It is a universal gRPC service framework suitable for model as a service. The data is input-output like JSON. Different services only need to add different model class files. At the same time, extension also becomes easy, as it can be inherited or replaced.
- see the Quickstart section below to get started with the basic functionality
- see the
/examplesfolder for more in-depthpywalkthroughs
- grpcio >= 1.57.0
- grpcio_health_checking >= 1.56.0
- protobuf >= 4.24.3
- scikit_learn >= 1.2.1
You can install the latest published version from PyPI using pip:
pip install basegrpc
- Input / output: please refer to the struct of gRPC
- Model file:
/examples/logic.mdl - Model class files: the file name and class name must be the same(
Logic.pyandclass Logic)
import pickle
import numpy as np
class Logic(object):
def __init__(self):
self.model_path = "logic.mdl"
# load model
self.model = self.load_model()
def load_model(self):
with open(self.model_path, "rb") as pickle_file:
model = pickle.load(pickle_file)
return model
def predict(self, data) -> list:
result = np.array([])
try:
data_X = data['data_X']
result = self.model.predict(data_X)
except Exception as e:
print(e)
return result.tolist()from basegrpc.grpc_service import BaseService
from basegrpc.grpc_module import ml_module
def run():
ml_module.module_name = "LogicV1"
grpc_service = BaseService()
grpc_service.run()
if __name__ == '__main__':
run()
# output:
# [PID 11732] gRPC load module: LogicV1
# [PID 11732] gRPC server start...from basegrpc.grpc_client import GrpcClient
if __name__ == "__main__":
grpc_client = GrpcClient(bind_address='127.0.0.1:50051')
data = {"data_X": [[5.1, 3.5]]}
response = grpc_client.request(request_data=data)
print("response: ", response)
health_status = grpc_client.health_check()
print("health status: ", health_status)
# output:
# [0.0]
# status: SERVING