[2] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[3253] | 3 | * Copyright (C) 2002-2010 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[2] | 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.Xml;
|
---|
| 25 | using System.Globalization;
|
---|
| 26 | using System.Text;
|
---|
[2285] | 27 | using System.Linq;
|
---|
[3376] | 28 | using HeuristicLab.Common;
|
---|
[3253] | 29 | using HeuristicLab.Core;
|
---|
| 30 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 31 | using HeuristicLab.Data;
|
---|
[2] | 32 |
|
---|
[3253] | 33 | namespace HeuristicLab.Problems.DataAnalysis {
|
---|
| 34 | [Item("Dataset", "Represents a dataset containing data that should be analyzed.")]
|
---|
| 35 | [StorableClass]
|
---|
| 36 | public sealed class Dataset : NamedItem, IStringConvertibleMatrix {
|
---|
[2319] | 37 | public Dataset()
|
---|
[3458] | 38 | : base() {
|
---|
| 39 | this.Name = string.Empty;
|
---|
| 40 | this.data = new double[0, 0];
|
---|
| 41 | this.variableNames = new string[0];
|
---|
| 42 | this.SortableView = false;
|
---|
[2319] | 43 | }
|
---|
[2] | 44 |
|
---|
[3264] | 45 | public Dataset(IEnumerable<string> variableNames, double[,] data)
|
---|
[3458] | 46 | : this() {
|
---|
[2319] | 47 | Name = "-";
|
---|
[3264] | 48 | if (variableNames.Count() != data.GetLength(1)) {
|
---|
| 49 | throw new ArgumentException("Number of variable names doesn't match the number of columns of data");
|
---|
| 50 | }
|
---|
[3442] | 51 | this.data = data;
|
---|
| 52 | this.variableNames = variableNames.ToArray();
|
---|
[2038] | 53 | }
|
---|
| 54 |
|
---|
[3294] | 55 | [Storable]
|
---|
[3308] | 56 | private string[] variableNames;
|
---|
[2319] | 57 | public IEnumerable<string> VariableNames {
|
---|
| 58 | get { return variableNames; }
|
---|
[3330] | 59 | private set {
|
---|
[3321] | 60 | if (variableNames != value) {
|
---|
| 61 | variableNames = value.ToArray();
|
---|
| 62 | OnColumnNamesChanged();
|
---|
| 63 | }
|
---|
| 64 | }
|
---|
[333] | 65 | }
|
---|
| 66 |
|
---|
[3294] | 67 | [Storable]
|
---|
[3308] | 68 | private double[,] data;
|
---|
| 69 | private double[,] Data {
|
---|
[3253] | 70 | get { return data; }
|
---|
[237] | 71 | set {
|
---|
[3253] | 72 | if (data != value) {
|
---|
| 73 | if (value == null) throw new ArgumentNullException();
|
---|
| 74 | this.data = value;
|
---|
| 75 | OnReset(EventArgs.Empty);
|
---|
[237] | 76 | }
|
---|
[2] | 77 | }
|
---|
| 78 | }
|
---|
| 79 |
|
---|
[3253] | 80 | // elementwise access
|
---|
| 81 | public double this[int rowIndex, int columnIndex] {
|
---|
| 82 | get { return data[rowIndex, columnIndex]; }
|
---|
[2319] | 83 | set {
|
---|
[3253] | 84 | if (!value.Equals(data[rowIndex, columnIndex])) {
|
---|
| 85 | data[rowIndex, columnIndex] = value;
|
---|
| 86 | OnDataChanged(new EventArgs<int, int>(rowIndex, columnIndex));
|
---|
[3321] | 87 | OnItemChanged(rowIndex, columnIndex);
|
---|
[3253] | 88 | }
|
---|
[2285] | 89 | }
|
---|
[2] | 90 | }
|
---|
[3839] | 91 | public double this[string variableName, int rowIndex] {
|
---|
| 92 | get {
|
---|
| 93 | int columnIndex = GetVariableIndex(variableName);
|
---|
| 94 | return data[rowIndex, columnIndex];
|
---|
| 95 | }
|
---|
| 96 | set {
|
---|
| 97 | int columnIndex = GetVariableIndex(variableName);
|
---|
