rodentQC

[eilidhmacnicol]

This PR puts in place the foundation to extend MRIQC to other populations, including rodents.

Still to do:

• Improve segmentation
• Test alternative EPI masking methods
• Finalise Initial head mask parameters for rats
• Improving the reportlets for rodent FOVs (although maybe this requires some more work from the niworkflows end) -> moving to NiWorkflows
• Add a test for rats (WDYT?) -> will open a separate issue for this

For future, it might be worth establishing a dictionary or a set of JSON files that set template choices for each species available in TemplateFlow.

NB: I tried to push to this as a feature branch, rather than straight to master, but I didn't have the credentials.

[pep8speaks]

Hello @eilidhmacnicol, Thank you for updating!

Cheers! There are no style issues detected in this Pull Request. 🍻 To test for issues locally, pip install flake8 and then run flake8 niworkflows.

Comment last updated at 2022-03-25 14:56:36 UTC

[oesteban]

Looks awesome - thanks for this!

[oesteban]

Hey @eilidhmacnicol, I've just cut 22.0.0rc1 and rebased this branch to current master. This is to say, I'm happy to help here and do whatever you need to get this to the finish line.

It feels that the better segmentation, the tests, and the better parameters of the head mask can/should be addressed in separate PRs. The improvements to the reportlets pertain to niworkflows/nirodents.

[eilidhmacnicol]

Hi @oesteban - I've added the option for Atropos segmentation, and it's looking much better now.

I agree that segmentation and headmasking can be improved in later commits when we encounter further test data. Marking as ready to merge now.

[oesteban]

One more for style -- can you run

black mriqc/

?

[eilidhmacnicol]

@oesteban - tests are looking good, but I've just realised an issue: I see the NiWorkflows version has been pinned, but NiRodents and artsBrainExtraction currently uses a feature branch of NiWorkflows so this will need to be updated for a Docker image.

[oesteban]

We will need to finalize spitting out the brain extraction from niworkflows in order to be able to use the new package :(

The alternative is too dirty for an official MRIQC release.

[eilidhmacnicol]

Okay - some very good news. The NiWorkflows pin issue was (thankfully) a false alarm. I ran an image built on mriqc 22.0.0rc1 with the necessary changes from here and nipreps/nirodents#55 built on top and it ran with no issues.

At this stage, the blockers for this PR's merge are:

related to NiRodents

• FIX: Add unlink() with missing_ok for Python 3.7 niworkflows#695 and the subsequent minor release of NiWorkflows
• Cut a new tag for NiRodents on PyPi pinning the new NiWorkflows release (depends on pin: bump niworkflows to 1.5.x nirodents#55)
• Pin the NiRodents version here

related to DiPy

• dipy 1.4.1 breaks NiPype (See I am currently using the latest version of nipype and am running into the following error: nipy/nipype#2402), so unless nipype >= 1.7.1 then dipy <= 1.4.0 for rodents

[eilidhmacnicol]

I think the tests failing are unrelated to the changes made here and might be beyond the scope of this PR. WDYT @oesteban? Is it time to merge?

[oesteban]

I think the tests failing are unrelated to the changes made here and might be beyond the scope of this PR. WDYT @oesteban? Is it time to merge?

You are going to break master, but considering the type of error it is fair to say that the problem is beyond the scope of this PR. So go ahead and merge.

The reason why this is failing is that we probably need to pin jinja2 with some upper bound version. We should do that quickly after your PR so that dependency errors don't start snowballing.

[oesteban]

Actually, you can't merge, I believe. I'm giving you permissions right now - please merge yourself (I believe one should merge their own PRs whenever that is possible/allowed).

[welcome]

Thanks for opening this pull request and congratulations for taking it to the finish line! It looks like this is your first time contributing to MRIQC. 😄
We invite you to list yourself as an MRIQC contributor. To learn more about what that entails and how we credit our contributors, please check out the contributing guidelines. If your name is not already on the list, please insert it, in alphabetical order of (i) lastname and (ii) firstname, into the .maint/CONTRIBUTORS.md file.
Of course, if you want to opt-out this time, there is no problem at all with adding your name later. You will be always welcome to add it in the future whenever you feel it should be listed.