LabelTimeSpectroGenerator.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: window-slider in c:\\users\\hp\\anaconda3\\lib\\site-packages (0.8)\n"
     ]
    }
   ],
   "source": [
    "!pip install window-slider"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import csv\n",
    "from scipy.stats import kurtosis,skew,iqr\n",
    "from scipy import signal\n",
    "from window_slider import Slider\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "acc = '_acc'\n",
    "gyro = '_gyro'\n",
    "source = 'E:/games/IISc-May-2019/HAPT Data Set/MyLab/'\n",
    "dest = 'E:/games/IISc-May-2019/HAPT Data Set/TimeSpectroSet/60-60/'\n",
    "fs =50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(231, 30)\n",
      "(231, 30)\n",
      "(151, 30)\n",
      "(151, 30)\n",
      "(137, 30)\n",
      "(137, 30)\n",
      "(238, 30)\n",
      "(238, 30)\n",
      "(274, 30)\n",
      "(274, 30)\n",
      "(261, 30)\n",
      "(261, 30)\n",
      "(40, 30)\n",
      "(40, 30)\n",
      "(30, 30)\n",
      "(30, 30)\n",
      "(48, 30)\n",
      "(48, 30)\n",
      "(42, 30)\n",
      "(42, 30)\n",
      "(56, 30)\n",
      "(56, 30)\n",
      "(42, 30)\n",
      "(42, 30)\n"
     ]
    }
   ],
   "source": [
    "for i in range(1,13):\n",
    "    accfile = source + str(i) + acc + '.csv'\n",
    "    gyrofile = source + str(i) + gyro + '.csv'\n",
    "    df1 = pd.read_csv(accfile)\n",
    "    df2 = pd.read_csv(gyrofile)\n",
    "    ax = df1['ax']\n",
    "    ay = df1['ay']\n",
    "    az = df1['az']\n",
    "    wx = df2['wx']\n",
    "    wy = df2['wy']\n",
    "    wz = df2['wz']\n",
    "    #------------------------------------------------------------------------------------------------------------------- \n",
    "    #Time domain\n",
    "    timeAcc = np.empty([0,30])\n",
    "    timeGyro = np.empty([0,30])\n",
    "\n",
    "    size = 512\n",
    "    overlap = 256\n",
    "\n",
    "    \n",
    "    for k in range(2):\n",
    "        slider1 = Slider(size, overlap)\n",
    "        slider2 = Slider(size, overlap)\n",
    "        slider3 = Slider(size, overlap)\n",
    "        if(k == 0):\n",
    "            slider1.fit(ax)\n",
    "            slider2.fit(ay)\n",
    "            slider3.fit(az)\n",
    "        elif(k == 1):\n",
    "            slider1.fit(wx)\n",
    "            slider2.fit(wy)\n",
    "            slider3.fit(wz)        \n",
    "        while True:\n",
    "            window_data1 = slider1.slide()\n",
    "            window_data2 = slider2.slide()\n",
    "            window_data3 = slider3.slide()\n",
    "\n",
    "            # do your stuff\n",
    "            mean1 = np.mean(window_data1)\n",
    "            variance1 = np.var(window_data1)\n",
    "            median1 = np.median(window_data1)\n",
    "            rms1 = np.sqrt(np.mean(window_data1**2))\n",
    "            skewdness1 = skew(window_data1)\n",
    "            kurtos1 = kurtosis(window_data1)\n",
    "            ptp1 = np.ptp(window_data1)\n",
    "            rang1 = np.max(window_data1) - np.min(window_data1)\n",
    "            interq_range1 = iqr(window_data1)\n",
    "            crest_factor1 = np.max(window_data1)/rms1\n",
    "\n",
    "            mean2 = np.mean(window_data2)\n",
    "            variance2 = np.var(window_data2)\n",
    "            median2 = np.median(window_data2)\n",
    "            rms2 = np.sqrt(np.mean(window_data2**2))\n",
    "            skewdness2 = skew(window_data2)\n",
    "            kurtos2 = kurtosis(window_data2)\n",
    "            ptp2 = np.ptp(window_data2)\n",
    "            rang2 = np.max(window_data2) - np.min(window_data2)\n",
    "            interq_range2 = iqr(window_data2)\n",
    "            crest_factor2 = np.max(window_data2)/rms2\n",
