Merge "[FileRepo] Split out store/purge functions for thumbnails and made them skip...
[lhc/web/wiklou.git] / includes / SiteStats.php
index 0a7359b..abb1130 100644 (file)
@@ -5,7 +5,7 @@
  */
 class SiteStats {
        static $row, $loaded = false;
-       static $admins, $jobs;
+       static $jobs;
        static $pageCount = array();
        static $groupMemberCounts = array();
 
@@ -13,6 +13,9 @@ class SiteStats {
                self::load( true );
        }
 
+       /**
+        * @param $recache bool
+        */
        static function load( $recache = false ) {
                if ( self::$loaded && !$recache ) {
                        return;
@@ -32,6 +35,9 @@ class SiteStats {
                self::$loaded = true;
        }
 
+       /**
+        * @return Bool|ResultWrapper
+        */
        static function loadAndLazyInit() {
                wfDebug( __METHOD__ . ": reading site_stats from slave\n" );
                $row = self::doLoad( wfGetDB( DB_SLAVE ) );
@@ -49,7 +55,7 @@ class SiteStats {
                        // clean schema with mwdumper.
                        wfDebug( __METHOD__ . ": initializing damaged or missing site_stats\n" );
 
-                       SiteStatsInit::doAllAndCommit( false );
+                       SiteStatsInit::doAllAndCommit( wfGetDB( DB_SLAVE ) );
 
                        $row = self::doLoad( wfGetDB( DB_MASTER ) );
                }
@@ -60,67 +66,88 @@ class SiteStats {
                return $row;
        }
 
+       /**
+        * @param $db DatabaseBase
+        * @return Bool|ResultWrapper
+        */
        static function doLoad( $db ) {
                return $db->selectRow( 'site_stats', '*', false, __METHOD__ );
        }
 
+       /**
+        * @return int
+        */
        static function views() {
                self::load();
                return self::$row->ss_total_views;
        }
 
+       /**
+        * @return int
+        */
        static function edits() {
                self::load();
                return self::$row->ss_total_edits;
        }
 
+       /**
+        * @return int
+        */
        static function articles() {
                self::load();
                return self::$row->ss_good_articles;
        }
 
+       /**
+        * @return int
+        */
        static function pages() {
                self::load();
                return self::$row->ss_total_pages;
        }
 
+       /**
+        * @return int
+        */
        static function users() {
                self::load();
                return self::$row->ss_users;
        }
 
+       /**
+        * @return int
+        */
        static function activeUsers() {
                self::load();
                return self::$row->ss_active_users;
        }
 
+       /**
+        * @return int
+        */
        static function images() {
                self::load();
                return self::$row->ss_images;
        }
 
-       /**
-        * @deprecated Use self::numberingroup('sysop') instead
-        */
-       static function admins() {
-               wfDeprecated(__METHOD__);
-               return self::numberingroup('sysop');
-       }
-
        /**
         * Find the number of users in a given user group.
-        * @param string $group Name of group
-        * @return int
+        * @param $group String: name of group
+        * @return Integer
         */
-       static function numberingroup($group) {
+       static function numberingroup( $group ) {
                if ( !isset( self::$groupMemberCounts[$group] ) ) {
                        global $wgMemc;
                        $key = wfMemcKey( 'SiteStats', 'groupcounts', $group );
                        $hit = $wgMemc->get( $key );
                        if ( !$hit ) {
                                $dbr = wfGetDB( DB_SLAVE );
-                               $hit = $dbr->selectField( 'user_groups', 'COUNT(*)',
-                                       array( 'ug_group' => $group ), __METHOD__ );
+                               $hit = $dbr->selectField(
+                                       'user_groups',
+                                       'COUNT(*)',
+                                       array( 'ug_group' => $group ),
+                                       __METHOD__
+                               );
                                $wgMemc->set( $key, $hit, 3600 );
                        }
                        self::$groupMemberCounts[$group] = $hit;
@@ -128,44 +155,62 @@ class SiteStats {
                return self::$groupMemberCounts[$group];
        }
 
