Emotion Recognition - YOLO.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "5f3ef90b",
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "module 'numpy' has no attribute 'typeDict'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "File \u001b[1;32mC:\\ProgramData\\Anaconda3\\envs\\cv_open\\lib\\site-packages\\numpy\\__init__.py:320\u001b[0m, in \u001b[0;36m__getattr__\u001b[1;34m(attr)\u001b[0m\n\u001b[0;32m    317\u001b[0m     \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mtesting\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Tester\n\u001b[0;32m    318\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m Tester\n\u001b[1;32m--> 320\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mAttributeError\u001b[39;00m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mmodule \u001b[39m\u001b[38;5;132;01m{!r}\u001b[39;00m\u001b[38;5;124m has no attribute \u001b[39m\u001b[38;5;124m\"\u001b[39m\n\u001b[0;32m    321\u001b[0m                      \u001b[38;5;124m\"\u001b[39m\u001b[38;5;132;01m{!r}\u001b[39;00m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;241m.\u001b[39mformat(\u001b[38;5;18m__name__\u001b[39m, attr))\n",
      "\u001b[1;31mAttributeError\u001b[0m: module 'numpy' has no attribute 'typeDict'"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\envs\\cv_open\\lib\\site-packages\\requests\\__init__.py:109: RequestsDependencyWarning: urllib3 (1.26.15) or chardet (None)/charset_normalizer (3.1.0) doesn't match a supported version!\n",
      "  warnings.warn(\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n",
      "WARNING:root:Limited tf.compat.v2.summary API due to missing TensorBoard installation.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Frames per second :  29.97002997002997 FPS\n",
      "Frame count :  93268.0\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>File Name</th>\n",
       "      <th>Start time</th>\n",
       "      <th>End time</th>\n",
       "      <th>session_name</th>\n",
       "      <th># of students in video</th>\n",
       "      <th>which assignments</th>\n",
       "      <th># of students with wrong answers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>6.25</td>\n",
       "      <td>9.00</td>\n",
       "      <td>Information Overload &amp; Solutions for Informati...</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>9.22</td>\n",
       "      <td>10.09</td>\n",
       "      <td>Chunking</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>10.10</td>\n",
       "      <td>10.29</td>\n",
       "      <td>Omission</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>10.30</td>\n",
       "      <td>10.51</td>\n",
       "      <td>Queuing</td>\n",
       "      <td>13</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>10.52</td>\n",
       "      <td>11.17</td>\n",
       "      <td>Capitulation</td>\n",
       "      <td>16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>11.40</td>\n",
       "      <td>12.19</td>\n",
       "      <td>Filtering &amp; Giving examples about who uses RecSys</td>\n",
       "      <td>16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>16.20</td>\n",
       "      <td>19.45</td>\n",
       "      <td>RecSys</td>\n",
       "      <td>17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>19.46</td>\n",
       "      <td>22.14</td>\n",
       "      <td>How RecSys works (Netflix Example)</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>22.15</td>\n",
       "      <td>24.20</td>\n",
       "      <td>How it works</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>24.22</td>\n",
       "      <td>28.07</td>\n",
       "      <td>Traditional Recommendation Algs</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>28.33</td>\n",
       "      <td>29.47</td>\n",
       "      <td>Context-Aware Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>29.48</td>\n",
       "      <td>31.00</td>\n",
       "      <td>Multi-Criteria Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>31.01</td>\n",
       "      <td>32.00</td>\n",
       "      <td>Group Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>32.02</td>\n",
       "      <td>32.27</td>\n",
       "      <td>Cross-Domain Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>32.29</td>\n",
       "      <td>33.20</td>\n",
       "      <td>Human Factor Based Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>33.22</td>\n",
       "      <td>35.07</td>\n",
       "      <td>Multi-Stakeholder Recommendations</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>36.11</td>\n",
       "      <td>36.47</td>\n",
       "      <td>Recommendations Tasks</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>36.50</td>\n",
       "      <td>40.29</td>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>18</td>\n",
       "      <td>hw6</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>40.29</td>\n",
       "      <td>46.13</td>\n",
       "      <td>Top-N Recommendation</td>\n",
       "      <td>18</td>\n",
       "      <td>hw 6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>46.16</td>\n",
       "      <td>47.26</td>\n",
       "      <td>Relevance &amp; Ranking Metrics (Intro)</td>\n",
       "      <td>18</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>47.26</td>\n",
       "      <td>51.52</td>\n",
       "      <td>Relevance Metrics</td>\n",
       "      <td>18</td>\n",
       "      <td>hw2, hw6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          File Name  Start time  End time  \\\n",
       "0   ITMD 525_03.MOV        6.25      9.00   \n",
       "1   ITMD 525_03.MOV        9.22     10.09   \n",
       "2   ITMD 525_03.MOV       10.10     10.29   \n",
       "3   ITMD 525_03.MOV       10.30     10.51   \n",
       "4   ITMD 525_03.MOV       10.52     11.17   \n",
       "5   ITMD 525_03.MOV       11.40     12.19   \n",
       "6   ITMD 525_03.MOV       16.20     19.45   \n",
       "7   ITMD 525_03.MOV       19.46     22.14   \n",
       "8   ITMD 525_03.MOV       22.15     24.20   \n",
       "9   ITMD 525_03.MOV       24.22     28.07   \n",
       "10  ITMD 525_03.MOV       28.33     29.47   \n",
       "11  ITMD 525_03.MOV       29.48     31.00   \n",
       "12  ITMD 525_03.MOV       31.01     32.00   \n",
       "13  ITMD 525_03.MOV       32.02     32.27   \n",
       "14  ITMD 525_03.MOV       32.29     33.20   \n",
       "15  ITMD 525_03.MOV       33.22     35.07   \n",
       "16  ITMD 525_03.MOV       36.11     36.47   \n",
       "17  ITMD 525_03.MOV       36.50     40.29   \n",
       "18  ITMD 525_03.MOV       40.29     46.13   \n",
       "19  ITMD 525_03.MOV       46.16     47.26   \n",
       "20  ITMD 525_03.MOV       47.26     51.52   \n",
       "\n",
       "                                         session_name  # of students in video  \\\n",
       "0   Information Overload & Solutions for Informati...                      13   \n",
       "1                                            Chunking                      13   \n",
       "2                                            Omission                      13   \n",
       "3                                             Queuing                      13   \n",
       "4                                        Capitulation                      16   \n",
