Fix downloadable HTML zip being one HTML page

[Gold856]

One local build later, it turns out that RTD will actually run the build commands in the source directory, unless you override the build command, where it just places you at the root repo, with the difference not being super obvious.

Just to be sure this works in RTD-land, I'm first going to add it as a post_build command, which will execute on PRs and will actually give you a non-functional link that you can fix up to then use to download the docs. After we confirm that it spits out the ZIP correctly, then we can integrate it properly.

[Gold856]

https://frc-docs--3159.org.readthedocs.build/_/downloads/en/3159/htmlzip/ (created by fusing the base PR docs domain with the path to the zip, which you can get by clicking the RTD flyout and copying the link from Downloads > HTML) is an archive that has the multipage HTML files. The post_build command had to be modified a tiny bit because I couldn't just build directly into the htmlzip directory, but that's where the normal htmlzip build output goes, so that should be fine. (emphasis on should) If you think it looks good, let me know, and I can uncomment out the build block and remove the post_build command.