ai-catalyst

AI_Catalyst Framework

A production-ready AI components framework with enterprise-grade security, resilience, and async performance optimizations.

🚀 Features

Core Components

• 🤖 Three-Tier LLM System - Local/Network/OpenAI with concurrent failover (3x faster)
• 🔒 PII Detection & Scrubbing - Enterprise-grade data protection with audit logging
• 📁 Multi-Format File Processing - Async streaming for JSON, JSONL, CSV, TXT
• 🗄️ Database Patterns - Async PostgreSQL with connection pooling
• ⚙️ Configuration Management - Encrypted config with key rotation
• 📊 System Monitoring - Health checks and performance metrics

Security & Resilience

• 🔐 Secure Configuration - AES-256 encryption for sensitive data
• 🔑 API Key Vault - Secure storage with automatic rotation
• 🚦 Rate Limiting - Token bucket algorithm with configurable limits
• 📋 Audit Logging - Comprehensive compliance tracking
• 🔄 Retry Logic - Exponential backoff with circuit breakers
• 💚 Health Monitoring - Continuous service availability checks

📦 Installation

pip install ai-catalyst

⚡ Quick Start

Basic Usage

import asyncio
from ai_catalyst import LLMProvider, PIIProcessor, FileProcessor

async def main():
    # LLM with automatic failover
    llm = LLMProvider()
    response = await llm.generate_async("Explain quantum computing")
    print(f"Response: {response['content']}")
    
    # PII scrubbing with audit logging
    pii = PIIProcessor()
    clean_text = await pii.scrub_text_async("Call John at 555-1234")
    print(f"Scrubbed: {clean_text}")  # "Call <NAME> at <PHONE>"
    
    # Async file processing
    processor = FileProcessor()
    async for item in processor.process_file_async("data.json"):
        print(f"Processed: {item['data']}")

asyncio.run(main())

Production Setup with Security

from ai_catalyst import (
    LLMProvider, APIKeyVault, RateLimiter, 
    AuditLogger, SecureConfigManager
)

# Setup secure environment
key_vault = APIKeyVault()
key_vault.store_key('openai', 'your-api-key')

rate_limiter = RateLimiter()
audit_logger = AuditLogger(log_file='audit.log')

# Initialize with security components
llm = LLMProvider(
    key_vault=key_vault,
    rate_limiter=rate_limiter,
    audit_logger=audit_logger
)

🏗️ Architecture

Async-First Design

• Non-blocking I/O - All operations are async by default
• Concurrent Processing - Parallel execution with semaphore control
• Backward Compatibility - Sync wrappers for existing code
• Resource Management - Proper cleanup and session handling

Security Model

• Zero-Trust - All operations are logged and rate-limited
• Encryption - Sensitive data encrypted at rest
• Audit Trail - Complete compliance logging
• Access Control - API key management with rotation

Resilience Patterns

• Circuit Breakers - Automatic failure detection and recovery
• Retry Logic - Exponential backoff with jitter
• Health Checks - Continuous service monitoring
• Graceful Degradation - Fallback strategies for all components

📊 Performance

🔧 Configuration

Environment Variables

# API Keys
export OPENAI_API_KEY="your-openai-key"
export ANTHROPIC_API_KEY="your-anthropic-key"

# Security
export AI_CATALYST_MASTER_KEY="your-encryption-key"

Configuration File

# config.yaml
llm:
  timeout: 30
  model: "gpt-4"
  
security:
  rate_limits:
    openai: "paid"  # or "free"
  audit_log: "audit.log"
  
processing:
  batch_size: 100
  max_concurrency: 10

🧪 Testing

# Install with dev dependencies
pip install ai-catalyst[dev]

# Run tests
pytest

# Run with coverage
pytest --cov=ai_catalyst

📚 Documentation

• Migration Guide - Upgrading to async
• Security Guide - Enterprise security setup
• Performance Guide - Optimization tips
• API Reference - Complete API documentation

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests for new functionality
4. Ensure all tests pass
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details.

🆘 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Security: Report security issues privately to security@example.com