This question is not for the discussion of whether or not the singleton design pattern is desirable, is an anti-pattern, or for any religious wars, but to discuss how this pattern is best implemented in Python in such a way that is most pythonic. In this instance I define ‘most pythonic’ to mean that it follows the ‘principle of least astonishment’.

I have multiple classes which would become singletons (my use-case is for a logger, but this is not important). I do not wish to clutter several classes with added gumph when I can simply inherit or decorate.

Best methods:

Method 1: A decorator

def singleton(class_):
    instances = {}
    def getinstance(*args, **kwargs):
        if class_ not in instances:
            instances[class_] = class_(*args, **kwargs)
        return instances[class_]
    return getinstance

@singleton
class MyClass(BaseClass):
    pass

Pros

• Decorators are additive in a way that is often more intuitive than multiple inheritance.

Cons

• While objects created using MyClass() would be true singleton objects, MyClass itself is a a function, not a class, so you cannot call class methods from it. Also for m = MyClass(); n = MyClass(); o = type(n)(); then m == n && m != o && n != o

Method 2: A base class

class Singleton(object):
    _instance = None
    def __new__(class_, *args, **kwargs):
        if not isinstance(class_._instance, class_):
            class_._instance = object.__new__(class_, *args, **kwargs)
        return class_._instance

class MyClass(Singleton, BaseClass):
    pass

Pros

• It’s a true class

Cons

• Multiple inheritance – eugh! __new__ could be overwritten during inheritance from a second base class? One has to think more than is necessary.

Method 3: A metaclass

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]

#Python2
class MyClass(BaseClass):
    __metaclass__ = Singleton

#Python3
class MyClass(BaseClass, metaclass=Singleton):
    pass

Pros

• It’s a true class
• Auto-magically covers inheritance
• Uses __metaclass__ for its proper purpose (and made me aware of it)

Cons

• Are there any?

Method 4: decorator returning a class with the same name

def singleton(class_):
    class class_w(class_):
        _instance = None
        def __new__(class_, *args, **kwargs):
            if class_w._instance is None:
                class_w._instance = super(class_w,
                                    class_).__new__(class_,
                                                    *args,
                                                    **kwargs)
                class_w._instance._sealed = False
            return class_w._instance
        def __init__(self, *args, **kwargs):
            if self._sealed:
                return
            super(class_w, self).__init__(*args, **kwargs)
            self._sealed = True
    class_w.__name__ = class_.__name__
    return class_w

@singleton
class MyClass(BaseClass):
    pass

Pros

• It’s a true class
• Auto-magically covers inheritance

Cons

• Is there not an overhead for creating each new class? Here we are creating two classes for each class we wish to make a singleton. While this is fine in my case, I worry that this might not scale. Of course there is a matter of debate as to whether it aught to be too easy to scale this pattern…
• What is the point of the _sealed attribute
• Can’t call methods of the same name on base classes using super() because they will recurse. This means you can’t customize __new__ and can’t subclass a class that needs you to call up to __init__.

Method 5: a module

a module file singleton.py

Pros

• Simple is better than complex

Cons

Use a Metaclass

I would recommend Method #2, but you’re better off using a metaclass than a base class. Here is a sample implementation:

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]

class Logger(object):
    __metaclass__ = Singleton

Or in Python3

class Logger(metaclass=Singleton):
    pass

If you want to run __init__ every time the class is called, add

        else:
            cls._instances[cls].__init__(*args, **kwargs)

to the if statement in Singleton.__call__.

A few words about metaclasses. A metaclass is the class of a class; that is, a class is an instance of its metaclass. You find the metaclass of an object in Python with type(obj). Normal new-style classes are of type type. Logger in the code above will be of type class 'your_module.Singleton', just as the (only) instance of Logger will be of type class 'your_module.Logger'. When you call logger with Logger(), Python first asks the metaclass of Logger, Singleton, what to do, allowing instance creation to be pre-empted. This process is the same as Python asking a class what to do by calling __getattr__ when you reference one of it’s attributes by doing myclass.attribute.

A metaclass essentially decides what the definition of a class means and how to implement that definition. See for example http://code.activestate.com/recipes/498149/, which essentially recreates C-style structs in Python using metaclasses. The thread What are some (concrete) use-cases for metaclasses? also provides some examples, they generally seem to be related to declarative programming, especially as used in ORMs.

