<?php
class FacebookBridge extends BridgeAbstract {

	const MAINTAINER = 'teromene, logmanoriginal';
	const NAME = 'Facebook';
	const URI = 'https://www.facebook.com/';
	const CACHE_TIMEOUT = 300; // 5min
	const DESCRIPTION = 'Input a page title or a profile log. For a profile log,
 please insert the parameter as follow : myExamplePage/132621766841117';

	const PARAMETERS = array(
		'User' => array(
			'u' => array(
				'name' => 'Username',
				'required' => true
			),
			'media_type' => array(
				'name' => 'Media type',
				'type' => 'list',
				'required' => false,
				'values' => array(
					'All' => 'all',
					'Video' => 'video',
					'No Video' => 'novideo'
				),
				'defaultValue' => 'all'
			),
			'skip_reviews' => array(
				'name' => 'Skip reviews',
				'type' => 'checkbox',
				'required' => false,
				'defaultValue' => false,
				'title' => 'Feed includes reviews when checked'
			)
		),
		'Group' => array(
			'g' => array(
				'name' => 'Group',
				'type' => 'text',
				'required' => true,
				'exampleValue' => 'https://www.facebook.com/groups/743149642484225',
				'title' => 'Insert group name or facebook group URL'
			)
		)
	);

	private $authorName = '';
	private $groupName = '';

	public function getURI() {
		$uri = self::URI;

		switch($this->queriedContext) {

			case 'Group':
				$uri .= 'groups/' . $this->sanitizeGroup(filter_var($this->getInput('g'), FILTER_SANITIZE_URL));
				break;

		}

		return $uri .= '?_fb_noscript=1';
	}

	public function collectData() {

		switch($this->queriedContext) {

			case 'Group':
				$this->collectGroupData();
				break;

			case 'User':
				$this->collectUserData();
				break;

			default:
				returnClientError('Unknown context: "' . $this->queriedContext . '"!');

		}

	}

	#region Group

	private function collectGroupData() {

		$header = array('Accept-Language: ' . getEnv('HTTP_ACCEPT_LANGUAGE') . "\r\n");

		$html = getSimpleHTMLDOM($this->getURI(), $header)
			or returnServerError('Failed loading facebook page: ' . $this->getURI());

		if(!$this->isPublicGroup($html)) {
			returnClientError('This group is not public! RSS-Bridge only supports public groups!');
		}

		defaultLinkTo($html, substr(self::URI, 0, strlen(self::URI) - 1));

		$this->groupName = $this->extractGroupName($html);

		$posts = $html->find('div.userContentWrapper')
			or returnServerError('Failed finding posts!');

		foreach($posts as $post) {

			$item = array();

			$item['uri'] = $this->extractGroupURI($post);
			$item['title'] = $this->extractGroupTitle($post);
			$item['author'] = $this->extractGroupAuthor($post);
			$item['content'] = $this->extractGroupContent($post);
			$item['timestamp'] = $this->extractGroupTimestamp($post);
			$item['enclosures'] = $this->extractGroupEnclosures($post);

			$this->items[] = $item;

		}

	}

	private function sanitizeGroup($group) {

		if(filter_var(
			$group,
			FILTER_VALIDATE_URL,
			FILTER_FLAG_HOST_REQUIRED | FILTER_FLAG_PATH_REQUIRED)) {
			// User provided a URL

			$urlparts = parse_url($group);

			if($urlparts['host'] !== parse_url(self::URI)['host']
			&& 'www.' . $urlparts['host'] !== parse_url(self::URI)['host']) {

				returnClientError('The host you provided is invalid! Received "'
				. $urlparts['host']
				. '", expected "'
				. parse_url(self::URI)['host']
				. '"!');

			}

			return explode('/', $urlparts['path'])[2];

		} elseif(strpos($group, '/') !== false) {
			returnClientError('The group you provided is invalid: ' . $group);
		} else {
			return $group;
		}

	}

	private function isPublicGroup($html) {

		// Facebook redirects to the groups about page for non-public groups
		$about = $html->find('#pagelet_group_about', 0);

		return !($about);

