We Keep Coding

PDFbox 1.7.0 - How to keep existing image while adding new one with PDFBox?

I'm using PDFBox 1.7.0 (I do not have a choice for the version due to old version in production server). I am trying to add an image to an existing PDF which has already a logo.
When I add the new image, the old one disappears like it is replaced.
// Use for convert mm to dots
// ... 72 dots per inch
static final int DEFAULT_USER_SPACE_UNIT_DPI = 72;
// ... mm -> inch -> dots
static final float MM_TO_UNITS = 1 / (10 * 2.54f) * DEFAULT_USER_SPACE_UNIT_DPI;
/**
* Add a given image to a specific page of a PDF
* #param document PDF document to manipulate
* #param input image inputStream
* #param pdfpage page number to target
* #param x image position (en mm)
* #param y image position (en mm)
* #param width max width of the image (mm)
* #param height max height of the image (en mm)
* #param opacity opacity level of the image (fraction)
*/
void addImageToPage (PDDocument document, InputStream input, int pdfpage, int x, int y, int width, int height, float opacity) throws IOException {
if (input != null) {
// Convert inputstream to usable BufferedImage
BufferedImage tmp_image = ImageIO.read (input);
// User TYPE_4BYTE_ABGR to fix PDFBox issue with transparent PNG
BufferedImage image = new BufferedImage (tmp_image.getWidth(), tmp_image.getHeight(), BufferedImage.TYPE_4BYTE_ABGR);
// Prepare the image
image.createGraphics().drawRenderedImage (tmp_image, null);
PDXObjectImage ximage = new PDPixelMap (document, image);
// Resize the image
int iWidth = ximage.getWidth();
int iHeight = ximage.getHeight();
if (width / height > iWidth / iHeight) {
ximage.setWidth (Math.round (width * MM_TO_UNITS));
ximage.setHeight (Math.round ((iHeight * width / iWidth) * MM_TO_UNITS));
} else {
ximage.setWidth (Math.round ((iWidth * height / iHeight) * MM_TO_UNITS));
ximage.setHeight (Math.round (height * MM_TO_UNITS));
}
// Retrieve the page to update
PDPage page = (PDPage)document.getDocumentCatalog().getAllPages().get (pdfpage);
PDResources resources = page.findResources();
// Get graphics states
Map graphicsStates = resources.getGraphicsStates();
if (graphicsStates == null) {
graphicsStates = new HashMap();
}
// Set graphics states configurations
PDExtendedGraphicsState extendedGraphicsState = new PDExtendedGraphicsState();
// Set the opacity of the image
extendedGraphicsState.setNonStrokingAlphaConstant (opacity);
graphicsStates.put ("TransparentState", extendedGraphicsState);
// Restore graphics states
resources.setGraphicsStates (graphicsStates);
// Retrieve the content stream
PDPageContentStream contentStream = new PDPageContentStream (document, page, true, true);
// Activate transparency options
contentStream.appendRawCommands ("/TransparentState gs\n");
contentStream.endMarkedContentSequence();
// Insert image
contentStream.drawImage (
ximage,
(float) x * MM_TO_UNITS,
(float) y * MM_TO_UNITS
);
// close the stream
contentStream.close();
}
}
I expected to have the new image within the page, but the existing image inside the page has disappeared instead of the new one.
Example of used PDF : http://www.mediafire.com/folder/g6p7c2b5ob1c7/PDFBox_issue

There are several bugs in 1.7... one I mentioned in a comment (turns out it doesn't affect you), the other one is that the resources does some caching but isn't managed properly… long story short, you need to save and restore your xobject resources like this:
Map<String, PDXObject> xObjectsMap = page.getResources().getXObjects(); // save xobjects
…
PDXObjectImage ximage = new PDPixelMap (document, image);
String imgName = page.getResources().addXObject(ximage, "Im");
cs.drawImage(ximage, 0, 0); // bug happens here, old xobjects gets lost
xObjectsMap.put(imgName, ximage);
page.getResources().setXObjects(xObjectsMap); // restore xobjects
This is really just a workaround… there may be more bad surprises coming. You should not use old versions. They no longer spark joy. You should thank them for their service and then let them go without guilt.

