maintenance/cleanupTitles.php

   1 <?php
   2 /*
   3  * Script to clean up broken, unparseable titles.
   4  *
   5  * Usage: php cleanupTitles.php [--dry-run]
   6  * Options:
   7  *   --dry-run  don't actually try moving them
   8  *
   9  * Copyright (C) 2005 Brion Vibber <brion@pobox.com>
  10  * http://www.mediawiki.org/
  11  *
  12  * This program is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU General Public License as published by
  14  * the Free Software Foundation; either version 2 of the License, or
  15  * (at your option) any later version.
  16  *
  17  * This program is distributed in the hope that it will be useful,
  18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  20  * GNU General Public License for more details.
  21  *
  22  * You should have received a copy of the GNU General Public License along
  23  * with this program; if not, write to the Free Software Foundation, Inc.,
  24  * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  25  * http://www.gnu.org/copyleft/gpl.html
  26  *
  27  * @author Brion Vibber <brion at pobox.com>
  28  * @package MediaWiki
  29  * @subpackage maintenance
  30  */
  31
  32 $options = array( 'dry-run' );
  33
  34 require_once( 'commandLine.inc' );
  35 require_once( 'FiveUpgrade.inc' );
  36
  37 class TitleCleanup extends FiveUpgrade {
  38         function TitleCleanup( $dryrun = false ) {
  39                 parent::FiveUpgrade();
  40
  41                 $this->maxLag = 10; # if slaves are lagged more than 10 secs, wait
  42                 $this->dryrun = $dryrun;
  43         }
  44
  45         function cleanup() {
  46                 $this->runTable( 'page', 'WHERE page_namespace=0',
  47                         array( &$this, 'processPage' ) );
  48         }
  49
  50         function init( $count, $table ) {
  51                 $this->processed = 0;
  52                 $this->updated = 0;
  53                 $this->count = $count;
  54                 $this->startTime = wfTime();
  55                 $this->table = $table;
  56         }
  57
  58         function progress( $updated ) {
  59                 $this->updated += $updated;
  60                 $this->processed++;
  61                 if( $this->processed % 100 != 0 ) {
  62                         return;
  63                 }
  64                 $portion = $this->processed / $this->count;
  65                 $updateRate = $this->updated / $this->processed;
  66
  67                 $now = wfTime();
  68                 $delta = $now - $this->startTime;
  69                 $estimatedTotalTime = $delta / $portion;
  70                 $eta = $this->startTime + $estimatedTotalTime;
  71
  72                 global $wgDBname;
  73                 printf( "%s %s: %6.2f%% done on %s; ETA %s [%d/%d] %.2f/sec <%.2f%% updated>\n",
  74                         $wgDBname,
  75                         wfTimestamp( TS_DB, intval( $now ) ),
  76                         $portion * 100.0,
  77                         $this->table,
  78                         wfTimestamp( TS_DB, intval( $eta ) ),
  79                         $this->processed,
  80                         $this->count,
  81                         $this->processed / $delta,
  82                         $updateRate * 100.0 );
  83                 flush();
  84         }
  85
  86         function runTable( $table, $where, $callback ) {
  87                 $fname = 'CapsCleanup::buildTable';
  88
  89                 $count = $this->dbw->selectField( $table, 'count(*)', '', $fname );
  90                 $this->init( $count, 'page' );
  91                 $this->log( "Processing $table..." );
  92
  93                 $tableName = $this->dbr->tableName( $table );
  94                 $sql = "SELECT * FROM $tableName $where";
  95                 $result = $this->dbr->query( $sql, $fname );
  96
  97                 while( $row = $this->dbr->fetchObject( $result ) ) {
  98                         $updated = call_user_func( $callback, $row );
  99                 }
 100                 $this->log( "Finished $table... $this->updated of $this->processed rows updated" );
 101                 $this->dbr->freeResult( $result );
 102         }
 103
 104         function processPage( $row ) {
 105                 global $wgContLang;
 106
 107                 $current = Title::makeTitle( $row->page_namespace, $row->page_title );
 108                 $display = $current->getPrefixedText();
 109
 110                 $verified = UtfNormal::cleanUp( $display );
 111                 $title = Title::newFromText( $verified );
 112
 113                 if( is_null( $title ) ) {
 114                         $this->log( "page $row->page_id ($display) is illegal." );
