Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/sources/HeuristicLab.DataPreprocessing/3.4/Implementations/TransactionalPreprocessingData.cs @ 11130

Last change on this file since 11130 was 11114, checked in by mkommend, 10 years ago

#2206: Merged data preprocessing into the trunk.

File size: 11.8 KB
RevLine 
[10163]1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2013 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 *
5 * This file is part of HeuristicLab.
6 *
7 * HeuristicLab is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * HeuristicLab is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
15 * GNU General Public License for more details.
16 *
17 * You should have received a copy of the GNU General Public License
18 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
19 */
20#endregion
21
22using System;
[10168]23using System.Collections;
[10163]24using System.Collections.Generic;
[11002]25using System.Globalization;
26using System.Linq;
[10185]27using HeuristicLab.Common;
[10163]28using HeuristicLab.Core;
[10994]29using HeuristicLab.Data;
[10163]30using HeuristicLab.Problems.DataAnalysis;
31
[10182]32namespace HeuristicLab.DataPreprocessing {
[10978]33  [Item("PreprocessingData", "Represents data used for preprocessing.")]
34  public class TransactionalPreprocessingData : PreprocessingData, ITransactionalPreprocessingData {
[10550]35
[10978]36    private class Snapshot {
37      public IList<IList> VariableValues { get; set; }
38      public IList<string> VariableNames { get; set; }
[10550]39
[10994]40      public IntRange TrainingPartition { get; set; }
41      public IntRange TestPartition { get; set; }
[10978]42      public IList<ITransformation> Transformations { get; set; }
43      public DataPreprocessingChangedEventType ChangedType { get; set; }
[10550]44
[10978]45      public int ChangedColumn { get; set; }
46      public int ChangedRow { get; set; }
47    }
[10814]48
[10550]49    private const int MAX_UNDO_DEPTH = 5;
50
[10978]51    private readonly IList<Snapshot> undoHistory = new List<Snapshot>();
52    private readonly Stack<DataPreprocessingChangedEventType> eventStack = new Stack<DataPreprocessingChangedEventType>();
[10550]53
[10805]54    public bool IsInTransaction { get { return eventStack.Count > 0; } }
55
[10609]56    public TransactionalPreprocessingData(IDataAnalysisProblemData problemData)
57      : base(problemData) {
[10185]58    }
[10187]59
[11114]60    protected TransactionalPreprocessingData(TransactionalPreprocessingData original, Cloner cloner)
[10609]61      : base(original, cloner) {
[11068]62    }
[10163]63
[10550]64    private void SaveSnapshot(DataPreprocessingChangedEventType changedType, int column, int row) {
[10978]65      if (IsInTransaction) return;
[10580]66
[10814]67      var currentSnapshot = new Snapshot {
68        VariableValues = CopyVariableValues(variableValues),
69        VariableNames = new List<string>(variableNames),
[10994]70        TrainingPartition = new IntRange(TrainingPartition.Start, TrainingPartition.End),
71        TestPartition = new IntRange(TestPartition.Start, TestPartition.End),
[10814]72        Transformations = new List<ITransformation>(transformations),
73        ChangedType = changedType,
74        ChangedColumn = column,
75        ChangedRow = row
76      };
[10550]77
78      if (undoHistory.Count >= MAX_UNDO_DEPTH)
79        undoHistory.RemoveAt(0);
80
81      undoHistory.Add(currentSnapshot);
82    }
83
[10163]84    #region NamedItem abstract Member Implementations
85
[10185]86    public override IDeepCloneable Clone(Cloner cloner) {
[10586]87      return new TransactionalPreprocessingData(this, cloner);
[10163]88    }
89
90    #endregion
91
[10586]92    #region Overridden IPreprocessingData Members
[10163]93
[11068]94    public override T GetCell<T>(int columnIndex, int rowIndex) {
[10991]95      return (T)variableValues[columnIndex][rowIndex];
96    }
97
[10586]98    public override void SetCell<T>(int columnIndex, int rowIndex, T value) {
[10550]99      SaveSnapshot(DataPreprocessingChangedEventType.ChangeItem, columnIndex, rowIndex);
[10991]100      variableValues[columnIndex][rowIndex] = value;
[10805]101      if (!IsInTransaction)
[10580]102        OnChanged(DataPreprocessingChangedEventType.ChangeItem, columnIndex, rowIndex);
[10367]103    }
104
[11068]105    public override string GetCellAsString(int columnIndex, int rowIndex) {
[10991]106      return variableValues[columnIndex][rowIndex].ToString();
107    }
108
