Object-oriented programming is one of the most effective approaches to writing software. In object-oriented programming you write classes that represent real-world things and situations, and you create objects based on these classes. When you write a class, you define the general behavior that a whole category of objects can have.
When you create individual objects from the class, each object is automatically equipped with the general behavior; you can then give each object whatever unique traits you desire. Making an object from a class is called instantiation, and you work with instances of a class.
Understanding object-oriented programming will help you see the world as a programmer does. It’ll help you really know your code, not just what’s happening line by line, but also the bigger concepts behind it.
Knowing the logic behind classes will train you to think logically so you can write programs that effectively address almost any problem you encounter. Classes also make life easier for you and the other programmers you’ll work with as you take on increasingly complex challenges. When you and other programmers write code based on the same kind of logic, you’ll be able to understand each other’s work. Your programs will make sense to many collaborators, allowing everyone to accomplish more.
Creating and Using a Class
You can model almost anything using classes. Let’s start by writing a simple class, Dog, that represents a dog—not one dog in particular, but any dog.
What do we know about most pet dogs?
- Well, they all have a name and age.
- We also know that most dogs sit and roll over.
Those two pieces of information (name and age) and those two behaviors (sit and roll over) will go in our Dog class because they’re common to most dogs. This class will tell Python how to make an object representing a dog. After our class is written, we’ll use it to make individual instances, each of which represents one specific dog.
Creating the Dog Class
Each instance created from the Dog class will store a name and an age, and we’ll give each dog the ability to sit() and roll_over():
class Dog:
"""A simple attempt to model a dog."""
def __init__(self, name, age):
"""Initialize name and age attributes."""
self.name = name
self.age = age
def sit(self):
"""Simulate a dog sitting in response to a command."""
print(f"{self.name} is now sitting.")
def roll_over(self):
"""Simulate rolling over in response to a command."""
print(f"{self.name} rolled over!")
At Class Dog: we define a class called Dog. By convention, capitalized names refer to classes in Python. There are no parentheses in the class definition because we’re creating this class from scratch.
The init() Method
A function that’s part of a class is a method. Everything you learned about functions applies to methods as well; the only practical difference for now is the way we’ll call methods. The __init__() method at w is a special method that Python runs automatically whenever we create a new instance based on the Dog class. This method has two leading underscores and two trailing underscores, a convention that helps prevent Python’s default method names from conflicting with your method names. Make sure to use two underscores on each side of __init__(). If you use just one on each side, the method won’t be called automatically when you use your class, which can result in errors that are difficult to identify.
We define the __init__() method to have three parameters: self, name, and age.
The self parameter is required in the method definition, and it must come first before the other parameters. It must be included in the definition because when Python calls this method later (to create an instance of Dog), the method call will automatically pass the self argument. Every method call associated with an instance automatically passes self, which is a reference to the instance itself; it gives the individual instance access to the attributes and methods in the class.
When we make an instance of Dog, Python will call the __init__() method from the Dog class. We’ll pass Dog() a name and an age as arguments; self is passed automatically, so we don’t need to pass it. Whenever we want to make an instance from the Dog class, we’ll provide values for only the last two parameters, name and age.
Each of the two variables defined have the prefix self. Any variable prefixed with self is available to every method in the class, and we’ll also be able to access these variables through any instance created from the class.
The line self.name = name takes the value associated with the parameter name and assigns it to the variable name, which is then attached to the instance being created. The same process happens with self.age = age. Variables that are accessible through instances like this are called attributes.
The Dog class has two other methods defined: sit() and roll_over(). Because these methods don’t need additional information to run, we just define them to have one parameter, self. The instances we create later will have access to these methods. In other words, they’ll be able to sit and roll over.
For now, sit() and roll_over() don’t do much. They simply print a message saying the dog is sitting or rolling over. But the concept can be extended to realistic situations: if this class were part of an actual computer game, these methods would contain code to make an animated dog sit and roll over. If this class was written to control a robot, these methods would direct movements that cause a robotic dog to sit and roll over.
