Will pymupdf support PAX attribution?

[xsank]

Some paper use PAX as cite
eg:

α∗
th
{'kind': 4, 'xref': 314, 'from': Rect(144.49099731445312, 105.655029296875, 148.4770050048828, 123.5889892578125), 'nameddest': 'PAX%40.%2Fpapers%2F224_Paper.pax%401', 'id': ''}
e1
{'kind': 4, 'xref': 316, 'from': Rect(249.94500732421875, 505.87701416015625, 257.9150085449219, 520.0850219726562), 'nameddest': 'PAX%40.%2Fpapers%2F224_Paper.pax%402', 'id': ''}
(Head
r sum-
{'kind': 4, 'xref': 317, 'from': Rect(264.44500732421875, 623.175048828125, 291.1260070800781, 636.9190063476562), 'nameddest': 'PAX%40.%2Fpapers%2F224_Paper.pax%403', 'id': ''}

Pymupdf could find the from rect, but missing the to point

I could not find any related docs in the website：https://pymupdf.readthedocs.io/en/latest/search.html?q=PAX&check_keywords=yes&area=default

will pymupdf support this feature? thanks

[JorjMcKie]

This is already supported!
You are showing so-called "named" destination links. They do not have a "to" attribute because the symbolic name given under "nameddest" has all that information.
Please use a variant of the import statement import fitz_new as fitz to activate full named destination support.
This "rebased" implementation will replace the current, "classic" implementation during the first half of January.

[xsank]

nameddest only contains the information of PAX@./papers/224_Paper.pax@5

[xsank]

In fact, cite have the to attribution,

[JorjMcKie]

In fact, cite have the to attribution,

What is cite?

[JorjMcKie]

Or are you referring to the HTML tag <cite>?
If yes:
We are talking of PDFs here - there is not such thing in PDF. What you seem to have in your PDF are named stinations. And those are supported by PyMuPDF as mentioned.

[JorjMcKie]

I am going to move this to the "Discussions" tab. This is no issue.

[xsank]

 an
{'kind': 4, 'xref': 6, 'from': Rect(229.4770050048828, 636.8400268554688, 235.45399475097656, 649.3610229492188), 'page': 0, 'to': Point(87.006, 80.829), 'zoom': 0.0, 'nameddest': 'Hfootnote.3', 'id': ''}
(Sinha et al.,
for scholarly
{'kind': 4, 'xref': 7, 'from': Rect(445.02099609375, 501.156005859375, 493.64898681640625, 512.0999755859375), 'page': 7, 'to': Point(301.161, 635.532), 'zoom': 0.0, 'nameddest': 'cite.Sinha2015AnOO', 'id': ''}
 2015)
data in
{'kind': 4, 'xref': 8, 'from': Rect(497.67999267578125, 501.156005859375, 519.5980224609375, 512.0999755859375), 'page': 7, 'to': Point(301.161, 635.532), 'zoom': 0.0, 'nameddest': 'cite.Sinha2015AnOO', 'id': ''}

like this. the cite, footnote, PAX are all PDF attributions. so that the pdf reader could jump the specific position of the page.

[JorjMcKie]

You obviously used import fitz_new this time.
If you extract the page links via page.get_links() in this implementation, then indeed the method already does the destination resolution for your comfort, so you will find "to", "page" and other attributes.

But you still misunderstand: the words in the name, like "footnote", "cite" have no meaning in PDF! It is just an identifying string - not more. Could also be "abracadabra123_xyz" without any adverse effect.

So I hope you will accept that your initial requirement is already satisfied - no enhancement necessary.

[xsank]

cite, footnote have the to point, but PAX not.

[JorjMcKie]

Well, if you still have concerns then please attach that problem PDF here.

[JorjMcKie]

Let me try another explanation, because you seem to be unaware of what PDF offers as hyperlink kinds:

• GoTo: jump to another page inside the same document
• GoToR: jump to some page in another document
• Launch: start an application or open a document (no jumping to a particular point)
• URI: jump to a resource given in HTML syntax. Can be again a file or an internet resource, etc.)
• Named: jump to a named destination. Details are encoded in the definition of the symbolic name.

All of the above are supported by PyMuPDF.

• GoToE: similar to GoToR, but the document is contained as an attachment inside the current document. This is not supported by PyMuPDF.

There is nothing outside the above list which can be used as a hyperlink in PDF.
Again: things like "cite", "quote", "footnote" are undefined in PDF. The same is true for "PAX" ... whatever that may be.

[xsank]

this discussions could be closed, i found the full points in the doc.resolve_names(), thanks again finally.