Det match updates

[mmccrackan]

Replaces #1515 which was similar. This does the following:

• Adds a parameter called allow_unassigned_to_assigned which will either allow or prevent matches from bg=-1 in the design or pointing data to bg != -1 in the pointing or tuneset data. We want to prevent such matches when matching from design to pointing, since the detector types with bg=-1 should always have that value. When matching between pointing and tunesets, we allow bg = -1 to other bg values since some detectors may be assigned bg = -1 in det_cal due to the conditions of when the pointing tune was taken but may be recovered in later tunesets (this likely won't happen since it would ideally match the pointing to a bg != -1 detector and copy the bg from the design).
• Like Det match updates #1515, forces det_types of UNRT, SQID, and BARE to always be bg = -1.
• Prevents mismatch between assigned bgs (i.e matching bg=4 to bg=6).
• Makes det_row, det_col, and det_rhombus optional for LF.
• Adds det_match_solutions code to coords.
• Adds start_time and stop_time to update_det_match to simplify epoch handling for detsets. Future update will add epochs to the databases.

Remainder of LF matching updates will be in a separate PR after testing and confirming it works on better pointing data (also needs #1603).

[mhasself]

Thanks -- mostly documentation / tidying requests, but I understand if you need to press forward. Great to have this side of things covered in sotodlib now.

[mmccrackan]

Thanks -- mostly documentation / tidying requests, but I understand if you need to press forward. Great to have this side of things covered in sotodlib now.

Okay, fixed imports, updated the docstrings and documentation to add a description of the solutions code and what is meant by solutions and pointing, and added and tested the solutions script to update_det_match and the CLI interface.