source: branches/3040_VectorBasedGP/HeuristicLab.Problems.DataAnalysis/3.4/Dataset.cs @ 17364

Last change on this file since 17364 was 17364, checked in by pfleck, 11 months ago

#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.
File size: 18.8 KB
Line 
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
22using System;
23using System.Collections;
24using System.Collections.Generic;
25using System.Collections.ObjectModel;
26using System.Linq;
27using HEAL.Attic;
28using HeuristicLab.Common;
29using HeuristicLab.Core;
30using HeuristicLab.Data;
31
32namespace HeuristicLab.Problems.DataAnalysis {
33  [Item("Dataset", "Represents a dataset containing data that should be analyzed.")]
34  [StorableType("49F4D145-50D7-4497-8D8A-D190CD556CC8")]
35  public class Dataset : NamedItem, IDataset {
36    [StorableConstructor]
37    protected Dataset(StorableConstructorFlag _) : base(_) { }
38    protected Dataset(Dataset original, Cloner cloner)
39      : base(original, cloner) {
40      // no need to clone the variable values because these can't be modified
41      variableValues = new Dictionary<string, IList>(original.variableValues);
42      variableNames = new List<string>(original.variableNames);
43      rows = original.rows;
44    }
45
46    public override IDeepCloneable Clone(Cloner cloner) { return new Dataset(this, cloner); }
47
48    public Dataset()
49      : base() {
50      Name = "-";
51      VariableNames = Enumerable.Empty<string>();
52      variableValues = new Dictionary<string, IList>();
53      rows = 0;
54    }
55
56    /// <summary>
57    /// Creates a new dataset. The variableValues are not cloned.
58    /// </summary>
59    /// <param name="variableNames">The names of the variables in the dataset</param>
60    /// <param name="variableValues">The values for the variables (column-oriented storage). Values are not cloned!</param>
61    public Dataset(IEnumerable<string> variableNames, IEnumerable<IList> variableValues)
62      : this(variableNames, variableValues, cloneValues: true) {
63    }
64
65    protected Dataset(IEnumerable<string> variableNames, IEnumerable<IList> variableValues, bool cloneValues = false) {
66      Name = "-";
67
68      if (variableNames.Any()) {
69        this.variableNames = new List<string>(variableNames);
70      } else {
71        this.variableNames = Enumerable.Range(0, variableValues.Count()).Select(x => "Column " + x).ToList();
72      }
73      // check if the arguments are consistent (no duplicate variables, same number of rows, correct data types, ...)
74      CheckArguments(this.variableNames, variableValues);
75
76      rows = variableValues.First().Count;
77
78      if (cloneValues) {
79        this.variableValues = CloneValues(this.variableNames, variableValues);
80      } else {
81        this.variableValues = new Dictionary<string, IList>(this.variableNames.Count);
82        for (int i = 0; i < this.variableNames.Count; i++) {
83          var variableName = this.variableNames[i];
84          var values = variableValues.ElementAt(i);
85          this.variableValues.Add(variableName, values);
86        }
87      }
88    }
89
90    public Dataset(IEnumerable<string> variableNames, double[,] variableValues) {
91      Name = "-";
92      if (variableNames.Count() != variableValues.GetLength(1)) {
93        throw new ArgumentException("Number of variable names doesn't match the number of columns of variableValues");
94      }
95      if (variableNames.Distinct().Count() != variableNames.Count()) {
96        var duplicateVariableNames = variableNames.GroupBy(v => v).Where(g => g.Count() > 1).Select(g => g.Key).ToList();
97        string message = "The dataset cannot contain duplicate variables names: " + Environment.NewLine;
98        foreach (var duplicateVariableName in duplicateVariableNames)
99          message += duplicateVariableName + Environment.NewLine;
100        throw new ArgumentException(message);
101      }
102
103      rows = variableValues.GetLength(0);
104      this.variableNames = new List<string>(variableNames);
105
106      this.variableValues = new Dictionary<string, IList>(variableValues.GetLength(1));
107      for (int col = 0; col < variableValues.GetLength(1); col++) {
108        string columName = this.variableNames[col];
109        var values = new List<double>(variableValues.GetLength(0));
110        for (int row = 0; row < variableValues.GetLength(0); row++) {
111          values.Add(variableValues[row, col]);
112        }
113        this.variableValues.Add(columName, values);
114      }
115    }
116
117    public static Dataset FromRowData(IEnumerable<string> variableNames, double[,] data) {
118      var colWise = new List<IList>(data.GetLength(1));
119      for (var col = 0; col < data.GetLength(1); col++) {
120        var column = new List<double>(data.GetLength(0));
