📚 RAG Chat Bot

A simple Retrieval-Augmented Generation (RAG) chatbot using LangChain, HuggingFace embeddings, and the Groq LLM. It indexes local .txt files and answers questions based only on the content of those files.

---

📦 Features

• Loads .txt documents from a folder
• Splits content into chunks
• Generates vector embeddings using HuggingFace
• Stores embeddings in memory
• Uses Groq LLM for answering questions
• Returns document-based answers (with optional fallback to LLM if needed)

---

🏗️ Project Structure

rag_chat_bot/
│
├── rag_demo.py           # Main script to run the RAG chatbot
├── documents/            # Folder containing your .txt files
│   ├── python_basics.txt
│   ├── machine_learning.txt
│   └── rag_technology.txt
├── .venv/                # Optional: Your virtual environment
└── README.md             # This file

---

⚙️ Setup Instructions

1. Clone the repository

git clone https://github.com/darunnatarajan/rag_chat_bot.git
cd rag_chat_bot

2. Create a virtual environment and activate it

python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

If you don’t have a requirements.txt, you can use:

pip install langchain langchain-huggingface huggingface-hub groq

---

🔑 Environment Variables

Create a .env file or set environment variables manually with your API key for Groq.

export GROQ_API_KEY="your-groq-api-key"

Or on Windows (PowerShell):

$env:GROQ_API_KEY = "your-groq-api-key"

---

🚀 Run the Chat Bot

python rag_demo.py

You will see:

==================================================
RAG System Ready! Ask your questions:
Available topics: Python, Machine Learning, RAG
==================================================

Example prompts:

• What is Python?
• Who created Python?
• What does RAG stand for?

Type 'quit' to exit the chat.

---

🧠 How It Works

1. Document Loading: Loads .txt files from the documents/ folder.

2. Chunking: Breaks each file into manageable pieces.

3. Embedding: Converts text chunks into vectors using HuggingFace embeddings.

4. Vector Store: Stores those vectors in memory for fast retrieval.

5. Querying: When you ask a question, it:

   • Retrieves the most relevant chunks
   • Passes them to the Groq LLM to generate an answer

---

📝 Add Your Own Files

Place your .txt files into the documents/ folder. The bot will automatically load and index them on startup.

---

✅ Example Output

Question: What is Python?
Answer: Python is a high-level programming language known for its simplicity and readability.

---

📌 Notes

• The LLM (Groq) may fall back to its own internal knowledge only if nothing is retrieved — this can be disabled if you want document-only answers.
• You can customize chunk size, embedding model, or LLM settings in the script.

---

📄 License

MIT License.