Add dtype to ScaleIntensityRange transforms

[Nic-Ma]

Fixes #3234 .

Description

This PR added dtype arg to ScaleIntensityRange transforms, same as other intensity scale transforms which may change data type during computation, especially for int or long input.

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]

/build

[Nic-Ma]

I also fixed the data in / out type issue in histogram_normalize transform.

Thanks.

[Nic-Ma]

/black

[Nic-Ma]

/build

[myron]

Thanks for this!

[Nic-Ma]

/integration-test

[wyli]

I feel in many cases now it's good to use the CastToType to indicate type casting, perhaps this needs more discussion? dtype parameter may be useful to control the memory footprint or float precision of the transform computations.

[Nic-Ma]

I feel in many cases now it's good to use the CastToType to indicate type casting, perhaps this needs more discussion?

Hi @wyli ,

I agree with you, we should use CastToType for most cases. Maybe this PR changed many things that confused you.
Actually, here I didn't add dtype to more transforms, just added to ScaleIntensityRange, because data type may be changed during scaling, just followed the same pattern as ScaleIntensity transform.
All the other changes are to add support for dtype=None.
Could you please help take a look again?

Thanks.

[wyli]

Thanks, we agreed to have this based on the discussions.