An Introduction to Object Oriented Programming with Python

Posted on August 07, 2021 - 11 minutes of reading

A brief introduction to the Object Oriented paradigm with Python, where we will see how the language can be used in this context and how it differs from other languages aimed at using this paradigm.

pythonOOP

Index

1 Intro
- 1.1 Naming conventions in Python
2 Classes
3 Polymorphism
4 Encapsulation
5 Relations between classes
6 References
7 Conclusion

1 Intro

Python differs greatly from other programming languages that it supports, or that are complete object-oriented. This post is a small guide with some language details focused on this paradigm.

1.1 Naming conventions in Python

These conventions have a name that we can use to refer to how we are naming certain objects in our program:

PascalCase - means that all words start with a capital letter and nothing is used to separate them, as in: MyClass, Class, MyObject, MyProgramVeryCool. This is the convention used for classes in Python;

camelCase - the only difference between camelCase to PascalCase is the first letter. In camelCase, the first letter will always be lower case and the rest of the words must start with a capital letter. As in: myFunction, SumFunction, etc... This convention is not used in Python;

snake_case - this is the pattern used in Python to define anything that is not a class. All letters will be lowercase and separated by an underscore, as in: my_variable, cool_function, sum.

So the patterns used in Python are: snake_case for anything and PascalCase for classes.

2 Classes

Classes provide a way to organize data and functionality together. Creating a new class creates a new “type” of object, allowing new “instances” of that type to be produced. Each instance of the class can have attributes attached to it to maintain its state. Instances of the class can also have methods (defined by the class) for modifying their state.

2.1 Declare class

class Person:
    pass

2.2 Abstract Class

A generic class that will not be instantiated, can have concrete and abstract methods. For that, we must import the module ABC (abstract base class).

from abc import ABC, abstractmethod

class A(ABC):
    @abstractmethod
    def speak(self):
        pass

class B(A):
    def speak(self):
        print(f'B speaking...')

abst = B()
abst.speak()

2.3 Attributes

Attributes can be defined directly inside the class (Class Attribute) or within the methods created in the class (Instance Attributes).

The difference is that the class attribute can be accessed and changed by the class. but instances can only access it but cannot change it.

If a value is assigned to an attribute of the instance of the same name that it inherited from the class, a new instance attribute will be created while the class attribute remains unchanged.

class Cls:
    num = 1

inst = Cls()
print(inst.__dict__)
# {} (object doesn't have any instance attributes yet...)
print(inst.num)
# 1 (and as it doesn't have the "num" attribute it goes to a higher level where it finds the class attribute...)
inst.num = 2
print(inst.__dict__)
# {'num': 2} (after assignment now the object has its own attribute...)
print(inst.num)
# 2 (so it no longer accesses the attribute of the same name of the class...)
print(cls.num)
# 1 (where the attribute still has initial value.)

2.4 Methods

2.4.1 Instance Methods

Any method created within the class, which when instantiated the object will receive its methods.

self All instance methods of the class receive self as their first attribute, it refers to the instance that was created from the class.

class Person:
     def eat(self, food):
         print(f'Person is eating {food}.')

Class attributes can be declared in the constructor, assigning the values that are passed by the parameter, as well as being created directly in the class and being accessible to all instances created through self.

2.4.2 Class methods

In addition to the instance methods that are available for each object instantiated from the class, methods of the class can also be created where they are accessible only to the class itself and not for objects instantiated from the class. For this behavior, we must first pass a decorator called @classmethod and the first parameter becomes the (cls) class itself.

class Person:
    current_year = 2021
    
    def __init__(self, name, age):
        self.name = name
        self.age = age

    @classmethod
    def person_birth(cls, name, birth):
        age = class.current_year - birth
        return cls(name, age)

2.4.3 Static methods

They are methods that do not receive the context of the instance and the class, they could even be created outside the class. You must add a @staticmethod decorator before the method signature.

from random import randint
class Person:
    def __init__(self, name, age):
        self.name = name
        self.age = age

    @staticmethod
    def generate_id():
        rand = randint(10000, 19999)
        return rand

2.4.3 Abstract methods

These are methods that don't have a body just the method signature, are usually created within an abstract class to be overridden on instances that inherit from that class.

from abc import ABC, abstractmethod
class Example(ABC):
    @abstractmethod
    def withdraw(self, value):
        pass

2.4.4 Magic Methods

They are special methods you can define to add "magic" to classes. They are always surrounded by two underscores (eg __init__ or __lt__). To see all magic methods I recommend this guide, Here are some of the most used ones:

__new__ It is the first method to be called when instantiating an object. It takes the class and then any other arguments and will pass it to __init__.

class Person:
     def __new__(cls, *args, **kwargs):
         pass

__init__ The class initializer is called once the class is instantiated.

class Person:
     def __init__(self, name):
         self.name = name

__call__ Allows an instance of a class to be called as a function. Essentially, this means that Person() is the same as Person.__call__(). Note that __call__ takes a variable number of arguments, this means that you define __call__ as you would any other function, using as many arguments as you like.

class Person:
     def __call__(self, *args, **kwargs):
         print(args)
         print(kwargs)

p = Person()
p(1, 2, 3, name='Gui')
# (1, 2, 3)
# {'name': 'gui'}

__setattr__ It is an encapsulation solution. It allows you to define the behavior for assigning to an attribute, regardless of the existence or not of this attribute, which means you can define custom rules for any changes in attribute values.

class Pessoa:
    def __init__(self, name):
        self.name = name
        
    def __setattr__(self, key, value):
        if key != 'name':
            self.__dict__[key] = value

2.5 Metaclasses

In python, everything is an object, including classes. Metaclasses are like "classes" that create classes. type is a metaclass too.

