use of org.apache.pdfbox.cos.COSNumber in project pdfbox by apache.
the class PDFMergerUtility method appendDocument.
/**
* append all pages from source to destination.
*
* @param destination the document to receive the pages
* @param source the document originating the new pages
*
* @throws IOException If there is an error accessing data from either
* document.
*/
public void appendDocument(PDDocument destination, PDDocument source) throws IOException {
if (source.getDocument().isClosed()) {
throw new IOException("Error: source PDF is closed.");
}
if (destination.getDocument().isClosed()) {
throw new IOException("Error: destination PDF is closed.");
}
PDDocumentCatalog destCatalog = destination.getDocumentCatalog();
PDDocumentCatalog srcCatalog = source.getDocumentCatalog();
if (isDynamicXfa(srcCatalog.getAcroForm())) {
throw new IOException("Error: can't merge source document containing dynamic XFA form content.");
}
PDDocumentInformation destInfo = destination.getDocumentInformation();
PDDocumentInformation srcInfo = source.getDocumentInformation();
mergeInto(srcInfo.getCOSObject(), destInfo.getCOSObject(), Collections.<COSName>emptySet());
// use the highest version number for the resulting pdf
float destVersion = destination.getVersion();
float srcVersion = source.getVersion();
if (destVersion < srcVersion) {
destination.setVersion(srcVersion);
}
int pageIndexOpenActionDest = -1;
if (destCatalog.getOpenAction() == null) {
// PDFBOX-3972: get local dest page index, it must be reassigned after the page cloning
PDDestinationOrAction openAction = srcCatalog.getOpenAction();
PDDestination openActionDestination = null;
if (openAction instanceof PDActionGoTo) {
openActionDestination = ((PDActionGoTo) openAction).getDestination();
} else if (openAction instanceof PDDestination) {
openActionDestination = (PDDestination) openAction;
}
if (openActionDestination instanceof PDPageDestination) {
PDPage page = ((PDPageDestination) openActionDestination).getPage();
if (page != null) {
pageIndexOpenActionDest = srcCatalog.getPages().indexOf(page);
}
}
destCatalog.setOpenAction(srcCatalog.getOpenAction());
}
PDFCloneUtility cloner = new PDFCloneUtility(destination);
mergeAcroForm(cloner, destCatalog, srcCatalog);
COSArray destThreads = (COSArray) destCatalog.getCOSObject().getDictionaryObject(COSName.THREADS);
COSArray srcThreads = (COSArray) cloner.cloneForNewDocument(destCatalog.getCOSObject().getDictionaryObject(COSName.THREADS));
if (destThreads == null) {
destCatalog.getCOSObject().setItem(COSName.THREADS, srcThreads);
} else {
destThreads.addAll(srcThreads);
}
PDDocumentNameDictionary destNames = destCatalog.getNames();
PDDocumentNameDictionary srcNames = srcCatalog.getNames();
if (srcNames != null) {
if (destNames == null) {
destCatalog.getCOSObject().setItem(COSName.NAMES, cloner.cloneForNewDocument(srcNames));
} else {
cloner.cloneMerge(srcNames, destNames);
}
}
PDDocumentNameDestinationDictionary destDests = destCatalog.getDests();
PDDocumentNameDestinationDictionary srcDests = srcCatalog.getDests();
if (srcDests != null) {
if (destDests == null) {
destCatalog.getCOSObject().setItem(COSName.DESTS, cloner.cloneForNewDocument(srcDests));
} else {
cloner.cloneMerge(srcDests, destDests);
}
}
PDDocumentOutline destOutline = destCatalog.getDocumentOutline();
PDDocumentOutline srcOutline = srcCatalog.getDocumentOutline();
if (srcOutline != null) {
if (destOutline == null || destOutline.getFirstChild() == null) {
PDDocumentOutline cloned = new PDDocumentOutline((COSDictionary) cloner.cloneForNewDocument(srcOutline));
destCatalog.setDocumentOutline(cloned);
} else {
// search last sibling for dest, because /Last entry is sometimes wrong
PDOutlineItem destLastOutlineItem = destOutline.getFirstChild();
