Unable to add signatures to a PDF that has been altered with pypdf

[ag-gaphp]

This is somewhat related to my other discussion, but different enough I think it warrants a new post.

I'm trying to remove some placeholders in my PDF, store their rect, and then re-add those as signatures using pyHanko. My other discussion is more about trying to sort out why pyHanko is detecting field widgets that other toolkits have deleted, while this discussion is more about wondering what I am doing wrong with pyHanko when adding these (or perhaps there is something wrong with the PDF file?)

I'm putting together another test where I remove all the widgets using pypdf instead of PyMuPDF, but I get a different error now:

Entries in field list must be dictionary objects, not <class 'pyhanko.pdf_utils.generic.NullObject'> -- skipping.
# the above line repeats 158 times
ERROR:  'NullObject' object has no attribute 'get'

Here is the sample file after having all the widgets removed using pypdf: test.pdf

This is the snippet:

from pyhanko.pdf_utils.incremental_writer import IncrementalPdfFileWriter
from pyhanko.sign.fields import SigFieldSpec, append_signature_field

# in my full script, I store the rect and page of my placeholders with pypdf to populate the boxes dict
# this is just a copy/pasted example to display the issue and what the dict looks like when populated
boxes = {
    'sigPrimary1': {
        'page': 10,
        'box': (30.84, 190.262, 241.881, 203.299)
    },
    'init29': {
        'page': 10,
        'box': (391.974, 190.262, 495.015, 203.299)
    },
    'sigSecondary1': {
        'page': 10,
        'box': (30.84, 133.541, 241.881, 146.578)
    }
}

with open('test.pdf', 'rb+') as sig_doc:
    writer = IncrementalPdfFileWriter(sig_doc, strict=False)
    for name in boxes.keys():
        _dict = boxes[name]
        try:
            append_signature_field(writer, SigFieldSpec(
                                        sig_field_name=name,
                                        on_page=_dict["page"],
                                        box=_dict["box"]
                                    ))

        except Exception as e:
            print("ERROR: ", e)

    writer.write_in_place()

I guess I'm curious if there's some other way I should be adding signatures in pyHanko. I've tried strict and non-strict mode, but there is no difference that I can see for this scenario.

[ag-gaphp]

I think I sorted this out in another way. In pypdf's method that deletes widgets, it is setting them also to NullObjects, which is why I think this particular error occurs.

I tweaked the method in pypdf to sip over the NullObject part, and now pyHanko is giving me the same results as in my other discussion. Going to close this one in favor of the other.