Agent Function Call Help

[dejiezhen]

Hi,

I am trying to test the agent function call example but I am having trouble. https://github.com/microsoft/autogen/blob/main/notebook/agentchat_function_call.ipynb

I copied the code onto my notebook and gave it the task:
"Generate a chart based on google stock data from the past 5 years"

Steps to reproduce

config_list = [
    {
        'model': 'gpt-4',
        'api_key': openai.api_key
    },
    {
        'model': 'gpt-35-turbo',
        'api_key': openai.api_key,
        'base_url': openai.api_key,
        'api_type': 'azure',
        'api_version': '2023-08-01-preview',
    },
    {
        'model': 'gpt-35-turbo-16k',
        'api_key': openai.api_key,
        'base_url': 'openai.api_base
        'api_type': 'azure',
        'api_version': '2023-08-01-preview',
    },
]

llm_config = {
    "functions": [
        {
            "name": "python",
            "description": "run cell in ipython and return the execution result.",
            "parameters": {
                "type": "object",
                "properties": {
                    "cell": {
                        "type": "string",
                        "description": "Valid Python cell to execute.",
                    }
                },
                "required": ["cell"],
            },
        },
        {
            "name": "sh",
            "description": "run a shell script and return the execution result.",
            "parameters": {
                "type": "object",
                "properties": {
                    "script": {
                        "type": "string",
                        "description": "Valid shell script to execute.",
                    }
                },
                "required": ["script"],
            },
        },
    ],
    "config_list": config_list,
    "timeout": 120,
}
chatbot = autogen.AssistantAgent(
    name="chatbot",
    system_message="For coding tasks, only use the functions you have been provided with. Reply TERMINATE when the task is done.",
    llm_config=llm_config,
)

# create a UserProxyAgent instance named "user_proxy"
user_proxy = autogen.UserProxyAgent(
    name="user_proxy",
    is_termination_msg=lambda x: x.get("content", "") and x.get("content", "").rstrip().endswith("TERMINATE"),
    human_input_mode="NEVER",
    max_consecutive_auto_reply=10,
    code_execution_config={"work_dir": "coding", 'use_docker':"python:3"},
)

# define functions according to the function description
from IPython import get_ipython

def exec_python(cell):
    ipython = get_ipython()
    result = ipython.run_cell(cell)
    log = str(result.result)
    if result.error_before_exec is not None:
        log += f"\n{result.error_before_exec}"
    if result.error_in_exec is not None:
        log += f"\n{result.error_in_exec}"
    return log

def exec_sh(script):
    return user_proxy.execute_code_blocks([("sh", script)])

# register the functions
user_proxy.register_function(
    function_map={
        "python": exec_python,
        "sh": exec_sh,
    }
)

task = "Generate a chart based on google stock data from the past 5 years"

# start the conversation
user_proxy.initiate_chat(
    chatbot,
    message=task
)

Dependency versions:

pydantic: 2.5.2
pyautogen: 0.2.0b2
langchain: 0.0.351

When I run the notebook, I am getting KeyError: 'pydantic_fields_set'

Can you give me some pointers? I want to use this library to generate charts when I give it data. Is there an updated code? Thank you!

[rickyloynd-microsoft]

@kevin666aa

[ekzhu]

Can you try using the code from the latest main branch? You can do that by pip install -e . once you have cloned the repo.

[tianshangwuyun]

I have the same error KeyError: 'pydantic_fields_set'
How should I handle it to solve this problem
It may be an error caused by the incompatibility of Langchain, as deleting Langchain is normal, but I need the convenient tools it provides

pydantic: 2.6.0
pyautogen: 0.2.8
langchain: 0.1.4

[tianshangwuyun]

This error is caused by an error when retrieving cache content

[ekzhu]

Could you take a look at the updated function call documentation? The notebook you referenced has already been updated

https://microsoft.github.io/autogen/docs/Use-Cases/agent_chat#tool-calling