2613 saver/loader consistencies

[wyli]

fixes #2613
fixes #1629
fixes #1357
fixes #1713 (by upgrading itk)
fixes #1711

Description

• using the numpy array layout for the get_data of itkreader,pilreader, so that the loader results are consistent
• itkreader's affine is converted to nibabel's convention to be consistent
• python -m tests.test_distributed_sampler still gives messages such as UserWarning: semaphore_tracker: There appear to be 1 leaked semaphores to clean up at shutdown because of itk 5.2.0 semaphore leaks addressed by workaround ac7944e

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

/black

[madil90]

/build

[wyli]

I run a few tutorials/demo with this PR and all work fine. Additional integration tests show that loadimaged with the 'itkreader' or 'nibabelreader' option gives the same results.

MONAI/tests/test_integration_segmentation_3d.py

Lines 254 to 257 in 7835727

	if idx == 1:
	_readers = ("itkreader", "itkreader")
	elif idx == 2:
	_readers = ("itkreader", "nibabelreader")

I think this gives us some confidence in the code. Could you please review @ericspod @rijobro @Nic-Ma? The current itkreader in the dev branch is broken #1711, which means we don't have any basic working loader for DICOM.

[Nic-Ma]

Hi @wyli ,

Thanks for your quick update and test.
I just came back from a short vacation last night and still checking the ITK APIs related to the data loading and thinking about how to develop the saving feature for other formats, will review this PR ASAP.

Thanks.

[Nic-Ma]

Looks good to except for several minor comments.

Thanks.