ECG_Mortara.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "fdf64122",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# **ECG HEARTBEAT CLASSIFICATION WITH ML - Mortara**\n",
    "\n",
    "by Erica Brisigotti (2097202), Ekaterina Chueva (2072050), Sofia Pacheco Garcia (2070771), Nadillia Sahputra (2070770)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6fdecb94-b3ec-485b-a4bc-b55dfb42a796",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import wfdb\n",
    "import wfdb.processing\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "from matplotlib import pyplot as plt\n",
    "from scipy import fftpack, stats, signal\n",
    "import pywt\n",
    "from scipy import stats\n",
    "import time\n",
    "import copy\n",
    "import pandas.errors as pe\n",
    "\n",
    "import sklearn\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics import balanced_accuracy_score\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.utils.class_weight import compute_sample_weight\n",
    "\n",
    "import tensorflow as tf\n",
    "import xgboost\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.svm import SVC\n",
    "from scikeras.wrappers import KerasClassifier\n",
    "import joblib\n",
    "from joblib import Parallel, delayed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b3ea9f17-2084-446d-8575-28196d4bbcbf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# firstly, we set\n",
    "# - a saving option\n",
    "FILE_SAVE = False\n",
    "# - a second saving option for redudant yet maybe worth doing for small datasets\n",
    "EXTRA_SAVE_FOR_SMALL_DATA = False\n",
    "# - an option to show the structure of the datasets anayzed (both in general and for the specific chosen example)\n",
    "SHOW_STRUCTURE = False \n",
    "# - the color of the records from a healthier patient in plots\n",
    "ch = 'forestgreen'\n",
    "# - the color of the annotations from a healthier patient in plots\n",
    "cha = 'navy'\n",
    "# - an additional color for a healthier patient in plots\n",
    "che = 'deepskyblue'\n",
    "# - the color of the records from a more sick patient in plots\n",
    "cs = 'darkorange'\n",
    "# -  the color of the annotaitons from a more sick patient in plots\n",
    "csa = 'red' \n",
    "# - an additional color for a more sick patient in plots\n",
    "cse = 'maroon'\n",
    "# - an option for the classification algorithm (between xgboost and ??)\n",
    "ML_ALGORITHM = 'xgboost'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "id": "09dedc10",
   "metadata": {},
   "outputs": [],
   "source": [
    "# we also define a function that allows to quickly display the distribution of its input\n",
    "def distribution(arg, a1, a2, title='arg'):\n",
    "    '''\n",
    "    function that returns the distribution of arg (which name/meaning is encoded in title)\n",
    "    '''\n",
    "    average = np.median(arg)\n",
    "    std = np.std(arg)\n",
    "    fig, ax = plt.subplots(1,1, figsize=(12,4))\n",
    "    info = ax.hist(arg, bins=np.arange(np.min(arg), np.max(arg)+1, 1), align='left', density=True, edgecolor='none', \n",
    "                   color='black', alpha=0.8, label = 'distribution of '+title)\n",
    "    ax.vlines(average, 0, 1, color='magenta', label='average')\n",
    "    ax.vlines(average+std, 0, 1, color='gold', label='1 $\\sigma$')\n",
    "    ax.vlines(average-std, 0, 1, color='gold')\n",
    "    ax.set_xlabel(title)\n",
    "    ax.set_ylabel('density')\n",
    "    ax.set_xlim(a1, a2)\n",
    "    ax.set_ylim(0, 1.1*np.max(info[0]))\n",
    "    ax.set_title('distribution of '+title)\n",
    "    ax.legend()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7fc85ecf",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 1A. Data Preprocessing Example"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2241d43c",
   "metadata": {},
   "source": [
    "We set the parameters for the preprocessing part by we choosing two random datasets, *one from the first half and one from second half of the set of patients. Such a split is justified by the knowledge that the second half of patients is generally more sick (since it was purposefully introduced to insert more sick patients).*\n",
    "\n",
    "To reach such a goal, we upload the two WFDB records using the <code>rdrecord</code> function. They are uploaded as <code>wfdb.Record</code> objects, with many different attributes. Of such attributes, we are interested in the sampling frequency <code>.fs</code> and in the time series, which is stored under as <code>.p_signal</code>.\n",
    "\n",
    "Similarly, the annotations for the two patients are uploaded using the <code>rdann</code> function, that retuns <code>wfdb.Annotation</code> objects with many attributes. Of such attributes, we are interested in the annotation indexes (saved as <code>.sample</code> ) and symbols (under <code>.symbol</code>)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fcf48f82",
   "metadata": {},
   "outputs": [],
   "source": [
    "t_start_all = time.time()\n",
    "\n",
    "address_1 = \"Hour1RawData\"\n",
    "address_2 = \"Hour3RawData\"\n",
    "record_1 = wfdb.rdrecord('input_files/'+address_1)\n",
    "record_2 = wfdb.rdrecord('input_files/'+address_2)\n",
    "leads = record_1.sig_name\n",
    "\n",
    "if SHOW_STRUCTURE:\n",
    "    print('the structure of the wfdb.Record object is:')\n",
    "    display(record_1.__dict__)\n",
    "    print('\\nabout patient '+address_1)\n",
    "    with open(\"input_files/\"+address_1+\".hea\", 'r') as file:\n",
    "        t = file.read().replace('#',\"\")\n",
    "        hea = t.split('\\n')\n",
    "        print('\\n',hea[3],\": the first two data are info about the age and gender\")\n",
    "        temp = hea[3].split(\" \")\n",
    "        print(temp[1],\" is the age of the patient and \",temp[2],\"is the gender of the patient\")\n",
    "        print(hea[4],\": are the meds the patient is on\")\n",
    "        print(hea[5],\": are the notes for the patient\")\n",
    "    print('\\nabout patient '+address_2)\n",
    "    with open(\"input_files/\"+address_1+\".hea\", 'r') as file:\n",
    "        t = file.read().replace('#',\"\")\n",
    "        hea = t.split('\\n')\n",
    "        print('\\n',hea[3],\": the first two data are info about the age and gender\")\n",
    "        temp = hea[3].split(\" \")\n",
    "        print(temp[1],\" is the age of the patient and \",temp[2],\"is the gender of the patient\")\n",
    "        print(hea[4],\": are the meds the patient is on\")\n",
    "        print(hea[5],\": are the notes for the patient\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a4865385",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# we choose to plot just 2 out of the 12 leads\n",
    "s0_1 = record_1.p_signal[:,0]\n",
    "s1_1 = record_1.p_signal[:,7]\n",
    "s0_2 = record_2.p_signal[:,0]\n",
    "s1_2 = record_2.p_signal[:,7]\n",
    "freq = record_1.fs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63d6019d",
   "metadata": {},
   "source": [
    "We plot the records and annotations in both leads for both patients:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "637df436",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x800 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig1, ax1 = plt.subplots(2,2,figsize = (15, 8), sharex = True)\n",
    "start_point = 25100\n",
    "end_point = start_point+500\n",
    "\n",
    "ax1[0][0].plot(s0_1, '-', label='I', color=ch)\n",
    "ax1[0][0].set_xlim(start_point, end_point)\n",
    "ax1[0][0].set_ylim(min(s0_1[start_point:end_point])-0.5, max(s0_1[start_point:end_point])+0.5)\n",
    "ax1[0][0].set_ylabel('I [mV]', fontsize=12)\n",
    "ax1[0][0].set_title('patient '+address_1, fontsize=14)\n",
    "ax1[0][0].legend(ncol=2, fontsize=10)\n",
    "ax1[0][0].grid()\n",
    "\n",
    "ax1[0][1].plot(s0_2, '-', label='I', color=cs)\n",
    "ax1[0][1].set_ylim(min(s0_2[start_point:end_point])-0.5, max(s0_2[start_point:end_point])+0.5)\n",
    "ax1[0][1].set_ylabel('I [mV]', fontsize=12)\n",
    "ax1[0][1].set_title('patient '+address_2, fontsize=14)\n",
    "ax1[0][1].legend(ncol=2, fontsize =10)\n",
    "ax1[0][1].grid()\n",
    "\n",
    "ax1[1][0].plot(s1_1, '-', label='V4', color=ch)\n",
    "ax1[1][0].set_ylim(min(s1_1[start_point:end_point])-0.5, max(s1_1[start_point:end_point])+0.5)\n",
    "ax1[1][0].set_xlabel('index', fontsize=12)\n",
    "ax1[1][0].set_ylabel('V4 [mV]', fontsize=12)\n",
    "ax1[1][0].legend(ncol=2, fontsize=10)\n",
    "ax1[1][0].grid()\n",
    "\n",
    "ax1[1][1].plot(s1_2, '-', label='V4', color=cs)\n",
    "ax1[1][1].set_ylim(min(s1_2[start_point:end_point])-0.5, max(s1_2[start_point:end_point])+0.5)\n",
    "ax1[1][1].set_xlabel('index', fontsize=12)\n",
    "ax1[1][1].set_ylabel('V4 [mv]', fontsize=12)\n",
    "ax1[1][1].legend(ncol=2, fontsize=10)\n",
    "ax1[1][1].grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f4c5c929",
   "metadata": {},
   "source": [
    "We implement filtering since the signal is visibly noisy and includes a baseline wonder. We filter with both fast Fourier transform and Wavelet transform tecniques, which are applied serially.\n",
    "\n",
    "First of all, for the **Fast Fourier Transform**, we want to display and analyze the frequency spectra (which is extracted by the <code>power_fft</code> function)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "84685f81",
   "metadata": {},
   "outputs": [],
   "source": [
    "def power_fft(vect, timestamp):\n",
    "    '''\n",
    "    function that returns the fft, absolute value of fft and range of\n",
    "    frequencies (vect_all_freq) used in the fft\n",
    "    '''\n",
    "    vect_fft = fftpack.fft(vect)\n",
    "    vect_power = np.abs(vect_fft)\n",
    "    vect_all_freq = fftpack.fftfreq(vect.size, d = timestep)\n",
    "    return vect_fft, vect_power, vect_all_freq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3a27c0b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "timestep = 1/freq\n",
    "norm_fft_1, norm_power_1, norm_all_freq_1 = power_fft(s0_1, timestep)\n",
    "norm_fft_2, norm_power_2, norm_all_freq_2 = power_fft(s0_2, timestep)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "5ba171e8",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x300 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig2, ax2 = plt.subplots(1, 2, figsize = (15, 3), sharex = True)\n",
    "\n",
    "ax2[0].plot(norm_all_freq_1, norm_power_1, '-', label='I spectrum ', color=ch)\n",
    "ax2[0].set_xlim(0,max(norm_all_freq_1))\n",
    "ax2[0].set_ylim(10**(-2), 10**5)\n",
    "ax2[0].set_yscale('log')\n",
    "ax2[0].set_xlabel('frequency [Hz]', fontsize=12)\n",
    "ax2[0].set_ylabel('power', fontsize=12)\n",
    "ax2[0].set_title('patient '+address_1, fontsize=14)\n",
    "ax2[0].legend(loc='upper right', ncol=2, fontsize=10)\n",
    "ax2[0].grid()\n",
    "\n",
    "ax2[1].plot(norm_all_freq_2, norm_power_2, '-', label='I spectrum ', color=cs)\n",
    "ax2[1].set_ylim(10**(-2),10**5)\n",
    "ax2[1].set_yscale('log')\n",
    "ax2[1].set_xlabel('frequency [Hz]', fontsize=12)\n",
    "ax2[1].set_ylabel('power', fontsize=12)\n",
    "ax2[1].set_title('patient '+address_2, fontsize=14)\n",
    "ax2[1].legend(loc='upper right', ncol=2, fontsize=10)\n",
    "ax2[1].grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "249372b0",
   "metadata": {},
   "source": [
    "We notice a peak at around 60 Hz, which is due to the power supply of the recording device and must be discarded. We also see high values of the spectrum for low frequency, which should be discarded as well.\n",
    "\n",
    "Therefore, we decide to implement a band-pass filter: the precise cut frequencies are chosen in the interval $[0.1-55]$ Hz."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "51e35baa",
   "metadata": {},
   "outputs": [],
   "source": [
    "down_freq_lim = np.array([0.1, 1, 2])\n",
    "up_freq_lim = np.array([25, 35, 45])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4ea8ab8c",
   "metadata": {},
   "source": [
    "We notice a peak at around 60 Hz, which is due to the power supply of the recording device and must be discarded. We also see high values of the spectrum for low frequency, which should be discarded as well.\n",
    "\n",
    "Therefore, we decide to implement a band-pass filter: the precise cut frequencies are chosen in the interval $[0.1-50]$ Hz.\n",
    "\n",
    "The cut frequency are chosen visually based on the results of their filtering (which is done by the <code>filtered_power</code>) that are shown in the plot below.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f5670b2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def filtered_power(vect_fft, vect_all_freq, down_freq_lim, up_freq_lim):\n",
    "    '''\n",
    "    function that filtered the signla starting from its fft, by limiting\n",
    "    the range of frequencies from vect_all_freq to the range\n",
    "    [down_freq_lim, up_freq_lim]\n",
    "    '''\n",
    "    vect_filtered_fft = vect_fft.copy()\n",
    "    vect_filtered_fft[np.abs(vect_all_freq) >= up_freq_lim] = 0\n",
    "    vect_filtered_fft[np.abs(vect_all_freq) <= down_freq_lim] = 0\n",
    "    vect_filtered = (fftpack.ifft(vect_filtered_fft)).real\n",
    "    return vect_filtered"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4983833f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x1500 with 9 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig3, ax3 = plt.subplots(len(down_freq_lim), len(up_freq_lim), figsize = (15, 15), sharex=True, sharey=True)\n",
    "\n",
    "start_point = 25150\n",
    "end_point = start_point+500\n",
    "\n",
    "for i in range(len(down_freq_lim)):\n",
    "    for j in range(len(up_freq_lim)):\n",
    "        norm_filtered_1 = filtered_power(norm_fft_1, norm_all_freq_1, down_freq_lim[i], up_freq_lim[j])\n",
    "        x_1 = np.arange(0, len(norm_filtered_1), 1)\n",
    "        norm_filtered_2 = filtered_power(norm_fft_2, norm_all_freq_2, down_freq_lim[i], up_freq_lim[j])\n",
    "        temp = np.concatenate([norm_filtered_1[start_point:end_point],norm_filtered_2[start_point:end_point]])\n",
    "        ax3[i][j].plot(norm_filtered_1, '-', label=address_1+' FFT', color=ch)\n",
    "        ax3[i][j].plot(norm_filtered_2, '-', label=address_2+' FFT', color=cs)\n",
    "        ax3[i][j].grid()\n",
    "        ax3[i][j].legend(fontsize=10, ncol=2)\n",
    "        ax3[i][j].set_xlim(start_point, end_point)\n",
    "        ax3[i][j].set_ylim(1.3*(min(temp)-0.5), 1.3*(max(temp)+0.5))\n",
    "        if i==0:\n",
    "            ax3[i][j].set_title(\"to \"+str(up_freq_lim[j])+\" Hz\", fontsize=14)\n",
    "        if j==0:\n",
    "            ax3[i][j].set_ylabel('from\\n'+str(down_freq_lim[i])+\" Hz\", fontsize=14, rotation=0, ha=\"right\")\n",
    "        if i==(len(up_freq_lim)-1):\n",
    "            ax3[i][j].set_xlabel('index', fontsize=12)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2692aed0",
   "metadata": {},
   "source": [
    "From the plots we can see that:\n",
    "- for small upper limits (e.g. 10 Hz), the filtered signal is too smooth and there is loss of important information\n",
    "- for intermediate upper limits (e.g. 30 Hz), the filtered signal is still quite noisy\n",
    "- for nigher upper limits (e.g. 50 Hz), the filtered signal is smoother without major loss of information\n",
    "- there is no visible variation coming from the different lower limits\n",
    "\n",
    "Therefore, we decide to set:\n",
    "- the lower cut frequency as the lowest value, 0.1 Hz\n",
    "- the upper cut frequency as the highest value available, 50 Hz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "74d932ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "down_freq_lim = 1\n",
    "up_freq_lim = 50"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de67fabe",
   "metadata": {},
   "source": [
    "We filter the results shown above with **disctere wavelet transform**. We use a sym4 wavelet and a maximum possible decomposition level.\n",
    "\n",
    "Lastly, to be safe we want to **eliminate the baseline wander** of the signal to align the isoelectric line. We do this by substracting a low order polynomial function from the signal. </span> Just like before, we use the <code>stats.zscore</code> function to standardize the output. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "67117d09",
   "metadata": {},
   "outputs": [],
   "source": [
    "def isoelectric_signal(vect, freq, down_freq_lim, up_freq_lim):\n",
    "    '''\n",
    "    function that removes baseline wander, performs band-pass filetring using FFT and DWT filtering\n",
    "    input: raw signal, sampling frequency, lower/upper limits of band-pass filter\n",
    "    output: filtered signal with mean=0 and standard_deviation=1.\n",
    "    '''\n",
    "    timestep = 1/freq\n",
    "    vect_fft = fftpack.fft(vect)\n",
    "    vect_power = np.abs(vect_fft)\n",
    "    vect_all_freq = fftpack.fftfreq(vect.size, d = timestep)\n",
    "    \n",
    "    vect_filtered_fft = vect_fft.copy()\n",
    "    vect_filtered_fft[np.abs(vect_all_freq) >= up_freq_lim] = 0\n",
    "    vect_filtered_fft[np.abs(vect_all_freq) <= down_freq_lim] = 0\n",
    "    vect_filtered = (fftpack.ifft(vect_filtered_fft)).real\n",
    "    \n",
    "    \n",
    "    w = pywt.Wavelet('sym4')\n",
    "    maxlev = pywt.dwt_max_level(len(vect), w.dec_len)\n",
    "    threshold = 0.04 # Threshold for filtering\n",
    "    coeffs = pywt.wavedec(vect_filtered, 'sym4', level=maxlev)\n",
    "    for i in range(1, len(coeffs)):\n",
    "        coeffs[i] = pywt.threshold(coeffs[i], threshold*max(coeffs[i]))\n",
    "    datarec = pywt.waverec(coeffs, 'sym4')\n",
    "    \n",
    "    y1 = stats.zscore(datarec)\n",
    "    x=range(len(y1))\n",
    "\n",
    "    coefs_filt_1=np.polyfit(x, y1, 8, rcond=None, full=False, w=None, cov=False)\n",
    "\n",
    "    isoelectric_signal_1=y1-np.polyval(coefs_filt_1, x)\n",
    "    isoelectric_signal_1 = stats.zscore(isoelectric_signal_1)\n",
    "\n",
    "    return (isoelectric_signal_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "6264c1a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1500x800 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig4, ax4 = plt.subplots(2,2,figsize = (15, 8), sharex = True)\n",
    "start_point = 25000\n",
    "end_point = start_point+500\n",
    "\n",
    "f_s0_1 = isoelectric_signal(s0_1,freq,down_freq_lim, up_freq_lim)\n",
    "f_s0_2 = isoelectric_signal(s0_2,freq,down_freq_lim, up_freq_lim)\n",
    "\n",
    "f_s1_1 = isoelectric_signal(s1_1,freq,down_freq_lim, up_freq_lim)\n",
    "f_s1_2 = isoelectric_signal(s1_2,freq,down_freq_lim, up_freq_lim)\n",
    "\n",
    "ax4[0][0].plot(s0_1, '-', label='I unfiltered', color=ch)\n",
    "ax4[0][0].plot(f_s0_1, '-', label='I filtered', color=che)\n",
    "ax4[0][0].set_xlim(start_point, end_point)\n",
    "temp = np.concatenate([s0_1[start_point:end_point], f_s0_1[start_point:end_point]])\n",
    "ax4[0][0].set_ylim(1.3*min(temp), 1.3*max(temp))\n",
    "ax4[0][0].set_ylabel('I [mV]', fontsize=12)\n",
    "ax4[0][0].set_title('patient '+address_1, fontsize=14)\n",
    "ax4[0][0].legend(ncol=3, fontsize=10)\n",
    "ax4[0][0].grid()\n",
    "\n",
    "ax4[0][1].plot(s0_2, '-', label='I unfiltered', color=cs)\n",
    "ax4[0][1].plot(f_s0_2, '-', label='I filtered', color=cse)\n",
    "temp = np.concatenate([s0_2[start_point:end_point], f_s0_2[start_point:end_point]])\n",
    "ax4[0][1].set_ylim(1.3*min(temp), 1.3*max(temp))\n",
    "ax4[0][1].set_ylabel('I [mV]', fontsize=12)\n",
    "ax4[0][1].set_title('patient '+address_2, fontsize=14)\n",
    "ax4[0][1].legend(ncol=3, fontsize=10)\n",
    "ax4[0][1].grid()\n",
    "\n",
    "ax4[1][0].plot(s1_1, '-', label='V4 unfiltered', color=ch)\n",
    "ax4[1][0].plot(f_s1_1, '-', label='V4 filtered', color=che)\n",
    "temp = np.concatenate([s1_1[start_point:end_point], f_s1_1[start_point:end_point]])\n",
    "ax4[1][0].set_ylim(1.3*min(temp), 1.3*max(temp))\n",
    "ax4[1][0].set_xlabel('index', fontsize=12)\n",
    "ax4[1][0].set_ylabel('V5 [mV]', fontsize=12)\n",
    "ax4[1][0].legend(ncol=3, fontsize=10)\n",
    "ax4[1][0].grid()\n",
    "\n",
    "ax4[1][1].plot(s1_2, '-', label='V4 unfiltered', color=cs)\n",
    "ax4[1][1].plot(f_s1_2, '-', label='V4 filtered', color=cse)\n",
