mirror of
https://github.com/bentoml/OpenLLM.git
synced 2026-03-06 08:08:03 -05:00
fix(yapf): align weird new lines break [generated] [skip ci] (#284)
fix(yapf): align weird new lines break Signed-off-by: aarnphm-ec2-dev <29749331+aarnphm@users.noreply.github.com>
This commit is contained in:
@@ -40,13 +40,7 @@ if t.TYPE_CHECKING:
|
||||
from openllm.client import BaseAsyncClient
|
||||
|
||||
class ResponseComparator(JSONSnapshotExtension):
|
||||
|
||||
def serialize(self,
|
||||
data: SerializableData,
|
||||
*,
|
||||
exclude: PropertyFilter | None = None,
|
||||
matcher: PropertyMatcher | None = None,
|
||||
) -> SerializedData:
|
||||
def serialize(self, data: SerializableData, *, exclude: PropertyFilter | None = None, matcher: PropertyMatcher | None = None,) -> SerializedData:
|
||||
if openllm.utils.LazyType(ListAny).isinstance(data):
|
||||
data = [d.unmarshaled for d in data]
|
||||
else:
|
||||
@@ -55,7 +49,6 @@ class ResponseComparator(JSONSnapshotExtension):
|
||||
return orjson.dumps(data, option=orjson.OPT_INDENT_2 | orjson.OPT_SORT_KEYS).decode()
|
||||
|
||||
def matches(self, *, serialized_data: SerializableData, snapshot_data: SerializableData) -> bool:
|
||||
|
||||
def convert_data(data: SerializableData) -> openllm.GenerationOutput | t.Sequence[openllm.GenerationOutput]:
|
||||
try:
|
||||
data = orjson.loads(data)
|
||||
@@ -83,8 +76,7 @@ class ResponseComparator(JSONSnapshotExtension):
|
||||
return (len(s.responses) == len(t.responses) and all([_s == _t for _s, _t in zip(s.responses, t.responses)]) and
|
||||
eq_config(s.marshaled_config, t.marshaled_config))
|
||||
|
||||
return len(serialized_data) == len(snapshot_data) and all(
|
||||
[eq_output(s, t) for s, t in zip(serialized_data, snapshot_data)])
|
||||
return len(serialized_data) == len(snapshot_data) and all([eq_output(s, t) for s, t in zip(serialized_data, snapshot_data)])
|
||||
|
||||
@pytest.fixture()
|
||||
def response_snapshot(snapshot: SnapshotAssertion):
|
||||
@@ -133,14 +125,8 @@ class LocalHandle(_Handle):
|
||||
return self.process.poll() is None
|
||||
|
||||
class HandleProtocol(t.Protocol):
|
||||
|
||||
@contextlib.contextmanager
|
||||
def __call__(*,
|
||||
model: str,
|
||||
model_id: str,
|
||||
image_tag: str,
|
||||
quantize: t.AnyStr | None = None,
|
||||
) -> t.Generator[_Handle, None, None]:
|
||||
def __call__(*, model: str, model_id: str, image_tag: str, quantize: t.AnyStr | None = None,) -> t.Generator[_Handle, None, None]:
|
||||
...
