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                 wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
  72                 $row = self::doLoad( wfGetDB( DB_SLAVE ) );
  73
  74                 if ( !self::isSane( $row ) ) {
  75                         // Might have just been initialized during this request? Underflow?
  76                         wfDebug( __METHOD__ . ": site_stats damaged or missing on slave\n" );
  77                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  78                 }
  79
  80                 if ( !self::isSane( $row ) ) {
  81                         // Normally the site_stats table is initialized at install time.
  82                         // Some manual construction scenarios may leave the table empty or
  83                         // broken, however, for instance when importing from a dump into a
  84                         // clean schema with mwdumper.
  85                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  86
  87                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
  88
  89                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  90                 }
  91
  92                 if ( !self::isSane( $row ) ) {
  93                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  94                 }
  95                 return $row;
  96         }
  97
  98         /**
  99          * @param DatabaseBase $db
 100          * @return bool|ResultWrapper
 101          */
 102         static function doLoad( $db ) {
 103                 return $db->selectRow( 'site_stats', array(
 104                                 'ss_row_id',
 105                                 'ss_total_edits',
 106                                 'ss_good_articles',
 107                                 'ss_total_pages',
 108                                 'ss_users',
 109                                 'ss_active_users',
 110                                 'ss_images',
 111                         ), false, __METHOD__ );
 112         }
 113
 114         /**
 115          * Return the total number of page views. Except we don't track those anymore.
 116          * Stop calling this function, it will be removed some time in the future. It's
 117          * kept here simply to prevent fatal errors.
 118          *
 119          * @deprecated since 1.25
 120          * @return int
 121          */
 122         static function views() {
 123                 wfDeprecated( __METHOD__, '1.25' );
 124                 return 0;
 125         }
 126
 127         /**
 128          * @return int
 129          */
 130         static function edits() {
 131                 self::load();
 132                 return self::$row->ss_total_edits;
 133         }
 134
 135         /**
 136          * @return int
 137          */
 138         static function articles() {
 139                 self::load();
 140                 return self::$row->ss_good_articles;
 141         }
 142
 143         /**
 144          * @return int
 145          */
 146         static function pages() {
 147                 self::load();
 148                 return self::$row->ss_total_pages;
 149         }
 150
 151         /**
 152          * @return int
 153          */
 154         static function users() {
 155                 self::load();
 156                 return self::$row->ss_users;
 157         }
 158
 159         /**
 160          * @return int
 161          */
 162         static function activeUsers() {
 163                 self::load();
 164                 return self::$row->ss_active_users;
 165         }
 166
 167         /**
 168          * @return int
 169          */
 170         static function images() {
 171                 self::load();
 172                 return self::$row->ss_images;
 173         }
 174
 175         /**
 176          * Find the number of users in a given user group.
 177          * @param string $group Name of group
 178          * @return int
 179          */
 180         static function numberingroup( $group ) {
 181                 if ( !isset( self::$groupMemberCounts[$group] ) ) {
 182                         global $wgMemc;
 183                         $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
 184                         $hit = $wgMemc->get( $key );
 185                         if ( !$hit ) {
 186                                 $dbr = wfGetDB( DB_SLAVE );
 187                                 $hit = $dbr->selectField(
 188                                         'user_groups',
 189                                         'COUNT(*)',
 190                                         array( 'ug_group' => $group ),
 191                                         __METHOD__
 192                                 );
 193                                 $wgMemc->set( $key, $hit, 3600 );
 194                         }
 195                         self::$groupMemberCounts[$group] = $hit;
 196                 }
 197                 return self::$groupMemberCounts[$group];
 198         }
 199
 200         /**
 201          * @return int
 202          */
 203         static function jobs() {
 204                 if ( !isset( self::$jobs ) ) {
 205                         $dbr = wfGetDB( DB_SLAVE );
 206                         self::$jobs = array_sum( JobQueueGroup::singleton()->getQueueSizes() );
 207                         /**
 208                          * Zero rows still do single row read for row that doesn't exist,
 209                          * but people are annoyed by that
 210                          */
 211                         if ( self::$jobs == 1 ) {
 212                                 self::$jobs = 0;
 213                         }
 214                 }
 215                 return self::$jobs;
 216         }
 217
 218         /**
 219          * @param int $ns
 220          *
 221          * @return int
 222          */
 223         static function pagesInNs( $ns ) {
 224                 if ( !isset( self::$pageCount[$ns] ) ) {
 225                         $dbr = wfGetDB( DB_SLAVE );
 226                         self::$pageCount[$ns] = (int)$dbr->selectField(
 227                                 'page',
 228                                 'COUNT(*)',
 229                                 array( 'page_namespace' => $ns ),
 230                                 __METHOD__
 231                         );
 232                 }
 233                 return self::$pageCount[$ns];
 234         }
 235
 236         /**
 237          * Is the provided row of site stats sane, or should it be regenerated?
 238          *
 239          * Checks only fields which are filled by SiteStatsInit::refresh.
