Contents

1. When to use classes
2. Anatomy of a class
3. Inheritance

When to use classes

While Python doesn’t force you to use classes, they’re useful in a number of situations:

1. When you want to centralize storage of multiple data structures. For example, you may want one class to organize data that might be spread across several different dictionaries.
2. To extend built-in data structures. For example, let’s say you want to keep track of counts of the number of times you’ve seen different words in a file, so you have a dictionary with words as keys and counts as values. You may want to keep track of the sum of all the values so you know the total count of how many words you’ve counted. You can create something that inherits from a dictionary but adds a wait to count the total.

Anatomy of a class

For now, let’s consider a class whose goal it is to count upwards from zero and reset the count if requested. This class is contained in counter.py.

class Counter(object):
    """Simulate a simple hand counter."""

    def __init__(self):
        """Initialize a counter to zero."""
        self.count = 0

    def __str__(self):
        """Return a string of the count."""
        return str(self.count)

    def __repr__(self):
        """Return a machine-readable string of the count."""
        return "<Counter: {}>".format(self.count)

    def increment(self, amount):
        """Increment the counter."""
        self.count += amount

    def reset(self):
        """Reset the counter to zero."""
        self.count = 0

Usage of this would look like:

>>> c = Counter() # Initialize a new counter
>>> print(repr(c))
<Counter: 0>
>>> print(c) # Same as str(c)
0
>>> c.increment(1)
>>> c
<Counter: 1>
>>> c.increment(2)
>>> c
<Counter: 3>
>>> c.reset()
>>> c
<Counter: 0>

A few things to note:

1. Unlike many object-oriented languages, there’s no formal way to define public/private methods and members. Informally, underscores are used: count is public, while _count is private.
2. Similarly, the use of get/set functions for simple classes is discouraged.
3. Every method must take self as the first argument.
4. The definition Counter(object) says that Counter inherits from object, which is the standard base type. In order to get Python to use its “new-style” way of implementing classes, you must always explicitly tell it which class to inherit from, even if it’s the default one, object.
5. Special method names are surrounded by double underscores. __init__ is the constructor, __str__ provides a “pretty” string representation, and __repr__ provides a “machine-readable” string representation.

Inheritance

When inheriting from another class, you need to specify which class it is and override whichever methods you want to. You do not have to override __init__, but if you do make sure to explicitly call the base class’s __init__.

To see which methods you want to override in a built-in class, take a look at the Python data model documentation.

Here’s an example of overriding __setitem__ to allow a dictionary to track the sum over all its values. You can download the full example, totaldict.py.

class TotalDict(dict):
    """Track the total value of all dictionary keys."""

    def __init__(self):
        dict.__init__(self)
        self.total = 0

    def __setitem__(self, key, value):
        prev_value = self[key] if key in self else 0
        dict.__setitem__(self, key, value)
        self.total += value - prev_value

Here’s an example of this class in use:

>>> t = TotalDict()
>>> t.total
0
>>> t['a'] = 5
>>> t.total
5
>>> t['a'] = 3
>>> t.total
3
>>> t['b'] = 5
>>> t.total
8
>>> t['a'] = 7
>>> t.total
12
>>> t['a'] = -100
>>> t.total
-95
>>> t.values()
[-100, 5]
>>> sum(t.values())
-95