+       /**
+        * @return int
+        */
        static function jobs() {
                if ( !isset( self::$jobs ) ) {
                        $dbr = wfGetDB( DB_SLAVE );
-                       self::$jobs = $dbr->estimateRowCount('job');
+                       self::$jobs = $dbr->estimateRowCount( 'job' );
                        /* Zero rows still do single row read for row that doesn't exist, but people are annoyed by that */
-                       if (self::$jobs == 1) {
+                       if ( self::$jobs == 1 ) {
                                self::$jobs = 0;
                        }
                }
                return self::$jobs;
        }
 
+       /**
+        * @param $ns int
+        *
+        * @return int
+        */
        static function pagesInNs( $ns ) {
                wfProfileIn( __METHOD__ );
                if( !isset( self::$pageCount[$ns] ) ) {
                        $dbr = wfGetDB( DB_SLAVE );
-                       $pageCount[$ns] = (int)$dbr->selectField( 'page', 'COUNT(*)', array( 'page_namespace' => $ns ), __METHOD__ );
+                       self::$pageCount[$ns] = (int)$dbr->selectField(
+                               'page',
+                               'COUNT(*)',
+                               array( 'page_namespace' => $ns ),
+                               __METHOD__
+                       );
                }
                wfProfileOut( __METHOD__ );
-               return $pageCount[$ns];
+               return self::$pageCount[$ns];
        }
 
-       /** Is the provided row of site stats sane, or should it be regenerated? */
+       /**
+        * Is the provided row of site stats sane, or should it be regenerated?
+        *
+        * @param $row
+        *
+        * @return bool
+        */
        private static function isSane( $row ) {
                if(
                        $row === false
-                       or $row->ss_total_pages < $row->ss_good_articles
-                       or $row->ss_total_edits < $row->ss_total_pages
-                       or $row->ss_users       < $row->ss_admins
+                       || $row->ss_total_pages < $row->ss_good_articles
+                       || $row->ss_total_edits < $row->ss_total_pages
                ) {
                        return false;
                }
                // Now check for underflow/overflow
                foreach( array( 'total_views', 'total_edits', 'good_articles',
-               'total_pages', 'users', 'admins', 'images' ) as $member ) {
+               'total_pages', 'users', 'images' ) as $member ) {
                        if(
-                                  $row->{"ss_$member"} > 2000000000
-                               or $row->{"ss_$member"} < 0
+                               $row->{"ss_$member"} > 2000000000
+                               || $row->{"ss_$member"} < 0
                        ) {
                                return false;
                        }
@@ -174,11 +219,10 @@ class SiteStats {
        }
 }
 
-
 /**
- *
+ * Class for handling updates to the site_stats table
  */
-class SiteStatsUpdate {
+class SiteStatsUpdate implements DeferrableUpdate {
 
        var $mViews, $mEdits, $mGood, $mPages, $mUsers;
 
@@ -190,6 +234,11 @@ class SiteStatsUpdate {
                $this->mUsers = $users;
        }
 
+       /**
+        * @param $sql
+        * @param $field
+        * @param $delta
+        */
        function appendUpdate( &$sql, $field, $delta ) {
                if ( $delta ) {
                        if ( $sql ) {
@@ -204,7 +253,6 @@ class SiteStatsUpdate {
        }
 
        function doUpdate() {
-               $fname = 'SiteStatsUpdate::doUpdate';
                $dbw = wfGetDB( DB_MASTER );
 
                $updates = '';
@@ -220,22 +268,37 @@ class SiteStatsUpdate {
                        $sql = "UPDATE $site_stats SET $updates";
 
                        # Need a separate transaction because this a global lock
-                       $dbw->begin();
-                       $dbw->query( $sql, $fname );
-                       $dbw->commit();
+                       $dbw->begin( __METHOD__ );
+                       $dbw->query( $sql, __METHOD__ );
+                       $dbw->commit( __METHOD__ );
                }
        }
 
