vector<string> PDFModifiedPage::WriteModifiedResourcesDict(PDFParser* inParser,PDFDictionary* inResourcesDictionary,ObjectsContext& inObjectContext,PDFDocumentCopyingContext* inCopyingContext)
{
vector<string> formResourcesNames;
MapIterator<PDFNameToPDFObjectMap> resourcesDictionaryIt = inResourcesDictionary->GetIterator();
// create modified page object
DictionaryContext* dict = mWriter->GetObjectsContext().StartDictionary();
// copy all elements of the page to the new page object, but the "Contents" and "Resources" elements
while(resourcesDictionaryIt.MoveNext())
{
if(resourcesDictionaryIt.GetKey()->GetValue() != "XObject")
{
dict->WriteKey(resourcesDictionaryIt.GetKey()->GetValue());
inCopyingContext->CopyDirectObjectAsIs(resourcesDictionaryIt.GetValue());
}
}
// now write a new xobject entry.
dict->WriteKey("XObject");
DictionaryContext* xobjectDict = inObjectContext.StartDictionary();
PDFObjectCastPtr<PDFDictionary> existingXObjectDict(inParser->QueryDictionaryObject(inResourcesDictionary,"XObject"));
string imageObjectName;
if(existingXObjectDict.GetPtr())
{
// i'm having a very sophisticated algo here to create a new unique name.
// i'm making sure it's different in one letter from any name, using a well known discrete math proof method
MapIterator<PDFNameToPDFObjectMap> itExisting = existingXObjectDict->GetIterator();
unsigned long i=0;
while(itExisting.MoveNext())
{
string name = itExisting.GetKey()->GetValue();
xobjectDict->WriteKey(name);
inCopyingContext->CopyDirectObjectAsIs(itExisting.GetValue());
imageObjectName.push_back((char)(GetDifferentChar((name.length() >= i+1) ? name[i]:0x39)));
++i;
}
inObjectContext.EndLine();
}
PDFFormXObjectVector::iterator itForms = mContenxts.begin();
imageObjectName.push_back('_');
for(int i=0;itForms != mContenxts.end();++i,++itForms)
{
string formObjectName = imageObjectName + Int(i).ToString();
xobjectDict->WriteKey(formObjectName);
xobjectDict->WriteObjectReferenceValue((*itForms)->GetObjectID());
formResourcesNames.push_back(formObjectName);
}
inObjectContext.EndDictionary(xobjectDict);
inObjectContext.EndDictionary(dict);
return formResourcesNames;
}
void showXObjectsPerPageInfo(PDFParser& parser,PDFObjectCastPtr<PDFDictionary> xobjects)
{
RefCountPtr<PDFName> key;
PDFObjectCastPtr<PDFIndirectObjectReference> value;
MapIterator<PDFNameToPDFObjectMap> it = xobjects->GetIterator();
while(it.MoveNext())
{
key = it.GetKey();
value = it.GetValue();
cout << "XObject named " << key->GetValue().c_str() << " is object " << value->mObjectID << " of type ";
PDFObjectCastPtr<PDFStreamInput> xobject(parser.ParseNewObject(value->mObjectID));
PDFObjectCastPtr<PDFDictionary> xobjectDictionary(xobject->QueryStreamDictionary());
PDFObjectCastPtr<PDFName> typeOfXObject = xobjectDictionary->QueryDirectObject("Subtype");
cout << typeOfXObject->GetValue().c_str() << "\n";
}
}
Handle<Value> DocumentCopyingContextDriver::GetCopiedObjects(const Arguments& args)
{
HandleScope scope;
DocumentCopyingContextDriver* copyingContextDriver = ObjectWrap::Unwrap<DocumentCopyingContextDriver>(args.This());
if(!copyingContextDriver->CopyingContext)
{
ThrowException(Exception::TypeError(String::New("copying context object not initialized, create using pdfWriter.createPDFCopyingContext or PDFWriter.createPDFCopyingContextForModifiedFile")));
return scope.Close(Undefined());
}
// create an object that will serve as the map
Local<Object> result = Object::New();
MapIterator<ObjectIDTypeToObjectIDTypeMap> it = copyingContextDriver->CopyingContext->GetCopiedObjectsMappingIterator();
