The backend for a search engine on the proceedings of the Greek parliament.
Features:
* Stemming, tokenization, stopword removal.
* Inverted index for searching the contents of speeches.
* Keyword extraction from speeches.
* Finding similarity between two speeches, two speakers or finding the
  speech pairs with the greatest similarity.
* Search using Latent Semantic Analysis.
* A model which attempts to predict the party of the speaker, when given the contents of
  a speech.