I am considering the task of converting the templates from the gene articles in Wikipedia (http://en.wikipedia.org/wiki/Portal:Gene_Wiki) to use/create wikidata assertions.  This involves an extensive update of the template structure as well as the code for the bot that keeps them in sync with external public databases.  (https://bitbucket.org/sulab/pygenewiki)

More specifically I'm thinking about working with a Google Summer of Code student on this project.

Given a time frame of now through August, would it make sense for us to pursue this objective directly in the context of wikidata (through the public API).  Or would it be better for us to install our own version of the wikibase software (kept in sync with code updates) and develop the new gene wiki bot code locally with the aim of switching to the public API later?  Or is it too early to consider this project?

I want to get involved and support wikidata with this important data, but I'm hesitant to ramp up development (especially with a student) in a moving target situation.

Any thoughts?