Fix for spatial_neighbors breaks when transform='spectral'

[selmanozleyen]

Fixes #1014

[selmanozleyen]

thanks for the feedback but fau doesn't work with python 3.10.

[selmanozleyen]

@Intron7 should I merge this branch?

[flying-sheep]

looks good to me, but @Intron7 should check if it works in his example.

[codecov]

Codecov Report

❌ Patch coverage is 31.25000% with 11 lines in your changes missing coverage. Please review.
✅ Project coverage is 23.94%. Comparing base (d5ff8dd) to head (0086dac).

Files with missing lines	Patch %	Lines
src/squidpy/gr/_build.py	31.25%	11 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1028      +/-   ##
==========================================
+ Coverage   23.91%   23.94%   +0.02%     
==========================================
  Files          43       43              
  Lines        6368     6370       +2     
  Branches     1063     1064       +1     
==========================================
+ Hits         1523     1525       +2     
  Misses       4828     4828              
  Partials       17       17              

Files with missing lines	Coverage Δ
src/squidpy/gr/_build.py	19.75% <31.25%> (+0.66%)	⬆️

🚀 New features to boost your workflow:

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

[Intron7]

why np.float64?

[selmanozleyen]

because original function used np.float64. Didn't want to change old behaviour

[Intron7]

can we check if this is needed

[selmanozleyen]

well I can't tell tbh because there was no tests for the spectral transform lol. We can just use float32 I guess. But the better way would be to ask ths users of spatial_neighbours

[selmanozleyen]

@Intron7 ok the results seem similar when I checked them. I also added some numerical tests to ensure stability. Also when I recalled that this was just a normalization of weights in a graph so it should not be very sensitive to precision in theory.