Robust Python

Chapter 14. Runtime Checking With pydantic

The central theme of robust code is making it easier to detect errors. Errors are an inevitable part of developing complex systems; you can’t avoid them. By writing your own types, you create a vocabulary that makes it harder to introduce inconsistencies. Using type annotations provides you a safety net, letting you catch mistakes as you are developing. Both of these are examples of shifting errors left; instead of finding errors during testing (or worse, in production), you find them earlier, ideally as you develop code.

However, not every error is easily found through code inspection and static analysis. There is a whole class of errors that will only be detectable at runtime. Any time you interact with data supplied from outside your program (such as databases, config files, network requests), you run the risk of inputting invalid data. Your code can be rock-solid in how you retrieve and parse data, but there’s not much you can do to prevent users from passing in invalid data.

Your first inclination might be to write a lot of validation logic: if statements and checks to see if all of the data passed in is correct. The problem is that validation logic is often complex, sprawling, and tough to understand at a glance. The more comprehensive your validation, the worse it gets. If your goal is to find errors, reading all the code (and tests) will be your best shot. In that case, you need to minimize the amount of code you look at. ...