More tests!

[cglosser]

More tests are always useful, but it'd be great to lock down the functionality of a couple key areas:

1. The AIM grid is a complex structure (it maps between real world coordinates, grid coordinates, and a grid index). Employing rapidcheck to ensure that things like a idx2grid(grid2idx(x)) transformation gives x for arbitrary grids would be useful.
2. Locking down results for various propagators. This test is probably most easily accomplished by building a collection of input/output tables analytically for various propagators and system geometries, constructing that same simulation in QuEST, and comparing the outputs to make sure each propagator works correctly.