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                         $dbr = wfGetDB( DB_SLAVE );
  52                         self::$row = $dbr->selectRow( 'site_stats', '*', false, __METHOD__ );
  53                 }
  54
  55                 self::$loaded = true;
  56         }
  57
  58         /**
  59          * @return Bool|ResultWrapper
  60          */
  61         static function loadAndLazyInit() {
  62                 wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
  63                 $row = self::doLoad( wfGetDB( DB_SLAVE ) );
  64
  65                 if( !self::isSane( $row ) ) {
  66                         // Might have just been initialized during this request? Underflow?
  67                         wfDebug( __METHOD__ . ": site_stats damaged or missing on slave\n" );
  68                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  69                 }
  70
  71                 if( !self::isSane( $row ) ) {
  72                         // Normally the site_stats table is initialized at install time.
  73                         // Some manual construction scenarios may leave the table empty or
  74                         // broken, however, for instance when importing from a dump into a
  75                         // clean schema with mwdumper.
  76                         wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
  77
  78                         SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
  79
  80                         $row = self::doLoad( wfGetDB( DB_MASTER ) );
  81                 }
  82
  83                 if( !self::isSane( $row ) ) {
  84                         wfDebug( __METHOD__ . ": site_stats persistently nonsensical o_O\n" );
  85                 }
  86                 return $row;
  87         }
  88
  89         /**
  90          * @param $db DatabaseBase
  91          * @return Bool|ResultWrapper
  92          */
  93         static function doLoad( $db ) {
  94                 return $db->selectRow( 'site_stats', '*', false, __METHOD__ );
  95         }
  96
  97         /**
  98          * @return int
  99          */
 100         static function views() {
 101                 self::load();
 102                 return self::$row->ss_total_views;
 103         }
 104
 105         /**
 106          * @return int
 107          */
 108         static function edits() {
 109                 self::load();
 110                 return self::$row->ss_total_edits;
 111         }
 112
 113         /**
 114          * @return int
 115          */
 116         static function articles() {
 117                 self::load();
 118                 return self::$row->ss_good_articles;
 119         }
 120
 121         /**
 122          * @return int
 123          */
 124         static function pages() {
 125                 self::load();
 126                 return self::$row->ss_total_pages;
 127         }
 128
 129         /**
 130          * @return int
 131          */
 132         static function users() {
 133                 self::load();
 134                 return self::$row->ss_users;
 135         }
 136
 137         /**
 138          * @return int
 139          */
 140         static function activeUsers() {
 141                 self::load();
 142                 return self::$row->ss_active_users;
 143         }
 144
 145         /**
 146          * @return int
 147          */
 148         static function images() {
 149                 self::load();
 150                 return self::$row->ss_images;
 151         }
 152
 153         /**
 154          * Find the number of users in a given user group.
 155          * @param $group String: name of group
 156          * @return Integer
 157          */
 158         static function numberingroup( $group ) {
 159                 if ( !isset( self::$groupMemberCounts[$group] ) ) {
 160                         global $wgMemc;
 161                         $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
 162                         $hit = $wgMemc->get( $key );
 163                         if ( !$hit ) {
 164                                 $dbr = wfGetDB( DB_SLAVE );
 165                                 $hit = $dbr->selectField(
 166                                         'user_groups',
 167                                         'COUNT(*)',
 168                                         array( 'ug_group' => $group ),
 169                                         __METHOD__
 170                                 );
 171                                 $wgMemc->set( $key, $hit, 3600 );
 172                         }
 173                         self::$groupMemberCounts[$group] = $hit;
 174                 }
 175                 return self::$groupMemberCounts[$group];
 176         }
 177
 178         /**
 179          * @return int
 180          */
 181         static function jobs() {
 182                 if ( !isset( self::$jobs ) ) {
 183                         $dbr = wfGetDB( DB_SLAVE );
 184                         self::$jobs = $dbr->estimateRowCount( 'job' );
 185                         /* Zero rows still do single row read for row that doesn't exist, but people are annoyed by that */
 186                         if ( self::$jobs == 1 ) {
 187                                 self::$jobs = 0;
 188                         }
 189                 }
 190                 return self::$jobs;
 191         }
 192
 193         /**
 194          * @param $ns int
 195          *
 196          * @return int
 197          */
 198         static function pagesInNs( $ns ) {
 199                 wfProfileIn( __METHOD__ );
 200                 if( !isset( self::$pageCount[$ns] ) ) {
 201                         $dbr = wfGetDB( DB_SLAVE );
 202                         self::$pageCount[$ns] = (int)$dbr->selectField(
 203                                 'page',
 204                                 'COUNT(*)',
 205                                 array( 'page_namespace' => $ns ),
 206                                 __METHOD__
 207                         );
 208                 }
 209                 wfProfileOut( __METHOD__ );
 210                 return self::$pageCount[$ns];
 211         }
 212
 213         /**
 214          * Is the provided row of site stats sane, or should it be regenerated?
