bedrock-swarm 0.2.0.dev0

A framework for building multi-agent systems using AWS Bedrock

🤖 Bedrock Swarm

A powerful framework for building multi-agent systems using AWS Bedrock. Create, manage, and orchestrate AI agents powered by state-of-the-art language models.

✨ Features

• 🚀 Easy-to-use API - Create and manage AI agents with just a few lines of code
• 🤝 Multi-Agent Support - Build complex systems with multiple cooperating agents
• 🔧 Extensible Tool System - Create and add custom tools to enhance agent capabilities
• 💾 Memory Management - Built-in conversation memory for persistent context
• 🔒 Type Safety - Full type hints and runtime type checking
• 📚 Comprehensive Documentation - Detailed guides and API reference

🛠️ Installation

# Basic installation
pip install bedrock-swarm

# With development dependencies
pip install "bedrock-swarm[dev]"

# With documentation dependencies
pip install "bedrock-swarm[docs]"

🚀 Quick Start

Here's a simple example using an agent with the built-in CurrentTimeTool:

from bedrock_swarm.agency.agency import Agency
from bedrock_swarm.config import AWSConfig
from bedrock_swarm.tools.time import CurrentTimeTool

# Create AWS config
aws_config = AWSConfig(region="us-east-1")

# Create agency
agency = Agency(aws_config=aws_config)

# Create an agent with the time tool
agent = agency.add_agent(
    name="time_agent",
    model_id="us.anthropic.claude-3-5-sonnet-20241022-v2:0",
    tools=[CurrentTimeTool()],
    instructions="You are a helpful assistant that can tell the current time in different formats and timezones.",
)

# Create a thread
thread = agency.create_thread("time_agent")

# Example conversation
messages = [
    "What time is it now?",
    "What time is it in UTC?",
    "What's the current date in YYYY-MM-DD format?",
]

for message in messages:
    response = agency.execute(thread.thread_id, message)
    print(f"User: {message}")
    print(f"Assistant: {response.content}\n")

🎯 Examples

Check out our examples directory for ready-to-use examples:

1. Simple Time Example - A basic example showing how to:
   • Configure AWS and create an agent
   • Use the CurrentTimeTool for time-related queries
   • Handle different time formats and timezones

To run the examples:

1. Install with development dependencies:

pip install "bedrock-swarm[dev]"

2. Set up your environment variables:

export AWS_REGION=us-west-2
export AWS_PROFILE=default

3. Run an example:

python examples/simple_time.py

🛠️ Built-in Tools

CurrentTimeTool

A versatile tool for getting the current time and date in various formats and timezones:

from bedrock_swarm.tools.time import CurrentTimeTool

time_tool = CurrentTimeTool()

# Get current time in ISO format
current_time = time_tool.execute()  # Returns: "2024-03-14T15:09:26"

# Get time in specific format
formatted_time = time_tool.execute(
    format="%Y-%m-%d %H:%M:%S"
)  # Returns: "2024-03-14 15:09:26"

# Get time in specific timezone
utc_time = time_tool.execute(
    timezone="UTC",
    format="%H:%M:%S"
)  # Returns: "15:09:26"

# Get formatted date in specific timezone
tokyo_date = time_tool.execute(
    timezone="Asia/Tokyo",
    format="%A, %B %d, %Y"
)  # Returns: "Thursday, March 14, 2024"

🔧 Creating Custom Tools

You can create your own tools by extending the BaseTool class:

from typing import Any, Dict
from bedrock_swarm.tools.base import BaseTool

class MyCustomTool(BaseTool):
    @property
    def name(self) -> str:
        return "my_tool"

    @property
    def description(self) -> str:
        return "Description of my tool"

    def get_schema(self) -> Dict[str, Any]:
        return {
            "name": self.name,
            "description": self.description,
            "parameters": {
                "type": "object",
                "properties": {
                    "param1": {"type": "string"},
                },
                "required": ["param1"]
            }
        }

    def _execute_impl(self, **kwargs: Any) -> str:
        # Implement your tool logic here
        return f"Tool executed with {kwargs}"

📖 Documentation

Visit our documentation for:

• Installation Guide
• Quick Start Guide
• Core Concepts
• API Reference
• Examples

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

📄 License

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

📬 Contact

• GitHub Issues: github.com/sandrich/bedrock-swarm/issues
• Documentation: bedrock-swarm.readthedocs.io

Logging

Bedrock Swarm includes built-in logging capabilities to help you monitor and debug your applications. By default, logging is set to WARNING level.

from bedrock_swarm import configure_logging

# Enable debug logging
configure_logging(level="DEBUG")

For detailed information about logging configuration and best practices, see the Logging Guide.