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                                 $dbr = wfGetDB( DB_REPLICA );
 190
 191                                 $setOpts += Database::getCacheSetOptions( $dbr );
 192
 193                                 return $dbr->selectField(
 194                                         'user_groups',
 195                                         'COUNT(*)',
 196                                         [ 'ug_group' => $group ],
 197                                         __METHOD__
 198                                 );
 199                         },
 200                         [ 'pcTTL' => $cache::TTL_PROC_LONG ]
 201                 );
 202         }
 203
 204         /**
 205          * @return int
 206          */
 207         static function jobs() {
 208                 if ( !isset( self::$jobs ) ) {
 209                         try{
 210                                 self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 211                         } catch ( JobQueueError $e ) {
 212                                 self::$jobs = 0;
 213                         }
 214                         /**
 215                          * Zero rows still do single row read for row that doesn't exist,
 216                          * but people are annoyed by that
 217                          */
 218                         if ( self::$jobs == 1 ) {
 219                                 self::$jobs = 0;
 220                         }
 221                 }
 222                 return self::$jobs;
 223         }
 224
 225         /**
 226          * @param int $ns
 227          *
 228          * @return int
 229          */
 230         static function pagesInNs( $ns ) {
 231                 if ( !isset( self::$pageCount[$ns] ) ) {
 232                         $dbr = wfGetDB( DB_REPLICA );
 233                         self::$pageCount[$ns] = (int)$dbr->selectField(
 234                                 'page',
 235                                 'COUNT(*)',
 236                                 [ 'page_namespace' => $ns ],
 237                                 __METHOD__
 238                         );
 239                 }
 240                 return self::$pageCount[$ns];
 241         }
 242
 243         /**
 244          * Is the provided row of site stats sane, or should it be regenerated?
 245          *
 246          * Checks only fields which are filled by SiteStatsInit::refresh.
 247          *
 248          * @param bool|object $row
 249          *
 250          * @return bool
 251          */
 252         private static function isSane( $row ) {
 253                 if ( $row === false
 254                         || $row->ss_total_pages < $row->ss_good_articles
 255                         || $row->ss_total_edits < $row->ss_total_pages
 256                 ) {
 257                         return false;
 258                 }
 259                 // Now check for underflow/overflow
 260                 foreach ( [
 261                         'ss_total_edits',
 262                         'ss_good_articles',
 263                         'ss_total_pages',
 264                         'ss_users',
 265                         'ss_images',
 266                 ] as $member ) {
 267                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 268                                 return false;
 269                         }
 270                 }
 271                 return true;
 272         }
 273 }
 274
 275 /**
 276  * Class designed for counting of stats.
 277  */
 278 class SiteStatsInit {
 279
 280         // Database connection
 281         private $db;
 282
 283         // Various stats
 284         private $mEdits = null, $mArticles = null, $mPages = null;
 285         private $mUsers = null, $mFiles = null;
 286
 287         /**
 288          * Constructor
 289          * @param bool|IDatabase $database
 290          * - boolean: Whether to use the master DB
 291          * - IDatabase: Database connection to use
 292          */
 293         public function __construct( $database = false ) {
 294                 if ( $database instanceof IDatabase ) {
 295                         $this->db = $database;
 296                 } elseif ( $database ) {
 297                         $this->db = wfGetDB( DB_MASTER );
 298                 } else {
 299                         $this->db = wfGetDB( DB_REPLICA, 'vslow' );
 300                 }
 301         }
 302
 303         /**
 304          * Count the total number of edits
 305          * @return int
 306          */
 307         public function edits() {
 308                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 309                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 310                 return $this->mEdits;
 311         }
 312
 313         /**
 314          * Count pages in article space(s)
 315          * @return int
 316          */
 317         public function articles() {
 318                 global $wgArticleCountMethod;
 319
 320                 $tables = [ 'page' ];
 321                 $conds = [
 322                         'page_namespace' => MWNamespace::getContentNamespaces(),
 323                         'page_is_redirect' => 0,
 324                 ];
 325
 326                 if ( $wgArticleCountMethod == 'link' ) {
 327                         $tables[] = 'pagelinks';
 328                         $conds[] = 'pl_from=page_id';
 329                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 330                         // To make a correct check for this, we would need, for each page,
 331                         // to load the text, maybe uncompress it, maybe decode it and then
 332                         // check if there's one comma.
 333                         // But one thing we are sure is that if the page is empty, it can't
 334                         // contain a comma :)
 335                         $conds[] = 'page_len > 0';
 336                 }
 337
 338                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 339                         $conds, __METHOD__ );
 340                 return $this->mArticles;
 341         }
 342
 343         /**
 344          * Count total pages
 345          * @return int
 346          */
 347         public function pages() {
 348                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 349                 return $this->mPages;
 350         }
 351
 352         /**
 353          * Count total users
 354          * @return int
 355          */
 356         public function users() {
 357                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 358                 return $this->mUsers;
 359         }
 360
 361         /**
 362          * Count total files
 363          * @return int
 364          */
 365         public function files() {
 366                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 367                 return $this->mFiles;
 368         }
 369
 370         /**
 371          * Do all updates and commit them. More or less a replacement
 372          * for the original initStats, but without output.
 373          *
 374          * @param IDatabase|bool $database
 375          * - boolean: Whether to use the master DB
 376          * - IDatabase: Database connection to use
 377          * @param array $options Array of options, may contain the following values
 378          * - activeUsers boolean: Whether to update the number of active users (default: false)
 379          */
 380         public static function doAllAndCommit( $database, array $options = [] ) {
 381                 $options += [ 'update' => false, 'activeUsers' => false ];
 382
 383                 // Grab the object and count everything
 384                 $counter = new SiteStatsInit( $database );
 385
 386                 $counter->edits();
 387                 $counter->articles();
 388                 $counter->pages();
 389                 $counter->users();
 390                 $counter->files();
 391
 392                 $counter->refresh();
 393
 394                 // Count active users if need be
 395                 if ( $options['activeUsers'] ) {
 396                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 397                 }
 398         }
 399
 400         /**
 401          * Refresh site_stats
 402          */
 403         public function refresh() {
 404                 $values = [
 405                         'ss_row_id' => 1,
 406                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 407                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 408                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 409                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 410                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 411                 ];
 412
 413                 $dbw = wfGetDB( DB_MASTER );
 414                 $dbw->upsert( 'site_stats', $values, [ 'ss_row_id' ], $values, __METHOD__ );
 415         }
 416 }