* Break apart CrawlerRetreiver * Break apart HttpFetcher into an interface and impl for testing sanity * Add special logic for Lemmy, Mediawiki and Discourse to not waste requests on paths that aren't interesting.
Updated the experiment runner, as well as the script.