importDump.php

Go to the documentation of this file.

<?php
use MediaWiki\MediaWikiServices;
 
require_once __DIR__ . '/Maintenance.php';
 
class BackupReader extends Maintenance {
    public $reportingInterval = 100;
    public $pageCount = 0;
    public $revCount = 0;
    public $dryRun = false;
    public $uploads = false;
    protected $uploadCount = 0;
    public $imageBasePath = false;
    public $nsFilter = false;
 
    function __construct() {
        parent::__construct();
        $gz = in_array( 'compress.zlib', stream_get_wrappers() )
            ? 'ok'
            : '(disabled; requires PHP zlib module)';
        $bz2 = in_array( 'compress.bzip2', stream_get_wrappers() )
            ? 'ok'
            : '(disabled; requires PHP bzip2 module)';
 
        $this->addDescription(
            <<<TEXT
This script reads pages from an XML file as produced from Special:Export or
dumpBackup.php, and saves them into the current wiki.
 
Compressed XML files may be read directly:
  .gz $gz
  .bz2 $bz2
  .7z (if 7za executable is in PATH)
 
Note that for very large data sets, importDump.php may be slow; there are
alternate methods which can be much faster for full site restoration:
<https://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
TEXT
        );
        $this->stderr = fopen( "php://stderr", "wt" );
        $this->addOption( 'report',
            'Report position and speed after every n pages processed', false, true );
        $this->addOption( 'namespaces',
            'Import only the pages from namespaces belonging to the list of ' .
            'pipe-separated namespace names or namespace indexes', false, true );
        $this->addOption( 'rootpage', 'Pages will be imported as subpages of the specified page',
            false, true );
        $this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
        $this->addOption( 'debug', 'Output extra verbose debug information' );
        $this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
        $this->addOption(
            'no-updates',
            'Disable link table updates. Is faster but leaves the wiki in an inconsistent state'
        );
        $this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
        $this->addOption( 'skip-to', 'Start from nth page by skipping first n-1 pages', false, true );
        $this->addOption( 'username-prefix', 'Prefix for interwiki usernames', false, true );
        $this->addOption( 'no-local-users',
            'Treat all usernames as interwiki. ' .
            'The default is to assign edits to local users where they exist.',
            false, false
        );
        $this->addArg( 'file', 'Dump file to import [else use stdin]', false );
    }
 
    public function execute() {
        if ( wfReadOnly() ) {
            $this->fatalError( "Wiki is in read-only mode; you'll need to disable it for import to work." );
        }
 
        $this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
        if ( !$this->reportingInterval ) {
            $this->reportingInterval = 100; // avoid division by zero
        }
 
        $this->dryRun = $this->hasOption( 'dry-run' );
        $this->uploads = $this->hasOption( 'uploads' ); // experimental!
        if ( $this->hasOption( 'image-base-path' ) ) {
            $this->imageBasePath = $this->getOption( 'image-base-path' );
        }
        if ( $this->hasOption( 'namespaces' ) ) {
            $this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
        }
 
        if ( $this->hasArg() ) {
            $this->importFromFile( $this->getArg() );
        } else {
            $this->importFromStdin();
        }
 
        $this->output( "Done!\n" );
        $this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges,\n" );
        $this->output( "and initSiteStats.php to update page and revision counts\n" );
    }
 
    function setNsfilter( array $namespaces ) {
        if ( count( $namespaces ) == 0 ) {
            $this->nsFilter = false;
 
            return;
        }
        $this->nsFilter = array_unique( array_map( [ $this, 'getNsIndex' ], $namespaces ) );
    }
 
    private function getNsIndex( $namespace ) {
        $contLang = MediaWikiServices::getInstance()->getContentLanguage();
        $result = $contLang->getNsIndex( $namespace );
        if ( $result !== false ) {
            return $result;
        }
        $ns = intval( $namespace );
        if ( strval( $ns ) === $namespace && $contLang->getNsText( $ns ) !== false ) {
            return $ns;
        }
        $this->fatalError( "Unknown namespace text / index specified: $namespace" );
    }
 
    private function skippedNamespace( $obj ) {
        $title = null;
        if ( $obj instanceof Title ) {
            $title = $obj;
        } elseif ( $obj instanceof Revision ) {
            $title = $obj->getTitle();
        } elseif ( $obj instanceof WikiRevision ) {
            $title = $obj->title;
        } else {
            throw new MWException( "Cannot get namespace of object in " . __METHOD__ );
        }
 
        if ( is_null( $title ) ) {
            // Probably a log entry
            return false;
        }
 
