Subsetting of datasets with 2d coordinates

[konstntokas]

• Improved xcube_resampling.utils.clip_dataset_by_bbox to support datasets with
  2D coordinates. This function is also used internally by
  xcube_resampling.rectify.rectify_dataset.
• Added reproject_bbox, a utility to reproject a bounding box from one CRS to another.
• Added bbox_overlap, a utility that computes the fraction of the source bounding box
  overlapped by the target bounding box.
  Closes Implent subsetting of a irregular dataset #3

Checklist:

• Add unit tests and/or doctests in docstrings
• Add docstrings and API docs for any new/modified user-facing classes and functions
• New/modified features documented in docs/
• Changes documented in CHANGES.md
• CI checks pass
• Test coverage remains or increases (target 100%)

[codecov]

Codecov Report

❌ Patch coverage is 98.37398% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 98.81%. Comparing base (0ec6c04) to head (25284b8).
⚠️ Report is 7 commits behind head on main.

Files with missing lines	Patch %	Lines
xcube_resampling/rectify.py	94.28%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #12      +/-   ##
==========================================
- Coverage   98.85%   98.81%   -0.05%     
==========================================
  Files          21       21              
  Lines        1924     2021      +97     
==========================================
+ Hits         1902     1997      +95     
- Misses         22       24       +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.

[forman]

Please also check API of xcube.core.geom for commonalities.

[forman]

Add types

[konstntokas]

done

[forman]

load will modify the original dataset which is not always wanted, i.e., users loose chunking. Either make it an option, but preferrably we should stay in dask world if it is a dask array. If this isn't possible, please comment in code, but anyway, use DataArray.compute() or DataArray.values to not modify the original dataset.

[konstntokas]

Keeping it in dask array, will load the coordinates twice in

rows = np.any(mask, axis=1)
cols = np.any(mask, axis=0)

also when this is implemented in dask. I therefore wanted to specifically load the mask and get the indices using purely numpy.

I change to mask = mask.values, since the array is used in numpy functions only. Further, I added a comment into the code.

[forman]

OK!

[forman]

You should add required type aliases from constants so the documentation can refer to them.

[konstntokas]

I added them: see https://xcube-dev.github.io/xcube-resampling/api/#xcube_resampling.constants

[konstntokas]

Please also check API of xcube.core.geom for commonalities.

I looked at it shortly. I will not be able to work on this this week. I put it on my list and will merge this module with the utils.py in xcube-resampling in the next release.