search

HBClab / NiBetaSeries

Nipype implementation of BetaSeries Correlations (Beta)
https://nibetaseries.readthedocs.io
MIT License
32 stars 28 forks source link

Trouble getting singularity image to work #282

Closed sjw42 closed 4 years ago

sjw42 commented 4 years ago

Describe the bug I'm sure this is an issue with our hpc system and not a bug, but I thought that someone may have insight into what is going wrong. I sincerely apologize if submitting here as an issue is not appropriate. I built a singularity container for NiBetaSeries v0.4.2 and got the error pasted below when trying to use it: Traceback (most recent call last): File "/opt/miniconda-latest/bin/nibs", line 8, in sys.exit(main()) File "/opt/miniconda-latest/lib/python3.7/site-packages/nibetaseries/cli/run.py", line 237, in main atlas_img=os.path.abspath(opts.atlas_img), File "/opt/miniconda-latest/lib/python3.7/posixpath.py", line 378, in abspath path = os.fspath(path) TypeError: expected str, bytes or os.PathLike object, not NoneType

To Reproduce I used the command pasted below to point to the container and run the program. I thought that passing the cleanenv flag would avoid conflicts with any environment variables. I also tried unset PYTHONPATH with no luck.

singularity run --cleanenv /gpfs/group/sjw42/default/sjw42_collab/sw/singularity/nibetaseries/nibetaseries-v0.4.2.simg \ nibs -c white_matter \ --participant-label 01 \ -w /gpfs/group/sjw42/default/ASH/SEMA/work \ /gpfs/group/sjw42/default/ASH/SEMA/bids \ fmriprep \ /gpfs/group/sjw42/default/ASH/SEMA/bids/derivatives \ participant

Expected behavior A clear and concise description of what you expected to happen.

Desktop (please complete the following information): Redhat

nibetaseries version v0.4.2

Additional context Thanks for any thoughts or suggestions for troubleshooting, and sorry again if I should be seeking help elsewhere.

jdkent commented 4 years ago

Hi @sjw42,

Thanks for opening an issue, I see the problem arises when one does not specify --atlas-img and/or --atlas-lut in the command. Your use case is one that should work, so I will change that code and release a new version shortly (e.g., today or tomorrow).

jdkent commented 4 years ago

Hi @sjw42,

That particular issue should be fixed with release v0.4.3 which should be available in the next half hour, could you try again with that version and see if you run into the same error?

sjw42 commented 4 years ago

Will do! Thanks very much for fixing so quickly (and for making such a great tool!) I'll give it a test run tomorrow morning and will let you know how it goes.

sjw42 commented 4 years ago

Just tested v0.4.3 and it worked fine without the -a and -l flags. I also tested a second time with an atlas and lookup table added to see what that output looked like and got a few lines of the error pasted below. I noticed the same error in the sample output on the tutorial pages of the readthedocs site, so maybe these errors can just be ignored? Thanks again for your help!

/opt/miniconda-latest/lib/python3.7/site-packages/nibetaseries/interfaces/nilearn.py:130: RuntimeWarning: invalid value encountered in greater n_lines = int(np.sum(connmat > 0) / 2)

jdkent commented 4 years ago

Thanks for trying again!

You can safely ignore that warning, but it is annoying to see, so it would be good to fix.