page.page_contents_add stalls forever about pikepdf

[jbarlow83]

What happens if you try...

        if idx == 0:
            watermark_file = Pdf.open('watermark.pdf')
            watermark = watermark_file.pages[0]
            watermark_cs = watermark.Contents.read_bytes()
            for pdf_page in pdf.pages:
                pdf_page.page_contents_add(contents=watermark_cs)

Also, is this the latest version?

[DanielBalsam]

Thanks for the response!

Same problem unfortunately. I am on the latest version.

[jbarlow83]

The code sample is not executable - some missing variables and definitions. So I am to some extent guessing at the intent.

I'm not sure how it's able to execute, because my approximate to it immediately gave this error:

TypeError: page_contents_add(): incompatible function arguments. The following argument types are supported:
    1. (self: pikepdf._qpdf.Object, contents: pikepdf._qpdf.Object, prepend: bool=False) -> None

It is no longer necessary to hold source files open. You can do it more like this:

import pikepdf
from pikepdf import Pdf
import io

wm = 'tests/resources/formxobject.pdf'

def test(input_files, output_stream):
    merger = Pdf.new()

    for idx, input_file in enumerate(input_files):
        with Pdf.open(input_file) as pdf:
            if idx == 0:
                watermark_file = Pdf.open(wm)
                watermark = watermark_file.pages[0]

                for pdf_page in pdf.pages:
                    pdf_page.page_contents_add(contents=watermark.Contents)

            merger.pages.extend(pdf.pages)

    merger.save(output_stream)
    return True

input_files = ['tests/resources/sandwich.pdf', 'tests/resources/graph.pdf']
with open('issue43.pdf', 'wb') as output_stream:
    test(input_files, output_stream)

Another point to note is that unless the watermark is entirely vector art, it will not transfer when used that way. You have to transfer any resources referenced in the content stream as well, such as fonts or images, to the target page. The optimal way to do this is use QPDF's new feature to convert a page to a Form XObject (a "page in a box") but I haven't had a chance to integrate that API. Have at look weave.py in ocrmypdf for an example of watermarking pages with text.

[jbarlow83]

Please check again with the latest version. It's also possible this was related to a GIL issue that was fixed in a dependency.

[jbarlow83]

Closing due to age