Preserve Table of Contents

[rien333]

What a great little app!

However, after processing a pdf with krop, the table of contents "metadata" seems to be deleted. Is there any way to retain it? (makes sense for ereaders too, navigating to a specific part is especially cumbersome on slower and smaller devices if you can't just select a chapter)

[kupiqu]

I second this!

[arminstraub]

It would be great to be able to preserve tables of contents and links within the PDF (it's been on my personal wishlist for a long time). Unfortunately, I don't think this is possible using the PDF library that krop currently uses. If anyone has an idea how to possibly implement this, I would love to learn about it!

[rien333]

Unfortunately, I don't think this is possible using the PDF library that krop currently uses

That would be poppler, right? Also, I still really enjoy using krop. Simple, but well executed.

[arminstraub]

Thank you for the kind words! I'm glad you find krop useful despite this shortcoming. Poppler is used for displaying the PDF, but the cropping is done using PyPDF2.

[chrthi]

After looking a bit through PyPDF2, I was able to preserve links within a PDF with this change:

diff --git a/krop/mainwindow.py b/krop/mainwindow.py
index fd1ae32..e8adadf 100644
--- a/krop/mainwindow.py
+++ b/krop/mainwindow.py
@@ -413,6 +413,7 @@ class MainWindow(QKMainWindow):
             pdf = PdfFile()
             pdf.loadFromFile(inputFileName)
             cropper = PdfCropper()
+            cropper.copyDocumentRoot(pdf)
             for nr in pages:
                 c = self.viewer.cropValues(nr)
                 cropper.addPageCropped(pdf, nr, c, alwaysinclude, rotation)
diff --git a/krop/pdfcropper.py b/krop/pdfcropper.py
index 679c6fc..21a0df1 100644
--- a/krop/pdfcropper.py
+++ b/krop/pdfcropper.py
@@ -56,6 +56,9 @@ class AbstractPdfCropper:
     def addPageCropped(self, pdffile, pagenumber, croplist, rotate=0):
         pass

+    def copyDocumentRoot(self, pdffile):
+        pass
+

 class PyPdfFile(AbstractPdfFile):
     """Implementation of PdfFile using pyPdf"""
@@ -110,6 +113,15 @@ class PyPdfCropper(AbstractPdfCropper):
         if rotate != 0:
             page.rotateClockwise(rotate)

+    def copyDocumentRoot(self, pdffile):
+        # Sounds promising in PyPDF2 (see PdfFileWriter.cloneDocumentFromReader),
+        # but doesn't seem to produce a readable PDF:
+        # self.output.cloneReaderDocumentRoot(pdffile.reader)
+        # Instead, this copies at least the named destinations for links:
+        for dest in pdffile.reader.namedDestinations.values():
+            self.output.addNamedDestinationObject(dest)
+
+
 def optimizePdfGhostscript(oldfilename, newfilename):
     import subprocess
     subprocess.check_call(('gs', '-sDEVICE=pdfwrite', '-sOutputFile=' + newfilename,

It seems PyPDF2 has a special method to copy all such metadata at once named cloneReaderDocumentRoot, but that gave me a document with a lot of empty pages and only the links. So copying just the named destinations for links was the best I could come up with for now. If you would like to experiment further, I suggest a python debugger or using ptpython interactively in a script like this:

#!/usr/bin/env python3
from PyPDF2 import PdfFileReader
from ptpython.repl import embed

if __name__=="__main__":
    with open("test.pdf", "rb") as infile:
        pdf = PdfFileReader(infile)
        embed(globals(), locals())

This prepares a reader and then drops you to an interactive REPL with useful autocompletion.

[edumco]

@arminstraub the previous comment seems to solve this issue. Am I wrong?

[arminstraub]

@chrthi Thank you so much for offering this solution to preserving links! It didn't work on all the PDFs that I tested it with, but it was definitely better than nothing. It should be part of the next release of krop.

Once I have some time (hard these days...), I am planning to add support for pikepdf to krop which hopefully will make these sorts of things easier to work with.

[64kramsystem]

After looking a bit through PyPDF2, I was able to preserve links within a PDF with this change:

diff --git a/krop/mainwindow.py b/krop/mainwindow.py
...

Just tried this, but unfortunately doesn't work! The links are not preserved.