 215          *
 216          * @param $row
 217          *
 218          * @return bool
 219          */
 220         private static function isSane( $row ) {
 221                 if(
 222                         $row === false
 223                         || $row->ss_total_pages < $row->ss_good_articles
 224                         || $row->ss_total_edits < $row->ss_total_pages
 225                 ) {
 226                         return false;
 227                 }
 228                 // Now check for underflow/overflow
 229                 foreach( array( 'total_views', 'total_edits', 'good_articles',
 230                 'total_pages', 'users', 'images' ) as $member ) {
 231                         if(
 232                                 $row->{"ss_$member"} > 2000000000
 233                                 || $row->{"ss_$member"} < 0
 234                         ) {
 235                                 return false;
 236                         }
 237                 }
 238                 return true;
 239         }
 240 }
 241
 242 /**
 243  * Class for handling updates to the site_stats table
 244  */
 245 class SiteStatsUpdate implements DeferrableUpdate {
 246         protected $views = 0;
 247         protected $edits = 0;
 248         protected $pages = 0;
 249         protected $goodPages = 0;
 250         protected $users = 0;
 251         protected $images = 0;
 252
 253         // @TODO: deprecate this
 254         function __construct( $views, $edits, $good, $pages = 0, $users = 0 ) {
 255                 $this->views = $views;
 256                 $this->edits = $edits;
 257                 $this->goodPages = $good;
 258                 $this->pages = $pages;
 259                 $this->users = $users;
 260         }
 261
 262         /**
 263          * @param $deltas Array
 264          * @return SiteStatsUpdate
 265          */
 266         public static function factory( array $deltas ) {
 267                 $update = new self( 0, 0, 0 );
 268
 269                 $fields = array( 'views', 'edits', 'pages', 'goodPages', 'users', 'images' );
 270                 foreach ( $fields as $field ) {
 271                         if ( isset( $deltas[$field] ) && $deltas[$field] ) {
 272                                 $update->$field = $deltas[$field];
 273                         }
 274                 }
 275
 276                 return $update;
 277         }
 278
 279         /**
 280          * @param $sql
 281          * @param $field
 282          * @param $delta
 283          */
 284         function appendUpdate( &$sql, $field, $delta ) {
 285                 if ( $delta ) {
 286                         if ( $sql ) {
 287                                 $sql .= ',';
 288                         }
 289                         if ( $delta < 0 ) {
 290                                 $sql .= "$field=$field-1";
 291                         } else {
 292                                 $sql .= "$field=$field+1";
 293                         }
 294                 }
 295         }
 296
 297         public function doUpdate() {
 298                 $dbw = wfGetDB( DB_MASTER );
 299
 300                 $updates = '';
 301
 302                 $this->appendUpdate( $updates, 'ss_total_views', $this->views );
 303                 $this->appendUpdate( $updates, 'ss_total_edits', $this->edits );
 304                 $this->appendUpdate( $updates, 'ss_good_articles', $this->goodPages );
 305                 $this->appendUpdate( $updates, 'ss_total_pages', $this->pages );
 306                 $this->appendUpdate( $updates, 'ss_users', $this->users );
 307                 $this->appendUpdate( $updates, 'ss_images', $this->images );
 308
 309                 if ( $updates ) {
 310                         $site_stats = $dbw->tableName( 'site_stats' );
 311                         $sql = "UPDATE $site_stats SET $updates";
 312
 313                         # Need a separate transaction because this a global lock
 314                         $dbw->begin( __METHOD__ );
 315                         $dbw->query( $sql, __METHOD__ );
 316                         $dbw->commit( __METHOD__ );
 317                 }
 318         }
 319
 320         /**
 321          * @param $dbw DatabaseBase
 322          * @return bool|mixed
 323          */
 324         public static function cacheUpdate( $dbw ) {
 325                 global $wgActiveUserDays;
 326                 $dbr = wfGetDB( DB_SLAVE, array( 'SpecialStatistics', 'vslow' ) );
 327                 # Get non-bot users than did some recent action other than making accounts.
