Nighttime data identification does not work for multi-night missions

[MBARIMike]

The nighttime identification logic in resample.py is too simplistic. It was designed for the Diamond missions where there is just one sunset and and most one sunrise during a mission. As seen in one of the MBTS line missions (http://stoqs.mbari.org/p/lblVNGs) no biolume_proxy_ parameters are produced because no nighttime data are identified, e.g.:

INFO 2023-09-02 21:00:50,065 resample.py select_nighttime_bl_raw():406 [1] Sun is up at start, but no sunset in this mission
INFO 2023-09-02 21:00:50,065 resample.py select_nighttime_bl_raw():417 [1] No sunset during this mission. No biolume_raw data will be extracted.
INFO 2023-09-02 21:00:50,065 resample.py add_biolume_proxies():634 [1] No nighttime_bl_raw data to compute adinos, diatoms, hdinos proxies

The logic should be changed loop over multiple days to find all the nighttime data.

[MBARIMike]

The dorado 2020.337.00 mission was reprocessed with the code change made in #13 and loaded into the stoqs_all_dorado_t test database. Adding "_t" to the http://stoqs.mbari.org/p/lblVNGs request shows that the biolume_proxy_ parameters are now processed, e.g.:

[MBARIMike]

This still needs to be pulled to production, but that will be done along with the other corrections as in #3, #4, and #5.