Closed ljhearne closed 6 years ago
So it looks like singularity can't build the image properly from docker hub...
If it possible for you to use docker to build the image and then convert it to a singularity container. You would need a machine you can run docker on (like a local laptop).
$ docker build --privileged -t --rm tigrlab/fmriprep_ciftify:
Otherwise, we can try (tomorrow) to put the image up on singularity hub?
Also, I just built a newer version fo the container (with the fixes to some of the ciftify_recon_all options) which I haven't pushed yet (will tomorrow). I would suggest using the version number instead of "latest" to keep track of what version you have.
Thanks for letting me know! We'll update the docs with this issue
Thanks Erin, I'll download docker locally and build the singularity container that way.
I'll report back if I still get the same error.
edit: Seems like it was a docker issue before - I fixed by following the comments here. I'm using Singularity on our HPC and the help arguments are working, so I'm assuming this has fixed the problem! Great!
Hi Erin,
Thanks for the fantastic tool / all the work you have put in.
I'm trying to get ciftify up and running on the local HPC. To do so I created a singularity container using:
singularity pull docker://tigrlab/fmriprep_ciftify:latest
Then to check it was working I open a shell in the container and ask for help on a ciftify function:
Singularity shell --cleanenv <path-to-img>
(within the container):~> ciftify_recon_all --help
However, I get an error when trying to import some pandas modules (below). I have been following the scripts on https://github.com/edickie/bids-on-scinet/blob/master/examples/sbatch_fmriprep_ciftify_p08.sh so it doesn't seem like I need to import anything else other than the container... Let me know if you have any thoughts!