Python之元类

类型对象负责创建对象实例,控制对象行为。那么类型对象又由谁来创建呢?

 

元类(metaclass)——类型的类型

New-Style Class的默认类型是type

>>> class Data(object):
...     pass
... 
>>> 
>>> Data.__class__
<type 'type'>
>>> 
>>> type.__class__
<type 'type'>

#最终的类型是type,包括type自己

关键字class会被编译成元类创建类型对象指令

>>> Data = type("Data",(object),{"x":1})
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: type() argument 2 must be tuple, not type
>>> Data = type("Data",(object,),{"x":1})#class的实际行为
>>> 
>>> Data.x
1
>>> Data.__class__
<type 'type'>
>>> 
>>> Data.__base__
<type 'object'>
class type(object)
 |  type(object) -> the object's type
 |  type(name, bases, dict) -> a new type

 

正因为class和def一样是指令,可以在任何地方创建类型对象。

>>> def test():
...     class Data(object):pass
...     return Data
... 
>>> 
>>> Data = test()
>>> 
>>> Data.__name__
'Data'
>>> 
>>> type(Data)
<type 'type'>
>>> 
>>> Data()
<__main__.Data object at 0x7f822e1d2490>

元类、类型以及实例的创建顺序:

class = metaclass(....) #元类创建类型
instance = class(...)   #类型创建实例

instance.__class__ is class #实例的类型
class.__class__ is metaclass #类型的类型

 

__metaclass__

除了使用默认元类type以外,还可以用__metaclass__属性指定自定义元类,以便对类型对象创建过程进行干预。

#!/usr/bin/env python26
#-*- coding:utf-8 -*-

class InjectMeta(type):
        def __new__(cls,name,bases,attrs):
                t = type.__new__(cls,name,bases,attrs)

                def print_id(self):print hex(id(self))

                t.print_id = print_id #为类型对象添加实例方法
                t.s = "Hello,world!" #添加静态字段

                return t

class Data(object):
        __metaclass__ = InjectMeta #显示指定元类

print Data.__metaclass__
print Data.__class__
print Data.s

print dir(Data)

Data().print_id()

输出:

<class '__main__.InjectMeta'>
<class '__main__.InjectMeta'>
Hello,world!
['__class__', '__delattr__', '__dict__', '__doc__', '__format__', '__getattribute__', '__hash__', '__init__', '__metaclass__', '__module__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', 'print_id', 's']
0x7f3687df7250

自定义元类通常都从type继承,习惯以Meta结尾。和继承有点类似。

只需注意__new和__init__方法参数的区别:

http://www.cnblogs.com/gsblog/p/3368304.html

当解释器创建类型对象时,会按以下顺序查找__metaclass__属性:

class.__metaclass__->bases.__metaclass__->module.__metaclass__->type

 magic

对象行为由类型决定。实例不过存储状态数据。控制类型对象的创建,也就意味着可以让对象的实际行为和代码存在极大的差异。这是魔法的力量。

静态类(static class):不允许创建实例。通常作为工具类(Utility)存在

[root@typhoeus79 20131014]# more static_class.py
#!/usr/bin/env python26
#-*- coding:utf-8 -*-

class StaticClassMeta(type):
        def __new__(cls,name,bases,attr):
                t = type.__new__(cls,name,bases,attr)

                def ctor(cls,*args,**kwargs):
                        raise RuntimeError("Cannot be created a instance of the static class!")
                t.__new__ = staticmethod(ctor)

                return t
class Data(object):
        __metaclass__ = StaticClassMeta

Data()
[root@typhoeus79 20131014]# ./static_class.py   
Traceback (most recent call last):
  File "./static_class.py", line 16, in <module>
    Data()
  File "./static_class.py", line 9, in ctor
    raise RuntimeError("Cannot be created a instance of the static class!")
RuntimeError: Cannot be created a instance of the static class!

 

密封类(sealed class):禁止被继承

>>> class SealedClassMeta(type):
...     _types = set()
...     def __init__(cls,name,bases,attrs):
...             if cls._types & set(bases):
...                     raise SyntaxError("Cannot inherit from a sealed class!")
...             cls._types.add(cls)
... 
>>> 
>>> class A(object):                    
...     __metaclass__ = SealedClassMeta
... 
>>> class B(A):pass
... 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 5, in __init__
SyntaxError: Cannot inherit from a sealed class!

 

posted @ 2013-10-14 14:26  小郭学路  阅读(561)  评论(0编辑  收藏  举报