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->addDescription(
  53                         <<<TEXT
  54 This script reads pages from an XML file as produced from Special:Export or
  55 dumpBackup.php, and saves them into the current wiki.
  56
  57 Compressed XML files may be read directly:
  58   .gz $gz
  59   .bz2 $bz2
  60   .7z (if 7za executable is in PATH)
  61
  62 Note that for very large data sets, importDump.php may be slow; there are
  63 alternate methods which can be much faster for full site restoration:
  64 <https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
  65 TEXT
  66                 );
  67                 $this->stderr = fopen( "php://stderr", "wt" );
  68                 $this->addOption( 'report',
  69                         'Report position and speed after every n pages processed', false, true );
  70                 $this->addOption( 'namespaces',
  71                         'Import only the pages from namespaces belonging to the list of ' .
  72                         'pipe-separated namespace names or namespace indexes', false, true );
  73                 $this->addOption( 'rootpage', 'Pages will be imported as subpages of the specified page',
  74                         false, true );
  75                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
  76                 $this->addOption( 'debug', 'Output extra verbose debug information' );
  77                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
  78                 $this->addOption(
  79                         'no-updates',
  80                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
  81                 );
  82                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
  83                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
  84         }
  85
  86         public function execute() {
  87                 if ( wfReadOnly() ) {
  88                         $this->error( "Wiki is in read-only mode; you'll need to disable it for import to work.", true );
  89                 }
  90
  91                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
  92                 if ( !$this->reportingInterval ) {
  93                         $this->reportingInterval = 100; // avoid division by zero
  94                 }
  95
  96                 $this->dryRun = $this->hasOption( 'dry-run' );
  97                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
  98                 if ( $this->hasOption( 'image-base-path' ) ) {
  99                         $this->imageBasePath = $this->getOption( 'image-base-path' );
 100                 }
 101                 if ( $this->hasOption( 'namespaces' ) ) {
 102                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 103                 }
 104
 105                 if ( $this->hasArg() ) {
 106                         $this->importFromFile( $this->getArg() );
 107                 } else {
 108                         $this->importFromStdin();
 109                 }
 110
 111                 $this->output( "Done!\n" );
 112                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges\n" );
 113         }
 114
 115         function setNsfilter( array $namespaces ) {
 116                 if ( count( $namespaces ) == 0 ) {
 117                         $this->nsFilter = false;
 118
 119                         return;
 120                 }
 121                 $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
 122         }
 123
 124         private function getNsIndex( $namespace ) {
 125                 global $wgContLang;
 126                 $result = $wgContLang->getNsIndex( $namespace );
 127                 if ( $result !== false ) {
 128                         return $result;
 129                 }
 130                 $ns = intval( $namespace );
 131                 if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
 132                         return $ns;
 133                 }
 134                 $this->error( "Unknown namespace text / index specified: $namespace", true );
 135         }
 136
 137         /**
 138          * @param Title|Revision $obj
 139          * @return bool
 140          */
 141         private function skippedNamespace( $obj ) {
 142                 $title = null;
 143                 if ( $obj instanceof Title ) {
 144                         $title = $obj;
 145                 } elseif ( $obj instanceof Revision ) {
 146                         $title = $obj->getTitle();
 147                 } elseif ( $obj instanceof WikiRevision ) {
 148                         $title = $obj->title;
 149                 } else {
 150                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 151                 }
 152
 153                 if ( is_null( $title ) ) {
 154                         // Probably a log entry
 155                         return false;
 156                 }
 157
 158                 $ns = $title->getNamespace();
 159
 160                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 161         }
 162
 163         function reportPage( $page ) {
 164                 $this->pageCount++;
 165         }
 166
 167         /**
 168          * @param Revision $rev
 169          */
 170         function handleRevision( $rev ) {
 171                 $title = $rev->getTitle();
 172                 if ( !$title ) {
 173                         $this->progress( "Got bogus revision with null title!" );
 174
 175                         return;
 176                 }
 177
 178                 if ( $this->skippedNamespace( $title ) ) {
 179                         return;
 180                 }
 181
 182                 $this->revCount++;
 183                 $this->report();
 184
 185                 if ( !$this->dryRun ) {
 186                         call_user_func( $this->importCallback, $rev );
