I'd love to read something on how Perplexity+R1 integrates sources into the reasoning part.