preprocesspanel.java

Java 编写的多种数据挖掘算法 包括聚类、分类、预处理等

	  m_EditBut.setEnabled(true);
	  m_Log.statusMessage("OK");
	  // Fire a propertychange event
	  m_Support.firePropertyChange("", null, null);
	}
      };
      if (SwingUtilities.isEventDispatchThread()) {
	r.run();
      } else {
	SwingUtilities.invokeAndWait(r);
      }
    } catch (Exception ex) {
      JOptionPane.showMessageDialog(this,
				    "Problem setting base instances:\n"
				    + ex.getMessage(),
				    "Instances",
				    JOptionPane.ERROR_MESSAGE);
    }
  }

  /**
   * Gets the working set of instances.
   *
   * @return the working instances
   */
  public Instances getInstances() {

    return m_Instances;
  }
  
  /**
   * Adds a PropertyChangeListener who will be notified of value changes.
   *
   * @param l a value of type 'PropertyChangeListener'
   */
  public void addPropertyChangeListener(PropertyChangeListener l) {

    m_Support.addPropertyChangeListener(l);
  }

  /**
   * Removes a PropertyChangeListener.
   *
   * @param l a value of type 'PropertyChangeListener'
   */
  public void removePropertyChangeListener(PropertyChangeListener l) {

    m_Support.removePropertyChangeListener(l);
  }
  
  /**
   * Passes the dataset through the filter that has been configured for use.
   */
  protected void applyFilter(final Filter filter) {

    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	public void run() {
	  try {

	    if (filter != null) {
	    
	      if (m_Log instanceof TaskLogger) {
		((TaskLogger)m_Log).taskStarted();
	      }
	      m_Log.statusMessage("Passing dataset through filter "
				  + filter.getClass().getName());
	      int classIndex = m_AttVisualizePanel.getColoringIndex();
	      if ((classIndex < 0) && (filter instanceof SupervisedFilter)) {
		throw new IllegalArgumentException("Class (colour) needs to " +
						   "be set for supervised " +
						   "filter.");
	      }
	      Instances copy = new Instances(m_Instances);
	      copy.setClassIndex(classIndex);
	      filter.setInputFormat(copy);
	      Instances newInstances = filter.useFilter(copy, filter);
	      if (newInstances == null || newInstances.numAttributes() < 1) {
		throw new Exception("Dataset is empty.");
	      }
	      m_Log.statusMessage("Saving undo information");
	      addUndoPoint();
	      m_AttVisualizePanel.setColoringIndex(copy.classIndex());
	      // if class was not set before, reset it again after use of filter
	      if (m_Instances.classIndex() < 0)
		newInstances.setClassIndex(-1);
	      m_Instances = newInstances;
	      setInstances(m_Instances);
	      if (m_Log instanceof TaskLogger) {
		((TaskLogger)m_Log).taskFinished();
	      }
	    }
	    
	  } catch (Exception ex) {
	
	    if (m_Log instanceof TaskLogger) {
	      ((TaskLogger)m_Log).taskFinished();
	    }
	    // Pop up an error optionpane
	    JOptionPane.showMessageDialog(PreprocessPanel.this,
					  "Problem filtering instances:\n"
					  + ex.getMessage(),
					  "Apply Filter",
					  JOptionPane.ERROR_MESSAGE);
	    m_Log.logMessage("Problem filtering instances: " + ex.getMessage());
	    m_Log.statusMessage("Problem filtering instances");
	  }
	  m_IOThread = null;
	}
      };
      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't apply filter at this time,\n"
				    + "currently busy with other IO",
				    "Apply Filter",
				    JOptionPane.WARNING_MESSAGE);
    }
  }

