// Cloned by test2 on 10 Nov 2022 from World "XOR multi-layer network" by "Coding Train" project 
// Please leave this clone trail here.
 

// XOR multi-layer network

// Port from:
// https://github.com/CodingTrain/Toy-Neural-Network-JS/tree/master/examples/xor
// with modifications 

// libraries from:
// https://github.com/CodingTrain/Toy-Neural-Network-JS/tree/master/lib
// ported to here:
// https://ancientbrain.com/uploads.php?userid=codingtrain 


// Theory:
// A sum of 10 random small numbers (pos or neg, can cancel out) is likely to be small 
// A sum of 50000 random small numbers (pos or neg, can cancel out) is likely to be much larger 
// proof?

// nohidden = 10 will give x close to 0 (y=1/2)
// nohidden = 50000 will give all sorts of x, but likely - large or + large (y = 0 or 1)


//=== Tweaker's box ============================================

// number of nodes in each layer:
const noinput = 2;
const nohidden = 10000;
const nooutput = 1;

// define the exemplars to learn from:
let training_data = [
  {    inputs: [0, 0],    outputs: [0]  },
  {    inputs: [0, 1],    outputs: [1]  },
  {    inputs: [1, 0],    outputs: [1]  },
  {    inputs: [1, 1],    outputs: [0]  }
];

var nn;     // global var 

const learningrate = 0.2; 

// train this number of times per draw()
const notrain = 10;

// Take screenshot on this step:
AB.screenshotStep  = 200;   


// divide 0,1 into squares 
// show all squares or just the corner squares:
var showall = true;

const canvassize    = 400;
const squaresize    = 40;

const cols          = 10 ;
const rows          = 10;


// Matrix.randomize() is changed to point to this. Must be defined by user of Matrix. 

function randomWeight()
{
   return ( AB.randomFloatAtoB ( -0.5, 0.5 ) );
            // Coding Train default is -1 to 1
}    

//=== End of tweaker's box ============================================




function setup() 
{
  createCanvas (canvassize, canvassize);
  
   $.getScript ( "/uploads/codingtrain/matrix.js", function()
   {
        $.getScript ( "/uploads/codingtrain/nn.js", function()
        {
            nn = new NeuralNetwork ( noinput, nohidden, nooutput );
        });
   });
}



function draw() 
{
  // check if libraries loaded yet:
  if ( typeof nn == 'undefined' ) return;
  nn.setLearningRate ( learningrate );

  background ('#ffffcc'); 

 logthem();

  // train n times 
  for (let i = 0; i < notrain ; i++) 
  {
    let data = random ( training_data );
    nn.train ( data.inputs, data.outputs );
  }

// draw either some squares or all squares:

  if ( showall )
  {
    // redraw all squares each time round
    for (let i = 0; i < cols; i++) 
        for (let j = 0; j < rows; j++) 
            drawsquare ( i, j );
  }
  else
  {
    // redraw just the 4 squares 
    for ( let i = 0; i < cols; i = i + cols-1 ) 
        for ( let j = 0; j < rows; j = j + rows-1 ) 
            drawsquare ( i, j );
  }
}  

     
// copy this function logthem
// call it in draw at the right place 
// check all good in console
// change console.log to AB.msg 

function logthem()
{
      let inputs  = [0, 0];      let y  = nn.predict(inputs);
      let inputs2 = [0, 1];      let y2 = nn.predict(inputs2);
      let inputs3 = [1, 0];      let y3 = nn.predict(inputs3);
      let inputs4 = [1, 1];      let y4 = nn.predict(inputs4);
      
   AB.msg ( "<span style='color:red'><b> " +
   "[0,0] leads to " +  y[0].toFixed(2) + "<br>" +
   "[0,1] leads to " + y2[0].toFixed(2) + "<br>" +
   "[1,0] leads to " + y3[0].toFixed(2) + "<br>" +
   "[1,1] leads to " + y4[0].toFixed(2) + "<br>" +
   "</b></span>" );
}
 
 
function drawsquare ( i, j )
{
 //   i = 8;
 //   j = 8;
      let x1 = i / cols;
      let x2 = j / rows;
      let inputs = [x1, x2];
      let y = nn.predict(inputs);
     // console.log ( "input (" +x1 + "," + x2 + ") output " + y );
      
      strokeWeight(2);
      stroke('black');
  //    stroke ('red');
      fill ( y * 255 );      // 0 is black, 1 is white 
      
      rect ( i * squaresize, j * squaresize, squaresize, squaresize );
}