When you're working on something intricate, the owner's manual is never enough-- that's why car manufacturers sell the workshop manuals that technicians use on a subscription basis. When I wanted to do a bunch of wiring work in my car, I realized I'd need the service manual, and wanted to download the entire thing, but since each page is its own HTML page and the manual has thousands of pages, it'd take forever to do manually. Since I couldn't find a downloader, I wrote one!

‍

Turns out, with some reverse engineering and a tree traversal, you can download the entire thing from the table of contents, which this project does. I wrote apps for both Toyota and Ford (surprisingly popular) to PDF the entire thing from a bunch of webpages using TypeScript and Playwright (similar to Selenium). The methodology is simple, but the technology serving these manuals is 15-20 years old with a few patches added here and there-- Ford's doesn't even require authentication to download individual pages/images, just to search through them! This age adds all sorts of complications, from oddly nested <i> blocks (6-8 of them between pieces of content) to insanely complex cookie requirements to authenticate for certain diagrams.

‍

My favorite part of these projects, by far, is how many people have been able to dive through the developer tools to get the needed information and run the app. I'm so honored to see people getting some good use out of my work!