<?php

# Use the rsscreator classes to generate the final merged feed

include("rsscreator.class.php");

# Set our RSS version to 1.0 by default

if(is_null($version)){ $version = '1.0'; }

if(!isset($debug)){

	$debug = true;

	$debug = false;
	
	}

else{ $debug = true; }

if($debug){

	header('Content-type: text/xml');
	
	}


# feedvalidator.org won't validate unless we use the following content type

else{ header('Content-type: application/rss+xml'); }

# Here's my array of feeds

$url_list = array(
	"http://www.profane.eu.org/recent_resources_rss.php",
	"http://www.medev.ac.uk/static/news_rss_feed.php",
	"http://minnesota.ncl.ac.uk/medicaleducation/index.rdf",
	"http://daringfireball.net/index.xml",
	#"http://www.ltsn-01.ac.uk/static/events_rss_feed.php",
);


# We use this to store feed items under a hash key of each feed URL so we still have
# items associated with channels, though we don't use this currently

$url_list_of_hashes = array();

# Hash which uses item dc:dates as keys and URLs as values so that we can sort on this later

$date_hash = array();

# Hash using URL as a key and item hash as value, so that in combo with the above we can output
# in date order

$flat_item_list = array();

# foreach feed url

foreach($url_list as $this_url){

	# Create an entry in the hash for this feed

	$url_list_of_hashes[$this_url] = array();

	# XML for this feed

	$this_xml = '';

	# Open the feed and read in the XML

#	$fd_handle = curl_init($this_url); 

#	$fd_handle = fopen($this_url, "r");

#	while(!feof($fd_handle)){

#		$this_xml.= fread($fd_handle, 1024);

#		}

#	ob_start();

#	curl_exec($fd_handle);

#	curl_close($fd_handle);

#	$this_xml = ob_get_contents();

#	ob_end_clean();

	# I've used Snoopy because fopen could not see past the proxy..

	$this_xml = get_url($this_url);

	if($this_xml == false){ break; }

	# Create our parser

	$xml_parser = xml_parser_create();

	# Use an element handler method for parsing

	xml_set_element_handler($xml_parser, "start_element", "end_element");

	xml_set_character_data_handler($xml_parser, "character_data");	

	# Parse this XML

	xml_parse($xml_parser, $this_xml);

#	fclose($fd_handle);

	xml_parser_free($xml_parser);

	}

# Functions for parsing run

# Flag to tell when we're in an item node

$inside_item = false;

# hash for each item

$this_item = array();

$tag = '';

# holders for item values

$title = '';

$description = '';

$link = '';

$date = '';

function start_element($parser, $name, $attrs){

	global $inside_item, $tag, $title, $description, $link, $date, $this_url, $url_list_of_hashes; 

	if($inside_item){

		$tag = $name;

		}

	elseif($name == "ITEM"){ 

		# If we are at an item node then start parsing

		$inside_item = true; 

		# Create a key for this item in url_list of hashes

		$url_list_of_hashes[$this_url] = array();

		}

	}

function end_element($parser, $name){

	global $inside_item, $tag, $title, $link, $description, $date, $this_item, $this_url, 
	$url_list_of_hashes, $flat_item_list, $date_hash;

	# Hash for our item

	$this_item = array();

	if($name == "ITEM"){

		# Grab ur item values
		
		$this_item['title'] = trim($title);

		$title = ''; 				

		# Use strip tags to remove HTML

		$this_item['description'] = trim(strip_tags($description));

		$description = ''; 				

		$this_item['link'] = trim($link);

		$link = ''; 				

		$this_item['date'] = trim($date);

		$date = ''; 				

		$url_list_of_hashes[$this_url][$this_item['link']] = $this_item;

		# Put the item hash into our flat list hash using the link URL as a key

		$flat_item_list[$this_item['link']] = $this_item;

		# Use the date as a key and as a value set a list of all the URLs posted on that date
	
		if(isset($date_hash[$this_item['date']])){

			array_push($date_hash[$this_item['date']], $this_item['link']);

			}

		else{

			$date_hash[$this_item['date']] = array();
						
			array_push($date_hash[$this_item['date']], $this_item['link']);

			}

		$inside_item = false;	

		}


	}

function character_data($parser, $data){

	global $inside_item, $tag, $title, $description, $link, $date;

	if($inside_item){

		switch($tag){

			case "TITLE": $title .= $data;

			break;

			case "DESCRIPTION": $description .= $data;

			break;

			case "LINK": $link .= $data;

			break;


			case "DC:DATE": $date .= $data;

			break;

			}

		}

	}


# Get a list of our dates

$these_keys = array_keys($date_hash);

# Sort them into order so the latest are first

rsort($these_keys);

# Create an RSSCreator object

$rss = new UniversalRSSCreator();

# Set the channel info

$rss->title = 'Combined BIOME news plus OMNI and VetGate latest additions';

$rss->description = 'This is a combined RSS feed made up of RSS feeds from BIOME including latest news and latest additions to the OMNI and VetGate gateways';

$rss->link = 'http://www.biome.ac.uk/';

$rss->RSSURL = "http://".$_SERVER["SERVER_NAME"].htmlspecialchars($_SERVER["REQUEST_URI"]);

# Now foreach date print out the list of items for that day using the URLs as keys
# This ensures de-duping too. :)

foreach($these_keys as $this_key){

	foreach($date_hash[$this_key] as $this_link){

		$item = new RSSItem();

		$this_item = $flat_item_list[$this_link];

		# Tricky set of regex's where order is critical to sort out ampersand encoding
		# should check its not rsscreator that's messing things up here

		$this_item['description'] = preg_replace("/&amp;/","&", $this_item['description']);

		$this_item['description'] = preg_replace("/&amp;amp;/","&", $this_item['description']);

		$this_item['description'] = preg_replace("/\b&\b/"," &amp; ", $this_item['description']);

		$item->title = $this_item['title'];
		$item->description = $this_item['description'];
		$item->link = $this_item['link'];
		$item->date = $this_item['date'];

		$rss->addItem($item);

		}

	}

$output = $rss->createRSS($version);

# Call an external script to parse out MS meta characters which break validation

include('strip_ms_chars.php');

function get_url($this_url){

	$error_level_stored = error_reporting() ;
  
	error_reporting( $error_level_stored & ~ E_NOTICE ) ;
		
	require_once("magpierss-0.72/extlib/Snoopy.class.inc");
	
	$snoopy = new Snoopy; 

	$snoopy->read_timeout = 5;

	if( ! $snoopy->fetch( $this_url ) || ! $snoopy->results ) {
	
		if (!empty($snoopy->error)) {
	
			$this->_setErrors('Could not open '
			.$this_url
			." snoopy status=".$snoopy->error);
	
		return false;
	

		} 

	else {
		
		$this->_setErrors('Could not open '
		.$this_url);
	
		return false;

		}
	}
	
	$xml = $snoopy->results ;
	
	error_reporting( $error_level_stored ) ;

	return $xml;

	}

?>