#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2018 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.Generic;
using System.IO;
using System.Linq;
using System.Text.RegularExpressions;
using HeuristicLab.Random;
using ProtoBuf;
using static HeuristicLab.Analysis.FitnessLandscape.QAPDirectedWalk;

namespace WalkExporter {
  class Program {
    public static readonly string[] SBF = new[] { "Sharpness", "Bumpiness", "Flatness" };
    public static readonly string[] RUG = new[] {
      "AC1", "CorrelationLength" };
    public static readonly string[] IAL = new[] {
      "InformationContent", "DensityBasinInformation", "PartialInformationContent",
      "InformationStability", "Diversity", "Regularity", "TotalEntropy", "SymmetricInformationContent",
      "SymmetricDensityBasinInformation", "SymmetricTotalEntropy", "PeakInformationContent", "PeakDensityBasinInformation",
      "PeakTotalEntropy", "PeakSymmetricInformationContent", "PeakSymmetricDensityBasinInformation", "PeakSymmetricTotalEntropy" };
    public static readonly string[] IALREG = new[] {
      "InformationContent", "DensityBasinInformation", "PartialInformationContent",
      "InformationStability", "Diversity", "Regularity", "TotalEntropy",
      "PeakInformationContent", "PeakDensityBasinInformation", "PeakTotalEntropy"
    };
    public static readonly string[] IALSYM = new[] {
      "PartialInformationContent", "InformationStability", "Diversity", "Regularity",
      "SymmetricInformationContent", "SymmetricDensityBasinInformation", "SymmetricTotalEntropy",
      "PeakSymmetricInformationContent", "PeakSymmetricDensityBasinInformation", "PeakSymmetricTotalEntropy"
    };
    

    static void Main(string[] args) {
      //AnalyzeRandomWalkIdentification();
      //AnalyzeDirectedWalkIdentification();
      var provider = new HeuristicLab.Problems.Instances.QAPLIB.QAPLIBInstanceProvider();
      var tai30a = provider.LoadData(provider.GetDataDescriptors().Single(x => x.Name == "tai30a"));
      RandomWalk.ConfinedRandomWalkAnalysis(tai30a);
      var esc32f = provider.LoadData(provider.GetDataDescriptors().Single(x => x.Name == "esc32f"));
      RandomWalk.ConfinedRandomWalkAnalysis(esc32f);
    }

    private static void AnalyzeRandomWalkIdentification() {
      string[] RUG_IAL = RUG.Concat(IAL).ToArray();

      var trainFiles = GetFiles(@"randwalk_kb_train_(?<eff>\d+)");
      var testFiles = GetFiles(@"randwalk_kb_test_(?<eff>\d+)");

      var filename = "randwalk_results.csv";

      var features = new[] { (Name: "RUG", Set: RUG), (Name: "IAL", Set: IAL),
        (Name: "IALREG", Set: IALREG), (Name: "IALSYM", Set: IALSYM), (Name: "RUG_IAL", Set: RUG_IAL) };

      using (var writer = File.CreateText(filename)) {
        CompareMatching(trainFiles, testFiles, features, "randwalk", writer);
      }
    }

    private static void AnalyzeDirectedWalkIdentification() {
      string[] RUG_IAL = RUG.Concat(IAL).ToArray();
      string[] SBF_RUG = SBF.Concat(RUG).ToArray();
      string[] SBF_IAL = SBF.Concat(IAL).ToArray();
      string[] SBF_IALREG = SBF.Concat(IALREG).ToArray();
      string[] SBF_IALSYM = SBF.Concat(IALSYM).ToArray();

      var features = new[] { (Name: "RUG", Set: RUG), (Name: "IAL", Set: IAL),
        (Name: "IALREG", Set: IALREG), (Name: "IALSYM", Set: IALSYM), (Name: "SBF", Set: SBF),
        (Name: "RUG_IAL", Set: RUG_IAL), (Name: "SBF_RUG", Set: SBF_RUG), (Name: "SBF_IAL", Set: SBF_IAL),
        (Name: "SBF_IALREG", Set: SBF_IALREG), (Name: "SBF_IALSYM", Set: SBF_IALSYM) };


      var trainFiles = GetFiles(@"rrdw_best_kb_train_(?<eff>\d+)_qap");
      var testFiles = GetFiles(@"rrdw_best_kb_test_(?<eff>\d+)_qap");

      var filename = "rrdw_best_results.csv";