+       /**
+        * @param $dbw DatabaseBase
+        * @return bool|mixed
+        */
        public static function cacheUpdate( $dbw ) {
-               $dbr = wfGetDB( DB_SLAVE, array( 'SpecialStatistics', 'vslow') );
+               global $wgActiveUserDays;
+               $dbr = wfGetDB( DB_SLAVE, array( 'SpecialStatistics', 'vslow' ) );
                # Get non-bot users than did some recent action other than making accounts.
                # If account creation is included, the number gets inflated ~20+ fold on enwiki.
-               $activeUsers = $dbr->selectField( 'recentchanges', 'COUNT( DISTINCT rc_user_text )',
-                       array( 'rc_user != 0', 'rc_bot' => 0, "rc_log_type != 'newusers' OR rc_log_type IS NULL" ),
-                       __METHOD__ );
-               $dbw->update( 'site_stats',
-                       array( 'ss_active_users' => intval($activeUsers) ),
-                       array( 'ss_row_id' => 1 ), __METHOD__
+               $activeUsers = $dbr->selectField(
+                       'recentchanges',
+                       'COUNT( DISTINCT rc_user_text )',
+                       array(
+                               'rc_user != 0',
+                               'rc_bot' => 0,
+                               "rc_log_type != 'newusers' OR rc_log_type IS NULL",
+                               "rc_timestamp >= '{$dbw->timestamp( wfTimestamp( TS_UNIX ) - $wgActiveUserDays*24*3600 )}'",
+                       ),
+                       __METHOD__
+               );
+               $dbw->update(
+                       'site_stats',
+                       array( 'ss_active_users' => intval( $activeUsers ) ),
+                       array( 'ss_row_id' => 1 ),
+                       __METHOD__
                );
                return $activeUsers;
        }
@@ -246,7 +309,7 @@ class SiteStatsUpdate {
  */
 class SiteStatsInit {
 
-       // Db connection
+       // Database connection
        private $db;
 
        // Various stats
@@ -254,15 +317,21 @@ class SiteStatsInit {
 
        /**
         * Constructor
-        * @param $useMaster bool Whether to use the master db
+        * @param $database Boolean or DatabaseBase:
+        * - Boolean: whether to use the master DB
+        * - DatabaseBase: database connection to use
         */
-       public function __construct( $useMaster = false ) {
-               $this->db = wfGetDB( $useMaster ? DB_MASTER : DB_SLAVE );
+       public function __construct( $database = false ) {
+               if ( $database instanceof DatabaseBase ) {
+                       $this->db = $database;
+               } else {
+                       $this->db = wfGetDB( $database ? DB_MASTER : DB_SLAVE );
+               }
        }
 
        /**
         * Count the total number of edits
-        * @return int
+        * @return Integer
         */
        public function edits() {
                $this->mEdits = $this->db->selectField( 'revision', 'COUNT(*)', '', __METHOD__ );
@@ -271,36 +340,56 @@ class SiteStatsInit {
        }
 
        /**
-        * Count pages in article space
-        * @return int
+        * Count pages in article space(s)
+        * @return Integer
         */
        public function articles() {
-               global $wgContentNamespaces;
-               $this->mArticles = $this->db->selectField( 'page', 'COUNT(*)', array( 'page_namespace' => $wgContentNamespaces, 'page_is_redirect' => 0, 'page_len > 0' ), __METHOD__ );
+               global $wgArticleCountMethod;
+
+               $tables = array( 'page' );
+               $conds = array(
+                       'page_namespace' => MWNamespace::getContentNamespaces(),
+                       'page_is_redirect' => 0,
+               );
+
+               if ( $wgArticleCountMethod == 'link' ) {
+                       $tables[] = 'pagelinks';
+                       $conds[] = 'pl_from=page_id';
+               } elseif ( $wgArticleCountMethod == 'comma' ) {
+                       // To make a correct check for this, we would need, for each page,
+                       // to load the text, maybe uncompress it, maybe decode it and then
+                       // check if there's one comma.
+                       // But one thing we are sure is that if the page is empty, it can't
+                       // contain a comma :)
+                       $conds[] = 'page_len > 0';
+               }
+
+               $this->mArticles = $this->db->selectField( $tables, 'COUNT(DISTINCT page_id)',
+                       $conds, __METHOD__ );
                return $this->mArticles;
        }
 
