Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2839_HiveProjectManagement/HeuristicLab.Problems.DataAnalysis/3.4/Dataset.cs @ 18236

Last change on this file since 18236 was 16092, checked in by jkarder, 6 years ago

#2839: merged [16057-16091/trunk] into branch

File size: 15.0 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 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 HeuristicLab.Common;
28using HeuristicLab.Core;
29using HeuristicLab.Data;
30using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
31
32namespace HeuristicLab.Problems.DataAnalysis {
33  [Item("Dataset", "Represents a dataset containing data that should be analyzed.")]
34  [StorableClass]
35  public class Dataset : NamedItem, IDataset {
36    [StorableConstructor]
37    protected Dataset(bool deserializing) : base(deserializing) { }
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 ModifiableDataset ToModifiable() {
118      return new ModifiableDataset(variableNames, variableNames.Select(v => variableValues[v]), true);
119    }
120
121    /// <summary>
122    /// Shuffle a dataset's rows
123    /// </summary>
124    /// <param name="random">Random number generator used for shuffling.</param>
125    /// <returns>A shuffled copy of the current dataset.</returns>
126    public Dataset Shuffle(IRandom random) {
127      var values = variableNames.Select(x => variableValues[x]).ToList();
128      return new Dataset(variableNames, values.ShuffleLists(random));
129    }
130
131
132
133    #region Backwards compatible code, remove with 3.5
134    private double[,] storableData;
135    //name alias used to suppport backwards compatibility
136    [Storable(Name = "data", AllowOneWay = true)]
137    private double[,] StorableData { set { storableData = value; } }
138
139    [StorableHook(HookType.AfterDeserialization)]
140    private void AfterDeserialization() {
141      if (variableValues == null) {
142        rows = storableData.GetLength(0);
143        variableValues = new Dictionary<string, IList>();
144        for (int col = 0; col < storableData.GetLength(1); col++) {
145          string columName = variableNames[col];
146          var values = new List<double>(rows);
147          for (int row = 0; row < rows; row++) {
148            values.Add(storableData[row, col]);
149          }
150          variableValues.Add(columName, values);
151        }
152        storableData = null;
153      }
154    }
155    #endregion
156
157    [Storable(Name = "VariableValues")]
158    protected Dictionary<string, IList> variableValues;
159
160    protected List<string> variableNames;
161    [Storable]
162    public IEnumerable<string> VariableNames {
163      get { return variableNames; }
164      protected set {
165        if (variableNames != null) throw new InvalidOperationException();
166        variableNames = new List<string>(value);
167      }
168    }
169    public IEnumerable<string> DoubleVariables {
170      get { return variableValues.Where(p => p.Value is IList<double>).Select(p => p.Key); }
171    }
172
173    public IEnumerable<string> StringVariables {
174      get { return variableValues.Where(p => p.Value is IList<string>).Select(p => p.Key); }
175    }
176
177    public IEnumerable<string> DateTimeVariables {
178      get { return variableValues.Where(p => p.Value is IList<DateTime>).Select(p => p.Key); }
179    }
180
181    public IEnumerable<double> GetDoubleValues(string variableName) {
182      return GetValues<double>(variableName);
183    }
184    public IEnumerable<string> GetStringValues(string variableName) {
185      return GetValues<string>(variableName);
186    }
187    public IEnumerable<DateTime> GetDateTimeValues(string variableName) {
188      return GetValues<DateTime>(variableName);
189    }
190
191    public ReadOnlyCollection<double> GetReadOnlyDoubleValues(string variableName) {
192      var values = GetValues<double>(variableName);
193      return new ReadOnlyCollection<double>(values);
194    }
195    public double GetDoubleValue(string variableName, int row) {
196      var values = GetValues<double>(variableName);
197      return values[row];
198    }
199    public IEnumerable<double> GetDoubleValues(string variableName, IEnumerable<int> rows) {
200      return GetValues<double>(variableName, rows);
201    }
202
203    public string GetStringValue(string variableName, int row) {
204      var values = GetValues<string>(variableName);
205      return values[row];
206    }
207
208    public IEnumerable<string> GetStringValues(string variableName, IEnumerable<int> rows) {
209      return GetValues<string>(variableName, rows);
210    }
211    public ReadOnlyCollection<string> GetReadOnlyStringValues(string variableName) {
212      var values = GetValues<string>(variableName);
213      return new ReadOnlyCollection<string>(values);
214    }
215
216    public DateTime GetDateTimeValue(string variableName, int row) {
217      var values = GetValues<DateTime>(variableName);
218      return values[row];
219    }
220    public IEnumerable<DateTime> GetDateTimeValues(string variableName, IEnumerable<int> rows) {
221      return GetValues<DateTime>(variableName, rows);
222    }
223    public ReadOnlyCollection<DateTime> GetReadOnlyDateTimeValues(string variableName) {
224      var values = GetValues<DateTime>(variableName);
225      return new ReadOnlyCollection<DateTime>(values);
226    }
227    private IEnumerable<T> GetValues<T>(string variableName, IEnumerable<int> rows) {
228      var values = GetValues<T>(variableName);
229      return rows.Select(x => values[x]);
230    }
231    private IList<T> GetValues<T>(string variableName) {
232      IList list;
233      if (!variableValues.TryGetValue(variableName, out list))
234        throw new ArgumentException("The variable " + variableName + " does not exist in the dataset.");
235      IList<T> values = list as IList<T>;
