💬 ChatDB50: Natural Language Interface for SQL & NoSQL Databases

ChatDB50 is a full-stack web application that allows users to query both SQL and NoSQL (MongoDB) databases using natural language. It leverages a Large Language Model (LLM) to convert user-friendly English queries into executable database-specific queries and presents results in an intuitive format.

---

🚀 Key Features

• Accepts natural language input and converts it to SQL or MongoDB queries
• Executes queries on MySQL and MongoDB databases
• Displays results in a clean, readable format
• Powered by an LLM API for query translation

---

⚙️ Prerequisites

Ensure the following tools and environments are installed:

1. Python (≥3.8)

• Recommended: Use a virtual environment (venv or conda)

2. Node.js & npm (for frontend)

• Download Node.js

3. MongoDB

• Install MongoDB Community Edition or connect to a cloud MongoDB instance (MongoDB Atlas)

4. MySQL

• Install MySQL server and create a sample database

5. OpenAI API Key (or other LLM provider)

• Generate your API key from OpenAI
• Save it securely in your backend or config

---

📁 Project Structure

ChatDB50/
│
├── backend/
│   └── backend.py             # Flask backend API and query execution logic
│
├── frontend/
│   ├── public/                # Public assets (index.html, favicon, etc.)
│   ├── src/                   # Frontend source code
│   │   ├── App.js
│   │   ├── components/
│   │   │   └── QueryInput.js
│   │   └── utils/
│   │       ├── jsonToTableString.ts
│   │       ├── patternMatcher.ts
│   │       └── queryExecuter.ts
│   ├── components.json
│   ├── index.html
│   ├── package.json
│   ├── package-lock.json
│   ├── tailwind.config.js
│   ├── postcss.config.js
│   ├── vite.config.ts
│   ├── tsconfig.app.json
│   ├── tsconfig.json
│   ├── tsconfig.node.json
│   └── eslint.config.js
│
└── README.md

---

🛠️ Setup Instructions

🐍 Backend (Flask + LLM)

1. Clone the Repository

   git clone https://github.com/yourusername/chatdb50.git
   cd chatdb50/backend

2. Create Virtual Environment

   python -m venv venv
   source venv/bin/activate  # or venv\Scripts\activate on Windows

3. Run Backend

   python backend.py

🖥️ Frontend (React + Vite + TailwindCSS)

1. Navigate to frontend directory

   cd ../frontend

2. Install Dependencies

   npm install

3. Start the Development Server

   npm run preview

---

💡 Example Usage

For accessing SQL database, need to specifically add keyword "SQL" (case-insensitive) in the Natural Language Query. Similarly for NoSQL, need to add keyword "NoSQL" or "MongoDB" (case-insensitive)

• SQL Input: SQL Show me all customers who purchased in January

• NoSQL Input: NoSQL Show me all customers who purchased in January

• Translated SQL: SELECT * FROM purchases WHERE MONTH(purchase_date) = 1

• Translated MongoDB:

  db.purchases.find({ purchase_date: { $gte: ISODate("2023-01-01"), $lt: ISODate("2023-02-01") } })

---

🧠 Prompt Engineering Tips

If you're customizing prompts for your LLM:

prompt = f"""
Convert this natural language request into a {'MongoDB' if db_type == 'mongo' else 'SQL'} query:
"{user_input}"

Schema hint: {schema_info}
"""

---

🔮 Future Improvements

• Add support for voice-based queries
• User authentication and session-based history
• Add support for PostgreSQL, Cassandra, Firebase
• Fine-tune LLM prompts for improved precision and error handling

---

---