maintenance/refreshLinks.php

   1 <?php
   2 /**
   3  * Refresh link tables.
   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  * @ingroup Maintenance
  21  */
  22
  23 require_once( dirname( __FILE__ ) . '/Maintenance.php' );
  24
  25 class RefreshLinks extends Maintenance {
  26         public function __construct() {
  27                 parent::__construct();
  28                 $this->mDescription = "Refresh link tables";
  29                 $this->addOption( 'dfn-only', 'Delete links from nonexistent articles only' );
  30                 $this->addOption( 'new-only', 'Only affect articles with just a single edit' );
  31                 $this->addOption( 'redirects-only', 'Only fix redirects, not all links' );
  32                 $this->addOption( 'old-redirects-only', 'Only fix redirects with no redirect table entry' );
  33                 $this->addOption( 'm', 'Maximum replication lag', false, true );
  34                 $this->addOption( 'e', 'Last page id to refresh', false, true );
  35                 $this->addArg( 'start', 'Page_id to start from, default 1', false );
  36                 $this->setBatchSize( 100 );
  37         }
  38
  39         public function execute() {
  40                 $max = $this->getOption( 'm', 0 );
  41                 if ( !$this->hasOption( 'dfn-only' ) ) {
  42                         $start = $this->getArg( 0, 1 );
  43                         $new = $this->getOption( 'new-only', false );
  44                         $end = $this->getOption( 'e', 0 );
  45                         $redir = $this->getOption( 'redirects-only', false );
  46                         $oldRedir = $this->getOption( 'old-redirects-only', false );
  47                         $this->doRefreshLinks( $start, $new, $max, $end, $redir, $oldRedir );
  48                 }
  49                 $this->deleteLinksFromNonexistent( $max, $this->mBatchSize );
  50         }
  51
  52         /**
  53          * Do the actual link refreshing.
  54          * @param $start int Page_id to start from
  55          * @param $newOnly bool Only do pages with 1 edit
  56          * @param $maxLag int Max DB replication lag
  57          * @param $end int Page_id to stop at
  58          * @param $redirectsOnly bool Only fix redirects
  59          * @param $oldRedirectsOnly bool Only fix redirects without redirect entries
  60          */
  61         private function doRefreshLinks( $start, $newOnly = false, $maxLag = false,
  62                                                 $end = 0, $redirectsOnly = false, $oldRedirectsOnly = false ) {
  63                 global $wgParser, $wgUseTidy;
  64
  65                 $reportingInterval = 100;
  66                 $dbr = wfGetDB( DB_SLAVE );
  67                 $start = intval( $start );
  68
  69                 // Give extensions a chance to optimize settings
  70                 wfRunHooks( 'MaintenanceRefreshLinksInit', array( $this ) );
  71
  72                 # Don't generate extension images (e.g. Timeline)
  73                 $wgParser->clearTagHooks();
  74
  75                 # Don't use HTML tidy
  76                 $wgUseTidy = false;
  77
  78                 $what = $redirectsOnly ? "redirects" : "links";
  79
  80                 if ( $oldRedirectsOnly ) {
  81                         # This entire code path is cut-and-pasted from below.  Hurrah.
  82
  83                         $conds = array(
  84                                 "page_is_redirect=1",
  85                                 "rd_from IS NULL"
  86                         );
  87
  88                         if ( $end == 0 ) {
  89                                 $conds[] = "page_id >= $start";
  90                         } else {
  91                                 $conds[] = "page_id BETWEEN $start AND $end";
  92                         }
  93
  94                         $res = $dbr->select(
  95                                 array( 'page', 'redirect' ),
  96                                 'page_id',
  97                                 $conds,
  98                                 __METHOD__,
  99                                 array(),
 100                                 array( 'redirect' => array( "LEFT JOIN", "page_id=rd_from" ) )
 101                         );
 102                         $num = $dbr->numRows( $res );
 103                         $this->output( "Refreshing $num old redirects from $start...\n" );
 104
 105                         $i = 0;
 106
 107                         foreach ( $res as $row ) {
 108                                 if ( !( ++$i % $reportingInterval ) ) {
 109                                         $this->output( "$i\n" );
 110                                         wfWaitForSlaves();
 111                                 }
 112                                 $this->fixRedirect( $row->page_id );
 113                         }
 114                 } elseif ( $newOnly ) {
 115                         $this->output( "Refreshing $what from " );
 116                         $res = $dbr->select( 'page',
 117                                 array( 'page_id' ),
 118                                 array(
 119                                         'page_is_new' => 1,
 120                                         "page_id >= $start" ),
 121                                 __METHOD__
 122                         );
 123                         $num = $dbr->numRows( $res );
 124                         $this->output( "$num new articles...\n" );
 125
 126                         $i = 0;
 127                         foreach ( $res as $row ) {
 128                                 if ( !( ++$i % $reportingInterval ) ) {
 129                                         $this->output( "$i\n" );
 130                                         wfWaitForSlaves();
 131                                 }
 132                                 if ( $redirectsOnly ) {
 133                                         $this->fixRedirect( $row->page_id );
 134                                 } else {
 135                                         self::fixLinksFromArticle( $row->page_id );
 136                                 }
 137                         }
 138                 } else {
 139                         if ( !$end ) {
 140                                 $maxPage = $dbr->selectField( 'page', 'max(page_id)', false );
 141                                 $maxRD = $dbr->selectField( 'redirect', 'max(rd_from)', false );
 142                                 $end = max( $maxPage, $maxRD );
 143                         }
 144                         $this->output( "Refreshing redirects table.\n" );
 145                         $this->output( "Starting from page_id $start of $end.\n" );
 146
 147                         for ( $id = $start; $id <= $end; $id++ ) {
