Splitting a large Pdf file with PDFBox gets large result files

Question

I am processing some large pdf files, (up to 100MB and about 2000 pages), with pdfbox. Some of the pages contain a QR code, I want to split those files into smaller ones with the pages from one QR code to the next. I got this, but the result file sizes are the same as the source file. I mean, if I cut a 100MB pdf file into a ten files I am getting ten files 100MB each.

This is the code:

 PDDocument documentoPdf = 
        PDDocument.loadNonSeq(new File("myFile.pdf"), 
                           new RandomAccessFile(new File("./tmp/temp"), "rw"));

    int numPages = documentoPdf.getNumberOfPages();
    List pages = documentoPdf.getDocumentCatalog().getAllPages();

    int previusQR = 0;
    for(int i =0; i<numPages; i++){
       PDPage page = (PDPage) pages.get(i);
       BufferedImage firstPageImage =    
           page.convertToImage(BufferedImage.TYPE_USHORT_565_RGB , 200);

       String qrText = readQRWithQRCodeMultiReader(firstPageImage, hintMap);

       if(qrText != null and i!=0){
         PDDocument outputDocument = new PDDocument();
         for(int j = previusQR; j<i; j++){
           outputDocument.importPage((PDPage)pages.get(j));
          }
         File f = new File("./splitting_files/"+previusQR+".pdf");
         outputDocument.save(f);
         outputDocument.close();
         documentoPdf.close();
    }

I also tried the following code for storing the new file:

PDDocument outputDocument = new PDDocument();

for(int j = previusQR; j<i; j++){
 PDStream src = ((PDPage)pages.get(j)).getContents();
 PDStream streamD = new PDStream(outputDocument);
 streamD.addCompression();

 PDPage newPage = new PDPage(new   
           COSDictionary(((PDPage)pages.get(j)).getCOSDictionary()));
 newPage.setContents(streamD);

 byte[] buf = new byte[10240];
 int amountRead = 0;
 InputStream is = null;
 OutputStream os = null;
 is = src.createInputStream();
 os = streamD.createOutputStream();
 while((amountRead = is.read(buf,0,10240)) > -1) {
    os.write(buf, 0, amountRead);
  }

 outputDocument.addPage(newPage);
}

File f = new File("./splitting_files/"+previusQR+".pdf");

outputDocument.save(f);
outputDocument.close();

But this code creates files which lacks some content and also have the same size than the original.

How can I create smaller pdfs files from a larger one? Is it posible with PDFBox? Is there any other library with which I can transform a single page into an image (for qr recognition), and also allows me to split a big pdf file into smaller ones?

Thx!

Answer 1

Thx! Tilman you are right, the PDFSplit command generates smaller files. I checked the PDFSplit code out and found that it removes the page links to avoid not needed resources.

Code extracted from Splitter.class :

private void processAnnotations(PDPage imported) throws IOException
    {
        List<PDAnnotation> annotations = imported.getAnnotations();
        for (PDAnnotation annotation : annotations)
        {
            if (annotation instanceof PDAnnotationLink)
            {
                PDAnnotationLink link = (PDAnnotationLink)annotation;   
                PDDestination destination = link.getDestination();
                if (destination == null && link.getAction() != null)
                {
                    PDAction action = link.getAction();
                    if (action instanceof PDActionGoTo)
                    {
                        destination = ((PDActionGoTo)action).getDestination();
                    }
                }
                if (destination instanceof PDPageDestination)
                {
                    // TODO preserve links to pages within the splitted result  
                    ((PDPageDestination) destination).setPage(null);
                }
            }
            else
            {
                // TODO preserve links to pages within the splitted result  
                annotation.setPage(null);
            }
        }
    }

So eventually my code looks like this:

PDDocument documentoPdf = 
        PDDocument.loadNonSeq(new File("docs_compuestos/50.pdf"), new RandomAccessFile(new File("./tmp/t"), "rw"));

        int numPages = documentoPdf.getNumberOfPages();
        List pages = documentoPdf.getDocumentCatalog().getAllPages();


        int previusQR = 0;
        for(int i =0; i<numPages; i++){
            PDPage firstPage = (PDPage) pages.get(i);
            String qrText ="";


            BufferedImage firstPageImage = firstPage.convertToImage(BufferedImage.TYPE_USHORT_565_RGB , 200);


            firstPage =null;

            try {
                qrText = readQRWithQRCodeMultiReader(firstPageImage, hintMap);
            } catch (NotFoundException e) {
                e.printStackTrace();
            } finally {
                firstPageImage = null;
            }


        if(i != 0 && qrText!=null){
                    PDDocument outputDocument = new PDDocument();
                    outputDocument.setDocumentInformation(documentoPdf.getDocumentInformation());
                    outputDocument.getDocumentCatalog().setViewerPreferences(
                            documentoPdf.getDocumentCatalog().getViewerPreferences());


                    for(int j = previusQR; j<i; j++){
                        PDPage importedPage = outputDocument.importPage((PDPage)pages.get(j));

                        importedPage.setCropBox( ((PDPage)pages.get(j)).findCropBox() );
                        importedPage.setMediaBox( ((PDPage)pages.get(j)).findMediaBox() );
                        // only the resources of the page will be copied
                        importedPage.setResources( ((PDPage)pages.get(j)).getResources() );
                        importedPage.setRotation( ((PDPage)pages.get(j)).findRotation() );

                        processAnnotations(importedPage);


                    }


                    File f = new File("./splitting_files/"+previusQR+".pdf");

                    previusQR = i;

                    outputDocument.save(f);
                    outputDocument.close();
                }
            }


        }

Thank you very much!!