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 use MediaWiki\MediaWikiServices;
  28
  29 require_once __DIR__ . '/Maintenance.php';
  30
  31 /**
  32  * Maintenance script that imports XML dump files into the current wiki.
  33  *
  34  * @ingroup Maintenance
  35  */
  36 class BackupReader extends Maintenance {
  37         public $reportingInterval = 100;
  38         public $pageCount = 0;
  39         public $revCount = 0;
  40         public $dryRun = false;
  41         public $uploads = false;
  42         protected $uploadCount = 0;
  43         public $imageBasePath = false;
  44         /** @var array|false */
  45         public $nsFilter = false;
  46
  47         function __construct() {
  48                 parent::__construct();
  49                 $gz = in_array( 'compress.zlib', stream_get_wrappers() )
  50                         ? 'ok'
  51                         : '(disabled; requires PHP zlib module)';
  52                 $bz2 = in_array( 'compress.bzip2', stream_get_wrappers() )
  53                         ? 'ok'
  54                         : '(disabled; requires PHP bzip2 module)';
  55
  56                 $this->addDescription(
  57                         <<<TEXT
  58 This script reads pages from an XML file as produced from Special:Export or
  59 dumpBackup.php, and saves them into the current wiki.
  60
  61 Compressed XML files may be read directly:
  62   .gz $gz
  63   .bz2 $bz2
  64   .7z (if 7za executable is in PATH)
  65
  66 Note that for very large data sets, importDump.php may be slow; there are
  67 alternate methods which can be much faster for full site restoration:
  68 <https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
  69 TEXT
  70                 );
  71                 $this->stderr = fopen( "php://stderr", "wt" );
  72                 $this->addOption( 'report',
  73                         'Report position and speed after every n pages processed', false, true );
  74                 $this->addOption( 'namespaces',
  75                         'Import only the pages from namespaces belonging to the list of ' .
  76                         'pipe-separated namespace names or namespace indexes', false, true );
  77                 $this->addOption( 'rootpage', 'Pages will be imported as subpages of the specified page',
  78                         false, true );
  79                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
  80                 $this->addOption( 'debug', 'Output extra verbose debug information' );
  81                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
  82                 $this->addOption(
  83                         'no-updates',
  84                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
  85                 );
  86                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
  87                 $this->addOption( 'skip-to', 'Start from nth page by skipping first n-1 pages', false, true );
  88                 $this->addOption( 'username-prefix', 'Prefix for interwiki usernames', false, true );
  89                 $this->addOption( 'no-local-users',
  90                         'Treat all usernames as interwiki. ' .
  91                         'The default is to assign edits to local users where they exist.',
  92                         false, false
  93                 );
  94                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
  95         }
  96
  97         public function execute() {
  98                 if ( wfReadOnly() ) {
  99                         $this->fatalError( "Wiki is in read-only mode; you'll need to disable it for import to work." );
 100                 }
 101
 102                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
 103                 if ( !$this->reportingInterval ) {
 104                         $this->reportingInterval = 100; // avoid division by zero
 105                 }
 106
 107                 $this->dryRun = $this->hasOption( 'dry-run' );
 108                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
 109                 if ( $this->hasOption( 'image-base-path' ) ) {
 110                         $this->imageBasePath = $this->getOption( 'image-base-path' );
 111                 }
 112                 if ( $this->hasOption( 'namespaces' ) ) {
 113                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 114                 }
 115
 116                 if ( $this->hasArg( 0 ) ) {
 117                         $this->importFromFile( $this->getArg( 0 ) );
 118                 } else {
 119                         $this->importFromStdin();
 120                 }
 121
 122                 $this->output( "Done!\n" );
 123                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges,\n" );
 124                 $this->output( "and initSiteStats.php to update page and revision counts\n" );
 125         }
 126
 127         function setNsfilter( array $namespaces ) {
 128                 if ( count( $namespaces ) == 0 ) {
 129                         $this->nsFilter = false;
 130
 131                         return;
 132                 }
 133                 $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
 134         }
 135
 136         private function getNsIndex( $namespace ) {
 137                 $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 138                 $result = $contLang->getNsIndex( $namespace );
 139                 if ( $result !== false ) {
 140                         return $result;
 141                 }
 142                 $ns = intval( $namespace );
 143                 if ( strval( $ns ) === $namespace && $contLang->getNsText( $ns ) !== false ) {
 144                         return $ns;
 145                 }
 146                 $this->fatalError( "Unknown namespace text / index specified: $namespace" );
 147         }
 148
 149         /**
 150          * @param Title|Revision $obj
 151          * @throws MWException
 152          * @return bool
 153          */
 154         private function skippedNamespace( $obj ) {
 155                 $title = null;
 156                 if ( $obj instanceof Title ) {
 157                         $title = $obj;
 158                 } elseif ( $obj instanceof Revision ) {
 159                         $title = $obj->getTitle();
 160                 } elseif ( $obj instanceof WikiRevision ) {
 161                         $title = $obj->title;
 162                 } else {
 163                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 164                 }
 165
 166                 if ( is_null( $title ) ) {
 167                         // Probably a log entry
 168                         return false;
 169                 }
 170
 171                 $ns = $title->getNamespace();
 172
 173                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 174         }
 175
 176         function reportPage( $page ) {
 177                 $this->pageCount++;
 178         }
 179
 180         /**
 181          * @param Revision $rev
 182          */
 183         function handleRevision( $rev ) {
 184                 $title = $rev->getTitle();
 185                 if ( !$title ) {
 186                         $this->progress( "Got bogus revision with null title!" );
