Thanks, I've turned to a more elaborate, but probably more fun route.
I installed Fdroid and then Termux and inside Termux I got emacs,
git, and nodejs appearing to work (still need to get evil mode going).
So I can either save the page from the browser and then manipulate
the saved html, or fetch the page without a browser and use an html
parser or something like `cheerio` to generate a DOM that I can then
query.
My other option is to write my own html parser in like C or something.
That sounds like a lotta work, sigh.