MKDocs Documentation Sites

Project contains a separated MKDocs docker container. This container has its own Dockerfile. This Dockerfile is similar to the [one] created for Jupyterlab.

Multi-stage Build

Docker allows performing a multi-stage build to minimize the Docker image size.

Docker Base Image

Although Alpine is a great Linux distro for creating a Docker base Image, there are other possible valid options such as debian slim (debian:stable-slim) or debian buster slim (debian:buster-slim). These base images are light and allow using very well known tools such as apt-get.

FROM python:3.8-slim-buster AS compile-image

There are already python base images such as python:3.8-slim-buster. Having Python installed, it makes unnecessary to install Miniconda distribution and it is possible to specifically install a Python Package Manager and a Python Virtual Environment Manager.

Python Package Manager

Similarly to Jupyterlab, this project uses both Conda and Mamba package manager. Conda is a package manager.

environment.yml

- nbconvert 5.4.1
- nbextensions 0.5.1
- Pandoc 1.19.2.1
- MiKTeX 2.9

# Install the package as normal:
COPY environment.yml .
RUN conda env create -f environment.yml

Install Mamba

# Python packages
RUN conda config --add channels conda-forge && \
    conda install --yes --quiet \
    mamba && conda clean -yt

Working Directory

Add MKDocks Plugins

Add ToC Sidebar Plugin

Add mkdocs toc sidebar plugin mkdocs Table of Contents sidebar plugin

Tasks

• Add mkdocs toc sidebar plugin

References

• Pretty Little Python Secrets - EP 1 - Installing Python Tools/ Libraries the Right Way - Marcello Salvati
• Pretty Little Python Secrets - EP 2 - Python Development & Packaging as Beautiful as a Poem - Marcello Salvati
• Conda Myths and misconceptions