Add parameter unpacking for transforms

[Spenhouet]

Signed-off-by: Sebastian Penhouet sebastian.penhouet@airamed.de

Fixes #2426

Description

Transforms can return any number of values. This pull request enables that these return values are passed into the next transform in a Compose as parameters (unpacked arguments).
This is especially useful in the case where the input and output of transforms always matches like in the case of always passing along the image array and the meta data dict.

Status

Ready

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• 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.

[wyli]

/integration-test

[Spenhouet]

mypy issues:

monai/engines/utils.py:144:16: error: Incompatible types in assignment (expression has type "Union[List[Dict[Any, Any]], Dict[Any, Any]]", variable has type "Dict[Any, Any]")  [assignment]
monai/data/image_dataset.py:127:38: error: "None" object is not iterable  [misc]
monai/data/dataset.py:1062:20: error: Incompatible types in assignment (expression has type "Union[List[Dict[str, Any]], Dict[str, Any]]", variable has type "Dict[str, Any]")  [assignment]

Strange that I did not get them on my local development. I will look into it.

[Nic-Ma]

Hi @Spenhouet ,

Thanks for your quick PR, I think mypy usually has big difference between versions, please update your local mypy to 0.902 and try again.

Thanks.

[wyli]

/integration-test

[Spenhouet]

@wyli @Nic-Ma I fixed the mypy complains (and also the usage in ImageDataset).

[wyli]

thanks!

[Spenhouet]

Btw. I did set the unpack flag to False by default to not break any current behavior.

[Nic-Ma]

Hi @Spenhouet ,

Thanks very much for your quick PR!
I think this assert is not correct here, some MONAI sampling transforms return a list of dictionaries instead of a single dict.
I will submit a quick PR to fix it soon, CC @wyli @ericspod @rijobro .
Others of this PR look good to me.

Thanks.