23 tests for hot swappable numpy backends demo use

[kenjim21]

added jax backend compatibility for mueler functions and corresponding testing

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 38.70968% with 76 lines in your changes missing coverage. Please review.

Project coverage is 74.35%. Comparing base (3b8fdc3) to head (b6797eb).

Files	Patch %	Lines
tests/test_mueller.py	37.83%	69 Missing :warning:
tests/test_katsu_math.py	46.15%	7 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #24 +/- ## =========================================== - Coverage 98.38% 74.35% -24.04% =========================================== Files 3 3 Lines 186 308 +122 =========================================== + Hits 183 229 +46 - Misses 3 79 +76 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[Jashcraf]

Ah looks like adding Jax as an optional dependency and not explicitly testing it hurts code coverage. I think that's fine for now, but just writing a note here to consider a way of structuring a test file to maintain the code coverage. Maybe removing it from the automated codecov reporting or just configuring the tests on the remote machine

[kenjim21]

For the codecov we can def look at adding all the possible autodiff libraries that are being supported to a special designation for testing so that while it's not on the requirements.txt, it'll be install for the automated testing

[Jashcraf]

Oh man did it run with the test uncommented?

[kenjim21]

This is the margin of difference between the matrices when I print them

[Jashcraf]

So that value in the lower-left is concerning, but not huge - can you make an issue to investigate this difference? For now I think this is fine for the PR

[Jashcraf]

3…2…1…

LGTM!

[kenjim21]

Ya i'll add an issue. it looks like for any of the non zero values minus the top row there's its almost consistently off at around the 6th decimal or further