Optimize Point in Face Computation

[philipc2]

Overview

• Optimizes Grid.get_faces_containing_point()
• Uses a SciPy KDTree, with similar benefits as seen in Refactor & Optimize Remapping (NN & IDW) #1244
• Enables batched point queries by passing in an array of points, allowing for a single call to the KDTree
• Implements a spherical winding number approach for determining if a point contains a face

[Copilot]

Pull Request Overview

This PR optimizes how grid faces containing a point are found by introducing a SciPy KDTree cull and Numba-accelerated spherical winding‐number tests, and refactors legacy implementations.

• Adds uxarray/grid/point_in_face.py with Numba functions for single and batched point‐in‐face queries using spherical winding numbers.
• Updates Grid.get_faces_containing_point to a keyword‐only API that delegates to the new batched query path.
• Replaces legacy geometry routines in uxarray/grid/geometry.py with parallelized Euclidean radius and winding‐number implementations and removes deprecated functions.

Reviewed Changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
uxarray/grid/point_in_face.py	New Numba routines for fast face‐in‐point queries, batched support
uxarray/grid/grid.py	Refactored get_faces_containing_point signature and call path
uxarray/grid/geometry.py	Replaced old point‐in‐face and radius logic with parallel versions
test/test_point_in_face.py	Added tests for new point‐in‐face behavior and return_counts
test/test_grid.py	Removed outdated face containment tests
test/test_geometry.py	Updated tests to call _face_contains_point and normalize coords

Comments suppressed due to low confidence (3)

uxarray/grid/geometry.py:1082

• The code uses math.sqrt but math is not imported in this module. Add import math at the top.

return math.sqrt(global_max2)

uxarray/grid/grid.py:2610

• The function _lonlat_rad_to_xyz is not imported in this file. Add from uxarray.grid.coordinates import _lonlat_rad_to_xyz to the imports.

point_xyz = _lonlat_rad_to_xyz(lon, lat)

uxarray/grid/point_in_face.py:216

• [nitpick] Casting lists to int64 may cause a later implicit conversion to INT_DTYPE (e.g., int32) in the Numba call. Cast directly to INT_DTYPE for consistency and to avoid extra copies.

flat_cands = np.concatenate([np.array(lst, dtype=np.int64) for lst in cand_lists])

[Copilot]

Pull Request Overview

This PR optimizes the point‐in‐face computation by introducing a SciPy KDTree for candidate culling and a new spherical winding‑number approach accelerated with Numba.

• Implements new functions (_face_contains_point, _batch_point_in_face, etc.) for efficient point‐in‐face querying.
• Updates Grid.get_faces_containing_point to support batched queries and both Cartesian and spherical inputs.
• Refactors related utility and test functions to align with the new implementation.

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
uxarray/grid/point_in_face.py	Introduces optimized functions with KDTree culling and winding tests.
uxarray/grid/grid.py	Updates the API and docstring for get_faces_containing_point.
uxarray/grid/geometry.py	Revises max face radius calculation with parallel Numba implementation.
uxarray/grid/coordinates.py	Adds _prepare_points_for_kdtree and updates error handling.
test/*	Updates and adds tests to cover the new behavior and edge cases.

Comments suppressed due to low confidence (2)

uxarray/grid/coordinates.py:873

• The error message contains a typo ('One one can be provided'). Please change it to 'Only one can be provided at a time' for clarity.

Both Cartesian (xyz) and Spherical (lonlat) coordinates were provided. One one can be provided at a time.

uxarray/grid/geometry.py:1082

• The math module is used here but it is not imported. Add 'import math' at the top of the module to ensure math.sqrt is available.

return math.sqrt(global_max2)