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->addOption( 'skip-to', 'Start from nth page by skipping first n-1 pages', false, true );
  84                 $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
  85         }
  86
  87         public function execute() {
  88                 if ( wfReadOnly() ) {
  89                         $this->error( "Wiki is in read-only mode; you'll need to disable it for import to work.", true );
  90                 }
  91
  92                 $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
  93                 if ( !$this->reportingInterval ) {
  94                         $this->reportingInterval = 100; // avoid division by zero
  95                 }
  96
  97                 $this->dryRun = $this->hasOption( 'dry-run' );
  98                 $this->uploads = $this->hasOption( 'uploads' ); // experimental!
  99                 if ( $this->hasOption( 'image-base-path' ) ) {
 100                         $this->imageBasePath = $this->getOption( 'image-base-path' );
 101                 }
 102                 if ( $this->hasOption( 'namespaces' ) ) {
 103                         $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
 104                 }
 105
 106                 if ( $this->hasArg() ) {
 107                         $this->importFromFile( $this->getArg() );
 108                 } else {
 109                         $this->importFromStdin();
 110                 }
 111
 112                 $this->output( "Done!\n" );
 113                 $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges,\n" );
 114                 $this->output( "and initSiteStats.php to update page and revision counts\n" );
 115         }
 116
 117         function setNsfilter( array $namespaces ) {
 118                 if ( count( $namespaces ) == 0 ) {
 119                         $this->nsFilter = false;
 120
 121                         return;
 122                 }
 123                 $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
 124         }
 125
 126         private function getNsIndex( $namespace ) {
 127                 global $wgContLang;
 128                 $result = $wgContLang->getNsIndex( $namespace );
 129                 if ( $result !== false ) {
 130                         return $result;
 131                 }
 132                 $ns = intval( $namespace );
 133                 if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
 134                         return $ns;
 135                 }
 136                 $this->error( "Unknown namespace text / index specified: $namespace", true );
 137         }
 138
 139         /**
 140          * @param Title|Revision $obj
 141          * @return bool
 142          */
 143         private function skippedNamespace( $obj ) {
 144                 $title = null;
 145                 if ( $obj instanceof Title ) {
 146                         $title = $obj;
 147                 } elseif ( $obj instanceof Revision ) {
 148                         $title = $obj->getTitle();
 149                 } elseif ( $obj instanceof WikiRevision ) {
 150                         $title = $obj->title;
 151                 } else {
 152                         throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
 153                 }
 154
 155                 if ( is_null( $title ) ) {
 156                         // Probably a log entry
 157                         return false;
 158                 }
 159
 160                 $ns = $title->getNamespace();
 161
 162                 return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 163         }
 164
 165         function reportPage( $page ) {
 166                 $this->pageCount++;
 167         }
 168
 169         /**
 170          * @param Revision $rev
 171          */
 172         function handleRevision( $rev ) {
 173                 $title = $rev->getTitle();
 174                 if ( !$title ) {
 175                         $this->progress( "Got bogus revision with null title!" );
 176
 177                         return;
 178                 }
 179
 180                 if ( $this->skippedNamespace( $title ) ) {
 181                         return;
 182                 }
 183
 184                 $this->revCount++;
 185                 $this->report();
 186
 187                 if ( !$this->dryRun ) {
 188                         call_user_func( $this->importCallback, $rev );
 189                 }
 190         }
 191
 192         /**
 193          * @param Revision $revision
 194          * @return bool
 195          */
 196         function handleUpload( $revision ) {
 197                 if ( $this->uploads ) {
 198                         if ( $this->skippedNamespace( $revision ) ) {
 199                                 return false;
 200                         }
 201                         $this->uploadCount++;
 202                         // $this->report();
 203                         $this->progress( "upload: " . $revision->getFilename() );
 204
 205                         if ( !$this->dryRun ) {
 206                                 // bluuuh hack
 207                                 // call_user_func( $this->uploadCallback, $revision );
 208                                 $dbw = $this->getDB( DB_MASTER );
 209
 210                                 return $dbw->deadlockLoop( [ $revision, 'importUpload' ] );
