ApiClient.request_sync rejects successful synchronous responses as non-awaitable

[FU-max-boop]

Summary

ApiClient.request_sync() is a synchronous convenience wrapper, but it passes the already-computed return value from the synchronous ApiClient.request() into asyncio.run_until_complete(). Any successful non-awaitable response therefore raises TypeError instead of being returned.

The async client needs the event-loop bridge; the sync client does not.

Minimal reproduction

from unittest.mock import patch

from qdrant_client.http.api_client import ApiClient

client = ApiClient("http://localhost:6333")
with patch.object(client, "request", return_value={"ok": True}):
    client.request_sync(
        type_=dict,
        method="GET",
        url="/collections",
    )

Current result:

TypeError: An asyncio.Future, a coroutine or an awaitable is required

Expected result:

{"ok": True}

Root cause

The sync and async classes currently share the same implementation:

return get_event_loop().run_until_complete(self.request(type_=type_, **kwargs))

For AsyncApiClient, self.request(...) is a coroutine. For ApiClient, it is an ordinary synchronous result.

Proposed fix

Have ApiClient.request_sync() directly return self.request(...), while leaving AsyncApiClient.request_sync() unchanged. Add a regression covering both typed responses and type_=None.

I searched open and closed issues and pull requests for request_sync / ApiClient; I did not find an existing report or implementation. I am preparing a focused patch against dev.

[Harsh23Kashyap]

I'll take this. Replacing the run_until_complete wrapper with a direct self.request(...) return on the sync client only (async stays the same), and adding a regression that mocks request to return a plain value. PR shortly.

[Harsh23Kashyap]

Opened PR #1367 with the one-line fix and a unit regression. Sync side now returns self.request(...) directly; async side is unchanged. Pure unit test, no live server needed.