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
  26 /**
  27  * Static accessor class for site_stats and related things
  28  */
  29 class SiteStats {
  30         /** @var bool|stdClass */
  31         private static $row;
  32
  33         /** @var bool */
  34         private static $loaded = false;
  35
  36         /** @var int */
  37         private static $jobs;
  38
  39         /** @var int[] */
  40         private static $pageCount = [];
  41
  42         static function unload() {
  43                 self::$loaded = false;
  44         }
  45
  46         static function recache() {
  47                 self::load( true );
  48         }
  49
  50         /**
  51          * @param bool $recache
  52          */
  53         static function load( $recache = false ) {
  54                 if ( self::$loaded && !$recache ) {
  55                         return;
  56                 }
  57
  58                 self::$row = self::loadAndLazyInit();
  59
  60                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  61                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  62                         # Update schema
  63                         $u = new SiteStatsUpdate( 0, 0, 0 );
  64                         $u->doUpdate();
  65                         self::$row = self::doLoad( wfGetDB( DB_REPLICA ) );
  66                 }
  67
  68                 self::$loaded = true;
  69         }
  70
  71         /**
  72          * @return bool|stdClass
  73          */
  74         static function loadAndLazyInit() {
  75                 global $wgMiserMode;
  76
  77                 wfDebug( __METHOD__ . ": reading site_stats from replica DB\n" );
  78                 $row = self::doLoad( wfGetDB( DB_REPLICA ) );
  79
  80                 if ( !self::isSane( $row ) ) {
  81                         // Might have just been initialized during this request? Underflow?
  82                         wfDebug( __METHOD__ . ": site_stats damaged or missing on replica DB\n" );
  83                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  84                 }
  85
  86                 if ( !$wgMiserMode && !self::isSane( $row ) ) {
  87                         // Normally the site_stats table is initialized at install time.
  88                         // Some manual construction scenarios may leave the table empty or
  89                         // broken, however, for instance when importing from a dump into a
  90                         // clean schema with mwdumper.
  91                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  92
  93                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_REPLICA ) );
  94
  95                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  96                 }
  97
  98                 if ( !self::isSane( $row ) ) {
  99                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
 100                 }
 101                 return $row;
 102         }
 103
 104         /**
 105          * @param IDatabase $db
 106          * @return bool|stdClass
 107          */
 108         static function doLoad( $db ) {
 109                 return $db->selectRow( 'site_stats', [
 110                                 'ss_row_id',
 111                                 'ss_total_edits',
 112                                 'ss_good_articles',
 113                                 'ss_total_pages',
 114                                 'ss_users',
 115                                 'ss_active_users',
 116                                 'ss_images',
 117                         ], [], __METHOD__ );
 118         }
 119
 120         /**
 121          * Return the total number of page views. Except we don't track those anymore.
 122          * Stop calling this function, it will be removed some time in the future. It's
 123          * kept here simply to prevent fatal errors.
 124          *
 125          * @deprecated since 1.25
 126          * @return int
 127          */
 128         static function views() {
 129                 wfDeprecated( __METHOD__, '1.25' );
 130                 return 0;
 131         }
 132
 133         /**
 134          * @return int
 135          */
 136         static function edits() {
 137                 self::load();
 138                 return self::$row->ss_total_edits;
 139         }
 140
 141         /**
 142          * @return int
 143          */
 144         static function articles() {
 145                 self::load();
 146                 return self::$row->ss_good_articles;
 147         }
 148
 149         /**
 150          * @return int
 151          */
 152         static function pages() {
 153                 self::load();
 154                 return self::$row->ss_total_pages;
 155         }
 156
 157         /**
 158          * @return int
 159          */
 160         static function users() {
 161                 self::load();
 162                 return self::$row->ss_users;
 163         }
 164
 165         /**
 166          * @return int
 167          */
 168         static function activeUsers() {
 169                 self::load();
 170                 return self::$row->ss_active_users;
 171         }
 172
 173         /**
 174          * @return int
 175          */
 176         static function images() {
 177                 self::load();
 178                 return self::$row->ss_images;
 179         }
 180
 181         /**
 182          * Find the number of users in a given user group.
 183          * @param string $group Name of group
 184          * @return int
 185          */
 186         static function numberingroup( $group ) {
 187                 $cache = ObjectCache::getMainWANInstance();
 188                 return $cache->getWithSetCallback(
 189                         wfMemcKey( 'SiteStats', 'groupcounts', $group ),
 190                         $cache::TTL_HOUR,
 191                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 192                                 global $wgDisableUserGroupExpiry;
 193                                 $dbr = wfGetDB( DB_REPLICA );
 194
 195                                 $setOpts += Database::getCacheSetOptions( $dbr );
 196
 197                                 return $dbr->selectField(
 198                                         'user_groups',
 199                                         'COUNT(*)',
 200                                         [
 201                                                 'ug_group' => $group,
 202                                                 $wgDisableUserGroupExpiry ?
