maintenance/importDump.php

   1 <?php
   2 /**
   3  * Import XML dump files into the current wiki.
   4  *
   5  * Copyright © 2005 Brion Vibber <brion@pobox.com>
   6  * https://www.mediawiki.org/
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  * @ingroup Maintenance
  25  */
  26
  27 require_once __DIR__ . '/Maintenance.php';
  28
  29 /**
  30  * Maintenance script that imports XML dump files into the current wiki.
  31  *
  32  * @ingroup Maintenance
  33  */
  34 class BackupReader extends Maintenance {
  35         public $reportingInterval = 100;
  36         public $pageCount = 0;
  37         public $revCount = 0;
  38         public $dryRun = false;
  39         public $uploads = false;
  40         public $imageBasePath = false;
  41         public $nsFilter = false;
  42
  43         function __construct() {
  44                 parent::__construct();
  45                 $gz = in_array( 'compress.zlib', stream_get_wrappers() )
  46                         ? 'ok'
  47                         : '(disabled; requires PHP zlib module)';
  48                 $bz2 = in_array( 'compress.bzip2', stream_get_wrappers() )
  49                         ? 'ok'
  50                         : '(disabled; requires PHP bzip2 module)';
  51
  52                 $this->mDescription = <<<TEXT
  53 This script reads pages from an XML file as produced from Special:Export or
  54 dumpBackup.php, and saves them into the current wiki.
  55
  56 Compressed XML files may be read directly:
  57   .gz $gz
  58   .bz2 $bz2
  59   .7z (if 7za executable is in PATH)
  60
  61 Note that for very large data sets, importDump.php may be slow; there are
  62 alternate methods which can be much faster for full site restoration:
  63 <https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
  64 TEXT;
  65                 $this->stderr = fopen( "php://stderr", "wt" );
  66                 $this->addOption( 'report',
  67                         'Report position and speed after every n pages processed', false, true );
  68                 $this->addOption( 'namespaces',
  69                         'Import only the pages from namespaces belonging to the list of ' .
  70                         'pipe-separated namespace names or namespace indexes', false, true );
  71                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
  72                 $this->addOption( 'debug', 'Output extra verbose debug information' );
  73                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
  74                 $this->addOption(
  75                         'no-updates',
  76                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
  77                 );
  78                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
  79                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
  80         }
  81
  82         public function execute() {
  83                 if ( wfReadOnly() ) {
  84                         $this->error( "Wiki is in read-only mode; you'll need to disable it for import to work.", true );
  85                 }
  86
  87                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
  88                 if ( !$this->reportingInterval ) {
  89                         $this->reportingInterval = 100; // avoid division by zero
  90                 }
  91
  92                 $this->dryRun = $this->hasOption( 'dry-run' );
  93                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
  94                 if ( $this->hasOption( 'image-base-path' ) ) {
  95                         $this->imageBasePath = $this->getOption( 'image-base-path' );
  96                 }
  97                 if ( $this->hasOption( 'namespaces' ) ) {
  98                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
  99                 }
 100
 101                 if ( $this->hasArg() ) {
 102                         $this->importFromFile( $this->getArg() );
 103                 } else {
 104                         $this->importFromStdin();
 105                 }
 106
 107                 $this->output( "Done!\n" );
 108                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges\n" );
 109         }
 110
 111         function setNsfilter( array $namespaces ) {
 112                 if ( count( $namespaces ) == 0 ) {
 113                         $this->nsFilter = false;
 114                         return;
 115                 }
 116                 $this->nsFilter = array_unique( array_map( array( $this, 'getNsIndex' ), $namespaces ) );
 117         }
 118
 119         private function getNsIndex( $namespace ) {
 120                 global $wgContLang;
 121                 if ( ( $result = $wgContLang->getNsIndex( $namespace ) ) !== false ) {
 122                         return $result;
 123                 }
 124                 $ns = intval( $namespace );
 125                 if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
 126                         return $ns;
 127                 }
 128                 $this->error( "Unknown namespace text / index specified: $namespace", true );
 129         }
 130
 131         /**
 132          * @param Title|Revision $obj
 133          * @return bool
 134          */
 135         private function skippedNamespace( $obj ) {
 136                 if ( $obj instanceof Title ) {
 137                         $ns = $obj->getNamespace();
 138                 } elseif ( $obj instanceof Revision ) {
 139                         $ns = $obj->getTitle()->getNamespace();
 140                 } elseif ( $obj instanceof WikiRevision ) {
 141                         $ns = $obj->title->getNamespace();
 142                 } else {
 143                         echo wfBacktrace();
 144                         $this->error( "Cannot get namespace of object in " . __METHOD__, true );
 145                 }
 146                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 147         }
 148
 149         function reportPage( $page ) {
 150                 $this->pageCount++;
 151         }
 152
 153         /**
 154          * @param Revision $rev
