/docker-notebook-template

Primary LanguageJupyter Notebook

docker-notebook-template

Run jupyter notebook from a Docker container.

Compared to pip/conda environments and VM - (my opinions)

Pros:

  • clean working environment compared to multiple virtual environments
  • container has access to host resources
  • faster than a VM
  • isolated environment gives more realistic production results (doesn't really apply to a notebook, but dependency issues tend to bubble up when you build in a bare environment often)

Cons:

  • extra steps to getting your work changes and results to host (copying files from container, I added some a script for this)
  • slower than running in a virtual environment on your machine

How to use (if you are confused these are basic steps)

  1. Copy your project files to the /src folder
  2. Copy your requirements.txt file to /
  3. Run the run.sh script
    • check the port setup in the docker-compose file if you have other services running - it is 9999 by default
  4. Connect to your jupyter notebook
  5. Customize the copy.sh script to pull your work out of the container when you're done
  6. Stop the container with the stop.sh script
    • you can also remove the container with remove.sh