Is there a way to use this library to obtain full concrete syntax tree for whitespace-accurate unparsing?

I need to rewrite URLs in `srcset` attributes, and I want to be as conservative as possible. Can this be used to obtain full concrete syntax tree that can be "mechanically" unparsed without the knowledge of the grammar?