ci: auto fixes from pre-commit.ci

For more information, see https://pre-commit.ci

src/_bentoml_impl/client/http.py

@@ -327,12 +327,14 @@ def _call(
         kwargs: dict[str, t.Any],
         *,
         headers: t.Mapping[str, str] | None = None,
-    ) -> t.Any: ...
+    ) -> t.Any:
+        ...
 
     @abstractmethod
     def _get_stream(
         self, endpoint: ClientEndpoint, args: t.Any, kwargs: t.Any
-    ) -> t.Any: ...
+    ) -> t.Any:
+        ...
 
 
 class SyncHTTPClient(HTTPClient[httpx.Client]):

src/_bentoml_impl/serde.py

@@ -28,16 +28,20 @@ class Serde(abc.ABC):
     media_type: str
 
     @abc.abstractmethod
-    def serialize_model(self, model: IODescriptor) -> bytes: ...
+    def serialize_model(self, model: IODescriptor) -> bytes:
+        ...
 
     @abc.abstractmethod
-    def deserialize_model(self, model_bytes: bytes, cls: type[T]) -> T: ...
+    def deserialize_model(self, model_bytes: bytes, cls: type[T]) -> T:
+        ...
 
     @abc.abstractmethod
-    def serialize(self, obj: t.Any) -> bytes: ...
+    def serialize(self, obj: t.Any) -> bytes:
+        ...
 
     @abc.abstractmethod
-    def deserialize(self, obj_bytes: bytes) -> t.Any: ...
+    def deserialize(self, obj_bytes: bytes) -> t.Any:
+        ...
 
     async def parse_request(self, request: Request, cls: type[T]) -> T:
         """Parse a input model from HTTP request"""

src/_bentoml_sdk/api.py

@@ -87,10 +87,12 @@ def __attrs_post_init__(self) -> None:
             self.output_spec.media_type = DEFAULT_STREAM_MEDIA_TYPE
 
     @t.overload
-    def __get__(self: T, instance: None, owner: type) -> T: ...
+    def __get__(self: T, instance: None, owner: type) -> T:
+        ...
 
     @t.overload
-    def __get__(self, instance: object, owner: type) -> t.Callable[P, R]: ...
+    def __get__(self, instance: object, owner: type) -> t.Callable[P, R]:
+        ...
 
     def __get__(self: T, instance: t.Any, owner: type) -> t.Callable[P, R] | T:
         from pydantic.fields import FieldInfo
@@ -216,7 +218,8 @@ def _flatten_model_schema(model: type[IODescriptor]) -> dict[str, t.Any]:
 
 
 @t.overload
-def api(func: t.Callable[t.Concatenate[t.Any, P], R]) -> APIMethod[P, R]: ...
+def api(func: t.Callable[t.Concatenate[t.Any, P], R]) -> APIMethod[P, R]:
+    ...
 
 
 @t.overload
@@ -230,7 +233,8 @@ def api(
     batch_dim: int | tuple[int, int] = ...,
     max_batch_size: int = ...,
     max_latency_ms: int = ...,
-) -> t.Callable[[t.Callable[t.Concatenate[t.Any, P], R]], APIMethod[P, R]]: ...
+) -> t.Callable[[t.Callable[t.Concatenate[t.Any, P], R]], APIMethod[P, R]]:
+    ...
 
 
 def api(

src/_bentoml_sdk/service/dependency.py

@@ -54,10 +54,12 @@ def get(
         return _dependent_cache[key]
 
     @t.overload
-    def __get__(self, instance: None, owner: t.Any) -> Dependency[T]: ...
+    def __get__(self, instance: None, owner: t.Any) -> Dependency[T]:
+        ...
 