 328                 # If account creation is included, the number gets inflated ~20+ fold on enwiki.
 329                 $activeUsers = $dbr->selectField(
 330                         'recentchanges',
 331                         'COUNT( DISTINCT rc_user_text )',
 332                         array(
 333                                 'rc_user != 0',
 334                                 'rc_bot' => 0,
 335                                 "rc_log_type != 'newusers' OR rc_log_type IS NULL",
 336                                 "rc_timestamp >= '{$dbw->timestamp( wfTimestamp( TS_UNIX ) - $wgActiveUserDays*24*3600 )}'",
 337                         ),
 338                         __METHOD__
 339                 );
 340                 $dbw->update(
 341                         'site_stats',
 342                         array( 'ss_active_users' => intval( $activeUsers ) ),
 343                         array( 'ss_row_id' => 1 ),
 344                         __METHOD__
 345                 );
 346                 return $activeUsers;
 347         }
 348 }
 349
 350 /**
 351  * Class designed for counting of stats.
 352  */
 353 class SiteStatsInit {
 354
 355         // Database connection
 356         private $db;
 357
 358         // Various stats
 359         private $mEdits, $mArticles, $mPages, $mUsers, $mViews, $mFiles = 0;
 360
 361         /**
 362          * Constructor
 363          * @param $database Boolean or DatabaseBase:
 364          * - Boolean: whether to use the master DB
 365          * - DatabaseBase: database connection to use
 366          */
 367         public function __construct( $database = false ) {
 368                 if ( $database instanceof DatabaseBase ) {
 369                         $this->db = $database;
 370                 } else {
 371                         $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
 372                 }
 373         }
 374
 375         /**
 376          * Count the total number of edits
 377          * @return Integer
 378          */
 379         public function edits() {
 380                 $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
 381                 $this->mEdits += $this->db->selectField( 'archive', 'COUNT(*)', '', __METHOD__ );
 382                 return $this->mEdits;
 383         }
 384
 385         /**
 386          * Count pages in article space(s)
 387          * @return Integer
 388          */
 389         public function articles() {
 390                 global $wgArticleCountMethod;
 391
 392                 $tables = array( 'page' );
 393                 $conds = array(
 394                         'page_namespace' => MWNamespace::getContentNamespaces(),
 395                         'page_is_redirect' => 0,
 396                 );
 397
 398                 if ( $wgArticleCountMethod == 'link' ) {
 399                         $tables[] = 'pagelinks';
 400                         $conds[] = 'pl_from=page_id';
 401                 } elseif ( $wgArticleCountMethod == 'comma' ) {
 402                         // To make a correct check for this, we would need, for each page,
 403                         // to load the text, maybe uncompress it, maybe decode it and then
 404                         // check if there's one comma.
