RF: Redirect nipype.workflows to niflow.nipype1.workflows

[effigies]

Summary

Imports from niflow-nipype1-workflows, adding the available modules to sys.modules under their old names, so that code that currently imports via import nipype.workflows.X or from nipype.workflows.Y import Z should continue to work.

Closes #2991.

Acknowledgment

• (Mandatory) I acknowledge that this contribution will be available under the Apache 2 license.

[codecov]

Codecov Report

Merging #3067 into master will increase coverage by 0.48%.
The diff coverage is 54.28%.

@@            Coverage Diff             @@
##           master    #3067      +/-   ##
==========================================
+ Coverage    67.7%   68.19%   +0.48%     
==========================================
  Files         344      297      -47     
  Lines       44112    39775    -4337     
  Branches     5563     5211     -352     
==========================================
- Hits        29866    27124    -2742     
+ Misses      13481    11941    -1540     
+ Partials      765      710      -55

Flag	Coverage Δ
#smoketests	51.78% <48.57%> (+1.48%)	⬆️
#unittests	65.44% <54.28%> (+0.32%)	⬆️

Impacted Files	Coverage Δ
nipype/interfaces/fsl/epi.py	63.22% <ø> (ø)	⬆️
nipype/interfaces/fsl/dti.py	65.79% <ø> (ø)	⬆️
nipype/workflows/__init__.py	100% <100%> (ø)
nipype/interfaces/mrtrix/convert.py	17.85% <11.11%> (-1.14%)	⬇️
nipype/pipeline/engine/workflows.py	76.93% <0%> (-2.03%)	⬇️
nipype/interfaces/freesurfer/base.py	78.81% <0%> (-1.7%)	⬇️
nipype/interfaces/fsl/base.py	86.66% <0%> (-1.12%)	⬇️
nipype/interfaces/fsl/preprocess.py	81.76% <0%> (-0.29%)	⬇️
nipype/interfaces/dipy/tracks.py	32.22% <0%> (+0.55%)	⬆️
... and 1 more

---

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 1ccd70b...5c9e571. Read the comment docs.

[effigies]

I feel like this would be a good one to get in for RC1, as it has the chance to break things for people, so finding those before the final release would be good.

[satra]

i like it and a question.

how are we dealing with circularity? we are installing nipype and these niflows require nipype.

[effigies]

Niflows depend on nipype, nipype will detect them at runtime, so there isn't an installation circularity.

For the purpose of testing on docker, it installs the latest release and then installs the current source on top of that. We could just as easily post-install the niflows.

[mgxd]

LGTM, would be nice to include an extra_requires that automatically installs niflow.nipype1.workflows but this might be circular

[effigies]

Yeah, that would be circular, though it's possible pip is more relaxed for extras. I also don't want to encourage people to continue using this API, so making it a little difficult is fine with me.