jonadev95/miniconda-python3-ssl-docker repository overview

⁠miniconda-python3-ssl-docker

A Miniconda⁠ Jupyter⁠ container, secured with SSL.

This image includes a Miniconda installation and automatically generates SSL certificates on startup.

conda and pip are added to the path, wget is installed to grab data from everywhere.

There's also a smaller image available, which is based on miniconda instead of anaconda.

Miniconda is like anaconda, but instead of having a lot of packages preinstalled, only python and conda are present (jupyter will be installed during image building), but all packages which are available in anaconda are available through the conda command.

⁠Usage

I recommend to mount a folder from the docker host, to be able move data from the container to the host and vice versa.

⁠Run

$ docker run -d -p 8888:8888 \
    -v `pwd`/jupyter-mount:/home/conda/jupyter-mount \
    --name miniconda \
    jonadev95/miniconda-python3-ssl-docker

⁠Get the login token

In order to be able to log in you need the token generated by jupyter, so execute

$ docker logs miniconda

and search for the line

[C 21:54:40.449 NotebookApp] 
    
    Copy/paste this URL into your browser when you connect for the first time,
    to login with a token:
        https://0.0.0.0:8888/?token=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Now either

• Copy the whole URL, replace 0.0.0.0 with the docker host IP and follow that link
• Just copy the token, go to https://$HOST_IP:8888 and use the token to login

Both methods will log you in perfectly, but when using the latter method firefox will save the token as a password so you don't need to look it up again

⁠Install additional packages

Jupyter Notebooks got a buil-in terminal, which makes it easy to install additional packages. Just use pip or conda to install whatever you want, or grab datasets with wget.

⁠Note

Root access/sudo is disabled for security reasons, which means you're not able to install packages via apt-get.

⁠TODO

• do not regenerate SSL certificates on container restarts
• use the same token for the jupyter notebook on container restarts