update to 3.2 version of full-text-rss, issue #694

[github/wallabag/wallabag.git] / inc / 3rdparty / makefulltextfeed.php

<?php\r
// Full-Text RSS: Create Full-Text Feeds\r
// Author: Keyvan Minoukadeh\r
// Copyright (c) 2013 Keyvan Minoukadeh\r
// License: AGPLv3\r
// Version: 3.2\r
// Date: 2013-05-13\r
// More info: http://fivefilters.org/content-only/\r
// Help: http://help.fivefilters.org\r
\r
/*\r
This program is free software: you can redistribute it and/or modify\r
it under the terms of the GNU Affero General Public License as published by\r
the Free Software Foundation, either version 3 of the License, or\r
(at your option) any later version.\r
\r
This program is distributed in the hope that it will be useful,\r
but WITHOUT ANY WARRANTY; without even the implied warranty of\r
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the\r
GNU Affero General Public License for more details.\r
\r
You should have received a copy of the GNU Affero General Public License\r
along with this program.  If not, see <http://www.gnu.org/licenses/>.\r
*/\r
\r
// Usage\r
// -----\r
// Request this file passing it a web page or feed URL in the querystring: makefulltextfeed.php?url=example.org/article\r
// For more request parameters, see http://help.fivefilters.org/customer/portal/articles/226660-usage\r
\r
error_reporting(E_ALL ^ E_NOTICE);\r
ini_set("display_errors", 1);\r
@set_time_limit(120);\r
\r
// Deal with magic quotes\r
if (get_magic_quotes_gpc()) {\r
	$process = array(&$_GET, &$_POST, &$_REQUEST);\r
	while (list($key, $val) = each($process)) {\r
		foreach ($val as $k => $v) {\r
			unset($process[$key][$k]);\r
			if (is_array($v)) {\r
				$process[$key][stripslashes($k)] = $v;\r
				$process[] = &$process[$key][stripslashes($k)];\r
			} else {\r
				$process[$key][stripslashes($k)] = stripslashes($v);\r
			}\r
		}\r
	}\r
	unset($process);\r
}\r
\r
// set include path\r
set_include_path(realpath(dirname(__FILE__).'/libraries').PATH_SEPARATOR.get_include_path());\r
\r
require_once dirname(__FILE__).'/makefulltextfeedHelpers.php';\r
\r
////////////////////////////////\r
// Load config file\r
////////////////////////////////\r
require dirname(__FILE__).'/config.php';\r
\r
////////////////////////////////\r
// Prevent indexing/following by search engines because:\r
// 1. The content is already public and presumably indexed (why create duplicates?)\r
// 2. Not doing so might increase number of requests from search engines, thus increasing server load\r
// Note: feed readers and services such as Yahoo Pipes will not be affected by this header.\r
// Note: Using Disallow in a robots.txt file will be more effective (search engines will check\r
// that before even requesting makefulltextfeed.php).\r
////////////////////////////////\r
header('X-Robots-Tag: noindex, nofollow');\r
\r
////////////////////////////////\r
// Check if service is enabled\r
////////////////////////////////\r
if (!$options->enabled) {\r
	die('The full-text RSS service is currently disabled');\r
}\r
\r
////////////////////////////////\r
// Debug mode?\r
// See the config file for debug options.\r
////////////////////////////////\r
$debug_mode = false;\r
if (isset($_GET['debug'])) {\r
	if ($options->debug === true || $options->debug == 'user') {\r
		$debug_mode = true;\r
	} elseif ($options->debug == 'admin') {\r
		session_start();\r
		$debug_mode = (@$_SESSION['auth'] == 1);\r
	}\r
	if ($debug_mode) {\r
		header('Content-Type: text/plain; charset=utf-8');\r
	} else {\r
		if ($options->debug == 'admin') {\r
			die('You must be logged in to the <a href="admin/">admin area</a> to see debug output.');\r
		} else {\r
			die('Debugging is disabled.');\r
		}\r
	}\r
}\r
\r
////////////////////////////////\r
// Check for APC\r
////////////////////////////////\r
$options->apc = $options->apc && function_exists('apc_add');\r
if ($options->apc) {\r
	debug('APC is enabled and available on server');\r
} else {\r
	debug('APC is disabled or not available on server');\r
}\r
\r
////////////////////////////////\r
// Check for smart cache\r
////////////////////////////////\r
$options->smart_cache = $options->smart_cache && function_exists('apc_inc');\r
\r
////////////////////////////////\r
// Check for feed URL\r
////////////////////////////////\r
if (!isset($_GET['url'])) {\r
	die('No URL supplied');\r
}\r
$url = trim($_GET['url']);\r
if (strtolower(substr($url, 0, 7)) == 'feed://') {\r
