Closed DVSneuro closed 1 week ago
This happens because pyAFQ downloads its templates/waypoints/etc. on runtime, but may be addressed by the integration of pyAFQ download
into the image in PennLINC/qsirecon#17.
@36000: does pyAFQ download
also run the templateflow code to download the MNI template? That might still be an issue in this scenario.
We now include the pyafq templates in the docker image. Closing for now
Summary
We were trying to use the
mrtrix_multishell_msmt_pyafq_tractometry
recon specification on our local computing cluster and we are getting network connection errors (see below). The computing nodes don't connect out to the internet, so I'm wondering if there are files we can download to prevent QSIprep from trying to retrieve files from the internet?Additional details
What were you trying to do?
We were trying to run
mrtrix_multishell_msmt_pyafq_tractometry
.What did you expect to happen?
We expected the program to run, but it seems to require a network connection to retrieve other files.
What actually happened?
Network errors:
OSError: [Errno 101] Network is unreachable
The full contents of crash file is pasted below:Reproducing the bug
Here's the code to reproduce the bug:
Note that the dataset is BIDS valid. We had to use
--skip_bids_validation
because we were getting issues similar to this Neurostars post with a large dataset (https://neurostars.org/t/clarifying-file-reading-in-bids-validator-vs-fmriprep/5195/8).In case it's helpful, these data are on OpenNeuro: https://openneuro.org/datasets/ds005123/versions/1.0.1
Thanks for any help!
Best wishes, David