Como coger solo paginas publicadas

[ferg0nar]

Tienes toda la razón, sorry...

es un codigo antiguo que he ido ajustando a los requerimientos actuales de google news.

Código:

header("Content-type: text/xml; charset=UTF-8");
header("Content-encoding: UTF-8");
 
// No direct access
defined( '_JEXEC' ) or die( 'Restricted access' );

$lg = &JFactory::getLanguage();
$language = $lg->getLocale();
$language = str_replace(Array('.UTF-8','@euro'),'',$language[4]);

$config =& JFactory::getConfig();
$use_suffix = $config->getValue('config.sef_suffix');
$offset = $config->getValue('config.offset');
list($hours, $minutes) = explode('.', $offset);
$minutes = 60 * "0.$minutes";
if ($hours >= 0) {
  $hours = "+$hours";
}
if (strlen($hours) == 2) {
  $hours = substr($hours, 0, 1) . '0' . substr($hours, 1, 1);
}
if (strlen($minutes) == 1) {
  $minutes = "0$minutes";
}
$timezone = "$hours:$minutes";

$threedaysago = date('Y-m-d h:i:s',(time()-172800));
 
$db =& JFactory::getDBO();

$get_all_cfg = 'SELECT * FROM #__gnsg_cfg';

$db->setQuery($get_all_cfg);
$rows = $db->loadAssocList();

$cfg = Array();
if (is_array($rows))
	foreach ($rows as $row)
		$cfg[$row['key']] = $row['value'];

if ($cfg['language'])
	$language = $cfg['language'];


switch ($cfg['sef-component']) {
	
	case 'sh404' :
			$query = '
				SELECT 
					oldurl,newurl 
				FROM 
					`jos_redirection` 
				WHERE 
					`newurl` 
				REGEXP 
					"option=com_content&Itemid=[0-9]*&catid=[0-9]*&id=[0-9]*&lang=[a-zA-Z]*&view=article"
				LIMIT 
					5000
			';
			$gg_q = @mysql_query($query);
			$db->setQuery( $query );
			$rows = $db->loadAssocList();
			for ( $a=0; $a<sizeof($rows); $a++ ) {
				$row = $rows[$a];
				$rm = preg_match('/option=com_content&Itemid=[0-9]*&catid=[0-9]*&id=([0-9]*)&lang=[a-zA-Z]*&view=article/',$row['newurl'],$matches);
				
				$lookfor_content = '
					SELECT
						a.created,
						a.title AS title
						a.metakey AS metakey
						a.publish_up AS publish_up
					FROM
						#__content AS a
					WHERE
						a.id = '.$matches[1].' AND
						a.created >="'.$threedaysago.'"
					ORDER BY
						a.created DESC
					LIMIT 1
				';
				
				$db->setQuery( $lookfor_content );
				$thise = $db->loadAssocList();
				if ($thise[0]['created']) {
					$rows[$a] = Array(
						'id' 		=> $matches[1],
						'cont' 		=> '',
						'created' 	=> $thise[0]['created'],
						'cat'		=> '',
						'catid'		=> '',
						'title'		=> $thise[0]['title'],
						'new'		=> $row['oldurl']
					);
				} else {
					$rows[$a]=Array();
				}
			}
		break;
	case 'no-sef' :
	case 'no-modrewrite' :
	case 'core' : 
	default		:
			$query = '
				SELECT 
					a.id,
					a.alias AS cont,
					a.created,
					b.alias AS cat,
					a.catid AS catid,
					c.alias AS `sec`,
					a.sectionid AS secid,
					a.title AS title,			
					a.metakey AS metakey,
					a.publish_up AS publish_up
				FROM 
					#__content AS a, 
					#__categories AS b,
					#__sections as c
				WHERE 
					c.id = a.sectionid AND
					b.id = a.catid  AND
					a.created >="'.$threedaysago.'"
				ORDER BY a.created DESC 
				LIMIT 50000
			';
			$db->setQuery( $query );
			$rows = $db->loadAssocList();
		break;
}

echo '<?xml version="1.0" encoding="UTF-8"?>'.chr(10);
echo '<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"'.chr(10);
echo '        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">'.chr(10);

foreach ($rows as $row) {
	
	if (!$row['id'])
		continue;
	
	if ($cfg['keywords']!='')
		$keyz = htmlspecialchars($cfg['keywords']);
	else
		$keyz = htmlspecialchars($row['metakey']);
	
	switch ($cfg['sef-component']) {
		
		case 'sh404' :
				$url = JURI::root().$row['new'];
			break;
		case 'no-sef':
				$url = JURI::root().'index.php?option=com_content&amp;view=article&amp;id='.$row['id'].':'.$row['cont'].'&amp;catid='.$row['catid'].':'.$row['cat'];
			break;
		case 'no-modrewrite' :
				$url = JURI::root().'noticias/'.$row['cat'].'/'.$row['id'].'-'.$row['cont'].'';
			break;
		case 'core' : 
		default		:
				$find_menu = '
					SELECT 
						alias
					FROM
						#__menu
					WHERE
						link LIKE "index.php?option=com_content&view=section&id='.$row['secid'].'"
				';
				$db->setQuery( $find_menu );
				$menus = $db->loadAssocList();
				if ($menus[0]['alias']) {
					$prefix = $menus[0]['alias'].'/'.$row['catid'].'-';
				} else {
					$prefix = '';
				}
				
				$find_menu = '
					SELECT 
						alias
					FROM
						#__menu
					WHERE
						link LIKE "index.php?option=com_content&view=categoru&id='.$row['catid'].'" OR 
						link LIKE "index.php?option=com_content&view=category&layout=blog&id='.$row['catid'].'" 
				';
				$db->setQuery( $find_menu );
				$menus = $db->loadAssocList();
				if ($menus[0]['alias']) {
					$row['cat'] = $menus[0]['alias'];
				}
				
				if ($use_suffix) {
					$url = JURI::root().$prefix.$row['cat'].'/'.$row['id'].'-'.$row['cont'].'.html';
				} else {
						$url = JURI::root().$prefix.$row['cat'].'/'.$row['id'].'-'.$row['cont'].'/';
				}
			break;
	}
	
	echo '
	<url>
		<loc>'.$url.'</loc>
		<news:news>
			<news:publication>
				<news:name>'.$cfg['pub-name'].'</news:name>
				<news:language>'.es.'</news:language>
			</news:publication>
			<news:publication_date>'.substr($row['created'],0,10).'</news:publication_date>
			<news:title>'.htmlspecialchars($row['title']).'</news:title>
			<news:keywords>'.$keyz.'</news:keywords>
		</news:news>
	</url>';
}

echo '</urlset>'.chr(10);

JRequest::getInt('no_html', 0);

$no_html = intval(JRequest::getInt('no_html', 0));
if ($scriptname != 'index2.php' || $no_html != 1) {
	die();
}