Major Description
There are three org.apache.pdfbox.filter.Filter implementations which are unimplemented. These are:
CCITTFaxDecodeFilter
DCTFilter
RunLengthDecodeFilter
They all contain calls to System.err with messages like
Warning: DCTFilter.decode is not implemented yet, skipping this stream.
In my code I iterate over all images in a PDF and try to obtain their raw, undecoded content. I use code like this:
private byte [] getUnDecodedImageBytes(COSStream st) throws IOException {
ByteArrayOutputStream baos = new ByteArrayOutputStream();
IOUtil.writeStream(st.getUnfilteredStream(), baos);
return baos.toByteArray();
}
The getUnfilteredStream() method, when called on JPG embedded images seems to try to invoke the DCTFilter. If I have a large ebook file with lots of JPG images - this yields LOTS of text to the Standard error output which can't be suppressed.
PDFBox uses commons-logging all over the place. Why not push those warnings to the log. They are non-critical. In my particular case when I use the above method I get an empty array. If I do, I resort to another method:
private byte [] getDecodedImageBytes(COSStream st) throws IOException {
ByteArrayOutputStream baos = new ByteArrayOutputStream();
PDXObjectImage ximage = (PDXObjectImage)PDXObject.createXObject( st );
ximage.write2OutputStream(baos);
return baos.toByteArray();
}
This seems to work, even for those images where getUnfilteredStream returns an empty stream.
I don't quite understand what's the difference, since I would expect a method labelled 'getUnfilteredStream' to return the stream as-it-is in the PDF file, without using any Filters. Moreover such a failure would imply that the library simply cannot process JPG images in PDF files, which is not the case because write2OutputStream works OK. So I don't know where the real problem lies. Maybe someone with more PDFBox knowledge could take a look.
Still, my patch only moves those warnings to the log, where I can suppress them. This is simple and fixes the immediate problem in my application.