Userful Docker Snippets to Bootstrap Python Jupyter Projects

Here’s a few useful scripts for I use to bootstrap a new project with a dedicated JupyterLab instance using Docker. If you’re on a Mac, copy these files and call it in your terminal.

bootstrap.sh

I use this script to create a new folder with a Dockerfile, a build script, and a start script.

if [ -z ${1} ]; then NOTEBOOK_NAME="my-project"; else NOTEBOOK_NAME=$1; fi

mkdir $NOTEBOOK_NAME
cd $NOTEBOOK_NAME
# export NOTEBOOK_NAME=${PWD##*/} # use the current folder's name as the project name
echo "Creating new project:" $NOTEBOOK_NAME
echo '# Start from a core stack version
FROM jupyter/datascience-notebook:9f9e5ca8fe5a
# Install in the default python3 environment
' > Dockerfile
echo 'docker build --rm -t jupyter/'${NOTEBOOK_NAME}' .' > build.sh
echo 'docker run -it -v /$(pwd):/home/jovyan -p 8888:8888 jupyter/'$NOTEBOOK_NAME':latest start.sh jupyter lab' > run.sh
source build.sh
source run.sh
echo "Finished bootstrap new folder:" $NOTEBOOK_NAME

To run it in terminal you can do sh boostrap.sh my-custom-name where my-custom-name is an optional argument to name your project. After running this the first time, you only have to run sh run.sh afterwards.

Dockerfile

In the bootstrap script above, I’m not installing any extra packages. If you wish to start the image with some packages or requirements.txt instead, you can refer to the recipes contributed by the community here.

# https://jupyter-docker-stacks.readthedocs.io/en/latest/using/recipes.html#using-pip-install-or-conda-install-in-a-child-docker-image
# Start from a core stack version
FROM jupyter/datascience-notebook:9f9e5ca8fe5a
# Install in the default python3 environment
RUN pip install 'opencv-python'

build.sh

This will build the Docker image from the docker file.

docker build --rm -t jupyter/my-project-name .

run.sh

This will start the docker container from the built image. You need to run build first before running this.

Note: the -v /$(pwd):/home/jovyan is important here. It mounts the current working folder to the Docker environment, so any changes you make to the folder will be updated in the docker container and vice versa. If you don’t have this, you won’t be able to call code or assets from your computer, and changes to your files will be lost when you kill the session.

Here I use jupyter lab out of personal preference, but you can replace it with jupyter notebook instead.

docker run -it -v /$(pwd):/home/jovyan -p 8888:8888 jupyter/my-project-name:latest start.sh jupyter lab