 115                         $this->moveIllegalPage( $row );
 116                         return $this->progress( 1 );
 117                 }
 118
 119                 if( !$title->equals( $current ) ) {
 120                         $this->log( "page $row->page_id ($display) doesn't match self." );
 121                         $this->moveInconsistentPage( $row, $title );
 122                         return $this->progress( 1 );
 123                 }
 124
 125                 $this->progress( 0 );
 126         }
 127
 128         function moveIllegalPage( $row ) {
 129                 $legal = 'A-Za-z0-9_/\\\\-';
 130                 $legalized = preg_replace_callback( "!([^$legal])!",
 131                         array( &$this, 'hexChar' ),
 132                         $row->page_title );
 133                 if( $legalized == '.' ) $legalized = '(dot)';
 134                 if( $legalized == '_' ) $legalized = '(space)';
 135                 $legalized = 'Broken/' . $legalized;
 136
 137                 $title = Title::newFromText( $legalized );
 138                 if( is_null( $title ) ) {
 139                         $clean = 'Broken/id:' . $row->page_id;
 140                         $this->log( "Couldn't legalize; form '$legalized' still invalid; using '$clean'" );
 141                         $title = Title::newFromText( $clean );
 142                 } elseif( $title->exists() ) {
 143                         $clean = 'Broken/id:' . $row->page_id;
 144                         $this->log( "Legalized for '$legalized' exists; using '$clean'" );
 145                         $title = Title::newFromText( $clean );
 146                 }
 147
 148                 $dest = $title->getDbKey();
 149                 if( $this->dryrun ) {
 150                         $this->log( "DRY RUN: would rename $row->page_id ($row->page_namespace,'$row->page_title') to ($row->page_namespace,'$dest')" );
 151                 } else {
 152                         $this->log( "renaming $row->page_id ($row->page_namespace,'$row->page_title') to ($row->page_namespace,'$dest')" );
 153                         $dbw =& wfGetDB( DB_MASTER );
 154                         $dbw->update( 'page',
 155                                 array( 'page_title' => $dest ),
 156                                 array( 'page_id' => $row->page_id ),
 157                                 'cleanupTitles::moveInconsistentPage' );
 158                 }
 159         }
 160
 161         function moveInconsistentPage( $row, $title ) {
 162                 if( $title->exists() || $title->getInterwiki() ) {
 163                         if( $title->getInterwiki() ) {
 164                                 $prior = $title->getPrefixedDbKey();
 165                         } else {
 166                                 $prior = $title->getDbKey();
 167                         }
 168                         $clean = 'Broken/' . $prior;
 169                         $verified = Title::makeTitleSafe( $row->page_namespace, $clean );
 170                         if( $verified->exists() ) {
 171                                 $blah = "Broken/id:" . $row->page_id;
 172                                 $this->log( "Couldn't legalize; form '$clean' exists; using '$blah'" );
 173                                 $verified = Title::makeTitleSafe( $row->page_namespace, $blah );
 174                         }
 175                         $title = $verified;
 176                 }
 177                 if( is_null( $title ) ) {
 178                         die( "Something awry; empty title.\n" );
 179                 }
 180                 $dest = $title->getDbKey();
 181                 if( $this->dryrun ) {
 182                         $this->log( "DRY RUN: would rename $row->page_id ($row->page_namespace,'$row->page_title') to ($row->page_namespace,'$dest')" );
 183                 } else {
 184                         $this->log( "renaming $row->page_id ($row->page_namespace,'$row->page_title') to ($row->page_namespace,'$dest')" );
 185                         $dbw =& wfGetDB( DB_MASTER );
 186                         $dbw->update( 'page',
 187                                 array( 'page_title' => $dest ),
 188                                 array( 'page_id' => $row->page_id ),
 189                                 'cleanupTitles::moveInconsistentPage' );
 190                 }
 191         }
 192
 193         function hexChar( $matches ) {
 194                 return sprintf( "\\x%02x", ord( $matches[1] ) );
 195         }
 196 }
 197
 198 $wgUser->setName( 'Conversion script' );
 199 $caps = new TitleCleanup( isset( $options['dry-run'] ) );
 200 $caps->cleanup();
 201
 202 ?>