Making an Instance from a Class
Think of a class as a set of instructions for how to make an instance. The class Dog is a set of instructions that tells Python how to make individual instances representing specific dogs.
Let’s make an instance representing a specific dog:
my_dog = Dog('Willie', 6)
print(f"My dog's name is {my_dog.name}.")
print(f"My dog is {my_dog.age} years old.")
The line my_dog = Dog('Willie', 6) tells Python to create a dog whose name is 'Willie' and whose age is 6. When Python reads this line, it calls the __init__() method in Dog with the arguments 'Willie' and 6. The __init__() method creates an instance representing this particular dog and sets the name and age attributes using the values we provided. Python then returns an instance representing this dog. We assign that instance to the variable my_dog. The naming convention is helpful here: we can usually assume that a capitalized name like Dog refers to a class, and a lowercase name like my_dog refers to a single instance created from a class.
Accessing Attributes
To access the attributes of an instance, you use dot notation. we can access the value of my_dog’s attribute name by writing my_dog.name.
Dot notation is used often in Python. This syntax demonstrates how Python finds an attribute’s value. Here Python looks at the instance my_dog and then finds the attribute name associated with my_dog. This is the same attribute referred to as self.name in the class Dog.
Calling Methods
After we create an instance from the class Dog, we can use dot notation to call any method defined in Dog.
my_dog.sit()
my_dog.roll_over()
To call a method, give the name of the instance and the method you want to call, separated by a dot. When Python reads my_dog.sit(), it looks for the method sit() in the class Dog and runs that code.
This syntax is quite useful. When attributes and methods have been given appropriately descriptive names like name, age, sit(), and roll_over(), we can easily infer what a block of code, even one we’ve never seen before, is supposed to do.
Creating Multiple Instances
You can create as many instances from a class as you need. Let’s create a second dog called your_dog:
my_dog = Dog('Willie', 6)
your_dog = Dog('Lucy', 3)
print(f"My dog's name is {my_dog.name}.")
print(f"My dog is {my_dog.age} years old.")
my_dog.sit()
print(f"\nYour dog's name is {your_dog.name}.")
print(f"Your dog is {your_dog.age} years old.")
your_dog.sit()
My dog's name is Willie.
My dog is 6 years old.
Willie is now sitting.
Your dog's name is Lucy.
Your dog is 3 years old.
Lucy is now sitting.
Each dog is a separate instance with its own set of attributes, capable of the same set of actions. Even if we used the same name and age for the second dog, Python would still create a separate instance from the Dog class. You can make as many instances from one class as you need, as long as you give each instance a unique variable name or it occupies a unique spot in a list or dictionary.
Working with Classes and Instances
You can use classes to represent many real-world situations. Once you write a class, you’ll spend most of your time working with instances created from that class. One of the first tasks you’ll want to do is modify the attributes associated with a particular instance. You can modify the attributes of an instance directly or write methods that update attributes in specific ways.
The Car Class
Let’s write a new class representing a car. Our class will store information about the kind of car we’re working with, and it will have a method that summarizes this information:
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
my_new_car = Car('audi', 'a4', 2019)
print(my_new_car.get_descriptive_name())
2019 Audi A4
To make the class more interesting, let’s add an attribute that changes over time. We’ll add an attribute that stores the car’s overall mileage.
Setting a Default Value for an Attribute
When an instance is created, attributes can be defined without being passed in as parameters. These attributes can be defined in the __init__() method, where they are assigned a default value.
Let’s add an attribute called odometer_reading that always starts with a value of 0. We’ll also add a method read_odometer() that helps us read each car’s odometer:
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
self.odometer_reading = 0
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
def read_odometer(self):
"""Print a statement showing the car's mileage."""
print(f"This car has {self.odometer_reading} miles on it.")
my_new_car = Car('audi', 'a4', 2019)
print(my_new_car.get_descriptive_name())
my_new_car.read_odometer()
2019 Audi A4
This car has 0 miles on it.
