Many of those scripts will most likely fail within a few weeks, as their data extraction logic is way too simplistic and based on unstable and non-semantic HTML structures (i.e. doc.cssselect('small a')[0] ).