maintenance/importDump.php

   1 <?php
   2 /**
   3  * Import XML dump files into the current wiki.
   4  *
   5  * Copyright © 2005 Brooke Vibber <bvibber@wikimedia.org>
   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\Linker\LinkTarget;
  28 use MediaWiki\Maintenance\Maintenance;
  29 use MediaWiki\Permissions\UltimateAuthority;
  30 use MediaWiki\User\User;
  31
  32 // @codeCoverageIgnoreStart
  33 require_once __DIR__ . '/Maintenance.php';
  34 // @codeCoverageIgnoreEnd
  35
  36 /**
  37  * Maintenance script that imports XML dump files into the current wiki.
  38  *
  39  * @ingroup Maintenance
  40  */
  41 class BackupReader extends Maintenance {
  42         /** @var int */
  43         public $reportingInterval = 100;
  44         /** @var int */
  45         public $pageCount = 0;
  46         /** @var int */
  47         public $revCount = 0;
  48         /** @var bool */
  49         public $dryRun = false;
  50         /** @var bool */
  51         public $uploads = false;
  52         /** @var int */
  53         protected $uploadCount = 0;
  54         /** @var string|false */
  55         public $imageBasePath = false;
  56         /** @var array|false */
  57         public $nsFilter = false;
  58         /** @var resource|false */
  59         public $stderr;
  60         /** @var callable|null */
  61         protected $importCallback;
  62         /** @var callable|null */
  63         protected $logItemCallback;
  64         /** @var callable|null */
  65         protected $uploadCallback;
  66         /** @var float */
  67         protected $startTime;
  68
  69         public function __construct() {
  70                 parent::__construct();
  71                 $gz = in_array( 'compress.zlib', stream_get_wrappers() )
  72                         ? 'ok'
  73                         : '(disabled; requires PHP zlib module)';
  74                 $bz2 = in_array( 'compress.bzip2', stream_get_wrappers() )
  75                         ? 'ok'
  76                         : '(disabled; requires PHP bzip2 module)';
  77
  78                 $this->addDescription(
  79                         <<<TEXT
  80 This script reads pages from an XML file as produced from Special:Export or
  81 dumpBackup.php, and saves them into the current wiki.
  82
  83 Compressed XML files may be read directly:
  84   .gz $gz
  85   .bz2 $bz2
  86   .7z (if 7za executable is in PATH)
  87
  88 Note that for very large data sets, importDump.php may be slow; there are
  89 alternate methods which can be much faster for full site restoration:
  90 <https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
  91 TEXT
  92                 );
  93                 $this->stderr = fopen( "php://stderr", "wt" );
  94                 $this->addOption( 'report',
  95                         'Report position and speed after every n pages processed', false, true );
  96                 $this->addOption( 'namespaces',
  97                         'Import only the pages from namespaces belonging to the list of ' .
  98                         'pipe-separated namespace names or namespace indexes', false, true );
  99                 $this->addOption( 'rootpage', 'Pages will be imported as subpages of the specified page',
 100                         false, true );
 101                 $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
 102                 $this->addOption( 'debug', 'Output extra verbose debug information' );
 103                 $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
 104                 $this->addOption(
 105                         'no-updates',
 106                         'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
 107                 );
 108                 $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
 109                 $this->addOption( 'skip-to', 'Start from nth page by skipping first n-1 pages', false, true );
 110                 $this->addOption( 'username-prefix',
 111                         'Prefix for interwiki usernames; a trailing ">" will be added. Default: "imported>"',
 112                         false, true );
 113                 $this->addOption( 'no-local-users',
 114                         'Treat all usernames as interwiki. ' .
