clusterer.java

MacroWeka扩展了著名数据挖掘工具weka

/*
 *    This program 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 2 of the License, or
 *    (at your option) any later version.
 *
 *    This program 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 this program; if not, write to the Free Software
 *    Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*
 *    Clusterer.java
 *    Copyright (C) 2004 Stefan Mutter
 *
 */

package weka.gui.beans;


import java.util.Vector;
import java.util.Enumeration;
import java.util.Hashtable;
import javax.swing.JPanel;
import javax.swing.JLabel;
import javax.swing.JTextField;
import java.awt.BorderLayout;
import java.awt.event.MouseAdapter;
import java.awt.event.MouseEvent;
import java.awt.event.InputEvent;
import java.awt.*;
import java.io.Serializable;
import java.io.Reader;
import java.io.BufferedReader;
import java.io.FileReader;
import java.io.File;
import javax.swing.ImageIcon;
import javax.swing.SwingConstants;
import java.beans.EventSetDescriptor;

import weka.core.Instance;
import weka.core.Instances;
import weka.clusterers.*;
import weka.filters.Filter;
import weka.filters.unsupervised.attribute.Remove;
//import weka.classifiers.*;
//import weka.classifiers.rules.ZeroR;
import weka.gui.Logger;

/**
 * Bean that wraps around weka.clusterers
 *
 * @author <a href="mailto:[e-mail]">Stefan Mutter</a>
 * @version $Revision: 1.1 $
 * @see JPanel
 * @see BeanCommon
 * @see Visible
 * @see WekaWrapper
 * @see Serializable
 * @see UserRequestAcceptor
 * @see TrainingSetListener
 * @see TestSetListener
 */
public class Clusterer extends JPanel implements BeanCommon, Visible, WekaWrapper, EventConstraints, Serializable, UserRequestAcceptor, TrainingSetListener, TestSetListener{

  protected BeanVisual m_visual = 
    new BeanVisual("Clusterer",
		   BeanVisual.ICON_PATH+"EM.gif",
		   BeanVisual.ICON_PATH+"EM_animated.gif");

  private static int IDLE = 0;
  private static int BUILDING_MODEL = 1;
  private static int CLUSTERING = 2;

  private int m_state = IDLE;

  private Thread m_buildThread = null;

  /**
   * Global info for the wrapped classifier (if it exists).
   */
  protected String m_globalInfo;

  /**
   * Objects talking to us
   */
  private Hashtable m_listenees = new Hashtable();

  /**
   * Objects listening for batch clusterer events
   */
  private Vector m_batchClustererListeners = new Vector();


  /**
   * Objects listening for graph events
   */
  private Vector m_graphListeners = new Vector();

  /**
   * Objects listening for text events
   */
  private Vector m_textListeners = new Vector();

  /**
   * Holds training instances for batch training. 
   */
  private Instances m_trainingSet;
  private transient Instances m_testingSet;
  private weka.clusterers.Clusterer m_Clusterer = new EM();
  

  private transient Logger m_log = null;

  private Double m_dummy = new Double(0.0);

  /**
   * Global info (if it exists) for the wrapped classifier
   *
   * @return the global info
   */
  public String globalInfo() {
    return m_globalInfo;
  }

  /**
   * Creates a new <code>Clusterer</code> instance.
   */
  public Clusterer() {
    setLayout(new BorderLayout());
    add(m_visual, BorderLayout.CENTER);
    setClusterer(m_Clusterer);
  }

  /**
   * Set the clusterer for this wrapper
   *
   * @param c a <code>weka.clusterers.Clusterer</code> value
   */
  public void setClusterer(weka.clusterers.Clusterer c) {
    boolean loadImages = true;
    if (c.getClass().getName().
	compareTo(m_Clusterer.getClass().getName()) == 0) {
      loadImages = false;
    } else {
      // clusterer has changed so any batch training status is now
      // invalid
      m_trainingSet = null;
    }
    m_Clusterer = c;
    String clustererName = c.getClass().toString();
    clustererName = clustererName.substring(clustererName.
					      lastIndexOf('.')+1, 
					      clustererName.length());
    if (loadImages) {
      if (!m_visual.loadIcons(BeanVisual.ICON_PATH+clustererName+".gif",
		       BeanVisual.ICON_PATH+clustererName+"_animated.gif")) {
	useDefaultVisual();
      }
    }
    m_visual.setText(clustererName);

    // get global info
    m_globalInfo = KnowledgeFlowApp.getGlobalInfo(m_Clusterer);
  }


  /**
   * Returns true if this clusterer has an incoming connection that is
   * a batch set of instances
   *
   * @return a <code>boolean</code> value
   */
  public boolean hasIncomingBatchInstances() {
    if (m_listenees.size() == 0) {
      return false;
    }
    if (m_listenees.containsKey("trainingSet") ||
	m_listenees.containsKey("testSet") ||
        m_listenees.containsKey("dataSet")) {
      return true;
    }
    return false;
  }

  /**
   * Get the clusterer currently set for this wrapper
   *
   * @return a <code>weka.clusterers.Clusterer</code> value
   */
  public weka.clusterers.Clusterer getClusterer() {
    return m_Clusterer;
  }

  /**
   * Sets the algorithm (clusterer) for this bean
   *
   * @param algorithm an <code>Object</code> value
   * @exception IllegalArgumentException if an error occurs
   */
  public void setWrappedAlgorithm(Object algorithm) 
    {

    if (!(algorithm instanceof weka.clusterers.Clusterer)) { 
      throw new IllegalArgumentException(algorithm.getClass()+" : incorrect "
					 +"type of algorithm (Clusterer)");
    }
    setClusterer((weka.clusterers.Clusterer)algorithm);
  }

