Netflix Insights Chatbot

This repository contains a chatbot application designed to provide insights into Netflix engagement trends for the year 2023. The system consists of a Flask backend and a React frontend, utilizing advanced AI models, RAG pipeline and text-to-sql to handle user queries about Netflix titles, technical specifications, and engagement trends.

Table of Contents

• Features
• Tech Stack
• Setup Instructions
• Backend
• Frontend
• Docker Deployment
• API Endpoints

Features

• GPT-4o powered Chat Interface: The chatbot leverages GPT-4o model for natural conversation.
• Vector Search: Uses FAISS for efficient similarity-based queries.
• OAuth2 Authentication: Supports Google login via OAuth2 for secure user access.
• Rate Limiting: Protects the API with rate limiting for a smoother experience.
• SQL-based querying: Provides insights into Netflix titles based on user queries through a powerful text-to-SQL pipeline.
• AI Agents: Utilizes intelligent agents to handle technical queries, provide guided prompts, and deliver structured information.
• Structured Responses: Returns related questions to help users explore similar topics.

Tech Stack

Backend:

• Python (Flask)
• LangChain
• PostgreSQL (TimescaleDB for vector storage)
• JWT-based Authentication
• OAuth2 with Google
• SQLAlchemy ORM

Frontend:

• ReactJS
• React Router
• Axios for HTTP requests
• JWT for authentication

Setup Instructions

Prerequisites

Make sure you have the following installed on your machine:

• Docker & Docker Compose
• Node.js (if running frontend locally)
• Python 3.9+

Environment Variables

Note: If you need the required environment variables, please feel free to reach out to me directly.

Create a .env file in the backend and frontend directories with the following variables:

Backend

SECRET_KEY=<your-secret-key>
JWT_SECRET_KEY=<your-jwt-secret-key>
GOOGLE_OAUTH_CLIENT_ID=<your-google-client-id>
GOOGLE_OAUTH_CLIENT_SECRET=<your-google-client-secret>
OPENAI_API_KEY=<your-openai-api-key>
EXTERNAL_DB_URL=<your-external-db-url>
CHAT_DB_URL=<your-chat-db-url>
FRONTEND_URL=<frontend-url>

Frontend

REACT_APP_BACKEND_URL=http://localhost:5001/api

Backend

The backend is located in the backend folder and provides the agent setup for processing user input and retrieving insights from the database.

To Run Locally:

1. Install dependencies:

   cd backend
   pip install -r requirements.txt

2. Start the Flask server:

   python app.py

Frontend

The frontend is located in the frontend folder and provides a sleek chat interface for interacting with the chatbot.

To Run Locally:

1. Install dependencies:

   cd frontend
   npm install

2. Start the React development server:

   npm start

Docker Deployment

You can easily deploy the entire system using Docker and Docker Compose.

1. Build and run the containers:

   docker-compose up --build

2. Access the frontend at http://localhost:3000 and the backend at http://localhost:5001/api.

API Endpoints

Backend API Endpoints:

• POST /api/chat: Send a user message to the agent and receive a response.
• GET /api/history: Fetch the chat history of a user for a given chat window.
• GET /api/chat-windows: Retrieve all chat windows for a user.
• POST /api/chat-windows: Create a new chat window.

License

This project is licensed under the MIT License.