Pagedest (addlink) not working as intended with positional parameters .

[chukuwa-isobe]

If I create a link with positional, it jumps to an unintended page.

Environment

$ python -m platform
Windows-10-10.0

$ python -c "import PyPDF2;print(PyPDF2.__version__)"
2.10.7

Code + PDF

I want to use the "fit" option, so I write it positional.

from PyPDF2 import PdfWriter, PdfReader
from PyPDF2.generic import RectangleObject

in_file = r"D:\xxxxxxxxxxx\in.pdf"
out_file = r"D:\xxxxxxxxxxx\out.pdf"
writer = PdfWriter()
reader = PdfReader(in_file)

for current_page in reader.pages:
    writer.add_page(current_page)

writer.addLink(0, 1, [97, 758, 162, 768], [1, 1, 1], "/XYZ", 0, 0, 0)  # <-- here

with open(out_file, "wb") as o:
    writer.write(o)

Then when I click the link it jumps from page 1 to page "3".
When I open the link properties with Acrobat, the link destination is set to page "2", and after closing the properties with OK without updating anything, clicking the link again correctly jumps to page "2".
It's hard to open the properties of all created links and click OK, so please give me an idea how to solve it.

Traceback

This is the complete Traceback I see:

TODO

[pubpub-zz]

@chukuwa-isobe,
can you provde an example of pdf file (in and out) with the issue please

[chukuwa-isobe]

Dear pubpub-zz,

Upload in.pdf and out.pdf.
Two links are created on the first page in out.pdf.
If I click on the first link, it jump to page 3 (issue...), but if I click on the second link, it jump to page 2.
The pages specified should be the same.

from PyPDF2 import PdfFileWriter, PdfFileReader
from PyPDF2.generic import RectangleObject

inFile = r"xxxxxxxxxxxxxxxx\in.pdf"
outFile = r"xxxxxxxxxxxxxxxx\out.pdf"
pdf_writer = PdfFileWriter()
pdf_reader = PdfFileReader(open(inFile, "rb"))

num_of_pages = pdf_reader.getNumPages()

for page in range(num_of_pages):
    current_page = pdf_reader.getPage(page)
    pdf_writer.addPage(current_page)

pdf_writer.addLink(0, 1, [57, 721, 186, 731], [1, 1, 1], "/XYZ", 0, 0, 0)  # 1st BOX
pdf_writer.addLink(
    pagenum=0,
    pagedest=1,
    rect=RectangleObject([98, 708, 130, 718]),
    border=[1, 1, 1],
    fit="/XYZ",
)  # 2nd BOX

with open(outFile, "wb") as o:
    pdf_writer.write(o)

Is it not a good way to write?

in.pdf
out.pdf

[corsair20141]

Is there any progress on this?

When I try to run the "Add Link" example(s) from the documentation they don't work either...

[ssjkamei]

Is this a feature that has been removed from PdfWriter?
When I looked at the changelog, I saw that addLink was changed to add_link, but I couldn't find any mention of its deprecation.

[stefan6419846]

To follow PEP8, lots of names have been changed - in this case, addLink to add_link as well. Nevertheless, we additionally refactored the annotations handling as well to provide one generic interface for adding any type of annotation to the writer.

For the link-specific annotations, see https://pypdf.readthedocs.io/en/stable/user/adding-pdf-annotations.html#link

[ssjkamei]

@stefan6419846 Thank you very much. I will check it out.

[ssjkamei]

I have confirmed that it still occurs.

from pypdf import PdfWriter
from pypdf.generic import Fit
from pypdf.annotations import Link

out_file = r"out.pdf"
writer = PdfWriter()

for i in range(4):
    writer.add_blank_page(width=596, height=842)
    free_text_annotation = FreeText(
        text="Page" + str(i + 1),
        rect=(50, 770, 200, 800),
        bold=True,
    )
    writer.add_annotation(i, free_text_annotation)

annotation = Link(
    target_page_index=1,
    rect=[60, 730, 200, 750],
    border=[1, 1, 1],
    fit=Fit(fit_type="/XYZ"),
)  # 1st BOX to page 3(Issue)
writer.add_annotation(page_number=0, annotation=annotation)
annotation = Link(
    target_page_index=1,
    rect=RectangleObject([90, 700, 200, 720]),
    border=[1, 1, 1]
)  # 2nd BOX to page 2
writer.add_annotation(page_number=0, annotation=annotation)
annotation = Link(
    target_page_index=1,
    rect=[120, 670, 200, 690],
    border=[1, 1, 1],
    fit=Fit(fit_type="/XYZ", fit_args=(0, 0, 0))
)  # 3rd BOX to page 3(Issue)
writer.add_annotation(page_number=0, annotation=annotation)

with open(out_file, "wb") as o:
    writer.write(o)

[ssjkamei]

IndirectObject should be specified, but a page number was specified. It seems that page numbers are specified only for remote pages.

Wrong: /Dest [ 1 /XYZ 0.0 0.0 0.0 ]
Correct: /Dest [ 3 0 R /XYZ 0.0 0.0 0.0 ]

Below is an excerpt from the PDF 1.7 specification.

12.3.2.2 Explicit Destinations
Table 151 shows the allowed syntactic forms for specifying a destination explicitly in a PDF file. In each case, "page" is an indirect reference to a page object (except in a remote go-to action; see 12.6.4.3, "Remote Go-To Actions"). All coordinate values (left, right, top, and bottom) shall be expressed in the default user space coordinate system. The page's corp box, the corresponding side of the crop box shall be used instead; see 14.11.2, "Page Boundaries," for further discussion of the corp box.)
No page object can be specified for a destination associated with a remote go-to action (see 12.6.4.3 "Remote Go-To Actions") because the destination page is in a different PDF document. In this case, the page parameter specifies an integer page number within the remote document instead of a page object in document.

[ssjkamei]

It seems to be necessary to reference the IndirectObject of the page and set it when add_annotation is executed, but I think it would be difficult to make adjustments unless the reference to the IndirectObject is determined at write time, instead of being determined there.

I haven't read the code in much detail, so if I were to implement it, it would be quite a bit later. I am hoping someone will implement this.

[stefan6419846]

This might be related to #2443 and #2450. The PR is on hold, as it breaks external links.