oogifu

CLASSEMENT DE LA LIBERTÉ DE LA PRESSE 2011/2012

2012-01-30T21:40:00.003Z

Hmm... CLASSEMENT DE LA LIBERTÉ DE LA PRESSE 2011/2012

Scala, JavaFX, TableColumns

2012-01-19T16:07:00.000Z

In my previous post we created a bunch of columns using the following code:

val maxMemColumn = new TableColumn[MyTableRow, Long]("Max")
maxMemColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, Long]("maxMem"))

It would be nice to abstract this call, create a template for it. Looking at the TableColumn API, another way to bind a cell to an object value is to use setCellValueFactory on the TableColumn:

firstNameCol.setCellValueFactory(new Callback, ObservableValue>() {
  public ObservableValue call(CellDataFeatures p) {
    // p.getValue() returns the Person instance for a particular TableView row
	return p.getValue().firstNameProperty();
  }
});

Note that firstNameProperty is of type ObservableValue. We can therefore remove all the other properties from MyTableRow and just leave the ones which are of type ObservableValue... We are then left with:

sealed class MyTableRow2 {
  val totMem = new SimpleLongProperty(0L)
  val maxMem = new SimpleLongProperty(0L)
  val freeMem = new SimpleLongProperty(0L)
  val usedMem = new SimpleLongProperty(0L)
}

Now, let's try to write a generic create column Scala function (newCol_2 below):

private def newCol_2[R, V](displayName: String, get: (R) => ObservableValue[V]): TableColumn[R, V] = {
  val column = new TableColumn[R, V](displayName)
  column.setCellValueFactory(new Callback[CellDataFeatures[R, V], ObservableValue[V]] {
    override def call(p: CellDataFeatures[R, V]) : ObservableValue[V] = get(p.getValue())
  })
  column
}

the columns are now initialised like this:

val maxMemColumn = newCol_2("Max", {r:MyTableRow2 => r.maxMem})
val totalMemColumn = newCol_2("Tot", {r:MyTableRow2 => r.totMem})
val freeMemColumn = newCol_2("Free", {r:MyTableRow2 => r.freeMem})
val usedMemColumn = newCol_2("Used", {r:MyTableRow2 => r.usedMem})

Memory Pie Chart in TableView with Scala and JavaFX

2012-01-14T20:08:00.000Z

In my series (Scala, JavaFX 2) - another code snippet to display a memory pie chart in a tableview with a custom renderer. Below is the code to achieve this:

The imports in used:

import javafx.application.Application
import javafx.beans.property.SimpleLongProperty
import javafx.beans.property.SimpleObjectProperty
import javafx.scene.control.cell.PropertyValueFactory
import javafx.scene.control.TableColumn
import javafx.scene.control.TableView
import javafx.scene.layout.BorderPane
import javafx.scene.paint.Color
import javafx.scene.Scene
import javafx.stage.Stage
import javafx.util.Callback
import javafx.scene.control.TableCell

Then the class representing one line in the table (Note the use of a tuple2 what will later be used for the chart - not really needed as the row can be accessed in the custom renderer - but other posts will follow...):

sealed class MyTableRow {
  val totMem = new SimpleLongProperty(0L)
  def totMemProperty = totMem
  def getTotMem : Long = totMem.get
  def setTotMem(value: Long) = totMem.set(value)
  
  val maxMem = new SimpleLongProperty(0L)
  def maxMemProperty = maxMem
  def getMaxMem : Long = maxMem.get
  def setMaxMem(value: Long) = maxMem.set(value)
  
  val freeMem = new SimpleLongProperty(0L)
  def freeMemProperty = freeMem
  def getFreeMem : Long = freeMem.get
  def setFreeMem(value: Long) = freeMem.set(value)
  
  val usedMem = new SimpleLongProperty(0L)
  def usedMemProperty = usedMem
  def getUsedMem : Long = usedMem.get
  def setUsedMem(value: Long) = usedMem.set(value)
  
  val chartData = new SimpleObjectProperty((0L, 0L))
  def chartDataProperty = chartData
  def getChartData : Tuple2[Long, Long] = chartData.get
  def setChartData(value: Tuple2[Long, Long]) = chartData.set(value)
}