    "temp = np.concatenate([s1_2[start_point:end_point], f_s1_2[start_point:end_point]])\n",
    "ax4[1][1].set_ylim(1.3*min(temp), 1.3*max(temp))\n",
    "ax4[1][1].set_xlabel('index', fontsize=12)\n",
    "ax4[1][1].set_ylabel('V5 [mv]', fontsize=12)\n",
    "ax4[1][1].legend(ncol=3, fontsize=10)\n",
    "ax4[1][1].grid()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d5b7003",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 1B) Data Preprocessing on the entire dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e84e67a2",
   "metadata": {},
   "source": [
    "We avoid saving the unfiltered signal in a .csv format, but the option is there if needed: in such a case, the data relative to each patient is saved in a dataframe, in which each columns correspond to one lead."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "1b53f0dc",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "r  = np.genfromtxt(\"input_files/RECORDS\", dtype='str')\n",
    "records = list(range(1,(len(r)+1)))\n",
    "\n",
    "if EXTRA_SAVE_FOR_SMALL_DATA:\n",
    "    t1 = time.time()\n",
    "    for path in records:\n",
    "        # Load the record\n",
    "        path = str(path)\n",
    "        record = wfdb.rdrecord(\"input_files/Hour\"+path+\"RawData\")\n",
    "        signal = record.sig_name\n",
    "        df = pd.DataFrame()\n",
    "        for i in range(0,len(signal)):\n",
    "            df[signal[i]] = record.p_signal[:, i]\n",
    "\n",
    "        # Save the DataFrame to a CSV file\n",
    "        outfn = 'output_files/'+'data'+path+'.csv'\n",
    "        df.to_csv(outfn, index=False)\n",
    "    t2 = time.time()\n",
    "    print('the metadata saving took',t2-t1,'seconds')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f967a87",
   "metadata": {},
   "source": [
    "Finally, we save the filtered signal in a .csv format: the data relative to each patient is saved in a dataframe, in which each columns correspond to one lead."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0f8bb3ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the filtered data saving took 1444.082210302353 seconds\n"
     ]
    }
   ],
   "source": [
    "r  = np.genfromtxt(\"input_files/RECORDS\", dtype='str')\n",
    "records = list(range(1,(len(r)+1)))\n",
    "FILE_SAVE = True\n",
    "if FILE_SAVE:\n",
    "    t1 = time.time()\n",
    "    # for every file provided\n",
    "    for path in records:\n",
    "        path = str(path)\n",
    "        # we load the record\n",
    "        record = wfdb.rdrecord(\"input_files/Hour\"+path+\"RawData\")\n",
    "        # we add the names of the missing leads (which can be computed from the 8 that are already measured)\n",
    "        signal = record.sig_name \n",
    "        moreleads = ['III', 'aVR', 'aVL', 'aVF']\n",
    "        signal= np.append(signal,moreleads)\n",
    "        # we add the missing leads\n",
    "        psig = record.p_signal\n",
    "        III = record.p_signal[:,2]-record.p_signal[:,1]\n",
    "        aVR = (-record.p_signal[:,1]-record.p_signal[:,2])/2\n",
    "        aVL = (record.p_signal[:,1]-III)/2\n",
    "        aVF = (record.p_signal[:,2]+III)/2\n",
    "        # and fix their shape\n",
    "        III = np.reshape(III,(len(psig),1))\n",
    "        aVR = np.reshape(aVR,(len(psig),1))\n",
    "        aVL = np.reshape(aVL,(len(psig),1))\n",
    "        aVF = np.reshape(aVF,(len(psig),1))\n",
    "        psig = np.hstack((psig,III,aVR,aVL,aVF))\n",
    "        # we initialize a dataframe in which we're going to save the filtered signal\n",
    "        df = pd.DataFrame()\n",
    "        mask = np.arange(0, record.p_signal.shape[0], 1)\n",
    "        \n",
    "        # for every lead\n",
    "        for i in range(0,len(signal)):\n",
    "            # we save the relative signal\n",
    "            df[signal[i]] = isoelectric_signal(psig[mask, i],freq,down_freq_lim, up_freq_lim)            \n",
    "\n",
    "        # Save the DataFrame to a CSV file\n",
    "        outfn = \"output_files/\"+\"data\"+path+'_2filtered.csv'\n",
    "        df.to_csv(outfn, index=False)\n",
    "    t2 = time.time()\n",
    "    print('the filtered data saving took',t2-t1,'seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3f2d41dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# for future purposes, we add the new leads to the stored list of leads \n",
    "leads = record_1.sig_name\n",
    "if len(leads)<=8:\n",
    "    moreleads = ['III', 'aVR', 'aVL', 'aVF']\n",
    "    leads= np.append(leads,moreleads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "690a328b",
   "metadata": {},
   "outputs": [],
   "source": [
    "main_lead = 'II'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "28e621c4",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 2) R-peaks detection"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3d91cc7-cbb5-4dc2-9023-8f56d38c0ba0",
   "metadata": {
    "tags": []
   },
   "source": [
    "Since we have the filtered signal $y$, we can implement an algorithm for the detection of the R peaks of the wave based on source [3].\n",
    " \n",
    "Our algorithm finds the peaks by minimizing the error $\\varepsilon$ between the parabola fit of consecutive signal windows and the actual signal. We use the quadratic mean parabola-fit error $\\varepsilon (i)$ : \n",
    " \n",
    "$$ \\varepsilon(i)=\\frac{1}{2 w+1} \\sum_{j=i-w}^{i+w}\\left(y(j)-y^{\\prime}(j)\\right)^2  \\; \\; \\text{for} \\; \\;\n",
    " i \\in \\{w, w+1, w+2, \\cdots, N-w-1\\} $$ \n",
    " \n",
    "where:  \n",
    "- $j \\in\\{i-w, i-w+1, \\cdots, i, i+1, i+2, \\cdots, i+w-1, i+w\\}$\n",
    " \n",
    "- $w$ is half of the window size, i.e. is a positive number such that $2w +1 = m< N $. Particularly, $m=2w+1$ is the length of both $y(j)$ and $y\\prime(j)$ and of the window that should contain the R-peak\n",
    " \n",
    "- $y(j)$ is the sequence defined by fragmenting the filtered signal $y$\n",
    " \n",
    "- $y\\prime(j)=a(i-j)^2+v$ is the sequence defined by the parabola fit, which has vertex at index $i$: \n",
    "\n",
    " \n",
    "The window size for R peak detection is defined by the average duration of the QRS complex (from 70 to 100 ms). \n",
    "Rounding down the QRS size (to 60 ms) and diving by 2 to get half of the window size, we get $ W = 0.060 \\text{s} \\: / \\: 2 = 0.030 \\text{s} $.\n",
    "This can be converted into indexes/AU by multiplying by the sampling frequency of the dataset $f_s$\n",
    "$$w = 0.030 \\: \\text{s} \\cdot f_s$$ \n",
    "     \n",
    "     \n",
    "So, our goal is to calculate the parameters that make $y^{\\prime}(j)$ to best fit $y(j)$. We simplify the task by changing variable to $x=(i-j)^2$. In this way, we can treat the problem as a linear regression $y^{\\prime}(j)=ax+v$, where the vertex of the parabola will be located at $(i,v)$. \n",
    " \n",
    "The two fitting parameters are computed such that:\n",
    " \n",
    "$$a=\\frac{m \\sum_{j=l_1}^{l_2} x(j) y(j)-\\sum_{j=l_1}^{l_2} x(j) \\sum_{j=l_1}^{l_2} y(j)}{m \\sum_{j=l_1}^{l_2} x^2(j)-\\left(\\sum_{j=l_1}^{l_2} x(j)\\right)^2}$$ \n",
    " \n",
    "$$v=\\frac{\\sum_{j=l_1}^{l_2} x^2(j) \\sum_{j=l_1}^{l_2} y(j)-\\sum_{j=l_1}^{l_2} x(j) y(j) \\sum_{j=l_1}^{l_2} x(j)}{m \\sum_{j=l_1}^{l_2} x^2(j)-\\left(\\sum_{j=l_1}^{l_2} x(j)\\right)^2}$$ \n",
    " \n",
    "where $l_1=i-w$, $l_2=i+w$ are the limits of the sums. \n",
    " \n",
    "In this way, we generate a parabola $y^{\\prime}$ that can adjust to the values in the time window (which will go through the entire signal). When the processed signal $y(i)$ reaches a peak, the fitting algorithm finds a parabola such that the vertex of the parabola will coincide with the maximum or minimum point of the wave. If the assumption for a parabola-shaped peak is correct, the fit is be very close to the actual signal. \n",
    " \n",
    "Nevertheless, such a fit can also resemble parts of the signal that are similar to straight lines since the corresponding values of $\\varepsilon(i)$ are low. To avoid finding \"false\" r-peaks, we place a lower bound to the $a$ coefficient of the parabola to differenciate a straight line ($a \\rightarrow 0$) from a wave."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c78ad23e",
   "metadata": {},
   "outputs": [],
   "source": [
    "peak_width=0.018\n",
    "freq=1000\n",
    "w = int(peak_width*freq) #number of samples in half of the window\n",
    "m = 2*w+1 #number of samples in a window"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "aa61ac04",
   "metadata": {},
   "outputs": [],
   "source": [
    "def yfit(yn, fs=freq, peak_width=peak_width): \n",
    "    '''    \n",
    "    This function receives as input parameters yn (ECG filtered signal), fs (frequency with which the signal was sampled) \n",
    "    and peak_width (approximate half-width of an R peak).     \n",
    "    The function performs a linear regression for each window of the signal, obtaining the coefficients of the linear fit.    \n",
    "    It also calculates the quadratic mean parabola-fit error 𝜀(𝑖) for each window.\n",
    "    Three arrays are obtained as an output: the first one, eps_i, containins the mean parabola fit error for each window.\n",
    "    The other two, a and v, correspond to the coefficients of the fit for each window.   \n",
    "    '''\n",
    "    \n",
    "    w = int(peak_width*fs) #number of samples in half of the window\n",
    "    m = 2*w+1 #number of samples in a window\n",
    "    \n",
    "    #i is a sequence that goes from w(half size of a window) to N-w being N the total number of data\n",
    "    i = np.arange(w,len(yn)-w)\n",
    "    #j is a matrix, each row is the array of the indexes of a window\n",
    "    j = np.zeros((len(i),m-1))\n",
    "    for k in range(len(i)):\n",
    "        j[k] = np.arange(k,k+m-1)\n",
    "\n",
    "    #x is the change of variable such that instead of a 2nd order fit, we can perform a linear fit\n",
    "    x = (j[0]-i[0])**2\n",
    "    \n",
    "    #yj is a matrix, each row is an array with the signal (data) values for each of the indexes of a window\n",
    "    yj = yn[j.reshape(j.shape[0]*j.shape[1])].values.reshape(j.shape[0], j.shape[1])\n",
    "\n",
    "    #a and v are the parameters of the fit such that yj_fit=ax+v. They both are arrays containing the parameters for each i.\n",
    "    a = ((m*np.sum((x*yj), axis=1))-(np.sum(x)*np.sum(yj, axis=1)))/(m*np.sum(x**2)-(np.sum(x))**2)\n",
    "    v = ((np.sum(x**2)*np.sum(yj, axis=1))-(np.sum((x*yj), axis=1)*np.sum(x)))/(m*np.sum(x**2)-(np.sum(x))**2)\n",
    "    \n",
    "    #yj_fit=ax+v is a matrix that contains in each row the fitted signal for each window \n",
    "    yj_fit = (np.tile(a.reshape(len(a)), (m-1,1)).T)*x + (np.tile(v.reshape(len(v)), (m-1,1)).T)\n",
    "\n",
    "    #eps_i is the Quadratic mean parabola-fit error  𝜀(𝑖) for each i\n",
    "    eps_i = (1/(2*w+1))*np.sum(((yj-yj_fit)**2), axis=1)\n",
    "    \n",
    "    return(eps_i, a, v)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "658b72ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "def R_fit(yn, fs=freq, peak_width=peak_width): \n",
    "    '''    \n",
    "    This function receives as input parameters yn (ECG filtered signal), fs (frequency with which the signal was sampled) \n",
    "    and peak_width (approximate half-width of an R peak). \n",
    "    \n",
    "    The function returns an array R_i with the indexes of the R peaks, found by impossing the corresponding conditions     \n",
    "    on the output vectors of the yfit function on the signal (a and v fit coefficients and the error).\n",
    "    \n",
    "    Conditions:\n",
    "    1) The parabola fit error must be a local minimum    \n",
    "    2) We impose a lower bound in the 'a' coefficient of the linear regression to get rid of false Rpeaks on straight lines.\n",
    "    '''\n",
    "    \n",
    "    w = int(peak_width*fs) #number of samples in half of the window\n",
    "    m = 2*w+1 #number of samples in a window\n",
    "    \n",
    "    y_fit = yfit(yn, fs, peak_width)\n",
    "    i = np.arange(0,len(y_fit[1]),1, dtype=int)\n",
    "    a = y_fit[1]\n",
    "    v = y_fit[2]\n",
    "    \n",
    "    #CONDITIONS\n",
    "    #condition 1: it checks that the R points are minima of the Quadratic mean parabola-fit error  𝜀(𝑖)\n",
    "    c = np.arange(len(yn)-m+1)[2:]\n",
    "    cond_min_1 = y_fit[0][c-2]>y_fit[0][c-1]\n",
    "    cond_min_2 = y_fit[0][c-1]<y_fit[0][c]\n",
    "    cond_min_eps = cond_min_1 & cond_min_2\n",
    "    #condition 2: threshold for the a coefficient of the fit to avoid cases in which 𝜀(𝑖)->0 in straight lines\n",
    "    R_a_mask = abs(y_fit[1])>1.65e-2\n",
    "\n",
    "    # reshaping to fill in with zeros at the beginning and end (since during the first and last w there are no peaks)\n",
    "    zeros = np.zeros(w+1)!=0\n",
    "    zeros2 = np.zeros(w)!=0\n",
    "    cond_1 = np.concatenate((zeros, cond_min_eps, zeros))\n",
    "    cond_2 = np.concatenate((zeros2, R_a_mask, zeros2))\n",
    "    a_fixed=np.concatenate((zeros2, y_fit[1], zeros2))\n",
    "    v_fixed=np.concatenate((zeros2, y_fit[2], zeros2))\n",
    "\n",
    "    # using masks to find the desired results\n",
    "    R_i_mask=cond_1&cond_2\n",
    "    R_i=(np.where(R_i_mask))[0]\n",
    "    a=a_fixed[R_i]\n",
    "    v=v_fixed[R_i]\n",
    "    \n",
    "    return R_i, a, v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "4ce062ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "def align_to_longest(r, test, level):\n",
    "    '''\n",
    "    function that returns a one-to-one correspondance between:\n",
    "    # - the vector of reference rpeaks\n",
    "    # - the peaks to test\n",
    "    # the third input (level) is a treshold to be set based on the RR distance\n",
    "    '''\n",
    "    \n",
    "    # we make sure that everything is a numpy array\n",
    "    r = np.array(r)\n",
    "    test = np.array(test)\n",
    "    \n",
    "    # we make deep copies that we can safely modify during the loop without loss of data\n",
    "    r_peaks = copy.copy(r)\n",
    "    test_peaks = copy.copy(test) \n",
    "    # and we initialize the vectors in which we are going to save the results\n",
    "    # we initialize to weird numbers so that we can make sure that all values have been changed by the end of the algorithm\n",
    "    bool_record = np.ones(len(r), dtype=int)*(-55)\n",
    "    idx_record = np.ones(len(r))*(-55)\n",
    "\n",
    "    # for every peak in the longest peak\n",
    "    for i in range(len(r)):\n",
    "        # we make sure that there are still peaks to be compared with it\n",
    "        if (len(test_peaks)>0):\n",
    "            # we get the current element to examine, which every time is the first element of the array (since we later\n",
    "            # delete the elements aftering examining them)\n",
    "            r_index = r_peaks[0]\n",
    "            # calculate the distance from all the remaining test peaks\n",
    "            distance = np.abs(test_peaks - r_index)\n",
    "            # and sort such distances to find the minimum distance and the index (of rpeaks) corresponding to it\n",
    "            sort_distance = pd.DataFrame({'distance': distance}).sort_values(by=['distance'])\n",
    "            # we set the corresponding index to be the \n",
    "            min_distance = sort_distance.distance.values[0]\n",
    "            # we get the index corresponding to the minimum distance\n",
    "            min_distance_idx = sort_distance.index.values[0]\n",
    "            # If the minimum distance is appropriate\n",
    "            if min_distance<level:\n",
    "                # we save the positive finding\n",
    "                bool_record[i] = 1\n",
    "                # and the index corresponding to that finding\n",
    "                idx_record[i] = test_peaks[min_distance_idx]\n",
    "                # and remove the information corresponding to this configuration,\n",
    "                # to avoid repetitions and grant a one-to-one correspondance\n",
    "                test_peaks = np.delete(test_peaks, min_distance_idx)\n",
    "                r_peaks = np.delete(r_peaks, 0 )\n",
    "            # otherwise (if the closest value is too far away to be considered close)\n",
    "            else:\n",
    "                # we save the negative finding\n",
    "                bool_record[i] = 0\n",
    "                # and signal the lack of a corresponding index with a non-numeric value\n",
    "                idx_record[i] = np.nan\n",
    "                # we forget about the data \n",
    "                #test_peaks = np.delete(test_peaks, min_distance_idx)\n",
    "                r_peaks = np.delete(r_peaks, 0 )\n",
    "        # if all peaks have already been compared, we for sure have a negative result\n",
    "        else:\n",
    "            # we save the negative finding\n",
    "                bool_record[i] = 0\n",
    "                # and signal the lack of a corresponding index with a non-numeric value\n",
    "                idx_record[i] = np.nan\n",
    "    return bool_record, idx_record"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "f1dafb66",
   "metadata": {},
   "outputs": [],
   "source": [
    "def align(r, ann_idx, ann_symb, level):\n",
    "    '''\n",
    "    function that returns a one-to-one correspondance between:\n",
    "    - the vector of r-peaks r\n",
    "    - the annotations indexes ann_idx\n",
    "    while estabilishing this correspondance, the third input (ann_symbol) is altered identically to ann_idx\n",
    "    since the desired outcome of the code is to find the ann_symb corresponding to the r-peaks. \n",
    "    the last input, level, is a treshold which correspond to the maximum distance for compatibility\n",
    "    '''\n",
    "    if len(r) != len(np.unique(r)):\n",
    "        print('input has repetition!!! but we took care of it inside the function')\n",
    "        r = np.unique(r)\n",
    "                     \n",
    "    # we make deep copies that we can safely modify during the loop without loss of data\n",
    "    rpeaks = copy.copy(r)\n",
    "    ann_indexes = copy.copy(ann_idx) \n",
    "    ann_symbols = copy.copy(ann_symb)\n",
    "    # lastly, we initialize the vectors in which we are going to save the results\n",
    "    # - the r-peaks for which we find a corresponding annotation\n",
    "    real_r_peaks = list()\n",
    "    # - the annotation (symbols) for which we find a corresponding rpeak\n",
    "    real_ann_symbols = list()\n",
    "    ann_aligned = list()\n",
    "    # - a dummy initialization needed to enter the while-loop\n",
    "    sort_distance = np.zeros(100)\n",
    "    # - a register for the rpeaks (that will work, later on, as a boolean mask to be applied to the input r)\n",
    "    r_register = np.zeros(len(r),dtype=bool)\n",
    "    # - a list in which we save every peak once (if found in both rpeak and ann, the rpeak value is kept)\n",
    "    val = []\n",
    "    # - to lists where we keep record of where the values in val come from\n",
    "    ann_bool = []\n",
    "    rpeak_bool = []\n",
    "    \n",
    "    i=-1\n",
    "    while ((i<=len(ann_idx)+1) & (len(sort_distance)>1) & (len(ann_indexes)>=1)):\n",
    "        i +=1\n",
    "        # we get the current element to examin\n",
    "        ann_index = ann_indexes[0]\n",
    "        # calculate the distance of the many (remaining) rpeaks from it\n",
    "        distance = np.abs(rpeaks - ann_index)\n",
    "        # and sort such distances to find the minimum distance and the index (of rpeaks) corresponding to it\n",
    "        sort_distance = pd.DataFrame({'distance': distance}).sort_values(by=['distance'])\n",
    "        # we set the corresponding index to be the \n",
    "        min_distance = sort_distance.distance.values[0]\n",
    "        # If the minimum distance is appropriate\n",
    "        if min_distance<level:            \n",
    "            # we get the index corresponding to the minimum distance\n",
    "            min_distance_idx = sort_distance.index.values[0]\n",
    "            # we save the rpeak corresponding to minimum distance\n",
