This repository contains a collection of Dockerfiles and supporting files for building various containers for SageMath and its components (GAP, Singular, PARI/GP, ...).
The containers are available on dockerhub.
Each image is tagged with the version of SageMath installed in that image (e.g. 8.0).
Tags with a version like X.Y-N represent revisions to the image itself i.e.
changes that were made to the Dockerfile it was built from. For example while
sagemath:8.0 is the first image to contain SageMath 8.0, sagemath:8.0-1 is
the first revised SageMath 8.0 image.
sagemath/sagemath (roughly 3.6 GB)
This container contains a basic installation of the latest version of SageMath, built from sources on the latest Ubuntu. Commands are run as the Sage user. The SageMath distribution includes several programs like GAP, Singular, PARI/GP, R, ... which are available in the path.
docker pull sagemath/sagemath
or simply continue to the next step.
To run Sage:
docker run -it sagemath/sagemath
Other software included in this image can be run similarly:
docker run -it sagemath/sagemath gap
docker run -it sagemath/sagemath gp # PARI/GP
docker run -it sagemath/sagemath maxima
docker run -it sagemath/sagemath R
docker run -it sagemath/sagemath singular
To run the Jupyter Notebook interface (for Sage, ...):
docker run -p 8888:8888 sagemath/sagemath-jupyter
Alternatively, to run the legacy Sage notebook server:
docker run -p 8080:8080 sagemath/sagemath sage -notebook
You can then connect your web browser to the printed out typically, namely
http://localhost:8888 for the Jupyter notebook and http://localhost:8080 for
the legacy notebook. For the legacy notebook the webbrowser will ask for a
login and password which are respectively admin and sage.
Note Running the sagemath-jupyter container is equivalent to running the
sagemath/sagemath base docker container with the following command:
docker run -p 127.0.0.1:8888:8888 sagemath/sagemath sage -notebook=jupyter --no-browser --ip='*' --port=8888
The --ip option is required by the Jupyter notebook to allow connections to
the notebook through the Docker network.
Prequisites: network access to download Sage (http/https)
docker build --tag="sagemath/sagemath" sagemath
This container is similar to the previous one, except that SageMath is built from the latest unstable release version of Sage, retrieved by cloning the develop branch from github.
TODO: include git-trac
To download and start it:
docker run -it sagemath/sagemath-develop
docker build --tag="sagemath/sagemath-develop" sagemath-develop
If you want to have a container already set up for the Jupyter enviroment, you can use sagemath/sagemath-jupyter. It is based on sagemath/sagemath.
docker run -p 8888:8888 sagemath/sagemath-jupyter
makes the Jupyter notebook accessible via localhost:8888, while
docker run sagemath/sagemath-jupyter
makes it accessible under the container's ip address on port 8888. You can
see the ip address of the container using docker inspect. This is useful if
you want to have more than one notebook server running. Typically this will
be something like:
172.17.0.1
where the fourth field may be incremented depending on the number of running containers on the host.
docker build --tag="sagemath/sagemath-jupyter" sagemath-jupyter
This container, built on top of sagemath-develop, is meant to run instances of the Sage patchbot running securely in a sandbox, to ensure that the untrusted code it fetches and run cannot harm the host machine.
docker run -t --name="patchbot" -d sagemath/sagemath-patchbot
pid=$(docker inspect -f '{{.State.Pid}}' patchbot )
ip=$(docker inspect -f '{{.NetworkSettings.IPAddress}}' patchbot )
trac_ip=$(getent hosts trac.sagemath.org | awk '{ print $1 }')
patchbot_ip=$(getent hosts patchbot.sagemath.org | awk '{ print $1 }')
nsenter -t $pid -n iptables -A FORWARD -s ${ip} -d ${trac_ip} -j ACCEPT
nsenter -t $pid -n iptables -A FORWARD -s ${ip} -d ${patchbot_ip} -j ACCEPT
nsenter -t $pid -n iptables -A FORWARD -s ${ip} -j REJECT --reject-with icmp-host-prohibited
while true; docker run --tty=true --rm=true sagemath/sagemath-patchbot; done
docker build --tag="sagemath/sagemath-patchbot" sagemath-patchbot
sagemath/sagemath with latex, the commonly used (GAP) packages, etc.
Same as sagemath-jupyter, but based on sagemath-fat