test: add xfail for 3d transform on 4d data

[oesteban]

Summary

• add regression test for applying a single 3D transform to a 4D dataset

Testing

• pytest nitransforms/tests/test_resampling.py::test_apply_single_3d_on_4d -q

---

https://chatgpt.com/codex/tasks/task_e_689f54b6c3bc8330a3230878ab5c4454

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.74%. Comparing base (0c2ad9f) to head (1428106).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #282   +/-   ##
=======================================
  Coverage   96.74%   96.74%           
=======================================
  Files          16       16           
  Lines        1994     1994           
  Branches      267      267           
=======================================
  Hits         1929     1929           
  Misses         41       41           
  Partials       24       24           

Flag	Coverage Δ
unittests	96.72% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[effigies]

At present, this could fail because checking the values after applying the transform is wrong. I think a much simpler test is:

Suggested change

	@pytest.mark.xfail(reason="gh-281: applying a single 3D transform to 4D data")
	def test_apply_single_3d_on_4d():
	"""Apply one 3D transform across all timepoints of a 4D dataset."""
	nvols = 5
	data = np.zeros((10, 5, 5, nvols), dtype=np.float32)
	for i in range(nvols):
	data[i + 1, 2, 2, i] = i + 1
	img = nb.Nifti1Image(data, np.eye(4))
	mat = np.eye(4)
	mat[0, 3] = -1.0
	ref = nb.Nifti1Image(np.zeros((10, 5, 5), dtype=np.uint8), np.eye(4))
	xfm = nitl.Affine(mat, reference=ref)
	moved = apply(xfm, img, order=0)
	moved_data = np.asanyarray(moved.dataobj)
	assert moved_data.shape == data.shape
	for i in range(nvols):
	assert moved_data[i + 2, 2, 2, i] == i + 1
	assert moved_data[i + 1, 2, 2, i] == 0
	@pytest.mark.xfail(reason="gh-281: applying a single 3D transform to 4D data", strict=True)
	def test_apply_single_3d_on_4d():
	"""Apply one 3D transform across all timepoints of a 4D dataset."""
	img4d = nb.Nifti1Image(np.zeros((2, 3, 4, 5), np.eye(4))
	ref3d = img4d.slicer[..., 0]
	xfm = nitl.Affine(np.eye(4))
	# Just a smoke test. When this passes, we should validate the values.
	apply(xfm, img, reference=ref3d)

I'm adding strict=True so that the test doesn't just sit around XPASSing when we should start validating values after that point.