ENH: Add CIFTI-2 support

[effigies]

This builds on @satra's PR (#126).

It also supersedes #210, although that could be done separately. We should probably release PyBIDS 0.11 once this is working with master.

Still some things to be done, but this gets through run-level models on my test data.

TODO:

• Set up tests with ds003
• Test second-level
• Fix/skip plotting for CIFTI-2
• Figure out model maps with CIFTI-2
• Probably some other bits I haven't thought of yet.

[pep8speaks]

Hello @effigies, Thank you for updating!

• In the file fitlins/interfaces/nistats.py:

Line 177:100: E501 line too long (100 > 99 characters)

• In the file setup.py:

Line 3:1: F401 'os' imported but unused

To test for issues locally, pip install flake8 and then run flake8 fitlins.

Comment last updated at 2020-07-14 23:29:17 UTC

[effigies]

@satra I'm currently able to get this to fit at the run level and do a fixed-effects combination at the subject level. The reporting needs work (subject level shows run 1) and the group level is crashing due to a failure to filter inputs properly, but feel free to give this a try and see if you're getting sensible results at the first/second level.

[codecov-commenter]

Codecov Report

Merging #239 into master will decrease coverage by 5.18%.
The diff coverage is 44.44%.

@@            Coverage Diff             @@
##           master     #239      +/-   ##
==========================================
- Coverage   78.13%   72.95%   -5.19%     
==========================================
  Files          19       19              
  Lines        1084     1198     +114     
  Branches      191      214      +23     
==========================================
+ Hits          847      874      +27     
- Misses        148      227      +79     
- Partials       89       97       +8     

Flag	Coverage Δ
#ds003	72.95% <44.44%> (-5.19%)	⬇️

Impacted Files	Coverage Δ
fitlins/workflows/base.py	64.13% <ø> (ø)
fitlins/interfaces/visualizations.py	61.71% <10.20%> (-32.11%)	⬇️
fitlins/cli/run.py	83.84% <33.33%> (-0.77%)	⬇️
fitlins/interfaces/nistats.py	70.23% <41.79%> (-14.66%)	⬇️
fitlins/__init__.py	76.47% <75.00%> (-1.31%)	⬇️
fitlins/interfaces/bids.py	73.68% <85.71%> (+0.52%)	⬆️
fitlins/interfaces/abstract.py	100.00% <100.00%> (ø)
fitlins/viz/reports.py	90.00% <100.00%> (+0.44%)	⬆️

---

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 ab47bf2...0b6188c. Read the comment docs.

[effigies]

The current version of ds003 that I'm testing on doesn't have usable CIFTI-2 files. I'm planning on re-running, but I'm not sure it's worth holding this up for that.

@adelavega Can you test neuroscout on this branch, or do I need to merge into master first?

[bpinsard]

Thanks for that nice feature!
Does smoothing works on the CIFTI space, prior to computing the maps?

[effigies]

No, I don't have smoothing implemented for CIFTI, yet.