       "5   Filtering & Giving examples about who uses RecSys                      16   \n",
       "6                                              RecSys                      17   \n",
       "7                  How RecSys works (Netflix Example)                      18   \n",
       "8                                        How it works                      18   \n",
       "9                     Traditional Recommendation Algs                      18   \n",
       "10                      Context-Aware Recommendations                      18   \n",
       "11                     Multi-Criteria Recommendations                      18   \n",
       "12                              Group Recommendations                      18   \n",
       "13                       Cross-Domain Recommendations                      18   \n",
       "14                 Human Factor Based Recommendations                      18   \n",
       "15                  Multi-Stakeholder Recommendations                      18   \n",
       "16                              Recommendations Tasks                      18   \n",
       "17                                  Rating Prediction                      18   \n",
       "18                               Top-N Recommendation                      18   \n",
       "19                Relevance & Ranking Metrics (Intro)                      18   \n",
       "20                                  Relevance Metrics                      18   \n",
       "\n",
       "   which assignments  # of students with wrong answers  \n",
       "0                NaN                               NaN  \n",
       "1                NaN                               NaN  \n",
       "2                NaN                               NaN  \n",
       "3                NaN                               NaN  \n",
       "4                NaN                               NaN  \n",
       "5                NaN                               NaN  \n",
       "6                NaN                               NaN  \n",
       "7                NaN                               NaN  \n",
       "8                NaN                               NaN  \n",
       "9                NaN                               NaN  \n",
       "10               NaN                               NaN  \n",
       "11               NaN                               NaN  \n",
       "12               NaN                               NaN  \n",
       "13               NaN                               NaN  \n",
       "14               NaN                               NaN  \n",
       "15               NaN                               NaN  \n",
       "16               NaN                               NaN  \n",
       "17               hw6                               8.0  \n",
       "18              hw 6                               3.0  \n",
       "19               NaN                               NaN  \n",
       "20          hw2, hw6                               3.0  "
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import math, time\n",
    "import tensorflow\n",
    "import cv2\n",
    "from keras.models import load_model\n",
    "from time import time\n",
    "from datetime import datetime\n",
    "from PIL import Image\n",
    "from IPython.display import display, HTML\n",
    "from tqdm import tqdm\n",
    "import itertools as it\n",
    "from IPython.display import display, HTML\n",
    "from matplotlib import pyplot as plt\n",
    "from scipy.stats import mode\n",
    "import warnings \n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "\n",
    "\n",
    "filename=\"ITMD 525_03.MOV\"   #filename\n",
    "cap = cv2.VideoCapture(filename) #Read Frame\n",
    "fps=cap.get(cv2.CAP_PROP_FPS) #Extract the frame per second (fps) \n",
    "print('Frames per second : ', fps,'FPS')\n",
    "frame_count = cap.get(cv2.CAP_PROP_FRAME_COUNT)#get overall no# of frames in the video file\n",
    "print('Frame count : ', frame_count)\n",
    "sessions=pd.read_excel('session_info.xlsx', sheet_name='Sheet2')\n",
    "sessions=sessions.rename(columns=lambda x:x.strip())\n",
    "ses_crop=sessions.iloc[:, 0:7]#cropping the session info table for columns of interest \n",
    "ses_crop\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0a763fdb",
   "metadata": {},
   "outputs": [],
   "source": [
    "#selecting only sessions for which we have homework values \n",
    "ses_crop=ses_crop[np.isnan(ses_crop['# of students with wrong answers'])==False]\n",
    "ses_crop.reset_index(drop=True, inplace=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "64dba826",
   "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>File Name</th>\n",
       "      <th>Start time</th>\n",
       "      <th>End time</th>\n",
       "      <th>session_name</th>\n",
       "      <th># of students in video</th>\n",
       "      <th>which assignments</th>\n",
       "      <th># of students with wrong answers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>36.50</td>\n",
       "      <td>40.29</td>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>18</td>\n",
       "      <td>hw6</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>40.29</td>\n",
       "      <td>46.13</td>\n",
       "      <td>Top-N Recommendation</td>\n",
       "      <td>18</td>\n",
       "      <td>hw 6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>47.26</td>\n",
       "      <td>51.52</td>\n",
       "      <td>Relevance Metrics</td>\n",
       "      <td>18</td>\n",
       "      <td>hw2, hw6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         File Name  Start time  End time          session_name  \\\n",
       "0  ITMD 525_03.MOV       36.50     40.29     Rating Prediction   \n",
       "1  ITMD 525_03.MOV       40.29     46.13  Top-N Recommendation   \n",
       "2  ITMD 525_03.MOV       47.26     51.52     Relevance Metrics   \n",
       "\n",
       "   # of students in video which assignments  # of students with wrong answers  \n",
       "0                      18               hw6                               8.0  \n",
       "1                      18              hw 6                               3.0  \n",
       "2                      18          hw2, hw6                               3.0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ses_crop.shape\n",
    "ses_crop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3ac4e74d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "start time in disected form: [(0.5, 36.0), (0.28999999999999915, 40.0), (0.259999999999998, 47.0)]\n",
      "\n",
      "\n",
      "Start time in Time Format: ['00:36:50', '00:40:29', '00:47:26']\n",
      "\n",
      "End time in disected form: [(0.28999999999999915, 40.0), (0.13000000000000256, 46.0), (0.5200000000000031, 51.0)]\n",
      "\n",
      "\n",
      "End time in Time Format: ['00:40:29', '00:46:13', '00:51:52']\n"
     ]
    }
   ],
   "source": [
    "#let's convert the start time and end time columns into proper time format i.e \"%H:%M:%S\"\n",
    "from time import time\n",
    "import math, time\n",
    "\n",
    "starttime=[]\n",
    "for i in range(len(ses_crop)):\n",
    "    starttime.append(math.modf(ses_crop['Start time'][i])) # (0.5678000000000338, 1234.0)\n",
