tweeper.php: make json_to_xml() do strictly what its name says

[tweeper.git] / tweeper.php

diff --git a/tweeper.php b/tweeper.php
index 81f5f85..0034a0d 100644 (file)
--- a/tweeper.php
+++ b/tweeper.php
@@ -18,7 +18,11 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-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');
 
@@ -161,7 +165,7 @@ class Tweeper {
   private function load_stylesheet($host) {
     $stylesheet = "file://" . __DIR__ . "/rss_converter_" . $host . ".xsl";
     if (FALSE === file_exists($stylesheet)) {
-      trigger_error("Conversion to RSS not supported: $host", E_USER_ERROR);
+      trigger_error("Conversion to RSS not supported for $host ($stylesheet not found)", E_USER_ERROR);
       return NULL;
     }
 
@@ -178,39 +182,50 @@ class Tweeper {
     return $xsltProcessor;
   }
 
-  private function json_to_xml($html, $json_match_expr, $rootName) {
-    // pre-process, convert json to XML
-    $ret = preg_match($json_match_expr, $html, $matches);
-    if ($ret !== 1) {
-      trigger_error("Cannot match expression: $json_match_expr\n", E_USER_ERROR);
-      return NULL;
-    }
-
-    $data = json_decode($matches[1]);
+  private function json_to_xml($json, $rootName) {
+    // 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 
+    $data = json_decode($json, $assoc = true);
     if (!$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');
+    // extract the json data from the html code
+    $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);
+      return NULL;
+    }
+
+    return $this->json_to_xml($matches[1], 'instagram');
+  }
+
+  private function preprocess_html_facebook_com($html) {
+    $html = str_replace('<!--', '', $html);
+    $html = str_replace('-->', '', $html);
+    return $html;
   }
 
   private function html_to_xml($html, $host) {
@@ -244,7 +259,10 @@ 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 +272,12 @@ 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 +294,7 @@ class Tweeper {
 
 function is_cli()
 {
-  return (substr(php_sapi_name(), 0, 3) == 'cli');
+  return (php_sapi_name() === "cli");
 }
 
 function usage($argv)
@@ -340,7 +363,7 @@ if (is_cli()) {
 }
 
 if (!isset($options['src_url'])) {
-  fwrite($ERROR_STREAM, usage($argv));
+  fwrite($ERROR_STREAM, usage(is_cli() ? $argv : NULL));
   exit(1);
 }