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         /** @var int[] */
  40         private static $groupMemberCounts = [];
  41
  42         static function recache() {
  43                 self::load( true );
  44         }
  45
  46         /**
  47          * @param bool $recache
  48          */
  49         static function load( $recache = false ) {
  50                 if ( self::$loaded && !$recache ) {
  51                         return;
  52                 }
  53
  54                 self::$row = self::loadAndLazyInit();
  55
  56                 # This code is somewhat schema-agnostic, because I'm changing it in a minor release -- TS
  57                 if ( !isset( self::$row->ss_total_pages ) && self::$row->ss_total_pages == -1 ) {
  58                         # Update schema
  59                         $u = new SiteStatsUpdate( 0, 0, 0 );
  60                         $u->doUpdate();
  61                         self::$row = self::doLoad( wfGetDB( DB_SLAVE ) );
  62                 }
  63
  64                 self::$loaded = true;
  65         }
  66
  67         /**
  68          * @return bool|ResultWrapper
  69          */
  70         static function loadAndLazyInit() {
  71                 global $wgMiserMode;
  72
  73                 wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
  74                 $row = self::doLoad( wfGetDB( DB_SLAVE ) );
  75
  76                 if ( !self::isSane( $row ) ) {
  77                         // Might have just been initialized during this request? Underflow?
  78                         wfDebug( __METHOD__ . ": site_stats damaged or missing on slave\n" );
  79                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  80                 }
  81
  82                 if ( !$wgMiserMode && !self::isSane( $row ) ) {
  83                         // Normally the site_stats table is initialized at install time.
  84                         // Some manual construction scenarios may leave the table empty or
  85                         // broken, however, for instance when importing from a dump into a
  86                         // clean schema with mwdumper.
  87                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  88
  89                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
  90
  91                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  92                 }
  93
  94                 if ( !self::isSane( $row ) ) {
  95                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  96                 }
  97                 return $row;
  98         }
  99
 100         /**
 101          * @param IDatabase $db
 102          * @return bool|ResultWrapper
 103          */
 104         static function doLoad( $db ) {
 105                 return $db->selectRow( 'site_stats', [
 106                                 'ss_row_id',
 107                                 'ss_total_edits',
 108                                 'ss_good_articles',
 109                                 'ss_total_pages',
 110                                 'ss_users',
 111                                 'ss_active_users',
 112                                 'ss_images',
 113                         ], false, __METHOD__ );
 114         }
 115
 116         /**
 117          * Return the total number of page views. Except we don't track those anymore.
 118          * Stop calling this function, it will be removed some time in the future. It's
 119          * kept here simply to prevent fatal errors.
 120          *
 121          * @deprecated since 1.25
 122          * @return int
 123          */
 124         static function views() {
 125                 wfDeprecated( __METHOD__, '1.25' );
 126                 return 0;
 127         }
 128
 129         /**
 130          * @return int
 131          */
 132         static function edits() {
 133                 self::load();
 134                 return self::$row->ss_total_edits;
 135         }
 136
 137         /**
 138          * @return int
 139          */
 140         static function articles() {
 141                 self::load();
 142                 return self::$row->ss_good_articles;
 143         }
 144
 145         /**
 146          * @return int
 147          */
 148         static function pages() {
 149                 self::load();
 150                 return self::$row->ss_total_pages;
 151         }
 152
 153         /**
 154          * @return int
 155          */
 156         static function users() {
 157                 self::load();
 158                 return self::$row->ss_users;
 159         }
 160
 161         /**
 162          * @return int
 163          */
 164         static function activeUsers() {
 165                 self::load();
 166                 return self::$row->ss_active_users;
 167         }
 168
 169         /**
 170          * @return int
 171          */
 172         static function images() {
 173                 self::load();
 174                 return self::$row->ss_images;
 175         }
 176
 177         /**
 178          * Find the number of users in a given user group.
