NaNs were messing up phase_change

[tjlane]

Using meteor.phaseboost, it was possible to get NaNs in the output metadata:

2025-02-04 20:01:49 [info     ] convergence criteria:          max_iterations=5.0 phase_tolerance=0.001
2025-02-04 20:01:51 [info     ]   iteration 0001               negentropy=0.0034 phase_change=0.8369 tv_weight=0.01
2025-02-04 20:01:53 [info     ]   iteration 0002               negentropy=0.0073 phase_change=nan tv_weight=0.01
2025-02-04 20:01:55 [info     ]   iteration 0003               negentropy=0.0109 phase_change=nan tv_weight=0.01
2025-02-04 20:01:58 [info     ]   iteration 0004               negentropy=0.0139 phase_change=nan tv_weight=0.01
2025-02-04 20:02:00 [info     ]   iteration 0005               negentropy=0.0164 phase_change=nan tv_weight=0.01
2025-02-04 20:02:02 [info     ]   iteration 0006               negentropy=0.0184 phase_change=nan tv_weight=0.01

During iterative TV, it's possible that we compare arrays of native/derivative data without the same indices. In this case, NaNs are generated at the positions of mismatched HKL indices, but dealt with gracefully in downstream code... except for the average_phase_diff_in_degrees function in utils.

This PR updates that function to simply ignore NaN values. Based on feedback from @kmdalton, we proactively filter out non-common indices, which precludes the generation of the offending NaNs in the first place.

A regression test added to enforce this.

[tjlane]

merge #97 before merging this PR

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.63%. Comparing base (8769c2f) to head (fbe5e49).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #98   +/-   ##
=======================================
  Coverage   95.62%   95.63%           
=======================================
  Files          15       15           
  Lines         938      940    +2     
=======================================
+ Hits          897      899    +2     
  Misses         41       41           

Flag	Coverage Δ
unittests	95.63% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[kmdalton]

can you either

• filter outside the while loop? or
• calculate the indices used for filtration outside the loop?

I'm sure it doesn't affect the code's correctness, but it unnecessarily costly as written. Not a big deal if it's a hassle.