Skip to content

README.md

Latest commit

 

History

History
94 lines (71 loc) · 2.05 KB

README.md

File metadata and controls

94 lines (71 loc) · 2.05 KB

AI2SQL – Llama 3 Edition

A sophisticated natural language to SQL converter using FastAPI, LangChain, and Groq's Llama 3 API, featuring advanced query validation and security measures.

Features

Core Features

  • Natural language to SQL conversion
  • Session-based schema persistence
  • Interactive schema management
  • Query history tracking
  • Copy-to-clipboard functionality
  • Dark mode interface

Advanced Security

Query Validation

  • Pattern-Based Security

    • SQL injection prevention
    • System command blocking
    • Multiple statement detection
    • Hex value detection
    • Time-delay attack prevention

  • Natural Language Processing

    • Informal text detection
    • Non-SQL query filtering
    • Query complexity analysis
    • SQL keyword validation

  • Input Sanitization

    • String literal validation
    • Numeric overflow prevention
    • Special character filtering

Setup

  1. Create virtual environment and activate:

    python -m venv venv
.\venv\Scripts\activate  # Windows
source venv/bin/activate  # Unix

  2. Install dependencies:

    pip install -r requirements.txt

  3. Create .env file with your Groq API key:

    GROQ_API_KEY=your_api_key_here

  4. Run the application:

    python app.py

Usage

  1. First visit: Enter your database schema
  2. Ask questions in natural language
  3. Get instant SQL translations
  4. View and copy previous queries from history
  5. Edit or clear schema as needed

Query Examples

✅ Valid Queries:

  • "Show all users"
  • "How many orders were placed today?"
  • "Find total sales by category"

❌ Invalid Queries:

  • "Hi, who are you?"
  • "What's the weather today?"
  • "SELECT * FROM users; DROP TABLE users;"

Technology Stack

  • Frontend: HTML/CSS/JS, Dark Mode Theme
  • Backend: FastAPI, LangChain
  • AI: Groq's Llama 3 API
  • Security: NLTK, TensorFlow for query validation
  • Database: PostgreSQL/MySQL support

Contributing

Contributions are welcome! Please read our contributing guidelines before submitting pull requests.

License

MIT License