	$url = 'http://'.substr($url, 7);\r
}\r
if (!preg_match('!^https?://.+!i', $url)) {\r
	$url = 'http://'.$url;\r
}\r
\r
$url = filter_var($url, FILTER_SANITIZE_URL);\r
$test = filter_var($url, FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);\r
// deal with bug http://bugs.php.net/51192 (present in PHP 5.2.13 and PHP 5.3.2)\r
if ($test === false) {\r
	$test = filter_var(strtr($url, '-', '_'), FILTER_VALIDATE_URL, FILTER_FLAG_SCHEME_REQUIRED);\r
}\r
if ($test !== false && $test !== null && preg_match('!^https?://!', $url)) {\r
	// all okay\r
	unset($test);\r
} else {\r
	die('Invalid URL supplied');\r
}\r
debug("Supplied URL: $url");\r
\r
/////////////////////////////////\r
// Redirect to hide API key\r
/////////////////////////////////\r
if (isset($_GET['key']) && ($key_index = array_search($_GET['key'], $options->api_keys)) !== false) {\r
	$host = $_SERVER['HTTP_HOST'];\r
	$path = rtrim(dirname($_SERVER['SCRIPT_NAME']), '/\\');\r
	$_qs_url = (strtolower(substr($url, 0, 7)) == 'http://') ? substr($url, 7) : $url;\r
	$redirect = 'http://'.htmlspecialchars($host.$path).'/makefulltextfeed.php?url='.urlencode($_qs_url);\r
	$redirect .= '&key='.$key_index;\r
	$redirect .= '&hash='.urlencode(sha1($_GET['key'].$url));\r
	if (isset($_GET['html'])) $redirect .= '&html='.urlencode($_GET['html']);\r
	if (isset($_GET['max'])) $redirect .= '&max='.(int)$_GET['max'];\r
	if (isset($_GET['links'])) $redirect .= '&links='.urlencode($_GET['links']);\r
	if (isset($_GET['exc'])) $redirect .= '&exc='.urlencode($_GET['exc']);\r
	if (isset($_GET['format'])) $redirect .= '&format='.urlencode($_GET['format']);\r
	if (isset($_GET['callback'])) $redirect .= '&callback='.urlencode($_GET['callback']);\r
	if (isset($_GET['l'])) $redirect .= '&l='.urlencode($_GET['l']);\r
	if (isset($_GET['xss'])) $redirect .= '&xss';\r
	if (isset($_GET['use_extracted_title'])) $redirect .= '&use_extracted_title';\r
	if (isset($_GET['content'])) $redirect .= '&content='.urlencode($_GET['content']);\r
	if (isset($_GET['summary'])) $redirect .= '&summary='.urlencode($_GET['summary']);\r
	if (isset($_GET['debug'])) $redirect .= '&debug';\r
	if ($debug_mode) {\r
		debug('Redirecting to hide access key, follow URL below to continue');\r
		debug("Location: $redirect");\r
	} else {\r
		header("Location: $redirect");\r
	}\r
	exit;\r
}\r
\r
///////////////////////////////////////////////\r
// Set timezone.\r
// Prevents warnings, but needs more testing -\r
// perhaps if timezone is set in php.ini we\r
// don't need to set it at all...\r
///////////////////////////////////////////////\r
if (!ini_get('date.timezone') || !@date_default_timezone_set(ini_get('date.timezone'))) {\r
	date_default_timezone_set('UTC');\r
}\r
\r
///////////////////////////////////////////////\r
// Check if the request is explicitly for an HTML page\r
///////////////////////////////////////////////\r
$html_only = (isset($_GET['html']) && ($_GET['html'] == '1' || $_GET['html'] == 'true'));\r
\r
///////////////////////////////////////////////\r
// Check if valid key supplied\r
///////////////////////////////////////////////\r
$valid_key = false;\r
if (isset($_GET['key']) && isset($_GET['hash']) && isset($options->api_keys[(int)$_GET['key']])) {\r
	$valid_key = ($_GET['hash'] == sha1($options->api_keys[(int)$_GET['key']].$url));\r
}\r
$key_index = ($valid_key) ? (int)$_GET['key'] : 0;\r
if (!$valid_key && $options->key_required) {\r
	die('A valid key must be supplied');\r
}\r
if (!$valid_key && isset($_GET['key']) && $_GET['key'] != '') {\r
	die('The entered key is invalid');\r
}\r
\r
if (file_exists('custom_init.php')) require 'custom_init.php';\r
\r
///////////////////////////////////////////////\r
// Check URL against list of blacklisted URLs\r
///////////////////////////////////////////////\r
if (!url_allowed($url)) die('URL blocked');\r
\r
///////////////////////////////////////////////\r
// Max entries\r
// see config.php to find these values\r
///////////////////////////////////////////////\r
if (isset($_GET['max'])) {\r
	$max = (int)$_GET['max'];\r
	if ($valid_key) {\r
		$max = min($max, $options->max_entries_with_key);\r
	} else {\r
		$max = min($max, $options->max_entries);\r
	}\r
} else {\r
	if ($valid_key) {\r
		$max = $options->default_entries_with_key;\r
	} else {\r
		$max = $options->default_entries;\r
	}\r
}\r
\r
///////////////////////////////////////////////\r
// Link handling\r
///////////////////////////////////////////////\r
