ECG Autoencoder Convolutional.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\"\n",
    "import tensorflow as tf \n",
    "tf.set_random_seed(22)\n",
    "from keras.backend.tensorflow_backend import set_session \n",
    "config = tf.ConfigProto() \n",
    "config.gpu_options.per_process_gpu_memory_fraction = 0.8\n",
    "set_session(tf.Session(config = config))\n",
    "import pandas as pd\n",
    "import wfdb\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Good beats from Physionet beats annotation guidlines\n",
    "good_beats = ['N', 'L', 'R', 'B', 'A', 'a', 'J', 'S', 'V', 'r',\n",
    "        'F', 'e', 'j', 'n', 'E', '/', 'f', 'Q', '?']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Beat extraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "twoch_df = pd.DataFrame()\n",
    "firstch_df = pd.DataFrame()\n",
    "\n",
    "# 120 samples to the left and right from the beat label \n",
    "half_qrs=120\n",
    "\n",
    "for filename in os.listdir('mitdb'):\n",
    "    # Exclusion of some datasets\n",
    "    if filename.endswith(\".dat\") and not filename.startswith(('102','104','107','217')):\n",
    "        ann = wfdb.rdann('mitdb/' + filename.strip('.dat'), 'atr')\n",
    "        record = wfdb.rdsamp('mitdb/' + filename.strip('.dat'))\n",
    "        \n",
    "        # We want just the signals\n",
    "        data = record[0]\n",
    "        \n",
    "        # Prepare containers\n",
    "        signals, classes = [], []\n",
    "        firstch_signals = []\n",
    "        firstch_class = []\n",
    "        \n",
    "        # Beat extraction\n",
    "        for it, beat in enumerate(ann.symbol):\n",
    "            if beat in good_beats:\n",
    "                \n",
    "                sta = ann.sample[it] - half_qrs\n",
    "                end = ann.sample[it] + half_qrs\n",
    "                qrs = data[sta : end, :]\n",
    "            \n",
    "                # This may happen at the edges\n",
    "                if len(qrs) != 2 * half_qrs: continue\n",
    "                \n",
    "                # Extracting only first channel signals and classes \n",
    "                firstch_beat_class = '{}_{}'.format(record[1]['sig_name'][0], beat)\n",
    "                firstch_signals.append(qrs[:,0].reshape(240,))\n",
    "                firstch_class.append(firstch_beat_class)\n",
    "            \n",
    "                # Keep the channel type in the class name (both channels)\n",
    "                for ch in range(2):\n",
    "                    beat_class = '{}_{}'.format(record[1]['sig_name'][ch], beat)\n",
    "                    signals.append(qrs[:, ch].reshape(240,))\n",
    "                    classes.append(beat_class)\n",
    "                \n",
    "        new_frame = pd.DataFrame({'qrs_data' : signals,\n",
    "                              'qrs_type' : classes})\n",
    "        \n",
    "        new_frame_firstch = pd.DataFrame({'qrs_data' : firstch_signals,\n",
    "                              'qrs_type' : firstch_class})\n",
    "        # Cumulate\n",
    "        twoch_df = twoch_df.append(new_frame)\n",
    "        firstch_df = firstch_df.append(new_frame_firstch)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preparation of datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "training = pd.DataFrame()\n",
    "\n",
    "# We extract 120 samples to the left and right from the beat label. Total: 240 samples per beat\n",
    "half_qrs=120\n",
    "\n",
    "for filename in os.listdir('mitdb'):\n",
    "    if filename.endswith(\".dat\"):\n",
    "        ann = wfdb.rdann('mitdb/' + filename.strip('.dat'), 'atr')\n",
    "        record = wfdb.rdsamp('mitdb/' + filename.strip('.dat'))\n",
    "        \n",
    "        # Signals are stored in position 0 of record\n",
    "        data = record[0]\n",
    "        \n",
    "        # Prepare containers\n",
    "        signals, classes = [], []\n",
    "        \n",
    "        # Beat extraction\n",
    "        for it, beat in enumerate(ann.symbol):\n",
    "            if beat in good_beats:\n",
    "                \n",
    "                sta = ann.sample[it] - half_qrs\n",
    "                end = ann.sample[it] + half_qrs\n",
    "                qrs = data[sta : end, :]\n",
    "            \n",
    "                # This may happen at the edges\n",
    "                if len(qrs) != 2 * half_qrs: continue\n",
    "                \n",
    "                # Extracting only first channel signals and classes \n",
    "                beat_class = '{}_{}'.format(record[1]['sig_name'][0], beat)\n",
    "                signals.append(qrs[:,0].reshape(240,))\n",
    "                classes.append(beat_class)\n",
    "                \n",
    "        frame = pd.DataFrame({'beat' : signals,\n",
    "                              'label' : classes})\n",
    "\n",
    "        # Append\n",
    "        training = training.append(frame)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(109459, 2)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MLII_N    72945\n",
       "MLII_L     8071\n",
       "MLII_R     7257\n",
       "MLII_V     7080\n",
       "MLII_/     3619\n",
       "V5_/       3407\n",
       "MLII_A     2536\n",
       "V5_N       2082\n",
       "MLII_F      798\n",
       "V5_f        722\n",
       "MLII_f      260\n",
       "MLII_j      229\n",
       "MLII_a      150\n",
       "MLII_E      106\n",
       "MLII_J       81\n",
       "V5_V         49\n",
       "V5_Q         18\n",
       "MLII_e       16\n",
       "MLII_Q       15\n",
       "V5_A         10\n",
       "V5_F          4\n",
       "MLII_S        2\n",
       "V5_J          2\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MLII_N    72945\n",
       "MLII_L     8071\n",
       "MLII_R     7257\n",
       "MLII_V     7080\n",
       "MLII_A     2536\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Exclude all classes except N, L, R, V, A\n",
    "\n",
    "training = training[training['label']!='V5_J']\n",
    "training = training[training['label']!='MLII_S']\n",
    "training = training[training['label']!='V5_F']\n",
    "training = training[training['label']!='V5_A']\n",
    "training = training[training['label']!='MLII_Q']\n",
    "training = training[training['label']!='MLII_e']\n",
    "training = training[training['label']!='V5_Q']\n",
    "training = training[training['label']!='V5_V']\n",
    "training = training[training['label']!='MLII_J']\n",
    "training = training[training['label']!='MLII_E']\n",
    "training = training[training['label']!='MLII_a']\n",
    "training = training[training['label']!='MLII_j']\n",
    "training = training[training['label']!='MLII_f']\n",
    "training = training[training['label']!='V5_f']\n",
    "training = training[training['label']!='MLII_F']\n",
    "training = training[training['label']!='V5_N']\n",
    "training = training[training['label']!='V5_/']\n",
    "training = training[training['label']!='MLII_/']\n",
    "\n",
    "\n",
    "\n",
    "training['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_labels = training['label'].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### We take 2000 beats from each class for train autoencoders' datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "autoenc_train = pd.DataFrame()\n",
    "firstch_classes = pd.DataFrame(firstch_df['qrs_type'].value_counts())\n",
    "\n",
    "for row in firstch_classes.itertuples():\n",
    "    if row[0] in ['MLII_N', 'MLII_L', 'MLII_R', 'MLII_V', 'MLII_A']:\n",
    "        autoenc_train = autoenc_train.append(firstch_df[firstch_df['qrs_type']==row[0]].iloc[:2000,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "autoenc_train_labels = autoenc_train['qrs_type'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "autoenc_train = np.vstack(autoenc_train['qrs_data']).reshape(10000,240,1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Autoencoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras\n",
    "from keras.layers import Layer, Input,Conv1D,MaxPooling1D,UpSampling1D, Flatten, Dense, Reshape\n",
    "from keras.models import Model\n",
    "from keras.optimizers import RMSprop, Nadam\n",
    "import keras.models as models\n",
    "from keras import backend as K\n",
    "from keras.layers.advanced_activations import LeakyReLU\n",
    "from keras.layers import Dropout\n",
    "from keras.utils.vis_utils import model_to_dot\n",
    "from IPython.display import SVG\n",
    "from keras.utils import plot_model\n",
    "from sklearn.model_selection import train_test_split\n",
    "import random\n",
    "import matplotlib as mpl\n",
    "import matplotlib.cm as cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters\n",
    "batch_size = 25\n",
    "epochs = 20\n",
    "input_sig = Input(shape = (240,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv_autoencoder(input_sig):\n",
    "    \n",
    "    # Part 1: Encoder\n",
    "    conv1 = Conv1D(30, 3, activation='relu', padding='same')(input_sig)\n",
    "    pool1 = MaxPooling1D(2)(conv1)\n",
    "    conv2 = Conv1D(60, 3, activation='relu', padding='same')(pool1)\n",
    "    pool2 = MaxPooling1D(2)(conv2)\n",
    "    flat = Flatten()(pool2)\n",
    "    dense1 = Dense(25)(flat)\n",
    "    dense2 = Dense(10)(dense1) #15\n",
    "                         \n",
    "    # Part 2: Decoder\n",
    "    dense3 = Dense(25)(dense2)\n",
    "    dense4 = Dense(60*60)(dense3)\n",
    "    reshape = Reshape((60,60))(dense4)\n",
    "    up1 = UpSampling1D(2)(reshape)\n",
    "    conv5 = Conv1D(30, 3, activation='relu', padding='same')(up1)\n",
    "    up2 = UpSampling1D(2)(conv5) \n",
    "    decoded = Conv1D(1, 3, activation='relu', padding='same')(up2)\n",
    "    return decoded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "autoencoder = Model(input_sig, conv_autoencoder(input_sig))\n",
    "autoencoder.compile(loss='mean_squared_error', optimizer = Nadam())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 240, 1)            0         \n",
      "_________________________________________________________________\n",
      "conv1d_13 (Conv1D)           (None, 240, 30)           120       \n",
      "_________________________________________________________________\n",
      "max_pooling1d_7 (MaxPooling1 (None, 120, 30)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_14 (Conv1D)           (None, 120, 60)           5460      \n",
      "_________________________________________________________________\n",
      "max_pooling1d_8 (MaxPooling1 (None, 60, 60)            0         \n",
      "_________________________________________________________________\n",
      "flatten_4 (Flatten)          (None, 3600)              0         \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 25)                90025     \n",
      "_________________________________________________________________\n",
