aegis-framework 0.1.14

A comprehensive, extensible AI agent framework with local LLM integration

Aegis Framework

A comprehensive, extensible AI agent framework with local LLM integration.

Version 0.1.14

Features

• Modular agent architecture
• Local LLM integration with Ollama
• Simplified public interfaces
• Extensible design patterns
• Real-time task monitoring
• Database insights tracking
• Flexible model selection

Installation

pip install aegis-framework

For web interface support:

pip install aegis-framework[web]

Quick Start

from aegis_framework import MasterAIAgent, DesignAgent

# Create a basic agent
agent = MasterAIAgent(model="gemma2:9b")

# Ask questions
response = agent.answer_question("What are the key principles of AI?")
print(response)

# Generate designs
designer = DesignAgent()
design = designer.generate_new_design()
print(design)

Creating Custom Agents

from aegis_framework import MasterAIAgent

class DataAnalysisAgent(MasterAIAgent):
    def __init__(self, model: str = "codellama"):
        super().__init__(model=model)
        self.agent_task_map.update({
            "data_analysis": [
                "analyze data",
                "run analysis",
                "statistical test"
            ]
        })
    
    def analyze_data(self, data: str) -> str:
        return self.generate_response(f"Analyze this data: {data}")

# Use custom agent
analyst = DataAnalysisAgent()
result = analyst.analyze_data("Your data here")

Requirements

• Python 3.7+
• Ollama (for local LLM support)

Optional Dependencies

• Flask & Flask-SocketIO (for web interface)
• fuzzywuzzy (for enhanced text matching)
• sqlite3-api (for database insights)

Documentation

For detailed documentation and examples, visit our GitHub repository.

License

MIT License. See LICENSE file for details.

Contact

• Author: Metis Analytics
• Email: [e-mail]