1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 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 |
|
---|
22 | using System;
|
---|
23 | using System.Collections.Generic;
|
---|
24 | using System.Drawing;
|
---|
25 | using System.Linq;
|
---|
26 | using HeuristicLab.Common;
|
---|
27 | using HeuristicLab.Core;
|
---|
28 | using HeuristicLab.Data;
|
---|
29 | using HEAL.Attic;
|
---|
30 | using HeuristicLab.Random;
|
---|
31 |
|
---|
32 | namespace HeuristicLab.DataPreprocessing {
|
---|
33 | [Item("Data Grid", "Represents a data grid.")]
|
---|
34 | [StorableType("DC6AE5CE-B0FA-4C8C-BDBB-D490C6DE4174")]
|
---|
35 | public class DataGridContent : PreprocessingContent, IStringConvertibleMatrix, IViewShortcut {
|
---|
36 | public static new Image StaticItemImage {
|
---|
37 | get { return HeuristicLab.Common.Resources.VSImageLibrary.Table; }
|
---|
38 | }
|
---|
39 |
|
---|
40 | public int Rows {
|
---|
41 | get { return PreprocessingData.Rows; }
|
---|
42 | set { }
|
---|
43 | }
|
---|
44 |
|
---|
45 | public int Columns {
|
---|
46 | get { return PreprocessingData.Columns; }
|
---|
47 | set { }
|
---|
48 | }
|
---|
49 |
|
---|
50 | public IEnumerable<string> ColumnNames {
|
---|
51 | get { return PreprocessingData.VariableNames; }
|
---|
52 | set { }
|
---|
53 | }
|
---|
54 |
|
---|
55 | public IEnumerable<string> RowNames {
|
---|
56 | get { return Enumerable.Range(1, Rows).Select(n => n.ToString()); }
|
---|
57 | set { throw new NotSupportedException(); }
|
---|
58 | }
|
---|
59 |
|
---|
60 | public bool SortableView {
|
---|
61 | get { return true; }
|
---|
62 | set { throw new NotSupportedException(); }
|
---|
63 | }
|
---|
64 |
|
---|
65 | public bool ReadOnly {
|
---|
66 | get { return false; }
|
---|
67 | }
|
---|
68 |
|
---|
69 | public IDictionary<int, IList<int>> Selection {
|
---|
70 | get { return PreprocessingData.Selection; }
|
---|
71 | set { PreprocessingData.Selection = value; }
|
---|
72 | }
|
---|
73 |
|
---|
74 | #region Constructor, Cloning & Persistence
|
---|
75 | public DataGridContent(IFilteredPreprocessingData preprocessingData)
|
---|
76 | : base(preprocessingData) {
|
---|
77 | }
|
---|
78 |
|
---|
79 | public DataGridContent(DataGridContent original, Cloner cloner)
|
---|
80 | : base(original, cloner) {
|
---|
81 | }
|
---|
82 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
83 | return new DataGridContent(this, cloner);
|
---|
84 | }
|
---|
85 |
|
---|
86 | [StorableConstructor]
|
---|
87 | protected DataGridContent(StorableConstructorFlag _) : base(_) { }
|
---|
88 | #endregion
|
---|
89 |
|
---|
90 | public void DeleteRows(IEnumerable<int> rows) {
|
---|
91 | PreprocessingData.DeleteRowsWithIndices(rows);
|
---|
92 | }
|
---|
93 |
|
---|
94 | public void DeleteColumn(int column) {
|
---|
95 | PreprocessingData.DeleteColumn(column);
|
---|
96 | }
|
---|
97 |
|
---|
98 | public bool Validate(string value, out string errorMessage, int columnIndex) {
|
---|
99 | return PreprocessingData.Validate(value, out errorMessage, columnIndex);
|
---|
100 | }
|
---|
101 |
|
---|
102 | public string GetValue(int rowIndex, int columnIndex) {
|
---|
