51工具盒子

依楼听风雨
笑看云卷云舒,淡观潮起潮落

Python 实现动态的工厂模式

本文讨论的主要是,如何把工厂模式生产的产品,即抽象类的子类信息动态地保存到一个表中 ,而不是直接简单粗暴地使用 if-else 来判断。可能是 Java 给我的影响比较大,我还比较喜欢用抽象类抽象方法这些东西,至于好坏应该由实际情况来决定,这里就不详细展开讨论了。

那么本文所使用的示例来自于我一个分析 QQ 聊天记录文件的项目,因为它对于群聊的聊天记录和私聊的聊天记录格式不同,需要分别处理,所以就到了本文所讨论的工厂模式。

Hook {#Hook}

通过 __init_subclass__ 来实现注册的功能。

|---------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 | from typing import Type from abc import ABC, abstractmethod class Shape(ABC): _shapes: dict[str, Type['Shape']] = {} def __init_subclass__(cls, variety: str) -> None: super().__init_subclass__() cls._shapes[variety] = cls @classmethod def make(cls, variety: str) -> 'Shape': return cls._shapes[variety]() @abstractmethod def draw(self): raise NotImplementedError() class Circle(Shape, variety='circle'): def draw(self): print('Drawing a circle') class Rectangle(Shape, variety='rectangle'): def draw(self): print('Drawing a rectangle') if __name__ == '__main__': Shape.make('rectangle').draw() Shape.make('circle').draw() |

装饰器 {#装饰器}

这里不是介绍装饰器 ,只是通过装饰器的形式来保存子类信息。我这里省略了 Parser 中的其它方法,只保留抽象方法。

|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 | from abc import ABC, abstractmethod from typing import Type, Callable class Parser(ABC): @abstractmethod def extract_id(self, line: str) -> str: raise NotImplementedError() @abstractmethod def get_display_name(self, extracted_id: str) -> str: raise NotImplementedError() class ParserFactory: def __init__(self) -> None: self.parsers: dict[str, Type[Parser]] = {} def register(self, parser_name: str) -> Callable[[Type[Parser]], Type[Parser]]: def wrapper(parser: Type[Parser]) -> Type[Parser]: self.parsers[parser_name] = parser return parser return wrapper def get_instance(self, name: str) -> Parser: if parser := self.parsers.get(name): return parser() raise NameError(f'unknown parser name: {name}') factory = ParserFactory() @factory.register('group') class GroupParser(Parser): def extract_id(self, line: str) -> str: return 'GROUP ID' def get_display_name(self, extracted_id: str) -> str: return f'GROUP NAME {extracted_id}' @factory.register('private') class PrivateParser(Parser): def extract_id(self, line: str) -> str: return 'PRIVATE ID' def get_display_name(self, extracted_id: str) -> str: return f'PRIVATE NAME {extracted_id}' def process(parser_name: str) -> None: parser = factory.get_instance(parser_name) print(parser.extract_id('ABC')) print(parser.get_display_name('DEF')) if __name__ == '__main__': process('private') process('group') |

这种方式比较直观,但需要额外创建一个 factory 全局变量。如果你非常讨厌全局变量,我还有另一种方式。

|------------------------------------------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 | from typing import Type class ParserFactory: parsers: dict[str, Type[Parser]] = {} def __init__(self, name: str) -> None: self.name = name def __call__(self, parser: Type[Parser]) -> Type[Parser]: # 如果对使用 self 访问类属性感到不舒服,你可以 # 用 self.__class__.parsers 或者 ParserFactory.parsers self.parsers[self.name] = parser return parser @classmethod def get_instance(cls, name: str) -> Parser: if parser := cls.parsers.get(name): return parser() raise NameError(f'unknown parser name: {name}') @ParserFactory('group') class GroupParser(Parser): def extract_id(self, line: str) -> str: return 'GROUP ID' def get_display_name(self, extracted_id: str) -> str: return f'GROUP NAME {extracted_id}' @ParserFactory('private') class PrivateParser(Parser): def extract_id(self, line: str) -> str: return 'PRIVATE ID' def get_display_name(self, extracted_id: str) -> str: return f'PRIVATE NAME {extracted_id}' def process(parser_name: str) -> None: parser = ParserFactory.get_instance(parser_name) print(parser.extract_id('ABC')) print(parser.get_display_name('DEF')) |

当然,也可以定义一个类方法 register 使代码语义更清晰,这里不再演示。

元类 {#元类}

我超,原!