    "\n",
    "            mean3 = np.mean(window_data3)\n",
    "            variance3 = np.var(window_data3)\n",
    "            median3 = np.median(window_data3)\n",
    "            rms3 = np.sqrt(np.mean(window_data3**2))\n",
    "            skewdness3 = skew(window_data3)\n",
    "            kurtos3 = kurtosis(window_data3)\n",
    "            ptp3 = np.ptp(window_data3)\n",
    "            rang3 = np.max(window_data3) - np.min(window_data3)\n",
    "            interq_range3 = iqr(window_data3)\n",
    "            crest_factor3 = np.max(window_data3)/rms3\n",
    "\n",
    "            time_dom = np.array([mean1,mean2,mean3,variance1,variance2,variance3,median1,median2,median3,\n",
    "                                 rms1,rms2,rms3,skewdness1,skewdness2,skewdness3,kurtos1,kurtos2,kurtos3,\n",
    "                                 ptp1,ptp2,ptp3,rang1,rang2,rang3,interq_range1,interq_range2,interq_range3, \n",
    "                                 crest_factor1,crest_factor2,crest_factor3])\n",
    "            if( k == 0 ):\n",
    "                timeAcc = np.append(timeAcc, [time_dom], axis=0)\n",
    "            elif ( k == 1):\n",
    "                timeGyro = np.append(timeGyro, [time_dom], axis=0)\n",
    "\n",
    "            if slider1.reached_end_of_list(): break\n",
    "\n",
    "#     print(timeAcc.shape)\n",
    "#     print(timeGyro.shape)\n",
    "            \n",
    "        \n",
    "    \n",
    "    #Frequency Domain\n",
    "    Afx, Atx, Ax = signal.spectrogram(ax, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    Afy, Aty, Ay = signal.spectrogram(ay, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    Afz, Atz, Az = signal.spectrogram(az, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    \n",
    "    Wfx, Wtx, Wx = signal.spectrogram(wx, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    Wfy, Wty, Wy = signal.spectrogram(wy, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    Wfz, Wtz, Wz = signal.spectrogram(wz, fs=fs,window=signal.get_window('hann',512),noverlap=256,nperseg=512)\n",
    "    \n",
    "    Ax = Ax.T\n",
    "    Ay = Ay.T\n",
    "    Az = Az.T\n",
    "    Wx = Wx.T\n",
    "    Wy = Wy.T\n",
    "    Wz = Wz.T\n",
    "#     print(Ax.shape)\n",
    "    resA = np.concatenate((Ax, Ay, Az), axis = 1)\n",
    "    resW = np.concatenate((Wx, Wy, Wz), axis = 1)\n",
    "#     print(resA.shape)\n",
    "    \n",
    "    for j in range(resA.shape[0]):\n",
    "        sA = sorted(resA[j])\n",
    "        sW = sorted(resW[j])\n",
    "        featureAcc = sA[0:15] + sA[resA.shape[1]-15:resA.shape[1]]    #For 100 features set 50 to 25\n",
    "        featureAcc.extend( timeAcc[j] )\n",
    "        featureGyr = sW[0:15] + sW[resA.shape[1]-15:resA.shape[1]]    #For 100 features set 50 to 25\n",
    "        featureGyr.extend( timeGyro[j] )\n",
    "        final = np.concatenate((featureAcc,featureGyr))\n",
    "        final = np.append( final, i )\n",
    "#         print(final)\n",
    "        with open (dest+str(i)+'_Time_SpectroFeatures.csv','a',newline='') as out_file:\n",
    "            writer = csv.writer(out_file)\n",
    "            writer.writerow(final)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from window_slider import Slider\n",
    "# import numpy\n",
    "# list = numpy.array([0, 1, 2, 3, 4, 5, 6, 7])\n",
    "# bucket_size = 3\n",
    "# overlap_count = 1\n",
    "# slider = Slider(bucket_size,overlap_count)\n",
    "# slider.fit(list)       \n",
    "# while True:\n",
    "#     window_data = slider.slide()\n",
    "#     # do your stuff\n",
    "#     print(window_data)\n",
    "#     if slider.reached_end_of_list(): break"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}