Support TreeDoc.dedupeNodes(); update TDNode.hashCode() to not includ…

…e key
eBay · Dec 5, 2022 · 29d8064 · 29d8064
1 parent bb7db42
commit 29d8064
Show file tree

Hide file tree

Showing 6 changed files with 75 additions and 29 deletions.
diff --git a/pom.xml b/pom.xml
@@ -85,7 +85,7 @@
       <dependency>
         <groupId>org.projectlombok</groupId>
         <artifactId>lombok</artifactId>
-        <version>1.18.16</version>
+        <version>1.18.24</version>
         <scope>provided</scope>
       </dependency>
       <dependency>
@@ -124,7 +124,7 @@
             <path>
               <groupId>org.projectlombok</groupId>
               <artifactId>lombok</artifactId>
-              <version>1.18.2</version>
+              <version>1.18.24</version>
             </path>
           </annotationProcessorPaths>
         </configuration>
@@ -182,7 +182,7 @@
           <plugin>
             <groupId>org.apache.maven.plugins</groupId>
             <artifactId>maven-surefire-plugin</artifactId>
-            <version>2.21.0</version>
+            <version>2.12.4</version>
             <configuration>
               <skipTests>true</skipTests>
             </configuration>
@@ -218,7 +218,7 @@
           <plugin>
             <groupId>org.apache.maven.plugins</groupId>
             <artifactId>maven-javadoc-plugin</artifactId>
-            <version>3.1.0</version>
+            <version>3.4.1</version>
             <executions>
               <execution>
                 <id>attach-javadocs</id>

diff --git a/treedoc/src/main/java/org/jsonex/treedoc/TDNode.java b/treedoc/src/main/java/org/jsonex/treedoc/TDNode.java
@@ -27,6 +27,7 @@
 
 import static org.jsonex.core.util.LangUtil.orElse;
 import static org.jsonex.core.util.ListUtil.last;
+import static org.jsonex.core.util.ListUtil.map;
 
 /** A Node in TreeDoc */
 @RequiredArgsConstructor
@@ -249,11 +250,11 @@ public StringBuilder toString(StringBuilder sb, boolean includeRootKey, boolean
   }
 
   public List<Object> childrenValueAsList() {
-    return getChildren() == null ? Collections.emptyList() : ListUtil.map(getChildren(), c -> c.getValue());
+    return getChildren() == null ? Collections.emptyList() : map(getChildren(), c -> c.getValue());
   }
 
   public List<List<Object>> childrenValueAsListOfList() {
-    return getChildren() == null ? Collections.emptyList() : ListUtil.map(getChildren(), c -> c.childrenValueAsList());
+    return getChildren() == null ? Collections.emptyList() : map(getChildren(), c -> c.childrenValueAsList());
   }
 
   @Override public boolean equals(Object o) {
@@ -267,7 +268,8 @@ public List<List<Object>> childrenValueAsListOfList() {
     return Objects.equals(key, tdNode.key) && Objects.equals(value, tdNode.value) && Objects.equals(children, tdNode.children);
   }
 
+  /** Hash code of value and children, key is not included */
   @Override public int hashCode() {
-    return hash.getOrCompute(() -> Objects.hash(key, value, children));
+    return hash.getOrCompute(() -> Objects.hash(value, children, map(children, TDNode::getKey)));
   }
 }
diff --git a/treedoc/src/main/java/org/jsonex/treedoc/TreeDoc.java b/treedoc/src/main/java/org/jsonex/treedoc/TreeDoc.java
@@ -6,10 +6,9 @@
 import lombok.experimental.Accessors;
 
 import java.net.URI;
-import java.util.Collection;
-import java.util.HashMap;
-import java.util.Map;
+import java.util.*;
 
+import static java.lang.String.format;
 import static org.jsonex.core.util.LangUtil.doIfNotNull;
 import static org.jsonex.core.util.ListUtil.mapKeys;
 
@@ -64,4 +63,32 @@ public static TreeDoc merge(Collection<TDNode> nodes) {
 
     return result;
   }
+
+  /**
+   * Dedupe the nodes that are with the same contends by comparing the hash
+   */
+  public void dedupeNodes() {
+    Map<Integer, TDNode> hashToNodeMap = new LinkedHashMap<>();
+    Queue<TDNode> queue = new LinkedList<>();
+    queue.add(root);
+    while(!queue.isEmpty()) {
+      TDNode n = queue.poll();
+      if (n.isLeaf())
+        continue;  // Don't dedupe leaf node and array node
+      int hash = n.hashCode();
+      TDNode refNode = hashToNodeMap.get(hash);
+      if ( refNode != null && refNode.getType() == TDNode.Type.MAP) {
+        if (refNode.getChild(TDNode.ID_KEY) == null)
+          refNode.createChild(TDNode.ID_KEY).setValue(format("%x", hash));
+        n.children.clear();;
+        n.setType(TDNode.Type.MAP).createChild(TDNode.REF_KEY).setValue(format("#%x", hash));
+      } else {
+        hashToNodeMap.put(hash, n);
+        if (n.hasChildren())
+          for (TDNode cn : n.children) {
+            queue.add(cn);
+          }
+      }
+    }
+  }
 }
diff --git a/treedoc/src/test/java/org/jsonex/treeedoc/TDPathTest.java b/treedoc/src/test/java/org/jsonex/treeedoc/TDPathTest.java
diff --git a/treedoc/src/test/resources/org/jsonex/treedoc/test.json b/treedoc/src/test/resources/org/jsonex/treedoc/test.json
@@ -0,0 +1,17 @@
+{
+  "a": 1,
+  "b": {
+    "c": [1,2,3,4],
+    "d": {
+      "e": 1,
+      "f": 2
+    }
+  },
+  "e": {
+    "c": [1, 2, 3, 4],
+    "d": {
+      "e": 1,
+      "f": 2
+    }
+  }
+}
diff --git a/treedoc/src/test/resources/org/jsonex/treedoc/test_deduped.json b/treedoc/src/test/resources/org/jsonex/treedoc/test_deduped.json
@@ -0,0 +1,19 @@
+{
+  a:1,
+  b:{
+    c:[
+      1,
+      2,
+      3,
+      4
+    ],
+    d:{
+      e:1,
+      f:2
+    },
+    $id:"5a9c3cc0"
+  },
+  e:{
+    $ref:"#5a9c3cc0"
+  }
+}