 240          *
 241          * @param bool|object $row
 242          *
 243          * @return bool
 244          */
 245         private static function isSane( $row ) {
 246                 if ( $row === false
 247                         || $row->ss_total_pages < $row->ss_good_articles
 248                         || $row->ss_total_edits < $row->ss_total_pages
 249                 ) {
 250                         return false;
 251                 }
 252                 // Now check for underflow/overflow
 253                 foreach ( array(
 254                         'ss_total_edits',
 255                         'ss_good_articles',
 256                         'ss_total_pages',
 257                         'ss_users',
 258                         'ss_images',
 259                 ) as $member ) {
 260                         if ( $row->$member > 2000000000 || $row->$member < 0 ) {
 261                                 return false;
 262                         }
 263                 }
 264                 return true;
 265         }
 266 }
 267
 268 /**
 269  * Class designed for counting of stats.
 270  */
 271 class SiteStatsInit {
 272
 273         // Database connection
 274         private $db;
 275
 276         // Various stats
 277         private $mEdits = null, $mArticles = null, $mPages = null;
 278         private $mUsers = null, $mFiles = null;
 279
 280         /**
 281          * Constructor
 282          * @param bool|DatabaseBase $database
 283          * - Boolean: whether to use the master DB
 284          * - DatabaseBase: database connection to use
 285          */
 286         public function __construct( $database = false ) {
 287                 if ( $database instanceof DatabaseBase ) {
 288                         $this->db = $database;
 289                 } else {
 290                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 291                 }
 292         }
 293
 294         /**
 295          * Count the total number of edits
 296          * @return int
 297          */
 298         public function edits() {
 299                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 300                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 301                 return $this->mEdits;
 302         }
 303
 304         /**
 305          * Count pages in article space(s)
 306          * @return int
 307          */
 308         public function articles() {
 309                 global $wgArticleCountMethod;
 310
 311                 $tables = array( 'page' );
 312                 $conds = array(
 313                         'page_namespace' => MWNamespace::getContentNamespaces(),
 314                         'page_is_redirect' => 0,
 315                 );
 316
 317                 if ( $wgArticleCountMethod == 'link' ) {
 318                         $tables[] = 'pagelinks';
 319                         $conds[] = 'pl_from=page_id';
 320                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 321                         // To make a correct check for this, we would need, for each page,
 322                         // to load the text, maybe uncompress it, maybe decode it and then
 323                         // check if there's one comma.
 324                         // But one thing we are sure is that if the page is empty, it can't
 325                         // contain a comma :)
 326                         $conds[] = 'page_len > 0';
 327                 }
 328
 329                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 330                         $conds, __METHOD__ );
 331                 return $this->mArticles;
 332         }
 333
 334         /**
 335          * Count total pages
 336          * @return int
 337          */
 338         public function pages() {
 339                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 340                 return $this->mPages;
 341         }
 342
 343         /**
 344          * Count total users
 345          * @return int
 346          */
 347         public function users() {
 348                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 349                 return $this->mUsers;
 350         }
 351
 352         /**
 353          * Count total files
 354          * @return int
 355          */
 356         public function files() {
 357                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 358                 return $this->mFiles;
 359         }
 360
 361         /**
 362          * Do all updates and commit them. More or less a replacement
 363          * for the original initStats, but without output.
 364          *
 365          * @param DatabaseBase|bool $database
 366          * - Boolean: whether to use the master DB
 367          * - DatabaseBase: database connection to use
 368          * @param array $options Array of options, may contain the following values
 369          * - activeUsers Boolean: whether to update the number of active users (default: false)
 370          */
 371         public static function doAllAndCommit( $database, array $options = array() ) {
 372                 $options += array( 'update' => false, 'activeUsers' => false );
 373
 374                 // Grab the object and count everything
 375                 $counter = new SiteStatsInit( $database );
 376
 377                 $counter->edits();
 378                 $counter->articles();
 379                 $counter->pages();
 380                 $counter->users();
 381                 $counter->files();
 382
 383                 $counter->refresh();
 384
 385                 // Count active users if need be
 386                 if ( $options['activeUsers'] ) {
 387                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 388                 }
 389         }
 390
 391         /**
 392          * Refresh site_stats
 393          */
 394         public function refresh() {
 395                 $values = array(
 396                         'ss_row_id' => 1,
 397                         'ss_total_edits' => ( $this->mEdits === null ? $this->edits() : $this->mEdits ),
 398                         'ss_good_articles' => ( $this->mArticles === null ? $this->articles() : $this->mArticles ),
 399                         'ss_total_pages' => ( $this->mPages === null ? $this->pages() : $this->mPages ),
 400                         'ss_users' => ( $this->mUsers === null ? $this->users() : $this->mUsers ),
 401                         'ss_images' => ( $this->mFiles === null ? $this->files() : $this->mFiles ),
 402                 );
 403
 404                 $dbw = wfGetDB( DB_MASTER );
 405                 $dbw->upsert( 'site_stats', $values, array( 'ss_row_id' ), $values, __METHOD__ );
 406         }
 407 }