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  * @file
  21  * @ingroup Maintenance
  22  */
  23
  24 require_once __DIR__ . '/Maintenance.php';
  25
  26 /**
  27  * Maintenance script to refresh link tables.
  28  *
  29  * @ingroup Maintenance
  30  */
  31 class RefreshLinks extends Maintenance {
  32         public function __construct() {
  33                 parent::__construct();
  34                 $this->mDescription = "Refresh link tables";
  35                 $this->addOption( 'dfn-only', 'Delete links from nonexistent articles only' );
  36                 $this->addOption( 'new-only', 'Only affect articles with just a single edit' );
  37                 $this->addOption( 'redirects-only', 'Only fix redirects, not all links' );
  38                 $this->addOption( 'old-redirects-only', 'Only fix redirects with no redirect table entry' );
  39                 $this->addOption( 'm', 'Maximum replication lag', false, true );
  40                 $this->addOption( 'e', 'Last page id to refresh', false, true );
  41                 $this->addArg( 'start', 'Page_id to start from, default 1', false );
  42                 $this->setBatchSize( 100 );
  43         }
  44
  45         public function execute() {
  46                 $max = $this->getOption( 'm', 0 );
  47                 if ( !$this->hasOption( 'dfn-only' ) ) {
  48                         $start = $this->getArg( 0, 1 );
  49                         $new = $this->getOption( 'new-only', false );
  50                         $end = $this->getOption( 'e', 0 );
  51                         $redir = $this->getOption( 'redirects-only', false );
  52                         $oldRedir = $this->getOption( 'old-redirects-only', false );
  53                         $this->doRefreshLinks( $start, $new, $max, $end, $redir, $oldRedir );
  54                 }
  55                 $this->deleteLinksFromNonexistent( $max, $this->mBatchSize );
  56         }
  57
  58         /**
  59          * Do the actual link refreshing.
  60          * @param $start int Page_id to start from
  61          * @param $newOnly bool Only do pages with 1 edit
  62          * @param $maxLag int Max DB replication lag
  63          * @param $end int Page_id to stop at
  64          * @param $redirectsOnly bool Only fix redirects
  65          * @param $oldRedirectsOnly bool Only fix redirects without redirect entries
  66          */
  67         private function doRefreshLinks( $start, $newOnly = false, $maxLag = false,
  68                                                 $end = 0, $redirectsOnly = false, $oldRedirectsOnly = false ) {
  69                 global $wgParser, $wgUseTidy;
  70
  71                 $reportingInterval = 100;
  72                 $dbr = wfGetDB( DB_SLAVE );
  73                 $start = intval( $start );
  74
  75                 // Give extensions a chance to optimize settings
  76                 wfRunHooks( 'MaintenanceRefreshLinksInit', array( $this ) );
  77
  78                 # Don't generate extension images (e.g. Timeline)
  79                 $wgParser->clearTagHooks();
  80
  81                 # Don't use HTML tidy
  82                 $wgUseTidy = false;
  83
  84                 $what = $redirectsOnly ? "redirects" : "links";
  85
  86                 if ( $oldRedirectsOnly ) {
  87                         # This entire code path is cut-and-pasted from below.  Hurrah.