      "dense_14 (Dense)             (None, 10)                260       \n",
      "_________________________________________________________________\n",
      "dense_15 (Dense)             (None, 25)                275       \n",
      "_________________________________________________________________\n",
      "dense_16 (Dense)             (None, 3600)              93600     \n",
      "_________________________________________________________________\n",
      "reshape_4 (Reshape)          (None, 60, 60)            0         \n",
      "_________________________________________________________________\n",
      "up_sampling1d_7 (UpSampling1 (None, 120, 60)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_15 (Conv1D)           (None, 120, 30)           5430      \n",
      "_________________________________________________________________\n",
      "up_sampling1d_8 (UpSampling1 (None, 240, 30)           0         \n",
      "_________________________________________________________________\n",
      "conv1d_16 (Conv1D)           (None, 240, 1)            91        \n",
      "=================================================================\n",
      "Total params: 195,261\n",
      "Trainable params: 195,261\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "autoencoder.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_X,valid_X,train_ground,valid_ground = train_test_split(autoenc_train,\n",
    "                                                             autoenc_train, \n",
    "                                                             test_size=0.3, \n",
    "                                                             random_state=13)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7000 samples, validate on 3000 samples\n",
      "Epoch 1/20\n",
      "7000/7000 [==============================] - 6s 848us/step - loss: 0.2346 - val_loss: 0.2373\n",
      "Epoch 2/20\n",
      "7000/7000 [==============================] - 5s 734us/step - loss: 0.2295 - val_loss: 0.2369\n",
      "Epoch 3/20\n",
      "7000/7000 [==============================] - 5s 733us/step - loss: 0.2294 - val_loss: 0.2368\n",
      "Epoch 4/20\n",
      "7000/7000 [==============================] - 5s 739us/step - loss: 0.2288 - val_loss: 0.2360\n",
      "Epoch 5/20\n",
      "7000/7000 [==============================] - 5s 735us/step - loss: 0.2288 - val_loss: 0.2361\n",
      "Epoch 6/20\n",
      "7000/7000 [==============================] - 5s 741us/step - loss: 0.2282 - val_loss: 0.2369\n",
      "Epoch 7/20\n",
      "7000/7000 [==============================] - 5s 747us/step - loss: 0.2282 - val_loss: 0.2355\n",
      "Epoch 8/20\n",
      "7000/7000 [==============================] - 5s 741us/step - loss: 0.2280 - val_loss: 0.2365\n",
      "Epoch 9/20\n",
      "7000/7000 [==============================] - 5s 751us/step - loss: 0.2285 - val_loss: 0.2359\n",
      "Epoch 10/20\n",
      "7000/7000 [==============================] - 5s 750us/step - loss: 0.2279 - val_loss: 0.2353\n",
      "Epoch 11/20\n",
      "7000/7000 [==============================] - 5s 749us/step - loss: 0.2279 - val_loss: 0.2355\n",
      "Epoch 12/20\n",
      "7000/7000 [==============================] - 5s 758us/step - loss: 0.2279 - val_loss: 0.2352\n",
      "Epoch 13/20\n",
      "7000/7000 [==============================] - 5s 754us/step - loss: 0.2277 - val_loss: 0.2355\n",
      "Epoch 14/20\n",
      "7000/7000 [==============================] - 5s 753us/step - loss: 0.2277 - val_loss: 0.2354\n",
      "Epoch 15/20\n",
      "7000/7000 [==============================] - 5s 757us/step - loss: 0.2278 - val_loss: 0.2356\n",
      "Epoch 16/20\n",
      "7000/7000 [==============================] - 6s 793us/step - loss: 0.2278 - val_loss: 0.2354\n",
      "Epoch 17/20\n",
      "7000/7000 [==============================] - 5s 785us/step - loss: 0.2278 - val_loss: 0.2353\n",
      "Epoch 18/20\n",
      "7000/7000 [==============================] - 6s 829us/step - loss: 0.2277 - val_loss: 0.2352\n",
      "Epoch 19/20\n",
      "7000/7000 [==============================] - 5s 750us/step - loss: 0.2277 - val_loss: 0.2352\n",
      "Epoch 20/20\n",
      "7000/7000 [==============================] - 5s 757us/step - loss: 0.2277 - val_loss: 0.2352\n"
     ]
    }
   ],
   "source": [
    "autoencoder_train = autoencoder.fit(train_X, train_ground, batch_size=batch_size,epochs=epochs,verbose=1,\n",
    "                                    validation_data=(valid_X, valid_ground))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEICAYAAABfz4NwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xl8FFW2wPHfSSBA2B8oAoFElpEdAjFssgiILALDOmhcR8yg44zbQ9E46nNERxF3niP6eI5DFFncFRlGkIijSIAAEkC2gJCwPghg2JKc90dXepqQpZPudGc538+nP6m6davqVHWnTtetW9WiqhhjjDEhwQ7AGGNM+WAJwRhjDGAJwRhjjMMSgjHGGMASgjHGGIclBGOMMYAlBONHIhIqIqdEpKU/6waTiLQREb/3zRaRISKS5jG+TUT6eVO3FOt6S0QeKe38RSz3KRF529/LNcFTLdgBmOARkVMeo+HAWSDHGf+dqiaWZHmqmgPU8XfdqkBVr/DHckRkCnCjqg70WPYUfyzbVH6WEKowVXUfkJ1voFNU9Z+F1ReRaqqaHYjYjDGBZ01GplBOk8D7IvKeiJwEbhSR3iLyvYgcF5EMEXlFRKo79auJiIpIlDM+z5m+REROish3InJ5Ses604eLyE8ikikir4rItyJyayFxexPj70Rkh4gcE5FXPOYNFZEXReSoiOwChhWxfxJEZH6+stki8oIzPEVEtjjbs9P59l7YsvaJyEBnOFxE/u7Ethnoka/uoyKyy1nuZhEZ7ZR3Bl4D+jnNcUc89u0THvNPdbb9qIh8JCJNvdk3xRGRsU48x0VkuYhc4THtERFJF5ETIrLVY1t7icg6p/ygiMz0dn2mDKiqvewFkAYMyVf2FHAOGIXry0Mt4EqgJ66zy1bAT8DdTv1qgAJRzvg84AgQA1QH3gfmlaLupcBJYIwz7X7gPHBrIdviTYwfA/WBKOD/8rYduBvYDEQAjYAk179JgetpBZwCanss+xAQ44yPcuoIMAg4DXRxpg0B0jyWtQ8Y6Aw/D3wNNAQigdR8dScBTZ335AYnhibOtCnA1/ninAc84QwPdWLsBtQE/htY7s2+KWD7nwLedobbO3EMct6jR4BtznBHYA9wmVP3cqCVM7wGuN4Zrgv0DPb/QlV+2RmCKc4qVf1UVXNV9bSqrlHV1aqaraq7gDnAgCLmX6Sqyap6HkjEdSAqad3rgBRV/diZ9iKu5FEgL2N8RlUzVTUN18E3b12TgBdVdZ+qHgX+UsR6dgE/4kpUANcAx1Q12Zn+qaruUpflwFdAgReO85kEPKWqx1R1D65v/Z7rXaCqGc578i6uZB7jxXIB4oC3VDVFVc8A04EBIhLhUaewfVOUycAnqrrceY/+giup9ASycSWfjk6z425n34ErsbcVkUaqelJVV3u5HaYMWEIwxfnZc0RE2onI5yJyQEROAE8CjYuY/4DHcBZFX0gurG4zzzhUVXF9oy6QlzF6tS5c32yL8i5wvTN8gzOeF8d1IrJaRP5PRI7j+nZe1L7K07SoGETkVhHZ4DTNHAfaeblccG2fe3mqegI4BjT3qFOS96yw5ebieo+aq+o24AFc78MhpwnyMqfqbUAHYJuI/CAiI7zcDlMGLCGY4uTvcvkGrm/FbVS1HvAYriaRspSBqwkHABERLjyA5edLjBlAC4/x4rrFLgCGiEhzXGcK7zox1gIWAc/gas5pAPzDyzgOFBaDiLQCXgfuBBo5y93qsdziusim42qGylteXVxNU/u9iKskyw3B9Z7tB1DVearaF1dzUSiu/YKqblPVybiaBWcBi0Wkpo+xmFKyhGBKqi6QCfwiIu2B3wVgnZ8B3UVklIhUA+4BLimjGBcA94pIcxFpBDxUVGVVPQCsAt4GtqnqdmdSDSAMOAzkiMh1wOASxPCIiDQQ130ad3tMq4ProH8YV268A9cZQp6DQETeRfQCvAfcLiJdRKQGrgPzN6pa6BlXCWIeLSIDnXVPw3XdZ7WItBeRq531nXZeubg24CYRaeycUWQ625brYyymlCwhmJJ6ALgF1z/7G7gu/pYpVT0I/AZ4ATgKtAbW47pvwt8xvo6rrX8Trguei7yY511cF4ndzUWqehy4D/gQ14XZCbgSmzcex3WmkgYsAd7xWO5G4FXgB6fOFYBnu/syYDtwUEQ8m37y5v8SV9PNh878LXFdV/CJqm7Gtc9fx5WshgGjnesJNYDncF33OYDrjCTBmXUEsEVcvdieB36jqud8jceUjriaY42pOEQkFFcTxQRV/SbY8RhTWdgZgqkQRGSY04RSA/gTrt4pPwQ5LGMqFUsIpqK4CtiFqzniWmCsqhbWZGSMKQVrMjLGGAPYGYIxxhhHhXq4XePGjTUqKirYYRhjTIWydu3aI6paVFdtoIIlhKioKJKTk4MdhjHGVCgiUtwd94A1GRljjHFYQjDGGANYQjDGGOOoUNcQjDGBdf78efbt28eZM2eCHYrxQs2aNYmIiKB69cIeZVU0SwjGmELt27ePunXrEhUVheshs6a8UlWOHj3Kvn37uPzyy4ufoQDWZGSMKdSZM2do1KiRJYMKQERo1KiRT2dzlhCMMUWyZFBx+PpeVYmE8Nprr/HJJ59gj+kwxpjCVfqEkJOTw5tvvsmYMWOIjY1lyZIllhiMqSCOHj1Kt27d6NatG5dddhnNmzd3j587593PJtx2221s27atyDqzZ88mMTHRHyFz1VVXkZKS4pdlBVqlTwihoaGsXbuWuXPncuTIEUaMGEGfPn1YtmyZJQZj/CwxMZGoqChCQkKIiory+SDbqFEjUlJSSElJYerUqdx3333u8bCwMMB1MTU3t/AfWfvf//1frrjiiiLX8/vf/564OJ9/J6jCq/QJAaBatWrubwlz5sxh//79DB06lP79+7NixYpgh2dMpZCYmEh8fDx79uxBVdmzZw/x8fF+++btaceOHXTo0IG4uDg6duxIRkYG8fHxxMTE0LFjR5588kl33bxv7NnZ2TRo0IDp06fTtWtXevfuzaFDhwB49NFHeemll9z1p0+fTmxsLFdccQX/+te/APjll18YP348HTp0YMKECcTExBR7JjBv3jw6d+5Mp06deOSRRwDIzs7mpptucpe/8sorALz44ot06NCBLl26cOONN/p9n3mjSiSEPGFhYdxxxx1s376d2bNns2vXLgYNGsSgQYNYtWpVsMMzpkJLSEggKyvrgrKsrCwSEhIKmcM3W7du5b777iM1NZXmzZvzl7/8heTkZDZs2MCyZctITU29aJ7MzEwGDBjAhg0b6N27N3Pnzi1w2arKDz/8wMyZM93J5dVXX+Wyyy4jNTWVP/3pT6xfv77I+Pbt28ejjz7KihUrWL9+Pd9++y2fffYZa9eu5ciRI2zatIkff/yRm2++GYDnnnuOlJQUNm7cyGuvvebj3imdKpUQ8tSoUYO77rqLnTt38tJLL5Gamkq/fv0YOnQo3333XbDDM6ZC2rt3b4nKfdW6dWtiYmLc4++99x7du3ene/fubNmypcCEUKtWLYYPHw5Ajx49SEtLK3DZ48aNu6jOqlWrmDx5MgBdu3alY8eORca3evVqBg0aROPGjalevTo33HADSUlJtGnThm3btvHHP/6RpUuXUr9+fQA6duzIjTfeSGJiYqlvLPNVlUwIeWrWrMk999zDrl27eP7551m/fj19+vRhxIgRrFmzJtjhGVOhtGzZskTlvqpdu7Z7ePv27bz88sssX76cjRs3MmzYsAL74+dddwDX9cXs7OwCl12jRo1i65RWo0aN2LhxI/369WP27Nn87ne/A2Dp0qVMnTqVNWvWEBsbS05Ojl/X640qnRDyhIeH88ADD7B7926eeeYZVq9eTWxsLGPGjKmwvQWMCbQZM2YQHh5+QVl4eDgzZswo83WfOHGCunXrUq9ePTIyMli6dKnf19G3b18WLFgAwKZNmwo8A/HUs2dPVqxYwdGjR8nOzmb+/PkMGDCAw4cPo6pMnDiRJ598knXr1pGTk8O+ffsYNGgQzz33HEeOHLmo+S0Q7NEVHurUqcP06dO56667eOWVV5g1axbR0dGMGzeOqVOncuWVV9KgQYNghxlwqkpOTg7VqtnHxRQur5dOQkICe/fupWXLlsyYMSMgvXe6d+9Ohw4daNeuHZGRkfTt29fv6/jDH/7AzTffTIcOHdyvvOaegkRERPDnP/+ZgQMHoqqMGjWKkSNHsm7dOm6//XZUFRHh2WefJTs7mxtuuIGTJ0+Sm5vLf/7nf1K3bl2/b0NxKtRvKsfExGggfyDn+PHjvPjii7z44oucPHkSgDZt2hATE0NMTAxXXnkl0dHRQXnjAuXIkSNce+211KtXj+XLl9tdq1XMli1baN++fbDDKBeys7PJzs6mZs2abN++naFDh7J9+/Zy90WpoPdMRNaqakwhs7iVry0pZxo0aMB//dd/8cADD/D999+TnJxMcnIy3377LfPnzwdct4q3a9fOnSRiYmLo1q3bRafOFdHRo0cZMmQIGzduRFVZvHgxEyZMCHZYxgTFqVOnGDx4MNnZ2agqb7zxRrlLBr6yM4RSOnjwIGvXrnUniTVr1nDgwAHAdSGqQ4cOXHnllRckiWD1HCiNY8eOMXjwYFJTU/nwww+ZNm0a586dY/PmzRVqO4xv7Ayh4rEzhCBo0qQJI0aMYMSIEe6y9PT0CxLEJ5984u7n3KNHD7788ksaN24crJC9dvz4ca655ho2b97Mxx9/zLBhw8jJyWHUqFG8+eab3HXXXcEO0RhTBuwMoQypKnv37mXZsmX84Q9/oG3btixbtowmTZoEO7RCZWZmcs0115CSksKHH37IyJEjAde2DBw4kK1bt7Jjx45Kfd3E/JudIVQ8vpwhWLfTMiQiREZGMmXKFD777DN27tzJwIEDSU9PD3ZoBTpx4gTDhg0jJSWFRYsWuZMBuLblueee49ChQ8yaNSuIURpjyoolhAAZPHgwX375Jfv27aN///5ldvdmaZ08eZLhw4eTnJzMggULGD169EV1evbsyYQJE3j++efd10uMMZWHJYQA6tevH8uWLePIkSP079+fXbt2BTskwNV7YsSIEaxevZr58+fz61//utC6Tz/9NGfPnr3g4WHGlJWrr776opvMXnrpJe68884i56tTpw7guq5XWM+4gQMHUlwT9EsvvXTBDWIjRozg+PHj3oRepCeeeILnn3/e5+X4myWEAOvVqxfLly/n5MmT9O/fn59++imo8fzyyy+MHDmS7777jvfee4/x48cXWb9t27bEx8czZ86cgMU+c+ZM7rzzTk6cOBGQ9Zny4/rrr3d38c4zf/58rr/+eq/mb9asGYsWLSr1+vMnhC+++KJS35xqCSEIunfvztdff825c+fo378/mzdvDkocWVlZjBo1ilWrVjFv3jwmTpzo1XyPPfYYtWrVcj/Otyx9/PHHPPjgg/z1r3+lR48exT5h0lQuEyZM4PPPP3f/GE5aWhrp6en069fPfV9A9+7d6dy5Mx9//PFF86elpdGpUycATp8+zeTJk2nfvj1jx47l9OnT7np33nmn+9HZjz/+OACvvPIK6enpXH311Vx99dUAREVFceTIEQBeeOEFOnXqRKdOndyPzk5LS6N9+/bccccddOzYkaFDh16wnoKkpKTQq1cvunTpwtixYzl27Jh7/XmPw857qN7KlSvdPxAUHR3tvmHWb1S1wrx69OihlUlqaqo2bdpUGzdurOvXrw/ourOysnTw4MEaEhKi8+bNK/H8TzzxhAL63XfflUF0Lrt27dL69etrjx49dNmyZdq8eXMNCwvT2bNna25ubpmt1/xbamqqe/iee+7RAQMG+PV1zz33FBvDyJEj9aOPPlJV1WeeeUYfeOABVVU9f/68ZmZmqqrq4cOHtXXr1u7PRe3atVVVdffu3dqxY0dVVZ01a5bedtttqqq6YcMGDQ0N1TVr1qiq6tGjR1VVNTs7WwcMGKAbNmxQVdXIyEg9fPiwO5a88eTkZO3UqZOeOnVKT548qR06dNB169bp7t27NTQ01P3/PHHiRP373/9+0TY9/vjjOnPmTFVV7dy5s3799deqqvqnP/3JvU+aNm2qZ86cUVXVY8eOqarqddddp6tWrVJV1ZMnT+r58+eLfM/yAMnqxTHWzhCCqH379iQlJVGrVi0GDRoUsCesnj59mjFjxrB8+XLefvvtUj1r5oEHHqBJkyY8+OCDZfLLc2fPnmXSpEkALFiwgCFDhpCSksLgwYP5/e9/z6RJk8jMzPT7ek3549ls5NlcpKo88sgjdOnShSFDhrB//34OHjxY6HKSkpLcPzzTpUsXunTp4p62YMECunfvTnR0NJs3by72wXWrVq1i7Nix1K5dmzp16jBu3Di++eYbAC6//HK6desGFP2IbXB18z5+/DgDBgwA4JZbbiEpKckdY1xcHPPmzXPfEd23b1/uv/9+XnnlFY4fP+73O6XtxrQga9OmDUlJSQwaNIghQ4awZMkS+vTpU2