Merge pull request #635 from veraPDF/struct_tree_loop

Add warnings for struct tree loop
veraPDF · Jan 24, 2024 · cdf0bf9 · cdf0bf9
2 parents aa9cd8a + 05b8408
commit cdf0bf9
Show file tree

Hide file tree

Showing 3 changed files with 38 additions and 4 deletions.
diff --git a/...dation-model/src/main/java/org/verapdf/gf/model/impl/operator/markedcontent/GFOp_BDC.java b/...dation-model/src/main/java/org/verapdf/gf/model/impl/operator/markedcontent/GFOp_BDC.java
@@ -25,7 +25,6 @@
 import org.verapdf.cos.COSKey;
 import org.verapdf.cos.COSObjType;
 import org.verapdf.cos.COSObject;
-import org.verapdf.exceptions.LoopedException;
 import org.verapdf.gf.model.impl.pd.util.PDResourcesHandler;
 import org.verapdf.model.baselayer.Object;
 import org.verapdf.model.operator.Op_BDC;
@@ -37,12 +36,16 @@
 import java.util.List;
 import java.util.Set;
 import java.util.HashSet;
+import java.util.logging.Level;
+import java.util.logging.Logger;
 
 /**
  * @author Timur Kamalov
  */
 public class GFOp_BDC extends GFOpMarkedContent implements Op_BDC {
 
+	private static final Logger LOGGER = Logger.getLogger(GFOp_BDC.class.getCanonicalName());
+
 	/** Type name for {@code GFOp_BDC} */
     public static final String OP_BDC_TYPE = "Op_BDC";
 	public final StructureElementAccessObject structureElementAccessObject;
@@ -101,7 +104,8 @@ public String getStructParentLang() {
 		while (baseLang == null && parent != null) {
 			COSKey key = parent.getObjectKey();
 			if (keys.contains(key)) {
-				throw new LoopedException("Struct tree loop found");
+				LOGGER.log(Level.WARNING, "Struct tree loop found");
+				break;
 			}
 			if (key != null) {
 				keys.add(key);

diff --git a/validation-model/src/main/java/org/verapdf/gf/model/impl/pd/GFPDStructElem.java b/validation-model/src/main/java/org/verapdf/gf/model/impl/pd/GFPDStructElem.java
@@ -27,7 +27,6 @@
 import org.verapdf.cos.COSString;
 import org.verapdf.cos.COSKey;
 import org.verapdf.cos.COSArray;
-import org.verapdf.exceptions.LoopedException;
 import org.verapdf.gf.model.impl.containers.StaticContainers;
 import org.verapdf.gf.model.impl.cos.GFCosActualText;
 import org.verapdf.gf.model.impl.cos.GFCosAlt;
@@ -46,11 +45,16 @@
 import org.verapdf.tools.TaggedPDFHelper;
 
 import java.util.*;
+import java.util.logging.Level;
+import java.util.logging.Logger;
 
 /**
  * @author Maksim Bezrukov
  */
 public class GFPDStructElem extends GFPDStructTreeNode implements PDStructElem {
+
+	private static final Logger LOGGER = Logger.getLogger(GFPDStructElem.class.getCanonicalName());
+
 	/**
 	 * Type name for {@code GFPDStructElem}
 	 */
@@ -264,7 +268,8 @@ public String getparentLang() {
 		while (baseLang == null && parent != null) {
 			key = parent.getObject().getObjectKey();
 			if (keys.contains(key)) {
-				throw new LoopedException("Struct tree loop found");
+				LOGGER.log(Level.WARNING, "Struct tree loop found");
+				break;
 			}
 			if (key != null) {
 				keys.add(key);

diff --git a/...on-model/src/main/java/org/verapdf/gf/model/impl/pd/gfse/contents/GFSEGroupedContent.java b/...on-model/src/main/java/org/verapdf/gf/model/impl/pd/gfse/contents/GFSEGroupedContent.java
@@ -21,6 +21,7 @@
 package org.verapdf.gf.model.impl.pd.gfse.contents;
 
 import org.verapdf.as.ASAtom;
+import org.verapdf.cos.COSKey;
 import org.verapdf.cos.COSObject;
 import org.verapdf.model.operator.Operator;
 import org.verapdf.model.selayer.SEGroupedContent;
@@ -29,10 +30,16 @@
 import org.verapdf.tools.TaggedPDFConstants;
 import org.verapdf.tools.TaggedPDFRoleMapHelper;
 
+import java.util.HashSet;
 import java.util.List;
+import java.util.Set;
+import java.util.logging.Level;
+import java.util.logging.Logger;
 
 public abstract class GFSEGroupedContent extends GFSEContentItem implements SEGroupedContent {
 
+    private static final Logger LOGGER = Logger.getLogger(GFSEGroupedContent.class.getCanonicalName());
+
     protected final List<Operator> operators;
     protected final COSObject parentStructElem;
     protected final String parentsTags;
@@ -87,12 +94,21 @@ public Boolean getisArtifact() {
 
     protected Boolean hasParentWithStandardType(String standardType) {
         TaggedPDFRoleMapHelper taggedPDFRoleMapHelper = StaticResources.getRoleMapHelper();
+        Set<COSKey> keys = new HashSet<>();
         if (parentStructElem != null && taggedPDFRoleMapHelper != null) {
             PDStructElem structElem = new PDStructElem(parentStructElem, taggedPDFRoleMapHelper.getRoleMap());
             while (structElem != null) {
                 if (standardType.equals(PDStructElem.getStructureElementStandardType(structElem))) {
                     return true;
                 }
+                COSKey key = structElem.getObject().getObjectKey();
+                if (keys.contains(key)) {
+                    LOGGER.log(Level.WARNING, "Struct tree loop found");
+                    break;
+                }
+                if (key != null) {
+                    keys.add(key);
+                }
                 structElem = structElem.getParent();
             }
         }
@@ -127,11 +143,20 @@ public Boolean getisTaggedContent() {
     }
 
     private boolean isTaggedContent() {
+        Set<COSKey> keys = new HashSet<>();
         PDStructElem structElem = new PDStructElem(parentStructElem, null);
         while (structElem != null) {
             if (structElem.getType() == ASAtom.STRUCT_TREE_ROOT) {
                 return true;
             }
+            COSKey key = structElem.getObject().getObjectKey();
+            if (keys.contains(key)) {
+                LOGGER.log(Level.WARNING, "Struct tree loop found");
+                break;
+            }
+            if (key != null) {
+                keys.add(key);
+            }
             structElem = structElem.getParent();
         }
         return false;