        /**
         * Count total pages
-        * @return int
+        * @return Integer
         */
        public function pages() {
                $this->mPages = $this->db->selectField( 'page', 'COUNT(*)', '', __METHOD__ );
                return $this->mPages;
        }
-       
+
        /**
         * Count total users
-        * @return int
+        * @return Integer
         */
        public function users() {
                $this->mUsers = $this->db->selectField( 'user', 'COUNT(*)', '', __METHOD__ );
                return $this->mUsers;
        }
-       
+
        /**
         * Count views
-        * @return int
+        * @return Integer
         */
        public function views() {
                $this->mViews = $this->db->selectField( 'page', 'SUM(page_counter)', '', __METHOD__ );
@@ -309,7 +398,7 @@ class SiteStatsInit {
 
        /**
         * Count total files
-        * @return int
+        * @return Integer
         */
        public function files() {
                $this->mFiles = $this->db->selectField( 'image', 'COUNT(*)', '', __METHOD__ );
@@ -318,14 +407,22 @@ class SiteStatsInit {
 
        /**
         * Do all updates and commit them. More or less a replacement
-        * for the original initStats, but without the calls to wfOut()
-        * @param $update bool Whether to update the current stats or write fresh
-        * @param $noViews bool When true, do not update the number of page views
-        * @param $activeUsers Whether to update the number of active users
+        * for the original initStats, but without output.
+        *
+        * @param $database DatabaseBase|bool
+        * - Boolean: whether to use the master DB
+        * - DatabaseBase: database connection to use
+        * @param $options Array of options, may contain the following values
+        * - update Boolean: whether to update the current stats (true) or write fresh (false) (default: false)
+        * - views Boolean: when true, do not update the number of page views (default: true)
+        * - activeUsers Boolean: whether to update the number of active users (default: false)
         */
-       public static function doAllAndCommit( $update, $noViews = false, $activeUsers = false ) {
+       public static function doAllAndCommit( $database, array $options = array() ) {
+               $options += array( 'update' => false, 'views' => true, 'activeUsers' => false );
+
                // Grab the object and count everything
-               $counter = new SiteStatsInit( false );
+               $counter = new SiteStatsInit( $database );
+
                $counter->edits();
                $counter->articles();
                $counter->pages();
@@ -333,18 +430,21 @@ class SiteStatsInit {
                $counter->files();
 
                // Only do views if we don't want to not count them
-               if( !$noViews )
+               if( $options['views'] ) {
                        $counter->views();
+               }
 
                // Update/refresh
-               if( $update )
+               if( $options['update'] ) {
                        $counter->update();
-               else
+               } else {
                        $counter->refresh();
+               }
 
                // Count active users if need be
-               if( $activeUsers )
+               if( $options['activeUsers'] ) {
                        SiteStatsUpdate::cacheUpdate( wfGetDB( DB_MASTER ) );
+               }
        }
 
        /**
@@ -369,15 +469,16 @@ class SiteStatsInit {
 
        /**
         * Return three arrays of params for the db queries
-        * @return array
+        * @return Array
         */
        private function getDbParams() {
-               $values = array( 'ss_total_edits' => $this->mEdits,
-                                               'ss_good_articles' => $this->mArticles,
-                                               'ss_total_pages' => $this->mPages,
-                                               'ss_users' => $this->mUsers,
-                                               'ss_admins' => SiteStats::numberingroup( 'sysop' ), // @todo make this go away
-                                               'ss_images' => $this->mFiles );
+               $values = array(
+                       'ss_total_edits' => $this->mEdits,
+                       'ss_good_articles' => $this->mArticles,
+                       'ss_total_pages' => $this->mPages,
+                       'ss_users' => $this->mUsers,
+                       'ss_images' => $this->mFiles
+               );
                $conds = array( 'ss_row_id' => 1 );
                $views = array( 'ss_total_views' => $this->mViews );
                return array( $values, $conds, $views );