WordCram using the names of Malware

[jag...@gmail.com]

I'd like to make an SVG showing growing malware threat names. A quick way to visually track what's fresh and emerging and what's not.

Problem is, when I load the malware names into a "word", anything with numerics, punctuation or spaces, gets tossed out of the resulting image. Is there any ways to change the behaviour of "Word word = new Word(name, frequency);" to allow the name to have weird-ish characters? I'm looking to use something like:

F-UPA-AAH_(ED)    0.81    f
BrowseFox.e    0.54    m
Adware-BrowseFox    0.49    m
PUP-FNK    0.48    f
PUP-FPS    0.47    m
Downloader-FSH!432718827A11    0.03    m
RDN/Generic_PWS.y!bbt    0.03    m
RDN/Generic.dx!dh3    0.02    m
Generic.Tra!7b7c92f2255a_(ED)    0.01    m
VBS/Autorun.worm.aadd!lnk    0.01    f

Needless to say, I hacked the firstnamesUsingWordPresets sample. The Solution is probably newbie error and I probably need  to switch to UTF8 or the like but I'm clueless on how to do that. :-)

Thanks

Jacques

[Dan Bernier]

It's hard to say what's going wrong - this should work.

I know that the firstnamesUsingWordPresets example strips off the numbers and the "m" or "f" (denoting gender), so it might be that that's your problem.

Could you paste in your code?

[jag...@gmail.com]

It is just a clone of the sample code. The ParseNames routines split the input record on TAB characters. It does look however, like the internal code does a regex on the NAME passed to it and removes all non-word characters.

Thanks for any hints you can throw my way.

Jacques

names.txt is just:
BrowseFox    1.00    m
T-UPA-AAH_(ED)    0.81    f

BrowseFox.e    0.54    m
Adware-BrowseFox    0.49    m
PUP-FNK    0.48    f
PUP-FPS    0.47    m
Downloader-FSH!432718827A11    0.03    m
RDN/Generic_PWS.y!bbt    0.03    m
RDN/Generic.dx!dh3    0.02    m
Generic.Tra!7b7c92f2255a_(ED)    0.01    m
VBS/Autorun.worm.aadd!lnk    0.01    f

Other    0.17    f
Artemis_Contribution    0.05    f


here's the code (nothing new or special here):

import wordcram.*;

WordCram wc;
Word[] names;

void setup() {
  colorMode(HSB);
  size(800, 600);
  background(255);
 
  loadNames();
  makeWordCram();
}

void loadNames() {
  String[] nameData = loadStrings("names.txt");
  names = new Word[nameData.length];
  for (int i = 0; i < names.length; i++) {
    names[i] = parseName(nameData[i]);
  }
}

void makeWordCram() {
   try {
      new WordCram(this).
      fromWords(names).
      toSvg("test.svg", width, height).
      drawAll();
  }
  catch (java.io.FileNotFoundException x) {
    println(x.getMessage());
  }

}

void draw() {
  while(wc.hasMore()) {
    wc.drawNext();
  }
    println("done");
    /* noLoop();*/
  
}

void mouseClicked() {
  background(255);
  makeWordCram();
  loop();
}


// Each row looks like:
// Mary\t1.0\tf
// ...or:
// {name} {tab} {frequency} {tab} {'f' for females, 'm' for males}
Word parseName(String data) {
  String[] parts = split(data, '\t');
 
  String name = parts[0];
  float frequency = float(parts[1]);
  boolean isFemale = "f".equals(parts[2]);
  println(name,frequency,isFemale);

 
  Word word = new Word(name, frequency);
 

  if (isFemale) {
    word.setColor(#f36d91); // pink
  }
  else {
    word.setColor(#476dd5); // blue
  }
 
  return word;
}

[Dan Bernier]

I don't think the problem is the regex in the internal code - that's only for when WordCram parses your text for you; here, you're creating your words, so it shouldn't come into play.

You have a lot of parts here that aren't all coming together correctly, and I think that's probably where your trouble lies. I was able to fix a few things, and got it to run, with the numbers included.

The things I changed are:

* removed the toSvg("test.svg") stuff

* in makeWordCram, it's instantiating a WordCram, calling drawAll on it, and then throwing it away; then, draw() tries to use the global wc instance. Instead, I took the drawAll() out of makeWordCram(), and had it set wc, so the draw() method would work.

* Many of your words were weighted very low, so they wouldn't come out - I weighted them all at 1 (just to test), and used sizedByWeight(15, 15) to draw them all at 15pt.

Here's the code I used:

import wordcram.*;

WordCram wc;

Word[] names;

void setup() {

  colorMode(HSB);

  size(800, 600);

  background(255);

  

  loadNames();

  makeWordCram();

}

void loadNames() { 

  String[] nameData = loadStrings("../../names.txt");

  names = new Word[nameData.length];

  for (int i = 0; i < names.length; i++) {

    names[i] = parseName(nameData[i]);

  }

}

void makeWordCram() {

      wc = new WordCram(this).

      sizedByWeight(15, 15).

      fromWords(names);

      //toSvg("test.svg", width, height); //.

      //drawAll();

}

void draw() {

  while(wc.hasMore()) {

    wc.drawNext();

  }

    /* noLoop();*/

   

}

void mouseClicked() {

  background(255);

  makeWordCram();

  loop();

}

// Each row looks like:

// Mary\t1.0\tf

// ...or:

// {name} {tab} {frequency} {tab} {'f' for females, 'm' for males}

Word parseName(String data) {

  String[] parts = split(data, '\t');

  

  String name = parts[0];

  float frequency = float(parts[1]); frequency = 1.0;

  boolean isFemale = "f".equals(parts[2]);

  println(name,frequency,isFemale);

  

  Word word = new Word(name, frequency);

  

  if (isFemale) {

    word.setColor(#f36d91); // pink

  }

  else {

    word.setColor(#476dd5); // blue

  }

  

  return word;

}