Enable processing CW/FM interleaving data

[leewujung]

This addresses #1506 and #1507 in one go, since they are tightly linked.

@ctuguinay : I think I've only gotten to store ping-by-ping filter coefficients (FIL1 datagram), but it's been a while so my memory is pretty faded. Please check. Thanks a lot for picking this up!

[leewujung]

@ctuguinay : I have made a number of changes to clean a few things up - the main things are:

• the logic surrounding calibrate using using a single or multiple different filter_time in calibration/api.py::_compute_cal()
  • there are many remaining TODOs that I don't think I can get to in the next few days, so I think we can go ahead to merge this (because the code works) to get a release ready, and then debug/fix in a later release
• simplify _retrieve_correct_beam_group_EK80 to return only 1 beam group name
  • I think _retrieve_correct_beam_group_EK60 and _retrieve_correct_beam_group_EK80 can be merged to leave just one function, but am also running out of time to do that -- will fix in a later release

I also have a question about this code block: I couldn't immediately figure out why this block would not work when there is only 1 filter_time in the Vendor-specific group, but it fails test_ek80_BB_complex_multiplex_NaNs_and_non_NaNs for both files:
https://github.com/leewujung/echopype/blob/62aeaf19100ed39a07ca4e2e4b7c049b0ef54531/echopype/calibrate/api.py#L127-L164
For the first file "NYOS2105-D20210525-T213648.raw" , instead of producing Sv with 4 channels, it only produces data for 2 channels. I think something is not quite right for the stack operation, but I couldn't figure out why dropna("pairs") drops those channels. Maybe you know?