package org.neuroph.nnet.learning.kmeans;

import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import org.neuroph.core.data.DataSet;
import org.neuroph.core.data.DataSetRow;

/**
 * 
 *   1. Pick an initial set of K centroids (this can be random or any other means)
 *   2. For each data point, assign it to the member of the closest centroid according to the given distance function
 *   3. Adjust the centroid position as the mean of all its assigned member data points. Go back to (2) until the membership isn't change and centroid position is stable.
 *
 * @author Zoran Sevarac
 * @author Uros Stojkic
 */
public class KMeansClustering {

    /**
     * Data/points to cluster
     */
    private DataSet dataSet;
    
    private KVector[] dataVectors;
    
    /**
     * Total number of clusters
     */
    private int numberOfClusters;
    
    /**
     * Clusters
     */
    private Cluster[] clusters;
        
    StringBuilder log=new StringBuilder();

    public KMeansClustering(DataSet dataSet) {
        this.dataSet = dataSet;
        this.dataVectors = new KVector[dataSet.size()];
        // iterate dataset and create dataVectors field
        this.dataVectors = new KVector[dataSet.size()];
         // iterate dataset and create dataVectors field
        int i=0;
        for(DataSetRow row : dataSet.getRows()) {
            KVector vector = new KVector(row.getInput());
            this.dataVectors[i]=vector;
            i++;
            
        }        
    }

    public KMeansClustering(DataSet dataSet, int numberOfClusters) {
        this.dataSet = dataSet;
        this.numberOfClusters = numberOfClusters;
        this.dataVectors = new KVector[dataSet.size()];
         // iterate dataset and create dataVectors field
        int i=0;
        for(DataSetRow row : dataSet.getRows()) {
            KVector vector = new KVector(row.getInput());
            this.dataVectors[i]=vector;
            i++;
            
        }
    }

    // find initial values for centroids/clusters
    // dont need to return string this is just for debuging and output in dialog
    // forgy and random partitions
    // http://en.wikipedia.org/wiki/K-means_clustering
    public void initClusters() {               
        
        ArrayList idxList = new ArrayList<>();
        
        for(int i=0; i0) {         // that cointains data
                double[] avgSum = cluster.getAvgSum();  // calculate avg sum
                cluster.getCentroid().setValues(avgSum);  // and set new centroid to avg sum
            }                                  
        }
    }

    public DataSet getDataSet() {
        return dataSet;
    }

    public void setDataSet(DataSet vectors) {
        this.dataSet = vectors;
    }

    public void setNumberOfClusters(int numberOfClusters) {
        this.numberOfClusters = numberOfClusters;
    }

    public Cluster[] getClusters() {
        return clusters;
    }
    
    public String getLog() {
        return log.toString();
    }
}