Skip to main content

A python-sdk client for interacting with AgentHub's ATTP protocol (Agent Tool Transport Protocol)

Project description

ATTP Client

A Python SDK client for interacting with AgentHub's ATTP (Agent Tool Transport Protocol).

Overview

The ATTP Client provides a comprehensive Python interface for connecting to and interacting with AgentHub's ATTP protocol. It enables real-time communication with AI agents, tool management, and inference operations through a high-performance async/await interface.

Features

  • Async/Await Support: Built with modern Python async patterns for optimal performance
  • Agent Communication: Direct communication with AI agents through the ATTP protocol
  • Tool Management: Register, unregister, and manage tools in organized catalogs
  • Inference API: Invoke AI agent inference with configurable parameters
  • Real-time Messaging: Stream responses and handle real-time agent interactions
  • Event Handling: Register custom event handlers for connect/disconnect and custom routes
  • Authentication: Secure token-based authentication with organization support

Installation

pip install attp-client

Or with Poetry:

poetry add attp-client

Quick Start

Basic Connection

import asyncio
from attp_client.client import ATTPClient

async def main():
    # Initialize the client
    client = ATTPClient(
        agt_token="your_agent_token_here",
        organization_id=1,
        connection_url="attp://localhost:6563"  # Optional, defaults to localhost
    )
    
    # Connect to AgentHub
    await client.connect()
    
    # Your code here...
    
    # Close the connection
    await client.close()

if __name__ == "__main__":
    asyncio.run(main())

Agent Inference

from attp_client.interfaces.inference.message import IMessageDTOV2
from attp_client.interfaces.inference.enums.message_type import MessageTypeEnum
from uuid import UUID

# Invoke inference by agent ID
response = await client.inference.invoke_inference(
    agent_id=17,
    input_configuration={},
    messages=[
        IMessageDTOV2(
            content="Hello, how can you help me?",
            message_type=MessageTypeEnum.USER_MESSAGE,
            chat_id=UUID("your-chat-id-here")
        )
    ],
    stream=False,
    timeout=30
)

print("Agent response:", response)

Tool Management

# Register a tool
tool_id = await client.tools.register(
    catalog_name="my_catalog",
    name="example_tool",
    description="An example tool for demonstration",
    schema_id="tool_schema_v1",
    return_direct=False,
    timeout_ms=20000
)

# Access a specific catalog
catalog = await client.catalog("my_catalog")

# Unregister a tool
await client.tools.unregister("my_catalog", str(tool_id))

Event Handling

def on_connect():
    print("Connected to AgentHub!")

def on_disconnect():
    print("Disconnected from AgentHub!")

def handle_custom_event(data):
    print(f"Received custom event: {data}")

# Register event handlers
client.add_event_handler("", "connect", on_connect)
client.add_event_handler("", "disconnect", on_disconnect)
client.add_event_handler("custom:pattern", "custom", handle_custom_event)

Configuration

Client Parameters

  • agt_token (str): Your AgentHub authentication token
  • organization_id (int): Your organization ID
  • connection_url (str, optional): ATTP server URL (default: "attp://localhost:6563")
  • max_retries (int, optional): Maximum connection retry attempts (default: 20)
  • limits (Limits, optional): Connection limits configuration
  • logger (Logger, optional): Custom logger instance

Connection Limits

from attp_core.rs_api import Limits

limits = Limits(max_payload_size=50000)
client = ATTPClient(
    agt_token="...",
    organization_id=1,
    limits=limits
)

API Reference

ATTPClient

The main client class for ATTP communication.

Methods

  • connect(): Establish connection to AgentHub
  • close(): Close the connection
  • catalog(catalog_name: str): Access a specific tool catalog
  • add_event_handler(pattern: str, route_type: RouteType, callback: Callable): Register event handlers

Properties

  • inference: Access to the inference API
  • tools: Access to the tools manager
  • router: Access to the low-level router

AttpInferenceAPI

Handles AI agent inference operations.

Methods

  • invoke_inference(): Invoke inference for a specific agent
  • invoke_chat_inference(): Invoke inference for a chat session

ToolsManager

Manages tool registration and organization.

Methods

  • register(): Register a new tool
  • unregister(): Remove tool registration

Error Handling

The client includes comprehensive error handling:

from attp_client.errors import (
    AttpException,
    DeadSessionError,
    NotFoundError,
    SerializationError,
    UnauthenticatedError
)

try:
    await client.connect()
except UnauthenticatedError:
    print("Authentication failed - check your token")
except DeadSessionError:
    print("Session has died - reconnection required")
except AttpException as e:
    print(f"ATTP error: {e}")

Requirements

  • Python 3.11 - 3.13
  • Dependencies:
    • pydantic (>=2.11.7,<3.0.0)
    • attp-core (==0.1.10)
    • msgpack (>=1.1.1,<2.0.0)
    • ascender-framework (>=2.0rc7,<3.0)

Development

Setup

# Clone the repository
git clone https://github.com/AscenderTeam/attp-client.git
cd attp-client

# Install dependencies
poetry install

# Run tests
python -m pytest tests/

Project Structure

src/
├── attp_client/           # Main package
│   ├── client.py         # Main client class
│   ├── inference.py      # AI inference API
│   ├── tools.py          # Tool management
│   ├── catalog.py        # Catalog operations
│   ├── router.py         # Message routing
│   ├── session.py        # Session management
│   ├── errors/           # Exception classes
│   ├── interfaces/       # Protocol interfaces
│   ├── misc/             # Utilities
│   ├── types/            # Type definitions
│   └── utils/            # Helper utilities
└── tests/                # Test suite

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

For issues and questions, please use the GitHub Issues page or contact the Ascender Team.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

attp_client-0.0.10.tar.gz (18.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

attp_client-0.0.10-py3-none-any.whl (27.1 kB view details)

Uploaded Python 3

File details

Details for the file attp_client-0.0.10.tar.gz.

File metadata

  • Download URL: attp_client-0.0.10.tar.gz
  • Upload date:
  • Size: 18.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/2.1.3 CPython/3.11.13 Linux/6.16.6-arch1-1

File hashes

Hashes for attp_client-0.0.10.tar.gz
Algorithm Hash digest
SHA256 9af3868c943323ef91b0d06f5f639311165a5d404a184e0473f713b9afc5845a
MD5 cb0443b717cdf3f2b05abf3ec3435ce0
BLAKE2b-256 83dd7b7744a0dfdf42f5a1b2228047ce96c07515af97c818242bbe6600613af7

See more details on using hashes here.

File details

Details for the file attp_client-0.0.10-py3-none-any.whl.

File metadata

  • Download URL: attp_client-0.0.10-py3-none-any.whl
  • Upload date:
  • Size: 27.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/2.1.3 CPython/3.11.13 Linux/6.16.6-arch1-1

File hashes

Hashes for attp_client-0.0.10-py3-none-any.whl
Algorithm Hash digest
SHA256 6a03cb914e22bd8c797d41fc81581009198bf2bc6a9123c89aa2aa1514ebe6eb
MD5 b8e7f03dceab3fe8897921fadc4b5090
BLAKE2b-256 dfb8fdf7381baa6749c920a4df96c60898dee846734804adc37627ef380edd41

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page