[11068]109    public override string GetVariableName(int columnIndex) {
[10991]110      return variableNames[columnIndex];
111    }
112
[11068]113    public override int GetColumnIndex(string variableName) {
[10991]114      return variableNames.IndexOf(variableName);
115    }
116
[11068]117    public override bool IsType<T>(int columnIndex) {
[10991]118      return variableValues[columnIndex] is List<T>;
119    }
120
121    [Obsolete("use the index based variant, is faster")]
[11068]122    public override IList<T> GetValues<T>(string variableName, bool considerSelection) {
[10991]123      return GetValues<T>(GetColumnIndex(variableName), considerSelection);
124    }
125
[11068]126    public override IList<T> GetValues<T>(int columnIndex, bool considerSelection) {
127      if (considerSelection) {
[10991]128        var list = new List<T>();
[11068]129        foreach (var rowIdx in selection[columnIndex]) {
[10991]130          list.Add((T)variableValues[columnIndex][rowIdx]);
131        }
132        return list;
[11068]133      } else {
[10991]134        return (IList<T>)variableValues[columnIndex];
135      }
136    }
137
[10586]138    public override void SetValues<T>(int columnIndex, IList<T> values) {
139      SaveSnapshot(DataPreprocessingChangedEventType.ChangeColumn, columnIndex, -1);
[11068]140      if (IsType<T>(columnIndex)) {
[10991]141        variableValues[columnIndex] = (IList)values;
[11068]142      } else {
[10991]143        throw new ArgumentException("The datatype of column " + columnIndex + " must be of type " + variableValues[columnIndex].GetType().Name + " but was " + typeof(T).Name);
144      }
[10805]145      if (!IsInTransaction)
[10580]146        OnChanged(DataPreprocessingChangedEventType.ChangeColumn, columnIndex, -1);
[10181]147    }
148
[11002]149    public override bool SetValue(string value, int columnIndex, int rowIndex) {
150      bool valid = false;
151      if (IsType<double>(columnIndex)) {
152        double val;
153        valid = double.TryParse(value, out val);
154        SetValueIfValid(columnIndex, rowIndex, valid, val);
155      } else if (IsType<string>(columnIndex)) {
156        valid = value != null;
157        SetValueIfValid(columnIndex, rowIndex, valid, value);
158      } else if (IsType<DateTime>(columnIndex)) {
159        DateTime date;
160        valid = DateTime.TryParse(value, out date);
161        SetValueIfValid(columnIndex, rowIndex, valid, date);
162      } else {
163        throw new ArgumentException("column " + columnIndex + " contains a non supported type.");
164      }
165
166      if (!IsInTransaction)
167        OnChanged(DataPreprocessingChangedEventType.ChangeColumn, columnIndex, -1);
168
169      return valid;
170    }
[11068]171
172    public override bool Validate(string value, out string errorMessage, int columnIndex) {
173      if (columnIndex < 0 || columnIndex > VariableNames.Count()) {
[11002]174        throw new ArgumentOutOfRangeException("column index is out of range");
175      }
176
177      bool valid = false;
178      errorMessage = string.Empty;
179      if (IsType<double>(columnIndex)) {
180        double val;
181        valid = double.TryParse(value, out val);
182        if (!valid) {
183          errorMessage = "Invalid Value (Valid Value Format: \"" + FormatPatterns.GetDoubleFormatPattern() + "\")";
184        }
185      } else if (IsType<string>(columnIndex)) {
186        valid = value != null;
187        if (!valid) {
188          errorMessage = "Invalid Value (string must not be null)";
189        }
190      } else if (IsType<DateTime>(columnIndex)) {
191        DateTime date;
192        valid = DateTime.TryParse(value, out date);
193        if (!valid) {
194          errorMessage = "Invalid Value (Valid Value Format: \"" + CultureInfo.CurrentCulture.DateTimeFormat + "\"";
195        }
196      } else {
197        throw new ArgumentException("column " + columnIndex + " contains a non supported type.");
198      }
199
200      return valid;
201    }
202
203    private void SetValueIfValid<T>(int columnIndex, int rowIndex, bool valid, T value) {
204      if (valid)
205        SetCell<T>(columnIndex, rowIndex, value);
206    }
207
208    public override bool AreAllStringColumns(IEnumerable<int> columnIndices) {
209      return columnIndices.All(x => IsType<string>(x));
210    }
211
212    public override void DeleteRowsWithIndices(IEnumerable<int> rows) {
[11098]213      SaveSnapshot(DataPreprocessingChangedEventType.AddRow, -1, -1);
214      foreach (int rowIndex in rows.OrderByDescending(x => x)) {
