generateSitemap.php

Go to the documentation of this file.

<?php
use MediaWiki\MainConfigNames;
use MediaWiki\Title\Title;
use MediaWiki\WikiMap\WikiMap;
use Wikimedia\Rdbms\IDatabase;
use Wikimedia\Rdbms\IResultWrapper;
 
require_once __DIR__ . '/Maintenance.php';
 
class GenerateSitemap extends Maintenance {
    private const GS_MAIN = -2;
    private const GS_TALK = -1;
 
    public $url_limit;
 
    public $size_limit;
 
    public $fspath;
 
    public $urlpath;
 
    public $compress;
 
    public $skipRedirects;
 
    public $limit = [];
 
    public $priorities = [];
 
    public $namespaces = [];
 
    public $timestamp;
 
    public $dbr;
 
    public $findex;
 
    public $file;
 
    private $identifier;
 
    public function __construct() {
        parent::__construct();
        $this->addDescription( 'Creates a sitemap for the site' );
        $this->addOption(
            'fspath',
            'The file system path to save to, e.g. /tmp/sitemap; defaults to current directory',
            false,
            true
        );
        $this->addOption(
            'urlpath',
            'The URL path corresponding to --fspath, prepended to filenames in the index; '
                . 'defaults to an empty string',
            false,
            true
        );
        $this->addOption(
            'compress',
            'Compress the sitemap files, can take value yes|no, default yes',
            false,
            true
        );
        $this->addOption( 'skip-redirects', 'Do not include redirecting articles in the sitemap' );
        $this->addOption(
            'identifier',
            'What site identifier to use for the wiki, defaults to $wgDBname',
            false,
            true
        );
        $this->addOption(
            'namespaces',
            'Only include pages in these namespaces in the sitemap, ' .
            'defaults to the value of wgSitemapNamespaces if not defined.',
            false, true, false, true
        );
    }
 
    public function execute() {
        $this->setNamespacePriorities();
        $this->url_limit = 50000;
        $this->size_limit = ( 2 ** 20 ) * 10;
 
        # Create directory if needed
        $fspath = $this->getOption( 'fspath', getcwd() );
        if ( !wfMkdirParents( $fspath, null, __METHOD__ ) ) {
            $this->fatalError( "Can not create directory $fspath." );
        }
 
        $dbDomain = WikiMap::getCurrentWikiDbDomain()->getId();
        $this->fspath = realpath( $fspath ) . DIRECTORY_SEPARATOR;
        $this->urlpath = $this->getOption( 'urlpath', "" );
        if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
            $this->urlpath .= '/';
        }
        $this->identifier = $this->getOption( 'identifier', $dbDomain );
        $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
        $this->skipRedirects = $this->hasOption( 'skip-redirects' );
        $this->dbr = $this->getReplicaDB();
        $this->generateNamespaces();
        $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
        $encIdentifier = rawurlencode( $this->identifier );
        $this->findex = fopen( "{$this->fspath}sitemap-index-{$encIdentifier}.xml", 'wb' );
        $this->main();
    }
 
    private function setNamespacePriorities() {
        $sitemapNamespacesPriorities = $this->getConfig()->get( MainConfigNames::SitemapNamespacesPriorities );
 
        // Custom main namespaces
        $this->priorities[self::GS_MAIN] = '0.5';
        // Custom talk namespaces
        $this->priorities[self::GS_TALK] = '0.1';
        // MediaWiki standard namespaces
        $this->priorities[NS_MAIN] = '1.0';
        $this->priorities[NS_TALK] = '0.1';
        $this->priorities[NS_USER] = '0.5';
        $this->priorities[NS_USER_TALK] = '0.1';
        $this->priorities[NS_PROJECT] = '0.5';
        $this->priorities[NS_PROJECT_TALK] = '0.1';
        $this->priorities[NS_FILE] = '0.5';
        $this->priorities[NS_FILE_TALK] = '0.1';
        $this->priorities[NS_MEDIAWIKI] = '0.0';
        $this->priorities[NS_MEDIAWIKI_TALK] = '0.1';
        $this->priorities[NS_TEMPLATE] = '0.0';
        $this->priorities[NS_TEMPLATE_TALK] = '0.1';
        $this->priorities[NS_HELP] = '0.5';
        $this->priorities[NS_HELP_TALK] = '0.1';
        $this->priorities[NS_CATEGORY] = '0.5';
        $this->priorities[NS_CATEGORY_TALK] = '0.1';
 