 155          */
 156         function handleRevision( $rev ) {
 157                 $title = $rev->getTitle();
 158                 if ( !$title ) {
 159                         $this->progress( "Got bogus revision with null title!" );
 160                         return;
 161                 }
 162
 163                 if ( $this->skippedNamespace( $title ) ) {
 164                         return;
 165                 }
 166
 167                 $this->revCount++;
 168                 $this->report();
 169
 170                 if ( !$this->dryRun ) {
 171                         call_user_func( $this->importCallback, $rev );
 172                 }
 173         }
 174
 175         /**
 176          * @param Revision $revision
 177          * @return bool
 178          */
 179         function handleUpload( $revision ) {
 180                 if ( $this->uploads ) {
 181                         if ( $this->skippedNamespace( $revision ) ) {
 182                                 return;
 183                         }
 184                         $this->uploadCount++;
 185                         // $this->report();
 186                         $this->progress( "upload: " . $revision->getFilename() );
 187
 188                         if ( !$this->dryRun ) {
 189                                 // bluuuh hack
 190                                 // call_user_func( $this->uploadCallback, $revision );
 191                                 $dbw = wfGetDB( DB_MASTER );
 192                                 return $dbw->deadlockLoop( array( $revision, 'importUpload' ) );
 193                         }
 194                 }
 195         }
 196
 197         function handleLogItem( $rev ) {
 198                 if ( $this->skippedNamespace( $rev ) ) {
 199                         return;
 200                 }
 201                 $this->revCount++;
 202                 $this->report();
 203
 204                 if ( !$this->dryRun ) {
 205                         call_user_func( $this->logItemCallback, $rev );
 206                 }
 207         }
 208
 209         function report( $final = false ) {
 210                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 211                         $this->showReport();
 212                 }
 213         }
 214
 215         function showReport() {
 216                 if ( !$this->mQuiet ) {
 217                         $delta = microtime( true ) - $this->startTime;
 218                         if ( $delta ) {
 219                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 220                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 221                         } else {
 222                                 $rate = '-';
 223                                 $revrate = '-';
 224                         }
 225                         # Logs dumps don't have page tallies
 226                         if ( $this->pageCount ) {
 227                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 228                         } else {
 229                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 230                         }
 231                 }
 232                 wfWaitForSlaves();
 233                 // XXX: Don't let deferred jobs array get absurdly large (bug 24375)
 234                 DeferredUpdates::doUpdates( 'commit' );
 235         }
 236
 237         function progress( $string ) {
 238                 fwrite( $this->stderr, $string . "\n" );
 239         }
 240
 241         function importFromFile( $filename ) {
 242                 if ( preg_match( '/\.gz$/', $filename ) ) {
 243                         $filename = 'compress.zlib://' . $filename;
 244                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 245                         $filename = 'compress.bzip2://' . $filename;
 246                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 247                         $filename = 'mediawiki.compress.7z://' . $filename;
 248                 }
 249
 250                 $file = fopen( $filename, 'rt' );
 251                 return $this->importFromHandle( $file );
 252         }
 253
 254         function importFromStdin() {
 255                 $file = fopen( 'php://stdin', 'rt' );
 256                 if ( self::posix_isatty( $file ) ) {
 257                         $this->maybeHelp( true );
 258                 }
 259                 return $this->importFromHandle( $file );
 260         }
 261
 262         function importFromHandle( $handle ) {
 263                 $this->startTime = microtime( true );
 264
 265                 $source = new ImportStreamSource( $handle );
 266                 $importer = new WikiImporter( $source );
 267
 268                 if ( $this->hasOption( 'debug' ) ) {
 269                         $importer->setDebug( true );
 270                 }
 271                 if ( $this->hasOption( 'no-updates' ) ) {
 272                         $importer->setNoUpdates( true );
 273                 }
 274                 $importer->setPageCallback( array( &$this, 'reportPage' ) );
 275                 $this->importCallback = $importer->setRevisionCallback(
 276                         array( &$this, 'handleRevision' ) );
 277                 $this->uploadCallback = $importer->setUploadCallback(
 278                         array( &$this, 'handleUpload' ) );
 279                 $this->logItemCallback = $importer->setLogItemCallback(
 280                         array( &$this, 'handleLogItem' ) );
 281                 if ( $this->uploads ) {
 282                         $importer->setImportUploads( true );
 283                 }
 284                 if ( $this->imageBasePath ) {
 285                         $importer->setImageBasePath( $this->imageBasePath );
 286                 }
 287
 288                 if ( $this->dryRun ) {
 289                         $importer->setPageOutCallback( null );
 290                 }
 291
 292                 return $importer->doImport();
 293         }
 294 }
 295
 296 $maintClass = 'BackupReader';
 297 require_once RUN_MAINTENANCE_IF_MAIN;