. */ // Usage // ----- // Request this file passing it your feed in the querystring: makefulltextfeed.php?url=mysite.org // The following options can be passed in the querystring: // * URL: url=[feed or website url] (required, should be URL-encoded - in php: urlencode($url)) // * URL points to HTML (not feed): html=true (optional, by default it's automatically detected) // * API key: key=[api key] (optional, refer to config.php) // * Max entries to process: max=[max number of items] (optional) error_reporting(E_ALL ^ E_NOTICE); ini_set("display_errors", 1); @set_time_limit(120); // set include path set_include_path(realpath(dirname(__FILE__).'/libraries').PATH_SEPARATOR.get_include_path()); // Autoloading of classes allows us to include files only when they're // needed. If we've got a cached copy, for example, only Zend_Cache is loaded. function __autoload($class_name) { static $mapping = array( // Include SimplePie for RSS/Atom parsing 'SimplePie' => 'simplepie/simplepie.class.php', 'SimplePie_Misc' => 'simplepie/simplepie.class.php', // Include FeedCreator for RSS/Atom creation 'FeedWriter' => 'feedwriter/FeedWriter.php', 'FeedItem' => 'feedwriter/FeedItem.php', // Include Readability for identifying and extracting content from URLs 'Readability' => 'readability/Readability.php', // Include Humble HTTP Agent to allow parallel requests and response caching 'HumbleHttpAgent' => 'humble-http-agent/HumbleHttpAgent.php', // Include IRI class for resolving relative URLs 'IRI' => 'iri/iri.php', // Include Zend Cache to improve performance (cache results) 'Zend_Cache' => 'Zend/Cache.php' ); if (isset($mapping[$class_name])) { //echo "Loading $class_name\n
"; require_once $mapping[$class_name]; return true; } else { return false; } } //////////////////////////////// // Load config file if it exists //////////////////////////////// // the config values below should be set in config.php (rename config-sample.php if config.php doesn't exist). // the values below will only be used if config.php doesn't exist. $options->enabled = true; $options->restrict = false; $options->default_entries = 5; $options->max_entries = 10; $options->rewrite_relative_urls = true; $options->caching = false; $options->cache_dir = dirname(__FILE__).'/cache'; $options->message_to_prepend = ''; $options->message_to_append = ''; $options->blocked_urls = array(); $options->alternative_url = ''; $options->error_message = '[unable to retrieve full-text content]'; $options->api_keys = array(); $options->default_entries_with_key = 5; $options->max_entries_with_key = 10; $options->message_to_prepend_with_key = ''; $options->message_to_append_with_key = ''; $options->error_message_with_key = '[unable to retrieve full-text content]'; $options->cache_directory_level = 0; if (file_exists(dirname(__FILE__).'/config.php')) { require_once(dirname(__FILE__).'/config.php'); } ////////////////////////////////////////////// // Convert $html to UTF8 // (uses HTTP headers and HTML to find encoding) // adapted from http://stackoverflow.com/questions/910793/php-detect-encoding-and-make-everything-utf-8 ////////////////////////////////////////////// function convert_to_utf8($html, $header=null) { $accept = array( 'type' => array('application/rss+xml', 'application/xml', 'application/rdf+xml', 'text/xml', 'text/html'), 'charset' => array_diff(mb_list_encodings(), array('pass', 'auto', 'wchar', 'byte2be', 'byte2le', 'byte4be', 'byte4le', 'BASE64', 'UUENCODE', 'HTML-ENTITIES', 'Quoted-Printable', '7bit', '8bit')) ); $encoding = null; if ($html || $header) { if (is_array($header)) $header = implode("\n", $header); if (!$header || !preg_match_all('/^Content-Type:\s+([^;]+)(?:;\s*charset=["\']?([^;"\'\n]*))?