chore(sdk): able to specify service name (#4377)

examples/quickstart/service.py

@@ -10,7 +10,7 @@ def preprocess(self, input_series: np.ndarray) -> np.ndarray:
         return input_series
 
 
-@bentoml.service(resources={"cpu": "200m", "memory": "512Mi"})
+@bentoml.service(name="iris_classifier", resources={"cpu": "200m", "memory": "512Mi"})
 class IrisClassifier:
     iris_model = bentoml.models.get("iris_clf:latest")
     preprocessing = bentoml.depends(Preprocessing)

src/_bentoml_impl/server/app.py

@@ -172,7 +172,7 @@ def __call__(self, is_main: bool = False) -> Starlette:
 
     @property
     def name(self) -> str:
-        return self.service.inner.__name__
+        return self.service.name
 
     @property
     def middlewares(self) -> list[Middleware]:

src/_bentoml_sdk/service/config.py

@@ -188,6 +188,7 @@ class LoggingSchema(TypedDict, total=False):
 
 
 class ServiceConfig(TypedDict, total=False):
+    name: str
     traffic: TrafficSchema
     backlog: Annotated[int, Ge(64)]
     max_runner_connections: Posint

src/_bentoml_sdk/service/factory.py

@@ -140,6 +140,8 @@ def schema(self) -> dict[str, t.Any]:
 
     @property
     def name(self) -> str:
+        if name := self.config.get("name"):
+            return name
         return self.inner.__name__.lower()
 
     @property

src/bentoml/_internal/configuration/v2/__init__.py

@@ -63,6 +63,7 @@
     },
 }
 _SERVICE_CONFIG = {
+    s.Optional("name", default=None): s.Or(str, None),
     s.Optional("batching"): {
         s.Optional("enabled"): bool,
         s.Optional("max_batch_size"): s.And(int, ensure_larger_than_zero),