//
//  OpenForecast - open source, general-purpose forecasting package.
//  Copyright (C) 2002-2004  Steven R. Gould. GNU public license 
//
import java.util.Date;
import java.util.Iterator;

import org.jfree.chart.ChartFactory;
import org.jfree.chart.ChartPanel;
import org.jfree.chart.JFreeChart;
import org.jfree.chart.renderer.xy.StandardXYItemRenderer;
import org.jfree.chart.renderer.xy.XYItemRenderer;
import org.jfree.chart.plot.XYPlot;
import org.jfree.data.time.TimeSeries;
import org.jfree.data.time.Quarter;
import org.jfree.data.time.TimeSeriesDataItem;
import org.jfree.data.xy.XYDataset;
import org.jfree.data.time.TimeSeriesCollection;
import org.jfree.ui.ApplicationFrame;

import net.sourceforge.openforecast.DataPoint;
import net.sourceforge.openforecast.DataSet;
import net.sourceforge.openforecast.Forecaster;
import net.sourceforge.openforecast.ForecastingModel;
import net.sourceforge.openforecast.Observation;
import net.sourceforge.openforecast.models.SimpleExponentialSmoothingModel;
import net.sourceforge.openforecast.models.DoubleExponentialSmoothingModel;
import net.sourceforge.openforecast.models.TripleExponentialSmoothingModel;


/**
 * An example of a time series chart, showing an initial series of
 * observations (the first data series), overlaid with forecasts produced
 * using a variety of different forecasting models. This both demonstrates
 * the use of JFreeChart, as well as provides a graphical demonstration of
 * single exponential smoothing. It clearly shows the weakness of single
 * exponential smoothing when applied to a series of points exhibiting both
 * a trend and seasonality. When double and triple exponential smoothing
 * have been implemented, they should be added to this demo.
 */
public class ExponentialSmoothingChartDemo extends ApplicationFrame
{
    /** The set of data points for which forecast values are required. */
    private TimeSeries fc;
    
    /**
     * A demonstration application showing a quarterly time series
     * along with the forecast values.
     * @param title the frame title.
     */
    public ExponentialSmoothingChartDemo(String title)
    {
        super(title);
        
        // Create a title...
        String chartTitle = "Quarterly time series with forecast";
        XYDataset dataset = createDataset();
        
        JFreeChart chart
            = ChartFactory.createTimeSeriesChart(chartTitle,
                                                 "Date",
                                                 "Quarterly Sales (Units sold)",
                                                 dataset,
                                                 true,  // Legend
                                                 true,  // Tooltips
                                                 false);// URLs
        
        XYPlot plot = chart.getXYPlot();
        XYItemRenderer renderer = plot.getRenderer();
        if (renderer instanceof StandardXYItemRenderer)
            {
                StandardXYItemRenderer r = (StandardXYItemRenderer) renderer;
                r.setBaseShapesVisible(true);
                r.setShapesFilled(Boolean.TRUE);
            }
        
        ChartPanel chartPanel = new ChartPanel(chart);
        chartPanel.setPreferredSize(new java.awt.Dimension(600, 500));
        setContentPane(chartPanel);
    }
    
    /**
     * Creates a dataset, consisting of two series of monthly data.
     * @return the dataset.
     */
    public XYDataset createDataset()
    {
        TimeSeries observations
            = new TimeSeries("Quarterly Sales", Quarter.class);
        
        observations.add(new Quarter(1,1990), 362.0);
        observations.add(new Quarter(2,1990), 385.0);
        observations.add(new Quarter(3,1990), 432.0);
        observations.add(new Quarter(4,1990), 341.0);
        observations.add(new Quarter(1,1991), 382.0);
        observations.add(new Quarter(2,1991), 409.0);
        observations.add(new Quarter(3,1991), 498.0);
        observations.add(new Quarter(4,1991), 387.0);
        observations.add(new Quarter(1,1992), 473.0);
        observations.add(new Quarter(2,1992), 513.0);
        observations.add(new Quarter(3,1992), 582.0);
        observations.add(new Quarter(4,1992), 474.0);
        observations.add(new Quarter(1,1993), 544.0);
        observations.add(new Quarter(2,1993), 582.0);
        observations.add(new Quarter(3,1993), 681.0);
        observations.add(new Quarter(4,1993), 557.0);
        observations.add(new Quarter(1,1994), 628.0);
        observations.add(new Quarter(2,1994), 707.0);
        observations.add(new Quarter(3,1994), 773.0);
        observations.add(new Quarter(4,1994), 592.0);
        observations.add(new Quarter(1,1995), 627.0);
        observations.add(new Quarter(2,1995), 725.0);
        observations.add(new Quarter(3,1995), 854.0);
        observations.add(new Quarter(4,1995), 661.0);
        