# type(name = class name, bases = inherited classes, namespace = attributes and methods)
M = type('metaclass', (), {})

class Meta(type):
      def __new__(mcs, name, bases, namespace):
          if name == 'A':
              return type.__new__(mcs, name, bases, namespace)

          if 'attr_cls' in namespace:
              del namespace['attr_cls']

          return type.__new__(mcs, name, bases, namespace)

2.6 Enum (Python 3.4)

Enum is a python class for creating enumerations, which are a set of symbolic names (members) bound to constant and unique values. The members of an enumeration may be compared to these symbolic names, and the enumeration itself can be iterated.

from enum import enum

Class directions (Enum):
      right = 0
      left = 1
      up = 2
      down = 3

def move(direction):
      if not isinstance(direction, Directions):
          raise ValueError('Cannot move in this direction')

      return f'Moving {direction.name} to position {direction.value}'

3 Polymorphism

In python, the only polymorphism the language supports is overlaid, which is the principle that allows classes to be derived from the same superclass and have the same methods (of the same signature) but different behaviors.

Same signature = Same quantity and type of parameters.

from abc import ABC, abstractmethod

class A(ABC):
    @abstractmethod
    def speak(self, msg):
        pass

class B(A):
    def speak(self, msg):
        print(f'B is speaking {msg}')

class C(A):
    def speak(self, msg):
        print(f'C is speaking {msg}')

b = B()
c = C()
b.speak('about skating')
# B is speaking about skating
c.speak('about soccer')
# C is speaking about soccer

4 Encapsulation

Encapsulation is one of the pillars of object orientation. Serves to protect class data. Encapsulating an application's data means preventing them from being accessed improperly. For this, a structure is created where modifiers such as public, protected, private are used to restrict access to this data. And methods that can be used by any other class, without causing inconsistencies in development are commonly called getters and setters.

4.1 Modifiers

In python we don't have modifiers to restrict class data access. Therefore, when naming attributes and methods, a PEP8 convention follows:

name    = public
_name   = protected
__name  = private

In practice the attributes or methods can still be accessed and modified, what changes are that when created with __ python does not let the value be reassigned for this variable, it ends up creating another one with the same name in the instance: obj.__nome.
And to access the value of the original variable you must put the name of the class before it: obj._Classe__nome.

class Data:
    def __init__(self):
        self.publics = {}
        self._protected = {}
        self.__private = {}

    # getter to get value out of class/object
    @property
    def protected(self):
        return self._protected

    # setter to set a value outside class/object
    @protected.setter
    def protected(self, value):
        self._protected = value

4.2 Getters & Setters

These methods are called as soon as an object is instantiated from the class, they serve as a filter. The getter method gets a value for the instance attribute. And it must contain the @property decorator before its signature. The setter method sets a value for the attribute. It must contain a decorator with the same name as the attribute followed by a setter, @name.setter.

class Product:
    def __init__(self, name):
        self.name = name

    # getter
    @property
    def name(self):
        return self._name

    # Regulator
    @name.setter
    def name(self, name):
        self._name = name.title()

4.3 Type Hints (Python 3.5)

Python has dynamic typing, where types are checked during execution. And also strong typing, all operations between different types must be explicitly defined and undefined operations between types will result in an error. For this, the use of type hints can be of great help to programmers with a background in some statically typed language like Java, Haskell, and Go, which has to develop something bigger than a simple Python script.

Some advantages of its use:

Serve as a form of documentation;
Optimizations at compile time;
Security when analyzing the program;
Code-complete better;
It's easier to find in a large codebase.

from typing import Union

# Defining variable types
x: int = 10
y: float = 10.5
z: bool = False

# Defining types of arguments and return of a function
def greeting(name: str) -> str:
    return 'Hello ' + name

# Defining return with more than one type with Union of the typing module
def fn(p1: float, p2: int) -> Union[int, float]:
    return p1 * p2

5 Relations between classes

Association (Use another class)
Aggregation (Has another class)
Composition (Owns another class)
Inheritance (Inherits from another class)

5.1 Association

It describes a link that occurs between classes. The most common way to implement association is to have one object as an attribute of another, in this example below:

class Writer:
    def __init__(self, name):
        self.__name = name
        self.__tool = None

    @property
    def name(self):
        return self.__name

    @property
    def tool(self):
        return self.__tool

    @ferramenta.setter
    def tool(self, tool):
        self.__tool = tool
    

class Pen:
    def __init__(self, tag):
        self.__brand = brand

    @property
    def tag(self):
        return self.__brand

    def write(self):
        print('Pen is writing...')
        
    
writer = Writer('John')
pen = Pen('Bic')

writer.tool = pen
writer.tool.write()

5.2 Aggregation

It is a special type of association where one tries to demonstrate that the information of an object (called object-todo) need to be supplemented by information contained in one or more objects of another class (called part-objects) known as whole/part. But these parts can exist separately.