    "            real_r_peaks.append( rpeaks[ min_distance_idx ] )\n",
    "            # And the annotation index\n",
    "            ann_aligned.append(ann_indexes[0])\n",
    "            # and save the corresponding annotation\n",
    "            real_ann_symbols.append( ann_symbols[0] )\n",
    "            # we find where the peak was originally\n",
    "            asmr = np.where(r == rpeaks[ min_distance_idx ])\n",
    "            # to fill in the boolean mask \n",
    "            r_register[asmr] = True\n",
    "            # we save the results:\n",
    "            val.append(rpeaks[min_distance_idx])\n",
    "            ann_bool.append(1)\n",
    "            rpeak_bool.append(1)\n",
    "            # and remove the annotations (both index and symbol) and rpeak corresponding to this configuration,\n",
    "            # to avoid repetitions and grant a one-to-one correspondance\n",
    "            rpeaks = np.delete(rpeaks, min_distance_idx)\n",
    "            ann_indexes = np.delete(ann_indexes, 0 )\n",
    "            ann_symbols = np.delete(ann_symbols, 0 )\n",
    "            \n",
    "        # otherwise (if the closest value is too far away to be considered close)\n",
    "        else:\n",
    "            val.append(ann_indexes[0])\n",
    "            ann_bool.append(1)\n",
    "            rpeak_bool.append(0)\n",
    "            \n",
    "            # we forget about the data corresponding to this annotation since there's no correspondance in the rpeak\n",
    "            ann_indexes = np.delete(ann_indexes, 0 )\n",
    "            ann_symbols = np.delete(ann_symbols, 0 )\n",
    "    if len(rpeaks)>0:\n",
    "        val.extend(rpeaks) # I'm really not sure about this line\n",
    "        ann_bool.extend(np.zeros(len(rpeaks),dtype=int))\n",
    "        rpeak_bool.extend(np.ones(len(rpeaks),dtype=int))\n",
    "\n",
    "#     print(len(r_register),'is the length of the boolean mask r_register, just like the lenght of the input ',len(r))\n",
    "#     print(sum(r_register),' is number of Trues in len(r_register), just like the length of real_r_peaks', \n",
    "#           len(real_r_peaks))\n",
    "\n",
    "    return ann_aligned, real_r_peaks, real_ann_symbols, np.array(val), np.array(ann_bool), np.array(rpeak_bool), r_register"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b2709d51",
   "metadata": {},
   "outputs": [],
   "source": [
    "def bool_mask_removing_repetitions(a2):\n",
    "    '''\n",
    "    the function produces a boolean mask that helps keep track of the changes that happen from a to np.unique(a)\n",
    "    '''\n",
    "    # we get the unique values and how many times each of them is present\n",
    "    a, counts = np.unique(a2, return_counts=True)\n",
    "    # we intialize the boolean mask in output, which dimensions must match the ones of the input and is initally all positives\n",
    "    output = np.ones(len(a2), dtype=bool)\n",
    "    # for the elements that are repeated\n",
    "    tocheck = a[counts >1]\n",
    "    for tc in tocheck:\n",
    "        # we get the indexes of the actual repetitions (i.e. all indexes beside the first occurrence)\n",
    "        idx = np.where(a2 == tc)[0][1:]\n",
    "        # and make sure to negate the boolean mask in such indexes\n",
    "        output[idx] = np.zeros(len(idx),dtype=bool)\n",
    "    return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "dac1cb82",
   "metadata": {},
   "outputs": [],
   "source": [
    "def rpeaks_per_patient(x_d, data, lev1, n_leads, annotations=False, ann_idx=None, ann_symbol=None, lev2=None,):\n",
    "    '''\n",
    "    The following function aligns the r-peaks found in the 12 leads and decides which ones to keep based on how many times \n",
    "    (out of 12) they are found. The inputs are:\n",
    "    - x_d: the dictionary, in which is key is a lead name and corresponds to a list of r-peaks\n",
    "    - lev1: an input theshold (to be set based on the RR distance - see align_to_longest function)\n",
    "    - n_leads : minimum number of matching leads for a peak to be accepted\n",
    "    - annotations: an option tg=hat, if True, filters also to match annotations and will return two accuracy scores\n",
    "    - ann_idx: a numpy array of annotations indexes\n",
    "    - ann_symbol: a list of annotations symbols (its length must match the one of ann_idx\n",
    "    - l\n",
    "    '''\n",
    "    # we extract the names fl\n",
    "    names = list(x_d.keys())\n",
    "    # we find the longest one\n",
    "    l = [len(x_d[name]) for name in names]\n",
    "    # we find the corresponding index\n",
    "    best_idx = np.argmax(l)\n",
    "    # and we call the corresponding lead \"best_lead\"\n",
    "    best_lead = names[best_idx]\n",
    "    # we now create a dataframe in which we are going to save everything\n",
    "    # it is going to have as many rows has the lead with most peaks\n",
    "    peaks = pd.DataFrame({ str(best_lead) : x_d[best_lead] })\n",
    "    # we now want to fill the dataframe with the remaining information\n",
    "    # for each one, we try to align the results to the best_lead\n",
    "    reference = x_d[best_lead]\n",
    "    \n",
    "    # for every lead (including the best one, so we make sure to save all the indexes in the same place)\n",
    "    for ol in names:\n",
    "        # we extract the test peaks\n",
    "        test = x_d[ol]\n",
    "        # to feed them to a function that returns the columns of the dataframe to be filled\n",
    "        # the first column contains booleans that signal if a correspondance for the peak was found in the dataset\n",
    "        # the second column contains the indexes that are potentially found in the correspondance (otherwise, they are NaN)\n",
    "        peaks[ol], peaks[ol+'_idx'] =  align_to_longest(reference, test, lev1)\n",
    "    # we calculate the number of times in which each peak is found\n",
    "    peaks['n'] = peaks[names].sum(axis=1)\n",
    "    # and decide if the peaks must be kept based on an input theshold\n",
    "    peaks['to_keep'] = peaks['n']>=n_leads   \n",
    "    # We delete the peaks that we don't keep\n",
    "    true_peaks = peaks[peaks['to_keep']==True]\n",
    "    true_peaks = true_peaks.reset_index(drop = True)\n",
    "    # we find the average location of each peak, selecting only the only ones to keep\n",
    "    name_idx = [name+'_idx' for name in names]\n",
    "    true_peaks['loc'] = (true_peaks[name_idx].mean(axis=1)).astype(int)\n",
    "    rpeaks = true_peaks['loc'].to_numpy()\n",
    "    where_repetitions = bool_mask_removing_repetitions(rpeaks)\n",
    "    rpeaks = np.unique(rpeaks)\n",
    "    \n",
    "    # if we can compare with annotations\n",
    "    if annotations:\n",
    "        # we compare with annotations\n",
    "        ann_aligned, r_aligned, ann_symbols_aligned, val, ann_bool, rpeak_bool, r_register = align(rpeaks, ann_idx, ann_symbol, lev2)\n",
    "\n",
    "        # we count the:\n",
    "        # - true positives\n",
    "        true_positive = np.sum( ( (np.array(ann_bool)==1) & (np.array(rpeak_bool)==1) ))\n",
    "        # - false positives\n",
    "        false_positive = np.sum( ( (np.array(ann_bool)==0) & (np.array(rpeak_bool)==1) ))\n",
    "        # - false negatives\n",
    "        false_negative = np.sum( ( (np.array(ann_bool)==1) & (np.array(rpeak_bool)==0) ))\n",
    "        # - true negatives\n",
    "        true_negative = len(data)-len(val)\n",
    "        # and concatenate the results in a (non-normalized) matrix\n",
    "        conmat_patient = np.array([[true_positive, false_negative],[false_positive,true_negative]]) \n",
    "        return r_aligned, peaks['to_keep'].to_numpy(), best_lead, where_repetitions, r_register, ann_symbols_aligned, conmat_patient\n",
    "    else:\n",
    "        return rpeaks, peaks['to_keep'].to_numpy(), best_lead, where_repetitions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "2262a482",
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_record_corrected(path, coef_parab):\n",
    "    '''\n",
    "    function that processes a single file by applying the previously defined functions\n",
    "    '''\n",
    "    # Load the record\n",
    "    path = str(path)\n",
    "    # we create a dictionary in which we going to save the r-peaks obtained in every lead\n",
    "    xd = {}\n",
    "    # and we do the same for saving the coefficients of the parabola fits\n",
    "    ecia = {}\n",
    "    lia = {}\n",
    "    # we initialize the estimates of the RR length\n",
    "    stat_RR = np.zeros(len(leads))\n",
    "    stat_RR2 = np.zeros(len(leads))\n",
    "    \n",
    "    # for every lead\n",
    "    for i,lead in enumerate(leads):\n",
    "        # we read the corresponiding data\n",
    "        data = pd.read_csv(\"output_files/data\"+path+\"_2filtered.csv\")[lead]\n",
    "        # we find its R-peaks \n",
    "        rpeak, temp1, temp2 = R_fit(data)\n",
    "        # we save the corresponding coefficients for the parabola fits\n",
    "        ecia[lead] = temp1\n",
    "        lia[lead] = temp2\n",
    "        # we compute and estimator of th RR lenght for the peaks we just found\n",
    "        stat_RR[i] = np.median(rpeak[1:]-rpeak[0:-1], )\n",
    "        stat_RR2[i] = np.mean(rpeak[1:]-rpeak[0:-1])\n",
    "        # and save the peaks we just found in a dictionary\n",
    "        xd[lead] = rpeak\n",
    "        \n",
    "    # we calculate an estimator of all the RR lengths and use to deduce the maximum distance for peak compatibility\n",
    "    # (which should generally be smaller than the half of the average RR interval)\n",
    "    # in this precise case, we choose the median as the estimator since the distribution is quite asymmetric \n",
    "    # (so the difference between mean and median is huge)\n",
    "    # we need to get rid of the NaN that are returned by np.median if no peak is found in a lead\n",
    "    lev1 = np.nanmedian(stat_RR)/2\n",
    "    print('maximum distance for compatibility of same peak in different leads:',lev1)\n",
    "    \n",
    "\n",
    "\n",
    "    rpeaks, pp, best_lead, where_repetitions = rpeaks_per_patient(xd, data, lev1, 8)\n",
    "    # we select the data relative to the best lead (now that we know which is)\n",
    "    ii = xd[best_lead]\n",
    "    a = ecia[best_lead]\n",
    "    v = lia[best_lead]\n",
    "\n",
    "    # we add the R peaks to the dataframe while also adapting them to the frequency reduction that we're about to implement\n",
    "    rpeaks = np.array(rpeaks)\n",
    "    final = (rpeaks/4).astype(int)\n",
    "    df = pd.DataFrame({'R': final}) \n",
    "    outfn = \"output_files/\"+\"data\"+path+'_12Rpeak.csv'\n",
    "    df.to_csv(outfn, index=False)\n",
    "    \n",
    "    # we save together the parameters for the parabola\n",
    "    temp = pd.DataFrame({'i': ii,'v':v, 'a':a, 'path': [path]*len(a) })\n",
    "    # crop them to have a one-to-one correspondence with the R peaks\n",
    "    temp = (((temp[pp]).reset_index(drop=True))[where_repetitions].reset_index(drop=True))\n",
    "    \n",
    "    # and adjust the index column to match the future frequency reduction\n",
    "    temp['i'] = (temp['i']/4).astype(int)\n",
    "    coef_parab = (pd.concat([coef_parab, temp])).reset_index(drop=True)\n",
    "    \n",
    "    print('process_record_corrected: match in len of output files: ',len(temp)==len(final))\n",
    "    if len(temp) !=len(final):\n",
    "        print(\"current coef_parab and rpeaks lengths don't match\")\n",
    "        print('input lenghts :',len(ii),len(a),len(v) )\n",
    "        print('pp starts from',len(pp),'and end to ',sum(pp))\n",
    "        print('where_repetitions starts from',len(where_repetitions),'and end to ',sum(where_repetitions))\n",
    "        print('r_register starts from',len(r_register),'and end to ',sum(r_register))\n",
    "        print('output length :', len(coef_parab) )\n",
    "    \n",
    "    return final, coef_parab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "a8744e3d",
   "metadata": {},
   "outputs": [],
   "source": [
    "FILE_SAVE=True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "c3beeaf1",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "maximum distance for compatibility of same peak in different leads: 311.25\n",
      "process_record_corrected: match in len of output files:  True\n",
      "1\n",
      "maximum distance for compatibility of same peak in different leads: 371.0\n",
      "process_record_corrected: match in len of output files:  True\n",
      "2\n",
      "maximum distance for compatibility of same peak in different leads: 404.0\n",
      "process_record_corrected: match in len of output files:  True\n",
      "3\n",
      "maximum distance for compatibility of same peak in different leads: 333.75\n",
      "process_record_corrected: match in len of output files:  True\n",
      "4\n",
      "maximum distance for compatibility of same peak in different leads: 320.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "5\n",
      "maximum distance for compatibility of same peak in different leads: 341.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "6\n",
      "maximum distance for compatibility of same peak in different leads: 265.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "7\n",
      "maximum distance for compatibility of same peak in different leads: 309.0\n",
      "process_record_corrected: match in len of output files:  True\n",
      "8\n",
      "maximum distance for compatibility of same peak in different leads: 351.75\n",
      "process_record_corrected: match in len of output files:  True\n",
      "9\n",
      "maximum distance for compatibility of same peak in different leads: 349.25\n",
      "process_record_corrected: match in len of output files:  True\n",
      "10\n",
      "maximum distance for compatibility of same peak in different leads: 411.625\n",
      "process_record_corrected: match in len of output files:  True\n",
      "11\n",
      "maximum distance for compatibility of same peak in different leads: 411.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "12\n",
      "maximum distance for compatibility of same peak in different leads: 374.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "13\n",
      "maximum distance for compatibility of same peak in different leads: 374.25\n",
      "process_record_corrected: match in len of output files:  True\n",
      "14\n",
      "maximum distance for compatibility of same peak in different leads: 351.25\n",
      "process_record_corrected: match in len of output files:  True\n",
      "15\n",
      "maximum distance for compatibility of same peak in different leads: 592.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "16\n",
      "maximum distance for compatibility of same peak in different leads: 590.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "17\n",
      "maximum distance for compatibility of same peak in different leads: 621.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "18\n",
      "maximum distance for compatibility of same peak in different leads: 649.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "19\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\nadil\\anaconda3\\envs\\torch\\lib\\site-packages\\numpy\\core\\fromnumeric.py:3432: RuntimeWarning: Mean of empty slice.\n",
      "  return _methods._mean(a, axis=axis, dtype=dtype,\n",
      "C:\\Users\\nadil\\anaconda3\\envs\\torch\\lib\\site-packages\\numpy\\core\\_methods.py:190: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  ret = ret.dtype.type(ret / rcount)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "maximum distance for compatibility of same peak in different leads: 644.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "20\n",
      "maximum distance for compatibility of same peak in different leads: 613.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "21\n",
      "maximum distance for compatibility of same peak in different leads: 606.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "22\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\nadil\\anaconda3\\envs\\torch\\lib\\site-packages\\numpy\\core\\fromnumeric.py:3432: RuntimeWarning: Mean of empty slice.\n",
      "  return _methods._mean(a, axis=axis, dtype=dtype,\n",
      "C:\\Users\\nadil\\anaconda3\\envs\\torch\\lib\\site-packages\\numpy\\core\\_methods.py:190: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  ret = ret.dtype.type(ret / rcount)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "maximum distance for compatibility of same peak in different leads: 599.5\n",
      "process_record_corrected: match in len of output files:  True\n",
      "23\n",
      "maximum distance for compatibility of same peak in different leads: 586.0\n",
      "process_record_corrected: match in len of output files:  True\n",
      "24\n",
      "Delta>0 for all peaks : False\n",
      "The R peak data saving took 95.4 minutes\n"
     ]
    }
   ],
   "source": [
    "if FILE_SAVE==True: \n",
    "    # we initialize the names of files to be processed\n",
    "    r  = np.genfromtxt(\"input_files/RECORDS\", dtype='str')\n",
    "    records = list(range(1,(len(r)+1)))\n",
    "    # and a database in which we're going to save the parameters of the \n",
    "    coef_parab = pd.DataFrame({'i': [], 'a':[], 'v':[], 'path': []})\n",
    "\n",
    "    if __name__ == \"__main__\":\n",
    "        t1 = time.time()\n",
    "        # for every file\n",
    "        for path in records:\n",
    "            # we process the data\n",
    "            rpeaks, coef_parab = process_record_corrected(path, coef_parab)\n",
    "            # we reset the dataframe indexes (to avoid future bugs)\n",
    "            coef_parab=coef_parab.reset_index(drop = True)\n",
    "            \n",
    "            print(path)\n",
    "        \n",
    "        # we reset the dataframe indexes (to avoid future bugs)\n",
    "        coef_parab=coef_parab.reset_index(drop = True)\n",
    "        # we make sure that the peaks found are correct (the corresponding parabola fits intersect with the x axis)\n",
    "        coef_parab['Delta'] = -4*coef_parab['a']*coef_parab['v']\n",
    "        print('Delta>0 for all peaks :',np.all(coef_parab['Delta']>0))\n",
    "#         print(\"rpeaks - coef_parab['i'] :\", np.mean(rpeaks - coef_parab['i']), '±', np.std(rpeaks - coef_parab['i']) )\n",
    "        t2 = time.time()\n",
    "        print('The R peak data saving took', round((t2-t1)/60,2), 'minutes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "abcf8790",
   "metadata": {},
   "outputs": [],
   "source": [
    "# outfn = \"output_files/coef_parab.csv\"\n",
    "# coef_parab.to_csv(outfn, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d2ddd183",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>path</th>\n",
       "      <th>Delta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>556.0</td>\n",
       "      <td>0.024642</td>\n",
       "      <td>-5.623284</td>\n",
       "      <td>1</td>\n",
       "      <td>0.554277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>695.0</td>\n",
       "      <td>0.021833</td>\n",
       "      <td>-5.121537</td>\n",
       "      <td>1</td>\n",
       "      <td>0.447271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>830.0</td>\n",
       "      <td>0.021400</td>\n",
       "      <td>-4.966572</td>\n",
       "      <td>1</td>\n",
       "      <td>0.425133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>964.0</td>\n",
       "      <td>0.022563</td>\n",
       "      <td>-5.065584</td>\n",
       "      <td>1</td>\n",
       "      <td>0.457178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1096.0</td>\n",
       "      <td>0.021240</td>\n",
       "      <td>-4.996527</td>\n",
       "      <td>1</td>\n",
       "      <td>0.424505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70897</th>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>24</td>\n",
       "      <td>0.421116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70898</th>\n",
       "      <td>847668.0</td>\n",
       "      <td>-0.029394</td>\n",
       "      <td>6.413937</td>\n",
       "      <td>24</td>\n",
