PDF to Image conversion skips signature image created with Adobe Reader

[Tlaloc-Es]

Hello,

I'm trying to convert a PDF document into an image using the render method in version 4.2.0 of the library. However, the signature image created with Adobe Reader is not being rendered in the resulting image.

Here's the code snippet I'm using:

pdf_bitmap = page.render(
    scale=3,
    rotation=0,
    crop=(0,0,0,0),
    grayscale=False,
    optimize_mode=None,
    draw_annots=True
)

Is there a way to include the signature image in the resulting image? Any help or suggestion would be greatly appreciated. Thank you.

[mara004]

Maybe the image is part of an AcroForm or something? Try calling pdf.init_forms() directly after constructing the PdfDocument. If this does not fix your problem, I'd need to see the PDF in question.

[Tlaloc-Es]

Here's the corrected text with minor edits and fixed typos:

We tried the following code:

pdf = pdfium.PdfDocument(source)
pdf.init_form()
page = pdf.get_page(0)
pdf_bitmap = page.render(
    scale=3,
    rotation=0,
    crop=(0,0,0,0),
    grayscale=False,
    optimize_mode=None,
    draw_annots=True
)
pdf_bitmap.to_pil().save(target)

However, it did not work and it's possible that the PDF is an AcroForm. How can I check this? Unfortunately, I cannot share the document as it's confidential. PDF2IMG works but it's too slow. I am not sure if this will help identify why the signature is not appearing

[mara004]

If the above does not work, then it's probably not about forms, but you could run
print( pdfium.internal.FormTypeToStr.get( pdf.get_formtype() ) ) to confirm.

If you can't share a PDF that demonstrates the problem, then I'm afraid we can't help you.

Apart from that, what's PDF2Miner? Is this another PDF rendering engine? Could you share a link?

[Tlaloc-Es]

Hi, sorry the library is PDF2IMG was a typo, the function returns None.

[mara004]

Hmm, could you maybe provide/craft a non-confidential PDF that shows the problem (or whoever has such a document or is capable of creating one) ?
Otherwise, I'll have to close this issue. Also, this sounds like a problem with PDFium itself, not with the Python bindings.

[Tlaloc-Es]

I apologize, but I was unable to create the PDF file as it was created in an office and I am not familiar with the process. However, I found that another library which also uses pdfium had faced a similar issue during implementation. You may find this information useful, and I recommend taking a look at this link: pvginkel/PdfiumViewer#87.

[mara004]

Thanks, the linked test file (archive url) shows the problem. I'll investigate.

[mara004]

I found and fixed an issue with the multi-page renderer not initializing the formenv in worker processes. The signature of the file mentioned in #191 (comment) (which is part of an AcroForm) now displays correctly with the multi-page renderer (as used by the pypdfium2 render CLI).
But this does not look like your issue, because you're using the single-page renderer, and your document has no forms according to #191 (comment).

[mara004]

pypdfium2 render SampleSignedPDFDocument.pdf -o out/ prints
Unsupported PDF feature: Signature annotation for me (triggered by pdfium.PdfUnspHandler().setup()).

So even if it seems to work for this particular document, it might be a feature not fully supported by PDFium.
Can you check if PDFium in Chrome/Chromium renders your signature PDF correctly? If not, you'd need to it report upstream.

[Tlaloc-Es]

Thank you, mara004. You work very quickly. Unfortunately, I was unable to check the information you provided due to the high security environment in which the PDFs are stored. Therefore, I suggest we close this issue for now. If I am able to access the information in the future, I will be sure to share it with you.

[mara004]

Thanks, feel free to reopen if you get access to the PDF again.
And thanks for making me aware of the forms issue with the document-level renderer, although it was somewhat unrelated.