RAG Workshop

A hands-on workshop for learning Retrieval-Augmented Generation (RAG) through implementing chunking and retrieval strategies. Build a production-ready RAG system in 4 hours.

Quick Start

MacOS

# Verify Python 3.12
python3 --version  # Should show Python 3.12.x

# Create and activate virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install dependencies
pip3 install -r requirements.txt
pip3 install -e

# Verify setup
python3 scripts/verify_setup.py

# Run the workshop application
python3 -m nicegui_app.main

Linux/WSL

Inside WSL - make sure you are within the home directory

cd ~

# update + install python3.12
sudo apt update
sudo apt install -y python3.12-full
# Verify Python 3.12
python3 --version  # Should show Python 3.12.x

# Create and activate virtual environment
python3 -m venv .venv
source .venv/bin/activate

# Install dependencies
pip3 install -r requirements.txt
pip3 install -e

# Verify setup
python3 scripts/verify_setup.py

# Run the workshop application
python3 -m nicegui_app.main

Windows (Non-WSL)

TBD

# Verify Python 3.12
python --version  # Should show Python 3.12.x

# Create and activate virtual environment
python -m venv .venv
.venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Verify setup
python scripts\verify_setup.py

# Run the workshop application
python -m nicegui_app.main

(Optional) Using uv (same instructions for all operating systems)

if not using UV, skip

# install uv; use homebrew, curl or whatever works on Windows

uv sync
uv run python scripts/verify_setup.py
uv run python -m nicegui_app.main

Environment Variables

Set these in a .env file or export them:

export GOOGLE_API_KEY="your-key-here"           # For Google AI Studio
export GOOGLE_CLOUD_PROJECT="your-project"      # For Vertex AI
export GOOGLE_CLOUD_LOCATION="us-central1"      # For Vertex AI
export OPENAI_API_KEY="your-key-here"           # For OpenAI
export ANTHROPIC_API_KEY="your-key-here"        # For Anthropic

# LiteLLM Proxy Configuration (if using a proxy)
export OPENAI_API_BASE="https://your-litellm-proxy:port"  # LiteLLM proxy endpoint
export SSL_VERIFY="false"                        # Set to "false" to disable SSL verification

LiteLLM Support

The workshop now supports LiteLLM as a proxy for accessing multiple LLM providers through a unified OpenAI-compatible API. This is useful for:

• Corporate environments with centralized LLM access
• Cost tracking and rate limiting across teams
• Environments requiring SSL inspection bypass

Configuration

Models are configured via Hydra configs in configs/models/. To use LiteLLM:

1. Set OPENAI_API_BASE to your LiteLLM proxy endpoint
2. Configure models using the litellm: prefix (e.g., litellm:global.anthropic.claude-sonnet-4-5-20250929-v1:0)
3. If using self-signed certificates, set SSL_VERIFY="false" environment variable

Exporting WhatsApp Chat

Follow instructions for your device. Place the exported text file at chats/default_chat.txt or use the included chats/example_chat.txt.

---

Workshop Guide

See the Participant Guide for the full hands-on walkthrough with step-by-step instructions, exercises, and troubleshooting.

---

License

MIT