utils/prepare_dataset_keras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**File name**: prepare_dataset.ipynb<br>\n",
    "**Authors**: Anton Karazeev &lt;[anton.karazeev@gmail.com](mailto:anton.karazeev@gmail.com)&gt;, Anton Lukashchuk &lt;[academik116@gmail.com](mailto:academik116@gmail.com)&gt;\n",
    "\n",
    "This file is part of REDE project (https://github.com/akarazeev/REDE)\n",
    "\n",
    "**Description**: content of this file was used to build a dataset called REDE - Reverse Engineering in Dispersion Engineering. Files from utils/matlab_data/ were used as raw data. These data contain information about system's simulations with given set of parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import interpolate\n",
    "import os\n",
    "import tqdm\n",
    "import pickle\n",
    "import scipy.io as spio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "from prepare_dataset_keras import preproc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def df_by_filepath(filepath):\n",
    "    # Load data.\n",
    "    mat = spio.loadmat(filepath, squeeze_me=True)\n",
    "\n",
    "    struct = mat['str']\n",
    "    header = ['id']\n",
    "    header.extend(struct[0][1].dtype.names)\n",
    "    header\n",
    "\n",
    "    # Create DataFrame.\n",
    "    dataset = []\n",
    "\n",
    "    for i in range(len(struct)):\n",
    "        tmp = [int(struct[i][0])]\n",
    "        tmp.extend([float(struct[i][1][name]) for name in header[1:]])\n",
    "        dataset.append(tmp)\n",
    "\n",
    "    df_data = pd.DataFrame(data=dataset, columns=header)\n",
    "    return df_data, struct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>id</th>\n",
       "      <th>gap</th>\n",
       "      <th>width1</th>\n",
       "      <th>height</th>\n",
       "      <th>radius1</th>\n",
       "      <th>width2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1000</th>\n",
       "      <td>1227</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>8.800000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1001</th>\n",
       "      <td>1228</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>9.200000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1002</th>\n",
       "      <td>1229</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>9.600000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1003</th>\n",
       "      <td>1230</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000020</td>\n",
       "      <td>1.000000e-06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1004</th>\n",
       "      <td>1231</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>8.000000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1005</th>\n",
       "      <td>1232</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>8.400000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1006</th>\n",
       "      <td>1233</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>8.800000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1007</th>\n",
       "      <td>1234</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>9.200000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1008</th>\n",
       "      <td>1235</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>9.600000e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1009</th>\n",
       "      <td>1236</td>\n",
       "      <td>4.000000e-07</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.500000e-07</td>\n",
       "      <td>0.000022</td>\n",
       "      <td>1.000000e-06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        id           gap    width1        height   radius1        width2\n",
