scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-11-15 09:12:18 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 311

test: cover the responses/ routing prefix in the native cache point case

Browse source
This commit is contained in:
mateo-berri 2026-09-06 03:33:20 -07:00
parent bf16fba1ba
commit 07083bd4f6
1 changed files with 7 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
tests/test_litellm/responses/test_responses_prompt_management.py
View file

@ -364,11 +364,16 @@ class TestResponsesAPIPromptManagement:
assert sent_input[1] == reasoning_item
assert sent_input[2]["id"] == "msg_1"
def test_native_responses_provider_places_role_targeted_cache_point(self):
@pytest.mark.parametrize(
"model",
["bedrock_mantle/openai.gpt-5.6-sol", "bedrock_mantle/responses/openai.gpt-5.6-sol"],
)
def test_native_responses_provider_places_role_targeted_cache_point(self, model: str):
"""A provider serving Responses natively gets no second pass, so the point lands on this one.
Handing a role-targeted point forward here sends it to the chat-completions bridge, which
never runs for this request, and the call goes upstream with nothing marked for caching.
The ``responses/`` routing prefix reaches the same deployment, so it has to read the same way.
"""
system_message = cast(AllMessageValues, {"role": "system", "content": "Analyze the request"})
user_message = cast(AllMessageValues, {"role": "user", "content": "Check for security issues"})
@ -389,7 +394,7 @@ class TestResponsesAPIPromptManagement:
litellm.responses(
input=cast(ResponseInputParam, [system_message, user_message]),
model="bedrock_mantle/openai.gpt-5.6-sol",
model=model,
litellm_logging_obj=_make_hook_backed_logging_obj(),
cache_control_injection_points=[{"location": "message", "role": "system"}],
aws_region_name="us-east-1",