        // Custom priorities
        if ( $sitemapNamespacesPriorities !== false ) {
            foreach ( $sitemapNamespacesPriorities as $namespace => $priority ) {
                $float = floatval( $priority );
                if ( $float > 1.0 ) {
                    $priority = '1.0';
                } elseif ( $float < 0.0 ) {
                    $priority = '0.0';
                }
                $this->priorities[$namespace] = $priority;
            }
        }
    }
 
    private function generateNamespaces() {
        // Use the namespaces passed in via command line arguments if they are set.
        $sitemapNamespacesFromConfig = $this->getOption( 'namespaces' );
        if ( is_array( $sitemapNamespacesFromConfig ) && count( $sitemapNamespacesFromConfig ) > 0 ) {
            $this->namespaces = $sitemapNamespacesFromConfig;
 
            return;
        }
 
        // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
        $sitemapNamespaces = $this->getConfig()->get( MainConfigNames::SitemapNamespaces );
        if ( is_array( $sitemapNamespaces ) ) {
            $this->namespaces = $sitemapNamespaces;
 
            return;
        }
 
        $res = $this->dbr->newSelectQueryBuilder()
            ->select( [ 'page_namespace' ] )
            ->from( 'page' )
            ->groupBy( 'page_namespace' )
            ->orderBy( 'page_namespace' )
            ->caller( __METHOD__ )->fetchResultSet();
 
        foreach ( $res as $row ) {
            $this->namespaces[] = $row->page_namespace;
        }
    }
 
    private function priority( $namespace ) {
        return $this->priorities[$namespace] ?? $this->guessPriority( $namespace );
    }
 
    private function guessPriority( $namespace ) {
        return $this->getServiceContainer()->getNamespaceInfo()->isSubject( $namespace )
            ? $this->priorities[self::GS_MAIN]
            : $this->priorities[self::GS_TALK];
    }
 
    private function getPageRes( $namespace ) {
        return $this->dbr->newSelectQueryBuilder()
            ->select( [ 'page_namespace', 'page_title', 'page_touched', 'page_is_redirect', 'pp_propname' ] )
            ->from( 'page' )
            ->leftJoin( 'page_props', null, [ 'page_id = pp_page', 'pp_propname' => 'noindex' ] )
            ->where( [ 'page_namespace' => $namespace ] )
            ->caller( __METHOD__ )->fetchResultSet();
    }
 
    public function main() {
        $services = $this->getServiceContainer();
        $contLang = $services->getContentLanguage();
        $langConverter = $services->getLanguageConverterFactory()->getLanguageConverter( $contLang );
 
        fwrite( $this->findex, $this->openIndex() );
 
        foreach ( $this->namespaces as $namespace ) {
            $res = $this->getPageRes( $namespace );
            $this->file = false;
            $this->generateLimit( $namespace );
            $length = $this->limit[0];
            $i = $smcount = 0;
 
            $fns = $contLang->getFormattedNsText( $namespace );
            $this->output( "$namespace ($fns)\n" );
            $skippedRedirects = 0; // Number of redirects skipped for that namespace
            $skippedNoindex = 0; // Number of pages with __NOINDEX__ switch for that NS
            foreach ( $res as $row ) {
                if ( $row->pp_propname === 'noindex' ) {
                    $skippedNoindex++;
                    continue;
                }
 
                if ( $this->skipRedirects && $row->page_is_redirect ) {
                    $skippedRedirects++;
                    continue;
                }
 
