Changeset 17364 for branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis

Timestamp:

11/22/19 10:06:42 (5 years ago)

Author:

pfleck

Message:

#3040

• Added double vectors for Dataset. Extended the type-checks for DataAnalysisProblemData.
• Added a small benchmark instance with data containing vectors. Adapted the ArtificialRegressionDataDescriptor to be able to specify non-double values.

Location:

branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis/3.4

Files:

• Dataset.cs (modified) (8 diffs)
• Implementation/DataAnalysisProblemData.cs (modified) (1 diff)
• Interfaces/IDataset.cs (modified) (2 diffs)

branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis/3.4/Dataset.cs

@@ -142 +142 @@
         else if (iter.Current[i] is string s)
           transposed.Add(new List<string>() { s });
+        else if (iter.Current[i] is IReadOnlyList<double> dv)
+          transposed.Add(new List<IReadOnlyList<double>>() { dv });
         else throw new NotSupportedException(string.Format("Variable {0} has type {1}. This is not supported when converting from row-wise data.", vnames[i], iter.Current[i].GetType()));
       }
@@ -212 +214 @@
       get { return variableValues.Where(p => p.Value is IList<double>).Select(p => p.Key); }
     }
-
     public IEnumerable<string> StringVariables {
       get { return variableValues.Where(p => p.Value is IList<string>).Select(p => p.Key); }
     }
-
     public IEnumerable<string> DateTimeVariables {
       get { return variableValues.Where(p => p.Value is IList<DateTime>).Select(p => p.Key); }
     }
+    public IEnumerable<string> DoubleVectorVariables {
+      get { return variableValues.Where(p => p.Value is IList<IReadOnlyList<double>>).Select(p => p.Key); }
+    }
 
     public IEnumerable<double> GetDoubleValues(string variableName) {
@@ -229 +232 @@
     public IEnumerable<DateTime> GetDateTimeValues(string variableName) {
       return GetValues<DateTime>(variableName);
+    }
+    public IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName) {
+      return GetValues<IReadOnlyList<double>>(variableName);
     }
 
@@ -247 +253 @@
       return values[row];
     }
-
     public IEnumerable<string> GetStringValues(string variableName, IEnumerable<int> rows) {
       return GetValues<string>(variableName, rows);
@@ -267 +272 @@
       return new ReadOnlyCollection<DateTime>(values);
     }
+
+    public IReadOnlyList<double> GetDoubleVectorValue(string variableName, int row) {
+      var values = GetValues<IReadOnlyList<double>>(variableName);
+      return values[row];
+    }
+    public IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName, IEnumerable<int> rows) {
+      return GetValues<IReadOnlyList<double>>(variableName, rows);
+    }
+    public ReadOnlyCollection<IReadOnlyList<double>> GetReadOnlyDoubleVectorValues(string variableName) {
+      var values = GetValues<IReadOnlyList<double>>(variableName);
+      return new ReadOnlyCollection<IReadOnlyList<double>>(values);
+    }
+
     private IEnumerable<T> GetValues<T>(string variableName, IEnumerable<int> rows) {
       var values = GetValues<T>(variableName);
@@ -298 +316 @@
     }
     protected static bool IsAllowedType(Type type) {
-      return type == typeof(double) || type == typeof(string) || type == typeof(DateTime);
+      return type == typeof(double) || type == typeof(string) || type == typeof(DateTime) || type == typeof(IReadOnlyList<double>);
     }
 
@@ -343 +361 @@
       if (dateTimeValues != null) return new List<DateTime>(dateTimeValues);
 
+      var doubleVectorValues = values as IList<IReadOnlyList<double>>;
+      if (doubleVectorValues != null) return doubleVectorValues.Select(x => new List<double>(x)).Cast<IReadOnlyList<double>>().ToList();
+
       throw new ArgumentException(string.Format("Unsupported variable type {0}.", GetElementType(values)));
     }
@@ -381 +402 @@
     }
     string IStringConvertibleMatrix.GetValue(int rowIndex, int columnIndex) {
+      var value = variableValues[variableNames[columnIndex]][rowIndex];
+      if (value is IReadOnlyList<double> doubleVector)
+        return $"[{string.Join(", ", doubleVector)}]";
       return variableValues[variableNames[columnIndex]][rowIndex].ToString();
     }

branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/DataAnalysisProblemData.cs

@@ -158 +158 @@
       if (allowedInputVariables == null) throw new ArgumentNullException("The allowed input variables must not be null.");
 
-      if (allowedInputVariables.Except(dataset.DoubleVariables).Except(dataset.StringVariables).Any())
-        throw new ArgumentException("All allowed input variables must be present in the dataset and of type double or string.");
-
-      var variables = dataset.VariableNames.Where(variable => dataset.VariableHasType<double>(variable) || dataset.VariableHasType<string>(variable));
+      if (allowedInputVariables.Except(dataset.DoubleVariables).Except(dataset.StringVariables).Except(dataset.DoubleVectorVariables).Any())
+        throw new ArgumentException("All allowed input variables must be present in the dataset and of type double, string or double-vector.");
+
+      var variables = dataset.VariableNames.Where(variable =>
+        dataset.VariableHasType<double>(variable) || dataset.VariableHasType<string>(variable) || dataset.VariableHasType<IReadOnlyList<double>>(variable));
       var inputVariables = new CheckedItemList<StringValue>(variables.Select(x => new StringValue(x).AsReadOnly()));
       foreach (StringValue x in inputVariables)

branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis/3.4/Interfaces/IDataset.cs

@@ -34 +34 @@
     IEnumerable<string> StringVariables { get; }
     IEnumerable<string> DateTimeVariables { get; }
+    IEnumerable<string> DoubleVectorVariables { get; }
 
     bool ContainsVariable(string variablename);
@@ -48 +49 @@
     ReadOnlyCollection<string> GetReadOnlyStringValues(string VariableName);
 
-    System.DateTime GetDateTimeValue(string variableName, int row);
+    DateTime GetDateTimeValue(string variableName, int row);
     IEnumerable<DateTime> GetDateTimeValues(string variableName);
     IEnumerable<DateTime> GetDateTimeValues(string variableName, IEnumerable<int> rows);
     ReadOnlyCollection<DateTime> GetReadOnlyDateTimeValues(string variableName);
+
+    IReadOnlyList<double> GetDoubleVectorValue(string variableName, int row);
+    IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName);
+    IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName, IEnumerable<int> rows);
+    ReadOnlyCollection<IReadOnlyList<double>> GetReadOnlyDoubleVectorValues(string variableName);
   }
 }