Overview of the SIOC browser

I promised to do it about 2 months ago, here’s finally an overview of my SIOC browser.

As we’ll work on it with Uldis for our BlogTalk presentation with John, here’s the time for a few informations.

So, the browser is only a part of a more important architecture which consists in:

  • SIOC data from people creating it with some of the currently available exporters. Data could be produced by weblogs (dotclear example), forums (anime.ie example using Drupal) or on-line services, as TalkDigger;
  • Uldis SIOC crawler to get the data locally;
  • an RDF Store (I’m currently using 3store – will try to blog about it soon – but I’ve started with Joseki) to store this data;
  • the browser as a query / visualization interface.

(While writing this post, Uldis shown me this other view of the SIOC architecture)

The browser itself, consists of a set of query pages (well, currently only 3 of them 😉 ) to visualize the content of the store with an user-friendly (I hope) AJAX interface.

Each query is define in a XML config file, which contains:

  • A title
  • A page description
  • The main SPARQL query
  • Query filters
  • Formating results function
  • Item SPARQL query
  • Item formating results function

So the files contains both parts of SPARQL and PHP code, as it can be seen in comments.xml file.

These config files are rendered by the browser engine – a PHP5 class that dynamically create rendering functions using part of code contained in the config files – that displays the browsing page, runs the main query without filters, using the formatting function to display results. Then, users can apply filters, that will reload the query using AJAX, or select result items to launch the Item query, where results will be formatted using the other method defined in the config file.

Regarding the interface with the store, queries are send over HTTP using SPARQL, and the results should be JSON formatted, to be parsed with a PHP extension.

So what’s need to be done ?

A few ideas:

  • Move XML config files to RDF ones ?;
  • Find cool queries;
  • Graphical queries / interfaces;
  • Add more data 🙂

We should then release the browser, so that you can play with your own data. Hope everything will be ready for BlogTalk !

Leave a Reply

Your email address will not be published. Required fields are marked *