source: branches/2895_PushGP_GenealogyAnalysis/HeuristicLab.Problems.ProgramSynthesis/Push/Analyzer/PushExpressionFrequencyAnalyzer.cs @ 16099

namespace HeuristicLab.Problems.ProgramSynthesis {
  using System;
  using System.Collections.Generic;
  using System.Linq;

  using HeuristicLab.Analysis;
  using HeuristicLab.Common;
  using HeuristicLab.Core;
  using HeuristicLab.Data;
  using HeuristicLab.Encodings.IntegerVectorEncoding;
  using HeuristicLab.Operators;
  using HeuristicLab.Optimization;
  using HeuristicLab.Parameters;
  using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;

  /// <summary>
  /// An operater that tracks the frequencies of distinct push expressions in a push solution
  /// </summary>
  [Item("PushExpressionFrequencyAnalyzer", "An operater that tracks the frequencies of distinct push expressions in a push solution")]
  [StorableClass]
  public class PushExpressionFrequencyAnalyzer : SingleSuccessorOperator, IPushExpressionAnalyzer {

    private const string PUSH_CONFIGURATION_PARAMETER_NAME = "PushConfiguration";
    private const string INTEGER_VECTOR_PARAMETER_NAME = "IntegerVector";
    private const string RESULTS_PARAMETER_NAME = "Results";
    private const string EXPRESSION_FREQUENCIES_PARAMETER_NAME = "InstructionFrequencies";
    private const string AGGREGATE_STACK_TYPES_PARAMETER_NAME = "Aggregate stack types";
    private const string IN_EXPRESSION_GROUP_NAME = "IN";
    private const string RANDOM_PARAMETER_NAME = "Random";
    private const string RESULT_PARAMETER_NAME = "Instruction frequencies";
    private const string RESULT_PARAMETER_DESCRIPTION = "Relative frequency of instructions aggregated over the whole population.";
    private const string Y_AXIS_TITLE = "Relative Instruction Frequency";

    public PushExpressionFrequencyAnalyzer() {
      Parameters.Add(new LookupParameter<IReadOnlyPushConfiguration>(PUSH_CONFIGURATION_PARAMETER_NAME, "The current specified push configuration."));
      Parameters.Add(new ScopeTreeLookupParameter<IntegerVector>(INTEGER_VECTOR_PARAMETER_NAME, "The integer vectors to analyze."));
      Parameters.Add(new ScopeTreeLookupParameter<PlushVector>("Plush", "The plush vectors to analyze."));
      Parameters.Add(new LookupParameter<DataTable>(EXPRESSION_FREQUENCIES_PARAMETER_NAME, "The data table to store the instruction frequencies."));
      Parameters.Add(new LookupParameter<ResultCollection>(RESULTS_PARAMETER_NAME, "The result collection where the symbol frequencies should be stored."));

      Parameters.Add(new FixedValueParameter<BoolValue>(AGGREGATE_STACK_TYPES_PARAMETER_NAME, "Determines if expressions should be aggregated by their primary stack type.", new BoolValue(true)));
    }

    [StorableConstructor]
    public PushExpressionFrequencyAnalyzer(bool deserializing) : base(deserializing) { }

    public PushExpressionFrequencyAnalyzer(PushExpressionFrequencyAnalyzer origin, Cloner cloner) : base(origin, cloner) { }

    public override IDeepCloneable Clone(Cloner cloner) {
      return new PushExpressionFrequencyAnalyzer(this, cloner);
    }

    public bool EnabledByDefault { get { return true; } }

    public ILookupParameter<IReadOnlyPushConfiguration> PushConfigurationParameter {
      get { return (ILookupParameter<IReadOnlyPushConfiguration>)Parameters[PUSH_CONFIGURATION_PARAMETER_NAME]; }
    }

    public IScopeTreeLookupParameter<PlushVector> PlushVectorParameter {
      get { return (IScopeTreeLookupParameter<PlushVector>)Parameters["Plush"]; }
    }

