//const modelURL = 'https://teachablemachine.withgoogle.com/models/r8wsgg5mm/';

// the json file (model topology) has a reference to the bin file (model weights)

const checkpointURL = modelURL + "model.json";

// the metatadata json file contains the text labels of your model and additional information

const metadataURL = modelURL + "metadata.json";

​

const size = 300;

const flip = true; // whether to flip the webcam

let webcam;

let model;

let totalClasses;

let myCanvas;

let ctx;

​

let  finesnd,badsnd;

let cclassname = "";

​

​

async function loadsnds(){

​

​

   // soundFormats('mp3', 'ogg');

 finesnd = loadSound('sounds/fine.mp3');

 badsnd = loadSound('sounds/bad.mp3');

  //sample.playMode(playMode);

​

​

}

​

​

// A function that loads the model from the checkpoint

async function load() {

  model = await tmPose.load(checkpointURL, metadataURL);

  totalClasses = model.getTotalClasses();

  console.log("Number of classes, ", totalClasses);

}

​

async function loadWebcam() {

  webcam = new tmPose.Webcam(size, size, flip); // can change width and height

  await webcam.setup(); // request access to the webcam

  await webcam.play();

  window.requestAnimationFrame(loopWebcam);

}

​

async function setup() {

  myCanvas = createCanvas(size, size);

  ctx = myCanvas.elt.getContext("2d");

  // Call the load function, wait until it finishes loading

 await loadsnds();

  await load();

  await loadWebcam();

}

​

​

async function loopWebcam(timestamp) {

  webcam.update(); // update the webcam frame

  await predict();

  window.requestAnimationFrame(loopWebcam);

}

​

async function predict() {

  // Prediction #1: run input through posenet

  // predict can take in an image, video or canvas html element

  const flipHorizontal = false;

  const { pose, posenetOutput } = await model.estimatePose(

    webcam.canvas,

    flipHorizontal

  );

  // Prediction 2: run input through teachable machine assification model

  const prediction = await model.predict(

    posenetOutput,

    flipHorizontal,

    totalClasses

  );

​

  // console.log('prediction: ', prediction);

  // Sort prediction array by probability

  // So the first classname will have the highest probability

  const sortedPrediction = prediction.sort((a, b) => - a.probability + b.probability);

​

  // Show the result

  const res = select('#res'); // select <span id="res">

  //res.html(sortedPrediction[0].className);

​

  // Show the probability

  const prob = select('#prob'); // select <span id="prob">

  //prob.html(sortedPrediction[0].probability.toFixed(2));

​

  // draw the keypoints and skeleton

  if (pose) {

    drawPose(pose);

    textSize(70);

    textStyle(BOLD);

    fill(255);

    textAlign(CENTER);

    text(sortedPrediction[0].className,width/2,120);

     textSize(20);

     text( sortedPrediction[0].probability.toFixed(2),width/2,80);

    let prop = sortedPrediction[0].probability.toFixed(3);

    // avoid permanet trigger

    if(cclassname != sortedPrediction[0].className && sortedPrediction[0].className !="neutral" && prop >.9){

      if(sortedPrediction[0].className == "down"){

        badsnd.play();

      }

       if(sortedPrediction[0].className == "up"){

          finesnd.play();

      }

      cclassname = sortedPrediction[0].className;

    }

    text( "#1 Arms up / #2 to your knees / #3 neutral ",width/2,height-80);

  }

}

​

function drawPose(pose) {

  if (webcam.canvas) {

    ctx.drawImage(webcam.canvas, 0, 0);

    // draw the keypoints and skeleton

    if (pose) {

      const minPartConfidence = 0.5;

      tmPose.drawKeypoints(pose.keypoints, minPartConfidence, ctx);

      tmPose.drawSkeleton(pose.keypoints, minPartConfidence, ctx);

    }

  }

}

​