#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2012 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
 *
 * This file is part of HeuristicLab.
 *
 * HeuristicLab is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * HeuristicLab is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
 */
#endregion

using System;
using System.Collections;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using HeuristicLab.Common;
using HeuristicLab.Core;
using HeuristicLab.Data;
using HeuristicLab.Optimization;
using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;

namespace HeuristicLab.Encodings.ParameterConfigurationTreeEncoding {
  [StorableClass]
  public class ParameterConfigurationTree : ParameterizedValueConfiguration, IEnumerable {

    [Storable]
    private DoubleValue quality;
    public DoubleValue Quality {
      get { return quality; }
      set {
        if (quality != value) {
          quality = value;
          OnQualityChanged();
        }
      }
    }

    [Storable]
    private DoubleArray normalizedQualityAverages;
    public DoubleArray NormalizedQualityAverages {
      get { return normalizedQualityAverages; }
      set {
        if (normalizedQualityAverages != value) {
          normalizedQualityAverages = value;
        }
      }
    }

    [Storable]
    private DoubleArray normalizedQualityDeviations;
    public DoubleArray NormalizedQualityDeviations {
      get { return normalizedQualityDeviations; }
      set {
        if (normalizedQualityDeviations != value) {
          normalizedQualityDeviations = value;
        }
      }
    }

    [Storable]
    private DoubleArray normalizedEvaluatedSolutions;
    public DoubleArray NormalizedEvaluatedSolutions {
      get { return normalizedEvaluatedSolutions; }
      set {
        if (normalizedEvaluatedSolutions != value) {
          normalizedEvaluatedSolutions = value;
        }
      }
    }

    [Storable]
    private DoubleArray bestQualities;
    public DoubleArray BestQualities {
      get { return bestQualities; }
      set {
        if (bestQualities != value) {
          bestQualities = value;
        }
      }
    }

    [Storable]
    private DoubleArray averageQualities;
    public DoubleArray AverageQualities {
      get { return averageQualities; }
      set { averageQualities = value; }
    }

    [Storable]
    private DoubleArray worstQualities;
    public DoubleArray WorstQualities {
      get { return worstQualities; }
      set { worstQualities = value; }
    }

    [Storable]
    private DoubleArray qualityVariances;
    public DoubleArray QualityVariances {
      get { return qualityVariances; }
      set { qualityVariances = value; }
    }

    [Storable]
    private DoubleArray qualityStandardDeviations;
    public DoubleArray QualityStandardDeviations {
      get { return qualityStandardDeviations; }
      set { qualityStandardDeviations = value; }
    }

    [Storable]
    private ItemList<TimeSpanValue> averageExecutionTimes;
    public ItemList<TimeSpanValue> AverageExecutionTimes {
      get { return averageExecutionTimes; }
      set { averageExecutionTimes = value; }
    }

    [Storable]
    private DoubleArray averageEvaluatedSolutions;
    public DoubleArray AverageEvaluatedSolutions {
      get { return averageEvaluatedSolutions; }
      set { averageEvaluatedSolutions = value; }
    }

    [Storable]
    private IntValue repetitions;
    public IntValue Repetitions {
      get { return repetitions; }
      set { repetitions = value; }
    }

    [Storable]
    protected RunCollection runs;
    public RunCollection Runs {
      get { return runs; }
      set { runs = value; }
    }

    [Storable]
    protected IDictionary<string, IItem> parameters;
    public IDictionary<string, IItem> Parameters {
      get { return parameters; }
      set { parameters = value; }
    }

    public ParameterizedValueConfiguration AlgorithmConfiguration {
      get {
        return this.ParameterConfigurations.ElementAt(0).ValueConfigurations.First() as ParameterizedValueConfiguration;
      }
    }

    public ParameterizedValueConfiguration ProblemConfiguration {
      get {
        return this.ParameterConfigurations.ElementAt(1).ValueConfigurations.First() as ParameterizedValueConfiguration;
      }
    }