 115                         'The default is to assign edits to local users where they exist.',
 116                         false, false
 117                 );
 118                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
 119         }
 120
 121         public function execute() {
 122                 if ( $this->getServiceContainer()->getReadOnlyMode()->isReadOnly() ) {
 123                         $this->fatalError( "Wiki is in read-only mode; you'll need to disable it for import to work." );
 124                 }
 125
 126                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
 127                 if ( !$this->reportingInterval ) {
 128                         // avoid division by zero
 129                         $this->reportingInterval = 100;
 130                 }
 131
 132                 $this->dryRun = $this->hasOption( 'dry-run' );
 133                 $this->uploads = $this->hasOption( 'uploads' );
 134
 135                 if ( $this->hasOption( 'image-base-path' ) ) {
 136                         $this->imageBasePath = $this->getOption( 'image-base-path' );
 137                 }
 138                 if ( $this->hasOption( 'namespaces' ) ) {
 139                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 140                 }
 141
 142                 if ( $this->hasArg( 0 ) ) {
 143                         $this->importFromFile( $this->getArg( 0 ) );
 144                 } else {
 145                         $this->importFromStdin();
 146                 }
 147
 148                 $this->output( "Done!\n" );
 149                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges,\n" );
 150                 $this->output( "and initSiteStats.php to update page and revision counts\n" );
 151         }
 152
 153         private function setNsfilter( array $namespaces ) {
 154                 if ( count( $namespaces ) == 0 ) {
 155                         $this->nsFilter = false;
 156
 157                         return;
 158                 }
 159                 $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
 160         }
 161
 162         private function getNsIndex( $namespace ) {
 163                 $contLang = $this->getServiceContainer()->getContentLanguage();
 164                 $result = $contLang->getNsIndex( $namespace );
 165                 if ( $result !== false ) {
 166                         return $result;
 167                 }
 168                 $ns = intval( $namespace );
 169                 if ( strval( $ns ) === $namespace && $contLang->getNsText( $ns ) !== false ) {
 170                         return $ns;
 171                 }
 172                 $this->fatalError( "Unknown namespace text / index specified: $namespace" );
 173         }
 174
 175         /**
 176          * @param LinkTarget|null $title
 177          * @return bool
 178          */
 179         private function skippedNamespace( $title ) {
 180                 if ( $title === null ) {
 181                         // Probably a log entry
 182                         return false;
 183                 }
 184
 185                 $ns = $title->getNamespace();
 186
 187                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 188         }
 189
 190         public function reportPage( $page ) {
 191                 $this->pageCount++;
 192         }
 193
 194         public function handleRevision( WikiRevision $rev ) {
 195                 $title = $rev->getTitle();
 196                 if ( !$title ) {
 197                         $this->progress( "Got bogus revision with null title!" );
 198
 199                         return;
 200                 }
 201
 202                 if ( $this->skippedNamespace( $title ) ) {
 203                         return;
 204                 }
 205
 206                 $this->revCount++;
 207                 $this->report();
 208
 209                 if ( !$this->dryRun ) {
 210                         ( $this->importCallback )( $rev );
 211                 }
 212         }
 213
 214         /**
 215          * @param WikiRevision $revision
 216          * @return bool
 217          */
 218         public function handleUpload( WikiRevision $revision ) {
 219                 if ( $this->uploads ) {
 220                         if ( $this->skippedNamespace( $revision->getTitle() ) ) {
 221                                 return false;
 222                         }
 223                         $this->uploadCount++;
 224                         // $this->report();
 225                         $this->progress( "upload: " . $revision->getFilename() );
 226
 227                         if ( !$this->dryRun ) {
 228                                 // bluuuh hack
 229                                 // ( $this->uploadCallback )( $revision );
 230                                 $importer = $this->getServiceContainer()->getWikiRevisionUploadImporter();
 231                                 $statusValue = $importer->import( $revision );
 232
 233                                 return $statusValue->isGood();
 234                         }
 235                 }
 236
 237                 return false;
 238         }
 239
 240         public function handleLogItem( WikiRevision $rev ) {
 241                 if ( $this->skippedNamespace( $rev->getTitle() ) ) {
 242                         return;
 243                 }
 244                 $this->revCount++;
 245                 $this->report();
 246
 247                 if ( !$this->dryRun ) {
 248                         ( $this->logItemCallback )( $rev );
 249                 }
 250         }
 251
