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