Then the main application:

sealed class MemTable extends Application {
  def runInJFXThread(f: () => Unit) = javafx.application.Platform.runLater(new Runnable() {def run() {f()}})
  
  val timer = new java.util.Timer
  val table = new TableView[MyTableRow]
  val tableData = javafx.collections.FXCollections.observableArrayList[MyTableRow]()
  table.setItems(tableData)
  
  val singleRow = new MyTableRow
  tableData.addAll(singleRow)
  
  override def start(stage: Stage) = {
    val borderPane = new BorderPane
    val scene = new Scene(borderPane, 400, 300, Color.WHITE)
    borderPane.setCenter(table)
    setupUI
    stage.setTitle("Memory Table")
    stage.setScene(scene)
    stage.show
    timer.schedule(new HeapSnapshotTask, 0L, 1000L)
  }
  
  sealed class HeapSnapshotTask extends java.util.TimerTask {
    val inKb = 1024L
    override def run = {
      val r = Runtime.getRuntime
      val totalMemory = r.totalMemory / inKb
      val freeMemory = r.freeMemory / inKb
      val maxMemory = r.maxMemory / inKb
      val usedMemory = totalMemory - freeMemory
      
      runInJFXThread{() => {
        singleRow.totMem.set(totalMemory)
        singleRow.maxMem.set(maxMemory)
        singleRow.freeMem.set(freeMemory)
        singleRow.usedMem.set(usedMemory)
        singleRow.chartData.set((freeMemory, usedMemory))
      }}
      
      println("T = " + totalMemory + ", F = " + freeMemory + ", M = " + maxMemory + ", U = " + usedMemory)
    }
  }
  
  private def setupUI = {
    val memColumns = new TableColumn[MyTableRow, AnyRef]("Memory")
    
    val maxMemColumn = new TableColumn[MyTableRow, Long]("Max")
    maxMemColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, Long]("maxMem"))
    val totalMemColumn = new TableColumn[MyTableRow, Long]("Total")
    totalMemColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, Long]("totMem"))
    val freeMemColumn = new TableColumn[MyTableRow, Long]("Free")
    freeMemColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, Long]("freeMem"))
    val usedMemColumn = new TableColumn[MyTableRow, Long]("Used")
    usedMemColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, Long]("usedMem"))
    val chartColumn = new TableColumn[MyTableRow, AnyRef]("Chart")
    chartColumn.setCellValueFactory(new PropertyValueFactory[MyTableRow, AnyRef]("chartData"))

    val chartRenderer = new Callback[TableColumn[MyTableRow, AnyRef], TableCell[MyTableRow, AnyRef]] {
      override def call(col : TableColumn[MyTableRow, AnyRef]) : TableCell[MyTableRow, AnyRef] = {
        new TableCell[MyTableRow, AnyRef] {
          import javafx.scene.chart.PieChart
          import javafx.collections.FXCollections
          import javafx.scene.layout.Pane
          import javafx.scene.control.Label
   
          val usedMemData = new PieChart.Data("Used", 0.0)
          val freeMemData = new PieChart.Data("Free", 0.0)
          val pieChartData = FXCollections.observableArrayList(usedMemData, freeMemData)
          val pieChart = new PieChart(pieChartData)
          pieChart.setPrefSize(150, 150)
          pieChart.setLabelsVisible(true)
          pieChart.setLegendVisible(false)
          val borderPane = new BorderPane
          val southPane = new Pane
          setGraphic(borderPane)
          borderPane.setCenter(pieChart)
          borderPane.setBottom(southPane)
          val southLabel = new Label
          southPane.getChildren().addAll(southLabel)
          override def updateItem(value : AnyRef, empty : Boolean) = {
            val row = getTableRow
            if (row != null && value != null) {
              val datum = value.asInstanceOf[Tuple2[Long, Long]]
              val free = datum._1
              val used = datum._2
              usedMemData.setPieValue(used)
              freeMemData.setPieValue(free)
              val desc = "F="+free+" U="+used
              southLabel.setText(desc)
            } // if
          } // updateItem
        }
      }
    } // chartRenderer

    chartColumn.setCellFactory(chartRenderer)
    
    memColumns.getColumns().addAll(maxMemColumn, totalMemColumn, freeMemColumn, usedMemColumn)
    
    table.getColumns().addAll(memColumns, chartColumn)
  }
}

object MemTable {
  def main(args: Array[String]): Unit = {
    javafx.application.Application.launch(classOf[MemTable])
  }
}

That's it for now

Callback a la scala

2012-01-13T17:59:00.001Z

The first (and not only) un-natural element when moving from Java to Scala is that callback via interfaces and anonymous classes are not needed anymore. Let's walk through a simple example, using the Spring-like SQL template First in Java:

import java.sql.Connection;
import java.sql.ResultSet;
import java.sql.Statement;

import javax.sql.DataSource;

interface ICallbackHandler {
  void process(ResultSet rs);
}

final class JSqlTemplate {
  DataSource ds = null;
 
  public void query(final String sql, final ICallbackHandler cb) throws Exception {
    final Connection conn = ds.getConnection();
    final Statement stmt = conn.createStatement();
    final ResultSet rs = stmt.executeQuery(sql);
    while(rs.next()) cb.process(rs);
      rs.close();
      stmt.close();
      conn.close();
    }
}

public class DBCallbackTest {
  public static void main(String[] args) throws Exception {
    new JSqlTemplate().query("select * from whatever", new ICallbackHandler() {
      public void process(ResultSet rs) {
        // get stuff here
      }
    });
  }
}

In Scala, there is no need to define a ICallbackHandler .. you can just go for:

import javax.sql.DataSource
import java.sql.ResultSet

sealed class SSqlTemplate(dataSource: DataSource)  {
  def query(sql: String, rowHandler: (ResultSet) => Unit) = {
    val conn = dataSource.getConnection
    val stmt = conn.createStatement
    val rs = stmt.executeQuery(sql)
    while (rs.next) rowHandler(rs)
    rs.close
    stmt.close
    conn.close
  }
}

object SDbCallback {
  def main(args: Array[String]): Unit = {
    val ds:DataSource = null
    new SSqlTemplate(ds).query("select * from whatever",
        {(rs: ResultSet) =>
          // get stuff here
       })
  }
}

In response to 'Generating every combination without duplicates'

2012-01-13T15:41:00.000Z

Peter, a friend of mine, posted this : Java riddle... Here is my scala solution using Scalaz:

import scalaz._
import Scalaz._

object Combi1 {
  def combine[A](xs: List[A]): List[List[A]] = xs.replicate[List](xs.size).sequence
  
  def main(args: Array[String]): Unit = {
    val list = List(1, 1, 2, 3, 3, 3)
    def isCorrect(list: List[Int]) : Boolean = {
      list.count(i => i == 1) == 2 &&
      list.count(i => i == 2) == 1 &&
      list.count(i => i == 3) == 3
    }
    val combi = combine(list).filter(l => isCorrect(l)).distinct
    println(combi.size + " elements")
    combi.foreach(l => println(l))
  } 
}

Memory Pie Chart (Scala, JavaFX2)

2012-01-10T19:59:00.000Z

Another simple code snippet using Scala and JavaFX2: A memory visualiser..

