DuoGuard-Nemo

A multi-layered security system for LLM applications, built on top of NeMo Guardrails.

Features

• NeMo Guardrails integration for content filtering
• OpenAI content generation with security checks
• Comprehensive security logging
• Easy-to-use SDK interface
• Configurable security rules

Installation

pip install duoguard-nemo

Quick Start

import asyncio
from duoguard_nemo import DuoGuardNemoSDK

async def main():
    # Initialize the SDK
    sdk = DuoGuardNemoSDK(
        config_path="config",  # Path to NeMo Guardrails config
        openai_api_key="your-api-key",  # Or use OPENAI_API_KEY env var
        log_path="logs/security.log"
    )
    
    # Process a message
    result = await sdk.process_message(
        messages=[{"role": "user", "content": "Write an article about healthy eating."}],
        system_prompt="You are an AI Article writer..."
    )
    
    if result["success"]:
        print("Generated content:", result["response"])
    else:
        print("Error:", result["error"])

if __name__ == "__main__":
    asyncio.run(main())

Configuration

NeMo Guardrails Config

Place your NeMo Guardrails configuration in the config directory. The SDK will use this for content filtering.

Environment Variables

• OPENAI_API_KEY: Your OpenAI API key (optional if provided during initialization)

Security Logging

Security decisions and events are logged to logs/security.log by default. Each log entry includes:

• Timestamp
• Event type
• Message content
• Decision details
• Error information (if any)

Development

1. Clone the repository
2. Install development dependencies:

   pip install -e ".[dev]"
   

3. Run tests:

   pytest
   

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.