我在Django上构建SPA,我有一个巨大的函数,有很多if语句用于检查我的对象字段的状态名称.像这样:
if self.state == 'new': do some logicif self.state == 'archive': do some logic
等等.我现在阅读好书“Fluent python”,我提到了@singledispatch装饰器,它看起来很棒,但它只能用不同类型的参数如str,int等覆盖函数.
问题是,如果在python或Django方式中分离逻辑就像在我的庞大函数中使用像singledispatch那样的重叠函数吗?
解决方法:
有,但你必须写它.一种可能性是创建一个基于instance.state或任何选择的state_attr进行调度的descriptor:
class StateDispatcher(object): def __init__(self, state_attr='state'): self.registry = {} self._state_attr = state_attr def __get__(self, instance, owner): if instance is None: return self method = self.registry[getattr(instance, self._state_attr)] return method.__get__(instance, owner) def register(self, state): def decorator(method): self.registry[state] = method return method return decorator
https://docs.python.org/3/howto/descriptor.html#functions-and-methods:
To support method calls, functions include the
__get__()
method for binding methods during attribute access. This means that all functions are non-data descriptors which return bound or unbound methods depending whether they are invoked from an object or a class.
在您的有状态类中,您可以创建一个调度程序并注册方法:
class StateMachine(object): dispatcher = StateDispatcher() state = None @dispatcher.register('test') def test(self): print('Hello, World!', self.state) @dispatcher.register('working') def do_work(self): print('Working hard, or hardly working?', self.state)
让我们看看它的实际效果:
>>> sm = StateMachine()>>> sm.state = 'test'>>> sm.dispatcher()Hello, World! test>>> sm.state = 'working'>>> sm.dispatcher()Working hard, or hardly working? working>>> sm.state = None>>> sm.dispatcher()Traceback (most recent call last): ... File "dispatcher.py", line 11, in __get__ method = self.registry[getattr(instance, self._state_attr)]KeyError: None
请注意,这是一种基于状态调度的非常恶劣的方法,因为对于未来的代码读者来说,整个机制将很难遵循.
另一种调度文本状态的方法是对方法名称中的状态进行编码,并在调度函数中根据该方法选择正确的方法.许多python类使用此模式(例如ast.NodeVisitor
):
class StateMachine(object): def dispatch(self, *args, **kwgs): getattr(self, 'do_{}'.format(self.state))(*args, **kwgs) def do_new(self): print('new') def do_archive(self): print('archive')sm = StateMachine()sm.state = 'new'sm.dispatch()sm.state = 'archive'sm.dispatch()
来源:http://www.icode9.com/content-1-231851.html
联系客服