TimeSeries_LSTM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "wP33DGJgdU4v"
   },
   "source": [
    "## **Importing the libraries for preprocessing the data.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jdSfbiCthvGx"
   },
   "outputs": [],
   "source": [
    "import numpy \n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import math\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "GtyWYClUdwid"
   },
   "source": [
    "## **Importing the libraries for model building.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "c6VcrKBVhvJj",
    "outputId": "ac4f40ea-3ed3-4c51-9a16-f08a684026c4"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "KOgnnb3qd1Mm"
   },
   "source": [
    "As Data is a Univariate Time series dataset. So, here I took only \"Feature\" column for processing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "iTF81f2hhvMN"
   },
   "outputs": [],
   "source": [
    "dataframe = pd.read_csv('train_csv.csv', usecols=[2], engine='python')\n",
    "dataset = dataframe.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 402
    },
    "colab_type": "code",
    "id": "9RbByKkhhvOZ",
    "outputId": "9add243b-fb44-43ab-e609-9ccd9900b573"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>237.97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>57.93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90.53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>260.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>49.70</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>469.72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>406.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>128.91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>85.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>121.12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>80 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    feature\n",
       "0    237.97\n",
       "1     57.93\n",
       "2     90.53\n",
       "3    260.75\n",
       "4     49.70\n",
       "..      ...\n",
       "75   469.72\n",
       "76   406.79\n",
       "77   128.91\n",
       "78    85.28\n",
       "79   121.12\n",
       "\n",
       "[80 rows x 1 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "N2CnoND9hvQx"
   },
   "outputs": [],
   "source": [
    "# MinMax scaling for 0 to 1\n",
    "scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "dataset = scaler.fit_transform(dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "K3zmr-dUidum",
    "outputId": "fd8f51c7-8f41-4d02-d057-5e04900fcdec"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "80"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "BFeasKMxidzT",
    "outputId": "75b832ae-7e52-4b34-e931-decda3610502"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "64 16\n"
     ]
    }
   ],
   "source": [
    "# splitting the 80% training and 20% testing\n",
    "train_size = int(len(dataset) * 0.8)\n",
    "test_size = len(dataset) - train_size\n",
    "train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]\n",
    "print(len(train), len(test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Here Sequences Data creation  is done with \"no_seq\" as number of sequence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "WikaxtJsid1x"
   },
   "outputs": [],
   "source": [
    "# function for pre processing data for RNN.\n",
    "def create_dataset(dataset, no_seq=1):\n",
    "    dataX, dataY = [], []\n",
    "    for i in range(len(dataset)-no_seq-1):\n",
    "        a = dataset[i:(i+no_seq), 0]\n",
    "        dataX.append(a)\n",
    "        dataY.append(dataset[i + no_seq, 0])\n",
    "    return np.array(dataX), np.array(dataY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "z0QBDZPPjJRV"
   },
   "outputs": [],
   "source": [
    "# with number of sequence=1, splitting of data takes place. \n",
    "no_seq = 1\n",
    "trainX, trainY = create_dataset(train, no_seq)\n",
    "testX, testY = create_dataset(test, no_seq)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 50
    },
    "colab_type": "code",
    "id": "xPzFAQv6un4o",
    "outputId": "e2dfa82e-2525-4c42-95a1-892758b9ff6f"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "62 62\n",
      "14 14\n"
     ]
    }
   ],
   "source": [
    "print(len(trainX),len(trainY))\n",
    "print(len(testX),len(testY))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "eLLIr5bFjQsS"
   },
   "outputs": [],
   "source": [
    "trainX = numpy.reshape(trainX, (trainX.shape[0], 1, trainX.shape[1]))\n",