  /**
   * Queries the user for a file to save instances as, then saves the
   * instances in a background process. This is done in the IO
   * thread, and an error message is popped up if the IO thread is busy.
   */
  public void saveWorkingInstancesToFileQ() {
    
    if (m_IOThread == null) {
      m_FileChooser.setAcceptAllFileFilterUsed(false);
      int returnVal = m_FileChooser.showSaveDialog(this);
      if (returnVal == JFileChooser.APPROVE_OPTION) {
	File sFile = m_FileChooser.getSelectedFile();
	if (m_FileChooser.getFileFilter() == m_arffFileFilter) {
	  if (!sFile.getName().toLowerCase().endsWith(Instances.FILE_EXTENSION)) {
	    sFile = new File(sFile.getParent(), sFile.getName() 
			     + Instances.FILE_EXTENSION);
	  }
	  File selected = sFile;
	  saveInstancesToFile(selected, m_Instances, true);
	} else if (m_FileChooser.getFileFilter() == m_csvFileFilter) {
	  if (!sFile.getName().toLowerCase().endsWith(CSVLoader.FILE_EXTENSION)) {
	    sFile = new File(sFile.getParent(), sFile.getName() 
			     + CSVLoader.FILE_EXTENSION);
	  }
	  File selected = sFile;
	  saveInstancesToFile(selected, m_Instances, false);
	} else if (m_FileChooser.getFileFilter() == m_c45FileFilter) {	 
	  File selected = sFile;
	  saveInstancesToC45File(selected, m_Instances);
	} else if (m_FileChooser.getFileFilter() == m_bsiFileFilter) {
	  if (!sFile.getName().toLowerCase().
	      endsWith(Instances.SERIALIZED_OBJ_FILE_EXTENSION)) {
	    sFile = new File(sFile.getParent(), sFile.getName() 
			     + Instances.SERIALIZED_OBJ_FILE_EXTENSION);
	  }
	  File selected = sFile;
	  saveSerializedInstancesToFile(selected, m_Instances);
	}
      }
      FileFilter temp = m_FileChooser.getFileFilter();
      m_FileChooser.setAcceptAllFileFilterUsed(true);
      m_FileChooser.setFileFilter(temp);
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't save at this time,\n"
				    + "currently busy with other IO",
				    "Save Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
  
  /**
   * Queries the user for a file to load instances from, then loads the
   * instances in a background process. This is done in the IO
   * thread, and an error message is popped up if the IO thread is busy.
   */
  public void setInstancesFromFileQ() {
    
    if (m_IOThread == null) {
      int returnVal = m_FileChooser.showOpenDialog(this);
      if (returnVal == JFileChooser.APPROVE_OPTION) {
	File selected = m_FileChooser.getSelectedFile();
	try {
	  addUndoPoint();
	} catch (Exception ignored) {}
	setInstancesFromFile(selected);
      }
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }

  /**
   * Loads instances from an SQL query the user provided with the
   * SqlViewerDialog, then loads the instances in a background process. This is
   * done in the IO thread, and an error message is popped up if the IO thread
   * is busy.
   * @param url           the database URL
   * @param user          the user to connect as
   * @param pw            the password of the user
   * @param query         the query for retrieving instances from
   */
  public void setInstancesFromDBQ(String url, String user, 
                                  String pw, String query) {
    if (m_IOThread == null) {
      try {
	InstanceQuery InstQ = new InstanceQuery();
        InstQ.setDatabaseURL(url);
        InstQ.setUsername(user);
        InstQ.setPassword(pw);
        InstQ.setQuery(query);
	
        // we have to disconnect, otherwise we can't change the DB!
        if (InstQ.isConnected())
          InstQ.disconnectFromDatabase();

	InstQ.connectToDatabase();      
	try {
	  addUndoPoint();
	} catch (Exception ignored) {}
	setInstancesFromDB(InstQ);
      } catch (Exception ex) {
	JOptionPane.showMessageDialog(this,
				      "Problem connecting to database:\n"
				      + ex.getMessage(),
				      "Load Instances",
				      JOptionPane.ERROR_MESSAGE);
      }
      
    } else {
      JOptionPane.showMessageDialog(this,
				     "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
    
