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