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
 115                         return;
 116                 }
 117                 $this->nsFilter = array_unique( array_map( array( $this, 'getNsIndex' ), $namespaces ) );
 118         }
 119
 120         private function getNsIndex( $namespace ) {
 121                 global $wgContLang;
 122                 if ( ( $result = $wgContLang->getNsIndex( $namespace ) ) !== false ) {
 123                         return $result;
 124                 }
 125                 $ns = intval( $namespace );
 126                 if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
 127                         return $ns;
 128                 }
 129                 $this->error( "Unknown namespace text / index specified: $namespace", true );
 130         }
 131
 132         /**
 133          * @param Title|Revision $obj
 134          * @return bool
 135          */
 136         private function skippedNamespace( $obj ) {
 137                 if ( $obj instanceof Title ) {
 138                         $ns = $obj->getNamespace();
 139                 } elseif ( $obj instanceof Revision ) {
 140                         $ns = $obj->getTitle()->getNamespace();
 141                 } elseif ( $obj instanceof WikiRevision ) {
 142                         $ns = $obj->title->getNamespace();
 143                 } else {
 144                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 145                 }
 146
 147                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 148         }
 149
 150         function reportPage( $page ) {
 151                 $this->pageCount++;
 152         }
 153
 154         /**
 155          * @param Revision $rev
 156          */
 157         function handleRevision( $rev ) {
 158                 $title = $rev->getTitle();
 159                 if ( !$title ) {
 160                         $this->progress( "Got bogus revision with null title!" );
 161
 162                         return;
 163                 }
 164
 165                 if ( $this->skippedNamespace( $title ) ) {
 166                         return;
 167                 }
 168
 169                 $this->revCount++;
 170                 $this->report();
 171
 172                 if ( !$this->dryRun ) {
 173                         call_user_func( $this->importCallback, $rev );
 174                 }
 175         }
 176
 177         /**
 178          * @param Revision $revision
 179          * @return bool
 180          */
 181         function handleUpload( $revision ) {
 182                 if ( $this->uploads ) {
 183                         if ( $this->skippedNamespace( $revision ) ) {
 184                                 return false;
 185                         }
 186                         $this->uploadCount++;
 187                         // $this->report();
 188                         $this->progress( "upload: " . $revision->getFilename() );
 189
 190                         if ( !$this->dryRun ) {
 191                                 // bluuuh hack
 192                                 // call_user_func( $this->uploadCallback, $revision );
 193                                 $dbw = wfGetDB( DB_MASTER );
 194
 195                                 return $dbw->deadlockLoop( array( $revision, 'importUpload' ) );
 196                         }
 197                 }
 198
 199                 return false;
 200         }
 201
 202         function handleLogItem( $rev ) {
 203                 if ( $this->skippedNamespace( $rev ) ) {
 204                         return;
 205                 }
 206                 $this->revCount++;
 207                 $this->report();
 208
 209                 if ( !$this->dryRun ) {
 210                         call_user_func( $this->logItemCallback, $rev );
 211                 }
 212         }
 213
 214         function report( $final = false ) {
 215                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 216                         $this->showReport();
 217                 }
 218         }
 219
 220         function showReport() {
 221                 if ( !$this->mQuiet ) {
 222                         $delta = microtime( true ) - $this->startTime;
 223                         if ( $delta ) {
 224                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 225                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 226                         } else {
 227                                 $rate = '-';
 228                                 $revrate = '-';
 229                         }
 230                         # Logs dumps don't have page tallies
 231                         if ( $this->pageCount ) {
 232                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 233                         } else {
 234                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 235                         }
 236                 }
 237                 wfWaitForSlaves();
 238                 // XXX: Don't let deferred jobs array get absurdly large (bug 24375)
 239                 DeferredUpdates::doUpdates( 'commit' );
 240         }
 241
 242         function progress( $string ) {
 243                 fwrite( $this->stderr, $string . "\n" );
 244         }
 245
 246         function importFromFile( $filename ) {
 247                 if ( preg_match( '/\.gz$/', $filename ) ) {
 248                         $filename = 'compress.zlib://' . $filename;
 249                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 250                         $filename = 'compress.bzip2://' . $filename;
 251                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 252                         $filename = 'mediawiki.compress.7z://' . $filename;
 253                 }
 254
 255                 $file = fopen( $filename, 'rt' );
 256
 257                 return $this->importFromHandle( $file );
 258         }
 259
 260         function importFromStdin() {
 261                 $file = fopen( 'php://stdin', 'rt' );
 262                 if ( self::posix_isatty( $file ) ) {
 263                         $this->maybeHelp( true );
 264                 }
 265
 266                 return $this->importFromHandle( $file );
 267         }
 268
 269         function importFromHandle( $handle ) {
 270                 $this->startTime = microtime( true );
 271
 272                 $source = new ImportStreamSource( $handle );
 273                 $importer = new WikiImporter( $source, $this->getConfig() );
 274
 275                 if ( $this->hasOption( 'debug' ) ) {
 276                         $importer->setDebug( true );
 277                 }
 278                 if ( $this->hasOption( 'no-updates' ) ) {
 279                         $importer->setNoUpdates( true );
 280                 }
 281                 $importer->setPageCallback( array( &$this, 'reportPage' ) );
 282                 $this->importCallback = $importer->setRevisionCallback(
 283                         array( &$this, 'handleRevision' ) );
 284                 $this->uploadCallback = $importer->setUploadCallback(
 285                         array( &$this, 'handleUpload' ) );
 286                 $this->logItemCallback = $importer->setLogItemCallback(
 287                         array( &$this, 'handleLogItem' ) );
 288                 if ( $this->uploads ) {
 289                         $importer->setImportUploads( true );
 290                 }
 291                 if ( $this->imageBasePath ) {
 292                         $importer->setImageBasePath( $this->imageBasePath );
 293                 }
 294
 295                 if ( $this->dryRun ) {
 296                         $importer->setPageOutCallback( null );
 297                 }
 298
 299                 return $importer->doImport();
 300         }
 301 }
 302
 303 $maintClass = 'BackupReader';
 304 require_once RUN_MAINTENANCE_IF_MAIN;