private void UpdateStructParentTreeEntries(int?pageStructParentIndex, IDictionary <int, PdfMcr> mcrs)
{
// element indexes in parentsOfPageMcrs shall be the same as mcid of one of their kids.
// See "Finding Structure Elements from Content Items" in pdf spec.
PdfArray parentsOfPageMcrs = new PdfArray();
int currentMcid = 0;
foreach (KeyValuePair <int, PdfMcr> entry in mcrs)
{
PdfMcr mcr = entry.Value;
if (mcr is PdfObjRef)
{
int structParent = KeyIntoStructParentIndex((int)entry.Key);
parentTree.AddEntry(structParent, ((PdfStructElem)mcr.GetParent()).GetPdfObject());
}
else
{
// if for some reason some mcr where not registered or don't exist, we ensure that the rest
// of the parent objects were placed at correct index
while (currentMcid++ < mcr.GetMcid())
{
parentsOfPageMcrs.Add(PdfNull.PDF_NULL);
}
parentsOfPageMcrs.Add(((PdfStructElem)mcr.GetParent()).GetPdfObject());
}
}
if (parentsOfPageMcrs.Size() > 0)
{
parentsOfPageMcrs.MakeIndirect(structTreeRoot.GetDocument());
parentTree.AddEntry(pageStructParentIndex, parentsOfPageMcrs);
structTreeRoot.GetDocument().CheckIsoConformance(parentsOfPageMcrs, IsoKey.TAG_STRUCTURE_ELEMENT);
parentsOfPageMcrs.Flush();
}
}
private bool UpdateStructParentTreeForContentStreamEntries(IDictionary <int, PdfMcr> mcrsOfContentStream, int
pageStructParentIndex)
{
// element indices in parentsOfMcrs shall be the same as mcid of one of their kids.
// See "Finding Structure Elements from Content Items" in pdf spec.
PdfArray parentsOfMcrs = new PdfArray();
int currentMcid = 0;
foreach (KeyValuePair <int, PdfMcr> entry in mcrsOfContentStream)
{
PdfMcr mcr = entry.Value;
PdfDictionary parentObj = ((PdfStructElem)mcr.GetParent()).GetPdfObject();
if (!parentObj.IsIndirect())
{
continue;
}
// if for some reason some mcrs were not registered or don't exist, we ensure that the rest
// of the parent objects were placed at correct index
while (currentMcid++ < mcr.GetMcid())
{
parentsOfMcrs.Add(PdfNull.PDF_NULL);
}
parentsOfMcrs.Add(parentObj);
}
if (!parentsOfMcrs.IsEmpty())
{
parentsOfMcrs.MakeIndirect(structTreeRoot.GetDocument());
parentTree.AddEntry(pageStructParentIndex, parentsOfMcrs);
structTreeRoot.GetDocument().CheckIsoConformance(parentsOfMcrs, IsoKey.TAG_STRUCTURE_ELEMENT);
parentsOfMcrs.Flush();
return(true);
}
return(false);
}
public virtual void RegisterMcr(PdfMcr mcr)
{
SortedDictionary <int, PdfMcr> pageMcrs = pageToPageMcrs.Get(mcr.GetPageObject().GetIndirectReference());
if (pageMcrs == null)
{
pageMcrs = new SortedDictionary <int, PdfMcr>();
pageToPageMcrs[mcr.GetPageObject().GetIndirectReference()] = pageMcrs;
}
if (mcr is PdfObjRef)
{
PdfDictionary obj = ((PdfDictionary)mcr.GetPdfObject()).GetAsDictionary(PdfName.Obj);
if (obj == null || obj.IsFlushed())
{
throw new PdfException(PdfException.WhenAddingObjectReferenceToTheTagTreeItMustBeConnectedToNotFlushedObject
);
}
PdfNumber n = obj.GetAsNumber(PdfName.StructParent);
if (n != null)
{
pageMcrs[StructParentIndexIntoKey(n.IntValue())] = mcr;
}
else
{
throw new PdfException(PdfException.StructParentIndexNotFoundInTaggedObject);
}
}
else
{
pageMcrs[mcr.GetMcid()] = mcr;
}
}
/// <returns>the topmost parent added to set. If encountered flushed element - stops and returns this flushed element.
