style: drop explanatory docstrings from converted-stream helpers and tests

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm/caching/caching_handler.py

@@ -109,8 +109,6 @@ def _is_chat_completion_cached_dict(cached_result: dict) -> bool:
 
 
 def _stream_replay_requested(kwargs: Mapping[str, object]) -> bool:
-    """True when the caller must receive a stream, including when a deployment hook downgraded
-    `kwargs["stream"]` to False for the provider call."""
     return kwargs.get("stream", False) is True or converted_stream_requested(kwargs)
 
 

litellm/utils.py

@@ -847,8 +847,6 @@ def _is_streaming_response_for_correlation(result: object) -> bool:
 
 
 def _is_converted_stream_result(result: object) -> bool:
-    """True if `result` is a lazy stream wrapper the caller must iterate, even when a deployment
-    hook downgraded `kwargs["stream"]` to False for the provider call."""
     from litellm.litellm_core_utils.streaming_handler import CustomStreamWrapper
     from litellm.responses.streaming_iterator import BaseResponsesAPIStreamingIterator
 

tests/test_litellm/caching/test_caching_handler.py

@@ -697,8 +697,6 @@ async def test_cache_hit_records_the_looked_up_key_as_the_preset_cache_key(monke
 
 @pytest.mark.asyncio
 async def test_converted_stream_cache_hit_replayed_as_plain_object_logs_at_hit_time(monkeypatch):
-    """A converted-stream Anthropic Messages request that hits a non-stream cache entry gets a plain dict back,
-    so the success callbacks must fire now; nothing else will fire them."""
     import litellm
     from litellm.caching.caching import Cache
     from litellm.types.utils import CallTypes

tests/test_litellm/test_utils.py

@@ -5312,8 +5312,6 @@ async def test_wrapper_async_restores_originating_task_context_after_success(mon
 
 
 class _ConvertStreamDeploymentHook(CustomLogger):
-    """Headroom-style interception: downgrade stream=True to a non-streaming provider call."""
-
     async def async_pre_call_deployment_hook(
         self, kwargs: dict[str, object], call_type: CallTypes | None
     ) -> dict[str, object] | None:
@@ -5372,8 +5370,6 @@ def _assert_cache_hit_logged_as_stream(capture: _SuccessKwargsCapture, success_k
 async def test_wrapper_async_logs_converted_chat_stream_with_standard_logging_object(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    """Regression LIT-7729: the fake CustomStreamWrapper hit the non-streaming success path, which
-    built no standard_logging_object and deduped the wrapper's own end-of-stream dispatch."""
     capture: Final = _install_converted_stream_callbacks(monkeypatch)
 
     response: Final = await litellm.acompletion(
@@ -5400,8 +5396,6 @@ async def test_wrapper_async_logs_converted_chat_stream_with_standard_logging_ob
 async def test_wrapper_async_logs_converted_responses_stream_with_standard_logging_object(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    """Regression LIT-7729, Responses surface: the fake MockResponsesAPIStreamingIterator took the
-    same non-streaming success path and lost its standard_logging_object."""
     from litellm.responses.streaming_iterator import BaseResponsesAPIStreamingIterator
 
     capture: Final = _install_converted_stream_callbacks(monkeypatch)
@@ -5446,8 +5440,6 @@ async def test_wrapper_async_logs_converted_responses_stream_with_standard_loggi
 async def test_wrapper_async_replays_cached_converted_chat_stream_as_stream(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    """A cache hit for a converted stream must replay as a stream: the caller still iterates the
-    result even though the deployment hook set kwargs["stream"] to False."""
     capture: Final = _install_converted_stream_callbacks(monkeypatch)
     monkeypatch.setattr(litellm, "cache", Cache(type="local"))
     request: Final = {
@@ -5476,7 +5468,6 @@ async def test_wrapper_async_replays_cached_converted_chat_stream_as_stream(
 async def test_wrapper_async_replays_cached_converted_responses_stream_as_stream(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    """Responses surface of the cache-hit replay: the hit must come back as a streaming iterator."""
     from litellm.responses.streaming_iterator import BaseResponsesAPIStreamingIterator
 
     capture: Final = _install_converted_stream_callbacks(monkeypatch)