Support visualization of model streaming chunks in AutoGen Studio

[victordibia]

Agentchat now supports streaming of tokens via ModelClientStreamingChunkEvent . This PR is to track progress on supporting that in the AutoGen Studio UI.

What

• Verify declarative support for streaming chunks in AGS
• update backend to handle ModelClientStreamingChunkEvent (do not save it)
• update fronted UI to appropriately display ModelClientStreamingChunkEvent

[husseinmozannar]

curious how to do this for tool calls or structured output?

[victordibia]

Good question.

• Currently tool calls seem to not yield any chunk events. This makes sense.
• I just tried with structured output and I seem to be seeing a bug.

Likely that @ekzhu is already aware of this.

from pydantic import BaseModel
# Define a tool that searches the web for information.
async def web_search(query: str) -> str:
    """Find information on the web"""
    return "AutoGen is a programming framework for building multi-agent applications."


class AgentResponse(BaseModel):
    content: str 


# Create an agent that uses the OpenAI GPT-4o model.
model_client = OpenAIChatCompletionClient(
    model="gpt-4o",
    response_format=AgentResponse, 
)
streaming_assistant = AssistantAgent(
    name="assistant",
    model_client=model_client,
    tools=[web_search],
    system_message="Use tools to solve tasks.",
    model_client_stream=True,
)
 
async for message in streaming_assistant.on_messages_stream(  # type: ignore
     [TextMessage(content="Find information on AutoGen", source="user")],
    cancellation_token=CancellationToken(),
):
    print(message)

TypeError: You tried to pass a BaseModel class to chat.completions.create(); You must use beta.chat.completions.parse() instead

[jackgerrits]

I believe that is this bug: #5568