#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2016 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.Linq;

namespace HeuristicLab.Analysis {
  public class ClusteringHelper<T> {
    private readonly int k;

    private List<KeyValuePair<T, double>> instances;
    private List<KeyValuePair<T, double>> excluded;
    private int[] clusterValues;
    private int[] rankedMap;

    private ClusteringHelper(int K) {
      this.k = K;
    }

    /// <summary>
    /// Helps in clustering data which is available as key-value pairs.
    /// It is possible to specify an exclude function to omit certain points from clustering.
    /// These points will be assigned cluster with id k. All other points will be assigned a
    /// cluster id in the range [0;k-1].
    /// </summary>
    /// <param name="k">The maximum number of clusters that should be created.</param>
    /// <param name="values">The data which links a certain item with a double value that is to be clustered.</param>
    /// <param name="excludeFunc">The function that allows excluding certing data points which will receive cluster id k.</param>
    /// <returns>A reference to the helper class to allow fluent calls.</returns>
    public static ClusteringHelper<T> Cluster(int k, IEnumerable<KeyValuePair<T, double>> values, Func<KeyValuePair<T, double>, bool> excludeFunc = null) {
      if (excludeFunc == null) excludeFunc = _ => false;
      var helper = new ClusteringHelper<T>(k);
      helper.Initialize(values, excludeFunc);
      if (helper.instances.Count == 0) {
        helper.clusterValues = new int[0];
        helper.rankedMap = new int[k + 1];
      } else {
        var centroids = CkMeans1D.Cluster(helper.instances.Select(x => x.Value).ToArray(), k, out helper.clusterValues);
        helper.rankedMap = new int[k + 1];
        var rank = 0;
        foreach (var c in centroids) helper.rankedMap[c.Value] = rank++;
      }
      // excluded are always ranked last
      helper.rankedMap[k] = k;
      return helper;
    }

    /// <summary>
    /// Returns the clustered data by 
    /// </summary>
    /// <returns></returns>
    public IEnumerable<KeyValuePair<T, Tuple<double, int>>> GetByInstance() {
      return GetClustered().Select(x => new KeyValuePair<T, Tuple<double, int>>(x.Item1.Key, Tuple.Create(x.Item1.Value, rankedMap[x.Item2])));
    }

    public IEnumerable<KeyValuePair<int, List<KeyValuePair<T, double>>>> GetByCluster() {
      return GetClustered().GroupBy(x => x.Item2)
        .Select(x => new KeyValuePair<int, List<KeyValuePair<T, double>>>(rankedMap[x.Key], x.Select(y => y.Item1).ToList()));
    }

    private void Initialize(IEnumerable<KeyValuePair<T, double>> values, Func<KeyValuePair<T, double>, bool> excludeFunc) {
      instances = new List<KeyValuePair<T, double>>();
      excluded = new List<KeyValuePair<T, double>>();
      foreach (var v in values) {
        if (!excludeFunc(v)) instances.Add(v);
        else excluded.Add(v);
      }
    }

    private IEnumerable<Tuple<KeyValuePair<T, double>, int>> GetClustered() {
      for (var i = 0; i < clusterValues.Length; i++) {
        yield return Tuple.Create(instances[i], clusterValues[i]);
      }
      foreach (var ex in excluded)
        yield return Tuple.Create(ex, k);
    }
  }
}