Add MagenticOne API

[gagb]

Enables a simple API

import asyncio
from autogen_ext.models.openai import OpenAIChatCompletionClient
from autogen_ext.teams.magentic_one import MagenticOne
from autogen_agentchat.ui import Console

async def example_usage_hil():
    client = OpenAIChatCompletionClient(model="gpt-4o")
    # to enable human-in-the-loop mode, set hil_mode=True
    m1 = MagenticOne(client=client, hil_mode=True)
    task = "Write a Python script to fetch data from an API."
    result = await Console(m1.run_stream(task=task))
    print(result)


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

[husseinmozannar]

Nice!

Let's create a doc/blog that has contents similar to the m1 package readme https://github.com/microsoft/autogen/tree/main/python/packages/autogen-magentic-one

let's also add a hil_mode (human-in-loop mode) (default is False) where we optionally add the user proxy as part of the agent team like the m1 example https://github.com/microsoft/autogen/blob/main/python/packages/autogen-magentic-one/examples/example.py

[gagb]

Added:

• hil_mode
• caution statement
• architecture section
• Logging
• Azure instructions
• BING API KEY instructions

Do you think logging, azure, bing are necessary? This is what docs currently look like.

[husseinmozannar]

Logging
Azure instructions
BING API KEY instructions

are no longer necessary! so I think this is good to go

[husseinmozannar]

I pushed a small commit to disable ocr of the websurfer, I am only seeing refusals for OCR so it is a wasted API call

[husseinmozannar]

@gagb Maybe also point to this API in the readme of the magentic-one package and say this is a preferred way to use magentic-one. I wonder if it's also worth a reference in the main readme of autogen or in other prominent places

[ekzhu]

Use autogen_core.models.ChatCompletionClient base class instead of concrete class for the argument type

[gagb]

I did this because we only tested with openai. Do you still think I should change it?

[ekzhu]

Ah I see. Though this client has already been used by many for Ollama models. Also there is AzureOpenAIChatCompletionClient, which we also tested on. So I think we should still update it to use the base class.

Instead of rely on type, I think we need to validate the model capabilities in the constructor -- unless it has already been done in the base class. Sorry on mobile so a bit hard to switch page here.

We can also validate the model name, and raise warning if the model is something we haven't tested.

[gagb]

Okay good point. I will do this.

[gagb]

Done.

1. I added a private method to the class that validates the client.
   

2. I had to expose the _create_args in BaseOpenAIChatCompletion to achieve this.
   

[ekzhu]

@gagb Maybe also point to this API in the readme of the magentic-one package and say this is a preferred way to use magentic-one. I wonder if it's also worth a reference in the main readme of autogen or in other prominent places

I agree. I think we should update the MagenticOne page to point to this API.

Would be good to have this team as a preset in Studio also. @victordibia

[gagb]

What if we added a simple CLI that looks like this if you install autogen_ext?

pip install autogen_ext
m1 "find me a nice shwarma place in my neighborhood"

[ekzhu]

What if we added a simple CLI that looks like this if you install autogen_ext?

pip install autogen_ext

m1 "find me a nice shwarma place in my neighborhood"

This rocks!!