Quite surprised that you're scraping HN and parse the DOM. Maybe should try use the API? https://github.com/HackerNews/API