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 = array();
  38
  39         /** @var int[] */
  40         private static $groupMemberCounts = array();
  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 DatabaseBase $db
 102          * @return bool|ResultWrapper
 103          */
 104         static function doLoad( $db ) {
 105                 return $db->selectRow( 'site_stats', array(
 106                                 'ss_row_id',
 107                                 'ss_total_edits',
 108                                 'ss_good_articles',
 109                                 'ss_total_pages',
 110                                 'ss_users',
 111                                 'ss_active_users',
 112                                 'ss_images',
 113                         ), false, __METHOD__ );
 114         }
 115
 116         /**
 117          * Return the total number of page views. Except we don't track those anymore.
 118          * Stop calling this function, it will be removed some time in the future. It's
 119          * kept here simply to prevent fatal errors.
 120          *
 121          * @deprecated since 1.25
 122          * @return int
 123          */
 124         static function views() {
 125                 wfDeprecated( __METHOD__, '1.25' );
 126                 return 0;
 127         }
 128
 129         /**
 130          * @return int
 131          */
 132         static function edits() {
 133                 self::load();
 134                 return self::$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( $group ) {
 183                 if ( !isset( self::$groupMemberCounts[$group] ) ) {
 184                         global $wgMemc;
 185                         $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
 186                         $hit = $wgMemc->get( $key );
 187                         if ( !$hit ) {
 188                                 $dbr = wfGetDB( DB_SLAVE );
 189                                 $hit = $dbr->selectField(
 190                                         'user_groups',
 191                                         'COUNT(*)',
 192                                         array( 'ug_group' => $group ),
 193                                         __METHOD__
 194                                 );
 195                                 $wgMemc->set( $key, $hit, 3600 );
 196                         }
 197                         self::$groupMemberCounts[$group] = $hit;
 198                 }
 199                 return self::$groupMemberCounts[$group];
 200         }
 201
 202         /**
 203          * @return int
 204          */
 205         static function jobs() {
 206                 if ( !isset( self::$jobs ) ) {
 207                         $dbr = wfGetDB( DB_SLAVE );
 208                         self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 209                         /**
 210                          * Zero rows still do single row read for row that doesn't exist,
 211                          * but people are annoyed by that
 212                          */
 213                         if ( self::$jobs == 1 ) {
 214                                 self::$jobs = 0;
 215                         }
 216                 }
 217                 return self::$jobs;
 218         }
 219
 220         /**
 221          * @param int $ns
 222          *
 223          * @return int
 224          */
 225         static function pagesInNs( $ns ) {
 226                 if ( !isset( self::$pageCount[$ns] ) ) {
 227                         $dbr = wfGetDB( DB_SLAVE );
 228                         self::$pageCount[$ns] = (int)$dbr->selectField(
 229                                 'page',
 230                                 'COUNT(*)',
 231                                 array( 'page_namespace' => $ns ),
 232                                 __METHOD__
 233                         );
 234                 }
 235                 return self::$pageCount[$ns];
 236         }
 237
 238         /**
 239          * Is the provided row of site stats sane, or should it be regenerated?
 240          *
 241          * Checks only fields which are filled by SiteStatsInit::refresh.
 242          *
 243          * @param bool|object $row
 244          *
 245          * @return bool
 246          */
 247         private static function isSane( $row ) {
 248                 if ( $row === false
 249                         || $row->ss_total_pages < $row->ss_good_articles
 250                         || $row->ss_total_edits < $row->ss_total_pages
 251                 ) {
 252                         return false;
 253                 }
 254                 // Now check for underflow/overflow
 255                 foreach ( array(
 256                         'ss_total_edits',
 257                         'ss_good_articles',
 258                         'ss_total_pages',
 259                         'ss_users',
 260                         'ss_images',
 261                 ) as $member ) {
 262                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 263                                 return false;
 264                         }
 265                 }
 266                 return true;
 267         }
 268 }
 269
 270 /**
 271  * Class designed for counting of stats.
 272  */
 273 class SiteStatsInit {
 274
 275         // Database connection
 276         private $db;
 277
 278         // Various stats
 279         private $mEdits = null, $mArticles = null, $mPages = null;
 280         private $mUsers = null, $mFiles = null;
 281
 282         /**
 283          * Constructor
 284          * @param bool|DatabaseBase $database
 285          * - Boolean: whether to use the master DB
 286          * - DatabaseBase: database connection to use
 287          */
 288         public function __construct( $database = false ) {
 289                 if ( $database instanceof DatabaseBase ) {
 290                         $this->db = $database;
 291                 } else {
 292                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 293                 }
 294         }
 295
 296         /**
 297          * Count the total number of edits
 298          * @return int
 299          */
 300         public function edits() {
 301                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 302                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 303                 return $this->mEdits;
 304         }
 305
 306         /**
 307          * Count pages in article space(s)
 308          * @return int
 309          */
 310         public function articles() {
 311                 global $wgArticleCountMethod;
 312
 313                 $tables = array( 'page' );
 314                 $conds = array(
 315                         'page_namespace' => MWNamespace::getContentNamespaces(),
 316                         'page_is_redirect' => 0,
 317                 );
 318
 319                 if ( $wgArticleCountMethod == 'link' ) {
 320                         $tables[] = 'pagelinks';
 321                         $conds[] = 'pl_from=page_id';
 322                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 323                         // To make a correct check for this, we would need, for each page,
 324                         // to load the text, maybe uncompress it, maybe decode it and then
 325                         // check if there's one comma.
 326                         // But one thing we are sure is that if the page is empty, it can't
 327                         // contain a comma :)
 328                         $conds[] = 'page_len > 0';
 329                 }
 330
 331                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 332                         $conds, __METHOD__ );
 333                 return $this->mArticles;
 334         }
 335
 336         /**
 337          * Count total pages
 338          * @return int
 339          */
 340         public function pages() {
 341                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 342                 return $this->mPages;
 343         }
 344
 345         /**
 346          * Count total users
 347          * @return int
 348          */
 349         public function users() {
 350                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 351                 return $this->mUsers;
 352         }
 353
 354         /**
 355          * Count total files
 356          * @return int
 357          */
 358         public function files() {
 359                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 360                 return $this->mFiles;
 361         }
 362
 363         /**
 364          * Do all updates and commit them. More or less a replacement
 365          * for the original initStats, but without output.
 366          *
 367          * @param DatabaseBase|bool $database
 368          * - Boolean: whether to use the master DB
 369          * - DatabaseBase: database connection to use
 370          * @param array $options Array of options, may contain the following values
 371          * - activeUsers Boolean: whether to update the number of active users (default: false)
 372          */
 373         public static function doAllAndCommit( $database, array $options = array() ) {
 374                 $options += array( 'update' => false, 'activeUsers' => false );
 375
 376                 // Grab the object and count everything
 377                 $counter = new SiteStatsInit( $database );
 378
 379                 $counter->edits();
 380                 $counter->articles();
 381                 $counter->pages();
 382                 $counter->users();
 383                 $counter->files();
 384
 385                 $counter->refresh();
 386
 387                 // Count active users if need be
 388                 if ( $options['activeUsers'] ) {
 389                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 390                 }
 391         }
 392
 393         /**
 394          * Refresh site_stats
 395          */
 396         public function refresh() {
 397                 $values = array(
 398                         'ss_row_id' => 1,
 399                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 400                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 401                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 402                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 403                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 404                 );
 405
 406                 $dbw = wfGetDB( DB_MASTER );
 407                 $dbw->upsert( 'site_stats', $values, array( 'ss_row_id' ), $values, __METHOD__ );
 408         }
 409 }