| 98 | if (!value.Equals(data[rowIndex, columnIndex])) {
|
---|
| 99 | data[rowIndex, columnIndex] = value;
|
---|
| 100 | OnDataChanged(new EventArgs<int, int>(rowIndex, columnIndex));
|
---|
| 101 | OnItemChanged(rowIndex, columnIndex);
|
---|
| 102 | }
|
---|
| 103 | }
|
---|
| 104 | }
|
---|
[3253] | 105 | // access to full columns
|
---|
| 106 | public double[] this[string variableName] {
|
---|
[3308] | 107 | get { return GetVariableValues(GetVariableIndex(variableName), 0, Rows); }
|
---|
[1287] | 108 | }
|
---|
| 109 |
|
---|
[3294] | 110 | public double[] GetVariableValues(int variableIndex, int start, int end) {
|
---|
[2311] | 111 | if (start < 0 || !(start <= end))
|
---|
| 112 | throw new ArgumentException("Start must be between 0 and end (" + end + ").");
|
---|
[3308] | 113 | if (end > Rows || end < start)
|
---|
| 114 | throw new ArgumentException("End must be between start (" + start + ") and dataset rows (" + Rows + ").");
|
---|
[2311] | 115 |
|
---|
| 116 | double[] values = new double[end - start];
|
---|
| 117 | for (int i = 0; i < end - start; i++)
|
---|
[3253] | 118 | values[i] = data[i + start, variableIndex];
|
---|
[2311] | 119 | return values;
|
---|
| 120 | }
|
---|
| 121 |
|
---|
[3294] | 122 | public double[] GetVariableValues(string variableName, int start, int end) {
|
---|
| 123 | return GetVariableValues(GetVariableIndex(variableName), start, end);
|
---|
[2311] | 124 | }
|
---|
| 125 |
|
---|
[2319] | 126 | #region Variable name methods
|
---|
[3294] | 127 | public string GetVariableName(int variableIndex) {
|
---|
[2319] | 128 | return variableNames[variableIndex];
|
---|
| 129 | }
|
---|
| 130 |
|
---|
[3294] | 131 | public int GetVariableIndex(string variableName) {
|
---|
[2319] | 132 | for (int i = 0; i < variableNames.Length; i++) {
|
---|
| 133 | if (variableNames[i].Equals(variableName)) return i;
|
---|
| 134 | }
|
---|
| 135 | throw new ArgumentException("The variable name " + variableName + " was not found.");
|
---|
| 136 | }
|
---|
| 137 |
|
---|
[1287] | 138 | public void SetVariableName(int variableIndex, string name) {
|
---|
[3308] | 139 | if (name == null) throw new ArgumentNullException("Cannot set variable name to null for variable at index " + variableIndex + " variableIndex");
|
---|
[3294] | 140 | if (variableNames.Contains(name)) throw new ArgumentException("The data set already contains a variable with name " + name + ".");
|
---|
[3308] | 141 | if (variableIndex < 0 || variableIndex >= variableNames.Length) throw new ArgumentException(" Cannot set name of not existent variable at index " + variableIndex + ".");
|
---|
[1287] | 142 | variableNames[variableIndex] = name;
|
---|
[3321] | 143 | OnColumnNamesChanged();
|
---|
[1287] | 144 | }
|
---|
| 145 |
|
---|
[2319] | 146 | #endregion
|
---|
[2310] | 147 |
|
---|
[3253] | 148 | #region variable statistics
|
---|
[3294] | 149 | public double GetMean(string variableName) {
|
---|
| 150 | return GetMean(GetVariableIndex(variableName));
|
---|
[2] | 151 | }
|
---|
| 152 |
|
---|
[3294] | 153 | public double GetMean(string variableName, int start, int end) {
|
---|
| 154 | return GetMean(GetVariableIndex(variableName), start, end);
|
---|
[2319] | 155 | }
|
---|
| 156 |
|
---|
[3294] | 157 | public double GetMean(int variableIndex) {
|
---|