while (destLastOutlineItem.getNextSibling() != null) {
destLastOutlineItem = destLastOutlineItem.getNextSibling();
}
for (PDOutlineItem item : srcOutline.children()) {
// get each child, clone its dictionary, remove siblings info,
// append outline item created from there
COSDictionary clonedDict = (COSDictionary) cloner.cloneForNewDocument(item);
clonedDict.removeItem(COSName.PREV);
clonedDict.removeItem(COSName.NEXT);
PDOutlineItem clonedItem = new PDOutlineItem(clonedDict);
destLastOutlineItem.insertSiblingAfter(clonedItem);
destLastOutlineItem = destLastOutlineItem.getNextSibling();
}
}
}
PageMode destPageMode = destCatalog.getPageMode();
PageMode srcPageMode = srcCatalog.getPageMode();
if (destPageMode == null) {
destCatalog.setPageMode(srcPageMode);
}
COSDictionary destLabels = (COSDictionary) destCatalog.getCOSObject().getDictionaryObject(COSName.PAGE_LABELS);
COSDictionary srcLabels = (COSDictionary) srcCatalog.getCOSObject().getDictionaryObject(COSName.PAGE_LABELS);
if (srcLabels != null) {
int destPageCount = destination.getNumberOfPages();
COSArray destNums;
if (destLabels == null) {
destLabels = new COSDictionary();
destNums = new COSArray();
destLabels.setItem(COSName.NUMS, destNums);
destCatalog.getCOSObject().setItem(COSName.PAGE_LABELS, destLabels);
} else {
destNums = (COSArray) destLabels.getDictionaryObject(COSName.NUMS);
}
COSArray srcNums = (COSArray) srcLabels.getDictionaryObject(COSName.NUMS);
if (srcNums != null) {
for (int i = 0; i < srcNums.size(); i += 2) {
COSNumber labelIndex = (COSNumber) srcNums.getObject(i);
long labelIndexValue = labelIndex.intValue();
destNums.add(COSInteger.get(labelIndexValue + destPageCount));
destNums.add(cloner.cloneForNewDocument(srcNums.getObject(i + 1)));
}
}
}
COSStream destMetadata = (COSStream) destCatalog.getCOSObject().getDictionaryObject(COSName.METADATA);
COSStream srcMetadata = (COSStream) srcCatalog.getCOSObject().getDictionaryObject(COSName.METADATA);
if (destMetadata == null && srcMetadata != null) {
PDStream newStream = new PDStream(destination, srcMetadata.createInputStream(), (COSName) null);
mergeInto(srcMetadata, newStream.getCOSObject(), new HashSet<>(Arrays.asList(COSName.FILTER, COSName.LENGTH)));
destCatalog.getCOSObject().setItem(COSName.METADATA, newStream);
}
COSDictionary destOCP = (COSDictionary) destCatalog.getCOSObject().getDictionaryObject(COSName.OCPROPERTIES);
COSDictionary srcOCP = (COSDictionary) srcCatalog.getCOSObject().getDictionaryObject(COSName.OCPROPERTIES);
if (destOCP == null && srcOCP != null) {
destCatalog.getCOSObject().setItem(COSName.OCPROPERTIES, cloner.cloneForNewDocument(srcOCP));
}
mergeOutputIntents(cloner, srcCatalog, destCatalog);
// merge logical structure hierarchy if logical structure information is available in both source pdf and
// destination pdf
boolean mergeStructTree = false;
int destParentTreeNextKey = -1;
COSDictionary destParentTreeDict = null;
COSDictionary srcParentTreeDict;
COSArray destNumbersArray = null;
COSArray srcNumbersArray = null;
PDMarkInfo destMark = destCatalog.getMarkInfo();
PDStructureTreeRoot destStructTree = destCatalog.getStructureTreeRoot();
PDMarkInfo srcMark = srcCatalog.getMarkInfo();
PDStructureTreeRoot srcStructTree = srcCatalog.getStructureTreeRoot();
if (destStructTree != null) {
PDNumberTreeNode destParentTree = destStructTree.getParentTree();
destParentTreeNextKey = destStructTree.getParentTreeNextKey();
if (destParentTree != null) {
destParentTreeDict = destParentTree.getCOSObject();
destNumbersArray = (COSArray) destParentTreeDict.getDictionaryObject(COSName.NUMS);
if (destNumbersArray != null) {
if (destParentTreeNextKey < 0) {
destParentTreeNextKey = destNumbersArray.size() / 2;
}
