Flux scaling error fix

[tjgalvin]

I tracked down an error in the flux calibration stage in some data I
have been working with. If there was a nod and shuffle source observed,
then the NEXP, SEXP, NSUBEXPS header fields will still contain this
information when other observing modes are used. When the pipeline uses
the ’ns’:True parameter, signifying that N&S has been used, the exptime
is incorrectly re-calculated using these N&S parameters for exposures
made with other modes - like ClassicEquall for the standard star. This
can obviously throw off the calibration as the observed flux is scaled
in by exptime.

The fix checks the observing mode of a frame. If it is not a N&S, then
falls back to the exptime in original header, which has not been edited.

[chrusher]

This is fixing the bug in the wrong place - the pipeline should check the headers in run_slitlet_mef (and run_sky_sub and run_cosmic_rays) to determine which files are nod and shuffle and which are not.

[tjgalvin]

More robust checking could definitely be thrown in during those routines as well, and you are right. Avoiding the incorrect ToS being written would be the far better solution. I needed a quick fix, but want to report the error so the important people know. I think it is is the run_slitlet_mef that actually computes the ToS stored in the header. The other tasks I don't think contribute to the problem as it is the incorrect ToS that gives the incorrect flux scaling - but it has been a while since I looked at the code.

I may have some time to put together a better fix.