PCA.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import librosa\n",
    "import matplotlib.pyplot as plt\n",
    "import librosa.display\n",
    "from IPython.display import Audio\n",
    "import os\n",
    "import re\n",
    "\n",
    "from keras.preprocessing.image import load_img \n",
    "from keras.preprocessing.image import img_to_array \n",
    "from keras.applications.vgg16 import preprocess_input \n",
    "\n",
    "from keras.applications.vgg16 import VGG16 \n",
    "from keras.models import Model\n",
    "\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "import pickle\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = '/Users/shoheisuzuki/workspace/euclidean-eukalyptus/work/music-recommendation/Spectogram_Images/'\n",
    "# change the working directory to the path where the images are located\n",
    "os.chdir(path)\n",
    "\n",
    "songs = []\n",
    "# creates a ScandirIterator aliased as files\n",
    "with os.scandir(path) as files:\n",
    "  # loops through each file in the directory\n",
    "    for file in files:\n",
    "        if file.name.endswith('.png'):\n",
    "          # adds only the image files to the flowers list\n",
    "            songs.append(file.name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(r'/Users/shoheisuzuki/workspace/euclidean-eukalyptus/work/music-recommendation/model/audio_features.pkl', 'rb') as file:\n",
    "    data = pickle.load(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4834, 4096)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get a list of the filenames\n",
    "filenames = np.array(list(data.keys()))\n",
    "\n",
    "# get a list of just the features\n",
    "feature = np.array(list(data.values()))\n",
    "feature.shape\n",
    "# (210, 1, 4096)\n",
    "\n",
    "# reshape so that there are 210 samples of 4096 vectors\n",
    "feature = feature.reshape(-1,4096)\n",
    "feature.shape\n",
    "\n",
    "# # get the unique labels (from the flower_labels.csv)\n",
    "# df = pd.read_csv('flower_labels.csv')\n",
    "# label = df['label'].tolist()\n",
    "# unique_labels = list(set(label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.        , 0.27688515, 0.        , ..., 1.9151807 , 0.        ,\n",
       "        0.        ],\n",
       "       [0.        , 2.6396756 , 0.        , ..., 0.        , 0.        ,\n",
       "        0.        ],\n",
       "       [0.        , 2.4752893 , 0.        , ..., 0.97800446, 0.        ,\n",
       "        0.        ],\n",
       "       ...,\n",
       "       [0.        , 2.3806176 , 0.        , ..., 1.444176  , 0.        ,\n",
       "        0.        ],\n",
       "       [0.        , 2.6879911 , 0.        , ..., 2.1829455 , 0.        ,\n",
       "        0.        ],\n",
       "       [0.        , 2.410195  , 0.7575766 , ..., 0.54020345, 0.        ,\n",
       "        0.        ]], dtype=float32)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>4086</th>\n",
       "      <th>4087</th>\n",
       "      <th>4088</th>\n",
       "      <th>4089</th>\n",
       "      <th>4090</th>\n",
       "      <th>4091</th>\n",
       "      <th>4092</th>\n",
       "      <th>4093</th>\n",
       "      <th>4094</th>\n",
       "      <th>4095</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.276885</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.525118</td>\n",
       "      <td>1.358259</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.603625</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.632590</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.574090</td>\n",
       "      <td>5.254221</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.915181</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.639676</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.443202</td>\n",
       "      <td>1.828559</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.724163</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.982517</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.223240</td>\n",
       "      <td>3.025865</td>\n",
       "      <td>4.277346</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.475289</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.703385</td>\n",
       "      <td>1.551151</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.414777</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.316081</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.689943</td>\n",
       "      <td>3.871218</td>\n",
       "      <td>4.733443</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.978004</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.155072</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.213467</td>\n",
       "      <td>2.855101</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.501060</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.347623</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.847612</td>\n",
       "      <td>4.460057</td>\n",
       "      <td>5.816047</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.786599</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.419908</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.596359</td>\n",
       "      <td>0.921858</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.341099</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.390624</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.033546</td>\n",
       "      <td>6.367900</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.749969</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</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>4829</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.134456</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.257037</td>\n",
       "      <td>1.926351</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.060061</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3.980004</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.593295</td>\n",
       "      <td>3.709148</td>\n",
       "      <td>6.243112</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.915098</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4830</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.763911</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.350312</td>\n",
       "      <td>2.407132</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.330967</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.757872</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.071882</td>\n",
       "      <td>3.476911</td>\n",
       "      <td>5.220980</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.952804</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4831</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.380618</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.237662</td>\n",
