Implementing a Python Singleton with Decorators

In Python, a singleton is a design pattern that ensures a class has only one instance, and it provides a global point of access to that instance. This is useful when managing shared resources such as database connections, configuration objects, or logging systems where multiple instantiations would lead to inefficiencies or inconsistencies.

In this blog post, we'll discuss a Python implementation of the singleton pattern using a decorator. We'll walk through the code and explain how it works, focusing on the _SingletonWrapper class and the singleton decorator function.

What is a Decorator?

A decorator in Python is a function that modifies or extends the behavior of another function or class. It allows you to "wrap" a function or a class, adding functionality to it without changing its structure.

The Singleton Pattern Explained

The singleton pattern guarantees that a class will have only one instance. When the singleton instance is requested, the existing instance is returned rather than creating a new one. This is particularly important for certain application-level services that need to be shared across the entire program.

The Code Breakdown

Let's break down the code to see how it enforces the singleton behavior.

1. The _SingletonWrapper Class

1class _SingletonWrapper:
2    """
3    A singleton wrapper class. Its instances would be created
4    for each decorated class.
5    """
6
7    def __init__(self, cls):
8        self.__wrapped__ = cls
9        self._instance = None

The _SingletonWrapper class serves as a wrapper around the original class that is being decorated.

• The __init__ method takes a class (cls) and stores it in the __wrapped__ attribute.
• It also initializes _instance to None. This is the attribute that will store the single instance of the wrapped class.

2. The __call__ Method

1    def __call__(self, *args, **kwargs):
2        """Returns a single instance of decorated class"""
3        if self._instance is None:
4            self._instance = self.__wrapped__(*args, **kwargs)
5        return self._instance

The __call__ method is a special method in Python that makes an instance of a class callable. It is triggered when the instance is called like a function. In this case, when the singleton object is called, it checks if an instance already exists.

• If _instance is None (i.e., no instance has been created yet), it creates a new instance of the decorated class.
• If an instance already exists, it returns the same instance, ensuring that only one instance of the class is ever created.

3. The singleton Decorator

1def singleton(cls):
2    """
3    A singleton decorator. Returns a wrapper object. A call on that object
4    returns a single instance object of decorated class. Use the __wrapped__
5    attribute to access the decorated class directly in unit tests.
6    """
7    return _SingletonWrapper(cls)

This function acts as the actual decorator. When a class is decorated with @singleton, it wraps that class in an instance of _SingletonWrapper. Now, whenever the decorated class is instantiated, the wrapped version will return a single instance.

Using the Singleton Decorator

Let's look at an example of how this decorator would be used.

 1@singleton
 2class Logger:
 3    def __init__(self):
 4        self.log = []
 5
 6    def write_log(self, message):
 7        self.log.append(message)
 8
 9    def read_log(self):
10        return self.log

In this case, the Logger class is decorated with the @singleton decorator. Now, no matter how many times you try to instantiate Logger, you'll always get the same instance:

1logger1 = Logger()
2logger2 = Logger()
3
4logger1.write_log("Log message 1")
5print(logger2.read_log())  # Output: ['Log message 1']
6
7print(logger1 is logger2)  # Output: True

As shown above, both logger1 and logger2 refer to the same instance, demonstrating the singleton behavior.

Advantages of Using a Singleton

1. Global Access: A singleton allows for centralized control and management of an instance across an entire application. This is useful when you have shared resources like database connections or logging services.

2. Efficiency: Since the same instance is reused, the singleton pattern can reduce memory usage and speed up object access, especially for resource-heavy objects.

3. Ease of Testing: With the __wrapped__ attribute, you can directly access the original, undecorated class in unit tests. This allows you to test the class independently of the singleton behavior, making the code easier to test and maintain.

When to Avoid Singletons

While the singleton pattern can be useful, it is not always appropriate for every situation. Overusing singletons can lead to tight coupling and make code difficult to test and maintain. They also introduce global state, which can lead to issues in multi-threaded environments or when scaling applications.

Conclusion

The singleton pattern is a powerful tool when applied correctly. Using a decorator like the one we've explored here allows you to create clean, reusable code while enforcing singleton behavior. As with any design pattern, it's important to understand the context in which it's most beneficial and apply it judiciously.