package org.mychart

import javafx.application.Application

import javafx.scene.control.Label
import javafx.scene.Scene
import javafx.scene.paint.Color
import javafx.scene.layout.BorderPane
import javafx.stage.Stage
import javafx.scene.chart.PieChart
import javafx.collections.FXCollections
import javafx.scene.layout.Pane

sealed class RealtimeMemory extends Application {
  val timer = new java.util.Timer
  val usedMemData = new PieChart.Data("Used", 0.0)
  val freeMemData = new PieChart.Data("Free", 0.0)
//  val totalMemData = new PieChart.Data("Total", 0.0)
  val pieChartData = FXCollections.observableArrayList(usedMemData, freeMemData)
  val chart = new PieChart(pieChartData)
  val southPane = new Pane
  val southLabel = new Label
  
  override def start(stage: Stage) = {
    val borderPane = new BorderPane
    val scene = new Scene(borderPane, 800, 600, Color.BEIGE)
    stage.setTitle("JFX Xiphias Pricing Client")
    stage.setScene(scene)
    chart.setLabelsVisible(true)
    chart.setLegendVisible(false)
    borderPane.setCenter(chart)
    borderPane.setBottom(southPane)
    
    southPane.getChildren.addAll(southLabel)
    
    stage.show
    
    timer.schedule(new HeapSnapshotTask, 0L, 1000L)
  }
  
  def runInJFXThread(f: () => Unit) = 
    javafx.application.Platform.runLater( new Runnable() { def run() {f()} } )
  
  sealed class HeapSnapshotTask extends java.util.TimerTask {
    override def run = {
      val r = Runtime.getRuntime
      val totalMemory = r.totalMemory / 1024 / 1024 
      val freeMemory = r.freeMemory / 1024 / 1024
      val maxMemory = r.maxMemory / 1024 / 1024
      val usedMemory = totalMemory - freeMemory
      
      usedMemData.setPieValue(usedMemory)
      freeMemData.setPieValue(freeMemory)
//      totalMemData.setPieValue(totalMemory)
      
      runInJFXThread(()=>southLabel.setText(
          "Max: " +maxMemory+", Used:"+usedMemory+
          ", Free:"+freeMemory+", Total:"+totalMemory))
    }
  }
}

object RealtimeMemory {  
  def main(args: Array[String]): Unit = {
    javafx.application.Application.launch(classOf[RealtimeMemory])
  }
}

Small crypto code for Android

2012-01-07T15:27:00.000Z

I'd like to share a simple class that I wrote a while ago as a cryptographic utility for Android. It uses the Bouncy Castle crypto lib for JDK 6 and the Apache Base 64 encoder.


package org.oogifu.crypto;

import org.apache.base64.Base64;
import org.bouncycastle.crypto.BlockCipher;
import org.bouncycastle.crypto.BufferedBlockCipher;
import org.bouncycastle.crypto.engines.AESFastEngine;
import org.bouncycastle.crypto.paddings.PKCS7Padding;
import org.bouncycastle.crypto.paddings.PaddedBufferedBlockCipher;
import org.bouncycastle.crypto.params.KeyParameter;
import org.bouncycastle.util.Arrays;

public final class BCEncryptionUtils {
  private static final BCEncryptionUtils instance = new BCEncryptionUtils();
 
  private transient byte rawKey[] = null;
 
  private BCEncryptionUtils() {
  }
 
  public static BCEncryptionUtils getInstance() {
    return instance;
  }
 
  public String getCipherInfo() {
    return "AES-256 provided By Bouncy Castle";
  }

  public String getFormat() {
    return "256";
  }
 
  public void setMasterPassword(String password) throws Exception {
    final byte[] masterPassword = new byte[32];
    Arrays.fill(masterPassword, (byte)0);
     
    if (password.length() > 32) password = password.substring(0, 32);
       
    final byte[] source = password.getBytes();
       
    System.arraycopy(source, 0, masterPassword, 0, source.length);
       
    rawKey = masterPassword;
  }

  public String decryptFromBase64(final String edata64) throws Exception {
    final byte[] edb = Base64.decodeBase64(edata64);
    final byte[] data = decode(edb);
    return new String(data).trim();
  }

  public String encryptToBase64(final String data) throws Exception {
    final byte[] newData = data.getBytes("UTF-8");
    return Base64.encodeBase64String(encode(newData));
  }
 
  private byte[] encode(final byte[] inputBytes) throws Exception {
    final BufferedBlockCipher cipher = getCipher(true);
    final byte[] outputBytes = new byte[cipher.getOutputSize(inputBytes.length)];
  