    "testX = numpy.reshape(testX, (testX.shape[0], 1, testX.shape[1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  Model Building and Training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "2E2P_YICjQvb"
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(LSTM(units=50, return_sequences=True,input_shape=(1, no_seq)))\n",
    "model.add(LSTM(units=50, return_sequences=False))\n",
    "model.add(Dense(units=25))\n",
    "model.add(Dense(units=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "eYd3D98ypWss"
   },
   "outputs": [],
   "source": [
    "model.compile(loss='mean_squared_error', optimizer='adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "uVDuOmkbpWw4",
    "outputId": "765b7dea-b578-4c31-a42f-90f1d6ec0e67"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.callbacks.History at 0x17d743ccf28>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# replace verbose 0 to 1 to see the loss function performance.\n",
    "model.fit(trainX, trainY, epochs=250, batch_size=32, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 281
    },
    "colab_type": "code",
    "id": "czjG-9vBpW3K",
    "outputId": "7f41661f-063d-4c60-9197-0d149023813a"
   },
   "outputs": [
    {
     "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": [
    "plt.title('Loss / Mean Squared Error')\n",
    "plt.plot(model.history.history['loss'], label='loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction and RMSE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 50
    },
    "colab_type": "code",
    "id": "p4Fgka43pW2P",
    "outputId": "c782f8f5-ec94-4658-a891-33ef12426b67"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Score: 83.17 RMSE\n",
      "Test Score: 122.90 RMSE\n"
     ]
    }
   ],
   "source": [
    "trainPredict = model.predict(trainX)\n",
    "testPredict = model.predict(testX)\n",
    "\n",
    "trainPredict = scaler.inverse_transform(trainPredict)\n",
    "trainY = scaler.inverse_transform([trainY])\n",
    "testPredict = scaler.inverse_transform(testPredict)\n",
    "testY = scaler.inverse_transform([testY])\n",
    "\n",
    "trainScore = math.sqrt(mean_squared_error(trainY[0], trainPredict[:,0]))\n",
    "print('Train Score: %.2f RMSE' % (trainScore))\n",
    "testScore = math.sqrt(mean_squared_error(testY[0], testPredict[:,0]))\n",
    "print('Test Score: %.2f RMSE' % (testScore))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "colab_type": "code",
    "id": "rORfCV47pWwI",
    "outputId": "5772a908-4780-445b-8453-313e605465fb"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD4CAYAAAAXUaZHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOydd5gkV3W339s5TU67O7O7s1HSBq3CIgkUEIgkQAIbgwUOsjHJn0gfYCwMMk6yhDE2yDaPDQZb+JPAYIKEZJCEkEABlLVabc67szs5d+6uvt8ft6q6Os305NmZ+z7PPrPTsbqn6lenfufcc4SUEo1Go9EsLVwLvQEajUajmX20uGs0Gs0SRIu7RqPRLEG0uGs0Gs0SRIu7RqPRLEE8C70BAM3NzbKzs3OhN0Oj0WjOKp577rkBKWVLufsWhbh3dnby7LPPLvRmaDQazVmFEOJEpfu0LaPRaDRLEC3uGo1GswTR4q7RaDRLEC3uGo1GswTR4q7RaDRLEC3uGo1GswTR4q7RaDRLEC3uGo1GM02yRo7vPH2SrJFb6E0pQYu7RqPRTJPHDg9w8w9289jhgYXelBK0uGs0Gs00OTkYB+DEQGyBt6QULe4ajUYzTU4NKXE/bor8YkKLu0aj0UyTU8NK1E8OaXHXaDSaJcOpoQQAxwe1LaPRaDRLhi4zcu8aSmDk5AJvTSFa3DUajWYajCYyjCWzrG8JkzZydI8mFnqTCtDirtFoNNPASqZesbEZyFfOLBa0uGs0Gs00sCwZS9wXW8WMFneNRqOZBl3DyoZ5RWcjPo+LE0OLK6mqxV2j0WimwamhODV+D/UhL6sbgpwY0JG7RqPRnPWcGk7Q0RhCCEFnU5gTi6zWXYu7RqPRTINTQ3FWNwQBWNMU4sRgDCkXTzmkFneNRqOZIlJKuoYTrG4MAdDZFCaeNhiIphd4y/JocddoNJopMhhLk8gYduS+tkmJ/IlFtFJVi7tGo9FMEavGvaNBifrapjCwuMohtbhrNBrNFDlllkFatkx7fRC3S3BSR+4ajUZz9pKP3JUt4/O4WFUf0JG7RqPRnM10DSdoCvsI+z32bYutHFKLu0aj0UyRruG4HbVbrGkM6YSqRqPRnM2cGorTYfrtFp1NYUbiGUbjmQXaqkK0uGs0Gs0UMHKS0yMJVjcUivsaqxxykfSY0eKu0Wg0U6B3LEnGkKxuLLRlOs1yyBOLJKmqxV2j0WimgNUNsiRyb1xcC5m0uGs0Gs0UKC6DtAj63LTV+nXkrtFoNGcjp4bjCAHtReIOaqWqFneNRqM5Czk1lKCtJoDf4y65b21jSCdUNRqN5mykazhekky16GwO0zuWIp7OzvNWlVK1uAsh3EKIF4QQ95m/NwohHhJCHDJ/Njge+xkhxGEhxAEhxBvnYsM1Go1mIegaTtgNw4ppjvgAGFkEte5Tidw/Buxz/H4z8LCUchPwsPk7QogtwA3AVuBNwFeFEKXXLxqNRnOWkTFydI8m7Fa/xVhWTSqbm8/NKktV4i6E6ADeAvy74+a3AXea/78TeLvj9u9IKVNSymPAYeCS2dlcjUajWTgGoilyElbUVRJ3JamprDGfm1WWaiP3LwOfBpynozYpZTeA+bPVvL0dOOV4XJd5WwFCiA8IIZ4VQjzb398/5Q3XaDSa+SaVURIY9JWXTr/XVfC4hWRScRdCvBXok1I+V+VrijK3lQwWlFJ+TUq5U0q5s6WlpcqX1mg0moUjbSjR9rnLO82LyZbxTP4QLgeuF0K8GQgAtUKI/wf0CiFWSim7hRArgT7z8V3AasfzO4Azs7nRGo1GsxCkTdH2eSpE7meTLSOl/IyUskNK2YlKlP5cSvm7wL3AjebDbgTuMf9/L3CDEMIvhFgHbAKenvUt12g0mnkmNam4m5H7IrBlqoncK3E78F0hxB8BJ4F3Akgp9wghvgvsBbLATVLKhT+NaTQazQyxI3f3JJ77WWLL2EgpHwUeNf8/CFxT4XG3ArfOcNs