app.py

import asyncio
from llama_index.core.agent.workflow import AgentWorkflow, ReActAgent
from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI

def add(a: int, b: int) -> int:
    """Add two numbers."""
    return a + b

def multiply(a: int, b: int) -> int:
    """Multiply two numbers."""
    return a * b

async def main():
    llm = HuggingFaceInferenceAPI(model_name="Qwen/Qwen2.5-Coder-32B-Instruct")

    # Workaround: llama_index's astream_chat calls _async_client.close()
    # after each streaming response, killing the httpx connection. ReAct
    # agents need multiple LLM calls per run, so subsequent steps fail.
    async def _noop():
        pass
    llm._async_client.close = _noop

    multiply_agent = ReActAgent(
        name="multiply_agent",
        description="Is able to multiply two integers",
        system_prompt="A helpful assistant that can use a tool to multiply numbers.",
        tools=[multiply],
        llm=llm,
    )

    addition_agent = ReActAgent(
        name="add_agent",
        description="Is able to add two integers",
        system_prompt="A helpful assistant that can use a tool to add numbers.",
        tools=[add],
        llm=llm,
    )

    workflow = AgentWorkflow(
        agents=[multiply_agent, addition_agent],
        root_agent="multiply_agent",
    )

    response = await workflow.run(user_msg="Can you add 5 and 3?")
    print(response)

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