[3308] | 158 | return GetMean(variableIndex, 0, Rows);
|
---|
[2319] | 159 | }
|
---|
| 160 |
|
---|
[3294] | 161 | public double GetMean(int variableIndex, int start, int end) {
|
---|
| 162 | return GetVariableValues(variableIndex, start, end).Average();
|
---|
[2319] | 163 | }
|
---|
| 164 |
|
---|
[3294] | 165 | public double GetRange(string variableName) {
|
---|
| 166 | return GetRange(GetVariableIndex(variableName));
|
---|
[2319] | 167 | }
|
---|
| 168 |
|
---|
[3294] | 169 | public double GetRange(int variableIndex) {
|
---|
[3308] | 170 | return GetRange(variableIndex, 0, Rows);
|
---|
[2319] | 171 | }
|
---|
| 172 |
|
---|
[3294] | 173 | public double GetRange(string variableName, int start, int end) {
|
---|
| 174 | return GetRange(GetVariableIndex(variableName), start, end);
|
---|
[2319] | 175 | }
|
---|
| 176 |
|
---|
[3294] | 177 | public double GetRange(int variableIndex, int start, int end) {
|
---|
| 178 | var values = GetVariableValues(variableIndex, start, end);
|
---|
| 179 | return values.Max() - values.Min();
|
---|
[2319] | 180 | }
|
---|
| 181 |
|
---|
[3294] | 182 | public double GetMax(string variableName) {
|
---|
| 183 | return GetMax(GetVariableIndex(variableName));
|
---|
[2319] | 184 | }
|
---|
| 185 |
|
---|
[3294] | 186 | public double GetMax(int variableIndex) {
|
---|
[3308] | 187 | return GetMax(variableIndex, 0, Rows);
|
---|
[2319] | 188 | }
|
---|
| 189 |
|
---|
[3294] | 190 | public double GetMax(string variableName, int start, int end) {
|
---|
| 191 | return GetMax(GetVariableIndex(variableName), start, end);
|
---|
[2319] | 192 | }
|
---|
| 193 |
|
---|
[3294] | 194 | public double GetMax(int variableIndex, int start, int end) {
|
---|
| 195 | return GetVariableValues(variableIndex, start, end).Max();
|
---|
[2319] | 196 | }
|
---|
| 197 |
|
---|
[3294] | 198 | public double GetMin(string variableName) {
|
---|
| 199 | return GetMin(GetVariableIndex(variableName));
|
---|
[2319] | 200 | }
|
---|
| 201 |
|
---|
[3294] | 202 | public double GetMin(int variableIndex) {
|
---|
[3308] | 203 | return GetMin(variableIndex, 0, Rows);
|
---|
[2319] | 204 | }
|
---|
| 205 |
|
---|
[3294] | 206 | public double GetMin(string variableName, int start, int end) {
|
---|
| 207 | return GetMin(GetVariableIndex(variableName), start, end);
|
---|
[2319] | 208 | }
|
---|
| 209 |
|
---|
[3294] | 210 | public double GetMin(int variableIndex, int start, int end) {
|
---|
| 211 | return GetVariableValues(variableIndex, start, end).Min();
|
---|
[2319] | 212 | }
|
---|
[2368] | 213 |
|
---|
[3294] | 214 | public int GetMissingValues(string variableName) {
|
---|
| 215 | return GetMissingValues(GetVariableIndex(variableName));
|
---|
[2368] | 216 | }
|
---|
[3294] | 217 | public int GetMissingValues(int variableIndex) {
|
---|
[3308] | 218 | return GetMissingValues(variableIndex, 0, Rows);
|
---|
[2368] | 219 | }
|
---|
| 220 |
|
---|
[3294] | 221 | public int GetMissingValues(string variableName, int start, int end) {
|
---|
| 222 | return GetMissingValues(GetVariableIndex(variableName), start, end);
|
---|
[2368] | 223 | }
|
---|
| 224 |
|
---|
[3294] | 225 | public int GetMissingValues(int variableIndex, int start, int end) {
|
---|
| 226 | return GetVariableValues(variableIndex, start, end).Count(x => double.IsNaN(x));
|
---|
[2368] | 227 | }
|
---|
