add support for the get_transform analog to set_transform

[pattonw]

given a set of scale and translation transforms, defined on both the group and individual individual images, each image will have a final scale and translation. These can now be conveniently retrieved with a little helper function that fetches all appropriate translations and applies them consecutively.

small test added to cover changes.

[ziw-liu]

Thanks for your contribution! I've left my comments inline.

[ziw-liu]

Re: CI failure: please refer to the contributing guide.

[pattonw]

passed the precommit checks 👍

[JoOkuma]

@ziw-liu, If get_effective_translation[or scale] returns a single transformation object would it be more friendly to return a tuple of floats or an array-like object?

@pattonw, I agree with the other comments from @ziw-liu

[ziw-liu]

If get_effective_translation[or scale] returns a single transformation object would it be more friendly to return a tuple of floats or an array-like object?

If the dataset does not have all 5 axes, what should the length of the returned sequence be? I'm guessing that @pattonw's use cases are mostly 3D instead of 5D?

Edit: I remembered it wrong

[pattonw]

If get_effective_translation[or scale] returns a single transformation object would it be more friendly to return a tuple of floats or an array-like object?

If the dataset does not have all 5 axes, what should the length of the returned sequence be? I'm guessing that @pattonw's use cases are mostly 3D instead of 5D?

Isn't it as simple as returning the .scale or .translation property of the TransformationMeta? Isn't the length always equal to the number of dimensions in the array?

[ziw-liu]

Isn't it as simple as returning the .scale or .translation property of the TransformationMeta? Isn't the length always equal to the number of dimensions in the array?

You're right! I mistakenly thought that Position.scale behaves differently (I was thinking of this one).

[pattonw]

Separated into 2 unit tests, return plain list of floats instead of TransformationMeta classes, and removed the error I introduced in scale property to fetch the effective scale of the first dataset instead of just the root.

[pattonw]

fixed test docstring and fixed case that was failing testing

[ziw-liu]

Looks great! Thanks @pattonw for your contribution!

[ziw-liu]

Ready to merge if @JoOkuma doen't have further comments.

[JoOkuma]

@ziw-liu, I don't have any additional comments.
Thanks for the contribution @pattonw