Skip no-op astype in _delayed_read_window (#1624)

[brendancol]

Closes #1624.

Summary

• Gate the per-chunk arr.astype(target_dtype) in _delayed_read_window on a real dtype mismatch.
• After Fix read_geotiff_dask losing nodata mask in non-first chunks (#1597) #1601 widened the call site to always pass target_dtype, every dask chunk ran astype even when arr.dtype == target_dtype. numpy.astype defaults to copy=True and so allocated a chunk-sized same-dtype buffer and memcpy on every read.
• The read_geotiff_dask drops nodata mask when sentinel only appears in non-first chunks #1597 mask path still promotes uint to float64 inline, so the contract that every chunk lands in the dask-declared dtype is unchanged.

Test plan

• New regression test test_dask_no_op_astype_1624.py wraps read_to_array so it returns an ndarray subclass that records astype calls, then asserts no same-dtype astype survives the per-chunk return path.
• Test fails on main with the expected error message; passes after the fix.
• Existing test_dask_int_nodata_chunks_1597.py (the regression suite from Fix read_geotiff_dask losing nodata mask in non-first chunks (#1597) #1601) still passes -- the int-with-sentinel promotion is intact.
• Full xrspatial/geotiff/tests/ suite passes apart from three pre-existing TestPalette failures on main that are unrelated to this change.

[Copilot]

Pull request overview

This PR addresses a performance regression in the GeoTIFF dask read path by avoiding redundant per-chunk astype() calls when the chunk dtype already matches the dask-declared dtype, while preserving the existing integer-nodata promotion behavior introduced for #1597/#1601.

Changes:

• Skip the per-chunk arr.astype(target_dtype) in _delayed_read_window when arr.dtype == target_dtype to avoid unnecessary chunk-sized allocations/copies.
• Add a regression test that instruments the per-chunk read_to_array return value to detect same-dtype astype() calls.
• Add coverage to ensure the uint16 nodata-sentinel path still promotes to float64.

Reviewed changes

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

File	Description
xrspatial/geotiff/__init__.py	Gates per-chunk astype() on a real dtype mismatch to avoid no-op copy costs.
xrspatial/geotiff/tests/test_dask_no_op_astype_1624.py	Adds regression tests to confirm no-op casts are skipped while nodata promotion remains intact.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[brendancol]

Addressed both Copilot review comments in 4c66bb1:

1. test_float32_chunks_avoid_redundant_copy — deleted. The wrapper around _delayed_read_window never recorded any astype calls, the same_dtype_casts list was unused, and the only assertion was on output dtype. test_astype_skipped_when_dtypes_match already covers the same case correctly via an ndarray subclass that records astype invocations. Confirmed it fails on the pre-fix code with AssertionError: Same-dtype astype still runs per chunk (dtype=float32, calls=[dtype('float32')]) and passes after the fix.

2. test_uint16_mask_path_still_promotes — the previously-unused arr is now exercised: the test asserts that pixels equal to the 65535 sentinel become NaN and every other pixel matches arr.astype(float64) byte-for-byte. Catches mask-branch regressions instead of just dtype drift.

pyflakes clean. All three remaining tests pass.