wolfram-jupyter

Wolfram offers a free development license for the Wolfram Engine. With this project, you can use it to explore the combination of the Wolfram Language and Jupyter notebooks.

Setup

First, obtain a license file by running scripts/activate.sh.

💡 If you already got a mathpass file, you can just copy it to the Licensing subfolder. However, mathpass files from a direct installation of wolframscript (not in a Docker container) will not work.

$ ./scripts/activate.sh
Please enter "Quit" after logging in using your Wolfram ID and password.

The Wolfram Engine requires one-time activation on this computer.

Visit https://wolfram.com/engine/free-license to get your free license.

Wolfram ID: [email protected]
Password:
Wolfram Engine activated. See https://www.wolfram.com/wolframscript/ for more information.
Wolfram Language 13.2.0 Engine for Linux x86 (64-bit)
Copyright 1988-2022 Wolfram Research, Inc.

In[1]:= Quit

Then, build the Docker image by running scripts/build.sh.

$ ./scripts/build.sh
[+] Building 32.9s (21/21) FINISHED
 => [internal] load build definition from Dockerfile                            0.0s
 => => transferring dockerfile: 1.66kB                                          0.0s
 => [internal] load .dockerignore                                               0.0s
 => => transferring context: 2B                                                 0.0s
[...]
 => => naming to docker.io/library/wolfram-jupyter                              0.0s

Running

Choose a directory you want to work on your notebooks in and supply the path as an argument to scripts/run.sh (this script is generated during the build step).

$ ./scripts/run.sh ~/wolfram-jupyter-notebooks
[I 15:57:53.959 NotebookApp] Writing notebook server cookie secret to /home/you/.local/share/jupyter/runtime/notebook_cookie_secret
[I 15:57:54.108 NotebookApp] Serving notebooks from local directory: /mnt/jupyter
[I 15:57:54.108 NotebookApp] Jupyter Notebook 6.4.10 is running at:
[I 15:57:54.108 NotebookApp] http://2838bb30a2d6:8888/?token=bed22856cdc79ab5b64aa2e9e4eea30b53f5efbbb6347fd0
[I 15:57:54.108 NotebookApp]  or http://127.0.0.1:8888/?token=bed22856cdc79ab5b64aa2e9e4eea30b53f5efbbb6347fd0
[I 15:57:54.108 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
[C 15:57:54.111 NotebookApp] 
    
    To access the notebook, open this file in a browser:
        file:///home/you/.local/share/jupyter/runtime/nbserver-7-open.html
    Or copy and paste one of these URLs:
        http://2838bb30a2d6:8888/?token=bed22856cdc79ab5b64aa2e9e4eea30b53f5efbbb6347fd0
     or http://127.0.0.1:8888/?token=bed22856cdc79ab5b64aa2e9e4eea30b53f5efbbb6347fd0

Note that the file URL and the first web URL output by Jupyter will not work, because they are internal to the Docker container.

💡 The mathpass file has to be mounted and can not be permanently integrated into the Docker image during the build step, because it has an expiration date and the Wolfram Engine will automatically renew it when it is close to expiring.

Development

Use the activation script as described above to obtain a license file. After that, you can start to change stuff and if you want to, use the build script to generate a run script to test your changes. Or manually build an image (see the build script file for a reference build command).

Planned features

(1) Mount folders that store Jupyter settings to make them persistent. Need to find out which folders.

(2) Automatically open the correct URL in a browser to access Jupyter as running outside of Docker would do.