while(it.MoveNext())
result->Set(String::New(ObjectIDTypeObject(it.GetKey()).ToString().c_str()),Number::New(it.GetValue()));
return scope.Close(result);
}
EStatusCode DCTDecodeFilterTest::ModifyImageObject(PDFWriter* inWriter,ObjectIDType inImageObject)
{
EStatusCode status = eSuccess;
PDFDocumentCopyingContext* modifiedFileContext = inWriter->CreatePDFCopyingContextForModifiedFile();
do
{
// get image source dictionary
PDFObjectCastPtr<PDFStreamInput> imageStream(inWriter->GetModifiedFileParser().ParseNewObject(inImageObject));
RefCountPtr<PDFDictionary> imageDictionary(imageStream->QueryStreamDictionary());
// strt object for modified image
inWriter->GetObjectsContext().StartModifiedIndirectObject(inImageObject);
DictionaryContext* newImageDictionary = inWriter->GetObjectsContext().StartDictionary();
MapIterator<PDFNameToPDFObjectMap> it = imageDictionary->GetIterator();
// copy all but "Filter" and "Length"
ObjectIDTypeList indirectObjects;
while (it.MoveNext())
{
if(it.GetKey()->GetValue() == "Filter" || it.GetKey()->GetValue() == "Length")
continue;
newImageDictionary->WriteKey(it.GetKey()->GetValue());
EStatusCodeAndObjectIDTypeList result = modifiedFileContext->CopyDirectObjectWithDeepCopy(it.GetValue());
if(result.first != eSuccess)
{
status = result.first;
break;
}
indirectObjects.insert(indirectObjects.end(),result.second.begin(),result.second.end());
}
if(status != eSuccess)
break;
// start image stream for this dictionary (make sure it's unfiltered)
PDFStream* newImageStream = inWriter->GetObjectsContext().StartUnfilteredPDFStream(newImageDictionary);
// copy source stream through read filter
IByteReader* sourceImage = modifiedFileContext->GetSourceDocumentParser()->StartReadingFromStream(imageStream.GetPtr());
if(!sourceImage)
{
cout<<"failed to read DCT stream\n";
status = eFailure;
break;
}
OutputStreamTraits traits(newImageStream->GetWriteStream());
status = traits.CopyToOutputStream(sourceImage);
// finalize stream
inWriter->GetObjectsContext().EndPDFStream(newImageStream);
delete newImageStream;
// late check for status so stream is deleted
if(status != eSuccess)
break;
// copy remaining indirect objects from image stream dictionary
status = modifiedFileContext->CopyNewObjectsForDirectObject(indirectObjects);
}
while (false);
delete modifiedFileContext;
return status;
}
ObjectIDType DCTDecodeFilterTest::FindDCTDecodedImageObject(PDFParser* inParser)
{
ObjectIDType imageObject = 0;
do
{
// find image by looking for the first image in the first page
RefCountPtr<PDFDictionary> firstPage = inParser->ParsePage(0);
if(!firstPage)
break;
PDFObjectCastPtr<PDFDictionary> resourceDictionary(inParser->QueryDictionaryObject(firstPage.GetPtr(),"Resources"));
if(!resourceDictionary)
break;
PDFObjectCastPtr<PDFDictionary> xobjectDictionary(inParser->QueryDictionaryObject(resourceDictionary.GetPtr(), "XObject"));
if(!xobjectDictionary)
break;
MapIterator<PDFNameToPDFObjectMap> it = xobjectDictionary->GetIterator();
while(it.MoveNext())
{
if(it.GetValue()->GetType() == PDFObject::ePDFObjectIndirectObjectReference)
{
PDFObjectCastPtr<PDFStreamInput> image(
inParser->ParseNewObject(((PDFIndirectObjectReference*)it.GetValue())->mObjectID));
RefCountPtr<PDFDictionary> imageDictionary = image->QueryStreamDictionary();
PDFObjectCastPtr<PDFName> objectType = imageDictionary->QueryDirectObject("Subtype");
if(!objectType || objectType->GetValue() != "Image")
continue;