      //using (var writer = File.CreateText(filename)) {
      //  CompareMatching(trainFiles, testFiles, features, "(rr)-dw", writer);
      //}
      //trainFiles = GetFiles(@"rldw_best_kb_train_(?<eff>\d+)_qap");
      //testFiles = GetFiles(@"rldw_best_kb_test_(?<eff>\d+)_qap");

      //filename = "rldw_best_results.csv";

      //using (var writer = File.CreateText(filename)) {
      //  CompareMatching(trainFiles, testFiles, features, "(rl)-dw", writer);
      //}

      trainFiles = GetFiles(@"lldw_best_kb_train_(?<eff>\d+)_qap");
      testFiles = GetFiles(@"lldw_best_kb_test_(?<eff>\d+)_qap");

      filename = "lldw_best_results.csv";

      using (var writer = File.CreateText(filename)) {
        CompareMatching(trainFiles, testFiles, features, "(ll)-dw", writer);
      }

      trainFiles = GetFiles(@"lidw_best_kb_train_(?<eff>\d+)_qap");
      testFiles = GetFiles(@"lidw_best_kb_test_(?<eff>\d+)_qap");

      filename = "lidw_best_results.csv";

      using (var writer = File.CreateText(filename)) {
        CompareMatching(trainFiles, testFiles, features, "(li)-dw", writer);
      }
    }

    private static List<(string Filename, int Effort)> GetFiles(string pattern) {
      // randwalk_kb_(train|test)_{n}.buf
      // {type}dw_best_kb_(train|test)_{n}_qap.buf

      return Directory.EnumerateFiles(".").Where(x => x.EndsWith(".buf"))
        .Select(x => {
          var match = Regex.Match(Path.GetFileName(x), pattern);
          if (match.Success) {
            return (Filename: x, Effort: int.Parse(match.Groups["eff"].Value));
          }
          return (Filename: "", Effort: -1);
        }).Where(x => !string.IsNullOrEmpty(x.Filename)).ToList();
    }

    private static void CompareMatching(List<(string Filename, int Effort)> trainFiles,
        List<(string Filename, int Effort)> testFiles, (string Name, string[] Set)[] featuresets,
        string type,
        StreamWriter writer) {
      var random = new MersenneTwister(42);
      var header = string.Format("{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6}\t{7}\t{8}\t{9}\t{10}\t{11}",
          "Dimension", "FSet", "Type", "TrainEff", "TestEff", "ExCnt", "ExRnk", "ClsCnt", "ClsRnk", "TotCnt", "TrainEffSolEquiv", "TestEffSolEquiv");
      writer.WriteLine(header);
      Console.WriteLine(header);

      foreach (var features in featuresets) {
        foreach (var dim in new[] { 20, 30, 40 }) {
          foreach (var a in trainFiles) {
            Knowledgebase train = null;
            using (var stream = File.OpenRead(a.Filename))
              train = Serializer.Deserialize<Knowledgebase>(stream);

            train.Problems.RemoveAll(x => x.Dimension != dim);
            if (train.Problems.Count == 0) throw new InvalidOperationException("Dimension does not exist: " + dim);

            var standardizer = InstancesStandardizer.CreateAndApply(train.Problems, features.Set);

            foreach (var b in testFiles) {
              Knowledgebase test = null;
              using (var stream = File.OpenRead(b.Filename))
                test = Serializer.Deserialize<Knowledgebase>(stream);

              test.Problems.RemoveAll(x => x.Dimension != dim);
              standardizer.Apply(test.Problems);
              // MATCH

              var match = EvaluateMatch(random, train, test, new HashSet<string>(features.Set));

              //correlation analysis
              //var corr = AnalyzeFeatureCorrelation(features.Set, train, test);

              string output = string.Format("{0}\t{1}\t{2}\t{3}\t{4}\t{5}\t{6:F2}\t{7}\t{8:F2}\t{9}\t{10:F2}\t{11:F2}",
                dim, features.Name, type, a.Effort, b.Effort, match.ExactCount,
                match.ExactAverageRank, match.ClsCount, match.ClsAverageRank, match.TotalCount,
                match.TrainingDescriptionEffort, match.TestDescriptionEffort);
              writer.WriteLine(output);
              Console.WriteLine(output);
            }
          }
        }
      }
    }

    private static MatchResult EvaluateMatch(MersenneTwister random, Knowledgebase train, Knowledgebase test, ISet<string> features) {
      var result = new MatchResult();

