A brief introduction to MyPy

Disclaimer: I wrote this to crystalize my own learning. I don't claim to be an expert.

---

Synopsis

A brief introduction to MyPy.

Contents:

• Why bother with MyPy.
• Two example scripts.
• Two exercise scripts.

---

Why Bother With MyPy

Python is dynamically typed, but sometimes this can make life harder.

---

Allows you to spot bugs sooner:

Debugging is easy in this trivial example, but in a bigger codebase adding typechecking to this function would reduce risk of accidentally calling it with args of the wrong type:

def get_reachable_hosts(allhosts, badhosts):
    """Do something for every host that you can
    reach from a list of hosts.
    """
    goodhosts = allhosts - badhosts
    for goodhost in goodhosts:
        ...  # do stuff

get_reachable_hosts(['foo', 'bar', 'baz'], set('foo'))

---

Programatically OK ≠ Intended Usage:

Classic example is:

def greet(name):
    print(f'Hello {name}')

greet({42: 'Deep thought'})
# Python handles this
# even if it's not sensible.
>>> Hello {42: 'Deep thought'}

---

Sets typing in the code rather than the docstring:

Docstring relies on human proofing for being up to date.

def function_has_evolved(input_):
   """Do something to a string.

   Args (str): Do something to a string
   """
   return [func(i) for i in input_]

---

Ultimately though, MyPy is optional. You can have both static and dynamic typing for different bits of the same script!

---

How to get a copy of this repo:

Either:

1. Use this link to download and unzip in a location of your choice.
2. On the command line: git clone https://github.com/wxtim/learn-mypy.git. Then cd learn-mypy.

---

Examples of MyPy

1. Brief introduction to the syntax. (Answers)
2. An example where type annotation would have made my life better. (Answers)
3. Exercise 1: Make a broken function fail MyPy validation (Answers)
4. Exercise 2: Use MyPy on some example legacy code. (Answers)