🏍️ Shop Assistant RAG Chatbot

A smart AI-powered assistant that helps users explore a shop's product catalog using natural language queries.

---

📌 Overview

This project demonstrates how to build a RAG (Retrieval-Augmented Generation) based Shop Assistant Chatbot using:

• 🧠 Google Gemini for LLM responses and embeddings
• 📆 Pinecone for vector search
• �� MySQL for product data storage
• ⚖️ FastAPI backend
• 🎨 Streamlit frontend

Users can browse, search, and chat about products via a sidebar chatbot while viewing a catalog in real-time!

---

🔥 Features

👉 Chatbot trained on actual product descriptions 👉 Live filtering by brand, gender, and price 👉 Modular and scalable backend + frontend 👉 Google Gemini + Pinecone-powered retrieval 👉 Full-stack Python implementation

---

🧱 Folder Structure

CustomerCareChatbot/
│
├── backend/              # FastAPI backend
│   ├── main.py
│   ├── routes/
│   └── services/
│
├── frontend/             # Streamlit frontend
│   └── app.py
│
├── embedding/            # Data loading and Pinecone sync
│   ├── data_insertion.py
│   └── sync_pinecone.py
│
├── .env
├── README.md
├── uv.toml
└── architecture.pdf      # System architecture diagram

---

⚙️ Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/snsupratim/ShopAssistantChatbot2.0.git
cd ShopAssistantChatbot2.0

2️⃣ Install uv and Setup Virtual Environment

pip install uv
uv venv
uv pip install -r requirements.txt

Or use uv pip install fastapi streamlit mysql-connector-python langchain-google-genai langchain-pinecone pinecone-client if no requirements.txt

---

3️⃣ Prepare .env file

Create a .env file in the root directory:

DB_PASSWORD=your_mysql_password
GOOGLE_API_KEY=your_gemini_api_key
PINECONE_API_KEY=your_pinecone_api_key

---

4️⃣ Insert Product Data into MySQL

python embedding/data_insertion.py

---

5️⃣ Generate Embeddings & Sync to Pinecone

python embedding/sync_pinecone.py

---

6️⃣ Start FastAPI Backend

uvicorn backend.main:app --reload

---

7️⃣ Start Streamlit Frontend

streamlit run frontend/app.py

---

🧪 Example Questions to Try

• "Do you have any red dresses for women?"
• "Show me Puma shoes under 3000."
• "What are the Nike products available?"
• "Tell me about affordable items for men."
• "Any stylish sneakers in black?"
• "What's the price of the Adidas t-shirt?"

---

📊 Architecture & Workflow

Please see the detailed architecture diagram and modular workflow in this PDF:

📄 Download architecture.pdf

---

💡 Tech Stack

Tool	Purpose
FastAPI	REST API backend
Streamlit	Frontend & Chat UI
MySQL	Product database
Pinecone	Vector DB for RAG
Google Gemini	Embedding + Chat Model
LangChain	LLM orchestration

---

🙌 Credits

• Developed by Supratim Nag
• Inspired by modern RAG-based AI systems

---

📽️ YouTube Demo

🔵 Watch the video tutorial: 👉 https://youtube.com/@sn-develpoment

---

📝 License

This project is licensed under the MIT License.