if (isset($_GET['links']) && in_array($_GET['links'], array('preserve', 'footnotes', 'remove'))) {\r
	$links = $_GET['links'];\r
} else {\r
	$links = 'preserve';\r
}\r
\r
///////////////////////////////////////////////\r
// Favour item titles in feed?\r
///////////////////////////////////////////////\r
$favour_feed_titles = true;\r
if ($options->favour_feed_titles == 'user') {\r
	$favour_feed_titles = !isset($_GET['use_extracted_title']);\r
} else {\r
	$favour_feed_titles = $options->favour_feed_titles;\r
}\r
\r
///////////////////////////////////////////////\r
// Include full content in output?\r
///////////////////////////////////////////////\r
if ($options->content === 'user') {\r
	if (isset($_GET['content']) && $_GET['content'] === '0') {\r
		$options->content = false;\r
	} else {\r
		$options->content = true;\r
	}\r
}\r
\r
///////////////////////////////////////////////\r
// Include summaries in output?\r
///////////////////////////////////////////////\r
if ($options->summary === 'user') {\r
	if (isset($_GET['summary']) && $_GET['summary'] === '1') {\r
		$options->summary = true;\r
	} else {\r
		$options->summary = false;\r
	}\r
}\r
\r
///////////////////////////////////////////////\r
// Exclude items if extraction fails\r
///////////////////////////////////////////////\r
if ($options->exclude_items_on_fail === 'user') {\r
	$exclude_on_fail = (isset($_GET['exc']) && ($_GET['exc'] == '1'));\r
} else {\r
	$exclude_on_fail = $options->exclude_items_on_fail;\r
}\r
\r
///////////////////////////////////////////////\r
// Detect language\r
///////////////////////////////////////////////\r
if ($options->detect_language === 'user') {\r
	if (isset($_GET['l'])) {\r
		$detect_language = (int)$_GET['l'];\r
	} else {\r
		$detect_language = 1;\r
	}\r
} else {\r
	$detect_language = $options->detect_language;\r
}\r
\r
$use_cld = extension_loaded('cld') && (version_compare(PHP_VERSION, '5.3.0') >= 0);\r
\r
/////////////////////////////////////\r
// Check for valid format\r
// (stick to RSS (or RSS as JSON) for the time being)\r
/////////////////////////////////////\r
if (isset($_GET['format']) && $_GET['format'] == 'json') {\r
	$format = 'json';\r
} else {\r
	$format = 'rss';\r
}\r
\r
/////////////////////////////////////\r
// Should we do XSS filtering?\r
/////////////////////////////////////\r
if ($options->xss_filter === 'user') {\r
	$xss_filter = isset($_GET['xss']);\r
} else {\r
	$xss_filter = $options->xss_filter;\r
}\r
if (!$xss_filter && isset($_GET['xss'])) {\r
	die('XSS filtering is disabled in config');\r
}\r
\r
/////////////////////////////////////\r
// Check for JSONP\r
// Regex from https://gist.github.com/1217080\r
/////////////////////////////////////\r
$callback = null;\r
if ($format =='json' && isset($_GET['callback'])) {\r
	$callback = trim($_GET['callback']);\r
	foreach (explode('.', $callback) as $_identifier) {\r
		if (!preg_match('/^[a-zA-Z_$][0-9a-zA-Z_$]*(?:\[(?:".+"|\'.+\'|\d+)\])*?$/', $_identifier)) {\r
			die('Invalid JSONP callback');\r
		}\r
	}\r
	debug("JSONP callback: $callback");\r
}\r
\r
//////////////////////////////////\r
// Enable Cross-Origin Resource Sharing (CORS)\r
//////////////////////////////////\r
if ($options->cors) header('Access-Control-Allow-Origin: *');\r
\r
//////////////////////////////////\r
// Check for cached copy\r
//////////////////////////////////\r
if ($options->caching) {\r
	debug('Caching is enabled...');\r
	$cache_id = md5($max.$url.(int)$valid_key.$links.(int)$favour_feed_titles.(int)$options->content.(int)$options->summary.(int)$xss_filter.(int)$exclude_on_fail.$format.$detect_language.(int)isset($_GET['pubsub']));\r
	$check_cache = true;\r
	if ($options->apc && $options->smart_cache) {\r
		apc_add("cache.$cache_id", 0, 10*60);\r
		$apc_cache_hits = (int)apc_fetch("cache.$cache_id");\r
		$check_cache = ($apc_cache_hits >= 2);\r
		apc_inc("cache.$cache_id");\r
		if ($check_cache) {\r
			debug('Cache key found in APC, we\'ll try to load cache file from disk');\r
		} else {\r
			debug('Cache key not found in APC');\r
		}\r
	}\r
	if ($check_cache) {\r
		$cache = get_cache();\r
		if ($data = $cache->load($cache_id)) {\r
			if ($debug_mode) {\r
				debug('Loaded cached copy');\r
				exit;\r
			}\r
			if ($format == 'json') {\r
				if ($callback === null) {\r
					header('Content-type: application/json; charset=UTF-8');\r
				} else {\r
					header('Content-type: application/javascript; charset=UTF-8');\r
