Prototype and Presentation

Last Friday, just minutes before my presentation, I completed a (very) rough prototype of the platform. It successfully processes input XML files into fully indexed SOLR documents with pre-serialized representations attached. As such, I was able to give a very short demo and talk a bit about future plans.

RCOS Fall 2010 Presentation 2

View more presentations from GraylinKim.

Unfortunately the demo was too short to really demonstrate how powerful the simple process was. Following a very simple protocol an object was indexed across every available field and sub-objects were branched off into their own fully indexed documents. These multiple documents and full indexes pushed into SOLR now allow for fairly sophisticated queries:

• Find all bills sponsored by Senator Adams and voted against by Alesi.
• Find all bills moved to 3rd Reading in the last 7 days.
• Find all bills unanimously approved in May 2010.
• Find all actions on all bills voted down last session.

This might not seems so impressive, and indeed its not (SOLR does all the heavy lifting) but its worth noting that this process would work equally well regardless of the complexity and structure of the incoming documents. Now that this proof of concept is done, I can begin to build on this platform to code best practices and enhancements into the processing.

I'm going to be pushing some of the Open Legislation data through the system sometime this week and push it up as a demo application. Hopefully this process will help me refine the set of XML input flags and shake out bugs.