       "      <td>0.754124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70899</th>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>24</td>\n",
       "      <td>0.772745</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70900</th>\n",
       "      <td>847925.0</td>\n",
       "      <td>-0.026892</td>\n",
       "      <td>5.604619</td>\n",
       "      <td>24</td>\n",
       "      <td>0.602885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70901</th>\n",
       "      <td>848053.0</td>\n",
       "      <td>-0.021183</td>\n",
       "      <td>2.349211</td>\n",
       "      <td>24</td>\n",
       "      <td>0.199056</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>70902 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              i         a         v  path     Delta\n",
       "0         556.0  0.024642 -5.623284     1  0.554277\n",
       "1         695.0  0.021833 -5.121537     1  0.447271\n",
       "2         830.0  0.021400 -4.966572     1  0.425133\n",
       "3         964.0  0.022563 -5.065584     1  0.457178\n",
       "4        1096.0  0.021240 -4.996527     1  0.424505\n",
       "...         ...       ...       ...   ...       ...\n",
       "70897  847539.0 -0.022921  4.593035    24  0.421116\n",
       "70898  847668.0 -0.029394  6.413937    24  0.754124\n",
       "70899  847797.0 -0.033397  5.784591    24  0.772745\n",
       "70900  847925.0 -0.026892  5.604619    24  0.602885\n",
       "70901  848053.0 -0.021183  2.349211    24  0.199056\n",
       "\n",
       "[70902 rows x 5 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_parab = pd.read_csv(\"output_files/coef_parab.csv\")\n",
    "coef_parab"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d2b838a",
   "metadata": {},
   "source": [
    "Now that we've analyzed in details the positions of the peaks, we can artificially reduce the frequency of the dataset to improve computation time. We take care only of the datasets since the R peaks were already adapted during saving."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "bd0d0380",
   "metadata": {},
   "outputs": [],
   "source": [
    "Resaving filtered file with new frequency, 250Hz\n",
    "if FILE_SAVE==True: \n",
    "    mask = list(np.arange(0, record_1.p_signal.shape[0], 4))\n",
    "    r  = np.genfromtxt(\"input_files/RECORDS\", dtype='str')\n",
    "    records = [1] #list(range(2,(len(r)+1)))\n",
    "    for path in records:\n",
    "        path = str(path)\n",
    "        datac = pd.DataFrame()\n",
    "        data = pd.read_csv(\"output_files/\"+\"data\"+path+'_2filtered.csv')\n",
    "        for i in range(0,len(leads)):\n",
    "            mask = np.arange(0,len(data),4)\n",
    "            datac[leads[i]] = data.iloc[mask, i]\n",
    "            outfn = \"output_files/\"+\"data\"+path+'_filtered.csv'\n",
    "            datac.to_csv(outfn, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4a32b632",
   "metadata": {},
   "outputs": [],
   "source": [
    "new_freq = int(freq/4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d69329e-624f-4d08-ada2-57735db623cf",
   "metadata": {
    "tags": []
   },
   "source": [
    "# 3) Segmentation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63e8ee8e-1d61-4a30-ad45-735ae34149b9",
   "metadata": {
    "tags": []
   },
   "source": [
    "The following segmentation algorithm is based on the known ratios that can be computed between:\n",
    "- the average length of the RR interval (~ 800 ms from 600-1000 ms) \n",
    "- the average length of the PR interval (~ 160 ms from 120-200 ms)\n",
    "- the average length of the QT interval (~ 410 ms from 360-460 ms)\n",
    "\n",
    "Such ratios are used to find the actual length of thePR and QT interval ($L_\\text{PR}^i$, $L_\\text{QT}^i$) based on the actual length of RR interval $L_\\text{RR}^i$ (which can be easily computed). \n",
    "\n",
    "$$ L_\\text{PR}^i = L_\\text{RR}^i \\cdot \\frac{410 \\; \\text{ms}}{800 \\; \\text{ms}} $$\n",
    "\n",
    "$$ L_\\text{QT}^i = L_\\text{RR}^i \\cdot \\frac{160 \\; \\text{ms}}{800 \\; \\text{ms}} $$\n",
    "\n",
    "Once $L_\\text{PR}^i$ and $L_\\text{QT}^i$ are computed, the hearbeat is identified as:\n",
    "\n",
    "$$ \\left[ x_\\text{R-peak}^i - L_\\text{PR}^i  , x_\\text{R-peak}^i + L_\\text{QT}^i \\right]  $$\n",
    "\n",
    "The length of such interval is $\\frac{(410+160) \\; \\text{ms}}{800 \\; \\text{ms}} = \\frac{57}{80} \\sim 71.25 \\% $."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "22f37275",
   "metadata": {},
   "outputs": [],
   "source": [
    "def segmentation_R(annotations, cp):\n",
    "    '''\n",
    "    function that performs a relative window segmentation based on the relative known proportions between RR, PR and QT \n",
    "    '''\n",
    "    ann = annotations\n",
    "    \n",
    "    # we initialize the variables we need for the while loop\n",
    "    patient_start = np.ones(2)*-3\n",
    "    j = -1\n",
    "    while np.any(patient_start<0):\n",
    "        j += 1\n",
    "        # we calculate the RR intervals\n",
    "        RR_intervals = (ann[(j+1):]-ann[j:-1])\n",
    "        # firstly, we crop the RR interval to match the relative proportion of RR:QT=0.8:0.41 \n",
    "        QT_size = (RR_intervals*0.41/0.8).astype(int)\n",
    "        PR_size = (RR_intervals*0.16/0.8).astype(int)\n",
    "        \n",
    "        patient_start = ann[j:-1] - PR_size\n",
    "        patient_end = ann[j:-1] + QT_size\n",
    "        \n",
    "    # j is the number of annotations ignored from the beginning\n",
    "    # i is the number of annotations ignored from the end\n",
    "    # once we've made sure that the annotations are aligned, we delete the first and last one just to be safe\n",
    "    RR_intervals = RR_intervals[1:-1]\n",
    "    QT_size = QT_size[1:-1]\n",
    "    PR_size = PR_size[1:-1]    \n",
    "    \n",
    "    patient_start = ann[j+1:-2] - PR_size\n",
    "    patient_end = ann[j+1:-2] + QT_size\n",
    "\n",
    "    cp = cp[j+1:-2]\n",
    "\n",
    "\n",
    "    patient_R_peak = ann[j+1:-2]\n",
    "    #patient_ann = list(annotations['symbol'][j+1:-2])\n",
    "        \n",
    "    return RR_intervals, QT_size, PR_size, patient_start, patient_end, patient_R_peak, cp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "9344509e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the segmentation algorithm took 0.33 seconds\n",
      "overflow of i from end 0.0 %\n",
      "overflow of i from start 0.07766 %\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>667</td>\n",
       "      <td>763</td>\n",
       "      <td>694</td>\n",
       "      <td>136</td>\n",
       "      <td>695.0</td>\n",
       "      <td>0.021833</td>\n",
       "      <td>-5.121537</td>\n",
       "      <td>0.447271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>804</td>\n",
       "      <td>898</td>\n",
       "      <td>830</td>\n",
       "      <td>134</td>\n",
       "      <td>830.0</td>\n",
       "      <td>0.021400</td>\n",
       "      <td>-4.966572</td>\n",
       "      <td>0.425133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>938</td>\n",
       "      <td>1031</td>\n",
       "      <td>964</td>\n",
       "      <td>132</td>\n",
       "      <td>964.0</td>\n",
       "      <td>0.022563</td>\n",
       "      <td>-5.065584</td>\n",
       "      <td>0.457178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1070</td>\n",
       "      <td>1163</td>\n",
       "      <td>1096</td>\n",
       "      <td>132</td>\n",
       "      <td>1096.0</td>\n",
       "      <td>0.021240</td>\n",
       "      <td>-4.996527</td>\n",
       "      <td>0.424505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1148</td>\n",
       "      <td>1434</td>\n",
       "      <td>1228</td>\n",
       "      <td>402</td>\n",
       "      <td>1228.0</td>\n",
       "      <td>0.021803</td>\n",
       "      <td>-4.962970</td>\n",
       "      <td>0.432822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70821</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70822</th>\n",
       "      <td>24</td>\n",
       "      <td>1819</td>\n",
       "      <td>847417</td>\n",
       "      <td>847523</td>\n",
       "      <td>847447</td>\n",
       "      <td>150</td>\n",
       "      <td>847407.0</td>\n",
       "      <td>-0.017340</td>\n",
       "      <td>4.541612</td>\n",
       "      <td>0.315007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70823</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70824</th>\n",
       "      <td>24</td>\n",
       "      <td>1821</td>\n",
       "      <td>847691</td>\n",
       "      <td>847778</td>\n",
       "      <td>847715</td>\n",
       "      <td>124</td>\n",
       "      <td>847668.0</td>\n",
       "      <td>-0.029394</td>\n",
       "      <td>6.413937</td>\n",
       "      <td>0.754124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70825</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>70826 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     0     667     763        694        136     695.0   \n",
       "1            1     1     804     898        830        134     830.0   \n",
       "2            1     2     938    1031        964        132     964.0   \n",
       "3            1     3    1070    1163       1096        132    1096.0   \n",
       "4            1     4    1148    1434       1228        402    1228.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "70821       24  1818  847295  847385     847320        127  847277.0   \n",
       "70822       24  1819  847417  847523     847447        150  847407.0   \n",
       "70823       24  1820  847574  847657     847597        118  847539.0   \n",
       "70824       24  1821  847691  847778     847715        124  847668.0   \n",
       "70825       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  \n",
       "0      0.021833 -5.121537  0.447271  \n",
       "1      0.021400 -4.966572  0.425133  \n",
       "2      0.022563 -5.065584  0.457178  \n",
       "3      0.021240 -4.996527  0.424505  \n",
       "4      0.021803 -4.962970  0.432822  \n",
       "...         ...       ...       ...  \n",
       "70821 -0.021987  3.692612  0.324755  \n",
       "70822 -0.017340  4.541612  0.315007  \n",
       "70823 -0.022921  4.593035  0.421116  \n",
       "70824 -0.029394  6.413937  0.754124  \n",
       "70825 -0.033397  5.784591  0.772745  \n",
       "\n",
       "[70826 rows x 10 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "FILE_SAVE=True\n",
    "r  = np.genfromtxt(\"input_files/RECORDS\", dtype='str')\n",
    "records = list(range(1,(len(r)+1)))\n",
    "\n",
    "if FILE_SAVE:\n",
    "    t1 = time.time()\n",
    "    \n",
    "    # we want to save a dataframe of heartbeats: to do that, we initialize some temporary means of storage\n",
    "    # - the number that identifies the patient\n",
    "    number = np.array([0])\n",
    "    # - the number that identifies the heartbeat of each patientx\n",
    "    idx = np.array([0])\n",
    "    # - the index of the start of the heartbeat\n",
    "    start = np.array([0])\n",
    "    # - the index of the end of the heartbeat\n",
    "    end = np.array([0]) \n",
    "    # - the index of the R peak of the heartbeat\n",
    "    R_peak = np.array([0]) \n",
    "    # - the symbol for the annotation of the heartbeat\n",
    "    annot = [None] \n",
    "    # - the duration of the heartbeat\n",
    "    RR = np.array([0])\n",
    "    # - the parabola fit coefficients\n",
    "    ddff = pd.DataFrame({ 'i': [], 'a': [], 'v': [], 'path': [] })\n",
    "    \n",
    "    # we then fill such temporary means of storage by updating them after analyzing every file/patient\n",
    "    for path in records:\n",
    "        path = str(path)\n",
    "        # UPLOAD: since we don't use the signal at this step, it doesn't matter if it's filtered or not\n",
    "        # data = pd.read_csv(\"output_files/data\"+path+\".csv\")\n",
    "#         annotations = pd.read_csv(\"output_files/data\"+path+\"_Rpeak.csv\")\n",
    "        annotations = pd.read_csv(\"output_files/data\"+path+'_12Rpeak.csv')\n",
    "        ann = annotations['R'].values\n",
    "        \n",
    "        cp = coef_parab[ coef_parab['path']==int(path)]\n",
    "#         cp = cp.reset_index(drop = True)\n",
    "        \n",
    "        # and use my custom function for SEGMENTATION\n",
    "        patient_RR, QT_size, PR_size, patient_start, patient_end, patient_R_peak, cp = segmentation_R(ann, cp)\n",
    "        cp = cp.reset_index(drop = True)\n",
    "\n",
    "        #print(path,patient_start[0])\n",
    "        # we add the remaining pieces of information\n",
    "        # - the number of the patient is the same for all its heartbeats\n",
    "        patient_number = np.ones_like(patient_start)*int(path)\n",
    "        # - the number that identifies the heartbeat of each patient is a simple increasing vector of integers\n",
    "        patient_idx = np.arange(0,len(patient_start),1,dtype = int)\n",
    "#         # - the location of the R peaks (by cropping the annotations accordingly)\n",
    "#         patient_R_peak = ann[j:-i]\n",
    "#         # - the symbols for the annotations (by cropping the annotations accordingly)\n",
    "#         patient_ann = list(annotations['symbol'][j:-i])\n",
    "        \n",
    "        # lastly, we check that all the annotations are in between the corresponding starts and stops\n",
    "        check = np.all([(patient_R_peak>patient_start),(patient_R_peak<patient_end)])\n",
    "#         if not check:\n",
    "#             print(check)\n",
    "           \n",
    "        # and update the temporary means of storage\n",
    "        number = np.concatenate((number, patient_number))\n",
    "        idx = np.concatenate((idx, patient_idx))\n",
    "        start = np.concatenate((start, patient_start))\n",
    "        end = np.concatenate((end, patient_end))\n",
    "        R_peak = np.concatenate((R_peak, patient_R_peak))\n",
    "        #annot.extend(patient_ann)\n",
    "        RR = np.concatenate((RR,patient_RR))\n",
    "        ddff = pd.concat([ddff,cp])\n",
    "        ddff = ddff.reset_index(drop = True)\n",
    "    \n",
    "    # finally, we put together the heartbeat dataframe  \n",
    "    hb = pd.DataFrame({\"patient\": number[1:].astype(int), \"idx\": idx[1:], \"start\": start[1:].astype(int), \n",
    "                       \"end\": end[1:].astype(int), \"ann_index\": R_peak[1:].astype(int), \"RR_length\": RR[1:].astype(int) })\n",
    "    hb['i'] = ddff['i']\n",
    "    hb['a'] = ddff['a']\n",
    "    hb['v'] = ddff['v']\n",
    "    hb['Delta'] = -4*hb['a']*hb['v']\n",
    "    t2 = time.time()\n",
    "    print(\"the segmentation algorithm took\",round(t2-t1,2),'seconds')\n",
    "    outfn = 'output_files/heartbeats.csv'\n",
    "    hb.to_csv(outfn, index=False)\n",
    "    print('overflow of i from end',round(sum(hb['end']<hb['i'])*100/len(hb),5), '%')\n",
    "    print('overflow of i from start', round(sum(hb['start']>hb['i'])*100/len(hb),5), '%')\n",
    "else:\n",
    "    hb = pd.read_csv('output_files/heartbeats.csv')\n",
    "    \n",
    "display(hb)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0872b7d",
   "metadata": {},
   "source": [
    "To ensure that the modifications made on the dataframe are applied also to the parabola fit coefficient, we added the to the parabola fit coefficients to the hb dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af58f295",
   "metadata": {},
   "source": [
    "We discard the peaks that have been fit with parabolas that have the wrong concavities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "b7917493",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the percentage of discarded peaks (for mis-fitting) is: 0.00141 %\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>667</td>\n",
       "      <td>763</td>\n",
       "      <td>694</td>\n",
       "      <td>136</td>\n",
       "      <td>695.0</td>\n",
       "      <td>0.021833</td>\n",
       "      <td>-5.121537</td>\n",
       "      <td>0.447271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>804</td>\n",
       "      <td>898</td>\n",
       "      <td>830</td>\n",
       "      <td>134</td>\n",
       "      <td>830.0</td>\n",
       "      <td>0.021400</td>\n",
       "      <td>-4.966572</td>\n",
       "      <td>0.425133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>938</td>\n",
       "      <td>1031</td>\n",
       "      <td>964</td>\n",
       "      <td>132</td>\n",
       "      <td>964.0</td>\n",
       "      <td>0.022563</td>\n",
       "      <td>-5.065584</td>\n",
       "      <td>0.457178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1070</td>\n",
       "      <td>1163</td>\n",
       "      <td>1096</td>\n",
       "      <td>132</td>\n",
       "      <td>1096.0</td>\n",
       "      <td>0.021240</td>\n",
       "      <td>-4.996527</td>\n",
       "      <td>0.424505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1148</td>\n",
       "      <td>1434</td>\n",
       "      <td>1228</td>\n",
       "      <td>402</td>\n",
       "      <td>1228.0</td>\n",
       "      <td>0.021803</td>\n",
       "      <td>-4.962970</td>\n",
       "      <td>0.432822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70820</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70821</th>\n",
       "      <td>24</td>\n",
       "      <td>1819</td>\n",
       "      <td>847417</td>\n",
       "      <td>847523</td>\n",
       "      <td>847447</td>\n",
       "      <td>150</td>\n",
       "      <td>847407.0</td>\n",
       "      <td>-0.017340</td>\n",
       "      <td>4.541612</td>\n",
       "      <td>0.315007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70822</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70823</th>\n",
       "      <td>24</td>\n",
       "      <td>1821</td>\n",
       "      <td>847691</td>\n",
       "      <td>847778</td>\n",
       "      <td>847715</td>\n",
       "      <td>124</td>\n",
       "      <td>847668.0</td>\n",
       "      <td>-0.029394</td>\n",
       "      <td>6.413937</td>\n",
       "      <td>0.754124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70824</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>70825 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     0     667     763        694        136     695.0   \n",
       "1            1     1     804     898        830        134     830.0   \n",
       "2            1     2     938    1031        964        132     964.0   \n",
       "3            1     3    1070    1163       1096        132    1096.0   \n",
       "4            1     4    1148    1434       1228        402    1228.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "70820       24  1818  847295  847385     847320        127  847277.0   \n",
       "70821       24  1819  847417  847523     847447        150  847407.0   \n",
       "70822       24  1820  847574  847657     847597        118  847539.0   \n",
       "70823       24  1821  847691  847778     847715        124  847668.0   \n",
       "70824       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  \n",
       "0      0.021833 -5.121537  0.447271  \n",
       "1      0.021400 -4.966572  0.425133  \n",
       "2      0.022563 -5.065584  0.457178  \n",
       "3      0.021240 -4.996527  0.424505  \n",
       "4      0.021803 -4.962970  0.432822  \n",
       "...         ...       ...       ...  \n",
       "70820 -0.021987  3.692612  0.324755  \n",
       "70821 -0.017340  4.541612  0.315007  \n",
       "70822 -0.022921  4.593035  0.421116  \n",
       "70823 -0.029394  6.413937  0.754124  \n",
       "70824 -0.033397  5.784591  0.772745  \n",
       "\n",
       "[70825 rows x 10 columns]"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hb = pd.read_csv('output_files/heartbeats.csv')\n",
    "temptemp = len(hb)\n",
    "hb  = (hb[ (hb['Delta'])>0 ] ).reset_index(drop = True)\n",
    "hb.to_csv('output_files/heartbeats_right.csv')\n",
    "print('the percentage of discarded peaks (for mis-fitting) is:', np.round((1-(len(hb)/temptemp))*100,5),'%') \n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94e207bd",
   "metadata": {},
   "source": [
