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