    "print('\\nstart time in disected form:', starttime)\n",
    "    \n",
    "seconds_s=[round(x[0]*100) for x in starttime]\n",
    "secondsfromminutes_s=[x[1]*60 for x in starttime]\n",
    "\n",
    "total_seconds_s=[]\n",
    "for i in range(len(seconds_s)):\n",
    "    total_seconds_s.append(seconds_s[i]+secondsfromminutes_s[i])\n",
    "Start_time=[]\n",
    "for i in range(len(total_seconds_s)):\n",
    "    Start_time.append(time.strftime(\"%H:%M:%S\", time.gmtime(total_seconds_s[i])))\n",
    "print('\\n\\nStart time in Time Format:',Start_time)\n",
    "\n",
    "#doing the same for End time\n",
    "endtime=[]\n",
    "for i in range(len(ses_crop)):\n",
    "    endtime.append(math.modf(ses_crop['End time'][i])) # (0.5678000000000338, 1234.0)\n",
    "print('\\nEnd time in disected form:',endtime)\n",
    "\n",
    "seconds_e=[round(x[0]*100) for x in endtime]\n",
    "secondsfromminutes_e=[x[1]*60 for x in endtime]\n",
    "\n",
    "total_seconds_e=[]\n",
    "for i in range(len(seconds_e)):\n",
    "    total_seconds_e.append(seconds_e[i]+secondsfromminutes_e[i])\n",
    "End_time=[]\n",
    "for i in range(len(total_seconds_e)):\n",
    "    End_time.append(time.strftime(\"%H:%M:%S\", time.gmtime(total_seconds_e[i])))\n",
    "print('\\n\\nEnd time in Time Format:',End_time)\n",
    "\n",
    "ses_len=[]\n",
    "for i in range(len(total_seconds_s)):\n",
    "    ses_len.append(total_seconds_e[i]-total_seconds_s[i])\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ea79ac4e",
   "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>File Name</th>\n",
       "      <th>Start time</th>\n",
       "      <th>End time</th>\n",
       "      <th>session_name</th>\n",
       "      <th># of students in video</th>\n",
       "      <th>which assignments</th>\n",
       "      <th># of students with wrong answers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>00:36:50</td>\n",
       "      <td>00:40:29</td>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>18</td>\n",
       "      <td>hw6</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>00:40:29</td>\n",
       "      <td>00:46:13</td>\n",
       "      <td>Top-N Recommendation</td>\n",
       "      <td>18</td>\n",
       "      <td>hw 6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ITMD 525_03.MOV</td>\n",
       "      <td>00:47:26</td>\n",
       "      <td>00:51:52</td>\n",
       "      <td>Relevance Metrics</td>\n",
       "      <td>18</td>\n",
       "      <td>hw2, hw6</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         File Name Start time  End time          session_name  \\\n",
       "0  ITMD 525_03.MOV   00:36:50  00:40:29     Rating Prediction   \n",
       "1  ITMD 525_03.MOV   00:40:29  00:46:13  Top-N Recommendation   \n",
       "2  ITMD 525_03.MOV   00:47:26  00:51:52     Relevance Metrics   \n",
       "\n",
       "   # of students in video which assignments  # of students with wrong answers  \n",
       "0                      18               hw6                               8.0  \n",
       "1                      18              hw 6                               3.0  \n",
       "2                      18          hw2, hw6                               3.0  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#assigning the time values in correct format to cropped session dataframe\n",
    "\n",
    "ses_crop['Start time']=Start_time\n",
    "ses_crop['End time']=End_time\n",
    "ses_crop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8ab39d82",
   "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>session_name</th>\n",
       "      <th># of students in video</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Top-N Recommendation</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Relevance Metrics</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           session_name  # of students in video\n",
       "0     Rating Prediction                      18\n",
       "1  Top-N Recommendation                      18\n",
       "2     Relevance Metrics                      18"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#slicing session-wise attendance for comparison at later stage\n",
    "num_students=ses_crop.iloc[:,3:5]\n",
    "num_students"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "995e2185",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[66233.76623376623, 72797.2027972028, 85294.70529470529]\n",
      "\n",
      " [66383.61638361638, 72947.05294705294, 85444.55544455544]\n"
     ]
    }
   ],
   "source": [
    "#for each session we have to divide the video into 3 slices begin, dom, end\n",
    "#for each of sub-sessions, we have to determine start frame and end frame for slicing the sessions\n",
    "#let's determine startframe and endframe for begin session\n",
    "startframe_begin=[]\n",
    "for i in range(len(total_seconds_s)):\n",
    "    startframe_begin.append(fps*(total_seconds_s[i]))\n",
    "\n",
    "#after 5 seconds, begin subsession ends, therefore end frame of the begin subsession is located at (total_seconds_ofstarttime+5 seconds)*fps\n",
    "endframe_begin=[]\n",
    "for i in range(len(total_seconds_s)):\n",
    "    endframe_begin.append(fps*(total_seconds_s[i]+5))\n",
    "\n",
    "#endframe_begin = startframe_dom\n",
    "#endframe_dom = startframe_end\n",
    "\n",
    "#startframe_end & endframe_end can be calculated with by using total_seconds_e\n",
    "startframe_end=[]\n",
    "for i in range(len(total_seconds_e)):\n",
    "    startframe_end.append(fps*(total_seconds_e[i]-5))\n",
    "    \n",
    "endframe_end=[]\n",
    "for i in range(len(total_seconds_e)):\n",
    "    endframe_end.append(fps*total_seconds_e[i])\n",
    "\n",
    "my_dict={'session_name':ses_crop['session_name'],'startframe_begin':startframe_begin, 'endframe_begin':endframe_begin,'startframe_end':startframe_end,'endframe_end':endframe_end, 'session_length':ses_len}\n",
    "ses_timeframes=pd.DataFrame(my_dict)\n",
    "\n",
    "#reseting the index so that loop iteration does not face any probelms\n",
    "ses_timeframes.reset_index(drop=True, inplace=True)\n",
    "display(HTML(ses_timeframes.to_html()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "4d68c435",
   "metadata": {},
   "outputs": [],
   "source": [
    "height=int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)) #height\n",
    "width=int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)) #width\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5e9c308a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████████████| 93268/93268 [00:42<00:00, 2207.16it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Begin Sessions Number of Frames 450         session_name                                              frame\n",
      "0  Rating Prediction  [[[72, 133, 147], [72, 133, 147], [72, 133, 14...\n",
      "1  Rating Prediction  [[[72, 133, 147], [72, 133, 147], [72, 133, 14...\n",
      "2  Rating Prediction  [[[73, 134, 148], [73, 134, 148], [73, 134, 14...\n",
      "3  Rating Prediction  [[[71, 134, 148], [71, 134, 148], [71, 134, 14...\n",
      "4  Rating Prediction  [[[73, 134, 148], [73, 134, 148], [73, 134, 14...\n",
      "\n",
      "\n",
      "DOM Sessions Number of Frames 23946         session_name                                              frame\n",
