ppmi_motioncode.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import joblib\n",
    "import os\n",
    "from motion_code import MotionCode\n",
    "from data_processing import process_ppmi_data_for_motion_codes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = joblib.load('../notebooks/artifacts/labels.pkl')\n",
    "time_series = joblib.load('../notebooks/artifacts/time_series.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_split(x, split=0.2):\n",
    "    sp_idx = int(len(x)*split)\n",
    "    test_x, train_x = x[:sp_idx], x[sp_idx:]\n",
    "    return train_x, test_x\n",
    "\n",
    "def pad_zeros(x):\n",
    "    maxlen = max([len(item) for item in x])\n",
    "    new_x = np.zeros((len(x), maxlen))\n",
    "    for i, item in enumerate(x):\n",
    "        new_x[i, :len(item)] = item\n",
    "    return new_x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "time_series = pad_zeros(time_series)\n",
    "Y_train, Y_test = train_test_split(time_series)\n",
    "labels_train, labels_test = train_test_split(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(216, 256) (216, 256) (216,)\n",
      "(53, 256) (53, 256) (53,)\n"
     ]
    }
   ],
   "source": [
    "X_train, Y_train, labels_train = process_ppmi_data_for_motion_codes(Y_train, labels_train)\n",
    "X_test, Y_test, labels_test = process_ppmi_data_for_motion_codes(Y_test, labels_test)\n",
    "\n",
    "print(X_train.shape, Y_train.shape, labels_train.shape)\n",
    "print(X_test.shape, Y_test.shape, labels_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = MotionCode(m=20, Q=1, latent_dim=2, sigma_y=0.1)\n",
    "# Then we train model on the given X_train, Y_train, label_train set and saved it to a file named test_model.\n",
    "model_path = 'saved_models/' + 'test_model_ppmi'\n",
    "os.makedirs(model_path, exist_ok=True)\n",
    "# model.fit(X_train, Y_train, labels_train, model_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True label: [1, 2, 1, 0, 2, 1, 2, 0, 0, 0]\n",
      "Predicted label: [1, 2, 1, 0, 2, 1, 1, 0, 0, 0]\n"
     ]
    }
   ],
   "source": [
    "# Next we load the model and perform classification on 10 random test time series.\n",
    "model.load(model_path)\n",
    "num_test = labels_test.shape[0]\n",
    "random_indices = np.random.choice(np.arange(num_test), size=10, replace=False)\n",
    "predicted_labels = []\n",
    "for ind in random_indices:\n",
    "    predicted_labels.append(model.classify_predict(X_test[ind], Y_test[ind]))\n",
    "print('True label:', list(labels_test[random_indices]))\n",
    "print('Predicted label:', predicted_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                   "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accurary: 0.8113207547169812\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r"
     ]
    }
   ],
   "source": [
    "# Now we can go on to perform prediction on the whole test set and get accuracy\n",
    "acc = model.classify_predict_on_batches(X_test_list=X_test, Y_test_list=Y_test, true_labels=labels_test)\n",
    "print('Accurary:', acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from visualize import visualize_data_by_GP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = (X_train, Y_train, labels_train, X_test, Y_test, labels_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize inducing points for each train sample\n",
    "visualize_data_by_GP(data, num_motion=model.num_motion, m=10, Q=1, \n",
    "                     label_names=[f'lbl_{m}' for m in range(model.num_motion)], \n",
    "                     plot_path='out/gp_ppmi_cluster_visual.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cfps",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}