Metadata - display the already set ones

[leomayer]

I'm not quite clear if it makes sense and if its within the scope of this project. Nevertheless an idea:

Currently its quite hard to see what kind of Metadata is set. With the pdfstandard ua-2 the data is set in the xmp - xmln structure. I think except Adobe Acrobat none of the PDF-Viewers are currently displaying this data. I know that I could unzip the PDF and take a manual look into the source-file. Nevertheless this is an extra effort.

I would find it useful to have them displayed in the analysed structure as well. Honestly I don't know about the difficulty it is but would find it beneficial.

[davidcarlisle]

As hinted by the name, the original idea was to (just) show the structure tree but there is a natural pressure to show more.

In particular:

• associated files can be attached to the StructTreeRoot but as that is currently modelled as the root of the XML document, not as an element, it can't have attributes in that representation.
• General information such as PDF version might be useful
• all the metadata is essentially XML already so could be shown somewhere.

It's not clear if there ought to be separate tools to extract these things or if the output schema should be modified with a new higher level root and the current output moved under a <StructTreeRoot> element, leaving space for other XML elements as siblings of the Structure tree.

On the other hand we don't want to make this a full PDF data extractor, other tools exist for that eg

qpdf --json file.pdf

gives a JSON representation of a PDF file

@u-fischer @zauguin ?

[leomayer]

About a different tool for extracting the XMP-data: I would guess

exiftool -XMP -s2 -b file.pdf

is the better choice for displaying the metadata. Metadata - in primary sense of what is set by TeX. Since all of the data are set in the same <rdf:Description ...> it might be hard to extract them separately. At least in the example of the ISO/DIS 32000-2 they are split into different parts - hinting which belong to the Dublincore.

Anyway this is far off topic for the current request.

I use your provided tool to display easily the structure - since Adobe is a struggle for testing. Your tool is helping me on this part already.

[u-fischer]

You don't need to uncompress the pdf to see the XMP metadata, they are stored uncompressed in the PDF. Beside this you can export them to a file:

\DocumentMetadata{.... ,debug={xmp-export}}

will write the data into \jobname.xmpi. See texdoc l3pdfmeta.

[u-fischer]

if the output schema should be modified with a new higher level root and the current output moved under a element

@davidcarlisle that sounds like a good idea.