Assistants API streaming with tool calls - implementation guide

I spent a week getting streaming to work properly with the Assistants API, especially with tool calls. Here's a complete implementation:

from openai import OpenAI
client = OpenAI()
with client.beta.threads.runs.stream(
    thread_id=thread.id,
    assistant_id=assistant.id,
) as stream:
    for event in stream:
        if event.event == "thread.message.delta":
            for block in event.data.delta.content:
                if block.type == "text":
                    print(block.text.value, end="", flush=True)
        elif event.event == "thread.run.requires_action":
            tool_outputs = []
            for call in event.data.required_action.submit_tool_outputs.tool_calls:
                result = handle_tool_call(call)
                tool_outputs.append({
                    "tool_call_id": call.id,
                    "output": json.dumps(result)
                })
            # Submit and continue streaming
            stream.submit_tool_outputs(tool_outputs)

The key gotcha: you MUST submit tool outputs through the stream object, not through a separate API call.