Potential bug with pickling/unpickling TimeInterpolatedPotential

[adrn]

Reported by @TomWagg:

There may be a bug leftover from #540 in which TimeInterpolatedPotential instances do not get reconstructed properly after unpickling. Also, make sure test_time_interpolated.py is testing all of the things that potential_helpers.py does. And make sure the pickling test does a comparison of values after evaluating the unpickled and original potential instances.

[TomWagg]

Just clarifying from earlier - I remember now that this bug does still exist, I just put a workaround in cogsworth so I don't see it anymore. I basically just avoid pickling the potentials entirely in the starmap call for orbit integration so I avoid this bug.

Sorry for the confusion!