[paulmd]

I've been thinking about writing some website-to-JSON scrapers myself and this basically solves that problem (since I would have been going after CSS selectors or xpath anyway myself). Nice job.