Horne extraction error propagation

[hpparvi]

This PR adds uncertainty propagation to HorneExtract so that uncertainties from 2D input images are properly propagated to the extracted 1D spectrum. Turns out this was very straightforward since we're already calculating the uncertainty for the Horne extraction itself (it's 1/den without the normalisation). So, aside from handling uncertainty types and such, we actually needed to add one extra line of code... I also renamed img to flux so that the variable is not mixed with image, and so that the variable naming is more understandable at the end of the method.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.90%. Comparing base (e2baedb) to head (1ab2f37).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #296      +/-   ##
==========================================
+ Coverage   89.76%   89.90%   +0.13%     
==========================================
  Files          17       17              
  Lines        1817     1822       +5     
==========================================
+ Hits         1631     1638       +7     
+ Misses        186      184       -2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tepickering]

thanks for adding this! looks good except for one case that isn't covered by the new tests according to codecov. i would add that and then feel free to merge.

[tepickering]

it looks like coverage for this case is missing from the tests you added.

[hpparvi]

Fixed! The whole uncertainty type conversion could be simplified to three lines.

[tepickering]

well that's exciting. the devdeps test caught an upcoming numpy deprecation. maybe worth fixing here if it's easy...