This project relies heavily on scraping the web. Sometimes we scrape undocumented APIs, sometimes we scrape HTML pages. Needless to say, anything can happen. As a result, we have to be specific and consistent with where and how we provide validation.
In this project, we validate all data through types - custom Python classes.
Each data source has a types.py file that define (and validate!) the type. If something passes this type validator, we assume it's trusted.