Protobuf data exchange format

brief introduction

  ProtoBuf（Google Protocol Buffer） By google Sequential structured data formats used by companies for data exchange , With cross platform 、 Cross language 、 Scalability , The type is suitable for the commonly used XML And JSON,
But with a smaller transport volume 、 Higher coding 、 Decoding ability , Especially suitable for data storage 、 Network data transmission and so on to the storage volume 、 Areas with high real-time requirements . With .proto For the suffix , Has its own compiler protoc, protoc2 and protoc3 edition .
Simple understanding : comparison json More efficient , Efficient serialization and deserialization , A data exchange format that can compress data into smaller sizes ( For the time being, it is simply understood as more efficient and powerful json).gRPC The default is based on protobuf Data format for data exchange , therefore
protobuf You must learn before you become gRPC Prepare for learning . Write it well .protoc After file , Use protoc And according to the input parameters , Select the type of language to convert , Can be converted into the code of various programming languages .

protobuf Examples

   Next, there is no need to go over the grammar word by word , I pass the commonly used grammar through a protobuf Documents can be elaborated , Make a summary . Just go to the official for detailed documents . I'm sure to learn in half an hour ,
It's not as complicated as expected , He Xue json almost . The details are as follows :

//  notes : 1. // Single-line comments  2. /* */  Multiline comment 
syntax = "proto3"; //  Use... Explicitly protobuf edition   Do not write by default   It is proto2
package com.test; //  Definition package package

//  Official documents : https://developers.google.com/protocol-buffers/docs/proto3

import "./pk2.proto"; // Import protoW file 

option go_package = "proto"; //  If you use protoc Compiled into java Code   Then generate according to the package name   Ignore  com.test
option java_package = "com.hello"; //  If it is compiled as go Code   According to com.hello Package name generation 

//  Definition message Information 
message Keyword {
    
    string value = 1;
}

//  Define a message message Parameters 
message QueryData{
    

    //  Basic data type field definitions   Basic data types can cover  string bool int double bytes wait 
    string name = 1;
    string orderBy = 2;
    double weight = 3;
    float searchWeight = 4;
    int32 type = 5;
    uint64 id = 6;
    fixed32 type2 = 7;
    fixed64 type3 = 8;
    bool isOk = 9;
    bytes passWord = 10;

    //  Define an array   similar  []string
    repeated string cases = 11;

    //  Definition map type  string=>string
    map <string,string> myMap = 15;

    // oneof Optional fields   There can only be one or 0 individual 
    oneof OptionField {
    
        string selectWord = 12;
        string queryType = 13;
    }

    //  Define enumeration types 
    enum WeekDay {
    
        SUNDAY = 0; //  Enumeration type   The first value must be 0
        MONDAY = 1;
        TUESDAY = 2;
    }
    WeekDay selectDay = 14;

    // keyword  Compound field type   Quote the above defined Keyword type 
    Keyword kw = 16;
}

//  Define interface return message Format 
message QueryResponse {
    
    int32 code = 1;
    string msg = 2;
    string htmlData = 3;
}

//  Definition rpc service   Generally refer to gRPC More 
service QueryApi {
    

    // 1.grpc One yuan in rpc Interface   Very common   It can be understood as normal http The interface is almost 
    rpc Query (QueryData) returns (QueryResponse){
    };

    // 2. Server based " streaming "rpc Interface   After establishing the connection   The server can send data to the client 
    rpc ServerQueryStyle(QueryData) returns (stream QueryResponse) {
    };

    // 3. Server based " streaming "rpc Interface   Different from the first one   This way is based on " flow " transmission   Instead of sending a request response and then disconnecting 
    rpc ClientQueryStyle(stream QueryData) returns (QueryResponse) {
    };

    // 4. client-based , Server side " Two way flow "rpc Interface   Full duplex communication   Both parties can send data to each other 
    rpc ClientAndServerStreamQuery(stream QueryData) returns (stream QueryResponse){
    };
}