TEST: Get data files for Node/Workflow tests

[effigies]

• Get ds000114 files for test_newnode.
• Get fMRIPrep intermediate files for test_newnode_neuro.

Closes #14.

[effigies]

@djarecka Could you tar up the files you're using for testing the surface sampling workflow?

[effigies]

Looks like test_current_wf_2a is timing out. Any thoughts?

[effigies]

I've pulled down the relevant files for surface sampling. I'm getting the following printed in a loop:

_NODES_CHECK INPUT set_xfm_source False
_NODES_CHECK INPUT sampler False
_NODES_CHECK INPUT sampler False
_NODES_CHECK INPUT sampler False
_NODES_CHECK INPUT sampler False

I'm assuming this shouldn't happen dozens of times?

[djarecka]

@effigies - _NODES_CHECK INPUT sampler False can be printed multiple times. Do you still want em to create a tar from my files?

[effigies]

Might as well. Using my own data wasn't working for me.

[djarecka]

@effigies - I kept the directory structure, but included only the files that are used in the test:
https://www.dropbox.com/s/rrxc5kegeuzkbtr/fmriprep_test.tar?dl=0

[effigies]

Thanks. I'll try to have a look tonight. I think if we want to release today, what we should do is settle on the theoretical, environment-variable-based locations of these files, make sure that tests skip on the absence of the data, and just hit publish. Then we can make fixing the tests a 0.1.0 target.

[effigies]

(Also, in meetings all day, so very limited ability to do more than comment.)

[djarecka]

I'll try to look later why test_current_wf_2a is failing for serial. Hardly ever serial fails and others work... Want to also finish my PR

[djarecka]

also I can create a datalad repo from the fmriprep files, so we can test it with travis, but want to concentrate on other things first, will create an issue (#21) to remember

[effigies]

Yeah, we can skip datalad on that for now. I also want to reduce the FreeSurfer files to the minimum necessary to perform sampling.

[djarecka]

I don't know how to recreate problem with test_current_wf_2a..:(

[djarecka]

you called the dir PYDRA_TEST_DATA

[djarecka]

ok, I know what is wrong with this test, you didn't change paths: https://github.com/nipype/pydra/pull/19/files#diff-2f4c9a06e6e1c89b63af0c7ec864cbe5R1659
not sure why travis doesn't return the trait error

I was wrong that it worked for mp/cf/dask, travis didn't even try. Of course, it worked on my laptop ;-)

[codecov]

Codecov Report

Merging #19 into master will increase coverage by 7%.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master      #19   +/-   ##
=======================================
+ Coverage    76.4%   83.41%   +7%     
=======================================
  Files           6        6           
  Lines         856      856           
  Branches      228      228           
=======================================
+ Hits          654      714   +60     
+ Misses        141       84   -57     
+ Partials       61       58    -3

Flag	Coverage Δ
#unittests	83.41% <ø> (+7%)	⬆️

Impacted Files	Coverage Δ
pydra/__about__.py	100% <ø> (ø)	⬆️
pydra/engine/submitter.py	93.6% <0%> (+4.79%)	⬆️
pydra/engine/auxiliary.py	79.67% <0%> (+4.81%)	⬆️
pydra/engine/node.py	80.25% <0%> (+8.75%)	⬆️
pydra/engine/state.py	77.19% <0%> (+17.54%)	⬆️

---

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 e01f7e4...39b9563. Read the comment docs.

[djarecka]

@effigies - are you still working on this. If not, I'll merge it and rest of the test we can do in a new PR. We will have a lot of conflict anyway, since all PRs are working on the same files, so might be better to keep PR relatively small if possible

[effigies]

Sounds good. We can do a separate PR.