maintainer = "teromene"; $this->name = "Facebook"; $this->uri = "http://www.facebook.com/"; $this->description = "Input a page title or a profile log. For a profile log, please insert the parameter as follow : myExamplePage/132621766841117"; $this->update = "31/03/2016"; $this->parameters[] = '[ { "name" : "Username", "identifier" : "u", "required" : "required" } ]'; } public function collectData(array $param) { //Extract a string using start and end delimiters function ExtractFromDelimiters($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) $link = 'https://www.facebook.com'.$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' => '👍'); $len = count($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']); $http_options = array( 'http' => array( 'method' => 'POST', 'user_agent'=> ini_get('user_agent'), 'header'=>array("Content-type: application/x-www-form-urlencoded\r\nReferer: $captcha_action\r\nCookie: noscript=1\r\n"), 'content' => http_build_query($captcha_fields), ), ); $context = stream_context_create($http_options); $html = file_get_contents($captcha_action, false, $context); if ($html === FALSE) { $this->returnError('Failed to submit captcha response back to Facebook', 500); } unset($_SESSION['captcha_fields']); $html = str_get_html($html); } unset($_SESSION['captcha_fields']); unset($_SESSION['captcha_action']); } //Retrieve page contents if (is_null($html)) { if (isset($param['u'])) { if (!strpos($param['u'], "/")) { $html = $this->file_get_html('https://www.facebook.com/'.urlencode($param['u']).'?_fb_noscript=1') or $this->returnError('No results for this query.', 404); } else { $html = $this->file_get_html('https://www.facebook.com/pages/'.$param['u'].'?_fb_noscript=1') or $this->returnError('No results for this query.', 404); } } else { $this->returnError('You must specify a Facebook username.', 400); } } //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; $_SESSION['captcha_action'] = 'https://www.facebook.com'.$captcha->find('form', 0)->action; //Show captcha filling form to the viewer, proxying the captcha image $img = base64_encode(file_get_contents($captcha->find('img', 0)->src)); header('HTTP/1.1 500 '.Http::getMessageForCode(500)); header('Content-Type: text/html'); die('
'); } //No captcha? We can carry on retrieving page contents :) $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); $profilePic = 'https://graph.facebook.com/'.$param['u'].'/picture?width=200&height=200'; $this->name = $author; foreach($element->children() as $post) { $item = new \Item(); if (count($post->find('abbr')) > 0) { //Retrieve post contents $content = preg_replace('/(?i)>