Hi @JoeJimFlood this is great, I can take a closer look but a couple of things off the top of my head:

• I want to write some unit tests to confirm it works, and locks in those gains.
• I want to also make sure it doesn't mess with downstream summary outputs. I remember there was an issue where it doesn't like when you request the zone totals for a zone above the smallest zone. Need to track down that issue.

If you can tackle the first one before I can get to this PR, it would definitely expedite things!

@nick-fournier Thanks for reviewing it! I actually had already agreed to write a unit test, though figuring out exactly how has been a challenge. Right now I'm leaning towards running the CALM example with and without a geography in between the District and PUMA levels and checking that the outputs are the same.

@JoeJimFlood I fixed the surface issue of failed linting, that was easy, but I started running into some other issues with test failures. I think they have to do with linux/windows nuances. Which maybe that is what you might be running into? Will report back soon.

These look to just be linting errors, uv run ruff check --fix . will fix all but one, which is just a simple deletion of one line of dead code.