Changeset 13625

Timestamp:

02/18/16 19:36:09 (8 years ago)

Author:

bburlacu

Message:

#1772: Small refactor of the SchemaCreator to avoid generating duplicate schemas when non-strict matching is used.

File:

• branches/HeuristicLab.EvolutionTracking/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Tracking/SchemaDiversification/SchemaCreator.cs (modified) (5 diffs)

branches/HeuristicLab.EvolutionTracking/HeuristicLab.Problems.DataAnalysis.Symbolic/3.4/Tracking/SchemaDiversification/SchemaCreator.cs

@@ -23 +23 @@
 using System.Collections.Generic;
 using System.Linq;
+using System.Text;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -201 +202 @@
                      select v;
 
-      var schemas = new List<ISymbolicExpressionTree>(GenerateSchemas(vertices, MinimumSchemaLength));
+      var schemas = new List<ISymbolicExpressionTree>(GenerateSchemas(vertices, MinimumSchemaLength, StrictSchemaMatching));
 
       #region create schemas and add subscopes representing the individuals
@@ -223 +224 @@
     }
 
-    public static IEnumerable<ISymbolicExpressionTree> GenerateSchemas(IEnumerable<IGenealogyGraphNode<ISymbolicExpressionTree>> vertices, int minimumSchemaLength) {
+    public static IEnumerable<ISymbolicExpressionTree> GenerateSchemas(IEnumerable<IGenealogyGraphNode<ISymbolicExpressionTree>> vertices, int minimumSchemaLength, bool strict = true) {
       var anySubtreeSymbol = new AnySubtreeSymbol();
       //            var anyNodeSymbol = new AnyNodeSymbol();
       var groups = vertices.GroupBy(x => x.Parents.First()).OrderByDescending(g => g.Count()).ToList();
       var hash = new HashSet<string>();
-      var formatter = new SymbolicExpressionTreeStringFormatter { Indent = false, AppendNewLines = false };
+      //      var formatter = new SymbolicExpressionTreeStringFormatter { Indent = false, AppendNewLines = false };
       foreach (var g in groups) {
         var parent = g.Key;
@@ -256 +257 @@
         }
         if (replaced) {
-          var str = formatter.Format(schema.Root.GetSubtree(0).GetSubtree(0));
+          //          var str = formatter.Format(schema.Root.GetSubtree(0).GetSubtree(0));
+          var str = SubtreeToString(schema.Root.GetSubtree(0).GetSubtree(0), strict);
           if (hash.Contains(str)) continue;
           yield return schema;
@@ -283 +285 @@
       }
     }
+
+    private static string SubtreeToString(ISymbolicExpressionTreeNode node, bool strict = false) {
+      StringBuilder strBuilder = new StringBuilder();
+      // internal nodes or leaf nodes?
+      if (node is AnySubtree)
+        return "# ";
+
+      if (node.SubtreeCount > 0) {
+        strBuilder.Append("(");
+        // symbol on same line as '('
+        string label = string.Empty;
+        if (node is AnyNode)
+          label = "=";
+        else {
+          label = node.Symbol.Name;
+        }
+        strBuilder.Append(label + " ");
+        // each subtree expression on a new line
+        // and closing ')' also on new line
+        foreach (var subtree in node.Subtrees) {
+          strBuilder.Append(SubtreeToString(subtree, strict));
+        }
+        strBuilder.Append(") ");
+      } else {
+        // symbol in the same line with as '(' and ')'
+        var v = node as VariableTreeNode;
+        var c = node as ConstantTreeNode;
+        var w = node as AnyNode; // wildcard
+        string label = string.Empty;
+        if (w != null)
+          label = "=";
+        else if (v != null)
+          label = strict ? string.Format("{0:0.00}_{1}", v.Weight, v.VariableName) : string.Format("{0}", v.VariableName);
+        else if (c != null)
+          label = strict ? string.Format("{0:0.00}", c.Value) : "C";
+        strBuilder.Append(label);
+        if (node.Parent != null && node != node.Parent.Subtrees.Last())
+          strBuilder.Append(" ");
+        //strBuilder.Append(")");
+      }
+      return strBuilder.ToString();
+    }
   }
 }