Page merge: 'NullObject' object has no attribute 'get_data'

[stefan6419846]

Applying the fix from #2150 on the existing code and using a PDF where remove_text() and remove_images() have been called on raises an error.

Environment

Which environment were you using when you encountered the problem?

$ python -m platform
Linux-5.14.21-150400.24.81-default-x86_64-with-glibc2.31

$ python -c "import pypdf;print(pypdf._debug_versions)"
pypdf==3.15.5, crypt_provider=('pycryptodome', '3.18.0'), PIL=10.0.0

Code + PDF

This is a minimal, complete example that shows the issue:

from pypdf import PdfReader, PdfWriter

watermark = PdfReader("watermark.pdf").pages[0]

pdf_file = PdfWriter(clone_from="file.pdf")
for page in pdf_file.pages:
    page.merge_page(watermark, over=True)

The watermark is https://github.com/py-pdf/pypdf/files/12428857/watermark.pdf, the cleaned file abc.pdf

Traceback

This is the complete traceback I see:

Traceback (most recent call last):
  File "/home/stefan/temp/pdf/run1.py", line 7, in <module>
    page.merge_page(watermark, over=True)
  File "/home/stefan/temp/venv/lib/python3.9/site-packages/pypdf/_page.py", line 1044, in merge_page
    self._merge_page(page2, over=over, expand=expand)
  File "/home/stefan/temp/venv/lib/python3.9/site-packages/pypdf/_page.py", line 1124, in _merge_page
    original_content = self.get_contents()
  File "/home/stefan/temp/venv/lib/python3.9/site-packages/pypdf/_page.py", line 955, in get_contents
    return ContentStream(self[PG.CONTENTS].get_object(), pdf)
  File "/home/stefan/temp/venv/lib/python3.9/site-packages/pypdf/generic/_data_structures.py", line 1021, in __init__
    stream_data = stream.get_data()
AttributeError: 'NullObject' object has no attribute 'get_data'

[pubpub-zz]

@stefan6419846
PR is available if you want to try it

[stefan6419846]

I can confirm that the PR seems to indeed avoid this issue and generates a correctly watermarked PDF file.

[MartinThoma]

I've re-opened it as I want to improve the test test_get_contents_from_nullobject: Doing the merge_page shows immediately why returning None is desirable at this point.