 405                         // But one thing we are sure is that if the page is empty, it can't
 406                         // contain a comma :)
 407                         $conds[] = 'page_len > 0';
 408                 }
 409
 410                 $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
 411                         $conds, __METHOD__ );
 412                 return $this->mArticles;
 413         }
 414
 415         /**
 416          * Count total pages
 417          * @return Integer
 418          */
 419         public function pages() {
 420                 $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
 421                 return $this->mPages;
 422         }
 423
 424         /**
 425          * Count total users
 426          * @return Integer
 427          */
 428         public function users() {
 429                 $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
 430                 return $this->mUsers;
 431         }
 432
 433         /**
 434          * Count views
 435          * @return Integer
 436          */
 437         public function views() {
 438                 $this->mViews = $this->db->selectField( 'page', 'SUM(page_counter)', '', __METHOD__ );
 439                 return $this->mViews;
 440         }
 441
 442         /**
 443          * Count total files
 444          * @return Integer
 445          */
 446         public function files() {
 447                 $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
 448                 return $this->mFiles;
 449         }
 450
 451         /**
 452          * Do all updates and commit them. More or less a replacement
 453          * for the original initStats, but without output.
 454          *
 455          * @param $database DatabaseBase|bool
 456          * - Boolean: whether to use the master DB
 457          * - DatabaseBase: database connection to use
 458          * @param $options Array of options, may contain the following values
 459          * - update Boolean: whether to update the current stats (true) or write fresh (false) (default: false)
 460          * - views Boolean: when true, do not update the number of page views (default: true)
 461          * - activeUsers Boolean: whether to update the number of active users (default: false)
 462          */
 463         public static function doAllAndCommit( $database, array $options = array() ) {
 464                 $options += array( 'update' => false, 'views' => true, 'activeUsers' => false );
 465
 466                 // Grab the object and count everything
 467                 $counter = new SiteStatsInit( $database );
 468
 469                 $counter->edits();
 470                 $counter->articles();
 471                 $counter->pages();
 472                 $counter->users();
 473                 $counter->files();
 474
 475                 // Only do views if we don't want to not count them
 476                 if( $options['views'] ) {
 477                         $counter->views();
 478                 }
 479
 480                 // Update/refresh
 481                 if( $options['update'] ) {
 482                         $counter->update();
 483                 } else {
 484                         $counter->refresh();
 485                 }
 486
 487                 // Count active users if need be
 488                 if( $options['activeUsers'] ) {
 489                         SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
 490                 }
 491         }
 492
 493         /**
 494          * Update the current row with the selected values
 495          */
 496         public function update() {
 497                 list( $values, $conds ) = $this->getDbParams();
 498                 $dbw = wfGetDB( DB_MASTER );
 499                 $dbw->update( 'site_stats', $values, $conds, __METHOD__ );
 500         }
 501
 502         /**
 503          * Refresh site_stats. Erase the current record and save all
 504          * the new values.
 505          */
 506         public function refresh() {
 507                 list( $values, $conds, $views ) = $this->getDbParams();
 508                 $dbw = wfGetDB( DB_MASTER );
 509                 $dbw->delete( 'site_stats', $conds, __METHOD__ );
 510                 $dbw->insert( 'site_stats', array_merge( $values, $conds, $views ), __METHOD__ );
 511         }
 512
 513         /**
 514          * Return three arrays of params for the db queries
 515          * @return Array
 516          */
 517         private function getDbParams() {
 518                 $values = array(
 519                         'ss_total_edits' => $this->mEdits,
 520                         'ss_good_articles' => $this->mArticles,
 521                         'ss_total_pages' => $this->mPages,
 522                         'ss_users' => $this->mUsers,
 523                         'ss_images' => $this->mFiles
 524                 );
 525                 $conds = array( 'ss_row_id' => 1 );
 526                 $views = array( 'ss_total_views' => $this->mViews );
 527                 return array( $values, $conds, $views );
 528         }
 529 }