 148
 149                                 if ( !( $id % $reportingInterval ) ) {
 150                                         $this->output( "$id\n" );
 151                                         wfWaitForSlaves();
 152                                 }
 153                                 $this->fixRedirect( $id );
 154                         }
 155
 156                         if ( !$redirectsOnly ) {
 157                                 $this->output( "Refreshing links table.\n" );
 158                                 $this->output( "Starting from page_id $start of $end.\n" );
 159
 160                                 for ( $id = $start; $id <= $end; $id++ ) {
 161
 162                                         if ( !( $id % $reportingInterval ) ) {
 163                                                 $this->output( "$id\n" );
 164                                                 wfWaitForSlaves();
 165                                         }
 166                                         self::fixLinksFromArticle( $id );
 167                                 }
 168                         }
 169                 }
 170         }
 171
 172         /**
 173          * Update the redirect entry for a given page
 174          * @param $id int The page_id of the redirect
 175          */
 176         private function fixRedirect( $id ) {
 177                 $page = WikiPage::newFromID( $id );
 178                 $dbw = wfGetDB( DB_MASTER );
 179
 180                 if ( $page === null ) {
 181                         // This page doesn't exist (any more)
 182                         // Delete any redirect table entry for it
 183                         $dbw->delete( 'redirect', array( 'rd_from' => $id ),
 184                                 __METHOD__ );
 185                         return;
 186                 }
 187
 188                 $rt = $page->getRedirectTarget();
 189
 190                 if ( $rt === null ) {
 191                         // The page is not a redirect
 192                         // Delete any redirect table entry for it
 193                         $dbw->delete( 'redirect', array( 'rd_from' => $id ),
 194                                 __METHOD__ );
 195                 }
 196         }
 197
 198         /**
 199          * Run LinksUpdate for all links on a given page_id
 200          * @param $id int The page_id
 201          */
 202         public static function fixLinksFromArticle( $id ) {
 203                 global $wgParser, $wgContLang;
 204
 205                 $page = WikiPage::newFromID( $id );
 206
 207                 LinkCache::singleton()->clear();
 208
 209                 if ( $page === null ) {
 210                         return;
 211                 }
 212
 213                 $content = $page->getContent( REVISION::RAW );
 214                 if ( null === false ) {
 215                         return;
 216                 }
 217
 218                 $dbw = wfGetDB( DB_MASTER );
 219                 $dbw->begin( __METHOD__ );
 220
 221                 $contentHandler = $content->getContentHandler();
 222                 $updates = $contentHandler->getSecondaryDataUpdates( $content, $page->getTitle() );
 223                 DataUpdate::runUpdates( $updates );
 224
 225                 $dbw->commit( __METHOD__ );
 226         }
 227
 228         /**
 229          * Removes non-existing links from pages from pagelinks, imagelinks,
 230          * categorylinks, templatelinks, externallinks, interwikilinks, langlinks and redirect tables.
 231          *
 232          * @param $maxLag int
 233          * @param $batchSize int The size of deletion batches
 234          *
 235          * @author Merlijn van Deen <valhallasw@arctus.nl>
 236          */
 237         private function deleteLinksFromNonexistent( $maxLag = 0, $batchSize = 100 ) {
 238                 wfWaitForSlaves();
 239
 240                 $dbw = wfGetDB( DB_MASTER );
 241
 242                 $lb = wfGetLBFactory()->newMainLB();
 243                 $dbr = $lb->getConnection( DB_SLAVE );
 244                 $dbr->bufferResults( false );
 245
 246                 $linksTables = array( // table name => page_id field
 247                         'pagelinks' => 'pl_from',
 248                         'imagelinks' => 'il_from',
 249                         'categorylinks' => 'cl_from',
 250                         'templatelinks' => 'tl_from',
 251                         'externallinks' => 'el_from',
 252                         'iwlinks' => 'iwl_from',
 253                         'langlinks' => 'll_from',
 254                         'redirect' => 'rd_from',
 255                         'page_props' => 'pp_page',
 256                 );
 257
 258                 foreach ( $linksTables as $table => $field ) {
 259                         $this->output( "Retrieving illegal entries from $table... " );
 260
 261                         // SELECT DISTINCT( $field ) FROM $table LEFT JOIN page ON $field=page_id WHERE page_id IS NULL;
 262                         $results = $dbr->select( array( $table, 'page' ),
 263                                                   $field,
 264                                                   array( 'page_id' => null ),
 265                                                   __METHOD__,
 266                                                   'DISTINCT',
 267                                                   array( 'page' => array( 'LEFT JOIN', "$field=page_id" ) )
 268                         );
 269
 270                         $counter = 0;
 271                         $list = array();
 272                         $this->output( "0.." );
 273                         foreach ( $results as $row ) {
 274                                 $counter++;
 275                                 $list[] = $row->$field;
 276                                 if ( ( $counter % $batchSize ) == 0 ) {
 277                                         wfWaitForSlaves();
 278                                         $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 279
 280                                         $this->output( $counter . ".." );
 281                                         $list = array();
 282                                 }
 283                         }
 284                         $this->output( $counter );
 285                         if ( count( $list ) > 0 ) {
 286                                 $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 287                         }
 288                         $this->output( "\n" );
 289                         wfWaitForSlaves();
 290                 }
 291                 $lb->closeAll();
 292         }
 293 }
 294
 295 $maintClass = 'RefreshLinks';
 296 require_once( RUN_MAINTENANCE_IF_MAIN );