iTextSharp - Using PDFAction.GotoLocalPage in Merged PDF

Question

I have written some code that merges together multiple PDF's into a single PDF that I then display from the MemoryStream. This works great. What I need to do is add a table of contents to the end of the file with links to the start of each of the individual PDF's. I planned on doing this using the GotoLocalPage action which has an option for page numbers but it doesn't seem to work. If I change the action to the code below to one of the presset ones like PDFAction.FIRSTPAGE it works fine. Does this not work because I am using the PDFCopy object for the writer parameter of GotoLocalPage?

Document mergedDoc = new Document();
MemoryStream ms = new MemoryStream();
PdfCopy copy = new PdfCopy(mergedDoc, ms);
mergedDoc.Open();

MemoryStream tocMS = new MemoryStream();
Document tocDoc = null;
PdfWriter tocWriter = null;

for (int i = 0; i < filesToMerge.Length; i++)
{
     string filename = filesToMerge[i];

     PdfReader reader = new PdfReader(filename);
     copy.AddDocument(reader);

     // Initialise TOC document based off first file
     if (i == 0)
     {
          tocDoc = new Document(reader.GetPageSizeWithRotation(1));
          tocWriter = PdfWriter.GetInstance(tocDoc, tocMS);
          tocDoc.Open();
     }

     // Create link for TOC, added random number of 3 for now
     Chunk link = new Chunk(filename);
     PdfAction action = PdfAction.GotoLocalPage(3, new PdfDestination(PdfDestination.FIT), copy);
     link.SetAction(action);
     tocDoc.Add(new Paragraph(link));
}

// Add TOC to end of merged PDF
tocDoc.Close();
PdfReader tocReader = new PdfReader(tocMS.ToArray());
copy.AddDocument(tocReader);

copy.Close();

displayPDF(ms.ToArray());

I guess an alternative would be to link to a named element (instead of page number) but I can't see how to add an 'invisible' element to the start of each file before adding to the merged document?

Answer 1

I would just go with two passes. In your first pass, do the merge as you are but also record the filename and page number it should link to. In your second pass, use a PdfStamper which will give you access to a ColumnText that you can use general abstractions like Paragraph in. Below is a sample that shows this off:

Since I don't have your documents, the below code creates 10 documents with a random number of pages each just for testing purposes. (You obviously don't need to do this part.) It also creates a simple dictionary with a fake file name as the key and the raw bytes from the PDF as a value. You have a true file collection to work with but you should be able to adapt that part.

//Create a bunch of files, nothing special here
//files will be a dictionary of names and the raw PDF bytes
Dictionary<string, byte[]> Files = new Dictionary<string, byte[]>();
var r = new Random();
for (var i = 1; i <= 10; i++) {
    using (var ms = new MemoryStream()) {
        using (var doc = new Document()) {
            using (var writer = PdfWriter.GetInstance(doc, ms)) {
                doc.Open();

                //Create a random number of pages
                for (var j = 1; j <= r.Next(1, 5); j++) {
                    doc.NewPage();
                    doc.Add(new Paragraph(String.Format("Hello from document {0} page {1}", i, j)));
                }
                doc.Close();
            }
        }
        Files.Add("File " + i.ToString(), ms.ToArray());
    }
}

This next block merges the PDFs. This is mostly the same as your code except that instead of writing a TOC here I'm just keeping track of what I want to write in the future. Where I'm using file.value you'd use your full file path and where I'm using file.key you'd use your file's name instead.

//Dictionary of file names (for display purposes) and their page numbers
var pages = new Dictionary<string, int>();

//PDFs start at page 1
var lastPageNumber = 1;

//Will hold the final merged PDF bytes
byte[] mergedBytes;

//Most everything else below is standard
using (var ms = new MemoryStream()) {
    using (var document = new Document()) {
        using (var writer = new PdfCopy(document, ms)) {
            document.Open();

            foreach (var file in Files) {

                //Add the current page at the previous page number
                pages.Add(file.Key, lastPageNumber);

                using (var reader = new PdfReader(file.Value)) {
                    writer.AddDocument(reader);

                    //Increment our current page index
                    lastPageNumber += reader.NumberOfPages;
                }
            }
        }
    }
    mergedBytes = ms.ToArray();
}

This last block actually writes the TOC. If we use a PdfStamper we can create a ColumnText which allows us to use Paragraphs

//Will hold the final PDF
byte[] finalBytes;
using (var ms = new MemoryStream()) {
    using (var reader = new PdfReader(mergedBytes)) {
        using (var stamper = new PdfStamper(reader, ms)) {

            //The page number to insert our TOC into
            var tocPageNum = reader.NumberOfPages + 1;

            //Arbitrarily pick one page to use as the size of the PDF
            //Additional logic could be added or this could just be set to something like PageSize.LETTER
            var tocPageSize = reader.GetPageSize(1);

            //Arbitrary margin for the page
            var tocMargin = 20;

            //Create our new page
            stamper.InsertPage(tocPageNum, tocPageSize);

            //Create a ColumnText object so that we can use abstractions like Paragraph
            var ct = new ColumnText(stamper.GetOverContent(tocPageNum));

            //Set the working area
            ct.SetSimpleColumn(tocPageSize.GetLeft(tocMargin), tocPageSize.GetBottom(tocMargin), tocPageSize.GetRight(tocMargin), tocPageSize.GetTop(tocMargin));

            //Loop through each page
            foreach (var page in pages) {
                var link = new Chunk(page.Key);
                var action = PdfAction.GotoLocalPage(page.Value, new PdfDestination(PdfDestination.FIT), stamper.Writer);
                link.SetAction(action);
                ct.AddElement(new Paragraph(link));
            }

            ct.Go();
        }
    }
    finalBytes = ms.ToArray();
}