  /**
   * Queries the user for a URL to load instances from, then loads the
   * instances in a background process. This is done in the IO
   * thread, and an error message is popped up if the IO thread is busy.
   */
  public void setInstancesFromURLQ() {
    
    if (m_IOThread == null) {
      try {
	String urlName = (String) JOptionPane.showInputDialog(this,
			"Enter the source URL",
			"Load Instances",
			JOptionPane.QUESTION_MESSAGE,
			null,
			null,
			m_LastURL);
	if (urlName != null) {
	  m_LastURL = urlName;
	  URL url = new URL(urlName);
	  try {
	    addUndoPoint();
	  } catch (Exception ignored) {}
	  setInstancesFromURL(url);
	}
      } catch (Exception ex) {
	ex.printStackTrace();
	JOptionPane.showMessageDialog(this,
				      "Problem with URL:\n"
				      + ex.getMessage(),
				      "Load Instances",
				      JOptionPane.ERROR_MESSAGE);
      }
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
  
  /**
   * sets Instances generated via DataGenerators (pops up a Dialog)
   */
  public void generateInstances() {
    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	  public void run() {
	    try {
              // create dialog
              final DataGeneratorPanel generatorPanel = new DataGeneratorPanel();
              final JDialog dialog = new JDialog();
              final JButton generateButton = new JButton("Generate");
              final JCheckBox showOutputCheckBox = 
                                  new JCheckBox("Show generated data as text, incl. comments");

              showOutputCheckBox.setMnemonic('S');
              generatorPanel.setGenerator(m_DataGenerator);
              generatorPanel.setPreferredSize(
                  new Dimension(
                        300, 
                        (int) generatorPanel.getPreferredSize().getHeight()));
              generateButton.setMnemonic('G');
              generateButton.setToolTipText("Generates the dataset according the settings.");
              generateButton.addActionListener(new ActionListener(){
                  public void actionPerformed(ActionEvent evt){
                    // generate
                    generatorPanel.execute();
                    boolean generated = (generatorPanel.getInstances() != null);
                    if (generated)
                      setInstances(generatorPanel.getInstances());

                    // close dialog
                    dialog.dispose();

                    // get last generator
                    m_DataGenerator = generatorPanel.getGenerator();

                    // display output?
                    if ( (generated) && (showOutputCheckBox.isSelected()) )
                      showGeneratedInstances(generatorPanel.getOutput());
                }
              });
              dialog.setTitle("DataGenerator");
              dialog.getContentPane().add(generatorPanel, BorderLayout.CENTER);
              dialog.getContentPane().add(generateButton, BorderLayout.EAST);
              dialog.getContentPane().add(showOutputCheckBox, BorderLayout.SOUTH);
              dialog.pack();
              
              // display dialog
              dialog.setVisible(true);
	    } 
            catch (Exception ex) {
	      ex.printStackTrace();
	      m_Log.logMessage(ex.getMessage());
	    }
	    m_IOThread = null;
	  }
	};
      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    } 
    else {
      JOptionPane.showMessageDialog(this,
				    "Can't generate data at this time,\n"
				    + "currently busy with other IO",
				    "Generate Data",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
  
  /**
   * displays a dialog with the generated instances from the DataGenerator
   */
  protected void showGeneratedInstances(String data) {
    final JDialog dialog = new JDialog();
    final JButton saveButton = new JButton("Save");
    final JButton closeButton = new JButton("Close");
    final JTextArea textData = new JTextArea(data);
    final JPanel panel = new JPanel();
    panel.setLayout(new FlowLayout(FlowLayout.RIGHT));
    textData.setEditable(false);
    textData.setFont(
        new Font("Monospaced", Font.PLAIN, textData.getFont().getSize()));