       "1000  1227  4.000000e-07  0.000002  7.500000e-07  0.000020  8.800000e-07\n",
       "1001  1228  4.000000e-07  0.000002  7.500000e-07  0.000020  9.200000e-07\n",
       "1002  1229  4.000000e-07  0.000002  7.500000e-07  0.000020  9.600000e-07\n",
       "1003  1230  4.000000e-07  0.000002  7.500000e-07  0.000020  1.000000e-06\n",
       "1004  1231  4.000000e-07  0.000002  7.500000e-07  0.000022  8.000000e-07\n",
       "1005  1232  4.000000e-07  0.000002  7.500000e-07  0.000022  8.400000e-07\n",
       "1006  1233  4.000000e-07  0.000002  7.500000e-07  0.000022  8.800000e-07\n",
       "1007  1234  4.000000e-07  0.000002  7.500000e-07  0.000022  9.200000e-07\n",
       "1008  1235  4.000000e-07  0.000002  7.500000e-07  0.000022  9.600000e-07\n",
       "1009  1236  4.000000e-07  0.000002  7.500000e-07  0.000022  1.000000e-06"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_data, struct = df_by_filepath('matlab_data/keras_set1_1848+15-18+part_23-26.mat')\n",
    "df_data[1000:1010]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                                                           | 0/10 [00:00<?, ?it/s]C:\\Users\\lukashch\\AppData\\Local\\Continuum\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:38: DeprecationWarning: The binary mode of fromstring is deprecated, as it behaves surprisingly on unicode inputs. Use frombuffer instead\n",
      "100%|██████████████████████████████████████████████████████████████████████████████████| 10/10 [00:01<00:00,  7.14it/s]\n"
     ]
    }
   ],
   "source": [
    "# Generate dataset.\n",
    "Disp = []\n",
    "parameters_list = []\n",
    "images = []\n",
    "\n",
    "for filepath in ['matlab_data/keras_set1_1848+15-18+part_23-26.mat']:\n",
    "    df_data, struct = df_by_filepath(filepath)\n",
    "\n",
    "#     for i in tqdm.tqdm(range(len(struct))):\n",
    "    for i in tqdm.tqdm(range(10)):\n",
    "        # Parameters.\n",
    "        sample_id = int(struct[i][0])\n",
    "        parameters = df_data[df_data['id'] == sample_id].values[0][1:]\n",
    "        ## get rid off bad samples with small radius\n",
    "        if parameters[3]<17e-6:\n",
    "            continue\n",
    "        parameters_list.append(parameters)\n",
    "\n",
    "        # Frequencies and modes.\n",
    "        freqs, modes = struct[i][2][:, 0].real, struct[i][2][:, 2].real\n",
    "#         frequencies_modes_list.append((freqs, modes))\n",
    "\n",
    "        # Images.\n",
    "        omega_total, delta_omega_total, freq_grid, Dint = preproc(freqs, modes)\n",
    "        # Frequencies and modes write to the array to save them\n",
    "#         freqs, modes = struct[i][2][:, 0].real, struct[i][2][:, 2].real\n",
    "        Disp.append((freq_grid, Dint))\n",
    "        \n",
    "        fig = plt.figure(figsize=(2, 1))\n",
    "        fig.add_subplot(111)\n",
    "        plt.xlim((150, 400))\n",
    "        plt.ylim((-500, 500))\n",
    "        plt.axis('off')\n",
    "        img = plt.scatter(omega_total * 1e-12, delta_omega_total * 1e-9, s=0.01)\n",
    "        fig.canvas.draw()\n",
    "\n",
    "        # Images -> arrays -> black and white.\n",
    "        data = np.fromstring(fig.canvas.tostring_rgb(), dtype=np.uint8, sep='')\n",
    "        data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))\n",
    "        data = data[5:-5, 20:-13, 0]\n",
    "        data = (data < 200) * 255\n",
    "        data = data.astype('uint8')\n",
    "        images.append(data)\n",
    "        plt.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5070"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(images)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4847, 3073,  346, 1706, 3268, 4485, 3885, 4607, 4734, 3770])"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = np.random.randint(1,len(images),10)\n",
