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