brO3PmDGPHjuWf//wnc+fO5aabbirVcurUqcPjjz/OXXfdxWeffcaoUaP8Gue0adNITk7mgw8+oFWrVgA0btyYzz77jFmzZvHwww+zdu1aFixYcMEz8U3ZyWsWCbQxY8Zw3333sW7dOrKysujRowfg+oW2w4cPs3btWqpXr05UVFSBj7wuzu7du3n++edZs2YNDRs25NZbby3VcvLkPTobXE8tKK7JqDCff/45SUlJfPrpp8yYMYNNmzYxffp0Ro4cyRdffEHfvn1ZunQp7dq1K3Ws+dkZQjkQFRVFUlISTZo0YejQoaxcubJM1nP27FnGjx/P0qVLefPNN7n11lt9Wt6UKVP41a9+xfTp0/36zPiFCxfy6quvcu+99zJ27NgLpoWEhDBt2jS++eYbsrOz6dOnDy+//LL9PnYlVqdOHa6++mp++9vfXnAxOTMzk0svvZTq1auzYsUK9uzZU+Ry+vfvz7vvvgvAjz/+yMaNGwHX/Te1a9emfv36HDx4kCVLlrjnqVu3boHt9P369eOjjz4iKyuLX375hQ8//JB+/fqVeNvq169Pw4YN3WcXf//73xkwYAC5ubn8/PPPXH311Tz77LNkZmZy6tQpdu7cSefOnXnooYe48sor2bp1a4nXWRRLCOVEREQEK1euJDIykuHDh/OPf/zDr8s/d+4cEydO5IsvvuCNN97g9ttv93mZ1atX5+mnnyY1NZW//e1vfojS9Vu5t99+Oz179uTZZ58ttF7v3r1JSUlh+PDh3HvvvYwbN859Mc5UPtdffz0bNmy4ICHExcWRnJxM586deeedd4r9pnznnXdy6tQp2rdvz2OPPeY+0+jatSvR0dG0a9eOG2644YJHZ8fHxzNs2DD3ReU83bt359ZbbyU2NpaePXsyZcoUoqOjS7Vtf/vb35g2bRpdunQhJSWFxx57jJycHG688UY6d+5MdHQ0f/zjH2nQoAEvvfQSnTp1okuXLlSvXt39629+482FhvLyqmwXlQty6NAh7dq1q4aFhemnn37ql2WePXtWx4wZo4D+93//t1+WmSc3N1d79eqlzZo1019++cWnZZ0+fVq7deumDRs21LS0NK/X/8ILL2j16tU1MjKyTC9yV0UFXaA05ZsvF5W9OhADw4BtwA5gegHT7wdSgY3AV0CkUx4JrANSgM3AVI95woA5wE/AVmB8cXFUhYSg6urx0KNHD61evbouXrzY6/lyc3M1PT1dV65cqW+99ZZOnz5dx48fr23btlVAX3vttTKJNykpSQF9+umnfVrO7373OwVKlQhXr16tUVFRWq1aNZ05c6bm5OT4FItxsYRQ8ZRpQgBCgZ1AK+cgvgHokK/O1UC4M3wn8L7++6BfwxmuA6QBzZzx/wKecoZDgMbFxVJVEoKq6vHjx7V3794aGhqq7733nrs8NzdXMzIyNCkpSefOnasPP/ywTpgwQbt27aq1a9dWwP2qVq2a/upXv9KRI0fqO++8U6bxjho1SuvVq3dBF72SSExMVEAffPDBUsdw7NgxHTdunAI6cuRIPXLkSKmXZVwsIVQ8ZZ0QegNLPcYfBh4uon408G0B5Y2AvR4J4WegtjdB5r2qUkJQVT1x4oT2799fQ0JCdPTo0RodHa116tS56KDftm1bHTFihN5zzz362muv6Zdffqk7d+4ssI9yWdm8ebOGhITovffeW+J5t2zZorVr19a+ffvquXPnfIojNzdXX331VQ0LC9OIiAj95ptvfFpeVZeammr3fFQgubm5PiWEYh9/LSITgGGqOsUZvwnoqap3F1L/NeCAqj7ljLcAPgfaANNUdbaINAA2AQuBgc4ZyN2qelEnYhGJB+IBWrZs2aO4ngSVTVZWFrfddhvr1q2jbdu27lebNm1o27YtkZGR5eZXm6ZMmcI777zDtm3buPzyy72aJysri549e3LgwAHWr19PRESEX2JZt24dkyZNIi0tjT//+c889NBDhIRYH4qS2r17N3Xr1qVRo0b286nlnKpy9OhRTp48edH/n7ePv/ZrQhCRG4G7gQGqejbftGbAR8AoIAc4DExU1UUicj8QrapFdoqvaL+HUNXs37+ftm3bMnbsWBITE72a57e//S1vv/02S5Ys4dprr/VrPCdOnCA+Pp7333+fa6+9lnnz5lWIHygqT86fP8++fft86pdvAqdmzZpERERc9KuG3iYEvzUZAUOALcClRSxrLjABEOAXIMQpbwFsLi6WqtZkVBE9/PDDCujatWuLrfv2228roI8++miZxZObm6t//etfNSwsTFu2bKk//PBDma3LmPIKP15DqAbsAi7n3xeVO+arE42r2adtvvIIoJYz3BBXj6LOzvh8YJAzfCuwsLhYLCGUf8ePH9dGjRrpkCFDiqy3adMmrVWrlg4cOFCzs7PLPK41a9ZoZGSkhoWF6euvv27t4qZK8VtCcC2LEc7BfCeQ4JQ9CYx2hv8JHMTVvTQF+MQpvwZXV9QNzt94j2VGAkn8u6tqy+LisIRQMbz44osK6NKlSwucfvLkSW3Xrp02adJE09PTAxbXkSNHdNiwYQroTTfd5PN9E8ZUFN4mBPtNZeN3Z8+epV27djRo0IC1a9decDFXVbnpppt47733WLZsGYMGDQpobLm5uTz11FM88cQTdOrUicWLF9O2bduAxmBMoNlvKpugqVGjBjNmzCAlJcX97Jg8b731FomJiTzxxBMBTwbgehbSY489xpIlS9i/fz8xMTF8+OGHAY+jKvnll1+KfAqpKUe8OY0oLy9rMqo4cnJytHv37hoZGamnT59WVdX169drjRo19JprrgnIdYPipKWl6ZVXXqmATps2LaD3bVRmeX3hZ82apddcc42GhYUpoL169dJZs2bpnj17gh1ilYM/ryGUl5clhIpl2bJlCuisWbM0MzNT27Rpo82aNdODBw8GOzS3M2fO6F133aWA9u/fXzMyMoIdUoV08uRJ/fjjj3Xq1KkaGRnpvnGyQ4cO+sADD+hTTz2l0dHR7vLevXvrCy+8oHv37g126FWCtwnBriGYMnXttdeyZs0a+vfvz2effcby5cvp379/sMO6yLx584iPj6d+/fosWLCgVI8yLsjx48dZtWoV58+fZ+jQodSuXdsvy/WGqvL999/z7rvvsnPnTqKiomjVqhWtW7emVatWtGrVirp165Z62ampqSxZsoQlS5bwzTffcP78eerUqcPgwYMZPnw4w4YNIzIy8oL5duzYwcKFC1mwYAEpKSkA9OnTh0mTJjFhwgSaN2/u83abi/ntPoTy9LIzhIpn/fr1KiIK6DPPPBPscIq0ceNGbdu2rYaGhuqsWbNK1TX16NGj+tFHH+l9992n0dHR7m0HtFatWjpx4kRdsGCBnjp1qgy2wGXLli366KOPaqtWrRTQmjVrateuXbVhw4YXPPYE0MaNG2tsbKxOnjxZExIS9H/+5390xYoVumfPnoua9TIzM/WDDz7QO+64Q1u0aOFeRqdOnXTatGn61Vdf6dmzZ72O86efftIZM2Zo165d3cvq27evvvzyy7pv3z5/75YqDTtDMOVFQkICBw8eZM6cOeX+8REnTpzgtttu44MPPmD8+PHMnTuXevXqFVr/8OHDJCUlsXLlSlauXMmmTZtQVWrWrEmvXr0YMGAAAwcORFVZtGgRixYt4tChQ4SHh3PdddcxceJERowYQXh4uE9xp6enM3/+fBITE1m3bh0hISEMGjSIuLg4xo0b596GY8eOsXv3bnbu3MmuXbsueO3Zs4ecnBz3MvN+hax169acPn2ab7/9luzsbOrWrcuQIUPcZwEtWrTwKXaAn376yX3mkPfDNVdddRWTJk1i/PjxNGvWzOd1VGV+e3RFeWIJwQSCqvLCCy/w0EMP0bp1axYvXkynTp0AOHDggPvgv3LlSvdv74aHh9OnTx8GDBjAgAEDiI2NveCnFPPk5OSQlJTEwoULWbx48QXJYdKkSQwfPtzr5JCZmckHH3xAYmIiy5cvR1WJiYnhhhtuYPLkyTRt2rRE252dnc3PP/9cYLLIzc3lmmuuYfjw4fTp04ewsLASLbsktm3b5k4OmzZtQkTo1q0bLVq0oGnTpjRt2pRmzZq5h5s2bcqll15abp7pVR5ZQjDGR0lJSfzmN7/hxIkTjB8/nh9++IFt27YBrp917Nu3r/sMoEePHiU+SOYlhwULFrB48WIOHz5M7dq13WcOBSWHs2fPsmTJEhITE/n00085e/YsrVu3Ji4ujhtuuIErrrjCb9tfHmzZsoWFCxfyr3/9i4yMDNLT0zly5MhF9UJCQrj00ksLTRgtWrTgiiuuoE6dOkHYiuCzhGCMH2RkZHDzzTeTnJzsTgADBgyge/fufv1Gmp2d7U4OH3zwgTs5jBo1iokTJ9KwYUPeffddFi1axPHjx7nkkkuYPHkycXFxxMbGVqknkZ47d46DBw+6E0RGRsZFwxkZGRw8eJD8x7eIiAjat29Pu3bt3K/27dtz2WWXVep9aAnBGD9S1YAdMLKzs1m5cqW7WSnvG3Ht2rUZO3YscXFxDBkyxJpIipGdnc2hQ4fIyMhgz549bN0qRmJLAAAS00lEQVS6la1bt7Jlyxa2bt3KqVOn3HXr1at3QYLIG27duvVFTw71lJOTQ1ZWFqdOnSrydebMGUJCQggNDSU0NNQ9XNjfgsqGDx9OaGhoqfaFJQRjKoHs7Gy+/vprMjMzGTZsWEC7rVZmqkp6ero7OXgmivT0dHe9atWq0aZNG5o3b17ggf/06dMBi/n06dPUrFmzVPNaQjDGmFI4ceKEO0nkvQ4cOEDt2rWpU6dOsa+C6tWsWZPc3Fxyc3PJyckhJyfHPZz/b2HTunfvXupeet4mBDvnNMYYD/Xq1SM2NpbY2NhghxJw5btTuDHGmICxhGCMMQawhGCMMcZhCcEYYwxgCcEYY4zDEoIxxhjAEoIxxhhHpU8IiYmJREVFERISQlRUFImJicEOyRhjyqVKfWNaYmIi8fHxZGVlAbBnzx7i4+MBiIuLC2ZoxhhT7lTqM4SEhAR3MsiTlZVFQkJCkCIyxpjyq1InhL1795ao3BhjqrJKnRBatmxZonJjjKnKKnVCmDFjxkW/OBUeHs6MGTOCFJExxpRflTohxMXFMWfOHCIjIxERIiMjmTNnjl1QNsaYAtjvIRhjTCXn7e8hVOozBGOMMd6zhGCMMQawhGCMMcZhCcEYYwzgZUIQkWEisk1EdojI9AKm3y8iqSKyUUS+EpFIpzxSRNaJSIqIbBaRqQXM+4mI/Oj7phhjjPFFsQlBREKB2cBwoANwvYh0yFdtPRCjql2ARcBzTnkG0FtVuwE9geki0sxj2eOAUz5vhTHGGJ95c4YQC+xQ1V2qeg6YD4zxrKCqK1Q176FB3wMRTvk5VT3rlNfwXJ+I1AHuB57ybROMMcb4gzcJoTnws8f4PqesMLcDS/JGRKSFiGx0lvGsqqY7k/4MzAKyLl7Ev4lIvIgki0jy4cOHvQjXGGNMafj1orKI3AjEADPzylT1Z6cpqQ1wi4g0EZFuQGtV/bC4ZarqHFWNUdWYSy65xJ/hGmOM8eDN7yHsB1p4jEc4ZRcQkSFAAjDAo5nITVXTnYvH/YBLgBgRSXNiuFREvlbVgSXeAmOMMX7hzRnCGqCtiFwuImHAZOATzwoiEg28AYxW1UMe5REiUssZbghcBWxT1ddVtZmqRjllP1kyMMaY4Cr2DEFVs0XkbmApEArMVdXNIvIkkKyqn+BqIqoDLBQRgL2qOhpoD8wSEQUEeF5VN5XRthhjjPGBPdzOGGMqOXu4nTHGmBKxhGCMMQawhGCMMcZhCcEYYwxgCcEYY4zDEoIxxhjAEoIxxhiHJQRjjDGAJQRjjDEOSwjGGGMASwjGGGMclhCMMcYAlhCMMcY4LCEYY4wBLCEYY4xxWEIwxhgDWEIwxhjjsIRgjDEGsIRgjDHGYQnBGGMMYAnBGGOMwxKCMcYYwBKCMcYYhyUEY4wxgCUEY4wxDksIxhhjAEsIxhhjHJYQjDHGAJYQjDHGOCwhGGOMASwhGGOMcVhCMMYYA3iZEERkmIhsE5EdIjK9gOn3i0iqiGwUka9EJNIpjxSRdSKSIiKbRWSqUx4uIp+LyFan/C/+3SxjjDElVWxCEJFQYDYwHOgAXC8iHfJVWw/EqGoXYBHwnFOeAfRW1W5AT2C6iDRzpj2vqu2AaKCviAz3eWuMMcaUmjdnCLHADlXdparngPnAGM8KqrpCVbOc0e+BCKf8nKqedcpr5K1PVbNUdUVeHWBd3jzGGGOCw5uE0Bz42WN8n1NWmNuBJXkjItJCRDY6y3hWVdM9K4tIA2AU8FVBCxOReBFJFpHkw4cPexGuMcaY0vDrRWURuRGIAWbmlanqz05TUhvgFhFp4lG/GvAe8Iqq7ipomao6R1VjVDXmkksu8We4xhhjPHiTEPYDLTzGI5yyC4jIECABGO3RTOTmnBn8CPTzKJ4DbFfVl0oStDHGGP/zJiGsAdqKyOUiEgZMBj7xrCAi0cAbuJLBIY/yCBGp5Qw3BK4CtjnjTwH1gXv9sSHGGGN8U2xCUNVs4G5gKbAFWKCqm0XkSREZ7VSbCdQBFjpdTPMSRntgtYhsAFbi6lm0SUQicJ1NdADyuqVO8e+mGWOMKQlR1WDH4LWYmBhNTk4OdhjGGFOhiMhaVY0prp7dqWyMMQawhGCMMcZhCcEYYwxgCcEYY4zDEoIxxhjAEoIxxhiHJQRjjDGAJQRjjDEOSwjGGGMASwjGGGMclhCMMcYAlhCMMcY4LCEYY4wBLCEYY4xxWEIwxhgDWEIwxhjjsIRgjDEGsIRgjDHGYQnBGGMMYAnBGGOMwxKCMcYYwBKCMcYYhyUEY4wxgCUEY4wxDksIxhhjAEsIxhhjHJYQjDHGAJYQjDHGOCwhGGOMASwhGGOMcVhCMMYYA1hCMMYY4/AqIYjIMBHZJiI7RGR6AdPvF5FUEdkoIl+JSKRTHiki60QkRUQ2i8hUj3l6iMgmZ5mviIj4b7OMMcaUVLEJQURCgdnAcKADcL2IdMhXbT0Qo6pdgEXAc055BtBbVbsBPYHpItLMmfY6cAfQ1nkN83FbjDHG+MCbM4RYYIeq7lLVc8B8YIxnBVVdoapZzuj3QIRTfk5VzzrlNfLWJyJNgXqq+r2qKvAO8Guft8YYY0ypeZMQmgM/e4zvc8oKczuwJG9ERFqIyEZnGc+qaroz/z5vliki8SKSLCLJhw8f9iJcY4wxpeHXi8oiciMQA8zMK1PVn52mpDbALSLSpCTLVNU5qhqjqjGXXHKJP8M1xhjjwZuEsB9o4TEe4ZRdQESGAAnAaI9mIjfnzOBHoJ8zf0RxyzTGGBM43iSENUBbEblcRMKAycAnnhVEJBp4A1cyOORRHiEitZzhhsBVwDZVzQBOiEgvp3fRzcDHftkiY4wxpVKtuAqqmi0idwNLgVBgrqpuFpEngWRV/QRXE1EdYKHTe3Svqo4G2gOzREQBAZ5X1U3Oou8C3gZq4brmsARjjDFBI65OPhVDTEyMJicnBzsMY4ypUERkrarGFFfP7lQ2xhgDWEIoVmJiIlFRUYSEhBAVFUViYmKwQzLGmDJR7DWEqiwxMZH4+Hiyslz33O3Zs4f4+HgA4uLighmaMcb4nZ0hFCEhIcGdDPJkZWWRkJAQpIiMMabsWEIowt69e0tUXhBrcjLGVBSWEIrQsmXLEpXnl9fktGfPHlTV3eRkScEYUx5ZQijCjBkzCA8Pv6AsPDycGTNmeDW/NTkZYyoSSwhFiIuLY86cOURGRiIiREZGMmfOHK8vKPujyckYYwLFbkwrQ1FRUezZs+ei8sjISNLS0gIfkDGmSrIb08oBX5ucjDEmkCwhlCFfm5yMMSaQrMnIGGMqOWsyMsYYUyKWEIwxxgCWEIwxxjgsIZRz9ugLY0ygWEIox/zx6AtLKMYYb1kvo3LM1xvb8j++G1z3QVjXV2OqFm97GVlCKMdCQkIo6P0REXJzc4ud3+6UNsaAdTutFHx92qo9S8kYUxKWEMoxXx994WtCMcZULZYQyjFfH31hz1IyxpSEJYRyLi4ujrS0NHJzc0lLSyvRxWB7lpL1sjKmJOyisqm0rJeVMS52UdlUefaLdcaUjCUEU2lZLytjSsYSgqm0rJeVMSVjCcFUWtbLypiSsYRgKi3rZWVMyVhCMEXytdtmsLt9+tJt15iqplqwAzDlV/5um3lPWwW8OrD6Or8xJrDsPgRTKF8fjmcP1zOmfPDrfQgiMkxEtonIDhGZXsD0+0UkVUQ2ishXIhLplHcTke9EZLMz7Tce8wwWkXUikiIiq0SkTUk20JQ9X7tt+qPbZ7CbnIypSopNCCISCswGhgMdgOtFpEO+auuBGFXtAiwCnnPKs4CbVbUjMAx4SUQaONNeB+JUtRvwLvCorxtj/MvXbpu+zu+PHwgKNktopiLx5gwhFtihqrtU9RwwHxjjWUFVV6hq3i2h3wMRTvlPqrrdGU4HDgGX5M0G1HOG6wPpvmyI8T9fu236On95uNPYlwN6ZUhopopR1SJfwATgLY/xm4DXiqj/GvBoAeWxwBYgxBnvBxwF9gGpQL1ClhcPJAPJLVu2VBNY8+bN08jISBURjYyM1Hnz5gVsfhFRXF8cLniJSEk3o1TmzZun4eHhF6w7PDzc622IjIwsMP7IyMgSxeDL/jdGVRVI1mKO9arq34QA3IjrDKFGvvKmwDagl0fZB0BPZ3ia5zoKe/Xo0aMs95kpZ/xxQA3m+n1NaL4mpMrAEqJ/+DMh9AaWeow/DDxcQL0hzhnApfnK6wHrgAkeZZcAOz3GWwKpxcViCaFqCfYB0dcDuq8JJdgJ0R98OaD74/23hOLiz4RQDdgFXA6EARuAjvnqRAM7gbb5ysOAr4B7C1jmEeBXzvjtwOLiYrGEUPUE8x/a1wOyrwc0fzSZBXP/BbvJLdhfKMoTvyUE17IYAfzkHPQTnLIngdHO8D+Bg0CK8/pE/92EdN6jPAXo5kwbC2xyEszXQKvi4rCEYAIp2N9Qy8MBMZjxB/sMTTW419D8MX8evyaE8vKyhGACzb5hB+8MJ9gJxdftD/b8niwhGFMJBLOXVrCvgQQ7IVb0+T1ZQjCmiqvo37DzlhGsi9K+bn+w5/dkCcGYKi7Y37DzYghmL59gXgMJ9vyeLCEYY4Le7bMiC/Y1ALuGYAnBmHIl2N/wgy3YvYQC3cvIHn9tjDGVnF8ff22MMabys4RgjDEGsIRgjDHGYQnBGGMMYAnBGGOMo0L1MhKRw8DFv9runca4nrBaXll8vrH4fGPx+aa8xxepqpcUV6lCJQRfiEiyN92ugsXi843F5xuLzzflPT5vWZORMcYYwBKCMcYYR1VKCHOCHUAxLD7fWHy+sfh8U97j80qVuYZgjDGmaFXpDMEYY0wRLCEYY4wBKmFCEJFhIrJNRHaIyPQCptcQkfed6atFJCqAsbUQkRUikioim0XkngLqDBSRTBFJcV6PBSo+Z/1pIrLJWfdFj5YVl1ec/bdRRLoHMLYrPPZLioicEJF789UJ6P4TkbkickhEfvQo+w8RWSYi252/DQuZ9xanznYRuSWA8c0Uka3O+/ehiDQoZN4iPwtlGN8TIrLf4z0cUci8Rf6vl2F873vEliYiKYXMW+b7z++8eUZ2RXkBocBOoBUQBmwAOuSrcxfwV2d4MvB+AONrCnR3husCPxUQ30DgsyDuwzSgcRHTRwBLAAF6AauD+F4fwHXDTdD2H9Af6A786FH2HDDdGZ4OPFvAfP8B7HL+NnSGGwYovqFANWf42YLi8+azUIbxPQH8pxfvf5H/62UVX77ps4DHgrX//P2qbGcIscAOVd2lqueA+cCYfHXGAH9zhhcBg0VEAhGcqmao6jpn+CSwBWgeiHX70RjgHXX5HmggIk2DEMdgYKeqlvbOdb9Q1STg//IVe37G/gb8uoBZrwWWqer/qeoxYBkwLBDxqeo/VDXbGf0eiPD3er1VyP7zhjf/6z4rKj7nuDEJeM/f6w2WypYQmgM/e4zv4+IDrruO80+RCTQKSHQenKaqaGB1AZN7i8gGEVkiIh0DGpjrp/r+ISJrRSS+gOne7ONAmEzh/4jB3H8ATVQ1wxk+ADQpoE552Y+/xXXGV5DiPgtl6W6nSWtuIU1u5WH/9QMOqur2QqYHc/+VSmVLCBWCiNQBFgP3quqJfJPX4WoG6Qq8CnwU4PCuUtXuwHDg9yLSP8DrL5aIhAGjgYUFTA72/ruAutoOymXfbhFJALKBxEKqBOuz8DrQGugGZOBqlimPrqfos4Ny/7+UX2VLCPuBFh7jEU5ZgXVEpBpQHzgakOhc66yOKxkkquoH+aer6glVPeUMfwFUF5HGgYpPVfc7fw8BH+I6NffkzT4ua8OBdap6MP+EYO8/x8G8ZjTn76EC6gR1P4rIrcB1QJyTtC7ixWehTKjqQVXNUdVc4M1C1hvs/VcNGAe8X1idYO0/X1S2hLAGaCsilzvfIicDn+Sr8wmQ16NjArC8sH8If3PaHP8H2KKqLxRS57K8axoiEovrPQpIwhKR2iJSN28Y18XHH/NV+wS42elt1AvI9GgeCZRCv5kFc/958PyM3QJ8XECdpcBQEWnoNIkMdcrKnIgMAx4ERqtqViF1vPkslFV8ntekxhayXm/+18vSEGCrqu4raGIw959Pgn1V298vXL1gfsLVAyHBKXsS14cfoCaupoYdwA9AqwDGdhWu5oONQIrzGgFMBaY6de4GNuPqNfE90CeA8bVy1rvBiSFv/3nGJ8BsZ/9uAmIC/P7WxnWAr+9RFrT9hysxZQDncbVj347rmtRXwHbgn8B/OHVjgLc85v2t8zncAdwWwPh24Gp/z/sM5vW6awZ8UdRnIUDx/d35bG3EdZBvmj8+Z/yi//VAxOeUv533mfOoG/D95++XPbrCGGMMUPmajIwxxpSSJQRjjDGAJQRjjDEOSwjGGGMASwjGGGMclhCMMcYAlhCMMcY4/h/oywhD3NbJGgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "loss = autoencoder_train.history['loss']\n",
