Improve error message when scipy is missing for NDPointIndex

[Sakshee-D]

This PR improves the error message raised when using NDPointIndex without scipy installed.

Previously, the missing dependency error was raised indirectly during index creation. This change wraps the scipy import in ScipyKDTreeAdapter and raises a more actionable ImportError explaining that scipy is required.

Manual testing

• Verified behavior when scipy is not installed
• Verified no behavior change when scipy is installed

Related to #11047

[welcome]

Thank you for opening this pull request! It may take us a few days to respond here, so thank you for being patient.
If you have questions, some answers may be found in our contributing guidelines.

[keewis]

thanks for the PR, @Sakshee-D.

This looks good to me, but we might need a test? This might require more changes, though: right now, the dedicated test module (test_ndpoint_index.py) is skipped if scipy is not available. Are you up for that? If not, I can make the changes for you.

[Sakshee-D]

Thanks for the suggestion.
Since test_nd_point_index.py is skipped entirely when SciPy isn’t available, I will add a small test in a new test file under tests/ (without importorskip("scipy")). The test will use monkeypatch to simulate SciPy being missing and display the improved error message when NDPointIndex is constructed.

I’ll include this in the PR so you can review it there and suggest any changes if needed.

[keewis]

instead of the monkey patching it should be enough to check on the environments that don't install scipy (also needs from xarray.tests import has_scipy):

Suggested change

	def test_ndpointindex_missing_scipy(monkeypatch):
	@pytest.mark.skipif(has_scipy)
	def test_ndpointindex_missing_scipy():

[Sakshee-D]

Thanks for pointing that out .
I missed this earlier. I will update the test to rely on the no-SciPy CI environments and call ScipyKDTreeAdapter directly to keep it minimal.