Seems to work great with schema embedded in the document, but it does not pick up json/ld script tag microdata at the bottom of the document.