Robust Python

Chapter 6. Customizing Your Typechecker

Typecheckers are one of your best resources for building robust codebases. Jukka Lehtosalo, the lead developer of mypy, offers a beautifully concise definition of typecheckers: “In essence, [a typechecker] provides verified documentation.”¹ Type annotations provide documentation about your codebase, allowing other developers the ability to reason about your intentions. Typecheckers use those annotations to verify that the documentation matches the behavior.

As such, a typechecker is invaluable. Confucius once said, “The mechanic, who wishes to do his work well, must first sharpen his tools.”² This chapter is all about sharpening your typechecker. Great coding techniques can get you far, but it’s your surrounding tooling that takes you to the next level. Don’t stop with just learning your editor, compiler, or operating system. Learn your typechecker too. I will show you some of the more useful options to get the most out of your tools.

Configuring Your Typechecker

I will focus on one of the most popular typecheckers out there: mypy. When you run a typechecker in an IDE (such as PyCharm), it typically runs mypy underneath the hood (although many IDEs will allow you to change the default typechecker). Anytime you configure mypy (or whatever your default typechecker is), your IDE will use that configuration as well.

Mypy offers quite a few configuration options to control the typechecker’s strictness, or the amount of errors reported. The stricter ...