/im', $header, $match, PREG_SET_ORDER)) { // error parsing the response } else { $match = end($match); // get last matched element (in case of redirects) if (!in_array(strtolower($match[1]), $accept['type'])) { // type not accepted // TODO: avoid conversion } if (isset($match[2])) $encoding = trim($match[2], '"\''); } if (!$encoding) { if (preg_match('/^<\?xml\s+version=(?:"[^"]*"|\'[^\']*\')\s+encoding=("[^"]*"|\'[^\']*\')/s', $html, $match)) { $encoding = trim($match[1], '"\''); } elseif(preg_match('/]+)/i', $html, $match)) { if (isset($match[1])) $encoding = trim($match[1]); } } if (!$encoding) { $encoding = 'utf-8'; } else { if (!in_array($encoding, array_map('strtolower', $accept['charset']))) { // encoding not accepted // TODO: avoid conversion } if (strtolower($encoding) != 'utf-8') { if (strtolower($encoding) == 'iso-8859-1') { // replace MS Word smart qutoes $trans = array(); $trans[chr(130)] = '‚'; // Single Low-9 Quotation Mark $trans[chr(131)] = 'ƒ'; // Latin Small Letter F With Hook $trans[chr(132)] = '„'; // Double Low-9 Quotation Mark $trans[chr(133)] = '…'; // Horizontal Ellipsis $trans[chr(134)] = '†'; // Dagger $trans[chr(135)] = '‡'; // Double Dagger $trans[chr(136)] = 'ˆ'; // Modifier Letter Circumflex Accent $trans[chr(137)] = '‰'; // Per Mille Sign $trans[chr(138)] = 'Š'; // Latin Capital Letter S With Caron $trans[chr(139)] = '‹'; // Single Left-Pointing Angle Quotation Mark $trans[chr(140)] = 'Œ'; // Latin Capital Ligature OE $trans[chr(145)] = '‘'; // Left Single Quotation Mark $trans[chr(146)] = '’'; // Right Single Quotation Mark $trans[chr(147)] = '“'; // Left Double Quotation Mark $trans[chr(148)] = '”'; // Right Double Quotation Mark $trans[chr(149)] = '•'; // Bullet $trans[chr(150)] = '–'; // En Dash $trans[chr(151)] = '—'; // Em Dash $trans[chr(152)] = '˜'; // Small Tilde $trans[chr(153)] = '™'; // Trade Mark Sign $trans[chr(154)] = 'š'; // Latin Small Letter S With Caron $trans[chr(155)] = '›'; // Single Right-Pointing Angle Quotation Mark $trans[chr(156)] = 'œ'; // Latin Small Ligature OE $trans[chr(159)] = 'Ÿ'; // Latin Capital Letter Y With Diaeresis $html = strtr($html, $trans); } $html = SimplePie_Misc::change_encoding($html, $encoding, 'utf-8'); /* if (function_exists('iconv')) { // iconv appears to handle certain character encodings better than mb_convert_encoding $html = iconv($encoding, 'utf-8', $html); } else { $html = mb_convert_encoding($html, 'utf-8', $encoding); } */ } } } return $html; } function makeAbsolute($base, $elem) { $base = new IRI($base); foreach(array('a'=>'href', 'img'=>'src') as $tag => $attr) { $elems = $elem->getElementsByTagName($tag); for ($i = $elems->length-1; $i >= 0; $i--) { $e = $elems->item($i); //$e->parentNode->replaceChild($articleContent->ownerDocument->createTextNode($e->textContent), $e); if ($e->hasAttribute($attr)) { // Trim leading and trailing white space. I don't really like this but // unfortunately it does appear on some sites. e.g. $url = trim(str_replace('%20', ' ', $e->getAttribute($attr))); $url = str_replace(' ', '%20', $url); if (!preg_match('!https?://!i', $url)) { $absolute = IRI::absolutize($base, $url); if ($absolute) { $e->setAttribute($attr, $absolute); } } } } } } //////////////////////////////// // Check if service is enabled //////////////////////////////// if (!$options->enabled) { die('The full-text RSS service is currently disabled'); } //////////////////////////////// // Check for feed URL //////////////////////////////// if (!isset($_GET['url'])) { die('No URL supplied'); } $url = $_GET['url']; if (!preg_match('!^https?://.+!i', $url)) { $url = 'http://'.$url; } $valid_url = filter_var($url, FILTER_VALIDATE_URL); if ($valid_url !== false && $valid_url !== null && preg_match('!