Querying DOM nodes is inherently more complicated than a regex on unparsed HTML.