Enforce consistency between CG table potentials and cut off values

Hoomd table potentials only take an r_min and an r_max (r_cut), and split the entire arrays provided for energy and force between them, so we need to make sure that the potential files are being sliced according to the r_cut used in the simulation. The table potentials (derived in MSIBI) might go out to r values larger than we may want to use in the sim, so we can't just pass in the whole column from these files (which is the current behavior).

Performing some array slicing in `_create_hoomd_sim_from_snapshot` should solve this issue.