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 $wgUser, $wgParser, $wgUseTidy;
  64
  65                 $reportingInterval = 100;
  66                 $dbr = wfGetDB( DB_SLAVE );
  67                 $start = intval( $start );
  68
  69                 # Don't generate TeX PNGs (lack of a sensible current directory causes errors anyway)
  70                 $wgUser->setOption( 'math', MW_MATH_SOURCE );
  71
  72                 # Don't generate extension images (e.g. Timeline)
  73                 if ( method_exists( $wgParser, "clearTagHooks" ) ) {
  74                         $wgParser->clearTagHooks();
  75                 }
  76
  77                 # Don't use HTML tidy
  78                 $wgUseTidy = false;
  79
  80                 $what = $redirectsOnly ? "redirects" : "links";
  81
  82                 if ( $oldRedirectsOnly ) {
  83                         # This entire code path is cut-and-pasted from below.  Hurrah.
  84                         $res = $dbr->query(
  85                                 "SELECT page_id " .
  86                                 "FROM page " .
  87                                 "LEFT JOIN redirect ON page_id=rd_from " .
  88                                 "WHERE page_is_redirect=1 AND rd_from IS NULL AND " .
  89                                 ( $end == 0 ? "page_id >= $start"
  90                                                    : "page_id BETWEEN $start AND $end" ),
  91                                 __METHOD__
  92                         );
  93                         $num = $dbr->numRows( $res );
  94                         $this->output( "Refreshing $num old redirects from $start...\n" );
  95
  96                         $i = 0;
  97
  98                         foreach ( $res as $row ) {
  99                                 if ( !( ++$i % $reportingInterval ) ) {
 100                                         $this->output( "$i\n" );
 101                                         wfWaitForSlaves( $maxLag );
 102                                 }
 103                                 $this->fixRedirect( $row->page_id );
 104                         }
 105                 } elseif ( $newOnly ) {
 106                         $this->output( "Refreshing $what from " );
 107                         $res = $dbr->select( 'page',
 108                                 array( 'page_id' ),
 109                                 array(
 110                                         'page_is_new' => 1,
 111                                         "page_id >= $start" ),
 112                                 __METHOD__
 113                         );
 114                         $num = $dbr->numRows( $res );
 115                         $this->output( "$num new articles...\n" );
 116
 117                         $i = 0;
 118                         foreach ( $res as $row ) {
 119                                 if ( !( ++$i % $reportingInterval ) ) {
 120                                         $this->output( "$i\n" );
 121                                         wfWaitForSlaves( $maxLag );
 122                                 }
 123                                 if ( $redirectsOnly )
 124                                         $this->fixRedirect( $row->page_id );
 125                                 else
 126                                         self::fixLinksFromArticle( $row->page_id );
 127                         }
 128                 } else {
 129                         if ( !$end ) {
 130                                 $maxPage = $dbr->selectField( 'page', 'max(page_id)', false );
 131                                 $maxRD = $dbr->selectField( 'redirect', 'max(rd_from)', false );
 132                                 $end = max( $maxPage, $maxRD );
 133                         }
 134                         $this->output( "Refreshing redirects table.\n" );
 135                         $this->output( "Starting from page_id $start of $end.\n" );
 136
 137                         for ( $id = $start; $id <= $end; $id++ ) {
 138
 139                                 if ( !( $id % $reportingInterval ) ) {
 140                                         $this->output( "$id\n" );
 141                                         wfWaitForSlaves( $maxLag );
 142                                 }
 143                                 $this->fixRedirect( $id );
 144                         }
 145
 146                         if ( !$redirectsOnly ) {
 147                                 $this->output( "Refreshing links table.\n" );
 148                                 $this->output( "Starting from page_id $start of $end.\n" );
 149
 150                                 for ( $id = $start; $id <= $end; $id++ ) {
 151
 152                                         if ( !( $id % $reportingInterval ) ) {
 153                                                 $this->output( "$id\n" );
 154                                                 wfWaitForSlaves( $maxLag );
 155                                         }
 156                                         self::fixLinksFromArticle( $id );
 157                                 }
 158                         }
 159                 }
 160         }
 161
 162         /**
 163          * Update the redirect entry for a given page
 164          * @param $id int The page_id of the redirect
 165          */
