Introduction

Abstraction is the process of hiding the real implementation of an application from the user and emphasizing only on usage of it.

Abstraction in Python is achieved by using Abstract Base Classes (ABCs) 1. The abc module in Python library provides the infrastructure for defining custom abstract base classes.

Abstract classes are classes that contain one or more abstract methods and generally provides incomplete functionality. An abstract method is a method that is declared, but contains no implementation. Abstract classes cannot be instantiated, and require subclasses to provide implementations for the abstract methods.

There are many built-in ABCs in Python. ABCs for Data structures like Iterator, Generator, Set, Mapping etc. are defined in collections.abc module.

Implementation

The abc module defines ABCMeta class which is a metaclass 2 (or) ABC helper class for defining abstract base class and @abstractmethod decorator 3 for defining abstract methods.

Let’s create a Abstract Class Shape for Rectangle and Square which implements the area() abstract method.

Abstraction using ABCMeta

NOTE: All abstract methods defined in Abstaction Class has to be implemented by subclass or else TypeError exception is raised.

Abstraction using ABC

ABC helper class already has ABCMeta as its metaclass. Simply replace the metaclass=ABCMeta with ABC from the above implementation.

FootNotes

  1. https://docs.python.org/3/library/abc.html ↩︎

  2. More about Meta Classes: http://harish678.github.io/posts/metaclass/ ↩︎

  3. More about Decorators: http://harish678.github.io/posts/decorators/ ↩︎