Cohort 3, chapter 19:

[dsollberger]

To aid slidedeck navigation

• separated long slides into separate slides
• folded longer code blocks into details
• hid "TODO" elements

The previous cohort correctly noted that some codes do not work at the moment. I commented out what wasn't working at the moment and set "eval = TRUE" for the rest of the code.

[jonthegeek]

Trying to fix the build.

[dsollberger]

Thanks for checking in on this push request. I cannot seem to push new edits to this fork (the usethis::pr_push() no longer goes through for me). As you mentioned, setting eval = FALSE to each code chunk that had map in it should restore this to a buildable state. Could you do that change for the following code chunks?

• 20-27
• 20-31
• 20-34

[jonthegeek]

Thanks for checking in on this push request. I cannot seem to push new edits to this fork (the usethis::pr_push() no longer goes through for me). As you mentioned, setting eval = FALSE to each code chunk that had map in it should restore this to a buildable state. Could you do that change for the following code chunks?

I'll take care of it! Some tips in case this happens again:

• It's easy to miss, but try to remember to pull (the button on the GitHub tab of RStudio is the easiest way to do it) before making any changes to something you've already submitted. That's likely what you're running into.
• If you miss that and run into issues like you're having, you have to undo your most recent commit. I have to google this every time (and specifically "taught" ChatGPT to respond with a simple version of it if I ask), but you run this in a terminal (the terminal tab in RStudio works): git reset --soft HEAD~1 (the "1" means "go back 1 commit", basically; you can use a higher number if you know you need to go back farther, or just do them one at a time). When you do that, your work will (generally? I'm hedging in case there are situations I'm not thinking of) still be there, but not "checked in". That way you can check how it compares to the changes you've pulled.
• In theory, you should be able to make these changes directly in the WebUI. I'll be doing that for this one, in the review tab of this PR.

Hope that helps!

[jonthegeek]

This turned out to be a complicated artifact of the way we currently build books! Back in Chapter 13, both a "class" object (a data.frame) and a "map" object (a function) are defined. Because we build the book as a single Rmd, those objects still exist in later chapters. That map() call in chapter 19 was the first thing to run into them.

I've added code to 13 to clean them up, and this is yet another argument for me to finally update everything to the new Quarto format (see https://DSLC.io/wapir for an example), which won't have this issue (each chapter is self-contained)! I know there's also a bookdown setting to avoid it, but there are other benefits to the Quarto format, so I'll keep working on getting there!

Anyhoo, this SHOULD work with my next push! I'm building now locally to make sure.

[jonthegeek]

Looks like this took care of it!

This is dealt with.

[lgibson7]

Thanks @jonthegeek !