For now, the parsing is very crude, and parses only the document url. The meta queries for the import are ignored.