	}

	private function extractGroupName($html) {

		$ogtitle = $html->find('meta[property="og:title"]', 0)
			or returnServerError('Unable to find group title!');

		return htmlspecialchars_decode($ogtitle->content, ENT_QUOTES);

	}

	private function extractGroupURI($post) {

		$elements = $post->find('a')
			or returnServerError('Unable to find URI!');

		foreach($elements as $anchor) {

			// Find the one that is a permalink
			if(strpos($anchor->href, 'permalink') !== false) {
				return $anchor->href;
			}

		}

		return null;

	}

	private function extractGroupContent($post) {

		$content = $post->find('div.userContent', 0)
			or returnServerError('Unable to find user content!');

		return $content->innertext . $content->next_sibling()->innertext;

	}

	private function extractGroupTimestamp($post) {

		$element = $post->find('abbr[data-utime]', 0)
			or returnServerError('Unable to find timestamp!');

		return $element->getAttribute('data-utime');

	}

	private function extractGroupAuthor($post) {

		$element = $post->find('img', 0)
			or returnServerError('Unable to find author information!');

		return $element->{'aria-label'};

	}

	private function extractGroupEnclosures($post) {

		$elements = $post->find('div.userContent', 0)->next_sibling()->find('img');

		$enclosures = array();

		foreach($elements as $enclosure) {
			$enclosures[] = $enclosure->src;
		}

		return empty($enclosures) ? null : $enclosures;

	}

	private function extractGroupTitle($post) {

		$element = $post->find('h5', 0)
			or returnServerError('Unable to find title!');

		if(strpos($element->plaintext, 'shared') === false) {

			$content = strip_tags($this->extractGroupContent($post));

			return $this->extractGroupAuthor($post)
			. ' posted: '
			. substr(
					$content,
					0,
					strpos(wordwrap($content, 64), "\n")
				)
			. '...';

		}

		return $element->plaintext;

	}

	#endregion

	private function collectUserData(){

		//Extract a string using start and end delimiters
		function extractFromDelimiters($string, $start, $end){
			if(strpos($string, $start) !== false) {
				$section_retrieved = substr($string, strpos($string, $start) + strlen($start));
				$section_retrieved = substr($section_retrieved, 0, strpos($section_retrieved, $end));
				return $section_retrieved;
			}

			return false;
		}

		//Utility function for cleaning a Facebook link
		$unescape_fb_link = function($matches){
			if(is_array($matches) && count($matches) > 1) {
				$link = $matches[1];
				if(strpos($link, '/') === 0)
					$link = self::URI . $link;
				if(strpos($link, 'facebook.com/l.php?u=') !== false)
					$link = urldecode(extractFromDelimiters($link, 'facebook.com/l.php?u=', '&'));
				return ' href="' . $link . '"';
			}
		};

		//Utility function for converting facebook emoticons
		$unescape_fb_emote = function($matches){
			static $facebook_emoticons = array(
				'smile' => ':)',
				'frown' => ':(',
				'tongue' => ':P',
				'grin' => ':D',
				'gasp' => ':O',
				'wink' => ';)',
				'pacman' => ':<',
				'grumpy' => '>_<',
				'unsure' => ':/',
				'cry' => ':\'(',
				'kiki' => '^_^',
				'glasses' => '8-)',
				'sunglasses' => 'B-)',
				'heart' => '<3',
				'devil' => ']:D',
				'angel' => '0:)',
				'squint' => '-_-',
				'confused' => 'o_O',
				'upset' => 'xD',
				'colonthree' => ':3',
				'like' => '&#x1F44D;');
			$len = count($matches);
			if ($len > 1)
				for ($i = 1; $i < $len; $i++)
					foreach ($facebook_emoticons as $name => $emote)
						if ($matches[$i] === $name)
							return $emote;
			return $matches[0];
		};

		$html = null;

		//Handle captcha response sent by the viewer
		if (isset($_POST['captcha_response'])) {
			if (session_status() == PHP_SESSION_NONE)
				session_start();
			if (isset($_SESSION['captcha_fields'], $_SESSION['captcha_action'])) {
				$captcha_action = $_SESSION['captcha_action'];
				$captcha_fields = $_SESSION['captcha_fields'];
				$captcha_fields['captcha_response'] = preg_replace('/[^a-zA-Z0-9]+/', '', $_POST['captcha_response']);

