Add continuum check to arc lamp standards

[sybenzvi]

The easiest way to check for stray light is elevated continuum (particularly in the r and z cameras) during arc cals. Probably this can be done by cutting out all prominent lines and looking at the resulting integrated flux in each camera.

[sbailey]

@jose-bermejo @sybenzvi pinging you about this Nightwatch enhancement request. There was a faint light left on the past few nights (20231215 and 20231216) and this test might have been able to catch it.

[deisenstein]

My Jupyter code is here on NERSC:

/global/u1/e/eisenste/public/CalibrationLamps.ipynb

What this does is compare between two sets of cals. Ideally, one holds one set "fixed" at some version thought to be good. I'm sure one could make different choices about how to separate the lines from the continuum, and surely write better python.

I note that the lamps have changed over time, such that I have had to occasionally change the reference night.

[sybenzvi]

Thanks @deisenstein , we based the calibration standards now in Nightwatch off your notebook. There is a continuum check for the LEDs and an EW check for the arc lines, but we didn't implement the median-filtered plots of the arc lamps, which is useful for picking out stray light. We also correlated the drifts in the light levels to the air temperature in the dome but didn't implement automatic adjustments in the standards yet.