This time when Python calls the __init__() method to create a new instance, it stores the make, model, and year values as attributes like it did in the previous example. Then Python creates a new attribute called odometer_reading and sets its initial value to 0. We also have a new method called read_odometer() that makes it easy to read a car’s mileage.
Not many cars are sold with exactly 0 miles on the odometer, so we need a way to change the value of this attribute.
Modifying Attribute Values
You can change an attribute’s value in three ways: you can change the value directly through an instance, set the value through a method, or increment the value (add a certain amount to it) through a method.
Modifying an Attribute’s Value Directly
The simplest way to modify the value of an attribute is to access the attribute directly through an instance.
my_new_car.odometer_reading = 23
my_new_car.read_odometer()
2019 Audi A4
This car has 0 miles on it.
Sometimes you’ll want to access attributes directly like this, but other times you’ll want to write a method that updates the value for you.
Modifying an Attribute’s Value Through a Method
It can be helpful to have methods that update certain attributes for you. Instead of accessing the attribute directly, you pass the new value to a method that handles the updating internally.
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
self.odometer_reading = 0
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
def read_odometer(self):
"""Print a statement showing the car's mileage."""
print(f"This car has {self.odometer_reading} miles on it.")
def update_odometer(self, mileage):
"""Set the odometer reading to the given value."""
self.odometer_reading = mileage
my_new_car = Car('audi', 'a4', 2019)
print(my_new_car.get_descriptive_name())
my_new_car.update_odometer(23)
my_new_car.read_odometer()
2019 Audi A4
This car has 23 miles on it.
We can extend the method update_odometer() to do additional work every time the odometer reading is modified. Let’s add a little logic to make sure no one tries to roll back the odometer reading:
def update_odometer(self, mileage):
"""
Set the odometer reading to the given value.
Reject the change if it attempts to roll the odometer back.
"""
if mileage >= self.odometer_reading:
self.odometer_reading = mileage
else:
print("You can't roll back an odometer!")
Now update_odometer() checks that the new reading makes sense before modifying the attribute.
Incrementing an Attribute’s Value Through a Method
Sometimes you’ll want to increment an attribute’s value by a certain amount rather than set an entirely new value. Say we buy a used car and put 100 miles on it between the time we buy it and the time we register it. We can define a method increment_odometer() that allows us to pass this incremental amount and add that value to the odometer reading:
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
self.odometer_reading = 0
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
def read_odometer(self):
"""Print a statement showing the car's mileage."""
print(f"This car has {self.odometer_reading} miles on it.")
def update_odometer(self, mileage):
"""
Set the odometer reading to the given value.
Reject the change if it attempts to roll the odometer back.
"""
if mileage >= self.odometer_reading:
self.odometer_reading = mileage
else:
print("You can't roll back an odometer!")
def increment_odometer(self, miles):
"""Add the given amount to the odometer reading."""
self.odometer_reading += miles
my_used_car = Car('subaru', 'outback', 2015)
print(my_used_car.get_descriptive_name())
my_used_car.update_odometer(23_500)
my_used_car.read_odometer()
my_used_car.increment_odometer(100)
my_used_car.read_odometer()
2015 Subaru Outback
This car has 23500 miles on it.
This car has 23600 miles on it.
You can easily modify this method to reject negative increments so no one uses this function to roll back an odometer.
You can use methods like this to control how users of your program update values such as an odometer reading, but anyone with access to the program can set the odometer reading to any value by accessing the attribute directly. Effective security takes extreme attention to detail in addition to basic checks like those shown here.
Inheritance
You don’t always have to start from scratch when writing a class. If the class you’re writing is a specialized version of another class you wrote, you can use inheritance. When one class inherits from another, it takes on the attributes and methods of the first class. The original class is called the parent class, and the new class is the child class. The child class can inherit any or all of the attributes and methods of its parent class, but it’s also free to define new attributes and methods of its own.