103 | return PreprocessingData.GetCellAsString(columnIndex, rowIndex);
|
---|
104 | }
|
---|
105 |
|
---|
106 | public bool SetValue(string value, int rowIndex, int columnIndex) {
|
---|
107 | return PreprocessingData.SetValue(value, columnIndex, rowIndex);
|
---|
108 | }
|
---|
109 |
|
---|
110 | public event DataPreprocessingChangedEventHandler Changed {
|
---|
111 | add { PreprocessingData.Changed += value; }
|
---|
112 | remove { PreprocessingData.Changed -= value; }
|
---|
113 | }
|
---|
114 |
|
---|
115 | #region unused stuff/not implemented but necessary due to IStringConvertibleMatrix
|
---|
116 | #pragma warning disable 0067
|
---|
117 | // Is not used since DataGridContentView overrides dataGridView_CellValidating and uses
|
---|
118 | // DataGridLogic#Validate(string value, out string errorMessage, int columnIndex)
|
---|
119 | public bool Validate(string value, out string errorMessage) {
|
---|
120 | errorMessage = string.Empty;
|
---|
121 | return true;
|
---|
122 | }
|
---|
123 |
|
---|
124 | public event EventHandler ColumnsChanged;
|
---|
125 | public event EventHandler RowsChanged;
|
---|
126 | public event EventHandler ColumnNamesChanged;
|
---|
127 | public event EventHandler RowNamesChanged;
|
---|
128 | public event EventHandler SortableViewChanged;
|
---|
129 | public event EventHandler<EventArgs<int, int>> ItemChanged;
|
---|
130 | public event EventHandler Reset;
|
---|
131 | #pragma warning restore 0067
|
---|
132 | #endregion
|
---|
133 |
|
---|
134 | #region Manipulations
|
---|
135 | private void ReplaceIndicesByValue(IDictionary<int, IList<int>> cells, Func<int, double> doubleAggregator = null,
|
---|
136 | Func<int, DateTime> dateTimeAggregator = null, Func<int, string> stringAggregator = null) {
|
---|
137 | PreprocessingData.InTransaction(() => {
|
---|
138 | foreach (var column in cells) {
|
---|
139 | if (doubleAggregator != null && PreprocessingData.VariableHasType<double>(column.Key)) {
|
---|
140 | var value = doubleAggregator(column.Key);
|
---|
141 | foreach (int index in column.Value)
|
---|
142 | PreprocessingData.SetCell<double>(column.Key, index, value);
|
---|
143 | } else if (dateTimeAggregator != null && PreprocessingData.VariableHasType<DateTime>(column.Key)) {
|
---|
144 | var value = dateTimeAggregator(column.Key);
|
---|
145 | foreach (int index in column.Value)
|
---|
146 | PreprocessingData.SetCell<DateTime>(column.Key, index, value);
|
---|
147 | } else if (stringAggregator != null && PreprocessingData.VariableHasType<string>(column.Key)) {
|
---|
148 | var value = stringAggregator(column.Key);
|
---|
149 | foreach (int index in column.Value)
|
---|
150 | PreprocessingData.SetCell<string>(column.Key, index, value);
|
---|
151 | }
|
---|
152 | }
|
---|
153 | });
|
---|
154 | }
|
---|
155 |
|
---|
156 | private void ReplaceIndicesByValues(IDictionary<int, IList<int>> cells, Func<int, IEnumerable<double>> doubleAggregator = null,
|
---|
157 | Func<int, IEnumerable<DateTime>> dateTimeAggregator = null, Func<int, IEnumerable<string>> stringAggregator = null) {
|
---|
158 | PreprocessingData.InTransaction(() => {
|
---|
