Add kwarg to not check spatial consistency

[fepegar]

Related to #734.

Description
The new kwarg allows to avoid checks for spatial consistency across images within a subject. This is useful when we are certain that the images are in the same physical space but different voxel space or different orientation. See the @Spenhouet's issue for more information:

• RandomAffine and RandomElasticDeformation on differently sized and oriented images #734

The recommended usage in this case is

RandomAffine(center='origin', check_shape=False)

Checklist

• I have read the CONTRIBUTING docs and have a developer setup (especially important are pre-commitand pytest)
• Non-breaking change (would not break existing functionality)
• Breaking change (would cause existing functionality to change)
• Tests added or modified to cover the changes
• Integration tests passed locally by running pytest
• In-line docstrings updated
• Documentation updated, tested running make html inside the docs/ folder
• This pull request is ready to be reviewed
• If the PR is ready and there are multiple commits, I have squashed them and force-pushed

[codecov]

Codecov Report

Merging #735 (16c0722) into main (b28cdc0) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main     #735   +/-   ##
=======================================
  Coverage   93.80%   93.80%           
=======================================
  Files          72       72           
  Lines        4440     4442    +2     
=======================================
+ Hits         4165     4167    +2     
  Misses        275      275           

Impacted Files	Coverage Δ
...o/transforms/augmentation/spatial/random_affine.py	96.34% <100.00%> (+0.04%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b28cdc0...16c0722. Read the comment docs.

[fepegar]

@Spenhouet could you please take a look? I used this code on your data and it worked fine:

from pathlib import Path
import torchio as tio

ims_dir = Path('/tmp/example_images')
subject = tio.Subject(
    im=tio.ScalarImage(ims_dir / 't1.nii.gz'),
    lm=tio.LabelMap(ims_dir / 'thalamic_reorient.nrrd'),
)
transform = tio.RandomAffine(center='origin', check_shape=False)
transformed_subject = transform(subject)
transformed_subject.im.save('/tmp/t1_tr.nii.gz')
transformed_subject.lm.save('/tmp/t1_thalamus_seg.nii.gz')

thalamic_reorient.nrrd is the thalamus label map reoriented into some random orientation, to make sure the transform is robust to different orientations.

[Spenhouet]

@fepegar I tested it with more labels with different resolution and shape.
It works nicely! Thanks a lot!

[fepegar]

No worries! Do you mind if I add you to the list of contributors?

[fepegar]

I'll take the upvote as a yes!

@all-contributors please add @Spenhouet for ideas

[allcontributors]

@fepegar

I've put up a pull request to add @Spenhouet! 🎉