Error extracting highlights

[edanerg]

When adding a highlight from firefox, this error appears in the docker debug log.

Traceback (most recent call last):
File "/home/nonroot/pdfding/pdf/service.py", line 292, in set_highlights_and_comments
date_time_string = f'{annotation_object["/CreationDate"].split(':')[-1]}-+00:00'
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/nonroot/.venv/lib/python3.12/site-packages/pypdf/generic/_data_structures.py", line 472, in getitem
return dict.getitem(self, key).get_object()
^^^^^^^^^^^^^^^^^^^^^^^^^^^
KeyError: '/CreationDate'

[edanerg]

This is based on v1.4.1

If you run on master you will also hit:
Traceback (most recent call last):
File "/home/nonroot/pdfding/pdf/services/pdf_services.py", line 176, in set_highlights_and_comments
annotation_type = annotation_object["/Subtype"]
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^
File "/home/nonroot/.venv/lib/python3.12/site-packages/pypdf/generic/_data_structures.py", line 472, in getitem
return dict.getitem(self, key).get_object()
^^^^^^^^^^^^^^^^^^^^^^^^^^^
KeyError: '/Subtype'

[edanerg]

I think this is because I used the pdf in macOS preview, and there are some existing annotations that lacked certain keys.

Seems like the logic in pdf_service.py is that even if one of the annotations has an error, the entire annotation extraction fails.

Should protect against this case?

[mrmn2]

Hi, yes I also think that it's probably related to that. I never encountered any problems when extracting highlights.

You are right, it would be better if the extraction continues for other annotations if a problem occurs.