0Go1mUWF77kvBltFoNBqNworc/RXEPeBdPAlVLe4ajUZTJVUnVDM6ctdoNJqzBstuqei5L6JSSC3uGo1GUyWTRe4+z+JJqGpx12g0miqZLKHqdgm8bkFS2zIajUZz9jBZ5A7KmtGRu0aj0ZxFpCapcweVVNWlkBqNRnMWMdkiJjDFfRGsUNXirtFoNFWSNnJ43QKXq1x/RIXfq20ZjUajOatIZ3MTRu2gbRmNRqM560hncxMmU8ESdx25azQazVlDdeLu1p67RqPRnE2kjSrE3attGY1GozmrqN5z15G7RqPRnDWksjl8nvIj9iz0IiaNRqM5y6jKltHVMhqNRnN2kc4a+CezZbx6EZNGo9GcVaSzOXuUXiW0LaPRaDRnGWlDL2LSaDSaJUdVde5m+wEp5TxtVXm0uGs0Gk2VVLtCVUrIGFrcNRqN5qyg2jp3YMGtGS3uGo1GUyXVrVBdHHNUtbhrNBpNlaSqtGWABR+1p8Vdo9FoqqRazx105K7R2BwbiPH0saGF3owF51DvOEZuYZNxmlKklKSN3OSLmMz2BAu9kEmLu2bR8E8PH+L/3PXcQm/GgtI7luSNX/4lD+7pWehN0RSRzUmknHg4NmAvctIJVY3GZDCWZiCaZjSRWehNWTAGo2lyEvqjqYXeFE0R9nBsbctoNFPDEvXjA7EF3pKFI5bOAhBPL/wKR00h1QzHBocto8Vdo1HY4j64jMU9pcV9sWKL+6Qtf83IXVfLaDQKS9yPLePI3RL1uCnymsVDukpbJuDVtoxGYyOl1OIORK3IfYGjPk0paUP9TaqZoQpa3DUaQImaVf63nD13K2JPaFtm0ZGq2nPX1TIajc1IXEXtQa+bYwOxBe+ot1DELFsmrW2ZxYZly/irjdx1nbtGk/fbt3fUMZbMMhxfnuWQOqG6eKnWc/efLZ67ECIghHhaCLFLCLFHCPGX5u2NQoiHhBCHzJ8Njud8RghxWAhxQAjxxrn8AJqlwZgp7hesrgeWr+9uJ1S1uC860kZ14m7ZNmeDLZMCXiul3AFcALxJCHEZcDPwsJRyE/Cw+TtCiC3ADcBW4E3AV4UQE9cOLRJ+8HwXB3vHF3ozliUjWtwBR0JVi/uio9o6d5dL4HO7Fn/kLhVR81ev+U8CbwPuNG+/E3i7+f+3Ad+RUqaklMeAw8Als7rVc8TnfvQyd/36xEJvxrLEtmXa63C7xLJNqlpee0J77ouOam0ZMEftnQ2euxDCLYR4EegDHpJSPgW0SSm7AcyfrebD24FTjqd3mbcVv+YHhBDPCiGe7e/vn8lnmBWklCQyBmNJfVAtBFZCtSnio6MhyLFlupApmtK2zGKlWlsGlO9+NtgySCkNKeUFQAdwiRBi2wQPF+Veosxrfk1KuVNKubOlpaW6rZ1D0kYOKWE8uTwTeQvNaCKD1y0Iet10NoWXb+SuSyEXLdWWQoKqmFn0towTKeUI8CjKS+8VQqwEMH/2mQ/rAlY7ntYBnJnxls4xSfMSSkfuC8NoIkNd0IcQgnXN4WVbDmmVQsbS2Vn5/B/7zgv8+2NHZ/w6mupLIcGK3Be5uAshWoQQ9eb/g8DrgP3AvcCN5sNuBO4x/38vcIMQwi+EWAdsAp6e7Q2fbaw+EONa3BeE0USauqAHgHXNYeJpg/7x5dcZ0SqFzMnZKaV7ZH8fTx4ZnPHraKbqubsXvLeMp4rHrATuNCteXMB3pZT3CSF+BXxXCPFHwEngnQBSyj1CiO8Ce4EscJOUctFfY1qRezSlbZmFYDSRoT7kA6CzOQyoipnW2sBCbta841y8lEgbBLzTLzRLZVUOaTmeJOeCKXnuHhfJBY7cJxV3KeVLwIVlbh8ErqnwnFuBW2e8dfOIlfzQkfvCMBLP0GYK+bomJe7HB2Ncur5pITdr3ommstQFvYwmMsQzBg2TP6UiQ7E0gBb3WaLaUkiwqmXOgoTqcsCK3MeTs+N1aqbGaCJDfdALQHtDEK9bcHSZJVWNnCSZydFS4wdmXg45GFXiPhBNkdNj+2ZMOpvD7RJ4qhF371mWUF3KJM3I3cipksi55sZvPs39L3XP+fucLYwmMtSa4u52CdY0hpZdxYxlyTRHlD0VS81sP7SmOWVzkuF4emYbpyFt5KqK2sGM3LW4Lw6SDkGfa2vGyEl+cbCfZ0/oYdCgvo/xpLIjLNY1hzk+EF/ArZp/LDFvqVH21Exr3a3IHfTYvtkgnc1V5beDJe7allkUJB2ryea61t06kUS1vw/k+8rUh/Li3tkU5vhgbFnZCdaIvZaIactkZrZ/DDgEXfvuMyc1JXF3nx0rVJcDzsh9rmvdrYgsqqftAPm+MgWRe0uYVDZHz1hyoTZr3rHKIJtrlC0z88hdi/tsks5OwZY5G+rclwvzacsktLgXMFomcrcqZpZTAzHbljEj95mK+0A0TWNYnSj6tLjPmLSRq2oBE2hbZlHhrEmda1smbl5u67JLxYiZ7HNG7s5a9+WCnVA1q2VmOkd1IJpidWOIsM+tI/dZIJ01pmbL6Mh9cZDSkfuCMVrGlllRG8DvcS2rihlrf7Aj9xlWbQ1G0zSHfbTU+LW4zwJTTaims7kFLavW4m6SmsfI3RZ3HbkD+YRqXdBn3+Zy5XvMLBcsG6Yp4kOImTcPG4imaI74aanx0ze+fHIXc0Vqip679ZyFQou7STJj4BIgxNxH7jqhWojV7tcZuUO+Yma5YCVUw34PIa97Rp57LicZiqVpiujIfbaYWuRuzlHV4j4/9I4lOdwXLXtfMqP6eET8nrkX90xe3JdTqV8lRhMZQj53yYGztjnEqaEExjL5jqyEasjrJujzzGhI9mgiQzYnVeQeqU7cs0aOn+3t1Su0K5A2pmbLwMKO2ltW4v6Fn+7nprueL3tfMqMy4bUB75yLe9IRkcX0xB1GEpmSqB1UxUzayHFmJLEAWzX/xNNZAl4XHreLkG9mkftgTIl5U8RHa22AsWS2oCKsHI8e6Od933qWA3rUZFmmVAppifsC1rovK3EfiqUrLsO2IveagGfuq2Ucgq6tGauXe6m4z2XFTDJjLLrvPprKEvapXn4zFff+cbWfW5G7um3i6N06IegqrvJMyZbxaltmXomlshWTVMlsziHu82PLgE6qQmVxX9ec7w4529x6/z5u/ObiGjMQTxuE/Xlxn0lC1RJqK6EKk7cgGEvoKVATMbUVqtqWmVfGk9mKTcGSGQO/x0VNwMv4HPd0d9oy44sselwIRuPlxb21xk/I556THjOnhuMV8y8LRTSVJeRTEV9ohp671VfGSqjC5JH7mHnFOh+N885GprqICXTkPm/E0lmyOWn3ZXaSMiP3eUmopnXk7kQN6igVdyEEa+eoYiaWyjKayCz4KkIn8XTWjtyDM7RlBqIpXAIaQj5aqxR3a73BZN78cmUqnrs1ZEV77vOEVY1QLjJRnrtr/m0ZHbkzkkiXjdwB1jXPTetf6288EF08rXBjqUJbZmbirloPuF2CxrCqm5+sBcGYFvcJmeoiJtC2zLwRnWCyfMpOqHoZT2bmtBwskTbwuITapmUeuSczBslMzh6xV8zapjAnh+JkjdmNgKwqpYFFVP8dS2UJF9gyM4vcm81Eqsftoinsq8KWUd9JcoG7GS5WplYKqROq80bGyNl2TDkvM5nJEfCohGrGkHP6R0mkDfvAW+6euxUt1laK3JvCZHOS07NcDmmdVBfT4p7ShOpMPPcUTZH8CbO5ilp3y5bRnnspRk5i5CQ+d3Uzba0Vqgt5FbRsxD3mENGytkzWwO91URtQB9fYHJZDxjOGneRa7pF7ub4yTuaqHNKy6BbTEItoQeTuJp4xpn0FORBN0xT227+31gaqqJYxxV1Xy5Rgz0/VCdX5YTiW5v6Xuumroue309sut/MmM4YZuSuRmUvfPZHOEvF7CPncROe4MmexY7f7rSjuIQBODM5exUwqa9iT7KuxZcaTGbb/xQM8sr9v1rahHMUJVSmnb5EMOmwZUM3I+ic5TqyAJrmIksyLhamLu7ZlZsTJoTg33f08u0+PTvpY5zzK8gnVnJ1QhTkW94xByKcqc5Z7QrVSXxmLloifsM89q5G7c1+oJnLvHk0ynsxycA5XbqayBhlD5m0Zs9piOuWQibRBLG0U2DItNX76o6kJrwTsahkduZeQMtR3Uv0iJmuFqrZlpoV1IFQjkM4IuVyiKpXNJ1RhbjtDxtMGAZ+byDxU5ix2yg3qcCKEoLO5unLIruHqonunFTZQhbgPx1RFjbWtc0Hc6itj2TLmvj2dpKr1mVqckXuNn4whK36GVNawrxJ0QrUUK3L3T7X9gI7cp0fYb0U3kx8AUWfkXvR4KaXqLWOuUIW5tmUMQl43NTpyLztir5jOpvCk5ZAHesa54guP8NyJ4Unf0/mdV5NQtVpWzKW4W9U7zoQqTC+5ORjLL2CymKzW3bm/64RqKVO1Zax6eC3u0yRk9uGIVSGQEyVUrT9AoS0zt5F7yIzcdUJVfc/WFVM5OptDnBpOkJmgHPJov1pt2j06eVWNJe6NVZQHAgyb1tGcirsZfDh7y8A0I/fxfOsBCyuBX6nWfczx2bS4l2LlaKoVdyHEgo/aO6vF3aoscHqolXBGa8UHjLWKbN4SqhnTltGRO2OJDLUBD26z7r8cnU1hjJyka7iycFuDtKs5WVon+s6mUFWLmOY3clf7dNBr2jLT2D+cHSEtJmtB4PxsehFTKXbkXqUtA+YcVb1CdXp43C78HldVbXOdkXvxzmtVB/i9LiL+ubVlDLP9QcjrIeKf+/bCi52ReJq6Cn67RTUNxGxxr0IMrbUFnc1hoqnspElLK+k7VqW4R1PZKa+qdQ7qgBlG7tF8R0iLycTdWsDk97i0uJdhqrYMqM6Q2paZARG/Z0q2jBClFQjWzhzwuHG7xJz2l7HeO+RT/v5yj9xHExnqg+VXp1pYte4TCWbfmBKtaq7irH1hXZN63YHxiaP3oSkmVL/2iyO89Z8eL9vDqPI2FSZU7XzSNIR2IJoi4vfY/U0AavweAl5Xxeog68TVVhvQtkwZ7ITqVMRd2zIzI+SvrgdHNGXgc7sIl1nWbVUHWAfDXPZ0t5K5QYcts5wn31Qa1OGkKeyjxu+ZUNx7RlXkXs1VnGXdWCeNycohR6Zoy3SPJommsuztHqvq8ZA/6UfsOnf1czqrVAej6QJLBpQH3FLjr7gmZNQWd7+ulilDaoqeO1jiriP3aRP2VRf9RlMZwn43QZ+71JaxInezNnUum4dZUVHQqxKqRk4u64NpNJGZ1JaxyiGPTbCQqXcKtkzU9txNcZ8kqWolVMeS1Z2ILaF84eTklTsW1tWEVSSQr3OfXuTutGQsWiL+ypG7Gcy01gT0CtUyTMuW8bi15z4Twv7q+l7HUgaRgIdgmcHD+WoZK3Kfu57u1ntbi5iAOe8fv5gZqyJyB1jbNHF3yN4pJFStZf6ttUoAJ6t1txKqRk5WdfKwxP35kyOTPtYiZu4XEccKVZieuA9G0zSFS62uiQZljyWy+Nwu6kLeRdUGebEwLVvGq22ZGaGW8FdXLRP2ecpOuCmO3OfWc8/bMlbZ5XIth5RSMlJhUEcx65rDdA3Hy/rY48mMLY7V5l/Cfo/dCnfSyD2Wtg/qaqyZ6UbuQuT3Qb/HhatMfqgaBqIpmmtKI/fWmsCE1TK1QS8h78wmQC1V8tUy1TUOA23LzJiI31NVuVgslbWTTMUJI0vcrX4Qc2nLJJ22zBRW2C5F4mmDbE5W7CvjpLMpTE6WX4Xa6/CRq62WiQQ8eN0uGkO+CT33XE6t6lzbpHrcVCvuQkDXcIK+8cn7HoHZy93nQQhVEiqEKJsfmgwjJxmKp2muELkPxzNlT5BjyQy1wfzxsZzzQOWYap07mLaMFvfpE/JVXy0TNpt1lSRUHYuYALun+1yQt2U8eXFfppF7NatTLTonKIfsNStlwj531WWx1nffHPFP2DxsLJkhJ/P+fLXifsHqegBeqNKaUU3DCqPC4DTmqA7H00gJTeU895rKNpRab+Al6HOTk5AxtLg7sXrETCbu8UycX3f/GimlWeeubZlpE/G77UvyiYim8p0YK9kyVuReG/DYdb/VkjVyVbUJti6zg+YKVVi+Pd1H4xP3lXGyzm79Wxq5W5UyG1ojVZdCWuJuNdSqhJVMtU4uk9W6Z4wc8bTBFRub8boFz1dpzVi2oZPpTGOyhLtSQhXK21BW7sPKO+lyyEKqidyllPzpY3/K+x98P9/e/20CXveUymFnm0nFXQixWgjxiBBinxBijxDiY+btjUKIh4QQh8yfDY7nfEYIcVgIcUAI8ca5/AChKSRUw353WVsmZXvueVsmnc1NKRnyjceP8dq//8WkE4MSZRKqyzdyV4nKSoM6nDSEvNQGypdDWguY1jeHq7LTxpP51rrNkYlbEFjJ1GptGev+lho/W1fV8cKJaiP3/KAOi+A0hmQ7B2MXYyWQy4p7Mktt0GtfveqFTIVUs0L12/u/zaOnHmVVeBVffOaLxMTBRW/LZIFPSinPAy4DbhJCbAFuBh6WUm4CHjZ/x7zvBmAr8Cbgq0KI6rMQUyTityYnTbwzRh22TGnkXmrLwNRWqZ4ajjMQTXFgkrawCe2524xNwZYRQrC+JcLRgWjJfb1jSWoDHpoj/uosunRh5D4wQStcqyNktbaMs4XxhWvqeen0yIQ9cSyiqay9gMli1iP3CfrLWG0ggmaAM9/iLqWc9VGKs4kl7l53+TYZB4YO8KVnv8RVHVfxveu/R0dNB7vSd5CUg/O5mQVMKu5Sym4p5fPm/8eBfUA78DbgTvNhdwJvN///NuA7UsqUlPIYcBi4ZLY33MJepj3B5biUklg6S43fKoUsFADrxOCM3GFq4m7ZAZOVvzmrZSxbZrmKe77d78QrVC02tEQ40lfOc0+yoi5A2O8hkTEwchP7xdFkobgnM7mKfwPLluloCOJ2iaoj99qgl4vWNJDM5NjfPXkf+LjjhGMxPXG3Wg+UfqfWZKbeooVMUkq7WmahbJnvPdvFq27/edUCf7Q/OqfT0opJmfNTrYS3k3gmzqd+8Snq/HX89eV/Ta2vlq+85isYMk2q8T9IGQsz7WtKnrsQohO4EHgKaJNSdoM6AQCt5sPagVOOp3WZtxW/1geEEM8KIZ7t7++f+pabVNPTPZ42kFI9NujzlKmWyeF2Cbzu4si9+p3Hev/Jyt8SaQOXUGVSfo8bn9u1bPvLTDaoo5j1LWF6xpIlf+uesRRttQFbHCdLqiqLLp9QBSo2ELNWp9aHfNQFvZOKu3U1Um9G7gAvnJrcd4+nDLuHu0W5q8zJGIim8LhE2e/U53HRHPGXiHsio6qW6oJeR+Q+v1H0/p5x+sZTdrviiTBykt/46pP85b1752HLFOlsrmIv99uevo0TYye4/crbaQw0ArC+fj2X130YETjFbU/dNm/b6aRqcRdCRIDvAx+XUk60rrrcdUtJKCWl/JqUcqeUcmdLS0u1m1GClYSaKMJxNmUK+dxkDFlwqaxG7OW/iunUn8dscZ88cg85St4iAc+yHbU3FE/jdQu7u+dkbGiJAHCsvzB67x1NKnEPTN4C2hqxZ/2NJ2uoNRxP43YJagMeU9wn3iecM2Hb64O01vh5vsoe88XfQ8jnIZ6ZqueuBmOXizABVtUHODNaKO5j5meqDTgi93mudbdyG1aPoIk43BdlNJHhgT0982YfpbO5ssnUPQN7+NHhH/G+7e/jkpWFBsWmyKtID1/Cjw7/iExu/o/xqsRdCOFFCftdUsofmDf3CiFWmvevBKwBk13AasfTO4Azs7O5pYTM8rGJInfrvog/7yk6o3c1HNvRZMkekj11cT82ELN92nIkMkZBQ6eIf/n2dB+Kps2FRJXb/TrZ0KJ87yP9ed/dyEn6oylW1AbsaHwicc/3TVd/g8nEfSiWoT7oRQhBbRWRu1PchRBcuKaeF05NnlQtn1B1T2g3FtM9muDZE8Nl/XaLlXUBukcKWyfnrSTPgiVULXHvj06+LmBXl/o+o6ksjx6Y/lX/VKgk7kdHjwJw/YbrS+7ze10YidUY0qAn2jPn21hMNdUyAvgGsE9K+Q+Ou+4FbjT/fyNwj+P2G4QQfiHEOmAT8PTsbXIh1qX4RFUF9gHt99jLup1zIpOZXEHkXjtNW8Y6qF6c4GBOpAsTZ8u5p/tQLE1juLIQFbOmKYTbJezBHKAiVSMnaav1EzFP9BPZXNaJNGL+jfO2THlxH4mnaTAXBFVjy1hWk1UBdNGaBk4MxidscWDlhIoj93AZz/2BPT2cd8tP+fw9LxcMJnlgTw/XfuUxekaTfOyaTRXfa2VdkO7iyD2ZPyHZx8d8i3us+sj9pa4RIn4PTWEf9700Z3FjAWmjvLj3xJRot4XbSu7ze1zITBMAp8ZPldw/11QTuV8O/B7wWiHEi+a/NwO3A68XQhwCXm/+jpRyD/BdYC/wU+AmKeWc7SmWUE5U3xy1bRm3HbnHC8S9MJqebkL1svWNuF0T1zZbU5gslvMc1cFY+R4olfB73KxuCHLEYctYZZBttQHboqtmX7BOBA0hH26XmNCWaTDr8OuC3knr3EcTGcI+t52/uXCNqhB+cQK7Tq0IpWwpZCJjkHMkiJ87MUwqa3DXUyd59d89yp/9cDef/eFuPvhfz9HREOT+j17JG7auqPheK+sCRFPZgsDF+ky1AW/ZK9v5YMiyZaqYjPVS1yjb2+t407YVPLyvrySwi6ay/MODB2Y14ZrO5sqWQXbHuqn31xP0BEvu83vc5NKLWNyllI9LKYWU8nwp5QXmv/+VUg5KKa+RUm4yfw45nnOrlHKDlPIcKeVP5vIDRKq4FHfaMuVmU1rzU4tfc2rinqWlxs+5K2om9N2LbZnlPEdVRe7VizuYFTOOyN1anWpVy0C1Fp0SbLdL0Bj2TRC5Z+xqnrqgpypbxln9c35HHZ5JTvh2L/cyCVXID5MBODkYZ31LhEf/5GreubOD7z17irueOsn7r1zHD/74cnuxVyVW