RefCountPtr<PDFObject> filters = imageDictionary->QueryDirectObject("Filter");
if(!filters)
break;
if(filters->GetType() == PDFObject::ePDFObjectName &&
((PDFName*)filters.GetPtr())->GetValue() == "DCTDecode")
{
imageObject = ((PDFIndirectObjectReference*)it.GetValue())->mObjectID;
break;
}
PDFArray* filtersArray = (PDFArray*)filters.GetPtr();
if(filtersArray->GetLength() == 1)
{
PDFObjectCastPtr<PDFName> firstDecoder(filtersArray->QueryObject(0));
if(firstDecoder->GetValue() == "DCTDecode")
{
imageObject = ((PDFIndirectObjectReference*)it.GetValue())->mObjectID;
break;
}
}
}
}
}
while (false);
return imageObject;
}
PDFHummus::EStatusCode PDFModifiedPage::WritePage()
{
EStatusCode status = EndContentContext(); // just in case someone forgot to close the latest content context
do {
if (status != eSuccess || !mIsDirty) {
break;
}
// allocate an object ID for the new contents stream (for placing the form)
// we first create the modified page object, so that we can define a name for the new form xobject
// that is unique
ObjectsContext& objectContext = mWriter->GetObjectsContext();
ObjectIDType newContentObjectID = objectContext.GetInDirectObjectsRegistry().AllocateNewObjectID();
ObjectIDType newEncapsulatingObjectID = 0;
// create a copying context, so we can copy the page dictionary, and modify its contents + resources dict
PDFDocumentCopyingContext* copyingContext = mWriter->CreatePDFCopyingContextForModifiedFile();
// get the page object
ObjectIDType pageObjectID = copyingContext->GetSourceDocumentParser()->GetPageObjectID(mPageIndex);
PDFObjectCastPtr<PDFDictionary> pageDictionaryObject = copyingContext->GetSourceDocumentParser()->ParsePage(mPageIndex);
MapIterator<PDFNameToPDFObjectMap> pageDictionaryObjectIt = pageDictionaryObject->GetIterator();
// create modified page object
objectContext.StartModifiedIndirectObject(pageObjectID);
DictionaryContext* modifiedPageObject = mWriter->GetObjectsContext().StartDictionary();
// copy all elements of the page to the new page object, but the "Contents", "Resources" and "Annots" elements
while (pageDictionaryObjectIt.MoveNext())
{
if (pageDictionaryObjectIt.GetKey()->GetValue() != "Resources" &&
pageDictionaryObjectIt.GetKey()->GetValue() != "Contents" &&
pageDictionaryObjectIt.GetKey()->GetValue() != "Annots")
{
modifiedPageObject->WriteKey(pageDictionaryObjectIt.GetKey()->GetValue());
copyingContext->CopyDirectObjectAsIs(pageDictionaryObjectIt.GetValue());
}
}
// Write new annotations entry, joining existing annotations, and new ones (from links attaching or what not)
if (pageDictionaryObject->Exists("Annots") || mWriter->GetDocumentContext().GetAnnotations().size() > 0)
{
modifiedPageObject->WriteKey("Annots");
objectContext.StartArray();
// write old annots, if any exist
if(pageDictionaryObject->Exists("Annots")) {
PDFObjectCastPtr<PDFArray> anArray(copyingContext->GetSourceDocumentParser()->QueryDictionaryObject(pageDictionaryObject.GetPtr(), "Annots"));
SingleValueContainerIterator<PDFObjectVector> refs = anArray->GetIterator();
while (refs.MoveNext())
copyingContext->CopyDirectObjectAsIs(refs.GetItem());
}
// write new annots from links
ObjectIDTypeSet& annotations = mWriter->GetDocumentContext().GetAnnotations();
if (annotations.size() > 0)
{
ObjectIDTypeSet::iterator it = annotations.begin();
for (; it != annotations.end(); ++it)
objectContext.WriteNewIndirectObjectReference(*it);
}
annotations.clear();
objectContext.EndArray(eTokenSeparatorEndLine);
}