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAADdCAYAAACrB4nBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAADiNJREFUeJzt3X+sX3V9x/Hna20pgiNQB6S2ZGDSOMkSwTTIxrIw0A1/xPIHLBizNUuT/uM23EwUt79M/EOTRdgSg2nE2S1OYPiDhjgdqxBjsiEXYQhWLTIGXTuKExT3R231vT++p1mF+/V77v3+aj/f5yO5+X7Puefb8z49977u577v55xvqgpJ0qnvl+ZdgCRpMgx0SWqEgS5JjTDQJakRBrokNcJAl6RGGOiS1IixAj3JNUm+k+SJJDdNqihJ0spltRcWJVkDfBd4M3AAeBB4Z1V9a3LlSZL6WjvGay8DnqiqJwGS3A5sA4YG+mlZX6dz5hi7lKTF8yLPf7+qzh213TiBvgl45oTlA8AbX7pRkp3AToDTOYM35uoxdilJi+df6q7/7LPdOD30LLPuZf2bqtpVVVuraus61o+xO0nSLzJOoB8ALjhheTNwcLxyJEmrNU6gPwhsSXJRktOAG4A9kylLkrRSq+6hV9WxJH8MfBlYA3yyqh6fWGWSpBUZ54+iVNUXgS9OqBZJ0hi8UlSSGmGgS1IjDHRJaoSBLkmNMNAlqREGuiQ1wkCXpEYY6JLUCANdkhphoEtSIwx0SWqEgS5JjTDQJakRBrokNcJAl6RGGOiS1AgDXZIaYaBLUiMMdElqhIEuSY0w0CWpESMDPcknkxxO8tgJ6zYkuTfJ/u7xnOmWKUkapc8I/VPANS9ZdxOwt6q2AHu7ZUnSHI0M9Kr6KvCDl6zeBuzunu8Grp1wXZKkFVptD/38qjoE0D2eN2zDJDuTLCVZOsqRVe5OkjTK1P8oWlW7qmprVW1dx/pp706SFtZqA/3ZJBsBusfDkytJkrQaqw30PcD27vl24O7JlCNJWq0+0xY/A/wr8NokB5LsAD4MvDnJfuDN3bIkaY7Wjtqgqt455FNXT7gWSdIYvFJUkhphoEtSIwx0SWqEgS5JjTDQJakRBrokNcJAl6RGGOiS1AgDXZIaYaBLUiMMdElqhIEuSY0w0CWpEQa6JDXCQJekRhjoktQIA12SGmGgS1IjDHRJaoSBLkmNMNAlqREGuiQ1YmSgJ7kgyX1J9iV5PMmN3foNSe5Nsr97PGf65UqShukzQj8GvLeqXgdcDrw7ycXATcDeqtoC7O2WJUlzMjLQq+pQVX2je/4isA/YBGwDdneb7QaunVaRkqTRVtRDT3IhcCnwAHB+VR2CQegD5w15zc4kS0mWjnJkvGolSUP1DvQkrwQ+C7ynqn7U93VVtauqtlbV1nWsX02NkqQeegV6knUMwvzTVfW5bvWzSTZ2n98IHJ5OiZKkPvrMcglwG7Cvqj56wqf2ANu759uBuydfniSpr7U9trkC+APgm0ke6db9BfBh4M4kO4CngeunU6IkqY+RgV5VXwMy5NNXT7YcSdJqeaWoJDXCQJekRhjoktQIA12SGmGgS1IjDHRJaoSBLkmNMNAlqREGuiQ1wkCXpEYY6JLUCANdkhphoEtSIwx0SWqEgS5JjTDQJakRBrokNcJAl6RGGOiS1AgDXZIaYaBLUiNGBnqS05N8Pcm/J3k8yQe79RcleSDJ/iR3JDlt+uVKkoZZ22ObI8BVVfXjJOuAryX5J+DPgZur6vYkHwd2ALdOsVZpqC8ffGTeJYzl9159ybxLUANGjtBr4Mfd4rruo4CrgLu69buBa6dSoSSpl1499CRrkjwCHAbuBb4HvFBVx7pNDgCbhrx2Z5KlJEtHOTKJmiVJy+jTcqGqfgpckuRs4PPA65bbbMhrdwG7AM7KhmW3kX6RU72d0sekjtHWzWJb0SyXqnoBuB+4HDg7yfEfCJuBg5MtTZK0En1muZzbjcxJ8grgTcA+4D7gum6z7cDd0ypSkjRan5bLRmB3kjUMfgDcWVX3JPkWcHuSDwEPA7dNsU4tgEVorUzbSv8PbdG0ZWSgV9WjwKXLrH8SuGwaRUmSVs4rRSWpEb1muUiTNO3WyqnSRjgZWky2aNriCF2SGmGgS1IjbLloJrxw5uUmdSyzbN2sZl8tnbOTnSN0SWqEgS5JjbDloqkZpxXgr+n9rfT/ataza4btz3M8eY7QJakRBrokNcKWiybKC1VOfidLi6bPv+vXx8o4QpekRhjoktQIWy4am22Wts2zRWNbZmUcoUtSIwx0SWqELRetim0WDdP3XE+qNeOFS//PEbokNcJAl6RG2HJRb7ZZNEnDvj5sxayeI3RJakTvQE+yJsnDSe7pli9K8kCS/UnuSHLa9MqUJI2ykpbLjcA+4Kxu+SPAzVV1e5KPAzuAWydcn+bMNotmrc/X0DhtmZZbMb1G6Ek2A28DPtEtB7gKuKvbZDdw7TQKlCT107flcgvwPuBn3fKrgBeq6li3fADYtNwLk+xMspRk6ShHxipWkjTcyJZLkrcDh6vqoSRXHl+9zKa13OurahewC+CsbFh2G53aWvhVVaeWacyQaaEV06eHfgXwjiRvBU5n0EO/BTg7ydpulL4ZODi9MiVJo4xsuVTVB6pqc1VdCNwAfKWq3gXcB1zXbbYduHtqVUqSRhrnwqL3A7cn+RDwMHDbZErSvM36TYSlSVn0VsyKAr2q7gfu754/CVw2+ZIkSavhlaKS1Ajv5aJVORl/3ZSGWZRWjCN0SWqEgS5JjbDlIsCZLVpM027FzLr94ghdkhphoEtSI2y5qDdntmhRTKoVM+uZMI7QJakRBrokNcKWywJzZou0MtNoxUyy/eIIXZIaYaBLUiMMdElqhD10/UJOVZRGO/H7ZJ5TGx2hS1IjDHRJaoQtlwXjVEVpuqYxtXHNxn6vcYQuSY0w0CWpEbZc9DLObJEmb5yZMH31CvQkTwEvAj8FjlXV1iQbgDuAC4GngN+vquenUqUkaaSVtFx+p6ouqaqt3fJNwN6q2gLs7ZYlSXMyTstlG3Bl93w3cD/w/jHr0RQ4s0U6uUzjre+g/wi9gH9O8lCSnd2686vqEED3eN5yL0yyM8lSkqWjHBmrWEnScH1H6FdU1cEk5wH3Jvl23x1U1S5gF8BZ2VCrqFGS1EOvQK+qg93j4SSfBy4Dnk2ysaoOJdkIHJ5inZoyZ7ZI8zf8+/CJXq8f2XJJcmaSXz7+HPhd4DFgD7C922w7cHevPUqSpqLPCP184PNJjm//D1X1pSQPAncm2QE8DVw/vTIlSaOMDPSqehJ4/TLr/we4ehpFSZJWzkv/JakRBrokNcJ7uTTKi4mkxeMIXZIaYaBLUiNsuSwwLyaS2uIIXZIaYaBLUiMMdElqhIEuSY0w0CWpEc5yaYgXE0mLzRG6JDXCQJekRthyWTBeTCS1yxG6JDXCQJekRhjoktQIA12SGmGgS1IjnOXSEGewSIut1wg9ydlJ7kry7ST7kvxGkg1J7k2yv3s8Z9rFSpKG69ty+WvgS1X1a8DrgX3ATcDeqtoC7O2WJUlzMjLQk5wF/DZwG0BV/