    #region constructors and cloning
    public ParameterConfigurationTree(IAlgorithm algorithm, IProblem problem)
      : base(null, algorithm.GetType(), false) {
      this.Optimize = false;
      this.IsOptimizable = false;
      this.parameters = new Dictionary<string, IItem>();
      this.Name = algorithm.ItemName;

      var algproblemitem = new AlgorithmProblemItem();
      algproblemitem.AlgorithmParameter.Value = algorithm;
      algproblemitem.ProblemParameter.Value = problem;
      this.discoverValidValues = false;

      this.parameterConfigurations.Add(new SingleValuedParameterConfiguration("Algorithm", algproblemitem.AlgorithmParameter));
      this.parameterConfigurations.Add(new SingleValuedParameterConfiguration("Problem", algproblemitem.ProblemParameter));

      // problems can be modified in the list of problem instances, so the parameters which are not Optimize=true, 
      // must not be modifiable in the parameter configuration tree. otherwise the parameter values would be ambiguous
      ProblemConfiguration.ValuesReadOnly = true;
    }
    public ParameterConfigurationTree() { }
    [StorableConstructor]
    protected ParameterConfigurationTree(bool deserializing) : base(deserializing) { }
    protected ParameterConfigurationTree(ParameterConfigurationTree original, Cloner cloner)
      : base(original, cloner) {
      this.quality = cloner.Clone(original.quality);
      this.normalizedQualityAverages = cloner.Clone(original.normalizedQualityAverages);
      this.normalizedQualityDeviations = cloner.Clone(original.normalizedQualityDeviations);
      this.normalizedEvaluatedSolutions = cloner.Clone(original.normalizedEvaluatedSolutions);
      this.bestQualities = cloner.Clone(original.BestQualities);
      this.averageQualities = cloner.Clone(original.averageQualities);
      this.worstQualities = cloner.Clone(original.worstQualities);
      this.qualityStandardDeviations = cloner.Clone(original.qualityStandardDeviations);
      this.qualityVariances = cloner.Clone(original.qualityVariances);
      this.averageExecutionTimes = cloner.Clone(original.averageExecutionTimes);
      this.averageEvaluatedSolutions = cloner.Clone(original.averageEvaluatedSolutions);
      this.repetitions = cloner.Clone(original.repetitions);
      this.runs = cloner.Clone(original.runs);
      this.parameters = new Dictionary<string, IItem>();
      foreach (var p in original.parameters) {
        this.parameters.Add(p.Key, cloner.Clone(p.Value));
      }
    }
    public override IDeepCloneable Clone(Cloner cloner) {
      return new ParameterConfigurationTree(this, cloner);
    }
    [StorableHook(HookType.AfterDeserialization)]
    private void AfterDeserialization() {
      if (ProblemConfiguration != null) ProblemConfiguration.ValuesReadOnly = true;
    }
    #endregion

    public virtual void CollectResultValues(IDictionary<string, IItem> values) {
      values.Add("RunsAverageExecutionTimes", AverageExecutionTimes);
      values.Add("RunsAverageEvaluatedSolutions", AverageEvaluatedSolutions);
      values.Add("Repetitions", Repetitions);
      values.Add("RunsBestQualities", BestQualities);
      values.Add("RunsAverageQualities", AverageQualities);
      values.Add("RunsWorstQualities", WorstQualities);
      values.Add("RunsQualityVariances", QualityVariances);
      values.Add("RunsQualityStandardDeviations", QualityStandardDeviations);
      values.Add("QualitiesNormalized", NormalizedQualityAverages);
      values.Add("AverageQualityNormalized", Quality);
      values.Add("Runs", Runs);
    }

    public virtual void CollectParameterValues(IDictionary<string, IItem> values) {
      foreach (var p in parameters) {
        values.Add(p);
      }
    }

