How to extract internal links using PyPDF

[swathiJayav]

I am trying to extract internal links. This is what I have.

    import pypdf

    internal_links = []
    with open(<pdf_path>, 'rb') as pdf_file:
        reader = pypdf.PdfReader(pdf_file)
        for page_num in range(len(reader.pages)):
            page = reader.pages[page_num]
            if "/Annots" in page:
                for annot in page["/Annots"]:
                    annot_obj = annot.get_object()
                    if annot_obj["/Subtype"] == "/Link":
                        if "/Dest" in annot_obj:
                            dest = annot_obj["/Dest"]
                            print(f"page: {page_num} dest: {dest}")

I'm unable to get the page number from dest. I looked in the pypdf codebase and also in the returned object, but can't find a way to correlate it with a page number.

[pubpub-zz]

I convert this thread into a discussion as this is not an issue

[pubpub-zz]

this should provide with the page number in your case:

pg_num = [ p.indirect_reference for p in reader.flattened_pages].index(dest)

be aware that all links are not using directly ["/Dest"] but some could also use ["/A"] with /Goto actions, and also some documents may store the destinations in the reader.root_object["/Names"]["/Dests"]

[swathiJayav]

We are on 5.0.1 (latest), and still not able to access reader.root_object["/Names"]["/Dests"]

[stefan6419846]

reader.root_object should always exist: https://pypdf.readthedocs.io/en/latest/modules/PdfReader.html#pypdf.PdfReader.root_object If it does not for your PDF file or there are errors, please provide it for further analysis and possibly a stacktrace. Otherwise there is not much to help you with.

[swathiJayav]

Actually, I am able to access it, but don't see anything with this key: reader.root_object["/Names"]. I'm guessing this only works if pg_num = [ p.indirect_reference for p in reader.flattened_pages].index(dest) didn't work.

Is the recommended way to find the linked page number to try all three of these approaches until one works? Does this cover all the cases we need to check for?

1. pg_num = [ p.indirect_reference for p in reader.flattened_pages].index(dest)
2. use ["/A"] with /Goto actions
3. check reader.root_object["/Names"]["/Dests"]

[stefan6419846]

The following should work:

dest = annot_obj["/Dest"]
pg_ref = dest[0]
pg_num = [p.indirect_reference for p in reader.flattened_pages].index(pg_ref)

For the corresponding syntax, see table 149 of the PDF 2.0 specification.

Full code (slightly modified):

import pypdf

internal_links = []
with pypdf.PdfReader('Boeing.pdf') as reader:
    for page_num, page in enumerate(reader.pages):
        page = reader.pages[page_num]
        if "/Annots" in page:
            for annot in page["/Annots"]:
                annot_obj = annot.get_object()
                if annot_obj["/Subtype"] == "/Link":
                    if "/Dest" in annot_obj:
                        dest = annot_obj["/Dest"][0]
                        pg_num = [ p.indirect_reference for p in reader.flattened_pages].index(dest)
                        print(f"page: {page_num} pg_num: {pg_num}")

And reader.root_object works fine here:

>>> from pypdf import PdfReader
>>> reader = PdfReader('Boeing.pdf')
>>> reader.root_object
{'/Type': '/Catalog', '/Pages': IndirectObject(3, 0, 134954891051504)}
>>> 

[swathiJayav]

I'm able to get it working when /Dest is available, thank you.

But for cases where it's a named link (specified with /GoTo), it's not clear how to correlate this with a page number. In the image below, annot_obj['/A']['/D'] does not match with anything in reader.root_object["/Names"]["/Dests"] . Attached is the document as well (taken from pypdf tests).

pdflatex-outline.pdf

[stefan6419846]

Retrieve the mapping of names to pages:

dests = reader.root_object["/Names"]["/Dests"]
kids = dests["/Kids"]
reader.pages[0]  # Populate `reader.flattened_pages`
_pages = [p.indirect_reference for p in reader.flattened_pages]
name_to_page_number = {}
for kid in kids:
    names = kid["/Names"]
    name_count = len(names)
    for i in range(0, name_count, 2):
        page_reference = names[i + 1]["/D"][0]
        name_to_page_number[names[i]] = _pages.index(page_reference)

Handle the annotation:

if "/A" in annot_obj:
    dest = annot_obj["/A"]["/D"]
    pg_num = name_to_page_number[dest]
    print(dest, pg_num)

[swathiJayav]

This works, thank you so much!

[swathiJayav]

Final answer:

For non-named links: #2911 (reply in thread)
`
For named links: #2911 (reply in thread)