JP-2075: fixed flux scaling issues in wfss_contam #8416
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #8416 +/- ##
==========================================
+ Coverage 56.38% 56.47% +0.09%
==========================================
Files 387 387
Lines 38716 38730 +14
==========================================
+ Hits 21830 21874 +44
+ Misses 16886 16856 -30 ☔ View full report in Codecov by Sentry. |
|
Have you run a regression test? I believe we have 1 or 2 tests of wfss_contam in the regtest suite. Of course we'll expect differences/failures due to these updates. |
|
Didn't realize there were any wfss_contam regtests; run started here |
Test results show zero failures, which normally is a good thing, but I expected differences, because we do have 1 regtest for this: test_nircam_wfss_contam.py. However, I see that that test module is set to be skipped, because of the ridiculously long run time. Can you try running that one test locally in your PR branch environment? |
|
Sure, I started that just now edit: see JIRA ticket for results |
Resolves JP-2075
Closes #6016
This PR fixes a difference in the overall signal level of simulated contamination spectra generated by the step relative to the spectra in the grism image being corrected. The flux scaling issues were found to be caused by a combination of two problems: First, background fluxes were being propagated into the grism image when the input direct image was not background subtracted, leading to unphysical broadening in the cross-dispersion direction for the model dispersed sources. Second, flux oversampling in the dispersed_pixel() function was leading to double-counting of flux; dividing the result by the oversampling factor resolved this.
Checklist for maintainers
CHANGES.rstwithin the relevant release sectionHow to run regression tests on a PR