Help running a piece of code (ConvChainJava)

daddydean

Hey guys , I am trying to run this: https://github.com/mxgmn/ConvChain

Which has a processing version over here: https://github.com/buckle2000/ConvChainJava

How am I supposed to set it up correctly? I cant quite get it to work, im getting some errors.

I am assuming I was to run main.java...

Thx! :)

jeremydouglass

Re:

How am I supposed to set it up correctly?
I cant quite get it to work, im getting some errors.

How have you set it up currently?

What specific error are you getting?

GoToLoop

• That ConvChain got some problems if you wish to compile it using the PDE. :-SS
• 1st, it doesn't have a ".pde" file as an entry point for setup() & draw(). @-)
• Then that "Main.java" needs to be converted to a proper "Main.pde".
• 2nd, both "Generator.java" & "Pattern.java" are using lambdas. ~O)
• However, the PDE doesn't accept lambda syntax yet, even for ".java" files! ~X(
• Therefore, those lambdas need to be converted to their corresponding old style anonymous instantiations.
• Which can be a challenging to find out the right way to do it for each lambda type. :-&
• I can give it a try later. But I think I'm gonna have to fork that project for it. #:-S

koogs

it's a java project that uses the processing library, you need to compile it using ant or mvn or a makefile or an ide or something, none of which is provided (pet hate of mine)

JAR=/path/to/processing-3.3/core/library/core.jar
javac -cp ${JAR} ./src/ConvChain/Generator.java ./src/ConvChain/Pattern.java ./src/Main.java

that's enough to get you the class files... i haven't got as far as running it...

koogs

(works in netbeans if you add the core.jar to the project)

(still can't get it to run from the command line)

koogs

ok, quick port to non-java8 here:

HOWEVER, the results look different from when i run it in netbeans so there might be a logical error somewhere. (original code left as comments)

(you'll need the images from the original copied into data folder)

//import ConvChain.Generator;
//import processing.core.PApplet;
//import processing.core.PImage;

//public class Main extends PApplet {
//    public static void main(String[] args) {
//        PApplet.main(Main.class.getCanonicalName());
//    }

public void settings() {
  size(256, 128);
  noSmooth();
}

boolean[][] toArray(PImage image) {
  boolean[][] result = new boolean[image.width][image.height];
  for (int x = 0; x < image.width; ++x) {
    for (int y = 0; y < image.height; ++y) {
      int colour = image.pixels[y * image.width + x];
      if (colour == -16777216) {
        result[x][y] = false; // black
      } else {
        result[x][y] = true; // white
      }
    }
  }
  return result;
}

PImage toPImage(boolean[][] array, PImage result) {
  if (result == null) {
    result = createImage(array.length, array[0].length, RGB);
  }

  for (int x = 0; x < array.length; ++x) {
    for (int y = 0; y < array[0].length; ++y) {
      if (array[x][y]) {
        result.pixels[y * result.width + x] = color(1);
      } else {
        result.pixels[y * result.width + x] = color(0);
      }
    }
  }
  result.updatePixels();
  return result;
}


PImage image;
Generator generator;
PImage output;

void updateOutput() {
  output = toPImage(generator.field, output);
}

public void setup() {
  colorMode(RGB, 1, 1, 1);
  image = loadImage("Less Rooms.bmp");
  generator = new Generator(toArray(image), 3, 1.2, 32, 32);
  updateOutput();
}

public void draw() {
  image(image, 0, 0, 128, 128);
  image(output, 128, 0, 128, 128);
}

