Changeset 7805 for branches/ProblemInstancesRegressionAndClassification/HeuristicLab.Problems.Instances.Regression/3.4/RegressionInstanceProvider.cs

Timestamp:

05/14/12 15:21:28 (12 years ago)

Author:

sforsten

Message:

#1784: changes have been applied, according to the review comments of mkommend

File:

• branches/ProblemInstancesRegressionAndClassification/HeuristicLab.Problems.Instances.Regression/3.4/RegressionInstanceProvider.cs (modified) (5 diffs)

branches/ProblemInstancesRegressionAndClassification/HeuristicLab.Problems.Instances.Regression/3.4/RegressionInstanceProvider.cs

@@ -23 +23 @@
 using System.Collections;
 using System.Collections.Generic;
+using System.Globalization;
 using System.IO;
 using System.Linq;
@@ -29 +30 @@
 
 namespace HeuristicLab.Problems.Instances.Regression {
-  public abstract class RegressionInstanceProvider : IProblemInstanceProvider<RegressionProblemData> {
-    public RegressionProblemData LoadData(string path) {
-      TableFileParser csvFileParser = new TableFileParser();
-      csvFileParser.Parse(path);
+  public abstract class RegressionInstanceProvider : IProblemInstanceProvider<IRegressionProblemData> {
 
-      Dataset dataset = new Dataset(csvFileParser.VariableNames, csvFileParser.Values);
-      string targetVar = csvFileParser.VariableNames.Last();
-      IEnumerable<string> allowedInputVars = csvFileParser.VariableNames.Where(x => !x.Equals(targetVar));
+    public IRegressionProblemData LoadData(string path) {
+      NumberFormatInfo numberFormat;
+      DateTimeFormatInfo dateFormat;
+      char separator;
+      TableFileParser.DetermineFileFormat(path, out numberFormat, out dateFormat, out separator);
 
-      RegressionProblemData regData = new RegressionProblemData(dataset, allowedInputVars, targetVar);
+      IRegressionProblemData regData = LoadData(new FileStream(path, FileMode.Open), numberFormat, dateFormat, separator);
+
       int pos = path.LastIndexOf('\\');
       if (pos < 0)
@@ -46 +47 @@
         regData.Name = path.Substring(pos, path.Length - pos);
       }
+      return regData;
+    }
+
+    protected IRegressionProblemData LoadData(Stream stream, NumberFormatInfo numberFormat, DateTimeFormatInfo dateFormat, char separator) {
+      TableFileParser csvFileParser = new TableFileParser();
+
+      csvFileParser.Parse(stream, numberFormat, dateFormat, separator);
+
+      Dataset dataset = new Dataset(csvFileParser.VariableNames, csvFileParser.Values);
+      string targetVar = csvFileParser.VariableNames.Last();
+      IEnumerable<string> allowedInputVars = csvFileParser.VariableNames.Where(x => !x.Equals(targetVar));
+
+      RegressionProblemData regData = new RegressionProblemData(dataset, allowedInputVars, targetVar);
 
       int trainingPartEnd = csvFileParser.Rows * 2 / 3;
@@ -52 +66 @@
       regData.TestPartition.Start = trainingPartEnd;
       regData.TestPartition.End = csvFileParser.Rows;
+
       return regData;
     }
 
-    public void SaveData(RegressionProblemData instance, string path) {
+    public void SaveData(IRegressionProblemData instance, string path) {
       StringBuilder strBuilder = new StringBuilder();
 
@@ -80 +95 @@
 
     public abstract IEnumerable<IDataDescriptor> GetDataDescriptors();
-    public abstract RegressionProblemData LoadData(IDataDescriptor descriptor);
+    public abstract IRegressionProblemData LoadData(IDataDescriptor descriptor);
 
     public abstract string Name { get; }