http://bentasker.i2p/posts/blog/software-development/automatically-preserving-linked-urls-to-defend-against-link-rot.html
The xpath filter allows you to instead specify where in the page it should look for links. For example, on www.bentasker.co.uk all post content appears under a div who's parent div is labelled with itemprop="articleBody text" : So, my xpath filter looks for that and then extracts any links that occur under a p element, with the result that the script will only extract links from within the post body.