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         static $row, $loaded = false;
  28         static $jobs;
  29         static $pageCount = array();
  30         static $groupMemberCounts = array();
  31
  32         static function recache() {
  33                 self::load( true );
  34         }
  35
  36         /**
  37          * @param $recache bool
  38          */
  39         static function load( $recache = false ) {
  40                 if ( self::$loaded && !$recache ) {
  41                         return;
  42                 }
  43
  44                 self::$row = self::loadAndLazyInit();
  45
  46                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  47                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  48                         # Update schema
  49                         $u = new SiteStatsUpdate( 0, 0, 0 );
  50                         $u->doUpdate();
  51                         self::$row = self::doLoad( wfGetDB( DB_SLAVE ) );
  52                 }
  53
  54                 self::$loaded = true;
  55         }
  56
  57         /**
  58          * @return Bool|ResultWrapper
  59          */
  60         static function loadAndLazyInit() {
  61                 wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
  62                 $row = self::doLoad( wfGetDB( DB_SLAVE ) );
  63
  64                 if ( !self::isSane( $row ) ) {
  65                         // Might have just been initialized during this request? Underflow?
  66                         wfDebug( __METHOD__ . ": site_stats damaged or missing on slave\n" );
  67                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  68                 }
  69
  70                 if ( !self::isSane( $row ) ) {
  71                         // Normally the site_stats table is initialized at install time.
  72                         // Some manual construction scenarios may leave the table empty or
  73                         // broken, however, for instance when importing from a dump into a
  74                         // clean schema with mwdumper.
  75                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  76
  77                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
  78
  79                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  80                 }
  81
  82                 if ( !self::isSane( $row ) ) {
  83                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  84                 }
  85                 return $row;
  86         }
  87
  88         /**
  89          * @param $db DatabaseBase
  90          * @return Bool|ResultWrapper
  91          */
  92         static function doLoad( $db ) {
  93                 return $db->selectRow( 'site_stats', array(
  94                                 'ss_row_id',
  95                                 'ss_total_views',
  96                                 'ss_total_edits',
  97                                 'ss_good_articles',
  98                                 'ss_total_pages',
  99                                 'ss_users',
 100                                 'ss_active_users',
 101                                 'ss_images',
 102                         ), false, __METHOD__ );
 103         }
 104
 105         /**
 106          * @return int
 107          */
 108         static function views() {
 109                 self::load();
 110                 return self::$row->ss_total_views;
 111         }
 112
 113         /**
 114          * @return int
 115          */
 116         static function edits() {
 117                 self::load();
 118                 return self::$row->ss_total_edits;
 119         }
 120
 121         /**
 122          * @return int
 123          */
 124         static function articles() {
 125                 self::load();
 126                 return self::$row->ss_good_articles;
 127         }
 128
 129         /**
 130          * @return int
 131          */
 132         static function pages() {
 133                 self::load();
 134                 return self::$row->ss_total_pages;
 135         }
 136
 137         /**
 138          * @return int
 139          */
 140         static function users() {
 141                 self::load();
 142                 return self::$row->ss_users;
 143         }
 144
 145         /**
 146          * @return int
 147          */
 148         static function activeUsers() {
 149                 self::load();
 150                 return self::$row->ss_active_users;
 151         }
 152
 153         /**
 154          * @return int
 155          */
 156         static function images() {
 157                 self::load();
 158                 return self::$row->ss_images;
 159         }
 160
 161         /**
 162          * Find the number of users in a given user group.
 163          * @param string $group name of group
 164          * @return Integer
 165          */
 166         static function numberingroup( $group ) {
 167                 if ( !isset( self::$groupMemberCounts[$group] ) ) {
 168                         global $wgMemc;
 169                         $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
 170                         $hit = $wgMemc->get( $key );
 171                         if ( !$hit ) {
 172                                 $dbr = wfGetDB( DB_SLAVE );
 173                                 $hit = $dbr->selectField(
 174                                         'user_groups',
 175                                         'COUNT(*)',
 176                                         array( 'ug_group' => $group ),
 177                                         __METHOD__
 178                                 );
 179                                 $wgMemc->set( $key, $hit, 3600 );
 180                         }
 181                         self::$groupMemberCounts[$group] = $hit;
 182                 }
 183                 return self::$groupMemberCounts[$group];
 184         }
 185
 186         /**
 187          * @return int
 188          */
 189         static function jobs() {
 190                 if ( !isset( self::$jobs ) ) {
 191                         $dbr = wfGetDB( DB_SLAVE );
 192                         self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 193                         /* Zero rows still do single row read for row that doesn't exist, but people are annoyed by that */
 194                         if ( self::$jobs == 1 ) {
 195                                 self::$jobs = 0;
 196                         }
 197                 }
 198                 return self::$jobs;
 199         }
 200
 201         /**
 202          * @param $ns int
 203          *
 204          * @return int
 205          */
 206         static function pagesInNs( $ns ) {
 207                 wfProfileIn( __METHOD__ );
 208                 if ( !isset( self::$pageCount[$ns] ) ) {
 209                         $dbr = wfGetDB( DB_SLAVE );
 210                         self::$pageCount[$ns] = (int)$dbr->selectField(
 211                                 'page',
 212                                 'COUNT(*)',
 213                                 array( 'page_namespace' => $ns ),
 214                                 __METHOD__
 215                         );
 216                 }
 217                 wfProfileOut( __METHOD__ );
 218                 return self::$pageCount[$ns];
 219         }
 220
 221         /**
 222          * Is the provided row of site stats sane, or should it be regenerated?