				}\r
			} else {\r
				header('Content-type: text/xml; charset=UTF-8');\r
				header('X-content-type-options: nosniff');\r
			}\r
			if (headers_sent()) die('Some data has already been output, can\'t send RSS file');\r
			if ($callback) {\r
				echo "$callback($data);";\r
			} else {\r
				echo $data;\r
			}\r
			exit;\r
		}\r
	}\r
}\r
\r
//////////////////////////////////\r
// Set Expires header\r
//////////////////////////////////\r
if (!$debug_mode) {\r
	header('Expires: ' . gmdate('D, d M Y H:i:s', time()+(60*10)) . ' GMT');\r
}\r
\r
//////////////////////////////////\r
// Set up HTTP agent\r
//////////////////////////////////\r
global $http;\r
$http = new HumbleHttpAgent();\r
$http->debug = $debug_mode;\r
$http->userAgentMap = $options->user_agents;\r
$http->headerOnlyTypes = array_keys($options->content_type_exc);\r
$http->rewriteUrls = $options->rewrite_url;\r
\r
//////////////////////////////////\r
// Set up Content Extractor\r
//////////////////////////////////\r
global $extractor;\r
$extractor = new ContentExtractor(dirname(__FILE__).'/site_config/custom', dirname(__FILE__).'/site_config/standard');\r
$extractor->debug = $debug_mode;\r
SiteConfig::$debug = $debug_mode;\r
SiteConfig::use_apc($options->apc);\r
$extractor->fingerprints = $options->fingerprints;\r
$extractor->allowedParsers = $options->allowed_parsers;\r
\r
////////////////////////////////\r
// Get RSS/Atom feed\r
////////////////////////////////\r
if (!$html_only) {\r
	debug('--------');\r
	debug("Attempting to process URL as feed");\r
	// Send user agent header showing PHP (prevents a HTML response from feedburner)\r
	$http->userAgentDefault = HumbleHttpAgent::UA_PHP;\r
	// configure SimplePie HTTP extension class to use our HumbleHttpAgent instance\r
	SimplePie_HumbleHttpAgent::set_agent($http);\r
	$feed = new SimplePie();\r
	// some feeds use the text/html content type - force_feed tells SimplePie to process anyway\r
	$feed->force_feed(true);\r
	$feed->set_file_class('SimplePie_HumbleHttpAgent');\r
	//$feed->set_feed_url($url); // colons appearing in the URL's path get encoded\r
	$feed->feed_url = $url;\r
	$feed->set_autodiscovery_level(SIMPLEPIE_LOCATOR_NONE);\r
	$feed->set_timeout(20);\r
	$feed->enable_cache(false);\r
	$feed->set_stupidly_fast(true);\r
	$feed->enable_order_by_date(false); // we don't want to do anything to the feed\r
	$feed->set_url_replacements(array());\r
	// initialise the feed\r
	// the @ suppresses notices which on some servers causes a 500 internal server error\r
	$result = @$feed->init();\r
	//$feed->handle_content_type();\r
	//$feed->get_title();\r
	if ($result && (!is_array($feed->data) || count($feed->data) == 0)) {\r
		die('Sorry, no feed items found');\r
	}\r
	// from now on, we'll identify ourselves as a browser\r
	$http->userAgentDefault = HumbleHttpAgent::UA_BROWSER;\r
}\r
\r
////////////////////////////////////////////////////////////////////////////////\r
// Our given URL is not a feed, so let's create our own feed with a single item:\r
// the given URL. This basically treats all non-feed URLs as if they were\r
// single-item feeds.\r
////////////////////////////////////////////////////////////////////////////////\r
$isDummyFeed = false;\r
if ($html_only || !$result) {\r
	debug('--------');\r
	debug("Constructing a single-item feed from URL");\r
	$isDummyFeed = true;\r
	unset($feed, $result);\r
	// create single item dummy feed object\r
	$feed = new DummySingleItemFeed($url);\r
}\r
\r
////////////////////////////////////////////\r
// Create full-text feed\r
////////////////////////////////////////////\r
$output = new FeedWriter();\r
$output->setTitle(strip_tags($feed->get_title()));\r
$output->setDescription(strip_tags($feed->get_description()));\r
$output->setXsl('css/feed.xsl'); // Chrome uses this, most browsers ignore it\r
if ($valid_key && isset($_GET['pubsub'])) { // used only on fivefilters.org at the moment\r
	$output->addHub('http://fivefilters.superfeedr.com/');\r
	$output->addHub('http://pubsubhubbub.appspot.com/');\r
	$output->setSelf('http://'.$_SERVER['HTTP_HOST'].$_SERVER['REQUEST_URI']);\r
}\r
$output->setLink($feed->get_link()); // Google Reader uses this for pulling in favicons\r
if ($img_url = $feed->get_image_url()) {\r
	$output->setImage($feed->get_title(), $feed->get_link(), $img_url);\r
}\r
\r
////////////////////////////////////////////\r
// Loop through feed items\r
////////////////////////////////////////////\r
