python子类增加属性_python – 使用动态base_class在__new__子类中设置属性

我有一个类A,它子类str通过kwargs添加元数据

class A(str):

def __new__(cls, value, **kwargs):

obj = str.__new__(cls, value)

obj.__dict__.update(kwargs)

return obj

喜欢使用：

x = A('test', meta=10)

x.meta

10

现在,我也想处理unicode类型,所以我采用了以下方法.动态检查类型并将其用作base_class

class B(object):

def __new__(cls, value, **kwargs):

base_class = str if isinstance(value, str) else unicode

new_type = type(cls.__name__, (base_class,), dict(cls.__dict__))

obj = base_class.__new__(new_type, value)

obj.__dict__.update(kwargs)

return obj

但是,这似乎是不正确的做法.给我一个

TypeError: descriptor ‘dict‘ for ‘B’ objects doesn’t apply to ‘B’ object

B('test', meta=10)

-----------------------------------------------------------------------

TypeError Traceback (most recent call last)

in ()

----> 1 B('tes', a=10).a

in __new__(cls, value, **kw)

11 new_type = type(cls.__name__, (base_class,), dict(cls.__dict__))

12 obj = base_class.__new__(new_type, value)

---> 13 obj.__dict__.update(kw)

14 return obj

15

TypeError: descriptor '__dict__' for 'B' objects doesn't apply to 'B' object

我可以使用setattr代替C语言,这很好.

class C(object):

def __new__(cls, value, **kw):

base_class = str if isinstance(value, str) else unicode

new_type = type(cls.__name__, (base_class,), dict(cls.__dict__))

obj = base_class.__new__(new_type, value)

for k, v in kw.items():

setattr(obj, k, v)

return obj

x = C('test', meta=10)

x.meta

10

你能帮我理解我在B级上错过了什么以及如何使用__dict__吗？