 166         private function fixRedirect( $id ) {
 167                 $title = Title::newFromID( $id );
 168                 $dbw = wfGetDB( DB_MASTER );
 169
 170                 if ( is_null( $title ) ) {
 171                         // This page doesn't exist (any more)
 172                         // Delete any redirect table entry for it
 173                         $dbw->delete( 'redirect', array( 'rd_from' => $id ),
 174                                 __METHOD__ );
 175                         return;
 176                 }
 177                 $article = new Article( $title );
 178
 179                 $rt = $article->followRedirect();
 180
 181                 if ( !$rt || !is_object( $rt ) ) {
 182                         // $title is not a redirect
 183                         // Delete any redirect table entry for it
 184                         $dbw->delete( 'redirect', array( 'rd_from' => $id ),
 185                                 __METHOD__ );
 186                 } else {
 187                         $article->updateRedirectOn( $dbw, $rt );
 188                 }
 189         }
 190
 191         /**
 192          * Run LinksUpdate for all links on a given page_id
 193          * @param $id int The page_id
 194          */
 195         public static function fixLinksFromArticle( $id ) {
 196                 global $wgParser;
 197
 198                 $title = Title::newFromID( $id );
 199                 $dbw = wfGetDB( DB_MASTER );
 200
 201                 LinkCache::singleton()->clear();
 202
 203                 if ( is_null( $title ) ) {
 204                         return;
 205                 }
 206                 $dbw->begin();
 207
 208                 $revision = Revision::newFromTitle( $title );
 209                 if ( !$revision ) {
 210                         return;
 211                 }
 212
 213                 $options = new ParserOptions;
 214                 $parserOutput = $wgParser->parse( $revision->getText(), $title, $options, true, true, $revision->getId() );
 215                 $update = new LinksUpdate( $title, $parserOutput, false );
 216                 $update->doUpdate();
 217                 $dbw->commit();
 218         }
 219
 220         /*
 221          * Removes non-existing links from pages from pagelinks, imagelinks,
 222          * categorylinks, templatelinks and externallinks tables.
 223          *
 224          * @param $maxLag
 225          * @param $batchSize The size of deletion batches
 226          *
 227          * @author Merlijn van Deen <valhallasw@arctus.nl>
 228          */
 229         private function deleteLinksFromNonexistent( $maxLag = 0, $batchSize = 100 ) {
 230                 wfWaitForSlaves( $maxLag );
 231
 232                 $dbw = wfGetDB( DB_MASTER );
 233
 234                 $lb = wfGetLBFactory()->newMainLB();
 235                 $dbr = $lb->getConnection( DB_SLAVE );
 236                 $dbr->bufferResults( false );
 237
 238                 $linksTables = array( // table name => page_id field
 239                         'pagelinks' => 'pl_from',
 240                         'imagelinks' => 'il_from',
 241                         'categorylinks' => 'cl_from',
 242                         'templatelinks' => 'tl_from',
 243                         'externallinks' => 'el_from',
 244                 );
 245
 246                 foreach ( $linksTables as $table => $field ) {
 247                         $this->output( "Retrieving illegal entries from $table... " );
 248
 249                         // SELECT DISTINCT( $field ) FROM $table LEFT JOIN page ON $field=page_id WHERE page_id IS NULL;
 250                         $results = $dbr->select( array( $table, 'page' ),
 251                                                   $field,
 252                                                   array( 'page_id' => null ),
 253                                                   __METHOD__,
 254                                                   'DISTINCT',
 255                                                   array( 'page' => array( 'LEFT JOIN', "$field=page_id" ) )
 256                         );
 257
 258                         $counter = 0;
 259                         $list = array();
 260                         $this->output( "0.." );
 261
 262                         foreach ( $results as $row ) {
 263                                 $counter++;
 264                                 $list[] = $row->$field;
 265                                 if ( ( $counter % $batchSize ) == 0 ) {
 266                                         wfWaitForSlaves( 5 );
 267                                         $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 268
 269                                         $this->output( $counter . ".." );
 270                                         $list = array();
 271                                 }
 272                         }
 273                         $this->output( $counter );
 274                         if ( count( $list ) > 0 ) {
 275                                 $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 276                         }
 277                         $this->output( "\n" );
 278                 }
 279                 $lb->closeAll();
 280         }
 281 }
 282
 283 $maintClass = 'RefreshLinks';
 284 require_once( RUN_MAINTENANCE_IF_MAIN );