开闭原则

开：对扩展开放
闭：对修改关闭

对象增强方式

一般来说实现对象增强的方式有3种：

• 继承
• 代理
• 装饰器

装饰器模式

函数装饰

#include <iostream>
#include <functional>
using namespace std;

template <typename R, typename... Args>
struct Logger3
{
    Logger3(function<R(Args...)> func) : func_(func)
    {
    }

    R operator()(Args... args)
    {
        cout << "start" << endl;
        R result = func_(args...);
        cout << "end" << endl;
        return result;
    }

    function<R(Args...)> func_;
};

template <typename R, typename... Args>
auto make_logger3(R (*func)(Args...))
{
    return Logger3<R, Args...>(std::function<R(Args...)>(func));
}

double add(double a, double b)
{
    return a + b;
}

long my_abs(int a)
{
    return a > 0 ? a : -a;
}

int main()
{
    auto caller = make_logger3(add);
    auto result = caller(2.1, 3.2);
    cout << result << endl;

    auto caller2 = make_logger3(my_abs);
    auto result2 = caller2(-100);
    cout << result2 << endl;
}

python 装饰器

装饰器本质是个python函数，是由闭包支持。

• 常规写法

@decorator
def func(*args, **kwargs):
    do_something()

• 应用

  • 函数执行时间统计

    import time
    def calculate_function_run_time(func):
        def call_fun(*args, **kwargs):
            start_time = time.time()
            f = func(*args, **kwargs)
            end_time = time.time()
            print('{0} run time：{1} ms'.format(func.__name__, 1000 * int(end_time - start_time)))
            return f
      
        return call_fun
    
    
    @calculate_function_run_time
    def test1():
        for i in range(100000000):
            i = i + 1
    
    if __name__ == "__main__":
        test1()
    

  • 引入日志

  • 执行函数前的预备处理

  • 执行函数后的清理功能

  • 权限校验

    user_list = [
    {'name': 'sb1', 'passwd': '123'},
    {'name': 'sb2', 'passwd': '123'},
    {'name': 'sb3', 'passwd': '123'},
    {'name': 'sb4', 'passwd': '123'}]
    
    client_dic = {'username': None, 'login': False}
    
    def auth_func(func):
        def wrapper(*args, **kwargs):
            if client_dic['username'] and client_dic['login']:
                res = func(*args, **kwargs)
                return res
    
            user_name = 'sb1'
            password = '123'
    
        # 对比列表, 检查用户名和密码是否正确
        for user_dic in user_list:
            if user_name == user_dic['name'] and password == user_dic['passwd']:
                client_dic['username'] = user_dic['name']
                client_dic['login'] = True
                res = func(*args, **kwargs)
                return res
        else:
            print('用户名或者密码错误!')
    
    return wrapper
    
    
    @auth_func
    def index():
        print("欢迎来到主页")
    
    
    @auth_func
    def home(name):
        print("欢迎回家:%s" % name)
    
    
    if __name__ == '__main__':
        index()
        home('sb1')
    

  • 缓存

  • 事物处理

Java IO流中的装饰器

image.png

和代理模式的异同

两者都是对类的方法进行扩展，但装饰器模式强调的是增强自身，在被装饰之后nenggo

参考资料

1. https://www.jianshu.com/p/d80b6b4b76fc
2. https://blog.csdn.net/qq_33531400/article/details/79324551
3. https://mp.weixin.qq.com/s?__biz=Mzg5ODAwODM0Mg%3D%3D&mid=2247483679&idx=1&sn=0469fc64b030ad57660294bba1a1c453&chksm=c0685566f71fdc7059d6f92ecc82097d8a556c7e8d97c3a5a5e6b3b6bcc1e08735a17d828343&token=1492230032&lang=zh_CN