In this situation, if you use your Method #2, and a subclass defines a __new__ method, it will be executed every time you call SubClassOfSingleton() — because it is responsible for calling the method that returns the stored instance. With a metaclass, it will only be called once, when the only instance is created. You want to customize what it means to call the class, which is decided by it’s type.

In general, it makes sense to use a metaclass to implement a singleton. A singleton is special because is created only once, and a metaclass is the way you customize the creation of a class. Using a metaclass gives you more control in case you need to customize the singleton class definitions in other ways.

Your singletons won’t need multiple inheritance (because the metaclass is not a base class), but for subclasses of the created class that use multiple inheritance, you need to make sure the singleton class is the first / leftmost one with a metaclass that redefines __call__ This is very unlikely to be an issue. The instance dict is not in the instance’s namespace so it won’t accidentally overwrite it.

You will also hear that the singleton pattern violates the “Single Responsibility Principle” — each class should do only one thing. That way you don’t have to worry about messing up one thing the code does if you need to change another, because they are separate and encapsulated. The metaclass implementation passes this test. The metaclass is responsible for enforcing the pattern and the created class and subclasses need not be aware that they are singletons. Method #1 fails this test, as you noted with “MyClass itself is a a function, not a class, so you cannot call class methods from it.”

Python 2 and 3 Compatible Version

Writing something that works in both Python2 and 3 requires using a slightly more complicated scheme. Since metaclasses are usually subclasses of type type, it’s possible to use one to dynamically create an intermediary base class at run time with it as its metaclass and then use that as the baseclass of the public Singleton base class. It’s harder to explain than to do, as illustrated next:

# works in Python 2 & 3
class _Singleton(type):
    """ A metaclass that creates a Singleton base class when called. """
    _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]

class Singleton(_Singleton('SingletonMeta', (object,), {})): pass

class Logger(Singleton):
    pass

An ironic aspect of this approach is that it’s using subclassing to implement a metaclass. One possible advantage is that, unlike with a pure metaclass, isinstance(inst, Singleton) will return True.

Corrections

On another topic, you’ve probably already noticed this, but the base class implementation in your original post is wrong. _instances needs to be referenced on the class, you need to use super() or you’re recursing, and __new__ is actually a static method that you have to pass the class to, not a class method, as the actual class hasn’t been created yet when it is called. All of these things will be true for a metaclass implementation as well.

class Singleton(object):
  _instances = {}
  def __new__(class_, *args, **kwargs):
    if class_ not in class_._instances:
        class_._instances[class_] = super(Singleton, class_).__new__(class_, *args, **kwargs)
    return class_._instances[class_]

class MyClass(Singleton):
  pass

c = MyClass()

Decorator Returning A Class

I originally was writing a comment but it was too long, so I’ll add this here. Method #4 is better than the other decorator version, but it’s more code than needed for a singleton, and it’s not as clear what it does.

The main problems stem from the class being it’s own base class. First, isn’t it weird to have a class be a subclass of a nearly identical class with the same name that exists only in its __class__ attribute? This also means that you can’t define any methods that call the method of the same name on their base class with super() because they will recurse. This means your class can’t customize __new__, and can’t derive from any classes that need __init__ called on them.

When to use the singleton pattern

Your use case is one of the better examples of wanting to use a singleton. You say in one of the comments “To me logging has always seemed a natural candidate for Singletons.” You’re absolutely right.

When people say singletons are bad, the most common reason is they are implicit shared state. While with global variables and top-level module imports are explicit shared state, other objects that are passed around are generally instantiated. This is a good point, with two exceptions.

The first, and one that gets mentioned in various places, is when the singletons are constant. Use of global constants, especially enums, is widely accepted, and considered sane because no matter what, none of the users can mess them up for any other user. This is equally true for a constant singleton.

The second exception, which get mentioned less, is the opposite — when the singleton is only a data sink, not a data source (directly or indirectly). This is why loggers feel like a “natural” use for singletons. As the various users are not changing the loggers in ways other users will care about, there is not really shared state. This negates the primary argument against the singleton pattern, and makes them a reasonable choice because of their ease of use for the task.

