Fix NPE, when structure element references a page outside the page tree

For some reason OCR in Apryse SDK produces such files. There is a
proper page in the page tree. But the structure elements reference some
entirely different page, which is not referenced in /Pages...

Now in such cases we will assume, that no marked content was found.

Author: Eswcvlad

src/main/java/com/itextpdf/rups/view/itext/StructureTree.java

@@ -47,6 +47,7 @@ This file is part of the iText (R) project.
 import com.itextpdf.kernel.pdf.PdfName;
 import com.itextpdf.kernel.pdf.PdfNumber;
 import com.itextpdf.kernel.pdf.PdfObject;
+import com.itextpdf.kernel.pdf.PdfPage;
 import com.itextpdf.rups.controller.PdfReaderController;
 import com.itextpdf.rups.model.ObjectLoader;
 import com.itextpdf.rups.model.TreeNodeFactory;
@@ -134,15 +135,27 @@ TreeModel recalculateTreeModel() {
         return new DefaultTreeModel(root);
     }
 
-    private Map<Integer, MarkedContentInfo> indexMarkedContentOnPage(PdfDictionary page) {
-        final PdfIndirectReference ref = page.getIndirectReference();
+    private Map<Integer, MarkedContentInfo> indexMarkedContentOnPage(PdfDictionary pageDict) {
+        final PdfIndirectReference ref = pageDict.getIndirectReference();
         Map<Integer, MarkedContentInfo> result = this.mciByPage.get(ref);
         if (result != null) {
             return result;
         }
-        final MarkedContentInfoGatherer gatherer = new MarkedContentInfoGatherer();
-        gatherer.processPageContent(this.loader.getFile().getPdfDocument().getPage(page));
-        result = gatherer.getMarkedContentIndex();
+        final PdfPage page = this.loader.getFile().getPdfDocument().getPage(pageDict);
+        if (page != null) {
+            final MarkedContentInfoGatherer gatherer = new MarkedContentInfoGatherer();
+            gatherer.processPageContent(page);
+            result = gatherer.getMarkedContentIndex();
+        } else {
+            /*
+             * This can happen in weird cases, when there is a page in the
+             * document, that is referenced within a structure element, but it
+             * is absent in the page tree for some reason... OCR in Apryse SDK
+             * does this, oddly enough. So we will just assume, that we could
+             * not find anything there.
+             */
+            result = Map.of();
+        }
         this.mciByPage.put(ref, result);
         return result;
     }

src/test/java/com/itextpdf/rups/view/itext/StructureTreeTest.java

@@ -127,6 +127,35 @@ void structElemsGeneratedByAcrobatTest()
         Assertions.assertEquals(0, secondTextNode.getChildCount());
     }
 
+    @Test
+    void structElemsGeneratedByApryseSdkOcrTest()
+            throws IOException, ExecutionException, InterruptedException {
+        /*
+         * For some reason when OCR in Apryse SDK creates the structure elem
+         * tree, the elements there reference a page, which is absent in the
+         * overall page tree. We shouldn't crash in such cases.
+         */
+        final PdfFile pdfFile = PdfFile.open(
+                new File(SOURCE_DIR + "ApryseSdkOcrTest.pdf")
+        );
+
+        final StructureTreeNode rootNode = getStructureTreeRootNode(pdfFile);
+        Assertions.assertEquals(1, rootNode.getChildCount());
+
+        final StructureTreeNode docNode = (StructureTreeNode) rootNode.getChildAt(0);
+        Assertions.assertEquals("/Document", docNode.toString());
+        Assertions.assertEquals(1, docNode.getChildCount());
+        final StructureTreeNode floatNode = (StructureTreeNode) docNode.getChildAt(0);
+        Assertions.assertEquals("/Float", floatNode.toString());
+        Assertions.assertEquals(1, floatNode.getChildCount());
+        final StructureTreeNode figureNode = (StructureTreeNode) floatNode.getChildAt(0);
+        Assertions.assertEquals("/Figure", figureNode.toString());
+        Assertions.assertEquals(1, figureNode.getChildCount());
+        final StructureTreeNode mcidNode = (StructureTreeNode) figureNode.getChildAt(0);
+        Assertions.assertEquals("0", mcidNode.toString());
+        Assertions.assertEquals(0, mcidNode.getChildCount());
+    }
+
     private static StructureTreeNode getStructureTreeRootNode(IPdfFile pdfFile)
             throws ExecutionException, InterruptedException {