b08x/docker-stacks

Docker images containing IRuby Notebook applications

Docker Stacks: Jupyter Python Ruby

The RubyData Jupyter Notebook Docker Stack is designed to provide a seamless environment for data processing and analysis using Ruby and Jupyter notebooks. This stack includes a curated selection of essential Ruby gems to facilitate efficient data extraction, transformation, and loading (ETL) processes.

RubyData Docker stacks are a set of ready-to-run Docker images containing Jupyter and data tools for Rubyists.

This repository is based on rubydata/docker-stacks.

Quick Start

Minimal Image

LLM Image

LlamaStuff Image

NLP Image

minimal --> llamaindex --> nlp

minimal -> audio

Contains gems necessary to build an application working with Large Language Models (LLMs).

Question Answering System: Building a system that can answer questions based on a knowledge base stored in Chroma.

Text Summarization or Generation Tool: Leveraging LLMs to summarize larger pieces of text or generate new content based on prompts.

Intelligent Search: Creating a search interface allowing the LLMs to understand natural language queries and find suitable matches within the knowledge base.

Building Images

Using the Rakefile

The Rakefile defines the following key tasks:

• pull/base_image/{image} - Pulls the base image for each supported image (minimal, datascience, etc.)
• build/{image} - Builds the Docker image, depends on pulling the base image
• tag/{image} - Tags the built image with the Git commit sha
• push/{image} - Pushes the image to the Docker registry

It also defines composite tasks:

• build-all - Builds all images in parallel
• tag-all - Tags all images
• push-all - Pushes all images

These tasks leverage variables defined at the top like ALL_IMAGES, BASE_IMAGES, OWNER to programmatically handle building, tagging and pushing multiple images.

Individual tasks are invoked for each image, while the composite tasks iterate over ALL_IMAGES to perform the action for all of them.

This Rakefile automation handles common repetitive Docker image operations and allows building, tagging and pushing multiple images with simple rake commands.

License

MIT License