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