      "0  Rating Prediction  [[[77, 139, 151], [77, 139, 151], [77, 139, 15...\n",
      "1  Rating Prediction  [[[76, 138, 150], [76, 138, 150], [76, 138, 15...\n",
      "2  Rating Prediction  [[[76, 138, 150], [76, 138, 150], [76, 138, 15...\n",
      "3  Rating Prediction  [[[76, 138, 150], [76, 138, 150], [76, 138, 15...\n",
      "4  Rating Prediction  [[[74, 138, 150], [74, 138, 150], [74, 138, 15...\n",
      "\n",
      "\n",
      "End Sessions Number of Frames 449         session_name                                              frame\n",
      "0  Rating Prediction  [[[76, 133, 154], [76, 133, 154], [76, 133, 15...\n",
      "1  Rating Prediction  [[[76, 133, 154], [76, 133, 154], [76, 133, 15...\n",
      "2  Rating Prediction  [[[76, 131, 156], [76, 131, 156], [76, 131, 15...\n",
      "3  Rating Prediction  [[[76, 132, 155], [76, 132, 155], [76, 132, 15...\n",
      "4  Rating Prediction  [[[76, 131, 156], [76, 131, 156], [79, 131, 15...\n",
      "Execution Time:  56  sec\n"
     ]
    }
   ],
   "source": [
    "#let's now break down the video files into sessions as given in excel file\n",
    "#further each session is going to be further divided into sub-sessions: begin, dom, end\n",
    "\n",
    "\n",
    "tic=time.time()\n",
    "\n",
    "#initializing lists for storing session-wise values for each subsession\n",
    "begin_list = []\n",
    "dom_list = []\n",
    "end_list =[]\n",
    "useless=[]\n",
    "c =0\n",
    "\n",
    "for i in tqdm(range(int(frame_count))):\n",
    "    #following if-elif statements are only meants for determining value of c for iterating through the ses_timeframes table shown above\n",
    "    \n",
    "    if 66233.766234  <= i <= 72797.202797:\n",
    "        c=0\n",
    "    elif 72797.202797  <= i <= 83106.893107:\n",
    "        c=1\n",
    "    elif 85294.705295  <= i <= 93266.733267:\n",
    "        c=2\n",
    "  \n",
    "       \n",
    "    #read the frame from given video file\n",
    "    ret, frame = cap.read()\n",
    "    \n",
    "    frame=cv2.resize(frame, (width,height))  #resize the frame\n",
    "    \n",
    "    if i>=ses_timeframes['startframe_begin'][c] and i<=ses_timeframes['endframe_begin'][c]:  \n",
    "        begin_list.append([ses_timeframes['session_name'][c],frame])            #output\n",
    "    elif i>=ses_timeframes['endframe_begin'][c] and i<=ses_timeframes['startframe_end'][c]:\n",
    "        dom_list.append([ses_timeframes['session_name'][c],frame]) \n",
    "    elif i>=ses_timeframes['startframe_end'][c] and i<=ses_timeframes['endframe_end'][c]:  \n",
    "        end_list.append([ses_timeframes['session_name'][c],frame]) \n",
    "   # else:\n",
    "    #    useless.append(frame)\n",
    "        \n",
    "#creating data frames out of lists\n",
    "df_begin = pd.DataFrame(begin_list, columns=[\"session_name\", \"frame\"])\n",
    "df_dom = pd.DataFrame(dom_list, columns=[\"session_name\", \"frame\"])\n",
    "df_end = pd.DataFrame(end_list, columns=[\"session_name\", \"frame\"]) \n",
    "\n",
    "#printing heads for begin, dom and end dfs\n",
    "print('Begin Sessions Number of Frames', len(df_begin),  df_begin.head())\n",
    "print('\\n\\nDOM Sessions Number of Frames', len(df_dom), df_dom.head())\n",
    "print('\\n\\nEnd Sessions Number of Frames', len(df_end), df_end.head())  \n",
    "\n",
    "tac=time.time()\n",
    "print('Execution Time: ',round(tac-tic),' sec')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c5ea2cff",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>frame</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_name</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Rating Prediction</th>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Top-N Recommendation</th>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Relevance Metrics</th>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>frame</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_name</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Rating Prediction</th>\n",
       "      <td>6264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Top-N Recommendation</th>\n",
       "      <td>10010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Relevance Metrics</th>\n",
       "      <td>7672</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>frame</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>session_name</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Rating Prediction</th>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Top-N Recommendation</th>\n",
       "      <td>149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Relevance Metrics</th>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#just checking the number of frames per subsession \n",
    "begin_sessions=df_begin.groupby(['session_name'], sort=False).count()\n",
    "display(HTML(begin_sessions.to_html()))\n",
    "dom_sessions=df_dom.groupby(['session_name'],sort=False).count()\n",
    "display(HTML(dom_sessions.to_html()))\n",
    "end_sessions=df_end.groupby(['session_name'],sort=False).count()\n",
    "display(HTML(end_sessions.to_html()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "5a2354c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "### YOU LOOK ONLY ONCE ###\n",
    "#creating function for applying YOLO algorithm on each frame to detect faces\n",
    "\n",
    "def Prob_Emo_YOLO (image):\n",
    "    # Load image and convert to blob\n",
    "    \n",
    "    blob = cv2.dnn.blobFromImage(image, scale_factor, input_size, swapRB=True)\n",
    "\n",
    "    # Set input blob for the network\n",
    "    net.setInput(blob)\n",
    "\n",
    "    # Get output layer names\n",
    "    output_layer_names = net.getUnconnectedOutLayersNames()\n",
    "\n",
    "    # Forward pass through the network\n",
    "    outputs = net.forward(output_layer_names)\n",
    "\n",
    "    # Get bounding boxes, confidences, and class IDs for detected faces\n",
    "    confidences = []\n",
    "    class_ids = []\n",
    "    boxes=[]\n",
    "    probs = []\n",
    "\n",
    "    for output in outputs:\n",
    "        for detection in output:\n",
    "            scores = detection[5:]\n",
    "            class_id = np.argmax(scores)\n",
    "            confidence = scores[class_id]\n",
    "            if confidence > 0.5 and classes[class_id] == 'face':\n",
    "                center_x = int(detection[0] * image.shape[1])\n",
    "                center_y = int(detection[1] * image.shape[0])\n",
    "                w = int(detection[2] * image.shape[1])\n",
    "                h = int(detection[3] * image.shape[0])\n",
    "                x = int(center_x - width / 2)\n",
    "                y = int(center_y - height / 2)\n",
    "                \n",
    "                if y - 10 > 0:\n",
    "                    y = y - 10\n",
    "                    roi = image[y:y + h, x:x + w]\n",
    "                    roi_gray = cv2.cvtColor(roi, cv2.COLOR_BGR2GRAY)\n",
    "                    roi_gray = cv2.resize(roi_gray, (48, 48))\n",
    "                    roi_gray = roi_gray.astype('float') / 255.0\n",
    "                    roi_gray = np.asarray(roi_gray)\n",
    "                    roi_gray = cv2.resize(roi_gray, (64, 64))\n",
