pdf-association
[Question] Availability of complete RELAX-NG schemas to validate PDF XMP metadata
ISO 16684-2:2014 describes the use of RELAX-NG schemas to validate XMP metadata. The standard is not PDF focused, but provides a sample schema that validates basic XMP properties that are also used in PDF (e.g. dc:title, dc:creator, xmp:ModifyDate, etc.). Since its license is permissive I actually extracted it and made it available publicly. Extensions that support all the other PDF properties, including those described by PDF/A compliances, are not included in that schema (eg. pdf:Trapped, pdfaid:part, etc.). I noticed recently that extensions schema (in "PDF/A extension schema container schema" format) describing properties that are available in cross formats (eg. basic PDF, PDF/A, PDF/X, etc.) were published, but the main question is: are there RELAX-NG schemas that describes all the XMP properties mentioned in PDF, PDF/A, PDF/X, etc. standards? The same question was also submitted to Adobe XMP github project, to no avail.
To be very clear: PDF/A-4 only recommends ("should") the use of ISO 16684-2 RELAX-NG schema (as an Associated File with AFRelationship == Schema) for custom XMP metadata. There is also an open question as to whether this would then force such PDF/A-4 files to always be conformance level PDF/A-4f due to the AF...
This is unlike all prior versions (PDF/A-1, PDF/A-2, and PDF/A-3) which all required ("shall") that all XMP custom extensions have their full schema declared in XMP Extension Schema syntax (as documented in the PDF/A-1, PDF/A-2, and PDF/A-3 specs) in an XMP metadata stream (so not as a standalone schema). Templates for this are available from: https://pdfa.org/free-xmp-extension-schema-templates/
I should have explained more what I wanted to do, as templates in [1] are not enough for me: basically I wanted to validate the whole XMP packet using the method described in ISO 16684-2:2014, which suggests a procedure to normalize the XMP packet so there won't be ambiguities when validating it against a RELAX-NG grammar. I implemented the procedure myself a couple of years ago (I believe it's still the only public implementation around of that standard), even though I have still to check some edge cases. Now I would like to use that procedure and a standalone RELAX-NG grammar describing the XMP properties in use in PDF/A profiles to validate the XMP packet for these compliances. I understand that the XMP metadata can be extended through extensions, and such standalone grammars won't incluse those, but as a first step it would be enough to validate against a non extended grammar. It's my believing that such grammars still don't exist, at least not as open source/downloadable for free, but it would not be a big task to craft those, based on the basic free grammar I extracted from ISO 16684-2. In few weeks I could be actually working on the task, unless you point me to other public resources I was not aware of.
[1] https://pdfa.org/free-xmp-extension-schema-templates/
Also: this is not really PDF/A-4 specific, as I want to validate PDF/A-1, PDF/A-2, PDF/A-3, so I believe you can remove that tag.
OK - I misunderstood what you meant...
AFAIK (and as you state) there are no official or unofficial RELAX-NG schemas for any of the PDF subsets - but very happy if anyone in the community wishes to offer up theirs...
As promised the WIP schema template (with links to current schema snapshots) can be found on this repository:
https://github.com/ceztko/XMP-RNG-Schema