I'm curious if there's a built-in mechanism for traversing a view of a sentence or a document with knowledge of both named entities as well as tokens. For example, say I'd like to iterate over a sentence yielding the lemma of each non-stopword token, unless it's part of a named entity span, in which case I'd like to yield the surface form of the NE span, and then continuing on to the token immediately following the NE span. Is there a built-in way to handle this? I can't find any relevant examples in the documentation. Thanks for your help.