Python 是否可以在基类中定义的装饰器生成的子类中检索值？

我有一个子类方法，我想有条件地短路。我试图做的是类似这样的事情，只是我想把验证逻辑放到基类中

class BaseClass(object):
    def getvalue(self):
        return True
    def validate(self):
        validated = self.getvalue()
        return validated

class ExtendedClass1(BaseClass):
    def do_some_work(self):
        validated = self.validate()
        if not validated:
            print "Not validated."
            return
        print "Things are validated if the method got this far.", validated

class ExtendedClass2(BaseClass):
    def do_some_work(self):
        validated = self.validate()
        if not validated:
            print "Not validated."
            return
        print "Things are validated if the method got this far.", validated

class ExtendedClass3(BaseClass):
    def do_some_work(self):
        print "This one doesn't require validation."

work1 = ExtendedClass1()
work1.do_some_work()

work2 = ExtendedClass2()
work2.do_some_work()

work3 = ExtendedClass3()
work3.do_some_work()

在下面的示例中，我能够将一些重复的代码转换为装饰器模式

class BaseClass(object):

    def validate(input_function):
        def wrapper(*args,**kwargs):
            validated = True
            if not validated:
                print "Not validated."
                return
            input_function(*args, **kwargs)
        return wrapper

    validate = staticmethod(validate)

class ExtendedClass1(BaseClass):
    @BaseClass.validate
    def do_some_work(self):
        print "Things are validated if the method got this far."

class ExtendedClass2(BaseClass):
    @BaseClass.validate
    def do_some_work(self):
        print "Things are validated if the method got this far."

class ExtendedClass3(BaseClass):
    def do_some_work(self):
        print "This one doesn't require validation."

work1 = ExtendedClass1()
work1.do_some_work()

work2 = ExtendedClass2()
work2.do_some_work()

work3 = ExtendedClass3()
work3.do_some_work()

但是，我需要在decorator中调用基类的方法来执行验证工作，并在子类中检索（validated）的值。在这个示例之后，我修改了decorator，试图让它调用self.getvalue（）。此时它不会出错，但也不起作用，因为self.getvalue（）不会返回True。这看起来似乎比它的价值更麻烦，但现在我很好奇这是否可能

class BaseClass(object):

    def getvalue(self):
        return True

    def validate(self):
        def wrap(input_function):
            def wrapper(*args,**kwargs):
                validated = self.getvalue()
                if not validated:
                    print "Not validated."
                    return
                input_function(*args, **kwargs)
            return wrapper
        return wrap

    validate = staticmethod(validate)

class ExtendedClass1(BaseClass):
    @BaseClass.validate
    def do_some_work(self):
        print "Things are validated if the method got this far."#, validated

class ExtendedClass2(BaseClass):
    @BaseClass.validate
    def do_some_work(self):
        print "Things are validated if the method got this far."#, validated

class ExtendedClass3(BaseClass):
    def do_some_work(self):
        print "This one doesn't require validation."

work1 = ExtendedClass1()
work1.do_some_work()

work2 = ExtendedClass2()
work2.do_some_work()

work3 = ExtendedClass3()
work3.do_some_work()

是否可以使用decorator设置一个属性，然后在以后检索它

                ...
                self.validated = True
                if not self.validated:
                    print "Not validated."
                    return
                ...
print work1.validated
                ...

AttributeError: 'ExtendedClass1' object has no attribute 'validated'

从本质上讲，我想把这个转变为：

class ExtendedClass1(BaseClass):
    def do_some_work(self):
        validated = self.validate()
        if not validated:
            print "Not validated."
            return
        print "Things are validated if the method got this far.", validated

为此：

class ExtendedClass1(BaseClass):
    @BaseClass.validate
    def do_some_work(self):
        print "Things are validated if the method got this far.", validated

使用Zaur Nasibov发布的建议，这个例子满足了我的用例。我仍然有兴趣知道@validate是否可以作为一种方法而不是一个独立的函数来实现，但这就完成了工作

class BaseClass(object):
    def getvalue(self):
        return True

def validate(func):
    def wrapped(self, *args, **kwargs):
        validated = self.getvalue()
        self.validated = validated
        if not validated:
            print "Not validated."
            return
        func(self, *args, **kwargs)
    return wrapped

class ExtendedClass1(BaseClass):
    @validate
    def do_some_work(self,input):
        print "Things are validated if the method got this far.", self.validated, input

class ExtendedClass2(BaseClass):
    @validate
    def do_some_work(self):
        print "Things are validated if the method got this far.", self.validated

class ExtendedClass3(BaseClass):
    def do_some_work(self):
        print "This one doesn't require validation."#, self.validated

work1 = ExtendedClass1()
work1.do_some_work(input="some text")

work2 = ExtendedClass2()
work2.do_some_work()

work3 = ExtendedClass3()
work3.do_some_work()