    int outputLen = cipher.processBytes(inputBytes, 0, inputBytes.length, outputBytes, 0);
    outputLen += cipher.doFinal(outputBytes, outputLen);
     
    return outputBytes;
  }
 
  private byte[] decode(final byte[] inputBytes) throws Exception {
    final BufferedBlockCipher cipher = getCipher(false);
  
    final byte[] outputBytes = new byte[cipher.getOutputSize(inputBytes.length)];
   
    int outputLen = cipher.processBytes(inputBytes, 0, inputBytes.length, outputBytes, 0);
    outputLen += cipher.doFinal(outputBytes, outputLen);
     
    return outputBytes;
  }
 
  private BufferedBlockCipher getCipher(final boolean forEncryption) {
    final BlockCipher aesEngine = new AESFastEngine();
    final BufferedBlockCipher cipher = new PaddedBufferedBlockCipher(aesEngine, new PKCS7Padding());
    cipher.init(forEncryption, new KeyParameter(rawKey));
    return cipher;
  }
}

Scala Lazy Types

2011-11-27T00:24:00.001Z

I have been working on a pricing engine [PE] (a piece of software that calculates bond prices using different models) for the last year, written in Java 6. A PE is essentially a DAG (Directed Acyclic Graph), and there are many ways to implement the relationships between the pricing nodes: remote nodes if you fancy distributed processes, matrix for the parent-child edges, or a pure OO approach where a pricing node has a direct reference to its parents. An example of such a naive implementation would look like this:

public abstract class AbsNode {
  protected List parents = new LinkedList<>();
  protected double price, yield;
  public abstract void calculate();
}

class FutureNode extends AbsNode {
  @Override public void calculate() {
  }
}

class GrossBasisNode extends AbsNode {
  @Override public void calculate() {
    price = parents.get(0).price * 0.98 + 1.0;
  }
}

class BenNode extends AbsNode {
  @Override public void calculate() {
    yield = parents.get(0).yield * 3.14;
  }
}

This obviously is only there to demonstrate the following issues: (1) The code cannot guaranty that the child of a future node must only be a ben node, or that the child of a ben node can only be a ben node, (2) that price and yield should only exist for bonds, and price only for futures (3) you can do pre-check at runtime using instanceof but it is not very nice, (4) there are potential class cast exceptions, (5) there is no compile-time check. Learning Scala, I was wondering if the very much debated type system could come to the rescue. I came up with:

trait MyNode {
  type parentType
  var parents:List[parentType] = List()
  var calcPrice = Double.NaN
  def addParent(parent: parentType) = parents = parent :: parents
}
trait MyBond extends MyNode {
  var calcYield = Double.NaN
}
trait MyFuture extends MyNode
sealed class GenFuture extends MyFuture
sealed class GenBenBond extends MyBond {
  type parentType = MyBond
}
sealed class GenGbnBond extends MyBond {
  type parentType = MyFuture
}

Notice the override of the type parentType: in GenBenBond I enforce the type to be a bond, whereas for the GenGbnBond bond I enforce the type to be MyFuture. This s quite neat. It allows me to write this:

val fut = new GenFuture
val gbn = new GenGbnBond
val ben = new GenBenBond
gbn.addParent(fut)
ben.addParent(gbn)

but this

ben.addParent(fut)

does not compile... exactly what I want. The following amended code shows another good side-effect, the parent in calculate are of the right type:

trait MyNode {
  type parentType
  var parents:List[parentType] = List()
  var calcPrice = Double.NaN
  def addParent(parent: parentType) = parents = parent :: parents
  def calculate = {}
}
trait MyBond extends MyNode {
  var calcYield = Double.NaN
}
trait MyFuture extends MyNode
sealed class GenFuture extends MyFuture
sealed class GenBenBond extends MyBond {
  type parentType = MyBond
  override def calculate = {
    val bond = parents(0)
    bond.calcYield
    bond.calcPrice
  }
}
sealed class GenGbnBond extends MyBond {
  type parentType = MyFuture
  override def calculate = {
    val fut = parents(0)
    // only calcPrice available
    fut.calcPrice 
  }
}

Plus

2011-08-28T08:07:00.001Z

Google plus! is very good.
Give it a try.
I am not a Facebook user, so I cannot compare, but Plus' simple interface suits me.