Here is a quote from http://googletesting.blogspot.com/2008/08/root-cause-of-singletons.html:

Now, there is one kind of Singleton which is OK. That is a singleton where all of the reachable objects are immutable. If all objects are immutable than Singleton has no global state, as everything is constant. But it is so easy to turn this kind of singleton into mutable one, it is very slippery slope. Therefore, I am against these Singletons too, not because they are bad, but because it is very easy for them to go bad. (As a side note Java enumeration are just these kind of singletons. As long as you don’t put state into your enumeration you are OK, so please don’t.)

The other kind of Singletons, which are semi-acceptable are those which don’t effect the execution of your code, They have no “side effects”. Logging is perfect example. It is loaded with Singletons and global state. It is acceptable (as in it will not hurt you) because your application does not behave any different whether or not a given logger is enabled. The information here flows one way: From your application into the logger. Even thought loggers are global state since no information flows from loggers into your application, loggers are acceptable. You should still inject your logger if you want your test to assert that something is getting logged, but in general Loggers are not harmful despite being full of state.

class Foo(object):
     pass

some_global_variable = Foo()

Modules are imported only once, everything else is overthinking. Don’t use singletons and try not to use globals.

Use a module. It is imported only once. Define some global variables in it – they will be singleton’s ‘attributes’. Add some functions – the singleton’s ‘methods’.

You probably never need a singleton in Python. Just define all your data and functions in a module and you have a de-facto singleton.

If you really absolutely have to have a singleton class then I’d go with:

class My_Singleton(object):
    def foo(self):
        pass

my_singleton = My_Singleton()

To use:

from mysingleton import my_singleton
my_singleton.foo()

where mysingleton.py is your filename that My_Singleton is defined in. This works because after the first time a file is imported, Python doesn’t re-execute the code.

Here’s a one-liner for you:

singleton = lambda c: c()

Here’s how you use it:

@singleton
class wat(object):
    def __init__(self): self.x = 1
    def get_x(self): return self.x

assert wat.get_x() == 1

Your object gets instantiated eagerly. This may or may not be what you want.

Check out Stack Overflow question Is there a simple, elegant way to define singletons in Python? with several solutions.

I’d strongly recommend to watch Alex Martelli’s talks on design patterns in python: part 1 and part 2. In particular, in part 1 he talks about singletons/shared state objects.

Here’s my own implementation of singletons. All you have to do is decorate the class; to get the singleton, you then have to use the Instance method. Here’s an example:

   @Singleton
   class Foo:
       def __init__(self):
           print 'Foo created'

   f = Foo() # Error, this isn't how you get the instance of a singleton

   f = Foo.Instance() # Good. Being explicit is in line with the Python Zen
   g = Foo.Instance() # Returns already created instance

   print f is g # True

And here’s the code:

class Singleton:
    """
    A non-thread-safe helper class to ease implementing singletons.
    This should be used as a decorator -- not a metaclass -- to the
    class that should be a singleton.

    The decorated class can define one `__init__` function that
    takes only the `self` argument. Other than that, there are
    no restrictions that apply to the decorated class.

    To get the singleton instance, use the `Instance` method. Trying
    to use `__call__` will result in a `TypeError` being raised.

    Limitations: The decorated class cannot be inherited from.

    """

    def __init__(self, decorated):
        self._decorated = decorated

    def Instance(self):
        """
        Returns the singleton instance. Upon its first call, it creates a
        new instance of the decorated class and calls its `__init__` method.
        On all subsequent calls, the already created instance is returned.

        """
        try:
            return self._instance
        except AttributeError:
            self._instance = self._decorated()
            return self._instance

    def __call__(self):
        raise TypeError('Singletons must be accessed through `Instance()`.')

    def __instancecheck__(self, inst):
        return isinstance(inst, self._decorated)

Method 3 seems to be very neat, but if you want your program to run in both Python 2 and Python 3, it doesn’t work. Even protecting the separate variants with tests for the Python version fails, because the Python 3 version gives a syntax error in Python 2.

Thanks to Mike Watkins: http://mikewatkins.ca/2008/11/29/python-2-and-3-metaclasses/. If you want the program to work in both Python 2 and Python 3, you need to do something like:

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]

MC = Singleton('MC', (object), {})

class MyClass(MC):
    pass    # Code for the class implementation

I presume that ‘object’ in the assignment needs to be replaced with the ‘BaseClass’, but I haven’t tried that (I have tried code as illustrated).

