scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-27 06:13:07 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 72

fix(hosted_vllm): map developer items in _validate_input_param so compact requests are covered too

Browse source
This commit is contained in:
michelligabriele 2026-10-16 15:03:08 +02:00
parent 272038ceeb
commit be2489114f
No known key found for this signature in database
2 changed files with 21 additions and 15 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
litellm/llms/hosted_vllm/responses/transformation.py
View file

@ -93,21 +93,8 @@ class HostedVLLMResponsesAPIConfig(OpenAIResponsesAPIConfig):
return f"{api_base}/v1/responses"
def transform_responses_api_request(
self,
model: str,
input: str | ResponseInputParam,
response_api_optional_request_params: dict, # mutable-ok: overrides the base class signature
litellm_params: GenericLiteLLMParams,
headers: dict, # mutable-ok: overrides the base class signature
) -> dict: # mutable-ok: overrides the base class signature
return super().transform_responses_api_request(
model=model,
input=_developer_items_as_system(self._validate_input_param(input)),
response_api_optional_request_params=response_api_optional_request_params,
litellm_params=litellm_params,
headers=headers,
)
def _validate_input_param(self, input: str | ResponseInputParam) -> str | ResponseInputParam:
return _developer_items_as_system(super()._validate_input_param(input))
def supports_native_websocket(self) -> bool:
"""Hosted vLLM does not support native WebSocket for Responses API"""

19
tests/unit/llms/hosted_vllm/responses/test_hosted_vllm_responses.py
View file

@ -287,6 +287,25 @@ def test_hosted_vllm_responses_passes_string_input_and_instructions_through() ->
assert request["instructions"] == CODEX_INSTRUCTIONS
def test_hosted_vllm_compact_request_sends_developer_items_as_system() -> None:
url, data = HostedVLLMResponsesAPIConfig().transform_compact_response_api_request(
model="kimi-k3",
input=[
{"type": "message", "role": "developer", "content": CODEX_INSTRUCTIONS},
{"type": "message", "role": "user", "content": "summarize the session"},
],
response_api_optional_request_params={},
api_base=f"{VLLM_API_BASE}/responses",
litellm_params=GenericLiteLLMParams(),
headers={},
)
assert url == f"{VLLM_API_BASE}/responses/compact"
assert data["input"] == [
{"type": "message", "role": "system", "content": CODEX_INSTRUCTIONS},
{"type": "message", "role": "user", "content": "summarize the session"},
]
@pytest.fixture
def vllm_responses_route(respx_mock: respx.MockRouter, monkeypatch: pytest.MonkeyPatch) -> respx.Route:
monkeypatch.setattr(litellm, "disable_aiohttp_transport", True)