^https?://!', $valid_url)) { $url = filter_var($url, FILTER_SANITIZE_URL); } else { die('Invalid URL supplied'); } //////////////////////////////// // Redirect to alternative URL? //////////////////////////////// if ($options->alternative_url != '' && !isset($_GET['redir']) && mt_rand(0, 100) > 50) { $redirect = $options->alternative_url.'?redir=true&url='.urlencode($url); if (isset($_GET['html'])) $redirect .= '&html='.urlencode($_GET['html']); if (isset($_GET['key'])) $redirect .= '&key='.urlencode($_GET['key']); if (isset($_GET['max'])) $redirect .= '&max='.(int)$_GET['max']; if (isset($_GET['links'])) $redirect .= '&links='.$_GET['links']; header("Location: $redirect"); exit; } /////////////////////////////////////////////// // Check if the request is explicitly for an HTML page /////////////////////////////////////////////// $html_only = (isset($_GET['html']) && $_GET['html'] == 'true'); /////////////////////////////////////////////// // Check if valid key supplied /////////////////////////////////////////////// $valid_key = (isset($_GET['key']) && in_array($_GET['key'], $options->api_keys)); /////////////////////////////////////////////// // Check URL against list of blacklisted URLs // TODO: set up better system for this /////////////////////////////////////////////// foreach ($options->blocked_urls as $blockurl) { if (strstr($url, $blockurl) !== false) { die('URL blocked'); } } /////////////////////////////////////////////// // Max entries // see config.php to find these values /////////////////////////////////////////////// if (isset($_GET['max'])) { $max = (int)$_GET['max']; if ($valid_key) { $max = min($max, $options->max_entries_with_key); } else { $max = min($max, $options->max_entries); } } else { if ($valid_key) { $max = $options->default_entries_with_key; } else { $max = $options->default_entries; } } /////////////////////////////////////////////// // Link handling /////////////////////////////////////////////// if (($valid_key || !$options->restrict) && isset($_GET['links']) && in_array($_GET['links'], array('preserve', 'footnotes', 'remove'))) { $links = $_GET['links']; } else { $links = 'preserve'; } ///////////////////////////////////// // Check for valid format // (stick to RSS for the time being) ///////////////////////////////////// $format = 'rss'; ////////////////////////////////// // Check for cached copy ////////////////////////////////// if ($options->caching) { $frontendOptions = array( 'lifetime' => ($valid_key || !$options->restrict) ? 10*60 : 20*60, // cache lifetime of 10 or 20 minutes 'automatic_serialization' => false, 'write_control' => false, 'automatic_cleaning_factor' => 100, 'ignore_user_abort' => false ); $backendOptions = array( 'cache_dir' => ($valid_key) ? $options->cache_dir.'/rss-with-key/' : $options->cache_dir.'/rss/', // directory where to put the cache files 'file_locking' => false, 'read_control' => true, 'read_control_type' => 'strlen', 'hashed_directory_level' => $options->cache_directory_level, 'hashed_directory_umask' => 0777, 'cache_file_umask' => 0664, 'file_name_prefix' => 'ff' ); // getting a Zend_Cache_Core object $cache = Zend_Cache::factory('Core', 'File', $frontendOptions, $backendOptions); $cache_id = md5($max.$url.$valid_key.$links); if ($data = $cache->load($cache_id)) { header("Content-type: text/xml; charset=UTF-8"); if (headers_sent()) die('Some data has already been output, can\'t send RSS file'); echo $data; exit; } } ////////////////////////////////// // Set Expires header ////////////////////////////////// if ($valid_key) { header('Expires: ' . gmdate('D, d M Y H:i:s', time()+(60*10)) . ' GMT'); } else { header('Expires: ' . gmdate('D, d M Y H:i:s', time()+(60*20)) . ' GMT'); } ////////////////////////////////// // Set up HTTP agent ////////////////////////////////// $http = new HumbleHttpAgent(); if ($options->caching) { $frontendOptions = array( 'lifetime' => 30*60, // cache lifetime of 30 minutes 'automatic_serialization' => true, 'write_control' => false, 'automatic_cleaning_factor' => 100, 'ignore_user_abort' => false ); $backendOptions = array( 'cache_dir' => $options->cache_dir.'