I went through the exercise of reviewing the diffs between conus-dp and main to understand what MAAPY (you like that ?) inclusions looked like. Here's a quick summary:
Most of our changes are due to some linter that Eli must be using. That's good and bad. Good b/c we should use linters! Bad because it makes the changes seem like a lot more
FileIO.py has the most changes just like Eli expected. A lot of the changes add some regional pass-through args. This MR proposes some ways to handle that but be backward compatible so we don't have to change the functionality of the original code at all. If we use some type of strong config/settings file (like jupyterhub codebase and others like it do) we can have a powerful way to be able to run this code in "different contexts/runtimes/environments" without changing the main transform/algorithms.
Didn't do an intensive sweep so there might be more patterns we need to apply or get acrobatic about
