Jupyter-my-own-image/README.md

Jupyter Notebook Docker Setup

This repository contains Docker configuration for running Jupyter Notebook with various data science libraries pre-installed.

Features

• Based on Ubuntu Focal (20.04)
• Includes Jupyter Notebook with:
  • Pandas
  • OpenCV
  • Scikit-learn
  • Scikit-image
  • PyTorch
  • Imutils
  • Notebook export capabilities (nbconvert)
• Volume mounted for persistent notebook storage
• Two deployment profiles (build locally or pull from registry)

Directory Structure

.
├── Dockerfile          # Docker image definition
├── docker-compose.yml  # Services configuration
├── requirements.txt    # Python package dependencies
└── notebooks/          # Directory for your Jupyter notebooks (will be created)

Usage

Create notebooks directory

mkdir -p notebooks

Profile 1: Build locally

To build the Docker image locally and run the container:

docker-compose --profile build up

Profile 2: Pull from your registry

To pull the image from your private registry and run it:

REGISTRY_URL=your-registry-url docker-compose --profile registry up

Replace your-registry-url with your actual Docker registry URL.

Accessing Jupyter Notebook

Once the container is running, access Jupyter Notebook by opening a browser and navigating to:

http://localhost:8888

No password is required (for simplicity - you may want to add authentication for production use).

Customization

• To add more Python packages, modify the pip3 install command in the Dockerfile
• To change the Jupyter configuration, modify the command in docker-compose.yml

Notes

• The notebooks directory is mounted as a volume to persist your work
• For security in production environments, consider adding authentication to your Jupyter instance