Resample pop data

[timlinux]

This PR adds support for generating Wee Score x Population outputs to be able to validate that every permutation of the 3 population levels and 5 wee score classes (15 permutations in all) is calculated correctly so I created two 4x4 sample datasets (one for pop classes and one for wee scores) that will produce every combination when run through the wee processor I wrote. You can see my testbed outputs here.

This is how the test datasets are structured:

Population classes:

Wee scores:

The tests run automatically on GitHub so we can always be sure that we don't break things as we add more features in the future.