/http-responses/', // directory where to put the cache files 'file_locking' => false, 'read_control' => true, 'read_control_type' => 'strlen', 'hashed_directory_level' => $options->cache_directory_level, 'hashed_directory_umask' => 0777, 'cache_file_umask' => 0664, 'file_name_prefix' => 'ff' ); $httpCache = Zend_Cache::factory('Core', 'File', $frontendOptions, $backendOptions); $http->useCache($httpCache); } //////////////////////////////// // Tidy config //////////////////////////////// if (function_exists('tidy_parse_string')) { $tidy_config = array( 'clean' => true, 'output-xhtml' => true, 'logical-emphasis' => true, 'show-body-only' => false, 'wrap' => 0, 'drop-empty-paras' => true, 'drop-proprietary-attributes' => false, 'enclose-text' => true, 'enclose-block-text' => true, 'merge-divs' => true, 'merge-spans' => true, 'char-encoding' => 'utf8', 'hide-comments' => true ); } //////////////////////////////// // Get RSS/Atom feed //////////////////////////////// if (!$html_only) { $feed = new SimplePie(); $feed->set_feed_url($url); $feed->set_autodiscovery_level(SIMPLEPIE_LOCATOR_NONE); $feed->set_timeout(20); $feed->enable_cache(false); $feed->set_stupidly_fast(true); $feed->enable_order_by_date(false); // we don't want to do anything to the feed $feed->set_url_replacements(array()); // initialise the feed // the @ suppresses notices which on some servers causes a 500 internal server error $result = @$feed->init(); //$feed->handle_content_type(); //$feed->get_title(); if ($result && (!is_array($feed->data) || count($feed->data) == 0)) { die('Sorry, no feed items found'); } } //////////////////////////////////////////////////////////////////////////////// // Extract content from HTML (if URL is not feed or explicit HTML request has been made) //////////////////////////////////////////////////////////////////////////////// if ($html_only || !$result) { unset($feed, $result); if ($response = $http->get($url)) { $effective_url = $response['effective_url']; $html = $response['body']; $html = convert_to_utf8($html, $response['headers']); } else { die('Error retrieving '.$url); } // Run through Tidy (if it exists). // This fixes problems with some sites which would otherwise // trouble DOMDocument's HTML parsing. if (function_exists('tidy_parse_string')) { $tidy = tidy_parse_string($html, $tidy_config, 'UTF8'); if (tidy_clean_repair($tidy)) { $html = $tidy->value; } } $readability = new Readability($html, $effective_url); if ($links == 'footnotes') $readability->convertLinksToFootnotes = true; $readability->init(); $readability->clean($readability->getContent(), 'select'); if ($options->rewrite_relative_urls) makeAbsolute($effective_url, $readability->getContent()); $title = $readability->getTitle()->textContent; $content = $readability->getContent()->innerHTML; if ($links == 'remove') { $content = preg_replace('!]*>!', '', $content); } if (!$valid_key) { $content = $options->message_to_prepend.$content; $content .= $options->message_to_append; } else { $content = $options->message_to_prepend_with_key.$content; $content .