Updated the new version for OAI

pgogy · Nov 23, 2012 · 4e02c82 · 4e02c82
1 parent 9c54773
commit 4e02c82
Show file tree

Hide file tree

Showing 9 changed files with 535 additions and 16 deletions.
diff --git a/index.php b/index.php
@@ -217,6 +217,7 @@ function getfeed(){
 		?>			
 		<div class="form_holder">
 			<p style="padding-top:10px">Ramanathan is an RSS parser which reads RSS feeds and then sets up submission of items from these feeds into a learning registry node</p>
+			<p>To submit an entire feed please see the <a href="submit_all.php">submit all page</a> or <a href="submit_all_oai.php">submit all page (OAI:DC)</a></p>
 			<p>Enter an RSS Feed you wish to examine - (If the feed contains "dc:rights" or "dc:subject" fields then these will be sent to the registry node)</p>
 			<form action = "" method ="POST" onsubmit="javascript:getfeed()">
 				<label>Enter RSS Feed Url</label><input id="url" type="text" size="150" />		

diff --git a/intro_ga.txt b/intro_ga.txt
@@ -2,17 +2,3 @@
 		<head>
 
 			<link href="style.css" rel="stylesheet" type="text/css" media="all" />
-
-			<script type="text/javascript">
-
-				 var _gaq = _gaq || [];
-				 _gaq.push(['_setAccount', 'UA-24466610-1']);
-				 _gaq.push(['_trackPageview']);
-
-				(function() {
-					var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
-					ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
-					var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
-				 })();
-
-			</script>
diff --git a/paradata_submit_url.php b/paradata_submit_url.php
@@ -122,7 +122,7 @@
 	curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1); 
 	curl_setopt($ch, CURLOPT_URL, "http://alpha.mimas.ac.uk/publish");
 	curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 50); 
-	curl_setopt($ch, CURLOPT_USERPWD, "");
+	curl_setopt($ch, CURLOPT_USERPWD, "fred:flintstone");
 	curl_setopt($ch, CURLOPT_TIMEOUT, 50); 
 	curl_setopt($ch, CURLOPT_MAXREDIRS, 10); 
 	curl_setopt($ch, CURLOPT_HTTPHEADER, array('Content-type: application/json')); 

