| Name | mas-framework JSON |
| Version |
0.1.14
JSON |
| download |
| home_page | None |
| Summary | A Multi-Agent System Framework |
| upload_time | 2025-01-09 21:38:29 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.11 |
| license | MIT |
| keywords |
ai
agent
framework
multi-agent
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# MAS AI - Experimental Multi-Agent System Framework
⚠️ **EXPERIMENTAL STATUS**: This project is in early development and APIs are subject to breaking changes. Not recommended for production use.
MAS AI is a Python framework for building multi-agent systems, focusing on reliable message passing and state management between agents.
## Core SDK Features
- **Simple Agent Creation** - Declarative agent definition with capabilities
- **Type-Safe State Management** - Immutable state management with Pydantic models
- **Message Routing** - Reliable agent-to-agent communication
- **Agent Discovery** - Find agents by capabilities
- **Lifecycle Management** - Controlled agent startup/shutdown
## Quick Start
1. Prerequisites:
```bash
# Required: Redis server for message transport
redis-server
# Install package
pip install mas-framework
```
2. Create an Agent:
```python
from mas.sdk.agent import Agent
from mas.sdk.decorators import agent
from mas.protocol import Message
from mas.sdk.state import AgentState
from pydantic import Field
# Optional: Define custom state model
class MyState(AgentState):
counter: int = Field(default=0)
name: str = Field(default="")
@agent(
agent_id="example_agent",
capabilities=["math", "storage"],
metadata={"version": "0.1.0"},
state_model=MyState # Optional custom state model
)
class ExampleAgent(Agent):
async def on_message(self, message: Message) -> None:
# Handle incoming messages
print(f"Got message: {message.payload}")
# Update agent's state
await self.update_state({
"counter": 42,
"name": "example"
})
# Access current state
current_state = self.state
print(f"Counter: {current_state.data['counter']}")
```
3. Run the Agent:
```python
import asyncio
from mas import mas_service
async def main():
async with mas_service() as context:
agent = await ExampleAgent.build(context)
try:
while True:
await asyncio.sleep(1)
finally:
await agent.stop()
if __name__ == "__main__":
asyncio.run(main())
```
## Key SDK Components
### State Management
Agents maintain immutable state with type-safe updates:
```python
# Define custom state model
class MyState(AgentState):
counter: int = Field(default=0)
status: str = Field(default="idle")
# Update state
await agent.update_state({
"counter": 42,
"status": "ready"
})
# Access state
current_state = agent.state
print(f"Counter: {current_state.data['counter']}")
# Reset state to initial values
await agent.reset_state()
# Subscribe to state changes
async def on_state_change(new_state: MyState) -> None:
print(f"State changed: {new_state.model_dump()}")
agent.subscribe_to_state(on_state_change)
```
### Message Handling
Pattern matching for message types:
```python
async def on_message(self, message: Message) -> None:
match message.message_type:
case MessageType.AGENT_MESSAGE:
await self.handle_agent_message(message)
case MessageType.DISCOVERY_RESPONSE:
await self.handle_discovery(message)
```
### Agent Discovery
Find other agents by capabilities:
```python
# Find agents with specific capabilities
await agent.runtime.discover_agents(capabilities=["math"])
```
### Lifecycle Hooks
```python
class MyAgent(Agent):
async def on_start(self) -> None:
"""Called when agent starts"""
await self.update_state({"status": "starting"})
async def on_stop(self) -> None:
"""Called when agent stops"""
await self.cleanup_resources()
```
## Current Limitations
As this is experimental software, there are several limitations:
- No authentication/authorization system yet
- Limited error recovery mechanisms
- Message delivery is not guaranteed
- No persistent storage (in-memory only)
- APIs may change without notice
- Limited testing in distributed environments
- No proper documentation yet
## Development Status
This project is under active development. Current focus areas:
- Stabilizing core APIs
- Improving error handling
- Adding authentication
- Adding persistent storage
- Documentation
- Testing infrastructure
## Contributing
This project is in experimental phase and we welcome feedback and contributions:
1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests: `pytest`
5. Submit a pull request
## License
MIT License - see LICENSE file for details
Raw data
{
"_id": null,
"home_page": null,
"name": "mas-framework",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.11",
"maintainer_email": null,
"keywords": "AI, agent, framework, multi-agent",
"author": null,
"author_email": "Lemuel Boyce <lemuel@vokality.com>",
"download_url": "https://files.pythonhosted.org/packages/f0/27/db60d301bc4bd6fbc6523a03059fa1ab08482894fbf9abf390f1cd64620c/mas_framework-0.1.14.tar.gz",
"platform": null,
