Improved copying of tagged pages

Improved separation process of the existing tag structure into two parts in order to insert copied page tags between them: now this process doesn't rebuilt the whole tag tree

Author: yulian-gaponenko

kernel/src/main/java/com/itextpdf/kernel/PdfException.java

@@ -23,6 +23,7 @@ public class PdfException extends RuntimeException {
     public static final String CannotCloseDocumentWithAlreadyFlushedPdfCatalog = "cannot.close.document.with.already.flushed.pdf.catalog";
     public static final String CannotConvertPdfArrayToRectanle = "cannot.convert.pdfarray.to.rectangle";
     public static final String CannotCopyFlushedObject = "cannot.copy.flushed.object";
+    public static final String CannotCopyFlushedTag = "cannot.copy.flushed.tag";
     public static final String CannotCopyObjectContent = "cannot.copy.object.content";
     public static final String CannotCopyIndirectObjectFromTheDocumentThatIsBeingWritten = "cannot.copy.indirect.object.from.the.document.that.is.being.written";
     public static final String CannotCopyToDocumentOpenedInReadingMode = "cannot.copy.to.document.opened.in.reading.mode";
@@ -158,6 +159,7 @@ public class PdfException extends RuntimeException {
     public static final String StdcfNotFoundEncryption = "stdcf.not.found.encryption";
     public static final String StructureElementShallContainParentObject = "structure.element.shall.contain.parent.object";
     public static final String TagCannotBeMovedToTheAnotherDocumentsTagStructure = "tag.cannot.be.moved.to.the.another.documents.tag.structure";
+    public static final String TagFromTheExistingTagStructureIsFlushedCannotAddCopiedPageTags = "tag.from.the.existing.tag.structure.is.flushed.cannot.add.copied.page.tags";
     public static final String TagTreePointerIsInInvalidStateItPointsAtFlushedElementUseMoveToRoot = "tagtreepointer.is.in.invalid.state.it.points.at.flushed.element.use.movetoroot";
     public static final String TagTreePointerIsInInvalidStateItPointsAtRemovedElementUseMoveToRoot = "tagtreepointer.is.in.invalid.state.it.points.at.removed.element.use.movetoroot";
     public static final String TextCannotBeNull = "text.cannot.be.null";

kernel/src/main/java/com/itextpdf/kernel/pdf/PdfDocument.java

@@ -906,17 +906,19 @@ public List<PdfPage> copyPagesTo(Set<Integer> pagesToCopy, PdfDocument toDocumen
         Map<PdfPage, PdfPage> page2page = new LinkedHashMap<>();
         Map<PdfPage, List<PdfOutline>> page2Outlines = new HashMap<>();
         Set<PdfOutline> outlinesToCopy = new HashSet<>();
+        int insertBefore = insertBeforePage;
+        int numberOfPagesBeforeCopying = toDocument.getNumberOfPages();
         for (Integer pageNum : pagesToCopy) {
             PdfPage page = getPage(pageNum);
             PdfPage newPage = page.copyTo(toDocument, copier);
             copiedPages.add(newPage);
             page2page.put(page, newPage);
-            if (insertBeforePage < toDocument.getNumberOfPages() + 1) {
-                toDocument.addPage(insertBeforePage, newPage);
+            if (insertBefore < toDocument.getNumberOfPages() + 1) {
+                toDocument.addPage(insertBefore, newPage);
             } else {
                 toDocument.addPage(newPage);
             }
-            insertBeforePage++;
+            insertBefore++;
             if (catalog.isOutlineMode()) {
                 List<PdfOutline> pageOutlines = page.getOutlines(false);
                 if (pageOutlines != null)
@@ -933,7 +935,7 @@ public List<PdfPage> copyPagesTo(Set<Integer> pagesToCopy, PdfDocument toDocumen
             if (tagStructureContext != null) {
                 tagStructureContext.removeAllConnectionsToTags();
             }
-            if (insertBeforePage > toDocument.getNumberOfPages()) {
+            if (insertBeforePage > numberOfPagesBeforeCopying) {
                 getStructTreeRoot().copyTo(toDocument, page2page);
             } else {
                 getStructTreeRoot().copyTo(toDocument, insertBeforePage, page2page);

kernel/src/main/java/com/itextpdf/kernel/pdf/tagging/PdfStructElem.java

@@ -207,7 +207,7 @@ public PdfStructElem addKid(int index, PdfStructElem kid) {
         if (type == InlineLevel || type == Illustration) {
             throw new PdfException(PdfException.InlineLevelOrIllustrationElementCannotContainKids, getPdfObject());
         }
-        addKidObject(index, kid.getPdfObject());
+        addKidObject(getPdfObject(), index, kid.getPdfObject());
         return kid;
     }
 