    "val_loss = autoencoder_train.history['val_loss']\n",
    "plt_epochs = range(epochs)\n",
    "plt.figure()\n",
    "plt.plot(plt_epochs, loss, 'ko', label='Training loss')\n",
    "plt.plot(plt_epochs, val_loss, 'k', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test how it is the output of a segment of every type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correlation N\n",
      " [[1.         0.98769604]\n",
      " [0.98769604 1.        ]]\n",
      "Correlation L\n",
      " [[1.         0.89279734]\n",
      " [0.89279734 1.        ]]\n",
      "Correlation R\n",
      " [[1.         0.96033074]\n",
      " [0.96033074 1.        ]]\n",
      "Correlation V\n",
      " [[1.         0.93428237]\n",
      " [0.93428237 1.        ]]\n",
      "Correlation A\n",
      " [[1.         0.91699909]\n",
      " [0.91699909 1.        ]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3Xd4W9X5wPHvkffeI3GcOHuHLEIgAcIqAcJeYe9RRoG2tKUDWvorbWkZZZSSsvcuI4QRIBCSkOGE7L0c20m895Z0fn8cyXY8ZVtXtpL38zx6JN17dHXsxHp11nuU1hohhBACwNbbFRBCCNF3SFAQQgjRSIKCEEKIRhIUhBBCNJKgIIQQopEEBSGEEI0kKAghhGgkQUGIHlJK7VVKndrb9RDCGyQoCCGEaCRBQQghRCMJCkIIIRpJUBBCCNFIgoIQQohGEhSEEEI0CuztCghxmAhSSoU2e27XWtt7rTZCdJO0FITwjgVATbPbH3u1NkJ0k5JNdoQQQrhJS0EIIUQjCQpCCCEaSVAQQgjRSIKCEEKIRn43JTUxMVFnZGT0djWEEMKvrF69ulBrndRZOb8LChkZGWRmZvZ2NYQQwq8opbI8KSfdR0IIIRpJUBBCCNFIgoIQQohGEhSEEEI0kqAghBCikQQFIYQQjSQoCCGEaCRBQQhvstdB5gtgr+/tmgjRLRIUhPCikm+fhvn3UPzjR71dFSG6RYKCEF6iG2rRy54CYPvaJb1cGyG6R4KCEF6y7rPniHcWUU0I+sA6nE7ZwEr4H8uCglIqVCm1Uim1Tim1SSn1pzbKXKuUKlBKrXXdbrSqPkJYzbHjK/JIoCD9DIY7drF8d2FvV0mILrOypVAHnKy1PgqYCMxWSk1vo9zbWuuJrttzFtZHCEv1q9xEbuQ4+o0+lkRVzqr1m3q7SkJ0mWVBQRuVrqdBrpu0p8VhKf9gNv11Pg2pkwkeMAkAW976Xq6VEF1n6ZiCUipAKbUWyAcWaq1XtFHsQqXUeqXUe0qp9Hauc7NSKlMplVlQUGBllYXoluz1iwGIHXEspI7DiSKmdEsv10qIrrM0KGitHVrricAAYJpSalyLIp8AGVrrCcBC4OV2rjNPaz1Vaz01KanTPSKE8LnavSuxaxsZ44+D4AiqA+MIrclDa2kcC//ik9lHWutSYBEwu8XxIq11nevpc8AUX9RHCG+LKlzHvqAMQsKiAKgPTSBWl1JQWdfJK4XoW6ycfZSklIp1PQ4DTgO2tijTr9nTcwBpbwv/ozXp9TspihrTdCwiiQRVTnZxde/VS4husLKl0A9YpJRaD6zCjCnMV0o9qJQ6x1XmZ67pquuAnwHXWlgfISxRmpdFHBU0JI1tPBYUnUIiZeyToCD8jGV7NGut1wOT2jh+f7PH9wH3WVUHIXyhYGcmsUDIgAmNx8LiUkhQ5WQVSVAQ/kVWNAvRQ7XZ6wBIHNo0JBYYlUykquVAYXFvVUuIbpGgIEQPBRRsIlsnk5aa0nQwIhmAssIDvVQrIbpHgoIQPRRfsY2soCEEBjT7c4owU6dVlayrEf5FgoIQPVFfTXJDLiWRIw497goKQbVFvVApIbpPgoIQPWAv2IENjSNx5KEnIhIBCKsvlmypwq9IUBCiB0pytgEQltp2SyGecirq7L6ulhDdJkFBiB6o2m/WY8YMaNFSCA6nISCcRFVGWXVDL9RMiO6RoCBEDzgLd3BAxzMgpXVOrobQBBJUGaU1sl+z8B8SFITogeCyvWTpVPrFhLU65wxPIoFySqWlIPyIBAUheiCmJov84AEE2FTrkxGJJKgKSmskKAj/IUFBiO6qKSHKUUZFREabpwPDY4iimrJq6T4S/kOCghDdVbQbAHvckDZPB4XHEqWqpftI+BUJCkJ0U23+TgCCk4a2eT4gLJpIaiiVloLwIxIUhOimsoIcAGKTB7ZdICSaQOWkuqrCh7USomckKAjRTdXFB6jTQfRLSWm7QGg0APVVJT6slRA9I0FBiG5yVhykgBj6x7WejgpAiAkK9uoyH9ZKiJ6RoCBEN9mq8inQsSREhLRdwBUUnDUSFIT/kKAgRDeF1BZSHhjf9hoFaOw+ctbKmILwHxIUhOimiIYiaoIT2i/gainY6svRWjKlCv9gWVBQSoUqpVYqpdYppTYppf7URpkQpdTbSqmdSqkVSqkMq+ojhFc57EQ5y2gIa53zqJGrpRCuq6mqd/ioYkL0jJUthTrgZK31UcBEYLZSanqLMjcAJVrrYcBjwN8trI8Q3lNdiA2Njkxuv0xIFABRVFMuqS6En7AsKGij0vU0yHVr2YY+F3jZ9fg94BSlVDsdtEL0HXWlZu/lwOh2pqMCBEehUUSpGsprJSgI/2DpmIJSKkAptRbIBxZqrVe0KJIGZANore1AGdBBJ60QfUNpvlm4Fhrbv/1CNhuOoAhXS0E22hH+wdKgoLV2aK0nAgOAaUqpcd25jlLqZqVUplIqs6BANkIXva+qaD8AkYkdBAXAGRxNFNVUSEtB+AmfzD7SWpcCi4DZLU7lAukASqlAIAZotdO51nqe1nqq1npqUlIHA3tC+Ii7+yg+aUDHBUOipPtI+BUrZx8lKaViXY/DgNOArS2KfQxc43p8EfCNlrl7wg84yg9SrsNISojrsJwKjSZSuo+EHwm08Nr9gJeVUgGY4POO1nq+UupBIFNr/THwPPCqUmonUAzMtbA+QniNri6mlCjSQzv+EwoIiyFKFcnsI+E3LAsKWuv1wKQ2jt/f7HEtcLFVdRDCMnVl1AZE0dlkOVtYDDGqhoo6aSkI/yArmoXohoC6chqCojovGBJFlJJ1CsJ/SFAQohtC7BXokBgPCpqNdmSgWfgLCQpCdFFtg4MwXYUtLLrzwqHRhFBPVXWN9RUTwgskKAjRRbmlNURTTXBkxzOPAAiNBUDXyEY7wj9IUBCii3KKKohUtYRFebD4PtyUCaxptfxGiD5JgoIQXVRQUAhAZEx854UjTW6kkDoJCsI/SFAQoouKivIBiIr1oKXgyqIaXl8keyoIvyBBQYguKisxLQVbWGznhSNMWpZYXUptg9PKagnhFRIUhOiiqvJi8yDUgympoTE4VBBJqkyS4gm/IEFBiC6qqXAFhRAPpqQqRV1oIkmqTNYqCL8gQUGILqizO9A1peaJJy0FwB6WRCJllElSPOEHJCgI0QUHSmuJwrUQzcOgoCOSSJSWgvATEhSE6ILc0hqiVZV5EuJB7iOAyGQTFCT/kfADEhSE6IKckmqiqcYZHAW2AI9eExidQgLllFfXWVw7IXpOgoIQXZBTUkO0qkaFedZ1BBASk0qgclJbVmhhzYTwDgkKQnRBbkkNSUG1qFAP1ii4BEabVc2OyjyrqiWE10hQEKILckpqSAis9Ww6qptrVTMVBdZUSggvkqAgRBfkltYQq6o9nnkENOY/CqjJt6hWQniPBAUhPNTgcHKgrIZoXdm1oBCVCkBojXQfib5PgoIQHjpYVkukriS6/iAkjfD8hSFRVNsiiKyToCD6PsuCglIqXSm1SCm1WSm1SSl1VxtlZimlypRSa123+62qjxA9tbuwigm2PeZJ2pQuvbYiOJlYu4wpiL4v0MJr24FfaK3XKKWigNVKqYVa680tyn2vtZ5jYT2E8Iqd+ZUcpXaZJ/0ndem1VSGpJFTnobVGKWVB7YTwDstaClrrA1rrNa7HFcAWIM2q9xPCarsKKpkatBudOKJrYwpAXXgqqaqYqnqHRbUTwjt8MqaglMoAJgEr2jh9rFJqnVLqM6XU2HZef7NSKlMplVlQIE1w0Tt25lUw0bYL1cWuIwBHZD8SKaO0otKCmgnhPZYHBaVUJPA+cLfWurzF6TXAIK31UcCTwIdtXUNrPU9rPVVrPTUpKcnaCgvRjpqCPcQ5S6D/5K6/OKY/NqWpKszxfsWE8CJLg4JSKggTEF7XWn/Q8rzWulxrXel6vAAIUkolWlknIbqjtLqeY2u/N0+GndLl1wfGDACgrliCgujbrJx9pIDngS1a60fbKZPqKodSapqrPrLDuehzduZXcm7AMsriJ0DC0C6/PiRhIAD2EgkKom+zcvbRDOAqYINSaq3r2G+BgQBa6/8AFwE/VUrZgRpgrpbdzUUfVLB7HWfYsige92C3Xh+RZIIC5RIURN9mWVDQWi8BOpx7p7V+CnjKqjoI4S0Bu78GIGbKJd16fXRsPBU6jIDKA96slhBeJyuahfCArWQ35SqKgJh+3Xp9aFAAecQTUn3QyzUTwrskKAjRCa01EVXZlIb2bJlNoS2J8FpJdSH6NgkKQnSioLKO/s6D2KMH9eg6lSHJREn+I9HHSVAQohPb9peQpgoJThrSo+s0RPQjxlkCDtmrWfRdEhSE6ETu3h0EKiexaSN7dB0V3R8bGke5DDaLvkuCghCdKMvdDkBk6rAeXSckIR2A4gN7e1gjIawjQUGITtiLXJlR4wf36DpRyRkAlB3c27MKCWEhCQpCdMDucBJckY1dBUNU/x5dK7F/BgDVhfu8UDMhrCFBQYgO5K37ktkspzpiANh69ufSLyWFKh2Co1RWNYu+S4KCEO1xOklZcB1KaYqO7156i+ZCggIpUAnYZFWz6MMkKAjRntpSAu3VvOg4k36Tz/TKJcuCUwivkVXNou+SoCBEe6rMhk62qCRCgwK8csm60GRiGvK9ci0hrCBBQYj2VBUCEBXfvXxHbQmJTSFGV5BfVuO1awrhTRIUhGhHVYnp5klOHeC1ayYlpRKiGlizW8YVRN8kQUGIdhzYb6aODh6U4bVrJqeYaa1bd2d57ZpCeJNH+ykopZIxm+b0x2yGsxHI1Fo7LaybEL2qKH8/w4BRQ3qWCK+5wMgEAPZky1oF0Td1GBSUUicBvwHigR+BfCAUOA8YqpR6D3hEa11udUWF8LWq4oOUEUVMZLj3LhoeD0BxQR5VdXYiQqzc/FCIruvsf+SZwE1a61Zfa5RSgcAc4DTgfQvqJkSv0VrjrCygNjiOGG9eOCwOgGhdwUdr93P5MQO9eXUheqyzMYVH2goIAFpru9b6Q621BARx2DlQVkuUowQdkeTdC4eZlsK4eAdPL9pJg0N6YEXf0llQWKuU+kopdYNSKtYnNRKiD1ibXUo8FYTGJHv3wq7uo9MHB5NbWsNHa/d79/pC9FBnQSEN+AcwE9imlPpIKTVXKRXW2YWVUulKqUVKqc1KqU1KqbvaKKOUUk8opXYqpdYrpSZ378cQwrvWZpeSqMqJSvDeGgUAAkMgKILBEXUMSYrgnVXZ3r2+ED3UYVDQWju01l9ora8D0oEXgHOBPUqp1zu5th34hdZ6DDAduF0pNaZFmTOA4a7bzcAz3fgZhPC6dVmFxKkKAiK93FIACI9H1ZRw4eQBrNxbzL6iau+/hxDd5PE6Ba11PbAZ2AKUA6M7KX9Aa73G9bjC9bqWO5+fC7yijeVArFLKy1/NhOiaBoeTnP2uTKYRid5/g7A4qC7m/ElpKAXvr3G9l9bw4lmw5hXvv6cQHuo0KLi6ge5VSq0B5rtec47W2uOuHqVUBjAJWNHiVBrQvP2cQ+vAgVLqZqVUplIqs6CgwNO3FaJbth2sINJeap5YFRRqSugfG8aJI5J4cekeCirqoGgnZC2BzBe9/55CeKjDoKCUWgYsAZIxU1NHaq3/qLXe6ukbKKUiMVNW7+7uegat9Tyt9VSt9dSkJC/PBhGiuZpSNu7ex1kBy83zWO8tXGsUHg81xQD8Yc4Yahuc/PGTTei9S835/WugQjKpit7RWUvhN0CG1vperfXqrl5cKRWECQiva60/aKNILmaswm2A69jhyV4Hix6C+qreroloz5uXccE3p3BH4Efooy6DNAvmPoTFQ7UJCkOTIrnr1OF8uv4A21d9CQHBpsyOL73/vkJ4oLOB5sVaa62UGqyUelQp9YFS6mP3raPXKqUU8DywRWv9aDvFPgauds1Cmg6Uaa0P30xh+5bDd3+HvUt6uyaiLU4H7F9DIbFsDpuKOvOf1rxPeDzUloLTrFH46YlDOfuo/kQcXMGW6Bk4o9Jg22fWvLcQnfB0jf2HmA/4TwBPV9vMAK4CNiil1rqO/RYYCKC1/g+wALNqeidQDVzn4bX9U7VJxUx9Ze/WQ7StZC/Ya3m84SoGnXAr40IirXmfsHjQThMYwuOx2RT//Ek8IdsK+WN+OoNUMFdWLsRZWUJIZJw1dRCiHZ4GhVqt9RNdubDWegmgOimjgdu7cl2/VlVk7utlCmKfVLANgB3OAZybbuFaTdcCNmpKGh+H5Jo5GJdfMpcfth8gaNMXvPbaM1xxy32YRrcQvuHplNR/KaUeUEodq5Sa7L5ZWrPDUWNLQcYU+qSCLQDsJI0JA7ya8ehQoa6AU1vadCxrKYTEMGL8dK656EJKQ9MYmLuANftK276GEBbxNCiMB24C/gY84rpZ1OF6GHPt5EWDBIU+KX8rRQFJ9EtOJio0yLr3CXUFnNqypmNZP8DAY8AWAEoRNvlSZtg28sWKDdbVoz37VkDhDt+/r+gTPA0KFwNDtNYnaq1Pct1OtrJihyVpKfRpOn8LW+xpTBkUb+0btQwKlQVQuA0GHddYJGT8eQQoTe3mT6ltcFhbn5Y+ug2+/INv31P0GZ4GhY2AJMTrKXdLQcYU+h6nA124nc2ONGYMS7D2vVoGhX0/mPtBM5rKpE6gNrw/xztWsmhrvrX1aammpLErTRx5PA0KscBWpdQXnk5JFW2oktlHfVZVATZHHft0MscO8XFQyFll1if0m9hURimCx83h+ICNfLNhj7X1aamuAkqy5MvLEcrT2UcPWFqLI4W7+6hB/tj6nBozoBsek0RCZIi17xUcASqgKSjkb4GkkRAYfEgx26izCF05j7pt31Bvn05woA+2VLfXgaPePC7cbuqaONz69xV9RmfbcSpXsrrvOivj/aodZpyOxlWsMqbQ99RVFhECDErrb/2bKWVaC41BYTNkzGxdbuCxOGzBjK/fxLJdhcwaaUHG1pbqKpoef/d32LYAblpkzcpu0Sd19tVjkVLqTqXUIXsGKqWClVInK6VeBq6xrnqHkZoSwBU7JSj0OVk5ZrOboQNb5WO0hjso1JRCeS4kt5F0ODAElTaZaQHbfbcZT/OgsG2Bud/T7ndCcRjqLCjMBhzAm0qp/a4Nc/YAO4DLgMe11i9ZXMfDg3s8ASQo9EH7ck3KrVGDfbRncmgM1JZDgSu3ZHLLrUYM28DpjLPt5fO1u9lyoFv5JLumeVBwy/rB+vcVfUZnuY9qtdb/1lrPAAYBpwCTtNaDtNY3aa1/9EktDwfu8YTQWBlT6IPyC/IAiI33QRcNQGi0aSnkbzbP2wkKDDyWAG3n2JAsHlqwBct7at1Bwb3ALn26ydnl9PG0WNFrurLJToNr4xxZYtkd7pZC3CBpKfQxTqemrNg17TPUwpXMzbm7j/K3QHAUxAxou1z6NABuH1bI9zsK+WS9xfki3UHhmFth+m0w9TqocwWvhhrI22zt+4te54PpDAJoainEDpSg0MfsyK8k1F5BfWCUWVHsC82DQvIoM/jclvB4SB3P5IpvmZAWzYOfbDIb8ljFHRTGXwSz/woDjzXPd30DX/4enj3BLLYThy0JCr5SmQ8oiJGg0NeszS4hWlWhwny4PjM01gSFwh2QOLLjstNuRuVt5Mnp5VTW2bn+pVVU1dnN63MyvVuvOte4RUiUuY8bBINmwtInYM2r4GyAHV949z1Fn9LZzmtPK6VmdFRGeCh/C8QPMd8QHXXSR9uHbD1YQbytmsAIH6apDo0xObAqD0LCkI7Ljr8EIpIZtO0Fnr58Mpv2l/HYwu2w5HF48QyztsBb3C0Fd1AAOO1PpqXrbDBpv7cu6Pw68sXHb3XWUtgO/FMptVcp9bBSapIvKnVYytsEKWPNYiCQP5o+ZHteBSlBtT5uKTQbu4gf2nHZoFAYex7sW84po1M4b2J/XluRRf2+VWahmTe37qyvBGWDoPCmYwOmNo0xjLvQdCU11LR/jX0r4G8DJamen+ps9tG/tNbHAicCRcALSqmtrjTaI3xSw8NBfRUU74aUcRAc3nRM9AnbDlaQEFDVNOPGF5oHhYRhnZePSjUf2DWlPJx1KWc5v8WRu86cq/Di4HNdhRn4bjnGccbf4fS/wKizwF4DG95t/xr7loHTbtJ3CL/j0ZiC1jpLa/13rfUkzPqE8wDJmOWp/C2AhtRxEOzazUumpfYJhZV1FFbWE6WrIMzH3Udu8Z10HwFEJJn7gxsIrM7jV+GfEuZw9f+Xe2FhW1kOfPN/ZjFd866jlgafaKapLnygadOolvI2mft8+YjwRx4FBaVUoFLqbKXU68BnwDbgAktrdjjJ22juU8Y2NcslKZ7vVOS1e2r7QdOHHmovh97oPorq39R67EiEa/3EQbO/QkpDdtM5b3QfrX0TFv8DcjM7Dgo2G8x5zAxy//BU03Fns116JSj4tc4Gmk9TSr0A5GA22fkUGKq1nqu1/sgXFTws5G1yzUUf2GxMQVoKPpG7Gh4ZAQfWtXl6W14FIdRjc9b3TvdRQifjCW7NWgpuDmzUE4j2Rkshz3Xdop0dBwWAlDHQ7ygTQMAE3b8Pgi3zwV5vEumBBAU/1VlL4T5gGTBaa32O1voNrbV0hneF1pC1zPwh2Wwy0Oxr+9ea+/aCwsEKBoW5Zu/4sqUQEm3uPQ0Kka6g4P7wDoqgImooB51xFB3Y2/P6HNzYrG6dBAWA1PEmQGkNOxeaqazr3zIBwWk342flOSaVh/ArnQ00n6y1fk5rXdLVCyulXlBK5SulNrZzfpZSqkwptdZ1u7+r79ElDTWQvcr3U0E3f2i6jya78ga6g4JsyekbRbsOvW9hW14F4xNcqSN82VIITzDjS833UOiIu6WQvxVsgXDuk4TP/iOFtgRK8rJ6Vhf3RAg3T4NCTYkZz9j5lTm28xvYv8Y8HnehuXfndhJ+w8rFay9hEup15Hut9UTX7UEL6wKbP4LnT21q2vqC0wlfPwjJY+GoueZYkMw+8qniXYfeN+N0arYfrGBMnKs/3JctheBw+NlamHy1Z+WDwkwXpLMBIlNg3IUEj51DaPwAgqrz2VfUg+5I90QI5fo4CIns/DWp4839gbWwa5FZqd9QBd8/YjYMGn2OOe8eX2hu04dmMZzokywLClrrxUCxVdfvsv6uJRb7fZjDr6bEfAObdGVT+gT37CMJCr7R2FLY3epUbmkNVfUOhkfWmgO+bCmA6RLqSlqNiETX61IaD6VnDCWFYm58eSVlNQ3dq4d7nMK9Hai7a6sjKWPN/arnobYUTvqd+b9dkgWnPWhmVEX1g+2ft37t0n/BMgkKfVVvp7k4Vim1Tin1mVJqbHuFlFI3K6UylVKZBQXdzLuSMMz8p3X3MftCjavXLbzZ9o6yTsF3HHYo2WseF+8+dIYMZjyhH0Ucs+MRs1LX0/793hLpmoEU1a/xUFRiOmGqnqji9Tzw7oruXTdvo2mFDJllnnvSfRQSBXGDYdfX5nc3YjZc8F+46gOY/lMzfnbUZbDjy0OnzNZXw8H1UFXQ8QI40Wt6MyisAQZprY8CngQ+bK+g1nqe1nqq1npqUlJS997NFmBmTPiypVDrSijbvFsiMAxQsk7BF8qyTXdL/8lmwVWLRV7b8iq4J/A9gutK4ar/+S5Dane5xxWimloK7gDxfuAfuGnnbSxa243u0ZxM87fh3nbTk6AAZqVzaAxc/aH5Pz7qTBh6ctP5SVeCdsLa15uO5a42A9EAZbldr6uwXK8FBa11uda60vV4ARCklEq09E37TTTfUhx2S9+mkWvf30O6JWw284d0JA7AbZkP799opi12xl4Hb1/VsyDu7joa/hNz32JcYdvBCiYG7UMNnA79PRzw7U3uoBCZ2nQsLgMA54BpDLflwoe3k1XUhVZobbn5mxh0XNPKak+6j8CsV7jDFVDakjAUMo6HzJea/uaylzedL8tu82Wid/VaUFBKpSpl1tIrpaa56tLOEkkv6T8J7LW++0Buq6UAcPSNZuD7SNvRavm/TXqEL39vpjK2lL0KiveYx1lLYcvH8OPrrct5qrhFUGgxA2nHgVIG65y2t8Lsi9pqKaRNges