| 228 |
|
---|
[2319] | 229 | #endregion
|
---|
| 230 |
|
---|
[3253] | 231 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
| 232 | Dataset clone = (Dataset)base.Clone(cloner);
|
---|
[3493] | 233 | clone.data = data;
|
---|
[3308] | 234 | clone.variableNames = (string[])variableNames.Clone();
|
---|
[3253] | 235 | return clone;
|
---|
[2319] | 236 | }
|
---|
| 237 |
|
---|
[3253] | 238 | #region events
|
---|
| 239 | public event EventHandler<EventArgs<int, int>> DataChanged;
|
---|
| 240 | private void OnDataChanged(EventArgs<int, int> e) {
|
---|
| 241 | var listeners = DataChanged;
|
---|
| 242 | if (listeners != null) listeners(this, e);
|
---|
[2319] | 243 | }
|
---|
[3253] | 244 | public event EventHandler Reset;
|
---|
| 245 | private void OnReset(EventArgs e) {
|
---|
| 246 | var listeners = Reset;
|
---|
| 247 | if (listeners != null) listeners(this, e);
|
---|
[2319] | 248 | }
|
---|
[3253] | 249 | #endregion
|
---|
[2] | 250 |
|
---|
[3253] | 251 | #region IStringConvertibleMatrix Members
|
---|
[237] | 252 |
|
---|
[3253] | 253 | public int Rows {
|
---|
| 254 | get {
|
---|
[3308] | 255 | return data.GetLength(0);
|
---|
[237] | 256 | }
|
---|
[3253] | 257 | set {
|
---|
| 258 | if (value == 0) throw new ArgumentException("Number of rows must be at least one (for variable names)");
|
---|
[3308] | 259 | if (value != Rows) {
|
---|
| 260 | var newValues = new double[value, Columns];
|
---|
| 261 | for (int row = 0; row < Math.Min(Rows, value); row++) {
|
---|
| 262 | for (int column = 0; column < Columns; column++) {
|
---|
[3253] | 263 | newValues[row, column] = data[row, column];
|
---|
| 264 | }
|
---|
[2] | 265 | }
|
---|
[3308] | 266 | Data = newValues;
|
---|
[2] | 267 | }
|
---|
| 268 | }
|
---|
| 269 | }
|
---|
| 270 |
|
---|
[3253] | 271 | public int Columns {
|
---|
| 272 | get {
|
---|
[3308] | 273 | return data.GetLength(1);
|
---|
[3253] | 274 | }
|
---|
| 275 | set {
|
---|
[3308] | 276 | if (value != Columns) {
|
---|
| 277 | var newValues = new double[Rows, value];
|
---|
[3253] | 278 | var newVariableNames = new string[value];
|
---|
[3308] | 279 | for (int row = 0; row < Rows; row++) {
|
---|
| 280 | for (int column = 0; column < Math.Min(value, Columns); column++) {
|
---|
[3253] | 281 | newValues[row, column] = data[row, column];
|
---|
| 282 | }
|
---|
| 283 | }
|
---|
[3294] | 284 | string formatString = new StringBuilder().Append('0', (int)Math.Log10(value) + 1).ToString(); // >= 100 variables => ###
|
---|
[3253] | 285 | for (int column = 0; column < value; column++) {
|
---|
[3308] | 286 | if (column < Columns)
|
---|
[3253] | 287 | newVariableNames[column] = variableNames[column];
|
---|
| 288 | else
|
---|
| 289 | newVariableNames[column] = "Var" + column.ToString(formatString);
|
---|
| 290 | }
|
---|
[3321] | 291 | VariableNames = newVariableNames;
|
---|
[3308] | 292 | Data = newValues;
|
---|
[2] | 293 | }
|
---|
| 294 | }
|
---|
| 295 | }
|
---|
| 296 |
|
---|
[3321] | 297 | [Storable]
|
---|
| 298 | private bool sortableView;
|
---|
| 299 | public bool SortableView {
|
---|
| 300 | get { return sortableView; }
|
---|
| 301 | set {
|
---|
| 302 | if (value != sortableView) {
|
---|
| 303 | sortableView = value;
|
---|
| 304 | OnSortableViewChanged();
|
---|
| 305 | }
|
---|
| 306 | }
|
---|
| 307 | }
|
