🛒 Smart Shop AI - LLM E-commerce Assistant

Smart Shop AI is an intelligent e-commerce platform that leverages Large Language Models (LLMs) and Vector Search to provide a personalized shopping experience. It features a conversational AI assistant, hybrid search capabilities, and a real-time analytics pipeline powered by Apache Kafka.

✨ Features

• 🤖 AI Shopping Assistant: Powered by Groq (Llama 3) and Pinecone, the chatbot understands natural language queries and provides product recommendations using Retrieval-Augmented Generation (RAG).
• 🔍 Hybrid Search: Combines Vector Similarity Search (semantic understanding) with Keyword Search fallback for precise product discovery.
• 🎯 Recommendation Engine:
  • Personalized Recommendations based on user history.
  • Frequently Viewed Together suggestions.
  • Trending Products based on real-time activity.
  • Similar Products using vector embeddings.
• 📊 Real-time Analytics Pipeline: Tracks user interactions (Product Views, Clicks, Add to Cart, Session Start) using Apache Kafka and processes them into a PostgreSQL Data Warehouse.
• ⚡ Modern Tech Stack: Built with Node.js, Express, React, and Docker.

📸 Screenshots

🛠️ Tech Stack

• Frontend: React, Vite, Tailwind CSS, Lucide Icons.
• Backend: Node.js, Express.js.
• Database:
  • SQLite (via Prisma): Primary Product Catalog.
  • PostgreSQL: Analytics Data Warehouse.
  • Pinecone: Vector Database for product embeddings.
  • Redis: Session Store & Caching.
• Infrastructure: Docker, Docker Compose, Apache Kafka, Zookeeper.
• AI/ML: Groq SDK (Llama 3), Transformers.js (Embeddings), Pinecone SDK.

📋 Prerequisites

Ensure you have the following installed:

• Node.js (v18 or higher)
• Docker & Docker Compose
• Git

You will also need:

• A Groq API Key (for LLM inference).
• A Pinecone API Key & Index (for vector search).

🚀 Installation & Setup

1. Clone the Repository

git clone https://github.com/your-username/llm-web-app.git
cd llm-web-app

2. Configure Environment Variables

Create a .env file in the root directory and add your keys (use .env.example as a reference):

PORT=3000
GROQ_API_KEY=your_groq_api_key
PINECONE_API_KEY=your_pinecone_api_key
PINECONE_INDEX=your_index_name
PINECONE_ENV=your_pinecone_environment
DATABASE_URL="file:./dev.db" # For SQLite Product Catalog
# Add Kafka/Redis/Postgres configs as needed

3. Start Infrastructure

Start the required services (Kafka, Zookeeper, Redis, Postgres) using Docker:

npm run docker:up

Wait a few moments for all containers to initialize.

4. Install Dependencies

Install backend and frontend dependencies:

# Backend
npm install

# Frontend
cd frontend
npm install
cd ..

5. Initialize Databases

Set up the Product Catalog (SQLite) and Analytics Warehouse (Postgres):

Product Catalog:

npx prisma db push
# Optional: Seed data if a seed script exists
# npx prisma db seed 

Analytics Warehouse:

npm run warehouse:init

6. Setup Vector Search

Initialize Pinecone index and sync product embeddings:

npm run pinecone:setup  # Creates index if not exists
npm run pinecone:sync   # Generates embeddings for products

🏃 Running the Application

You will need to run three processes (concurrently or in separate terminals):

1. Backend Server

npm run dev
# Server runs at http://localhost:3000

2. Frontend Application

cd frontend
npm run dev
# App runs at http://localhost:5173

3. Event Consumer (Analytics)

Starts the Kafka consumer to process user events:

npm run consumer:start

📂 Project Structure

• consumers/: Kafka consumer scripts for processing events.
• frontend/: React application source code.
• prisma/: Prisma schema and SQLite database file.
• scripts/: Utility scripts for Pinecone setup and data syncing.
• utils/: Helper functions for Kafka, Recommender, and AI logic.
• warehouse/: SQL schemas for the analytics data warehouse.
• server.js: Main Express backend server.
• docker-compose.yml: Docker configuration for infrastructure services.

🔗 API Endpoints

• GET /api/products: Fetch all products.
• POST /api/chat: Send a message to the AI assistant.
• GET /api/recommendations/trending: Get trending products.
• GET /api/recommendations/personalized/:userId: Get user-specific recommendations.
• POST /api/track/view: Track a product view event.

🤝 Contributing

Contributions are welcome! Please open an issue or submit a pull request.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Commit your Changes (git commit -m 'Add some AmazingFeature')
4. Push to the Branch (git push origin feature/AmazingFeature)
5. Open a Pull Request