 187                 }
 188         }
 189
 190         /**
 191          * @param Revision $revision
 192          * @return bool
 193          */
 194         function handleUpload( $revision ) {
 195                 if ( $this->uploads ) {
 196                         if ( $this->skippedNamespace( $revision ) ) {
 197                                 return false;
 198                         }
 199                         $this->uploadCount++;
 200                         // $this->report();
 201                         $this->progress( "upload: " . $revision->getFilename() );
 202
 203                         if ( !$this->dryRun ) {
 204                                 // bluuuh hack
 205                                 // call_user_func( $this->uploadCallback, $revision );
 206                                 $dbw = $this->getDB( DB_MASTER );
 207
 208                                 return $dbw->deadlockLoop( [ $revision, 'importUpload' ] );
 209                         }
 210                 }
 211
 212                 return false;
 213         }
 214
 215         function handleLogItem( $rev ) {
 216                 if ( $this->skippedNamespace( $rev ) ) {
 217                         return;
 218                 }
 219                 $this->revCount++;
 220                 $this->report();
 221
 222                 if ( !$this->dryRun ) {
 223                         call_user_func( $this->logItemCallback, $rev );
 224                 }
 225         }
 226
 227         function report( $final = false ) {
 228                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 229                         $this->showReport();
 230                 }
 231         }
 232
 233         function showReport() {
 234                 if ( !$this->mQuiet ) {
 235                         $delta = microtime( true ) - $this->startTime;
 236                         if ( $delta ) {
 237                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 238                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 239                         } else {
 240                                 $rate = '-';
 241                                 $revrate = '-';
 242                         }
 243                         # Logs dumps don't have page tallies
 244                         if ( $this->pageCount ) {
 245                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 246                         } else {
 247                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 248                         }
 249                 }
 250                 wfWaitForSlaves();
 251         }
 252
 253         function progress( $string ) {
 254                 fwrite( $this->stderr, $string . "\n" );
 255         }
 256
 257         function importFromFile( $filename ) {
 258                 if ( preg_match( '/\.gz$/', $filename ) ) {
 259                         $filename = 'compress.zlib://' . $filename;
 260                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 261                         $filename = 'compress.bzip2://' . $filename;
 262                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 263                         $filename = 'mediawiki.compress.7z://' . $filename;
 264                 }
 265
 266                 $file = fopen( $filename, 'rt' );
 267
 268                 return $this->importFromHandle( $file );
 269         }
 270
 271         function importFromStdin() {
 272                 $file = fopen( 'php://stdin', 'rt' );
 273                 if ( self::posix_isatty( $file ) ) {
 274                         $this->maybeHelp( true );
 275                 }
 276
 277                 return $this->importFromHandle( $file );
 278         }
 279
 280         function importFromHandle( $handle ) {
 281                 $this->startTime = microtime( true );
 282
 283                 $source = new ImportStreamSource( $handle );
 284                 $importer = new WikiImporter( $source, $this->getConfig() );
 285
 286                 if ( $this->hasOption( 'debug' ) ) {
 287                         $importer->setDebug( true );
 288                 }
 289                 if ( $this->hasOption( 'no-updates' ) ) {
 290                         $importer->setNoUpdates( true );
 291                 }
 292                 if ( $this->hasOption( 'rootpage' ) ) {
 293                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 294                         if ( !$statusRootPage->isGood() ) {
 295                                 // Die here so that it doesn't print "Done!"
 296                                 $this->error( $statusRootPage->getMessage()->text(), 1 );
 297                                 return false;
 298                         }
 299                 }
 300                 $importer->setPageCallback( [ $this, 'reportPage' ] );
 301                 $this->importCallback = $importer->setRevisionCallback(
 302                         [ $this, 'handleRevision' ] );
 303                 $this->uploadCallback = $importer->setUploadCallback(
 304                         [ $this, 'handleUpload' ] );
 305                 $this->logItemCallback = $importer->setLogItemCallback(
 306                         [ $this, 'handleLogItem' ] );
 307                 if ( $this->uploads ) {
 308                         $importer->setImportUploads( true );
 309                 }
 310                 if ( $this->imageBasePath ) {
 311                         $importer->setImageBasePath( $this->imageBasePath );
 312                 }
 313
 314                 if ( $this->dryRun ) {
 315                         $importer->setPageOutCallback( null );
 316                 }
 317
 318                 return $importer->doImport();
 319         }
 320 }
 321
 322 $maintClass = 'BackupReader';
 323 require_once RUN_MAINTENANCE_IF_MAIN;