类方法对象如何工作? [英] How does a classmethod object work?
问题描述
我很难理解类方法对象在Python中是如何工作的，尤其是在元类和__new__中.在我的特殊情况下，当我遍历赋予__new__的members时，我想获得一个classmethod成员的名称.
对于普通方法，名称只是存储在__name__属性中，但是对于类方法，显然没有这样的属性.我什至看不到如何调用classmethod，因为也没有__call__属性.
有人可以向我解释类方法的工作原理或为我提供一些文档吗?谷歌搜索使我无处可去.谢谢！
classmethod对象是描述符.您需要了解描述符的工作原理.
__get__的对象，该方法带有三个参数:self，instance和instance type.
在常规属性查找期间，如果查找对象A具有方法__get__，则会调用该方法，并将其返回的内容替换为对象A.这就是当您在对象上调用方法时，函数(也是描述符)如何成为绑定方法的方法.
class Foo(object):
def bar(self, arg1, arg2):
print arg1, arg2
foo = Foo()
# this:
foo.bar(1,2) # prints '1 2'
# does about the same thing as this:
Foo.__dict__['bar'].__get__(foo, type(foo))(1,2) # prints '1 2'
classmethod对象的工作方式相同.查找时，将调用其__get__方法.类方法的__get__丢弃与instance相对应的参数(如果有的话)，并且仅当在包装函数上调用__get__时才通过instance_type传递.
说明性涂鸦:
In [14]: def foo(cls):
....: print cls
....:
In [15]: classmethod(foo)
Out[15]: <classmethod object at 0x756e50>
In [16]: cm = classmethod(foo)
In [17]: cm.__get__(None, dict)
Out[17]: <bound method type.foo of <type 'dict'>>
In [18]: cm.__get__(None, dict)()
<type 'dict'>
In [19]: cm.__get__({}, dict)
Out[19]: <bound method type.foo of <type 'dict'>>
In [20]: cm.__get__({}, dict)()
<type 'dict'>
In [21]: cm.__get__("Some bogus unused string", dict)()
<type 'dict'>
有关描述符的更多信息，可以在这里(以及其他地方)找到: http://users.rcn.com/python/download/Descriptor.htm
对于获取由classmethod包装的函数名称的特定任务:
In [29]: cm.__get__(None, dict).im_func.__name__
Out[29]: 'foo'
I'm having trouble to understand how a classmethod object works in Python, especially in the context of metaclasses and in __new__. In my special case I would like to get the name of a classmethod member, when I iterate through the members that were given to __new__.
For normal methods the name is simply stored in a __name__ attribute, but for a classmethod there is apparently no such attribute. I don't even see how the classmethod is invoked, as there is no __call__ attribute either.
Can someone explain to me how a classmethod works or point me to some documentation? Googling led me nowhere. Thanks!
A classmethod object is a descriptor. You need to understand how descriptors work.
In a nutshell, a descriptor is an object which has a method __get__, which takes three arguments: self, an instance, and an instance type.
During normal attribute lookup, if a looked-up object A has a method __get__, that method gets called and what it returns is substituted in place for the object A. This is how functions (which are also descriptors) become bound methods when you call a method on an object.
class Foo(object):
def bar(self, arg1, arg2):
print arg1, arg2
foo = Foo()
# this:
foo.bar(1,2) # prints '1 2'
# does about the same thing as this:
Foo.__dict__['bar'].__get__(foo, type(foo))(1,2) # prints '1 2'
A classmethod object works the same way. When it gets looked up, its __get__ method gets called. The __get__ of a classmethod discards the argument corresponding to the instance (if there was one) and only passes along the instance_type when it calls __get__ on the wrapped function.
A illustrative doodle:
In [14]: def foo(cls):
....: print cls
....:
In [15]: classmethod(foo)
Out[15]: <classmethod object at 0x756e50>
In [16]: cm = classmethod(foo)
In [17]: cm.__get__(None, dict)
Out[17]: <bound method type.foo of <type 'dict'>>
In [18]: cm.__get__(None, dict)()
<type 'dict'>
In [19]: cm.__get__({}, dict)
Out[19]: <bound method type.foo of <type 'dict'>>
In [20]: cm.__get__({}, dict)()
<type 'dict'>
In [21]: cm.__get__("Some bogus unused string", dict)()
<type 'dict'>
More info on descriptors can be found here (among other places): http://users.rcn.com/python/download/Descriptor.htm
For the specific task of getting the name of the function wrapped by a classmethod:
In [29]: cm.__get__(None, dict).im_func.__name__
Out[29]: 'foo'
这篇关于类方法对象如何工作?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
