Docker image containing a few kernels preinstalled:
- C++
- Clojure
- Haskell
- Java - appears to only work in Jupyter Notebook, not Jupyter Lab
- JavaScript
- Python 2
- Python 3
- Ruby
- Scala (via Apache Toree)
(No, it does not contain Tensorflow and the likes).
You can run it with:
docker run -ti -p 8888:8888 kubuszok/jupyter-pack:latest
if you need to use Jupyter Notebook instead (e.g. Java kernel does not work with Jupyter Lab) use:
docker run -ti -p 8888:8888 -e LEGACY=1 kubuszok/jupyter-pack:latest
Out of the box it uses Java 8. If you want to switch to Java 9 (for Java notebook only, as the rest of kernels works poorly with JRE 9) use:
docker run -ti -p 8888:8888 -e JAVA=java-9-openjdk kubuszok/jupyter-pack:latest
This change is preserved between runs. To restore Java 8 (required by e.g. Scala):
docker run -ti -p 8888:8888 -e JAVA=java-8-openjdk kubuszok/jupyter-pack:latest
Some kernels (esp. Haskell) might take a while to start the first time you use them.
Image is based on Antergos - Antergos is based on Arch and this way I have access to both repositories with lastest versions as well as AURs which adds community support for many packages, that I would otherwise had to install manually.