Well, other than agreeing with the general Pythonic suggestion on having module-level global, how about this:

def singleton(class_):
    class class_w(class_):
        _instance = None
        def __new__(class2, *args, **kwargs):
            if class_w._instance is None:
                class_w._instance = super(class_w, class2).__new__(class2, *args, **kwargs)
                class_w._instance._sealed = False
            return class_w._instance
        def __init__(self, *args, **kwargs):
            if self._sealed:
                return
            super(class_w, self).__init__(*args, **kwargs)
            self._sealed = True
    class_w.__name__ = class_.__name__
    return class_w

@singleton
class MyClass(object):
    def __init__(self, text):
        print text
    @classmethod
    def name(class_):
        print class_.__name__

x = MyClass(111)
x.name()
y = MyClass(222)
print id(x) == id(y)

Output is:

111     # the __init__ is called only on the 1st time
MyClass # the __name__ is preserved
True    # this is actually the same instance

How about this:

def singleton(cls):
    instance=cls()
    cls.__new__ = cls.__call__= lambda cls: instance
    cls.__init__ = lambda self: None
    return instance

Use it as a decorator on a class that should be a singleton. Like this:

@singleton
class MySingleton:
    #....

This is similar to the singleton = lambda c: c() decorator in another answer. Like the other solution, the only instance has name of the class (MySingleton). However, with this solution you can still “create” instances (actually get the only instance) from the class, by doing MySingleton(). It also prevents you from creating additional instances by doing type(MySingleton)() (that also returns the same instance).

I’ll toss mine into the ring. It’s a simple decorator.

from abc import ABC

def singleton(real_cls):

    class SingletonFactory(ABC):

        instance = None

        def __new__(cls, *args, **kwargs):
            if not cls.instance:
                cls.instance = real_cls(*args, **kwargs)
            return cls.instance

    SingletonFactory.register(real_cls)
    return SingletonFactory

# Usage
@singleton
class YourClass:
    ...  # Your normal implementation, no special requirements.

Benefits I think it has over some of the other solutions:

• It’s clear and concise (to my eye ;D).
• Its action is completely encapsulated. You don’t need to change a single thing about the implementation of YourClass. This includes not needing to use a metaclass for your class (note that the metaclass above is on the factory, not the “real” class).
• It doesn’t rely on monkey-patching anything.
• It’s transparent to callers:
  • Callers still simply import YourClass, it looks like a class (because it is), and they use it normally. No need to adapt callers to a factory function.
  • What YourClass() instantiates is still a true instance of the YourClass you implemented, not a proxy of any kind, so no chance of side effects resulting from that.
  • isinstance(instance, YourClass) and similar operations still work as expected (though this bit does require abc so precludes Python <2.6).

One downside does occur to me: classmethods and staticmethods of the real class are not transparently callable via the factory class hiding it. I’ve used this rarely enough that I’ve never happen to run into that need, but it would be easily rectified by using a custom metaclass on the factory that implements __getattr__() to delegate all-ish attribute access to the real class.

A related pattern I’ve actually found more useful (not that I’m saying these kinds of things are required very often at all) is a “Unique” pattern where instantiating the class with the same arguments results in getting back the same instance. I.e. a “singleton per arguments”. The above adapts to this well and becomes even more concise:

def unique(real_cls):

    class UniqueFactory(ABC):

        @functools.lru_cache(None)  # Handy for 3.2+, but use any memoization decorator you like
        def __new__(cls, *args, **kwargs):
            return real_cls(*args, **kwargs)

    UniqueFactory.register(real_cls)
    return UniqueFactory

All that said, I do agree with the general advice that if you think you need one of these things, you really should probably stop for a moment and ask yourself if you really do. 99% of the time, YAGNI.

Code based on Tolli’s answer.

