test(rasterize): close public-API coverage gaps (deep-sweep pass 1)

[brendancol]

Summary

Closes documented but untested public-API surface in xrspatial.rasterize flagged by the test-coverage sweep on 2026-05-17. Adds 34 tests in xrspatial/tests/test_rasterize_coverage_2026_05_17.py. Sweep is test-only; no source files modified.

Coverage gaps closed:

• Cat 3 HIGH -- 1x1 single-pixel raster across all four backends. test_rasterize.py covers 1xN strips and Nx1 strips but never the width=1 AND height=1 case, so the polygon scanline, line Bresenham, and point-burn kernels ship without the single-cell degenerate case.
• Cat 4 HIGH -- like= template-raster parameter (rasterize.py:2038) is implemented by _extract_grid_from_like (line 1930) but no test exercises it. Pins dtype/bounds/coords inheritance, the three override branches (dtype, bounds, width/height), the three validation branches (not-DataArray, 3D, wrong dim names), and like= on all four backends. Mutation against rasterize.py:2183-2184 flipped the inheritance test red.
• Cat 4 HIGH -- resolution= happy path. Only the oversize-rejection error path was tested (test_oversize_resolution_rejected); pins the scalar branch, the tuple branch, the ceil-and-clamp-to-1 semantics, and resolution= on all four backends.
• Cat 4 HIGH -- non-empty GeometryCollection unpacking is documented at rasterize.py:1995 and implemented by _classify_geometries_loop (line 228) but only the empty-GC case was tested. Pins polygon+point and polygon+line+point collections plus parity across cupy / dask+numpy / dask+cupy.
• Cat 1 MEDIUM -- eager cupy all_touched=True parity vs eager numpy. The existing parity test only covered dask+cupy all_touched, leaving the direct GPU all_touched kernel untested.
• Cat 2 MEDIUM -- int32 dtype with default NaN fill silently casts to the int32-min sentinel. Pin the cast so any future raises-instead switch is visible as a code-review diff.

Test plan

• All 34 new tests pass locally on a CUDA host (numpy + cupy + dask+numpy + dask+cupy).
• Pre-existing 143 passing + 2 skipped tests in test_rasterize.py remain green.
• Mutation against the like= dtype inheritance branch (elif like_dtype is not None: final_dtype = like_dtype) flips the inheritance test red, confirming the new coverage catches a real regression.
• CI runs the cupy / dask+cupy branches on the GPU host (locally exercised but CI re-validates).