  88
  89                         $conds = array(
  90                                 "page_is_redirect=1",
  91                                 "rd_from IS NULL"
  92                         );
  93
  94                         if ( $end == 0 ) {
  95                                 $conds[] = "page_id >= $start";
  96                         } else {
  97                                 $conds[] = "page_id BETWEEN $start AND $end";
  98                         }
  99
 100                         $res = $dbr->select(
 101                                 array( 'page', 'redirect' ),
 102                                 'page_id',
 103                                 $conds,
 104                                 __METHOD__,
 105                                 array(),
 106                                 array( 'redirect' => array( "LEFT JOIN", "page_id=rd_from" ) )
 107                         );
 108                         $num = $res->numRows();
 109                         $this->output( "Refreshing $num old redirects from $start...\n" );
 110
 111                         $i = 0;
 112
 113                         foreach ( $res as $row ) {
 114                                 if ( !( ++$i % $reportingInterval ) ) {
 115                                         $this->output( "$i\n" );
 116                                         wfWaitForSlaves();
 117                                 }
 118                                 $this->fixRedirect( $row->page_id );
 119                         }
 120                 } elseif ( $newOnly ) {
 121                         $this->output( "Refreshing $what from " );
 122                         $res = $dbr->select( 'page',
 123                                 array( 'page_id' ),
 124                                 array(
 125                                         'page_is_new' => 1,
 126                                         "page_id >= $start" ),
 127                                 __METHOD__
 128                         );
 129                         $num = $res->numRows();
 130                         $this->output( "$num new articles...\n" );
 131
 132                         $i = 0;
 133                         foreach ( $res as $row ) {
 134                                 if ( !( ++$i % $reportingInterval ) ) {
 135                                         $this->output( "$i\n" );
 136                                         wfWaitForSlaves();
 137                                 }
 138                                 if ( $redirectsOnly ) {
 139                                         $this->fixRedirect( $row->page_id );
 140                                 } else {
 141                                         self::fixLinksFromArticle( $row->page_id );
 142                                 }
 143                         }
 144                 } else {
 145                         if ( !$end ) {
 146                                 $maxPage = $dbr->selectField( 'page', 'max(page_id)', false );
 147                                 $maxRD = $dbr->selectField( 'redirect', 'max(rd_from)', false );
 148                                 $end = max( $maxPage, $maxRD );
 149                         }
 150                         $this->output( "Refreshing redirects table.\n" );
 151                         $this->output( "Starting from page_id $start of $end.\n" );
 152
 153                         for ( $id = $start; $id <= $end; $id++ ) {
 154
 155                                 if ( !( $id % $reportingInterval ) ) {
 156                                         $this->output( "$id\n" );
 157                                         wfWaitForSlaves();
 158                                 }
 159                                 $this->fixRedirect( $id );
 160                         }
 161
 162                         if ( !$redirectsOnly ) {
 163                                 $this->output( "Refreshing links table.\n" );
 164                                 $this->output( "Starting from page_id $start of $end.\n" );
 165
 166                                 for ( $id = $start; $id <= $end; $id++ ) {
 167
 168                                         if ( !( $id % $reportingInterval ) ) {
 169                                                 $this->output( "$id\n" );
 170                                                 wfWaitForSlaves();
 171                                         }
 172                                         self::fixLinksFromArticle( $id );
 173                                 }
 174                         }
 175                 }
 176         }
 177
 178         /**
 179          * Update the redirect entry for a given page.
 180          *
 181          * This methods bypasses the "redirect" table to get the redirect target,
 182          * and parses the page's content to fetch it. This allows to be sure that
 183          * the redirect target is up to date and valid.
 184          * This is particularly useful when modifying namespaces to be sure the
 185          * entry in the "redirect" table points to the correct page and not to an
 186          * invalid one.