     @t.overload
-    def __get__(self, instance: t.Any, owner: t.Any) -> T: ...
+    def __get__(self, instance: t.Any, owner: t.Any) -> T:
+        ...
 
     def __get__(self, instance: t.Any, owner: t.Any) -> Dependency[T] | T:
         if instance is None:

src/_bentoml_sdk/service/factory.py

@@ -46,7 +46,8 @@
     HookF_ctx = t.TypeVar("HookF_ctx", bound=ContextFunc)
 
     class _ServiceDecorator(t.Protocol):
-        def __call__(self, inner: type[T]) -> Service[T]: ...
+        def __call__(self, inner: type[T]) -> Service[T]:
+            ...
 
 
 def with_config(
@@ -324,11 +325,13 @@ def serve_http(
 
 
 @t.overload
-def service(inner: type[T], /) -> Service[T]: ...
+def service(inner: type[T], /) -> Service[T]:
+    ...
 
 
 @t.overload
-def service(inner: None = ..., /, **kwargs: Unpack[Config]) -> _ServiceDecorator: ...
+def service(inner: None = ..., /, **kwargs: Unpack[Config]) -> _ServiceDecorator:
+    ...
 
 
 def service(inner: type[T] | None = None, /, **kwargs: Unpack[Config]) -> t.Any:

src/bentoml/_internal/bento/build_config.py

@@ -791,7 +791,8 @@ def __init__(
             python: PythonOptions | dict[str, t.Any] | None = ...,
             conda: CondaOptions | dict[str, t.Any] | None = ...,
             models: list[ModelSpec | str | dict[str, t.Any]] | None = ...,
-        ) -> None: ...
+        ) -> None:
+            ...
 
     def __attrs_post_init__(self) -> None:
         use_conda = not self.conda.is_empty()

src/bentoml/_internal/client/__init__.py

@@ -87,17 +87,18 @@ async def async_wait_until_server_ready(
 
     @t.overload
     @staticmethod
-    def from_url(
-        server_url: str, *, kind: None | t.Literal["auto"] = ...
-    ) -> Client: ...
+    def from_url(server_url: str, *, kind: None | t.Literal["auto"] = ...) -> Client:
+        ...
 
     @t.overload
     @staticmethod
-    def from_url(server_url: str, *, kind: t.Literal["http"] = ...) -> HTTPClient: ...
+    def from_url(server_url: str, *, kind: t.Literal["http"] = ...) -> HTTPClient:
+        ...
 
     @t.overload
     @staticmethod
-    def from_url(server_url: str, *, kind: t.Literal["grpc"] = ...) -> GrpcClient: ...
+    def from_url(server_url: str, *, kind: t.Literal["grpc"] = ...) -> GrpcClient:
+        ...
 
     @staticmethod
     def from_url(
@@ -204,19 +205,22 @@ async def wait_until_server_ready(
     @classmethod
     async def from_url(
         cls, server_url: str, *, kind: None | t.Literal["auto"] = ...
-    ) -> AsyncGrpcClient | AsyncHTTPClient: ...
+    ) -> AsyncGrpcClient | AsyncHTTPClient:
+        ...
 
     @t.overload
     @classmethod
     async def from_url(
         cls, server_url: str, *, kind: t.Literal["http"] = ...
-    ) -> AsyncHTTPClient: ...
+    ) -> AsyncHTTPClient:
+        ...
 
     @t.overload
     @classmethod
     async def from_url(
         cls, server_url: str, *, kind: t.Literal["grpc"] = ...
-    ) -> AsyncGrpcClient: ...
+    ) -> AsyncGrpcClient:
+        ...
 
     @classmethod
     async def from_url(
@@ -322,19 +326,22 @@ def wait_until_server_ready(
     @classmethod
     def from_url(
         cls, server_url: str, *, kind: None | t.Literal["auto"] = ...
-    ) -> SyncGrpcClient | SyncHTTPClient: ...
+    ) -> SyncGrpcClient | SyncHTTPClient:
+        ...
 
     @t.overload
     @classmethod
     def from_url(
         cls, server_url: str, *, kind: t.Literal["http"] = ...
-    ) -> SyncHTTPClient: ...
+    ) -> SyncHTTPClient:
+        ...
 
     @t.overload
     @classmethod
     def from_url(
         cls, server_url: str, *, kind: t.Literal["grpc"] = ...
-    ) -> SyncGrpcClient: ...
+    ) -> SyncGrpcClient:
+        ...
 
