Description
Using AutoDetectParser to call PDFParser causes the mime-type to be added twice. It should be added exactly once.
Proposed Fix:
parser/pdf/PDFParser.java should be changed from:
metadata.add(Metadata.CONTENT_TYPE, "application/pdf");
to:
metadata.set(Metadata.CONTENT_TYPE, "application/pdf");
as per other Tika bundled parsers.