这里有一个坑,就是 ParserMeta 需要继承自 abc.ABCMeta,因为 Parser 是继承自 abc.ABC 的抽象类,下文有详细解释。

|------------------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | from abc import ABCMeta from typing import Type, cast class ParserMeta(ABCMeta): parsers: dict[str, Type[Parser]] = {} def __new__(mcs, name, bases, attrs) -> Type[Parser]: assert Parser in bases assert '__parser_name__' in attrs t = cast(Type[Parser], super().__new__(mcs, name, bases, attrs)) mcs.parsers[attrs['__parser_name__']] = t return t @classmethod def get_instance(mcs, name: str) -> Parser: if parser := mcs.parsers.get(name): return parser() raise NameError(f'unknown parser name: {name}') class GroupParser(Parser, metaclass=ParserMeta): __parser_name__ = 'group' def extract_id(self, line: str) -> str: return 'GROUP ID' def get_display_name(self, extracted_id: str) -> str: return f'GROUP NAME {extracted_id}' class PrivateParser(Parser, metaclass=ParserMeta): __parser_name__ = 'private' def extract_id(self, line: str) -> str: return 'PRIVATE ID' def get_display_name(self, extracted_id: str) -> str: return f'PRIVATE NAME {extracted_id}' def process(parser_name: str) -> None: parser = ParserMeta.get_instance(parser_name) print(parser.extract_id('ABC')) print(parser.get_display_name('DEF')) |

如果直接让 ParserMeta 继承 type,那么会喜提下方的报错信息:

|-------------|---------------------------------------------------------------------------------------------------------------------------------------------| | 1 2 | TypeError: metaclass conflict: the metaclass of a derived class must be a (non-strict) subclass of the metaclasses of all its bases |

其实这句话我只看明白了大概意思,就是需要保证元类之间的继承关系,需要使 ParserMeta 继承与 abc.ABCMeta 即可解决。

具体的分析如下:首先,元类的本质是默认创建类的时候会调用 type,而指定元类后就调用指定的那个类。由于以下的继承关系(比较简单就不用专门的作图软件了,省略 PrivateParser,它和图示一样):

|-------------------|---------------------------------------------------------------------| | 1 2 3 4 5 | abc.ABC <- abc.ABCMeta | Parser | GroupParser <- ParserMeta |

如果 ParserMeta 不继承于 abc.ABCMeta,创建 GroupParser 时就无法判断使用 abc.ABCMeta 还是 ParserMeta,所以导致了上文中的报错。

题外话:依赖注入 {#题外话:依赖注入}

其实本文内容和依赖注入没什么关系,但还是比较想提一下。

依赖注入,就是入参传一个 interface,结束。

哈哈,其实没这么简单,大家可以参考维基百科给出的解释,就是 Spring 那一套,只不过注入的时候需要手动指定。

本文中给的情景就是一个很好的解释依赖注入的例子。如果 Parser 类中有很多具体方法和抽象方法,那可以把这些抽象方法单独提取出来一个 interface,然后把 Parser 变成具体类,构造时接受这个 interface 为参数,通过这个 interace 来提取 id 之类的信息。

但是,由于 Python 中不存在 interface 这个概念,实际上我这个 Parser 类中方法也不是太多,如果硬要使用依赖注入反而会让代码更复杂,所以最终我就没用。

赞(1)
未经允许不得转载:工具盒子 » Python 实现动态的工厂模式