$items = $feed->get_items(0, $max);\r
// Request all feed items in parallel (if supported)\r
$urls_sanitized = array();\r
$urls = array();\r
foreach ($items as $key => $item) {\r
	$permalink = htmlspecialchars_decode($item->get_permalink());\r
	// Colons in URL path segments get encoded by SimplePie, yet some sites expect them unencoded\r
	$permalink = str_replace('%3A', ':', $permalink);\r
	// validateUrl() strips non-ascii characters\r
	// simplepie already sanitizes URLs so let's not do it again here.\r
	//$permalink = $http->validateUrl($permalink);\r
	if ($permalink) {\r
		$urls_sanitized[] = $permalink;\r
	}\r
	$urls[$key] = $permalink;\r
}\r
debug('--------');\r
debug('Fetching feed items');\r
$http->fetchAll($urls_sanitized);\r
//$http->cacheAll();\r
\r
// count number of items added to full feed\r
$item_count = 0;\r
\r
foreach ($items as $key => $item) {\r
	debug('--------');\r
	debug('Processing feed item '.($item_count+1));\r
	$do_content_extraction = true;\r
	$extract_result = false;\r
	$text_sample = null;\r
	$permalink = $urls[$key];\r
	debug("Item URL: $permalink");\r
	$extracted_title = '';\r
	$feed_item_title = $item->get_title();\r
	if ($feed_item_title !== null) {\r
		$feed_item_title = strip_tags(htmlspecialchars_decode($feed_item_title));\r
	}\r
	$newitem = $output->createNewItem();\r
	$newitem->setTitle($feed_item_title);\r
	if ($valid_key && isset($_GET['pubsub'])) { // used only on fivefilters.org at the moment\r
		if ($permalink !== false) {\r
			$newitem->setLink('http://fivefilters.org/content-only/redirect.php?url='.urlencode($permalink));\r
		} else {\r
			$newitem->setLink('http://fivefilters.org/content-only/redirect.php?url='.urlencode($item->get_permalink()));\r
		}\r
	} else {\r
		if ($permalink !== false) {\r
			$newitem->setLink($permalink);\r
		} else {\r
			$newitem->setLink($item->get_permalink());\r
		}\r
	}\r
	//if ($permalink && ($response = $http->get($permalink, true)) && $response['status_code'] < 300) {\r
	// Allowing error codes - some sites return correct content with error status\r
	// e.g. prospectmagazine.co.uk returns 403\r
	if ($permalink && ($response = $http->get($permalink, true)) && ($response['status_code'] < 300 || $response['status_code'] > 400)) {\r
		$effective_url = $response['effective_url'];\r
		if (!url_allowed($effective_url)) continue;\r
		// check if action defined for returned Content-Type\r
		$mime_info = get_mime_action_info($response['headers']);\r
		if (isset($mime_info['action'])) {\r
			if ($mime_info['action'] == 'exclude') {\r
				continue; // skip this feed item entry\r
			} elseif ($mime_info['action'] == 'link') {\r
				if ($mime_info['type'] == 'image') {\r
					$html = "<a href=\"$effective_url\"><img src=\"$effective_url\" alt=\"{$mime_info['name']}\" /></a>";\r
				} else {\r
					$html = "<a href=\"$effective_url\">Download {$mime_info['name']}</a>";\r
				}\r
				$extracted_title = $mime_info['name'];\r
				$do_content_extraction = false;\r
			}\r
		}\r
		if ($do_content_extraction) {\r
			$html = $response['body'];\r
			// remove strange things\r
			$html = str_replace('</[>', '', $html);\r
			$html = convert_to_utf8($html, $response['headers']);\r
			// check site config for single page URL - fetch it if found\r
			$is_single_page = false;\r
			if ($single_page_response = getSinglePage($item, $html, $effective_url)) {\r
				$is_single_page = true;\r
				$effective_url = $single_page_response['effective_url'];\r
				// check if action defined for returned Content-Type\r
				$mime_info = get_mime_action_info($single_page_response['headers']);\r
				if (isset($mime_info['action'])) {\r
					if ($mime_info['action'] == 'exclude') {\r
						continue; // skip this feed item entry\r
					} elseif ($mime_info['action'] == 'link') {\r
						if ($mime_info['type'] == 'image') {\r
							$html = "<a href=\"$effective_url\"><img src=\"$effective_url\" alt=\"{$mime_info['name']}\" /></a>";\r
						} else {\r
							$html = "<a href=\"$effective_url\">Download {$mime_info['name']}</a>";\r
						}\r
						$extracted_title = $mime_info['name'];\r
						$do_content_extraction = false;\r
					}\r
				}\r
				if ($do_content_extraction) {\r
					$html = $single_page_response['body'];\r
					// remove strange things\r
					$html = str_replace('</[>', '', $html);\r
					$html = convert_to_utf8($html, $single_page_response['headers']);\r
