generateSitemap.php

Go to the documentation of this file.

<?php
require_once( dirname( __FILE__ ) . '/Maintenance.php' );

class GenerateSitemap extends Maintenance {
        const GS_MAIN = -2;
        const GS_TALK = -1;

        var $url_limit;

        var $size_limit;

        var $fspath;

        var $urlpath;

        var $compress;

        var $limit = array();

        var $priorities = array();

        var $namespaces = array();

        var $timestamp;

        var $dbr;

        var $findex;


        var $file;

        private $identifier;

        public function __construct() {
                parent::__construct();
                $this->mDescription = "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( 'identifier', 'What site identifier to use for the wiki, defaults to $wgDBname', false, true );
        }

        public function execute() {
                $this->setNamespacePriorities();
                $this->url_limit = 50000;
                $this->size_limit = pow( 2, 20 ) * 10;
                $this->fspath = self::init_path( $this->getOption( 'fspath', getcwd() ) );
                $this->urlpath = $this->getOption( 'urlpath', "" );
                if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
                        $this->urlpath .= '/';
                }
                $this->identifier = $this->getOption( 'identifier', wfWikiID() );
                $this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
                $this->dbr = wfGetDB( DB_SLAVE );
                $this->generateNamespaces();
                $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
                $this->findex = fopen( "{$this->fspath}sitemap-index-{$this->identifier}.xml", 'wb' );
                $this->main();
        }

        private function setNamespacePriorities() {
                global $wgSitemapNamespacesPriorities;

                // Custom main namespaces
                $this->priorities[self::GS_MAIN] = '0.5';
                // Custom talk namesspaces
                $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 ( $wgSitemapNamespacesPriorities !== false ) {
                        foreach ( $wgSitemapNamespacesPriorities 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 static function init_path( $fspath ) {
                if ( !isset( $fspath ) ) {
                        return null;
                }
                # Create directory if needed
                if ( $fspath && !is_dir( $fspath ) ) {
                        wfMkdirParents( $fspath, null, __METHOD__ ) or die( "Can not create directory $fspath.\n" );
                }

                return realpath( $fspath ) . DIRECTORY_SEPARATOR ;
        }

        function generateNamespaces() {
                // Only generate for specific namespaces if $wgSitemapNamespaces is an array.
                global $wgSitemapNamespaces;
                if ( is_array( $wgSitemapNamespaces ) ) {
                        $this->namespaces = $wgSitemapNamespaces;
                        return;
                }

                $res = $this->dbr->select( 'page',
                        array( 'page_namespace' ),
                        array(),
                        __METHOD__,
                        array(
                                'GROUP BY' => 'page_namespace',
                                'ORDER BY' => 'page_namespace',
                        )
                );

                foreach ( $res as $row )
                        $this->namespaces[] = $row->page_namespace;
        }

        function priority( $namespace ) {
                return isset( $this->priorities[$namespace] ) ? $this->priorities[$namespace] : $this->guessPriority( $namespace );
        }

        function guessPriority( $namespace ) {
                return MWNamespace::isMain( $namespace ) ? $this->priorities[self::GS_MAIN] : $this->priorities[self::GS_TALK];
        }

        function getPageRes( $namespace ) {
                return $this->dbr->select( 'page',
                        array(
                                'page_namespace',
                                'page_title',
                                'page_touched',
                        ),
                        array( 'page_namespace' => $namespace ),
                        __METHOD__
                );
        }

        public function main() {
                global $wgContLang;

                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 = $wgContLang->getFormattedNsText( $namespace );
                        $this->output( "$namespace ($fns)\n" );
                        foreach ( $res as $row ) {
                                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 ( $wgContLang->hasVariants() ) {
                                        $variants = $wgContLang->getVariants();
                                        foreach ( $variants as $vCode ) {
                                                if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
                                                $entry = $this->fileEntry( $title->getCanonicalURL( '', $vCode ), $date, $this->priority( $namespace ) );
                                                $length += strlen( $entry );
                                                $this->write( $this->file, $entry );
                                        }
                                }
                        }
                        if ( $this->file ) {
                                $this->write( $this->file, $this->closeFile() );
                                $this->close( $this->file );
                        }
                }
                fwrite( $this->findex, $this->closeIndex() );
                fclose( $this->findex );
        }

        function open( $file, $flags ) {
                $resource = $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
                if( $resource === false ) {
                        wfDebugDieBacktrace( __METHOD__ . " error opening file $file with flags $flags. Check permissions?" );
                }
                return $resource;
        }

        function write( &$handle, $str ) {
                if( $handle === true || $handle === false ) {
                        wfDebugDieBacktrace( __METHOD__ . " was passed a boolean as a file handle.\n" );
                }
                if ( $this->compress )
                        gzwrite( $handle, $str );
                else
                        fwrite( $handle, $str );
        }

        function close( &$handle ) {
                if ( $this->compress )
                        gzclose( $handle );
                else
                        fclose( $handle );
        }

        function sitemapFilename( $namespace, $count ) {
                $ext = $this->compress ? '.gz' : '';
                return "sitemap-{$this->identifier}-NS_$namespace-$count.xml$ext";
        }

        function xmlHead() {
                return '<?xml version="1.0" encoding="UTF-8"?>' . "\n";
        }

        function xmlSchema() {
                return 'http://www.sitemaps.org/schemas/sitemap/0.9';
        }

        function openIndex() {
                return $this->xmlHead() . '<sitemapindex xmlns="' . $this->xmlSchema() . '">' . "\n";
        }

        function indexEntry( $filename ) {
                return
                        "\t<sitemap>\n" .
                        "\t\t<loc>{$this->urlpath}$filename</loc>\n" .
                        "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
                        "\t</sitemap>\n";
        }

        function closeIndex() {
                return "</sitemapindex>\n";
        }

        function openFile() {
                return $this->xmlHead() . '<urlset xmlns="' . $this->xmlSchema() . '">' . "\n";
        }

        function fileEntry( $url, $date, $priority ) {
                return
                        "\t<url>\n" .
                        "\t\t<loc>$url</loc>\n" .
                        "\t\t<lastmod>$date</lastmod>\n" .
                        "\t\t<priority>$priority</priority>\n" .
                        "\t</url>\n";
        }

        function closeFile() {
                return "</urlset>\n";
        }

        function generateLimit( $namespace ) {
                // bug 17961: make a title with the longest possible URL in this namespace
                $title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );

                $this->limit = array(
                        strlen( $this->openFile() ),
                        strlen( $this->fileEntry( $title->getCanonicalURL(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
                        strlen( $this->closeFile() )
                );
        }
}

$maintClass = "GenerateSitemap";
require_once( RUN_MAINTENANCE_IF_MAIN );