Ok. I have given up trying to use PDFbox 1.7 for this part of the development. It requirement to many fixes to implements few things. It is not really maintainable for the future works. Thanks to everyone for the hints and helps.

unable to put stamp using itext7 using Java lanaguage on only skia generated pdf (shows inverted stamp)

I am unable to put stamp using itext7 using Java language on only skia generated pdf (skia is pdf library used by google; if someone has worked on google docs-> Clicks on Print -> Save as Pdf ). It Stamps incorrectly; if I stamp at top left position of pdf page then it would stamp at bottom left and show (inverted mirror) image and (inverted mirror) text. For all other pdfs it gives correct stamping.
It seems pdf generated by skia has missing meta -data.

Since you didn't share any code, nor any document, I created a PDF document from Google docs, and I used the code I wrote in answer to the Itextsharp 7 - Scaled and Centered Image as watermark question to add a Watermark in the center.
The result looked like this:
As you can see in the Document Properties, the original document was created using Skia/PDF m67; modified using iText® 7.1.3.
You need a Watermark in the top left, so I adapted the code like this:
public void createPdf(String src, String dest) throws IOException {
PdfDocument pdfDoc = new PdfDocument(
new PdfReader(src), new PdfWriter(dest));
Document document = new Document(pdfDoc);
PdfCanvas over;
PdfExtGState gs1 = new PdfExtGState();
gs1.setFillOpacity(0.5f);
int n = pdfDoc.getNumberOfPages();
Rectangle pagesize;
ImageData img = ImageDataFactory.create(IMG);
float iW = img.getWidth();
float iH = img.getHeight();
float x, y;
for (int i = 1; i <= n; i++)
{
PdfPage pdfPage = pdfDoc.getPage(i);
pagesize = pdfPage.getPageSize();
x = pagesize.getLeft();
y = pagesize.getTop() - iH;
over = new PdfCanvas(pdfDoc.getPage(i));
over.saveState();
over.setExtGState(gs1);
over.addImage(img, iW, 0, 0, iH, x, y);
over.restoreState();
}
document.close();
pdfDoc.close();
}
The result looks like this:
The image isn't mirrored; it's at the top-left position of the page. In short: there doesn't seem to be any problem with PDF's created with Skia/PDF m67.

2D barcode generation issue in Java

I am generating barcodes using iText API, everything looks good when it is linear barcodes when it is 2D barcodes then barcodes are placed into pdf document as images, hence reducing the quality of the barcode on low resolution printers and unable to scan the barcode. Below is the code
BarcodePDF417 pdf417 = new BarcodePDF417();
String text = "BarcodePDF417 barcode";
pdf417.setText(text);
Image img = pdf417.getImage();
document.add(img);
Now i am look for an alternative to draw barcode and i found palceBarcode method which might favor to the requirement.
I have seen the below code in BarcodePDF417 class in itext source and could not able to find out the way how to use it
public void placeBarcode(PdfContentByte cb, BaseColor foreground, float moduleHeight, float moduleWidth) {
paintCode();
int stride = (bitColumns + 7) / 8;
cb.setColorFill(foreground);
for (int k = 0; k < codeRows; ++k) {
int p = k * stride;
for (int j = 0; j < bitColumns; ++j) {
int b = outBits[p + j / 8] & 0xff;
b <<= j % 8;
if ((b & 0x80) != 0) {
cb.rectangle(j * moduleWidth, (codeRows - k - 1) * moduleHeight, moduleWidth, moduleHeight);
}
}
}
cb.fill();
}
Can anyone suggest the way to use the above method?
I have written code like below but getting dark page as a whole.
Rectangle pageSize = new Rectangle(w * 72, h * 72);
Document doc = new Document(pageSize, 1f, 1f, 1f, 1f);
PdfWriter writer = PdfWriter.getInstance(doc, getOutputStream());
doc.open();
PdfContentByte cb = writer.getDirectContent();
BarcodePDF417 pf = new BarcodePDF417();
pf.setText("BarcodePDF417 barcode");
pf.getImage();
Rectangle rc = pf.getBarcodeSize();
pf.placeBarcode(cb, BaseColor.BLACK, rc.getHeight(), rc.getWidth());
doc.close();

