Implement PDF/UA checks for checkpoint 02: role mapping

DEVSIX-7922

Author: introfog

pdfua/src/main/java/com/itextpdf/pdfua/checkers/PdfUA1Checker.java

@@ -31,16 +31,23 @@ This file is part of the iText (R) project.
 import com.itextpdf.kernel.pdf.PdfDictionary;
 import com.itextpdf.kernel.pdf.PdfDocument;
 import com.itextpdf.kernel.pdf.PdfName;
+import com.itextpdf.kernel.pdf.PdfObject;
 import com.itextpdf.kernel.pdf.PdfResources;
 import com.itextpdf.kernel.pdf.PdfStream;
 import com.itextpdf.kernel.pdf.tagging.PdfMcr;
+import com.itextpdf.kernel.pdf.tagging.PdfNamespace;
+import com.itextpdf.kernel.pdf.tagging.PdfStructTreeRoot;
+import com.itextpdf.kernel.pdf.tagging.StandardRoles;
+import com.itextpdf.kernel.pdf.tagutils.IRoleMappingResolver;
+import com.itextpdf.kernel.pdf.tagutils.TagStructureContext;
 import com.itextpdf.kernel.utils.IValidationChecker;
 import com.itextpdf.kernel.utils.ValidationContext;
 import com.itextpdf.pdfua.exceptions.PdfUAConformanceException;
 import com.itextpdf.pdfua.exceptions.PdfUAExceptionMessageConstants;
 
 import java.util.Collection;
 import java.util.HashSet;
+import java.util.Map;
 import java.util.Set;
 import java.util.Stack;
 
@@ -54,13 +61,16 @@ public class PdfUA1Checker implements IValidationChecker {
 
     private final PdfDocument pdfDocument;
 
+    private final TagStructureContext tagStructureContext;
+
     /**
      * Creates PdfUA1Checker instance with PDF document which will be validated against PDF/UA-1 standard.
      *
      * @param pdfDocument the document to validate
      */
     public PdfUA1Checker(PdfDocument pdfDocument) {
         this.pdfDocument = pdfDocument;
+        this.tagStructureContext = new TagStructureContext(pdfDocument);
     }
 
     /**
@@ -69,6 +79,7 @@ public PdfUA1Checker(PdfDocument pdfDocument) {
     @Override
     public void validateDocument(ValidationContext validationContext) {
         checkCatalog(validationContext.getPdfDocument().getCatalog());
+        checkStructureTreeRoot(validationContext.getPdfDocument().getStructTreeRoot());
         checkFonts(validationContext.getFonts());
     }
 
@@ -107,13 +118,14 @@ private Stack<Tuple2<PdfName, PdfDictionary>> getTagStack(Object data) {
     }
 
     private void checkOnOpeningBeginMarkedContent(Object obj, Object extra) {
+        Tuple2<PdfName, PdfDictionary> currentBmc = (Tuple2<PdfName, PdfDictionary>) extra;
+        checkStandardRoleMapping(currentBmc);
+
         Stack<Tuple2<PdfName, PdfDictionary>> stack = getTagStack(obj);
         if (stack.isEmpty()) {
             return;
         }
 
-        Tuple2<PdfName, PdfDictionary> currentBmc = (Tuple2<PdfName, PdfDictionary>) extra;
-
         boolean isRealContent = isRealContent(currentBmc);
         boolean isArtifact = PdfName.Artifact.equals(currentBmc.getFirst());
 
@@ -125,6 +137,15 @@ private void checkOnOpeningBeginMarkedContent(Object obj, Object extra) {
         }
     }
 
+    private void checkStandardRoleMapping(Tuple2<PdfName, PdfDictionary> tag) {
+        final PdfNamespace namespace = tagStructureContext.getDocumentDefaultNamespace();
+        final String role = tag.getFirst().getValue();
+        if (!StandardRoles.ARTIFACT.equals(role) && !tagStructureContext.checkIfRoleShallBeMappedToStandardRole(role, namespace)) {
+            throw new PdfUAConformanceException(
+                    MessageFormatUtil.format(PdfUAExceptionMessageConstants.TAG_MAPPING_DOESNT_TERMINATE_WITH_STANDARD_TYPE, role));
+        }
+    }
+
     private boolean isInsideArtifact(Stack<Tuple2<PdfName, PdfDictionary>> tagStack) {
         for (Tuple2<PdfName, PdfDictionary> tag : tagStack) {
             if (PdfName.Artifact.equals(tag.getFirst())) {
@@ -175,6 +196,19 @@ private void checkCatalog(PdfCatalog catalog) {
         }
     }
 
+    private void checkStructureTreeRoot(PdfStructTreeRoot structTreeRoot) {
+        PdfDictionary roleMap = structTreeRoot.getRoleMap();
+        for (Map.Entry<PdfName, PdfObject> entry : roleMap.entrySet()) {
+            final String role = entry.getKey().getValue();
+            final IRoleMappingResolver roleMappingResolver = pdfDocument.getTagStructureContext()
+                    .getRoleMappingResolver(role);
+
+            if (roleMappingResolver.currentRoleIsStandard()) {
+                throw new PdfUAConformanceException(PdfUAExceptionMessageConstants.ONE_OR_MORE_STANDARD_ROLE_REMAPPED);
+            }
+        }
+    }
+
     private void checkFonts(Collection<PdfFont> fontsInDocument) {
         Set<String> fontNamesThatAreNotEmbedded = new HashSet<>();
         for (PdfFont font : fontsInDocument) {

pdfua/src/main/java/com/itextpdf/pdfua/exceptions/PdfUAExceptionMessageConstants.java

@@ -26,6 +26,8 @@ This file is part of the iText (R) project.
  * Class that bundles all the error message templates as constants.
  */
 public final class PdfUAExceptionMessageConstants {
+    public static final String ONE_OR_MORE_STANDARD_ROLE_REMAPPED =
+            "One or more standard types are remapped.";
     public static final String TAG_HASNT_BEEN_ADDED_BEFORE_CONTENT_ADDING =
             "Tag hasn't been added before adding content to the canvas.";
     public static final String CONTENT_IS_NOT_REAL_CONTENT_AND_NOT_ARTIFACT =
@@ -43,6 +45,8 @@ public final class PdfUAExceptionMessageConstants {
             "Tagged content is present inside content marked as Artifact or vice versa.";
     public static final String SUSPECTS_ENTRY_IN_MARK_INFO_DICTIONARY_SHALL_NOT_HAVE_A_VALUE_OF_TRUE =
             "Suspects entry in mark info dictionary shall not have a value of true.";
+    public static final String TAG_MAPPING_DOESNT_TERMINATE_WITH_STANDARD_TYPE =
+            "\"{0}\" tag mapping does not terminate with a standard type.";
 
     private PdfUAExceptionMessageConstants() {
         // Empty constructor

pdfua/src/test/java/com/itextpdf/pdfua/PdfUACanvasTest.java

@@ -432,7 +432,7 @@ public void checkPoint_01_005_RectangleMarkedContentWithoutMcid() throws IOExcep
         PdfCanvas canvas = new PdfCanvas(pdfDoc.addNewPage());
         canvas
                 .saveState()
-                .openTag(new CanvasTag(PdfName.Rect))
+                .openTag(new CanvasTag(PdfName.Art))
                 .setFillColor(ColorConstants.RED);
         Exception e = Assert.assertThrows(PdfUAConformanceException.class, () -> {
             canvas.rectangle(new Rectangle(200, 200, 100, 100));