---

@tponthieux，您可以做的是设置被调用函数（方法）的属性，然后检索它：

简单示例（已更新））：

---

如果装饰方法已经过验证，那么只调用它如何？如果愿意，可以传递

validate

方法的返回值：

class BaseClass(object):
    def getvalue(self):
        return True

    def validate(input_function):
        def wrapper(self, *args, **kwargs):
            self.validated = self.getvalue()
            if not self.validated:
                print "Not validated."
                return
            input_function(self, validated=self.validated, *args, **kwargs)
        return wrapper

    validate = staticmethod(validate)

class ExtendedClass1(BaseClass):
    @BaseClass.validate
    def do_some_work(self, validated=None):
        print "Things are validated if the method got this far.", validated

class ExtendedClass2(BaseClass):
    @BaseClass.validate
    def do_some_work(self, validated=None):
        print "Things are validated if the method got this far.", validated

class ExtendedClass3(BaseClass):
    def do_some_work(self):
        print "This one doesn't require validation."

work1 = ExtendedClass1()
work1.do_some_work()

work2 = ExtendedClass2()
work2.do_some_work()

work3 = ExtendedClass3()
work3.do_some_work()

这里的关键是将

self

添加到

wrapper

函数中。所发生的情况是，您的修饰函数没有绑定到实例（并成为方法），而是由修饰器返回的函数（

wrapper

，在上面的示例中）得到绑定。因此，此函数将在调用时获得传入的

self

（实例）参数！重要的是要记住，

@decorator

所做的只是调用

decorator

传入要修饰的函数，然后用decorator返回的任何函数替换要修饰的函数。在您的示例中，这是

包装器

，对于类来说，这与装饰之前的原始函数没有区别

在上面的示例中，我明确声明了

self

。如果没有，我们也可以从

args

中获取它：

def validate(input_function):
    def wrapper(*args, **kwargs):
        print "args[0] is now the instance (conventionally called 'self')", args[0]
        self = args[0]
        self.validated = self.getvalue()
        if not self.validated:
            print "Not validated."
            return
        input_function(validated=self.validated, *args, **kwargs)

---

还要注意，我们将一个额外的关键字参数传递给wrapped方法，名为

validated

。这是完全可选的，您可以从示例中删除

validated=self.validated

和

validated=None

部分

这看起来很有希望，但我不确定如何让decorator代码访问基类的getvalue（）方法。有人想把这个技巧运用到问题示例中吗？好吧，另一个示例：）但我认为这个肯定会对你有所帮助。你永远不会调用包装的

func

，你只会从包装中返回它。这不会给您带来预期的结果；-）我想你对那里的

*args

和

**kwargs

感到困惑。您需要传递它们才能使用这些属性。我不完全确定您的答案是否符合OP的要求，但至少代码现在在语法上是正确的。：-）我不是反对票，但反对票可能是因为答案没有公布。需要注意的一点是@BaseClass.validate必须在decorator方法之外调用基类中描述的方法

validated=self.getvalue（）

我将编辑问题的第一个代码块，使其更清楚。Martijn，我为没有理由的否决表示歉意。你的答案绝对没有错，事实上我也从中吸取了教训。请你编辑一下，这样我就可以取消“-1”了。谢谢。啊，我首先读到的是“但是，我需要在decorator中调用基类的一个方法来执行验证工作，并在子类中检索（validated）的值。”但是稍后您需要设置self.validated。我将进行更新。@tponthieux:updated以包含OP中的

def getvalue（self）

方法，该方法被代码调整为调用该方法，而不是虚构的

def validate（self）

方法。代码示例按现在的方式执行。

def validate(input_function):
    def wrapper(*args, **kwargs):
        print "args[0] is now the instance (conventionally called 'self')", args[0]
        self = args[0]
        self.validated = self.getvalue()
        if not self.validated:
            print "Not validated."
            return
        input_function(validated=self.validated, *args, **kwargs)