X-Git-Url: https://git.ao2.it/tweeper.git/blobdiff_plain/01dfd03d97afd0d5d06a36af37d96892905340ad..2c7f7e5da9efb2905d9c87cb5a9a2d6c057f576c:/src/Tweeper.php?ds=sidebyside diff --git a/src/Tweeper.php b/src/Tweeper.php index cfa3812..f79bb08 100644 --- a/src/Tweeper.php +++ b/src/Tweeper.php @@ -6,7 +6,7 @@ namespace Tweeper; * @file * Tweeper - a Twitter to RSS web scraper. * - * Copyright (C) 2013-2018 Antonio Ospite + * Copyright (C) 2013-2020 Antonio Ospite * * This program is free software: you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by @@ -36,7 +36,9 @@ date_default_timezone_set('UTC'); */ class Tweeper { - private static $userAgent = "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:60.0) Gecko/20100101 Firefox/60.0"; + private static $userAgent = "APIs-Google (+https://developers.google.com/webmasters/APIs-Google.html)"; + private static $maxConnectionTimeout = 5; + private static $maxConnectionRetries = 5; /** * Create a new Tweeper object controlling optional settings. @@ -47,10 +49,18 @@ class Tweeper { * Enables showing the username in front of the content for multi-user * sites (enabled by default). Only some stylesheets supports this * functionality (twitter, instagram, pump.io). + * @param bool $show_multimedia + * Enables showing multimedia content (images, videos) directly in the + * item description (enabled by default). Only some stylesheets supports + * this functionality (twitter, instagram, dilbert). + * @param bool $verbose_output + * Enables showing non-fatal errors like XML parsing errors. */ - public function __construct($generate_enclosure = FALSE, $show_usernames = TRUE) { + public function __construct($generate_enclosure = FALSE, $show_usernames = TRUE, $show_multimedia = TRUE, $verbose_output = TRUE) { $this->generate_enclosure = $generate_enclosure; $this->show_usernames = $show_usernames; + $this->show_multimedia = $show_multimedia; + $this->verbose_output = $verbose_output; } /** @@ -86,24 +96,51 @@ class Tweeper { } /** + * Perform a cURL session multiple times when it fails with a timeout. + * + * @param resource $ch + * a cURL session handle. + */ + private static function curlExec($ch) { + $ret = FALSE; + $attempt = 0; + do { + $ret = curl_exec($ch); + if (FALSE === $ret) { + trigger_error(curl_error($ch), E_USER_WARNING); + } + } while (curl_errno($ch) == CURLE_OPERATION_TIMEDOUT && ++$attempt < Tweeper::$maxConnectionRetries); + + $response_code = curl_getinfo($ch, CURLINFO_RESPONSE_CODE); + if (FALSE === $response_code) { + trigger_error(curl_error($ch), E_USER_WARNING); + return FALSE; + } + + if ($response_code >= 400) { + trigger_error("HTTP reponse code $response_code", E_USER_WARNING); + return FALSE; + } + + return $ret; + } + + /** * Get the contents from a URL. */ - private static function getUrlContents($url) { + private static function getUrlContents($url, $user_agent = NULL) { $ch = curl_init($url); - curl_setopt_array($ch, array( + curl_setopt_array($ch, [ CURLOPT_HEADER => FALSE, + CURLOPT_CONNECTTIMEOUT => Tweeper::$maxConnectionTimeout, // Follow http redirects to get the real URL. CURLOPT_FOLLOWLOCATION => TRUE, + CURLOPT_COOKIEFILE => "", CURLOPT_RETURNTRANSFER => TRUE, - CURLOPT_SSL_VERIFYHOST => FALSE, - CURLOPT_SSL_VERIFYPEER => FALSE, - CURLOPT_HTTPHEADER => array('Accept-language: en'), - CURLOPT_USERAGENT => Tweeper::$userAgent, - )); - $contents = curl_exec($ch); - if (FALSE === $contents) { - trigger_error(curl_error($ch)); - } + CURLOPT_HTTPHEADER => ['Accept-language: en'], + CURLOPT_USERAGENT => isset($user_agent) ? $user_agent : Tweeper::$userAgent, + ]); + $contents = Tweeper::curlExec($ch); curl_close($ch); return $contents; @@ -112,22 +149,27 @@ class Tweeper { /** * Get the headers from a URL. */ - private static function getUrlInfo($url) { + private static function getUrlInfo($url, $user_agent = NULL) { $ch = curl_init($url); - curl_setopt_array($ch, array( + curl_setopt_array($ch, [ CURLOPT_HEADER => TRUE, CURLOPT_NOBODY => TRUE, + CURLOPT_CONNECTTIMEOUT => Tweeper::$maxConnectionTimeout, // Follow http redirects to get the real URL. CURLOPT_FOLLOWLOCATION => TRUE, CURLOPT_RETURNTRANSFER => TRUE, - CURLOPT_SSL_VERIFYHOST => FALSE, - CURLOPT_SSL_VERIFYPEER => FALSE, - CURLOPT_USERAGENT => Tweeper::$userAgent, - )); - curl_exec($ch); + CURLOPT_USERAGENT => isset($user_agent) ? $user_agent : Tweeper::$userAgent, + ]); + + $ret = Tweeper::curlExec($ch); + if (FALSE === $ret) { + curl_close($ch); + return FALSE; + } + $url_info = curl_getinfo($ch); if (FALSE === $url_info) { - trigger_error(curl_error($ch)); + trigger_error(curl_error($ch), E_USER_WARNING); } curl_close($ch); @@ -138,7 +180,7 @@ class Tweeper { * Generate an RSS element. */ public static function generateEnclosure($url) { - $supported_content_types = array( + $supported_content_types = [ "application/octet-stream", "application/ogg", "application/pdf", @@ -157,13 +199,17 @@ class Tweeper { "video/mp4", "video/mpeg", "video/ogg", - ); + ]; $url_info = Tweeper::getUrlInfo($url); + if (FALSE === $url_info) { + trigger_error("Failed to retrieve info for URL: " . $url, E_USER_WARNING); + return ''; + } $supported = in_array($url_info['content_type'], $supported_content_types); if (!$supported) { - error_log("Unsupported enclosure content type \"" . $url_info['content_type'] . "\" for URL: " . $url_info['url']); + trigger_error("Unsupported enclosure content type \"" . $url_info['content_type'] . "\" for URL: " . $url_info['url'], E_USER_WARNING); return ''; } @@ -171,10 +217,17 @@ class Tweeper { // See http://sourceforge.net/p/feedvalidator/bugs/72/ $http_url = preg_replace("/^https/", "http", $url_info['url']); + // When the server does not provide a Content-Length header, + // curl_getinfo() would return a negative value for + // "download_content_length", however RSS recommends to use 0 when the + // enclosure's size cannot be determined. + // See: https://www.feedvalidator.org/docs/error/UseZeroForUnknown.html + $length = max($url_info['download_content_length'], 0); + $dom = new DOMDocument(); $enc = $dom->createElement('enclosure'); $enc->setAttribute('url', $http_url); - $enc->setAttribute('length', $url_info['download_content_length']); + $enc->setAttribute('length', $length); $enc->setAttribute('type', $url_info['content_type']); return $enc; @@ -211,7 +264,7 @@ class Tweeper { $output .= " line $error->line"; - error_log($output); + trigger_error($output, E_USER_WARNING); } /** @@ -228,17 +281,17 @@ class Tweeper { $encoder = new XmlEncoder(); $normalizer = new ObjectNormalizer(); - $serializer = new Serializer(array($normalizer), array($encoder)); + $serializer = new Serializer([$normalizer], [$encoder]); - $serializer_options = array( + $serializer_options = [ 'xml_encoding' => "UTF-8", 'xml_format_output' => TRUE, 'xml_root_node_name' => $root_node_name, - ); + ]; $xml_data = $serializer->serialize($data, 'xml', $serializer_options); if (!$xml_data) { - trigger_error("Cannot serialize data", E_USER_ERROR); + trigger_error("Cannot serialize data", E_USER_WARNING); return NULL; } @@ -253,19 +306,25 @@ class Tweeper { $json_match_expr = '/window._sharedData = (.*);/'; $ret = preg_match($json_match_expr, $html, $matches); if ($ret !== 1) { - trigger_error("Cannot match expression: $json_match_expr\n", E_USER_ERROR); + trigger_error("Cannot match expression: $json_match_expr\n", E_USER_WARNING); return NULL; } $data = json_decode($matches[1], $assoc = TRUE); - // The "qe" object contains elements which will result in invalid XML - // element names, so remove it. + // Remove items that may contain elements which can result in invalid XML + // element names (for example names starting with a number). unset($data["qe"]); - - // The "knobs" object contains elements with undefined namespaces, so - // remove it to silence an error message. - unset($data["knobs"]); + unset($data["knobx"]); + unset($data["to_cache"]); + + // Stop here in case Instagram redirected to the login page, this can + // happen when too many consecutive requests have been made from the same + // IP. + if (array_key_exists("LoginAndSignupPage", $data["entry_data"])) { + trigger_error("Cannot open instagram page: redirected to Login page.