    saveButton.setMnemonic('S');
    saveButton.setToolTipText("Saves the output to a file");
    saveButton.addActionListener(new ActionListener(){
      public void actionPerformed(ActionEvent evt){
        JFileChooser filechooser = new JFileChooser();
        int result = filechooser.showSaveDialog(dialog);
        if (result == JFileChooser.APPROVE_OPTION) {
          try {
            BufferedWriter writer = new BufferedWriter(
                                      new FileWriter(
                                        filechooser.getSelectedFile()));
            writer.write(textData.getText());
            writer.flush();
            writer.close();
            JOptionPane.showMessageDialog(
              dialog, 
              "Output successfully saved to file '" 
              + filechooser.getSelectedFile() + "'!",
              "Information",
              JOptionPane.INFORMATION_MESSAGE);
          }
          catch (Exception e) {
            e.printStackTrace();
          }
          dialog.dispose();
        }
      }
    });
    closeButton.setMnemonic('C');
    closeButton.setToolTipText("Closes the dialog");
    closeButton.addActionListener(new ActionListener(){
      public void actionPerformed(ActionEvent evt){
        dialog.dispose();
      }
    });
    panel.add(saveButton);
    panel.add(closeButton);
    dialog.setTitle("Generated Instances (incl. comments)");
    dialog.getContentPane().add(new JScrollPane(textData), BorderLayout.CENTER);
    dialog.getContentPane().add(panel, BorderLayout.SOUTH);
    dialog.pack();

    // make sure, it's not bigger than 80% of the screen
    Dimension screen = Toolkit.getDefaultToolkit().getScreenSize();
    int width  = dialog.getWidth() > screen.getWidth()*0.8
                    ? (int) (screen.getWidth()*0.8) : dialog.getWidth();
    int height = dialog.getHeight() > screen.getHeight()*0.8 
                    ? (int) (screen.getHeight()*0.8) : dialog.getHeight();
    dialog.setSize(width, height);
    
    // display dialog
    dialog.setVisible(true);
  }

  /**
   * Saves the current instances in C45 names and data file format
   *
   * @param f a value of type 'File'
   * @param inst the instances to save
   */
  protected void saveInstancesToC45File(final File f, final Instances inst) {
    if (m_IOThread == null) {
      final int classIndex = m_AttVisualizePanel.getColoringIndex();
      if (inst.attribute(classIndex).isNumeric()) {	      
	JOptionPane.showMessageDialog(this,
				      "Can't save in C45 format,\n"
				      + "as the selected class is numeric.",
				      "Save Instances",
				      JOptionPane.ERROR_MESSAGE);
	return;
      }
      m_IOThread = new Thread() {
	  public void run() {
	    try {
	      m_Log.statusMessage("Saving to file...");
	      String name = f.getAbsolutePath();
	      if (name.lastIndexOf('.') != -1) {
		name = name.substring(0, name.lastIndexOf('.'));
	      }
	      File fData = new File(name+".data");
	      File fNames = new File(name+".names");
	      Writer w = new BufferedWriter(new FileWriter(fNames));
	      Writer w2 =  new BufferedWriter(new FileWriter(fData));	      
	      
	      // write the names file
	      for (int i = 0; i < inst.attribute(classIndex).numValues(); i++) {
		w.write(inst.attribute(classIndex).value(i));
		if (i < inst.attribute(classIndex).numValues()-1) {
		  w.write(",");
		} else {
		  w.write(".\n");
		}
	      }
	      for (int i = 0; i < inst.numAttributes(); i++) {
		if (i != classIndex) {
		  w.write(inst.attribute(i).name()+": ");
		  if (inst.attribute(i).isNumeric() || inst.attribute(i).isDate()) {
		    w.write("continuous.\n");
		  } else {
		    Attribute temp = inst.attribute(i);
		    for (int j = 0; j < temp.numValues(); j++) {
		      w.write(temp.value(j));
		      if (j < temp.numValues()-1) {
			w.write(",");
		      } else {
			w.write(".\n");
		      }
		    }
		  }
		}
	      }
	      w.close();