diff --git a/style.css b/style.css
@@ -284,4 +284,8 @@ div#links a{
 
 div#links a:hover{
 	text-decoration:underline;
+}
+
+a img{
+	border:none;
 }
diff --git a/submit_all.php b/submit_all.php
@@ -0,0 +1,24 @@
+<?PHP
+
+	echo file_get_contents("intro_ga.txt");
+
+?>
+		</head>
+		<?PHP
+
+			echo file_get_contents("post_title.txt");
+
+		?>			
+		<div class="form_holder">
+			<p style="padding-top:10px">Ramanathan is an RSS parser which reads RSS feeds and then sets up submission of items from these feeds into a learning registry node</p>
+			<p>This page submits an entire feed - for inidividual items please see <a href="index.php">individual rss item submission</a></p>
+			<p>Enter an RSS Feed you wish to submit (If the feed contains "dc:rights" or "dc:subject" fields then these will be sent to the registry node)</p>
+			<form action ="submit_all_feed.php" target="_blank" method ="POST" onsubmit="javascript:getfeed()">
+				<label>Enter RSS Feed Url</label><input id="url" name="url" type="text" size="150" />		
+				<input type="submit" value="submit" />
+			</form>
+		</div>
+		<div id="links">
+		</div>
+	</body>
+</html>
diff --git a/submit_all_feed.php b/submit_all_feed.php
@@ -0,0 +1,235 @@
+<?PHP
+
+	echo file_get_contents("intro_ga.txt");
+
+?>
+		</head>
+		<?PHP
+
+			echo file_get_contents("post_title.txt");
+
+		?>			
+		<div>
+		<?PHP
+
+			$ch = curl_init(); 
+			curl_setopt($ch, CURLOPT_HEADER, 0); 
+			curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1); 
+			curl_setopt($ch, CURLOPT_URL, $_POST['url']); 
+			curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 50); 
+			curl_setopt($ch, CURLOPT_TIMEOUT, 50); 
+			curl_setopt($ch, CURLOPT_MAXREDIRS, 10);
+
+			$response = curl_exec($ch);
+
+			$data = @simplexml_load_string($response);
+
+			if($data){
+
+				if(isset($data->channel->item)){
+
+					$return = new StdClass();
+
+				 	foreach ($data->channel->item as $item) {  
+
+						$sxe = new SimpleXMLElement("<item xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns:dct=\"http://purl.org/dc/terms/\"></item>");
+
+						$ns_dc = $item->children('http://purl.org/dc/elements/1.1/'); 
+
+						$rights = "";
+						$subject = "";
+
+						foreach($ns_dc as $key => $value){
+
+							@$sxe->addChild("dc:" . $key, "<![CDATA[" . $value . "]]>", "xmlns:dc=\"http://purl.org/dc/elements/1.1/\"");
+
+							if($key=="rights"){
+
+								$rights = (string)$value[0];
+
+							}
+
+							if($key=="subject"){
+
+								$subject .= $value . ",";
+
+							}
+
+						}
+
+						foreach($item as $key => $value){
+
+							@$sxe->addChild($key, "<![CDATA[" . $value . "]]>");
+
+						}						
+
+				          	$data_object = new StdClass();
+
+						$data_object->title = (string)$item->title[0];
+						$data_object->rights = $rights;
+						$data_object->subject = substr($subject,0,(strlen($subject)-1));
+						$data_object->package = $sxe->asXML();
+
+						$return->{$item->link} = $data_object;
+
+					 }
+
+				 }else{
+
+					echo "not a valid XML Feed";
+
+				 }
+
+			 }else{
+
+				echo "not a valid XML Feed";
+
+			 }
+
+			$content_info = array();
+
+			$content_info['resource_data_type'] = 'metadata'; 
+			$content_info['submitter'] = "ramanathan"; 
+			$content_info['curator'] = "ramanathan"; 
+			$content_info['active' ] = TRUE; 
+			$content_info['payload_schema'] = 'DC 1.1'; 
+
+			$opt_id_fields = array(
+					'curator',
+					'owner',
+					'signer',
+			);
+
+			$opt_res_fields = array(
+					'submitter_timestamp',
+					'submitter_TTL',
+					'keys',
+					'resource_TTL',
+					'payload_schema_locator',
+					'payload_schema_format',
+			);
+
+			$opt_sig_fields = array(
+				'signature',
+				'key_server',
+				'key_locations',
+				'key_owner',
+				'signing_method',
+			);
+
+			$opt_tos_fields = array(
+				'tos_submission_attribution',
+			);
+
+			// Make some parts of the PHP data structure
+
+			$identity = new StdClass;
+			$resource_data = new StdClass;
+
+			$identity->submitter_type = 'ramanathan';
+			$identity->submitter = $content_info['submitter'];
+
+			$resource_data->doc_type = 'resource_data';
+			$resource_data->doc_version = '0.23.0';
+			$resource_data->resource_data_type = $content_info['resource_data_type'];
+			$resource_data->active = $content_info['active'];
+			$resource_data->identity = $identity;
+
+			$resource_data->payload_placement = 'inline';
+			$resource_data->payload_schema = array($content_info['payload_schema']);
+
+			foreach($return as $data => $value){
+
+				echo "<p>URL (" . $data . ") submitted ";				
+
+				$content_info['resource_locator'] = $data; 
+				$resource_data->resource_locator = $data;
+
+				if($value->rights==""){
+					$tos = "None provided";
+				}else{
+					$tos = $value->rights;
+				}
+
+				$content_info['tos'] = $tos;
+
+				$tos = new StdClass;
+
+				$tos->submission_TOS = $content_info['tos'];
+
+				// Optional identity values.
+				foreach ($opt_id_fields as $field) {
+					if (array_key_exists($field, $content_info)) {
+						$identity->$field = $content_info[$field];
+					}
+				}
+
+				// Optional resource_data values.
+				foreach ($opt_res_fields as $field) {
+					if (array_key_exists($field, $content_info)) {
+						$resource_data->$field = $content_info[$field];
+					}
+				}
+
+				// Optional TOS values.
+				foreach ($opt_tos_fields as $field) {
+					if (array_key_exists($field, $content_info)) {
+						$tos->$field = $content_info[$field];
+					}
+				}
+
+				// Now the data structure is sort of finished, so add in some extra bits
+
+				$resource_data->TOS = $tos;
+
+				$resource_data->resource_data = htmlspecialchars_decode($data_object->package);
+
+				if($data_object->subject==""){
+					$keys = "None provided";
+				}else{
+					$keys = $data_object->subject;
+				}
+
+				$resource_data->keys = explode(",",$keys);
+
+				$submission = new StdClass;
+
+				$submission->documents[] = $resource_data;
+
+				$data_to_send = json_encode($submission);
+
+				// Curl is some PHP stuff to send data across the interwebs
+
+				$ch = curl_init();
+
+				curl_setopt($ch, CURLOPT_HEADER, 0); 
+				curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1); 
+				curl_setopt($ch, CURLOPT_URL, "http://alpha.mimas.ac.uk/publish");
+				curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 50); 
+				curl_setopt($ch, CURLOPT_USERPWD, "fred:flintstone");
+				curl_setopt($ch, CURLOPT_TIMEOUT, 50); 
+				curl_setopt($ch, CURLOPT_MAXREDIRS, 10); 
+				curl_setopt($ch, CURLOPT_HTTPHEADER, array('Content-type: application/json')); 
+				curl_setopt($ch, CURLOPT_POSTFIELDS, $data_to_send);
+
+				//CURL exec sends the data
+
+				$result = curl_exec($ch);
+
+				//Traps the error if something bad happens
+
+				$error = curl_error($ch);
+				curl_close($ch);
+
+				// Convert the data from LR json back into PHP so we can check it
+
+				$doc_data = json_decode($result);
+
+				echo "<a target='_blank' href='http://alpha.mimas.ac.uk/obtain?by_doc_ID=true&request_ID=" . $doc_data->document_results[0]->doc_ID . "'> Document ID : " . $doc_data->document_results[0]->doc_ID . "</a></p>";
+
+			}
+
+		?>
+		</div>
+	</body>
+</html>