Changeset 16382

Timestamp:

12/14/18 13:30:28 (6 years ago)

Author:

bburlacu

Message:

#2950: Change signature of ComputeSimilarity methods to accent a generic list of trees. This enables us to directly pass HL ItemAray or ItemList without overhead.

File:

• trunk/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Hashing/SymbolicExpressionTreeHash.cs (modified) (8 diffs)

trunk/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Hashing/SymbolicExpressionTreeHash.cs

@@ -21 +21 @@
 
 using System;
+using System.Collections.Generic;
 using System.Linq;
 using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
@@ -70 +71 @@
     }
 
-    // this will only work if lh and rh are sorted 
+    // this will only work if lh and rh are sorted
     private static double ComputeSimilarity(ulong[] lh, ulong[] rh) {
       double count = 0;
@@ -90 +91 @@
     }
 
-    public static double ComputeAverageSimilarity(ISymbolicExpressionTree[] trees, bool simplify = false, bool strict = false) {
-      var total = (double)trees.Length * (trees.Length - 1) / 2;
+    public static double ComputeAverageSimilarity(IList<ISymbolicExpressionTree> trees, bool simplify = false, bool strict = false) {
+      var total = (double)trees.Count * (trees.Count - 1) / 2;
       double avg = 0;
-      var hashes = new ulong[trees.Length][];
+      var hashes = new ulong[trees.Count][];
       // build hash arrays
-      for (int i = 0; i < trees.Length; ++i) {
+      for (int i = 0; i < trees.Count; ++i) {
         var nodes = trees[i].MakeNodes(strict);
         hashes[i] = (simplify ? nodes.Simplify(HashUtil.DJBHash) : nodes.Sort(HashUtil.DJBHash)).Select(x => x.CalculatedHashValue).ToArray();
@@ -101 +102 @@
       }
       // compute similarity matrix
-      for (int i = 0; i < trees.Length - 1; ++i) {
-        for (int j = i + 1; j < trees.Length; ++j) {
+      for (int i = 0; i < trees.Count - 1; ++i) {
+        for (int j = i + 1; j < trees.Count; ++j) {
           avg += ComputeSimilarity(hashes[i], hashes[j]);
         }
@@ -109 +110 @@
     }
 
-    public static double[,] ComputeSimilarityMatrix(ISymbolicExpressionTree[] trees, bool simplify = false, bool strict = false) {
-      var sim = new double[trees.Length, trees.Length];
-      var hashes = new ulong[trees.Length][];
+    public static double[,] ComputeSimilarityMatrix(IList<ISymbolicExpressionTree> trees, bool simplify = false, bool strict = false) {
+      var sim = new double[trees.Count, trees.Count];
+      var hashes = new ulong[trees.Count][];
       // build hash arrays
-      for (int i = 0; i < trees.Length; ++i) {
+      for (int i = 0; i < trees.Count; ++i) {
         var nodes = trees[i].MakeNodes(strict);
         hashes[i] = (simplify ? nodes.Simplify(HashUtil.DJBHash) : nodes.Sort(HashUtil.DJBHash)).Select(x => x.CalculatedHashValue).ToArray();
@@ -119 +120 @@
       }
       // compute similarity matrix
-      for (int i = 0; i < trees.Length - 1; ++i) {
-        for (int j = i + 1; j < trees.Length; ++j) {
+      for (int i = 0; i < trees.Count - 1; ++i) {
+        for (int j = i + 1; j < trees.Count; ++j) {
           sim[i, j] = sim[j, i] = ComputeSimilarity(hashes[i], hashes[j]);
         }
@@ -218 +219 @@
     #region tree simplification
     // these simplification methods rely on the assumption that child nodes of the current node have already been simplified
-    // (in other words simplification should be applied in a bottom-up fashion) 
+    // (in other words simplification should be applied in a bottom-up fashion)
     public static ISymbolicExpressionTree Simplify(ISymbolicExpressionTree tree) {
       ulong hashFunction(byte[] bytes) => HashUtil.JSHash(bytes);
@@ -248 +249 @@
     }
 
-    // simplify multiplications by reducing constants and div terms  
+    // simplify multiplications by reducing constants and div terms
     public static void SimplifyMultiplication(ref HashNode<ISymbolicExpressionTreeNode>[] nodes, int i) {
       var node = nodes[i];