Classes and dataclasses

A class bundles data together with the functions that operate on it. It's a template; each instance is a concrete thing made from it. Classes are how you model the "nouns" of a system — and a clean class is just good decomposition with state attached.

Defining a class

__init__ runs when you create an instance and sets up its data; methods are functions that act on self (the instance):

class Counter:
    def __init__(self, start=0):
        self.count = start      # an attribute on this instance

    def increment(self):
        self.count += 1

    def value(self):
        return self.count

c = Counter()       # make an instance
c.increment()
c.value()           # 1

self is the instance the method was called on; c.increment() passes c as self automatically. Each instance has its own count — two counters don't interfere.

dataclasses: classes that mostly hold data

Many classes just hold a few fields. A dataclass generates the boilerplate (__init__, a readable repr, equality) for you:

from dataclasses import dataclass

@dataclass
class Point:
    x: int
    y: int

p = Point(3, 4)
p              # Point(x=3, y=4)  — readable automatically
p == Point(3, 4)  # True — value equality, for free

Compare that to writing __init__ and __eq__ by hand. Reach for @dataclass whenever a class is primarily a typed bundle of fields.

Where to go next

Next: context managers — guaranteeing setup and cleanup with with.