source: branches/PushGP/HeuristicLab.PushGP/HeuristicLab.Problem.ProgramSynthesis.BenchmarkSuite/BenchmarkSuiteDataDescriptor.cs @ 15160

namespace HeuristicLab.BenchmarkSuite {
  using System.Collections.Generic;
  using System.IO;
  using System.IO.Compression;
  using System.Linq;
  using System.Reflection;
  using System.Text.RegularExpressions;


  public abstract class BenchmarkSuiteDataDescriptor : IBenchmarkSuiteDataDescriptor {
    private const string InstanceArchiveName = "HeuristicLab.BenchmarkSuite.Data.BenchmarkExamples.zip";
    private const string ResourcePath = @".*\.Data\.";
    private const string ExampleFieldDelimiter = ",";

    private Example[] examples;

    protected Example[] CloneExamples() {
      if (examples == null) examples = ParseData().ToArray();
      return examples.Select(e => (Example)e.Clone()).ToArray();
    }

    public abstract ProblemData CreateProblemData();

    protected abstract string FileName { get; }
    public abstract string Name { get; }
    public abstract string Description { get; }
    protected abstract int InputArgumentCount { get; }
    protected abstract int OutputArgumentCount { get; }

    protected abstract Example ParseExample(string[] input, string[] output);

    //private IEnumerable<Example> ParseData() {
    //  using (var file = GetType().Assembly.GetManifestResourceStream(InstanceArchiveName))
    //  using (var archive = new ZipArchive(file, ZipArchiveMode.Read)) {
    //    var entry = archive.Entries.SingleOrDefault(x => x.Name == FileName);

    //    using (var parser = new TextFieldParser(entry.Open())) {
    //      parser.TextFieldType = FieldType.Delimited;
    //      parser.SetDelimiters(ExampleFieldDelimiter);
    //      parser.HasFieldsEnclosedInQuotes = true;
    //      parser.TrimWhiteSpace = false;

    //      //Processing rows
    //      while (!parser.EndOfData) {
    //        var fields = parser.ReadFields();

    //        if (fields.Length != InputArgumentCount + OutputArgumentCount)
    //          throw new InvalidDataException("Number of values do not fit");

    //        var input = fields.Take(InputArgumentCount).ToArray();
    //        var output = fields.Skip(InputArgumentCount).ToArray();

    //        yield return ParseExample(input, output);
    //      }
    //    }
    //  }
    //}

    private IEnumerable<Example> ParseData() {
      using (var file = GetType().Assembly.GetManifestResourceStream(InstanceArchiveName))
      using (var archive = new ZipArchive(file, ZipArchiveMode.Read)) {
        var entry = archive.Entries.SingleOrDefault(x => x.Name == FileName);

        using (var reader = new StreamReader(entry.Open())) {
          var lines = CsvParser.Parse(reader, ',', '"');

          foreach (var line in lines) {
            var input = line.Take(InputArgumentCount).ToArray();
            var output = line.Skip(InputArgumentCount).ToArray();

            yield return ParseExample(input, output);
          }
        }
      }
    }

    protected static string GetResourceName(string fileName) {
      return Assembly
          .GetExecutingAssembly()
          .GetManifestResourceNames()
          .SingleOrDefault(x => Regex.Match(x, ResourcePath + fileName).Success);
    }
  }
}