 187          *
 188          * @param $id int The page ID to check
 189          */
 190         private function fixRedirect( $id ) {
 191                 $page = WikiPage::newFromID( $id );
 192                 $dbw = wfGetDB( DB_MASTER );
 193
 194                 if ( $page === null ) {
 195                         // This page doesn't exist (any more)
 196                         // Delete any redirect table entry for it
 197                         $dbw->delete( 'redirect', array( 'rd_from' => $id ),
 198                                 __METHOD__ );
 199                         return;
 200                 }
 201
 202                 $rt = null;
 203                 $content = $page->getContent( Revision::RAW );
 204                 if ( $content !== null ) {
 205                         $rt = $content->getUltimateRedirectTarget();
 206                 }
 207
 208                 if ( $rt === null ) {
 209                         // The page is not a redirect
 210                         // Delete any redirect table entry for it
 211                         $dbw->delete( 'redirect', array( 'rd_from' => $id ), __METHOD__ );
 212                         $fieldValue = 0;
 213                 } else {
 214                         $page->insertRedirectEntry( $rt );
 215                         $fieldValue = 1;
 216                 }
 217
 218                 // Update the page table to be sure it is an a consistent state
 219                 $dbw->update( 'page', array( 'page_is_redirect' => $fieldValue ),
 220                         array( 'page_id' => $id ), __METHOD__ );
 221         }
 222
 223         /**
 224          * Run LinksUpdate for all links on a given page_id
 225          * @param $id int The page_id
 226          */
 227         public static function fixLinksFromArticle( $id ) {
 228                 $page = WikiPage::newFromID( $id );
 229
 230                 LinkCache::singleton()->clear();
 231
 232                 if ( $page === null ) {
 233                         return;
 234                 }
 235
 236                 $content = $page->getContent( Revision::RAW );
 237                 if ( $content === null ) {
 238                         return;
 239                 }
 240
 241                 $dbw = wfGetDB( DB_MASTER );
 242                 $dbw->begin( __METHOD__ );
 243
 244                 $updates = $content->getSecondaryDataUpdates( $page->getTitle() );
 245                 DataUpdate::runUpdates( $updates );
 246
 247                 $dbw->commit( __METHOD__ );
 248         }
 249
 250         /**
 251          * Removes non-existing links from pages from pagelinks, imagelinks,
 252          * categorylinks, templatelinks, externallinks, interwikilinks, langlinks and redirect tables.
 253          *
 254          * @param $maxLag int
 255          * @param $batchSize int The size of deletion batches
 256          *
 257          * @author Merlijn van Deen <valhallasw@arctus.nl>
 258          */
 259         private function deleteLinksFromNonexistent( $maxLag = 0, $batchSize = 100 ) {
 260                 wfWaitForSlaves();
 261
 262                 $dbw = wfGetDB( DB_MASTER );
 263
 264                 $lb = wfGetLBFactory()->newMainLB();
 265                 $dbr = $lb->getConnection( DB_SLAVE );
 266                 $dbr->bufferResults( false );
 267
 268                 $linksTables = array( // table name => page_id field
 269                         'pagelinks' => 'pl_from',
 270                         'imagelinks' => 'il_from',
 271                         'categorylinks' => 'cl_from',
 272                         'templatelinks' => 'tl_from',
 273                         'externallinks' => 'el_from',
 274                         'iwlinks' => 'iwl_from',
 275                         'langlinks' => 'll_from',
 276                         'redirect' => 'rd_from',
 277                         'page_props' => 'pp_page',
 278                 );
 279
 280                 foreach ( $linksTables as $table => $field ) {
 281                         $this->output( "Retrieving illegal entries from $table... " );
 282
 283                         // SELECT DISTINCT( $field ) FROM $table LEFT JOIN page ON $field=page_id WHERE page_id IS NULL;
 284                         $results = $dbr->select(
 285                                 array( $table, 'page' ),
 286                                 $field,
 287                                 array( 'page_id' => null ),
 288                                 __METHOD__,
 289                                 'DISTINCT',
 290                                 array( 'page' => array( 'LEFT JOIN', "$field=page_id" ) )
 291                         );
 292
 293                         $counter = 0;
 294                         $list = array();
 295                         $this->output( "0.." );
 296                         foreach ( $results as $row ) {
 297                                 $counter++;
 298                                 $list[] = $row->$field;
 299                                 if ( ( $counter % $batchSize ) == 0 ) {
 300                                         wfWaitForSlaves();
 301                                         $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 302
 303                                         $this->output( $counter . ".." );
 304                                         $list = array();
 305                                 }
 306                         }
 307                         $this->output( $counter );
 308                         if ( count( $list ) > 0 ) {
 309                                 $dbw->delete( $table, array( $field => $list ), __METHOD__ );
 310                         }
 311                         $this->output( "\n" );
 312                         wfWaitForSlaves();
 313                 }
 314                 $lb->closeAll();
 315         }
 316 }
 317
 318 $maintClass = 'RefreshLinks';
 319 require_once RUN_MAINTENANCE_IF_MAIN;