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( 'rootpage', 'Pages will be imported as subpages of the specified page',
  72                         false, true );
  73                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
  74                 $this->addOption( 'debug', 'Output extra verbose debug information' );
  75                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
  76                 $this->addOption(
  77                         'no-updates',
  78                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
  79                 );
  80                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
  81                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
  82         }
  83
  84         public function execute() {
  85                 if ( wfReadOnly() ) {
  86                         $this->error( "Wiki is in read-only mode; you'll need to disable it for import to work.", true );
  87                 }
  88
  89                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
  90                 if ( !$this->reportingInterval ) {
  91                         $this->reportingInterval = 100; // avoid division by zero
  92                 }
  93
  94                 $this->dryRun = $this->hasOption( 'dry-run' );
  95                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
  96                 if ( $this->hasOption( 'image-base-path' ) ) {
  97                         $this->imageBasePath = $this->getOption( 'image-base-path' );
  98                 }
  99                 if ( $this->hasOption( 'namespaces' ) ) {
 100                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 101                 }
 102
 103                 if ( $this->hasArg() ) {
 104                         $this->importFromFile( $this->getArg() );
 105                 } else {
 106                         $this->importFromStdin();
 107                 }
 108
 109                 $this->output( "Done!\n" );
 110                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges\n" );
 111         }
 112
 113         function setNsfilter( array $namespaces ) {
 114                 if ( count( $namespaces ) == 0 ) {
 115                         $this->nsFilter = false;
 116
 117                         return;
 118                 }
 119                 $this->nsFilter = array_unique( array_map( array( $this, 'getNsIndex' ), $namespaces ) );
 120         }
 121
 122         private function getNsIndex( $namespace ) {
 123                 global $wgContLang;
 124                 $result = $wgContLang->getNsIndex( $namespace );
 125                 if ( $result !== false ) {
 126                         return $result;
 127                 }
 128                 $ns = intval( $namespace );
 129                 if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
 130                         return $ns;
 131                 }
 132                 $this->error( "Unknown namespace text / index specified: $namespace", true );
 133         }
 134
 135         /**
 136          * @param Title|Revision $obj
 137          * @return bool
 138          */
 139         private function skippedNamespace( $obj ) {
 140                 $title = null;
 141                 if ( $obj instanceof Title ) {
 142                         $title = $obj;
 143                 } elseif ( $obj instanceof Revision ) {
 144                         $title = $obj->getTitle();
 145                 } elseif ( $obj instanceof WikiRevision ) {
 146                         $title = $obj->title;
 147                 } else {
 148                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 149                 }
 150
 151                 if ( is_null( $title ) ) {
 152                         // Probably a log entry
 153                         return false;
 154                 }
 155
 156                 $ns = $title->getNamespace();
 157
 158                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 159         }
 160
 161         function reportPage( $page ) {
 162                 $this->pageCount++;
 163         }
 164
 165         /**
 166          * @param Revision $rev
 167          */
 168         function handleRevision( $rev ) {
 169                 $title = $rev->getTitle();
 170                 if ( !$title ) {
 171                         $this->progress( "Got bogus revision with null title!" );
 172
 173                         return;
 174                 }
 175
 176                 if ( $this->skippedNamespace( $title ) ) {
 177                         return;
 178                 }
 179
 180                 $this->revCount++;
 181                 $this->report();
 182
 183                 if ( !$this->dryRun ) {
 184                         call_user_func( $this->importCallback, $rev );
 185                 }
 186         }
 187
 188         /**
 189          * @param Revision $revision
 190          * @return bool
 191          */
 192         function handleUpload( $revision ) {
 193                 if ( $this->uploads ) {
 194                         if ( $this->skippedNamespace( $revision ) ) {
 195                                 return false;
 196                         }
 197                         $this->uploadCount++;
 198                         // $this->report();
