Lecture 25

Object-oriented Programming 3

Inheritance

MCS 260 Fall 2020
David Dumas

Reminders

Goals

More overloading

Recall operator overloading means writing code to give built-in operators custom behavior when applied to your classes.

Last time: Custom equality test with __eq__.

Now: Custom addition with __add__.

We continue with geom.py from Lecture 24.

How should we add two instances of Rectangle?

Idea: Define R+S to be the smallest rectangle that contains both R and S.

Inheritance

Complex programs may have many classes.

Often, some classes have a "is-a" relationship: One represents a more specific type of object than another.

e.g. Dresser is a FurnitureItem

More restrictive classes can have specialized functions (e.g. open_drawer(idx)) and attributes (e.g. ndrawers).

In OOP, is-a relationships are formalized through inheritance. The more specific class is a subclass of the more general one.

Subclasses inherit all methods and attributes from their superclass, but these can be changed or added to in the subclass definition.

Syntax: class Dresser(FurnitureItem):

Class hierarchy example

Class hierarchy example

In geom module?

Circle and Rectangle share a lot of behavior—should both be subclasses of another class?

This is worth considering, but we won't do it today.

What if we want to add a class Square? Since any square is a rectangle, we should make Square a subclass of Rectangle.

super()

In a method of a subclass, super() returns an modified view of the current object that behaves like an instance of the superclass.

e.g. In a Square object, super() returns a version of the same object that will act like a Rectangle.

super() is often used to call the superclass constructor.

__class__

Every object has an attribute __class__ that refers to its class.

In a method body, self.__class__.__name__ gives the name of the class as a string.