/**
 * Copyright (c) 2011, The University of Southampton and the individual contributors.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 *   *  Redistributions of source code must retain the above copyright notice,
 *      this list of conditions and the following disclaimer.
 *
 *   *  Redistributions in binary form must reproduce the above copyright notice,
 *      this list of conditions and the following disclaimer in the documentation
 *      and/or other materials provided with the distribution.
 *
 *   *  Neither the name of the University of Southampton nor the names of its
 *      contributors may be used to endorse or promote products derived from this
 *      software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
package org.openimaj.ml.clustering.kmeans;

import java.util.concurrent.ExecutorService;

import org.openimaj.knn.NearestNeighbours;
import org.openimaj.knn.NearestNeighboursFactory;
import org.openimaj.util.parallel.GlobalExecutorPool;

/**
 * Configuration for the K-Means algorithm.
 * 
 * @author Jonathon Hare (jsh2@ecs.soton.ac.uk)
 * 
 * @param <NN>
 *            The type of {@link NearestNeighbours} to use
 * @param <DATA>
 *            The type of data
 */
public class KMeansConfiguration<NN extends NearestNeighbours<DATA, ?, ?>, DATA> implements Cloneable {
        /**
         * The default number of samples per parallel assignment instance.
         */
        public static final int DEFAULT_BLOCK_SIZE = 50000;

        /**
         * The default number of iterations.
         */
        public static final int DEFAULT_NUMBER_ITERATIONS = 30;

        /**
         * The number of clusters
         */
        protected int K;

        /**
         * The factory for producing the {@link NearestNeighbours} objects used in
         * assignment.
         */
        protected NearestNeighboursFactory<? extends NN, DATA> factory;

        /**
         * The size of processing blocks for each thread
         */
        protected int blockSize;

        /**
         * The max number of iterations
         */
        protected int niters;

        /**
         * The threadpool for parallel processing
         */
        protected ExecutorService threadpool;

        /**
         * Create configuration for data that will create <code>K</code> clusters.
         * The algorithm will run for a maximum of
         * {@link #DEFAULT_NUMBER_ITERATIONS} iterations, and make use of all
         * available processors, processing with blocks of
         * {@link #DEFAULT_BLOCK_SIZE} vectors.
         * <p>
         * The specified {@link NearestNeighboursFactory} determines the actual type
         * of k-means that will be performed; it could be exact nearest-neighbours,
         * or it could be an approximate method, for example based on an ensemble of
         * kd-trees.
         * 
         * @param K
         *            number of clusters to be found
         * @param nnFactory
         *            the factory for producing the {@link NearestNeighbours}.
         */
        public KMeansConfiguration(int K, NearestNeighboursFactory<? extends NN, DATA> nnFactory) {
                this(K, nnFactory, DEFAULT_NUMBER_ITERATIONS, DEFAULT_BLOCK_SIZE, GlobalExecutorPool.getPool());
        }

        /**
         * Create configuration for data that will create <code>K</code> clusters.
         * The algorithm will run for a maximum of the given number of iterations,
         * and will make use of all available processors, processing with blocks of
         * {@link #DEFAULT_BLOCK_SIZE} vectors.
         * <p>
         * The specified {@link NearestNeighboursFactory} determines the actual type
         * of k-means that will be performed; it could be exact nearest-neighbours,
         * or it could be an approximate method, for example based on an ensemble of
         * kd-trees.
         * 
         * @param K
         *            number of clusters to be found
         * @param nnFactory
         *            the factory for producing the {@link NearestNeighbours}.
         * @param niters
         *            number of iterations
         */
        public KMeansConfiguration(int K, NearestNeighboursFactory<? extends NN, DATA> nnFactory, int niters) {
                this(K, nnFactory, niters, DEFAULT_BLOCK_SIZE, GlobalExecutorPool.getPool());
        }

        /**
         * Create configuration for data that will create <code>K</code> clusters.
         * The algorithm will run for a maximum of the given number of iterations,
         * and will make use of the provided threadpool, processing with blocks of
         * {@link #DEFAULT_BLOCK_SIZE} vectors.
         * <p>
         * The specified {@link NearestNeighboursFactory} determines the actual type
         * of k-means that will be performed; it could be exact nearest-neighbours,
         * or it could be an approximate method, for example based on an ensemble of
         * kd-trees.
         * 
         * @param K
         *            number of clusters to be found
         * @param nnFactory
         *            the factory for producing the {@link NearestNeighbours}.
         * @param threadpool
         *            threadpool to use for parallel processing
         * @param niters
         *            number of training iterations
         */
        public KMeansConfiguration(int K, NearestNeighboursFactory<? extends NN, DATA> nnFactory, int niters,
                        ExecutorService threadpool)
        {
                this(K, nnFactory, niters, DEFAULT_BLOCK_SIZE, threadpool);
        }

