Comment by marginalia_nu

BeautifulSoup and its clones does parsing pretty well. Just extracting the text out of HTML isn't incredibly hard, and metadata is too unreliable to ever be much use.