					debug("Retrieved single-page view from $effective_url");\r
				}\r
				unset($single_page_response);\r
			}\r
		}\r
		if ($do_content_extraction) {\r
			debug('--------');\r
			debug('Attempting to extract content');\r
			$extract_result = $extractor->process($html, $effective_url);\r
			$readability = $extractor->readability;\r
			$content_block = ($extract_result) ? $extractor->getContent() : null;\r
			$extracted_title = ($extract_result) ? $extractor->getTitle() : '';\r
			// Deal with multi-page articles\r
			//die('Next: '.$extractor->getNextPageUrl());\r
			$is_multi_page = (!$is_single_page && $extract_result && $extractor->getNextPageUrl());\r
			if ($options->multipage && $is_multi_page && $options->content) {\r
				debug('--------');\r
				debug('Attempting to process multi-page article');\r
				$multi_page_urls = array();\r
				$multi_page_content = array();\r
				while ($next_page_url = $extractor->getNextPageUrl()) {\r
					debug('--------');\r
					debug('Processing next page: '.$next_page_url);\r
					// If we've got URL, resolve against $url\r
					if ($next_page_url = makeAbsoluteStr($effective_url, $next_page_url)) {\r
						// check it's not what we have already!\r
						if (!in_array($next_page_url, $multi_page_urls)) {\r
							// it's not, so let's attempt to fetch it\r
							$multi_page_urls[] = $next_page_url;\r
							$_prev_ref = $http->referer;\r
							if (($response = $http->get($next_page_url, true)) && $response['status_code'] < 300) {\r
								// make sure mime type is not something with a different action associated\r
								$page_mime_info = get_mime_action_info($response['headers']);\r
								if (!isset($page_mime_info['action'])) {\r
									$html = $response['body'];\r
									// remove strange things\r
									$html = str_replace('</[>', '', $html);\r
									$html = convert_to_utf8($html, $response['headers']);\r
									if ($extractor->process($html, $next_page_url)) {\r
										$multi_page_content[] = $extractor->getContent();\r
										continue;\r
									} else { debug('Failed to extract content'); }\r
								} else { debug('MIME type requires different action'); }\r
							} else { debug('Failed to fetch URL'); }\r
						} else { debug('URL already processed'); }\r
					} else { debug('Failed to resolve against '.$effective_url); }\r
					// failed to process next_page_url, so cancel further requests\r
					$multi_page_content = array();\r
					break;\r
				}\r
				// did we successfully deal with this multi-page article?\r
				if (empty($multi_page_content)) {\r
					debug('Failed to extract all parts of multi-page article, so not going to include them');\r
					$_page = $readability->dom->createElement('p');\r
					$_page->innerHTML = '<em>This article appears to continue on subsequent pages which we could not extract</em>';\r
					$multi_page_content[] = $_page;\r
				}\r
				foreach ($multi_page_content as $_page) {\r
					$_page = $content_block->ownerDocument->importNode($_page, true);\r
					$content_block->appendChild($_page);\r
				}\r
				unset($multi_page_urls, $multi_page_content, $page_mime_info, $next_page_url, $_page);\r
			}\r
		}\r
		// use extracted title for both feed and item title if we're using single-item dummy feed\r
		if ($isDummyFeed) {\r
			$output->setTitle($extracted_title);\r
			$newitem->setTitle($extracted_title);\r
		} else {\r
			// use extracted title instead of feed item title?\r
			if (!$favour_feed_titles && $extracted_title != '') {\r
				debug('Using extracted title in generated feed');\r
				$newitem->setTitle($extracted_title);\r
			}\r
		}\r
	}\r
	if ($do_content_extraction) {\r
		// if we failed to extract content...\r
		if (!$extract_result) {\r
			if ($exclude_on_fail) {\r
				debug('Failed to extract, so skipping (due to exclude on fail parameter)');\r
				continue; // skip this and move to next item\r
			}\r
			//TODO: get text sample for language detection\r
			$html = $options->error_message;\r
			// keep the original item description\r
			$html .= $item->get_description();\r
		} else {\r
			$readability->clean($content_block, 'select');\r
			if ($options->rewrite_relative_urls) makeAbsolute($effective_url, $content_block);\r
			// footnotes\r
			if (($links == 'footnotes') && (strpos($effective_url, 'wikipedia.org') === false)) {\r
				$readability->addFootnotes($content_block);\r
			}\r
			// remove nesting: <div><div><div><p>test</p></div></div></div> = <p>test</p>\r