@ Override
  public void keyPressed() {
  generator.iterate();
  updateOutput();
}
//}

import java.util.HashMap;
import java.util.Random;

public class Generator {
  public double temperature;
  public Random random;
  public boolean[][] field;
  public final double DEFAULT_WEIGHT = 0.01;

  private int N;
  private int width;
  private int height;

  public int getWidth() {
    return width;
  }

  public int getHeight() {
    return height;
  }

  private Pattern[] patterns;
  private HashMap<Integer, Double> weights;

  public Generator(boolean[][] sample, int N, double temperature, int width, int height) {
    this.N = N;
    this.temperature = temperature;
    this.width = width;
    this.height = height;
    this.random = new Random();
    this.field = new boolean[width][height];
    this.weights = new HashMap<>();

    for (int y = 0; y < sample[0].length; y++) {
      for (int x = 0; x < sample.length; x++) {
        Pattern[] p = new Pattern[8];
        p[0] = new Pattern(sample, x, y, N);
        p[1] = p[0].getRotated();
        p[2] = p[1].getRotated();
        p[3] = p[2].getRotated();
        p[4] = p[0].getReflected();
        p[5] = p[1].getReflected();
        p[6] = p[2].getReflected();
        p[7] = p[3].getReflected();
        for (int k = 0; k < 8; k++) {
          weights.put(p[k].getIndex(), weights.getOrDefault(p[k].getIndex(), 0.) + 1.);
        }
      }
    }

    //weights.replaceAll((pattern, weight) -> {
    //    if (weight <= 0.)
    //        return DEFAULT_WEIGHT;
    //    else
    //        return weight;
    //});
    // replaced with:
    for (Integer i : weights.keySet()) {
      if (weights.get(i) <= 0) {
        weights.put(i, DEFAULT_WEIGHT);
      }
    }

    for (int y = 0; y < height; y++) {
      for (int x = 0; x < width; x++) {
        field[x][y] = random.nextBoolean();
      }
    }
  }

  double energyExp(int i, int j) {
    double value = 1.0;
    for (int y = j - N + 1; y <= j + N - 1; y++) {
      for (int x = i - N + 1; x <= i + N - 1; x++) {
        value *= weights.getOrDefault(new Pattern(field, x, y, N).getIndex(), DEFAULT_WEIGHT);
      }
    }
    return value;
  }

  void metropolis(int i, int j) {
    double p = energyExp(i, j);
    field[i][j] = !field[i][j];
    double q = energyExp(i, j);

    if (Math.pow(q / p, 1.0 / temperature) < random.nextDouble()) {
      field[i][j] = !field[i][j];
    }
  }

  public void iterate() {
    for (int k = 0; k < width * height; k++) {
      metropolis(random.nextInt(width), random.nextInt(height));
    }
  }
}


class Pattern {
  public boolean[][] data;

  private int getSize() {
    return data.length;
  }

  // new
  private void setValue(boolean value) {
    for (int j = 0; j < getSize(); j++) {
      for (int i = 0; i < getSize(); i++) {
        data[i][j] = value;
      }
    }
  }

  //public Pattern(int size, BiPredicate<Integer, Integer> f) {
  //  data = new boolean[size][size];
  //  for (int j = 0; j < getSize(); j++) {
  //    for (int i = 0; i < getSize(); i++) {
  //      data[i][j] = f.test(i, j);
  //    }
  //  }
  //}

  // new
  public Pattern(int size, boolean value) {
    data = new boolean[size][size];
    for (int j = 0; j < getSize(); j++) {
      for (int i = 0; i < getSize(); i++) {
        data[i][j] = value;
      }
    }
  }

  public Pattern(boolean[][] field, int x, int y, int size) {
    //this(size, (i, j) -> false);
    // replaced with
    this(size, false);
    //Set((i, j) ->
    //  field[(x + i + field.length) % field.length]
    //  [(y + j + field[0].length) % field[0].length]);
    // replaced with:
    for (int i = 0; i < size; i++) {
      for (int j = 0; j < size; j++) {
        int i1 = (x + i + field.length) % field.length;
        int i2 = (y + j + field[0].length) % field[0].length;
        data[i][j] = (field[i1][i2]); // NOT setValue(field[i1][i2]);
      }
    }
  }

