Run download.sh. Some links will go to external domains that you don't want to recursively scrape, so manually download those and put the mappings in mappings.txt. Then run replace.sh.
You might also want to download lecture questions and answers manually (from 6858.csail.mit.edu, which requires auth) as mhtml files (see the rq/ folder). The updated questions.html reflects those newer file paths.
To get Piazza questions, use https://gist.github.com/yasyf/06801c37f1785cc430b52edc87c648ea.