---|
| 308 |
|
---|
[3430] | 309 | public bool ReadOnly {
|
---|
| 310 | get { return false; }
|
---|
| 311 | }
|
---|
| 312 |
|
---|
[3308] | 313 | IEnumerable<string> IStringConvertibleMatrix.ColumnNames {
|
---|
| 314 | get { return this.VariableNames; }
|
---|
| 315 | set {
|
---|
| 316 | int i = 0;
|
---|
| 317 | foreach (string variableName in value) {
|
---|
| 318 | SetVariableName(i, variableName);
|
---|
| 319 | i++;
|
---|
| 320 | }
|
---|
[3321] | 321 | OnColumnNamesChanged();
|
---|
[3308] | 322 | }
|
---|
| 323 | }
|
---|
| 324 |
|
---|
[3311] | 325 | IEnumerable<string> IStringConvertibleMatrix.RowNames {
|
---|
| 326 | get { return new List<string>(); }
|
---|
| 327 | set { throw new NotImplementedException(); }
|
---|
| 328 | }
|
---|
| 329 |
|
---|
[3253] | 330 | public bool Validate(string value, out string errorMessage) {
|
---|
[3321] | 331 | double val;
|
---|
| 332 | bool valid = double.TryParse(value, out val);
|
---|
[3253] | 333 | errorMessage = string.Empty;
|
---|
[3321] | 334 | if (!valid) {
|
---|
| 335 | StringBuilder sb = new StringBuilder();
|
---|
| 336 | sb.Append("Invalid Value (Valid Value Format: \"");
|
---|
| 337 | sb.Append(FormatPatterns.GetDoubleFormatPattern());
|
---|
| 338 | sb.Append("\")");
|
---|
| 339 | errorMessage = sb.ToString();
|
---|
| 340 | }
|
---|
| 341 | return valid;
|
---|
[3253] | 342 | }
|
---|
[2] | 343 |
|
---|
[3253] | 344 | public string GetValue(int rowIndex, int columnIndex) {
|
---|
[3308] | 345 | return data[rowIndex, columnIndex].ToString();
|
---|
[2] | 346 | }
|
---|
| 347 |
|
---|
[3253] | 348 | public bool SetValue(string value, int rowIndex, int columnIndex) {
|
---|
[3308] | 349 | double v;
|
---|
| 350 | if (double.TryParse(value, out v)) {
|
---|
| 351 | data[rowIndex, columnIndex] = v;
|
---|
[3321] | 352 | OnDataChanged(new EventArgs<int, int>(rowIndex, columnIndex));
|
---|
| 353 | OnItemChanged(rowIndex, columnIndex);
|
---|
[3308] | 354 | return true;
|
---|
| 355 | } else return false;
|
---|
[237] | 356 | }
|
---|
| 357 |
|
---|
[3321] | 358 | public event EventHandler ColumnNamesChanged;
|
---|
| 359 | private void OnColumnNamesChanged() {
|
---|
| 360 | EventHandler handler = ColumnNamesChanged;
|
---|
| 361 | if (handler != null)
|
---|
| 362 | handler(this, EventArgs.Empty);
|
---|
| 363 | }
|
---|
| 364 | public event EventHandler RowNamesChanged;
|
---|
| 365 | private void OnRowNamesChanged() {
|
---|
| 366 | EventHandler handler = RowNamesChanged;
|
---|
| 367 | if (handler != null)
|
---|
| 368 | handler(this, EventArgs.Empty);
|
---|
| 369 | }
|
---|
| 370 | public event EventHandler SortableViewChanged;
|
---|
| 371 | private void OnSortableViewChanged() {
|
---|
| 372 | EventHandler handler = SortableViewChanged;
|
---|
| 373 | if (handler != null)
|
---|
| 374 | handler(this, EventArgs.Empty);
|
---|
| 375 | }
|
---|
[3253] | 376 | public event EventHandler<EventArgs<int, int>> ItemChanged;
|
---|
[3321] | 377 | private void OnItemChanged(int rowIndex, int columnIndex) {
|
---|
| 378 | if (ItemChanged != null)
|
---|
| 379 | ItemChanged(this, new EventArgs<int, int>(rowIndex, columnIndex));
|
---|
| 380 | OnToStringChanged();
|
---|
| 381 | }
|
---|
[2012] | 382 | #endregion
|
---|
[3493] | 383 |
|
---|
| 384 |
|
---|
[2] | 385 | }
|
---|
| 386 | }
|
---|