				$header = array("Content-type:
application/x-www-form-urlencoded\r\nReferer: $captcha_action\r\nCookie: noscript=1\r\n");
				$opts = array(
					CURLOPT_POST => 1,
					CURLOPT_POSTFIELDS => http_build_query($captcha_fields)
				);

				$html = getContents($captcha_action, $header, $opts);

				if($html === false) {
					returnServerError('Failed to submit captcha response back to Facebook');
				}
				unset($_SESSION['captcha_fields']);
				$html = str_get_html($html);
			}
			unset($_SESSION['captcha_fields']);
			unset($_SESSION['captcha_action']);
		}

		//Retrieve page contents
		if(is_null($html)) {
			$header = array('Accept-Language: ' . getEnv('HTTP_ACCEPT_LANGUAGE') . "\r\n");

			// Check if the user provided a fully qualified URL
			if (filter_var($this->getInput('u'), FILTER_VALIDATE_URL)) {

				$urlparts = parse_url($this->getInput('u'));

				if($urlparts['host'] !== parse_url(self::URI)['host']) {
					returnClientError('The host you provided is invalid! Received "'
					. $urlparts['host']
					. '", expected "'
					. parse_url(self::URI)['host']
					. '"!');
				}

				if(!array_key_exists('path', $urlparts)
				|| $urlparts['path'] === '/') {
					returnClientError('The URL you provided doesn\'t contain the user name!');
				}

				$user = explode('/', $urlparts['path'])[1];

				$html = getSimpleHTMLDOM(self::URI . urlencode($user) . '?_fb_noscript=1', $header)
						or returnServerError('No results for this query.');

			} else {

				// First character cannot be a forward slash
				if(strpos($this->getInput('u'), '/') === 0) {
					returnClientError('Remove leading slash "/" from the username!');
				}

				if(!strpos($this->getInput('u'), '/')) {
					$html = getSimpleHTMLDOM(self::URI . urlencode($this->getInput('u')) . '?_fb_noscript=1', $header)
						or returnServerError('No results for this query.');
				} else {
					$html = getSimpleHTMLDOM(self::URI . 'pages/' . $this->getInput('u') . '?_fb_noscript=1', $header)
						or returnServerError('No results for this query.');
				}

			}
		}

		//Handle captcha form?
		$captcha = $html->find('div.captcha_interstitial', 0);
		if (!is_null($captcha)) {
			//Save form for submitting after getting captcha response
			if (session_status() == PHP_SESSION_NONE)
				session_start();
			$captcha_fields = array();
			foreach ($captcha->find('input, button') as $input)
				$captcha_fields[$input->name] = $input->value;
			$_SESSION['captcha_fields'] = $captcha_fields;
			$_SESSION['captcha_action'] = $captcha->find('form', 0)->action;

			//Show captcha filling form to the viewer, proxying the captcha image
			$img = base64_encode(getContents($captcha->find('img', 0)->src));
			http_response_code(500);
			header('Content-Type: text/html');
			$message = <<<EOD
<form method="post" action="?{$_SERVER['QUERY_STRING']}">
	<h2>Facebook captcha challenge</h2>
	<p>Unfortunately, rss-bridge cannot fetch the requested page.<br />
		Facebook wants rss-bridge to resolve the following captcha:</p>
	<p><img src="data:image/png;base64,{$img}" /></p>
	<p><b>Response:</b> <input name="captcha_response" placeholder="please fill in" />
	<input type="submit" value="Submit!" /></p>
</form>
EOD;
			die($message);
		}

		//No captcha? We can carry on retrieving page contents :)
		//First, we check wether the page is public or not
		$loginForm = $html->find('._585r', 0);
		if($loginForm != null) {
			returnServerError('You must be logged in to view this page. This is not supported by RSS-Bridge.');
		}

		$element = $html
		->find('#pagelet_timeline_main_column')[0]
		->children(0)
		->children(0)
		->children(0)
		->next_sibling()
		->children(0);

		if(isset($element)) {

			$author = str_replace(' | Facebook', '', $html->find('title#pageTitle', 0)->innertext);
			$profilePic = 'https://graph.facebook.com/'
			. $this->getInput('u')
			. '/picture?width=200&amp;height=200';