			while ($content_block->childNodes->length == 1 && $content_block->firstChild->nodeType === XML_ELEMENT_NODE) {\r
				// only follow these tag names\r
				if (!in_array(strtolower($content_block->tagName), array('div', 'article', 'section', 'header', 'footer'))) break;\r
				//$html = $content_block->firstChild->innerHTML; // FTR 2.9.5\r
				$content_block = $content_block->firstChild;\r
			}\r
			// convert content block to HTML string\r
			// Need to preserve things like body: //img[@id='feature']\r
			if (in_array(strtolower($content_block->tagName), array('div', 'article', 'section', 'header', 'footer'))) {\r
				$html = $content_block->innerHTML;\r
			} else {\r
				$html = $content_block->ownerDocument->saveXML($content_block); // essentially outerHTML\r
			}\r
			//unset($content_block);\r
			// post-processing cleanup\r
			$html = preg_replace('!<p>[\s\h\v]*</p>!u', '', $html);\r
			if ($links == 'remove') {\r
				$html = preg_replace('!</?a[^>]*>!', '', $html);\r
			}\r
			// get text sample for language detection\r
			$text_sample = strip_tags(substr($html, 0, 500));\r
			$html = make_substitutions($options->message_to_prepend).$html;\r
			$html .= make_substitutions($options->message_to_append);\r
		}\r
	}\r
\r
	if ($valid_key && isset($_GET['pubsub'])) { // used only on fivefilters.org at the moment\r
		$newitem->addElement('guid', 'http://fivefilters.org/content-only/redirect.php?url='.urlencode($item->get_permalink()), array('isPermaLink'=>'false'));\r
	} else {\r
		$newitem->addElement('guid', $item->get_permalink(), array('isPermaLink'=>'true'));\r
	}\r
	// filter xss?\r
	if ($xss_filter) {\r
		debug('Filtering HTML to remove XSS');\r
		$html = htmLawed::hl($html, array('safe'=>1, 'deny_attribute'=>'style', 'comment'=>1, 'cdata'=>1));\r
	}\r
\r
	// add content\r
	if ($options->summary === true) {\r
		// get summary\r
		$summary = '';\r
		if (!$do_content_extraction) {\r
			$summary = $html;\r
		} else {\r
			// Try to get first few paragraphs\r
			if (isset($content_block) && ($content_block instanceof DOMElement)) {\r
				$_paras = $content_block->getElementsByTagName('p');\r
				foreach ($_paras as $_para) {\r
					$summary .= preg_replace("/[\n\r\t ]+/", ' ', $_para->textContent).' ';\r
					if (strlen($summary) > 200) break;\r
				}\r
			} else {\r
				$summary = $html;\r
			}\r
		}\r
		unset($_paras, $_para);\r
		$summary = get_excerpt($summary);\r
		$newitem->setDescription($summary);\r
		if ($options->content) $newitem->setElement('content:encoded', $html);\r
	} else {\r
		if ($options->content) $newitem->setDescription($html);\r
	}\r
\r
	// set date\r
	if ((int)$item->get_date('U') > 0) {\r
		$newitem->setDate((int)$item->get_date('U'));\r
	} elseif ($extractor->getDate()) {\r
		$newitem->setDate($extractor->getDate());\r
	}\r
\r
	// add authors\r
	if ($authors = $item->get_authors()) {\r
		foreach ($authors as $author) {\r
			// for some feeds, SimplePie stores author's name as email, e.g. http://feeds.feedburner.com/nymag/intel\r
			if ($author->get_name() !== null) {\r
				$newitem->addElement('dc:creator', $author->get_name());\r
			} elseif ($author->get_email() !== null) {\r
				$newitem->addElement('dc:creator', $author->get_email());\r
			}\r
		}\r
	} elseif ($authors = $extractor->getAuthors()) {\r
		//TODO: make sure the list size is reasonable\r
		foreach ($authors as $author) {\r
			// TODO: xpath often selects authors from other articles linked from the page.\r
			// for now choose first item\r
			$newitem->addElement('dc:creator', $author);\r
			break;\r
		}\r
	}\r
\r
	// add language\r
	if ($detect_language) {\r
		$language = $extractor->getLanguage();\r
		if (!$language) $language = $feed->get_language();\r
		if (($detect_language == 3 || (!$language && $detect_language == 2)) && $text_sample) {\r
			try {\r
				if ($use_cld) {\r
					// Use PHP-CLD extension\r
					$php_cld = 'CLD\detect'; // in quotes to prevent PHP 5.2 parse error\r
					$res = $php_cld($text_sample);\r
					if (is_array($res) && count($res) > 0) {\r
						$language = $res[0]['code'];\r
					}\r
				} else {\r
					//die('what');\r
					// Use PEAR's Text_LanguageDetect\r
					if (!isset($l))	{\r
					  $l = new Text_LanguageDetect();\r
					  $l->setNameMode(2); // return ISO 639-1 codes (e.g. "en")\r
					}\r
					$l_result = $l->detect($text_sample, 1);\r
					if (count($l_result) > 0) {\r
						$language = key($l_result);\r
