RAGgyBot

RAGgyBot is a Retrieval-Augmented Generation (RAG) chatbot built using FastAPI, Streamlit, and AWS services. It combines the power of embeddings, Chroma database retrieval, and Together AI for generating intelligent, context-aware responses. The application allows users to query a knowledge base and receive detailed, human-like answers.

Features

• Interactive Chatbot: Engages users with accurate, contextually relevant responses.
• RAG Integration: Utilizes retrieval-augmented generation to combine document retrieval and generative AI.
• Chroma Database: Stores and retrieves documents for query resolution.
• AWS S3 Integration: Downloads and manages Chroma databases from an S3 bucket.
• Streamlit Frontend: A user-friendly web interface for interacting with the chatbot.
• Dockerized Deployment: Streamlined deployment using Docker and Docker Compose.

---

Project Structure

RAG1_langchain/
├── app/
│   ├── main.py          # FastAPI backend for processing queries
│   ├── Dockerfile       # Dockerfile for the FastAPI service
├── streamlit/
│   ├── streamlit_app.py # Streamlit frontend for user interaction
│   ├── Dockerfile       # Dockerfile for the Streamlit service
├── docker-compose.yml   # Docker Compose configuration
└── README.md            # Project documentation

---

Requirements

• Python 3.8 or higher
• Docker and Docker Compose
• AWS credentials with access to the specified S3 bucket

---

Setup Instructions

1. Clone the Repository

git clone https://github.com/OppoTrain/RAG1_langchain.git
cd RAG1_langchain

2. Configure Environment Variables

Create a .env file in the root directory and add the following variables:

TOGETHER_API_KEY=<your_together_api_key>
AWS_ACCESS_KEY_ID=<your_aws_access_key>
AWS_SECRET_ACCESS_KEY=<your_aws_secret_key>
AWS_REGION=<your_aws_region>
S3_BUCKET_NAME=<your_s3_bucket_name>

3. Build and Run Services with Docker Compose

docker-compose up --build

• FastAPI service will run on http://localhost:8000
• Streamlit frontend will run on http://localhost:8501

---

Usage

1. Open the Streamlit app at http://localhost:8501.
2. Enter your query in the input box and click Get Answer.
3. View responses in the chat and refer to the conversation history.

---

Components

FastAPI Backend

• Hosts the API endpoint /synthesize/ to process user queries.
• Retrieves documents from Chroma database and generates responses using Together AI.

Streamlit Frontend

• Provides an intuitive web-based interface for user interaction.
• Displays conversation history and allows seamless communication with the backend.

Chroma Database

• Stores embeddings for document retrieval.
• Managed locally and synchronized with AWS S3.

---

Deployment

1. Ensure AWS credentials are correctly configured for accessing the S3 bucket.
2. Run docker-compose up to deploy both the backend and frontend.
3. Access the application via http://localhost:8501.

---

Development

Install Dependencies

pip install -r requirements.txt

Run Backend Locally

uvicorn app.main:app --reload

Run Frontend Locally

streamlit run streamlit/streamlit_app.py

---

Technologies Used

• FastAPI: Backend framework for API creation.
• Streamlit: Web application framework for the frontend.
• LangChain: Provides retrieval and embedding functionality.
• Chroma: Vector database for storing document embeddings.
• AWS S3: Cloud storage for managing Chroma database files.
• Docker Compose: Orchestrates multi-container deployment.

---

Future Enhancements

• Add support for additional retrieval methods (e.g., hybrid retrieval).
• Implement user authentication for personalized sessions.
• Extend the chatbot’s knowledge base with additional data sources.

---

License

This project is licensed under the MIT License.