_spatial_norm_callback implementation

[selmanozleyen]

hi @giovp
I am writing this issue to document one of the commits I will make in #696. At this point it will be hard for me to separate that commit in the PR and it will be a small discussion compared to how big the PR will be. I want to discuss this implementation.

moscot/src/moscot/base/problems/problem.py

Lines 615 to 634 in 0892c6b

	# TODO(@giovp): refactor
	@staticmethod
	def _spatial_norm_callback(
	term: Literal["x", "y"],
	adata: AnnData,
	adata_y: Optional[AnnData] = None,
	attr: Optional[Literal["X", "obsp", "obsm", "layers", "uns"]] = None,
	key: Optional[str] = None,
	) -> TaggedArray:
	if term == "y":
	if adata_y is None:
	raise ValueError("When `term` is `y`, `adata_y` cannot be `None`.")
	adata = adata_y
	if attr is None:
	raise ValueError("`attrs` cannot be `None` with this callback.")
	spatial = TaggedArray._extract_data(adata, attr=attr, key=key)
	logger.info(f"Normalizing spatial coordinates of `{term}`.")
	spatial = (spatial - spatial.mean()) / spatial.std()
	return TaggedArray(spatial, tag=Tag.POINT_CLOUD)

First Issue: Typo

first of all I will fix the type from attrs to attr

Second Issue: Standardization

but I also want to ask whether

spatial = (spatial - spatial.mean()) / spatial.std()

is a correct standardization, shouldn't it be

spatial = (spatial - spatial.mean(axis=0,keepdims=True)) / spatial.std(axis=0,keepdims=True)

since we'd like each feature to be gaussian units right? or am I missing something?

[giovp]

hi @selmanozleyen , yes the axis=0 is correct for mean, not for std as it would introduce a distortion in the axis, I've added the fix in #697