121        for (var row = 0; row < data.GetLength(0); row++) {
122          column.Add(data[row, col]);
123        }
124        colWise.Add(column);
125      }
126      return new Dataset(variableNames, colWise);
127    }
128
129    public static Dataset FromRowData(IEnumerable<string> variableNames, IEnumerable<IList> data) {
130      var vnames = variableNames.ToList();
131      var transposed = new List<IList>();
132      var iter = data.GetEnumerator();
133      if (!iter.MoveNext()) throw new ArgumentException("Data does not contain any rows", nameof(data));
134      for (var i = 0; i < iter.Current.Count; i++) {
135        if (i >= vnames.Count) throw new ArgumentException("There are more variables in data, than variable names.", nameof(variableNames));
136        if (iter.Current[i] == null) throw new ArgumentException("Null values are not supported.", nameof(data));
137        if (!IsAllowedType(iter.Current[i].GetType())) throw new ArgumentException("Data contains types that are not allowed.", nameof(data));
138        if (iter.Current[i] is double d)
139          transposed.Add(new List<double>() { d });
140        else if (iter.Current[i] is DateTime dt)
141          transposed.Add(new List<DateTime>() { dt });
142        else if (iter.Current[i] is string s)
143          transposed.Add(new List<string>() { s });
144        else if (iter.Current[i] is IReadOnlyList<double> dv)
145          transposed.Add(new List<IReadOnlyList<double>>() { dv });
146        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()));
147      }
148      if (transposed.Count < vnames.Count) throw new ArgumentException("There are less variables in data, than variable names.", nameof(variableNames));
149      while (iter.MoveNext()) {
150        for (var i = 0; i < iter.Current.Count; i++)
151          if (transposed[i].Add(iter.Current[i]) < 0)
152            throw new ArgumentException(string.Format("Variable {0} has invalid value ({1})", vnames[i], iter.Current[i]), nameof(data));
153      }
154      return new Dataset(vnames, transposed);
155    }
156
157    public ModifiableDataset ToModifiable() {
158      return new ModifiableDataset(variableNames, variableNames.Select(v => variableValues[v]), true);
159    }
160
161    /// <summary>
162    /// Shuffle a dataset's rows
163    /// </summary>
164    /// <param name="random">Random number generator used for shuffling.</param>
165    /// <returns>A shuffled copy of the current dataset.</returns>
166    public Dataset Shuffle(IRandom random) {
167      var values = variableNames.Select(x => variableValues[x]).ToList();
168      return new Dataset(variableNames, values.ShuffleLists(random));
169    }
170
171
172
173    #region Backwards compatible code, remove with 3.5
174    private double[,] storableData;
175    //name alias used to support backwards compatibility
176    [Storable(OldName = "data")]
177    private double[,] StorableData { set { storableData = value; } }
178
179    [StorableHook(HookType.AfterDeserialization)]
180    private void AfterDeserialization() {
181      if (variableValues == null) {
182        rows = storableData.GetLength(0);
183        variableValues = new Dictionary<string, IList>();
184        for (int col = 0; col < storableData.GetLength(1); col++) {
185          string columName = variableNames[col];
186          var values = new List<double>(rows);
187          for (int row = 0; row < rows; row++) {
188            values.Add(storableData[row, col]);
189          }
190          variableValues.Add(columName, values);
191        }
192        storableData = null;
193      }
194    }
195    #endregion
196
197    [Storable(Name = "VariableValues")]
198    protected Dictionary<string, IList> variableValues;
199
200    protected List<string> variableNames;
201    [Storable]
202    public IEnumerable<string> VariableNames {
203      get { return variableNames; }
204      protected set {
205        if (variableNames != null) throw new InvalidOperationException();
206        variableNames = new List<string>(value);
207      }
208    }
209
210    public bool ContainsVariable(string variableName) {
211      return variableValues.ContainsKey(variableName);
212    }
213    public IEnumerable<string> DoubleVariables {
214      get { return variableValues.Where(p => p.Value is IList<double>).Select(p => p.Key); }
215    }
216    public IEnumerable<string> StringVariables {
217      get { return variableValues.Where(p => p.Value is IList<string>).Select(p => p.Key); }
218    }
219    public IEnumerable<string> DateTimeVariables {
220      get { return variableValues.Where(p => p.Value is IList<DateTime>).Select(p => p.Key); }
221    }
222    public IEnumerable<string> DoubleVectorVariables {
223      get { return variableValues.Where(p => p.Value is IList<IReadOnlyList<double>>).Select(p => p.Key); }
