X-Git-Url: https://git.ao2.it/tweeper.git/blobdiff_plain/69e71256e0c4a30c57fe13f0ff5b3038e2c5f563..5d5f682cf2bd84c56190c8494bfeaf040b5585e8:/tweeper.php diff --git a/tweeper.php b/tweeper.php index 178176d..bed6350 100644 --- a/tweeper.php +++ b/tweeper.php @@ -18,7 +18,11 @@ * along with this program. If not, see . */ -require_once 'XML/Serializer.php'; +require_once 'Symfony/Component/Serializer/autoload.php'; + +use Symfony\Component\Serializer\Serializer; +use Symfony\Component\Serializer\Encoder\XmlEncoder; +use Symfony\Component\Serializer\Normalizer\ObjectNormalizer; date_default_timezone_set('UTC'); @@ -186,33 +190,43 @@ class Tweeper { return NULL; } - $data = json_decode($matches[1]); + // Apparenty the ObjectNormalizer used afterwards is not able to handle + // the stdClass object created by json_decode() with the default setting + // $assoc = false; so use $assoc = true + $data = json_decode($matches[1], $assoc = true); if (!$data) { return NULL; } + $encoder = new XmlEncoder(); + $normalizer = new ObjectNormalizer(); + $serializer = new Serializer(array($normalizer), array($encoder)); + $serializer_options = array ( - 'addDecl' => TRUE, - 'encoding' => "UTF-8", - 'indent' => ' ', - 'rootName' => $rootName, + 'xml_encoding' => "UTF-8", + 'xml_format_output' => TRUE, + 'xml_root_node_name' => $rootName, ); - $serializer = new XML_Serializer($serializer_options); - - $status = $serializer->serialize($data); - if (PEAR::isError($status)) { - trigger_error($status->getMessage(), E_USER_ERROR); + $xml_data = $serializer->serialize($data, 'xml', $serializer_options); + if (!$xml_data) { + trigger_error("Cannot serialize data", E_USER_ERROR); return NULL; } - return $serializer->getSerializedData(); + return $xml_data; } private function get_xml_instagram_com($html) { return $this->json_to_xml($html, '/window._sharedData = (.*);/', 'instagram'); } + private function preprocess_html_facebook_com($html) { + $html = str_replace('', '', $html); + return $html; + } + private function html_to_xml($html, $host) { $xmlDoc = new DOMDocument(); @@ -244,7 +258,10 @@ class Tweeper { return NULL; } - $xsltProcessor = $this->load_stylesheet($url["host"]); + // Strip the leading www. to be more forgiving on input URLs + $host = preg_replace('/^www\./', '', $url["host"]); + + $xsltProcessor = $this->load_stylesheet($host); if (NULL === $xsltProcessor) { return NULL; } @@ -254,7 +271,12 @@ class Tweeper { return NULL; } - $xmlDoc = $this->html_to_xml($html, $url["host"]); + $preprocess_html_host_method = 'preprocess_html_' . str_replace(".", "_", $host); + if (method_exists($this, $preprocess_html_host_method)) { + $html = call_user_func_array(array($this, $preprocess_html_host_method), array($html)); + } + + $xmlDoc = $this->html_to_xml($html, $host); if (NULL === $xmlDoc) { return NULL; } @@ -271,7 +293,7 @@ class Tweeper { function is_cli() { - return (substr(php_sapi_name(), 0, 3) == 'cli'); + return (php_sapi_name() === "cli"); } function usage($argv) @@ -340,7 +362,7 @@ if (is_cli()) { } if (!isset($options['src_url'])) { - fwrite($ERROR_STREAM, usage($argv)); + fwrite($ERROR_STREAM, usage(is_cli() ? $argv : NULL)); exit(1); }