// Write new contents entry, joining the existing contents with the new one. take care of various scenarios of the existing Contents
modifiedPageObject->WriteKey("Contents");
if (!pageDictionaryObject->Exists("Contents"))
{ // no contents
objectContext.WriteIndirectObjectReference(newContentObjectID);
}
else
{
objectContext.StartArray();
if (mEnsureContentEncapsulation)
{
newEncapsulatingObjectID = objectContext.GetInDirectObjectsRegistry().AllocateNewObjectID();
objectContext.WriteNewIndirectObjectReference(newEncapsulatingObjectID);
}
RefCountPtr<PDFObject> pageContent(copyingContext->GetSourceDocumentParser()->QueryDictionaryObject(pageDictionaryObject.GetPtr(), "Contents"));
if (pageContent->GetType() == PDFObject::ePDFObjectStream)
{
// single content stream. must be a refrence which points to it
PDFObjectCastPtr<PDFIndirectObjectReference> ref(pageDictionaryObject->QueryDirectObject("Contents"));
objectContext.WriteIndirectObjectReference(ref->mObjectID, ref->mVersion);
}
else if (pageContent->GetType() == PDFObject::ePDFObjectArray)
{
PDFArray* anArray = (PDFArray*)pageContent.GetPtr();
// multiple content streams
SingleValueContainerIterator<PDFObjectVector> refs = anArray->GetIterator();
PDFObjectCastPtr<PDFIndirectObjectReference> ref;
while (refs.MoveNext())
{
ref = refs.GetItem();
objectContext.WriteIndirectObjectReference(ref->mObjectID, ref->mVersion);
}
}
else {
// this basically means no content...or whatever. just ignore.
}
objectContext.WriteNewIndirectObjectReference(newContentObjectID);
objectContext.EndArray();
objectContext.EndLine();
}
// Write a new resource entry. copy all but the "XObject" entry, which needs to be modified. Just for kicks i'm keeping the original
// form (either direct dictionary, or indirect object)
ObjectIDType resourcesIndirect = 0;
ObjectIDType newResourcesIndirect = 0;
vector<string> formResourcesNames;
modifiedPageObject->WriteKey("Resources");
if (!pageDictionaryObject->Exists("Resources"))
{
// check if there's inherited dict. if so - write directly as a modified version
PDFObjectCastPtr<PDFDictionary> parentDict(
pageDictionaryObject->Exists("Parent") ?
copyingContext->GetSourceDocumentParser()->QueryDictionaryObject(pageDictionaryObject.GetPtr(), "Parent"):
NULL);
if(!parentDict) {
formResourcesNames = WriteNewResourcesDictionary(objectContext);
}
else {
PDFObjectCastPtr<PDFDictionary> inheritedResources = findInheritedResources(copyingContext->GetSourceDocumentParser(),parentDict.GetPtr());
if(!inheritedResources) {
formResourcesNames = WriteNewResourcesDictionary(objectContext);
} else {
formResourcesNames = WriteModifiedResourcesDict(copyingContext->GetSourceDocumentParser(), inheritedResources.GetPtr(), objectContext, copyingContext);
}
}
}
else
{
// resources may be direct, or indirect. if direct, write as is, adding the new form xobject, otherwise wait till page object ends and write then
PDFObjectCastPtr<PDFIndirectObjectReference> resourceDictRef(pageDictionaryObject->QueryDirectObject("Resources"));
if (!resourceDictRef)
{
PDFObjectCastPtr<PDFDictionary> resourceDict(pageDictionaryObject->QueryDirectObject("Resources"));
formResourcesNames = WriteModifiedResourcesDict(copyingContext->GetSourceDocumentParser(), resourceDict.GetPtr(), objectContext, copyingContext);
}
else
{
resourcesIndirect = resourceDictRef->mObjectID;
// later will write a modified version of the resources dictionary, with the new form.
// only modify the resources dict object if wasn't already modified (can happen when sharing resources dict between multiple pages).
