Metadata gets updated in the PDF structure, but doesn't reflect in the Adobe Reader (some properties)

[SirishaGorasa]

Issue:

I have been updating the metadata of my pdf document and saving it back to file using set_metadata() function. Before setting metadata, I remove complete metadata using

doc.scrub(attached_files=False, clean_pages=False, embedded_files=False, hidden_text=False, javascript=False, metadata=True,
                redactions=False, redact_images=0, remove_links=False, reset_fields=False, reset_responses=False, thumbnails=False, xml_metadata=False)

Here's how it looks inside pdf structure.

Note: setting xml_metadata to true or false in the `scrub()` method resulted in same thing.

<<
  /Author (Martie Shrader)
  /CreationDate (D:19040229023654Z)
  /Creator (Adobe\256 PageMaker\256 6.5)
  /Keywords (untagged keywords)
  /ModDate (D:20150217093420-07'00')
  /Producer (Acrobat PDFWriter 4.05  for Power Macintosh)
  /Subject (untagged subject)
  /Title (Upload Title)
  /Trapped null
>>

In this example, keywords property didn't reflect in the Adobe reader (Works for some pdfs).

Please do let me know if I am doing something wrong, or anything I need to do, to make it work.

Thanks in advance!

[JorjMcKie]

Which version are you using?
Platform presumably Mac OSX?

[SirishaGorasa]

Which version are you using?
Platform presumably Mac OSX?

PyMuPDF Details:

Name: PyMuPDF
Version: 1.18.15
Summary: PyMuPDF is a Python binding for the document renderer and toolkit MuPDF
Home-page: https://github.com/pymupdf/PyMuPDF
Author: Jorj McKie
Author-email: [email protected]
License: GNU AFFERO GPL 3.0

Platform: Windows 10, python 3.9.2

[JorjMcKie]

Ok, looks good.
Which code did you use to update the metadata?
Cannot reproduce. When I do

import fitz
doc = fitz.open()
page = doc.new_page()
m = doc.metadata
m["keywords"] = "kw1 kw2"
doc.set_metadata(m)
doc.save("x.pdf")

the file show no irregularities in whatever viewer.

[JorjMcKie]

If you want, please share the file and let me have a look.

[SirishaGorasa]

untagged.pdf

Could you please check if the same works on the attached pdf.

[JorjMcKie]

Hm, same thing happening for me 🤔.
All other viewer I am using show the keywords correctly: Firefox, PDF XChange, Foxit Reader, Nitro 5, evince (Linux).
And needless to mention: the PDF object also looks harmless:

Python 3.9.6 (tags/v3.9.6:db3ff76, Jun 28 2021, 15:26:21) [MSC v.1929 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license()" for more information.
>>> import fitz
>>> doc=fitz.open("untagged-mod.pdf")
>>> from pprint import pprint
>>> pprint(doc.metadata)
{'author': 'Martie Shrader',
 'creationDate': 'D:19040229023654Z',
 'creator': 'Adobe® PageMaker® 6.5',
 'encryption': None,
 'format': 'PDF 1.6',
 'keywords': 'kw1,kw2',
 'modDate': "D:20150217093420-07'00'",
 'producer': 'Acrobat PDFWriter 4.05  for Power Macintosh',
 'subject': '',
 'title': '',
 'trapped': ''}
>>> print(doc.xref_object(-1)) # the trailer
<<
  /Size 28
  /Info 14 0 R
  /Root 15 0 R
  /ID [ <182A58473815524CCFA528451E12F3C6> <5BE0747C168AEBD2D843611781631B79> ]
>>
>>> print(doc.xref_object(14)) # info object
<<
  /Author (Martie Shrader)
  /CreationDate (D:19040229023654Z)
  /Creator (Adobe\256 PageMaker\256 6.5)
  /Keywords (kw1,kw2)
  /ModDate (D:20150217093420-07'00')
  /Producer (Acrobat PDFWriter 4.05  for Power Macintosh)
  /Subject null
  /Title null
  /Trapped null
>>
>>> 

Out of good advice, I must say ...

[JorjMcKie]

Ha! I have the reason:
Because the file has XML metadata, Adobe Acrobat uses them for properties display.
There, keywords are empty. If you do doc.del_xml_metadata(), the display works!

[SirishaGorasa]

Thanks @JorjMcKie .
Is it recommended to remove xml metadata? Is there any way we can automatically generate xml metadata for pdf using our original metadata ?

[JorjMcKie]

Is it recommended to remove xml metadata?

Well, at least I recommend it. But you are on your own as per any risks losing potentially important information.
As I wrote in the documentation: I do not include XML manipulation features - on purpose.
If you want, develop something that interprets the PDF's XML metadata extracted via doc.get_xml_metadat(), and feeds information to whereever you like - e.g. the standard metadata.
You can also do the other way round and update XML metadata via doc.set_xml_metadata(xml source) with information you might have taken from doc.metadata.

But again: PyMuPDF is not in the business of dealing with XML syntax - use e.g. lxml if you need something there.

[SirishaGorasa]

Ok! But when we set the xml_metadata=True in the doc.scrub() method, Isn't it supposed to remove any xml metadata present in the document?

[JorjMcKie]

Isn't it supposed to remove any xml metadata present in the document?

Sure - and it does, just tested it, save the resulting file, added keywords information and, voilà, Adobe did show them.

[SirishaGorasa]

Oh yeah, I was just checking for some permissions before I scrub the data. That might be the issue in my case. Thanks for checking!