if (destParentTreeNextKey > 0 && srcStructTree != null) {
PDNumberTreeNode srcParentTree = srcStructTree.getParentTree();
if (srcParentTree != null) {
srcParentTreeDict = srcParentTree.getCOSObject();
srcNumbersArray = (COSArray) srcParentTreeDict.getDictionaryObject(COSName.NUMS);
if (srcNumbersArray != null) {
mergeStructTree = true;
}
}
}
}
}
if (destMark != null && destMark.isMarked() && !mergeStructTree) {
destMark.setMarked(false);
}
if (!mergeStructTree) {
destCatalog.setStructureTreeRoot(null);
}
}
Map<COSDictionary, COSDictionary> objMapping = new HashMap<>();
int pageIndex = 0;
for (PDPage page : srcCatalog.getPages()) {
PDPage newPage = new PDPage((COSDictionary) cloner.cloneForNewDocument(page.getCOSObject()));
newPage.setCropBox(page.getCropBox());
newPage.setMediaBox(page.getMediaBox());
newPage.setRotation(page.getRotation());
PDResources resources = page.getResources();
if (resources != null) {
// this is smart enough to just create references for resources that are used on multiple pages
newPage.setResources(new PDResources((COSDictionary) cloner.cloneForNewDocument(resources)));
} else {
newPage.setResources(new PDResources());
}
if (mergeStructTree) {
updateStructParentEntries(newPage, destParentTreeNextKey);
objMapping.put(page.getCOSObject(), newPage.getCOSObject());
List<PDAnnotation> oldAnnots = page.getAnnotations();
List<PDAnnotation> newAnnots = newPage.getAnnotations();
for (int i = 0; i < oldAnnots.size(); i++) {
objMapping.put(oldAnnots.get(i).getCOSObject(), newAnnots.get(i).getCOSObject());
}
// TODO update mapping for XObjects
}
destination.addPage(newPage);
if (pageIndex == pageIndexOpenActionDest) {
// PDFBOX-3972: reassign the page.
// The openAction is either a PDActionGoTo or a PDPageDestination
PDDestinationOrAction openAction = destCatalog.getOpenAction();
PDPageDestination pageDestination;
if (openAction instanceof PDActionGoTo) {
pageDestination = (PDPageDestination) ((PDActionGoTo) openAction).getDestination();
} else {
pageDestination = (PDPageDestination) openAction;
}
pageDestination.setPage(newPage);
}
++pageIndex;
}
if (mergeStructTree) {
updatePageReferences(srcNumbersArray, objMapping);
for (int i = 0; i < srcNumbersArray.size() / 2; i++) {
destNumbersArray.add(COSInteger.get(destParentTreeNextKey + i));
destNumbersArray.add(srcNumbersArray.getObject(i * 2 + 1));
}
destParentTreeNextKey += srcNumbersArray.size() / 2;
destParentTreeDict.setItem(COSName.NUMS, destNumbersArray);
PDNumberTreeNode newParentTreeNode = new PDNumberTreeNode(destParentTreeDict, COSBase.class);
destStructTree.setParentTree(newParentTreeNode);
destStructTree.setParentTreeNextKey(destParentTreeNextKey);
COSDictionary kDictLevel0 = new COSDictionary();
COSArray newKArray = new COSArray();
COSArray destKArray = destStructTree.getKArray();
COSArray srcKArray = srcStructTree.getKArray();
if (destKArray != null && srcKArray != null) {
updateParentEntry(destKArray, kDictLevel0);
newKArray.addAll(destKArray);
updateParentEntry(srcKArray, kDictLevel0);
newKArray.addAll(srcKArray);
}
kDictLevel0.setItem(COSName.K, newKArray);
kDictLevel0.setItem(COSName.P, destStructTree);
kDictLevel0.setItem(COSName.S, COSName.DOCUMENT);
destStructTree.setK(kDictLevel0);
}
}
use of org.apache.pdfbox.cos.COSNumber in project pdfbox by apache.
the class COSParser method parseCOSStream.
/**
* This will read a COSStream from the input stream using length attribute within dictionary. If
* length attribute is a indirect reference it is first resolved to get the stream length. This
* means we copy stream data without testing for 'endstream' or 'endobj' and thus it is no
* problem if these keywords occur within stream. We require 'endstream' to be found after
* stream data is read.
