5.20. First class

A class groups data and the functions that act on that data under a single name. Reaching for one is worthwhile when several pieces of state belong together and most of the functions in a script take that same set of values as arguments.

5.20.1. The class keyword

The keyword is class. The body holds an __init__ method that runs every time a new instance is created:

class Point:
    def __init__(self, x, y):
        self.x = x
        self.y = y

p = Point(3, 4)
print(p.x, p.y)

Output:

3 4

Point is the class; p is an instance of it. The parentheses on the call Point(3, 4) invoke __init__, which assigns the arguments to attributes on the new instance. self is the conventional name for the instance; self.x sets an attribute that the rest of the instance’s methods will read.

5.20.2. When to use a class

Reach for a class when these are true at the same time:

• A handful of related values travel together everywhere (a point’s x and y; a sensor reading’s value, unit, and timestamp).

• Several functions take that same set of values and operate on them.

• It is useful to keep more than one of these bundles alive at a time, distinct from each other.

If the data is one-off, a plain dict or tuple is usually enough. If you only need a single bundle of constants, module-level variables are fine. Classes are not the default container – they earn their place when state and behaviour want to stick together.

5.20.3. What __init__ does

__init__ is just a regular method. Anything assigned to self inside it becomes an attribute on the new instance:

class Sensor:
    def __init__(self, name, unit):
        self.name = name
        self.unit = unit
        self.readings = []

temp = Sensor("temperature", "C")
print(temp.name, temp.unit, temp.readings)

Output:

temperature C []

readings is initialised to an empty list – each new instance gets its own list, not a shared one. Anything not assigned to self is local to __init__ and disappears when it returns.