Please take a look at the BarcodePlacement example. In this example, we create three PDF417 barcodes:
PdfContentByte cb = writer.getDirectContent();
Image img = createBarcode(cb, "This is a 2D barcode", 1, 1);
document.add(new Paragraph(
String.format("This barcode measures %s by %s user units",
img.getScaledWidth(), img.getScaledHeight())));
document.add(img);
img = createBarcode(cb, "This is NOT a raster image", 3, 3);
document.add(new Paragraph(
String.format("This barcode measures %s by %s user units",
img.getScaledWidth(), img.getScaledHeight())));
document.add(img);
img = createBarcode(cb, "This is vector data drawn on a PDF page", 1, 3);
document.add(new Paragraph(
String.format("This barcode measures %s by %s user units",
img.getScaledWidth(), img.getScaledHeight())));
The result looks like this on the outside:
One particular barcode looks like this on the inside:
I'm adding this inside view to show that the 2D barcode is not added as a raster image (as was the case with the initial approach you've tried). It is a vector image consisting of a series of small rectangles. You can check this for yourself by taking a look at the barcode_placement.pdf file.
Please don't be confused because I use an Image object. If you look at the createBarcode() method, you can see that the Image is, in fact, a vector image:
public Image createBarcode(PdfContentByte cb, String text,
float mh, float mw) throws BadElementException {
BarcodePDF417 pf = new BarcodePDF417();
pf.setText("BarcodePDF417 barcode");
Rectangle size = pf.getBarcodeSize();
PdfTemplate template = cb.createTemplate(
mw * size.getWidth(), mh * size.getHeight());
pf.placeBarcode(template, BaseColor.BLACK, mh, mw);
return Image.getInstance(template);
}
The height and the width passed to the placeBarcode() method, define the height and the width of the small rectangles that are drawn. If you look at the inside view, you can see for instance:
0 21 3 1 re
This is a rectangle with x = 0, y = 21, width 3 and height 1.
When you ask the barcode for its size, you get the number of rectangles that will be drawn. Hence the dimensions of the barcode is:
Rectangle size = pf.getBarcodeSize();
float width = mw * size.getWidth();
float height = mh * size.getHeight();
Your assumption that size is a size in user units is only correct if mw and mh are equal to 1.
I use these values to create a PdfTemplate instance and I draw the barcode to this Form XObject. Most of the times, it's easier to work with the Image class than working with PdfTemplate, so I wrap the PdfTemplate inside an Image.
I can then add this Image to the document just like any other image. The main difference with ordinary images, is that this image is a vector image.

How to add overlay text with link annotations to existing pdf?

I would like to add a link in my overlay text. I've read that using Anchor will only work for documents made from scratch but not for existing pdfs. My code is adding an overlay text to every page. My goal is to make a portion of that text clickable. I don't know how to make a link annotation that is part of a phrase.
Here's my code:
int n = reader.getNumberOfPages();
// step 4: we add content
PdfImportedPage page;
PdfCopy.PageStamp stamp;
for (int j = 0; j < n; )
{
++j;
page = writer.getImportedPage(reader, j);
if (i == 1) {
stamp = writer.createPageStamp(page);
Rectangle mediabox = reader.getPageSize(j);
Rectangle crop = new Rectangle(mediabox);
writer.setCropBoxSize(crop);
// add overlay text
Paragraph p = new Paragraph();
p.setAlignment(Element.ALIGN_CENTER);
FONT_URL_OVERLAY.setColor(0, 191, 255);
// get current user
EPerson loggedin = context.getCurrentUser();
String eperson = null;
if (loggedin != null)
{
eperson = loggedin.getFullName();
}
else eperson = "Anonymous";
Phrase downloaded = new Phrase();
Chunk site = new Chunk("My Website",FONT_URL_OVERLAY);
site.setAction(new PdfAction("http://www.mywebsite.com"));
downloaded.add(new Chunk("Downloaded by [" + eperson + "] from ", FONT_OVERLAY));
downloaded.add(site);
downloaded.add(new Chunk(" on ", FONT_OVERLAY));
downloaded.add(new Chunk(new SimpleDateFormat("MMMM d, yyyy").format(new Date()), FONT_OVERLAY));
downloaded.add(new Chunk(" at ", FONT_OVERLAY));
downloaded.add(new Chunk(new SimpleDateFormat("h:mm a z").format(new Date()), FONT_OVERLAY));
p.add(downloaded);
ColumnText.showTextAligned(stamp.getOverContent(), Element.ALIGN_CENTER, p,
crop.getLeft(10), crop.getHeight() / 2 + crop.getBottom(), 90);
stamp.alterContents();
}
writer.addPage(page);
}
So my overlay would looked like this:
Downloaded by [Anonymous] from My Website on February 17, 2015 at 1:20 AM CST
How can I convert My Website to a link annotation? Searching here in SO, I found this post, but I don't know how to apply adding link annotation to a portion of my overlay text.
Thanks in advance.
EDIT: How to add a rotated overlay text with link annotations to existing pdf?
Thanks to Bruno Lowagie for going out of his way in answering my question. Although I originally asked how to add link annotations in an overlay text to existing pdfs, he also catered my questions in the comments section of his answer about setting the coordinates properly if the overlay text were rotated.

