ENH: Add command to extract annotated pages

[wolfram77]

Hello pdfly contributors, I hope you're doing well. My advisor has repeatedly asked for a way to filter annotated pages from a PDF (thesis). I managed to find a solution using pymupdf, but having a CLI tool for this would be helpful. Any suggestions on how to integrate this feature into pdfly?

[Lucas-C]

Hi @wolfram77 👋

In order to determine if pdfly could include this feature, could you please provide a little more details?

• would this feature only extract pages with annotations from an existing PDF?
• how do you suggest this to be added to pdfly? With a new extract-pages-with-annotations subcommand?

[wolfram77]

Hello @Lucas-C

Thanks for considering my request.

Yes, extracting only pages with annotations from a PDF would be useful, as it would help my guide and other professors filter out the pages they have commented on, especially when reviewing theses.

As for adding this to pdfly, perhaps a shorter annotated-pages command would work, with an optional output file - allowing the output to be something like input_file_annotated.pdf.

[Lucas-C]

Alright.

I think that keep the extract- prefix may be better, to be explicit in what this subcommand does, and for consistency with the existing extract- commands.

Would you like to submit a PR to implement this feature? 🙂

You will find some documentation on how to detect annotations using pypdf2 there:
https://pypdf2.readthedocs.io/en/3.x/user/reading-pdf-annotations.html

[wolfram77]

I tried the following, but it seems to include way too many pages than expected.

    input  = PdfReader(str(input_pdf))
    output = PdfWriter()
    # Copy only the pages with annotations
    for page in input.pages:
        if "/Annots" in page:
            output.add_page(page)
    # Save the output PDF
    output.write(output_pdf)

[Lucas-C]

That's a good start 🙂 👍

There are many kind of PDF annotations :

• file attachments
• Text / FreeText / Ink / Highligh text comments
• PDF signatures
• actions, triggered when opening the document or when clicking in an area...
• even links (internal or external) are annotations!

In order to distinguish between those, you will have to check their /Type

For more information, you can check the PDF specs: https://developer.adobe.com/document-services/docs/assets/5b15559b96303194340b99820d3a70fa/PDF_ISO_32000-2.pdf