    "To ensure that the modifications made on the dataframe are applied also to the parabola fit coefficient, we added the to the parabola fit coefficients to the hb dataframe."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea608908",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 4) Extraction of features\n",
    "In the previous step we extracted the RR length which is strictly related to the heartbeat frequency.\n",
    "\n",
    "We also noticed visually that parts of the data is extremely affected by noise, which appears either as extremely high peaks or local plateaus. These anomalous features can hinder the extraction of features, so we need to remove them beforehand.\n",
    "\n",
    "After removing the anomalous heartbeats, we will focus on proper feature extraction which will be executed in 3 rounds (DWT coefficient extraction + 2 rounds of visualizable features extraction)."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e64a2f6-e2ca-446b-b1c7-b51cc4c485d6",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## 4A) Discarding anomalous heartbeats"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d7d9a97",
   "metadata": {},
   "source": [
    "We now want to check that the distribution of the RR_length of the heartbeats found, *which was already averaged over the number of leads in which the peaks was identified*. Ideally, if the patient is healthy and the algorithm is identifying the R-peaks correctly, the data should be normally distributed around the average RR_lenght, with a relatively small width."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "id": "6a9c438e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# we also define a function that allows to quickly display the distribution of its input\n",
    "def distribution2(arg, a1, a2, title='arg', val = 0):\n",
    "    '''\n",
    "    function that returns the distribution of arg (which name/meaning is encoded in title)\n",
    "    '''\n",
    "    average = np.median(arg)\n",
    "    std = np.std(arg)\n",
    "    fig, ax = plt.subplots(1,1, figsize=(12,4))\n",
    "    info = ax.hist(arg, bins=np.arange(np.min(arg), np.max(arg)+1, 1), align='left', density=True, edgecolor='none', \n",
    "                   color='black', alpha=0.8, label = 'distribution of '+title)\n",
    "    most_common_length = val #(info[1])[np.argmax(info[0])]\n",
    "    diff_to_average = average - most_common_length\n",
    "    rr_down_lim = int(most_common_length - diff_to_average)\n",
    "    rr_up_lim = int(most_common_length + diff_to_average)\n",
    "\n",
    "    ax.vlines(most_common_length, 0,1, label = '')\n",
    "    ax.vlines(average, 0, 1, color='magenta', label='average')\n",
    "    ax.vlines(average+std, 0, 1, color='gold', label='1 $\\sigma$')\n",
    "    ax.vlines(average-std, 0, 1, color='gold')\n",
    "    ax.vlines(rr_down_lim, 0, 1, color='green', label = 'limits', linestyle='dashed')\n",
    "    ax.vlines(rr_up_lim, 0, 1, color='green', linestyle='dashed')\n",
    "    ax.set_xlabel(title)\n",
    "    ax.set_ylabel('density')\n",
    "    ax.set_xlim(a1, a2)\n",
    "    ax.set_ylim(0, 1.1*np.max(info[0]))\n",
    "    ax.set_title('distribution of '+title)\n",
    "    ax.legend()\n",
    "    plt.show()\n",
    "    return rr_down_lim, rr_up_lim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "id": "a699447e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "hb = pd.read_csv('output_files/heartbeats.csv')\n",
    "rr_down_lim, rr_up_lim = distribution2(hb['RR_length'], 0, 1200, title='RR_length', val = 160)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa9d8633",
   "metadata": {},
   "source": [
    "We observe that the distribution has long tails.\n",
    "Therefore, we decide to discard the heartbeats which values of RR_lenght fall into the tails of the RR_length distribution. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "id": "07bcb5c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "hb = pd.read_csv('output_files/heartbeats.csv')\n",
    "ol = len(hb)\n",
    "# - mean of the std distribution\n",
    "mean_of_rr= np.mean(hb['RR_length'])\n",
    "# - sigma of the std distribution\n",
    "sigma_of_rr=  np.std(hb['RR_length'])\n",
    "# we then compute the distances from the average in units of sigma \n",
    "hb['n_sigma_rr'] = np.abs(hb['RR_length'] - mean_of_rr)/sigma_of_rr # 2 sided\n",
    "\n",
    "# # finally, we can select the rows by keeping only the rows in which\n",
    "# hb = hb[(hb[\"n_sigma_rr\"] <= 0.1)].reset_index(drop=True)\n",
    "\n",
    "# print('percentage of peaks kept: ', round(len(hb)/ol*100,5),'%')\n",
    "# hb.to_csv('output_files/heartbeats_filtered0.csv', index=False)\n",
    "# hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "126656a1",
   "metadata": {},
   "source": [
    "We take only the left peak in the RR length distribution which corresponds to shorter RR length. This choice results in the best hearbeat shape. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "id": "960e523a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "percentage of peaks kept:  50.00141 %\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "      <th>n_sigma_rr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>667</td>\n",
       "      <td>763</td>\n",
       "      <td>694</td>\n",
       "      <td>136</td>\n",
       "      <td>695.0</td>\n",
       "      <td>0.021833</td>\n",
       "      <td>-5.121537</td>\n",
       "      <td>0.447271</td>\n",
       "      <td>0.216937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>804</td>\n",
       "      <td>898</td>\n",
       "      <td>830</td>\n",
       "      <td>134</td>\n",
       "      <td>830.0</td>\n",
       "      <td>0.021400</td>\n",
       "      <td>-4.966572</td>\n",
       "      <td>0.425133</td>\n",
       "      <td>0.219550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>938</td>\n",
       "      <td>1031</td>\n",
       "      <td>964</td>\n",
       "      <td>132</td>\n",
       "      <td>964.0</td>\n",
       "      <td>0.022563</td>\n",
       "      <td>-5.065584</td>\n",
       "      <td>0.457178</td>\n",
       "      <td>0.222164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1070</td>\n",
       "      <td>1163</td>\n",
       "      <td>1096</td>\n",
       "      <td>132</td>\n",
       "      <td>1096.0</td>\n",
       "      <td>0.021240</td>\n",
       "      <td>-4.996527</td>\n",
       "      <td>0.424505</td>\n",
       "      <td>0.222164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1603</td>\n",
       "      <td>1700</td>\n",
       "      <td>1630</td>\n",
       "      <td>138</td>\n",
       "      <td>1631.0</td>\n",
       "      <td>0.022747</td>\n",
       "      <td>-5.421619</td>\n",
       "      <td>0.493310</td>\n",
       "      <td>0.214323</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35409</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "      <td>0.228698</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35410</th>\n",
       "      <td>24</td>\n",
       "      <td>1819</td>\n",
       "      <td>847417</td>\n",
       "      <td>847523</td>\n",
       "      <td>847447</td>\n",
       "      <td>150</td>\n",
       "      <td>847407.0</td>\n",
       "      <td>-0.017340</td>\n",
       "      <td>4.541612</td>\n",
       "      <td>0.315007</td>\n",
       "      <td>0.198641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35411</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "      <td>0.240460</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35412</th>\n",
       "      <td>24</td>\n",
       "      <td>1821</td>\n",
       "      <td>847691</td>\n",
       "      <td>847778</td>\n",
       "      <td>847715</td>\n",
       "      <td>124</td>\n",
       "      <td>847668.0</td>\n",
       "      <td>-0.029394</td>\n",
       "      <td>6.413937</td>\n",
       "      <td>0.754124</td>\n",
       "      <td>0.232619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35413</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "      <td>0.210402</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>35414 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     0     667     763        694        136     695.0   \n",
       "1            1     1     804     898        830        134     830.0   \n",
       "2            1     2     938    1031        964        132     964.0   \n",
       "3            1     3    1070    1163       1096        132    1096.0   \n",
       "4            1     5    1603    1700       1630        138    1631.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "35409       24  1818  847295  847385     847320        127  847277.0   \n",
       "35410       24  1819  847417  847523     847447        150  847407.0   \n",
       "35411       24  1820  847574  847657     847597        118  847539.0   \n",
       "35412       24  1821  847691  847778     847715        124  847668.0   \n",
       "35413       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  n_sigma_rr  \n",
       "0      0.021833 -5.121537  0.447271    0.216937  \n",
       "1      0.021400 -4.966572  0.425133    0.219550  \n",
       "2      0.022563 -5.065584  0.457178    0.222164  \n",
       "3      0.021240 -4.996527  0.424505    0.222164  \n",
       "4      0.022747 -5.421619  0.493310    0.214323  \n",
       "...         ...       ...       ...         ...  \n",
       "35409 -0.021987  3.692612  0.324755    0.228698  \n",
       "35410 -0.017340  4.541612  0.315007    0.198641  \n",
       "35411 -0.022921  4.593035  0.421116    0.240460  \n",
       "35412 -0.029394  6.413937  0.754124    0.232619  \n",
       "35413 -0.033397  5.784591  0.772745    0.210402  \n",
       "\n",
       "[35414 rows x 11 columns]"
      ]
     },
     "execution_count": 238,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# most_common_length = (info[1])[np.argmax(info[0])]\n",
    "# diff_to_average = average - most_common_length\n",
    "# rr_down_lim = int(most_common_length - diff_to_average)\n",
    "# rr_up_lim = int(most_common_length + diff_to_average)\n",
    "\n",
    "# finally, we can select the rows by keeping only the rows in which\n",
    "hb = hb[(hb[\"RR_length\"] >= rr_down_lim)&(hb[\"RR_length\"] <= rr_up_lim)].reset_index(drop=True)\n",
    "\n",
    "print('percentage of peaks kept: ', round(len(hb)/ol*100,5),'%')\n",
    "hb.to_csv('output_files/heartbeats_filtered0.csv', index=False)\n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7489c32d-a20d-4638-8931-dcf37776ec76",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "We also noticed a correspondance between anomalous/noisy signal in a heartbeat and the corresponding average and standard deviation. \n",
    "\n",
    "Therefore, we examine the distributions of these quantities (heartbeat signal average, heartbeat signal standard deviation) by computing their estimators. From those, we decide to discard quantities that differ more than 1$\\sigma$ from the corresponding expectation value. This process is one-sided for the standard deviation (since it is non-negative), and two-sided for the average.\n",
    "\n",
    "The hearbeats are discarded by scrapping the corresponding lines in the hb dataframe. This step is essential to preserving the integrity of the heartbeats that are left. In fact, if this step was to be done before segmentation, the signal would be cropped based on the r-peaks: as a consequence, segmentation would lead to half-heartbeats what would be troublesome in feature extraction.\n",
    "\n",
    "This process was done for just for the main lead, from which most of the visualizable features were extracted."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "id": "c4f3db87",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "percentage of peaks kept:  42.48 %\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "      <th>n_sigma_rr</th>\n",
       "      <th>average</th>\n",
       "      <th>sigma</th>\n",
       "      <th>n_sigma_average</th>\n",
       "      <th>n_sigma_std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1603</td>\n",
       "      <td>1700</td>\n",
       "      <td>1630</td>\n",
       "      <td>138</td>\n",
       "      <td>1631.0</td>\n",
       "      <td>0.022747</td>\n",
       "      <td>-5.421619</td>\n",
       "      <td>0.493310</td>\n",
       "      <td>0.214323</td>\n",
       "      <td>-0.050164</td>\n",
       "      <td>1.253226</td>\n",
       "      <td>0.976512</td>\n",
       "      <td>0.104327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1739</td>\n",
       "      <td>1843</td>\n",
       "      <td>1768</td>\n",
       "      <td>147</td>\n",
       "      <td>1769.0</td>\n",
       "      <td>0.022400</td>\n",
       "      <td>-5.391631</td>\n",
       "      <td>0.483091</td>\n",
       "      <td>0.202561</td>\n",
       "      <td>-0.030056</td>\n",
       "      <td>1.214021</td>\n",
       "      <td>0.773147</td>\n",
       "      <td>-0.090741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1886</td>\n",
       "      <td>1989</td>\n",
       "      <td>1915</td>\n",
       "      <td>146</td>\n",
       "      <td>1915.0</td>\n",
       "      <td>0.022183</td>\n",
       "      <td>-5.358755</td>\n",
       "      <td>0.475495</td>\n",
       "      <td>0.203868</td>\n",
       "      <td>-0.030867</td>\n",
       "      <td>1.245937</td>\n",
       "      <td>0.781348</td>\n",
       "      <td>0.068060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3208</td>\n",
       "      <td>3310</td>\n",
       "      <td>3236</td>\n",
       "      <td>145</td>\n",
       "      <td>3236.0</td>\n",
       "      <td>0.022220</td>\n",
       "      <td>-5.284945</td>\n",
       "      <td>0.469724</td>\n",
       "      <td>0.205175</td>\n",
       "      <td>0.013554</td>\n",
       "      <td>1.208109</td>\n",
       "      <td>0.332079</td>\n",
       "      <td>-0.120153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>3353</td>\n",
       "      <td>3455</td>\n",
       "      <td>3381</td>\n",
       "      <td>145</td>\n",
       "      <td>3381.0</td>\n",
       "      <td>0.022099</td>\n",
       "      <td>-5.145189</td>\n",
       "      <td>0.454813</td>\n",
       "      <td>0.205175</td>\n",
       "      <td>-0.001347</td>\n",
       "      <td>1.196663</td>\n",
       "      <td>0.482788</td>\n",
       "      <td>-0.177106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30079</th>\n",
       "      <td>24</td>\n",
       "      <td>1816</td>\n",
       "      <td>847037</td>\n",
       "      <td>847128</td>\n",
       "      <td>847062</td>\n",
       "      <td>129</td>\n",
       "      <td>847019.0</td>\n",
       "      <td>-0.016930</td>\n",
       "      <td>3.502978</td>\n",
       "      <td>0.237228</td>\n",
       "      <td>0.226085</td>\n",
       "      <td>0.039249</td>\n",
       "      <td>0.083158</td>\n",
       "      <td>0.072199</td>\n",
       "      <td>-5.717411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30080</th>\n",
       "      <td>24</td>\n",
       "      <td>1817</td>\n",
       "      <td>847166</td>\n",
       "      <td>847257</td>\n",
       "      <td>847191</td>\n",
       "      <td>129</td>\n",
       "      <td>847149.0</td>\n",
       "      <td>-0.021423</td>\n",
       "      <td>3.853764</td>\n",
       "      <td>0.330233</td>\n",
       "      <td>0.226085</td>\n",
       "      <td>0.047095</td>\n",
       "      <td>0.094299</td>\n",
       "      <td>0.007148</td>\n",
       "      <td>-5.661976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30081</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "      <td>0.228698</td>\n",
       "      <td>0.056552</td>\n",
       "      <td>0.125824</td>\n",
       "      <td>0.102793</td>\n",
       "      <td>-5.505126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30082</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "      <td>0.240460</td>\n",
       "      <td>-0.012383</td>\n",
       "      <td>0.254795</td>\n",
       "      <td>0.594403</td>\n",
       "      <td>-4.863424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30083</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "      <td>0.210402</td>\n",
       "      <td>-0.043063</td>\n",
       "      <td>0.187300</td>\n",
       "      <td>0.904694</td>\n",
       "      <td>-5.199248</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30084 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     5    1603    1700       1630        138    1631.0   \n",
       "1            1     6    1739    1843       1768        147    1769.0   \n",
       "2            1     7    1886    1989       1915        146    1915.0   \n",
       "3            1    10    3208    3310       3236        145    3236.0   \n",
       "4            1    11    3353    3455       3381        145    3381.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "30079       24  1816  847037  847128     847062        129  847019.0   \n",
       "30080       24  1817  847166  847257     847191        129  847149.0   \n",
       "30081       24  1818  847295  847385     847320        127  847277.0   \n",
       "30082       24  1820  847574  847657     847597        118  847539.0   \n",
       "30083       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  n_sigma_rr   average     sigma  \\\n",
       "0      0.022747 -5.421619  0.493310    0.214323 -0.050164  1.253226   \n",
       "1      0.022400 -5.391631  0.483091    0.202561 -0.030056  1.214021   \n",
       "2      0.022183 -5.358755  0.475495    0.203868 -0.030867  1.245937   \n",
       "3      0.022220 -5.284945  0.469724    0.205175  0.013554  1.208109   \n",
       "4      0.022099 -5.145189  0.454813    0.205175 -0.001347  1.196663   \n",
       "...         ...       ...       ...         ...       ...       ...   \n",
       "30079 -0.016930  3.502978  0.237228    0.226085  0.039249  0.083158   \n",
       "30080 -0.021423  3.853764  0.330233    0.226085  0.047095  0.094299   \n",
       "30081 -0.021987  3.692612  0.324755    0.228698  0.056552  0.125824   \n",
       "30082 -0.022921  4.593035  0.421116    0.240460 -0.012383  0.254795   \n",
       "30083 -0.033397  5.784591  0.772745    0.210402 -0.043063  0.187300   \n",
       "\n",
       "       n_sigma_average  n_sigma_std  \n",
       "0             0.976512     0.104327  \n",
       "1             0.773147    -0.090741  \n",
       "2             0.781348     0.068060  \n",
       "3             0.332079    -0.120153  \n",
       "4             0.482788    -0.177106  \n",
       "...                ...          ...  \n",
       "30079         0.072199    -5.717411  \n",
       "30080         0.007148    -5.661976  \n",
       "30081         0.102793    -5.505126  \n",
       "30082         0.594403    -4.863424  \n",
       "30083         0.904694    -5.199248  \n",
       "\n",
       "[30084 rows x 15 columns]"
      ]
     },
     "execution_count": 241,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hb = pd.read_csv('output_files/heartbeats_filtered0.csv')\n",
    "# we initialize the values of average and standard deviation\n",
    "average = np.zeros(len(hb))\n",
    "sigma = np.zeros(len(hb))\n",
    "# we recover the information that we need from the hb dataframe as arrays\n",
    "start = np.array(hb[\"start\"]) \n",
    "end = np.array(hb[\"end\"]) \n",
    "# we extract the quantities we need\n",
    "path = \"blah\"\n",
    "for beat in range(len(hb)):\n",
    "    if path != hb[\"patient\"][beat]:\n",
    "        path = hb[\"patient\"][beat]\n",
    "        data = pd.read_csv(\"output_files/data\"+str(path)+\"_filtered.csv\")\n",
    "        y = np.array(data[main_lead])\n",
    "    \n",
    "    average[beat] = np.mean(y[start[beat]:(end[beat]+1)])\n",
    "    sigma[beat] = np.std(y[start[beat]:(end[beat]+1)])\n",
    "    if np.isnan(average[beat]) or np.isnan(sigma[beat]):\n",
    "        print((end[beat]+1)- start[beat])\n",
    "        \n",
    "# and add them to the dataframe\n",
    "hb['average'] = average\n",
    "hb['sigma'] = sigma\n",
    "\n",
    "# we can now select the rows based on the distribution\n",
    "# to do that, we compute the\n",
    "# - mean of the average distribution\n",
    "mean_of_average = np.mean(average)\n",
    "# - sigma of the average distribution\n",
