[Frontend][Core] Standardize request error handling with VLLMError hierarchy (#49665)

Signed-off-by: Zach Zhu <zzqshu@126.com>
This commit is contained in:
Zach Zhu
2026-07-29 04:20:40 +00:00
committed by GitHub
parent f37f03db4a
commit 58f9659397
46 changed files with 370 additions and 248 deletions
@@ -18,6 +18,7 @@ from tests.utils import RemoteOpenAIServer
from vllm.entrypoints.openai.chat_completion.protocol import (
ChatCompletionRequest,
)
from vllm.exceptions import VLLMValidationError
from vllm.sampling_params import SamplingParams
# any model with a chat template should work here
@@ -1074,7 +1075,7 @@ def test_chat_completion_request_n_parameter_exceeds_default_limit(
max_tokens=10,
)
with pytest.raises(ValueError, match="n must be at most"):
with pytest.raises(VLLMValidationError, match="n must be at most"):
request.to_sampling_params(
max_tokens=10,
default_sampling_params={},
@@ -1136,7 +1137,7 @@ def test_chat_completion_request_n_parameter_custom_limit(
max_tokens=10,
)
with pytest.raises(ValueError, match="n must be at most 128"):
with pytest.raises(VLLMValidationError, match="n must be at most 128"):
request_over.to_sampling_params(
max_tokens=10,
default_sampling_params={},
@@ -1160,7 +1161,7 @@ def test_chat_completion_request_n_parameter_massive_value(
max_tokens=1,
)
with pytest.raises(ValueError, match="n must be at most"):
with pytest.raises(VLLMValidationError, match="n must be at most"):
request.to_sampling_params(
max_tokens=1,
default_sampling_params={},