     @classmethod
     def from_url(

src/bentoml/_internal/cloud/schemas/modelschemas.py

@@ -343,9 +343,9 @@ class DeploymentTargetConfig:
     deployment_strategy: t.Optional[DeploymentStrategy] = attr.field(
         default=None
     )  # Specific
-    bento_deployment_overrides: t.Optional[ApiServerBentoDeploymentOverrides] = (
-        attr.field(default=None)
-    )
+    bento_deployment_overrides: t.Optional[
+        ApiServerBentoDeploymentOverrides
+    ] = attr.field(default=None)
     bento_request_overrides: t.Optional[BentoRequestOverrides] = attr.field(
         default=None
     )  # Put into image builder

src/bentoml/_internal/context.py

@@ -81,12 +81,12 @@ def mutablecopy(self) -> Metadata:
 
 class ServiceContext:
     def __init__(self) -> None:
-        self._request_var: contextvars.ContextVar[ServiceContext.RequestContext] = (
-            contextvars.ContextVar("request")
-        )
-        self._response_var: contextvars.ContextVar[ServiceContext.ResponseContext] = (
-            contextvars.ContextVar("response")
-        )
+        self._request_var: contextvars.ContextVar[
+            ServiceContext.RequestContext
+        ] = contextvars.ContextVar("request")
+        self._response_var: contextvars.ContextVar[
+            ServiceContext.ResponseContext
+        ] = contextvars.ContextVar("response")
         # A dictionary for storing global state shared by the process
         self.state: dict[str, t.Any] = {}
 

src/bentoml/_internal/external_typing/starlette.py

@@ -10,7 +10,8 @@
     from starlette.types import Send as ASGISend
 
     class AsgiMiddleware(t.Protocol):
-        def __call__(self, app: ASGIApp, **options: t.Any) -> ASGIApp: ...
+        def __call__(self, app: ASGIApp, **options: t.Any) -> ASGIApp:
+            ...
 
     __all__ = [
         "AsgiMiddleware",

src/bentoml/_internal/external_typing/tensorflow.py

@@ -38,30 +38,37 @@
     from tensorflow.python.eager.function import ConcreteFunction
 
     class GenericFunction(t.Protocol):
-        def __call__(self, *args: t.Any, **kwargs: t.Any): ...
+        def __call__(self, *args: t.Any, **kwargs: t.Any):
+            ...
 
     # NOTE: future proof when tensorflow decided to remove EagerTensor
     # and enable eager execution on Tensor by default. This class only serves
     # as type fallback.
-    class EagerTensor(Tensor): ...
+    class EagerTensor(Tensor):
+        ...
 
 
 class SignatureMap(t.Mapping[str, t.Any], Trackable):
     """A collection of SavedModel signatures."""
 
     _signatures: t.Dict[str, t.Union[ConcreteFunction, RestoredFunction, Function]]
 
-    def __init__(self) -> None: ...
+    def __init__(self) -> None:
+        ...
 
     def __getitem__(
         self, key: str
-    ) -> t.Union[ConcreteFunction, RestoredFunction, Function]: ...
+    ) -> t.Union[ConcreteFunction, RestoredFunction, Function]:
+        ...
 
-    def __iter__(self) -> t.Iterator[str]: ...
+    def __iter__(self) -> t.Iterator[str]:
+        ...
 
-    def __len__(self) -> int: ...
+    def __len__(self) -> int:
+        ...
 
-    def __repr__(self) -> str: ...
+    def __repr__(self) -> str:
+        ...
 