    "                    roi_gray = roi_gray.reshape(1, 64, 64, 1)\n",
    "                    prediction = emotion_model.predict(roi_gray)\n",
    "                    probs.append(prediction)\n",
    "   \n",
    "                \n",
    "    return probs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c33a019e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████| 450/450 [02:40<00:00,  2.80it/s]\n",
      "23602it [2:25:31,  2.70it/s]\n",
      "100%|████████████████████████████████████████████████████████████████████████████████| 449/449 [02:23<00:00,  3.14it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Execution Time:  9038  sec\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "import itertools as it\n",
    "tic = time.time()\n",
    "net = cv2.dnn.readNetFromDarknet('yolov3-face.cfg', 'yolov3-wider_16000.weights')\n",
    "\n",
    "net.setPreferableBackend(cv2.dnn.DNN_BACKEND_CUDA)\n",
    "net.setPreferableTarget(cv2.dnn.DNN_TARGET_CUDA)\n",
    "\n",
    "emotion_model = load_model('emotion_model.hdf5', compile = False)\n",
    "emotion_labels = [\"angry\", \"disgust\", \"fear\", \"happy\", \"sad\", \"surprise\", \"neutral\"]\n",
    "\n",
    "\n",
    "# Define classes to detect (in this case, only faces)\n",
    "classes = ['face']\n",
    "\n",
    "# Set input/output sizes and scaling factors\n",
    "input_size = (416, 416)\n",
    "scale_factor = 1/255\n",
    "\n",
    "#detect number of faces for each frame of begin subsessions\n",
    "face_list_begin=[]\n",
    "for i in tqdm(range(int(len(df_begin)))):\n",
    "    frame_1=df_begin['frame'][i]\n",
    "    prob_emo=Prob_Emo_YOLO(frame_1)\n",
    "    face_list_begin[len(face_list_begin):]=[(df_begin['session_name'][i], prob_emo)]\n",
    "    \n",
    "faces_begin = pd.DataFrame(face_list_begin, columns=[\"session_name\", \"Prob of Emotions\"])  \n",
    "\n",
    "#detect number of faces for each frame of begin subsessions\n",
    "face_list_dom=[]\n",
    "for i in tqdm(it.chain(range(0, 910), range(977, 7141), range(7210, 14511), range(14568, 15293), range(15414, 19962), range(19968, 21719),range(21743, 23946))):\n",
    "    frame_1=df_dom['frame'][i]\n",
    "    prob_emo=Prob_Emo_YOLO(frame_1)\n",
    "    face_list_dom[len(face_list_dom):]=[(df_dom['session_name'][i], prob_emo)]\n",
    "    \n",
    "faces_dom = pd.DataFrame(face_list_dom, columns=[\"session_name\", \"Prob of Emotions\"])  \n",
    "\n",
    "#detect number of faces for each frame of begin subsessions\n",
    "face_list_end=[]\n",
    "for i in tqdm(range(int(len(df_end)))):\n",
    "    frame_1=df_end['frame'][i]\n",
    "    prob_emo=Prob_Emo_YOLO(frame_1)\n",
    "    face_list_end[len(face_list_end):]=[(df_end['session_name'][i], prob_emo)]\n",
    "    \n",
    "faces_end = pd.DataFrame(face_list_end, columns=[\"session_name\", \"Prob of Emotions\"])  \n",
    "\n",
    "\n",
    "tac=time.time()\n",
    "print('Execution Time: ',round(tac-tic),' sec')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a97bcc28",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>session_name</th>\n",
       "      <th>emotion</th>\n",
       "      <th>avg_emotion</th>\n",
       "      <th>std_dev_emotion</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.120022856, 0.0001235502, 0.12921241, 0.103057295, 0.034890965, 0.21643983, 0.3962531]]</td>\n",
       "      <td>[[0.11696277, 0.00012123612, 0.05999251, 0.038711652, 0.011765267, 0.16196027, 0.13034709]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.11893397, 0.0001239974, 0.12713405, 0.104040526, 0.035074573, 0.21506603, 0.39962685]]</td>\n",
       "      <td>[[0.11407356, 0.00012092545, 0.0580106, 0.038420312, 0.011962884, 0.162122, 0.12800615]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.11595917, 0.00012403674, 0.1280999, 0.104239024, 0.03498844, 0.21356073, 0.40302873]]</td>\n",
       "      <td>[[0.111149, 0.00012723205, 0.06205665, 0.039685693, 0.0119917365, 0.16005158, 0.12665659]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12079915, 0.00014441984, 0.123549074, 0.102975816, 0.036400255, 0.21258794, 0.40354335]]</td>\n",
       "      <td>[[0.122440174, 0.00015856129, 0.055966597, 0.04426507, 0.01371543, 0.16064149, 0.12337304]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.11345344, 0.00013551772, 0.13152473, 0.10442599, 0.036373958, 0.21363275, 0.40045363]]</td>\n",
       "      <td>[[0.106205314, 0.0001369509, 0.06538927, 0.039762557, 0.013704724, 0.16000459, 0.1286323]]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbsAAAESCAYAAABkens4AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAVdUlEQVR4nO3df0zU9+HH8dcBcipyR2FwpxGCSd2UVGxEJ7ctjVMqtdTpxGVmRrFxmhlwQ1KtZFY7uwRDjb9WLcuyVZvVuLhMN3G0ZTRiNigqHYbaarqmDBI8sDHcCd96/LrvH42f9Sq2guDJ2+cj+ST9fN7vz93707R55nO/sAWDwaAAADBYRLgXAADASCN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYLyrcCxiK/v5+tba2KjY2VjabLdzLAQCESTAY1I0bNzRp0iRFRNz5/m1Uxq61tVXJycnhXgYA4AHR0tKiyZMn33F8VMYuNjZW0ucX53A4wrwaAEC4+P1+JScnW124k1EZu1svXTocDmIHAPjat7T4gAoAwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxRuWXygF8LnXr6XAvwdK0KyfcSwDuiDs7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADDePcVu165dstlsKiwstI7dvHlT+fn5SkhI0IQJE5Sbm6u2traQ85qbm5WTk6Px48crKSlJmzdvVm9v770sBQCAOxpy7M6fP6/f/va3Sk9PDzm+adMmnTp1SsePH1d1dbVaW1u1bNkya7yvr085OTnq7u5WTU2Njhw5osOHD2v79u1DvwoAAL7CkGLX2dmplStX6ne/+50eeeQR67jP59Pvf/977dmzR/Pnz1dGRoZee+011dTU6N1335Ukvf322/rggw/0xz/+UY8//rgWLVqkl156SQcPHlR3d/fwXBUAAF8wpNjl5+crJydHWVlZIcfr6+vV09MTcnzatGlKSUlRbW2tJKm2tlYzZsyQy+Wy5mRnZ8vv9+vSpUsDPl8gEJDf7w/ZAAC4W1GDPeHYsWN67733dP78+dvGvF6voqOjFRcXF3Lc5XLJ6/Vac74Yulvjt8YGUlJSol/96leDXSoAAJIGeWfX0tKiX/ziF3rjjTc0duzYkVrTbYqLi+Xz+aytpaXlvj03AGD0G1Ts6uvr1d7erlmzZikqKkpRUVGqrq7WgQMHFBUVJZfLpe7ubnV0dISc19bWJrfbLUlyu923fTrz1v6tOV9mt9vlcDhCNgAA7tagYrdgwQI1NjaqoaHB2mbPnq2VK1da/zxmzBhVVVVZ51y5ckXNzc3yeDySJI/Ho8bGRrW3t1tzKisr5XA4lJaWNkyXBQDA/wzqPbvY2Fg99thjIcdiYmKUkJBgHV+7dq2KiooUHx8vh8OhjRs3yuPxKDMzU5K0cOFCpaWladWqVSotLZXX69W2bduUn58vu90+TJcFAMD/DPoDKl9n7969ioiIUG5urgKBgLKzs3Xo0CFrPDIyUuXl5dqwYYM8Ho9iYmKUl5ennTt3DvdSAACQJNmCwWAw3IsYLL/fL6fTKZ/Px/t3eKilbj0d7iVYmnblhHsJeAjdbQ/4bUwAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIw3qNi9+uqrSk9Pl8PhkMPhkMfjUUVFhTV+8+ZN5efnKyEhQRMmTFBubq7a2tpCHqO5uVk5OTkaP368kpKStHnzZvX29g7P1QAAMIBBxW7y5MnatWuX6uvrdeHCBc2fP19LlizRpUuXJEmbNm3SqVOndPz4cVVXV6u1tVXLli2zzu/r61NOTo66u7tVU1OjI0eO6PDhw9q+ffvwXhUAAF9gCwaDwXt5gPj4eL388stavny5EhMTdfToUS1fvlySdPnyZU2fPl21tbXKzMxURUWFnnnmGbW2tsrlckmSysrK9Pzzz+vatWuKjo6+q+f0+/1yOp3y+XxyOBz3snxgVEvdejrcS7A07coJ9xLwELrbHgz5Pbu+vj4dO3ZMXV1d8ng8qq+vV09Pj7Kysqw506ZNU0pKimprayVJtbW1mjFjhhU6ScrOzpbf77fuDgcSCATk9/tDNgAA7tagY9fY2KgJEybIbrfrZz/7mU6cOKG0tDR5vV5FR0crLi4uZL7L5ZLX65Ukeb3ekNDdGr81diclJSVyOp3WlpycPNhlAwAeYoOO3be+9S01NDSorq5OGzZsUF5enj744IORWJuluLhYPp/P2lpaWkb0+QAAZoka7AnR0dF69NFHJUkZGRk6f/689u/frx//+Mfq7u5WR0dHyN1dW1ub3G63JMntduvcuXMhj3fr05q35gzEbrfLbrcPdqkAAEgahu/Z9ff3KxAIKCMjQ2PGjFFVVZU1duXKFTU3N8vj8UiSPB6PGhsb1d7ebs2prKyUw+FQWlravS4FAIABDerOrri4WIsWLVJKSopu3Liho0eP6syZM3rrrbfkdDq1du1aFRUVKT4+Xg6HQxs3bpTH41FmZqYkaeHChUpLS9OqVatUWloqr9erbdu2KT8/nzs3AMCIGVTs2tvbtXr1al29elVOp1Pp6el666239OSTT0qS9u7dq4iICOXm5ioQCCg7O1uHDh2yzo+MjFR5ebk2bNggj8ejmJgY5eXlaefOncN7VQAAfME9f88uHPieHfA5vmeHh92If88OAIDRgtgBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeMQOAGA8YgcAMB6xAwAYj9gBAIxH7AAAxiN2AADjETsAgPGIHQDAeIOKXUlJiebMmaPY2FglJSVp6dKlunLlSsicmzdvKj8/XwkJCZowYYJyc3PV1tYWMqe5uVk5OTkaP368kpKStHnzZvX29t771QAAMIBBxa66ulr5+fl69913VVlZqZ6eHi1cuFBdXV3WnE2bNunUqVM6fvy4qqur1draqmXLllnjfX19ysnJUXd3t2pqanTkyBEdPnxY27dvH76rAgDgC2zBYDA41JOvXbumpKQkVVdX64knnpDP51NiYqKOHj2q5cuXS5IuX76s6dOnq7a2VpmZmaqoqNAzzzyj1tZWuVwuSVJZWZmef/55Xbt2TdHR0V/7vH6/X06nUz6fTw6HY6jLB0a91K2nw70ES9OunHAvAQ+hu+3BPb1n5/P5JEnx8fGSpPr6evX09CgrK8uaM23aNKWkpKi2tlaSVFtbqxkzZlihk6Ts7Gz5/X5dunRpwOcJBALy+/0hGwAAd2vIsevv71dhYaG++93v6rHHHpMkeb1eRUdHKy4uLmSuy+WS1+u15nwxdLfGb40NpKSkRE6n09qSk5OHumwAwENoyLHLz8/X+++/r2PHjg3negZUXFwsn89nbS0tLSP+nAAAc0QN5aSCggKVl5fr7Nmzmjx5snXc7Xaru7tbHR0dIXd3bW1tcrvd1pxz586FPN6tT2vemvNldrtddrt9KEsFAGBwd3bBYFAFBQU6ceKE3nnnHU2ZMiVkPCMjQ2PGjFFVVZV17MqVK2pubpbH45EkeTweNTY2qr293ZpTWVkph8OhtLS0e7kWAAAGNKg7u/z8fB09elR//etfFRsba73H5nQ6NW7cODmdTq1du1ZFRUWKj4+Xw+HQxo0b5fF4lJmZKUlauHCh0tLStGrVKpWWlsrr9Wrbtm3Kz8/n7g0AMCIGFbtXX31VkjRv3ryQ46+99prWrFkjSdq7d68iIiKUm5urQCCg7OxsHTp0yJobGRmp8vJybdiwQR6PRzExMcrLy9POnTvv7UoAALiDe/qeXbjwPTvgc3zPDg+7+/I9OwAARgNiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMROwCA8YgdAMB4xA4AYDxiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMROwCA8YgdAMB4xA4AYDxiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMROwCA8YgdAMB4xA4AYDxiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMROwCA8YgdAMB4xA4AYDxiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMNOnZnz57V4sWLNWnSJNlsNp08eTJkPBgMavv27Zo4caLGjRunrKwsffTRRyFzrl+/rpUrV8rhcCguLk5r165VZ2fnPV0IAAB3MujYdXV1aebMmTp48OCA46WlpTpw4IDKyspUV1enmJgYZWdn6+bNm9aclStX6tKlS6qsrFR5ebnOnj2r9evXD/0qAAD4ClGDPWHRokVatGjRgGPBYFD79u3Ttm3btGTJEknS66+/LpfLpZMnT2rFihX68MMP9eabb+r8+fOaPXu2JOk3v/mNnn76ae3evVuTJk267XEDgYACgYC17/f7B7tsAMBDbFjfs/vkk0/k9XqVlZVlHXM6nZo7d65qa2slSbW1tYqLi7NCJ0lZWVmKiIhQXV3dgI9bUlIip9NpbcnJycO5bACA4YY1dl6vV5LkcrlCjrtcLmvM6/UqKSkpZDwqKkrx8fHWnC8rLi6Wz+eztpaWluFcNgDAcIN+GTMc7Ha77HZ7uJcBABilhvXOzu12S5La2tpCjre1tVljbrdb7e3tIeO9vb26fv26NQcAgOE0rLGbMmWK3G63qqqqrGN+v191dXXyeDySJI/Ho46ODtXX11tz3nnnHfX392vu3LnDuRwAACQN4WXMzs5O/ec//7H2P/nkEzU0NCg+Pl4pKSkqLCzUr3/9a02dOlVTpkzRCy+8oEmTJmnp0qWSpOnTp+upp57SunXrVFZWpp6eHhUUFGjFihUDfhITAIB7NejYXbhwQd///vet/aKiIklSXl6eDh8+rC1btqirq0vr169XR0eHvve97+nNN9/U2LFjrXPeeOMNFRQUaMGCBYqIiFBubq4OHDgwDJcDAMDtbMFgMBjuRQyW3++X0+mUz+eTw+EI93KAsEndejrcS7A07coJ9xLwELrbHvDbmAAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGI/YAQCMR+wAAMYjdgAA4xE7AIDxiB0AwHjEDgBgPGIHADAesQMAGC9ssTt48KBSU1M1duxYzZ07V+fOnQvXUgAAhgtL7P70pz+pqKhIO3bs0HvvvaeZM2cqOztb7e3t4VgOAMBwUeF40j179mjdunV69tlnJUllZWU6ffq0/vCHP2jr1q23zQ8EAgoEAta+z+eTJPn9/vuzYOAB1R/4v3AvwcL/jwiHW//dBYPBr54YvM8CgUAwMjIyeOLEiZDjq1evDv7gBz8Y8JwdO3YEJbGxsbGxsQ24tbS0fGV77vud3aeffqq+vj65XK6Q4y6XS5cvXx7wnOLiYhUVFVn7/f39un79uhISEmSz2UZ0vcCDyu/3Kzk5WS0tLXI4HOFeDhAWwWBQN27c0KRJk75yXlhexhwsu90uu90eciwuLi48iwEeMA6Hg9jhoeZ0Or92zn3/gMo3vvENRUZGqq2tLeR4W1ub3G73/V4OAOAhcN9jFx0drYyMDFVVVVnH+vv7VVVVJY/Hc7+XAwB4CITlZcyioiLl5eVp9uzZ+va3v619+/apq6vL+nQmgK9nt9u1Y8eO217iB3A7WzD4dZ/XHBmvvPKKXn75ZXm9Xj3++OM6cOCA5s6dG46lAAAMF7bYAQBwv/DbmAAA4xE7AIDxiB0AwHjEDkCI1NRU7du3L9zLAIYVsQNGuXnz5qmwsDDcywAeaMQOeAgEg0H19vaGexlA2BA7YATNmzdPP//5z7VlyxbFx8fL7XbrxRdftMY7Ojr005/+VImJiXI4HJo/f74uXrxoja9Zs0ZLly4NeczCwkLNmzfPGq+urtb+/ftls9lks9nU1NSkM2fOyGazqaKiQhkZGbLb7frnP/+pjz/+WEuWLJHL5dKECRM0Z84c/eMf/7gP/yaA8CJ2wAg7cuSIYmJiVFdXp9LSUu3cuVOVlZWSpB/96Edqb29XRUWF6uvrNWvWLC1YsEDXr1+/q8fev3+/PB6P1q1bp6tXr+rq1atKTk62xrdu3apdu3bpww8/VHp6ujo7O/X000+rqqpK//73v/XUU09p8eLFam5uHpFrBx4Uo+KvHgCjWXp6unbs2CFJmjp1ql555RVVVVVp3LhxOnfunNrb262f/Nq9e7dOnjypP//5z1q/fv3XPrbT6VR0dLTGjx8/4A+p79y5U08++aS1Hx8fr5kzZ1r7L730kk6cOKG//e1vKigouNdLBR5YxA4YYenp6SH7EydOVHt7uy5evKjOzk4lJCSEjH/22Wf6+OOPh+W5Z8+eHbLf2dmpF198UadPn9bVq1fV29urzz77jDs7GI/YASNszJgxIfs2m039/f3q7OzUxIkTdebMmdvOufX3GiMiIvTlX/Tr6em56+eOiYkJ2X/uuedUWVmp3bt369FHH9W4ceO0fPlydXd33/VjAqMRsQPCZNasWfJ6vYqKilJqauqAcxITE/X++++HHGtoaAgJaHR0tPr6+u7qOf/1r39pzZo1+uEPfyjp8zu9pqamIa0fGE34gAoQJllZWfJ4PFq6dKnefvttNTU1qaamRr/85S914cIFSdL8+fN14cIFvf766/roo4+0Y8eO2+KXmpqquro6NTU16dNPP1V/f/8dn3Pq1Kn6y1/+ooaGBl28eFE/+clPvnI+YApiB4SJzWbT3//+dz3xxBN69tln9c1vflMrVqzQf//7X7lcLklSdna2XnjhBW3ZskVz5szRjRs3tHr16pDHee655xQZGam0tDQlJiZ+5ftve/bs0SOPPKLvfOc7Wrx4sbKzszVr1qwRvU7gQcCf+AEAGI87OwCA8YgdAMB4xA4AYDxiBwAwHrEDABiP2AEAjEfsAADGI3YAAOMROwCA8YgdAMB4xA4AYLz/B3jRTliP9pGDAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "session_name\n",
       "Rating Prediction       neutral\n",
       "Relevance Metrics       neutral\n",
       "Top-N Recommendation    neutral\n",
       "Name: emotion, dtype: object"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#let's analyze which emotion is dominant for each session\n",
    "#there are two ways to do this\n",
    "#number#1: For a given frame, get average of all of probability ndarrays of emotion. Subsequently, assign a label for \n",
    "#emotion for each frame on the basis of maximum value of emotion. This method is implemented in this code cell.\n",
    "avg_emotion_b=[]\n",
    "std_emotion_b=[]\n",
    "emo_b=[]\n",
    "for i in range(len(faces_begin)):\n",
    "    list_b=faces_begin.iloc[i,1]\n",
    "    avg_emo = np.mean(list_b, axis=0)   #getting average of arrays of emotions for every frame\n",
    "    std_emo = np.std(list_b, axis=0)    #getting standard deviation of arrays of emotions for every frame\n",
    "    maxindex = int(np.argmax(avg_emo))\n",
    "    emotion = emotion_labels[maxindex]  #figuring out the emotion label based on highest value of emotion\n",
    "    emo_b.append(emotion)               #preparing list of emotion label  \n",
    "    avg_emotion_b.append(avg_emo)       #preparing list of ndarrays of average emotion\n",
    "    std_emotion_b.append(std_emo)       #preparing list of ndarrays of standard deviation of emotion for a frame\n",
    "    \n",
    "\n",
    "faces_begin['emotion']= emo_b\n",
    "faces_begin['avg_emotion']=avg_emotion_b #per frame\n",
    "faces_begin['std_dev_emotion']= std_emotion_b\n",
    "begin_emotion_framewise=faces_begin.drop(columns='Prob of Emotions')\n",
    "display(HTML(begin_emotion_framewise.head().to_html()))\n",
    "\n",
    "# Creating histogram\n",
    "fig, ax = plt.subplots(figsize =(5,3))\n",
    "ax.hist(emo_b )\n",
    " \n",
    "# Show plot\n",
    "plt.show()\n",
    "\n",
    "#finally let's get value of mode of emotion for each session\n",
    "begin_emotion_framewise.groupby(['session_name'])['emotion'].agg(pd.Series.mode)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "742fbace",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Rating Prediction:  1192\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Relevance Metrics:  1390\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Top-N Recommendation:  1325\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n"
     ]
    }
   ],
   "source": [
    "#let's try another way now\n",
    "#number#2:In this, we'll get ALL the emotion arrays for a given session\n",
    "#In other words, we'll be aggregating emotion arrayss for each of the frames of a given session\n",
    "#Then we'll find the average of emotion arrays for that session\n",
    "\n",
    "#let's generate list of lists of emotion ndarryas since each frame contains a list of emotion ndarryas\n",
    "# so aggregation will result in list of lists\n",
    "\n",
