This should be everything needed to run the notebooks underneath and also work as a decent-ish template for anyone trying to do the same.
When adding packages to the system that you want to "automagically" be installed then add them to the requirements.txt and re-launch with:
docker-compose up --build
Save custom settings in the .jupyter folder as they will be automatically copied in with the volume mounting. You can siphon these custom settings from the "Advanced Settings" pane in Jupyter Lab, then put the files in the same place they exist on the lab machine.
Don't put your secrets in notebooks you're saving. That's a horrible idea. Instead use the system of the python-dotenv package where a .env file is loaded into your Jupyter Notebooks automatically. See contained notebooks for an example. There are certainly more elegant ways to do this with Vault or whatever but nothing this simple.
Before committing your notebooks, be careful about what data is contained within them -- particularly if you are using a public repository. All of your work, including any sensitive data during the course of your investigations, will be a part of the notebooks. You can clear that using the following command:
jupyter nbconvert --clear-output --inplace notebook.ipynb
The nbconvert command also recognizes globbing so a simple *.ipynb can clear
all of your notebooks.
The docker-compose setup automatically loads in /data in the current working directory to /data in the /data directory of the docker container. Therefore if you need additional datasets (ex. historical ASN data for pyasn) place them in that directory to have them become available.