Add pydocstyle to the ruff rules (#52136)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Signed-off-by: Harry Mellor <hej.mellor@gmail.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 5
parent
804e5377aa
commit
8c1557a79c
@@ -454,7 +454,8 @@ async def test_derender_completion_empty_generate_responses(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_derender_completion_logprobs(client):
|
||||
"""token_id:N placeholders in logprobs are resolved; CompletionLogProbs
|
||||
flat-list structure is returned with non-empty tokens and text_offsets."""
|
||||
flat-list structure is returned with non-empty tokens and text_offsets.
|
||||
"""
|
||||
gr1 = await _render_completion(client, "Hello world")
|
||||
ids1 = gr1["token_ids"][:3]
|
||||
token_id = ids1[0]
|
||||
@@ -555,7 +556,7 @@ async def test_derender_chat_oversized_token_ids_rejected(client):
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_derender_chat_too_many_choices_rejected(client):
|
||||
"""choices count exceeding VLLM_MAX_N_SEQUENCES returns 400."""
|
||||
"""Choices count exceeding VLLM_MAX_N_SEQUENCES returns 400."""
|
||||
# Default VLLM_MAX_N_SEQUENCES is 16384; use a larger count.
|
||||
oversized_choices = [
|
||||
{"index": i, "token_ids": [42], "finish_reason": "stop"} for i in range(20_000)
|
||||
|
||||
@@ -121,7 +121,8 @@ async def _disagg(
|
||||
|
||||
def _tool_sig(response_choice: dict) -> list[tuple[str, dict]]:
|
||||
"""[(name, json normalized args)] so key ordering / whitespace don't
|
||||
cause false negatives."""
|
||||
cause false negatives.
|
||||
"""
|
||||
return [
|
||||
(tc["function"]["name"], json.loads(tc["function"]["arguments"]))
|
||||
for tc in (response_choice["message"].get("tool_calls") or [])
|
||||
@@ -143,7 +144,8 @@ async def _run_parity_case(
|
||||
client: httpx.AsyncClient, messages: list[dict], **extra
|
||||
) -> tuple[dict, dict]:
|
||||
"""Run the coupled request then feed its generated tokens into the
|
||||
disaggregated derender endpoint. Returns (coupled, disagg)."""
|
||||
disaggregated derender endpoint. Returns (coupled, disagg).
|
||||
"""
|
||||
coupled = await _coupled(client, messages, **extra)
|
||||
ch = coupled["choices"][0]
|
||||
chat_request = {"model": MODEL, "messages": messages, **extra}
|
||||
@@ -203,7 +205,8 @@ async def test_parity_tool_call(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_parity_reasoning_and_tool_call(client):
|
||||
"""Combined reasoning + tool call parity means the highest drift risk
|
||||
since it exercises both parser branches on the same output."""
|
||||
since it exercises both parser branches on the same output.
|
||||
"""
|
||||
messages = [{"role": "user", "content": "What's the weather in Paris?"}]
|
||||
coupled, disagg = await _run_parity_case(
|
||||
client,
|
||||
|
||||
@@ -41,7 +41,6 @@ def _make_stream_chunk(
|
||||
usage: dict | None = None,
|
||||
) -> GenerateStreamResponse:
|
||||
"""Build a GenerateStreamResponse SSE chunk."""
|
||||
|
||||
return GenerateStreamResponse(
|
||||
request_id=request_id,
|
||||
choices=[
|
||||
@@ -61,7 +60,6 @@ def _make_usage_chunk(
|
||||
request_id: str = "test-req",
|
||||
) -> GenerateStreamResponse:
|
||||
"""Build a usage only final SSE chunk (empty choices)."""
|
||||
|
||||
return GenerateStreamResponse(
|
||||
request_id=request_id,
|
||||
choices=[],
|
||||
@@ -531,7 +529,6 @@ async def client(server):
|
||||
|
||||
async def _render_chat(client) -> dict:
|
||||
"""Render a minimal chat request and return the GenerateRequest dict."""
|
||||
|
||||
resp = await client.post(
|
||||
"/v1/chat/completions/render",
|
||||
json={
|
||||
|
||||
@@ -337,7 +337,6 @@ async def test_chat_completion_render_multi_turn(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_chat_completion_render_with_stream_true(client):
|
||||
"""Render accepts stream params but still returns JSON (non-streamed)."""
|
||||
|
||||
response = await client.post(
|
||||
"/v1/chat/completions/render",
|
||||
json={
|
||||
@@ -456,7 +455,8 @@ async def test_chat_completion_render_with_sampling_params(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_completion_render_emits_token_offsets(client):
|
||||
"""With return_token_offsets, /v1/completions/render returns per-token
|
||||
(start, end) char offsets aligned with token_ids."""
|
||||
(start, end) char offsets aligned with token_ids.
|
||||
"""
|
||||
prompt = "Hello, world."