 211                         }
 212                 }
 213
 214                 return false;
 215         }
 216
 217         function handleLogItem( $rev ) {
 218                 if ( $this->skippedNamespace( $rev ) ) {
 219                         return;
 220                 }
 221                 $this->revCount++;
 222                 $this->report();
 223
 224                 if ( !$this->dryRun ) {
 225                         call_user_func( $this->logItemCallback, $rev );
 226                 }
 227         }
 228
 229         function report( $final = false ) {
 230                 if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
 231                         $this->showReport();
 232                 }
 233         }
 234
 235         function showReport() {
 236                 if ( !$this->mQuiet ) {
 237                         $delta = microtime( true ) - $this->startTime;
 238                         if ( $delta ) {
 239                                 $rate = sprintf( "%.2f", $this->pageCount / $delta );
 240                                 $revrate = sprintf( "%.2f", $this->revCount / $delta );
 241                         } else {
 242                                 $rate = '-';
 243                                 $revrate = '-';
 244                         }
 245                         # Logs dumps don't have page tallies
 246                         if ( $this->pageCount ) {
 247                                 $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
 248                         } else {
 249                                 $this->progress( "$this->revCount ($revrate revs/sec)" );
 250                         }
 251                 }
 252                 wfWaitForSlaves();
 253         }
 254
 255         function progress( $string ) {
 256                 fwrite( $this->stderr, $string . "\n" );
 257         }
 258
 259         function importFromFile( $filename ) {
 260                 if ( preg_match( '/\.gz$/', $filename ) ) {
 261                         $filename = 'compress.zlib://' . $filename;
 262                 } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 263                         $filename = 'compress.bzip2://' . $filename;
 264                 } elseif ( preg_match( '/\.7z$/', $filename ) ) {
 265                         $filename = 'mediawiki.compress.7z://' . $filename;
 266                 }
 267
 268                 $file = fopen( $filename, 'rt' );
 269
 270                 return $this->importFromHandle( $file );
 271         }
 272
 273         function importFromStdin() {
 274                 $file = fopen( 'php://stdin', 'rt' );
 275                 if ( self::posix_isatty( $file ) ) {
 276                         $this->maybeHelp( true );
 277                 }
 278
 279                 return $this->importFromHandle( $file );
 280         }
 281
 282         function importFromHandle( $handle ) {
 283                 $this->startTime = microtime( true );
 284
 285                 $source = new ImportStreamSource( $handle );
 286                 $importer = new WikiImporter( $source, $this->getConfig() );
 287
 288                 // Updating statistics require a lot of time so disable it
 289                 $importer->disableStatisticsUpdate();
 290
 291                 if ( $this->hasOption( 'debug' ) ) {
 292                         $importer->setDebug( true );
 293                 }
 294                 if ( $this->hasOption( 'no-updates' ) ) {
 295                         $importer->setNoUpdates( true );
 296                 }
 297                 if ( $this->hasOption( 'rootpage' ) ) {
 298                         $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
 299                         if ( !$statusRootPage->isGood() ) {
 300                                 // Die here so that it doesn't print "Done!"
 301                                 $this->error( $statusRootPage->getMessage()->text(), 1 );
 302                                 return false;
 303                         }
 304                 }
 305                 if ( $this->hasOption( 'skip-to' ) ) {
 306                         $nthPage = (int)$this->getOption( 'skip-to' );
 307                         $importer->setPageOffset( $nthPage );
 308                         $this->pageCount = $nthPage - 1;
 309                 }
 310                 $importer->setPageCallback( [ $this, 'reportPage' ] );
 311                 $this->importCallback = $importer->setRevisionCallback(
 312                         [ $this, 'handleRevision' ] );
 313                 $this->uploadCallback = $importer->setUploadCallback(
 314                         [ $this, 'handleUpload' ] );
 315                 $this->logItemCallback = $importer->setLogItemCallback(
 316                         [ $this, 'handleLogItem' ] );
 317                 if ( $this->uploads ) {
 318                         $importer->setImportUploads( true );
 319                 }
 320                 if ( $this->imageBasePath ) {
 321                         $importer->setImageBasePath( $this->imageBasePath );
 322                 }
 323
 324                 if ( $this->dryRun ) {
 325                         $importer->setPageOutCallback( null );
 326                 }
 327
 328                 return $importer->doImport();
 329         }
 330 }
 331
 332 $maintClass = 'BackupReader';
 333 require_once RUN_MAINTENANCE_IF_MAIN;