    public IScopeTreeLookupParameter<IntegerVector> IntegerVectorParameter {
      get { return (IScopeTreeLookupParameter<IntegerVector>)Parameters[INTEGER_VECTOR_PARAMETER_NAME]; }
    }

    public ILookupParameter<DataTable> ExpressionFrequenciesParameter {
      get { return (ILookupParameter<DataTable>)Parameters[EXPRESSION_FREQUENCIES_PARAMETER_NAME]; }
    }

    public ILookupParameter<IRandom> RandomParameter {
      get { return (ILookupParameter<IRandom>)Parameters[RANDOM_PARAMETER_NAME]; }
    }

    public ILookupParameter<ResultCollection> ResultsParameter {
      get { return (ILookupParameter<ResultCollection>)Parameters[RESULTS_PARAMETER_NAME]; }
    }

    public IValueParameter<BoolValue> AggregateStackTypesParameter {
      get { return (IValueParameter<BoolValue>)Parameters[AGGREGATE_STACK_TYPES_PARAMETER_NAME]; }
    }

    public bool AggregateStackTypes {
      get { return AggregateStackTypesParameter.Value.Value; }
      set { AggregateStackTypesParameter.Value.Value = value; }
    }

    public override IOperation Apply() {
      var config = PushConfigurationParameter.ActualValue;

      IReadOnlyList<PushProgram> pushPrograms = null;

      if (IntegerVectorParameter.ActualValue.Length > 0) {
        pushPrograms = IntegerVectorParameter.ActualValue.Select(iv => iv.ToPushProgram(config)).ToList();
      } else if (PlushVectorParameter.ActualValue.Length > 0) {
        pushPrograms = PlushVectorParameter.ActualValue.Select(pv => pv.PushProgram).ToList();
      } else {
        // nothing to do
        return base.Apply();
      }

      var results = ResultsParameter.ActualValue;
      var frequencies = ExpressionFrequenciesParameter.ActualValue;

      if (frequencies == null) {
        frequencies = new DataTable(
          RESULT_PARAMETER_NAME,
          RESULT_PARAMETER_DESCRIPTION) {
          VisualProperties = {
            YAxisTitle = Y_AXIS_TITLE
          }
        };

        ExpressionFrequenciesParameter.ActualValue = frequencies;
        results.Add(new Result(RESULT_PARAMETER_NAME, frequencies));
      }

      var allExpressions = pushPrograms
        .SelectMany(p => p.DepthLast())
        .Select(e => new {
          Expression = e,
          Attribute = ExpressionTable.TypeToAttributeTable[e.GetType()]
        });

      var expressionFrequencies = allExpressions
        .GroupBy(x => GetGroupName(x.Attribute, AggregateStackTypes))
        .ToDictionary(
          group => group.Key,
          group => group.Count());

      var totalNumberOfExpressions = Math.Max(1.0, expressionFrequencies.Values.Sum());

      // all rows must have the same number of values so we can just take the first
      var numberOfValues = frequencies.Rows.Select(r => r.Values.Count).FirstOrDefault();

      foreach (var pair in expressionFrequencies) {
        if (!frequencies.Rows.ContainsKey(pair.Key)) {
          var row = new DataRow(pair.Key, string.Empty, Enumerable.Repeat(0.0, numberOfValues)) {
            VisualProperties = {
              StartIndexZero = true,
            }
          };

          frequencies.Rows.Add(row);
        }

        frequencies.Rows[pair.Key].Values.Add(Math.Round(pair.Value / totalNumberOfExpressions, 3));
      }

      // add a zero for each data row that was not modified in the previous loop
      foreach (var row in frequencies.Rows.Where(r => r.Values.Count != numberOfValues + 1))
        row.Values.Add(0.0);

      return base.Apply();
    }

    private string GetGroupName(PushExpressionAttribute attribute, bool aggregateStackTypes) {
      if (aggregateStackTypes) {
        return attribute.IsInExpression
         ? IN_EXPRESSION_GROUP_NAME
         : attribute.StackType.ToString();
      }

      return attribute.Name;
    }
  }
}