[Feature Request] `ChatAgent` add code_execution capability

[Wendong-Fan]

Required prerequisites

• [X] I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• [ ] Consider asking first in a Discussion.

Motivation

each chat agent should be able to execute the code rather than only the EmbodiedAgent

Solution

transfer capability from EmbodiedAgent into ChatAgent, and remove EmbodiedAgent

enable_code_execution setting can be ‘auto’, ‘True’, ‘False’, need more research to check how OpenAI did

config reference: code_execution_config{work_dir (Optional, str), use_docker (Optional, list, str or bool), timeout (Optional, int), last_n_messages (Experimental, int or str) #The number of messages to look back for code execution}

Alternatives

No response

Additional context

No response

[onemquan]

OpenAI code-interpreter Case

from openai import OpenAI
client = OpenAI()

assistant = client.beta.assistants.create(
    # instructions: The system instructions that the assistant uses.
    instructions="You are a personal math tutor. When asked a math question, write and run code to answer the question.",
    model="gpt-3.5-turbo",
    temperature=0.0,
    tools=[{"type": "code_interpreter"}],
)

question = "Weng earns $12 an hour for babysitting. Yesterday, she just did 50 minutes of babysitting. How much did she earn?"

thread = client.beta.threads.create(
    messages=[
        {
            "role": "user",
            "content": question,
        }
    ]
)
run = client.beta.threads.runs.create_and_poll(
    thread_id=thread.id,
    assistant_id=assistant.id,
)

run_steps = client.beta.threads.runs.steps.list(
    thread_id=thread.id, run_id=run.id
)

for step in run_steps:
    print(step.step_details)

# MessageCreationStepDetails(message_creation=MessageCreation(message_id='msg_qnRfCmnCHoi0P5rh8JArdiBU'), type='message_creation')
# ToolCallsStepDetails(tool_calls=[CodeInterpreterToolCall(id='call_tAICvobwplzPZ6eZbjxAaqeq', code_interpreter=CodeInterpreter(input='# Constants\nhourly_rate = 12\nminutes_worked = 50\n\n# Convert minutes to hours\nhours_worked = minutes_worked / 60\n\n# Calculate earnings\nearnings = hours_worked * hourly_rate\nearnings', outputs=[]), type='code_interpreter')], type='tool_calls')
# MessageCreationStepDetails(message_creation=MessageCreation(message_id='msg_dQl0XqodCDfgkheVeazPcO25'), type='message_creation')

[Wendong-Fan]

Maybe we can also set code executor as one tool. @WHALEEYE is working on adding docker for the code execution to make it more safe, once the PR is created maybe @onemquan can also help do the review

cc @lightaime