                if ( $i++ === 0
                    || $i === $this->url_limit + 1
                    || $length + $this->limit[1] + $this->limit[2] > $this->size_limit
                ) {
                    if ( $this->file !== false ) {
                        $this->write( $this->file, $this->closeFile() );
                        $this->close( $this->file );
                    }
                    $filename = $this->sitemapFilename( $namespace, $smcount++ );
                    $this->file = $this->open( $this->fspath . $filename, 'wb' );
                    $this->write( $this->file, $this->openFile() );
                    fwrite( $this->findex, $this->indexEntry( $filename ) );
                    $this->output( "\t$this->fspath$filename\n" );
                    $length = $this->limit[0];
                    $i = 1;
                }
                $title = Title::makeTitle( $row->page_namespace, $row->page_title );
                $date = wfTimestamp( TS_ISO_8601, $row->page_touched );
                $entry = $this->fileEntry( $title->getCanonicalURL(), $date, $this->priority( $namespace ) );
                $length += strlen( $entry );
                $this->write( $this->file, $entry );
                // generate pages for language variants
                if ( $langConverter->hasVariants() ) {
                    $variants = $langConverter->getVariants();
                    foreach ( $variants as $vCode ) {
                        if ( $vCode == $contLang->getCode() ) {
                            continue; // we don't want default variant
                        }
                        $entry = $this->fileEntry(
                            $title->getCanonicalURL( [ 'variant' => $vCode ] ),
                            $date,
                            $this->priority( $namespace )
                        );
                        $length += strlen( $entry );
                        $this->write( $this->file, $entry );
                    }
                }
            }
 
            if ( $skippedNoindex > 0 ) {
                $this->output( "  skipped $skippedNoindex page(s) with __NOINDEX__ switch\n" );
            }
 
            if ( $this->skipRedirects && $skippedRedirects > 0 ) {
                $this->output( "  skipped $skippedRedirects redirect(s)\n" );
            }
 
            if ( $this->file ) {
                $this->write( $this->file, $this->closeFile() );
                $this->close( $this->file );
            }
        }
        fwrite( $this->findex, $this->closeIndex() );
        fclose( $this->findex );
    }
 
    private function open( $file, $flags ) {
        $resource = $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
        if ( $resource === false ) {
            throw new RuntimeException( __METHOD__
                . " error opening file $file with flags $flags. Check permissions?" );
        }
 
        return $resource;
    }
 
    private function write( &$handle, $str ) {
        if ( $handle === true || $handle === false ) {
            throw new InvalidArgumentException( __METHOD__ . " was passed a boolean as a file handle.\n" );
        }
        if ( $this->compress ) {
            gzwrite( $handle, $str );
        } else {
            fwrite( $handle, $str );
        }
    }
 
    private function close( &$handle ) {
        if ( $this->compress ) {
            gzclose( $handle );
        } else {
            fclose( $handle );
        }
    }
 
    private function sitemapFilename( $namespace, $count ) {
        $ext = $this->compress ? '.gz' : '';
 
        return "sitemap-{$this->identifier}-NS_$namespace-$count.xml$ext";
    }
 
    private function xmlHead() {
        return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
    }
 
    private function xmlSchema() {
        return 'http://www.sitemaps.org/schemas/sitemap/0.9';
    }
 
    private function openIndex() {
        return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
    }
 
    private function indexEntry( $filename ) {
        return "\t<sitemap>\n" .
            "\t\t<loc>" . wfGetServerUrl( PROTO_CANONICAL ) .
                ( substr( $this->urlpath, 0, 1 ) === "/" ? "" : "/" ) .
                "{$this->urlpath}$filename</loc>\n" .
            "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
            "\t</sitemap>\n";
    }
 
    private function closeIndex() {
        return "</sitemapindex>\n";
    }
 
    private function openFile() {
        return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
    }
 
    private function fileEntry( $url, $date, $priority ) {
        return "\t<url>\n" .
            // T36666: $url may contain bad characters such as ampersands.
            "\t\t<loc>" . htmlspecialchars( $url ) . "</loc>\n" .
            "\t\t<lastmod>$date</lastmod>\n" .
            "\t\t<priority>$priority</priority>\n" .
            "\t</url>\n";
    }
 
    private function closeFile() {
        return "</urlset>\n";
    }
 
    private function generateLimit( $namespace ) {
        // T19961: make a title with the longest possible URL in this namespace
        $title = Title::makeTitle( $namespace, str_repeat( "\u{28B81}", 63 ) . "\u{5583}" );
 
        $this->limit = [
            strlen( $this->openFile() ),
            strlen( $this->fileEntry(
                $title->getCanonicalURL(),
                wfTimestamp( TS_ISO_8601, wfTimestamp() ),
                $this->priority( $namespace )
            ) ),
            strlen( $this->closeFile() )
        ];
    }
}
 
$maintClass = GenerateSitemap::class;
require_once RUN_MAINTENANCE_IF_MAIN;