// in the case where it was alrady modified, create a new resources dictionary that's a copy, and use it instead, to avoid overwriting
// the previous modification
GetObjectWriteInformationResult res = objectContext.GetInDirectObjectsRegistry().GetObjectWriteInformation(resourcesIndirect);
if (res.first && res.second.mIsDirty)
{
newResourcesIndirect = objectContext.GetInDirectObjectsRegistry().AllocateNewObjectID();
modifiedPageObject->WriteObjectReferenceValue(newResourcesIndirect);
}
else
modifiedPageObject->WriteObjectReferenceValue(resourcesIndirect);
}
}
objectContext.EndDictionary(modifiedPageObject);
objectContext.EndIndirectObject();
if (resourcesIndirect != 0)
{
if (newResourcesIndirect != 0)
objectContext.StartNewIndirectObject(newResourcesIndirect);
else
objectContext.StartModifiedIndirectObject(resourcesIndirect);
PDFObjectCastPtr<PDFDictionary> resourceDict(copyingContext->GetSourceDocumentParser()->ParseNewObject(resourcesIndirect));
formResourcesNames = WriteModifiedResourcesDict(copyingContext->GetSourceDocumentParser(), resourceDict.GetPtr(), objectContext, copyingContext);
objectContext.EndIndirectObject();
}
// if required write encapsulation code, so that new stream is independent of graphic context of original
PDFStream* newStream;
PrimitiveObjectsWriter primitivesWriter;
if (newEncapsulatingObjectID != 0)
{
objectContext.StartNewIndirectObject(newEncapsulatingObjectID);
newStream = objectContext.StartPDFStream();
primitivesWriter.SetStreamForWriting(newStream->GetWriteStream());
primitivesWriter.WriteKeyword("q");
objectContext.EndPDFStream(newStream);
}
// last but not least, create the actual content stream object, placing the form
objectContext.StartNewIndirectObject(newContentObjectID);
newStream = objectContext.StartPDFStream();
primitivesWriter.SetStreamForWriting(newStream->GetWriteStream());
if (newEncapsulatingObjectID != 0) {
primitivesWriter.WriteKeyword("Q");
}
vector<string>::iterator it = formResourcesNames.begin();
for (; it != formResourcesNames.end(); ++it)
{
primitivesWriter.WriteKeyword("q");
primitivesWriter.WriteInteger(1);
primitivesWriter.WriteInteger(0);
primitivesWriter.WriteInteger(0);
primitivesWriter.WriteInteger(1);
primitivesWriter.WriteInteger(0);
primitivesWriter.WriteInteger(0);
primitivesWriter.WriteKeyword("cm");
primitivesWriter.WriteName(*it);
primitivesWriter.WriteKeyword("Do");
primitivesWriter.WriteKeyword("Q");
}
objectContext.EndPDFStream(newStream);
} while (false);
return status;
}
METHOD_RETURN_TYPE DocumentCopyingContextDriver::GetCopiedObjects(const ARGS_TYPE& args)
{
CREATE_ISOLATE_CONTEXT;
CREATE_ESCAPABLE_SCOPE;
DocumentCopyingContextDriver* copyingContextDriver = ObjectWrap::Unwrap<DocumentCopyingContextDriver>(args.This());
if(!copyingContextDriver->CopyingContext)
{
THROW_EXCEPTION("copying context object not initialized, create using pdfWriter.createPDFCopyingContext or PDFWriter.createPDFCopyingContextForModifiedFile");
SET_FUNCTION_RETURN_VALUE(UNDEFINED);
}
// create an object that will serve as the map
Local<Object> result = NEW_OBJECT;
MapIterator<ObjectIDTypeToObjectIDTypeMap> it = copyingContextDriver->CopyingContext->GetCopiedObjectsMappingIterator();
while(it.MoveNext())
result->Set(NEW_STRING(ObjectIDTypeObject(it.GetKey()).ToString().c_str()),NEW_NUMBER(it.GetValue()));
SET_FUNCTION_RETURN_VALUE(result);
}
METHOD_RETURN_TYPE PDFDictionaryDriver::ToJSObject(const ARGS_TYPE& args)
{
CREATE_ISOLATE_CONTEXT;
CREATE_ESCAPABLE_SCOPE;
PDFDictionaryDriver* driver = ObjectWrap::Unwrap<PDFDictionaryDriver>(args.This());
Local<Object> result = NEW_OBJECT;
MapIterator<PDFNameToPDFObjectMap> it = driver->TheObject->GetIterator();
while(it.MoveNext())
result->Set(NEW_STRING(it.GetKey()->GetValue().c_str()),PDFObjectDriver::CreateDriver(it.GetValue()));
SET_FUNCTION_RETURN_VALUE(result);
}