pdf.getDocumentInfo().title sometimes None

[clach04]

Test case in #511 (comment) (original no longer available test case Details in https://github.com/mstamy2/PyPDF3/issues/13).

pdf = PdfFileReader(f)
info = pdf.getDocumentInfo()
info.title  # reports None
info['/Title']  # works

[clach04]

@MartinThoma is this now resolved? Or can't reproduce as the other project appears to have been removed? I might have the original bug script around if that would be helpful.

[MartinThoma]

#13 is the issue you've linked to. We moved PyPDF2 from mstamy2 -> me (MartinThoma) -> py-pdf (a Github Organization) this week.

As PyPDF2 has been inactive for a long time, I need to clean up a lot of things (PRs, issues, code, documentation, ...).

I'm sorry that I didn't add more details. Could you please check if the issue still persists? If so, please create a MCVE (including a PDF): https://stackoverflow.com/help/minimal-reproducible-example . Then I'll re-open :-)

[clach04]

That's great news @MartinThoma! Thanks for the quick status update. How about any future issues get closed with a link to #657 as an explanation?

Back to this specific issue, this issue has nothing to do with #13, its specifically about the linked issue to a repo that no longer exists. I have the test code in claird#59 but the test pdf is only in the deleted repo (it was about 5Mb). @mstamy2 is the original owner, I noticed he responded to #657 so pinging him in case the repo is simply private rather than deleted as that's where I put all my notes. @mstamy2 do you still have access to the https://github.com/mstamy2/PyPDF3/issues/13?

I can try and check an old hard drive to see if I still have it but I won't have access to it for a while :-(

[clach04]

New test case created based on original :-)

Overview

I've seen a number of PDF files where the title attribute/property is reported as None but when then accessing /Title there is content. I've no idea if this is a problem with the pdf(s) or with PyPDF. There is a workaround (which may be an indication of a potential change to PyPDF but I'm unclear of what the correct thing to do here is)

Attached PDF is about 5Mb and is a sample of a document that exhibits this behavior, I did not create it (nor do I know how it was created) so the only information we have is the metadata inside.

Test case, along with workaround below:

Test PDF file

title_bug.pdf

Test case

EDIT inline version dumps PyPDF version (attached version does not).

inline and attached (rename to .py)
pdf_title_bug.txt

#!/usr/bin/env python
# -*- coding: windows-1252 -*-
# vim:ts=4:sw=4:softtabstop=4:smarttab:expandtab
#

import os
import sys

ver_to_test = 3
ver_to_test = 4
ver_to_test = 2

if ver_to_test == 4:
    from pypdf import PdfFileReader  # https://github.com/claird/PyPDF4
elif ver_to_test == 3:
    from PyPDF3 import PdfFileReader  # https://github.com/mstamy2/PyPDF3
else:
    from PyPDF2 import PdfFileReader  # https://github.com/py-pdf/PyPDF2 - nee https://github.com/mstamy2/PyPDF2 / https://pythonhosted.org/PyPDF2/
    import PyPDF2 as pypdf_lib


print('Python %s on %s' % (sys.version, sys.platform))
print(pypdf_lib.__version__)

filename = 'title_bug.pdf'
f = open(filename, 'rb')
pdf = PdfFileReader(f)
info = pdf.documentInfo
#print(info)
print('title attribute %r' % info.title)  # reports None
print('title getText() %r' % info.getText("/Title"))  # this is what .title property calls
print('title get() %r' % info.get("/Title"))  # this is part of what dict[] does
print('title get().getObject() %r' % info.get("/Title").getObject())  # this is what dict[] does
print('/Title dict entry %r' % info['/Title'])  # with test pdf works
print('title attribute %r' % info.title)  # Sanity check it is still None
print('title Workaround %r' % (info.title or info['/Title']))  # Workaround


f.close()

output

Python 2

Python 2.7.10 (default, May 23 2015, 09:40:32) [MSC v.1500 32 bit (Intel)] on win32
1.27.2
title attribute None
title getText() None
title get() IndirectObject(305, 0)
title get().getObject() u'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'
/Title dict entry u'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'
title attribute None
title Workaround u'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'

Python 3

Python 3.7.3 (v3.7.3:ef4ec6ed12, Mar 25 2019, 22:22:05) [MSC v.1916 64 bit (AMD64)] on win32
1.27.2
title attribute None
title getText() None
title get() IndirectObject(305, 0)
title get().getObject() 'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'
/Title dict entry 'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'
title attribute None
title Workaround 'DIY   ENTERTAINMENT Retro Gaming on Raspberry Pi- Understanding ROMs, RetroPie, Recalbox, and More'

[clach04]

@MartinThoma hopefully this helps. I really know nothing about PDF internals which is why I've not attempted a fix :-( I have a workaround that seems to be effective but not sure if it is reasonable.

If you need anything else from me on this please ping me (I recall I had other PDFs with similar behaviors, this was one of the smaller ones).

Thanks for picking up the torch on this and trying to organize collaboration

[MartinThoma]

this issue has nothing to do with #13, its specifically about the linked issue to a repo that no longer exists

The repo was moved issue 13 from the linked repo is #13 here.

[MartinThoma]

Thank you very much! I now hope that somebody will pick it up and dig into it :-)

[clach04]

updated test case with PyPDF2 version

[clach04]

Located another PDF, appears to be created with the same PDF generator.

Adobe Reader reports; PDF Version 1.3 (Acrobat 4.x). InfoKey: Producer, InfoValue: macOS Version 10.10 Quartz PDFContext

• https://github.com/ValveSoftware/steamlink-sdk/blob/87666bbd3c512fe0aef19cd024fd0f0bf0765fb4/external/qt-everywhere-opensource-src-5.9.1/qtwebengine/src/3rdparty/chromium/third_party/grpc/examples/objective-c/route_guide/Misc/Images.xcassets/second.imageset/second.pdf

This file is much smaller than the attached test PDF.

I also found some Microsoft Word generated PDFs but when I attempted to export/create PDFs from recent Word, the title worked fine.

[clach04]

@MartinThoma

this issue has nothing to do with #13, its specifically about the linked issue to a repo that no longer exists

The repo was moved issue 13 from the linked repo is #13 here.

This is not correct, I've copy/pasted the one email I have where someone posted to the issue I created (note issue, not a PR - in a completely different repo):

From: johns1c [email protected]
Sent: Wednesday, May 6, 2020 5:45 PM
To: mstamy2/PyPDF3 [email protected]
Cc: Chris Clark ; Author [email protected]
Subject: Re: [mstamy2/PyPDF3] pdf.getDocumentInfo().title sometimes None (#13)

Are you running Python 3 by any chance
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub, or unsubscribe.

[MartinThoma]

There is something really weird about that PDF:

1. When you upload it to http://pdf-analyser.edpsciences.org/result/1e54b64d it also gives no title
2. mutool clean -d title_bug.pdf title_bug.txt seems to be catched in an infinite loop (from mupdf-tools)