Extract Reading Order Sequence in a Tagged PDF

Question

I'm currently validating the correct order of the content in a Tagged PDF File.

Is there any way to extract the reading order numbers of Tagged PDF Files programmatically?

I've tried converting the tagged PDF to XML but I can't figure out which tags belong to a certain text.

I've tried the following Libraries:

• Syncfusion
• IText7

but I can't find any methods that get its reading order numbers.

Is it really possible? Thanks in advance!

Answer 1

You can extract the marked content tree of tagged pdf using the PdfPig (.NET) library. My understanding is that the reading order is indicated by the Marked-content identifier (MCID).

If a marked content element does not contain an MCID (like pagination elements), the MCID is set to -1.

Each MarkedContentElement will contain the letters, images and paths that belong to it:

        using UglyToad.PdfPig;
        [...]

        using (PdfDocument document = PdfDocument.Open(pathToFile))
        {
            for (int p = 0; p < document.NumberOfPages; p++)
            {
                var page = document.GetPage(p + 1);

                // extract the page's marked content
                var markedContents = page.GetMarkedContents(); 

                var orderedMarkedContents = markedContents
                       .OrderBy(mc => mc.MarkedContentIdentifier);

                foreach (var mc in orderedMarkedContents)
                {
                    // do something
                }
            }
        }

If you want to extract the result to XML, you can have a look at the PageXmlTextExporter class. Have a look at the wiki for more information on ITextExporter and IReadingOrderDetector.

Note: I am an active contributer to this library.