The __init__() Method for a Child Class
When you’re writing a new class based on an existing class, you’ll often want to call the __init__() method from the parent class. This will initialize any attributes that were defined in the parent __init__() method and make them available in the child class.
As an example, let’s model an electric car. An electric car is just a specific kind of car, so we can base our new ElectricCar class on the Car class we wrote earlier. Then we’ll only have to write code for the attributes and behavior specific to electric cars.
Let’s start by making a simple version of the ElectricCar class, which does everything the Car class does:
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
self.odometer_reading = 0
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
def read_odometer(self):
"""Print a statement showing the car's mileage."""
print(f"This car has {self.odometer_reading} miles on it.")
def update_odometer(self, mileage):
"""
Set the odometer reading to the given value.
Reject the change if it attempts to roll the odometer back.
"""
if mileage >= self.odometer_reading:
self.odometer_reading = mileage
else:
print("You can't roll back an odometer!")
def increment_odometer(self, miles):
"""Add the given amount to the odometer reading."""
self.odometer_reading += miles
class ElectricCar(Car):
"""Represent aspects of a car, specific to electric vehicles."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes of the parent class."""
super().__init__(manufacturer, model, year)
my_tesla = ElectricCar('tesla', 'model s', 2019)
print(my_tesla.get_descriptive_name())
2019 Tesla Model S
At class Car: we start with Car. When you create a child class, the parent class must be part of the current file and must appear before the child class in the file.
At class ElectricCar(Car): we define the child class, ElectricCar. The name of the parent class must be included in parentheses in the definition of a child class.
The __init__() method at def __init__(self, manufacturer, model, year): takes in the information required to make a Car
instance.
The super() function at super().__init__(manufacturer, model, year) is a special function that allows you to call a method from the parent class. This line tells Python to call the __init__() method from Car, which gives an ElectricCar instance all the attributes defined in that method. The name super comes from a convention of calling the parent class a superclass and the child class a subclass.
We test whether inheritance is working properly by trying to create an electric car with the same kind of information we’d provide when making a regular car. At my_tesla = ElectricCar('tesla', 'model s', 2019) we make an instance of the ElectricCar class and assign it to my_tesla. This line calls the __init__() method defined in ElectricCar, which in turn tells Python to call the __init__() method defined in the parent class Car. We provide the arguments 'tesla', 'model s', and 2019.
Aside from __init__(), there are no attributes or methods yet that are particular to an electric car. At this point we’re just making sure the electric car has the appropriate Car behaviors.
The ElectricCar instance works just like an instance of Car, so now we can begin defining attributes and methods specific to electric cars.
Defining Attributes and Methods for the Child Class
Once you have a child class that inherits from a parent class, you can add any new attributes and methods necessary to differentiate the child class from the parent class.
Let’s add an attribute that’s specific to electric cars (a battery, for example) and a method to report on this attribute. We’ll store the battery size and write a method that prints a description of the battery:
class Car:
--snip--
class ElectricCar(Car):
"""Represent aspects of a car, specific to electric vehicles."""
def __init__(self, manufacturer, model, year):
"""
Initialize attributes of the parent class.
Then initialize attributes specific to an electric car.
"""
super().__init__(manufacturer, model, year)
self.battery_size = 75
def describe_battery(self):
"""Print a statement describing the battery size."""
print(f"This car has a {self.battery_size}-kWh battery.")
my_tesla = ElectricCar('tesla', 'model s', 2019)
print(my_tesla.get_descriptive_name())
my_tesla.describe_battery()
2019 Tesla Model S
This car has a 75-kWh battery.
At self.battery_size = 75 we add a new attribute self.battery_size and set its initial value to, say, 75. This attribute will be associated with all instances created from the ElectricCar class but won’t be associated with any instances of Car. We also add a method called describe_battery() that prints information about the battery at v. When we call this method, we get a description that is clearly specific to an electric car.