You are using ColumnText.showAligned() which is sufficient to add a line of text without any special features, but if you want the anchor to work, you need to use ColumnText differently.
This is shown in the AddLinkAnnotation2 example:
public void manipulatePdf(String src, String dest) throws IOException, DocumentException {
PdfReader reader = new PdfReader(src);
PdfStamper stamper = new PdfStamper(reader, new FileOutputStream(dest));
PdfContentByte canvas = stamper.getOverContent(1);
Font bold = new Font(FontFamily.HELVETICA, 12, Font.BOLD);
Chunk chunk = new Chunk("The Best iText Questions on StackOverflow", bold);
chunk.setAnchor("http://pages.itextpdf.com/ebook-stackoverflow-questions.html");
Phrase p = new Phrase("Download ");
p.add(chunk);
p.add(" and discover more than 200 questions and answers.");
ColumnText ct = new ColumnText(canvas);
ct.setSimpleColumn(36, 700, 559, 750);
ct.addText(p);
ct.go();
stamper.close();
reader.close();
}
In this case, we define a rectangle for a ColumnText object, we add the Phrase to the column, and we go().
If you check the result, link_annotation2.pdf, you'll notice that you can click the words in bold.
There are no plans to support this in ColumnText.showTextAligned(). That is a convenience method that can be used as a short-cut for the handful of lines shown above, but there are some known limitations: lines are not wrapped, interactivity is ignored,...
Update 1: in the comment section, you asked an additional question about rotation the content and the link.
Rotating the content isn't difficult. There's even more than one way to do that. Rotating the link isn't trivial, as a link is a type of annotation, and annotations aren't part of the content.
Let's first take a look at AddLinkAnnotation3:
public void manipulatePdf(String src, String dest) throws IOException, DocumentException {
PdfReader reader = new PdfReader(src);
PdfStamper stamper = new PdfStamper(reader, new FileOutputStream(dest));
AffineTransform transform = AffineTransform.getRotateInstance(Math.PI / 6);
stamper.getWriter().setPageEvent(new AddAnnotation(stamper, transform));
PdfContentByte canvas = stamper.getOverContent(1);
Font bold = new Font(FontFamily.HELVETICA, 12, Font.BOLD);
Chunk chunk = new Chunk("The Best iText Questions on StackOverflow", bold);
chunk.setGenericTag("http://pages.itextpdf.com/ebook-stackoverflow-questions.html");
Phrase p = new Phrase("Download ");
p.add(chunk);
p.add(" and discover more than 200 questions and answers.");
canvas.saveState();
canvas.concatCTM(transform);
ColumnText ct = new ColumnText(canvas);
ct.setSimpleColumn(300, 0, 800, 400);
ct.addText(p);
ct.go();
canvas.restoreState();
stamper.close();
reader.close();
}
In this example, we define a tranformation of 30 degrees (Math.PI / 6):
AffineTransform transform = AffineTransform.getRotateInstance(Math.PI / 6);
We use this transformation when rendering the column:
canvas.saveState();
canvas.concatCTM(transform);
// render column
canvas.restoreState();
This rotates the content, but we didn't add any annotation yet. Instead, we define a page event:
stamper.getWriter().setPageEvent(new AddAnnotation(stamper, transform));
and we introduced a generic tag:
chunk.setGenericTag("http://pages.itextpdf.com/ebook-stackoverflow-questions.html");
To add the annotation, we use some magic in the page event implementation:
public class AddAnnotation extends PdfPageEventHelper {
protected PdfStamper stamper;
protected AffineTransform transform;
public AddAnnotation(PdfStamper stamper, AffineTransform transform) {
this.stamper = stamper;
this.transform = transform;
}
#Override
public void onGenericTag(PdfWriter writer, Document document, Rectangle rect, String text) {
float[] pts = {rect.getLeft(), rect.getBottom(), rect.getRight(), rect.getTop()};
transform.transform(pts, 0, pts, 0, 2);
float[] dstPts = {pts[0], pts[1], pts[2], pts[3]};
rect = new Rectangle(dstPts[0], dstPts[1], dstPts[2], dstPts[3]);
PdfAnnotation annot = PdfAnnotation.createLink(writer, rect, PdfAnnotation.HIGHLIGHT_INVERT, new PdfAction(text));
stamper.addAnnotation(annot, 1);
