Add Sampled Cross-Sections

[philipc2]

Closes #1318 #1319 #1320 #1296

Overview

• Adds a UxDataArray.cross_section() method, which can be used to sample along lines of constant lat/lon or arbitrary GCAs
• Moves existing cross_section methods under the subset namespace and begins deprecation
• Adds support for restricting the latitude or longitude range when doing a constant lat/lon subset
• Safeguards cases where no faces are found when conducting a subset
• Updates the Cross Section user guide notebook to use to showcase the new functionality, with vertical cross section examples.

[philipc2]

@rajeeja

I'm opening up this PR to isolate our interpolation/sampling from the vertical cross sections examples in #1317

I've also put together a more general approach and cleaned up the implementation in #1317

[philipc2]

Sampling along a transect is very similar in concept to the work I put together in #1271, however instead of sampling onto pixels in an Image, we sample points along the transect.

[rajeeja]

Oh, if possible can you create a PR or changes directly to my PR. Are we planning to close it in favor of this?

[philipc2]

Oh, if possible can you create a PR or changes directly to my PR. Are we planning to close it in favor of this?

Let's keep the API changes in this PR and use #1317 for applications to vertical cross sections (i.e. the updates to the notebooks, etc.)

Does that work? I'm happy to hop on sometime this week to pair program too.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[Copilot]

Pull Request Overview

This PR adds comprehensive cross-section functionality for sampling data along constant latitude/longitude lines and arbitrary great circle arcs. The changes consolidate cross-section methods under the subset namespace while deprecating existing methods and introduce a new sampled cross-section interface for data extraction.

• Adds new UxDataArray.cross_section() method for sampling along arbitrary great circle arcs and constant lat/lon lines
• Moves existing cross-section methods from cross_section to subset namespace with deprecation warnings
• Implements constrained latitude/longitude range support for subset operations

Reviewed Changes

Copilot reviewed 8 out of 11 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
uxarray/subset/grid_accessor.py	Added new constant_latitude/longitude methods and interval methods with enhanced error handling
uxarray/subset/dataarray_accessor.py	Added new constant_latitude/longitude methods with lon/lat range constraints
uxarray/cross_sections/sample.py	New file implementing sampling functions for geodesic, constant latitude, and constant longitude cross-sections
uxarray/cross_sections/grid_accessor.py	Deprecated existing methods with warnings and redirects to subset accessor
uxarray/cross_sections/dataarray_accessor.py	Replaced with new callable cross-section interface and deprecated old methods
test/test_cross_sections.py	Updated tests to use subset namespace and added new cross-section functionality tests
docs/user-guide/cross-sections.ipynb	Updated documentation to showcase new cross-section sampling functionality
docs/api.rst	Updated API documentation to reflect new structure

Comments suppressed due to low confidence (1)

[Copilot]

The docstring contains a grammatical error: 'selecting all faces are within' should be 'selecting all faces that are within'

[Copilot]

Setting __doc__ to __call__.__doc__ creates a circular reference since __call__ doesn't have its own docstring. This should either be removed or set to the actual docstring content.

Suggested change

__doc__ = __call__.__doc__

[rajeeja]

looks good to go, minor fixes. in comments.

[erogluorhan]

This looks great to me except a small comment I've put into the cross-sections notebook (can you see it?)

[rajeeja]

call still assumes 'n_face' exists, raise error here itself if not face-centered. Test for great clrcle case..

[rajeeja]

np.nan cannot be represented in integer dtype; this might raise or produce invalid values.