Python实现单例模式-CSDN博客

2019独角兽企业重金招聘Python工程师标准>>>

定义

An implementation of the singleton pattern must:

ensure that only one instance of the singleton class ever exists; and
provide global access to that instance.

单例模式就是保证类在整个使用过程中有且只有一个实例.

Python原理

要想保证类的使用过程中有且只有一个实例, 就需要对Python中类的实例化, 以及实例化后的使用过程有所了解. 这个过程主要涉及到3个方法:

__new__
__init__
__call__

这3个方法的调用过程是: 首先调用__new__, 来生成新的实例; 其次, 调用__init__对新实例进行相关设置; 最后, 如果类中实现了__call__方法的话, 那么类的实例也是一个可调用对象, 调用实例时instance()其实就是调用实例所属的类类的__call__方法.

清楚了3个方法的调用时机, 再结合单例模式的定义就不难实现Python版本的单例模式了.

Python实现

Python中实现单例模式主要有一下几种方式:

__call__
__new__

`call`

在python中一切都是对象, 类其实也是一个对象, 一个类是其元类(MetaClass)的实例(定义类的时候其实就会调用元类的__init__方法).

这样理解过来, 类的实例初始化实际上就是通过调用了类的元类的__call__方法实现的.

# coding: utf-8
"""
代码摘在mininet源代码
"""

class Singleton( type ):
    """Singleton pattern from Wikipedia
       See http://en.wikipedia.org/wiki/Singleton_Pattern
       Intended to be used as a __metaclass_ param, as shown for the class
       below."""

    def __new__( cls, *args, **kw ):
        return super( Singleton, cls ).__new__( cls, *args, **kw )
    
    def __init__( cls, name, bases, dict_ ):
        super( Singleton, cls ).__init__( name, bases, dict_ )
        cls.instance = None

    def __call__( cls, *args, **kw ):
        if cls.instance is None:
            cls.instance = super( Singleton, cls ).__call__( *args, **kw )
        return cls.instance


class MininetLogger( Logger, object ):
    """Mininet-specific logger
       Enable each mininet .py file to with one import:
       from mininet.log import [lg, info, error]
       ...get a default logger that doesn't require one newline per logging
       call.
       Inherit from object to ensure that we have at least one new-style base
       class, and can then use the __metaclass__ directive, to prevent this
       error:
       TypeError: Error when calling the metaclass bases
       a new-style class can't have only classic bases
       If Python2.5/logging/__init__.py defined Filterer as a new-style class,
       via Filterer( object ): rather than Filterer, we wouldn't need this.
       Use singleton pattern to ensure only one logger is ever created."""

    __metaclass__ = Singleton

    def __init__( self ):

        Logger.__init__( self, "mininet" )

        # create console handler
        ch = StreamHandlerNoNewline()
        # create formatter
        formatter = logging.Formatter( LOGMSGFORMAT )
        # add formatter to ch
        ch.setFormatter( formatter )
        # add ch to lg
        self.addHandler( ch )

        self.setLogLevel()

    def setLogLevel( self, levelname=None ):
        """Setup loglevel.
           Convenience function to support lowercase names.
           levelName: level name from LEVELS"""
        level = LOGLEVELDEFAULT
        if levelname is not None:
            if levelname not in LEVELS:
                raise Exception( 'unknown levelname seen in setLogLevel' )
            else:
                level = LEVELS.get( levelname, level )

        self.setLevel( level )
        self.handlers[ 0 ].setLevel( level )

    # pylint: disable=method-hidden
    # "An attribute inherited from mininet.log hide this method" (sic)
    # Not sure why this is occurring - this function definitely gets called.

    # See /usr/lib/python2.5/logging/__init__.py; modified from warning()
    def output( self, msg, *args, **kwargs ):
        """Log 'msg % args' with severity 'OUTPUT'.
           To pass exception information, use the keyword argument exc_info
           with a true value, e.g.
           logger.warning("Houston, we have a %s", "cli output", exc_info=1)
        """
        if self.manager.disable >= OUTPUT:
            return
        if self.isEnabledFor( OUTPUT ):
            self._log( OUTPUT, msg, args, kwargs )

    # pylint: enable=method-hidden

每次实例化MininetLogger时, 其实调用了Singleton的__call__方法, 这个可以在Singleton和MininetLogger类中相关方法添加打印日志进行验证.

当然Singleton还有其他的定义方式:

# https://stackoverflow.com/questions/6760685/creating-a-singleton-in-python
class Singleton(type):
    _instances = {}
    def __call__(cls, *args, **kwargs):
        if cls not in cls._instances:
            cls._instances[cls] = super(Singleton, cls).__call__(*args, **kwargs)
        return cls._instances[cls]

但是这种方式在多个类(A, B)同时指向Singleton为__metaclass__的话, 当实例化A后, 可以通过B._instances看到到A的instance.

`new`

每次实例化一个类的新实例的时候其实会调用类的__new__方法, 这样就可以通过在类的__new__方法中进行唯一实例的控制.

# https://stackoverflow.com/questions/6760685/creating-a-singleton-in-python

class Singleton(object):
    _instance = None
    def __new__(cls, *args, **kwargs):
        if not isinstance(cls._instance, cls):
            cls._instance = object.__new__(cls, *args, **kwargs)
        return cls._instance
        
class A(Singleton):
    pass
    
id(A()) == id(A())
# True