maintenance/dumpIterator.php

   1 <?php
   2 /**
   3  * Take page text out of an XML dump file and perform some operation on it.
   4  * Used as a base class for CompareParsers and PreprocessDump.
   5  * We implement below the simple task of searching inside a dump.
   6  *
   7  * Copyright © 2011 Platonides
   8  * https://www.mediawiki.org/
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  *
  25  * @file
  26  * @ingroup Maintenance
  27  */
  28
  29 require_once __DIR__ . '/Maintenance.php';
  30
  31 /**
  32  * Base class for interating over a dump.
  33  *
  34  * @ingroup Maintenance
  35  */
  36 abstract class DumpIterator extends Maintenance {
  37         private $count = 0;
  38         private $startTime;
  39         /** @var string|bool|null */
  40         private $from;
  41
  42         public function __construct() {
  43                 parent::__construct();
  44                 $this->addDescription( 'Does something with a dump' );
  45                 $this->addOption( 'file', 'File with text to run.', false, true );
  46                 $this->addOption( 'dump', 'XML dump to execute all revisions.', false, true );
  47                 $this->addOption( 'from', 'Article from XML dump to start from.', false, true );
  48         }
  49
  50         public function execute() {
  51                 if ( !( $this->hasOption( 'file' ) ^ $this->hasOption( 'dump' ) ) ) {
  52                         $this->fatalError( "You must provide a file or dump" );
  53                 }
  54
  55                 $this->checkOptions();
  56
  57                 if ( $this->hasOption( 'file' ) ) {
  58                         $revision = new WikiRevision( $this->getConfig() );
  59
  60                         $revision->setText( file_get_contents( $this->getOption( 'file' ) ) );
  61                         $revision->setTitle( Title::newFromText(
  62                                 rawurldecode( basename( $this->getOption( 'file' ), '.txt' ) )
  63                         ) );
  64                         $this->from = false;
  65                         $this->handleRevision( $revision );
  66
  67                         return;
  68                 }
  69
  70                 $this->startTime = microtime( true );
  71
  72                 if ( $this->getOption( 'dump' ) == '-' ) {
  73                         $source = new ImportStreamSource( $this->getStdin() );
  74                 } else {
  75                         $this->fatalError( "Sorry, I don't support dump filenames yet. "
  76                                 . "Use - and provide it on stdin on the meantime." );
  77                 }
  78                 $importer = new WikiImporter( $source, $this->getConfig() );
  79
  80                 $importer->setRevisionCallback(
  81                         [ $this, 'handleRevision' ] );
  82                 $importer->setNoticeCallback( function ( $msg, $params ) {
  83                         echo wfMessage( $msg, $params )->text() . "\n";
  84                 } );
  85
  86                 $this->from = $this->getOption( 'from', null );
  87                 $this->count = 0;
  88                 $importer->doImport();
  89
  90                 $this->conclusions();
  91
  92                 $delta = microtime( true ) - $this->startTime;
  93                 $this->error( "Done {$this->count} revisions in " . round( $delta, 2 ) . " seconds " );
  94                 if ( $delta > 0 ) {
  95                         $this->error( round( $this->count / $delta, 2 ) . " pages/sec" );
  96                 }
  97
  98                 # Perform the memory_get_peak_usage() when all the other data has been
  99                 # output so there's no damage if it dies. It is only available since
 100                 # 5.2.0 (since 5.2.1 if you haven't compiled with --enable-memory-limit)
 101                 $this->error( "Memory peak usage of " . memory_get_peak_usage() . " bytes\n" );
 102         }
 103
 104         public function finalSetup() {
 105                 parent::finalSetup();
 106
 107                 if ( $this->getDbType() == Maintenance::DB_NONE ) {
 108                         global $wgUseDatabaseMessages, $wgLocalisationCacheConf, $wgHooks;
 109                         $wgUseDatabaseMessages = false;
 110                         $wgLocalisationCacheConf['storeClass'] = LCStoreNull::class;
 111                         $wgHooks['InterwikiLoadPrefix'][] = 'DumpIterator::disableInterwikis';
 112                 }
 113         }
 114
 115         static function disableInterwikis( $prefix, &$data ) {
 116                 # Title::newFromText will check on each namespaced article if it's an interwiki.
 117                 # We always answer that it is not.
 118
 119                 return false;
 120         }
 121
 122         /**
 123          * Callback function for each revision, child classes should override
 124          * processRevision instead.
 125          * @param WikiRevision $rev
 126          */
 127         public function handleRevision( $rev ) {
 128                 $title = $rev->getTitle();
 129                 if ( !$title ) {
 130                         $this->error( "Got bogus revision with null title!" );
 131
 132                         return;
 133                 }
 134
 135                 $this->count++;
 136                 if ( $this->from !== false ) {
 137                         if ( $this->from != $title ) {
 138                                 return;
 139                         }
 140                         $this->output( "Skipped " . ( $this->count - 1 ) . " pages\n" );
 141
 142                         $this->count = 1;
 143                         $this->from = null;
 144                 }
 145
 146                 $this->processRevision( $rev );
 147         }
 148
 149         /* Stub function for processing additional options */
 150         public function checkOptions() {
 151         }
 152
 153         /* Stub function for giving data about what was computed */
 154         public function conclusions() {
 155         }
 156
 157         /* Core function which does whatever the maintenance script is designed to do */
 158         abstract public function processRevision( $rev );
 159 }
 160
 161 /**
 162  * Maintenance script that runs a regex in the revisions from a dump.
 163  *
 164  * @ingroup Maintenance
 165  */
 166 class SearchDump extends DumpIterator {
 167
 168         public function __construct() {
 169                 parent::__construct();
 170                 $this->addDescription( 'Runs a regex in the revisions from a dump' );
 171                 $this->addOption( 'regex', 'Searching regex', true, true );
 172         }
 173
 174         public function getDbType() {
 175                 return Maintenance::DB_NONE;
 176         }
 177
 178         /**
 179          * @param Revision $rev
 180          */
 181         public function processRevision( $rev ) {
 182                 if ( preg_match( $this->getOption( 'regex' ), $rev->getContent()->getTextForSearchIndex() ) ) {
 183                         $this->output( $rev->getTitle() . " matches at edit from " . $rev->getTimestamp() . "\n" );
 184                 }
 185         }
 186 }
 187
 188 $maintClass = SearchDump::class;
 189 require_once RUN_MAINTENANCE_IF_MAIN;