some eltk refactoring

[scott farrar]

Hi Bill

I ended up following most of your suggestions about eltk re-naming and refactoring. Here's what I did in a nutshell. Knowledge bases (KBs) in the Description Logic world consist of two logically separate components: the ontology (sometimes called a TBox) and the various assertions about concepts in the ontology (called an ABox). So, the standard description of such a KB is:

KB = <TBox, ABox> where TBox contains classes and relations and ABox contains assertions about instances.

anyway, that's not so important  other than that's how I modeled the eltk.ontology module.

-eltk.ontology.Ontology is basically the TBox

-eltk.ontology.LinkedData is basically the ABox

(Note that I use LinkedData b/c that's the big W3C project that I'm trying to adhere to, see linkeddata.org for this.)

So, you'll be creating a LinkedData graph when processing the CSV files. I've prepared a page to illustrate this:

http://uakari.ling.washington.edu/e-linguistics/eltk/linkeddata.html

Note as well, that Ontology and LinkedData are both subclasses of rdflib.Graph.Graph. I've created an abstract reader class and a reader factory as well.

The latest documentation for the elkt can be found here:

http://uakari.ling.washington.edu/e-linguistics/eltk.html

Scott

[W.P. McNeill]

Ukari is down right now so I can't see the documentation.

Is there a reader class the best exemplifies the new refactoring?
IGTReader maybe? Something else.