\n", E_USER_WARNING); + return NULL; + } $json = json_encode($data); @@ -293,15 +352,20 @@ class Tweeper { // If there is a host-specific method to get the XML data, use it! $get_xml_host_method = 'getXml' . Tweeper::toUpperCamelCase($host, '.'); if (method_exists($this, $get_xml_host_method)) { - $xml_data = call_user_func_array(array($this, $get_xml_host_method), array($html)); + $xml_data = call_user_func_array([$this, $get_xml_host_method], [$html]); + if (NULL === $xml_data) { + return NULL; + } $xmlDoc->loadXML($xml_data); } else { $xmlDoc->loadHTML($html); } - foreach (libxml_get_errors() as $xml_error) { - Tweeper::logXmlError($xml_error); + if ($this->verbose_output) { + foreach (libxml_get_errors() as $xml_error) { + Tweeper::logXmlError($xml_error); + } } libxml_clear_errors(); libxml_use_internal_errors($xml_errors_value); @@ -315,11 +379,15 @@ class Tweeper { private function loadStylesheet($host) { $stylesheet = "file://" . __DIR__ . "/rss_converter_" . $host . ".xsl"; if (FALSE === file_exists($stylesheet)) { - trigger_error("Conversion to RSS not supported for $host ($stylesheet not found)", E_USER_ERROR); + trigger_error("Conversion to RSS not supported for $host ($stylesheet not found)", E_USER_WARNING); return NULL; } - $stylesheet_contents = Tweeper::getUrlContents($stylesheet); + $stylesheet_contents = file_get_contents($stylesheet); + if (FALSE === $stylesheet_contents) { + trigger_error("Cannot open $stylesheet", E_USER_WARNING); + return NULL; + } $xslDoc = new DOMDocument(); $xslDoc->loadXML($stylesheet_contents); @@ -328,6 +396,7 @@ class Tweeper { $xsltProcessor->registerPHPFunctions(); $xsltProcessor->setParameter('', 'generate-enclosure', $this->generate_enclosure); $xsltProcessor->setParameter('', 'show-usernames', $this->show_usernames); + $xsltProcessor->setParameter('', 'show-multimedia', $this->show_multimedia); $xsltProcessor->importStylesheet($xslDoc); return $xsltProcessor; @@ -336,25 +405,25 @@ class Tweeper { /** * Convert the site content to RSS. */ - public function tweep($src_url, $host=NULL, $validate_scheme=TRUE) { + public function tweep($src_url, $host = NULL, $validate_scheme = TRUE) { $url = parse_url($src_url); if (FALSE === $url) { - trigger_error("Invalid URL: $src_url", E_USER_ERROR); + trigger_error("Invalid URL: $src_url", E_USER_WARNING); return NULL; } if (TRUE === $validate_scheme) { $scheme = $url["scheme"]; - if (!in_array($scheme, array("http", "https"))) { - trigger_error("unsupported scheme: $scheme", E_USER_ERROR); + if (!in_array($scheme, ["http", "https"])) { + trigger_error("unsupported scheme: $scheme", E_USER_WARNING); return NULL; } } - // if the host is not given derive it from the URL + // If the host is not given derive it from the URL. if (NULL === $host) { if (empty($url["host"])) { - trigger_error("Invalid host in URL: $src_url", E_USER_ERROR); + trigger_error("Invalid host in URL: $src_url", E_USER_WARNING); return NULL; } // Strip the leading www. to be more forgiving on input URLs. @@ -368,12 +437,13 @@ class Tweeper { $html = Tweeper::getUrlContents($src_url); if (FALSE === $html) { + trigger_error("Failed to retrieve $src_url", E_USER_WARNING); return NULL; } $preprocess_html_host_method = 'preprocessHtml' . Tweeper::toUpperCamelCase($host, '.'); if (method_exists($this, $preprocess_html_host_method)) { - $html = call_user_func_array(array($this, $preprocess_html_host_method), array($html)); + $html = call_user_func_array([$this, $preprocess_html_host_method], [$html]); } $xmlDoc = $this->htmlToXml($html, $host); @@ -383,7 +453,7 @@ class Tweeper { $output = $xsltProcessor->transformToXML($xmlDoc); if (FALSE === $output) { - trigger_error('XSL transformation failed.', E_USER_ERROR); + trigger_error('XSL transformation failed.', E_USER_WARNING); return NULL; }