|
||||
|
||||
@attr.define(init=False)
|
||||
@@ -148,9 +134,7 @@ class DockerHandle(_Handle):
|
||||
container_name: str
|
||||
docker_client: docker.DockerClient
|
||||
|
||||
def __init__(self, docker_client: docker.DockerClient, container_name: str, port: int,
|
||||
deployment_mode: t.Literal['container', 'local'],
|
||||
):
|
||||
def __init__(self, docker_client: docker.DockerClient, container_name: str, port: int, deployment_mode: t.Literal['container', 'local'],):
|
||||
self.__attrs_init__(port, deployment_mode, container_name, docker_client)
|
||||
|
||||
def status(self) -> bool:
|
||||
@@ -165,22 +149,14 @@ def _local_handle(model: str,
|
||||
quantize: t.Literal['int8', 'int4', 'gptq'] | None = None,
|
||||
*,
|
||||
_serve_grpc: bool = False,
|
||||
):
|
||||
):
|
||||
with openllm.utils.reserve_free_port() as port:
|
||||
pass
|
||||
|
||||
if not _serve_grpc:
|
||||
proc = openllm.start(model,
|
||||
model_id=model_id,
|
||||
quantize=quantize,
|
||||
additional_args=['--port', str(port)],
|
||||
__test__=True)
|
||||
proc = openllm.start(model, model_id=model_id, quantize=quantize, additional_args=['--port', str(port)], __test__=True)
|
||||
else:
|
||||
proc = openllm.start_grpc(model,
|
||||
model_id=model_id,
|
||||
quantize=quantize,
|
||||
additional_args=['--port', str(port)],
|
||||
__test__=True)
|
||||
proc = openllm.start_grpc(model, model_id=model_id, quantize=quantize, additional_args=['--port', str(port)], __test__=True)
|
||||
|
||||
yield LocalHandle(proc, port, deployment_mode)
|
||||
proc.terminate()
|
||||
@@ -201,7 +177,7 @@ def _container_handle(model: str,
|
||||
quantize: t.Literal['int8', 'int4', 'gptq'] | None = None,
|
||||
*,
|
||||
_serve_grpc: bool = False,
|
||||
):
|
||||
):
|
||||
envvar = openllm.utils.EnvVarMixin(model)
|
||||
|
||||
with openllm.utils.reserve_free_port() as port, openllm.utils.reserve_free_port() as prom_port:
|
||||
@@ -237,7 +213,7 @@ def _container_handle(model: str,
|
||||
'3000/tcp': port,
|
||||
'3001/tcp': prom_port
|
||||
},
|
||||
)
|
||||
)
|
||||
|
||||
yield DockerHandle(client, container.name, port, deployment_mode)
|
||||
|
||||
|
||||
@@ -16,11 +16,8 @@ model = 'flan_t5'
|
||||
model_id = 'google/flan-t5-small'
|
||||
|
||||
@pytest.fixture(scope='module')
|
||||
def flan_t5_handle(handler: HandleProtocol, deployment_mode: t.Literal['container', 'local'],
|
||||
clean_context: contextlib.ExitStack,
|
||||
):
|
||||
with openllm.testing.prepare(model, model_id=model_id, deployment_mode=deployment_mode,
|
||||
clean_context=clean_context) as image_tag:
|
||||
def flan_t5_handle(handler: HandleProtocol, deployment_mode: t.Literal['container', 'local'], clean_context: contextlib.ExitStack,):
|
||||
with openllm.testing.prepare(model, model_id=model_id, deployment_mode=deployment_mode, clean_context=clean_context) as image_tag:
|
||||
with handler(model=model, model_id=model_id, image_tag=image_tag) as handle:
|
||||
yield handle
|
||||
|
||||
|
||||
@@ -16,11 +16,8 @@ model = 'opt'
|
||||
model_id = 'facebook/opt-125m'
|
||||
|
||||
@pytest.fixture(scope='module')
|
||||
def opt_125m_handle(handler: HandleProtocol, deployment_mode: t.Literal['container', 'local'],
|
||||
clean_context: contextlib.ExitStack,
|
||||
):
|
||||
with openllm.testing.prepare(model, model_id=model_id, deployment_mode=deployment_mode,
|
||||
clean_context=clean_context) as image_tag:
|
||||
def opt_125m_handle(handler: HandleProtocol, deployment_mode: t.Literal['container', 'local'], clean_context: contextlib.ExitStack,):
|
||||
with openllm.testing.prepare(model, model_id=model_id, deployment_mode=deployment_mode, clean_context=clean_context) as image_tag:
|
||||
with handler(model=model, model_id=model_id, image_tag=image_tag) as handle:
|
||||
yield handle
|
||||
|
||||
|
||||
Reference in New Issue
Block a user