Add VDV generic extractor, and integrate that into the PDF extractor

This in itself doesn't produce useful output yet (we need more work on
VDV ticket parsing for that), but it is already producing enough to trigger
custom extractors on.