fix: walk ancestor chain, handle lazy-load attrs & relative URLs for images

fix: restore image extraction in HTML scraper

feat: extract dates from HTML-scraped articles

a try