👕 T-Shirt Inventory Chatbot with LangChain & Streamlit

This project is an AI-powered chatbot that allows users to ask questions in natural language about a t-shirt inventory stored in a MySQL database. The chatbot automatically translates these questions into SQL queries using LangChain, and displays the answer—no SQL knowledge needed!

---

🚀 Features

• 💬 Ask natural language questions (e.g., "How many red Nike shirts in size M?")
• 🤖 Automatic SQL generation using LangChain + LLM (LLaMA via Ollama)
• 🧠 Uses few-shot learning with semantic similarity to improve prompt quality
• 📊 SQL database with randomized t-shirt inventory and discounts
• 🖥️ Simple and interactive Streamlit web interface

---

🗂️ Project Structure

├── main.py # Streamlit UI for user interaction

├── Builder.py # Core logic: builds the LangChain SQL chain

├── few_shots.py # Few-shot examples to improve SQL generation

├── tshirt_db_creation.sql # SQL script to set up and populate the database

Required packages include:

• streamlit

• langchain

• langchain-community

• langchain-experimental

• huggingface-hub

• chromadb

• pymysql

• sentence-transformers

• ollama

Note: You must have Ollama installed and running a compatible model like llama3.

🎥 Demo Video

Quick.example.mp4

SQL QUERY Verification:

Answer verification using MySQL