In this aggregation example, the classes can exist separately but they work better when the cart has products.

class Product:
    def __init__(self, name, value):
        self.name = name
        self.value = value

class ShoppingCart:
    def __init__(self, ):
        self.products = []

    def insert_product(self, product):
        self.products.append(product)

    def product_list(self):
        for product in self.products:
            print(f'{product.name}: ${product.value}')

    def sum_total(self):
        total = 0
        for product in self.products:
            total += product.value
        return f'${total}'
    
    
cart = ShoppingCart()
p1 = Product('Hat', 50)
p2 = Product('underwear', 20)
p3 = Product('shoes', 150)
cart.insert_product(p1)
cart.insert_product(p2)
cart.insert_product(p3)
cart.list_products()
# Bone: USD 50
# underwear: $20
# tennis shoes: USD 150
print(cart.sum_total())
# USD 220

5.3 Composition

A composition also tries to represent a whole/part relationship. However, in composition, the object-whole is responsible for creating and destroying its parts. In a composition, the same part-object cannot be associated with more than one whole-object.

In the example below, the customer object has an address object in its address book, in this case when the customer object which is the all-object is deleted, the address object will also be deleted.

class Customer:
    def __init__(self, name):
        self.__name = name
        self.__addresses = []

    @property
    def name(self):
        return self.__name

    @name.setter
    def name(self, name):
        self.__name = name

    def insert_address(self, city):
        self.__addresses.append(Address(city))

    def address_list(self):
        for address in self.__addresses:
            print(address.city)

class Address:
    def __init__(self, city):
        self.__city = city

    @property
    def city(self):
        return self.__city

    @cidade.setter
    def city(self, city):
        self.__city = city

customer1 = Customer('John')
customer1.insert_address('Dublin')
print(client1.name)
# John
customer1.list_addresses()
# Dublin
from client1

5.4 Inheritance

An object can have methods and attributes of another class by inheritance, it means that the class has all features of the inherited class, besides being able to have your own.

One of the great advantages of using inheritance is code reuse. This reuse can be triggered when it is identified that the attribute or method of a class will be the same for the others. To inherit from a class, the name of the class is passed as a parameter.

# superclass
class Person:
    def __init__(self, name):
        self.__name = name
    
    @property
    def name(self):
        return self.__name

    @name.setter
    def name(self, name):
        self.__name = name

    def talk(self):
        print(f'{self.name} is speaking...')


# subclasses
class Customer(Person):
    def buy(self):
        print(f'{self.name} is buying...')


class Student(Person):
    def study(self):
        print(f'{self.name} is studying...')

5.4.1 Overriding methods

Inherited methods can be overridden within the class. To use the logic of the overridden method and add more lines of code it is necessary to pass to the super() method which refers to the class being inherited, to reference a specific class within the inheritance chain it is necessary to pass the name of the class with the method and self as a parameter.

class ClienteVip(Customer):
    def __init__(self, first name, last name):
        super().__init__(name)
        self.lastname = lastname

    def talk(self):
        Person.talk(self)   # in this case refers to the Person class specifically.
        super().falar()     # in this case it refers to the inherited class (Client).
        print(f'{self.name} {self.lastname} is vip...')

5.4.2 Multiple Inheritance

Multiple inheritances, in object orientation, is the concept of inheriting from two or more classes.

class A:
    def talk(self):
        print('Talking... I'm in A.')

class B(A):
    def talk(self):
        print('Speaking... I'm in B.')

class C(B, A):
    pass

c = C()
c.speak()
# Speaking... I'm in B.

5.4.3 Mixin

Especially in the context of Python, a mixin is a parent class that provides functionality to subclasses, but is not intended to be instantiated.

class LogMixin:
    """
    creates a log file with info and errors.
    """
    @staticmethod
    def write(msg):
        with open('log.log', 'a+') as f:
            f.write(msg)
            f.write('\n')

    def log_info(self, msg):
        """
        write an info message to file
        :param msg:
        :return:
        """
        self.write(f'INFO: {msg}')


    def log_error(self, msg):
        """
        write an error message to file
        :param msg:
        :return:
        """
        self.write(f'ERROR: {msg}')

6 References

7 Conclusion

Because it is a dynamically typed language and is not completely object-oriented, using this paradigm in Python, as we can see, has many differences compared to languages like Java for example. However, I hope this post has helped you to start using this paradigm and take advantage of it.

What did you think of this post? Do you have any suggestions or criticisms? Leave a reaction or a comment below. And thanks for visiting! 😉

Guilherme AlmeidaSoftware Engineer