 179          * @param string $group Name of group
 180          * @return int
 181          */
 182         static function numberingroup( $group ) {
 183                 $cache = ObjectCache::getMainWANInstance();
 184                 return $cache->getWithSetCallback(
 185                         wfMemcKey( 'SiteStats', 'groupcounts', $group ),
 186                         $cache::TTL_HOUR,
 187                         function ( $oldValue, &$ttl, array &$setOpts ) use ( $group ) {
 188                                 $dbr = wfGetDB( DB_SLAVE );
 189
 190                                 $setOpts += Database::getCacheSetOptions( $dbr );
 191
 192                                 return $dbr->selectField(
 193                                         'user_groups',
 194                                         'COUNT(*)',
 195                                         [ 'ug_group' => $group ],
 196                                         __METHOD__
 197                                 );
 198                         },
 199                         [ 'pcTTL' => 10 ]
 200                 );
 201         }
 202
 203         /**
 204          * @return int
 205          */
 206         static function jobs() {
 207                 if ( !isset( self::$jobs ) ) {
 208                         $dbr = wfGetDB( DB_SLAVE );
 209                         self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 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                 } else {
 293                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 294                 }
 295         }
 296
 297         /**
 298          * Count the total number of edits
 299          * @return int
 300          */
 301         public function edits() {
 302                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 303                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 304                 return $this->mEdits;
 305         }
 306
 307         /**
 308          * Count pages in article space(s)
 309          * @return int
 310          */
 311         public function articles() {
 312                 global $wgArticleCountMethod;
 313
 314                 $tables = [ 'page' ];
 315                 $conds = [
 316                         'page_namespace' => MWNamespace::getContentNamespaces(),
 317                         'page_is_redirect' => 0,
 318                 ];
 319
 320                 if ( $wgArticleCountMethod == 'link' ) {
 321                         $tables[] = 'pagelinks';
 322                         $conds[] = 'pl_from=page_id';
 323                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 324                         // To make a correct check for this, we would need, for each page,
 325                         // to load the text, maybe uncompress it, maybe decode it and then
 326                         // check if there's one comma.
 327                         // But one thing we are sure is that if the page is empty, it can't
 328                         // contain a comma :)
 329                         $conds[] = 'page_len > 0';
 330                 }
 331
 332                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 333                         $conds, __METHOD__ );
 334                 return $this->mArticles;
 335         }
 336
 337         /**
 338          * Count total pages
 339          * @return int
 340          */
 341         public function pages() {
 342                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 343                 return $this->mPages;
 344         }
 345
 346         /**
 347          * Count total users
 348          * @return int
 349          */
 350         public function users() {
 351                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 352                 return $this->mUsers;
 353         }
 354
 355         /**
 356          * Count total files
 357          * @return int
 358          */
 359         public function files() {
 360                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 361                 return $this->mFiles;
 362         }
 363
 364         /**
 365          * Do all updates and commit them. More or less a replacement
 366          * for the original initStats, but without output.
 367          *
 368          * @param IDatabase|bool $database
 369          * - boolean: Whether to use the master DB
 370          * - IDatabase: Database connection to use
 371          * @param array $options Array of options, may contain the following values
 372          * - activeUsers boolean: Whether to update the number of active users (default: false)
 373          */
 374         public static function doAllAndCommit( $database, array $options = [] ) {
 375                 $options += [ 'update' => false, 'activeUsers' => false ];
 376
 377                 // Grab the object and count everything
 378                 $counter = new SiteStatsInit( $database );
 379
 380                 $counter->edits();
 381                 $counter->articles();
 382                 $counter->pages();
 383                 $counter->users();
 384                 $counter->files();
 385
 386                 $counter->refresh();
 387
 388                 // Count active users if need be
 389                 if ( $options['activeUsers'] ) {
 390                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 391                 }
 392         }
 393
 394         /**
 395          * Refresh site_stats
 396          */
 397         public function refresh() {
 398                 $values = [
 399                         'ss_row_id' => 1,
 400                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 401                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 402                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 403                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 404                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 405                 ];
 406
 407                 $dbw = wfGetDB( DB_MASTER );
 408                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 409         }
 410 }