01190 & Patates

2011-08-21T17:23:00.001Z

Me voila sous 36 degres a Arbigny.. petit village de l'Ain ou il fait bon vivre... a l'aube d'une crise financiere mondiale, je reprends des forces et je fais pousser des patates...

A380 - Air France

2011-05-30T13:09:00.002Z

Intelligence and Rational Thinking

2011-03-19T13:14:00.000Z

As Daniel Kahneman at Princeton University puts it, intelligence is about brain power whereas rational thinking is about control.

French Quote Of The Day

2011-01-28T19:51:00.001Z

C'est parce que la vitesse de la lumiere est superieure a celle du son, que certains ont l'air brillant avant d'avoir l'air con!

Scala wins EU grant!

2011-01-17T22:37:00.000Z

Read this

Micro Credit

2010-12-27T11:30:00.000Z

Avec Babyloan

Chrome Web Store

2010-12-25T17:43:00.002Z

Very very impressive.. You can feel how irrelevant the desktop will be in a few years....

Gateau Polonais aux pommes

2010-12-19T14:21:00.003Z

Nothing to do today: snow all over the place, no golf, cant drive the car...So a recipe for a famous cake.

* 150g de farine
* 120g de sucre
* 4 pommes
* 2 oeufs
* 1/2 verre de lait
* 80g de matiere grasse
* levure

- Melanger sucre et matiere grasse
- Ajouter les oeufs, puis la farine et le lait
- Incorporer la levure
- Ajouter les morceaux de pommes (decoupe fine)

Faire cuire au four pendant 30/35 minutes a 180 degres.

Voila!

One chinese character per day

2010-11-08T20:19:00.000Z

1 character per day

Railway Children

2010-11-02T21:07:00.000Z

Every year, in the UK 100,000 under 16 run away from home

Brent

2010-10-04T19:32:00.002Z

Whoever is interested in p2y/y2p interpolations, or others root-finding techniques, should check-out the Brent's root-finding algorithm implemented as part of Apache Commons Math.

Gosling's interview

2010-09-29T19:34:00.001Z

My First Football Game

2010-09-29T19:19:00.002Z

I attended my first ever football game last night. After 10 years+ in London, I had the opportunity to go and watch Chelsea-Marseille at the Stamford Bridge stadium.
Chelsea won 2-0, but that is not the interesting part. The interesting part is that if you are a foreigner attending a game with the local supporters, be sure to wear the local shirt...in this case the Chelsea shirt...Avoid speaking any language that is not English.. actually, do not speak at all :-), and do not *ever* applaud a good action from the opposing team; otherwise, you'll find yourself in all kinds of trouble...
Still an enjoyable evening...
Supporters-wise, Marseille 10 - Chelsea 0

Ohhhh-Hisse

Android Password Manager File Transfer

2010-08-01T11:19:00.003Z

For the few people using my passwords manager for Android (http://janaudy.com/android/) - well, just me may be, here is an easy way to transfer the files from your desktop to your phone:

- Install https://www.dropbox.com for your desktop and on your phone (via Android Market)
- Install ASTRO file manager (http://www.androidtapp.com/astro-file-manager/) via Android Market
- Export the encrypted file per usual, drop it in the drop box
- Run ASTRO on the Android phone and copy the file from your local drop box folder to the root of the /sd card

Painless. No more cables.

I have also tried a free Bluetooth file transfer that also just works fine!

Jobs in San Francisco

2010-07-31T16:41:00.001Z

Just returned from a business trip in San Francisco.
A link to share for job openings: Craig's List

Interested in TSP?

2010-07-24T11:35:00.001Z

This is *the* place to visit: http://www.tsp.gatech.edu/