|
||||
response = await client.post(
|
||||
"/v1/completions/render",
|
||||
@@ -481,7 +481,8 @@ async def test_completion_render_emits_token_offsets(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_completion_render_default_no_token_offsets(client):
|
||||
"""Without the flag, token_offsets must be null (existing responses
|
||||
unchanged)."""
|
||||
unchanged).
|
||||
"""
|
||||
response = await client.post(
|
||||
"/v1/completions/render",
|
||||
json={
|
||||
@@ -498,7 +499,8 @@ async def test_completion_render_default_no_token_offsets(client):
|
||||
@pytest.mark.asyncio
|
||||
async def test_chat_render_emits_token_offsets(client):
|
||||
"""With return_token_offsets, /v1/chat/completions/render returns
|
||||
per-token offsets relative to the templated prompt string."""
|
||||
per-token offsets relative to the templated prompt string.
|
||||
"""
|
||||
response = await client.post(
|
||||
"/v1/chat/completions/render",
|
||||
json={
|
||||
@@ -672,7 +674,8 @@ async def test_chat_completion_render_assistant_tokens_mask_with_generation_tags
|
||||
client,
|
||||
):
|
||||
"""With a ``{% generation %}``-enabled template, the mask marks assistant
|
||||
tokens and the masked tokens decode to the assistant content."""
|
||||
tokens and the masked tokens decode to the assistant content.
|
||||
"""
|
||||
response = await client.post(
|
||||
"/v1/chat/completions/render",
|
||||
json={
|
||||
|
||||
@@ -16,7 +16,6 @@ VISION_MODEL_NAME = "Qwen/Qwen3-VL-2B-Instruct"
|
||||
@pytest.fixture(scope="module")
|
||||
def vision_server():
|
||||
"""Vision-capable server used for multimodal /render tests."""
|
||||
|
||||
args = [
|
||||
"--enforce-eager",
|
||||
"--max-model-len",
|
||||
@@ -48,7 +47,6 @@ async def test_chat_completion_render_with_base64_image_url(
|
||||
local_asset_server,
|
||||
):
|
||||
"""Render a multimodal chat request and verify tokens are returned."""
|
||||
|
||||
image = local_asset_server.get_image_asset("RGBA_comp.png")
|
||||
data_url = encode_image_url(image, format="PNG")
|
||||
|
||||
@@ -112,7 +110,6 @@ async def test_tokenize_matches_render_for_multimodal_input(
|
||||
local_asset_server,
|
||||
):
|
||||
"""`/tokenize` should match `/v1/chat/completions/render` token output."""
|
||||
|
||||
image = local_asset_server.get_image_asset("RGBA_comp.png")
|
||||
data_url = encode_image_url(image, format="PNG")
|
||||
|
||||
|
||||
@@ -1,7 +1,6 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
"""
|
||||
Roundtrip tests for multimodal serde used by the
|
||||
"""Roundtrip tests for multimodal serde used by the
|
||||
token_in_token_out generate endpoint.
|
||||
"""
|
||||
|
||||
|
||||
@@ -25,7 +25,8 @@ def _base_payload() -> dict:
|
||||
|
||||
def test_omitted_max_tokens_is_not_provided():
|
||||
"""Body without ``max_tokens`` must surface as 'not provided' so the
|
||||
server can apply its own default instead of the dataclass 16."""
|
||||
server can apply its own default instead of the dataclass 16.
|
||||
"""
|
||||
req = GenerateRequest.model_validate(_base_payload())
|
||||
# SamplingParams' dataclass default leaks through the parsed instance —
|
||||
# this is exactly the bug the server-side defaulting works around.
|
||||
@@ -35,7 +36,8 @@ def test_omitted_max_tokens_is_not_provided():
|
||||
|
||||
def test_explicit_max_tokens_is_provided():
|
||||
"""Even when the client picks the same value as the dataclass default,
|
||||
it must register as explicitly set so the server won't override it."""
|
||||
it must register as explicitly set so the server won't override it.
|
||||
"""
|
||||
payload = _base_payload()
|
||||
payload["sampling_params"] = {"max_tokens": 16}
|
||||
req = GenerateRequest.model_validate(payload)
|
||||
@@ -67,7 +69,8 @@ def test_json_roundtrip_preserves_provided_keys():
|
||||
def test_internal_instance_construction_treats_all_as_provided():
|
||||
"""When internal callers build ``GenerateRequest`` from a pre-resolved
|
||||
``SamplingParams`` instance, every field is considered explicitly set
|
||||
so server-side defaulting can't clobber values resolved upstream."""
|
||||
so server-side defaulting can't clobber values resolved upstream.
|
||||
"""
|
||||
sp = SamplingParams(max_tokens=500, temperature=0.0)
|
||||
req = GenerateRequest(token_ids=[1, 2, 3], sampling_params=sp)
|
||||
assert req.is_sampling_param_provided("max_tokens")
|
||||
|
||||
Reference in New Issue
Block a user