@@ -217,7 +217,7 @@ public PdfMcr addKid(PdfMcr kid) {
 
     public PdfMcr addKid(int index, PdfMcr kid) {
         getDocument().getStructTreeRoot().getMcrManager().registerMcr(kid);
-        addKidObject(index, kid.getPdfObject());
+        addKidObject(getPdfObject(), index, kid.getPdfObject());
         return kid;
     }
 
@@ -307,51 +307,50 @@ public void flush() {
         super.flush();
     }
 
-    @Override
-    protected boolean isWrappedObjectMustBeIndirect() {
-        return true;
-    }
-
-    protected PdfDocument getDocument() {
-        return getPdfObject().getIndirectReference().getDocument();
-    }
-
-    private void addKidObjectToStructElemList(PdfObject k, List<IPdfStructElem> list) {
-        if (k.isFlushed()) {
-            list.add(null);
-            return;
-        }
-
-        list.add(convertPdfObjectToIPdfStructElem(k));
-    }
-
-    private void addKidObject(int index, PdfObject kid) {
-        PdfDictionary pdfObject = getPdfObject();
-        if (pdfObject.isFlushed()) {
+    static void addKidObject(PdfDictionary parent, int index, PdfObject kid) {
+        if (parent.isFlushed()) {
             throw new PdfException(PdfException.CannotAddKidToTheFlushedElement);
         }
-        if (!pdfObject.containsKey(PdfName.P)) {
-            throw new PdfException(PdfException.StructureElementShallContainParentObject, pdfObject);
+        if (!parent.containsKey(PdfName.P)) {
+            throw new PdfException(PdfException.StructureElementShallContainParentObject, parent);
         }
-        PdfObject k = getK();
+        PdfObject k = parent.get(PdfName.K);
         if (k == null)
-            pdfObject.put(PdfName.K, kid);
+            parent.put(PdfName.K, kid);
         else {
             PdfArray a;
             if (k instanceof PdfArray) {
                 a = (PdfArray) k;
             } else {
                 a = new PdfArray();
                 a.add(k);
-                pdfObject.put(PdfName.K, a);
+                parent.put(PdfName.K, a);
             }
             if (index == -1)
                 a.add(kid);
             else
                 a.add(index, kid);
         }
         if (kid instanceof PdfDictionary && isStructElem((PdfDictionary) kid))
-            ((PdfDictionary) kid).put(PdfName.P, pdfObject);
+            ((PdfDictionary) kid).put(PdfName.P, parent);
+    }
+
+    @Override
+    protected boolean isWrappedObjectMustBeIndirect() {
+        return true;
+    }
+
+    protected PdfDocument getDocument() {
+        return getPdfObject().getIndirectReference().getDocument();
+    }
+
+    private void addKidObjectToStructElemList(PdfObject k, List<IPdfStructElem> list) {
+        if (k.isFlushed()) {
+            list.add(null);
+            return;
+        }
+
+        list.add(convertPdfObjectToIPdfStructElem(k));
     }
 
     private IPdfStructElem convertPdfObjectToIPdfStructElem(PdfObject obj) {

kernel/src/main/java/com/itextpdf/kernel/pdf/tagging/PdfStructTreeRoot.java

@@ -258,6 +258,15 @@ void addKidObject(PdfDictionary structElem) {
         addKidObject(-1, structElem);
     }
 
+    void addKidObject(int index, PdfDictionary structElem) {
+        if (index == -1)
+            getKidsObject().add(structElem);
+        else
+            getKidsObject().add(index, structElem);
+        if (PdfStructElem.isStructElem(structElem))
+            structElem.put(PdfName.P, getPdfObject());
+    }
+
     @Override
     protected boolean isWrappedObjectMustBeIndirect() {
         return true;
@@ -273,15 +282,6 @@ private void flushAllKids(IPdfStructElem elem) {
         }
     }
 
-    private void addKidObject(int index, PdfDictionary structElem) {
-        if (index == -1)
-            getKidsObject().add(structElem);
-        else
-            getKidsObject().add(index, structElem);
-        if (PdfStructElem.isStructElem(structElem))
-            structElem.put(PdfName.P, getPdfObject());
-    }
-
     private void ifKidIsStructElementAddToList(PdfObject kid, List<IPdfStructElem> kids) {
         if (kid.isFlushed()) {
             kids.add(null);