  /**
   * Returns the wrapped clusterer
   *
   * @return an <code>Object</code> value
   */
  public Object getWrappedAlgorithm() {
    return getClusterer();
  }


  /**
   * Accepts a training set and builds batch clusterer
   *
   * @param e a <code>TrainingSetEvent</code> value
   */
  public void acceptTrainingSet(final TrainingSetEvent e) {
    if (e.isStructureOnly()) {
      // no need to build a clusterer, instead just generate a dummy
      // BatchClustererEvent in order to pass on instance structure to
      // any listeners 
      BatchClustererEvent ce = 
	new BatchClustererEvent(this, m_Clusterer, 
				 new DataSetEvent(this, e.getTrainingSet()),
				 e.getSetNumber(), e.getMaxSetNumber(),1);

      notifyBatchClustererListeners(ce);
      return;
    }
    if (m_buildThread == null) {
      try {
	if (m_state == IDLE) {
	  synchronized (this) {
	    m_state = BUILDING_MODEL;
	  }
	  m_trainingSet = e.getTrainingSet();
	  final String oldText = m_visual.getText();
	  m_buildThread = new Thread() {
	      public void run() {
		try {
		  if (m_trainingSet != null) {  
		    m_visual.setAnimated();
		    m_visual.setText("Building clusters...");
		    if (m_log != null) {
		      m_log.statusMessage("Clusterer : building clusters...");
		    }
		    buildClusterer();
                    if(m_batchClustererListeners.size() > 0){
                        BatchClustererEvent ce = 
                            new BatchClustererEvent(this, m_Clusterer, 
				 new DataSetEvent(this, e.getTrainingSet()),
				 e.getSetNumber(), e.getMaxSetNumber(),1);
                        notifyBatchClustererListeners(ce);
                    }
		    if (m_Clusterer instanceof weka.core.Drawable && 
			m_graphListeners.size() > 0) {
		      String grphString = 
			((weka.core.Drawable)m_Clusterer).graph();
                      int grphType = ((weka.core.Drawable)m_Clusterer).graphType();
		      String grphTitle = m_Clusterer.getClass().getName();
		      grphTitle = grphTitle.substring(grphTitle.
						      lastIndexOf('.')+1, 
						      grphTitle.length());
		      grphTitle = "Set " + e.getSetNumber() + " ("
			+e.getTrainingSet().relationName() + ") "
			+grphTitle;
		      
		      GraphEvent ge = new GraphEvent(Clusterer.this, 
						     grphString, 
						     grphTitle,
                                                     grphType);
		      notifyGraphListeners(ge);
		    }

		    if (m_textListeners.size() > 0) {
		      String modelString = m_Clusterer.toString();
		      String titleString = m_Clusterer.getClass().getName();
		      
		      titleString = titleString.
			substring(titleString.lastIndexOf('.') + 1,
				  titleString.length());
		      modelString = "=== Clusterer model ===\n\n" +
			"Scheme:   " +titleString+"\n" +
			"Relation: "  + m_trainingSet.relationName() + 
			((e.getMaxSetNumber() > 1) 
			 ? "\nTraining Fold: "+e.getSetNumber()
			 :"")
			+ "\n\n"
			+ modelString;
		      titleString = "Model: " + titleString;

		      TextEvent nt = new TextEvent(Clusterer.this,
						   modelString,
						   titleString);
		      notifyTextListeners(nt);
		    }
		  }
		} catch (Exception ex) {
		  ex.printStackTrace();
		} finally {
		  m_visual.setText(oldText);
		  m_visual.setStatic();
		  m_state = IDLE;
		  if (isInterrupted()) {
		    // prevent any clusterer events from being fired
		    m_trainingSet = null;
		    if (m_log != null) {
		      m_log.logMessage("Build clusterer interrupted!");
		      m_log.statusMessage("OK");
		    }
		  } else {
		    // save header
		    m_trainingSet = new Instances(m_trainingSet, 0);
		  }
		  if (m_log != null) {
		    m_log.statusMessage("OK");
		  }
		  block(false);
		}
	      }	
	    };
	  m_buildThread.setPriority(Thread.MIN_PRIORITY);
	  m_buildThread.start();
	  // make sure the thread is still running before we block
	  //	  if (m_buildThread.isAlive()) {
	  block(true);
	    //	  }
	  m_buildThread = null;
	  m_state = IDLE;
	}
      } catch (Exception ex) {
	ex.printStackTrace();
      }
    }
  }

  /**
   * Accepts a test set for a batch trained clusterer
   *
   * @param e a <code>TestSetEvent</code> value
   */
  public void acceptTestSet(TestSetEvent e) {

    if (m_trainingSet != null) {
      try {
	if (m_state == IDLE) {
	  synchronized(this) {
	    m_state = CLUSTERING;
	  }
          m_testingSet = e.getTestSet();
	  if (m_trainingSet.equalHeaders(m_testingSet)) {
	    BatchClustererEvent ce = 
	      new BatchClustererEvent(this, m_Clusterer, 
				       new DataSetEvent(this, e.getTestSet()),
				  e.getSetNumber(), e.getMaxSetNumber(),0);

	    notifyBatchClustererListeners(ce);
	    
	  }
	  m_state = IDLE;
	}
      } catch (Exception ex) {