			$this->authorName = $author;

			foreach($element->children() as $cell) {
				// Manage summary posts
				if(strpos($cell->class, '_3xaf') !== false) {
					$posts = $cell->children();
				} else {
					$posts = array($cell);
				}

				// Optionally skip reviews
				if($this->getInput('skip_reviews')
				&& !is_null($cell->find('#review_composer_container', 0))) {
					continue;
				}

				foreach($posts as $post) {
					// Check media type
					switch($this->getInput('media_type')) {
						case 'all': break;
						case 'video':
							if(empty($post->find('[aria-label=Video]'))) continue 2;
							break;
						case 'novideo':
							if(!empty($post->find('[aria-label=Video]'))) continue 2;
							break;
						default: break;
					}

					$item = array();

					if(count($post->find('abbr')) > 0) {

						//Retrieve post contents
						$content = preg_replace(
							'/(?i)><div class=\"clearfix([^>]+)>(.+?)div\ class=\"userContent\"/i',
							'',
							$post);

						$content = preg_replace(
							'/(?i)><div class=\"_59tj([^>]+)>(.+?)<\/div><\/div><a/i',
							'',
							$content);

						$content = preg_replace(
							'/(?i)><div class=\"_3dp([^>]+)>(.+?)div\ class=\"[^u]+userContent\"/i',
							'',
							$content);

						$content = preg_replace(
							'/(?i)><div class=\"_4l5([^>]+)>(.+?)<\/div>/i',
							'',
							$content);

						//Remove html nodes, keep only img, links, basic formatting
						$content = strip_tags($content, '<a><img><i><u><br><p>');

						//Adapt link hrefs: convert relative links into absolute links and bypass external link redirection
						$content = preg_replace_callback('/ href=\"([^"]+)\"/i', $unescape_fb_link, $content);

						//Clean useless html tag properties and fix link closing tags
						foreach (array(
							'onmouseover',
							'onclick',
							'target',
							'ajaxify',
							'tabindex',
							'class',
							'style',
							'data-[^=]*',
							'aria-[^=]*',
							'role',
							'rel',
							'id') as $property_name)
								$content = preg_replace('/ ' . $property_name . '=\"[^"]*\"/i', '', $content);
						$content = preg_replace('/<\/a [^>]+>/i', '</a>', $content);

						//Convert textual representation of emoticons eg
						//"<i><u>smile emoticon</u></i>" back to ASCII emoticons eg ":)"
						$content = preg_replace_callback(
							'/<i><u>([^ <>]+) ([^<>]+)<\/u><\/i>/i',
							$unescape_fb_emote,
							$content
						);

						//Retrieve date of the post
						$date = $post->find('abbr')[0];
						if(isset($date) && $date->hasAttribute('data-utime')) {
							$date = $date->getAttribute('data-utime');
						} else {
							$date = 0;
						}

						//Build title from username and content
						$title = $author;
						if(strlen($title) > 24)
							$title = substr($title, 0, strpos(wordwrap($title, 24), "\n")) . '...';
						$title = $title . ' | ' . strip_tags($content);
						if(strlen($title) > 64)
							$title = substr($title, 0, strpos(wordwrap($title, 64), "\n")) . '...';

						$uri = self::URI . $post->find('abbr')[0]->parent()->getAttribute('href');

						//Build and add final item
						$item['uri'] = htmlspecialchars_decode($uri);
						$item['content'] = htmlspecialchars_decode($content);
						$item['title'] = $title;
						$item['author'] = $author;
						$item['timestamp'] = $date;
						$this->items[] = $item;
					}
				}
			}
		}
	}

	public function getName(){

		switch($this->queriedContext) {

			case 'User':
				if(!empty($this->authorName)) {
					return isset($this->extraInfos['name']) ? $this->extraInfos['name'] : $this->authorName
					. ' - Facebook Bridge';
				}
				break;

			case 'Group':
				if(!empty($this->groupName)) {
					return $this->groupName . ' - Facebook Bridge';
				}
				break;

		}

		return parent::getName();
	}
}