159 | foreach (var column in cells) {
|
---|
160 | if (doubleAggregator != null && PreprocessingData.VariableHasType<double>(column.Key)) {
|
---|
161 | var values = doubleAggregator(column.Key);
|
---|
162 | foreach (var pair in column.Value.Zip(values, (row, value) => new { row, value }))
|
---|
163 | PreprocessingData.SetCell<double>(column.Key, pair.row, pair.value);
|
---|
164 | } else if (dateTimeAggregator != null && PreprocessingData.VariableHasType<DateTime>(column.Key)) {
|
---|
165 | var values = dateTimeAggregator(column.Key);
|
---|
166 | foreach (var pair in column.Value.Zip(values, (row, value) => new { row, value }))
|
---|
167 | PreprocessingData.SetCell<DateTime>(column.Key, pair.row, pair.value);
|
---|
168 | } else if (stringAggregator != null && PreprocessingData.VariableHasType<string>(column.Key)) {
|
---|
169 | var values = stringAggregator(column.Key);
|
---|
170 | foreach (var pair in column.Value.Zip(values, (row, value) => new { row, value }))
|
---|
171 | PreprocessingData.SetCell<string>(column.Key, pair.row, pair.value);
|
---|
172 | }
|
---|
173 | }
|
---|
174 | });
|
---|
175 | }
|
---|
176 |
|
---|
177 | public void ReplaceIndicesByMean(IDictionary<int, IList<int>> cells, bool considerSelection = false) {
|
---|
178 | ReplaceIndicesByValue(cells,
|
---|
179 | col => PreprocessingData.GetMean<double>(col, considerSelection),
|
---|
180 | col => PreprocessingData.GetMean<DateTime>(col, considerSelection));
|
---|
181 | }
|
---|
182 |
|
---|
183 | public void ReplaceIndicesByMedianValue(IDictionary<int, IList<int>> cells, bool considerSelection = false) {
|
---|
184 | ReplaceIndicesByValue(cells,
|
---|
185 | col => PreprocessingData.GetMedian<double>(col, considerSelection),
|
---|
186 | col => PreprocessingData.GetMedian<DateTime>(col, considerSelection));
|
---|
187 | }
|
---|
188 |
|
---|
189 | public void ReplaceIndicesByMode(IDictionary<int, IList<int>> cells, bool considerSelection = false) {
|
---|
190 | ReplaceIndicesByValue(cells,
|
---|
191 | col => PreprocessingData.GetMode<double>(col, considerSelection),
|
---|
192 | col => PreprocessingData.GetMode<DateTime>(col, considerSelection),
|
---|
193 | col => PreprocessingData.GetMode<string>(col, considerSelection));
|
---|
194 | }
|
---|
195 |
|
---|
196 | public void ReplaceIndicesByRandomValue(IDictionary<int, IList<int>> cells, bool considerSelection = false) {
|
---|
197 | var rand = new FastRandom();
|
---|
198 | ReplaceIndicesByValues(cells,
|
---|
199 | col => {
|
---|
200 | double min = PreprocessingData.GetMin<double>(col, considerSelection);
|
---|
201 | double max = PreprocessingData.GetMax<double>(col, considerSelection);
|
---|
202 | double range = max - min;
|
---|
203 | return cells[col].Select(_ => rand.NextDouble() * range + min);
|
---|
204 | },
|
---|
205 | col => {
|
---|
206 | var min = PreprocessingData.GetMin<DateTime>(col, considerSelection);
|
---|
207 | var max = PreprocessingData.GetMax<DateTime>(col, considerSelection);
|
---|
208 | double range = (max - min).TotalSeconds;
|
---|
209 | return cells[col].Select(_ => min + TimeSpan.FromSeconds(rand.NextDouble() * range));
|
---|
210 | });