There’s no limit to how much you can specialize the ElectricCar class. You can add as many attributes and methods as you need to model an electric car to whatever degree of accuracy you need. An attribute or method that could belong to any car, rather than one that’s specific to an electric car, should be added to the Car class instead of the ElectricCar class. Then anyone who uses the Car class will have that functionality available as well, and the ElectricCar class will only contain code for the information and behavior specific to electric vehicles.
Overriding Methods from the Parent Class
You can override any method from the parent class that doesn’t fit what you’re trying to model with the child class. To do this, you define a method in the child class with the same name as the method you want to override in the parent class. Python will disregard the parent class method and only pay attention to the method you define in the child class.
Say the class Car had a method called fill_gas_tank(). This method is meaningless for an all-electric vehicle, so you might want to override this method. Here’s one way to do that:
class ElectricCar(car):
--snip--
def fill_gas_tank(self):
"""Electric cars don't have gas tanks."""
print("This car doesn't need a gas tank!")
Now if someone tries to call fill_gas_tank() with an electric car, Python will ignore the method fill_gas_tank() in Car and run this code instead. When you use inheritance, you can make your child classes retain what you need and override anything you don’t need from the parent class.
Instances as Attributes
When modeling something from the real world in code, you may find that you’re adding more and more detail to a class. You’ll find that you have a growing list of attributes and methods and that your files are becoming lengthy. In these situations, you might recognize that part of one class can be written as a separate class. You can break your large class into smaller classes that work together.
For example, if we continue adding detail to the ElectricCar class, we might notice that we’re adding many attributes and methods specific to the car’s battery. When we see this happening, we can stop and move those attributes and methods to a separate class called Battery. Then we can use a Battery instance as an attribute in the ElectricCar class:
class Car:
--snip--
class Battery:
"""A simple attempt to model a battery for an electric car."""
def __init__(self, battery_size=75):
"""Initialize the battery's attributes."""
self.battery_size = battery_size
def describe_battery(self):
"""Print a statement describing the battery size."""
print(f"This car has a {self.battery_size}-kWh battery.")
class ElectricCar(Car):
"""Represent aspects of a car, specific to electric vehicles."""
def __init__(self, manufacturer, model, year):
"""
Initialize attributes of the parent class.
Then initialize attributes specific to an electric car.
"""
super().__init__(manufacturer, model, year)
self.battery = Battery()
my_tesla = ElectricCar('tesla', 'model s', 2019)
print(my_tesla.get_descriptive_name())
my_tesla.battery.describe_battery()
2019 Tesla Model S
This car has a 75-kWh battery.
At class Battery: we define a new class called Battery that doesn’t inherit from any other class.
The __init__()method at def __init__(self, battery_size=75): has one parameter, battery_size, in addition to self. This is an optional parameter that sets the battery’s size to 75 if no value is provided. The method describe_battery() has been moved to this class as well def describe_battery(self):.
In the ElectricCar class, we now add an attribute called self.battery self.battery = Battery(). This line tells Python to create a new instance of Battery (with a default size of 75, because we’re not specifying a value) and assign that instance to the attribute self.battery. This will happen every time the __init__() method is called; any ElectricCar instance will now have a Battery instance created automatically.
We create an electric car and assign it to the variable my_tesla. When we want to describe the battery, we need to work through the car’s battery attribute:
my_tesla.battery.describe_battery()
This line tells Python to look at the instance my_tesla, find its battery attribute, and call the method describe_battery() that’s associated with the Battery instance stored in the attribute.