*
* @param dic dictionary that goes with this stream.
*
* @return parsed pdf stream.
*
* @throws IOException if an error occurred reading the stream, like problems with reading
* length attribute, stream does not end with 'endstream' after data read, stream too short etc.
*/
protected COSStream parseCOSStream(COSDictionary dic) throws IOException {
COSStream stream = document.createCOSStream(dic);
// read 'stream'; this was already tested in parseObjectsDynamically()
readString();
skipWhiteSpaces();
/*
* This needs to be dic.getItem because when we are parsing, the underlying object might still be null.
*/
COSNumber streamLengthObj = getLength(dic.getItem(COSName.LENGTH), dic.getCOSName(COSName.TYPE));
if (streamLengthObj == null) {
if (isLenient) {
LOG.warn("The stream doesn't provide any stream length, using fallback readUntilEnd, at offset " + source.getPosition());
} else {
throw new IOException("Missing length for stream.");
}
}
// get output stream to copy data to
try (OutputStream out = stream.createRawOutputStream()) {
if (streamLengthObj != null && validateStreamLength(streamLengthObj.longValue())) {
readValidStream(out, streamLengthObj);
} else {
readUntilEndStream(new EndstreamOutputStream(out));
}
}
String endStream = readString();
if (endStream.equals("endobj") && isLenient) {
LOG.warn("stream ends with 'endobj' instead of 'endstream' at offset " + source.getPosition());
// avoid follow-up warning about missing endobj
source.rewind(ENDOBJ.length);
} else if (endStream.length() > 9 && isLenient && endStream.substring(0, 9).equals(ENDSTREAM_STRING)) {
LOG.warn("stream ends with '" + endStream + "' instead of 'endstream' at offset " + source.getPosition());
// unread the "extra" bytes
source.rewind(endStream.substring(9).getBytes(ISO_8859_1).length);
} else if (!endStream.equals(ENDSTREAM_STRING)) {
throw new IOException("Error reading stream, expected='endstream' actual='" + endStream + "' at offset " + source.getPosition());
}
return stream;
}
use of org.apache.pdfbox.cos.COSNumber in project pdfbox by apache.
the class COSWriter method prepareIncrement.
private void prepareIncrement(PDDocument doc) {
if (doc != null) {
COSDocument cosDoc = doc.getDocument();
Map<COSObjectKey, Long> xrefTable = cosDoc.getXrefTable();
Set<COSObjectKey> keySet = xrefTable.keySet();
long highestNumber = doc.getDocument().getHighestXRefObjectNumber();
for (COSObjectKey cosObjectKey : keySet) {
COSBase object = cosDoc.getObjectFromPool(cosObjectKey).getObject();
if (object != null && cosObjectKey != null && !(object instanceof COSNumber)) {
objectKeys.put(object, cosObjectKey);
keyObject.put(cosObjectKey, object);
}
if (cosObjectKey != null) {
long num = cosObjectKey.getNumber();
if (num > highestNumber) {
highestNumber = num;
}
}
}
setNumber(highestNumber);
}
}
use of org.apache.pdfbox.cos.COSNumber in project pdfbox by apache.
the class COSArrayList method convertFloatCOSArrayToList.
/**
* This will take an array of COSNumbers and return a COSArrayList of
* java.lang.Float values.
*
* @param floatArray The existing float Array.
*
* @return The list of Float objects.
*/
public static List<Float> convertFloatCOSArrayToList(COSArray floatArray) {
List<Float> retval = null;
if (floatArray != null) {
List<Float> numbers = new ArrayList<>(floatArray.size());
for (int i = 0; i < floatArray.size(); i++) {
COSBase base = floatArray.getObject(i);
if (base instanceof COSNumber) {
numbers.add(((COSNumber) base).floatValue());
} else {
numbers.add(null);
}
}
retval = new COSArrayList<>(numbers, floatArray);
}
return retval;
}
use of org.apache.pdfbox.cos.COSNumber in project pdfbox by apache.
the class FDFAnnotation method getPage.
/**
* This will get the page number or null if it does not exist.
*
* @return The page number.
*/
public Integer getPage() {
Integer retval = null;
COSNumber page = (COSNumber) annot.getDictionaryObject(COSName.PAGE);
if (page != null) {
retval = page.intValue();
}
return retval;
}
Aggregations