    #region Events
    public event EventHandler QualityChanged;
    private void OnQualityChanged() {
      var handler = QualityChanged;
      if (handler != null) handler(this, EventArgs.Empty);
    }

    private void Quality_ValueChanged(object sender, EventArgs e) {
      OnQualityChanged();
    }
    #endregion

    public override void Parameterize(IParameterizedItem item) {
      this.parameters.Clear();
      var algorithm = item as IAlgorithm;
      var problem = algorithm.Problem;

      ProblemConfiguration.Parameterize(problem);
      AlgorithmConfiguration.Parameterize(algorithm);

      algorithm.CollectParameterValues(this.Parameters);
    }

    public Experiment GenerateExperiment(IAlgorithm algorithm, bool createBatchRuns, int repetitions) {
      Experiment experiment = new Experiment();
      foreach (ParameterizedValueConfiguration combination in this) {
        IAlgorithm clonedAlg = (IAlgorithm)algorithm.Clone();
        clonedAlg.Name = combination.ParameterInfoString;
        combination.Parameterize(clonedAlg);
        clonedAlg.StoreAlgorithmInEachRun = false;
        if (createBatchRuns) {
          BatchRun batchRun = new BatchRun(string.Format("BatchRun: {0}", combination.ParameterInfoString));
          batchRun.Optimizer = clonedAlg;
          batchRun.Repetitions = repetitions;
          experiment.Optimizers.Add(batchRun);
        } else {
          experiment.Optimizers.Add(clonedAlg);
        }
      }
      return experiment;
    }

    public Experiment GenerateExperiment(IAlgorithm algorithm) {
      return GenerateExperiment(algorithm, false, 0);
    }

    public IEnumerator GetEnumerator() {
      IEnumerator enumerator = new ParameterCombinationsEnumerator(this);
      enumerator.Reset();
      return enumerator;
    }

    /// <summary>
    /// returns the number of possible parameter combinations
    /// </summary>
    /// <param name="max">algorithm stops counting when max is reached. zero for infinite counting</param>
    /// <returns></returns>
    public long GetCombinationCount(long max) {
      long cnt = 0;
      foreach (var c in this) {
        cnt++;
        if (max > 0 && cnt >= max) {
          return cnt;
        }
      }
      return cnt;
    }

    public IOptimizable GetRandomOptimizable(IRandom random) {
      List<IOptimizable> allOptimizables = GetAllOptimizables();
      return allOptimizables[random.Next(allOptimizables.Count)];
    }

    public override string ToString() {
      return this.Name;
    }

    public IRun ToRun(bool clearParameters) {
      return ToRun(this.ParameterInfoString, clearParameters);
    }

    public IRun ToRun(string name, bool clearParameters) {
      IRun run = new Run();
      run.Name = name;
      this.CollectResultValues(run.Results);
      this.CollectParameterValues(run.Parameters);
      if (clearParameters) MetaOptimizationUtil.ClearParameters(run, this.GetOptimizedParameterNames());
      return run;
    }

    public override string ParameterInfoString {
      get {
        string algorithmInfo = this.AlgorithmConfiguration.ParameterInfoString;
        string problemInfo = this.ProblemConfiguration.ParameterInfoString;
        var sb = new StringBuilder();
        if (!string.IsNullOrEmpty(algorithmInfo)) {
          sb.Append("Algorithm (");
          sb.Append(algorithmInfo);
          sb.Append(")");
        }
        if (!string.IsNullOrEmpty(problemInfo)) {
          if (sb.Length > 0)
            sb.Append(", ");
          sb.Append("Problem( ");
          sb.Append(problemInfo);
          sb.Append(")");
        }
        return sb.ToString();
      }
    }

    public override void CollectOptimizedParameterNames(List<string> parameterNames, string prefix) {
      AlgorithmConfiguration.CollectOptimizedParameterNames(parameterNames, string.Empty);
      ProblemConfiguration.CollectOptimizedParameterNames(parameterNames, string.Empty);
    }
  }
}