215        foreach (IList column in variableValues) {
216          column.RemoveAt(rowIndex);
217        }
[11002]218      }
[11098]219      if (!IsInTransaction)
220        OnChanged(DataPreprocessingChangedEventType.DeleteRow, -1, -1);
[11002]221    }
222
[10586]223    public override void InsertRow(int rowIndex) {
[10550]224      SaveSnapshot(DataPreprocessingChangedEventType.DeleteRow, -1, rowIndex);
[11068]225      foreach (IList column in variableValues) {
[10991]226        Type type = column.GetType().GetGenericArguments()[0];
227        column.Insert(rowIndex, type.IsValueType ? Activator.CreateInstance(type) : null);
228      }
[10805]229      if (!IsInTransaction)
[10580]230        OnChanged(DataPreprocessingChangedEventType.AddRow, -1, rowIndex);
[10163]231    }
232
[10586]233    public override void DeleteRow(int rowIndex) {
[10550]234      SaveSnapshot(DataPreprocessingChangedEventType.AddRow, -1, rowIndex);
[11068]235      foreach (IList column in variableValues) {
[10991]236        column.RemoveAt(rowIndex);
237      }
[10805]238      if (!IsInTransaction)
[10580]239        OnChanged(DataPreprocessingChangedEventType.DeleteRow, -1, rowIndex);
[10163]240    }
241
[10586]242    public override void InsertColumn<T>(string variableName, int columnIndex) {
[10550]243      SaveSnapshot(DataPreprocessingChangedEventType.DeleteColumn, columnIndex, -1);
[10991]244      variableValues.Insert(columnIndex, new List<T>(Rows));
245      variableNames.Insert(columnIndex, variableName);
[10805]246      if (!IsInTransaction)
[10580]247        OnChanged(DataPreprocessingChangedEventType.AddColumn, columnIndex, -1);
[10163]248    }
249
[10586]250    public override void DeleteColumn(int columnIndex) {
[10550]251      SaveSnapshot(DataPreprocessingChangedEventType.AddColumn, columnIndex, -1);
[10991]252      variableValues.RemoveAt(columnIndex);
253      variableNames.RemoveAt(columnIndex);
[10805]254      if (!IsInTransaction)
[10580]255        OnChanged(DataPreprocessingChangedEventType.DeleteColumn, columnIndex, -1);
[10367]256    }
257
[11068]258    public override Dataset ExportToDataset() {
[10991]259      IList<IList> values = new List<IList>();
260
[11068]261      for (int i = 0; i < Columns; ++i) {
[10991]262        values.Add(variableValues[i]);
263      }
264
265      var dataset = new Dataset(variableNames, values);
266      return dataset;
267    }
268
[11068]269    public override void ClearSelection() {
[10991]270      Selection = new Dictionary<int, IList<int>>();
271    }
272
273    public override event EventHandler SelectionChanged;
274
[11068]275    protected override void OnSelectionChanged() {
[10991]276      var listeners = SelectionChanged;
277      if (listeners != null) listeners(this, EventArgs.Empty);
278    }
279
280
[10586]281    #endregion
[10221]282
[10586]283    #region TransactionalPreprocessingData members
[10221]284
[10547]285    public bool IsUndoAvailable {
[10550]286      get { return undoHistory.Count > 0; }
[10547]287    }
288
289    public void Undo() {
[10550]290      if (IsUndoAvailable) {
[10814]291        Snapshot previousSnapshot = undoHistory[undoHistory.Count - 1];
[10550]292        variableValues = previousSnapshot.VariableValues;
293        variableNames = previousSnapshot.VariableNames;
[10994]294        TrainingPartition = previousSnapshot.TrainingPartition;
295        TestPartition = previousSnapshot.TestPartition;
[10814]296        transformations = previousSnapshot.Transformations;
[10550]297        undoHistory.Remove(previousSnapshot);
298        OnChanged(previousSnapshot.ChangedType,
299          previousSnapshot.ChangedColumn,
300          previousSnapshot.ChangedRow);
301      }
[10547]302    }
303
[10665]304    public void InTransaction(Action action, DataPreprocessingChangedEventType type = DataPreprocessingChangedEventType.Any) {
305      BeginTransaction(type);
[10612]306      action();
307      EndTransaction();
308    }
309
[10665]310    public void BeginTransaction(DataPreprocessingChangedEventType type) {
311      SaveSnapshot(type, -1, -1);
[10805]312      eventStack.Push(type);
[10580]313    }
314
315    public void EndTransaction() {
[10805]316      if (eventStack.Count == 0)
[10580]317        throw new InvalidOperationException("There is no open transaction that can be ended.");
[10814]318
[10805]319      var @event = eventStack.Pop();
320      OnChanged(@event, -1, -1);
[10580]321    }
322
[10220]323    #endregion
[10163]324  }
325}
Note: See TracBrowser for help on using the repository browser.