    "sigma_of_average = np.std(average)\n",
    "# - mean of the std distribution\n",
    "mean_of_sigma = np.mean(sigma)\n",
    "# - sigma of the std distribution\n",
    "sigma_of_sigma=  np.std(sigma)\n",
    "# we then compute the distances from the average in units of sigma \n",
    "hb['n_sigma_average'] = np.abs(average - mean_of_average)/sigma_of_average # 2 sided\n",
    "hb['n_sigma_std'] = (sigma - mean_of_sigma)/sigma_of_sigma # 1 sided \n",
    "# finally, we can select the rows by keeping only the rows in which\n",
    "hb = hb[(hb[\"n_sigma_average\"] <= 1) & (hb[\"n_sigma_std\"] <=1)].reset_index(drop=True)\n",
    "print('percentage of peaks kept: ', round(len(hb)/ol*100,2),'%')\n",
    "hb.to_csv('output_files/heartbeats_filtered.csv', index=False)\n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1454176c-7079-44bd-9e16-04f12f40a3a5",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## 4B) Wavelet transform coefficients extraction"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7d6304f-c3ff-4dc9-94c8-765dab85ad76",
   "metadata": {},
   "source": [
    "The feature extraction using discrete wavelet transform (DWT) is based on the idea presented in the reference [2]. DWT decomposes a signal into detail signal and approximation signal and as a result we obtain for each decomposition level a set of coefficients. Information about QRS complex is mainly concentrated at detail levels 3-5, while information about T and P waves is located at detail levels 6 and 7 [2].\n",
    "\n",
    "We tried to implement the algorithm with similar decomposition levels, but we receive Warnings that decompostion level was too high. In the end, we settled for 4 decompostion levels, therefore obtaining information about QRS complex. We take approximation signal and detail levels 3-4, which in the end matches proposals discussed in [1]. The idea is that we can extaract features of P and T waves separetly and add them later.\n",
    "\n",
    "The algorithm for feature extraction using DWT is the following:\n",
    "\n",
    "1) We firstly determine a fixed size segment near R peak so that the number of decompostion coefficients (thus features) is aslo fixed.\n",
    "\n",
    "2) We take each segment (beat), perform DWT and return a row of features corresponding to this beat."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d32eb97",
   "metadata": {},
   "source": [
    "We perform this procedure on every lead and save the features in a matrix, where each row corresponds to a particular beat."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "59445fad",
   "metadata": {},
   "outputs": [],
   "source": [
    "def leads_dwt_coeff(hb, lead):\n",
    "    '''\n",
    "    function that returns the DWT coefficients of the input lead for all heartbeats\n",
    "    '''\n",
    "    # we retrieved the columns of the dataframe that we need\n",
    "    start = np.array(hb[\"start\"])\n",
    "    end = np.array(hb[\"end\"])\n",
    "    R = np.array(hb[\"ann_index\"])\n",
    "    # we examine the cases that we arbitrarily consider degenerate (to allow for a not-too-small fixed window)\n",
    "    cond = (hb['idx']==0)\n",
    "    \n",
    "    # we find the locations of the beginnings\n",
    "    j = np.where(hb['idx']==0)[0]\n",
    "    # and of the ends\n",
    "    k = np.concatenate([(j -1)[1:],[len(hb)-1]])\n",
    "    # and find a treshold that can grant us a not-too-small window based based on the histogram of the values\n",
    "    # (neglecting the lower part of the plt.hist(R[j]-start[j]) and the upper part of the plt.hist(end[k]-R[k])\n",
    "    temp_start = (R[j]-start[j])>=27\n",
    "    temp_end =  (end[k]-R[k])>=163\n",
    "    # and find the indexes corresponding to the degenerate cases\n",
    "    where_start = np.where(temp_start)[0]\n",
    "    where_end = np.where(temp_end)[0]\n",
    "    # so that we can neglect them in finding the fixed window\n",
    "    fixed_start = R - 30 #np.min((R[j]-start[j])[temp_start]).astype(int)\n",
    "    fixed_end = R + 90 #np.min((end[k]-R[k])[temp_end]).astype(int)\n",
    "\n",
    "    path = 'blah'\n",
    "    # for each patient\n",
    "    for i in range(0,len(hb)):\n",
    "        # we make sure to change the file/patient if we have exhausted the previous one\n",
    "        if path != str(hb[\"patient\"][i]):\n",
    "            path = str(hb[\"patient\"][i])\n",
    "            data = pd.read_csv(\"output_files/data\"+path+\"_filtered.csv\")\n",
    "            y = np.array(data[lead])\n",
    "            \n",
    "        # we select the signal in the fixed window\n",
    "        z = y[fixed_start[i]:fixed_end[i]]\n",
    "        # we extract the coefficients\n",
    "        coeffs = pywt.wavedec(z, 'db4', level=4, axis = 0)\n",
    "        # we initialize the output to zeros\n",
    "        if (i == 0):\n",
    "            l2 = len(coeffs[0])\n",
    "            l3 = len(coeffs[1])\n",
    "            l4 = len(coeffs[2])\n",
    "            # we extract the positions needed to make a single array\n",
    "            t1 = l2\n",
    "            t2 = t1 + l3\n",
    "            t3 = t2 + l4\n",
    "            features = np.zeros((len(hb), t3))\n",
    "        # if it's not a degenerate case, we save the result in a row\n",
    "        if i not in np.concatenate([where_start,where_end]):\n",
    "            features[i,0:t1]  = coeffs[0]\n",
    "            features[i,t1:t2] = coeffs[1]\n",
    "            features[i,t2:t3] = coeffs[2]\n",
    "        # otherwise we leave zeros\n",
    "\n",
    "    return features "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "55376705",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the extraction of wavelet features took 1.26 minutes\n"
     ]
    }
   ],
   "source": [
    "parallelization = True\n",
    "if parallelization:\n",
    "    n_jobs = -1 #to use all available CPU cores \n",
    "    time1 = time.time()\n",
    "    features_list = Parallel(n_jobs=n_jobs)(delayed(leads_dwt_coeff)(hb, lead) for lead in leads)\n",
    "    features = np.hstack(features_list)\n",
    "    time2 = time.time()\n",
    "    print(\"the extraction of wavelet features took\",round((time2-time1)/60,2),'minutes')\n",
    "else:\n",
    "    # If want to do it without parallelization\n",
    "    for i, lead in enumerate(leads):\n",
    "        f = leads_dwt_coeff(hb, lead)\n",
    "        if (i == 0):\n",
    "            features = f\n",
    "        else:\n",
    "            features = np.hstack((features, f))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f56edbc-6b66-451a-bbaa-820689e94e9b",
   "metadata": {
    "tags": []
   },
   "source": [
    "## 4C) Extraction of visualizable features:"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d04b7284-7b84-4b08-b435-b81cde8ae3c7",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "### First round of visualizable features\n",
    "\n",
    "We can now extract the features that were dependent on the ones we just extracted. They consist of:\n",
    "- the start and end of the QRS complex\n",
    "- the (real) position and height of the R peak\n",
    "- the difference (err) between the two estimates of the position of the R peak\n",
    "- the value and position of the maximum of the heartbeat\n",
    "- the value and position of the minimum of the heartbeat\n",
    "- the positive and negative integral of the heartbeat "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7bc5d69e",
   "metadata": {},
   "source": [
    "We already fitted the peaks found with parabolas and saved their parameters: we just need to convert the saved parameters (which were for linear regression because of the change of variable) to the classic parameters of the parabola\n",
    "\n",
    "$$  \\begin{cases} \n",
    "    \\begin{matrix}\n",
    "    y'(j) = a \\cdot x + v \\\\\n",
    "    x = (i-j)^2\n",
    "    \\end{matrix}\n",
    "    \\end{cases} \n",
    "    \\;\\;\\; \\Rightarrow \\;\\;\\; \n",
    "    y(j) = A\\cdot j^2 + B \\cdot j + C\n",
    "$$\n",
    "\n",
    "\n",
    "As a result, we get that\n",
    "\n",
    "$$  \\begin{cases} \n",
    "    \\begin{matrix}\n",
    "    A = a \\\\\n",
    "    B = - 2 \\cdot a \\cdot i \\\\\n",
    "    C = v +  a \\cdot i^2\n",
    "    \\end{matrix}\n",
    "    \\end{cases} \n",
    "    \\;\\;\\; \\Rightarrow \\;\\;\\; \n",
    "    \\Delta = -4 \\cdot a \\cdot v\n",
    "$$\n",
    "\n",
    "so that we can calculate the intesections of the parabola with the x axis and their distance from each other as:\n",
    "\n",
    "$$ x_\\pm = \\frac{-B \\pm \\sqrt{\\Delta}}{2A} \\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\;\\; \\delta_x = \\frac{\\sqrt{\\Delta}}{A}  $$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "id": "3b3838a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def x_int_points(i,a,v):\n",
    "    # we find the parabola coefficients from the fit coefficients\n",
    "    A = a\n",
    "    B = -2*a*i\n",
    "    C = v+a*i**2\n",
    "    Delta = np.sqrt(-4*a*v)\n",
    "    # we find the intersections\n",
    "    x_low = (-B-Delta)/(2*A)\n",
    "    x_up = (-B+Delta)/(2*A)\n",
    "    # and order them\n",
    "    temp = np.column_stack([x_low,x_up])\n",
    "    temp.sort(axis=1)\n",
    "    x_low = temp[:,0]\n",
    "    x_up = temp[:,1]\n",
    "    # and the distance from each other\n",
    "    delta_x = Delta/A\n",
    "    return x_low, x_up, delta_x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "id": "e7060f66",
   "metadata": {},
   "outputs": [],
   "source": [
    "def integral(y_beat):\n",
    "    reg_int = np.trapz(y_beat)\n",
    "    abs_int = np.trapz(np.abs(y_beat))\n",
    "    pos_int = (reg_int + abs_int)/2\n",
    "    neg_int = (abs_int - reg_int)/2\n",
    "    return pos_int, neg_int"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "id": "d5f96c8a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "start < start_fit & end_fit < end : True\n",
      "overflowing from the start: % 0.0\n",
      "overflowing from the end: % 0.0\n",
      "the extraction of visualizable features took 0.8 mins\n"
     ]
    }
   ],
   "source": [
    "hb = pd.read_csv('output_files/heartbeats_filtered.csv')\n",
    "t1 = time.time()\n",
    "\n",
    "# we find initialize the variable that we want to retrieve and store into arrays based on their meaning\n",
    "# - the position of the R_peak found as the local maximum/minimum\n",
    "x_R_peak = np.zeros(len(hb))    #np.array(hb[\"ann_index\"], dtype = int)\n",
    "# - the height of the R peak\n",
    "y_R_peak = np.zeros(len(hb))\n",
    "# - the coordinates of a point of the peak\n",
    "x_point = np.zeros(len(hb))\n",
    "y_point = np.zeros(len(hb))\n",
    "# - the difference between the annotation index and the actual position of the peak\n",
    "err = np.zeros(len(hb))\n",
    "# - the value of the maxima for the entire heartbeat\n",
    "Ms = np.zeros(len(hb))\n",
    "# - the index of the maxima for the entire heartbeat\n",
    "argMs = np.zeros(len(hb))\n",
    "# - the value of the minima for the entire heartbeat\n",
    "ms = np.zeros(len(hb))\n",
    "# - the index of the minima for the entire heartbeat\n",
    "argms = np.zeros(len(hb))\n",
    "# - the value of the postive integral for the entire heartbeat\n",
    "pos_int = np.zeros(len(hb))\n",
    "# - the value of the negative integral for the entire heartbeat\n",
    "neg_int = np.zeros(len(hb))\n",
    "\n",
    "# we then extract the information needed from the dataframe as arrays\n",
    "# - the start of the heartbeat\n",
    "start = np.array(hb[\"start\"]) \n",
    "# - the end of the heartbeat\n",
    "end = np.array(hb[\"end\"]) \n",
    "# - the RR intervals\n",
    "RR = np.array(hb[\"RR_length\"])\n",
    "# - the coefficients of the parabola fits\n",
    "i = hb['ann_index'].to_numpy()\n",
    "a = hb['a'].to_numpy()\n",
    "v = hb['v'].to_numpy()\n",
    "# - the average\n",
    "average = np.array(hb['average'])\n",
    "\n",
    "# we can now do the some of the vectorizable processes:\n",
    "# we exploit the known proportion QRS:RR = 0.09:0.8 to calculate the expected QRS semi-duration\n",
    "delta = np.array(RR*0.09/(0.8*2), dtype=int) \n",
    "\n",
    "# we save the instances in which the expected QRS semi-duration is null (will be needed later)\n",
    "skip = np.where(delta ==0)[0]\n",
    "okay = np.where(delta !=0)[0]\n",
    "# the expected QRS semi-duration is then used to find the \n",
    "start_fit = (hb['ann_index']-delta).astype(int)\n",
    "end_fit = (hb['ann_index']+delta).astype(int)\n",
    "\n",
    "# we make sure that the \n",
    "print('start < start_fit & end_fit < end :',np.all([start<start_fit, end>end_fit]))\n",
    "\n",
    "path = \"blah\"\n",
    "for beat in range(len(hb)):\n",
    "    if path != str(hb[\"patient\"][beat]):\n",
    "        path = str(hb[\"patient\"][beat])\n",
    "        data = pd.read_csv(\"output_files/data\"+path+\"_filtered.csv\")\n",
    "        y = np.array(data[main_lead])\n",
    "    \n",
    "    Ms[beat] = np.max(y[start[beat]:(end[beat]+1)])\n",
    "    ms[beat] = np.min(y[start[beat]:(end[beat]+1)])\n",
    "    argMs[beat] = np.argmax(y[start[beat]:(end[beat]+1)])\n",
    "    argms[beat] = np.argmin(y[start[beat]:(end[beat]+1)])\n",
    "    pos_int[beat], neg_int[beat] = integral(y[start[beat]:(end[beat]+1)])\n",
    "        \n",
    "    if (beat in okay):\n",
    "        # we select the beat\n",
    "        x0 = np.arange(start[beat],end[beat]+1,1)\n",
    "        y0 = np.array(y[start[beat]:(end[beat]+1)])\n",
    "        # we select the absolute of the net signal (obtained by subtracting the noise)\n",
    "        z = np.abs(y0-average[beat])\n",
    "        err[beat] = (np.argmax(z)+start[beat]) - x_R_peak[beat]\n",
    "        # we find the actual R peak (since the annotations aren't always accurate)\n",
    "        x_R_peak[beat] = np.argmax(z)+start[beat]\n",
    "        x_fit = np.arange(start_fit[beat],end_fit[beat]+1,1)\n",
    "        y_fit = np.array(y[start_fit[beat]:(end_fit[beat]+1)])\n",
    "        # we define\n",
    "        temp = np.abs(y_fit-average[beat])\n",
    "\n",
    "        if (y_fit[temp==np.max(temp)]>average[beat]):\n",
    "            cut_level = average[beat] #+ sigma[beat]\n",
    "            x_fit = x_fit[y_fit>=cut_level]\n",
    "            y_fit = y_fit[y_fit>=cut_level]\n",
    "            y_R_peak[beat] = np.max(y_fit)\n",
    "            x_R_peak[beat] = x_fit[np.argmax(y_fit)]\n",
    "            bool_mask = (y_fit>=average[beat])\n",
    "        else:\n",
    "            cut_level = average[beat] #- sigma[beat]\n",
    "            x_fit = x_fit[y_fit<=cut_level]\n",
    "            y_fit = y_fit[y_fit<=cut_level]\n",
    "            y_R_peak[beat] = np.min(y_fit)\n",
    "            x_R_peak[beat] = x_fit[np.argmin(y_fit)]\n",
    "            bool_mask = (y_fit<=average[beat])\n",
    "        x_point[beat] = (x_fit[bool_mask])[-1]\n",
    "        y_point[beat] = (y_fit[bool_mask])[-1]\n",
    "        \n",
    "# we finally take care the instances in which the expected QRS semi-duration is null\n",
    "# such instances are degenerate cases in which (x_point,y_point) will coincide with (x_R_peak,y_R_peak)\n",
    "x_point[skip] = x_R_peak[skip]\n",
    "y_point[skip] = y_R_peak[skip]\n",
    "\n",
    "#QRS_start, QRS_end, QRS = x_intersections_points(x_point, y_point, x_R_peak, y_R_peak, average)\n",
    "QRS_start, QRS_end, QRS = x_int_points(i,a,v)\n",
    "print('overflowing from the start: %',len(np.where(QRS_start<start)[0])/len(hb)*100)\n",
    "print('overflowing from the end: %',len(np.where(QRS_end>end)[0])/len(hb)*100)\n",
    "\n",
    "t2 = time.time()\n",
    "print(\"the extraction of visualizable features took\",round((t2-t1)/60,2),'mins')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "id": "bb850f73",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "      <th>...</th>\n",
       "      <th>QRS_end</th>\n",
       "      <th>QRS_length</th>\n",
       "      <th>x_R_peak</th>\n",
       "      <th>y_R_peak</th>\n",
       "      <th>pos_int</th>\n",
       "      <th>neg_int</th>\n",
       "      <th>arg_max</th>\n",
       "      <th>arg_min</th>\n",
       "      <th>max</th>\n",
       "      <th>min</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1603</td>\n",
       "      <td>1700</td>\n",
       "      <td>1630</td>\n",
       "      <td>138</td>\n",
       "      <td>1631.0</td>\n",
       "      <td>0.022747</td>\n",
       "      <td>-5.421619</td>\n",
       "      <td>0.493310</td>\n",
       "      <td>...</td>\n",
       "      <td>1645</td>\n",
       "      <td>30</td>\n",
       "      <td>1631</td>\n",
       "      <td>6.241203</td>\n",
       "      <td>27.286423</td>\n",
       "      <td>32.114297</td>\n",
       "      <td>28</td>\n",
       "      <td>33</td>\n",
       "      <td>6.241203</td>\n",
       "      <td>-2.306481</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1739</td>\n",
       "      <td>1843</td>\n",
       "      <td>1768</td>\n",
       "      <td>147</td>\n",
       "      <td>1769.0</td>\n",
       "      <td>0.022400</td>\n",
       "      <td>-5.391631</td>\n",
       "      <td>0.483091</td>\n",
       "      <td>...</td>\n",
       "      <td>1783</td>\n",
       "      <td>31</td>\n",
       "      <td>1769</td>\n",
       "      <td>6.324524</td>\n",
       "      <td>27.677816</td>\n",
       "      <td>30.861893</td>\n",
       "      <td>30</td>\n",
       "      <td>35</td>\n",
       "      <td>6.324524</td>\n",
       "      <td>-1.647094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1886</td>\n",
       "      <td>1989</td>\n",
       "      <td>1915</td>\n",
       "      <td>146</td>\n",
       "      <td>1915.0</td>\n",
       "      <td>0.022183</td>\n",
       "      <td>-5.358755</td>\n",
       "      <td>0.475495</td>\n",
       "      <td>...</td>\n",
       "      <td>1930</td>\n",
       "      <td>31</td>\n",
       "      <td>1916</td>\n",
       "      <td>6.176768</td>\n",
       "      <td>28.462620</td>\n",
       "      <td>31.806860</td>\n",
       "      <td>30</td>\n",
       "      <td>35</td>\n",
       "      <td>6.176768</td>\n",
       "      <td>-2.107186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3208</td>\n",
       "      <td>3310</td>\n",
       "      <td>3236</td>\n",
       "      <td>145</td>\n",
       "      <td>3236.0</td>\n",
       "      <td>0.022220</td>\n",
       "      <td>-5.284945</td>\n",
       "      <td>0.469724</td>\n",
       "      <td>...</td>\n",
       "      <td>3251</td>\n",
       "      <td>30</td>\n",
       "      <td>3236</td>\n",
       "      <td>6.224773</td>\n",
       "      <td>28.535936</td>\n",
       "      <td>27.017778</td>\n",
       "      <td>28</td>\n",
       "      <td>34</td>\n",
       "      <td>6.224773</td>\n",
       "      <td>-1.996578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>3353</td>\n",
       "      <td>3455</td>\n",
       "      <td>3381</td>\n",
       "      <td>145</td>\n",
       "      <td>3381.0</td>\n",
       "      <td>0.022099</td>\n",
       "      <td>-5.145189</td>\n",
       "      <td>0.454813</td>\n",
       "      <td>...</td>\n",
       "      <td>3396</td>\n",
       "      <td>30</td>\n",
       "      <td>3382</td>\n",
       "      <td>6.222979</td>\n",
       "      <td>27.152976</td>\n",
       "      <td>27.260977</td>\n",
       "      <td>29</td>\n",
       "      <td>34</td>\n",
       "      <td>6.222979</td>\n",
       "      <td>-1.870475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30079</th>\n",
       "      <td>24</td>\n",
       "      <td>1816</td>\n",
       "      <td>847037</td>\n",
       "      <td>847128</td>\n",
       "      <td>847062</td>\n",
       "      <td>129</td>\n",
       "      <td>847019.0</td>\n",
       "      <td>-0.016930</td>\n",
       "      <td>3.502978</td>\n",
       "      <td>0.237228</td>\n",
       "      <td>...</td>\n",
       "      <td>847076</td>\n",
       "      <td>-28</td>\n",
       "      <td>847058</td>\n",
       "      <td>-0.022090</td>\n",
       "      <td>4.122447</td>\n",
       "      <td>0.467157</td>\n",
       "      <td>84</td>\n",
       "      <td>0</td>\n",
       "      <td>0.323828</td>\n",
