preprocesspanel.java

:<<数据挖掘--实用机器学习技术及java实现>>一书的配套源程序

	m_Log.statusMessage("Passing through filter " + (i + 1) + ": "
			    + filters[i].getClass().getName());
	filters[i].setInputFormat(temp);
	temp = Filter.useFilter(temp, filters[i]);
      }
      if (m_Log instanceof TaskLogger) {
	((TaskLogger)m_Log).taskFinished();
      }
      // try to save a copy of the filters using serialization
      try {
	m_FiltersCopy = new SerializedObject(filters);
      } catch (Exception ex) {
	JOptionPane.showMessageDialog(this,
				      "Could not create copy of filters",
				      null,
				      JOptionPane.ERROR_MESSAGE);
      }
      return temp;
    } catch (UnassignedClassException ex) {
      // Pop up an error optionpane
      JOptionPane.showMessageDialog(this,
				    "Problem filtering instances:\n"
				    + ex.getMessage() + "\n\n"
                                    +"You should not use filters that require "
                                    +"a class attribute during preprocessing,\n"
                                    +"as this will bias any classification "
                                    +"results you obtain. Consider using a\n"
                                    +"FilteredClassifier instead.",
				    "Instance Filter",
				    JOptionPane.ERROR_MESSAGE);
      return null;
    } catch (Exception ex) {
      // Pop up an error optionpane
      JOptionPane.showMessageDialog(this,
				    "Problem filtering instances:\n"
				    + ex.getMessage(),
				    "Instance Filter",
				    JOptionPane.ERROR_MESSAGE);
      ex.printStackTrace();
      m_Log.logMessage(ex.getMessage());
      return null;
    }
  }
  
  /**
   * Applies the current filters and attribute selection settings and
   * sets the result as the working dataset. This is done in the IO
   * thread, and an error message is popped up if the IO thread is busy.
   */
  public void setWorkingInstancesFromFilters() {
    if (m_IOThread == null) {
      int res = 0;
      if (m_AttPanel.getSelectedAttributes().length == 0) {
	Object[] options = { "OK", "CANCEL" };
	res = JOptionPane.showOptionDialog(null, 
				     "Unselecting all attributes will result "
				     +"in an empty dataset.\nDo you want to "
				     +"continue?", "Warning", 
				     JOptionPane.DEFAULT_OPTION, 
				     JOptionPane.WARNING_MESSAGE,
				     null, options, options[1]);
	
      }
      if (res == 0) {
	m_IOThread = new Thread() {
	    public void run() {
	      Instances f = filterInstances(m_BaseInstances);
	      if (f != null) {
		if ((f == m_BaseInstances) && (f == m_WorkingInstances)) {
		  m_Log.logMessage("No changes from base relation");
		} else {
		  setWorkingInstances(f);
		}
	      }
	      m_IOThread = null;
	    }
	  };
	m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
	m_IOThread.start();
      }
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't apply filters at this time,\n"
				    + "currently busy with other IO",
				    "Apply filters",
				    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) {
      int returnVal = m_FileChooser.showSaveDialog(this);
      if (returnVal == JFileChooser.APPROVE_OPTION) {
	File sFile = m_FileChooser.getSelectedFile();
	if (!sFile.getName().toLowerCase().endsWith(Instances.FILE_EXTENSION)) {
	  sFile = new File(sFile.getParent(), sFile.getName() 
                           + Instances.FILE_EXTENSION);
	}
	File selected = sFile;
	saveInstancesToFile(selected, m_WorkingInstances);
      }
    } 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 setBaseInstancesFromFileQ() {
    
    if (m_IOThread == null) {
      int returnVal = m_FileChooser.showOpenDialog(this);
      if (returnVal == JFileChooser.APPROVE_OPTION) {
	File selected = m_FileChooser.getSelectedFile();
	setBaseInstancesFromFile(selected);
      }
    } 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 a database 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 setBaseInstancesFromDBQ() {
    if (m_IOThread == null) {
      try {
	InstanceQuery InstQ = 
	  (InstanceQuery)m_DatabaseQueryEditor.getValue();
	
	InstQ.connectToDatabase();      

	setBaseInstancesFromDB(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 setBaseInstancesFromURLQ() {
    
    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);
	  setBaseInstancesFromURL(url);
	}
      } catch (Exception ex) {
	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);
    }
  }
  
  
  /**
   * Saves the filtered instances to the supplied file.
   *
   * @param f a value of type 'File'
   * @param inst the instances to save
   */
  protected void saveInstancesToFile(final File f, final Instances inst) {
      
    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	public void run() {
	  try {
	    m_Log.statusMessage("Saving to file...");
	    Writer w = new BufferedWriter(new FileWriter(f));
	    Instances h = new Instances(inst, 0);
	    w.write(h.toString());
	    w.write("\n");
	    for (int i = 0; i < inst.numInstances(); i++) {
	      w.write(inst.instance(i).toString());
	      w.write("\n");
	    }
	    w.close();
	    m_Log.statusMessage("OK");
	  } 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 save at this time,\n"
				    + "currently busy with other IO",
				    "Save Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }

  /**
   * Pops up generic object editor with list of conversion filters
   *
   * @param f the File
   */
  private void converterQuery(final File f) {
    final GenericObjectEditor convEd = new GenericObjectEditor();

    try {
      convEd.setClassType(weka.core.converters.Loader.class);
      convEd.setValue(new weka.core.converters.CSVLoader());
      ((GenericObjectEditor.GOEPanel)convEd.getCustomEditor())
	.addOkListener(new ActionListener() {
	    public void actionPerformed(ActionEvent e) {
	      tryConverter((Loader)convEd.getValue(), f);
	    }
	  });
    } catch (Exception ex) {
    }

    PropertyDialog pd = new PropertyDialog(convEd, 100, 100);
  }

  /**
   * Applies the selected converter
   *
   * @param cnv the converter to apply to the input file
   * @param f the input file
   */
  private void tryConverter(final Loader cnv, final File f) {

    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	  public void run() {
	    try {
	      cnv.setSource(f);
	      Instances inst = cnv.getDataSet();
	      setBaseInstances(inst);
	    } catch (Exception ex) {
	      m_Log.statusMessage(cnv.getClass().getName()+" failed to load "
				 +f.getName());
	      m_IOThread = null;
	      converterQuery(f);
	    }
	    m_IOThread = null;
	  }
	};
      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    }
  }

  /**
   * Loads results from a set of instances contained in the supplied
   * file. This is started in the IO thread, and a dialog is popped up
   * if there's a problem.
   *
   * @param f a value of type 'File'
   */
  public void setBaseInstancesFromFile(final File f) {
      
    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	public void run() {
	  try {
	    m_Log.statusMessage("Reading from file...");
	    Reader r = new BufferedReader(new FileReader(f));
	    setBaseInstances(new Instances(r));
	    r.close();
	  } catch (Exception ex) {
	    m_Log.statusMessage("Problem reading " + f.getName() 
				+ " as an arff file.");
	    m_IOThread = null;
	    converterQuery(f);
	    /* m_Log.statusMessage("Problem reading " + f.getName());
	    JOptionPane.showMessageDialog(PreprocessPanel.this,
					  "Couldn't read from file:\n"
					  + f.getName() + "\n"
					  + ex.getMessage(),
					  "Load Instances",
					  JOptionPane.ERROR_MESSAGE); */
	  }
	  m_IOThread = null;
	}
      };
      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
  
  /**
   * Loads instances from a database
   *
   * @param iq the InstanceQuery object to load from (this is assumed
   * to have been already connected to a valid database).
   */
  public void setBaseInstancesFromDB(final InstanceQuery iq) {
    if (m_IOThread == null) {
      m_IOThread = new Thread() {
	public void run() {
	  
	  try {
	    m_Log.statusMessage("Reading from database...");
	    final Instances i = iq.retrieveInstances();
	    SwingUtilities.invokeAndWait(new Runnable() {
	      public void run() {
		setBaseInstances(new Instances(i));
	      }
	    });
	    iq.disconnectFromDatabase();
	  } catch (Exception ex) {
	    m_Log.statusMessage("Probelm executing DB query "+m_SQLQ);
	    JOptionPane.showMessageDialog(PreprocessPanel.this,
					  "Couldn't read from database:\n"
					  + ex.getMessage(),
					  "Load Instances",
					  JOptionPane.ERROR_MESSAGE);
	  }

	   m_IOThread = null;
	}
      };

      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    } else {
       JOptionPane.showMessageDialog(this,
				    "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }

  /**
   * Loads instances from a URL.
   *
   * @param u the URL to load from.
   */
  public void setBaseInstancesFromURL(final URL u) {

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

	  try {
	    m_Log.statusMessage("Reading from URL...");
	    Reader r = new BufferedReader(
		       new InputStreamReader(u.openStream()));
	    setBaseInstances(new Instances(r));
	    r.close();
	  } catch (Exception ex) {
	    m_Log.statusMessage("Problem reading " + u);
	    JOptionPane.showMessageDialog(PreprocessPanel.this,
					  "Couldn't read from URL:\n"
					  + u + "\n"
					  + ex.getMessage(),
					  "Load Instances",
					  JOptionPane.ERROR_MESSAGE);
	  }

	  m_IOThread = null;
	}
      };
      m_IOThread.setPriority(Thread.MIN_PRIORITY); // UI has most priority
      m_IOThread.start();
    } else {
      JOptionPane.showMessageDialog(this,
				    "Can't load at this time,\n"
				    + "currently busy with other IO",
				    "Load Instances",
				    JOptionPane.WARNING_MESSAGE);
    }
  }
  
  /**
   * Tests out the instance-preprocessing panel from the command line.
   *
   * @param args ignored
   */
  public static void main(String [] args) {

    try {
      final JFrame jf = new JFrame("Weka Knowledge Explorer: Preprocess");
      jf.getContentPane().setLayout(new BorderLayout());
      final PreprocessPanel sp = new PreprocessPanel();
      jf.getContentPane().add(sp, BorderLayout.CENTER);
      weka.gui.LogPanel lp = new weka.gui.LogPanel();
      sp.setLog(lp);
      jf.getContentPane().add(lp, BorderLayout.SOUTH);
      jf.addWindowListener(new WindowAdapter() {
	public void windowClosing(WindowEvent e) {
	  jf.dispose();
	  System.exit(0);
	}
      });
      jf.pack();
      jf.setSize(800, 600);
      jf.setVisible(true);
    } catch (Exception ex) {
      ex.printStackTrace();
      System.err.println(ex.getMessage());
    }
  }
}