feat: add Responses websocket model and stream_ws example

[seratch]

This pull request adds Responses WebSocket model support and a new examples/basic/stream_ws.py example to the Python Agents SDK.

The change introduces a responses_websocket_session implementation, extends model/provider/config plumbing to support a websocket-backed Responses path, and updates streamed runner behavior so the new model can run through existing agent interfaces. It also adds test coverage for config/model mapping, streamed execution, tracing, and websocket session behavior.

This is a user-facing runtime feature addition with public configuration/model-selection impact, so reviewers should pay attention to backward compatibility and streamed/non-streamed behavior alignment. Related docs reference: WebSocket mode guide.

import asyncio
from agents import Agent, responses_websocket_session

async def main() -> None:
    agent = Agent(
        name="Assistant",
        instructions="You are a helpful assistant.",
        model="gpt-5.2-codex",
    )
    async with responses_websocket_session() as session:
        result = session.run_streamed(agent, "Say hello in one sentence.")
        async for event in result.stream_events():
            if event.type == "raw_response_event" and event.data.type == "response.output_text.delta":
                print(event.data.delta, end="", flush=True)

        print()  # newline after streamed text

asyncio.run(main())