<?php
-define( 'GS_MAIN', -2 );
-define( 'GS_TALK', -1 );
/**
- * Creates a sitemap for the site
+ * Creates a sitemap for the site.
*
* Copyright © 2005, Ævar Arnfjörð Bjarmason, Jens Frank <jeluf@gmx.de> and
* Brion Vibber <brion@pobox.com>
* @see http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd
*/
-require_once( dirname( __FILE__ ) . '/Maintenance.php' );
+require_once( __DIR__ . '/Maintenance.php' );
+/**
+ * Maintenance script that generates a sitemap for the site.
+ *
+ * @ingroup Maintenance
+ */
class GenerateSitemap extends Maintenance {
+ const GS_MAIN = -2;
+ const GS_TALK = -1;
+
/**
* The maximum amount of urls in a sitemap file
*
*/
var $compress;
+ /**
+ * Whether or not to include redirection pages
+ *
+ * @var bool
+ */
+ var $skipRedirects;
+
/**
* The number of entries to save in each sitemap file
*
*/
var $file;
+ /**
+ * Identifier to use in filenames, default $wgDBname
+ *
+ * @var string
+ */
+ private $identifier;
+
/**
* Constructor
*/
$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 );
}
/**
if ( $this->urlpath !== "" && substr( $this->urlpath, -1 ) !== '/' ) {
$this->urlpath .= '/';
}
+ $this->identifier = $this->getOption( 'identifier', wfWikiID() );
$this->compress = $this->getOption( 'compress', 'yes' ) !== 'no';
+ $this->skipRedirects = $this->getOption( 'skip-redirects', false ) !== false ;
$this->dbr = wfGetDB( DB_SLAVE );
$this->generateNamespaces();
$this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
- $this->findex = fopen( "{$this->fspath}sitemap-index-" . wfWikiID() . ".xml", 'wb' );
+ $this->findex = fopen( "{$this->fspath}sitemap-index-{$this->identifier}.xml", 'wb' );
$this->main();
}
private function setNamespacePriorities() {
+ global $wgSitemapNamespacesPriorities;
+
// Custom main namespaces
- $this->priorities[GS_MAIN] = '0.5';
+ $this->priorities[self::GS_MAIN] = '0.5';
// Custom talk namesspaces
- $this->priorities[GS_TALK] = '0.1';
+ $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_HELP_TALK] = '0.1';
$this->priorities[NS_CATEGORY] = '0.5';
$this->priorities[NS_CATEGORY_TALK] = '0.1';
+
+ // Custom priorities
+ if ( $wgSitemapNamespacesPriorities !== false ) {
+ /**
+ * @var $wgSitemapNamespacesPriorities array
+ */
+ 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;
+ }
+ }
}
/**
* Create directory if it does not exist and return pathname with a trailing slash
+ * @param $fspath string
+ * @return null|string
*/
private static function init_path( $fspath ) {
if ( !isset( $fspath ) ) {
}
# Create directory if needed
if ( $fspath && !is_dir( $fspath ) ) {
- wfMkdirParents( $fspath ) or die( "Can not create directory $fspath.\n" );
+ wfMkdirParents( $fspath, null, __METHOD__ ) or die( "Can not create directory $fspath.\n" );
}
return realpath( $fspath ) . DIRECTORY_SEPARATOR ;
* @return String
*/
function guessPriority( $namespace ) {
- return MWNamespace::isMain( $namespace ) ? $this->priorities[GS_MAIN] : $this->priorities[GS_TALK];
+ return MWNamespace::isSubject( $namespace ) ? $this->priorities[self::GS_MAIN] : $this->priorities[self::GS_TALK];
}
/**
'page_namespace',
'page_title',
'page_touched',
+ 'page_is_redirect'
),
array( 'page_namespace' => $namespace ),
__METHOD__
$fns = $wgContLang->getFormattedNsText( $namespace );
$this->output( "$namespace ($fns)\n" );
+ $skippedRedirects = 0; // Number of redirects skipped for that namespace
foreach ( $res as $row ) {
+ if ($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() );
}
$title = Title::makeTitle( $row->page_namespace, $row->page_title );
$date = wfTimestamp( TS_ISO_8601, $row->page_touched );
- $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
+ $entry = $this->fileEntry( $title->getCanonicalURL(), $date, $this->priority( $namespace ) );
$length += strlen( $entry );
$this->write( $this->file, $entry );
// generate pages for language variants
$variants = $wgContLang->getVariants();
foreach ( $variants as $vCode ) {
if ( $vCode == $wgContLang->getCode() ) continue; // we don't want default variant
- $entry = $this->fileEntry( $title->getFullURL( '', $vCode ), $date, $this->priority( $namespace ) );
+ $entry = $this->fileEntry( $title->getCanonicalURL( '', $vCode ), $date, $this->priority( $namespace ) );
$length += strlen( $entry );
$this->write( $this->file, $entry );
}
}
}
+
+ if ($this->skipRedirects && $skippedRedirects > 0) {
+ $this->output( " skipped $skippedRedirects redirect(s)\n" );
+ }
+
if ( $this->file ) {
$this->write( $this->file, $this->closeFile() );
$this->close( $this->file );
* @return Resource
*/
function open( $file, $flags ) {
- return $this->compress ? gzopen( $file, $flags ) : fopen( $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 $resource;
}
/**
* gzwrite() / fwrite() wrapper
*/
function write( &$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
*/
function sitemapFilename( $namespace, $count ) {
$ext = $this->compress ? '.gz' : '';
- return "sitemap-" . wfWikiID() . "-NS_$namespace-$count.xml$ext";
+ return "sitemap-{$this->identifier}-NS_$namespace-$count.xml$ext";
}
/**
$this->limit = array(
strlen( $this->openFile() ),
- strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
+ strlen( $this->fileEntry( $title->getCanonicalURL(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
strlen( $this->closeFile() )
);
}
dépôts / lhc / web / wiklou.git / blobdiff