This looks like a lot of extra work, but now we can describe the battery in as much detail as we want without cluttering the ElectricCar class. Let’s add another method to Battery that reports the range of the car based on the battery size:
class Car:
--snip--
class Battery:
--snip--
def get_range(self):
"""Print a statement about the range this battery provides."""
if self.battery_size == 75:
range = 260
elif self.battery_size == 100:
range = 315
print(f"This car can go about {range} miles on a full charge.")
class ElectricCar(Car):
--snip--
my_tesla = ElectricCar('tesla', 'model s', 2019)
print(my_tesla.get_descriptive_name())
my_tesla.battery.describe_battery()
my_tesla.battery.get_range()
2019 Tesla Model S
This car has a 75-kWh battery.
This car can go about 260 miles on a full charge.
The new method get_range() performs some simple analysis. If the battery’s capacity is 75 kWh, get_range() sets the range to 260 miles, and if the capacity is 100 kWh, it sets the range to 315 miles. It then reports this value. When we want to use this method, we again have to call it through the car’s battery attribute at my_tesla.battery.get_range(). The output tells us the range of the car based on its battery size.
Modeling Real-World Objects
As you begin to model more complicated things like electric cars, you’ll wrestle with interesting questions. Is the range of an electric car a property of the battery or of the car? If we’re only describing one car, it’s probably fine to maintain the association of the method get_range() with the Battery class. But if we’re describing a manufacturer’s entire line of cars, we probably want to move get_range() to the ElectricCar class. The get_range() method would still check the battery size before determining the range, but it would report a range specific to the kind of car it’s associated with. Alternatively, we could maintain the association of the get_range() method with the battery but pass it a parameter such as car_model. The get_range() method would then report a range based on the battery size and car model.
This brings you to an interesting point in your growth as a programmer. When you wrestle with questions like these, you’re thinking at a higher logical level rather than a syntax-focused level. You’re thinking not about Python, but about how to represent the real world in code. When you reach this point, you’ll realize there are often no right or wrong approaches to modeling real-world situations. Some approaches are more efficient than others, but it takes practice to find the most efficient representations. If your code is working as you want it to, you’re doing well! Don’t be discouraged if you find you’re ripping apart your classes and rewriting them several times using different approaches. In the quest to write accurate, efficient code, everyone goes through this process.
Importing Classes
As you add more functionality to your classes, your files can get long, even when you use inheritance properly. In keeping with the overall philosophy of Python, you’ll want to keep your files as uncluttered as possible. To help, Python lets you store classes in modules and then import the classes you need into your main program.
Importing a Single Class
Let’s create a module containing just the Car class (car. py):
"""A class that can be used to represent a car."""
class Car:
"""A simple attempt to represent a car."""
def __init__(self, manufacturer, model, year):
"""Initialize attributes to describe a car."""
self.manufacturer = manufacturer
self.model = model
self.year = year
self.odometer_reading = 0
def get_descriptive_name(self):
"""Return a neatly formatted descriptive name."""
long_name = f"{self.year} {self.manufacturer} {self.model}"
return long_name.title()
def read_odometer(self):
"""Print a statement showing the car's mileage."""
print(f"This car has {self.odometer_reading} miles on it.")
def update_odometer(self, mileage):
"""
Set the odometer reading to the given value.
Reject the change if it attempts to roll the odometer back.
"""
if mileage >= self.odometer_reading:
self.odometer_reading = mileage
else:
print("You can't roll back an odometer!")
def increment_odometer(self, miles):
"""Add the given amount to the odometer reading."""
self.odometer_reading += miles
At the first line we include a module-level docstring that briefly describes the contents of this module. You should write a docstring for each module you create.
Now we can make a separate file called (my_car. py). This file will import the Car class and then create an instance from that class:
from car import Car
my_new_car = Car('audi', 'a4', 2019)
print(my_new_car.get_descriptive_name())
my_new_car.odometer_reading = 23
my_new_car.read_odometer()
The import statement tells Python to open the car module and import the class Car. Now we can use the Car class as if it were defined in this file. The output is the same as we saw earlier:
2019 Audi A4
This car has 23 miles on it.
