Add GBU test for an increase of more than the given percentage from the median background

[mgalloy]

This needs to be done like the GBU median difference test in ucomp_l1_check_gbu_median_diff.pro and can be done in the same routine.

Questions

• What are the correct percentage differences for the various wave regions?

Tasks

• put given percentage in wave region config files
• add new GBU condition in ucomp_gbu_conditions
• add ability to average backgrounds to ucomp_average_l1_files
• test for new condition in ucomp_l1_check_gbu_median_diff
• run test, but do not fail (but record values)

[detoma]

This code should be based on what we did for COMP because what we are doing currently for UCOMP is not working and leads to both false positive and false negative see #358

[detoma]

We use all wavelengths (only I) for this test

We use all the good extensions in the files, i.e. the extension that passed all the other gbu tests, to compute the median of the BKG. Our reference is the median of all the BKG of the good extensions.

The test is the following:

An I extension fail this test, if the difference between the extension BKG and the median BKG computed above is more than 50% of the median BKG.

We allow for the BKG to be significantly lower than the reference in case there is a cleaning of the objective lens and the BKG level drops during the observing day.

[detoma]

This test assume that all wavelengths have the same bkg. This may not be true as we move into new observing codes that will span a broader wavelength range.

I think we likely do not need this for the GBU and the improved sigma values and the maximum bkg will catch these sudden changes. I move this test to v1.2.