is empty, it can't
 340                         // contain a comma :)
 341                         $conds[] = 'page_len > 0';
 342                 }
 343
 344                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 345                         $conds, __METHOD__ );
 346                 return $this->mArticles;
 347         }
 348
 349         /**
 350          * Count total pages
 351          * @return int
 352          */
 353         public function pages() {
 354                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 355                 return $this->mPages;
 356         }
 357
 358         /**
 359          * Count total users
 360          * @return int
 361          */
 362         public function users() {
 363                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 364                 return $this->mUsers;
 365         }
 366
 367         /**
 368          * Count total files
 369          * @return int
 370          */
 371         public function files() {
 372                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 373                 return $this->mFiles;
 374         }
 375
 376         /**
 377          * Do all updates and commit them. More or less a replacement
 378          * for the original initStats, but without output.
 379          *
 380          * @param IDatabase|bool $database
 381          * - boolean: Whether to use the master DB
 382          * - IDatabase: Database connection to use
 383          * @param array $options Array of options, may contain the following values
 384          * - activeUsers boolean: Whether to update the number of active users (default: false)
 385          */
 386         public static function doAllAndCommit( $database, array $options = [] ) {
 387                 $options += [ 'update' => false, 'activeUsers' => false ];
 388
 389                 // Grab the object and count everything
 390                 $counter = new SiteStatsInit( $database );
 391
 392                 $counter->edits();
 393                 $counter->articles();
 394                 $counter->pages();
 395                 $counter->users();
 396                 $counter->files();
 397
 398                 $counter->refresh();
 399
 400                 // Count active users if need be
 401                 if ( $options['activeUsers'] ) {
 402                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 403                 }
 404         }
 405
 406         /**
 407          * Refresh site_stats
 408          */
 409         public function refresh() {
 410                 $values = [
 411                         'ss_row_id' => 1,
 412                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 413                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 414                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 415                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 416                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 417                 ];
 418
 419                 $dbw = wfGetDB( DB_MASTER );
 420                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 421         }
 422 }