Docker containers for the astro-informatics software stack:
- SSHT: Spin spherical harmonic transforms
- so3: Wigner transforms
- S2LET: Fast wavelets on the sphere
- massmappy: Mapping dark matter on the celestial sphere
This repository contains several versions of the stack:
-
mininal: Based on Ubuntu 16.04, the astro-informatics stack is installed in
/opt/astroinformaticsand the python modules are installed in the python environment. This can act as a clean base image. -
jupyter: Based on the jupyter/scipy-notebook, contains the same software as the minimal image, but hosts a jupyter server.
Obviously, you need to install Docker on your system, follow this link.
The following command will download the live jupyter notebook version of the stack:
docker run -it --rm -p 8888:8888 -P eiffl/astroinformatics-jupyter
Copy paste the notebook url shown in your terminal to your browser, it should
look something like: http://localhost:8888/?token=...
Feel free to try out the examples in the massmappy folder.
Note however that you cannot save any modifications you make if you start Docker with this simple command. To run a Docker image with a mounted local folder to save your work:
docker run -it --rm -v [path to local folder]:/home/jovyan/work -p 8888:8888 eiffl/astroinformatics-jupyter
You can then save files within the /home/jovyan/work inside the image.
This is the preferred option, it uses dockerhub to automatically build the images. See this link to see how to set it up.
Just go to the sub-directory of the specific image you want to build, and run:
docker build -t='eiffl/astroinformatics-jupyter' .
where the -t option here is used to set a specific tag for the image, should
be changed based on what version you are building.
To push the built image to docker hub:
docker push 'eiffl/astroinformatics-jupyter'
To run this command, you must first login to your dockerhub account with docker login.