Execute code chunks

Currently all the code chunks in the lesson are set to not evaluate. I assume this is because most of these functions such as `usethis::*` are not idempotent, and will throw errors if you execute them a second time.

However, I think this could be resolved with some thought, for example by generating a new temporary directory for the growing package each time the lesson is built.

The advantage of this would be:
* Better "testing" of the lesson. We will know when a chunk's code is wrong because it will be run and fail
* Better outputs. We can have the chunks print results that match what the user will see when they run the code themselves

No specific file, test, or entry point is named. Begin by locating the lesson's unevaluated code chunks and the process that builds the lesson, then determine how a fresh temporary package directory can be used for each build. Done means the chunks execute successfully, failures expose incorrect lesson code, and outputs match what users see.

Written by the indexing model from the issue text.