        $ns = $title->getNamespace();
 
        return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
    }
 
    function reportPage( $page ) {
        $this->pageCount++;
    }
 
    function handleRevision( $rev ) {
        $title = $rev->getTitle();
        if ( !$title ) {
            $this->progress( "Got bogus revision with null title!" );
 
            return;
        }
 
        if ( $this->skippedNamespace( $title ) ) {
            return;
        }
 
        $this->revCount++;
        $this->report();
 
        if ( !$this->dryRun ) {
            call_user_func( $this->importCallback, $rev );
        }
    }
 
    function handleUpload( $revision ) {
        if ( $this->uploads ) {
            if ( $this->skippedNamespace( $revision ) ) {
                return false;
            }
            $this->uploadCount++;
            // $this->report();
            $this->progress( "upload: " . $revision->getFilename() );
 
            if ( !$this->dryRun ) {
                // bluuuh hack
                // call_user_func( $this->uploadCallback, $revision );
                $dbw = $this->getDB( DB_MASTER );
 
                return $dbw->deadlockLoop( [ $revision, 'importUpload' ] );
            }
        }
 
        return false;
    }
 
    function handleLogItem( $rev ) {
        if ( $this->skippedNamespace( $rev ) ) {
            return;
        }
        $this->revCount++;
        $this->report();
 
        if ( !$this->dryRun ) {
            call_user_func( $this->logItemCallback, $rev );
        }
    }
 
    function report( $final = false ) {
        if ( $final xor ( $this->pageCount % $this->reportingInterval == 0 ) ) {
            $this->showReport();
        }
    }
 
    function showReport() {
        if ( !$this->mQuiet ) {
            $delta = microtime( true ) - $this->startTime;
            if ( $delta ) {
                $rate = sprintf( "%.2f", $this->pageCount / $delta );
                $revrate = sprintf( "%.2f", $this->revCount / $delta );
            } else {
                $rate = '-';
                $revrate = '-';
            }
            # Logs dumps don't have page tallies
            if ( $this->pageCount ) {
                $this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
            } else {
                $this->progress( "$this->revCount ($revrate revs/sec)" );
            }
        }
        wfWaitForSlaves();
    }
 
    function progress( $string ) {
        fwrite( $this->stderr, $string . "\n" );
    }
 
    function importFromFile( $filename ) {
        if ( preg_match( '/\.gz$/', $filename ) ) {
            $filename = 'compress.zlib://' . $filename;
        } elseif ( preg_match( '/\.bz2$/', $filename ) ) {
            $filename = 'compress.bzip2://' . $filename;
        } elseif ( preg_match( '/\.7z$/', $filename ) ) {
            $filename = 'mediawiki.compress.7z://' . $filename;
        }
 
        $file = fopen( $filename, 'rt' );
 
        return $this->importFromHandle( $file );
    }
 
    function importFromStdin() {
        $file = fopen( 'php://stdin', 'rt' );
        if ( self::posix_isatty( $file ) ) {
            $this->maybeHelp( true );
        }
 
        return $this->importFromHandle( $file );
    }
 
    function importFromHandle( $handle ) {
        $this->startTime = microtime( true );
 
        $source = new ImportStreamSource( $handle );
        $importer = new WikiImporter( $source, $this->getConfig() );
 
        // Updating statistics require a lot of time so disable it
        $importer->disableStatisticsUpdate();
 
        if ( $this->hasOption( 'debug' ) ) {
            $importer->setDebug( true );
        }
        if ( $this->hasOption( 'no-updates' ) ) {
            $importer->setNoUpdates( true );
        }
        if ( $this->hasOption( 'username-prefix' ) ) {
            $importer->setUsernamePrefix(
                $this->getOption( 'username-prefix' ),
                !$this->hasOption( 'no-local-users' )
            );
        }
        if ( $this->hasOption( 'rootpage' ) ) {
            $statusRootPage = $importer->setTargetRootPage( $this->getOption( 'rootpage' ) );
            if ( !$statusRootPage->isGood() ) {
                // Die here so that it doesn't print "Done!"
                $this->fatalError( $statusRootPage->getMessage()->text() );
                return false;
            }
        }
        if ( $this->hasOption( 'skip-to' ) ) {
            $nthPage = (int)$this->getOption( 'skip-to' );
            $importer->setPageOffset( $nthPage );
            $this->pageCount = $nthPage - 1;
        }
        $importer->setPageCallback( [ $this, 'reportPage' ] );
        $importer->setNoticeCallback( function ( $msg, $params ) {
            echo wfMessage( $msg, $params )->text() . "\n";
        } );
        $this->importCallback = $importer->setRevisionCallback(
            [ $this, 'handleRevision' ] );
        $this->uploadCallback = $importer->setUploadCallback(
            [ $this, 'handleUpload' ] );
        $this->logItemCallback = $importer->setLogItemCallback(
            [ $this, 'handleLogItem' ] );
        if ( $this->uploads ) {
            $importer->setImportUploads( true );
        }
        if ( $this->imageBasePath ) {
            $importer->setImageBasePath( $this->imageBasePath );
        }
 
        if ( $this->dryRun ) {
            $importer->setPageOutCallback( null );
        }
 
        return $importer->doImport();
    }
}
 
$maintClass = BackupReader::class;
require_once RUN_MAINTENANCE_IF_MAIN;