HLA-1161: MagErrAp2 and other magnitude errors in HAP catalogs have incorrect negative values

[s-goldman]

Resolves HLA-1161

This PR addresses an issue (#1704) where (1) magnitude errors are being calculated as negative numbers, and (2) these are used for threshold calculations resulting in bad sources with good flags.

With this code we switch to calculating the good_snr threshold based on the flux error (which is always positive). We also fix the negative magnitude error issue by setting the magnitude error to infinity is the flux (and consequently the magnitude) is less than zero.

I had to change the number of expected sources for one of the regression tests because it no longer recognizes two sources with magnitudes and fluxes = -9999.

I also added dates to our new Architecture Decision Record (ADR) document.

Checklist for maintainers

• added entry in CHANGELOG.rst within the relevant release section
• updated or added relevant tests
• updated relevant documentation
• added relevant milestone
• added relevant label(s)
• ran regression tests, post a link to the Jenkins job below.
  How to run regression tests on a PR

Jenkins Test (Dev build failing for Numpy 2.0 issue)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (71b66d8) 33.78% compared to head (f534432) 33.78%.

❗ Current head f534432 differs from pull request most recent head f91ab9c. Consider uploading reports for the commit f91ab9c to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1700      +/-   ##
==========================================
- Coverage   33.78%   33.78%   -0.01%     
==========================================
  Files         126      126              
  Lines       31169    31172       +3     
  Branches     5715     5715              
==========================================
+ Hits        10530    10531       +1     
- Misses      19450    19451       +1     
- Partials     1189     1190       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[s-goldman]

Had to change the number of expected sources in the point source catalog for test_svm_segment_total_cat() in test_svm_ibqk07.py. Currently investigating whether that is linked to these changes.

[nden]

I left a couple minor comments inline. I'm not familiar with the code but reading the discussion in the ticket, the changes here address that so I'm approving the PR.