Seemed pretty straightforward after a few tests. The only thing I noticed was regarding unique identifiers, the URL in the sources endpoint didn't seem to exactly match the URL attached to a headline? I could be wrong though, and not a big issue. Was more thinking about data cleanup and linking data easily (article/sources/authors etc).

Thanks very much for the feedback. Yes, I think your point is right. For some domains, I try to match the url with sources from "sources" list, and for others I extract source field directly from the url, I think there is a small bug when doing the former, thanks for pointing out. I will fix it soon.