224    }
225
226    public IEnumerable<double> GetDoubleValues(string variableName) {
227      return GetValues<double>(variableName);
228    }
229    public IEnumerable<string> GetStringValues(string variableName) {
230      return GetValues<string>(variableName);
231    }
232    public IEnumerable<DateTime> GetDateTimeValues(string variableName) {
233      return GetValues<DateTime>(variableName);
234    }
235    public IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName) {
236      return GetValues<IReadOnlyList<double>>(variableName);
237    }
238
239    public ReadOnlyCollection<double> GetReadOnlyDoubleValues(string variableName) {
240      var values = GetValues<double>(variableName);
241      return new ReadOnlyCollection<double>(values);
242    }
243    public double GetDoubleValue(string variableName, int row) {
244      var values = GetValues<double>(variableName);
245      return values[row];
246    }
247    public IEnumerable<double> GetDoubleValues(string variableName, IEnumerable<int> rows) {
248      return GetValues<double>(variableName, rows);
249    }
250
251    public string GetStringValue(string variableName, int row) {
252      var values = GetValues<string>(variableName);
253      return values[row];
254    }
255    public IEnumerable<string> GetStringValues(string variableName, IEnumerable<int> rows) {
256      return GetValues<string>(variableName, rows);
257    }
258    public ReadOnlyCollection<string> GetReadOnlyStringValues(string variableName) {
259      var values = GetValues<string>(variableName);
260      return new ReadOnlyCollection<string>(values);
261    }
262
263    public DateTime GetDateTimeValue(string variableName, int row) {
264      var values = GetValues<DateTime>(variableName);
265      return values[row];
266    }
267    public IEnumerable<DateTime> GetDateTimeValues(string variableName, IEnumerable<int> rows) {
268      return GetValues<DateTime>(variableName, rows);
269    }
270    public ReadOnlyCollection<DateTime> GetReadOnlyDateTimeValues(string variableName) {
271      var values = GetValues<DateTime>(variableName);
272      return new ReadOnlyCollection<DateTime>(values);
273    }
274
275    public IReadOnlyList<double> GetDoubleVectorValue(string variableName, int row) {
276      var values = GetValues<IReadOnlyList<double>>(variableName);
277      return values[row];
278    }
279    public IEnumerable<IReadOnlyList<double>> GetDoubleVectorValues(string variableName, IEnumerable<int> rows) {
280      return GetValues<IReadOnlyList<double>>(variableName, rows);
281    }
282    public ReadOnlyCollection<IReadOnlyList<double>> GetReadOnlyDoubleVectorValues(string variableName) {
283      var values = GetValues<IReadOnlyList<double>>(variableName);
284      return new ReadOnlyCollection<IReadOnlyList<double>>(values);
285    }
286
287    private IEnumerable<T> GetValues<T>(string variableName, IEnumerable<int> rows) {
288      var values = GetValues<T>(variableName);
289      return rows.Select(x => values[x]);
290    }
291    private IList<T> GetValues<T>(string variableName) {
292      IList list;
293      if (!variableValues.TryGetValue(variableName, out list))
294        throw new ArgumentException("The variable " + variableName + " does not exist in the dataset.");
295      IList<T> values = list as IList<T>;
296      if (values == null) throw new ArgumentException("The variable " + variableName + " is not a " + typeof(T) + " variable.");
297      return values;
298    }
299    public bool VariableHasType<T>(string variableName) {
300      return variableValues[variableName] is IList<T>;
301    }
302    protected Type GetVariableType(string variableName) {
303      IList list;
304      variableValues.TryGetValue(variableName, out list);
305      if (list == null)
306        throw new ArgumentException("The variable " + variableName + " does not exist in the dataset.");
307      return GetElementType(list);
308    }
309    protected static Type GetElementType(IList list) {
310      var type = list.GetType();
311      return type.IsGenericType ? type.GetGenericArguments()[0] : type.GetElementType();
312    }
313    protected static bool IsAllowedType(IList list) {
314      var type = GetElementType(list);
315      return IsAllowedType(type);
316    }
317    protected static bool IsAllowedType(Type type) {
318      return type == typeof(double) || type == typeof(string) || type == typeof(DateTime) || type == typeof(IReadOnlyList<double>);
319    }
320
321    protected static void CheckArguments(IEnumerable<string> variableNames, IEnumerable<IList> variableValues) {
322      if (variableNames.Count() != variableValues.Count()) {
323        throw new ArgumentException("Number of variable names doesn't match the number of columns of variableValues");