/// </returns>
private static PdfDictionary AddAllParentsToSet(PdfMcr mcr, ICollection <PdfObject> set)
{
IList <PdfDictionary> allParents = RetrieveParents(mcr, true);
set.AddAll(allParents);
return(allParents.IsEmpty() ? null : allParents[allParents.Count - 1]);
}
private bool UpdateStructParentTreeEntries(PdfPage page, IDictionary <int, PdfMcr> mcrs)
{
bool res = false;
// element indexes in parentsOfPageMcrs shall be the same as mcid of one of their kids.
// See "Finding Structure Elements from Content Items" in pdf spec.
PdfArray parentsOfPageMcrs = new PdfArray();
int currentMcid = 0;
foreach (KeyValuePair <int, PdfMcr> entry in mcrs)
{
PdfMcr mcr = entry.Value;
PdfDictionary parentObj = ((PdfStructElem)mcr.GetParent()).GetPdfObject();
if (!parentObj.IsIndirect())
{
continue;
}
if (mcr is PdfObjRef)
{
int structParent = KeyIntoStructParentIndex((int)entry.Key);
parentTree.AddEntry(structParent, parentObj);
res = true;
}
else
{
// if for some reason some mcr where not registered or don't exist, we ensure that the rest
// of the parent objects were placed at correct index
while (currentMcid++ < mcr.GetMcid())
{
parentsOfPageMcrs.Add(PdfNull.PDF_NULL);
}
parentsOfPageMcrs.Add(parentObj);
}
}
if (!parentsOfPageMcrs.IsEmpty())
{
int pageStructParentIndex;
if (page.IsFlushed())
{
PdfIndirectReference pageRef = page.GetPdfObject().GetIndirectReference();
if (!pageToStructParentsInd.ContainsKey(pageRef))
{
return(res);
}
pageStructParentIndex = (int)pageToStructParentsInd.JRemove(pageRef);
}
else
{
pageStructParentIndex = GetOrCreatePageStructParentIndex(page);
}
parentsOfPageMcrs.MakeIndirect(structTreeRoot.GetDocument());
parentTree.AddEntry(pageStructParentIndex, parentsOfPageMcrs);
res = true;
structTreeRoot.GetDocument().CheckIsoConformance(parentsOfPageMcrs, IsoKey.TAG_STRUCTURE_ELEMENT);
parentsOfPageMcrs.Flush();
}
return(res);
}
internal virtual void PutXObjectMcr(PdfIndirectReference xObjectIndRef, PdfMcr mcr)
{
SortedDictionary <int, PdfMcr> xObjectMcrs = pageResourceXObjects.Get(xObjectIndRef);
if (xObjectMcrs == null)
{
xObjectMcrs = new SortedDictionary <int, PdfMcr>();
pageResourceXObjects.Put(xObjectIndRef, xObjectMcrs);
}
pageResourceXObjects.Get(xObjectIndRef).Put(mcr.GetMcid(), mcr);
}
private static PdfObject GetStm(PdfMcr mcr)
{
/*
* Presence of Stm guarantees that the mcr belongs to XObject, absence of Stm guarantees that the mcr belongs to page content stream.
* See 14.7.4.2 Marked-Content Sequences as Content Items, Table 324 – Entries in a marked-content reference dictionary.
