(fix): raise error on non-integer floating types in iterables

[ilan-gold]

• Closes Unexpected Behavior: AnnData Allows Indexing with Float Arrays Without Error #1735
• Tests added
• Release note added (or unnecessary)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.53%. Comparing base (6ae9de8) to head (150d8aa).
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1746      +/-   ##
==========================================
- Coverage   86.99%   84.53%   -2.47%     
==========================================
  Files          40       40              
  Lines        6053     6057       +4     
==========================================
- Hits         5266     5120     -146     
- Misses        787      937     +150     

Files with missing lines	Coverage Δ
src/anndata/_core/index.py	95.09% <100.00%> (+0.12%)	⬆️

... and 11 files with indirect coverage changes

[flying-sheep]

Please fix the comparison so it doesn’t only work with float64:

>>> np.dtype(np.float64) == float
True
>>> np.dtype(np.float32) == float
False

(and parametrize the tests please)

[flying-sheep]

Suggested change

	@pytest.mark.parametrize(
	"index",
	[
	pytest.param(sparse.csr_matrix(np.random.random((1, 10))), id="sparse"),
	pytest.param([1.2, 3.4], id="list"),
	*(
	pytest.param(np.array([1.2, 2.3], dtype=dtype), id=f"ndarray-{dtype}")
	for dtype in [np.float32, np.float64]
	),
	],
	)
	def test_index_float_sequence_raises_error(index):
	with pytest.raises(IndexError, match=r"has floating point values"):
	gen_adata((10, 10))[index]
	@pytest.mark.parametrize(
	"index",
	[
	pytest.param(sparse.csr_matrix(np.random.random((1, 10))), id="sparse"),
	pytest.param([1.2, 3.4], id="list"),
	pytest.param(np.array([1.2, 2.3], dtype=dtype), id=f"ndarray-{dtype}")
	],
	)
	@pytest.mark.parametrize("dtype", [np.float32, np.float64])
	def test_index_float_sequence_raises_error(index, dtype):
	index = index.astype(dtype)
	with pytest.raises(IndexError, match=r"has floating point values"):
	gen_adata((10, 10))[index]

[ilan-gold]

@flying-sheep is the goal here to ensure dtype is checked on sparse matrices as well? I think it's enough to check the numpy arrays. Also we can't astype a list. Also dtype is not accessible outside of its parametrization/the test.