*/
class GenerateSitemap {
+ /**
+ * The maximum amount of urls in a sitemap file
+ *
+ * @link http://www.google.com/schemas/sitemap/0.84/sitemap.xsd
+ *
+ * @var int
+ */
+ var $url_limit;
+
+ /**
+ * The maximum size of a sitemap file
+ *
+ * @link http://www.google.com/webmasters/sitemaps/docs/en/protocol.html#faq_sitemap_size
+ *
+ * @var int
+ */
+ var $size_limit;
+
/**
* The path to prepend to the filename
*
* @var string
*/
var $path;
-
+
/**
* Whether or not to use compression
*
* @var bool
*/
var $compress;
-
+
/**
* The number of entries to save in each sitemap file
*
- * @var int
+ * @var array
*/
- var $limit;
+ var $limit = array();
/**
* Key => value entries of namespaces and their priorities
* @var object
*/
var $dbr;
-
+
/**
* A resource pointing to the sitemap index file
*
* @var resource
*/
var $findex;
-
+
/**
* A resource pointing to a sitemap file
*/
function GenerateSitemap( $fspath, $path, $compress ) {
global $wgDBname, $wgScriptPath;
-
+
+ $this->url_limit = 50000;
+ $this->size_limit = pow( 2, 20 ) * 10;
$this->fspath = isset( $fspath ) ? $fspath : '';
$this->path = isset( $path ) ? $path : $wgScriptPath;
$this->compress = $compress;
$this->stderr = fopen( 'php://stderr', 'wt' );
$this->dbr =& wfGetDB( DB_SLAVE );
$this->generateNamespaces();
- $this->generateTimestamp();
+ $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
$this->findex = fopen( "{$this->fspath}sitemap-index-$wgDBname.xml", 'wb' );
}
*/
function generateNamespaces() {
$fname = 'GenerateSitemap::generateNamespaces';
-
+
$res = $this->dbr->select( 'page',
array( 'page_namespace' ),
array(),
* If the namespace isn't listed on the priority list return the
* default priority for the namespace, varies depending on whether it's
* a talkpage or not.
- *
+ *
* @param int $namespace The namespace to get the priority for
*
* @return string
$fname = 'GenerateSitemap::getPageRes';
return $this->dbr->select( 'page',
- array(
+ array(
'page_namespace',
'page_title',
- 'page_is_redirect',
'page_touched',
),
array( 'page_namespace' => $namespace ),
* @access public
*/
function main() {
- global $wgDBname;
+ global $wgDBname, $wgContLang;
fwrite( $this->findex, $this->openIndex() );
-
- $this->generateLimit( NS_MAIN );
+
foreach ( $this->namespaces as $namespace ) {
$res = $this->getPageRes( $namespace );
$this->file = false;
+ $this->generateLimit( $namespace );
+ $length = $this->limit[0];
$i = $smcount = 0;
-
- $this->debug( $namespace );
+
+ $fns = $wgContLang->getFormattedNsText( $namespace );
+ $this->debug( "$namespace ($fns)" );
while ( $row = $this->dbr->fetchObject( $res ) ) {
- if ( $i++ % $this->limit === 0 ) {
+ 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 );
}
- $this->generateLimit( $namespace );
$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 ) );
$this->debug( "\t$filename" );
+ $length = $this->limit[0];
+ $i = 1;
}
$title = Title::makeTitle( $row->page_namespace, $row->page_title );
$date = wfTimestamp( TS_ISO_8601, $row->page_touched );
- $this->write( $this->file, $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) ) );
+ $entry = $this->fileEntry( $title->getFullURL(), $date, $this->priority( $namespace ) );
+ $length += strlen( $entry );
+ $this->write( $this->file, $entry );
}
if ( $this->file ) {
$this->write( $this->file, $this->closeFile() );
function open( $file, $flags ) {
return $this->compress ? gzopen( $file, $flags ) : fopen( $file, $flags );
}
-
+
/**
* gzwrite() / fwrite() wrapper
*/
function write( &$handle, $str ) {
if ( $this->compress )
gzwrite( $handle, $str );
- else
+ else
fwrite( $handle, $str );
}
*/
function sitemapFilename( $namespace, $count ) {
global $wgDBname;
-
+
$ext = $this->compress ? '.gz' : '';
-
+
return "sitemap-$wgDBname-NS_$namespace-$count.xml$ext";
}
* @return string
*/
function indexEntry( $filename ) {
- global $wgServer;
-
return
"\t<sitemap>\n" .
- "\t\t<loc>$wgServer{$this->path}/$filename</log>\n" .
- "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
+ "\t\t<loc>$filename</loc>\n" .
+ "\t\t<lastmod>{$this->timestamp}</lastmod>\n" .
"\t</sitemap>\n";
}
}
/**
- * According to the sitemap specification each sitemap must contain no
- * more than 50,000 urls and no more than 2^20 bytes (10MB), this
- * function calculates how many urls we can have in each file assuming
- * that we have the worst case of 63 four byte characters and 1 three
- * byte character in the title (63*4+1*3 = 255)
+ * Populate $this->limit
*/
function generateLimit( $namespace ) {
$title = Title::makeTitle( $namespace, str_repeat( "\xf0\xa8\xae\x81", 63 ) . "\xe5\x96\x83" );
-
- $olen = strlen( $this->openFile() );
- $elen = strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), '1.0' ) );
- $clen = strlen( $this->closeFile() );
-
- for ( $i = 1, $etot = $elen; ( $olen + $clen + $etot + $elen ) <= pow( 2, 20 ); ++$i )
- $etot += $elen;
-
- $this->limit = $i;
- }
- /**
- * Update $this->timestamp to the current time
- */
- function generateTimestamp() {
- $this->timestamp = wfTimestamp( TS_ISO_8601, wfTimestampNow() );
+ $this->limit = array(
+ strlen( $this->openFile() ),
+ strlen( $this->fileEntry( $title->getFullUrl(), wfTimestamp( TS_ISO_8601, wfTimestamp() ), $this->priority( $namespace ) ) ),
+ strlen( $this->closeFile() )
+ );
}
}
-if ( in_array( '--help', $argv ) )
- die(
+if ( in_array( '--help', $argv ) ) {
+ echo
"Usage: php generateSitemap.php [host] [options]\n" .
"\thost = hostname\n" .
"\toptions:\n" .
"\t\t--help\tshow this message\n" .
"\t\t--fspath\tThe file system path to save to, e.g /tmp/sitemap/\n" .
"\t\t--path\tThe http path to use, e.g. /wiki\n" .
- "\t\t--compress=[yes|no]\tcompress the sitemap files, default yes\n"
- );
+ "\t\t--compress=[yes|no]\tcompress the sitemap files, default yes\n";
+ die( -1 );
+}
if ( isset( $argv[1] ) && strpos( $argv[1], '--' ) !== 0 )
$_SERVER['SERVER_NAME'] = $argv[1];
dépôts / lhc / web / wiklou.git / blobdiff