*/
if (mcr is PdfMcrDictionary)
{
return(((PdfDictionary)mcr.GetPdfObject()).Get(PdfName.Stm, false));
}
return(null);
}
private bool UpdateStructParentTreeEntries(PdfPage page, ParentTreeHandler.PageMcrsContainer mcrs)
{
bool res = false;
foreach (KeyValuePair <int, PdfMcr> entry in mcrs.GetObjRefs())
{
PdfMcr mcr = entry.Value;
PdfDictionary parentObj = ((PdfStructElem)mcr.GetParent()).GetPdfObject();
if (!parentObj.IsIndirect())
{
continue;
}
int structParent = entry.Key;
parentTree.AddEntry(structParent, parentObj);
res = true;
}
int pageStructParentIndex;
foreach (KeyValuePair <PdfIndirectReference, SortedDictionary <int, PdfMcr> > entry in mcrs.GetPageResourceXObjects
())
{
PdfIndirectReference xObjectRef = entry.Key;
if (xObjectToStructParentsInd.ContainsKey(xObjectRef))
{
pageStructParentIndex = (int)xObjectToStructParentsInd.JRemove(xObjectRef);
if (UpdateStructParentTreeForContentStreamEntries(entry.Value, pageStructParentIndex))
{
res = true;
}
}
}
if (page.IsFlushed())
{
PdfIndirectReference pageRef = page.GetPdfObject().GetIndirectReference();
if (!pageToStructParentsInd.ContainsKey(pageRef))
{
return(res);
}
pageStructParentIndex = (int)pageToStructParentsInd.JRemove(pageRef);
}
else
{
pageStructParentIndex = GetOrCreatePageStructParentIndex(page);
}
if (UpdateStructParentTreeForContentStreamEntries(mcrs.GetPageContentStreamsMcrs(), pageStructParentIndex)
)
{
res = true;
}
return(res);
}
public virtual void UnregisterMcr(PdfMcr mcrToUnregister)
{
PdfDictionary pageDict = mcrToUnregister.GetPageObject();
if (pageDict == null)
{
// invalid mcr, ignore
return;
}
if (pageDict.IsFlushed())
{
throw new PdfException(PdfException.CannotRemoveMarkedContentReferenceBecauseItsPageWasAlreadyFlushed);
}
IDictionary <int, PdfMcr> pageMcrs = pageToPageMcrs.Get(pageDict.GetIndirectReference());
if (pageMcrs != null)
{
if (mcrToUnregister is PdfObjRef)
{
PdfDictionary obj = ((PdfDictionary)mcrToUnregister.GetPdfObject()).GetAsDictionary(PdfName.Obj);
if (obj != null && !obj.IsFlushed())
{
PdfNumber n = obj.GetAsNumber(PdfName.StructParent);
if (n != null)
{
pageMcrs.JRemove(StructParentIndexIntoKey(n.IntValue()));
structTreeRoot.SetModified();
return;
}
}
foreach (KeyValuePair <int, PdfMcr> entry in pageMcrs)
{
if (entry.Value.GetPdfObject() == mcrToUnregister.GetPdfObject())
{
pageMcrs.JRemove(entry.Key);
structTreeRoot.SetModified();
break;
}
}
}
else
{
pageMcrs.JRemove(mcrToUnregister.GetMcid());
structTreeRoot.SetModified();
}
}
}
public virtual int RemoveKid(IPdfStructElem kid)
{
if (kid is PdfMcr)
{
PdfMcr mcr = (PdfMcr)kid;
GetDocument().GetStructTreeRoot().GetParentTreeHandler().UnregisterMcr(mcr);
return(RemoveKidObject(mcr.GetPdfObject()));
}
else
{
if (kid is iText.Kernel.Pdf.Tagging.PdfStructElem)
{
return(RemoveKidObject(((iText.Kernel.Pdf.Tagging.PdfStructElem)kid).GetPdfObject()));
}
}
return(-1);
}
private static IList<PdfDictionary> RetrieveParents(PdfMcr mcr, bool all) {
IList<PdfDictionary> parents = new List<PdfDictionary>();
IStructureNode firstParent = mcr.GetParent();
PdfDictionary previous = null;
PdfDictionary current = firstParent is PdfStructElem ? ((PdfStructElem)firstParent).GetPdfObject() : null;
while (current != null && !PdfName.StructTreeRoot.Equals(current.GetAsName(PdfName.Type))) {
if (all) {
parents.Add(current);
}
previous = current;
current = previous.IsFlushed() ? null : previous.GetAsDictionary(PdfName.P);
}