 187
 188                         return;
 189                 }
 190
 191                 if ( $this->skippedNamespace( $title ) ) {
 192                         return;
 193                 }
 194
 195                 $this->revCount++;
 196                 $this->report();
 197
 198                 if ( !$this->dryRun ) {
 199                         call_user_func( $this->importCallback, $rev );
 200                 }
 201         }
 202
 203         /**
 204          * @param Revision $revision
 205          * @return bool
 206          */
 207         function handleUpload( $revision ) {
 208                 if ( $this->uploads ) {
 209                         if ( $this->skippedNamespace( $revision ) ) {
 210                                 return false;
 211                         }
 212                         $this->uploadCount++;
 213                         // $this->report();
 214                         // @phan-suppress-next-line PhanUndeclaredMethod
 215                         $this->progress( "upload: " . $revision->getFilename() );
 216
 217                         if ( !$this->dryRun ) {
 218                                 // bluuuh hack
 219                                 // call_user_func( $this->uploadCallback, $revision );
 220                                 $dbw = $this->getDB( DB_MASTER );
 221
 222                                 return $dbw->deadlockLoop( [ $revision, 'importUpload' ] );
 223                         }
 224                 }
 225
 226                 return false;
 227         }
 228
 229         function handleLogItem( $rev ) {
 230                 if ( $this->skippedNamespace( $rev ) ) {
 231                         return;
 232                 }
 233                 $this->revCount++;
 234                 $this->report();
 235
 236                 if ( !$this->dryRun ) {
 237                         call_user_func( $this->logItemCallback, $rev );
 238                 }
 239         }
 240
 241         function report( $final = false ) {
 242                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 243                         $this->showReport();
 244                 }
 245         }
 246
 247         function showReport() {
 248                 if ( !$this->mQuiet ) {
 249                         $delta = microtime( true ) - $this->startTime;
 250                         if ( $delta ) {
 251                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 252                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 253                         } else {
 254                                 $rate = '-';
 255                                 $revrate = '-';
 256                         }
 257                         # Logs dumps don't have page tallies
 258                         if ( $this->pageCount ) {
 259                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 260                         } else {
 261                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 262                         }
 263                 }
 264                 wfWaitForSlaves();
 265         }
 266
 267         function progress( $string ) {
 268                 fwrite( $this->stderr, $string . "\n" );
 269         }
 270
 271         function importFromFile( $filename ) {
 272                 if ( preg_match( '/\.gz$/', $filename ) ) {
 273                         $filename = 'compress.zlib://' . $filename;
 274                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 275                         $filename = 'compress.bzip2://' . $filename;
 276                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 277                         $filename = 'mediawiki.compress.7z://' . $filename;
 278                 }
 279
 280                 $file = fopen( $filename, 'rt' );
 281
 282                 return $this->importFromHandle( $file );
 283         }
 284
 285         function importFromStdin() {
 286                 $file = fopen( 'php://stdin', 'rt' );
 287                 if ( self::posix_isatty( $file ) ) {
 288                         $this->maybeHelp( true );
 289                 }
 290
 291                 return $this->importFromHandle( $file );
 292         }
 293
 294         function importFromHandle( $handle ) {
 295                 $this->startTime = microtime( true );
 296
 297                 $source = new ImportStreamSource( $handle );
 298                 $importer = new WikiImporter( $source, $this->getConfig() );
 299
 300                 // Updating statistics require a lot of time so disable it
 301                 $importer->disableStatisticsUpdate();
 302
 303                 if ( $this->hasOption( 'debug' ) ) {
 304                         $importer->setDebug( true );
 305                 }
 306                 if ( $this->hasOption( 'no-updates' ) ) {
 307                         $importer->setNoUpdates( true );
 308                 }
 309                 if ( $this->hasOption( 'username-prefix' ) ) {
 310                         $importer->setUsernamePrefix(
 311                                 $this->getOption( 'username-prefix' ),
 312                                 !$this->hasOption( 'no-local-users' )
 313                         );
 314                 }
 315                 if ( $this->hasOption( 'rootpage' ) ) {
 316                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 317                         if ( !$statusRootPage->isGood() ) {
 318                                 // Die here so that it doesn't print "Done!"
 319                                 $this->fatalError( $statusRootPage->getMessage()->text() );
 320                                 return false;
 321                         }
 322                 }
 323                 if ( $this->hasOption( 'skip-to' ) ) {
 324                         $nthPage = (int)$this->getOption( 'skip-to' );
 325                         $importer->setPageOffset( $nthPage );
 326                         $this->pageCount = $nthPage - 1;
 327                 }
 328                 $importer->setPageCallback( [ $this, 'reportPage' ] );
 329                 $importer->setNoticeCallback( function ( $msg, $params ) {
 330                         echo wfMessage( $msg, $params )->text() . "\n";
 331                 } );
 332                 $this->importCallback = $importer->setRevisionCallback(
 333                         [ $this, 'handleRevision' ] );
 334                 $this->uploadCallback = $importer->setUploadCallback(
 335                         [ $this, 'handleUpload' ] );
 336                 $this->logItemCallback = $importer->setLogItemCallback(
 337                         [ $this, 'handleLogItem' ] );
 338                 if ( $this->uploads ) {
 339                         $importer->setImportUploads( true );
 340                 }
 341                 if ( $this->imageBasePath ) {
 342                         $importer->setImageBasePath( $this->imageBasePath );
 343                 }
 344
 345                 if ( $this->dryRun ) {
 346                         $importer->setPageOutCallback( null );
 347                 }
 348
 349                 return $importer->doImport();
 350         }
 351 }
 352
 353 $maintClass = BackupReader::class;
 354 require_once RUN_MAINTENANCE_IF_MAIN;