1isRckbvlgg6q2XmO6E6HFPbMFm7hlTWYF/3GDtW1/PW81eRyBg8vK+v4DFffuggd/z8MPft6p617atky/TEek9uOrAAACAASURBVFgRLn8y9XtcyGwNXpdvcYr7Yifkq8aWcXjuZbrtpbKG7TWCmvAU8rmrtmWsy+qI38OFa+p58dRIxXK8RJnIXYt79WxojXBsIGZ/v87IvaaKhGrMcRVn0RKp3C1xKFYYuY8mMhP2XbFKCi0CXjdbVtVOeMK3t6lMnTsU7qsnhuKsaQzR0RDi1t/Yzi8//Rru+8gVfPYtW6rqe7KqLgDAGYfvPuqoWlqIUshU1rCPgcmauO3rHidjSHZ01HHJukZaa/wF1szhvnH+88njADxxZGDWtrGiLROfQNy9LsBFW3CVFvfpkF8IVM2luCOh6hT3TI6Ap/DAmkrzsETGIGdeVl+0poFoKsvhvtLFNlDGllmmCVXrgJ7Qltn13/DS9wpuWt8cJpXN2eLUO5rE7RI0R/z5hOoEJ3qr1YN1IoDKLQisrpVOz93IyQnbXYwm0tQFCyPwzW01HJugXXF+fmqRLeN1EyaB9xd/C5kEUkpOmeJusbIuyLb2uoqvXcwKU9x7CiL3/HeyEAlVK08Bk9syL5nJ1PNX1+N2Cd68fSWPHOhnPKlOun913z6CPjevOaeFJw8PFFhaMyFVwZbpifWwIlQpclfHeWtwFaeiWtynTMDrKjs6z0lhKaQ54aaoWsYZuYPZ9rfKEsWo4/Utj7XSZXgiY9hTdsD03Jdh5G6Nrmsss9gGgJwBP70ZfvTH0LvHvnlDqyqHtKyZ3rEkLRG/PR4RJrbTimeVQuXmYfG0KptsCOXFHSZepTpapj99e32Q3vFkxT4j+QqeQnEP+z28yfUMdc98GY4+ynA8QzSVLRD3qdJWG0AICsohxxIZQmaewOdWrYbnU9ytfSHodU+aUN11apTmiN++Arlux0rS2Rw/29fLw/v6+OXBfj7+us1ct2MVw/EM+3qqn4Y1EeVsmVgmxnh6fEJbBqDJv4qu8a5577R51ou71Rp1ImvD6pUd8rkJ+tRHdp4MkhnDPsta1AS8tI7uhuzkiyqsgzPid9PZFKIh5K24mCmRNgg6TiQ1/qn3sVkK2D1QKkXup5+DxBDkskrgDSWo65utckgVCfeMJWkze6b4PS7cLjFx/iWZv4qzsCL34oPPKs9z2jKQtzHKUU7cOxqCSElBdYuTfORe2hVym+uY+qV3DyfMjpgzEXev20Vrjb+gHNK5zUIIgvPc0936nje3Regfr9wKAlTkvqOjzi6fvXB1A6vqAnz/udP89f172dga4fdfuZbLNzYD8MTh2bFm0tlcyaAOq1JmMnFv9K0kkU0wmJzfVgRnvbiDOigmqgeOOnplW1FzsiihWhy5r/aO8ud9H4dn/n3S9897ph6ztrmhoscaT2ftqwdwJoSXl7jbkXulUsiDD4Bww3Vfge5d8MSXzcf7qA95CyL3tloVxakTvbuq5LozSm6p8ZMxZElEbiX5rARpbZWRe3E7hfYGlcTsGi4v7nFHqa6TkNfNdoe4nxxSHTHXNE1f3HnuP9kRGbFzFWD2cg8U5gnm03O3vudzVtSQNnL2oqpioqksh/ujnN9Rb9/mcgneumMVjx8e4MRgnM9ftwWv20VbbYBNrREePzw7glrOc7fEfWV4ZdnnWEUa9T51f9d416xsS7UsDXH3eSb0WVUvd/VFl5tNWVwKCbBe9OBCwvHHJn1/p6cPcNGaeg6Zq+iKUbaMM6GqDipj7/2QrjwjdKmRF/cKkfvBB2DNZXDxjbD1N+DRL0DvXtVArDls17r3jqVsHxnUFddk+Zewz43LlV84ZfVhKa6YsSLKxnB1tkwyY5DM5Eoi99UNSoxPVxD3/PzUooSqB7aK4+qX3j2cMsXder0p0/My/PhjfDD9rYKE6lgiS60jTxDwuue1WmbIjtxrgMoVM7u7RpESzl9dmGN46/lKPN+wpY0rN+VXu1++sZlnjg3NylVxuVLIaiP3Wo+qgZ/vpOrSEHf/xAM7oul8i9dy1TLlxL1DmOVVJ38FuYl3dOtS33oPy3ffVbSYKWPkyBjSTuqCOiFsEKdpuf8P4Pk7WS5YPUTK2jKjp6F3N2x6g/r9zX8PgTrTnsma5ZAxkhmD0UTGjtxBXcVNVi1THCFX6pZYyZapVOs+lihcwGSxoi6AS5SfIgV5W6a4cVhd7BhBkSYW6oDBw5zuH6a1xl8QHEyJXd8G4ILo46RHe2z7o9hKCnhd8+q5WwuYzllhiXt5391Kpu5wRO4A29vr+NI7d/C3v7m94PbLNzaTyBhVD0yZiHKee0+8B4GgJVS+fYol7mFXCwKhxX06hHwTD+yIJlVHSFBfuLASRtF+SAyTzObMsqU8K3K96j+JYRg4MOH7F1c7nN9RhxClSVXrUtcZodUEPGwSp9UvXc9M8kmXDkOxFO4KDa449KD6udkcBRBuhrd8CbpfhCfvYH2L8matiqRCcZ84/2KN2HPSUqF52Ei80JaZLHJ3th5w4nW7WFEboGtkssi9cLtqh3cDcKT9epAGuf4D0/fbjSzs/h60bceNwVuMR2z7o9iWCfrc6vjIpmHwyPTebwoMxdLUBDysrFP2VaXI/aWuUVY3Bkuu9oQQvOPijpK2C5eub8Ql4MlZ8N3L2TLd0W5agi14XeUb31m2jGF4aAu3aXGfDpFJIndntGYljOJpA77zbuQ9HyGdLS2FbMl2k5TmH+3EkxO+f7EtUxPwck5bTUn734Sj3a9z2zcIs06369lJPunSwaofd9ojNocehPo10HJu/ratb4dz3wq//CJbw6MAPGnWMa9wiPtk7RycU5gsWmr8vNf9E+oOFJZdWtaRNeM14vdM2Pa3kriD8t0ree6xtIHP7SoRj+DAy0RlgENN16j3Hzkwfb/96KMQ7YVXf5rB5p282/1zukdVgnasqDY/aHnuv/4qfPWVkJydipNKDMfVeoeJhomAautxflHUPhG1AS87Vtfz+GyIezZXMhy7J97Dikjl1cBW5J7KGqyuWa3FfTqoaUxm5C6l+ucgWnQpHvK5SacScOYFZI+KjoptmYZ0N7vkBnLhVmXNTEC5hTHnrKgpmRqUn59auIhpo8uM3EdOQGz2Fl4sZgajFRYwZZJKiDa9EYobir3pNpCSi/b/PQBPHlHJMqtaBsz8yyTVMiX2R8DDhz0/ZOfBf7SrckD1lakNePCYXqsQqjvkaCJTso+Boz99GXHvaAhV9NzjqazdAM+Jt28Xe2QnPd4OpCfAyuSR6Ufuu74NwQbY/EZGt/wea119JA88Qi4nGU9lSxZeJTIGHPk5GCkYPj6996ySoVia+pCPGr+qsy9XDjkYTXF6JMGOjupr+gGu2NjMrq7RGc9ELmfL9MZ6K9a4g1Pcc1rcp4tdIbHnR/B36yBauBw5VjQIIehzUz9+GHJZxOhJfGRKypxqk2c4mWslsfJSODGxuEfL1Cm31vhLyrrsyL2o/cAGcYaUR/mNyyV6r7g69fjjkInnLRkn9Wvgyk8SOXI/V7lf5uljquNFmyOhGgmY4p5NwZFHSl6i+EQPIMZO0yiihLPDStBMhh0LmCzqgl4iw3vh1pVgBgYWE0XuHQ1BesaSZYdRWNVcBRhZXD0vs1euI5aBVMNmNotT0xP35Bjsvw+2vQM8foI73sawjNCw/27GU1mkVHODLQJeN9l0Ck6Z/f6Gj039PafASDxDY0h10WytCZT13F/qUldrxX77ZLxqQzNGTvLU0aHJH1wBKWWJLSOlnLD1AOQDRkvch5JDxDKVF7PNNktD3C1bJtigPPK+PQX3W/NTLYJeN62x/QAImWON6C2M3DNJgsk+TslWRlt2wlgXjJys+P4xc0Sa02JoqfGTyBgFFkEiYw3HdpZCutggznC05RpV+nd6+Yh7U7kyyEMPgCcInVeUf+KrPgIN6/gb/51k0ilCPredTwGHLfPkP8F/vR369hc83ZrIVUD3iwDkcNlJR1B2QXFZY13Qy3nDj0A2AQd/WnBfxeEj/Qc5Xx7EyMmCEkSLeDpbUuPOwAHIJjjg3kg8bTAc2cR5rpP2LNcpsfceyCZhx7sBaKmv4wfGlXT0Pkx0UFmCxbbM2pR6f2BeInfrJNpS4y/rue/qGsElmNJqXICL1tYT8LpmZM1Yw7GdAeBoapSkkZxQ3D0ugUuoGc0dNR3A/JZDLg1x97mJZwxyLVvUDb2FE1qKo7Wgz8OqxCH79w2iu7DO3RTyk7KV/saL1W0TRO8TVWA4/cNytkww0UtYpOgKbIbWLcsmch8sF7lLqUog118N3tIuewB4A3DtF1iTO8173T9hRW2goB+8WvOQRj73n+qG7l0FT48mS/9WdO8ih5v/yV2N3P+/kFC5kpF4xq6UsagNetkaf0r9cvzxgvtGK1TLcN//5Zqn3sslYl9Z313NTy3apjMvAHDMu4l4OkuXbx0tYpQ1gWlEfru+A00boV3tyx63i4eC1+KWBu5ddwOUVMtsSZtXJd4QDM1t5D4cT9NonkStK95iXuoaZWNrpPRvNwl+j5tL1jXx+OEBjg/EePrYEPe9dIZfHam+/t1aWewU9+6YakpWqcYdlI3n97hJmpE7aHGfMmG/Bykh4WuAcCv07bPvS2dzpLM5Io6DJ+h1sTp1EFbuAGC96C6M3EdOAHBKttAbXA/+WjhZOakaLZOkWy172CaOlhV3py0jBg4CcNqzBjouhtPPT1p6ebaTMXKMJjKl4j5wUH33m98w8QtsfiOH6q/go54fcG64MNkX9nu4QuxCjJr+Zs9LBfc7R+zZnHmRWN0G/ivzWoSRUpEupnVUFLl3eMfZkD0C3jCcfKpgBfNoIkONmXTNf9gEdD2NK5fm674vMXby5ZKPE08bJfsPZ14EX4QB32riaYP9cg0AzdEpVq8Mn4ATj8OOGwpyGOmGjezzbad+390IcoXVMl43O4yXoeU8ldSew8g9mTGIpw07cm+t8Ze1ZV4+PTrlqN3iio1NHO6LcvXfP8q7/u1XfPjuF/j9bz5VcRXucCxd4NFb4u60ZSarcbfwe12kMoYt7vPpuy8JcbdapcbSWWjbUmDL2B0hHZ5ixCtYkzkOa68gE2plXXHkbu7Mp2Qr4ykJqyf23ctVYJz7wt/wfd9fkjn5lH2bVTtcUKc8oK4gjot2aN8JqVEYPFz9hz8LserHm4r7ylg2x6ZJxB04eOFn8ZDjjxL/WXB7xO/hd9w/JxdqhrZtBb548Yg9QF0tdL+It+NCdst1DIc6VaSLSqgW2zIXZ55T/3nVh5VtceZ5+76xRIa6okifrmfBSJN54+2k8XDprz5UmhNKZUsWMNH9IqzcQcDvJZE2eD6lxhCLIstxUl76rvp5/m8X3LyyPsj3xesJRE9yuWtPwSKmkFtygdwPnZdDQ+eceu7FC8VaawOMJ7MFdfb94yn6xlNsXVVe3JPZJF985os801O+lPiGS9bw+eu28KV37uBb772Ev7huCxlDVuw7c+N/PM3nfpQ/CacscXcsYuqJVynuHhepbI5aXy11/jot7lPFGrwQSxnQulX5rDm1czibell0cpoAKVi5g0TNOta7ugtLIYePI91++qlTl9prX6k80Fj5S7liTx8gNLQXv8hw8ZM3qeiJ8rYMAwcYJ0x3tgY6dqrblrjvXnF16sEHlSDXdUz6Gis6z+Vfjbdy8djDBSfeltwAr3U9z9i5N8CqC6H3ZbuypXjEHgDj3RDrJ7DmYjqbwjzivwZOPkm6/yixtFFiy2yLPU2vrEde8gF1g2MFc7m+Msq6EXgveDef8nyWYHoI7n4XpPP2SnHCHyOrTkorLyBkTmPaO+Jj1N1Q0EQNgBfvhn+7Cr7zO/DTP4On/g2euAO+/374l0vh0b+FzitVMtrBytoA341fQNLXwO+6f1aw3WvShwiLFLk1l0PjOhg5pbZpDrD2Bet7theUOSpm9nUrEd6ysrbk+VJK/ubXf8O39n6L9z34Pr7+0tfJycIr38FUFw1tu3jbhSu4anOLPczk5dOjJa8XTWXZfXqUAz3j9m2VInePy2MPxK6E3+O2Tw6rI/NbMbMkxD1kT+DJQut5KqIyo+9yq//WZ9XcQ1aeT7RmHevFmYJhHYycgIa1COFSkcWaV6nbK5REltgy8SFcsV7+O/cahJGGb98AybG8uHsdjx04xBnvaqJpA5o3g69myfvuQ9Ey4p4YVt9vuSqZMmxsreFO8XbigTb4yZ/YJ/NNp3+EW0h6N/02rDgf4oNKwCkdsQfkPfmVO7h4bSPfGN2JRJB+XiVW653baGRZN/YUvzB2EPfUqxORw3cfKSfuJ56AledDsJ7xpu3c0fAZ9Z4/+VP7IfGUUVgK2b9fJUBXXUjQbK1xcijOQGiTOllZpGPw4C0q6Bg8DM/9B/zk0/DQLWptRsM6uPJT8Bv/VvL9rawPMpbxsKv5Ol7vepa6TH4W6bqo8vtT7ZepyF0aqqhgDrCS0FbnTTtX5ZiluncCcf/ewe9xz5F7+MNtf8gb176RO164gw8//GFGkiM80/MMH3n4I1z/o+u55Ylb+NwTn8PIGaysC9AU9pUV95dPqxYHzrJVK6HqK/LcV4RW4BITS6iK3NW+Od/lkEtC3MP2wA5D2TJgRzjlWrx2pg+RwAdNmxgLd6oyOMPxhx4+gahfS0PIpyKL9ovA7a8o7rF0UZLOfO8n/VdyZ8dfQf8B+J/3kkypHTbgc3zt/Qfo9a1VwuNyQ/uFSz5yz7cecFTLHPm5EpHNb6rqNeqCXh7802sJvOU2FeU+fycYWVYf/x6/NLYz4m+HFeZydNOaKR6xByhvW7hgxXZ2djawJ15Hsv1V+PZ8F5CFkXvXM/iz4zySu0Bd0XVeUeC7l0TumaRadbxWVf60N4S4N7kDXvF+VZUzdgYppdnvxrH/mMlUVl1A2Ofm1FCcRMYg2XiO2pesKPrZb0J8AH7rm3DTU/BnZ+BTh+HTx+ATe+A934HXfhbq2ku+P6tl7g9dr0cAkd3/Zd/XPvY8R3IrSfib1QkC5iypWnwV11omct97Zoz2+mCJ5fVi34vc9vRtXNl+JR+/6ON84aov8NlLP8uvun/F6/7ndbz3gfeyq38XH9rxIT6040Pcf/R+bnniFnIyx9b2Ol4+XWrLWC0OxlNZO0GeLmPL9MZ6J7VkwPLc1fM7ajrojnWTyc3e+L+JWBribtsyWZUEQkCfqpiJOtrxWrQnD3JArgG3h5GgulyNRI/nX3BYRe4NYZ+K3D1+VWlQYaVqSbWMKe4jNZt4Um6HN38RDj/EhUe/htsl8jtJYhhifQwG1+Z7urfvVM/PlF/wshQoa8scfABCTXZFRzU0R/y4tv0mrL0cHv5r2P1d/PEe7jKuUULetlU9sETciyL35s3gC/OKTtUTaHfTtfjGjnOROFSYUD38EDnh5onctry4O3z3EnE//ZyKwDsvB1Ste/doAuPSP1ZXGk9/ncN9UVLZXOF4vO4X1RVc4waCPjfDZnTrWrFNvd7QUdVk7ok7YN2rYc2l6nlCQKQFQhNbBZAf2vHEUITHxYWIF76lTlI5g7aRF3kqd57yvRs61RPmKKlq9++xxV1tlzOpurd7jC2rCqP2gcQAn3z0k6wIreC2K2/DJVwIIbjh3Bv4r2v/i9esfg23XHYLD/7Wg9x0wU3cdMFNfOTCj/Djoz/mz5/8c7asDHOwd7ykqdiuU/kgz+oFlKpgy1Ql7k5bpmY1hjToifZU9+XMkCUi7o6Eqi+kfEJT3Esi91yOFfGD7DY6MXKSocBaAELjx9X9iWGV1GzopNGK3EH57t27IFU6YanElunbA8FGPDUrVLXMK/4INl/L9t57CHtFvnTPTKaOhdfnpzF17FQ9zItK+JYSg0U+KzkDDj0EG1+vrl6mghBw7RcgOQL3fpRsqI2HcxcpIQ/UKnEyxb3cSmIrcQmwvjlCfcjLjzM7ybpDfNr739QHHJUvhx5ivOUixggrcbfsuuOPIaVqGVwQXZ54AhCw5pWAGtqRMSR9nhVw3lvh2W/yq/3qMv2KTc35550xt8nlKsjPhNdcoP7T+7K6Uon1wavz9s5UWGXOUj01lOBe/5tVa4L990HPS/iyUX6dO0+tUq1dBW7fnCVVi1s8NIbVwHKr1j2RNjjaHy2wZIycwad+8SnG0mN8+TVfps5fmGjd1ryNL776i7zrnHcR8OQXuH3g/A9w0wU3ce+Re3kq8QU8bd/loz/7Ez7x6Ce4a99dgKqntxaKWWWrh0cOgCtli7uRM+iL91Up7oW2DMxfxczSEncr+m3dYte6l/TvHjmO34ixR3aSyBgMetvISDeBMdOHN5Of1K+lIey1e02z9lXKNuh6uuC9s0aOZCZXeFnduxfattJaF8iPb9v+W9RkB7nU46iEMcsgozXr84ud2s2k6hL23YdiKepDXntZP13PqsEck5VAVmLFdtj5XshlSG5/D1k8+f74K7bb4l4yYm+8V/nxK5VoulyCi9c08ERXiqe3fo7LXPtY+9RfqITseC/0vERi7WsAs6Y93KQS+McfJ5lRJbcFkfvxx5Uvb0bSHWZf99PDCXjlhyE5gth1N+uaw3RYbXyNjNreVWqbrHySENC6frta6HbmeXj8yypRal4VTJXmiB+PWbK5P3QJ1K9VswuOPwGQj9xdbpWMnavIPZamLpjfF9wuQVPYZ5cQH+gdJycpiNzv2ncXz/U+x+cu+xznNJ4zpff70I4P8bGLPsZQ+hTu8GH2Dr3MywMvc/vTt3PvwQfpGk7w5u1m//XhBM/0PMOtL36QwIrv23XuA4kBsjI7YY27hVUtA9gLmbS4TwGr+sE+oNu2wtARyCTsiNg+oLtV3fPLuU7i6SwJw8VJ2YpvxKwfNmvcaVhLY9hn95pm9WXg8sDRXxS8t9WN0o4GczlVZ9+6hZaIn8FoSg1z3vQGMsLLG8Sv808eOAhuH9k6Vcts5CTUtEHd6iXtu5e0Hjj4UyVaG66Z/ou+9nNwyQfh0g