Importing classes is an effective way to program. Picture how long this program file would be if the entire Car class were included. When you instead move the class to a module and import the module, you still get all the same functionality, but you keep your main program file clean and easy to read. You also store most of the logic in separate files; once your classes work as you want them to, you can leave those files alone and focus on the higher-level logic of your main program.
Storing Multiple Classes in a Module
You can store as many classes as you need in a single module, although each class in a module should be related somehow. The classes Battery and ElectricCar both help represent cars, so let’s add them to the module (car. py).
"""A set of classes used to represent gas and electric cars."""
class Car:
--snip--
class Battery:
"""A simple attempt to model a battery for an electric car."""
def __init__(self, battery_size=75):
"""Initialize the battery's attributes."""
self.battery_size = battery_size
def describe_battery(self):
"""Print a statement describing the battery size."""
print(f"This car has a {self.battery_size}-kWh battery.")
def get_range(self):
"""Print a statement about the range this battery provides."""
if self.battery_size == 75:
range = 260
elif self.battery_size == 100:
range = 315
print(f"This car can go about {range} miles on a full charge.")
class ElectricCar(Car):
"""Represent aspects of a car, specific to electric vehicles."""
def __init__(self, manufacturer, model, year):
"""
Initialize attributes of the parent class.
Then initialize attributes specific to an electric car.
"""
super().__init__(manufacturer, model, year)
self.battery = Battery()
Now we can make a new file called (my_electric_car. py), import the ElectricCar class, and make an electric car:
from car import ElectricCar
my_tesla = ElectricCar('tesla', 'model s', 2019)
print(my_tesla.get_descriptive_name())
my_tesla.battery.describe_battery()
my_tesla.battery.get_range()
This has the same output we saw earlier, even though most of the logic is hidden away in a module:
2019 Tesla Model S
This car has a 75-kWh battery.
This car can go about 260 miles on a full charge.
Importing Multiple Classes from a Module
You can import as many classes as you need into a program file. If we want to make a regular car and an electric car in the same file, we need to import both classes, Car and ElectricCar:
from car import Car, ElectricCar
my_beetle = Car('volkswagen', 'beetle', 2019)
print(my_beetle.get_descriptive_name())
my_tesla = ElectricCar('tesla', 'roadster', 2019)
print(my_tesla.get_descriptive_name())
2019 Volkswagen Beetle
2019 Tesla Roadster
You import multiple classes from a module by separating each class with a comma from car import Car, ElectricCar. Once you’ve imported the necessary classes, you’re free to make as many instances of each class as you need.
Importing an Entire Module
You can also import an entire module and then access the classes you need using dot notation. This approach is simple and results in code that is easy to read. Because every call that creates an instance of a class includes the module name, you won’t have naming conflicts with any names used in the current file.
Here’s what it looks like to import the entire car module and then create a regular car and an electric car:
import car
my_beetle = car.Car('volkswagen', 'beetle', 2019)
print(my_beetle.get_descriptive_name())
my_tesla = car.ElectricCar('tesla', 'roadster', 2019)
print(my_tesla.get_descriptive_name())
At import car we import the entire car module. We then access the classes we need through the module_name.ClassName syntax.
Importing All Classes from a Module
You can import every class from a module using the following syntax:
from module_name import *
This method is not recommended for two reasons. First, it’s helpful to be able to read the import statements at the top of a file and get a clear sense of which classes a program uses. With this approach it’s unclear which classes you’re using from the module. This approach can also lead to confusion with names in the file. If you accidentally import a class with the same name as something else in your program file, you can create errors that are hard to diagnose. We show this here because even though it’s not a recommended approach, you’re likely to see it in other people’s code at some point.
If you need to import many classes from a module, you’re better off importing the entire module and using the module_name.ClassName syntax. You won’t see all the classes used at the top of the file, but you’ll see clearly where the module is used in the program. You’ll also avoid the potential naming conflicts that can arise when you import every class in a module.