    "df1=faces_begin.groupby(['session_name'])['Prob of Emotions'].agg(pd.Series)\n",
    "\n",
    "\n",
    "rating_predict=df1.iloc[0]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_1 = []\n",
    "for sublist in rating_predict:\n",
    "    for item in sublist:\n",
    "        flat_list_1.append(item)\n",
    "        \n",
    "session_avg1=np.mean(flat_list_1, axis=0)\n",
    "maxindex = int(np.argmax(session_avg1))\n",
    "emo_ses_avg1 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\n\\nTotal number of Emotions Detected in Begin Session of Rating Prediction: \", len(flat_list_1))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg1 )\n",
    "\n",
    "#let's try the same thing for another session\n",
    "relevance_met=df1.iloc[1]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_2 = []\n",
    "for sublist in relevance_met:\n",
    "    for item in sublist:\n",
    "        flat_list_2.append(item)\n",
    "        \n",
    "session_avg2=np.mean(flat_list_2, axis=0)\n",
    "maxindex = int(np.argmax(session_avg2))\n",
    "emo_ses_avg2 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Relevance Metrics: \", len(flat_list_2))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg2 )\n",
    "\n",
    "#let's try the same thing for Top-N Recommendation session \n",
    "top_recom=df1.iloc[2]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_3 = []\n",
    "for sublist in top_recom:\n",
    "    for item in sublist:\n",
    "        flat_list_3.append(item)\n",
    "\n",
    "session_avg3=np.mean(flat_list_3, axis=0)\n",
    "maxindex = int(np.argmax(session_avg3))\n",
    "emo_ses_avg3 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Top-N Recommendation: \", len(flat_list_3))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg3 )\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "902504aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>session_name</th>\n",
       "      <th>emotion</th>\n",
       "      <th>avg_emotion</th>\n",
       "      <th>std_dev_emotion</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.121820465, 0.00019454703, 0.11202833, 0.0985593, 0.030196363, 0.22792722, 0.40927377]]</td>\n",
       "      <td>[[0.1254005, 0.00023810877, 0.062427755, 0.047767423, 0.017386189, 0.19817363, 0.15962201]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12314841, 0.0001724486, 0.11014336, 0.100952014, 0.030409133, 0.23234192, 0.40283275]]</td>\n",
       "      <td>[[0.12983069, 0.00020064082, 0.060464516, 0.04804496, 0.0167468, 0.2061102, 0.16405883]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12623037, 0.00017694148, 0.10965031, 0.09904104, 0.030569758, 0.22974862, 0.40458298]]</td>\n",
       "      <td>[[0.1287242, 0.00020860597, 0.056171108, 0.052046046, 0.016821215, 0.19530976, 0.15333271]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.1283689, 0.00013296232, 0.103868246, 0.09879019, 0.028763402, 0.22425064, 0.41582564]]</td>\n",
       "      <td>[[0.13967727, 0.00018520183, 0.05830961, 0.05903782, 0.018034605, 0.22323507, 0.164118]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12087871, 0.00015403035, 0.11200558, 0.1047276, 0.030022157, 0.22532257, 0.40688938]]</td>\n",
       "      <td>[[0.12749396, 0.00021912892, 0.06751391, 0.07299689, 0.01955367, 0.22501121, 0.15232337]]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "session_name\n",
       "Rating Prediction       neutral\n",
       "Relevance Metrics       neutral\n",
       "Top-N Recommendation    neutral\n",
       "Name: emotion, dtype: object"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#let's analyze which emotion is dominant for each session\n",
    "#there are two ways to do this\n",
    "#number#1: For a given frame, get average of all of probability ndarrays of emotion. Subsequently, assign a label for \n",
    "#emotion for each frame on the basis of maximum value of emotion. This method is implemented in this code cell.\n",
    "\n",
    "avg_emotion_d=[]\n",
    "std_emotion_d=[]\n",
    "emo_d=[]\n",
    "for i in range(len(faces_dom)):\n",
    "    list_d=faces_dom.iloc[i,1]\n",
    "    avg_emo = np.mean(list_d, axis=0)   #getting average of arrays of emotions for every frame\n",
    "    std_emo = np.std(list_d, axis=0)    #getting standard deviation of arrays of emotions for every frame\n",
    "    maxindex = int(np.argmax(avg_emo))\n",
    "    emotion = emotion_labels[maxindex]  #figuring out the emotion label based on highest value of emotion\n",
    "    emo_d.append(emotion)               #preparing list of emotion label  \n",
    "    avg_emotion_d.append(avg_emo)       #preparing list of ndarrays of average emotion\n",
    "    std_emotion_d.append(std_emo)       ##preparing list of ndarrays of standard deviation of emotion for a frame\n",
    "    \n",
    "#updating the dataframe\n",
    "faces_dom['emotion']= emo_d\n",
    "faces_dom['avg_emotion']=avg_emotion_d \n",
    "faces_dom['std_dev_emotion']= std_emotion_d\n",
    "dom_emotion_framewise=faces_dom.drop(columns='Prob of Emotions')\n",
    "display(HTML(dom_emotion_framewise.head().to_html()))\n",
    "\n",
    "# Creating histogram\n",
    "fig, ax = plt.subplots(figsize =(5,3))\n",
    "ax.hist(emo_d )\n",
    " \n",
    "# Show plot\n",
    "plt.show()\n",
    "\n",
    "#finally let's get value of mode of emotion for each session\n",
    "dom_emotion_framewise.groupby(['session_name'])['emotion'].agg(pd.Series.mode)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "fe3a5078",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Rating Prediction:  56977\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Relevance Metrics:  68160\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Top-N Recommendation:  111295\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n"
     ]
    }
   ],
   "source": [
    "#let's try another way now\n",
    "#In this, we'll get ALL the emotion arrays for a given session\n",
    "#In other words, we'll be aggregating emotion arrayss for each of the frames of a given session\n",
    "#Then we'll find the average of emotion arrays for that session\n",
    "\n",
    "#let's generate list of lists of emotion ndarryas since each frame contains a list of emotion ndarryas\n",
    "# so aggregation will result in list of lists\n",
    "\n",
    "df2=faces_dom.groupby(['session_name'])['Prob of Emotions'].agg(pd.Series)\n",
    "\n",
    "\n",
    "rating_predict=df2.iloc[0]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_4 = []\n",
    "for sublist in rating_predict:\n",
    "    for item in sublist:\n",
    "        flat_list_4.append(item)\n",
    "        \n",
    "session_avg4=np.mean(flat_list_4, axis=0)\n",
    "maxindex = int(np.argmax(session_avg4))\n",
    "emo_ses_avg4 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\n\\nTotal number of Emotions Detected in Begin Session of Rating Prediction: \", len(flat_list_4))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg4 )\n",
    "\n",
    "#let's try the same thing for another session\n",
    "relevance_met=df2.iloc[1]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_5 = []\n",
    "for sublist in relevance_met:\n",
    "    for item in sublist:\n",
    "        flat_list_5.append(item)\n",
    "        \n",
    "session_avg5=np.mean(flat_list_5, axis=0)\n",
