def _config_inference_apis(self):
self._inference_apis = []
for _, function in inspect.getmembers(
self.__class__,
predicate=lambda x: inspect.isfunction(x) or inspect.ismethod(
x),
):
if hasattr(function, "_is_api"):
api_name = getattr(function, "_api_name")
route = getattr(function, "_api_route", None)
api_doc = getattr(function, "_api_doc")
input_adapter = getattr(function, "_input_adapter")
output_adapter = getattr(function, "_output_adapter")
mb_max_latency = getattr(function, "_mb_max_latency")
mb_max_batch_size = getattr(function, "_mb_max_batch_size")
batch = getattr(function, "_batch")
# Bind api method call with self(BentoService instance)
user_func = function.__get__(self)
self._inference_apis.append(
InferenceAPI(
self,
api_name,
api_doc,
input_adapter=input_adapter,
user_func=user_func,
output_adapter=output_adapter,
mb_max_latency=mb_max_latency,
mb_max_batch_size=mb_max_batch_size,
batch=batch,
route=route,
))
def _config_inference_apis(self):
self._inference_apis = []
for _, function in inspect.getmembers(
self.__class__,
predicate=lambda x: inspect.isfunction(x) or inspect.ismethod(
x),
):
if hasattr(function, "_is_api"):
api_name = getattr(function, "_api_name")
route = getattr(function, "_api_route", None)
api_doc = getattr(function, "_api_doc")
http_methods = getattr(function, "_http_methods")
# TODO: Add this while scaling
# mb_max_latency = getattr(function, "_mb_max_latency")
# mb_max_batch_size = getattr(function, "_mb_max_batch_size")
# Bind api method call with self(BentoService instance)
user_func = function.__get__(self)
self._inference_apis.append(
InferenceAPI(
self,
api_name,
api_doc,
user_func=user_func,
http_methods=http_methods,
# TODO: Add this while scaling
# mb_max_latency=mb_max_latency,
# mb_max_batch_size=mb_max_batch_size,
route=route,
))