Importing a Module into a Module
Sometimes you’ll want to spread out your classes over several modules to keep any one file from growing too large and avoid storing unrelated classes in the same module. When you store your classes in several modules, you may find that a class in one module depends on a class in another module. When this happens, you can import the required class into the first module.
For example, let’s store the Car class in one module and the ElectricCar and Battery classes in a separate module. We’ll make a new module called (electric_car. py) and copy just the Battery and ElectricCar classes into this file:
"""A set of classes that can be used to represent electric cars."""
from car import Car
class Battery:
--snip--
class ElectricCar(Car):
--snip--
The class ElectricCar needs access to its parent class Car, so we import Car directly into the module. If we forget this line, Python will raise an error when we try to import the electric_car module. We also need to update the Car module (car. py) so it contains only the Car class:
"""A class that can be used to represent a car."""
class Car:
--snip--
Now we can import from each module separately and create whatever kind of car we need:
from car import Car
from electric_car import ElectricCar
my_beetle = Car('volkswagen', 'beetle', 2019)
print(my_beetle.get_descriptive_name())
my_tesla = ElectricCar('tesla', 'roadster', 2019)
print(my_tesla.get_descriptive_name())
2019 Volkswagen Beetle
2019 Tesla Roadster
Using Aliases
As we know aliases can be quite helpful when using modules to organize your projects’ code. You can use aliases when importing classes as well.
As an example, consider a program where you want to make a bunch of electric cars. It might get tedious to type (and read) ElectricCar over and over again. You can give ElectricCar an alias in the import statement:
from electric_car import ElectricCar as EC
Now you can use this alias whenever you want to make an electric car:
my_tesla = EC('tesla', 'roadster', 2019)
Finding Your Own Workflow
As you can see, Python gives you many options for how to structure code in a large project. It’s important to know all these possibilities so you can determine the best ways to organize your projects as well as understand other people’s projects.
When you’re starting out, keep your code structure simple. Try doing everything in one file and moving your classes to separate modules once everything is working. If you like how modules and files interact, try storing your classes in modules when you start a project. Find an approach that lets you write code that works, and go from there.
The Python Standard Library
The Python standard library is a set of modules included with every Python installation. Now that you have a basic understanding of how functions and classes work, you can start to use modules like these that other programmers have written. You can use any function or class in the standard library by including a simple import statement at the top of your file. Let’s look at one module, random, which can be useful in modeling many real-world situations.
One interesting function from the random module is randint(). This function takes two integer arguments and returns a randomly selected integer between (and including) those numbers.
Here’s how to generate a random number between 1 and 6:
>>> from random import randint
>>> randint(1, 6)
5
>>> randint(1, 6)
2
>>> randint(1, 6)
3
Another useful function is choice(). This function takes in a list or tuple and returns a randomly chosen element:
>>> from random import choice
>>> users = ['ahmed', 'ali', 'mohammed', 'omar', 'khalid']
>>> random_winning_user = choice(users)
>>> random_winning_user
'ali'
The random module shouldn’t be used when building security-related applications, but it’s good enough for many fun and interesting projects. You can also download modules from external sources.
Styling Classes
A few styling issues related to classes are worth clarifying, especially as your programs become more complicated.
Class names should be written in CamelCase. To do this, capitalize the first letter of each word in the name, and don’t use underscores. Instance and module names should be written in lowercase with underscores between words.
Every class should have a docstring immediately following the class definition. The docstring should be a brief description of what the class does, and you should follow the same formatting conventions you used for writing docstrings in functions. Each module should also have a docstring describing what the classes in a module can be used for.
You can use blank lines to organize code, but don’t use them excessively. Within a class you can use one blank line between methods, and within a module you can use two blank lines to separate classes.
If you need to import a module from the standard library and a module that you wrote, place the import statement for the standard library module first. Then add a blank line and the import statement for the module you wrote. In programs with multiple import statements, this convention makes it easier to see where the different modules used in the program come from.