|
---|
211 | }
|
---|
212 |
|
---|
213 | public void ReplaceIndicesByString(IDictionary<int, IList<int>> cells, string value) {
|
---|
214 | PreprocessingData.InTransaction(() => {
|
---|
215 | foreach (var column in cells) {
|
---|
216 | foreach (var rowIdx in column.Value) {
|
---|
217 | PreprocessingData.SetValue(value, column.Key, rowIdx);
|
---|
218 | }
|
---|
219 | }
|
---|
220 | });
|
---|
221 | }
|
---|
222 |
|
---|
223 |
|
---|
224 | public void ReplaceIndicesByLinearInterpolationOfNeighbours(IDictionary<int, IList<int>> cells) {
|
---|
225 | PreprocessingData.InTransaction(() => {
|
---|
226 | foreach (var column in cells) {
|
---|
227 | IList<Tuple<int, int>> startEndings = GetStartAndEndingsForInterpolation(column);
|
---|
228 | foreach (var tuple in startEndings) {
|
---|
229 | Interpolate(column, tuple.Item1, tuple.Item2);
|
---|
230 | }
|
---|
231 | }
|
---|
232 | });
|
---|
233 | }
|
---|
234 |
|
---|
235 | private List<Tuple<int, int>> GetStartAndEndingsForInterpolation(KeyValuePair<int, IList<int>> column) {
|
---|
236 | var startEndings = new List<Tuple<int, int>>();
|
---|
237 | var rowIndices = column.Value.OrderBy(x => x).ToList();
|
---|
238 | var count = rowIndices.Count;
|
---|
239 | int start = int.MinValue;
|
---|
240 | for (int i = 0; i < count; ++i) {
|
---|
241 | if (start == int.MinValue) {
|
---|
242 | start = IndexOfPrevPresentValue(column.Key, rowIndices[i]);
|
---|
243 | }
|
---|
244 | if (i + 1 == count || (i + 1 < count && rowIndices[i + 1] - rowIndices[i] > 1)) {
|
---|
245 | int next = IndexOfNextPresentValue(column.Key, rowIndices[i]);
|
---|
246 | if (start > 0 && next < PreprocessingData.Rows) {
|
---|
247 | startEndings.Add(new Tuple<int, int>(start, next));
|
---|
248 | }
|
---|
249 | start = int.MinValue;
|
---|
250 | }
|
---|
251 | }
|
---|
252 | return startEndings;
|
---|
253 | }
|
---|
254 |
|
---|
255 | private void Interpolate(KeyValuePair<int, IList<int>> column, int prevIndex, int nextIndex) {
|
---|
256 | int valuesToInterpolate = nextIndex - prevIndex;
|
---|
257 |
|
---|
258 | if (PreprocessingData.VariableHasType<double>(column.Key)) {
|
---|
259 | double prev = PreprocessingData.GetCell<double>(column.Key, prevIndex);
|
---|
260 | double next = PreprocessingData.GetCell<double>(column.Key, nextIndex);
|
---|
261 | double interpolationStep = (next - prev) / valuesToInterpolate;
|
---|
262 |
|
---|
263 | for (int i = prevIndex; i < nextIndex; ++i) {
|
---|
264 | double interpolated = prev + (interpolationStep * (i - prevIndex));
|
---|
265 | PreprocessingData.SetCell<double>(column.Key, i, interpolated);
|
---|
266 | }
|
---|
267 | } else if (PreprocessingData.VariableHasType<DateTime>(column.Key)) {
|
---|
268 | DateTime prev = PreprocessingData.GetCell<DateTime>(column.Key, prevIndex);
|
---|
269 | DateTime next = PreprocessingData.GetCell<DateTime>(column.Key, nextIndex);
|
---|
270 | double interpolationStep = (next - prev).TotalSeconds / valuesToInterpolate;
|
---|
271 |
|
---|
272 | for (int i = prevIndex; i < nextIndex; ++i) {
|
---|
273 | DateTime interpolated = prev.AddSeconds(interpolationStep * (i - prevIndex));