    "maxindex = int(np.argmax(session_avg5))\n",
    "emo_ses_avg5 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Relevance Metrics: \", len(flat_list_5))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg5 )\n",
    "\n",
    "#let's try the same thing for Top-N Recommendation session \n",
    "top_recom=df2.iloc[2]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_6 = []\n",
    "for sublist in top_recom:\n",
    "    for item in sublist:\n",
    "        flat_list_6.append(item)\n",
    "\n",
    "session_avg6=np.mean(flat_list_6, axis=0)\n",
    "maxindex = int(np.argmax(session_avg6))\n",
    "emo_ses_avg6 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Top-N Recommendation: \", len(flat_list_6))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg6 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "6f7f8e0d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>session_name</th>\n",
       "      <th>emotion</th>\n",
       "      <th>avg_emotion</th>\n",
       "      <th>std_dev_emotion</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.121820465, 0.00019454703, 0.11202833, 0.0985593, 0.030196363, 0.22792722, 0.40927377]]</td>\n",
       "      <td>[[0.1254005, 0.00023810877, 0.062427755, 0.047767423, 0.017386189, 0.19817363, 0.15962201]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12314841, 0.0001724486, 0.11014336, 0.100952014, 0.030409133, 0.23234192, 0.40283275]]</td>\n",
       "      <td>[[0.12983069, 0.00020064082, 0.060464516, 0.04804496, 0.0167468, 0.2061102, 0.16405883]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12623037, 0.00017694148, 0.10965031, 0.09904104, 0.030569758, 0.22974862, 0.40458298]]</td>\n",
       "      <td>[[0.1287242, 0.00020860597, 0.056171108, 0.052046046, 0.016821215, 0.19530976, 0.15333271]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.1283689, 0.00013296232, 0.103868246, 0.09879019, 0.028763402, 0.22425064, 0.41582564]]</td>\n",
       "      <td>[[0.13967727, 0.00018520183, 0.05830961, 0.05903782, 0.018034605, 0.22323507, 0.164118]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Rating Prediction</td>\n",
       "      <td>neutral</td>\n",
       "      <td>[[0.12087871, 0.00015403035, 0.11200558, 0.1047276, 0.030022157, 0.22532257, 0.40688938]]</td>\n",
       "      <td>[[0.12749396, 0.00021912892, 0.06751391, 0.07299689, 0.01955367, 0.22501121, 0.15232337]]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "session_name\n",
       "Rating Prediction       neutral\n",
       "Relevance Metrics       neutral\n",
       "Top-N Recommendation    neutral\n",
       "Name: emotion, dtype: object"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "avg_emotion_e=[]\n",
    "std_emotion_e=[]\n",
    "emo_e=[]\n",
    "for i in range(len(faces_end)):\n",
    "    list_e=faces_dom.iloc[i,1]\n",
    "    avg_emo = np.mean(list_e, axis=0)   #getting average of arrays of emotions for every frame\n",
    "    std_emo = np.std(list_e, axis=0)    #getting standard deviation of arrays of emotions for every frame\n",
    "    maxindex = int(np.argmax(avg_emo))\n",
    "    emotion = emotion_labels[maxindex]  #figuring out the emotion label based on highest value of emotion\n",
    "    emo_e.append(emotion)               #preparing list of emotion label  \n",
    "    avg_emotion_e.append(avg_emo)       #preparing list of ndarrays of average emotion\n",
    "    std_emotion_e.append(std_emo)       ##preparing list of ndarrays of standard deviation of emotion for a frame\n",
    "    \n",
    "#updating the dataframe\n",
    "faces_end['emotion']= emo_e\n",
    "faces_end['avg_emotion']=avg_emotion_e \n",
    "faces_end['std_dev_emotion']= std_emotion_e\n",
    "end_emotion_framewise=faces_end.drop(columns='Prob of Emotions')\n",
    "display(HTML(end_emotion_framewise.head().to_html()))\n",
    "\n",
    "# Creating histogram\n",
    "fig, ax = plt.subplots(figsize =(5,3))\n",
    "ax.hist(emo_e )\n",
    " \n",
    "# Show plot\n",
    "plt.show()\n",
    "\n",
    "#finally let's get value of mode of emotion for each session\n",
    "end_emotion_framewise.groupby(['session_name'])['emotion'].agg(pd.Series.mode)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "dac1be7c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Rating Prediction:  1192\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Relevance Metrics:  1390\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n",
      "\n",
      "Total number of Emotions Detected in Begin Session of Top-N Recommendation:  1325\n",
      "Avg emotion for Begin session of Rating Prediction:  neutral\n"
     ]
    }
   ],
   "source": [
    "#let's try another way now\n",
    "#In this, we'll get ALL the emotion values for a given session\n",
    "#In other words, we'll be aggregating emotion values for each of the frames of a given session\n",
    "#Then we'll find the mode of emotion values for that session\n",
    "\n",
    "#let's generate list of lists of emotion since each frame contains a list of emotion\n",
    "# so aggregation will result in list of lists\n",
    "\n",
    "df3=faces_begin.groupby(['session_name'])['Prob of Emotions'].agg(pd.Series)\n",
    "\n",
    "rating_predict=df3.iloc[0]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_7 = []\n",
    "for sublist in rating_predict:\n",
    "    for item in sublist:\n",
    "        flat_list_7.append(item)\n",
    "        \n",
    "session_avg7=np.mean(flat_list_7, axis=0)\n",
    "maxindex = int(np.argmax(session_avg7))\n",
    "emo_ses_avg7 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\n\\nTotal number of Emotions Detected in Begin Session of Rating Prediction: \", len(flat_list_7))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg7 )\n",
    "\n",
    "#let's try the same thing for another session\n",
    "relevance_met=df3.iloc[1]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_8 = []\n",
    "for sublist in relevance_met:\n",
    "    for item in sublist:\n",
    "        flat_list_8.append(item)\n",
    "        \n",
    "session_avg8=np.mean(flat_list_8, axis=0)\n",
    "maxindex = int(np.argmax(session_avg8))\n",
    "emo_ses_avg8 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Relevance Metrics: \", len(flat_list_8))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg8 )\n",
    "\n",
    "#let's try the same thing for Top-N Recommendation session \n",
    "top_recom=df3.iloc[2]\n",
    "\n",
    "#Let's disintegrate the list of lists such that it's only a list containing all emotion values of for given session\n",
    "flat_list_9 = []\n",
    "for sublist in top_recom:\n",
    "    for item in sublist:\n",
    "        flat_list_9.append(item)\n",
    "\n",
    "session_avg9=np.mean(flat_list_9, axis=0)\n",
    "maxindex = int(np.argmax(session_avg9))\n",
    "emo_ses_avg9 = emotion_labels[maxindex] \n",
    "        \n",
    "print(\"\\nTotal number of Emotions Detected in Begin Session of Top-N Recommendation: \", len(flat_list_9))\n",
    "print(\"Avg emotion for Begin session of Rating Prediction: \", emo_ses_avg9 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "56fd2cd3",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "OpenCV",
   "language": "python",
   "name": "cv_open"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}