= $options->message_to_append_with_key; } unset($readability, $html); $output = new FeedWriter(); //ATOM an option $output->setTitle($title); $output->setDescription("Content extracted from $url"); if ($format == 'atom') { $output->setChannelElement('updated', date(DATE_ATOM)); $output->setChannelElement('author', array('name'=>'Five Filters', 'uri'=>'http://fivefilters.org')); } $output->setLink($url); $newitem = $output->createNewItem(); $newitem->setTitle($title); $newitem->setLink($url); if ($format == 'atom') { $newitem->setDate(time()); $newitem->addElement('content', $content); } else { $newitem->setDescription($content); } $output->addItem($newitem); $output->genarateFeed(); exit; } //////////////////////////////////////////// // Create full-text feed //////////////////////////////////////////// $output = new FeedWriter(); $output->setTitle($feed->get_title()); $output->setDescription($feed->get_description()); $output->setLink($feed->get_link()); if ($img_url = $feed->get_image_url()) { $output->setImage($feed->get_title(), $feed->get_link(), $img_url); } if ($format == 'atom') { $output->setChannelElement('updated', date(DATE_ATOM)); $output->setChannelElement('author', array('name'=>'Five Filters', 'uri'=>'http://fivefilters.org')); } //////////////////////////////////////////// // Loop through feed items //////////////////////////////////////////// $items = $feed->get_items(0, $max); // Request all feed items in parallel (if supported) $urls_sanitized = array(); $urls = array(); foreach ($items as $key => $item) { $permalink = htmlspecialchars_decode($item->get_permalink()); $permalink = $http->validateUrl($permalink); if ($permalink) { $urls_sanitized[] = $permalink; } $urls[$key] = $permalink; } $http->fetchAll($urls_sanitized); $http->cacheAll(); foreach ($items as $key => $item) { $permalink = $urls[$key]; $newitem = $output->createNewItem(); $newitem->setTitle(htmlspecialchars_decode($item->get_title())); if ($permalink !== false) { $newitem->setLink($permalink); } else { $newitem->setLink($item->get_permalink()); } if ($permalink && $response = $http->get($permalink)) { $effective_url = $response['effective_url']; $html = $response['body']; $html = convert_to_utf8($html, $response['headers']); // Run through Tidy (if it exists). // This fixes problems with some sites which would otherwise // trouble DOMDocument's HTML parsing. (Although sometimes it fails // to return anything, so it's a bit of tradeoff.) if (function_exists('tidy_parse_string')) { $tidy = tidy_parse_string($html, $tidy_config, 'UTF8'); $tidy->cleanRepair(); $html = $tidy->value; } $readability = new Readability($html, $effective_url); if ($links == 'footnotes') $readability->convertLinksToFootnotes = true; $readability->init(); $readability->clean($readability->getContent(), 'select'); if ($options->rewrite_relative_urls) makeAbsolute($effective_url, $readability->getContent()); $html = $readability->getContent()->innerHTML; if ($links == 'remove') { $html = preg_replace('!]*>!', '', $html); } if (!$valid_key) { $html = $options->message_to_prepend.$html; $html .= $options->message_to_append; } else { $html = $options->message_to_prepend_with_key.$html; $html .= $options->message_to_append_with_key; } } else { if (!$valid_key) { $html = $options->error_message; } else { $html = $options->error_message_with_key; } $html .= $item->get_description(); } if ($format == 'atom') { $newitem->addElement('content', $html); $newitem->setDate((int)$item->get_date('U')); if ($author = $item->get_author()) { $newitem->addElement('author', array('name'=>$author->get_name())); } } else { $newitem->addElement('guid', $item->get_permalink(), array('isPermaLink'=>'true')); $newitem->setDescription($html); if ((int)$item->get_date('U') > 0) { $newitem->setDate((int)$item->get_date('U')); } if ($author = $item->get_author()) { $newitem->addElement('dc:creator', $author->get_name()); } } $output->addItem($newitem); unset($html); } // output feed if ($options->caching) { ob_start(); $output->genarateFeed(); $output = ob_get_contents(); ob_end_clean(); $cache->save($output, $cache_id); echo $output; } else { $output->genarateFeed(); } ?>