"description": "# MAS AI - Experimental Multi-Agent System Framework\n\n\u26a0\ufe0f **EXPERIMENTAL STATUS**: This project is in early development and APIs are subject to breaking changes. Not recommended for production use.\n\nMAS AI is a Python framework for building multi-agent systems, focusing on reliable message passing and state management between agents.\n\n## Core SDK Features\n\n- **Simple Agent Creation** - Declarative agent definition with capabilities\n- **Type-Safe State Management** - Immutable state management with Pydantic models\n- **Message Routing** - Reliable agent-to-agent communication\n- **Agent Discovery** - Find agents by capabilities\n- **Lifecycle Management** - Controlled agent startup/shutdown\n\n## Quick Start\n\n1. Prerequisites:\n\n```bash\n# Required: Redis server for message transport\nredis-server\n\n# Install package\npip install mas-framework\n```\n\n2. Create an Agent:\n\n```python\nfrom mas.sdk.agent import Agent\nfrom mas.sdk.decorators import agent\nfrom mas.protocol import Message\nfrom mas.sdk.state import AgentState\nfrom pydantic import Field\n\n# Optional: Define custom state model\nclass MyState(AgentState):\n counter: int = Field(default=0)\n name: str = Field(default=\"\")\n\n@agent(\n agent_id=\"example_agent\",\n capabilities=[\"math\", \"storage\"],\n metadata={\"version\": \"0.1.0\"},\n state_model=MyState # Optional custom state model\n)\nclass ExampleAgent(Agent):\n async def on_message(self, message: Message) -> None:\n # Handle incoming messages\n print(f\"Got message: {message.payload}\")\n\n # Update agent's state\n await self.update_state({\n \"counter\": 42,\n \"name\": \"example\"\n })\n\n # Access current state\n current_state = self.state\n print(f\"Counter: {current_state.data['counter']}\")\n```\n\n3. Run the Agent:\n\n```python\nimport asyncio\nfrom mas import mas_service\n\nasync def main():\n async with mas_service() as context:\n agent = await ExampleAgent.build(context)\n try:\n while True:\n await asyncio.sleep(1)\n finally:\n await agent.stop()\n\nif __name__ == \"__main__\":\n asyncio.run(main())\n```\n\n## Key SDK Components\n\n### State Management\n\nAgents maintain immutable state with type-safe updates:\n\n```python\n# Define custom state model\nclass MyState(AgentState):\n counter: int = Field(default=0)\n status: str = Field(default=\"idle\")\n\n# Update state\nawait agent.update_state({\n \"counter\": 42,\n \"status\": \"ready\"\n})\n\n# Access state\ncurrent_state = agent.state\nprint(f\"Counter: {current_state.data['counter']}\")\n\n# Reset state to initial values\nawait agent.reset_state()\n\n# Subscribe to state changes\nasync def on_state_change(new_state: MyState) -> None:\n print(f\"State changed: {new_state.model_dump()}\")\n\nagent.subscribe_to_state(on_state_change)\n```\n\n### Message Handling\n\nPattern matching for message types:\n\n```python\nasync def on_message(self, message: Message) -> None:\n match message.message_type:\n case MessageType.AGENT_MESSAGE:\n await self.handle_agent_message(message)\n case MessageType.DISCOVERY_RESPONSE:\n await self.handle_discovery(message)\n```\n\n### Agent Discovery\n\nFind other agents by capabilities:\n\n```python\n# Find agents with specific capabilities\nawait agent.runtime.discover_agents(capabilities=[\"math\"])\n```\n\n### Lifecycle Hooks\n\n```python\nclass MyAgent(Agent):\n async def on_start(self) -> None:\n \"\"\"Called when agent starts\"\"\"\n await self.update_state({\"status\": \"starting\"})\n\n async def on_stop(self) -> None:\n \"\"\"Called when agent stops\"\"\"\n await self.cleanup_resources()\n```\n\n## Current Limitations\n\nAs this is experimental software, there are several limitations:\n\n- No authentication/authorization system yet\n- Limited error recovery mechanisms\n- Message delivery is not guaranteed\n- No persistent storage (in-memory only)\n- APIs may change without notice\n- Limited testing in distributed environments\n- No proper documentation yet\n\n## Development Status\n\nThis project is under active development. Current focus areas:\n\n- Stabilizing core APIs\n- Improving error handling\n- Adding authentication\n- Adding persistent storage\n- Documentation\n- Testing infrastructure\n\n## Contributing\n\nThis project is in experimental phase and we welcome feedback and contributions:\n\n1. Fork the repository\n2. Create a feature branch\n3. Make your changes\n4. Run tests: `pytest`\n5. Submit a pull request\n\n## License\n\nMIT License - see LICENSE file for details\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A Multi-Agent System Framework",
"version": "0.1.14",
"project_urls": null,
"split_keywords": [
"ai",
" agent",
" framework",
" multi-agent"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "cf8eb32ffb6af8d67f4e62aa85a8f9c15439804df58c599f19c8eb011aa5a81d",
"md5": "63660fe9151ca77dc12edb9dc4ddf590",
"sha256": "288073e075ea12ce0e8f0388a7e3f3e28b18d9024d95219d41f711cc33bda947"
},
"downloads": -1,
"filename": "mas_framework-0.1.14-py3-none-any.whl",
"has_sig": false,
"md5_digest": "63660fe9151ca77dc12edb9dc4ddf590",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.11",
"size": 29892,
"upload_time": "2025-01-09T21:38:28",
"upload_time_iso_8601": "2025-01-09T21:38:28.799559Z",
"url": "https://files.pythonhosted.org/packages/cf/8e/b32ffb6af8d67f4e62aa85a8f9c15439804df58c599f19c8eb011aa5a81d/mas_framework-0.1.14-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "f027db60d301bc4bd6fbc6523a03059fa1ab08482894fbf9abf390f1cd64620c",
"md5": "d8c21e1a288f52faa39e2d68c766b9a7",
"sha256": "75477925000a04232ffc31765c49d04a1754ec1ef02a8db73b69e84f1b1efcbf"
},
"downloads": -1,
"filename": "mas_framework-0.1.14.tar.gz",
"has_sig": false,
"md5_digest": "d8c21e1a288f52faa39e2d68c766b9a7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.11",
"size": 42271,
"upload_time": "2025-01-09T21:38:29",
"upload_time_iso_8601": "2025-01-09T21:38:29.955375Z",
"url": "https://files.pythonhosted.org/packages/f0/27/db60d301bc4bd6fbc6523a03059fa1ab08482894fbf9abf390f1cd64620c/mas_framework-0.1.14.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-01-09 21:38:29",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "mas-framework"
}