aSqXgC2Abu7zXYD106rSEnSaH1G6K8BngP+NsnDST6R5Ezg/Ko6BNA9nrfci5PsTLKUZOkoRyZWuCTp5/UJ9LXAG4Bbq+pS4H9ZQXulqnZV1daq2rqO9assU5I0Sp9APwAcqKoHuuW7GAT8s0k2AnSPh6dToiSpj5GBXlX/DTyT5LXdqquBbwF7gO3duu3A3VOpUJLUS9956H8CfDrJacCTwB8x+GFwZ5IdwNPA9dMpUZLUR6pqdjtLnmPQg//+zHZ6cvgVFuuYPd72Ldoxz/t4f7Wqzh210UwDHSDJUlVtnelO52zRjtnjbd+iHfOpcrzey0WSGmGgS1Ij5hHou+awz3lbtGP2eNu3aMd8ShzvzHvokqTpsOUiSY0w0CWpETMN9CTXJPlOkieSNHe73SQXJLmvu2f840lu7NY3fe/4JGu6G7fd0y1flOSB7njv6C5Ia8aivT9Akj/rvp4fS/KZJKe3do6TfDLJ4SSPnbBu2XOagb/pcuzRJG+YX+U/b2aBnmQN8DHgLcDFwDuTXDyr/c/IMeC9VfU64HLg3d0xtn7v+BsZ3CP/uI8AN3fH+zywYy5VTc/CvD9Akk3AnwJbq+rXgTXADbR3jj8FXPOSdcPO6VuALd3HTuDWGdU40ixH6JcBT1TVk1X1E+B2BvdUb0ZVHaqqb3TPX2Twjb6Jhu8dn2Qz8DbgE91ygKsY3MQN2jveRXx/gLXAK5KsBc4ADtHYOa6qrwI/eMnqYed0G/B3NfBvwNnHb1Q4b7MM9E3AMycsH+jWNSnJhcClwAP0vHf8KeoW4H3Az7rlVwEvVNWxbrm18zzW+wOcaqrqv4C/YnC/pkPAD4GHaPscHzfsnJ60WTbLQM8y65qcM5nklcBngfdU1Y/mXc+0JHk7cLiqHjpx9TKbtnSex3p/gFNN1zfeBlwEvBo4k0HL4aVaOsejnLRf47MM9APABScsbwYOznD/M5FkHYMw/3RVfa5b3eq9468A3pHkKQYttKsYjNjP7n49h/bO86K9P8CbgP+oqueq6ijwOeA3afscHzfsnJ60WTbLQH8Q2NL9dfw0Bn9Y2TPD/U9d1z++DdhXVR894VNN3ju+qj5QVZur6kIG5/MrVfUu4D7gum6zZo4XFvL9AZ4GLk9yRvf1ffx4mz3HJxh2TvcAf9jNdrkc+OHx1szcVdXMPoC3At8Fvgf85Sz3PaPj+y0Gv3o9CjzSfbyVQV95L7C/e9ww71qncOxXAvd0z18DfB14AvhHYP2865vwsV4CLHXn+QvAOS2fY+CDwLeBx4C/B9a3do6BzzD4G8FRBiPwHcPOKYOWy8e6HPsmgxlAcz+GqvLSf0lqhVeKSlIjDHRJaoSBLkmNMNAlqREGuiQ1wkCXpEYY6JLUiP8DUNbQd1hbXYYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc0a66ebe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc129d3ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc0d48bf28>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc0d3714a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc0a66e400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc08fce470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc049bef60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc04928cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bc048370f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2bbfe938278>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for i in range(172, 180):\n",
    "for i in a:\n",
    "    plt.imshow(images[i])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert to np.array.\n",
    "images = np.array(images)\n",
    "Disp = np.array(Disp)\n",
    "parameters_list = np.array(parameters_list)\n",
    "\n",
    "with open('C:/Users/lukashch/Documents/GitHub/REDE/data/rede/raw/{}-5-parameters.pkl'.format(len(images)), 'wb') as f:\n",
    "    pickle.dump(parameters_list, f)\n",
    "\n",
    "with open('C:/Users/lukashch/Documents/GitHub/REDE/data/rede/raw/{}-frequencies_modes.pkl'.format(len(images)), 'wb') as f:\n",
    "    pickle.dump(Disp, f)\n",
    "\n",
    "with open('C:/Users/lukashch/Documents/GitHub/REDE/data/rede/raw/{}-62-111-images.pkl'.format(len(images)), 'wb') as f:\n",
    "    pickle.dump(images, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAm0AAAFNCAYAAABST1gVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XmcXXV9//HX2xBxVCCt4JIgi1vqbmpArK3FNWpbSflp3SsupcvPtTYq1Z/Vags21lq3KlULVsU1plHRiEVcqxIIGhFiKaAw0YJiQGVUiJ/fH+cM3kzuzNxJ5s6dO/N6Ph73MWf5nnM+9565937udzknVYUkSZLmt5sNOgBJkiRNz6RNkiRpCJi0SZIkDQGTNkmSpCFg0iZJkjQETNokSZKGgEmbNM8kOSJJJdmvh7LHJrlyLuKaK0kuT/LwHsuekOSLs3DMv07yjn3Y/sIkx+5rHF32+8kkT+/Dfp+S5NOzvV9J/WXSJu2DNsH4RZKDJyy/oE28jhhMZHNvJsnmfFNVf19Vz+6lbJLTkrxmwvb3rKpz+hDXo6vq9L3ZNslvJ/lykmuTXJPkS0mOavf73qp6ZEfZSrItyc06lr0myWn7/CQGZJj/H6XJmLRJ++4y4EnjM0nuDYwMLpz5yy/QuZHkQODjwJuAXwdWAK8Cfj7FZsuBJ+7l8RbceV2Iz0nDz6RN2nf/Dvxxx/zTgXd3FkhyUJJ3J7k6yXeSvHy8ViPJkiSvS/KDJJcCvzdh22ckuSjJj5NcmuRPewkqjX9KclVb2/KNJPdq152W5G1Jzmr3+7kkh3ds+xvtumuSbE/yRx3rRpL8Y/s8rk3yxSQjwOfbIjuT/CTJA9vmyy+1cVwDvDLJnZOcneSH7XN+b5JlPT6n2yTZlOS6JF8D7jxhfde4kxyT5PtJlnSU/cMk32inX5nkPR3rPtSWvzbJ55Pcs11+IvAU4MXtc/xYu/ymJt0k+yd5Q5Id7eMNSfZv1x2b5MokL2rPy/eSPGOK53tOkme30ye0r/XrkvwoyWVJHj3JpncDqKozqmpXVY1V1aer6hud+5qwzT8Ar+olWel4Hi9J8n3g39rlv5+mlnlnmlq++3Rs85Iko+3/2/YkD9vX1yvJ7yXZ2v4/XJHklR1hdvt/vFma99532v29O8lB7b7Ga+aeleS7wNnTvQ7SXDNpk/bdV4ADk9y9TQqeALxnQpk3AQcBdwJ+lybJG//y+RPg94FVwGrgcRO2vapdf2C7zT8l+c0e4nok8GCaL/BlbVw/7Fj/FODVwMHABcB7AZLcCjgLeB9wW5paxLeOJy7A64D7A79FU4vzYuCX7bEAllXVravqv9r5BwCXtvv6OyDAyTQ1O3cH7gi8sofnA/AW4GfAHYBntg+mi7uqvgL8FHhox76e3Jbt5pPAXdv9nD/+2lTVqe30P7TP8Q+6bPsy4BjgfsB9gaOBl3esvz3N/8IK4FnAW5L8Wo/P/wHAdppz9g/AO5OkS7lvA7uSnJ7k0T3ufwNwHXBCj7Hcnub8Hw6c2P5Pvgv4U+A2wNuBTW1SthJ4DnBUVR0ArAEub/ezL6/XT2neS8tofuz8eZK17bpu/48ntI+H0LwXbw28ecLz+l2a/8s1Pb4O0typqgX5oPnwuAr4Zg9lH0zzwXwj8LiO5fcD/gu4EPgG8IRBPy8f8+tB88XzcJovmZOBR9EkDvsBBRwBLKFplrpHx3Z/CpzTTp8N/FnHuke22+43yTE3As9vp48Frpyk3ENpvryPAW42Yd1pwPs75m8N7KJJoJ4AfGFC+bcDf0PzQ28MuG+X4x0xMW6aL8jvTvMargW2TnxNu5RbAtwA/EbHsr8HvthOTxp3O/0a4F3t9AE0X/iHt/OvBN4zSXzL2ud1UMdr95pu/wft9P8Aj+lYtwa4vON8jU14ja4Cjpnk2OcAz+54LS/pWHfLNq7bT7Lt3dtYr6T5bNsE3K5jX1/sKFvAXYDHAN8F9m9fr9Mm2fexwC+AW3Qs+xfg1RPKbadJgu7SPs+HA0snlJnN1+sNwD9N8f/4n8BfdMyvbP+n9usof6d9+Uzw4aOfj4Vc03YazRdoL75L8yE28Vf39cAfV9U92329oddmHC06/05Tc3MCE5pGaWpFbg58p2PZd2hqDqCpcbpiwrqbtDUlX2mb/HbSfLHuNvChm6o6m6YW4S3A/yY5NU1fp3FXdJT9CXBNG8vhwAPaJq6d7TGfQlPjcTBwC5ov2l51PjeS3DbJ+9umsutoaiWnfT7AITRfrpO9VlPFDc37+/i26e144Pyq2u21buNbkuSUJP/Txnd5u6qXGKF5DSee6+Ud8z+sqhs75q+nSZp78f3xiaq6vp3sum1VXVRVJ1TVocC92hjeMNXOq+pMms/DE3uI5eqq+lnH/OHAiya8/ncEllfVJcALaJLjq9rzP/6a7PXrleQBST6bptvBtcCfMfV56nas/YDbdSzb7f9Vmk8WbNJWVZ+n+RK6SZq+NJ9Kcl6SLyT5jbbs5dX09fjlhH18u6r+u53eQfML75C5eQYaJu2X/2U0CdWGCat/QPNr/vCOZYcBo+3092i+3DrXAU1/H+AjNE2St6uqZcCZNE2MvcT1xqq6P3BPmmbSdR2rbzpmklvTNHXtoPnS+lxVLet43Lqq/rx9Lj9jQl+y8cNNFsaE+ZPbZfepqgOBp/b4fK6mqTHq+lpNEzdV9S2aL+lHM3XT6JOB42hqhQ6iqYGhI8bJnue4Hex5rndMs01fVdXFND9k79VD8ZfTNFnecrrdTpi/Avi7Ca//LavqjDaG91XVb9O8NgW8tt1uX16v99HUIN6xqg4C3sbU56nbsW4E/neK5yXNGws2aZvEqcBz2y+xvwLe2uuGSY6mqS2ZSQ2DFpdnAQ+tqp92LqyqXcAHgb9LckCaDv9/ya/6vX0QeF6SQ9u+Oi/t2PzmNE1VVwM3th3PH0kPkhzV1kQspWkK/BlNE+i4x6S5LMTNafq2fbWqrqAZdXi3JE9LsrR9HJXk7lX1S5quB69PsrytlXpgm1xeTfPD507ThHYA8BOaDuIr2D2RnFT7Om6gGcxwyyT3oBn0MW7SuDvKvA94Hk2XiA9NEd/Pafr/3ZKmCbbT/zL1czwDeHmSQ9JcCuYV7NnHsa/SDMh4UZJD2/k70vTx+8p021Zz6ZJt7P7a9uJfgT9r/+eS5FbtQIEDkqxM8tD2/+RnNE2e4/+L+/J6HQBcU1U/az+jn9yxrtv/4xnAC5Mc2f5Q+XvgAxNq8qR5a9Ekbe0b9LeADyW5gKavyx163PYONM1fz2i/tKQ9VNX/VNWWSVY/lyZxuhT4Ik3y8K523b8Cm4Gv0/StvKmmrqp+TJNkfBD4Ec2X0qYeQzqw3fePaGqYfkhTYzfufTT91K6hGVjwlI5jPpLm8g87aJrkXkuTPELzg2cbcG677Wtp+sxdTzPQ4Ett89gxk8T1KuA3gWuBT7BnzeRUnkPTNPZ9mpqjfxtf0UPc0HxpHwucXVU/mOQY76Z5vUaBb7FnovNO4B7tc9zYZfvXAFto+sFuozmnr+lSrp9+TDNo4atJfkrzHL4JvKjH7V9OU/Pas/Z//09omuR/BFzCrwY17A+cQlNT+32aAR5/3a7bl9frL4C/TfJjmmTvgx3xdPt/fBfNZ/nnaWrGf0bz3pSGQqoWbk1wmgubfryq7tX25dleVZMmamkuJPnxqvpwx7IDaToDn1xVk/0yl4ZK+79+ZVW9fLqykqT5YdHUtFXVdcBlSR4PN13D6r5TbdM2G30UeLcJmyRJGqQFm7QlOYPmch0r01yc8Vk0zT/PSvJ1mst4HNeWPSrN/RsfD7w9yYXtbv6Ipu/LCWkuGHlBkvvN+ZORJEmL3oJuHpUkSVooFmxNmyRJ0kJi0iZJkjQEpr0x8DA6+OCD64gjjhh0GJIkSdM677zzflBV0168f0EmbUcccQRbtkx2uSxJkqT5I8ket9TrxuZRSZKkIWDSJkmSNARM2iRJkoaASZskSdIQMGmTJEkaAiZtkiRJQ8CkTZIkaQgM9DptSR4F/DOwBHhHVZ0yYf3+wLuB+wM/BJ5QVZfPdZySFoeNW0dZv3k7O3aOsXzZCOvWrGTtqhX7VHa2y/Vrn5Lmv4ElbUmWAG8BHgFcCZybZFNVfauj2LOAH1XVXZI8EXgt8IS5j1bSvprvycvGraOctGEbYzfsAmB05xgnbdgGsNdlZ7tcv/Y526/lTMoN+tjSMBlk8+jRwCVVdWlV/QJ4P3DchDLHAae30x8GHpYkcxijpGls3DrKg045myNf+gkedMrZbNw62rXMSRu2MbpzjOJXCcTEsr2W68c+12/eflOCM27shl2s37x9j2P3Wna2y/Vrn7P9Wg7yPM7k2OPlp/v/nUk5qZ8GmbStAK7omL+yXda1TFXdCFwL3KbbzpKcmGRLki1XX311H8KVFo+ZfJHNZkI0yORlx86xrs+x2/Jey852uX7tcyEloYNMVsfL9prcmQhqpgaZtHWrMau9KNMsrDq1qlZX1epDDpn2nqvSojSbtWIw+wnRIJOX5ctGupbrtrzXsrNdrl/7XEhJ6CCT1X7UMI6XNbkTDDZpuxK4Y8f8ocCOycok2Q84CLhmTqKThshsJmMzqamY7YRokMnLujUrGVm6ZLdlI0uXsG7Nyj227bXsbJfr1z4XUhI6yGS1HzWMJnfqNMik7VzgrkmOTHJz4InApgllNgFPb6cfB5xdVV1r2qTFaraTsZnUVMx2QjTI5GXtqhWcfPy9WbFshAArlo1w8vH37tqBvdeys12uX/tcSEnoIJPVftQw9iO5Gy9vgjd8BjZ6tKpuTPIcYDPNJT/eVVUXJvlbYEtVbQLeCfx7kktoatieOKh4pbnW6wi4qT7UO8vPpFZstMcEbd2albuNUITJE6LxWKd6Pr2W6+c+ex1l2GvZ2S7Xr2PD7L2WgzyPMzl2r/+/vZabyXun17Kzkdzt60hpR+LOH1mIFVerV6+uLVu2DDoMaa9N/FCF5kuiW03JkS/9RNeOngEuO+X3bpp/0Clnd/2SWLFshC+99KF7dezx8n6oa1j187IxMPl7p9eyvb5ve/0cmMk+Z/p8/BzYe0nOq6rV05Ub6MV1JXU3k1/Nvf5in+1asc7yfjhrWM1mrWU/ahj7Ucs327V3/bomoPZk0ibNsV4+sGbSN6YfyZiJmLR3ZruZe7aTOxieplntyaRNmkO9fmDN5FezyZi0sM12Ld9s197N1mVWbHKdnkmbNId6/cCaya9mMBmTNPsDTwbZNGuTa3cmbdIsmc1mz5n2K5OkmZjvTbM2uXZn0ibNgn41ey60DxxJw2VQTbM2uXZn0ibNgn41e0rSMJjtptn50OQ6H5m0SbPAZk9J6k0vCd58aHKdj5/TJm3SLLDZU5JmzyCbXOdzjZxJmzSFXn9t2ewpSbNrUE2uM6mRm2smbdIkZvJry2ZPSRqc2WxynckgiLlm0iZNYqa/tmz2lKT5qx+DIOaaSZs0ifn8a0uSNHOzPQhirt1s0AFI89Vkv6rmw68tSVJ/rF21gpOPvzcrlo0QYMWyEU4+/t7zoiXFmjYtSr0MMJjPv7YkSf0zX7u7mLRp0el1gIGDCyRJ84lJmxadmQwwmK+/tiRJi4992rToOMBAkjSMTNq06DjAQJI0jEzatOisW7OSkaVLdlvmAANJ0nxnnzYtOg4wkCQNI5M2LRi93icUHGAgSRo+Jm1aEGZyn1BJkoaRfdq0IEx1GQ9JkhYCkzYtCF7GQ5K00Jm0aUHwMh6SpIXOpE0LgpfxkCQtdA5E0ILgZTwkSQudSZsWDC/jIUlayGwelSRJGgLWtGnem8lFcyVJWqhM2jSvedFcSZIaNo9qXvOiuZIkNUzaNK950VxJkhombZrXvGiuJEkNkzbNa140V5KkhgMRNK950VxJkhombZr3vGiuJEk2j0qSJA0FkzZJkqQhYNImSZI0BOzTpoHx9lSSJPXOpE0D4e2pJEmamYE0jyb59SRnJfnv9u+vTVJuV5IL2semuY5T/ePtqSRJmplB9Wl7KfCfVXVX4D/b+W7Gqup+7eOxcxee+s3bU0mSNDODStqOA05vp08H1g4oDg2It6eSJGlmBpW03a6qvgfQ/r3tJOVukWRLkq8kMbFbQLw9lSRJM9O3gQhJPgPcvsuql81gN4dV1Y4kdwLOTrKtqv5nkuOdCJwIcNhhh804Xs0tb08lSdLMpKrm/qDJduDYqvpekjsA51TVlFUsSU4DPl5VH55u/6tXr64tW7bMTrCSJEl9lOS8qlo9XblBNY9uAp7eTj8d+I+JBZL8WpL92+mDgQcB35qzCCVJkuaRQSVtpwCPSPLfwCPaeZKsTvKOtszdgS1Jvg58FjilqkzaJEnSojSQi+tW1Q+Bh3VZvgV4djv9ZeDecxyaJEnSvOS9RyVJkoaASZskSdIQ8N6jmlXeBF6SpP4wadOs8SbwkiT1j82jmjXeBF6SpP4xadOs8SbwkiT1j0mbZo03gZckqX9M2jRrvAm8JEn940AEzRpvAi9JUv+YtGlWrV21wiRNkqQ+sHlUkiRpCJi0SZIkDQGTNkmSpCHQc5+2JL8GLAfGgMur6pd9i0qSJEm7mTJpS3IQ8H+BJwE3B64GbgHcLslXgLdW1Wf7HqUkSdIiN11N24eBdwO/U1U7O1ckuT/wtCR3qqp39itASZIkTZO0VdUjplh3HnDerEckSZKkPfQ0ECHJfyZ5zIRlp/YnJEmSJE3U60CEI4GXJDmqql7VLlvdp5g0D23cOuqdDiRJGqBeL/mxE3gYzQCEj7UDFLRIbNw6ykkbtjG6c4wCRneOcdKGbWzcOjro0CRJWjR6TdpSVTdW1V8AHwG+CNy2f2FpPlm/eTtjN+zabdnYDbtYv3n7gCKSJGnx6bV59G3jE1V1WpJtNJcC0SKwY+fYjJZLkqTZN9112n69nfxQxzTAZcBf9S0qzSvLl40w2iVBW75sZADRSJK0OE3XPHoesKX9u6Njeny5FoF1a1YysnTJbstGli5h3ZqVA4pIkqTFZ7rrtB05Pp1ka1Wt6n9Imm/GR4k6elSSpMHp+d6jQPUtCs17a1etMEmTJGmAeh09KkmSpAGabiDCX3bM3nbCPFX1+r5EJUmSpN1M1zx6QMf0v06YlyRJ0hyZbiDCq6ZaL0mSpLkxXfPoPYE7V9Wmdv6fgPFbWL25qs7vc3ySJEli+oEIpwA/6JhfA3wC+Czwin4FJUmSpN1N16ftDlX15Y7566rqIwBJ/rR/YUmSJKnTdDVtuw08qKpjOma9YbwkSdIcmS5p25HkARMXJjmG5rZWkiRJmgPTNY++BPhAktOA8UEH9weeDjyhj3FJkiSpw5Q1bVX1NeABwBLghPZxM+CYdp0kSZLmwLT3Hq2qq3Ck6IK1ceuoN4KXJGkITFnTluRjSf4gydIu6+6U5G+TPLN/4amfNm4d5aQN2xjdOUYBozvHOGnDNjZuHR10aJIkaYLpBiL8CfA7wMVJzk1yZpKzk1wKvB04r6re1fco1RfrN29n7IZduy0bu2EX6zdvH1BEkiRpMtPdxur7wIuBFyc5ArgDMAZ8u6qu73t06qsdO8dmtFySJA3OtH3axlXV5cDlfYtEc275shFGuyRoy5eNDCAaSZI0lemaR7WArVuzkpGlS3ZbNrJ0CevWrBxQRJIkaTI917Rp4RkfJeroUUmS5r+BJG1JHg+8Erg7cHRVbZmk3KOAf6a5Ttw7quqUOQtykVi7aoVJmiRJQ6Cn5tEkD0pyVpJvJ7k0yWXtCNK99U3geODzUxxzCfAW4NHAPYAnJbnHPhxTkiRpaPVa0/ZO4IXAecCuacpOq6ouAkgyVbGjgUuq6tK27PuB44Bv7evxJUmShk2vSdu1VfXJvkaypxXAFR3zV9LcUkuSJGnR6TVp+2yS9cAG4OfjC6vq/Mk2SPIZ4PZdVr2sqv6jh2N2q4arKY53InAiwGGHHdbD7iVJkoZHr0nbeA3X6o5lBTx0sg2q6uF7G1TrSuCOHfOHAjumON6pwKkAq1evnjS5kyRJGkY9JW1V9ZB+B9LFucBdkxwJjAJPBJ48gDgkSZIGrtfRowcleX2SLe3jH5MctLcHTfKHSa4EHgh8IsnmdvnyJGcCVNWNwHOAzcBFwAer6sK9PaYkSdIwS9X0LYlJPkJzmY7T20VPA+5bVcf3Mba9tnr16tqypeul3yRJkuaVJOdV1erpyvXap+3OVfV/OuZfleSCvQtNkiRJM9XrvUfHkvz2+EySBwF73mlckiRJfdFrTdufA6e