Without having looked into the code to closely, I would think maybe the code could, for each article, branch to a section meant to parse that periodical. For example, if the url was from domain nytimes.com - use the relevant section of the recipe for ny times, if the url was wsj.com, use wall street journal, etc.
I would think we agreed on a selected a set of reliable and diverse sources, we would find many have existing recipes. The initial google news recipe might be configured to use the sources with existing recipes and the other well-behaved sources that might be compatible with a generic recipe. Someone who limited their google news (for at least the account used with calibre) to those identified sources could be confident it would work.
Over time the remaining sources could be addressed. I'm just two days into the learning curve here... Would appreciate feedback or suggestion or a google news recipe!
Last edited by awitko; 10-31-2011 at 05:04 PM.
|