					}\r
				}\r
			} catch (Exception $e) {\r
				//die('error: '.$e);\r
				// do nothing\r
			}\r
		}\r
		if ($language && (strlen($language) < 7)) {\r
			$newitem->addElement('dc:language', $language);\r
		}\r
	}\r
\r
	// add MIME type (if it appeared in our exclusions lists)\r
	if (isset($mime_info['mime'])) $newitem->addElement('dc:format', $mime_info['mime']);\r
	// add effective URL (URL after redirects)\r
	if (isset($effective_url)) {\r
		//TODO: ensure $effective_url is valid witout - sometimes it causes problems, e.g.\r
		//http://www.siasat.pk/forum/showthread.php?108883-Pakistan-Chowk-by-Rana-Mubashir-�-25th-March-2012-Special-Program-from-Liari-(Karachi)\r
		//temporary measure: use utf8_encode()\r
		$newitem->addElement('dc:identifier', remove_url_cruft(utf8_encode($effective_url)));\r
	} else {\r
		$newitem->addElement('dc:identifier', remove_url_cruft($item->get_permalink()));\r
	}\r
\r
	// add categories\r
	if ($categories = $item->get_categories()) {\r
		foreach ($categories as $category) {\r
			if ($category->get_label() !== null) {\r
				$newitem->addElement('category', $category->get_label());\r
			}\r
		}\r
	}\r
\r
	// check for enclosures\r
	if ($options->keep_enclosures) {\r
		if ($enclosures = $item->get_enclosures()) {\r
			foreach ($enclosures as $enclosure) {\r
				// thumbnails\r
				foreach ((array)$enclosure->get_thumbnails() as $thumbnail) {\r
					$newitem->addElement('media:thumbnail', '', array('url'=>$thumbnail));\r
				}\r
				if (!$enclosure->get_link()) continue;\r
				$enc = array();\r
				// Media RSS spec ($enc): http://search.yahoo.com/mrss\r
				// SimplePie methods ($enclosure): http://simplepie.org/wiki/reference/start#methods4\r
				$enc['url'] = $enclosure->get_link();\r
				if ($enclosure->get_length()) $enc['fileSize'] = $enclosure->get_length();\r
				if ($enclosure->get_type()) $enc['type'] = $enclosure->get_type();\r
				if ($enclosure->get_medium()) $enc['medium'] = $enclosure->get_medium();\r
				if ($enclosure->get_expression()) $enc['expression'] = $enclosure->get_expression();\r
				if ($enclosure->get_bitrate()) $enc['bitrate'] = $enclosure->get_bitrate();\r
				if ($enclosure->get_framerate()) $enc['framerate'] = $enclosure->get_framerate();\r
				if ($enclosure->get_sampling_rate()) $enc['samplingrate'] = $enclosure->get_sampling_rate();\r
				if ($enclosure->get_channels()) $enc['channels'] = $enclosure->get_channels();\r
				if ($enclosure->get_duration()) $enc['duration'] = $enclosure->get_duration();\r
				if ($enclosure->get_height()) $enc['height'] = $enclosure->get_height();\r
				if ($enclosure->get_width()) $enc['width'] = $enclosure->get_width();\r
				if ($enclosure->get_language()) $enc['lang'] = $enclosure->get_language();\r
				$newitem->addElement('media:content', '', $enc);\r
			}\r
		}\r
	}\r
	$output->addItem($newitem);\r
	unset($html);\r
	$item_count++;\r
}\r
\r
// output feed\r
debug('Done!');\r
/*\r
if ($debug_mode) {\r
	$_apc_data = apc_cache_info('user');\r
	var_dump($_apc_data); exit;\r
}\r
*/\r
if (!$debug_mode) {\r
	if ($callback) echo "$callback("; // if $callback is set, $format also == 'json'\r
	if ($format == 'json') $output->setFormat(($callback === null) ? JSON : JSONP);\r
	$add_to_cache = $options->caching;\r
	// is smart cache mode enabled?\r
	if ($add_to_cache && $options->apc && $options->smart_cache) {\r
		// yes, so only cache if this is the second request for this URL\r
		$add_to_cache = ($apc_cache_hits >= 2);\r
		// purge cache\r
		if ($options->cache_cleanup > 0) {\r
			if (rand(1, $options->cache_cleanup) == 1) {\r
				// apc purge code adapted from from http://www.thimbleopensource.com/tutorials-snippets/php-apc-expunge-script\r
				$_apc_data = apc_cache_info('user');\r
				foreach ($_apc_data['cache_list'] as $_apc_item) {\r
				  if ($_apc_item['ttl'] > 0 && ($_apc_item['ttl'] + $_apc_item['creation_time'] < time())) {\r
					apc_delete($_apc_item['info']);\r
				  }\r
				}\r
			}\r
		}\r
	}\r
	if ($add_to_cache) {\r
		ob_start();\r
		$output->genarateFeed(false);\r
		$output = ob_get_contents();\r
		ob_end_clean();\r
		if ($html_only && $item_count == 0) {\r
			// do not cache - in case of temporary server glitch at source URL\r
		} else {\r
			$cache = get_cache();\r
			if ($add_to_cache) $cache->save($output, $cache_id);\r
		}\r
		echo $output;\r
	} else {\r
		$output->genarateFeed(false);\r
	}\r
	if ($callback) echo ');';\r
}\r
\r