2420 Add image_only option to several spatial transforms

[Nic-Ma]

Fixes #2420 .

Description

According to user's feedback, this PR added the image_only option for several spatial transforms, same as the LoadImage IO transform.

Status

Ready

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[Nic-Ma]

/black

[Nic-Ma]

/black

[wyli]

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

[Nic-Ma]

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.

[wyli]

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

[Nic-Ma]

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.