#decorator, modyfies new_cls
def _singleton(new_cls):
    instance = new_cls()                                              #2
    def new(cls):
        if isinstance(instance, cls):                                 #4
            return instance
        else:
            raise TypeError("I can only return instance of {}, caller wanted {}".format(new_cls, cls))
    new_cls.__new__  = new                                            #3
    new_cls.__init__ = lambda self: None                              #5
    return new_cls


#decorator, creates new class
def singleton(cls):
    new_cls = type('singleton({})'.format(cls.__name__), (cls,), {} ) #1
    return _singleton(new_cls)


#metaclass
def meta_singleton(name, bases, attrs):
    new_cls = type(name, bases, attrs)                                #1
    return _singleton(new_cls)

Explanation:

1. Create new class, inheriting from given cls
   (it doesn’t modify cls in case someone wants for example singleton(list))

2. Create instance. Before overriding __new__ it’s so easy.

3. Now, when we have easily created instance, overrides __new__ using method defined moment ago.

4. The function returns instance only when it’s what the caller expects, otherwise raises TypeError.
   The condition is not met when someone attempts to inherit from decorated class.

5. like __init__(self[, ...]), where self is the new instance and the remaining arguments are the same as were passed to __new__().

   instance is already initialized, so function replaces __init__ with function doing nothing.

See it working online

It is slightly similar to the answer by fab but not exactly the same.

The singleton contract does not require that we be able to call the constructor multiple times. As a singleton should be created once and once only, shouldn’t it be seen to be created just once? “Spoofing” the constructor arguably impairs legibility.

So my suggestion is just this:

class Elvis():
    def __init__(self):
        if hasattr(self.__class__, 'instance'):
            raise Exception()
        self.__class__.instance = self
        # initialisation code...

    @staticmethod
    def the():
        if hasattr(Elvis, 'instance'):
            return Elvis.instance
        return Elvis()

This does not rule out the use of the constructor or the field instance by user code:

if Elvis() is King.instance:

… if you know for sure that Elvis has not yet been created, and that King has.

But it encourages users to use the the method universally:

Elvis.the().leave(Building.the())

To make this complete you could also override __delattr__() to raise an Exception if an attempt is made to delete instance, and override __del__() so that it raises an Exception (unless we know the program is ending…)

Further improvements

My thanks to those who have helped with comments and edits, of which more are welcome. While I use Jython, this should work more generally, and be thread-safe.

try:
    # This is jython-specific
    from synchronize import make_synchronized
except ImportError:
    # This should work across different python implementations
    def make_synchronized(func):
        import threading
        func.__lock__ = threading.Lock()

        def synced_func(*args, **kws):
            with func.__lock__:
                return func(*args, **kws)

        return synced_func

class Elvis(object): # NB must be subclass of object to use __new__
    instance = None

    @classmethod
    @make_synchronized
    def __new__(cls, *args, **kwargs):
        if cls.instance is not None:
            raise Exception()
        cls.instance = object.__new__(cls, *args, **kwargs)
        return cls.instance

    def __init__(self):
        pass
        # initialisation code...

    @classmethod
    @make_synchronized
    def the(cls):
        if cls.instance is not None:
            return cls.instance
        return cls()

Points of note:

1. If you don’t subclass from object in python2.x you will get an old-style class, which does not use __new__
2. When decorating __new__ you must decorate with @classmethod or __new__ will be an unbound instance method
3. This could possibly be improved by way of use of a metaclass, as this would allow you to make the a class-level property, possibly renaming it to instance

One liner (I am not proud, but it does the job):

class Myclass:
  def __init__(self):
      # do your stuff
      globals()[type(self).__name__] = lambda: self # singletonify

If you don’t need lazy initialization of the instance of the Singleton, then the following should be easy and thread-safe:

class A:
    instance = None
    # Methods and variables of the class/object A follow
A.instance = A()

This way A is a singleton initialized at module import.

• If one wants to have multiple number of instances of the same class, but only if the args or kwargs are different, one can use this
• Ex.
  1. If you have a class handling serial communication, and to create an instance you want to send the serial port as an argument, then with traditional approach won’t work
  2. Using the above mentioned decorators, one can create multiple instances of the class if the args are different.
  3. For same args, the decorator will return the same instance which is already been created.

>>> from decorators import singleton
>>>
>>> @singleton
... class A:
...     def __init__(self, *args, **kwargs):
...         pass
...
>>>
>>> a = A(name='Siddhesh')
>>> b = A(name='Siddhesh', lname='Sathe')
>>> c = A(name='Siddhesh', lname='Sathe')
>>> a is b  # has to be different
False
>>> b is c  # has to be same
True
>>>

Maybe I missunderstand the singleton pattern but my solution is this simple and pragmatic (pythonic?). This code fullfills two goals

1. Make the instance of Foo accessiable everywhere (global).
2. Only one instance of Foo can exist.

This is the code.