}
}
We create an annotation, but before we do so, we perform a transformation on the rectangle. This makes sure that the text fits the rectangle with the text that needs to be clickable, but... this may not be what you expect:
You may have wanted the rectangle to be rotated, and that's possible, but it's more math. For instance: you could create a polygon that is a better fit: ITextShape Clickable Polygon or path
Fortunately, you don't need an angle of 30 degrees, you want to rotate the text with an angle of 90 degrees. In that case, you don't have the strange effect shown in the above screen shot.
Take a look at AddLinkAnnotation4
public class AddAnnotation extends PdfPageEventHelper {
protected PdfStamper stamper;
protected AffineTransform transform;
public AddAnnotation(PdfStamper stamper, AffineTransform transform) {
this.stamper = stamper;
this.transform = transform;
}
#Override
public void onGenericTag(PdfWriter writer, Document document, Rectangle rect, String text) {
float[] pts = {rect.getLeft(), rect.getBottom(), rect.getRight(), rect.getTop()};
transform.transform(pts, 0, pts, 0, 2);
float[] dstPts = {pts[0], pts[1], pts[2], pts[3]};
rect = new Rectangle(dstPts[0], dstPts[1], dstPts[2], dstPts[3]);
PdfAnnotation annot = PdfAnnotation.createLink(writer, rect, PdfAnnotation.HIGHLIGHT_INVERT, new PdfAction(text));
annot.setBorder(new PdfBorderArray(0, 0, 0));
stamper.addAnnotation(annot, 1);
}
}
As you can see, I've added a single line to remove the border (the border is there by default unless you redefine the PdfBorderArray).
The rest of the code is also almost identical. We now define an angle of Math.PI / 2 (90 degrees).
public void manipulatePdf(String src, String dest) throws IOException, DocumentException {
PdfReader reader = new PdfReader(src);
PdfStamper stamper = new PdfStamper(reader, new FileOutputStream(dest));
AffineTransform transform = AffineTransform.getRotateInstance(Math.PI / 2);
stamper.getWriter().setPageEvent(new AddAnnotation(stamper, transform));
PdfContentByte canvas = stamper.getOverContent(1);
Font bold = new Font(FontFamily.HELVETICA, 12, Font.BOLD);
Chunk chunk = new Chunk("The Best iText Questions on StackOverflow", bold);
chunk.setGenericTag("http://pages.itextpdf.com/ebook-stackoverflow-questions.html");
Phrase p = new Phrase("Download ");
p.add(chunk);
p.add(" and discover more than 200 questions and answers.");
canvas.saveState();
canvas.concatCTM(transform);
ColumnText ct = new ColumnText(canvas);
ct.setSimpleColumn(36, -559, 806, -36);
ct.addText(p);
ct.go();
canvas.restoreState();
stamper.close();
reader.close();
}
Note that the lower left corner of the page is the pivot point, hence we need to adapt the coordinates where we add the column, otherwise you'll rotate all the content outside the visible area of the page.
Update 2:
In yet another comment, you are asking about the coordinates you need to use when adding text in a rotated coordinate system.
I made this drawing:
In the top part, you add the word MIDDLE in the middle of a page, but that's not where it will appear: you are rotating everything by 90 degrees, hence the word MIDDLE will rotate outside your page (into the hatched area). The word will be in the PDF, but you'll never see it.
If you look at my code, you see that I use these coordinates:
ct.setSimpleColumn(36, -559, 806, -36);
This is outside the visible area (it's below the actual page dimensions), but as I rotate everything with 90 degrees, it rotates into the visible area.
If you look at my drawing, you can see that the page with coordinates (0, 0), (0, -595), (842, -598) and (842, 0) rotates by 90 degrees and thus gets the coincides with a page with coordinates (0, 0), (595, 0), (595, 842) and (0, 842). That's the type of Math we all learned in high school ;-)
You were adding text at position crop.getLeft(10), crop.getHeight() / 2 + crop.getBottom(). If you know that the text will be rotated by 90 degrees, you should use crop.getHeight() / 2 + crop.getBottom(), -crop.getLeft().
The best way to understand why, is to make a drawing.

