Allow desi_proc_night to restart mid processing in calibrations

[akremin]

desi_proc_night is designed to pick up where it left off based on information saved in the processing_table. However, there was a bug in the code that applies only to nights with some processing table entries and no science data where it was exiting due to an if statement checking whether there were new science exposures. The if statement was meant to speed up the exiting of the script so that it didn't need to run multiple computationally intensive functions just to arrive at the same conclusion. Therefore, I decided not to remove the logic, but improve it such that it desi_proc_night still quickly exits if no new science exposures have been observed while fixing the case that was failing before.

Case that was failing before:
We had processed 4 arcs but nothing else. (This arose due to a bad camera leading a job to fail. After Abhijeet edited the exposure table to flag the bad cameras I removed later entries in the processing table and tried to rerun, but that failed due to the above issue). This branch now works in that instance.

[coveralls]

coverage: 30.192% (-0.005%) from 30.197%
when pulling 46dae3f on daily_cal_no_science
into e98abca on main.