3/dgCXAOc0K+gJEmStLteR49eANw3yYHt/HV9jUqSJEm7mTJpS/LUqnpPkr+csByAqnp9H2OTJElSa7qatlu1fw/oss4L2EqSJM2RKZO2qnp7O/mZqvpS57p2MILmoY1bR1m/eTs7do6xfNkI69asZO2qFYMOS5Ik7YNeR4++qcdlGrCNW0c5acM2RneOUcDozjFO2rCNjVtHBx2aJEnaB9P1aXsg8FvAIRP6tR0ILOlnYNo76zdvZ+yGXbstG7thF+s3b7e2TZKkITZdn7abA7duy3X2a7sOeFy/gtLe27Gz++XzJlsuSZKGw3R92j4HfC7JaVX1nTmKSftg+bIRRrskaMuXjQwgGkmSNFt67dN2fZL1Sc5Mcvb4o6+Raa+sW7OSkaW7t1yPLF3CujUrBxSRJEmaDb0mbe8FLgaOBF4FXA6c26eYtA/WrlrBycffmxXLRgiwYtkIJx9/b/uzSZI05Hq9jdVtquqdSZ7f0WT6uX4Gpr23dtUKkzRJkhaYXpO2G9q/30vye8AO4ND+hCRJkqSJek3aXtPeLP5FNNdnOxB4Yd+ikiRJ0m56vWH8x9vJa4GH9C8cSZIkdTPdxXVfXFX/kORNdLnXaFU9r2+RSZIk6SbT1bRd1P7d0u9AJEmSNLnpLq77sXbyG1W1dQ7ikSRJUhe9Xqft9UkuTvLqJPfsa0SSJEnaQ09JW1U9BDgWuBo4Ncm2JC/vZ2CSJEn6lV5r2qiq71fVG4E/Ay4AXtG3qCRJkrSbnpK2JHdP8sokFwJvBr6MF9eVJEmaM71eXPffgDOAR1TVjj7GI0mSpC56vbjuMUlGgMP6HI8kSZK66LV59A9o+rF9qp2/X5JN/QxMkiRJv9LrQIRXAkcDOwGq6gLgiP6EJEmSpIl6TdpurKpr+xqJJEmSJtXrQIRvJnkysCTJXYHn0YwglSRJ0hzotabtucA9gZ/TjCK9DnhBv4KSJEnS7nodPXo98LL2IUmSpDk2ZdKW5GNATba+qh476xFpUhu3jrJ+83Z27Bxj+bIR1q1ZydpVKwYdliRJmgPT1bS9rv17PHB74D3t/JOAy/sUk7rYuHWUkzZsY+yGXQCM7hzjpA3bAEzcJElaBKZM2qrqcwBJXl1VD+5Y9bEkn+9rZNrN+s3bb0rYxo3dsIv1m7ebtEmStAj0OhDhkCR3Gp9JciRwSH9CUjc7do7NaLkkSVpYer3kxwuBc5Jc2s4fAZzYl4jU1fJlI4x2SdCWLxsZQDSSJGmu9VTTVlWfAu4KPL99rKyqT/czMO1u3ZqVjCxdstuykaVLWLdm5YAikiRJc6nXmjaq6ufA1/sYi6Yw3m/N0aOSJC1OPSdtGry1q1aYpEmStEj1OhBBkiRJA9RT0pbGU5O8op0/LMnR/Q1NkiRJ43qtaXsr8ECai+oC/Bh4S18ikiRJ0h567dP2gKr6zSRbAarqR0lu3se4JEmS1KHXmrYbkiyhvQ9pkkOAX+7tQZM8PsmFSX6ZZPUU5S5Psi3JBUm27O3xJEmShl2vNW1vBD4K3DbJ3wGPA16+D8f9Js39TN/eQ9mHVNUP9uFYkiRJQ6+npK2q3pvkPOBhQIC1VXXR3h50fNske7sLSZKkRWXKpC3Jr3fMXgWc0bmuqq7pV2CtAj6dpIC3V9WpfT6eJEnSvDRdTdt5NIlTgMOAH7XTy4DvAkdOtmGSzwC377LqZVX1Hz3G96Cq2pHktsBZSS6uqs9PcrwTae+Hethhh/W4e0mSpOEwZdJWVUcCJHkbsKmqzmznHw08fJptp1zfi6ra0f69KslHgaOBrklbWwt3KsDq1atrX48tSZI0n/Q6evSo8YQNoKo+Cfxuf0JqJLlVkgPGp4FH0gxgkCRJWnR6Tdp+kOTlSY5IcniSlwE/3NuDJvnDJFfSXLD3E0k2t8uXJxlPDm8HfDHJ14GvAZ+oqk/t7TElSZKGWa+X/HgS8Dc0l/2AponySZMXn1pVfbRjX53LdwCPaacvBe67t8eQJElaSHq95Mc1wPOTHAj8sqp+0t+wJEmS1KnXG8bfu72F1TbgwiTnJblXf0OTJEnSuF77tL0d+MuqOryqDgdeRDtSU5IkSf3Xa9J2q6r67PhMVZ0D3KovEUmSJGkPvQ5EuDTJ/wP+vZ1/KnBZf0KSJEnSRL3WtD0TOATYQDPq8xDgGf0KSpIkSbvrdfToj4Dn9TkWSZIkTWK6G8Zvmmp9VT12dsORJElSN9PVtD0QuAI4A/gqzc3iJUmSNMemS9puDzyC5u4HTwY+AZxRVRf2O7DFZOPWUdZv3s6OnWMsXzbCujUrWbtqxaDDkiRJ88iUAxGqaldVfaqqng4cA1wCnJPkuXMS3SKwcesoJ23YxujOMQoY3TnGSRu2sXHr6KBDkyRJ88i0o0eT7J/keOA9wP8F3kgzilSzYP3m7YzdsGu3ZWM37GL95u0DikiSJM1H0w1EOB24F/BJ4FVV9c05iWoR2bFzbEbLJUnS4jRdn7anAT8F7gY8L7lpHEKAqqoD+xjborB82QijXRK05ctGBhCNJEmar6br03azqjqgfRzY8TjAhG12rFuzkpGlS3ZbNrJ0CevWrBxQRJIkaT7q9TZW6pPxUaKOHpUkSVMxaZsH1q5aYZImSZKm1Ou9RyVJkjRAJm2SJElDwKRNkiRpCJi0SZIkDQGTNkmSpCFg0iZJkjQETNokSZKGgEmbJEnSEDBpkyRJGgImbZIkSUPApE2SJGkImLRJkiQNAZM2SZKkIWDSJkmSNARM2iRJkoaASZskSdIQMGmTJEkaAiZtkiRJQ8CkTZIkaQiYtEmSJA1kBMIgAAAMHElEQVQBkzZJkqQhYNImSZI0BEzaJEmShoBJmyRJ0hAwaZMkSRoC+w06gIVq49ZR1m/ezo6dYyxfNsK6NStZu2rFoMOSJElDyqStDzZuHeWkDdsYu2EXAKM7xzhpwzYAEzdJkrRXBtI8mmR9kouTfCPJR5Msm6Tco5JsT3JJkpfOdZx7a/3m7TclbOPGbtjF+s3bBxSRJEkadoPq03YWcK+qug/wbeCkiQWSLAHeAjwauAfwpCT3mNMo99KOnWMzWi5JkjSdgSRtVfXpqrqxnf0KcGiXYkcDl1TVpVX1C+D9wHFzFeO+WL5sZEbLJUmSpjMfRo8+E/hkl+UrgCs65q9sl81769asZGTpkt2WjSxdwro1KwcUkSRJGnZ9G4iQ5DPA7busellV/Udb5mXAjcB7u+2iy7Ka4ngnAicCHHbYYTOOdzaNDzZw9KgkSZotfUvaqurhU61P8nTg94GHVVW3ZOxK4I4d84cCO6Y43qnAqQCrV6+eNLmbK2tXrTBJkyRJs2ZQo0cfBbwEeGxVXT9JsXOBuyY5MsnNgScCm+YqRkmSpPlkUH3a3gwcAJyV5IIkbwNIsjzJmQDtQIXnAJuBi4APVtWFA4pXkiRpoAZycd2qussky3cAj+mYPxM4c67ikiRJmq/mw+hRSZIkTcOkTZIkaQiYtEmSJA0BkzZJkqQhYNImSZI0BEzaJEmShoBJmyRJ0hAwaZMkSRoCJm2SJElDwKRNkiRpCJi0SZIkDQGTNkmSpCFg0iZJkjQETNokSZKGgEmbJEnSEDBpkyRJGgImbZIkSUPApE2SJGkImLRJkiQNAZM2SZKkIWDSJkmSNARM2iRJkoaASZskSdIQMGmTJEkaAiZtkiRJQ8CkTZIkaQiYtEmSJA2B/QYdwLDZuHWU9Zu3s2PnGMuXjbBuzUrWrlox6LAkSdICZ9I2Axu3jnLShm2M3bALgNGdY5y0YRuAiZskSeorm0dnYP3m7TclbOPGbtjF+s3bBxSRJElaLEzaZmDHzrEZLZckSZotJm0zsHzZyIyWS5IkzRaTthlYt2YlI0uX7LZsZOkS1q1ZOaCIJEnSYuFAhBkYH2zg6FFJkjTXTNpmaO2qFSZpkiRpztk8KkmSNARM2iRJkoaASZskSdIQMGmTJEkaAiZtkiRJQ8CkTZIkaQiYtEmSJA0BkzZJkqQhkKoadAyzLsnVwHf6fJiDgR/0+RjaO56b+c3zM795fuYvz838ti/n5/CqOmS6QgsyaZsLSbZU1epBx6E9eW7mN8/P/Ob5mb88N/PbXJwfm0clSZKGgEmbJEnSEDBp23unDjoATcpzM795fuY3z8/85bmZ3/p+fuzTJkmSNASsaZMkSRoCJm1TSPKuJFcl+eYk649Ncm2SC9rHK+Y6xsUqyR2TfDbJRUkuTPL8LmWS5I1JLknyjSS/OYhYF6Mez4/vnwFIcoskX0vy9fbcvKpLmf2TfKB973w1yRFzH+ni1OP5OSHJ1R3vnWcPItbFKsmSJFuTfLzLur6+d/abzZ0tQKcBbwbePUWZL1TV789NOOpwI/Ciqjo/yQHAeUnOqqpvdZR5NHDX9vEA4F/av+q/Xs4P+P4ZhJ8DD62qnyRZCnwxySer6isdZZ4F/Kiq7pLkicBrgScMIthFqJfzA/CBqnrOAOITPB+4CDiwy7q+vnesaZtCVX0euGbQcWhPVfW9qjq/nf4xzRtoxYRixwHvrsZXgGVJ7jDHoS5KPZ4fDUD7fvhJO7u0fUzs3HwccHo7/WHgYUkyRyEuaj2eHw1IkkOB3wPeMUmRvr53TNr23QPbauxPJrnnoINZjNrq51XAVyesWgFc0TF/JSYOc26K8wO+fwaibd65ALgKOKuqJn3vVNWNwLXAbeY2ysWrh/MD8H/abh8fTnLHOQ5xMXsD8GLgl5Os7+t7x6Rt35xPc+uJ+wJvAjYOOJ5FJ8mtgY8AL6iq6yau7rKJv1jn0DTnx/fPgFTVrqq6H3AocHSSe00o4ntngHo4Px8Djqiq+wCf4Vc1O+qjJL8PXFVV501VrMuyWXvvmLTtg6q6brwau6rOBJYmOXjAYS0abX+PjwDvraoNXYpcCXT+Aj0U2DEXsWn68+P7Z/CqaidwDvCoCatueu8k2Q84CLuKzLnJzk9V/bCqft7O/itw/zkObbF6EPDYJJcD7wcemuQ9E8r09b1j0rYPktx+vK06ydE0r+cPBxvV4tC+7u8ELqqq109SbBPwx+0o0mOAa6vqe3MW5CLWy/nx/TMYSQ5JsqydHgEeDlw8odgm4Ont9OOAs8uLes6JXs7PhL65j6XpM6o+q6qTqurQqjoCeCLN++KpE4r19b3j6NEpJDkDOBY4OMmVwN/QdAqlqt5Gc0L+PMmNwBjwRD/Y5syDgKcB29q+HwB/DRwGN52fM4HHAJcA1wPPGECci1Uv58f3z2DcATg9yRKaRPmDVfXxJH8LbKmqTTQJ978nuYSmluCJgwt30enl/DwvyWNpRmlfA5wwsGjFXL53vCOCJEnSELB5VJIkaQiYtEmSJA0BkzZJkqQhYNImSZI0BEzaJEmSppDkXUmuSvLNHso+OMn5SW5M8rgu6w9MMprkzTONw6RN0ryRZFeSCzoeRww6ptmUZFWSdyR5Rsdz/EWSbe30KUlOmPhhnuScJKun2O/7k9y1/89AWrROY8+LUE/muzSXYXnfJOtfDXxub4LwOm2S5pOx9vY9XSXZr72f37D6a+A1VfV14N8A2qurP6SqftDOn7AX+/0Xmvsh/snshCmpU1V9fuKPyCR3Bt4CHEJzLdA/qaqLq+rydv0e9ydNcn/gdsCngEl/iE3GmjZJ81pb8/ShJB8DPt0uW5fk3PaG2a/qKPuyJNuTfCbJGUn+ql1+U01VkoPbRGn8xtzrO/b1p+3yY9ttPpzk4iTv7bh7w1FJvpzmRvdfS3JAki8kuV9HHF9Kcp8Jz+MA4D5twra3r8VjO2rotie5rF31BeDh7W1zJM2NU4HnVtX9gb8C3jpV4SQ3A/4RWLe3B/QNLmk+Gem4g8JlVfWH7fQDaRKea5I8ErgrcDTNzZk3JXkw8FOaq4+vovlsOx+Y6sbOAM+iub3ZUUn2B76U5NPtulXAPWnuV/sl4EFJvgZ8AHhCVZ2b5ECauzm8g6Y55AVJ7gbsX1XfmHCs1cC0/WFaT0jy2x3zdwFor7i+CSDJB2mbWKrql+0V2O/bw3OWtI+S3Br4LeBD7e85gP2n2ewvgDOr6oqObWbEpE3SfDJZ8+hZVTV+0+VHto+t7fytaZK4A4CPVtX1AEk29XC8RwL36egsfFC7r18AX6uqK9t9XQAcAVwLfK+qzoXmpvft+g8B/y/JOuCZNP1fJroDcHUPMQF8oKqeMz6T5JzOlUleTPNavaVj8VXAckzapLlwM2DnVN05ungg8DtJ/oLmc+vmSX5SVS/tdQcmbZKGwU87pgOcXFVv7yyQ5AXAZPflu5FfdQe5xYR9PbeqNk/Y17HAzzsW7aL5vEy3Y1TV9UnOAo4D/ojufVXGJhx7ryR5GPB44METVt2iPYakPquq65JcluTxVfWhtvvElN0fquop49Nt39XVM0nYwD5tkobPZuCZbfMESVYkuS3weeAPk4y0/cf+oGOby4H7t9OPm7CvP0+ytN3X3ZLcaopjXwwsT3JUW/6Ajn5k7wDeCJzbUSvY6SLaZs69leRwmn4zf1RVExO0uwEX7sv+JXWX5Azgv4CVSa5M8izgKcCzknyd5r13XFv2qCRX0vy4enuSWXtfWtMmaahU1aeT3B34r7ZfyE+Ap1bV+Uk+AFwAfIemc/641wEfTPI04OyO5e+gafY8v/2lfDWwdopj/yLJE4A3JRmhqdl6OPCTqjovyXW0o0K7bHtxkoOSHFBVP96rJ9/0m7sN8NH2ue+oqsckuR1Nc+n39nK/kqZQVU+aZNUelwFpu08cOs3+TqN7N4oppWqy1gRJGl5JXkmTTL1ujo63HDgH+I2q2mOof1vmhcCPq+ods3zsFwLXVdU7Z3O/kuYXm0claR8l+WPgq8DLJkvYWv/C7n3lZstO4PQ+7FfSPGJNmyRJ0hCwpk2SJGkImLRJkiQNAZM2SZKkIWDSJkmSNARM2iRJkoaASZskSdIQ+P8bmkGgzgMqFwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x14dd403fbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x, y = Disp[1]\n",
    "\n",
    "plt.figure(figsize=(10,5))\n",
    "plt.scatter(x, y)  # Plot each 300 data point.\n",
    "plt.xlabel(\"Frequency (THz)\")\n",
    "plt.ylabel(\"Mode deviation (GHz)\")\n",
    "plt.title(\"Modal spectral deviation in SiN resonator\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}