RAG-based Chatbot with Streamlit

A Retrieval-Augmented Generation (RAG) chatbot built with Streamlit and open-source models.

Features

• Retrieval-Augmented Generation (RAG) for context-aware responses.
• DeepSeek R1 as the LLM for generating answers.
• ChromaDB as the vector store for efficient document retrieval.
• Nomic Embed Text for embedding knowledge into a searchable format.
• Streamlit UI for easy interaction with the AI chatbot.

Project Structure

.
├── app.py
├── agents.py            # Main Streamlit App
├── nature.pdf           # Sample knowledge document
└── README.md            # Project Documentation

Installation & Setup

Clone the Repository

   git clone https://github.com/RAJA102002/ragbasedchatbot.git

Install Dependencies

Could you make sure you have Python 3.8+ installed?

• Install Python
• Install Ollama

Start Ollama (Locally) and other Dependencies

Ensure you have Ollama installed and running:

#Install Ollama

Download DeepSeek R1:

ollama pull deepseek-llm:latest

Download Nomic Embed models:

ollama pull nomic-embed-text:latest

Install required packages:

pip install chromadb

Now, set your OpenAI API key, to the requirement for the SDK:

export OPENAI_BASE_URL=http://localhost:11434/v1
export OPENAI_API_KEY=fake-key

Run the Application**

streamlit run agents.py

---

How It Works

1. Loads Knowledge – Uses sample.pdf, For Example nature.pdf for retrieval-based answering.
2. Embeds Data – Utilizes Nomic Embed Text for vectorized search.
3. Retrieves Relevant Info – Searches ChromaDB for the most relevant content.
4. Generates Responses – Feeds retrieved data into DeepSeek R1 for contextual answers.

---

Example Usage

1. Run the app and open the Streamlit UI.
2. Ask a question related to the uploaded document.
3. Get AI-generated responses based on retrieved knowledge!