- dependencies:/ scripts for installing common neuroimaging dependencies (run install_deps.sh)
- apps/: BIDS-Apps for processing & analysis (may move this to another repo in future)
- docker/: standardized environments for development and apps (may be deprecated by singularity soon)
- docker_templates/: collection of useful Dockerfile templates from outside our group
The install_deps.sh script will install each script in the dependencies folder, given a target path. Note that many of the scripts require a Debian-based linux distribution, has been only tested on Xenial Ubuntu. The Singularity file automates a singularity-hub build for khanlab/core, and call also be used from the command-line, e.g.:
singularity create -s 50000 khanlab.img
sudo singularity bootstrap khanlab.img Singularity
To pull the built image from singularity-hub, use (not yet tested):
singularity pull shub://khanlab/core:latest
-
Building folders with Dockerfiles can be built with make make build make run
-
Docker Hub organization for khanlab has been created, admin users should be able to: make push
khanlab/core: Nipype, neurodebian (FSL/AFNI), niftyreg
khanlab/vasst-dev: Sets up our vasst-dev repository
To do (Docker): -freesurfer versions -ANTS/C3D -camino -NODDI/DKE (MCR) -interactive version (itksnap, fslview fsleyes, 3d slicer, trackvis)?
Running X from OSX Docker:
First run XQuartz -> preferences, security -> allow connections from network clients, then run from a terminal:
ip=$(ifconfig en0 | grep inet | awk '$1=="inet" {print $2}')
xhost + $ip
docker run -ti -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$ip:0 --rm -v ~/work/scratch/nipype_tutorial/:/output miykael/nipype_course ipython