iText PDFDocument page size inaccurate

I am trying to add a header to existing pdf documents in Java with iText. I can add the header at a fixed place on the document, but all the documents are different page sizes, so it is not always at the top of the page. I have tried getting the page size so that I could calculate the position of the header, but it seems as if the page size is not actually what I want. On some documents, calling reader.getPageSize(i).getTop(20) will place the text in the right place at the top of the page, however, on some different documents it will place it half way down the page. Most of the pages have been scanned be a Xerox copier, if that makes a difference. Here is the code I am using:
PdfReader reader = new PdfReader(readFilePath);
PdfStamper stamper = new PdfStamper(reader, new FileOutputStream(writeFilePath));
BaseFont bf = BaseFont.createFont(BaseFont.HELVETICA, BaseFont.CP1252, BaseFont.NOT_EMBEDDED);
for (int i = 1; i <= reader.getNumberOfPages(); i++) {
PdfContentByte cb = stamper.getOverContent(i);
cb.beginText();
cb.setFontAndSize(bf, 14);
float x = reader.getPageSize(i).getWidth() / 2;
float y = reader.getPageSize(i).getTop(20);
cb.showTextAligned(PdfContentByte.ALIGN_CENTER, "Copy", x, y, 0);
cb.endText();
}
stamper.close();
PDF that works correctly
PDF that works incorrectly

Take a look at the StampHeader1 example. I adapted your code, introducing ColumnText.showTextAligned() and using a Phrase for the sake of simplicity (maybe you can change that part of your code too):
public void manipulatePdf(String src, String dest) throws IOException, DocumentException {
PdfReader reader = new PdfReader(src);
PdfStamper stamper = new PdfStamper(reader, new FileOutputStream(dest));
Phrase header = new Phrase("Copy", new Font(FontFamily.HELVETICA, 14));
for (int i = 1; i <= reader.getNumberOfPages(); i++) {
float x = reader.getPageSize(i).getWidth() / 2;
float y = reader.getPageSize(i).getTop(20);
ColumnText.showTextAligned(
stamper.getOverContent(i), Element.ALIGN_CENTER,
header, x, y, 0);
}
stamper.close();
reader.close();
}
As you have found out, this code assumes that no rotation was defined.
Now take a look at the StampHeader2 example. I'm using your "Wrong" file and I've added one extra line:
stamper.setRotateContents(false);
By telling the stamper not to rotate the content I'm adding, I'm adding the content using the coordinates as if the page isn't rotated. Please take a look at the result: stamped_header2.pdf. We added "Copy" at the top of the page, but as the page is rotated, we see the word appear on the side. The word is rotated because the page is rotated.
Maybe that's what you want, maybe it isn't. If it isn't, please take a look at StampHeader3 in which I calculate x and y differently, based on the rotation of the page:
if (reader.getPageRotation(i) % 180 == 0) {
x = reader.getPageSize(i).getWidth() / 2;
y = reader.getPageSize(i).getTop(20);
}
else {
x = reader.getPageSize(i).getHeight() / 2;
y = reader.getPageSize(i).getRight(20);
}
Now the word "Copy" appears on what is perceived as the "top of the page" (but in reality, it could be the side of the page): stamped_header3.pdf