      foreach (var x in train.Problems) {

        var ranked = test.Problems.Shuffle(random).Select(y => new {
          Instance = y,
          Distance = (from xx in x.Features.Where(f => features.Contains(f.Key))
                      from yy in y.Features.Where(f => features.Contains(f.Key))
                      where xx.Key == yy.Key
                      let vxx = xx.GetNumericValue()
                      let vyy = yy.GetNumericValue()
                      select (vxx - vyy) * (vxx - vyy)).Sum(),
        }).OrderBy(xx => xx.Distance).ToList();

        var exactRank = -1;
        var clsRank = -1;
        var count = 1;
        foreach (var r in ranked) {
          result.TestDescriptionEffort += r.Instance.DescriptionEffort;
          if (clsRank < 0 && r.Instance.Class == x.Class) {
            clsRank = count;
          }
          if (r.Instance.Name == x.Name) {
            exactRank = count;
            break;
          }
          count++;
        }
        result.TestDescriptionEffort /= test.Problems.Count;
        if (exactRank == 1) result.ExactCount++;
        if (clsRank == 1) result.ClsCount++;
        result.TotalCount++;

        result.TrainingDescriptionEffort += x.DescriptionEffort;
        result.ExactAverageRank += exactRank;
        result.ClsAverageRank += clsRank;

      }
      result.TrainingDescriptionEffort /= train.Problems.Count;
      result.ExactAverageRank /= train.Problems.Count;
      result.ClsAverageRank /= train.Problems.Count;

      return result;
    }

    private static double[,] AnalyzeFeatureCorrelation(string[] features, Knowledgebase train, Knowledgebase test) {
      var trainMat = new double[train.Problems.Count, features.Length];
      var testMat = new double[test.Problems.Count, features.Length];
      int trainCount = 0, testCount = 0;

      foreach (var x in train.Problems) {
        var xFeatures = x.GetNumericFeatures(features);
        foreach (var f in xFeatures.Select((v, i) => new { Index = i, Value = v })) {
          trainMat[trainCount, f.Index] = f.Value;
        }
        trainCount++;
      }
      foreach (var y in test.Problems) {
        var yFeatures = y.GetNumericFeatures(features);
        foreach (var f in yFeatures.Select((v, i) => new { Index = i, Value = v })) {
          testMat[testCount, f.Index] = f.Value;
        }
        testCount++;
      }

      double[,] corr;
      alglib.pearsoncorrm2(trainMat, testMat, out corr);
      return corr;
    }

    private static void DoRandomWalk() {
      var experiment = RandomWalk.PerformExperiment();
      Serializer.Serialize(File.Create("randwalk_trials_qap.buf"), experiment);
      foreach (var exp in Enumerable.Range(7, 18 - 6)) {
        var len = (int)Math.Pow(2, exp);
        var (training, test) = RandomWalk.GetKnowledgeBases(experiment, len);
        Serializer.Serialize(File.Create($"randwalk_kb_train_{exp}.buf"), training);
        Serializer.Serialize(File.Create($"randwalk_kb_test_{exp}.buf"), test);
      }
    }

    private static void DoDirectedWalk() {
      var (exp, train, test) = DirectedWalk.PerformExperiment(WalkType.RandomRandom);
      Save("rrdw_best", exp, train, test);
      (exp, train, test) = DirectedWalk.PerformExperiment(WalkType.RandomLocal);
      Save("rldw_best", exp, train, test);
      (exp, train, test) = DirectedWalk.PerformExperiment(WalkType.LocalLocal);
      Save("lldw_best", exp, train, test);
      (exp, train, test) = DirectedWalk.PerformExperiment(WalkType.LocalInverse);
      Save("lidw_best", exp, train, test);
    }

    private static void Save(string v, Experiment exp, Dictionary<int, Knowledgebase> train, Dictionary<int, Knowledgebase> test) {
      Serializer.Serialize(File.Create(v + "_experiment_qap.buf"), exp);
      foreach (var t in train.Keys) {
        Serializer.Serialize(File.Create(v + $"_kb_train_{t}_qap.buf"), train[t]);
        Serializer.Serialize(File.Create(v + $"_kb_test_{t}_qap.buf"), test[t]);
      }
    }
  }

  public class MatchResult {
    public int ExactCount { get; set; }
    public int ClsCount { get; set; }
    public int TotalCount { get; set; }
    public double ExactAverageRank { get; set; }
    public double ClsAverageRank { get; set; }

    public double TrainingDescriptionEffort { get; set; }
    public double TestDescriptionEffort { get; set; }
  }
}