Closes: #3530

Well, except it doesn't really fix the issue.

Without the : https://ppb.chymera.eu/b4e05b.log
With the : https://ppb.chymera.eu/9e1476.log

Basically instead of AttributeError: module 'networkx' has no attribute 'to_scipy_sparse_matrix' we now get IndexError: tuple index out of range... which, dunno, I guess solves the issue in the most narrow and unhelpful sense.

It seems our parameter usage is still correct, at least per the documentation of the new function... Any ideas?

Codecov Report

Base: 63.63% // Head: 48.26% // Decreases project coverage by -15.38%⚠️

Coverage data is based on head (8c0fe16) compared to base (7382b3d).
coverage: 4.00% of modified lines in pull request are covered.

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #3535       +/-   ##
===========================================
- Coverage   63.63%   48.26%   -15.38%     
===========================================
  Files         309      309               
  Lines       40873    40884       +11     
  Branches     5381     5381               
===========================================
- Hits        26011    19734     -6277     
- Misses      13838    21020     +7182     
+ Partials     1024      130      -894     

Impacted Files	Coverage Δ
nipype/interfaces/cmtk/cmtk.py	15.97% <0.00%> (-1.92%)	⬇️
nipype/interfaces/cmtk/nx.py	17.61% <0.00%> (-0.19%)	⬇️
nipype/pipeline/plugins/base.py	12.93% <0.00%> (-45.29%)	⬇️
nipype/interfaces/cmtk/nbs.py	31.91% <14.28%> (-1.06%)	⬇️
nipype/utils/imagemanip.py	16.66% <0.00%> (-83.34%)	⬇️
nipype/interfaces/dipy/stats.py	0.00% <0.00%> (-83.34%)	⬇️
nipype/interfaces/dipy/registration.py	0.00% <0.00%> (-83.34%)	⬇️
nipype/utils/functions.py	22.72% <0.00%> (-77.28%)	⬇️
nipype/utils/subprocess.py	14.07% <0.00%> (-72.60%)	⬇️
nipype/pipeline/plugins/linear.py	12.50% <0.00%> (-70.84%)	⬇️
... and 114 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

Caught a few more live ones, but there are still 20 test failures, at least they appear to be nipype-internal now:

https://ppb.chymera.eu/fdae94.log

The errors seem likely to stem from previously having 2D matrices with a single row (so you index by [0, idx]) and now having 1D vectors. We'll need to thoroughly purge matrices and switch to ndarrays.

Superseded by #3538