 223          *
 224          * @param $row
 225          *
 226          * @return bool
 227          */
 228         private static function isSane( $row ) {
 229                 if ( $row === false
 230                         || $row->ss_total_pages < $row->ss_good_articles
 231                         || $row->ss_total_edits < $row->ss_total_pages
 232                         || $row->ss_users < $row->ss_active_users
 233                 ) {
 234                         return false;
 235                 }
 236                 // Now check for underflow/overflow
 237                 foreach ( array(
 238                         'ss_total_views',
 239                         'ss_total_edits',
 240                         'ss_good_articles',
 241                         'ss_total_pages',
 242                         'ss_users',
 243                         'ss_images',
 244                 ) as $member ) {
 245                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 246                                 return false;
 247                         }
 248                 }
 249                 return true;
 250         }
 251 }
 252
 253 /**
 254  * Class designed for counting of stats.
 255  */
 256 class SiteStatsInit {
 257
 258         // Database connection
 259         private $db;
 260
 261         // Various stats
 262         private $mEdits, $mArticles, $mPages, $mUsers, $mViews, $mFiles = 0;
 263
 264         /**
 265          * Constructor
 266          * @param $database Boolean or DatabaseBase:
 267          * - Boolean: whether to use the master DB
 268          * - DatabaseBase: database connection to use
 269          */
 270         public function __construct( $database = false ) {
 271                 if ( $database instanceof DatabaseBase ) {
 272                         $this->db = $database;
 273                 } else {
 274                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 275                 }
 276         }
 277
 278         /**
 279          * Count the total number of edits
 280          * @return Integer
 281          */
 282         public function edits() {
 283                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 284                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 285                 return $this->mEdits;
 286         }
 287
 288         /**
 289          * Count pages in article space(s)
 290          * @return Integer
 291          */
 292         public function articles() {
 293                 global $wgArticleCountMethod;
 294
 295                 $tables = array( 'page' );
 296                 $conds = array(
 297                         'page_namespace' => MWNamespace::getContentNamespaces(),
 298                         'page_is_redirect' => 0,
 299                 );
 300
 301                 if ( $wgArticleCountMethod == 'link' ) {
 302                         $tables[] = 'pagelinks';
 303                         $conds[] = 'pl_from=page_id';
 304                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 305                         // To make a correct check for this, we would need, for each page,
 306                         // to load the text, maybe uncompress it, maybe decode it and then
 307                         // check if there's one comma.
 308                         // But one thing we are sure is that if the page is empty, it can't
 309                         // contain a comma :)
 310                         $conds[] = 'page_len > 0';
 311                 }
 312
 313                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 314                         $conds, __METHOD__ );
 315                 return $this->mArticles;
 316         }
 317
 318         /**
 319          * Count total pages
 320          * @return Integer
 321          */
 322         public function pages() {
 323                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 324                 return $this->mPages;
 325         }
 326
 327         /**
 328          * Count total users
 329          * @return Integer
 330          */
 331         public function users() {
 332                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 333                 return $this->mUsers;
 334         }
 335
 336         /**
 337          * Count views
 338          * @return Integer
 339          */
 340         public function views() {
 341                 $this->mViews = $this->db->selectField( 'page', 'SUM(page_counter)', '', __METHOD__ );
 342                 return $this->mViews;
 343         }
 344
 345         /**
 346          * Count total files
 347          * @return Integer
 348          */
 349         public function files() {
 350                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 351                 return $this->mFiles;
 352         }
 353
 354         /**
 355          * Do all updates and commit them. More or less a replacement
 356          * for the original initStats, but without output.
 357          *
 358          * @param $database DatabaseBase|bool
 359          * - Boolean: whether to use the master DB
 360          * - DatabaseBase: database connection to use
 361          * @param array $options of options, may contain the following values
 362          * - views Boolean: when true, do not update the number of page views (default: true)
 363          * - activeUsers Boolean: whether to update the number of active users (default: false)
 364          */
 365         public static function doAllAndCommit( $database, array $options = array() ) {
 366                 $options += array( 'update' => false, 'views' => true, 'activeUsers' => false );
 367
 368                 // Grab the object and count everything
 369                 $counter = new SiteStatsInit( $database );
 370
 371                 $counter->edits();
 372                 $counter->articles();
 373                 $counter->pages();
 374                 $counter->users();
 375                 $counter->files();
 376
 377                 // Only do views if we don't want to not count them
 378                 if ( $options['views'] ) {
 379                         $counter->views();
 380                 }
 381
 382                 $counter->refresh();
 383
 384                 // Count active users if need be
 385                 if ( $options['activeUsers'] ) {
 386                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 387                 }
 388         }
 389
 390         /**
 391          * Refresh site_stats.
 392          */
 393         protected function refresh() {
 394                 $values = array(
 395                         'ss_row_id' => 1,
 396                         'ss_total_edits' => $this->mEdits,
 397                         'ss_good_articles' => $this->mArticles,
 398                         'ss_total_pages' => $this->mPages,
 399                         'ss_users' => $this->mUsers,
 400                         'ss_images' => $this->mFiles,
 401                 ) + (
 402                         $this->mViews ? array( 'ss_total_views' => $this->mViews ) : array()
 403                 );
 404
 405                 $dbw = wfGetDB( DB_MASTER );
 406                 $dbw->upsert( 'site_stats', $values, array( 'ss_row_id' ), $values, __METHOD__ );
 407         }
 408 }