On the APCSP exam, there's actually very few points for your code itself. You get a point for identifying an abstraction and a point for identifying an algorithm, but the rest of the points (there are 8 total) are assigned based on your written responses. Analytical questions could be added to each exam–broad enough that no two students should have too much overlap, but narrow enough that it's clear what is expected of a student response.
From there, a few points could be assigned for the quality of the designed experiment, but a majority would be assigned based on the quality of the written response questions.
AP Physics follows a similar model on the regular paper exam–you have to prove some theory and design a complete experiment to do so. I believe it also included one light analysis question.
Lock someone in a room with some reagents, and expect them to produce extracted/cleaned output products of a certain testable purity and mass after N time. Make the process of getting from the reagents to the final product really involved, requiring at least one of every kind of basic reaction/setup taught in the course. (NileRed's "aspirin to tylenol" sequence might be a good example challenge: https://www.youtube.com/watch?v=b0Ejuew2riA)
You know this is "the practical approach", because this is essentially how gangs qualify their chemists. "If you can really make [illegal drug], then make us some! Without doing anything that'll get the police called! In four hours!"
That sort of practicum may be a reasonable test, but could give very different results than the normal written exam.
You propose a lab report, with example data, and I actually really like that idea, but I think it would fail on that last score. Looking back, AP chem is a lot of inorganic reactions (lots of memorization, some principles), acid base work, some organic chem, some lab techniques. It’s fairly scattered. A single (or several) lab reports would not capture all of it, especially given that chemistry doesn’t work perfectly predictably even with a post-graduate understanding. There’s a reason basic science lab courses do boring experiments, it’s because those work reliably.
Anyway, I like the idea, and want to hear more on how to do it, but a lab report would have to be supplemented by something else to really evaluate understanding in the sciences.