 203                                                         '1' :
 204                                                         'ug_expiry IS NULL OR ug_expiry >= ' . $dbr->addQuotes( $dbr->timestamp() )
 205                                         ],
 206                                         __METHOD__
 207                                 );
 208                         },
 209                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 210                 );
 211         }
 212
 213         /**
 214          * @return int
 215          */
 216         static function jobs() {
 217                 if ( !isset( self::$jobs ) ) {
 218                         try{
 219                                 self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 220                         } catch ( JobQueueError $e ) {
 221                                 self::$jobs = 0;
 222                         }
 223                         /**
 224                          * Zero rows still do single row read for row that doesn't exist,
 225                          * but people are annoyed by that
 226                          */
 227                         if ( self::$jobs == 1 ) {
 228                                 self::$jobs = 0;
 229                         }
 230                 }
 231                 return self::$jobs;
 232         }
 233
 234         /**
 235          * @param int $ns
 236          *
 237          * @return int
 238          */
 239         static function pagesInNs( $ns ) {
 240                 if ( !isset( self::$pageCount[$ns] ) ) {
 241                         $dbr = wfGetDB( DB_REPLICA );
 242                         self::$pageCount[$ns] = (int)$dbr->selectField(
 243                                 'page',
 244                                 'COUNT(*)',
 245                                 [ 'page_namespace' => $ns ],
 246                                 __METHOD__
 247                         );
 248                 }
 249                 return self::$pageCount[$ns];
 250         }
 251
 252         /**
 253          * Is the provided row of site stats sane, or should it be regenerated?
 254          *
 255          * Checks only fields which are filled by SiteStatsInit::refresh.
 256          *
 257          * @param bool|object $row
 258          *
 259          * @return bool
 260          */
 261         private static function isSane( $row ) {
 262                 if ( $row === false
 263                         || $row->ss_total_pages < $row->ss_good_articles
 264                         || $row->ss_total_edits < $row->ss_total_pages
 265                 ) {
 266                         return false;
 267                 }
 268                 // Now check for underflow/overflow
 269                 foreach ( [
 270                         'ss_total_edits',
 271                         'ss_good_articles',
 272                         'ss_total_pages',
 273                         'ss_users',
 274                         'ss_images',
 275                 ] as $member ) {
 276                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 277                                 return false;
 278                         }
 279                 }
 280                 return true;
 281         }
 282 }
 283
 284 /**
 285  * Class designed for counting of stats.
 286  */
 287 class SiteStatsInit {
 288
 289         // Database connection
 290         private $db;
 291
 292         // Various stats
 293         private $mEdits = null, $mArticles = null, $mPages = null;
 294         private $mUsers = null, $mFiles = null;
 295
 296         /**
 297          * Constructor
 298          * @param bool|IDatabase $database
 299          * - boolean: Whether to use the master DB
 300          * - IDatabase: Database connection to use
 301          */
 302         public function __construct( $database = false ) {
 303                 if ( $database instanceof IDatabase ) {
 304                         $this->db = $database;
 305                 } elseif ( $database ) {
 306                         $this->db = wfGetDB( DB_MASTER );
 307                 } else {
 308                         $this->db = wfGetDB( DB_REPLICA, 'vslow' );
 309                 }
 310         }
 311
 312         /**
 313          * Count the total number of edits
 314          * @return int
 315          */
 316         public function edits() {
 317                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 318                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 319                 return $this->mEdits;
 320         }
 321
 322         /**
 323          * Count pages in article space(s)
 324          * @return int
 325          */
 326         public function articles() {
 327                 global $wgArticleCountMethod;
 328
 329                 $tables = [ 'page' ];
 330                 $conds = [
 331                         'page_namespace' => MWNamespace::getContentNamespaces(),
 332                         'page_is_redirect' => 0,
 333                 ];
 334
 335                 if ( $wgArticleCountMethod == 'link' ) {
 336                         $tables[] = 'pagelinks';
 337                         $conds[] = 'pl_from=page_id';
 338                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 339                         // To make a correct check for this, we would need, for each page,
 340                         // to load the text, maybe uncompress it, maybe decode it and then
 341                         // check if there's one comma.
 342                         // But one thing we are sure is that if the page is empty, it can't
 343                         // contain a comma :)
 344                         $conds[] = 'page_len > 0';
 345                 }
 346
 347                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 348                         $conds, __METHOD__ );
 349                 return $this->mArticles;
 350         }
 351
 352         /**
 353          * Count total pages
 354          * @return int
 355          */
 356         public function pages() {
 357                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 358                 return $this->mPages;
 359         }
 360
 361         /**
 362          * Count total users
 363          * @return int
 364          */
 365         public function users() {
 366                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 367                 return $this->mUsers;
 368         }
 369
 370         /**
 371          * Count total files
 372          * @return int
 373          */
 374         public function files() {
 375                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 376                 return $this->mFiles;
 377         }
 378
 379         /**
 380          * Do all updates and commit them. More or less a replacement
 381          * for the original initStats, but without output.
 382          *
 383          * @param IDatabase|bool $database
 384          * - boolean: Whether to use the master DB
 385          * - IDatabase: Database connection to use
 386          * @param array $options Array of options, may contain the following values
 387          * - activeUsers boolean: Whether to update the number of active users (default: false)
 388          */
 389         public static function doAllAndCommit( $database, array $options = [] ) {
 390                 $options += [ 'update' => false, 'activeUsers' => false ];
 391
 392                 // Grab the object and count everything
 393                 $counter = new SiteStatsInit( $database );
 394
 395                 $counter->edits();
 396                 $counter->articles();
 397                 $counter->pages();
 398                 $counter->users();
 399                 $counter->files();
 400
 401                 $counter->refresh();
 402
 403                 // Count active users if need be
 404                 if ( $options['activeUsers'] ) {
 405                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 406                 }
 407         }
 408
 409         /**
 410          * Refresh site_stats
 411          */
 412         public function refresh() {
 413                 $values = [
 414                         'ss_row_id' => 1,
 415                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 416                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 417                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 418                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 419                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 420                 ];
 421
 422                 $dbw = wfGetDB( DB_MASTER );
 423                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 424         }
 425 }