SubDocumentsFilterTest should not use a sitemap file

[jnioche]

this currently uses the JSoupParser - which works only because we allow it to parse when no mimetype is available. We should have a minimalistic HTML doc to do the same OR use the RSS feed and generate subdocuments for each feed item

[jnioche]

The FeedParserBolt does not provide a DOM for the filters to operate on. We could demonstrate the functionality on a HTML doc instead or use JSoupParser on a feed.