Fix derivation of correct radius of influence when data layout is not standard

[adybbroe]

This is supposed to solve #554

Make use of the fact that longitudes are an xarray data array and use 'y' dimension for rows/scanlines

This makes the code resilient towards any other (awkward) data layout than the standard (where first dimension is usually the rows).

• [ ] Closes #554
• [x] Tests added
• [x] Tests passed
• [ ] Passes git diff origin/main **/*py | flake8 --diff
• [ ] Fully documented

[adybbroe]

It looks like I can assume that self.lons is an Xarray DataArray, and never (just) a numpy array, correct?

[djhoese]

It looks like I can assume that self.lons is an Xarray DataArray, and never (just) a numpy array, correct?

For Satpy, yes, but we can't (and shouldn't) make that assumption for pyresample.

[adybbroe]

It looks like I can assume that self.lons is an Xarray DataArray, and never (just) a numpy array, correct?

For Satpy, yes, but we can't (and shouldn't) make that assumption for pyresample.

Ok, got me there then! I was looking for tests on that, but found only with xarray, but possible that I just overlooked such tests...

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (6a8afc0) 94.11% compared to head (a7650f4) 94.14%. Report is 11 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #555      +/-   ##
==========================================
+ Coverage   94.11%   94.14%   +0.03%     
==========================================
  Files          82       84       +2     
  Lines       13078    13199     +121     
==========================================
+ Hits        12308    12426     +118     
- Misses        770      773       +3     

Flag	Coverage Δ
unittests	94.14% <100.00%> (+0.03%)	:arrow_up:

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

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

[coveralls]

coverage: 93.725% (+0.04%) from 93.69% when pulling a7650f4c5725ccb3557ea381b3e8eb396d149259 on adybbroe:finding-geocentric-resolution-resilient-to-data-layout into 6a8afc0085e0b4269f00991ab79fe1b3766bb817 on pytroll:main.

[adybbroe]

@djhoese I made a comment in the issue: https://github.com/pytroll/pyresample/issues/554#issuecomment-1820556077

Which solution do you propose I pursue? And does @gerritholl and @mraspaud have opinions as well?