|
---|
274 | PreprocessingData.SetCell<DateTime>(column.Key, i, interpolated);
|
---|
275 | }
|
---|
276 | }
|
---|
277 | }
|
---|
278 |
|
---|
279 | private int IndexOfPrevPresentValue(int columnIndex, int start) {
|
---|
280 | int offset = start - 1;
|
---|
281 | while (offset >= 0 && PreprocessingData.IsCellEmpty(columnIndex, offset)) {
|
---|
282 | offset--;
|
---|
283 | }
|
---|
284 |
|
---|
285 | return offset;
|
---|
286 | }
|
---|
287 |
|
---|
288 | private int IndexOfNextPresentValue(int columnIndex, int start) {
|
---|
289 | int offset = start + 1;
|
---|
290 | while (offset < PreprocessingData.Rows && PreprocessingData.IsCellEmpty(columnIndex, offset)) {
|
---|
291 | offset++;
|
---|
292 | }
|
---|
293 |
|
---|
294 | return offset;
|
---|
295 | }
|
---|
296 |
|
---|
297 | public void Shuffle(bool shuffleRangesSeparately) {
|
---|
298 | var random = new FastRandom();
|
---|
299 |
|
---|
300 | if (shuffleRangesSeparately) {
|
---|
301 | var ranges = new[] { PreprocessingData.TestPartition, PreprocessingData.TrainingPartition };
|
---|
302 | PreprocessingData.InTransaction(() => {
|
---|
303 | // process all given ranges - e.g. TrainingPartition, TestPartition
|
---|
304 | foreach (IntRange range in ranges) {
|
---|
305 | var indices = Enumerable.Range(0, PreprocessingData.Rows).ToArray();
|
---|
306 | var shuffledIndices = Enumerable.Range(range.Start, range.Size).Shuffle(random).ToArray();
|
---|
307 | for (int i = range.Start, j = 0; i < range.End; i++, j++)
|
---|
308 | indices[i] = shuffledIndices[j];
|
---|
309 |
|
---|
310 | ReOrderToIndices(indices);
|
---|
311 | }
|
---|
312 | });
|
---|
313 |
|
---|
314 | } else {
|
---|
315 | PreprocessingData.InTransaction(() => {
|
---|
316 | var indices = Enumerable.Range(0, PreprocessingData.Rows).ToArray();
|
---|
317 | indices.ShuffleInPlace(random);
|
---|
318 | ReOrderToIndices(indices);
|
---|
319 | });
|
---|
320 | }
|
---|
321 | }
|
---|
322 |
|
---|
323 | public void ReOrderToIndices(int[] indices) {
|
---|
324 | PreprocessingData.InTransaction(() => {
|
---|
325 | for (int i = 0; i < PreprocessingData.Columns; ++i) {
|
---|
326 | if (PreprocessingData.VariableHasType<double>(i))
|
---|
327 | ReOrderToIndices<double>(i, indices);
|
---|
328 | else if (PreprocessingData.VariableHasType<string>(i))
|
---|
329 | ReOrderToIndices<string>(i, indices);
|
---|
330 | else if (PreprocessingData.VariableHasType<DateTime>(i))
|
---|
331 | ReOrderToIndices<DateTime>(i, indices);
|
---|
332 | }
|
---|
333 | });
|
---|
334 | }
|
---|
335 |
|
---|
336 | private void ReOrderToIndices<T>(int columnIndex, int[] indices) {
|
---|
337 | var originalData = new List<T>(PreprocessingData.GetValues<T>(columnIndex));
|
---|
338 | if (indices.Length != originalData.Count) throw new InvalidOperationException("The number of provided indices does not match the values.");
|
---|
339 |
|
---|
340 | for (int i = 0; i < indices.Length; i++) {
|
---|
341 | T newValue = originalData[indices[i]];
|
---|
342 | PreprocessingData.SetCell<T>(columnIndex, i, newValue);
|
---|
343 | }
|
---|
344 | }
|
---|
345 | #endregion
|
---|
346 | }
|
---|
347 | }
|
---|