        fc = new TimeSeries("Forecast values", Quarter.class);
        fc.add(new Quarter(1,1990), 0.0);
        fc.add(new Quarter(2,1990), 0.0);
        fc.add(new Quarter(3,1990), 0.0);
        fc.add(new Quarter(4,1990), 0.0);
        fc.add(new Quarter(1,1991), 0.0);
        fc.add(new Quarter(2,1991), 0.0);
        fc.add(new Quarter(3,1991), 0.0);
        fc.add(new Quarter(4,1991), 0.0);
        fc.add(new Quarter(1,1992), 0.0);
        fc.add(new Quarter(2,1992), 0.0);
        fc.add(new Quarter(3,1992), 0.0);
        fc.add(new Quarter(4,1992), 0.0);
        fc.add(new Quarter(1,1993), 0.0);
        fc.add(new Quarter(2,1993), 0.0);
        fc.add(new Quarter(3,1993), 0.0);
        fc.add(new Quarter(4,1993), 0.0);
        fc.add(new Quarter(1,1994), 0.0);
        fc.add(new Quarter(2,1994), 0.0);
        fc.add(new Quarter(3,1994), 0.0);
        fc.add(new Quarter(4,1994), 0.0);
        fc.add(new Quarter(1,1995), 0.0);
        fc.add(new Quarter(2,1995), 0.0);
        fc.add(new Quarter(3,1995), 0.0);
        fc.add(new Quarter(4,1995), 0.0);
        fc.add(new Quarter(1,1996), 0.0);
        fc.add(new Quarter(2,1996), 0.0);
        fc.add(new Quarter(3,1996), 0.0);
        fc.add(new Quarter(4,1996), 0.0);
        fc.add(new Quarter(1,1997), 0.0);
        fc.add(new Quarter(2,1997), 0.0);
        fc.add(new Quarter(3,1997), 0.0);
        fc.add(new Quarter(4,1997), 0.0);
        fc.add(new Quarter(1,1998), 0.0);
        fc.add(new Quarter(2,1998), 0.0);
        fc.add(new Quarter(3,1998), 0.0);
        fc.add(new Quarter(4,1998), 0.0);
        
        DataSet initDataSet = getDataSet( observations, 0, 100 );
        initDataSet.setTimeVariable( "t" );
        initDataSet.setPeriodsPerYear( 4 );
        
        // Get "best fit" simple exponential smoothing model
        ForecastingModel sesModel
            = SimpleExponentialSmoothingModel.getBestFitModel(initDataSet);
        TimeSeries sesSeries
            = getForecastTimeSeries(sesModel,
                                    initDataSet,
                                    0, 30,
                                    "Simple Exponential Smoothing");
        
        // Get "best fit" double exponential smoothing model
        ForecastingModel desModel
            = DoubleExponentialSmoothingModel.getBestFitModel(initDataSet);
        TimeSeries desSeries
            = getForecastTimeSeries(desModel,
                                    initDataSet,
                                    0, 30,
                                    "Double Exponential Smoothing");
        
        // Get "best fit" triple exponential smoothing model
        ForecastingModel tesModel
            = TripleExponentialSmoothingModel.getBestFitModel(initDataSet);
        TimeSeries tesSeries
            = getForecastTimeSeries(tesModel,
                                    initDataSet,
                                    5, 28,
                                    "Triple Exponential Smoothing");

        TimeSeriesCollection dataset = new TimeSeriesCollection();
        dataset.addSeries(observations);
        dataset.addSeries(sesSeries);
        dataset.addSeries(desSeries);
        dataset.addSeries(tesSeries);
        
        return dataset;
    }
    
    /**
     * A helper function to convert data points (from startIndex to
     * endIndex) of a (JFreeChart) TimeSeries object into an
     * OpenForecast DataSet.
     * @param series the series of data points stored as a JFreeChart
     * TimeSeries object.
     * @param startIndex the index of the first data point required from the
     * series.
     * @param endIndex the index of the last data point required from the
     * series.
     * @return an OpenForecast DataSet representing the data points extracted
     * from the TimeSeries.
     */
    private DataSet getDataSet( TimeSeries series,
                                int startIndex, int endIndex )
    {
        DataSet dataSet = new DataSet();
        if ( endIndex > series.getItemCount() )
            endIndex = series.getItemCount();
        
        for ( int i=startIndex; i