+w3bdZ9SNOIeR7OHaF11ZVD2Rs9J8mx90nNno58x/wuizPXttSFRTl1Z7pt9mpqZucS1r2reiqUVWluPZ+wifsiwoKKXeBH4ARiqlcpRSNyilblVK3eoqchGwUSm1DngCmKut3lbK14PN7jGFlukTjv85RA+A+Xf7bGzhX1/t4PqXVlFSVY/TeeivWWvN8t1FfLxuPz/sKqK63u4+0faHd3dUF5sdvCJTYOWz8PBg2PBe03mnE167AObNgtw1ZkYLmDUe3VFfBZkvmu6V/pPMjBj3+AJwsKyW2oJdBNHQ/q5nfY37A7h5S0Ep84EeEEhUfCrJQTXsKaziv9+3Hlg/RE0JfPwzWPOKmeKaPs1ca9pNZg8Hbxkx2yT8W/aU+b+UswpGngUoaSn0UYFWXVhrfVkn558CnuqojNfFDzFN4wNrgausf7/aNrqPwKxVOOcJeO1C+OQuM0DX3iYrXrAzv5InvtmBw6k54R+LqGtwct3MDH59+iiqGxz86r11LNjQlCohMTKY+XfMJPX1k2Ds+TDr1x2/wcb3TZfQxA7+yXcsBO2AS14xq2YzX4CP7jCLnJJHQeleswDKFghvXNL0bTJ/kwko7X1QVebD8mdMX/bg45uOf/E70yK86n+myy4iyQxuuizcfJARyvVNNXlUxz9fX+HenS1uUNvnw+IIbKjk7HEJPPvdbi6fNpDk6NC2y657C9a8bB6nTWlaP+NtNpvZve3L30P2CvM3kX407P5WWgp9VG/PPvItm823g801pWZdQmBw63PDToGTf2e6UzrKONlFLVsBAH//fCthQQH858rJzBqZzKyRSTz73W5++vpqbnk1ky835XHv6SP56ucn8OxVU6iotfPv9z+Hgi1mkLCz1sJ3D8PnvwFHO1Mid34Nq/5r8vcMmGYGIOe+aT6I5t9jyrhX1J71CFQXmcDh7vZpr7WQ9QM8MRmWPAqLH246Xl9tPvQmX01N+gks3JxHWUAs9vLYFD4QAAAgAElEQVSmHEhfbMrjmEjX8yQ/CQrDT4frv2y/u8vVIv3Viak0OJz8+9u2F+wB5v9c3GDTknKvKbCKe/e2zBfNfeoEE+CkpdAnHVlBAcyA4sGNng129lRNacdz32f+HNKPgQX3QmnP/kDWZpdy/r+XMuL3n3HBv5eyco9ZIrKnsIqFm/O46fghzB7Xjycvm8SzV03hd2eOZtHWApbuLOJvF07g9pOGMSw5itPHpnL3qSNw7F5sLlyaxf8+/4K73vqRLQfaSFnQUIsu3AG1pTh2f9/6fPFu0y20/0c4+gYTmMH0ix93h0mzXLzHLHJSNrOhzNE3mTIn3AuBobDnu0OvWVtmuj3euJTqkES+CphJ/Z5lfLbG1WWy5zuw17AkeCbH/e1rbnolk9WFQezN2kNVnZ0DZTUs313EsVH5EDvIum/J3mazwcBj2j/vCgrpoXVcOHkAb6zcR355betyRbvMLKCp18M9m2DGXRZV2CVlHNiCzOp+ZTPddTEDpKXQRx0xQcHh1NQ2OCiPHw+OOv715sc8tnB7qwG5fUXVzH58Mde8sJIfdrU97r1kRyFnP7mEFbs7GRevLW1zleym/WUcLKs102TPfdqk3ph3oulzB/PNfMN7hwSK8toGvt6Sx40vr+KGl1axt9CMRdTZHby5ch+XzVtOfnkdV04fxMGyWm55NZOymgbeXLmPQJvismPSG6+llOKmE4bw6c9m8uK1R3PRlGabxm/6kFs3zOWqmLUU60icWrF3yTt8sekgZz+5hGl/+aoxQHy7LZ+VK5eitEkd8unb/2F5y9/J3qXm/pbFMOs3h54bd5G53/CemSsfP9Ss4zjtQdPtkz7N5O/PfKGptVC4A56ZCR/fSXlQAqcW3MM3IacQjJ1333udt959k/K1H1JnC+e6b0MYlhzJazccw9DBg4mwl3DpvB+49oVVhAQqhjp2+894gifcY1fVxdx20lAcTs28xW2MLax/B1AmFYUtwNKuS8BsEpQ6zkzySBhu/o1jBpgpqS3WjojeZ9mYQl/z+caD3P7GGgaqKhaHQNXeVfx3cxgf/JjDnScP59yJ/VEo7nxzDbklNZRWN3D7G2v46ucnEh8RTFWdnccWbmd9bhk/7iuhwaG56ZVM3rn1WEalNk3hW7qzkA25ZUSHBnFeeSHhLVoKry7P4v6PNhKgFBdPTee+M0cRfdM38PqF1H50D4tOfJ9ZWY8TtvpZasNSyD37Lebvj+KpRTtocGgSI4Opszs55dHvGJoUQU5JDdX1DiYNjOXZq6aQHBXKRVMGcPZTS3jwk818szWP08akkBzVum95eEoUw1NazEnftxxVtJ1RwK7+Z1FbncutAVu45oZTmLd4N3nltcxfv5+P1poFSRcHfMu0ICiNGc3MsuVMm7eEq2YM41enjyIsOACyluEIi2eHcwCD6h3mmEtZSCpVMZMJXPoySWGg0qaYE0GhTfPd5zxGbfaP2F6bi5p5N0HLnwBbIBtPe4PzPoXpwxJ54PLR6Ef+yr9D/03oJrP+Y75jOudOyeCh88cTHGiDrAyc2fMprarnYHkdH51STPD3u+F4i78l+5I7KNSUMGhQBHMm9OOtVdncdWwMUauegpP/AAFBZixh2KkQk+a7uvWfbFqL/SaY57EDzba0VflmAyHRZxwxQWFkahS/nj0KGyNxLIvht+PrOG38sfzhw4386r31/OOLbYQFBbCvuJpnrpjM4KQI5jyxhAufWUa93Ul5TQOV9XamDIzj7KP6c8PMwVz34ioufXY5f5gzhqSoEIoq6/jFu+sau+AnBe/HGZNOem0DUSGBPLpwO09+s5NTRiUzIC6MV5dn8e22fB44ewy22Ev5yd5/UPzOHYQFfs3/HDOYWb2B6LfPJd9+Ee/Fbafh5AcZN34iJdX1vLFiHxtzyzh2SAKnjklh5rBEXDN8GZcWw/mT0nh/TQ7BATZumDnY819Us37eoUfPNl01X/6esPr9/OYM0/d+26yhrNlXQnp8OInff4rOCSP2zAfgzbn8fcQ2frE0kO+2FTB3WjrnbfyGNfVDufVfS+gXE8o1x2WQX15HUlQIzy/ZwzHVM3k6+AmopylpoEtFbQN3vbWd7fk/479BjzB60Z/IjxhO3uz/ct1HhWQkBvPMlZMJCQ2C9GMI3fs9RcMvxnZgLUeddDdzpjSbPx+RjE038O0dEyl1hJD48vFmLGHild36/9QnuQfjXbPebpg5mI/W7mfXx/9kYtYLZp1OUKhZxDfncd/WLW0KZD5vxhPAjGeA6TqUoNCnKKtngXrb1KlTdWZmZs8u8uJZZqXrDV+itWbJzkJeWLKH2gYnN8wczKljzDzweYt38d7qHEb3iyY8OIALJg/g6IymWTDZxdVc/cJK9hQ2TSudmB7Ly5cMoqool4gPrmJhzUgeDb+b/rFhZGaVcOnUdP5y/jgCA2yszS7lF++sZVdBFVFUszrsdoJ1HXuip1F8wVsEle1h+OdXEFbjmhk09GS48gOPmvtVdXZ+3FfK2P7RxEW0MdDdnmdPNF1e4y6C8ReZDVyemAjTb4cdX8Dsv8PwUw/9Xdpr4cav4D/Hg72Gpacv4FcfbMJZmsMPoXfy/dBfUDTuBp78Zge7CqoIDrBR73AyKjWKhy+aQM7X8zhx9z/5TcSfKYo7iqFJkVw4ZQC//WAD2/IquPf0kUxIDmLDN2/zaPYw6ggmJNDGx3fMZGSqq6WTu9okdpt4Rdu/nw3vwfs3wO0roaoQXjoTLn7JzK46XNSWmb2Rf/J/cNydAMx9ZglP5F1FMsVm/UFVoem2uWtt1/Iu9VRZDrx4hplgkDrOBIMnJsI5T3qeFFD0iFJqtdZ6aqfljsig8NHtsOMr+OW2HtentsHBroJKymvs7CuuYvbYfsR8epNJI+xsIG/EXK7JPZ/qegc3nTCEK48Z2PiN3v36zzYeYHxaLMOW32cWtf10WdP0w7IcM8um8qCZ1nfK/TDtFjM42lBjxi1qSs23v9pSGDUHAkNcFy+HFc/ChEvan8bY0sNDzYfH2c2+ST4zo2n3uGGnwvnzTIsidYJZbzD2PDj7Xyb75bvXwLDTqD3jMep2fEvM53fAzd9B/4nU250UVdWREhXKwfJakqNCCAyw0eBw8vLS3SzfU0JJdQPrskuxOzUxYUE8PnciJ41sWiCVXVzNij3FpMeFccyQBDy2+1t45Vy49lOT1uGrB+De3RDRhWv0dVrDgwkw7WYo2gHDf0JWVRCDFt9DfmB/khwHUdppFqhNu6l36+p0wP+lwLG3mTEkYTlPg8IR0310iLgM8yFbX+3ZVogdCA0KYGz/GMhaxrHL74SYv5o5+a7tNlOSUvn84hM6fP35k1wB4IyH4eT7mzZUARMcJlxs0gTs/Nqk4v76QTOf39nGqtXUCTD3dTMv/83LIGsJ/PCk+UY25tyOf5iGGqgubApIbqPOahoI3vWN+cZXvBsumGcC0aCZptyYc+H0v8I3fyb0y18RqmxmyqOryyA40Ea/mDAA+seGNV4+KMDGjScM40bXr2ljbhmfrN/P9TMGk9Jinn16fDjp8d34N3NvZ1mZbxZQxQ85vAICmBZSWBzsXmTWaOz8ikFAZUgqv664gheD/0FpSH8+tZ/EFb1dV1uA+TdoZ58LYdQ2OKipd3Sttd9DR2hQcPVnlmZ5J8XBgXXw6gUm6dr7Nx66/3JXNoMPCjO3tgQEmhk5+34wt7pKCI02U17DYs2HXlU+fHyXCQYx6SZVxOkPmcVl71wNM+6G0/7U/vu7E5TFpB96/JhbTZAZfCI8fTQUulpYH//MZJ4ddZZ5rpT55ldVAEsfN9MQp1zbNA3VQ+PSYhiX5uXkdO7VwFUFJigMmeXd6/cVYXFNaVzOehSCI4kYejKJn+5l8cbPeanidL79ZBtHDUry/u+4qxKGmfUo4hBOp+apRTv5ems+m3LLcGrNqaNTSIkOZebwRE4fa+0YzBEaFDLMfcle7wSFrZ+amRQn/Q4W/cV8UAaGmm/d3ty4xZ3SYNBx7ZcJjoTXLzbf7N3dBEffBJ/eYz6oR5ze/uvdg8wtWwrh8U3dDeMvNgEib5NZDzDxytatrclXmwVljrrOWye+EhYPKsCk0KjMgwFH93aNrOH+EhIcZdYhKIUCHroogafiX+CSftFs+Ggjv35/Pf+7bYaZmdVbEoaarTydDt+Ob/RxH6/bz6MLtzNpYCw3Hj8Eh9PJ/PUHyMwqISkqRIKCJZoHBW/I32JaH8f9DFa/DBkzTO6drfN9v3HL8NPM2gftaBrACwyGM/4Bu741C+VuWdz2H6F7MVHLoNDchc+Z+60LTFCY3Ea6kPjBZuwhbxMMnN6jH8drbDazSc2OL8zzAZ12rfond1BIGnnIgHtQgI17ThsBmMO3vLqaR77cxn1n9mIywMTh4KiH0n3m/4ygtsHBP77Yxri0aN6/9ThsNvNv+LuzfLee5ohZvHYI9/643goKBdtMiyMo1HzgznnMLLwC327x6DbpitYzOoLDTVqNvI1mpk5bynIABdH9O3+PUWfCz7e2/6F/wX/h+s/71jfAiGQzXTN+qFllezhyT0vtIHXH6WNTuXL6QJ5dvJtVe3txc0R3qu62xhWcDt+luO8jKuvs/PS11eSW1vDbM0c3BgRfOzKDglKmtdDdoNB8xpa93qRoThppnkckmJlBY883OWX60orZ4acDCnZ/1/b5shwzMBwQ5Nn1ovu1fy48vqlF1lfEDDCJ9i5/2/Of0d+4WwqdJPn73Zlmbc1jC324Z3lLia6/mYPrW597/0Z4xwdJK/uAspoGbnw5k4l/+pLvthfwl/PHcdxQa7eW6ciR2X0E5gOrO4Ncq18ym5PfutQM9BbvMrOAWn4zix0Il77qjZp6T0SCWVG6e5HpPsnNNN+YR55hzhfv6rjryN/NedRs3nM4d1U0dh91HBTCggO45YQh/N+nW1i5p5hpg+NxOjUf/JjL8t1FDIgL48rpg0iMDLGurhEJkDQa9n5v0sk3l7XUTLVuqDUt8MNURW0Dlz77A7sKKrluRgazx/VjyqAuTE6xwJEdFHZ+Zb71dyX3y/YvTB/oimfhxHubZnq4Wwp93ZBZ8MPTJm23K2cRM+6C4+6C7JUw8+7erJ21POkW83fR/ZuSznXiimMG8d/vd/PT11bz5/PG8dnGg3yybj9x4UGU1jTw4tK9PHfN1EMWbHrd4BNMckN7fVM24aoiMxkAzBeXjJnWvX8ve/jzbWzLq+Cl66Zx4ohubjXsZUdm9xGYbhJ7rVkF6imtzcInMHP/i3aZ8QSUSfTlD4bMMi2bmDT45U6YegMs/Rf872YTJEbN6e0aip4YfzHctMijvEZhwQG8cdN0QoMCuO31NXyybj+/nj2KNX84jS/vPoGEiGCufG4F67JLravv4BPMVO7cZgtS8zc1Pd67xLr37mVr9pXw2oosrj0uo88EBDiSWwqRzRYzlWWbD/vk0WbgtGiXmU3Ucn59WY5ZCzDlWpNp8qmp5ltZXEaPF8H5zKAZZhrptJvMIrkz/2kSle38yuwG596dTvinwJAu7Tc9NCmSz+8+nq0HK4gLD2ZYciRgkiW+e+uxnPPUUu5880f+esF4ckqqySuvY/yAGCanxxET3jQuU1lnZ/66/aTFhXH88C58wGXMNH9DexabqdJVhWY2H0B02mEbFLTWPPjJZpKjQvjlTzzsZXjnarOT3cTLLa3bkRsU3PvdVubBm3PNgrPwBLj8HXjuVLjk5dZz7N3fZiZfA7PuM5uG2GvNNFB/ERgC5z3d9Nxmg9l/gxd+YhahWZ1GWfQ5UaFBbXYRJUSG8MRlE7n02eVc8dyKVudnDEtg9rh+ZBdX8+bKfVTUmtlC5xzVnztPHtY6A29bwmLNF5EdC83akdcuNF2xYXEw5jxY9ZxZad/eok4/9emGA6zNLuUfF00gIsSDj+H6apMCJ3W85XU7coNCpGvz87xNJiAMmWXy43z4U0BDQRuzMnJXQ0CIGZwNDIaT7vNdfa008Bi4+qOmDJZCuEwZFM+iX84it7SG5KgQkqNDWZ9Tyordxby9Kps/fLgRm4IzxvXjuhkZLN5ewLzvd/PphgO8esO0xlk0JVX1zPt+N2v3lXLqmBSun5HRlANs5JnwzZ9hyWOYv72tJnXK8NNg+dMmvUv+FkibbHYs9HOVdXb+8ukWRveL5oLJHk7sKN1n7uOsnyRhWVBQSr0AzAHytdatJoUr8z/iX8CZQDVwrdZ6jVX1acXdfeTemnPKdXBgPRS6gkHZvkPLa22CRr8JbW+v6e+GzOrtGog+qmW+qeOGJnLc0ETuPHkYhZX1hAUHEBNmupKmZsRz7YzBXPDvpdz3wQbunzOGH/eV8taqbEqq6xkYH86f528mu7ia++eMMXPxR80xQWHv901TxZNHm66lsHj49q9mfU1UP7hztf/slNeOf36xjYPltTx9xWQCPF2L4J4+H+thYssesHKg+SVgdgfnzwCGu243A89YWJfW3GkP3EEhLsPsRAVmX+WW22Nu+QQObpA0v0K4BAbYSI0JbQwIbvERwTx0/niyiqq54eVMnvluF0MSI/jkjpl8/fMTuXHmYF5atpdbXlvN11vyWFWVREPsEADyT/gbHP8L83cWEGQWSeZtNItNKw6YmXN+bPH2Al7+YS9XTx/E5IFdmHrqDgo+WPtjWUtBa71YKZXRQZFzgVe0yd29XCkVq5Tqp7U+YFWdDmGzmXEFd8sgZgCc8Euz9H7vEsjf3FTWXm8ykyaOhKOsHeQR4nBw3LBEXr5+GsEBNo5KjyE8uOmj5vdzxpASHcrDX2xl4WYz9fS2gKlcFFDNaW838KszruCWfkNN4THnwY+vwcx7zBe4ZU/C9NsgJLI3fqxu01rz7fYC7n13HcOTI/nNGV1ML1KyF4IiTKoWi/XmmEIa0PzreI7rWKugoJS6GdOaYODAgd6rQWSSSaEdGGoGmZWCY24x/Xc7vjSzjQq3Q85qk5/+8ndMtlIhRKc6mmZ50wlDuGL6QDbvL6emwcHewrHMr7qXU/aX89fPtpIaE8q5E9NMDq25b8Cw09i4ejHjts6H9W/B0Td6XpEVz5pZTSf/zgs/Vddprfnlu+t5f00O/WJCeeryyYdsS+uR0iyzJ4oPJoL4xSec1noeMA/MJjteu7A7x350/0N/2bEDzayiD242KyvBpK0YcbrX3lqII114cCBTXbOe3NNY6+wOrn5+Jb98dx2JkSHMGJaIffgZ/N+nW3hpWS0fBQ9hwDdPEj/lepQnKdkX/RW++5tJb9ILQSG/vJZHF27n/TU53H7SUO46ZUT3MtOW7PVZ2pjeXLyWCzRP3D/Adcx33DOQolss9HHvJ5C11EyXGzXHbIAjhLBUSGAA866eypDESK5/aRUvLt3Dja9k8tKyvVx73GB2DriQhJq9fLt0aecXqy6G7/8JIdFmkWptufU/QDN7C6s4+ZHveDszmxtmDuaXPxnZvYCgtU+DQm+2FD4G7lBKvQUcA5T5bDzBzb3DWct8P7HNYtVxdzYNQAshLBcTFsQbNx3D9S+t4k+fbCYoQPHn88Zx1fRBOLJr4fl/MH/RYr4vTaBfTCiXTksnOrSNBIfbvzCr96dcY8YiyrIhdKxPfgatNb/7cAMKWHjPCQxL9mDNRnuqCqGh2v+DglLqTWAWkKiUygEeAIIAtNb/ARZgpqPuxExJvc6qurSrsfuonZYCCgbP8mWNhBCYhXNv33Is2w5WMCw5snGBV0CiGYBObsjlhRVZ1NmdPP7Vds6fnEZSZCgZieGcOCKJ2PBgM2MwOg1Gn2uCQmk2pPgmKDy/ZA9Ldxbxf+eN61lAADOeAD6ZjgrWzj66rJPzGrjdqvf3iHutQss8MWGxpsl5OO7jK4SfCA0K4Kj0FvuRhMVCRBJ3D1X84tzT2Xqwgue+383bq7JpcJjhxv4xofzvxomk7PrapKSJdU1OKWsxzdyLymsb+M+3u7hoygA25JbxlwVbmD02lcuneTgxZst80xJIbWOfj8p8c+/+vLKYXww0WybKtR9AbBv/cMfdaaanCiH6loRhhJTtgQAb49JieHzuJB69ZCJ2p2bV3mJufiWTp158kT/ba01a+Igkk4nA/Y3by+rtTm59dTXLdhXx6vIsKuvsHJ0Rz+NzJ3q2UU7pPnj3GpPXaO7rrc/XuDZCCrcwW20zR26WVDDJ4S56EYac1Prcib8yM46EEH1LwrBWe6HYbIrgQBszhiXy36unMqJ6DbU6iBf3pfDqin0UBSaTn70TrVtPXnQ6ezah8cH5m1i2q4h7Tx9JclQIZ47rxyvXTyM0yMNppz88bcY+8ja1fb7aFRTCfBMUjuyWgs0G4y7o7VoIIboiYRj8+KqZTRQa3er0ccMSOTphF9urxvOnz3cDMDg4lsis7XybmcMlR5sxQ601H63dzx8/2cTUQXE8dP54kqO7tqHP/PX7eW35Pm46fjC3nzSM22YNbcrp5InqYrOve2AolOyBusrWC/NqisEWCCE9HJvw0JEdFIQQ/se9t3PxrrZTvVfmE1S0lTGnPMBLyUcTFRrExB8nUbH+E+Z+vJHSmnpOHZ3CF5vy+PvnWxndL5rvdxQyd95yFtx1/CHf8CtqG3h7VTbvZuaQGBXM5IFxHDc0kWMGx/P+mhzu+2ADkwbGcu/pZqe7LgUEMJlP7TVw4m/Meor8LZB+9KFlqotdaXl8k8FYgoIQwr+4g0LhjraDwo4vAVBDTmRWmmtwdu9AYp0lDI8P4qEFW3logdkx8Zyj+vPYpRNZurOQq19Yyb8X7eTnrv0N3lixj78u2EJFnZ0pg+IorW7g6UU7efKbnUSHBlJea+fYIQk8e/WU7q0/ANj4vtmga+LlJijkbWgdFGqKfTaeABIUhBD+JmEoBIaZXEgTLjn0XMF2+Py3Jr19v2abDbnWIn1y9WAOBPbnzZXZFFbW8cDZYwiwKU4YkcQFk9J4atFO6uxOEiKDeWjBVmYMS+DXs0cxYYCZBVVVZ2fBhgMs2pbPKaNSOGdif4ICuhkQyvebPGuzfmMmu4REtz2uUF3is/EEkKAghPA3AUEwYCrs+6H1ufn3mPOXvQm2ZgO94a6p5dXF9Esfys9PG9HqpX8+bxyBAYpnF5txiGkZ8Tx/zdGHdCdFhARy8dR0Lp6a3ur1Xfbja4CGcReZrqGUse0EhSITCH1EgoIQwv+kH2M25VnyOGz+0CSrdDpMappZ97WeZu7ufnFP72xDREggD190FHecNJyqejvDkyMJ7G4roDNVhbD0CZNCJ9HVHZY6Hta+YX6O5gGtphjCplpTjzYc2VNShRD+aeCxoB3w1R9NN9Ibl8La1wENY89rXT7MtXdBdftBofHSCeGM7hdtXUAAE9AaquCU+5uOpU0xu0AWbGs6prWpc7jvFtFKUBBC+J/0owEFaJj1WziwzuzeljzG7PHckgctBZ9x2GHdmzD6nEPrmuZqDeSsajpWXwnOBp8ONEtQEEL4n9AYSJ9mul9m/dp0H4XEtL8zYkgMKJtHLQXL7f3ejBOMv+jQ4wlDITQWcjObjvl44RrImIIQwl9d/bH5oAcYfir8apdZ5NUWm810IfWFlsKm/5ntRYedeuhxpcwAes7qpmM+TnEB0lIQQviroFAIDG56HhDU8QKvsPjebyk4nSZ764jZEBTW+nzaVLMVcE2ped4LLQUJCkKII0N4fO+3FIp2mjoMbSPfGsCIn5jA9r9bYf9akywPZPGaEEJ4XVi82Xe9O1pOE+2u/WvMff/JbZ9Pm2J2eVzwS9j+WdNxaSkIIYSXdbelUFcJDw+B9e90/bXbPjNbgbrlroagiLZnSLlNuwmumQ/nPWMG1FFNU2p9QFoKQogjQ1hc98YUyvdDbSlkvtg6rUZHqgrhzblw0u/hxHvNsdzVJl9TZ62Owceb+35HmdcE+O6jWloKQogjQ3i8yUjaUNN0rLDZvgy15bB1ATgaDn1ddaG537fMbOnpqco