 252         private function report( $final = false ) {
 253                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 254                         $this->showReport();
 255                 }
 256         }
 257
 258         private function showReport() {
 259                 if ( !$this->mQuiet ) {
 260                         $delta = microtime( true ) - $this->startTime;
 261                         if ( $delta ) {
 262                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 263                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 264                         } else {
 265                                 $rate = '-';
 266                                 $revrate = '-';
 267                         }
 268                         # Logs dumps don't have page tallies
 269                         if ( $this->pageCount ) {
 270                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 271                         } else {
 272                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 273                         }
 274                 }
 275                 $this->waitForReplication();
 276         }
 277
 278         private function progress( $string ) {
 279                 fwrite( $this->stderr, $string . "\n" );
 280         }
 281
 282         private function importFromFile( $filename ) {
 283                 if ( preg_match( '/\.gz$/', $filename ) ) {
 284                         $filename = 'compress.zlib://' . $filename;
 285                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 286                         $filename = 'compress.bzip2://' . $filename;
 287                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 288                         $filename = 'mediawiki.compress.7z://' . $filename;
 289                 }
 290
 291                 $file = fopen( $filename, 'rt' );
 292                 if ( $file === false ) {
 293                         $this->fatalError( error_get_last()['message'] ?? 'Could not open file' );
 294                 }
 295
 296                 return $this->importFromHandle( $file );
 297         }
 298
 299         private function importFromStdin() {
 300                 $file = fopen( 'php://stdin', 'rt' );
 301                 if ( self::posix_isatty( $file ) ) {
 302                         $this->maybeHelp( true );
 303                 }
 304
 305                 return $this->importFromHandle( $file );
 306         }
 307
 308         private function importFromHandle( $handle ) {
 309                 $this->startTime = microtime( true );
 310
 311                 $user = User::newSystemUser( User::MAINTENANCE_SCRIPT_USER, [ 'steal' => true ] );
 312
 313                 $source = new ImportStreamSource( $handle );
 314                 $importer = $this->getServiceContainer()
 315                         ->getWikiImporterFactory()
 316                         ->getWikiImporter( $source, new UltimateAuthority( $user ) );
 317
 318                 // Updating statistics require a lot of time so disable it
 319                 $importer->disableStatisticsUpdate();
 320
 321                 if ( $this->hasOption( 'debug' ) ) {
 322                         $importer->setDebug( true );
 323                 }
 324                 if ( $this->hasOption( 'no-updates' ) ) {
 325                         $importer->setNoUpdates( true );
 326                 }
 327                 $importer->setUsernamePrefix(
 328                         $this->getOption( 'username-prefix', 'imported' ),
 329                         !$this->hasOption( 'no-local-users' )
 330                 );
 331                 if ( $this->hasOption( 'rootpage' ) ) {
 332                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 333                         if ( !$statusRootPage->isGood() ) {
 334                                 // Die here so that it doesn't print "Done!"
 335                                 $this->fatalError( $statusRootPage );
 336                         }
 337                 }
 338                 if ( $this->hasOption( 'skip-to' ) ) {
 339                         $nthPage = (int)$this->getOption( 'skip-to' );
 340                         $importer->setPageOffset( $nthPage );
 341                         $this->pageCount = $nthPage - 1;
 342                 }
 343                 $importer->setPageCallback( [ $this, 'reportPage' ] );
 344                 $importer->setNoticeCallback( static function ( $msg, $params ) {
 345                         echo wfMessage( $msg, $params )->text() . "\n";
 346                 } );
 347                 $this->importCallback = $importer->setRevisionCallback(
 348                         [ $this, 'handleRevision' ] );
 349                 $this->uploadCallback = $importer->setUploadCallback(
 350                         [ $this, 'handleUpload' ] );
 351                 $this->logItemCallback = $importer->setLogItemCallback(
 352                         [ $this, 'handleLogItem' ] );
 353                 if ( $this->uploads ) {
 354                         $importer->setImportUploads( true );
 355                 }
 356                 if ( $this->imageBasePath ) {
 357                         $importer->setImageBasePath( $this->imageBasePath );
 358                 }
 359
 360                 if ( $this->dryRun ) {
 361                         $importer->setPageOutCallback( null );
 362                 }
 363
 364                 return $importer->doImport();
 365         }
 366 }
 367
 368 // @codeCoverageIgnoreStart
 369 $maintClass = BackupReader::class;
 370 require_once RUN_MAINTENANCE_IF_MAIN;
 371 // @codeCoverageIgnoreEnd