Changeset 15285 for branches/DataPreprocessing Cleanup/HeuristicLab.DataPreprocessing/3.4/Content/ManipulationContent.cs

Timestamp:

07/26/17 14:12:27 (7 years ago)

Author:

pfleck

Message:

#2809 removed ManipulationLogic

File:

• branches/DataPreprocessing Cleanup/HeuristicLab.DataPreprocessing/3.4/Content/ManipulationContent.cs (modified) (3 diffs)

branches/DataPreprocessing Cleanup/HeuristicLab.DataPreprocessing/3.4/Content/ManipulationContent.cs

@@ -20 +20 @@
 #endregion
 
+using System;
+using System.Collections.Generic;
 using System.Drawing;
+using System.Linq;
 using HeuristicLab.Common;
 using HeuristicLab.Core;
@@ -33 +36 @@
     }
 
-    [Storable]
-    public ManipulationLogic ManipulationLogic { get; private set; }
-
     #region Constructor, Cloning & Persistence
-    public ManipulationContent(IFilteredPreprocessingData preprocessingData, ManipulationLogic manipulationLogic)
+    public ManipulationContent(IFilteredPreprocessingData preprocessingData)
       : base(preprocessingData) {
-      ManipulationLogic = manipulationLogic;
     }
 
     public ManipulationContent(ManipulationContent original, Cloner cloner) :
       base(original, cloner) {
-      ManipulationLogic = cloner.Clone(original.ManipulationLogic);
     }
     public override IDeepCloneable Clone(Cloner cloner) {
@@ -54 +52 @@
       : base(deserializing) { }
     #endregion
+
+    public List<int> RowsWithMissingValuesGreater(double percent) {
+      List<int> rows = new List<int>();
+
+      for (int i = 0; i < PreprocessingData.Rows; ++i) {
+        int missingCount = PreprocessingData.GetRowMissingValueCount(i);
+        if (100f / PreprocessingData.Columns * missingCount > percent) {
+          rows.Add(i);
+        }
+      }
+
+      return rows;
+    }
+
+    public List<int> ColumnsWithMissingValuesGreater(double percent) {
+      List<int> columns = new List<int>();
+      for (int i = 0; i < PreprocessingData.Columns; ++i) {
+        int missingCount = PreprocessingData.GetMissingValueCount(i);
+        if (100f / PreprocessingData.Rows * missingCount > percent) {
+          columns.Add(i);
+        }
+      }
+
+      return columns;
+    }
+
+    public List<int> ColumnsWithVarianceSmaller(double variance) {
+      List<int> columns = new List<int>();
+      for (int i = 0; i < PreprocessingData.Columns; ++i) {
+        if (PreprocessingData.VariableHasType<double>(i)) {
+          double columnVariance = PreprocessingData.GetVariance<double>(i);
+          if (columnVariance < variance) {
+            columns.Add(i);
+          }
+        } else if (PreprocessingData.VariableHasType<DateTime>(i)) {
+          double columnVariance = (double)PreprocessingData.GetVariance<DateTime>(i).Ticks / TimeSpan.TicksPerSecond;
+          if (columnVariance < variance) {
+            columns.Add(i);
+          }
+        }
+      }
+      return columns;
+    }
+
+    public void DeleteRowsWithMissingValuesGreater(double percent) {
+      DeleteRows(RowsWithMissingValuesGreater(percent));
+    }
+
+    public void DeleteColumnsWithMissingValuesGreater(double percent) {
+      DeleteColumns(ColumnsWithMissingValuesGreater(percent));
+    }
+
+    public void DeleteColumnsWithVarianceSmaller(double variance) {
+      DeleteColumns(ColumnsWithVarianceSmaller(variance));
+    }
+
+    private void DeleteRows(List<int> rows) {
+      PreprocessingData.InTransaction(() => {
+        foreach (int row in rows.OrderByDescending(x => x)) {
+          PreprocessingData.DeleteRow(row);
+        }
+      });
+    }
+
+    private void DeleteColumns(List<int> columns) {
+      PreprocessingData.InTransaction(() => {
+        foreach (int column in columns.OrderByDescending(x => x)) {
+          PreprocessingData.DeleteColumn(column);
+        }
+      });
+    }
   }
 }