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