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         /** @var bool|resource */
  47         public $stderr;
  48         /** @var callable|null */
  49         protected $importCallback;
  50         /** @var callable|null */
  51         protected $logItemCallback;
  52         /** @var callable|null */
  53         protected $uploadCallback;
  54         /** @var int */
  55         protected $startTime;
  56
  57         function __construct() {
  58                 parent::__construct();
  59                 $gz = in_array( 'compress.zlib', stream_get_wrappers() )
  60                         ? 'ok'
  61                         : '(disabled; requires PHP zlib module)';
  62                 $bz2 = in_array( 'compress.bzip2', stream_get_wrappers() )
  63                         ? 'ok'
  64                         : '(disabled; requires PHP bzip2 module)';
  65
  66                 $this->addDescription(
  67                         <<<TEXT
  68 This script reads pages from an XML file as produced from Special:Export or
  69 dumpBackup.php, and saves them into the current wiki.
  70
  71 Compressed XML files may be read directly:
  72   .gz $gz
  73   .bz2 $bz2
  74   .7z (if 7za executable is in PATH)
  75
  76 Note that for very large data sets, importDump.php may be slow; there are
  77 alternate methods which can be much faster for full site restoration:
  78 <https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
  79 TEXT
  80                 );
  81                 $this->stderr = fopen( "php://stderr", "wt" );
  82                 $this->addOption( 'report',
  83                         'Report position and speed after every n pages processed', false, true );
  84                 $this->addOption( 'namespaces',
  85                         'Import only the pages from namespaces belonging to the list of ' .
  86                         'pipe-separated namespace names or namespace indexes', false, true );
  87                 $this->addOption( 'rootpage', 'Pages will be imported as subpages of the specified page',
  88                         false, true );
  89                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
  90                 $this->addOption( 'debug', 'Output extra verbose debug information' );
  91                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
  92                 $this->addOption(
  93                         'no-updates',
  94                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
  95                 );
  96                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
  97                 $this->addOption( 'skip-to', 'Start from nth page by skipping first n-1 pages', false, true );
  98                 $this->addOption( 'username-prefix', 'Prefix for interwiki usernames', false, true );
  99                 $this->addOption( 'no-local-users',
 100                         'Treat all usernames as interwiki. ' .
 101                         'The default is to assign edits to local users where they exist.',
 102                         false, false
 103                 );
 104                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
 105         }
 106
 107         public function execute() {
 108                 if ( wfReadOnly() ) {
 109                         $this->fatalError( "Wiki is in read-only mode; you'll need to disable it for import to work." );
 110                 }
 111
 112                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
 113                 if ( !$this->reportingInterval ) {
 114                         $this->reportingInterval = 100; // avoid division by zero
 115                 }
 116
 117                 $this->dryRun = $this->hasOption( 'dry-run' );
 118                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
 119                 if ( $this->hasOption( 'image-base-path' ) ) {
 120                         $this->imageBasePath = $this->getOption( 'image-base-path' );
 121                 }
 122                 if ( $this->hasOption( 'namespaces' ) ) {
 123                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 124                 }
 125
 126                 if ( $this->hasArg( 0 ) ) {
 127                         $this->importFromFile( $this->getArg( 0 ) );
 128                 } else {
 129                         $this->importFromStdin();
 130                 }
 131
 132                 $this->output( "Done!\n" );
 133                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges,\n" );
 134                 $this->output( "and initSiteStats.php to update page and revision counts\n" );
 135         }
 136
 137         function setNsfilter( array $namespaces ) {
 138                 if ( count( $namespaces ) == 0 ) {
 139                         $this->nsFilter = false;
 140
 141                         return;
 142                 }
 143                 $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
 144         }
 145
 146         private function getNsIndex( $namespace ) {
 147                 $contLang = MediaWikiServices::getInstance()->getContentLanguage();
 148                 $result = $contLang->getNsIndex( $namespace );
 149                 if ( $result !== false ) {
 150                         return $result;
 151                 }
 152                 $ns = intval( $namespace );
 153                 if ( strval( $ns ) === $namespace && $contLang->getNsText( $ns ) !== false ) {
 154                         return $ns;
 155                 }
 156                 $this->fatalError( "Unknown namespace text / index specified: $namespace" );
 157         }
 158
 159         /**
 160          * @param Title|Revision $obj
 161          * @throws MWException
 162          * @return bool
 163          */
 164         private function skippedNamespace( $obj ) {
 165                 $title = null;
 166                 if ( $obj instanceof Title ) {
 167                         $title = $obj;
 168                 } elseif ( $obj instanceof Revision ) {
 169                         $title = $obj->getTitle();
 170                 } elseif ( $obj instanceof WikiRevision ) {
 171                         $title = $obj->title;
 172                 } else {
 173                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 174                 }
 175
 176                 if ( is_null( $title ) ) {
 177                         // Probably a log entry
 178                         return false;
 179                 }
 180
 181                 $ns = $title->getNamespace();
 182
 183                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 184         }
 185
 186         function reportPage( $page ) {
 187                 $this->pageCount++;
 188         }
 189
 190         /**
 191          * @param Revision $rev
 192          */
 193         function handleRevision( $rev ) {
 194                 $title = $rev->getTitle();
 195                 if ( !$title ) {
 196                         $this->progress( "Got bogus revision with null title!" );
 197
 198                         return;
 199                 }
 200
 201                 if ( $this->skippedNamespace( $title ) ) {
