aegis-framework 0.1.13

A framework for creating multi-agent colonies

Aegis Multi-Agent Framework

A powerful framework for creating multi-agent AI colonies, with support for local LLM integration through Ollama.

Features

• 🤖 Multi-agent system architecture
• 🔄 Seamless Ollama LLM integration
• 🛠️ Easy-to-use API
• 📚 Comprehensive examples
• 🔌 Extensible design

Installation

Prerequisites

1. Python 3.8 or higher
2. Ollama (optional, for LLM integration)

Install from PyPI

pip install aegis-framework

Install from Source

git clone https://github.com/metisos/aegis_framework.git
cd aegis_framework
pip install -e .

Quick Start

Basic Usage (With Ollama LLM)

from aegis_framework import MasterAIAgent, OllamaLocalModel

# Initialize Ollama model
llm = OllamaLocalModel(model="llama2")  # or any other Ollama model

# Create an agent with LLM
agent = MasterAIAgent(name="LLM Agent", llm=llm)

# Ask questions
response = agent.answer_question("Explain quantum computing")
print(response)

Coding Assistant Example

The framework includes a powerful coding assistant example that can help with various programming tasks:

from aegis_framework import MasterAIAgent, OllamaLocalModel

def create_coding_agent():
    """Create an AI agent specialized for coding tasks"""
    llm = OllamaLocalModel(model="llama2")
    return MasterAIAgent(name="Code Assistant", llm=llm)

# Create the coding agent
agent = create_coding_agent()

# Example coding tasks
coding_questions = [
    "Write a Python function that implements binary search",
    "Create a Flask REST API endpoint",
    "Write unit tests for email validation"
]

# Get coding assistance
for question in coding_questions:
    print(f"\nQuestion: {question}")
    response = agent.answer_question(question)
    print(f"Response: {response}")

See coding_agent_example.py for a complete example with interactive mode.

Sample Scripts

The package includes several example scripts:

1. sample_usage.py: Basic usage examples
2. coding_agent_example.py: Coding assistant implementation

Configuration

Supported Ollama Models

You can use any model available in Ollama. Some recommended models:

• llama2
• codellama
• mistral
• gemma

To use a specific model:

llm = OllamaLocalModel(model="your_preferred_model")

Development

Running Tests

python -m pytest tests/

Contributing

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

• Author: Metis Analytics
• Email: [e-mail]
• GitHub: https://github.com/metisos/aegis_framework

Acknowledgments

• Thanks to the Ollama team for their excellent LLM runtime
• All contributors and users of the framework