/black

/black

I agree with the comment here #2426 (comment) probably we shouldn't do this for spacing and orientation.

Hi @wyli ,

Thanks for your comments.
I think your point is that: even only returning the image, the Spacing and Orientation transforms still need img and affine as input args, so they can't be composed in a chain, right?
I added image_only to them because I think the affine arg of Spacing and Orientation is optional, so if no need to provide affine matrix in some case(use default np.eye()), this image_only can be helpful, and it's non-breaking.
What do you think?

Thanks.

I think it's incorrect for Spacing and Orientation because these have to be in the context of affine/coordinate systems

I think it's incorrect for Spacing and Orientation because these have to be in the context of affine/coordinate systems

OK, I thought Spacing and Orientation can work without input affine because it's optional arg.
Let me remove the image_only arg in #2423.

Thanks.