if (!all) {
parents.Add(previous);
}
return parents;
}
private void RegisterMcr(PdfMcr mcr, bool registeringOnInit)
{
PdfDictionary mcrPageObject = mcr.GetPageObject();
if (mcrPageObject == null || (!(mcr is PdfObjRef) && mcr.GetMcid() < 0))
{
ILog logger = LogManager.GetLogger(typeof(iText.Kernel.Pdf.Tagging.ParentTreeHandler));
logger.Error(iText.IO.LogMessageConstant.ENCOUNTERED_INVALID_MCR);
return;
}
SortedDictionary <int, PdfMcr> pageMcrs = pageToPageMcrs.Get(mcrPageObject.GetIndirectReference());
if (pageMcrs == null)
{
pageMcrs = new SortedDictionary <int, PdfMcr>();
pageToPageMcrs.Put(mcrPageObject.GetIndirectReference(), pageMcrs);
}
if (mcr is PdfObjRef)
{
PdfDictionary obj = ((PdfDictionary)mcr.GetPdfObject()).GetAsDictionary(PdfName.Obj);
if (obj == null || obj.IsFlushed())
{
throw new PdfException(PdfException.WhenAddingObjectReferenceToTheTagTreeItMustBeConnectedToNotFlushedObject
);
}
PdfNumber n = obj.GetAsNumber(PdfName.StructParent);
if (n != null)
{
pageMcrs.Put(StructParentIndexIntoKey(n.IntValue()), mcr);
}
else
{
throw new PdfException(PdfException.StructParentIndexNotFoundInTaggedObject);
}
}
else
{
pageMcrs.Put(mcr.GetMcid(), mcr);
}
if (!registeringOnInit)
{
structTreeRoot.SetModified();
}
}
/// <returns>the topmost parent added to set. If encountered flushed element - stops and returns this flushed element.
/// </returns>
private static PdfDictionary AddAllParentsToSet(PdfMcr mcr, ICollection <PdfObject> set)
{
PdfDictionary elem = ((PdfStructElem)mcr.GetParent()).GetPdfObject();
set.Add(elem);
for (; ;)
{
if (elem.IsFlushed())
{
break;
}
PdfDictionary p = elem.GetAsDictionary(PdfName.P);
if (p == null || PdfName.StructTreeRoot.Equals(p.GetAsName(PdfName.Type)))
{
break;
}
else
{
elem = p;
set.Add(elem);
}
}
return(elem);
}
internal virtual void PutPageContentStreamMcr(int mcid, PdfMcr mcr)
{
pageContentStreams.Put(mcid, mcr);
}
public virtual PdfMcr AddKid(PdfMcr kid)
{
return(AddKid(-1, kid));
}
public virtual PdfMcr AddKid(int index, PdfMcr kid)
{
GetDocEnsureIndirectForKids().GetStructTreeRoot().GetParentTreeHandler().RegisterMcr(kid);
AddKidObject(GetPdfObject(), index, kid.GetPdfObject());
return(kid);
}
private void RegisterMcr(PdfMcr mcr, bool registeringOnInit)
{
PdfIndirectReference mcrPageIndRef = mcr.GetPageIndirectReference();
if (mcrPageIndRef == null || (!(mcr is PdfObjRef) && mcr.GetMcid() < 0))
{
ILog logger = LogManager.GetLogger(typeof(iText.Kernel.Pdf.Tagging.ParentTreeHandler));
logger.Error(iText.IO.LogMessageConstant.ENCOUNTERED_INVALID_MCR);
return;
}
ParentTreeHandler.PageMcrsContainer pageMcrs = pageToPageMcrs.Get(mcrPageIndRef);
if (pageMcrs == null)
{
pageMcrs = new ParentTreeHandler.PageMcrsContainer();
pageToPageMcrs.Put(mcrPageIndRef, pageMcrs);
}
PdfObject stm;
if ((stm = GetStm(mcr)) != null)
{
PdfIndirectReference stmIndRef;
PdfStream xObjectStream;
if (stm is PdfIndirectReference)
{
stmIndRef = (PdfIndirectReference)stm;
xObjectStream = (PdfStream)stmIndRef.GetRefersTo();
}
else
{
if (stm.GetIndirectReference() == null)
{
stm.MakeIndirect(structTreeRoot.GetDocument());
}
stmIndRef = stm.GetIndirectReference();
xObjectStream = (PdfStream)stm;
}
int?structParent = xObjectStream.GetAsInt(PdfName.StructParents);
if (structParent != null)
{
xObjectToStructParentsInd.Put(stmIndRef, structParent);
}
else
{
// TODO DEVSIX-3351 an error is thrown here because right now no /StructParents will be created.
