A Moment of Monoids

What’s functional programming and why should I care?

  1. Understanding functional programming will make you a better programmer
  2. Dropping FP jargon in conversations will make you seem smarter than you really are
  3. FP isn’t trivial, and learning it will expand your mind

There’s been a lot written about functional programming, most of it in languages like Haskell or Scala.

So let’s do some Python.

Semigroups

A semigroup is a nonempty set G with an associative binary operation.

Here’s a secret. You probably already know what a semigroup is.

You use them every time you add two numbers together or concatenate strings.

Following the definition of a semigroup above, let G be the set of all numbers and + (addition) be our binary operation.

Binary operation simply means a function that acts on two separate objects.

Since we know addition over numbers to be associative i.e. a + (b + c) = (a + b) + c, this means the set of numbers under addition is a semigroup.

Let’s look at some concrete examples in python.

import functools
import operator

# A semigroup is a nonempty set G...

stuff = [2, 3, 4]

# ...with an associative binary operation

multiply = operator.mul

# meaning we can compose those elements together
# almost as if we fold them one on top of another
# until we're left with a single thing

# that's what we're doing when we call `reduce` in this example

total = functools.reduce(multiply, stuff)

assert total == 24

letters = ['h', 'e', 'l', 'l', 'o']

greeting = functools.reduce(operator.add, letters)

assert greeting == 'hello'

# more often, we use the built-in sum function to reduce
# sets under addition

numbers = [1, 2, 3]

assert sum(numbers) == functools.reduce(operator.add, numbers)

Monoids

Definition I.1.1. For a multiplicative binary operation on G × G, we define the following properties: (i) Multiplication is associative if a(bc) = (ab)c for all a, b, c, ∈ G. (ii) Element e ∈ G is a two-sided identity if ae = ea = a for all a ∈ G.

A monoid is a semigroup with an identity.

Let’s say we were part of an e-commerce site and we had a csv that contained per-customer order totals for a given month.

We want to add up all the money each customer spent to figure out the total spent that month.

import csv
import io

january_order_totals = """
customer,order_total
sam,54.71
john,
andrea,72.11
""".strip()

reader = csv.DictReader(io.StringIO(january_order_totals))

cash_spent_per_customer = [row["order_total"] for row in reader]

print(cash_spent_per_customer)
['54.71', '', '72.11']

We have a minor problem in that we have an empty value – john didn’t spend any money in January.

We solve this by replacing the empty value with an identity.

By identity, we mean a value (a) that when combined with another value (b) will simply return the latter value (b).

For example, for the set of numbers under addition, the identity is 0 because for any number x, x + 0 = x.

The same is true for 1 for numbers under multiplication i.e. x * 1 = x for any number x.

For the set of strings under the concatenation operation, the identity is simply an empty string.

string = "hello, world"

identity = ""

string + identity == string
# cash_spent_per_customer is currently a collection of strings

# this comprehension attempts to convert each string to a floating point number UNLESS it is an empty string, in which case it evaluates to 0, the identity for numbers under addition

cash_spent_per_customer = [
    float(s) if s else 0 for s in cash_spent_per_customer
]

total = sum(cash_spent_per_customer)

print(f"The total spent in january was: {total}")
The total spent in january was: 126.82

We solved our data validation error by creating a monoid.

Cool.

Now you know what semigroups and monoids are.

Drop those terms at the next masquerade ball you attend and you will be the LIFE of the PARTY ;)

Just make you’re you’re actually wearing your mask to hide your identity when you do.