This:
"unparsed-text(document-uri(.))"
is finding the URI of the document containing the context node (which is by definition an XML document), and is then rereading the same document but this time without parsing it. I can just about, with a lot of imagination, think of cases where you might want to do that, but I certainly can't relate it to your problem as you described it.
Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference