Auto Generated Agent Chat: Collaborative Task Solving with Multiple Agents and Human Users#

AG2 offers conversable agents powered by LLM, tool, or human, which can be used to perform tasks collectively via automated chat. This framework allows tool use and human participation through multi-agent conversation. Please find documentation about this feature here.

In this notebook, we demonstrate an application involving multiple agents and human users to work together and accomplish a task. AssistantAgent is an LLM-based agent that can write Python code (in a Python coding block) for a user to execute for a given task. UserProxyAgent is an agent which serves as a proxy for a user to execute the code written by AssistantAgent. We create multiple UserProxyAgent instances that can represent different human users.

Requirements#

AG2 requires Python>=3.9. To run this notebook example, please install:

pip install autogen[openai]

# %pip install "autogen"

Set your API Endpoint#

The config_list_from_json function loads a list of configurations from an environment variable or a json file.

It first looks for an environment variable of a specified name (“OAI_CONFIG_LIST” in this example), which needs to be a valid json string. If that variable is not found, it looks for a json file with the same name. It filters the configs by models (you can filter by other keys as well).

The json looks like the following:

[
    {
        "model": "gpt-4",
        "api_key": "<your OpenAI API key here>"
    },
    {
        "model": "gpt-4",
        "api_key": "<your Azure OpenAI API key here>",
        "base_url": "<your Azure OpenAI API base here>",
        "api_type": "azure",
        "api_version": "2024-02-01"
    },
    {
        "model": "gpt-4-32k",
        "api_key": "<your Azure OpenAI API key here>",
        "base_url": "<your Azure OpenAI API base here>",
        "api_type": "azure",
        "api_version": "2024-02-01"
    }
]

You can set the value of config_list in any way you prefer. Please refer to this User Guide for full code examples of the different methods.

import autogen

config_list = autogen.config_list_from_json(
    "OAI_CONFIG_LIST",
    filter_dict={
        "model": ["gpt-4", "gpt4", "gpt-4-32k", "gpt-4-32k-0314", "gpt-4-32k-v0314"],
    },
)

Construct Agents#

We define ask_expert function to start a conversation between two agents and return a summary of the result. We construct an assistant agent named “assistant_for_expert” and a user proxy agent named “expert”. We specify human_input_mode as “ALWAYS” in the user proxy agent, which will always ask for feedback from the expert user.

def ask_expert(message):
    assistant_for_expert = autogen.AssistantAgent(
        name="assistant_for_expert",
        llm_config={
            "temperature": 0,
            "config_list": config_list,
        },
    )
    expert = autogen.UserProxyAgent(
        name="expert",
        human_input_mode="ALWAYS",
        code_execution_config={
            "work_dir": "expert",
            "use_docker": False,
        },  # Please set use_docker=True if docker is available to run the generated code. Using docker is safer than running the generated code directly.
    )

    expert.initiate_chat(assistant_for_expert, message=message)
    expert.stop_reply_at_receive(assistant_for_expert)
    # expert.human_input_mode, expert.max_consecutive_auto_reply = "NEVER", 0
    # final message sent from the expert
    expert.send("summarize the solution and explain the answer in an easy-to-understand way", assistant_for_expert)
    # return the last message the expert received
    return expert.last_message()["content"]

We construct another assistant agent named “assistant_for_student” and a user proxy agent named “student”. We specify human_input_mode as “TERMINATE” in the user proxy agent, which will ask for feedback when it receives a “TERMINATE” signal from the assistant agent. We set the functions in AssistantAgent and function_map in UserProxyAgent to use the created ask_expert function.

For simplicity, the ask_expert function is defined to run locally. For real applications, the function should run remotely to interact with an expert user.

assistant_for_student = autogen.AssistantAgent(
    name="assistant_for_student",
    system_message="You are a helpful assistant. Reply TERMINATE when the task is done.",
    llm_config={
        "timeout": 600,
        "cache_seed": 42,
        "config_list": config_list,
        "temperature": 0,
        "functions": [
            {
                "name": "ask_expert",
                "description": "ask expert when you can't solve the problem satisfactorily.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "message": {
                            "type": "string",
                            "description": "question to ask expert. Ensure the question includes enough context, such as the code and the execution result. The expert does not know the conversation between you and the user unless you share the conversation with the expert.",
                        },
                    },
                    "required": ["message"],
                },
            }
        ],
    },
)

student = autogen.UserProxyAgent(
    name="student",
    human_input_mode="TERMINATE",
    max_consecutive_auto_reply=10,
    code_execution_config={
        "work_dir": "student",
        "use_docker": False,
    },  # Please set use_docker=True if docker is available to run the generated code. Using docker is safer than running the generated code directly.
    function_map={"ask_expert": ask_expert},
)

Perform a task#

We invoke the initiate_chat() method of the student proxy agent to start the conversation. When you run the cell below, you will be prompted to provide feedback after the assistant agent sends a “TERMINATE” signal at the end of the message. The conversation will finish if you don’t provide any feedback (by pressing Enter directly). Before the “TERMINATE” signal, the student proxy agent will try to execute the code suggested by the assistant agent on behalf of the user.

# the assistant receives a message from the student, which contains the task description
student.initiate_chat(
    assistant_for_student,
    message="""Find $a + b + c$, given that $x+y \\neq -1$ and
\\begin{align}
    ax + by + c & = x + 7,\\
    a + bx + cy & = 2x + 6y,\\
    ay + b + cx & = 4x + y.
\\end{align}.
""",
)

When the assistant needs to consult the expert, it suggests a function call to ask_expert. When this happens, a line like the following will be displayed:

***** Suggested function Call: ask_expert *****