       "      <td>-0.086825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30080</th>\n",
       "      <td>24</td>\n",
       "      <td>1817</td>\n",
       "      <td>847166</td>\n",
       "      <td>847257</td>\n",
       "      <td>847191</td>\n",
       "      <td>129</td>\n",
       "      <td>847149.0</td>\n",
       "      <td>-0.021423</td>\n",
       "      <td>3.853764</td>\n",
       "      <td>0.330233</td>\n",
       "      <td>...</td>\n",
       "      <td>847204</td>\n",
       "      <td>-26</td>\n",
       "      <td>847188</td>\n",
       "      <td>-0.033164</td>\n",
       "      <td>4.850670</td>\n",
       "      <td>0.480880</td>\n",
       "      <td>5</td>\n",
       "      <td>91</td>\n",
       "      <td>0.326714</td>\n",
       "      <td>-0.062228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30081</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "      <td>...</td>\n",
       "      <td>847332</td>\n",
       "      <td>-25</td>\n",
       "      <td>847318</td>\n",
       "      <td>-0.061281</td>\n",
       "      <td>6.164660</td>\n",
       "      <td>1.018007</td>\n",
       "      <td>3</td>\n",
       "      <td>90</td>\n",
       "      <td>0.545477</td>\n",
       "      <td>-0.104907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30082</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "      <td>...</td>\n",
       "      <td>847611</td>\n",
       "      <td>-28</td>\n",
       "      <td>847603</td>\n",
       "      <td>0.106598</td>\n",
       "      <td>6.779662</td>\n",
       "      <td>7.486341</td>\n",
       "      <td>43</td>\n",
       "      <td>79</td>\n",
       "      <td>0.217286</td>\n",
       "      <td>-0.962193</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30083</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "      <td>...</td>\n",
       "      <td>847852</td>\n",
       "      <td>-26</td>\n",
       "      <td>847846</td>\n",
       "      <td>0.049305</td>\n",
       "      <td>5.120997</td>\n",
       "      <td>9.025870</td>\n",
       "      <td>93</td>\n",
       "      <td>98</td>\n",
       "      <td>0.205192</td>\n",
       "      <td>-0.701594</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30084 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     5    1603    1700       1630        138    1631.0   \n",
       "1            1     6    1739    1843       1768        147    1769.0   \n",
       "2            1     7    1886    1989       1915        146    1915.0   \n",
       "3            1    10    3208    3310       3236        145    3236.0   \n",
       "4            1    11    3353    3455       3381        145    3381.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "30079       24  1816  847037  847128     847062        129  847019.0   \n",
       "30080       24  1817  847166  847257     847191        129  847149.0   \n",
       "30081       24  1818  847295  847385     847320        127  847277.0   \n",
       "30082       24  1820  847574  847657     847597        118  847539.0   \n",
       "30083       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  ...  QRS_end  QRS_length  x_R_peak  \\\n",
       "0      0.022747 -5.421619  0.493310  ...     1645          30      1631   \n",
       "1      0.022400 -5.391631  0.483091  ...     1783          31      1769   \n",
       "2      0.022183 -5.358755  0.475495  ...     1930          31      1916   \n",
       "3      0.022220 -5.284945  0.469724  ...     3251          30      3236   \n",
       "4      0.022099 -5.145189  0.454813  ...     3396          30      3382   \n",
       "...         ...       ...       ...  ...      ...         ...       ...   \n",
       "30079 -0.016930  3.502978  0.237228  ...   847076         -28    847058   \n",
       "30080 -0.021423  3.853764  0.330233  ...   847204         -26    847188   \n",
       "30081 -0.021987  3.692612  0.324755  ...   847332         -25    847318   \n",
       "30082 -0.022921  4.593035  0.421116  ...   847611         -28    847603   \n",
       "30083 -0.033397  5.784591  0.772745  ...   847852         -26    847846   \n",
       "\n",
       "       y_R_peak    pos_int    neg_int  arg_max  arg_min       max       min  \n",
       "0      6.241203  27.286423  32.114297       28       33  6.241203 -2.306481  \n",
       "1      6.324524  27.677816  30.861893       30       35  6.324524 -1.647094  \n",
       "2      6.176768  28.462620  31.806860       30       35  6.176768 -2.107186  \n",
       "3      6.224773  28.535936  27.017778       28       34  6.224773 -1.996578  \n",
       "4      6.222979  27.152976  27.260977       29       34  6.222979 -1.870475  \n",
       "...         ...        ...        ...      ...      ...       ...       ...  \n",
       "30079 -0.022090   4.122447   0.467157       84        0  0.323828 -0.086825  \n",
       "30080 -0.033164   4.850670   0.480880        5       91  0.326714 -0.062228  \n",
       "30081 -0.061281   6.164660   1.018007        3       90  0.545477 -0.104907  \n",
       "30082  0.106598   6.779662   7.486341       43       79  0.217286 -0.962193  \n",
       "30083  0.049305   5.120997   9.025870       93       98  0.205192 -0.701594  \n",
       "\n",
       "[30084 rows x 26 columns]"
      ]
     },
     "execution_count": 245,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# before updating the dataframe with the first round of visualizable features, we need to make sure that the indexes \n",
    "# found for the start and end of the QRS complex don't exede the indexes for the start and end of the heartbeat\n",
    "# we find and fix the overflow from the start\n",
    "condition = (QRS_start < start)\n",
    "bad = np.where(condition)[0]\n",
    "QRS_start[bad] = start[bad]\n",
    "# and we find and fix the overflow from the end\n",
    "condition = (QRS_end > end)\n",
    "bad = np.where(condition)[0]\n",
    "QRS_end[bad] = end[bad]\n",
    "\n",
    "QRS_start = QRS_start.astype(int)\n",
    "QRS_end = QRS_end.astype(int)\n",
    "QRS = QRS.astype(int)\n",
    "x_R_peak = x_R_peak.astype(int)\n",
    "argMs = argMs.astype(int)\n",
    "argms = argms.astype(int)\n",
    "\n",
    "# so that we can update the dataframe\n",
    "hb['QRS_start'] = QRS_start\n",
    "hb[\"QRS_end\"] = QRS_end\n",
    "hb[\"QRS_length\"] = QRS\n",
    "hb['x_R_peak'] = x_R_peak\n",
    "hb[\"y_R_peak\"] = y_R_peak\n",
    "hb['pos_int'] = pos_int\n",
    "hb['neg_int'] = neg_int\n",
    "hb['arg_max'] = argMs\n",
    "hb['arg_min'] = argms\n",
    "hb['max'] = Ms\n",
    "hb['min'] = ms\n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf7874f0",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "### Second round of visualizable features extraction\n",
    "We can now also extract the features that were dependent on the ones we just extracted. They consist of:\n",
    "- the positive and negative integral of the QRS complex\n",
    "- the start and end of the P peak\n",
    "- the position and height of the P peak\n",
    "- the positive and negative integral of the P peak\n",
    "- the start and end of the T peak\n",
    "- the position and height of the T peak\n",
    "- the positive and negative integral of the T peak"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "id": "c4a8a7a0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# we extract the index/position of the R peak\n",
    "x_R_peak = hb['x_R_peak']\n",
    "\n",
    "# we initialize the arrays needed for the second for-loop\n",
    "# - the positive integral of the QRS complex\n",
    "r_pos_int = np.zeros(len(hb))\n",
    "# - the negative integral of the QRS complex\n",
    "r_neg_int = np.zeros(len(hb))\n",
    "# - the index of the start of the P peak\n",
    "p_start = np.zeros(len(hb), dtype=int)\n",
    "# - the index of the end of the P peak\n",
    "p_end = np.zeros(len(hb), dtype=int)\n",
    "# - the index/position of the P peak\n",
    "x_P_peak = np.zeros(len(hb), dtype=int)\n",
    "# - the value of the signal in the P peak\n",
    "y_P_peak = np.zeros(len(hb))\n",
    "# - the positive integral of P peak\n",
    "p_pos_int = np.zeros(len(hb))\n",
    "# - the negative integral of P peak\n",
    "p_neg_int = np.zeros(len(hb))\n",
    "# - the index of the start of the T peak\n",
    "t_start = np.zeros(len(hb), dtype=int)\n",
    "# - the index of the end of the T peak\n",
    "t_end = np.zeros(len(hb), dtype=int)\n",
    "# - the index/position of the T peak\n",
    "x_T_peak = np.zeros(len(hb), dtype=int)\n",
    "# - the value of the signal in the T peak\n",
    "y_T_peak = np.zeros(len(hb))\n",
    "# - the positive integral of the T peak\n",
    "t_pos_int = np.zeros(len(hb))\n",
    "# - the negative integral of the T peak\n",
    "t_neg_int = np.zeros(len(hb))\n",
    "\n",
    "# we compute everything that we can do outside of the following for-loops\n",
    "# particularly, we focus on finding the non-degenerate cases, based on the following conditions\n",
    "start_condition = (QRS_start > start)\n",
    "end_condition = (QRS_end < end)\n",
    "# we find the indexes that correspond to the condition\n",
    "good_start = np.where(start_condition)[0]\n",
    "good_end = np.where(end_condition)[0]\n",
    "# and also the indexes that don't satisfy the condition (\"degenerate cases\")\n",
    "bad_start = np.where(np.bitwise_not(start_condition))[0]\n",
    "bad_end = np.where(np.bitwise_not(end_condition))[0]\n",
    "# based on the start condition,I set the info for the P peak\n",
    "# in the non-degenerate cases:\n",
    "# we set the beginning of P peak as the start of the heartbeat\n",
    "p_start[good_start] = start[good_start]\n",
    "# and the end of the P peak as the start of the QRS comples\n",
    "p_end[good_start] = QRS_start[good_start]\n",
    "# for the degenerate cases:\n",
    "# we set both the beginning, end and P peak itself as the start of the heartbeat\n",
    "p_start[bad_start] = start[bad_start]\n",
    "p_end[bad_start] = start[bad_start]\n",
    "x_P_peak[bad_start] = start[bad_start]\n",
    "# based on the end condition, we set the info for the T peak\n",
    "# in the non-degenerate cases:\n",
    "# we set the beginning of T peak as the end of the QRS comples\n",
    "t_start[good_end] = QRS_end[good_end]\n",
    "# and the end of the T peak as the end of the heartbeat\n",
    "t_end[good_end] = end[good_end]\n",
    "# for the degenerate cases:\n",
    "# we set both the beginning, end and T peak itself as the end of the heartbeat\n",
    "t_start[bad_end] = end[bad_end]\n",
    "t_end[bad_end] = end[bad_end]\n",
    "x_T_peak[bad_end] = end[bad_end]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "394c9bad",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the algorithm took 0.77 minutes\n"
     ]
    }
   ],
   "source": [
    "######## another loop over beats <- for QRS integrals + P + T + FFT\n",
    "t1 = time.time()\n",
    "    \n",
    "path = \"blah\"\n",
    "for beat in range(len(hb)):\n",
    "    \n",
    "    if path != str(hb[\"patient\"][beat]):\n",
    "        path = str(hb[\"patient\"][beat])\n",
    "        data = pd.read_csv(\"output_files/data\"+path+\"_filtered.csv\")\n",
    "        y = np.array(data[main_lead])\n",
    "        \n",
    "    ###integrals for QRS    \n",
    "    y_r = y[QRS_start[beat]:QRS_end[beat]]\n",
    "    r_pos_int[beat], r_neg_int[beat] = integral(y_r)\n",
    "    \n",
    "    ### for P\n",
    "    if (beat in good_start):\n",
    "        y_p = y[p_start[beat]:p_end[beat]]\n",
    "        y_P_peak[beat] = np.max(y_p)\n",
    "        x_P_peak[beat] = np.argmax(y_p) + p_start[beat]\n",
    "        p_pos_int[beat], p_neg_int[beat] = integral(y_p)\n",
    "    else:\n",
    "        y_P_peak[beat] = y[x_P_peak[beat]]\n",
    "        p_pos_int[beat], p_neg_int[beat] = 0,0\n",
    "        \n",
    "    ### for T  \n",
    "    if (beat in good_end):\n",
    "        y_t = y[t_start[beat]:t_end[beat]]\n",
    "        y_T_peak[beat] = np.max(y_t)\n",
    "        x_T_peak[beat] = np.argmax(y_t) + t_start[beat]\n",
    "        t_pos_int[beat], t_neg_int[beat] = integral(y_t)\n",
    "    else:\n",
    "        y_T_peak[beat] = y[x_T_peak[beat]]\n",
    "        t_pos_int[beat], t_neg_int[beat] = 0,0\n",
    "        \n",
    "    ### for FFT\n",
    "#     y_beat = y[fixed_start[beat]:fixed_end[beat]]\n",
    "#     vect_fft, vect_power, vect_all_freq = power_fft(y_beat, timestep)\n",
    "#     temp = np.where((vect_all_freq>down_freq_lim) & (vect_all_freq<up_freq_lim))[0]\n",
    "#     f_keep = vect_power[temp]\n",
    "#     if beat==0:\n",
    "#         f = np.zeros((len(hb),len(f_keep)))\n",
    "\n",
    "#     f[beat, :len(f_keep)] = f_keep\n",
    "    \n",
    "t2 = time.time()\n",
    "print(\"the algorithm took\",round((t2-t1)/60,2),'minutes')  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "id": "a20d84b8",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "      <th>...</th>\n",
       "      <th>P_pos_int</th>\n",
       "      <th>P_neg_int</th>\n",
       "      <th>T_start</th>\n",
       "      <th>T_end</th>\n",
       "      <th>T_length</th>\n",
       "      <th>x_T_peak</th>\n",
       "      <th>y_T_peak</th>\n",
       "      <th>T_pos_int</th>\n",
       "      <th>T_neg_int</th>\n",
       "      <th>err</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1603</td>\n",
       "      <td>1700</td>\n",
       "      <td>1630</td>\n",
       "      <td>138</td>\n",
       "      <td>1631.0</td>\n",
       "      <td>0.022747</td>\n",
       "      <td>-5.421619</td>\n",
       "      <td>0.493310</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.296468</td>\n",
       "      <td>1645</td>\n",
       "      <td>1700</td>\n",
       "      <td>55</td>\n",
       "      <td>1666</td>\n",
       "      <td>0.163838</td>\n",
       "      <td>1.245540</td>\n",
       "      <td>5.121461</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1739</td>\n",
       "      <td>1843</td>\n",
       "      <td>1768</td>\n",
       "      <td>147</td>\n",
       "      <td>1769.0</td>\n",
       "      <td>0.022400</td>\n",
       "      <td>-5.391631</td>\n",
       "      <td>0.483091</td>\n",
       "      <td>...</td>\n",
       "      <td>0.292587</td>\n",
       "      <td>1.685074</td>\n",
       "      <td>1783</td>\n",
       "      <td>1843</td>\n",
       "      <td>60</td>\n",
       "      <td>1833</td>\n",
       "      <td>-0.008469</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.082809</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1886</td>\n",
       "      <td>1989</td>\n",
       "      <td>1915</td>\n",
       "      <td>146</td>\n",
       "      <td>1915.0</td>\n",
       "      <td>0.022183</td>\n",
       "      <td>-5.358755</td>\n",
       "      <td>0.475495</td>\n",
       "      <td>...</td>\n",
       "      <td>0.997703</td>\n",
       "      <td>1.459115</td>\n",
       "      <td>1930</td>\n",
       "      <td>1989</td>\n",
       "      <td>59</td>\n",
       "      <td>1988</td>\n",
       "      <td>0.048730</td>\n",
       "      <td>0.263409</td>\n",
       "      <td>4.748557</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3208</td>\n",
       "      <td>3310</td>\n",
       "      <td>3236</td>\n",
       "      <td>145</td>\n",
       "      <td>3236.0</td>\n",
       "      <td>0.022220</td>\n",
       "      <td>-5.284945</td>\n",
       "      <td>0.469724</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.872346</td>\n",
       "      <td>3251</td>\n",
       "      <td>3310</td>\n",
       "      <td>59</td>\n",
       "      <td>3266</td>\n",
       "      <td>0.138747</td>\n",
       "      <td>1.107563</td>\n",
       "      <td>4.883335</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>3353</td>\n",
       "      <td>3455</td>\n",
       "      <td>3381</td>\n",
       "      <td>145</td>\n",
       "      <td>3381.0</td>\n",
       "      <td>0.022099</td>\n",
       "      <td>-5.145189</td>\n",
       "      <td>0.454813</td>\n",
       "      <td>...</td>\n",
       "      <td>0.036477</td>\n",
       "      <td>1.844374</td>\n",
       "      <td>3396</td>\n",
       "      <td>3455</td>\n",
       "      <td>59</td>\n",
       "      <td>3441</td>\n",
       "      <td>0.031338</td>\n",
       "      <td>0.250444</td>\n",
       "      <td>7.208787</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30079</th>\n",
       "      <td>24</td>\n",
       "      <td>1816</td>\n",
       "      <td>847037</td>\n",
       "      <td>847128</td>\n",
       "      <td>847062</td>\n",
       "      <td>129</td>\n",
       "      <td>847019.0</td>\n",
       "      <td>-0.016930</td>\n",
       "      <td>3.502978</td>\n",
       "      <td>0.237228</td>\n",
       "      <td>...</td>\n",
       "      <td>0.757078</td>\n",
       "      <td>0.073372</td>\n",
       "      <td>847076</td>\n",
       "      <td>847128</td>\n",
       "      <td>52</td>\n",
       "      <td>847121</td>\n",
       "      <td>0.323828</td>\n",
       "      <td>2.190158</td>\n",
       "      <td>0.264696</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30080</th>\n",
       "      <td>24</td>\n",
       "      <td>1817</td>\n",
       "      <td>847166</td>\n",
       "      <td>847257</td>\n",
       "      <td>847191</td>\n",
       "      <td>129</td>\n",
       "      <td>847149.0</td>\n",
       "      <td>-0.021423</td>\n",
       "      <td>3.853764</td>\n",
       "      <td>0.330233</td>\n",
       "      <td>...</td>\n",
       "      <td>2.500342</td>\n",
       "      <td>0.005950</td>\n",
       "      <td>847204</td>\n",
       "      <td>847257</td>\n",
       "      <td>53</td>\n",
       "      <td>847249</td>\n",
       "      <td>0.215554</td>\n",
       "      <td>1.614151</td>\n",
       "      <td>0.094887</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30081</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>127</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "      <td>...</td>\n",
       "      <td>3.699306</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>847332</td>\n",
       "      <td>847385</td>\n",
       "      <td>53</td>\n",
       "      <td>847376</td>\n",
       "      <td>0.079204</td>\n",
       "      <td>2.192716</td>\n",
       "      <td>0.223051</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30082</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>118</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.749987</td>\n",
       "      <td>847611</td>\n",
       "      <td>847657</td>\n",
       "      <td>46</td>\n",
       "      <td>847617</td>\n",
       "      <td>0.217286</td>\n",
       "      <td>3.780919</td>\n",
       "      <td>6.167572</td>\n",
       "      <td>-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30083</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>141</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.922285</td>\n",
       "      <td>847852</td>\n",
       "      <td>847911</td>\n",
       "      <td>59</td>\n",
       "      <td>847904</td>\n",
       "      <td>0.205192</td>\n",
       "      <td>4.577482</td>\n",
       "      <td>3.439496</td>\n",
       "      <td>-7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30084 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     5    1603    1700       1630        138    1631.0   \n",
       "1            1     6    1739    1843       1768        147    1769.0   \n",
       "2            1     7    1886    1989       1915        146    1915.0   \n",
       "3            1    10    3208    3310       3236        145    3236.0   \n",
       "4            1    11    3353    3455       3381        145    3381.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "30079       24  1816  847037  847128     847062        129  847019.0   \n",
       "30080       24  1817  847166  847257     847191        129  847149.0   \n",
       "30081       24  1818  847295  847385     847320        127  847277.0   \n",
       "30082       24  1820  847574  847657     847597        118  847539.0   \n",
       "30083       24  1822  847811  847911     847839        141  847797.0   \n",
       "\n",
       "              a         v     Delta  ...  P_pos_int  P_neg_int  T_start  \\\n",
       "0      0.022747 -5.421619  0.493310  ...   0.000000   3.296468     1645   \n",
