什么是gRPC？

gRPC是由Google开发的高性能、开源的远程过程调用(RPC)框架，具有以下核心特性：

• 基于HTTP/2协议，支持双向流和头部压缩
• 使用Protocol Buffers作为接口定义语言(IDL)和序列化工具
• 支持多种编程语言（Python、Go、Java等）
• 自动生成客户端和服务端代码
• 提供认证、负载均衡、健康检查等高级功能

Python gRPC基础

1. 安装gRPC工具

pip install grpcio grpcio-tools

2. 定义Protocol Buffers接口

创建 calculator.proto 文件：

syntax = "proto3";

package calculator;

service Calculator {
  rpc Add (AddRequest) returns (AddResponse) {}
  rpc Multiply (MultiplyRequest) returns (MultiplyResponse) {}
  rpc SquareStream (stream Number) returns (stream Number) {}
}

message AddRequest {
  double num1 = 1;
  double num2 = 2;
}

message AddResponse {
  double result = 1;
}

message MultiplyRequest {
  double num1 = 1;
  double num2 = 2;
}

message MultiplyResponse {
  double result = 1;
}

message Number {
  double value = 1;
}

3. 生成Python代码

python -m grpc_tools.protoc -I. --python_out=. --grpc_python_out=. calculator.proto

此命令会生成两个文件：calculator_pb2.py（消息类）和calculator_pb2_grpc.py（服务类）

实现gRPC服务端

import grpc
from concurrent import futures
import calculator_pb2
import calculator_pb2_grpc
import time

class CalculatorService(calculator_pb2_grpc.CalculatorServicer):
    
    def Add(self, request, context):
        result = request.num1 + request.num2
        return calculator_pb2.AddResponse(result=result)
    
    def Multiply(self, request, context):
        result = request.num1 * request.num2
        return calculator_pb2.MultiplyResponse(result=result)
    
    def SquareStream(self, request_iterator, context):
        for number in request_iterator:
            squared = number.value ** 2
            yield calculator_pb2.Number(value=squared)

def serve():
    server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
    calculator_pb2_grpc.add_CalculatorServicer_to_server(
        CalculatorService(), server)
    server.add_insecure_port('[::]:50051')
    server.start()
    print("服务端已启动，监听端口 50051...")
    try:
        while True:
            time.sleep(86400)
    except KeyboardInterrupt:
        server.stop(0)

if __name__ == '__main__':
    serve()

实现gRPC客户端

import grpc
import calculator_pb2
import calculator_pb2_grpc

def run():
    channel = grpc.insecure_channel('localhost:50051')
    stub = calculator_pb2_grpc.CalculatorStub(channel)
    
    # 一元RPC调用
    add_response = stub.Add(calculator_pb2.AddRequest(num1=5, num2=3))
    print(f"5 + 3 = {add_response.result}")
    
    multiply_response = stub.Multiply(calculator_pb2.MultiplyRequest(num1=4, num2=6))
    print(f"4 * 6 = {multiply_response.result}")
    
    # 双向流式RPC
    numbers = [calculator_pb2.Number(value=i) for i in range(1, 6)]
    response_stream = stub.SquareStream(iter(numbers))
    
    print("双向流处理:")
    for number, response in zip(numbers, response_stream):
        print(f"客户端发送: {number.value} -> 服务端返回: {response.value}")

if __name__ == '__main__':
    run()

5 + 3 = 8.0
4 * 6 = 24.0
双向流处理:
客户端发送: 1 -> 服务端返回: 1.0
客户端发送: 2 -> 服务端返回: 4.0
客户端发送: 3 -> 服务端返回: 9.0
客户端发送: 4 -> 服务端返回: 16.0
客户端发送: 5 -> 服务端返回: 25.0