#!/usr/bin/env python3

class Foo:
    me = None

    def __init__(self):
        if Foo.me != None:
            raise Exception('Instance of Foo still exists!')

        Foo.me = self


if __name__ == '__main__':
    Foo()
    Foo()

Output

Traceback (most recent call last):
  File "./x.py", line 15, in <module>
    Foo()
  File "./x.py", line 8, in __init__
    raise Exception('Instance of Foo still exists!')
Exception: Instance of Foo still exists!

After struggling with this for some time I eventually came up with the following, so that the config object would only be loaded once, when called up from separate modules. The metaclass allows a global class instance to be stored in the builtins dict, which at present appears to be the neatest way of storing a proper program global.

import builtins

# -----------------------------------------------------------------------------
# So..... you would expect that a class would be "global" in scope, however
#   when different modules use this,
#   EACH ONE effectively has its own class namespace.  
#   In order to get around this, we use a metaclass to intercept
#   "new" and provide the "truly global metaclass instance" if it already exists

class MetaConfig(type):
    def __new__(cls, name, bases, dct):
        try:
            class_inst = builtins.CONFIG_singleton

        except AttributeError:
            class_inst = super().__new__(cls, name, bases, dct)
            builtins.CONFIG_singleton = class_inst
            class_inst.do_load()

        return class_inst

# -----------------------------------------------------------------------------

class Config(metaclass=MetaConfig):

    config_attr = None

    @classmethod
    def do_load(cls):
        ...<load-cfg-from-file>...

I can’t remember where I found this solution, but I find it to be the most ‘elegant’ from my non-Python-expert point of view:

class SomeSingleton(dict):
    __instance__ = None
    def __new__(cls, *args,**kwargs):
        if SomeSingleton.__instance__ is None:
            SomeSingleton.__instance__ = dict.__new__(cls)
        return SomeSingleton.__instance__

    def __init__(self):
        pass

    def some_func(self,arg):
        pass

Why do I like this? No decorators, no meta classes, no multiple inheritance…and if you decide you don’t want it to be a Singleton anymore, just delete the __new__ method. As I am new to Python (and OOP in general) I expect someone will set me straight about why this is a terrible approach?

This is my preferred way of implementing singletons:

class Test(object):
    obj = None

    def __init__(self):
        if Test.obj is not None:
            raise Exception('A Test Singleton instance already exists')
        # Initialization code here

    @classmethod
    def get_instance(cls):
        if cls.obj is None:
            cls.obj = Test()
        return cls.obj

    @classmethod
    def custom_method(cls):
        obj = cls.get_instance()
        # Custom Code here

This answer is likely not what you’re looking for. I wanted a singleton in the sense that only that object had its identity, for comparison to. In my case it was being used as a Sentinel Value. To which the answer is very simple, make any object mything = object() and by python’s nature, only that thing will have its identity.

#!python
MyNone = object()  # The singleton

for item in my_list:
    if item is MyNone:  # An Example identity comparison
        raise StopIteration

This solution causes some namespace pollution at the module level (three definitions rather than just one), but I find it easy to follow.

I’d like to be able to write something like this (lazy initialization), but unfortunately classes are not available in the body of their own definitions.

# wouldn't it be nice if we could do this?
class Foo(object):
    instance = None

    def __new__(cls):
        if cls.instance is None:
            cls.instance = object()
            cls.instance.__class__ = Foo
        return cls.instance

Since that isn’t possible, we can break out the initialization and the static instance in

Eager Initialization:

import random


class FooMaker(object):
    def __init__(self, *args):
        self._count = random.random()
        self._args = args


class Foo(object):
    def __new__(self):
        return foo_instance


foo_instance = FooMaker()
foo_instance.__class__ = Foo

Lazy initialization:

Eager Initialization:

import random


class FooMaker(object):
    def __init__(self, *args):
        self._count = random.random()
        self._args = args


class Foo(object):
    def __new__(self):
        global foo_instance
        if foo_instance is None:
            foo_instance = FooMaker()
        return foo_instance


foo_instance = None