[Profiler] Add per-session profiling controls to Python and Rust frontends (#57875)
Signed-off-by: czhu <conway.zhu@cohere.com> Signed-off-by: Conway Zhu <conway.zhu@cohere.com> Co-authored-by: aoshen02 <aoshen@inferact.ai>
This commit is contained in:
@@ -0,0 +1,86 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
|
||||
from unittest.mock import AsyncMock
|
||||
|
||||
import pytest
|
||||
from fastapi import FastAPI
|
||||
from fastapi.testclient import TestClient
|
||||
|
||||
from vllm.entrypoints.serve.profile.api_router import router
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def engine_client() -> AsyncMock:
|
||||
return AsyncMock()
|
||||
|
||||
|
||||
@pytest.fixture
|
||||
def client(engine_client: AsyncMock) -> TestClient:
|
||||
app = FastAPI()
|
||||
app.state.engine_client = engine_client
|
||||
app.include_router(router)
|
||||
return TestClient(app)
|
||||
|
||||
|
||||
def test_start_profile_without_body_is_backwards_compatible(
|
||||
client: TestClient, engine_client: AsyncMock
|
||||
) -> None:
|
||||
response = client.post("/start_profile")
|
||||
|
||||
assert response.status_code == 200
|
||||
engine_client.start_profile.assert_awaited_once_with()
|
||||
|
||||
|
||||
def test_start_profile_ignores_completion_payload(
|
||||
client: TestClient, engine_client: AsyncMock
|
||||
) -> None:
|
||||
response = client.post(
|
||||
"/start_profile",
|
||||
json={
|
||||
"model": "Qwen/Qwen3-0.6B",
|
||||
"prompt": "profile this request",
|
||||
"max_tokens": 16,
|
||||
"stream": False,
|
||||
},
|
||||
)
|
||||
|
||||
assert response.status_code == 200
|
||||
engine_client.start_profile.assert_awaited_once_with()
|
||||
|
||||
|
||||
def test_start_profile_forwards_session_overrides(
|
||||
client: TestClient, engine_client: AsyncMock
|
||||
) -> None:
|
||||
response = client.post(
|
||||
"/start_profile",
|
||||
json={
|
||||
"profile_prefix": "sharegpt_run-1",
|
||||
"delay_iterations": 5000,
|
||||
"max_iterations": 20,
|
||||
},
|
||||
)
|
||||
|
||||
assert response.status_code == 200
|
||||
engine_client.start_profile.assert_awaited_once_with(
|
||||
profile_prefix="sharegpt_run-1",
|
||||
delay_iterations=5000,
|
||||
max_iterations=20,
|
||||
)
|
||||
|
||||
|
||||
@pytest.mark.parametrize(
|
||||
"body",
|
||||
[
|
||||
{"profile_prefix": 1},
|
||||
{"delay_iterations": "1"},
|
||||
{"max_iterations": "20"},
|
||||
],
|
||||
)
|
||||
def test_start_profile_rejects_wrongly_typed_overrides(
|
||||
client: TestClient, engine_client: AsyncMock, body: dict
|
||||
) -> None:
|
||||
response = client.post("/start_profile", json=body)
|
||||
|
||||
assert response.status_code == 422
|
||||
engine_client.start_profile.assert_not_awaited()
|
||||
Reference in New Issue
Block a user