Setup & Usage Guide

• Configure virtual environment

  python -m pip install --upgrade pip
  python -m venv env
  ./env/Scripts/activate
  

• Installing gRPC toolkit

  > python -m pip install grpcio
  > python -m pip install grpcio-tools
  

• Sample code:

  • The sample code for each category of communication between client and server is available in grpc-dev branch, clone it out for reference.

• Create .proto protobuf file

  

• Generate stubs using below command

  > (env) python -m grpc_tools.protoc --proto_path=. --python_out=.\src\ --grpc_python_out=.\src\ .\protos\*.proto
  

• Generated Code from ProtoCompiler:

  • The protoc (Compiler) results in generating *_pb2.py & *_pb2_grpc.py for respective .proto file.

  • The generated code gets synthezied in *_pb2.py whereas gRPC specific code (data class) will be available in *_pb2_grpc.py.

  • Elements generated in gRPC specific code:

    • Stub (*Stub): The generated stub class will be used by gRPC clients. The stub objects are initialized with same names of corresponding methods defined as service in proto file.

    • Servicer (*Servicer): Each service defined in proto file will generate a Servicer class which will be implemented by Server and each method defined in service must be overriden by Srrver class.

    • Registration Function (add_*_to_server): Each service generates a function that needs to registered with grpc.Server object so that Server can address multiple services.

• Create Server service which uses generated stubs to take request as input and returns the response.

• Create Client service which uses generated stubs to send the request and wait for the response.

• Enable Server Reflection:

  • Install grpcio-reflection package

  (env) > python -m pip install grpcio-reflection
  

  • Import reflection module from grpc_reflection.v1alpha import reflection in server code.

  • Configure reflection service by adding the service names

    # Server Reflection
    server_names = (
        pb2.DESCRIPTOR.services_by_name['Unary'].full_name,
        reflection.SERVICE_NAME
    )
    reflection.enable_server_reflection(server_names, server)