ILog logger = LogManager.GetLogger(typeof(iText.Kernel.Pdf.Tagging.ParentTreeHandler));
logger.Error(iText.IO.LogMessageConstant.XOBJECT_HAS_NO_STRUCT_PARENTS);
}
pageMcrs.PutXObjectMcr(stmIndRef, mcr);
if (registeringOnInit)
{
xObjectStream.Release();
}
}
else
{
if (mcr is PdfObjRef)
{
PdfDictionary obj = ((PdfDictionary)mcr.GetPdfObject()).GetAsDictionary(PdfName.Obj);
if (obj == null || obj.IsFlushed())
{
throw new PdfException(PdfException.WhenAddingObjectReferenceToTheTagTreeItMustBeConnectedToNotFlushedObject
);
}
PdfNumber n = obj.GetAsNumber(PdfName.StructParent);
if (n != null)
{
pageMcrs.PutObjectReferenceMcr(n.IntValue(), mcr);
}
else
{
throw new PdfException(PdfException.StructParentIndexNotFoundInTaggedObject);
}
}
else
{
pageMcrs.PutPageContentStreamMcr(mcr.GetMcid(), mcr);
}
}
if (!registeringOnInit)
{
structTreeRoot.SetModified();
}
}
public virtual void UnregisterMcr(PdfMcr mcrToUnregister)
{
PdfDictionary pageDict = mcrToUnregister.GetPageObject();
if (pageDict == null)
{
// invalid mcr, ignore
return;
}
if (pageDict.IsFlushed())
{
throw new PdfException(PdfException.CannotRemoveMarkedContentReferenceBecauseItsPageWasAlreadyFlushed);
}
ParentTreeHandler.PageMcrsContainer pageMcrs = pageToPageMcrs.Get(pageDict.GetIndirectReference());
if (pageMcrs != null)
{
PdfObject stm;
if ((stm = GetStm(mcrToUnregister)) != null)
{
PdfIndirectReference xObjectReference = stm is PdfIndirectReference ? (PdfIndirectReference)stm : stm.GetIndirectReference
();
pageMcrs.GetPageResourceXObjects().Get(xObjectReference).JRemove(mcrToUnregister.GetMcid());
if (pageMcrs.GetPageResourceXObjects().Get(xObjectReference).IsEmpty())
{
pageMcrs.GetPageResourceXObjects().JRemove(xObjectReference);
xObjectToStructParentsInd.JRemove(xObjectReference);
}
structTreeRoot.SetModified();
}
else
{
if (mcrToUnregister is PdfObjRef)
{
PdfDictionary obj = ((PdfDictionary)mcrToUnregister.GetPdfObject()).GetAsDictionary(PdfName.Obj);
if (obj != null && !obj.IsFlushed())
{
PdfNumber n = obj.GetAsNumber(PdfName.StructParent);
if (n != null)
{
pageMcrs.GetObjRefs().JRemove(n.IntValue());
structTreeRoot.SetModified();
return;
}
}
foreach (KeyValuePair <int, PdfMcr> entry in pageMcrs.GetObjRefs())
{
if (entry.Value.GetPdfObject() == mcrToUnregister.GetPdfObject())
{
pageMcrs.GetObjRefs().JRemove(entry.Key);
structTreeRoot.SetModified();
break;
}
}
}
else
{
pageMcrs.GetPageContentStreamsMcrs().JRemove(mcrToUnregister.GetMcid());
structTreeRoot.SetModified();
}
}
}
}
public virtual void RegisterMcr(PdfMcr mcr)
{
RegisterMcr(mcr, false);
}
internal virtual void PutObjectReferenceMcr(int structParentIndex, PdfMcr mcr)
{
objRefs.Put(structParentIndex, mcr);
}
/// <summary>Gets the topmost non-root structure element parent.</summary>
/// <remarks>Gets the topmost non-root structure element parent. May be flushed.</remarks>
/// <param name="mcr">starting element</param>
/// <returns>
/// topmost non-root structure element parent, or
/// <see langword="null"/>
/// if it doesn't have any
/// </returns>
private static PdfDictionary GetTopmostParent(PdfMcr mcr)
{
return(RetrieveParents(mcr, false)[0]);
}