Access the fine grained document parsing result. This representation includes individual word locations on the page, as well as per-page handwriting detection. Available for PDF, image, Word, and HTML documents. Not available for table (xlsx/csv), text, or JSON documents.