RAG Agent

A Retrieval-Augmented Generation (RAG) agent built with Flask, LangChain, Hugging Face, and Redis.
It retrieves relevant context from your dataset and generates answers using Hugging Face models.

---

🚀 Features

• Flask backend with /chat endpoint
• RAG pipeline (vector search + LLM)
• Redis as vector database
• Configurable data source (data/product.txt)
• Session-based chat

---

📦 Setup Instructions

1. Clone the repository

git clone https:/github.com/Unfathomable-08/Rag-Agent.git
cd Rag-Agent

2. Create a virtual environment

python -m venv venv

3. Activate the virtual environment

Windows

venv\Scripts\activate

Linux/Mac

source venv/bin/activate

4. Install dependencies

pip install -r requirements.txt

5. Configure environment variables

Copy the example file:

cp .env.example .env

Open .env and fill in:

• Hugging Face API Token → Get it here
• Redis Cloud URL → Get it here

Example Redis URL format:

textredis:/default:<your-password>@<your-host>:<port>

6. Prepare data

You can keep or modify the file: textdata/product.txt

7. Build the vector database

python vector_builder.py

8. Start the server

python main.py

The Flask server will start locally.

📡 API Usage

Endpoint textPOST /chat

Request body

json{
  "session_id": "user123",
  "question": "What products do you have?"
}

Example using curl

curl -X POST http:/127.0.0.1:5000/chat \
  -H "Content-Type: application/json" \
  -d '{"session_id": "user123", "question": "What products do you have?"}'

Response

json{
  "answer": "We have a range of products including..."
}

🛠 Tech Stack

• Python 3
• Flask
• LangChain
• Hugging Face Inference API
• Redis (as vector DB)
• FAISS

📖 Notes

""" Replace data/product.txt with your own dataset for custom answers. Ensure Redis Cloud is running and accessible. Hugging Face token must have inference API permissions. """