  public Pattern getRotated() {
    //return new Pattern(getSize(), (x, y) -> data[getSize() - 1 - y][x]);
    // replaced with
    Pattern rotated = new Pattern(getSize(), false);
    for (int x = 0; x < getSize(); x++) {
      for (int y = 0; y < getSize(); y++) {
        rotated.data[x][y] = data[getSize() - 1 - y][x];
      }
    }
    return rotated;
  }

  public Pattern getReflected() {
    //return new Pattern(getSize(), (x, y) -> data[getSize() - 1 - x][y]);
    // replaced with
    Pattern reflected = new Pattern(getSize(), false);
    for (int x = 0; x < getSize(); x++) {
      for (int y = 0; y < getSize(); y++) {
        reflected.data[x][y] = data[getSize() - 1 - x][y];
      }
    }
    return reflected;
  }

  public int getIndex() {
    int result = 0;
    for (boolean[] row : data) {
      for (boolean datum : row) {
        result <<= 1;
        result += datum ? 1 : 0;
      }
    }
    return result;
  }
}

koogs

ok, fixed. error was line 221

daddydean

@jeremydouglass
Nevermind, obviously I was trying to run the main java file like a dummy.
@GoToLoop
Oh dear I hadnt anticipated it was going to be this complicated...im not familiar at all with this lamda stuff.

@koogs
You can run it in netbeans? (never heard of this) I'd like to run it in processing if possible...such a cool code.
Might be a bit beyond my skills....let me know if you figure it out

daddydean

There is also a javascript port of a later version: https://github.com/kchapelier/wavefunctioncollapse (which includes colour) and a python version: https://github.com/ikarth/wfc_python (colour)

koogs

let me know if you figure it out

you mean like in the post directly above that one? posted 90 minutes earlier?

that huge slab of code is the processing version.

daddydean

@koogs
Oh, haha! I thought you were just reposting the original main.java...so you actually have it working in processing on your end? Im not understanding what I would need to do to make it run....you have the pattern.java and generator.java together in that slab of code?
I have the original images in a folder called data, like you say.
I get a: unexpected token: href

koogs

tabs in processing are a falsehood, just there for organisational purposes. that code has the Pattern and Generator classes in it, which is enough.

cut and paste the entire thing into the editor, add the image using the menu or drag and drop and hit ctrl-r to run.

koogs

it's not very spectacular.

daddydean

@koogs I see. Its not working for me though, im getting the error
"unexpected token: href"..?
I have the same images in the data folder.

koogs

Line 69 should just be @Override. The forum software breaks things.

daddydean

Ah, think you can post it to pastebin? :3 Im getting a different error now: unexpected token > on line 107

koogs

Ah, ok, I think the pattern and generator classes have to be in files with .java extensions. I'll fork it on github.

koogs

here: https://github.com/acdean/ConvChainProcessing

download as zip file, unzip into your sketchbook and it should run.

daddydean

@koogs
Wow! You rock man! Just one question, I want to remove the sample image from the processing window (which I have done) but how do I centre the output image to fill up the entire window?
https://imgur.com/a/znK2J
Clicking the up and down arrows seems to make it iterate, not sure if there are other keys im missing. Great work.

koogs

line 66 is the drawing. image(output, 128, 0, 128, 128); top left corner is at 128, 0 and is 128x128 in size. see the reference.

line 70 - any key makes it iterate.

i'm still not sure what it does, what it's for...

GoToLoop

Oh, I was also attempting to convert it. But for now my custom version is buggy! 3:-O
I've tried to make the boolean[][] array to represent row × col instead, but I've failed! X_X

Nonetheless, gonna leave its download link here for now: 8-X
https://GitHub.com/GoToLoop/ConvChainP5/releases

daddydean

@koogs Got it now, thanks :) Looks great!
@GoToLoop I dont get whats wrong with it? Looks good to me, is it because you can't scale it bigger? How would I do that? ...ohh, changing the grid size is nice!!

daddydean

Also if anyone feels up to the task...the colour version is probably a lot cooler: https://github.com/kchapelier/wavefunctioncollapse
...this is the javascript version. Has other options like N amount and symmetry.