This project demonstrates how to build a local Retrieval-Augmented Generation (RAG) AI agent using Python. The agent runs entirely on your machine and leverages:

• Ollama for open-source LLMs and embeddings
• LangChain for orchestration
• SingleStore as the vector store

By the end we'll have a working Q&A system powered by the local data and models.

Ensure the following are installed and configured:

• Docker (running)
• Python 3.9+
• Ollama installed and working

1. Visit ollama.com and download the installer for your OS
2. Follow the on-screen instructions
3. Download the required models by running:

   ollama serve
   ollama pull llama3
   ollama pull mxbai-embed-large

1. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate  # On Windows use: venv\Scripts\activate

2. Install the required packages:

   pip install -r requirements.txt

The project uses a CSV file (pizza_reviews.csv) containing pizza reviews. The file should have the following columns:

• Title
• Date
• Rating
• Review

The Python script will automatically start a SingleStore instance using Docker. Just make sure Docker is running.

Start the application from the command line:

python3 main.py

Once running, you can ask questions about the pizza reviews. For example:

Enter your question about pizza (or 'exit' to quit): What do people think about the crust?

The system will:

1. Find relevant reviews using vector similarity search
2. Use the LLaMA model to generate an answer based on those reviews

.
├── README.md
├── requirements.txt
├── main.py
└── pizza_reviews.csv

The project uses the following main dependencies:

• langchain
• langchain-ollama
• langchain-singlestore
• singlestoredb
• pandas
• docker

Feel free to experiment with:

• Different LLMs or embedding models via Ollama
• Other datasets
• Custom prompt templates
• Adjusting the number of retrieved reviews (currently set to 2)

This project is licensed under the MIT License - see the LICENSE file for details.

This project is based on the tutorial "Build a Local AI Agent with Python, Ollama, LangChain and SingleStore" by Volodymyr Tkachuk, Engineering Manager at SingleStore. Original tutorial can be found at: SingleStore Blog.