Allow for Parallel Utilization

[kkappler]

This is desired and not required.

The natural parallel use case in this project is TF estimation, which is embarasingly paralelisable over each Fourier coefficient index. It is not clear however what the vision is for MTH5 specifically ... what aspect would parallize, ...

So, one way to do this is to use simple parallel python to do TF estimation (probably in the springtime).
This serves two purposes: 1. it provides breadcrumbs for a future user/collaborator for harnessing parallization 2. It will serve as a benchmark by which we can compare the effectiveness of dask, ray, or any of the prepackaged parallelizations