 199                         $this->progress( "upload: " . $revision->getFilename() );
 200
 201                         if ( !$this->dryRun ) {
 202                                 // bluuuh hack
 203                                 // call_user_func( $this->uploadCallback, $revision );
 204                                 $dbw = $this->getDB( DB_MASTER );
 205
 206                                 return $dbw->deadlockLoop( array( $revision, 'importUpload' ) );
 207                         }
 208                 }
 209
 210                 return false;
 211         }
 212
 213         function handleLogItem( $rev ) {
 214                 if ( $this->skippedNamespace( $rev ) ) {
 215                         return;
 216                 }
 217                 $this->revCount++;
 218                 $this->report();
 219
 220                 if ( !$this->dryRun ) {
 221                         call_user_func( $this->logItemCallback, $rev );
 222                 }
 223         }
 224
 225         function report( $final = false ) {
 226                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 227                         $this->showReport();
 228                 }
 229         }
 230
 231         function showReport() {
 232                 if ( !$this->mQuiet ) {
 233                         $delta = microtime( true ) - $this->startTime;
 234                         if ( $delta ) {
 235                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 236                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 237                         } else {
 238                                 $rate = '-';
 239                                 $revrate = '-';
 240                         }
 241                         # Logs dumps don't have page tallies
 242                         if ( $this->pageCount ) {
 243                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 244                         } else {
 245                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 246                         }
 247                 }
 248                 wfWaitForSlaves();
 249         }
 250
 251         function progress( $string ) {
 252                 fwrite( $this->stderr, $string . "\n" );
 253         }
 254
 255         function importFromFile( $filename ) {
 256                 if ( preg_match( '/\.gz$/', $filename ) ) {
 257                         $filename = 'compress.zlib://' . $filename;
 258                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 259                         $filename = 'compress.bzip2://' . $filename;
 260                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 261                         $filename = 'mediawiki.compress.7z://' . $filename;
 262                 }
 263
 264                 $file = fopen( $filename, 'rt' );
 265
 266                 return $this->importFromHandle( $file );
 267         }
 268
 269         function importFromStdin() {
 270                 $file = fopen( 'php://stdin', 'rt' );
 271                 if ( self::posix_isatty( $file ) ) {
 272                         $this->maybeHelp( true );
 273                 }
 274
 275                 return $this->importFromHandle( $file );
 276         }
 277
 278         function importFromHandle( $handle ) {
 279                 $this->startTime = microtime( true );
 280
 281                 $source = new ImportStreamSource( $handle );
 282                 $importer = new WikiImporter( $source, $this->getConfig() );
 283
 284                 if ( $this->hasOption( 'debug' ) ) {
 285                         $importer->setDebug( true );
 286                 }
 287                 if ( $this->hasOption( 'no-updates' ) ) {
 288                         $importer->setNoUpdates( true );
 289                 }
 290                 if ( $this->hasOption( 'rootpage' ) ) {
 291                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 292                         if ( !$statusRootPage->isGood() ) {
 293                                 // Die here so that it doesn't print "Done!"
 294                                 $this->error( $statusRootPage->getMessage()->text(), 1 );
 295                                 return false;
 296                         }
 297                 }
 298                 $importer->setPageCallback( array( &$this, 'reportPage' ) );
 299                 $this->importCallback = $importer->setRevisionCallback(
 300                         array( &$this, 'handleRevision' ) );
 301                 $this->uploadCallback = $importer->setUploadCallback(
 302                         array( &$this, 'handleUpload' ) );
 303                 $this->logItemCallback = $importer->setLogItemCallback(
 304                         array( &$this, 'handleLogItem' ) );
 305                 if ( $this->uploads ) {
 306                         $importer->setImportUploads( true );
 307                 }
 308                 if ( $this->imageBasePath ) {
 309                         $importer->setImageBasePath( $this->imageBasePath );
 310                 }
 311
 312                 if ( $this->dryRun ) {
 313                         $importer->setPageOutCallback( null );
 314                 }
 315
 316                 return $importer->doImport();
 317         }
 318 }
 319
 320 $maintClass = 'BackupReader';
 321 require_once RUN_MAINTENANCE_IF_MAIN;