Introduce indexing and slicing of TimePartition

[jwallwork23]

Closes #99.

Currently, indexing a TimePartition returns various pieces of information associated with that subinterval. The changes introduced in this PR are more logical, i.e., indexing gives a sub-TimePartition. Similarly for slicing.

[acse-ej321]

@jwallwork23 - Looks straightforward and slicing will be useful.

I was wondering if the following line of code might also be impacted by the refactoring?

num_timesteps = self.time_partition[subinterval].num_timesteps
if num_timesteps > N:

[jwallwork23]

Thanks @acse-ej321, good spot. In this snippet of code, we want to get the number of timesteps for a given subinterval.

This could be done by taking the sub-timepartition and evaluating num_timesteps, which sums num_timesteps_per_subinterval over all subintervals (in this case there is only one), which is what would be done with the snippet you quote.

An equivalent but perhaps clearer way is to just take the appropriate index of num_timesteps_per_subinterval, so that's what I've done in the latest commit 06da0ad.

(The pre-commit hook also seems to have put an unrelated formatting change in there, too.)