It has been one and a half year since I finished my thesis about image tagging but my interest on unstructured and structured data is still there. I recently came across a semantic web service, powered by Thomson Reuters. It is called OpenCalais which offers an API to access data related to your published content. On the official OpenCalais site I found the Tagaroo WordPress plugin and wanted to give it a try on my private blog. That is the reason why I publish this article. OpenCalais does not support German yet, so here is my first english post.
While writing this post I found out that:
- OpenCalais needs more then 64 words to start suggesting tags.
- After writing and doing any key-event on your keyboard tagaroo makes an AJAX request to OpenCalais
- Tagaroo seems to understand what I am writing and I assume analyse only on subjectives (not verbs)
- Tagaroo stays in the context of my article and does not suggest any unrelated tags
- The wysiwyg field and the cursor at the beginning freezes because the Tagaroo plugin does some JavaScript work
- In the XMLHttpRequest, the plugin sends the unstripped html content
- The response is a RDF format with some rdf xml elements like description, about, resource, type, socialtag, name, importance.
- The responses represent suggested tags for your content
- The suggested tags has to be added manually or can be ignored
Here is the screenshot of the plugin in action, when you write a wordpress article.
Here is an example of the rdf response xml structure.
<rdf:rdf xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:c="http://s.opencalais.com/1/pred/"> <rdf:description c:allowDistribution="false" c:allowSearch="false" c:calaisRequestID="4c587933-e4b5-820a-129c-26018f04900f" c:externalID="draft-1278862736" c:id="http://id.opencalais.com/4sOfLZ6dYUMOslB-UHK5vQ" rdf:about="http://d.opencalais.com/dochash-1/f980c9d1-5d23-394b-bbc3-af9bc9371a7e"> <rdf:type rdf:resource="http://s.opencalais.com/1/type/sys/DocInfo"> </rdf:type> <c:document> <![CDATA[<body><p> some text here </p> </body>]]> </c:document> <c:doctitle> </c:doctitle> <c:docdate> 2010-07-11 11:39:24.875 </c:docdate> <c:externalmetadata> <rdf:description> <c:caller> Tagaroo </c:caller> </rdf:description> </c:externalmetadata> <c:submitter> http://blog.nhanvu.de </c:submitter> </rdf:description> <rdf:description c:contentType="text/html" c:emVer="7.1.1103.5" c:langIdVer="DefaultLangId" c:language="English" c:processingVer="CalaisJob01" c:stagsVer="1.0.0-b1-2009-11-12_16:54:24" c:submissionDate="2010-07-11 11:39:24.656" rdf:about="http://d.opencalais.com/dochash-1/f980c9d1-5d23-394b-bbc3-af9bc9371a7e/meta"> <rdf:type rdf:resource="http://s.opencalais.com/1/type/sys/DocInfoMeta"> </rdf:type> <c:docid rdf:resource="http://d.opencalais.com/dochash-1/f980c9d1-5d23-394b-bbc3-af9bc9371a7e"> </c:docid> <c:submittercode> 7a5d6101-44b2-587e-4a77-de8802b9c7d7 </c:submittercode> <c:signature> </c:signature> </rdf:description> </rdf:rdf>
In order to use the Tagaroo plugin, you need to sign up for an OpenCalais API key.