       "      <td>0.482639</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.054648</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>4.733780</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.614663</td>\n",
       "      <td>4.179696</td>\n",
       "      <td>7.504448</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.444176</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4832</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.687991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.338039</td>\n",
       "      <td>0.954078</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.751172</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>5.132657</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.285458</td>\n",
       "      <td>3.959879</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.182945</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4833</th>\n",
       "      <td>0.0</td>\n",
       "      <td>2.410195</td>\n",
       "      <td>0.757577</td>\n",
       "      <td>0.154139</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5.577830</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>3.618307</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.124246</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.146846</td>\n",
       "      <td>3.534673</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.540203</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4834 rows × 4096 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      0         1         2         3         4     5         6     7     \\\n",
       "0      0.0  0.276885  0.000000  0.525118  1.358259   0.0  5.603625   0.0   \n",
       "1      0.0  2.639676  0.000000  1.443202  1.828559   0.0  5.724163   0.0   \n",
       "2      0.0  2.475289  0.000000  1.703385  1.551151   0.0  4.414777   0.0   \n",
       "3      0.0  1.155072  0.000000  0.213467  2.855101   0.0  5.501060   0.0   \n",
       "4      0.0  1.419908  0.000000  0.596359  0.921858   0.0  6.341099   0.0   \n",
       "...    ...       ...       ...       ...       ...   ...       ...   ...   \n",
       "4829   0.0  1.134456  0.000000  1.257037  1.926351   0.0  6.060061   0.0   \n",
       "4830   0.0  0.763911  0.000000  0.350312  2.407132   0.0  5.330967   0.0   \n",
       "4831   0.0  2.380618  0.000000  1.237662  0.482639   0.0  7.054648   0.0   \n",
       "4832   0.0  2.687991  0.000000  1.338039  0.954078   0.0  4.751172   0.0   \n",
       "4833   0.0  2.410195  0.757577  0.154139  0.000000   0.0  5.577830   0.0   \n",
       "\n",
       "      8     9     ...      4086  4087      4088      4089      4090      4091  \\\n",
       "0      0.0   0.0  ...  4.632590   0.0  0.000000  0.000000  2.574090  5.254221   \n",
       "1      0.0   0.0  ...  4.982517   0.0  0.000000  0.223240  3.025865  4.277346   \n",
       "2      0.0   0.0  ...  4.316081   0.0  0.000000  0.689943  3.871218  4.733443   \n",
       "3      0.0   0.0  ...  5.347623   0.0  0.000000  0.847612  4.460057  5.816047   \n",
       "4      0.0   0.0  ...  4.390624   0.0  0.000000  0.000000  3.033546  6.367900   \n",
       "...    ...   ...  ...       ...   ...       ...       ...       ...       ...   \n",
       "4829   0.0   0.0  ...  3.980004   0.0  0.000000  0.593295  3.709148  6.243112   \n",
       "4830   0.0   0.0  ...  4.757872   0.0  0.000000  0.071882  3.476911  5.220980   \n",
       "4831   0.0   0.0  ...  4.733780   0.0  0.000000  1.614663  4.179696  7.504448   \n",
       "4832   0.0   0.0  ...  5.132657   0.0  0.000000  0.000000  2.285458  3.959879   \n",
       "4833   0.0   0.0  ...  3.618307   0.0  0.124246  0.000000  2.146846  3.534673   \n",
       "\n",
       "      4092      4093  4094  4095  \n",
       "0      0.0  1.915181   0.0   0.0  \n",
       "1      0.0  0.000000   0.0   0.0  \n",
       "2      0.0  0.978004   0.0   0.0  \n",
       "3      0.0  1.786599   0.0   0.0  \n",
       "4      0.0  0.749969   0.0   0.0  \n",
       "...    ...       ...   ...   ...  \n",
       "4829   0.0  1.915098   0.0   0.0  \n",
       "4830   0.0  1.952804   0.0   0.0  \n",
       "4831   0.0  1.444176   0.0   0.0  \n",
       "4832   0.0  2.182945   0.0   0.0  \n",
       "4833   0.0  0.540203   0.0   0.0  \n",
       "\n",
       "[4834 rows x 4096 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame(data=feature)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = StandardScaler()\n",
    "df = scaler.fit_transform(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "cov_matrix = np.cov(df, rowvar=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# sns.heatmap(cov_matrix, cmap=sns.color_palette(\"vlag\", as_cmap=True), annot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "eigenvalues, eigenvectors = np.linalg.eigh(cov_matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "sorted_index = np.argsort(eigenvalues)[::-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "sorted_eigenvalues = eigenvalues[sorted_index]\n",
    "sorted_eigenvectors = eigenvectors[:,sorted_index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3.24296043e+02,  2.00791454e+02,  1.63453403e+02, ...,\n",
       "       -1.20829556e-14, -1.83988348e-14, -2.33848754e-14])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_eigenvalues"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 3600x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams['figure.figsize'] = (50,8)\n",
    "plt.bar(range(4096), [sorted_eigenvalues[0:i+1].sum()/np.sum(sorted_eigenvalues) * 100 for i, _ in enumerate(sorted_eigenvalues)], alpha=0.5)\n",
    "plt.scatter(range(4096), sorted_eigenvalues/np.sum(sorted_eigenvalues)*100, s=90)\n",
    "plt.axhline(95, c='r', ls='--')\n",
    "plt.xticks(range(4096), [f'PC{i}' for i in range(4096)])\n",
    "plt.ylabel('PVE (%)', fontsize=15)\n",
    "plt.tick_params('both', labelsize=15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "66.70788799832154"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_eigenvalues[0:100+1].sum()/np.sum(sorted_eigenvalues) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "95.00318759977964"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_eigenvalues[0:888+1].sum()/np.sum(sorted_eigenvalues) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "79.56460368065575"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_eigenvalues[0:250+1].sum()/np.sum(sorted_eigenvalues) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "72.3942588233053"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_eigenvalues[0:150+1].sum()/np.sum(sorted_eigenvalues) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "1ee71b2a6db0d2950e77713e5ec0375ec28161e0f84c6c3416ca97bf203ce59b"
  },
  "kernelspec": {
   "display_name": "Python 3.8.12 64-bit ('music-recommendation': conda)",
   "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": "3.8.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}