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

​

let model;

let totalClasses;

let myCanvas;

​

let classification = "None Yet";

let probability = "100";

let poser;

let video;

​

​

// 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 setup() {

  myCanvas = createCanvas(400, 400);

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

  videoCanvas = createCanvas(320, 240)

​

  await load();

  video = createCapture(VIDEO, videoReady);

  video.size(320, 240);

  video.hide();

​

}

​

function draw() {

  background(255);

  if(video) image(video,0,0);

  fill(255,0,0)

  textSize(18);

  text("Result:" + classification, 10, 40);

​

  text("Probability:" + probability, 10, 20)

  ///ALEX insert if statement here testing classification against apppropriate part of array for this time in your video

​

  textSize(8);

  if (poser) { //did we get a skeleton yet;

    for (var i = 0; i < poser.length; i++) {

      let x = poser[i].position.x;

      let y = poser[i].position.y;

      ellipse(x, y, 5, 5);

      text(poser[i].part, x + 4, y);

    }

  }

​

}

​

function videoReady() {

  console.log("Video Ready");

  predict();

}

​

​

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(

    video.elt, //webcam.canvas,

    flipHorizontal

  );

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

  const prediction = await model.predict(

    posenetOutput,

    flipHorizontal,

    totalClasses

  );

​

  // console.log(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);

​

  //communicate these values back to draw function with global variables

  classification = sortedPrediction[0].className;

  probability = sortedPrediction[0].probability.toFixed(2);

  if (pose) poser = pose.keypoints; // is there a skeleton

  predict();

}