236      if (values == null) throw new ArgumentException("The variable " + variableName + " is not a " + typeof(T) + " variable.");
237      return values;
238    }
239    public bool VariableHasType<T>(string variableName) {
240      return variableValues[variableName] is IList<T>;
241    }
242    protected Type GetVariableType(string variableName) {
243      IList list;
244      variableValues.TryGetValue(variableName, out list);
245      if (list == null)
246        throw new ArgumentException("The variable " + variableName + " does not exist in the dataset.");
247      return GetElementType(list);
248    }
249    protected static Type GetElementType(IList list) {
250      var type = list.GetType();
251      return type.IsGenericType ? type.GetGenericArguments()[0] : type.GetElementType();
252    }
253    protected static bool IsAllowedType(IList list) {
254      var type = GetElementType(list);
255      return IsAllowedType(type);
256    }
257    protected static bool IsAllowedType(Type type) {
258      return type == typeof(double) || type == typeof(string) || type == typeof(DateTime);
259    }
260
261    protected static void CheckArguments(IEnumerable<string> variableNames, IEnumerable<IList> variableValues) {
262      if (variableNames.Count() != variableValues.Count()) {
263        throw new ArgumentException("Number of variable names doesn't match the number of columns of variableValues");
264      } else if (!variableValues.All(list => list.Count == variableValues.First().Count)) {
265        throw new ArgumentException("The number of values must be equal for every variable");
266      } else if (variableNames.Distinct().Count() != variableNames.Count()) {
267        var duplicateVariableNames =
268          variableNames.GroupBy(v => v).Where(g => g.Count() > 1).Select(g => g.Key).ToList();
269        string message = "The dataset cannot contain duplicate variables names: " + Environment.NewLine;
270        foreach (var duplicateVariableName in duplicateVariableNames)
271          message += duplicateVariableName + Environment.NewLine;
272        throw new ArgumentException(message);
273      }
274      // check if all the variables are supported
275      foreach (var t in variableNames.Zip(variableValues, Tuple.Create)) {
276        var variableName = t.Item1;
277        var values = t.Item2;
278
279        if (!IsAllowedType(values)) {
280          throw new ArgumentException(string.Format("Unsupported type {0} for variable {1}.", GetElementType(values), variableName));
281        }
282      }
283    }
284
285    protected static Dictionary<string, IList> CloneValues(Dictionary<string, IList> variableValues) {
286      return variableValues.ToDictionary(x => x.Key, x => CloneValues(x.Value));
287    }
288
289    protected static Dictionary<string, IList> CloneValues(IEnumerable<string> variableNames, IEnumerable<IList> variableValues) {
290      return variableNames.Zip(variableValues, Tuple.Create).ToDictionary(x => x.Item1, x => CloneValues(x.Item2));
291    }
292
293    protected static IList CloneValues(IList values) {
294      var doubleValues = values as IList<double>;
295      if (doubleValues != null) return new List<double>(doubleValues);
296
297      var stringValues = values as IList<string>;
298      if (stringValues != null) return new List<string>(stringValues);
299
300      var dateTimeValues = values as IList<DateTime>;
301      if (dateTimeValues != null) return new List<DateTime>(dateTimeValues);
302
303      throw new ArgumentException(string.Format("Unsupported variable type {0}.", GetElementType(values)));
304    }
305
306    #region IStringConvertibleMatrix Members
307    [Storable]
308    private int rows;
309    public int Rows {
310      get { return rows; }
311      protected set { rows = value; }
312    }
313    int IStringConvertibleMatrix.Rows {
314      get { return Rows; }
315      set { throw new NotSupportedException(); }
316    }
317
318    public int Columns {
319      get { return variableNames.Count; }
320    }
321    int IStringConvertibleMatrix.Columns {
322      get { return Columns; }
323      set { throw new NotSupportedException(); }
324    }
325    bool IStringConvertibleMatrix.SortableView {
326      get { return false; }
327      set { throw new NotSupportedException(); }
328    }
329    bool IStringConvertibleMatrix.ReadOnly {
330      get { return true; }
331    }
332    IEnumerable<string> IStringConvertibleMatrix.ColumnNames {
333      get { return this.VariableNames; }
334      set { throw new NotSupportedException(); }
335    }
336    IEnumerable<string> IStringConvertibleMatrix.RowNames {
337      get { return Enumerable.Empty<string>(); }
338      set { throw new NotSupportedException(); }
339    }
340    string IStringConvertibleMatrix.GetValue(int rowIndex, int columnIndex) {
341      return variableValues[variableNames[columnIndex]][rowIndex].ToString();
342    }
343    bool IStringConvertibleMatrix.SetValue(string value, int rowIndex, int columnIndex) {
344      throw new NotSupportedException();
345    }
346    bool IStringConvertibleMatrix.Validate(string value, out string errorMessage) {
347      throw new NotSupportedException();
348    }
349
350    public virtual event EventHandler ColumnsChanged { add { } remove { } }
351    public virtual event EventHandler RowsChanged { add { } remove { } }
352    public virtual event EventHandler ColumnNamesChanged { add { } remove { } }
353    public virtual event EventHandler RowNamesChanged { add { } remove { } }
354    public virtual event EventHandler SortableViewChanged { add { } remove { } }
355    public virtual event EventHandler<EventArgs<int, int>> ItemChanged { add { } remove { } }
356    public virtual event EventHandler Reset { add { } remove { } }
357    #endregion
358  }
359}
Note: See TracBrowser for help on using the repository browser.