 202                         return;
 203                 }
 204
 205                 $this->revCount++;
 206                 $this->report();
 207
 208                 if ( !$this->dryRun ) {
 209                         call_user_func( $this->importCallback, $rev );
 210                 }
 211         }
 212
 213         /**
 214          * @param Revision $revision
 215          * @return bool
 216          */
 217         function handleUpload( $revision ) {
 218                 if ( $this->uploads ) {
 219                         if ( $this->skippedNamespace( $revision ) ) {
 220                                 return false;
 221                         }
 222                         $this->uploadCount++;
 223                         // $this->report();
 224                         // @phan-suppress-next-line PhanUndeclaredMethod
 225                         $this->progress( "upload: " . $revision->getFilename() );
 226
 227                         if ( !$this->dryRun ) {
 228                                 // bluuuh hack
 229                                 // call_user_func( $this->uploadCallback, $revision );
 230                                 $dbw = $this->getDB( DB_MASTER );
 231
 232                                 return $dbw->deadlockLoop( [ $revision, 'importUpload' ] );
 233                         }
 234                 }
 235
 236                 return false;
 237         }
 238
 239         function handleLogItem( $rev ) {
 240                 if ( $this->skippedNamespace( $rev ) ) {
 241                         return;
 242                 }
 243                 $this->revCount++;
 244                 $this->report();
 245
 246                 if ( !$this->dryRun ) {
 247                         call_user_func( $this->logItemCallback, $rev );
 248                 }
 249         }
 250
 251         function report( $final = false ) {
 252                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 253                         $this->showReport();
 254                 }
 255         }
 256
 257         function showReport() {
 258                 if ( !$this->mQuiet ) {
 259                         $delta = microtime( true ) - $this->startTime;
 260                         if ( $delta ) {
 261                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 262                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 263                         } else {
 264                                 $rate = '-';
 265                                 $revrate = '-';
 266                         }
 267                         # Logs dumps don't have page tallies
 268                         if ( $this->pageCount ) {
 269                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 270                         } else {
 271                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 272                         }
 273                 }
 274                 wfWaitForSlaves();
 275         }
 276
 277         function progress( $string ) {
 278                 fwrite( $this->stderr, $string . "\n" );
 279         }
 280
 281         function importFromFile( $filename ) {
 282                 if ( preg_match( '/\.gz$/', $filename ) ) {
 283                         $filename = 'compress.zlib://' . $filename;
 284                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 285                         $filename = 'compress.bzip2://' . $filename;
 286                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 287                         $filename = 'mediawiki.compress.7z://' . $filename;
 288                 }
 289
 290                 $file = fopen( $filename, 'rt' );
 291
 292                 return $this->importFromHandle( $file );
 293         }
 294
 295         function importFromStdin() {
 296                 $file = fopen( 'php://stdin', 'rt' );
 297                 if ( self::posix_isatty( $file ) ) {
 298                         $this->maybeHelp( true );
 299                 }
 300
 301                 return $this->importFromHandle( $file );
 302         }
 303
 304         function importFromHandle( $handle ) {
 305                 $this->startTime = microtime( true );
 306
 307                 $source = new ImportStreamSource( $handle );
 308                 $importer = new WikiImporter( $source, $this->getConfig() );
 309
 310                 // Updating statistics require a lot of time so disable it
 311                 $importer->disableStatisticsUpdate();
 312
 313                 if ( $this->hasOption( 'debug' ) ) {
 314                         $importer->setDebug( true );
 315                 }
 316                 if ( $this->hasOption( 'no-updates' ) ) {
 317                         $importer->setNoUpdates( true );
 318                 }
 319                 if ( $this->hasOption( 'username-prefix' ) ) {
 320                         $importer->setUsernamePrefix(
 321                                 $this->getOption( 'username-prefix' ),
 322                                 !$this->hasOption( 'no-local-users' )
 323                         );
 324                 }
 325                 if ( $this->hasOption( 'rootpage' ) ) {
 326                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 327                         if ( !$statusRootPage->isGood() ) {
 328                                 // Die here so that it doesn't print "Done!"
 329                                 $this->fatalError( $statusRootPage->getMessage()->text() );
 330                                 return false;
 331                         }
 332                 }
 333                 if ( $this->hasOption( 'skip-to' ) ) {
 334                         $nthPage = (int)$this->getOption( 'skip-to' );
 335                         $importer->setPageOffset( $nthPage );
 336                         $this->pageCount = $nthPage - 1;
 337                 }
 338                 $importer->setPageCallback( [ $this, 'reportPage' ] );
 339                 $importer->setNoticeCallback( function ( $msg, $params ) {
 340                         echo wfMessage( $msg, $params )->text() . "\n";
 341                 } );
 342                 $this->importCallback = $importer->setRevisionCallback(
 343                         [ $this, 'handleRevision' ] );
 344                 $this->uploadCallback = $importer->setUploadCallback(
 345                         [ $this, 'handleUpload' ] );
 346                 $this->logItemCallback = $importer->setLogItemCallback(
 347                         [ $this, 'handleLogItem' ] );
 348                 if ( $this->uploads ) {
 349                         $importer->setImportUploads( true );
 350                 }
 351                 if ( $this->imageBasePath ) {
 352                         $importer->setImageBasePath( $this->imageBasePath );
 353                 }
 354
 355                 if ( $this->dryRun ) {
 356                         $importer->setPageOutCallback( null );
 357                 }
 358
 359                 return $importer->doImport();
 360         }
 361 }
 362
 363 $maintClass = BackupReader::class;
 364 require_once RUN_MAINTENANCE_IF_MAIN;