        /**
         * Create configuration for data with <code>M</code> dimensions that will
         * create <code>K</code> clusters. The algorithm will run for a maximum of
         * the given number of iterations, and will make use of
         * <code>nThreads</code> processors, processing with blocks of
         * {@link #DEFAULT_BLOCK_SIZE} vectors.
         * <p>
         * The specified {@link NearestNeighboursFactory} determines the actual type
         * of k-means that will be performed; it could be exact nearest-neighbours,
         * or it could be an approximate method, for example based on an ensemble of
         * kd-trees.
         * 
         * @param K
         *            number of clusters to be found
         * @param nnFactory
         *            the factory for producing the {@link NearestNeighbours}.
         * @param threadpool
         *            threadpool to use for parallel processing
         * @param blockSize
         *            number of samples per parallel thread
         * @param niters
         *            number of training iterations
         */
        public KMeansConfiguration(int K, NearestNeighboursFactory<? extends NN, DATA> nnFactory, int niters,
                        int blockSize, ExecutorService threadpool)
        {
                this.K = K;
                this.factory = nnFactory;
                this.niters = niters;
                this.blockSize = blockSize;
                this.threadpool = (threadpool == null ? GlobalExecutorPool.getPool() : threadpool);
        }

        /**
         * A completely default configuration used primarily as a convenience
         * function for reading. The number of dimensions, number of clusters and
         * nearest-neighbours factory must be set before the configuration is used.
         */
        public KMeansConfiguration() {
                this(0, null);
        }

        /*
         * (non-Javadoc)
         * 
         * @see java.lang.Object#clone()
         */
        @SuppressWarnings("unchecked")
        @Override
        public KMeansConfiguration<NN, DATA> clone() {
                try {
                        return (KMeansConfiguration<NN, DATA>) super.clone();
                } catch (final CloneNotSupportedException e) {
                        throw new RuntimeException(e);
                }
        }

        /**
         * Get the number of clusters
         * 
         * @return the number of clusters
         */
        public int getK() {
                return K;
        }

        /**
         * Set the number of clusters
         * 
         * @param k
         *            the number of clusters
         */
        public void setK(int k) {
                K = k;
        }

        /**
         * Get the number of clusters
         * 
         * @return the number of clusters
         */
        public int numClusters() {
                return K;
        }

        /**
         * Set the number of clusters
         * 
         * @param k
         *            the number of clusters
         */
        public void setNumClusters(int k) {
                K = k;
        }

        /**
         * Get the number of samples processed in a batch by a thread. This needs to
         * be small enough that that the memory isn't exhausted, but big enough for
         * the thread to have enough data to work for a while.
         * 
         * @return the the number of samples processed in a batch by a thread
         */
        public int getBlockSize() {
                return blockSize;
        }

        /**
         * Set the number of samples processed in a batch by a thread. This needs to
         * be small enough that that the memory isn't exhausted, but big enough for
         * the thread to have enough data to work for a while.
         * 
         * @param blockSize
         *            the number of samples processed in a batch by a thread
         */
        public void setBlockSize(int blockSize) {
                this.blockSize = blockSize;
        }

        /**
         * Get the maximum allowed number of iterations.
         * 
         * @return the maximum allowed number of iterations.
         */
        public int getMaxIterations() {
                return niters;
        }

        /**
         * Set the maximum allowed number of iterations.
         * 
         * @param niters
         *            the maximum allowed number of iterations.
         */
        public void setMaxIterations(int niters) {
                this.niters = niters;
        }

        /**
         * Get the factory that produces the {@link NearestNeighbours} during
         * clustering.
         * 
         * @return the factory
         */
        public NearestNeighboursFactory<? extends NN, DATA> getNearestNeighbourFactory() {
                return factory;
        }

        /**
         * Set the factory that produces the {@link NearestNeighbours} during
         * clustering.
         * 
         * @param factory
         *            the factory to set
         */
        public void setNearestNeighbourFactory(NearestNeighboursFactory<? extends NN, DATA> factory) {
                this.factory = factory;
        }
}