324      } else if (!variableValues.All(list => list.Count == variableValues.First().Count)) {
325        throw new ArgumentException("The number of values must be equal for every variable");
326      } else if (variableNames.Distinct().Count() != variableNames.Count()) {
327        var duplicateVariableNames =
328          variableNames.GroupBy(v => v).Where(g => g.Count() > 1).Select(g => g.Key).ToList();
329        string message = "The dataset cannot contain duplicate variables names: " + Environment.NewLine;
330        foreach (var duplicateVariableName in duplicateVariableNames)
331          message += duplicateVariableName + Environment.NewLine;
332        throw new ArgumentException(message);
333      }
334      // check if all the variables are supported
335      foreach (var t in variableNames.Zip(variableValues, Tuple.Create)) {
336        var variableName = t.Item1;
337        var values = t.Item2;
338
339        if (!IsAllowedType(values)) {
340          throw new ArgumentException(string.Format("Unsupported type {0} for variable {1}.", GetElementType(values), variableName));
341        }
342      }
343    }
344
345    protected static Dictionary<string, IList> CloneValues(Dictionary<string, IList> variableValues) {
346      return variableValues.ToDictionary(x => x.Key, x => CloneValues(x.Value));
347    }
348
349    protected static Dictionary<string, IList> CloneValues(IEnumerable<string> variableNames, IEnumerable<IList> variableValues) {
350      return variableNames.Zip(variableValues, Tuple.Create).ToDictionary(x => x.Item1, x => CloneValues(x.Item2));
351    }
352
353    protected static IList CloneValues(IList values) {
354      var doubleValues = values as IList<double>;
355      if (doubleValues != null) return new List<double>(doubleValues);
356
357      var stringValues = values as IList<string>;
358      if (stringValues != null) return new List<string>(stringValues);
359
360      var dateTimeValues = values as IList<DateTime>;
361      if (dateTimeValues != null) return new List<DateTime>(dateTimeValues);
362
363      var doubleVectorValues = values as IList<IReadOnlyList<double>>;
364      if (doubleVectorValues != null) return doubleVectorValues.Select(x => new List<double>(x)).Cast<IReadOnlyList<double>>().ToList();
365
366      throw new ArgumentException(string.Format("Unsupported variable type {0}.", GetElementType(values)));
367    }
368
369    #region IStringConvertibleMatrix Members
370    [Storable]
371    private int rows;
372    public int Rows {
373      get { return rows; }
374      protected set { rows = value; }
375    }
376    int IStringConvertibleMatrix.Rows {
377      get { return Rows; }
378      set { throw new NotSupportedException(); }
379    }
380
381    public int Columns {
382      get { return variableNames.Count; }
383    }
384    int IStringConvertibleMatrix.Columns {
385      get { return Columns; }
386      set { throw new NotSupportedException(); }
387    }
388    bool IStringConvertibleMatrix.SortableView {
389      get { return false; }
390      set { throw new NotSupportedException(); }
391    }
392    bool IStringConvertibleMatrix.ReadOnly {
393      get { return true; }
394    }
395    IEnumerable<string> IStringConvertibleMatrix.ColumnNames {
396      get { return this.VariableNames; }
397      set { throw new NotSupportedException(); }
398    }
399    IEnumerable<string> IStringConvertibleMatrix.RowNames {
400      get { return Enumerable.Empty<string>(); }
401      set { throw new NotSupportedException(); }
402    }
403    string IStringConvertibleMatrix.GetValue(int rowIndex, int columnIndex) {
404      var value = variableValues[variableNames[columnIndex]][rowIndex];
405      if (value is IReadOnlyList<double> doubleVector)
406        return $"[{string.Join(", ", doubleVector)}]";
407      return variableValues[variableNames[columnIndex]][rowIndex].ToString();
408    }
409    bool IStringConvertibleMatrix.SetValue(string value, int rowIndex, int columnIndex) {
410      throw new NotSupportedException();
411    }
412    bool IStringConvertibleMatrix.Validate(string value, out string errorMessage) {
413      throw new NotSupportedException();
414    }
415
416    public virtual event EventHandler ColumnsChanged { add { } remove { } }
417    public virtual event EventHandler RowsChanged { add { } remove { } }
418    public virtual event EventHandler ColumnNamesChanged { add { } remove { } }
419    public virtual event EventHandler RowNamesChanged { add { } remove { } }
420    public virtual event EventHandler SortableViewChanged { add { } remove { } }
421    public virtual event EventHandler<EventArgs<int, int>> ItemChanged { add { } remove { } }
422    public virtual event EventHandler Reset { add { } remove { } }
423    #endregion
424  }
425}
Note: See TracBrowser for help on using the repository browser.