Changeset 10185 for branches/DataPreprocessing/HeuristicLab.DataPreprocessing

Timestamp:

12/04/13 13:51:07 (11 years ago)

Author:

pfleck

Message:

Implemented constructors of PreprocessingData.

File:

• branches/DataPreprocessing/HeuristicLab.DataPreprocessing/3.3/Implementations/PreprocessingData.cs (modified) (2 diffs)

branches/DataPreprocessing/HeuristicLab.DataPreprocessing/3.3/Implementations/PreprocessingData.cs

@@ -23 +23 @@
 using System.Collections;
 using System.Collections.Generic;
+using System.Linq;
+using HeuristicLab.Common;
 using HeuristicLab.Core;
 using HeuristicLab.DataPreprocessing;
@@ -31 +33 @@
   public class PreprocessingData : NamedItem, IPreprocessingData {
 
-    private Dictionary<string, IList> variableValues;
+    private IDictionary<string, IList> variableValues;
 
+    private IDictionary<string, int> variableNameIndices;
+
+    private double trainingToTestRatio;
+
+    private PreprocessingData(PreprocessingData original, Cloner cloner)
+      : base(original, cloner) {
+      variableValues = new Dictionary<string, IList>(variableValues);
+      variableNameIndices = new Dictionary<string, int>(variableNameIndices);
+    }
+    
     public PreprocessingData(IDataAnalysisProblemData problemData)
       : base() {
       Name = "-";
-      foreach (var s in problemData.Dataset.VariableNames) {
-        if (problemData.Dataset.IsType<double>(s)) {
 
-        } else if (problemData.Dataset.IsType<string>(s)) {
-
-        } else if (problemData.Dataset.IsType<DateTime>(s)) {
-
+      // create dictionary from variable name to index
+      variableNameIndices = new Dictionary<string, int>();
+      var variableNamesList = problemData.Dataset.VariableNames.ToList();
+      for (int i = 0; i < variableNamesList.Count; i++) {
+        variableNameIndices.Add(variableNamesList[i], i);
+      }
+      
+      // copy values
+      variableValues = new Dictionary<string, IList>();
+      foreach (var variableName in problemData.Dataset.VariableNames) {
+        if (problemData.Dataset.IsType<double>(variableName)) {
+          variableValues[variableName] = problemData.Dataset.GetDoubleValues(variableName).ToList();
+        } else if (problemData.Dataset.IsType<string>(variableName)) {
+          variableValues[variableName] = CreateColumn<string>(problemData.Dataset, variableNameIndices[variableName], x => x);
+        } else if (problemData.Dataset.IsType<DateTime>(variableName)) {
+          variableValues[variableName] = CreateColumn<DateTime>(problemData.Dataset, variableNameIndices[variableName], x => DateTime.Parse(x));
         } else {
-          throw new ArgumentException("The datatype of column " + s + "is of TODO");
+          throw new ArgumentException("The datatype of column " + variableName + " must be of type List<double>, List<string> or List<DateTime>");
         }
       }
-      throw new NotImplementedException();
+
+      trainingToTestRatio = (double)problemData.TrainingPartition.Size / problemData.TestPartition.Size;
+    }
+
+    private static IList CreateColumn<T>(Dataset ds, int column, Func<string, T> selector) {
+      var list = new List<T>(ds.Rows);
+      for (int row = 0; row < ds.Rows; row++) {
+        list[row] = selector(ds.GetValue(row, column));
+      }
+      return list;
     }
 
     #region NamedItem abstract Member Implementations
 
-    public override Common.IDeepCloneable Clone(Common.Cloner cloner) {
-      throw new NotImplementedException();
+    public override IDeepCloneable Clone(Cloner cloner) {
+      return new PreprocessingData(this, cloner);
     }