8c+/uMrLXw8ENkNZGEr/2pIxtf5qtRSQoCCGODO5+eXdrYf278NQUsyIaIPN5eOsyePoYqMhrel1VQdPjje97/n7ubTRzXUFh0wfgqG9/PKGPkKAghDgyNF/V7LDDor+Y5+7WQtEuc1+8C3Z90/S6KldLITzR5FbylPt1lQfN7mr/uxUGHA0jTu/+z+ADlgYFpdRspdQ2pdROpdRv2jh/rVKqQCm11nW70cr6CCGOYM1bChveNTudAZS7ZiSV7IWk0a4yRU2vcz9OHn3o8c40b2F8+XszPnD1xxAc0a3q+4plQUEpFQA8DZwBjAEuU0qNaaPo21rria7bc1bVRwhxhGveUshdbVJfhMY2TVMtzYLUcWALahpHAPONPyQGIlOgpsTz96sqABVgbtoJJ/8BgsO99/NYxMoh7WnATq31bgCl1FvAucBmC99TCCHa1rylUFNigkRwJJTlmsHlshyYMNdkJK1qHhQKICKh67OXqgogMtm1wY+CYad49cexipVBIQ1oPlSfAxzTRrkLlVInANuBe7TWrYb3lVI3AzcDDBw4sOVpIYToXPOWQm0phMWab/9lOeamnRA3CCISD/3wry6EiCTz+toyzxeyVRWaa13+jinvoz2We6q3B5o/ATK01hOAhcDLbRXSWs/TWk/VWk9NSkryaQWFEIeJwBCzcKy6xLQUwuLMt/jyHDOeABCXYT78D+k+KjKDzGHxgD50MVpHqgqagklojJd/GOtYGRRygeZ71g1wHWuktS7SWte5nj4HTLGwPkKII517VXNNqRlPiE4zASJ/izkfO8gEgObdR9Wub/xd2KgHaAoKfsbKoLAKGK6UGqyUCgbmAh83L6CU6tfs6TnAFgvrI4Q40rnHBWpLm1oKYKaa2oIgur8ZU3DPMnI6m7qBGrufPBxsrir0y6Bg2ZiC1tqulLoD+AIIAF7QWm9SSo6j4p4AAAmvSURBVD0IZGqtPwZ+ppQ6B7ADxcC1VtVHCCFM11CRaSmExR4aFGLTTd9/RKIJGo4GqKsw2342dh/Rfv6kynwTBJSC+iqz3WZEom9+Li+yNKGG1noBsKDFsfubPb4PuM/KOgghRKOweMjbbD7o3d1HYL79H3WZeezeD3nHwqbVzhFJJohA291H+3+E/54Mc9+EkbObup+kpSCEEH1YeDxUudJPhMWZ7qLAUJOk7tQ/ucq4gsKnv4CK/eZxRELH3UfLnzGzl/Z+L0FBCCH8RvN9CcJizW5tNyyE+MFNu7i5u3zcAQFM95F7n+eW3UeV+bDxA/M417WVpns1s3QfCSFEH9Z8BzP3bKJ+E1qUSWh6fMK9Jntq8mizz3NobOvuo2VPgrMBhp0Ge5eYvEoFzWYz+RkJCkKII0fzlkJobNtlwpt9uz/qMkgY2uxc/KHdR4U7TNfRxCvNFps7F5qAsHcpJI6UloIQQvRpbbUU2isTngDxQw49FxbX1H3kdMAnd0NQGJz6ANRXmuPZK2DfcphwsXfr7iMSFIQQR46WYwptCQgyK5AHTGudmiIsHioOmMffPQxZS+Dcf5scRzrJtDKWPQn1FTBohjU/g8V6O82FEEL4TrirdRAQDEEdZCw98xE48VdtvD7erHFoqIXvH4FxF8KkK8w5peCU+5tSZmTM9GrVfUVaCkKII4e7pRAa23GCuva6fsJceZEKt5vB5VFzDj0/+WrTfVS4HaJSvVNnH5OgIIQ4coTGmP0N2us66kzicGiohu2fm+fJow89rxSc+3TP6tjLJCgIIY4cSpkuoPYGmTuT5tpfee3rYAuE+KGty/hJiuz2yJiCEOLIEpF06FqErkgeY1ZAl+yFhOFNC94OI9JSEEIcWc550uy41h0BQZA6HnJWQfIo79arj5CWghDiyDJgas8+0Pu7upCSRndczk9JUBBCiK5wjyu0HGQ+TEhQEEKIrhh5Jky/HYae3Ns1sYSMKQghRFeERsPsh3q7FpaRloIQQohGEhSEEEI0kqAghBCikQQFIYQQjSwNCkqp2UqpbUqpnUqp37RxPkQp9bbr/AqlVIaV9RFCCNExy4KCUioAeBo4AxgDXKaUGtOi2A1AidZ6GPAY8Her6iOEEKJzVrYUpgE7tda7tdb1wFvAuS3KnAu87Hr8HnCKUn6eTUoIIfyYlUEhDchu9jzHdazNMlprO1AGtMpUpZS6WSmVqZTKLCgosKi6Qggh/GLxmtZ6HjAPQClVoJTK6ualEoFCr1XMd6TeviX19i1/rLc/1nmQJ4WsDAq5QHqz5wNcx9oqk6OUCgRigKKOLqq1TupuhZRSmVrrqd19fW+RevuW1Nu3/LHe/lhnT1nZfbQKGK6UGqyUCgbmAh+3KPMxcI3r8UXAN1prbWGdhBBCdMCyloLW2q6UugP4AggAXtBab1JKPQhkaq0/Bp4HXlVK7QSKMYFDCCFEL7F0TEFrvQBY0OLY/c0e1wLt7JBtiXk+fC9vknr7ltTbt/yx3v5YZ48o6a0RQgjhJmkuhBBCNJKgIIQQotFhGRT8NeeSB/U+QSm1RillV0pd1Bt1bIsH9f65UmqzUmq9UuprpZRH86Wt5kG9b1VKbVBKrVVKLWkjTUuv6KzezcpdqJTSSqlenzrpwe/6WtcapLWu2429Uc+WPPldK6Uucf3/3qSUesPXdfQ6rfVhdcPMdNoFDAGCgXXAmBZlbgP+43o8F3jbT+qdAUwAXgEu6u06d6HeJwHhrsc/9aPfd3Szx+cAn/tDvV3looDFwHJgal+vM3At8FRv/367Ue/hwI9AnOv5/7d3L6FxVXEcx7+/4kLQqlAjVCiosaE+qKmmUihihYJF8EU3SnUlghAsioJgXYgK1geRLooLFXQhiC+0UpQqtAqR0AjapyK2dlGpthRBo1EK+bu4x8lNTGdOTObeSfr7wCVz5x5mfnMzyX/OPck5F9Wde6bbfOwpzNU5l1rmjogjEbEXGKsj4Gnk5N4ZEX+m3SGKf2SsW07u30q75wCd8FcZOe9vgKcpJpj8q8pwp5GbudPk5L4f2BoRvwJExPGKM866+VgUZm3OpYrl5O5E0819H/BxWxPlycotqV/SIeB5YGNF2ZppmVvStcCSiNheZbAmct8j69MlxnclLZnieNVycvcAPZIGJQ1JWldZujaZj0XBOpSke4A+4IW6s+SKiK0R0Q08BjxRd55WJC0ABoBH6s4yTR8Bl0TEcuBTxnvyne4siktIa4C7gVckXVBrohmaj0VhOnMukTvnUgVycneirNyS1gKbgNsi4u+KsjUz3fP9FnBHWxPlaZV7IXA1sEvSEWAVsK3mweaW5zoiTpbeF68C11WUrZmc98hRYFtEnIqIH4HvKYrE3FX3oMZsbxSV+zBwKeODQ1dNatPPxIHmt+dC7lLb1+mcgeac872CYsBuad15p5l7aen2rRTTs3R87kntd1H/QHPOuV5cun0nMDQXzjWwDngj3b6Q4nLTorqzz+h11x2gTd/MWygq9iFgU7rvKYpPqQBnA+8APwC7gcvqzpyZeyXFJ5M/KHo2B+rOnJn7M+AX4Ju0bas7c2buLcCBlHlns1++nZR7Utvai0LmuX42nes96VwvqztzZm5RXK47COwD7qo780w3T3NhZmYN83FMwczM/icXBTMza3BRMDOzBhcFMzNrcFEwM7MGFwU7Y0laVJqV82dJP5X2v2zTc66Q9FqT412SPmnHc5vlaOtynGadLCJOAr0Akp4ERiLixTY/7ePAM00ynZB0TNLqiBhscxaz/3BPwWwKkkbS1zWSPpf0oaTDkjZL2iBpd1proTu165L0nqThtK2e4jEXAssjYk/av7HUM/k6HQf4ANhQ0Us1m8BFway1a4AHgCuAe4GeiLieYo6eB1ObLcBLEbESWJ+OTdYH7C/tPwr0R0QvcAMwmu7/Ku2bVc6Xj8xaG46IYwBpGu0d6f59FAsIAawFriwty3GepHMjYqT0OIuBE6X9QWBA0pvA+xFxNN1/HLh49l+GWWsuCmatlWd1HSvtjzH+M7QAWBURzRa1GaWYdwuAiNgsaTvF/DqDkm6OiO9Sm9HTPIZZW/nykdns2MH4pSQk9U7R5lvg8lKb7ojYFxHPAcPAsnSoh4mXmcwq46JgNjs2An1p5bCDFGMQE6RewPmlAeWHJO2XtBc4xfiKdDcBnbJqmp1hPEuqWYUkPQz8HhFTDUT/2+YL4PZI6/6aVck9BbNqvczEMYoJJHUBAy4IVhf3FMzMrME9BTMza3BRMDOzBhcFMzNrcFEwM7MGFwUzM2v4B+DzEkZbf0hPAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEWCAYAAABliCz2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xd4m9X1wPHvlWR57xXHjuPsnZABSQgjJOy9C2XTMtpCS+eP0r1LaWmh0BYKLVD2HgkbMiB7kL2d5cROvPe27u+PKyHZ8VBsTft8nieP1uv3PXbso6tzl9JaI4QQov+zBDsAIYQQgSEJXwghBghJ+EIIMUBIwhdCiAFCEr4QQgwQkvCFEGKAkIQvhBADhCR8ITqhlHpfKfXrTp6/RCl1RCllC0ZcQvSFJHwhOvc0cL1SSnV4/gbgOa11axBiEqJPlMy0FeJYSqlo4AhwkdZ6qfO5ZKAImKm13hjM+IToDWnhC9EJrXUD8DJwo8fTVwM7JNmLcCUJX4iuPQ1cqZSKcj6+0fmcEGFJSjpCdEMptQf4KbAG2AHkaK2PBjcqIXpHRhoI0b1nMC37McAHkuxFOJMWvhDdUErlAbuAYuC7WutXghqQEH0gCV+IHiilFgNTgEFa66YghyNEr0nCF0KIAUJG6QghxAAhCV8IIQYISfhCCDFASMIXQogBIqTG4aelpem8vLxghyGEEGFj3bp1pVrrdG+ODamEn5eXx9q1a4MdhhBChA2l1AFvj5WSjhBCDBCS8IUQYoCQhC+EEAOEJHwhhBggJOELIcQAIQlfCCEGCEn4QggxQEjCF8Jb+YugeHuwoxCi1yThC+Gtt74F7/4w2FEI0WshNdNWiJDlaIOaIqg5Ag0VEJ0c7IiEOG5+b+ErpaxKqS+UUgv8fS0h/KauBLQDdBvs+STY0QjRK4Eo6XwHkMKnCG81Re77O98LXhxC9IFfE75SKge4AHjCn9cRwu9qjprb1FGwb2lwYxGil/zdwv8b8CPA0dUBSqnblVJrlVJrS0pK/ByOEL3kauHnnQJ1xdDSGNx4hOgFvyV8pdSFQLHWel13x2mtH9daz9Baz0hP92pJZyECr+YIoGDwVOfjwqCGI0Rv+LOFPwe4WCm1H3gRmKeUetaP1xPCf2qKIDYdkoeax1WHgxuPEL3gt4Svtf6x1jpHa50HXAN8qrW+3l/XE8Kvao5AfCYk5JjH1ZLwRfiRiVdCeKP2CMRnQcJg81gSvghDAUn4WuvFWusLA3EtIfyi5gjEDwJ7jJl0JSUdEYakhS9ET9paobbYtPDBlHWkhS/CkCR8IXpSVwxoiMs0jxMGSwtfhCVJ+EL0pHyfuU3OM7eJ2dLCF2FJEr4QPSndZW7TRpnbhGxoKIfm+uDFJEQvSMIXoielu8EW7R6SmZBtbqtl8pUIL5LwheiBLt1FVWwev1ywnSW7SswELID6suAGJsRxkoQvRA8ainawtDyR/608wNefXsOmMuV8oSK4gQlxnCThC9EN3dJAVN0hiu1DWX3ffIalxfKzD50dto2VwQ1OiOMkCV+IbmzauB4LmtETp5EaF8mvL5nIgXq7eVFa+CLMSMIXohs7t60H4KQZMwGYOSyFnEGDANCS8EWYkYQvRDfaDm+iDQuRmWMAUEpx0ykjqNYxlJQcDXJ0QhwfSfhCdKGirpns+u2Ux440a+g4zR+XSaWOpbxUEr4IL5LwhejCqr2lTLHkQ/b0ds+nxNppjkigvqo0SJEJ0TuS8IXowu7tG0lU9SSPmn3Ma9bYFGiopKG5LQiRCdE7kvCF6ILjkNmd05Y745jXYpPSSKSWNfvLAx2WEL0mCV+ITmitSa/eQpMlGtLHHvN6cmomCaqOjQUyFl+ED0n4QnSitLaZ4Y79VMaPAov1mNcjYpNJVnUUVsoCaiJ8SMIXohP5JbUMV0U4Ukd3fkB0MjbaKKuQsfgifEjCF6ITBwqPkKEqicka0/kBUUkA1FbKSB0RPiThC9GJqoLtACRkH1u/B8y+tkBjdSla60CFJUSfSMIXohNtJWbTE5XWVUnHtPAjW6upamgJVFhC9IkkfCE6Ya/ahwMLpAzr/ABnCz+ROg5XNgQwMiF6TxK+EB00trSR0XyQ6qgssEV2fpCzhp+o6iisbAxgdEL0niR8ITooqmpkmDpCY3wXrXuAmBQAkqmhUFr4IkxIwheig8Pl9eSpI+jUkV0fZI9FR8SQaa2WhC/ChiR8ITooLj1CnGokKm1ot8ep2HRyImqlhi/ChiR8ITqoPXoAgPjMvO4PjMsg01pNcU2T/4MSwgck4QvRQVP5IQBsSTndHxibQYquoqaxNQBRCdF3kvCF6EBXOzcpTxjc/YFxGSQ6KqiWcfgiTNiCHYAQocZeV4QDC5a4Qd0fGJdBXFsVda0yLFOEB2nhC+GhzaGJayqmzp4K1h7aQ7HpKDQRTeW0OWR5BRH6JOEL4aG4ppFBlNIYndXzwXEZAKSrKmqlji/CgCR8ITwcrmggS5Wje6rfA8SahJ+mqqhulDq+CH2S8IXwcLiinixVTkRyDyN04MsWfhpVsoCaCAvSaSuEh5KSYmJUE7b07iddAe6ELy18ESakhS+Eh7rSgwDYvWnh2+NwWKNIU9VUN0gNX4Q+SfhCeGisPGLuxGX2fLBSOGLTSVeV0sIXYUESvhAemmvKzB3nevc9is0glWqZfCXCgiR8IZy01jjqys0DLxO+NTaVZFVDtQzLFGHAbwlfKRWllFqtlNqolNqqlPqVv64lhC9U1rcQ3VZjHji3MOyJikkm2VInLXwRFvw5SqcJmKe1rlVKRQCfK6Xe01qv9OM1hei1w5UNJKpa2qxRWCOivfui6BSSqJMavggLfmvha6PW+TDC+U/mn4uQdaiigSTqcEQmev9F0cnEUU9dvaynI0KfX2v4SimrUmoDUAx8pLVe1ckxtyul1iql1paUlPgzHCG6dbiygSRVi3JuX+gVZ62/rb7CT1EJ4Tt+Tfha6zat9QlADnCSUmpiJ8c8rrWeobWekZ6e7s9whOjW4YoGki31WGO9HKEDX+5tS0O5f4ISwocCMkpHa10JLALODcT1hOiNwsoG0qx1qOjjaeGbzl1bU6WfohLCd/w5SiddKZXkvB8NnAXs8Nf1hOgrU9KpgyjvRugAX5Z0JOGLcODPUTpZwNNKKSvmjeVlrfUCP15PiD45XNlAnKPW6yGZADg/DUS2VNPm0Fgtyk/RCdF3fkv4WutNwFR/nV8IX6pvbqWurhZ7VKP3s2zhy2OTVC31za3ER0X4KUIh+k5m2gqBqd8nUGceHE8LPzIBBxaSVC11TW3+CU4IH5GELwRmDH6iciX842jhWyy02BNJopa6ZlleQYQ2SfhC4OywxTlP8HgSPtAameRs4UvCF6FNEr4QmJJOiqUXLXzAEZlEErXUSsIXIU4SvhCYSVe5Mc3mwfEMywR0dLLU8EVYkIQvBKakMySqyTw4zha+ikkmiTrqpYYvQpwkfCEwLfwRthKwRkJkwnF9rSUmhSQlJR0R+mQTczHgtbQ5KKmuY6pjCYw+ByzH1w6yxaURqRqob2jwU4RC+Ia08MWAd6SqkdlqC7Et5TD56uP++og4M9u2rU6WVxChTRK+GPAOVzZwiXUZLfYEGHX2cX+9xblipq6XFTNFaJOELwa8wrIazrSsp3H4uWCLPP4TODt5VaMkfBHapIYvBjxHwRoSVT3N43u5ercz4Vsaq3wYlRC+Jy18MeClFC2lFQv2UfN7dwLXEsmNsuuVCG2S8MWAN7xyOTtt445v0TRPzhq+rVk6bUVok4QvBraGCvJa9pCfOLP354hMoA0Lka3VvotLCD+QhC8GNF2yE4CGlPG9P4lS1FvjiWqRhC9CmyR8MaDVHNoGgC1jTJ/O02BLJKZNOm1FaJOELwa0+sJtNOkIUrJH9ek8TbYE4hw1PopKCP+QhC8GNF2yi706i9z0+D6dp8WeSJyuQWvto8iE8D1J+GJAi67KJ18PJic5uk/naYlMJlnVUt8sSySL0CUJXwxMa56Apy4kobGQo/ZcIm3WPp3OEZlEInWyzaEIaZLwxcC0byns/wwLDmoTRvT5dDo6iXjVQF19ow+CE8I/JOGLgam2GJRp1TenTejz6VS0mXzVWF3a53MJ4S+S8MXAVFtM65gLmdv0F2IGj+vz6SyxJuE315T1+VxC+IskfDEw1ZVQbUthv84iNyWmz6ezxaYC0FonCV+ELq9Wy1RKZQBzgMFAA7AFWKu1dvgxNiH8o6UBmqop1WbtHF8k/IjETAAcNUf7fC4h/KXbhK+UOgO4F0gBvgCKgSjgUmCEUupV4C9aa5lTLsJHbTEA+5tisSgYlRnX51PaU4cCYKs+1OdzCeEvPbXwzwdu01of7PiCUsoGXAicBbzmh9iE8I+6EgB21EQxIj2OGHvft4WITkilWsdgry3o87mE8JeeftP/orU+0tkLWutW4E3fhySEnzlb+OvL7EwcleiTU8barezWaUTXFfrkfEL4Q0+dthuUUh8rpb6mlOrlYuFChJhaU2ffWRvNhMEJPjmlzWqhUKUT1ygJX4SunhJ+NvAAcAqwUyn1llLqGqVU3+ahCxFMzpJOGQlMzPZNCx+g2JJJQlMRyHo6IkR1m/C11m1a6w+01rcAQ4D/AJcA+5RSzwUiQCF8rraYRls8zUQw3kctfIAyWyZRjnpokK0ORWjyehy+1roZ2AZsB6qBvs9WESIY6oopJ4nhabEkREX47LQV9ixzp/KYMQ5ChIQeE75SaohS6odKqfXAAufXXKy1nub36ITwA0fNUQ61xHPyyFSfnrcmypnwq2SkjghNPY3DX46p47+MGZ65LiBRCeEvrU20lu2j2DGMU0am+fTUtVGDzR1p4YsQ1dOwzHuBz7Ts6iDCncMBpbtg9ePY64/yuuNGHhzu2xa+ik6mgUiiqw779LxC+Eq3CV9rvRRAKTUMuBvI8/warfXF/gwurLz5LcgYByffFexIRGe++B+8820AFkZfTGnaXJJi7D69RGyUjWriiG6s9Ol5hfAVb6cYvgk8CbwDyPo5ndm5EBrKJeGHqn1LIS6TzSf9ibvftfJ/s7J8fokYu40qYslskIQvQpO3Cb9Ra/2wXyMJZ83OoXhNsol1yCpYjWPILL67NpkhqQ5uOjnP55eIi7RR6YhBN1aifH52IfrO22GZDymlfqGUmq2Umub6190XOEf3LFJKbVNKbVVKfccH8YamaufsyiZZQy4k1RyBqoN8VDOUPcW1/OKi8URF9G1Lw87ERtqo0rFoaeGLEOVtC38ScAMwD3dJRzsfd6UV+L7Wer1SKh5Yp5T6SGu9rdfRhqpqZyedtPBDU8FqAP6Vn8rNJ+cxb2ymXy4TG2mlilh0Q6fLTwkRdN4m/KuA4c7JV17RWhcBRc77NUqp7Zghnv0w4bta+JLwQ1HrgZW0EUFT2kR+fP5Yv10n1m6jSsegGqv8dg0h+sLbks4WoNeLpyml8oCpwKreniOkSQs/pB3evYHdjmx+cskJRNp8X8pxcZV0LC210Nbqt+sI0VvetvCTgB1KqTVAk+tJb4ZlKqXiMOvl39PZRilKqduB2wFyc3O9DCfEuFr4rY3Q2gw23w73E71XXtdMZVkxEbGpzPHxRKuOXCUdABqrINa34/yF6CtvE/4venNypVQEJtk/p7V+vbNjtNaPA48DzJgxIzwneFV7LInbXAu2lODFItr55+I9XKtrSR/sv1KOS2ykjWrt3C6xsVISvgg5PS2toLSxpKdjOnseM3Z/u9b6wb6HGsKqPWZWNlVDjCT8UFBQXs/TKw7wncgG4pLT/X69uEibRwu/w0gd15+IkgGbInh6quEvUkrdrZRqV2tRStmVUvOUUk8DN3XxtXNwjuxRSm1w/jvfBzGHnupCiHGWCxplaGao+O3CbdgsEKtrIcr/+/fE2K1UaY+SjqePfg6PnSpr5Yug6qmkcy5wK/CCc3mFSiAa80bxIfA3rfUXnX2h1vpzGADzT1oaob4Uhp4CBz6XjtsQsf5gBR9sPcp9Zw5Bfd4K0f5P+HGRNqpdLXzPsfhFm2DFI6AdULwdMsf7PRYhOtPTBiiNWut/aK3nAEOB+cBUrfVQrfVtXSX7AcVVzslw1ogl4YeE19cfIirCwvVTnDtaBaSFb/No4Xsk/A9/ApHOjVZ2f+j3OIToyvFsgNKitS7SWss0Qk/F281t9gxzKwk/6FraHCzcVMSZ4zKJaXP+fwSghW+3WWiwxpsHrpJO1SGzjs/Jd0PmJNj9kd/jEKIrXid80YUSZ8LPOdHcyvIKQff57lIq6lu45IRs93aDAWjhA9ijYmlVNndJZ/sCczv+Uhh1FhxccWx9X4gAkYTfV8XbISkXEpyrL0oLP6ha2xw8+NEu0uIiOW10mru0EoAWPkBSrJ16S7z7utvfhozxkDYSRp0Nug3yFwUkFiE66jbhK6UeVUrNCVQwYal4B6SPg4gYUBZJ+EH2+Gd72Xy4il9fMsHMqnW1tAPUwk+KjqBWxZpWfMkuOLAcxjnnJ+acCFGJsEfKOiI4emrh7wL+rJTar5T6k1JqaiCCChttLWYXpYxxZnx1ZLwk/CA6VFHPQx/v5twJgzh/kvMTV6Bb+DERVOtYqC6CV242czKm32xetNpgxDxTx5fhmSIIehql85DWejZwOlAG/EcptcO5VPLogEQYysrywdFiPrKDGYkhCT9wVj3Wrjzyh3d3oBT87CKPYY8NleaTlz0+ICElRtup0DFQsBKKt8Jlj7nLfWDKOrVH4cimgMQjhCevavha6wNa6/u11lOBa4FLge1+jSwcHN1ibl1DMiPjpdM2UFqb4cOfwrK/AbDpUCULNxdx5+kjyE6Kdh/XWGnKKJbAdFclxUTwSuvpMPUG+NpHpqPW08gzAQXb3wlIPEJ48uqvQCllU0pdpJR6DngP2Alc7tfIwsHmVyE23dTwQUo6gXR0C7Q1w+H14HDw6KI9JETZ+Nopw9of11AZsPo9mBr+680n0XzBwzDkpGMPiMswZZ0Nz4OjLWBxCQE9d9qepZT6D3AIuA1YCIzQWl+jtX4rEAGGrMoC2P0BTLvRvTqmJPzAKVxvbpuq2bdzIx9sPcrNc4YRHxXR/rjGyoDV78G08AGqGlq6Pmj6TWbC3p5PAhSVEEZPLfwfA8uBcVrri7XWz2ut6wIQV+hb/7TpeHN1yMGxNfx37oF3fxjw0AaEw+vBYlYG+fCjBSRE2bils31qA9zCT4wxb/5VDd3sFTT6PPPJcMOzAYrKqbIA1j8T2GuKkNJTp+08rfUTWuuKQAUUNoo2waCJZgy+S2S8WVenuc6ssbPxRanV+svhdejhZ9BsiyOmeAPfO2s0ybGd7EMQ6BZ+tGnhV9Z308K32U3n7YHlgR2ts+E5ePtus8evGJBk4lVvNVRATIf1zkefY8Zf//d8s2ZKawPUFEFdWXBi7K+aatAlO3m3PJvVTUOZHbWf62cNPfY4rQNfw4/xIuED5MyAuhKoPBCAqJzqSs3t0a2Bu6YIKZLwe6uhAqKT2z839gL4ynNQtBHevsv9vGs0j/CN3R+i0Dx7JJvYnEmMsB7FZu3wq9zaBC9+1XziShsVsNCSos2njMruavjgXopj32ew8p/Q0uB+rbUJXr8DSvf4NriGcnMrCX/AkoTfW50lfICx58MJ15mWfuZE81yoJfzWZqgvD3YUvVa49L8c1qkMnnImJ4wfg2qqgeb69gft+QR2vgtzfwwn3RGw2BK/bOF3U8MH5+zsWDO09P172y+qdnQLbHoRtr3p2+DqnZ80JeEPWJLwe0PrrhM+wPyfm3LPiV+H2Aw44pHwD62FbccxwElreO9e2Ple32JuaYDH58LexfDJr+CRGaafIcy89OkaMo4uY13CWfz2simouEzzQm2HurRrUbtZ3zQzXAMkPtKGRfUwSgdMTNnT3DOBKw+6XyvLN7fFPp7qUu/RwpeZvgOSJPzeaKoxi2B1lfDjM+H7u2DGLaZj19XCryuF56+Gl280s0S9sXcRrPonLHuobzFXHoTCL8x5Nr1kWnv54TUs8NFFe9jxydPYlINzr7uHaLvV/KwBao62P7hkJyRkQ1RCQGO0WBSJ0RE91/ABcmeZWcBWO1QVuJ/3d8Iv2QH/OtWUksSAIgm/N1xL7naV8MHdqhw0yfyBNVSYIZpNNTDsdHjvR7D0ge5bWlrDkj+Z+wWr+laGqXUmxPxPTWchwFYflwx8rb4cDq5Ca8397+/ggQ92cm3iVnTaWOyDnJPd4gaZ29qOCX8HpPt/4/LOJMXYe67hA8y5B+74DFJHtW/hlzsTfukus16TrzSUm5+XowWOboZ1T/nu3CIsSMLvDW8Svsukq8wf7fPXwNbX4fQfwXWvwuSvwKe/hZeu73wUT+EX8M+Tzfrpk68x2+P1ZaJObbH7fmQCTLralIlCuazz5jfQT13A795Ywz8X5/O1GcmMatyEGnue+5j4ThK+w2FWqgxSwjct/B5q+ACRcc6hvUPMGHmXsj2AMom5LN835ZeWBmiphzHnQXQKjJhv3hTL9/X93CJsBK646S+tTfD+jyHvFJjoXu2hrqmVdzcXseVwFRFWC6My4zhvUhYJHWdi9sbxJPxBk2DKNbDxBXN/zj1gjTCLamVOMEn/lZvg5gXur3G0wVt3mRbuxX+HKV+FPR/Drvdh8lVdXqq51cHOIzWU1jYxZlA8gz3XlHElxBHzIGsKDJ0Dm1+mKX8pf9iVze7iGsYNSmBERhyFlQ3kpsRw/qQsYiOD9CtyYAXseh8FbF27mDtOu4h7h2xFbWk1E5dcolPMBCzPseWVB8yQ2PQxAQ8bzNDMslovEr5L4hDz/YJJ7mV7YchMswDbm3eaN/ub3zVvEL3l+nQ4+AS48K9QsQ8enmqGD8+8A8r3QlJewNYcEsER/gnfajejMRqr0BMuY83+Cl5eW8C7m4uob24jPtJGi8NBY4uDRxfl8825I4iwWoiwWThlZBqFlQ3UN7cxLiuelXvLSYuzM3ZQAvkltSzcXMQtJ+eRkRDlvl514fElfIB5PzOjdub91CR7MMspz/kOWCPh/f+DvUtg2Gnm+S+eNXX/K//rfhMbOb/LjTO01ryzqYj739vB4Ur38L6bZg/lrPGDaGxpY1ZZIXFWO1z/Oq0OTW1NJYnKyutvvMxTVZcxLiuB/608QFOrA6VM3vndu9u5Z/4orpoxpNvEr7Umv6SWBZuKSI21c8HkwSTHRFBR30JyTARKddjLftnDprNy/s/bPd1WvIuG93/OwcjRDMp/hTadSKqq5tcnVDPy7OGoZ39oOsNzZri/yGIxHeOeLfySHeY2Y1z3/y9+khkfxdbC41hELykXmqrM70hbi7k/5jw4tMZ80gPT0X7+A8cfTEsDPH4GjDnXPI5OMb9jKcMhbbT52xkxHx49EWZ/C87+7fFfQ4SN8E/4SlEQO5Hk/OXc8tgK1uyvIC7SxsVTBnPVjBym5SajNazeX853X9rAva9v9vzSLz8te963KNCYx6+sPcTj6S8TE2GlPO98Tl56PQ1jLiUa2FcXga28nuykaCwWRVNrGwXl9TQ0O8gvqSU1zs74rARSE7Ph2hc6j3/6zbD8YXjhGvPmddoPTN1+yCyYcJn7uMHTTGdrzRF3GQP44mAFv1mwjfUHKxmflcCPzh1DVmI0CzcV8vSKAzy9wkzs+XPEJuZGJPKrFzewam8ZxTVNvGnPY0zTZp665ffMHZNBa5uDwspGMhIi2Xy4igc/3MUv39nG/e/v5IQhSWQkRKKAxhYHu4traGg2i3/VNLUyqmkbG/UI2rDys7e2Emu3UtfcRm5KDGeNz2RabjLzx2UQZYXWz/6Kpamaz5Ov4M09LWw5XEVtYyu/aLyfc9QqxvMee/VgFgz9Dd9qeIxRDZvglVtg/2dw0UNgsbb/GcZntm/hF28zt2nBWcF7SEo0JTVNNDS3mY7lniQNMbeVBdBca+5njDefxOyxpjS1+nGY9Q2TqI/HzvfMiCXXKq6ekwXHXwJL/wxL/mg+RSx/xHx6ypM9j/qr8E/4wPOFmfyf5SOqmw/x20tncvm0bGLs7m9NKZg1PJVFP5hLcXUTAJUNzXy49ShZSVHERdrYVljN6aPTqWlqZVthNTaL4uSRqfxhwWZGF75FFM2s3LcRrNC442OiFZz72Gaa2EmM3cqYQfHsLanrdDjeyIw4/nLVFCbnJFJS04TFokiLizQvRkTBBX8xrfqqQ/DhT3FEJVN7/j8oLa0jJzmGwsoGHBEjGA5QuAHGnEtRVQP3v7eDNzcUkhYXyZ+umMwV03OwWkxr+qRhKVwyNZuG5jaiIiykvfl3yqqT2HSoksk5SUwfmkzjjlnMLHkFteU+2BWL7aKHyE2NAeDEvBSev20m6w9W8NaGQjYeqmJDgRlCaLMoRqTHkehcRmBE807u3P1L6qZ/k33T7mXZnlIOVzaQlRjN53tKKFv1Ihmr3+Nqyzc4PcfK9xtNeeGz1x7h08jLmJ6bzDjrYc7JX8X2UbdjO/FW8vKG8217JCzcBGueMD+rc//Yfu0il7hB7Ue57PsM0sYEdEkFT0NSzM+woKKe0ZnudfibWtt4dFE+724uYmR6HHfNG8nE7ERIdC7Psekl2P+5uZ82Em5515SrClbBmn+bjt3jTfibXzG31YfNbUyK+7UZt8JnD5pjhp5iSmGf/VkSfj/WLxL+bdddAy88y4LLIomY2MkUe6eoCOuXCS2XGCbnuBPCJSdkf3n/nAnuFvSrF0XCf0zH5ilWM2ElWdXSZo3iz9fOpK6plR1HatheVM3cMenMHZNOdISVYWlxlNU2sbWwmqeW7+fSfywj0mahscWBRcHYQQkU1zQxfWgS88dNJHHSX1i6eR8jC/7KwqqZrH1oB7DjyzhiaGRLpOLzpR9zoGIcf3p/B02tDu6bHc2ttY9hG/uQ+WjiYVquR8nJXgsjRrLkq2e4n8u6CF54zvzBWyPhnD+APebLl1VTNdNL3mT6RTd3X9td/B7shth1/2Ri+lAmTrkAkkYA8I25I3A8/zcsu3bxkuVnrCkYSauy0pw8iu+2reSH3/ordt0Ez/4YImIZd+m9EOvRCs2dbRKieUvOAAAgAElEQVT++Eth5p2dXz8+Ew6vNfeb6+DAMjjp9q7j9bNcV8Ivb5/w/7fiAA9/sptZw1NYta+MTc9UsuRHZxDhauEvf9jU8y94sH1id7XKj3eUVl2ZqdFb7WYpac9zASQMNq38ra+bIcSH1sC6p83EPFsn6xKJsNcvEn7KiJPAaieicC1MvMS3J9+3FFBmbfOCVaY1WXsEa0wKF00Z3MMXx3PyyDSumpHDU8v3U9fUyuCkaMpqm1l3oILRmXF8vqeUD7aa+nOEVXHDrF8yPz6SeWjS4iIpKK9nSHIMSTERFL81hMaD6/lZ/hYm5yTy92unMvSzH8Lu92HxH0y5oyt1xZDdYYfK3FmgrBCXCTWFpmQy+hz362v/Ax//0tSYR57Z9bnzP4XMSYA2/RGf/AquehpGnw1aYzm8DoafQXRVAaeVbYC8U7HNuBVevQX+e5ap51cXwZVPtk/2AGPON30fJ95mPqp1Jm6QmePQ1mpa923N3cfrZ66Ef7DcPftXa81zqw4yY2gyL94+m0U7irnlqTW8vaGQK6Zlmx254jLg1g8gLr39CaOdrfKG40z4294ARyvMvgtWPOI8V4d+p7n3QkS0WRYkIhpW/cu8eQ49+fiuJcJCv0j42CJNvXPtf8yQtqufcXeO9tXeJZA1GS562Kx/X3UYVj/mfYctZlz2PWd2Xk9uc2j2l9VR3dDC6Mz47kfF7JhJRv5i1g59n4Qz7sFuLTdlgKhEWP8/mH23KQV05GgzY+9ds1JdopPN6KCU4WbExra3Tcfh+EtNC881DHTLG10n0IZKM3v41O+b5FGy04wseeEauGOpKavUFcOYH8L4i+HNb5jkPfZ88/Uf/gxSR5if74gzjj2/PQZO62GJ6bgMQJuO2z0fmSULgpiwUmLtxNit7RL+ivwy9pXW8e355v9n7ph0xg6K57Gl+Vw2NRvLjW+ZN9aOyR7cv2vetvBfuQXis+DwOrOEw+SrTcKPTDj27yJ9DFz6D3N/6MmAMo2crn5+O9+D1JEBXZ+ov2tzaIqqGshJjun54D7qP2Ow5v0MBk81ow48J7H0RUsjHFptRs9kjDWjalzbGR5Hwu+O1VkPn5qb3PMQyKwpWOpLSNv+DPblf4HlfzfP3/CmedNb2sUojvoy0ynXMeGD+cOOH2Qmg214Fl6/zazj0lQLB1eaTwA73jEf8zvS2gw31W1muKfFCpnjTTyRcaalf3idOTZ7urnODW+4k/3Ey+F7W+GmtztP9t4afIK5XfoAbHrZjGiyRfb+fH2klCI3JYYCj4T/8toCEqMjOG9i1pfH3HH6cHYdrWXRzmLImd55sgfz5muP9y7htzabJblXPmqGdU6+ynT6Kmv7+n1nopNNw2nfZ52/3lgNL91gJg0KnyiuaeT6J1bxlcdWUtfU6vfr9Z+EP/x0M8IFzJLEvlC8zZQHsj2GAbq2MwxGh+DUG+Dc+2HiFbDldbMJy5RrzJos026CLa92/mbnmnQVl9H1uafdaDo6IxPg4CpT3nG0mJEhjVWmbNPRJ78yC39lT28/VDImBU75nqkfL3sYLBFmgpG/ZE83P5N1/zXDGs/6lf+u5aUhKTEUlJshso0tbXy8vZhzJwwiKsI9aufCyYPJTormX0vyez5hTLJ3JZ3ibeb/TTn/tCddZd780ka5S0PdGXGGeaOoLTafbj3fZPI/Mefeu1jW1PeBPcW1XPLIMr4oqOA7Z44KyJyX/pPwwXyMBd/9Mh5xDuHMmux+zjWZx0ct/OMSnQSz7jTjpVsbzKSzOfeY1052Lsf80vXwyW/a75fqGqMe203CH3s+3LXaTMgqWGmSdUSMWW0yKsl07HlytMHa/5oa+60fHFsqmHmHeaM8vNYke3+3uM/6jRntcs7vjn8kix8MSY7hYHk9Wms+311KbVMr500a1O6YCKuFr586jDX7K1h3oIdkHpPqXQu/aKO5vfRfcN4D7g165v/c3SDqzpSvmrr/gu/CMxebCYAuO98z5TLtMB39sgBbr+0tqeUrj62gpc3BJ2cc5OrCHpZZ8ZF+lvCdf1C9aeE7HPD539ovwnVks/konZTnfi4mxSS5vFP7FGqfDJ5mZmJOudZdS03MMbXuploztM5zYayqQ+Y2vpOSTke5M00/yMaXTEdeZByMuwh2vNt+GYZDa0xn66SrOu8viYiGW983n0jm/az336u3ErPhnk1w4tf8fy0v5KXF0NDSRn5JLe9uLiIxOoI5I9OOOe4rJw4hOSaCfy7e2/0Jo1Pcyxu3tXa9d3LRBohMNHX7mR4jlcZeYP71JH005J4MOxYACnYuNKW9tlbY9YHphxk8zSzr/KfhYb3MdjDUN7fy6Y6j3PzfNQC8dMdssstWmE/UXQ1K8KH+lfAjE0yrtDct/MIv4ONfwEcesz+PbDLLIXQcknjtCzDlK32LtS+UMq1qV2eby9x74e51ZvLMp79112ILVpuE4fnG1ZUhs8xtS517HfmJl0NzjekQddn1gakLj5jX9bmsEeYTycj5Xn9rfRKAPxhvnTcxi0ibhV++vY0Fm4o4f9IgIjpu0gLE2G3cODuPj7cfZffRLpI4mIZGQzlsfhUeHAf355mk27FvpWij+UTal5/FjFvM7YV/NSOgFv/RjFBrrDQzgC/8q2lsNJT7fkXPfqyxpY2rH1vBrU+tpbyumSdvPpER6XGmgZUyIiAx9K+Er5Rp5femhV+43txufhlKd5sW/5EtJuGHIqU6/6NWCi580LTmn74QVv/bjEsferJ366QMnmrGbQ+e6q7L550GMWmw8Aew6nHz3O4PzRj5IE1uCnXp8ZFce1Iun+8pJTbSyvfO6npdn5tOziPGbuW+NzbT2NLW+UHRKVBfAUvuNz/ziVeaTvvVHstst7WY39msKX0LftJVcNsiM8lt6vWwb4mZGGi1mzf4wSe4R05VyOJr3tBa86t3trHlcDV/vmoKK++bzwlDktxrJ6VKwu+duEHHro3ujcINZnijLQr+c45Z0KylLnQTfncSBsM3VpjZk5/+xvxRDvVy9mREFFz6TzOm3/WGYrXBV5419eD3fmg67Y5uca/PIjp1+2nDGZEey+8vm0R6fNd9GCmxdv505WTW7K/g+69sxOHopJYbk2LW2CnLh3EXw+WPwfC5Zn8D125fOxZAW1Pfh6QqZQYCKGXKedoBG583ZcxI50SypFzzCU9W2+yRK9m/sPog35g7giun5xDn6qCtPWryjLTwe6nXLfwvTF38uldh+BntJ1yFI3sMnPpdM8IGji8JTLry2Fbi0Nlwxb/N/Te+ASiYcPkxXyrcBidF88n353LepKwej71w8mDuO38sCzcV8cf3dxx7gGuGrG5zLwp3+r1mfsX6Z8wn0sX3m5FWo334Rpw1xb30wxiPVUqtEabfSFr4PXps6V6eWr6fr58yjB+d0+GTnmuzm9TADDToHxOvPMVnmfqy1t7XMZvrzQJT4y4064jkzTGdVHXFprUcrkbMNwuIVRf55pNKch7knGTmJuSdajpKhc/cdupwDlc08PjSvQxOjOLmOcPcL3qOCnMl/KGzzf/r9rfN/0XJdrjiyWMXl+sLpUxH7YpHjn0jSRnmbuEfXm9KTK2Npm8pGKPYQtCa/eU88MFOLpiUxU8uGHfsyrGuzW4C1MLvhwl/kPmI1FTj/fZ2Rzabj62DPZYesNrCO9mD+WO99F/mE4+vksCkq0zCn3Slb84nvqSU4ucXTaCoqpFfLdhGeV0zU3OTsVkVp0SnoACUlbbkEdTWt5gN00fMN8nYFmkm1nmusOorp//ItO5da/64pAyHrW+Y+6sfdw/dPbDCPbluACurbeKu59eTkxzNH66YdGyyB9PCt0SYNZQCoB8mfI+x+N4m/O1vA8oMN+tvcqb79nxTrzNvqJODOEqpH7NaFA9dM5X/e20TD3+658vnLx1Uwt+A0sgcvvroKnYdrWVKTiLDalP5m6PVTIw76Xbftu5dohLNBkMdJQ8ze0M0VJo+ncHTzOCHci8mkvVzWmt+8MpGKupbeP0bJ3a98VJ5vvnkbA1MKu6HNXzXptZF5o/go5+79wU9tM49mcql8qAZyTLlWu/GqQ909lg45btmnL3wi2i7lYevncrCb5/CS7fP4veXTaIa03hZW59JY4uDO04fjs1qoXHQdOq06RBeGXN6YANNcZacyvaYNZTyTjGlnDJJ+K+uO8SinSX8+LyxZgnsrpTvC9gIHejXLfwiM3Rwy2tQcQDO+jX871JT077NY2/YT507/Mz7SeBjFaIbEwabRDFzeCpfnZoGf4zgrLlncc7pc9uVB+qfOYOj+9dz7fuan1j3Mmt4KoOTokmJ9fMSx8nOhL/7Q7MESeYEU4se4C38QxX1/GbBNk7KS+Gm2XndH1xTFNCBIf0v4Sc4OxKrCszKlpEJsO1N05Hb2mBa+G0tZpRB4Qaz2uSce8yIAyFClT0GvvYB1rQxxwxGiLnyn1gbajj3vVJ+u9BMhIq1W7l8Wg4xHjtupcTaSYuLZG9pLa1tmkk5iZw5LrPd+j7HJXWkmdW72jl6K3OCaa3uX9a78/UDza0O7nr+C7SGB66ajCX/E7OxTGczwB0OUxLzZo0jH/FbwldK/Qe4ECjWWvtx5awO7DFmklBlgVlSYOwFpsPp7bvN2OIDy8xm4B/cZ/b7jE6BU78XsPCE6LXsLvpjYlKIjEnh79cO4a0NhdisioWbinh13SE0Zky/1tDU6gBMP4FVKZrbHERYFaMz45k4OJFvnjGCoamx3scTEWVW41zzhBmTnzbatPA3vWRGvtn9v9xvKNFac98bm9lQUMk/rpvG0DW/M6uWgpk70XE11KYqM1ikp1VMfcifLfyngEeAZ/x4jc4lDYGK/ebjUkK22dVn7EVQvhcemW7q+pUFkHMinPh10yklRJizWS1cMd18UvXcwc2loq6ZsromhqbGYlGK5fmlLNtTxraiahZsKuSTHcX84OzR5KbGMHt4auejSjqadqNJ+GmjzUghVz26Yp9p8Q8gT36+j1fXHeI780dx/qgYePPfZtjskc2mM9tzcyFwr0PUH1r4WuulSqk8f52/W4lDIH+RmaTiKtVYLGYYmT3eJP7c2WZxLyEGiORYO8kedf1TR6Vz6ijT6txTXMP1T6zm3tfNoIaZw1I4Y2wG47MSmDY02T0ztKOsKWZOhmueh2ul0rL8AZXw65paeWTRHk4fnc49Z46CjS+afo2zfwv/u8xsEtQx4TdUmNt+0sL3ilLqduB2gNzcXN+cNCnXLPYF7WvzFov5BT3wuWn1CyEAGJkRz+IfzuVodSOLd5bw78/28sf3zIzfjPhIHr52KrOGp3b+xTe94+5XcLXwS3cFIOrQ8dyqA1TWt3DPmaPMJ6Otb5iG57DTIWO8eyMgT67VT/tDC99bWuvHgccBZsyY4ZsFoZM83jg6dsZmTzUJf9xFPrmUEP1FVISVoamx3HRyLDednEdlfTNfFFTy63e2cc3jK5kwOIFU5z7LuSkx/OO6aWbTDs/ST1Si2SRo31Lv1t/vB4qqGvjn4nzmjExlam6yabnnf2pWinWtS7Tt7WNn/7tKOgFs4fe/cfjQftZaQoda5snfgetfk1E5QvQgKcbOGWMyeOfuU/jpBeOIjbRRVd/MiPRYPttdwp3PrmN/ad2xXzjqTDiwvOs1+/uR1jYHdz//Bc2tDn59iXNsyvYFZmcw16zn7BlmaenyDnseNEjC9w3XFPDIxGNn28ald70htxDiGHGRNr5+6nBevmM2b911Ck/cdCK/v2wSy/aUMvfPi3nisw6JbNTZJuHtWxqcgAPozx/uYu2BCn5/+SSztj2Yck5ynnvmfs6J5vbgyvZfXF9utqKMDNygEb8lfKXUC8AKYIxS6pBSKnBbEbla+LK4lxB+cc1JuSy7dx7nTMjktwu3887GQveLQ2aZwRG7PghegAGwfE8p/1qSz1dn5rpHRVXsN8uHT7jMXb5JH2tWO93fYXP4hnIzM9mbfSp8xG9X0lpfq7XO0lpHaK1ztNZP+utax4hOMhOupGwjhN9kJUbz0DVTOTEvme+/vJFVe52dkDa7WXl24wvuPXZ97ehW+PhXsOeToO2t+/dP95CVGMXPLxxvnlj9b3hoikn0k65yH2ixmGUn9n/ePtb68oB22EJ/LekAzLgVJl4R7CiE6NeiIqz8+8YZ5KRE8/Vn1ro3Yz/7d6ZV+9rXzYb3vrbyH/D5g/Ds5WZ/5QDbVljNir1l3HRynnum8tr/mlGAd35+7JDUvFPN7P+K/e7nGsoDWr+H/pzwz/oVTLkm2FEI0e8lxdj539dmkhpr5/onVrNgUyHEpsK8n5rhmSWdbOjSV0Wb3AMyKg/6/vw9eHxpPtERVq6dkWO2fyzdDcVbTSnHtV+Bp7xTza1nWac+sMsqQH9O+EKIgMlOiublO2czNiueu57/ghueXMWattHmxUNrzX7IW173zcVam82byKizzOPe7HDXB18crODNDYXcdHIeiYeXwFvfgv85d38bMa/zL0ofA7EZpqzjUl8mLXwhRHjKiI/ipdtn84OzR5NfXMtVrxZTZ4lnx8f/hTX/hjU+6sYr3WVmseadCrZos/dFgGit+eU727g0dgv3pKwwe/0CVB00a3hldrGznFKmjr/vM3cd39VpG0CS8IUQPmO3Wbhr3ig+/cFcbpiVxwbHcMY2fAFA68HV1BdsgpdugKUPuDdfP16uPS0GTe79Hta9tGhnMRsLKvlx0sdEvXePmVA16WrTXzHqrO5H3OSdAjWFZjx+c73ZDlJa+EKIcBcVYeU3l05kzulmH9xWZcOmmyl68hoc2xeafSiW/sn9BUe3wb/nwfZ3ej75kU2mZZ86wux/EaAWvtaaRz7dQ3ZSNBna2Tmt22DWN+DOZXD+A92fYNhp5nbfUrOSL0gNXwjRj2TPAMA26040ihEc5oXWuWyOnY3e+KJpIT8+1yT7w+vMLNWeFG0yo2As1oC28JfuLmX9wUruOH04qvYITL4GrnvNLJ2QkAWR8d2fIHUkxA2CTS/DG7eDPc79JhAgkvCFEP4z7DQ4+dsw5x6Uc6hi1OzbeKRiJqqmCF65CZpqYcpXYPBUKNne/fmaauHQasidZR67Wvh+Hovf0ubg1+9sZWhqDF+ZnATNtZA53iwj4S2lzFDxg8vN5ktXPBHQ7Q0hBBZPE0L0YxFRcPZvzP1pN0Hhei4//1xW1aZTtu1J4uwR2G58G2viYHj/Plj7H7MTVFe18L2LTYeta6nh+EHQUg9N1X7d1+LZlQfIL6njyZtmEFlf7Lx21vGf6Nzfmw2X2lrMp4IAk4QvhAiMmbcDoIBfXnYCPzj0O7aWtmJ/cjcv3J5KWvoYsw1p5QH3Bukd7f7AzKLPnW0ef7mH9VG/Jfym1jb+tSSfmcNSmDc2A/Y55xXED+rdCWPTfBfccZKSjhAi4GLsNv76nev53lfOpaCinq89vZbGZOe4/a4mamkNuz8yY92tEeY5V9L1Yx3/1XWHOFrdxN3znGvduzqJ4wf77Zr+IglfCBEUkTYrl5yQzcPXTGXzoUp+uKTJvFDcRR2/ZKdJ7J6r3X7ZwvfPSJ3WNgf/WpLPlJxE5ox0bgBT41worrct/CCShC+ECKqzJwzilxdP4J2ddZRb02k+sq3zAwtWmVtXOQcgPtPc1hQee7wPvL2xkILyBr51xkj3Hr81R0xZKTLOL9f0J6nhCyGC7sbZebS2aTZ8OIQxWxfzhH0z3z93XPu9dAtWmdmsniNbIuMhIgZqS3weU5tD84/F+YzJjOfMcZnuF2qKwrJ1D9LCF0KEiFtPGcbI+beSTQn5qxZy8SOfs+NItfuAgythyMz22wSCeRNw7Q/rQy+vLWBPcS13zx+JxeJxzWpJ+EII0We5c66G6BT+nruYi+pe5/pHP+K9zUVQVwrl+TDkpGO/KDYV6kvbP9dQAS2NPV/Q4eh0DH9lfTMPfLCTE/OSuWBSh+GTNUfCssMWJOELIUKJLRKmXkfikRV81/E0P41/l2+/+AUblzt3z3JNuPIUk2reEDw9eQ588quer7fwu/DgONj98ZdPldc1c/2Tq6hpbOGXF09w1+7BrNRZUwQJkvCFEKLv5v4Yrn8Nxl/CJS3vMTm5jRVL3qUZGwtKM489PibN7B7l0tIIpTu92xjlyBaTwF+8FhoqOVrdyFceW8Huo7U8fsMMJgzuMLa/ZLvZr3fQxL59j0EinbZCiNBijzVDLxNyUNve5rkTN1C3o5C9NaO46+VtrDxYy08vGO/eaSqmQ0mnqsDcluw05ZqONX9P9aVmIbbWBuqO5nPtazUcrWrkqVtOYvYI5zDMygKw2s2IINeWjVkn+P77DgBp4QshQlPGWBg5n6gtL5BatZVRM+Zzx2nDeXblQa7813IOljmXV45NNcsruJZbrjxgbpuqobqH4Zp1ZWYNH+C5D5exv7SOJ2460Z3stYb/XQYv32geF200QzKTu5gJHOIk4QshQtfka6D6MLQ1Yc2dxY/PH8e/b5zBwbJ6zn1oKY8u2kNrlDM5u0bqVBxwf3132yu2NkFzDc2ZUwAoOrCL7589xp3sAYo2QNluKFhp9qMt3GDW4e9u3fsQFp5RCyEGhrEXmGWEwQzJBM4an8nCb5/KKSPTeOCDnfz5c+cYfFdZp/IAKGdqK9lhavqPnAi/ToUPfwpAY0sb76zYBMAvljdTryM5PbORb87tsHrlltfA4qx8b3oZjm4xG5WHKanhCyFClz0GTvgqFKyGuIwvnx6SEsPjN85gwaZCXnptDyhYvnkXMzKmYK88CMl56MZqivZsYE3VYi4p3QXAxuUfcMuq06lramVk214uioQ5k8fSuD+L0zKajh2Rs+UN05/QUAHLHja7VA0Oz/o9SMIXQoS6c+8H7ej0pQsnD2ZK1OnwPLy8ZAM3L03gtYhNWKOTqW+MRu3+gg+3D+ISO+y3DmWItYYLxmdht1m4PLERPoELZ02C5hFQXeA+scNhNievPgQX/hXammDJ/ZA0FEbMD9A37nuS8IUQoc1iobvq85CcXAC+OTOJzIg8hq0r5cPG4URFRXNOy8f8aU4krIC8qfNh44v85pIJZuTOpi3mBLFpkDTEPQIHzBj+zS/DvJ/B6LPNc+Mu8tM3GDiS8IUQ4S0qCZSV0XFN/PiUIbC6isvnzzazYd98m9j9H0JCDiTnOUfz1Jo1eFyTtWJSITHH9AE018O2N2HZ32DGrXDq94P6rfmadNoKIcKbxQIxKWaUjmtz8KShZq9ZMC331BEQ55y0Vevcsaq+FJTVvGEkmk8JVB2C9c9A5kQ4/8/dj+EPQ5LwhRDhLybNtNjrnCN2YtMhdRTYnRuLp440zwHUHjW3daWmdW+xmJIOOGforjUdtRZrYL+HAJCEL4QIfzGppoX/ZcJPM4ncNaImdaS7hV+xHxb9wdy6thtMH2tm0y76vVk6Ydipgf4OAkISvhAi/MWlm5a7a/KVqzXvKut4Jvz1/4Mlf4R9S8wbBZiS0KSroXibGXc/pJNF2voBSfhCiPAXN8hsZO7qiI1OMbcjz4KIWMiabJK6sppZsy6uNwaAWXea2+zpYbmblTdklI4QIvzFZ0JLnZllG50MVmdqG3Yq3HfY3fkam2Y+CdiizCQqV0kHYNAkOOV7X66t0x9JwhdChL845w5UR7aYDlxPniNt4jJMwp98tVk4bfgZ7Y898xf+jTPIJOELIcKfazPzkh2QM6Pr4+Iygc2mbDP95kBEFlKkhi+ECH/xzm0IHS3tyzQduTpu+3HZpjuS8IUQ4S/OYyesjiUdTynDIDIR0sf5P6YQJCUdIUT4i04Ga6RZ5Ky7Fv7su2DKtWCzBy62ECItfCFE+FPK3crvroUfEW3WzRmgJOELIfoHV8dtdy38Ac6vCV8pda5SaqdSao9S6l5/XksIMcDFScLvid8SvlLKCjwKnAeMB65VSo331/WEEANcvHMsfnclnQHOny38k4A9Wuu9Wutm4EXgEj9eTwgxkLkmX0kLv0v+HKWTDXjsGcYhYGbHg5RStwO3A+Tm5voxHCFEvzbpCrMVoucQTdFO0DtttdaPa61naK1npKen9/wFQgjRmZThMPf/+t2mJb7kz4R/GBji8TjH+ZwQQogg8GfCXwOMUkoNU0rZgWuAt/14PSGEEN3wWw1fa92qlLoL+ACwAv/RWm/11/WEEEJ0z69LK2it3wXe9ec1hBBCeCfonbZCCCECQxK+EEIMEJLwhRBigJCEL4QQA4TSWgc7hi8ppUqAA7388jSg1IfhBEo4xh2OMYPEHWgSd2AM1Vp7NWs1pBJ+Xyil1mqtu9nMMjSFY9zhGDNI3IEmcYceKekIIcQAIQlfCCEGiP6U8B8PdgC9FI5xh2PMIHEHmsQdYvpNDV8IIUT3+lMLXwghRDck4QshxAARdgm/p43RlVKRSqmXnK+vUkrlBT7KY2LqKebTlFLrlVKtSqkrgxFjZ7yI+3tKqW1KqU1KqU+UUkODEWdHXsR9p1Jqs1Jqg1Lq81DZa7mnuD2Ou0IppZVSITF00Iuf981KqRLnz3uDUurrwYizI29+3kqpq52/41uVUs8HOkaf01qHzT/MMsv5wHDADmwExnc45pvAv5z3rwFeCoOY84DJwDPAlcH+OR9H3GcAMc773wj2z/o44k7wuH8x8H44xO08Lh5YCqwEZoRD3MDNwCPBjrUXcY8CvgCSnY8zgh13X/+FWwvfm43RLwGedt5/FZivVFD3POsxZq31fq31JsARjAC74E3ci7TW9c6HKzG7mgWbN3FXezyMBUJh5II3v9sAvwHuBxoDGVw3vI071HgT923Ao1rrCgCtdXGAY/S5cEv4nW2Mnt3VMVrrVqAKSA1IdJ3zJuZQdLxxfw14z68ReceruJVS31JK5QN/Ar4doNi602PcSqlpwBCt9cJABtYDb39PrnCW/l5VSky8Sf8AAAOxSURBVA3p5PVA8ybu0cBopdQypdRKpdS5AYvOT8It4YsQpJS6HpgBPBDsWLyltX5Uaz0C+D/gp8GOpydKKQvwIPD9YMfSC+8AeVrrycBHuD+BhzobpqwzF7gW+LdSKimoEfVRuCV8bzZG//IYpZQNSATKAhJd58J1M3ev4lZKnQn8BLhYa90UoNi6c7w/7xeBS/0akXd6ijsemAgsVkrtB2YBb4dAx22PP2+tdZnH78YTwPQAxdYdb35PDgFva61btNb7gF2YN4DwFexOhOPsaLEBe4FhuDtaJnQ45lu077R9OdRj9jj2KUKn09abn/VUTMfXqGDHe5xxj/K4fxGwNhzi7nD8YkKj09abn3eWx/3LgJVhEve5wNPO+2mYElBqsGPv0/cd7AB68R91PuadNh/4ifO5X2NamABRwCvAHmA1MDwMYj4R05qow3wa2RrsmL2M+2PgKLDB+e/tYMfsZdwPAVudMS/qLrGGUtwdjg2JhO/lz/sPzp/3RufPe2ywY/YyboUpo20DNgPXBDvmvv6TpRWEEGKACLcavhBCiF6ShC+EEAOEJHwhhBggJOELIcQAIQlfCCEGCEn4ol9SSqV6rM54RCl12OPxcj9dc6pS6sluXk9XSr3vj2sL4Q1bsAMQwh+01mXACQBKqV8CtVrrP/v5svcBv+0mphKlVJFSao7WepmfYxHiGNLCFwOOUqrWeTtXKbVEKfWWUmqvUuqPSqnrlFKrnevlj3Ael66Uek0ptcb5b04n54wHJmutNzofn+7xieIL5+sAbwLXBehbFaIdSfhioJsC3AmMA24ARmutT8Ks+XK385iHgL9qrU8ErnC+1tEMYIvH4x8A39JanwCcCjQ4n1/rfCxEwElJRwx0a7TWRQDO5ZI/dD6/GbPBC8CZwHiPbRUSlFJxWutaj/NkASUej5cBDyqlngNe11ofcj5fDAz2/bchRM8k4YuBznOFT4fHYwfuvw8LMEtr3d2mIw2YdZwA0Fr/USm1ELNeyzKl1Dla6x3OYxq6OIcQfiUlHSF69iHu8g5KqRM6OWY7MNLjmBFa681a6/uBNcBY50ujaV/6ESJgJOEL0bNvAzOcOzZtw9T823G23hM9OmfvUUptUUptAlpw7wZ2BhBKO1aJAURWyxTCR5RS3wVqtNaddeq6jlkKXKKd+6QKEUjSwhfCd/5J+z6BdpRS6cCDkuxFsEgLXwghBghp4QshxAAhCV8IIQYISfhCCDFASMIXQogBQhK+EEIMEP8PqVY9yUJ25PYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Input vs Output of the autoencoder\n",
    "convout1_f = K.function([autoencoder.input], [autoencoder.layers[14].output])\n",
    "C1 = convout1_f([autoenc_train[autoenc_train_labels=='MLII_N']])[0] #N, L, R, V, A\n",
    "C1 = np.squeeze(C1[0])\n",
    "C1.shape\n",
    "C2 = convout1_f([autoenc_train[autoenc_train_labels=='MLII_L']+1.5])[0] #N, L, R, V, A\n",
    "C2 = np.squeeze(C2[0])\n",
    "C2.shape\n",
    "C3 = convout1_f([autoenc_train[autoenc_train_labels=='MLII_R']+1])[0] #N, L, R, V, A\n",
    "C3 = np.squeeze(C3[0])\n",
    "C3.shape\n",
    "C4 = convout1_f([autoenc_train[autoenc_train_labels=='MLII_V']+2])[0] #N, L, R, V, A\n",
    "C4 = np.squeeze(C4[0])\n",
    "C4.shape\n",
    "C5 = convout1_f([autoenc_train[autoenc_train_labels=='MLII_A']+0.5])[0] #N, L, R, V, A\n",
    "C5 = np.squeeze(C5[0])\n",
    "C5.shape\n",
    "\n",
    "# Build X vector\n",
    "x=np.arange(240)\n",
    "x=x/360\n",
    "\n",
    "# Plot\n",
    "fig1, ax1 = plt.subplots()\n",
    "ax1.set_title(\"N\")\n",
    "ax1.set_xlabel('Time (s)')\n",
    "ax1.set_ylabel('V (mV)')\n",
    "C01 = autoenc_train[autoenc_train_labels=='MLII_N'][0]\n",
    "C01 = C01.flatten()\n",
    "ax1.plot(x, C01)\n",
    "ax1.plot(x, C1)\n",
    "corrcoefN = np.corrcoef(C01,C1)\n",
    "print('Correlation N\\n', corrcoefN)\n",
    "\n",
    "fig2, ax2 = plt.subplots()\n",
    "ax2.set_title(\"L\")\n",
    "ax2.set_xlabel('Time (s)')\n",
    "ax2.set_ylabel('V (mV)')\n",
    "C02 = autoenc_train[autoenc_train_labels=='MLII_L'][0]+1.5\n",
    "C02 = C02.flatten()\n",
    "ax2.plot(x, C02)\n",
    "ax2.plot(x, C2)\n",
    "corrcoefL = np.corrcoef(C02,C2)\n",
    "print('Correlation L\\n', corrcoefL)\n",
    "\n",
    "\n",
    "fig3, ax3 = plt.subplots()\n",
    "ax3.set_title(\"R\")\n",
    "ax3.set_xlabel('Time (s)')\n",
    "ax3.set_ylabel('V (mV)')\n",
    "C03 = autoenc_train[autoenc_train_labels=='MLII_R'][0]+1\n",
    "C03 = C03.flatten()\n",
    "ax3.plot(x, C03)\n",
    "ax3.plot(x, C3)\n",
    "corrcoefR = np.corrcoef(C03,C3)\n",
    "print('Correlation R\\n', corrcoefR)\n",
    "\n",
    "fig4, ax4 = plt.subplots()\n",
    "ax4.set_title(\"V\")\n",
    "ax4.set_xlabel('Time (s)')\n",
    "ax4.set_ylabel('V (mV)')\n",
    "C04 = autoenc_train[autoenc_train_labels=='MLII_V'][0]+2\n",
    "C04 = C04.flatten()\n",
    "ax4.plot(x, C04)\n",
    "ax4.plot(x, C4)\n",
    "corrcoefV = np.corrcoef(C04,C4)\n",
    "print('Correlation V\\n', corrcoefV)\n",
    "\n",
    "fig5, ax5 = plt.subplots()\n",
    "ax5.set_title(\"A\")\n",
    "ax5.set_xlabel('Time (s)')\n",
    "ax5.set_ylabel('V (mV)')\n",
    "C05 = autoenc_train[autoenc_train_labels=='MLII_A'][0]+0.5\n",
    "C05 = C05.flatten()\n",
    "ax5.plot(x, C05)\n",
    "ax5.plot(x, C5)\n",
    "corrcoefA = np.corrcoef(C05,C5)\n",
    "print('Correlation A\\n', corrcoefA)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_encoding_layers():\n",
    "    return [\n",
    "     Conv1D(30, 3, activation='relu', padding='same'),\n",
    "     MaxPooling1D(2),\n",
    "     Conv1D(60, 3, activation='relu', padding='same'),\n",
    "     MaxPooling1D(2),\n",
    "     Flatten(),\n",
    "     Dense(25),\n",
    "     Dense(10), #15\n",
    "    ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Build an encoder-only by copying the weights\n",
    "encoder = models.Sequential()\n",
    "encoder.add(Layer(input_shape=(240,1)))\n",
    "encoding_layers = create_encoding_layers()\n",
    "for l in encoding_layers:\n",
    "    encoder.add(l)\n",
    "encode = K.function([encoder.input], [encoder.output])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_repr = encode([autoenc_train])[0] #it was not train, but it was test, changed to avoid error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Exception ignored in: <bound method BaseSession._Callable.__del__ of <tensorflow.python.client.session.BaseSession._Callable object at 0x7fab64e329b0>>\n",
      "Traceback (most recent call last):\n",
      "  File \"/home/miguel/.local/lib/python3.6/site-packages/tensorflow/python/client/session.py\", line 1455, in __del__\n",
      "    self._session._session, self._handle, status)\n",
      "  File \"/home/miguel/.local/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py\", line 528, in __exit__\n",
      "    c_api.TF_GetCode(self.status.status))\n",
      "tensorflow.python.framework.errors_impl.InvalidArgumentError: No such callable handle: 124958832\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x1080 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(18,15))\n",
    "\n",
    "for i,label in enumerate(np.unique(autoenc_train_labels)): #it was not train, but it was test, changed to avoid error\n",
    "    plt.subplot(4, 3, i + 1)\n",
    "    plt.title(label)\n",
    "    encodings = train_repr[autoenc_train_labels == label] #it was not train, but it was test, changed to avoid error\n",
    "    \n",
    "    means = np.mean(encodings, axis=0)\n",
    "    stds = np.std(encodings, axis=0)\n",
    "    \n",
    "    bar_centers = np.arange(train_repr.shape[1])\n",
    "    plt.bar(bar_centers, means, width=0.8, align='center', yerr=stds, alpha=0.5)\n",
    "    plt.xticks(bar_centers, bar_centers)\n",
    "    plt.xlim((-0.5, bar_centers[-1] + 0.5))"
   ]
  },
  {
   "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}