Description

There are two straightforward ways to load demo data into the patient matching module:

1. run a previous sql dump (pro: easy, con: inflexible)
2. use a patient data loading program that accesses the OpenMRS API (pro: flexible, con: need to write)

We have a patient data loading program already (written by nribeka), and I would like to make it capable of loading multiple test cases, e.g.

1. 7500 patients with unique uids: no noise, 2500 links which are exact matches (as it is now)
2. 75000 patients with unique uids: no noise, 25000 links which are exact matches (previous case, times ten)
3. 7500 patients with unique uids: simple noise, 2500 links which match on all but one field

As of right now, a tool like this is not in the repository -- the goal is to polish existing code and introduce case (2), with the intent of reproducing a hypothesized garbage collection exception when the linkage process is run on a large (>10000) patient database.