 
 # This denotes all decorated function with `tf.function`

src/bentoml/_internal/frameworks/diffusers.py

@@ -653,9 +653,9 @@ def get_runnable(bento_model: bentoml.Model) -> t.Type[bentoml.Runnable]:
             type[diffusers.DiffusionPipeline], _str2cls(pipeline_class)
         )
 
-    scheduler_class: str | type[diffusers.SchedulerMixin] | None = (
-        bento_options.scheduler_class
-    )
+    scheduler_class: str | type[
+        diffusers.SchedulerMixin
+    ] | None = bento_options.scheduler_class
 
     if scheduler_class and isinstance(scheduler_class, str):
         scheduler_class = t.cast(
@@ -673,9 +673,9 @@ def get_runnable(bento_model: bentoml.Model) -> t.Type[bentoml.Runnable]:
     low_cpu_mem_usage: bool | None = bento_options.low_cpu_mem_usage
     variant: str | None = bento_options.variant
     _torch_dtype: str | torch.dtype | None = bento_options.torch_dtype
-    device_map: str | dict[str, int | str | torch.device] | None = (
-        bento_options.device_map
-    )
+    device_map: str | dict[
+        str, int | str | torch.device
+    ] | None = bento_options.device_map
     load_pretrained_extra_kwargs = bento_options.load_pretrained_extra_kwargs
 
     support_lora = True if issubclass(pipeline_class, LoraLoaderMixin) else False

src/bentoml/_internal/frameworks/transformers.py

@@ -56,14 +56,17 @@ class ModelDefaultMapping(t.TypedDict):
 
     class PreTrainedProtocol(t.Protocol):
         @property
-        def framework(self) -> str: ...
+        def framework(self) -> str:
+            ...
 
-        def save_pretrained(self, save_directory: str, **kwargs: t.Any) -> None: ...
+        def save_pretrained(self, save_directory: str, **kwargs: t.Any) -> None:
+            ...
 
         @classmethod
         def from_pretrained(
             cls, pretrained_model_name_or_path: str, *args: t.Any, **kwargs: t.Any
-        ) -> PreTrainedProtocol: ...
+        ) -> PreTrainedProtocol:
+            ...
 
     P = t.ParamSpec("P")
 
@@ -359,13 +362,15 @@ def delete_pipeline(task: str) -> None:
 @t.overload
 def load_model(
     bento_model: str | Tag | Model, **kwargs: t.Any
-) -> transformers.Pipeline: ...
+) -> transformers.Pipeline:
+    ...
 
 
 @t.overload
 def load_model(
     bento_model: str | Tag | Model, *args: t.Any, **kwargs: t.Any
-) -> TransformersPreTrained: ...
+) -> TransformersPreTrained:
+    ...
 
 
 def load_model(bento_model: str | Tag | Model, *args: t.Any, **kwargs: t.Any) -> t.Any:

src/bentoml/_internal/io_descriptors/base.py

@@ -79,7 +79,8 @@ class IODescriptor(ABC, _OpenAPIMeta, t.Generic[IOType]):
 
     if t.TYPE_CHECKING:
         # Populated by subclasses. Makes pyright happy.
-        def __init__(self, **kwargs: t.Any) -> None: ...
+        def __init__(self, **kwargs: t.Any) -> None:
+            ...
 
     __slots__ = (
         "_mime_type",
@@ -139,7 +140,8 @@ def sample(self, value: IOType) -> None:
     if t.TYPE_CHECKING:
 
         @classmethod
-        def from_sample(cls, sample: t.Any, **attrs: t.Any) -> t.Self: ...
+        def from_sample(cls, sample: t.Any, **attrs: t.Any) -> t.Self:
+            ...
 
     @abstractmethod
     def _from_sample(self, sample: t.Any) -> IOType:

src/bentoml/_internal/io_descriptors/multipart.py

@@ -327,12 +327,14 @@ async def from_proto(self, field: pb.Multipart) -> dict[str, t.Any]:
     @t.overload
     async def _to_proto_impl(
         self, obj: dict[str, t.Any], *, version: t.Literal["v1"]
-    ) -> pb.Multipart: ...
+    ) -> pb.Multipart:
+        ...
 
     @t.overload
     async def _to_proto_impl(
         self, obj: dict[str, t.Any], *, version: t.Literal["v1alpha1"]
-    ) -> pb_v1alpha1.Multipart: ...
+    ) -> pb_v1alpha1.Multipart:
+        ...
 
     async def _to_proto_impl(
         self, obj: dict[str, t.Any], *, version: str