       "1      0.022400 -5.391631  0.483091  ...   0.292587   1.685074     1783   \n",
       "2      0.022183 -5.358755  0.475495  ...   0.997703   1.459115     1930   \n",
       "3      0.022220 -5.284945  0.469724  ...   0.000000   2.872346     3251   \n",
       "4      0.022099 -5.145189  0.454813  ...   0.036477   1.844374     3396   \n",
       "...         ...       ...       ...  ...        ...        ...      ...   \n",
       "30079 -0.016930  3.502978  0.237228  ...   0.757078   0.073372   847076   \n",
       "30080 -0.021423  3.853764  0.330233  ...   2.500342   0.005950   847204   \n",
       "30081 -0.021987  3.692612  0.324755  ...   3.699306   0.000000   847332   \n",
       "30082 -0.022921  4.593035  0.421116  ...   0.000000   0.749987   847611   \n",
       "30083 -0.033397  5.784591  0.772745  ...   0.000000   2.922285   847852   \n",
       "\n",
       "        T_end  T_length  x_T_peak  y_T_peak  T_pos_int  T_neg_int  err  \n",
       "0        1700        55      1666  0.163838   1.245540   5.121461   -1  \n",
       "1        1843        60      1833 -0.008469   0.000000   7.082809   -1  \n",
       "2        1989        59      1988  0.048730   0.263409   4.748557   -1  \n",
       "3        3310        59      3266  0.138747   1.107563   4.883335    0  \n",
       "4        3455        59      3441  0.031338   0.250444   7.208787   -1  \n",
       "...       ...       ...       ...       ...        ...        ...  ...  \n",
       "30079  847128        52    847121  0.323828   2.190158   0.264696    4  \n",
       "30080  847257        53    847249  0.215554   1.614151   0.094887    3  \n",
       "30081  847385        53    847376  0.079204   2.192716   0.223051    2  \n",
       "30082  847657        46    847617  0.217286   3.780919   6.167572   -6  \n",
       "30083  847911        59    847904  0.205192   4.577482   3.439496   -7  \n",
       "\n",
       "[30084 rows x 43 columns]"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# we update the dataframe with the newly extracted visualizable features \n",
    "hb['r_pos_int'] = r_pos_int\n",
    "hb['r_neg_int'] = r_neg_int\n",
    "hb['P_start'] = p_start.astype(int)\n",
    "hb['P_end'] = p_end.astype(int)\n",
    "hb['P_length'] = (p_end - p_start).astype(int)\n",
    "hb['x_P_peak'] = x_P_peak.astype(int)\n",
    "hb['y_P_peak'] = y_P_peak\n",
    "hb['P_pos_int'] = p_pos_int\n",
    "hb['P_neg_int'] = p_neg_int\n",
    "hb['T_start'] = t_start.astype(int)\n",
    "hb['T_end'] = t_end.astype(int)\n",
    "hb['T_length'] = (t_end - t_start).astype(int)\n",
    "hb['x_T_peak'] = x_T_peak.astype(int)\n",
    "hb['y_T_peak'] = y_T_peak\n",
    "hb['T_pos_int'] = t_pos_int\n",
    "hb['T_neg_int'] = t_neg_int\n",
    "hb['err'] = hb['ann_index'] - hb['x_R_peak']\n",
    "\n",
    "hb.to_csv('output_files/heartbeats_full_absolute.csv', index=False)\n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "77218b7f",
   "metadata": {},
   "source": [
    "### Looking at some of the visualizable features\n",
    "\n",
    "To further check the behaviour of the two r-peak finding algorithm that we use (one in the beginning, one in feature extraction)\n",
    "we can take a look at the the difference (err) between the two estimates of the position of the R peak. \n",
    "\n",
    "If the two algorithms work well together, we expect the distirbution to be centered around 0 and be quite tight around it. Different behaviours could be caused by a bias (i.e. average shifted from 0) or by the algorithm not being very precise (i.e. wider distribution/standard deviation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "id": "5120e0dc",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1200x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "distribution(np.array(hb['err']), -5, 5,title='err')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1aa50cca",
   "metadata": {},
   "source": [
    "The R-peak finding algorithm and local minima/maxima algorithm return very similar results."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9f044a26",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Preparing extracted features for ML\n",
    "Finally, we can adapt the features extracted to be relative to the heartbeat (i.e. not absolute to the entire dataset) but also absolute in terms of time (since future datasets might have different sampling frequencies)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "id": "8a6c343f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>patient</th>\n",
       "      <th>idx</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>ann_index</th>\n",
       "      <th>RR_length</th>\n",
       "      <th>i</th>\n",
       "      <th>a</th>\n",
       "      <th>v</th>\n",
       "      <th>Delta</th>\n",
       "      <th>...</th>\n",
       "      <th>P_pos_int</th>\n",
       "      <th>P_neg_int</th>\n",
       "      <th>T_start</th>\n",
       "      <th>T_end</th>\n",
       "      <th>T_length</th>\n",
       "      <th>x_T_peak</th>\n",
       "      <th>y_T_peak</th>\n",
       "      <th>T_pos_int</th>\n",
       "      <th>T_neg_int</th>\n",
       "      <th>err</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>1603</td>\n",
       "      <td>1700</td>\n",
       "      <td>1630</td>\n",
       "      <td>0.552</td>\n",
       "      <td>1631.0</td>\n",
       "      <td>0.022747</td>\n",
       "      <td>-5.421619</td>\n",
       "      <td>0.493310</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.013186</td>\n",
       "      <td>0.168</td>\n",
       "      <td>0.388</td>\n",
       "      <td>0.220</td>\n",
       "      <td>0.252</td>\n",
       "      <td>0.163838</td>\n",
       "      <td>0.004982</td>\n",
       "      <td>0.020486</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>1739</td>\n",
       "      <td>1843</td>\n",
       "      <td>1768</td>\n",
       "      <td>0.588</td>\n",
       "      <td>1769.0</td>\n",
       "      <td>0.022400</td>\n",
       "      <td>-5.391631</td>\n",
       "      <td>0.483091</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001170</td>\n",
       "      <td>0.006740</td>\n",
       "      <td>0.176</td>\n",
       "      <td>0.416</td>\n",
       "      <td>0.240</td>\n",
       "      <td>0.376</td>\n",
       "      <td>-0.008469</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.028331</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>1886</td>\n",
       "      <td>1989</td>\n",
       "      <td>1915</td>\n",
       "      <td>0.584</td>\n",
       "      <td>1915.0</td>\n",
       "      <td>0.022183</td>\n",
       "      <td>-5.358755</td>\n",
       "      <td>0.475495</td>\n",
       "      <td>...</td>\n",
       "      <td>0.003991</td>\n",
       "      <td>0.005836</td>\n",
       "      <td>0.176</td>\n",
       "      <td>0.412</td>\n",
       "      <td>0.236</td>\n",
       "      <td>0.408</td>\n",
       "      <td>0.048730</td>\n",
       "      <td>0.001054</td>\n",
       "      <td>0.018994</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>3208</td>\n",
       "      <td>3310</td>\n",
       "      <td>3236</td>\n",
       "      <td>0.580</td>\n",
       "      <td>3236.0</td>\n",
       "      <td>0.022220</td>\n",
       "      <td>-5.284945</td>\n",
       "      <td>0.469724</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.011489</td>\n",
       "      <td>0.172</td>\n",
       "      <td>0.408</td>\n",
       "      <td>0.236</td>\n",
       "      <td>0.232</td>\n",
       "      <td>0.138747</td>\n",
       "      <td>0.004430</td>\n",
       "      <td>0.019533</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>11</td>\n",
       "      <td>3353</td>\n",
       "      <td>3455</td>\n",
       "      <td>3381</td>\n",
       "      <td>0.580</td>\n",
       "      <td>3381.0</td>\n",
       "      <td>0.022099</td>\n",
       "      <td>-5.145189</td>\n",
       "      <td>0.454813</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000146</td>\n",
       "      <td>0.007377</td>\n",
       "      <td>0.172</td>\n",
       "      <td>0.408</td>\n",
       "      <td>0.236</td>\n",
       "      <td>0.352</td>\n",
       "      <td>0.031338</td>\n",
       "      <td>0.001002</td>\n",
       "      <td>0.028835</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30037</th>\n",
       "      <td>24</td>\n",
       "      <td>1816</td>\n",
       "      <td>847037</td>\n",
       "      <td>847128</td>\n",
       "      <td>847062</td>\n",
       "      <td>0.516</td>\n",
       "      <td>847019.0</td>\n",
       "      <td>-0.016930</td>\n",
       "      <td>3.502978</td>\n",
       "      <td>0.237228</td>\n",
       "      <td>...</td>\n",
       "      <td>0.003028</td>\n",
       "      <td>0.000293</td>\n",
       "      <td>0.156</td>\n",
       "      <td>0.364</td>\n",
       "      <td>0.208</td>\n",
       "      <td>0.336</td>\n",
       "      <td>0.323828</td>\n",
       "      <td>0.008761</td>\n",
       "      <td>0.001059</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30038</th>\n",
       "      <td>24</td>\n",
       "      <td>1817</td>\n",
       "      <td>847166</td>\n",
       "      <td>847257</td>\n",
       "      <td>847191</td>\n",
       "      <td>0.516</td>\n",
       "      <td>847149.0</td>\n",
       "      <td>-0.021423</td>\n",
       "      <td>3.853764</td>\n",
       "      <td>0.330233</td>\n",
       "      <td>...</td>\n",
       "      <td>0.010001</td>\n",
       "      <td>0.000024</td>\n",
       "      <td>0.152</td>\n",
       "      <td>0.364</td>\n",
       "      <td>0.212</td>\n",
       "      <td>0.332</td>\n",
       "      <td>0.215554</td>\n",
       "      <td>0.006457</td>\n",
       "      <td>0.000380</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30039</th>\n",
       "      <td>24</td>\n",
       "      <td>1818</td>\n",
       "      <td>847295</td>\n",
       "      <td>847385</td>\n",
       "      <td>847320</td>\n",
       "      <td>0.508</td>\n",
       "      <td>847277.0</td>\n",
       "      <td>-0.021987</td>\n",
       "      <td>3.692612</td>\n",
       "      <td>0.324755</td>\n",
       "      <td>...</td>\n",
       "      <td>0.014797</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.148</td>\n",
       "      <td>0.360</td>\n",
       "      <td>0.212</td>\n",
       "      <td>0.324</td>\n",
       "      <td>0.079204</td>\n",
       "      <td>0.008771</td>\n",
       "      <td>0.000892</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30040</th>\n",
       "      <td>24</td>\n",
       "      <td>1820</td>\n",
       "      <td>847574</td>\n",
       "      <td>847657</td>\n",
       "      <td>847597</td>\n",
       "      <td>0.472</td>\n",
       "      <td>847539.0</td>\n",
       "      <td>-0.022921</td>\n",
       "      <td>4.593035</td>\n",
       "      <td>0.421116</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.003000</td>\n",
       "      <td>0.148</td>\n",
       "      <td>0.332</td>\n",
       "      <td>0.184</td>\n",
       "      <td>0.172</td>\n",
       "      <td>0.217286</td>\n",
       "      <td>0.015124</td>\n",
       "      <td>0.024670</td>\n",
       "      <td>-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30041</th>\n",
       "      <td>24</td>\n",
       "      <td>1822</td>\n",
       "      <td>847811</td>\n",
       "      <td>847911</td>\n",
       "      <td>847839</td>\n",
       "      <td>0.564</td>\n",
       "      <td>847797.0</td>\n",
       "      <td>-0.033397</td>\n",
       "      <td>5.784591</td>\n",
       "      <td>0.772745</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.011689</td>\n",
       "      <td>0.164</td>\n",
       "      <td>0.400</td>\n",
       "      <td>0.236</td>\n",
       "      <td>0.372</td>\n",
       "      <td>0.205192</td>\n",
       "      <td>0.018310</td>\n",
       "      <td>0.013758</td>\n",
       "      <td>-7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30042 rows × 43 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       patient   idx   start     end  ann_index  RR_length         i  \\\n",
       "0            1     5    1603    1700       1630      0.552    1631.0   \n",
       "1            1     6    1739    1843       1768      0.588    1769.0   \n",
       "2            1     7    1886    1989       1915      0.584    1915.0   \n",
       "3            1    10    3208    3310       3236      0.580    3236.0   \n",
       "4            1    11    3353    3455       3381      0.580    3381.0   \n",
       "...        ...   ...     ...     ...        ...        ...       ...   \n",
       "30037       24  1816  847037  847128     847062      0.516  847019.0   \n",
       "30038       24  1817  847166  847257     847191      0.516  847149.0   \n",
       "30039       24  1818  847295  847385     847320      0.508  847277.0   \n",
       "30040       24  1820  847574  847657     847597      0.472  847539.0   \n",
       "30041       24  1822  847811  847911     847839      0.564  847797.0   \n",
       "\n",
       "              a         v     Delta  ...  P_pos_int  P_neg_int  T_start  \\\n",
       "0      0.022747 -5.421619  0.493310  ...   0.000000   0.013186    0.168   \n",
       "1      0.022400 -5.391631  0.483091  ...   0.001170   0.006740    0.176   \n",
       "2      0.022183 -5.358755  0.475495  ...   0.003991   0.005836    0.176   \n",
       "3      0.022220 -5.284945  0.469724  ...   0.000000   0.011489    0.172   \n",
       "4      0.022099 -5.145189  0.454813  ...   0.000146   0.007377    0.172   \n",
       "...         ...       ...       ...  ...        ...        ...      ...   \n",
       "30037 -0.016930  3.502978  0.237228  ...   0.003028   0.000293    0.156   \n",
       "30038 -0.021423  3.853764  0.330233  ...   0.010001   0.000024    0.152   \n",
       "30039 -0.021987  3.692612  0.324755  ...   0.014797   0.000000    0.148   \n",
       "30040 -0.022921  4.593035  0.421116  ...   0.000000   0.003000    0.148   \n",
       "30041 -0.033397  5.784591  0.772745  ...   0.000000   0.011689    0.164   \n",
       "\n",
       "       T_end  T_length  x_T_peak  y_T_peak  T_pos_int  T_neg_int  err  \n",
       "0      0.388     0.220     0.252  0.163838   0.004982   0.020486   -1  \n",
       "1      0.416     0.240     0.376 -0.008469   0.000000   0.028331   -1  \n",
       "2      0.412     0.236     0.408  0.048730   0.001054   0.018994   -1  \n",
       "3      0.408     0.236     0.232  0.138747   0.004430   0.019533    0  \n",
       "4      0.408     0.236     0.352  0.031338   0.001002   0.028835   -1  \n",
       "...      ...       ...       ...       ...        ...        ...  ...  \n",
       "30037  0.364     0.208     0.336  0.323828   0.008761   0.001059    4  \n",
       "30038  0.364     0.212     0.332  0.215554   0.006457   0.000380    3  \n",
       "30039  0.360     0.212     0.324  0.079204   0.008771   0.000892    2  \n",
       "30040  0.332     0.184     0.172  0.217286   0.015124   0.024670   -6  \n",
       "30041  0.400     0.236     0.372  0.205192   0.018310   0.013758   -7  \n",
       "\n",
       "[30042 rows x 43 columns]"
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# List of column needed substraction and calculation change to temporal domain\n",
    "first_change = ['QRS_start', 'QRS_end', 'x_R_peak','P_start', 'P_end', 'x_P_peak', 'T_start', 'T_end', 'x_T_peak']\n",
    "# List for column needed calculation change calculation to temporal domain\n",
    "second_change = ['RR_length', 'QRS_length', 'pos_int', 'neg_int', 'arg_max', 'arg_min',  'r_pos_int', 'r_neg_int', \n",
    "                 'P_length',  'P_pos_int', 'P_neg_int', 'T_length', 'T_pos_int', 'T_neg_int']\n",
    "\n",
    "for col in first_change:\n",
    "    hb[col] = (hb[col] - hb['start'])/new_freq \n",
    "\n",
    "for col in second_change:\n",
    "    hb[col] = hb[col] /new_freq \n",
    "    \n",
    "hb.to_csv('output_files/heartbeats_full_relative.csv', index=False)\n",
    "hb"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe15a4a3",
   "metadata": {},
   "source": [
    "We can also remove the quantities that won't be ML features since they cannot be relavive to the heartbeat in any way.\n",
    "Lastly, we convert the dataframe into a numpy matrix, which is put side by side to the matrix with the DWT coefficients. \\\n",
    "In the ML features matrix, each row correspond to an heartbeat."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "id": "892196b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "features_names = list(hb.columns)\n",
    "no_features = ['patient', 'idx', 'start', 'end', 'ann_index']\n",
    "for nf in no_features:\n",
    "    features_names.remove(nf)\n",
    "\n",
    "hand_extracted_features = hb[features_names]\n",
    "hand_extracted_features = hand_extracted_features.to_numpy()\n",
    "ML_input = np.hstack((features, hand_extracted_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "id": "5dc10332",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savetxt('output_files/ML_input.txt',ML_input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bd54ab05",
   "metadata": {},
   "outputs": [],
   "source": [
    "ML_input = np.loadtxt('output_files/ML_input.txt')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3eecef3",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 5) Machine Learning"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "311edc61-51da-4ff3-b7e3-fa30fca97671",
   "metadata": {},
   "source": [
    "For this dataset, we just need to load the trained model (from ECG_PhysioNet) and have it predict based on the features obtained before:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "007c54c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_xgb_2 = XGBClassifier()\n",
    "model_xgb_2.load_model(\"multi_maxdepth5_nwindow3.json\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e36e89d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "num = 2072053\n",
    "np.random.seed(num)\n",
    "X = ML_input\n",
    "Y_pred = model_xgb_2.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "2638321b",
   "metadata": {},
   "outputs": [],
   "source": [
    "labels, freqs = np.unique(Y_pred, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a1d055d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 1] [29937   105]\n"
     ]
    }
   ],
   "source": [
    "print(labels, freqs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "ea5eaf06",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of sick beats to total beats classifed: 0.3495 %\n"
     ]
    }
   ],
   "source": [
    "print('Percentage of sick beats to total beats classifed:', np.round(freqs[1]/sum(freqs)*100,4), '%')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a25019c6",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# 6) Saving output for plotly display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "9936bb55",
   "metadata": {},
   "outputs": [],
   "source": [
    "hbb = copy.copy(hb)\n",
    "def change_values_vectorized(array):\n",
    "    array = np.where(array==0, 'N', array)\n",
    "    array = np.where(array=='1', 'V', array)\n",
    "    array = np.where(array=='2', 'R', array)\n",
    "    array = np.where(array=='3', 'A', array)\n",
    "    array = np.where(array=='4', 'other', array)\n",
    "    return array\n",
    "\n",
    "Y_label = change_values_vectorized(Y_pred)\n",
    "\n",
    "hbb['ann_symbol']=Y_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "9a15b2bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "hbb.to_csv('output_files/everything.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2040dfb",
   "metadata": {},
   "source": [
    "# References\n",
    "[0] St Petersburg INCART 12-lead Arrhythmia Database. https://physionet.org/content/incartdb/1.0.0/\n",
    "\n",
    "[1] A survey of heart anomaly detection using ambulatory electrocardiogram (ECG). Hongzu  Li and Pierre Boulanger. 2020\n",
    "\n",
    "[2] ECG signal classification using support vector machine based on wavelet multiresolution analysis. Ayman Rabee and Imad Barhumi. 2012.\n",
    "\n",
    "[3] Segmentation of the ECG Signal by Means of a Linear Regression Algorithm. Javier Aspuru et Al. 2019"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}