8CjhN923ZlY6SipVdxjmSqxc7ORo72x3jY9xq+mr2e8O5vwdNfg8M/Uw/Y+HrA0RnS9N1Ho6r6xRb3bFot33eIb4dzaMfqS8mtupirBr/HlRtMG0VKNVrQSCmriXxl1YraAIFgWC1GevrrEO2BV3962l+X2yXsbpq1Ib+6ujzxBDz/LRKRNfTSmC9HbOicO8+9zEKx1tp8rfveM4XJ1BNjJ7jjhTu4uuNqrt9w/bTe833b38ejv/1z3F23cGXwS9z3G/dxTsM53PbMXyPcUa4+p4WQz82RwR5ufuxmQOKpfZnxjKqWq7YMEkxbxvTcW0Ot+Fw+uqLzs5BpSYh7QbUMqIoZmYP+A2UP6JzwcFCuJplWUfdRuQr3kCnuVoRiJlSHY2nVH8YfgY5XwLFCcS/uCMnICcjEoG0LLTV+chIGYykI1LIn+AquNn6dX6TUfxAaNxAOBApei/aLoeu52f2SFhGD0XRhH/dDDyrrIji1viEFBBvgzX9HqF4dcPZ3uWI7IKFvb+mIPVvcz7dfZqfpu/9sXy//LN4N57xFCe6Td0BkBcHVKqIec4p7NkH6hDoZ2+J+5nnlx6/Ni3t7vTm0YyQBQnBkw42sFT28PbJbCfvPPq+i51d9FM65FsiL+2qrp0zbVuW7r3mlitxnwErTd68NeOHC31NFAwMHiK64DIBE2txPG9ap46K4WdpL31WlmDNg2NF6wKK1JmAnVPd2j1IT8NDREMTIGdzyxC343D5ueeUtBSuTp4oQgm3tdew5PYrP7eO2K28jno0SWPkDtq6qpb0hwOMj/8JwcpjfWXcLIHm0Wy1usxYwtYXaJn0fvzdvy7iEi46aDh25TwWfRw32tXu6t5qJtL69RNNZfB4XXssC6HmJ8dqNpPESz6jFEsdZiRg6qrzf4RMQqIdgPY1hH9mctKfDs/5q5Ycm8n3aS04eptdP27aSFgS/8l9BixzIL3wZOAgtm+2rCluQOnbC6MmSxS5LhYLIfeSU8pCrLIGcDJdLEPK5HeK+Tf3s2V1q0XW/qCJhf439/O3tdfjcLk4MxqkP+eE3v6Zsvv79sOl11AS8COGI3E3xFidVSaQt7laJpEPcgz43TWGfnai7N7OT07KZ80/+P/jl38MTX4GdfwSv/yt7NanVh2itJe7W53n1p0u7Zk6RlabvXhv0KOto2zsASLYrcS+I3FNj7Dt6gnteNDuYSqnKMH/8cfU3nCbO1gMWzlWqe8+MsWVlLUIIvr3/27zQ9wI3X3IzraHWab+nxbb2Ovb1jJMxcmxq2ES7/E08NXv52an78DY8wTC7+OTOT7K55pVko+fyUNc9pIxU1atTodCWgfkth1wS4g4Q8rvzQ68b16sopHdP4epRKaF7F/FGVS6ZSBuksjm6XO3qUni0y65xh3w1hzUphnWvBmRBbXNJ5G4tMmk5t0Tcn3DvJIsH9v5IdS0cPgbNm+3qDTupaibgii2gs57h4/DDPyYUPZGfnXroAfVzlsQdiiZz1a1WU5xMcS8Ysde9y/bbLQJeN9vNtrL1IZ+6Ynv3t1UuZMd7cLkEtQFvXtxN3z10RpXJ1gdNoTrxhDophJsKXr/D0df9l0dGeKjm7XhOPQmP/A2cf4OaOuUQbWsspN0N8uI/hHfeCetfM+PvySqHtE9Il38M1l5OZr2yx+w5qo2qHPInj/2Km7+/m1xOKusx2qOOm0dvL/8GPS+r/XwCykfufoZiKdLZHPt7xtmyqpaTYyf5yvNf4aqOq7hu/XXT/MSFbF1VSzqb41BvVHV5PHUJDa7zuO3p2+gS34PYVt5z7ntIZ3Nkhi5nLD3CT479hJ5YDyFPiFpfad19Mc5qGYBPXPwJ/vHqf5yV7Z+MJSPuYZ9jSIPbAy3nQN++wtWj4z0Q6yfZrOqfE2mDZMagy23WAQ8eUgJUr8Td2uls3739YtVT5Oij9vuWRO69e1Sk44/YZV2WuA8aIfYFL1K9SwaPKOuo+Ry7NUI0ZQrGqotUj5wD98/iN7QI+MUXiywv4wAAIABJREFUYdfd/KfxGXZkzcU4Bx9U31fz5ll7mxrnwA4h1GKmnt2FJauxQRg9VeC3W+xcq6yZhrApevVr4P0P2/55XdBbOLCj8wrq+p9nqzimhNLIqPr3taXJzvaGIKeHE4zE07zUNUJs6+9ApA22vB3e9i/gKjwkLVvGHtIRalSDS2YYtUO+HNLuK9N6Lvzh/+KtUVFxwhm5A77xkyQyBt1jSTjysLpvy9th190l3TfZey/86+Vw70cn3IaheGkDOcvOfP7kMPG0wXkrItzyxC14XV7+/LI/n5Ed48TqVfPymVFOjyQYimV519pP4RZuwu4GxrvewVgySzqbw4hvYG3Neu7ed7ddBlnNdoT8bqLJrMqnAevr17Omds0kz5odlo64+932FHlAeZN9ewsXiJgeq9GqxD2eVgONe72qRIn+g6ojpLkzW5eLduTu8akD3BFRl5TX9e21bSFr7JdVMZNIZ9ldd7V6j93m1J/mTXbUP25F7i4XnPtmVR44SeRz1hDth93fI73xzQzKWt6578MqMXjsF+UHc8yAkpm6bdtMzz2V7zFktvm1qlKc7OxUCc6GIrvAokTcL3g3OeHhfv9nWfnj96gEbCZWtpKloyHE6ZEETxweREq49LxO+NhL8K47VVBS5vEuUX515kyxxtoVd7G0eh+lisQ9ElOJwMN9UbXorPkceMs/gC8CD/9V/gX69qt5t74IvPSdilegCfP4K/6eW8yg6NEDatXsllV13HDuDdzyyltoC0/uc1fLuqYwYZ+bPadH7X7xV63fzN1vuZuPnHsH5NSAFRV5C965+Qb2De3j192/rsqSAdjcWkPayHGkv7SEeq5ZQuLusVsNAOqSeLwbEkN5y6TnJUDYvmUiY5DMGsS9DeCvU5fSRrrElrFr3UFZM4OH1BBnnIukPGpF4uARe5Vs0Ow1biWH4mmDww1XqcqQp7+uXq95U6nnDmqkXDoKx345a9/RgvLcf4CRonvnn/Cb6b9ksOVS+N9PqeTgLFoyYA3JdpzoV2yHTJxI7KQjmWqK+4rzS55/sRW5Vyvuqy7kaxffy+3ZG/D074EH/kzdvra0J31HQ5BUNsePXjxNxO9hx+p61ca4Ats76njhljewqa2m4mOmS7vpuRe3TLA6pNqRuy8M4VYa02qfP949oBb0bXitsp1e9VF1lXnyKUiOwnfeo1o2f/CXyiK97/+qY6OIoaIB5BbWuL1HD/ThdQs2t9Vy7bpruXbdtbP34VH5mS2ratl9epRdp0bwuV2cu6KWdXXr2NqmAr7TIwm7/cD1G66jxldDNBOtWtwti293V2mrg7lm6Yh78Xg1U2Cb40dUtDbeC89/C9q24g+rLzyRNkhlDAJeDzRvzEcY9Z1A3paxVtEBsP7V6qdZNVNgywwcULXwrVvsh7fU+B2Ru0Eu2AjrroT0uPKDfWH74LL6bACw7ioV+ey/b+ZfzkKTTasqkI2vo8e3ljHCHLjmG3DpHyvhrTSYY5pE/J78il+w2xC0xg4qcR86Cs/9p7KCylToNIZ9/Nmbz+W3Li7fwKxE3IH+jJ+7Pb+J+PhLKpp9/V+r1aJFWIL68/19XLa+KZ/on4CSTpOzxNZVtXz+ui28bkthNBww67mTmbxXLBvX0ZpVPXqME79SJ+WNZunqK/+PshF/9hfwgw+qvNU774SmDeq7GDoCj/9DyftbV8TlPHeA/T3jbGytKRlvN5tsa69jb/cYL5wc4byV+ffKl63G7fYDdf4w79ikks7VivuGlghBr5vdZfrYALzvzme47X/3lb1vpiwdcS+O1kyBXZk8SoM3ozrxxQfhbf9s+5iJjLosDHhdqmoibXaCMyP3sM+Nz+2yV1Sq190KoWb7RFCQULWSqW3b7Ic3O4YPJDKGeu8tbzPvVD5zU9iHEIVLrvH4YdPrYf//2k2xzlr2/FCtgLzsj/OtB2pCcO3t8KHH1WedRUpsmZZzwOVhVfII53ACvvFGSI3D2/+14mt84KoNKqouQ23Qmy+FNBlNZFQC1htUNeOXl/earcVKRk5y1ebmso+ZL1wuwR9evq6kXbXHrarPEo62u5naNXQIVb3V0vu4WrW61lyh6wur6p2TT8LBn8Ab/zZvSW14jWo3/Ng/KNvTQXG7XwvLzoS5saOcbFtVRzKT45kTQwV/78awj6DXTddwgnQ2p6YquVR7g7A3zHmN51X1+m6XYOuq2rJNypIZg18enLuZyUtH3H1FtkzNSgjUszZzhPf33qosmd/6Jqy60K6Lj5sJ1YDXDU2bzCcKFVGjamEbw7685w7KD193lYrcpSSWyuJxCbV6rXePqtJpXG8/vKXGz8C4yvxnc1KJ+7nXqWZVrWoH8bhdBavybM59q1pifpauVh1LZrjxG0+Revyf1IlswzXlm4bNMpFicff4oeVcLks/yae7PwFuL7z3ATUcZRpYkbtzPm7ZjpBlaG/IT5i6YuPCivtE+L2ugmEWsVAHqxik1pvj3NgzsOYyJeoWF/+Bqih6xfvAaods8YZb1WPv+7iqSIsPQSbBkHlFW2x/Bbxu+7ss7ikz21hJVSnhfMd8ViGEWdkULxiO3R5p57HffoyrV189pffYc2bMTqpavHByhLSR45J1k/cCmg5LRtxVKaQjwhUC2rbyFuMRtow/Adf+nb0wxFpGnEhnSWYtcd+gnlezssADbQj7GIoVRmmsf7Xy8wcO2hUYQpjNylrOKUiMWWPDrAMl4HWry/UbfwyXf9x+XHPET3/x7MhNrweX96y1ZnadGiF2+HH8/bvh0g+BEPn+3eG5sRrAMUfVyYrtdMrTxLwNSthbprZs3Uld0EvGkAWRbdle7hW2rT7kpb0+WNgsbJER9LoLmmqN+NtxCcm7VvSwiZMk1lxd+AS3l5ev/T59V95amhyPtKja/RNPwD9uVUPsb13B9fdsY5PoKjso3bJm5jpy39AStvVgR0dhj5p2s2y1eDi21z21fXd7ex2JjFGSVH3q2CBC5BP4s82SEfdwmQNatm7BI3I8u/I9cMn77dtdLkHQ6yaRMUhlcuqP22xG7qYlY9EY9hZ67mDWuwNHf0E05RiRZvaUcdJS42c8lbWTR/aszM4rCjzZ1toyU98Ddcqf339f6epA+0NKOPQzNT2o3DCR2CDsv19Vq8w2iWH41b/AyV+XvfvkUJw/9PyUUSIY29UkoKFYmhq/B79zOMosE/Z7SGVzZAzH0IYLfof7cq/irq1fh/rVM3p9S8Sdvnu14g7wpq0reM+la2atpG8uCPrcBZF7n1f1UXknDwJwrO7SgscbOcl7/v0pbv9pUUmkxUW/D7/7fbjuDnjTF+Caz/NUx3sZoK7s92YlVeda3D1uF+etrCXi97C+JVJwX0dDUCVUs7mCdr9T5fwKSdWnjw1x3oraqvebqTK1abOLmLDPoxYbGDk7SZW44A+448lxGs/5v+wsenzQnHBjR+6NZuRutTg1aQj57P4WNo3rVC38sV8Qy1ysyiDjQ2pRhyOZCtBi+ocnh9SqxJJxaiatNX4O9Y6X3nHuW+D+T6o+3q3nFt53/HH42V/mm5kJt/JBrUqbgw+ofuJIZTX93g/zJ7FqSY4qu6m2Xf1ze9SKxF9/FZ67U5X8eYLwez/Ie7D29j3Gm1zP8PXsW9l2KskVmyJqMHbxeL1ZxtlIzhqTl1r9Kj6c/jB/UlOa5JwqTnG3ygnHEpnSwdgVuP0dpRU6i42Ax11wZXIaJe6bBh+hX9bycnY1zj19X/cYY8lsWW8ZUNH8xtcV3HT/4MvIM2cKZ86arGsO0zuWmrNkspP3X7me7tFEyXZ0NIQYiWcYiqVnlNRd3xIh5FNJ1XeYSfp0NsfzJ4d59yVzV/O+dMTdPKDjKYO6kPpDRGs28q/G9fxNoNTftSJ3O6HqCymbZP3VBY8raB7mZP2rYc89dNZewbBvsyOZWiTu5uXliUHVWCpYQdxbTPsml5P5FZQA57xZifv++5S4Gxk12u3Jf1YLSWpWwVu/rBbj7L9fPe6nZhvYVRfC1Tcrb/++T8A33gC/8z+FXvPoaVUWuPF1pYnNkZNw5/X5dq/CrQR+7LQ6WLf9Flz4u3D/J+Cud8GN96rBJgC7/4d37f84J13tfJvrueTF01yxqZmhWKrsZfhsYvXujzrEveyIvWliWUqH+6Kcu6IWKSUj8Qz18yBE80XA5y6oljmTrSEhfQRzaZ6Ul3J4IF7w+GeODwFwpD+Wz2NNwlCZ1gMWn7n2vII5qnPJW84v393Rqpg5OhCbkbi7XYItK2sLKmZ2nx4hmclx6Rz57VCFuAshvgm8FeiTUm4zb2sE/hvoBI4D75JSDpv3fQb4I8AAPiqlfGBOtrwIuzNkOmuf7UtavDoImZedyYyRtwhe/5clj2sI+RiJZ8gauXyLWlBLxXd9h5v7b1YtBX5olkb9//bOPLqt+srjnytZsuUtXmInduLE2QjZQ8hGkrYEKAUKCQMtTWeg0GXaM9DTZTqdbmfapsvpNtNl5rQUukFb1m4kLGUv0IUSkpCUOBvZsB3b8Zp4izfpN3+89+QnS7IlW4qelN/nnBxJz5J1I0tX933vZquUAbtzNz4Mvihv+vKCbIYCio7eAUpt1QIUVhqdsXsfNCbzHXrckEN8xUa53ep/NSo0wHCsl/+XUernyYUCW7nWlMXwq3+Ce6+Dd//CmBm/61448ozRKTt1qZFwtiL71iPwy03GGcANPzGWh5yuNcrcCiqM6YvWrtP3bYOfXwW/vgFue8IYKfDslzmYtZgfTf0KqydN5sl9TXz1+sW0dQ8EPzTJwhrnYK+eirhib5ysnFnCvPJ8vvHEQS67sBylYCigknZ6nQpyskKrZVp7BqlnCvOo43D+Ko42h+rHlnP3BxSHmrqiVhrZiTR6wCIvOyt0AU4KsCqb3mzrYc4IySZelkyfxIM76vAHFG6X8Mpx4/ValSS9HWLT3O8BRu4++yzwnFJqHvCceRsRWQhsARaZj/mRiCRPXLURNtPddj1k1rqJJcv0DwZGjTKsKPP0iNI3qtfDp4/y+dwv8WzRuw0nPOttRiu5DWsEwdiyjDmqoDtCR+qCTUbj1IFHYd6VsOV++PcDRrmdJ4KjLJkd6tjBSBh/8GnjZ/ffZDSaNO6FDZ80SgLP1BuLlnfda8wE+cVVRnfsbY/D0pvg4luNL44bfwpXfjV0iXVhpRG1Z+XATy4z6p0X3cBtQ5+jrGwKm5dPo6t/iD8dbA4f95sE8myRu0XEFXvjxJvl4hs3LOHk6bN875nDQe09k5y7z+se7lDFyJU0Zxnvqfap6zliSw4qpdhxvD3Y/GXtPB2Ljt7BqI1iTsAKQgb9Kph0HS8jk6qvHGtnXnl+aCCXYMb8alRKvSQi1SMObwYuNa/fC7wAfMY8/qBSqh84LiJHgNVA5OWjCSS4as+WBOoeOffFhs/jpqd/iAG/KctEwT48bPLIP0ROIS/4l9FfeTlX3RQ+o8R6vEug1orcR5FlwNgdeeHI/oi1txt6dsVyYwTCeCmYCu9/3Fj0MH2V8UVhVfbMfhv84SPw6MeMCp28MsNhx6rRF1fD+7bDA++BBZs4s+7ztO16lhkluVwyp5TJ+dk8suekMQUwiWWQMDzErSeic0+MA15ZXcI/r5nBz/5ynAvM7tGMcu4eNw02597W08+reZeyft5iynJmUnfgjaD8cry1h9buAT759gs43NQVnqOKQkfPAEumJTdhOhFK87zkeFz0DQYm3Ei1ZNpwUnX25Dx2vdnB5uWViTAzKuO1eIpSqhHAvLTmb04D7PMs681jYYjIh0Vkp4jsbGmZeCVHsHY9JHK3jQYYgc/rDkbjsUTuISMIbBgr9qI/3u0SSvOzg5H7aLIMEF4xA4ZDr1o9McdukTMJrviSMbvGPsuksBJu2QZXbIWqNfCBP8affC27AD72Grx9K7UdRllnVUkubpdw3bIKnj3QzKBfhc5yTwKjnsUlIHK3+MxVF1Kan81XHjXGPGeSc8/xhCZU27oH2Fd6FVz3feaW5xNQcMLMI1mSzJpZpSyoKKSmYexWe6UU7b3RZRknICLBjuKJVnfZk6r7Gzvp7h9izezSsR84ARJdChmptitiDZ9S6m6l1Eql1MqysolXMFgOvDvGD3Su181pM1E62imXddoYVg6J8QbtGfCPqQ2W5WcHPyi5ESQiGC79CmtkOpe4XLDhE0Z0P6JqKF6sLzNrVO3m5dOCTRzJT6iag9hs74XR8i/jZZLPw9ZNi4K/O9Ocuz2h2tYzwGSzymlOmVGff7TZcO47jndQmudlTlkeCysLOdjUFdawM5LeAT8DQ+FDw5yGpbtPNHK3OlVfP3mGHcetL8Pk6e0wfud+SkQqAMxLa6tEPWAvIp4ONIzfvNixtOzeCLJMxMjdkxXcKh9b5D4Y9rP+oQD+gBrbuRcMyxDRZJlcbxb52VnBxcDpjuXcrQ1Cy6ZPYqY5tjbZpZCRZJmw0cwJ4urFU7ligXHi6uQoNF5yPC76zM9SIKBCciWzJ+cjYk6HxIjcV1YXI2IM4uod8Aerw6IRHEPheOduRO4TqXO3WDxtEvsbOnn5aBvVpbnBNYfJYrwWbwduNa/fCmyzHd8iItkiMguYB+yI8PiEM3rkHkmWcQWji9E0d6u8rb0nPKIe7cvDjt25R0uoWveLKMukIbXtvZTmeYOvjYiwebmh0KVClglbsZcgRIRvv2sZ33nX0uApfCbg87jpMztUO/sG8QdUMPnn87qZVuTjaEs3TWf6qG3vDVZ9WE1HYyVVrTNhp38hJipyB6OZ6eygnxcOtyRt5ICdWEohH8BInk4WkXrgS8A3gYdF5INALfBuAKVUjYg8DOwHhoA7lFLnpFg116pzt82Xaersw5vliuhQ7fJIzih6Wo7HTZ7XHTFyjzUatPT0LJeMOgWwzDZkLN2pbe8Z3vtpctu6atwiSe86tNYudttKIS1nEqlyaqKU5Hl598qJdb06DZ/HzaBfMegP0NptvHaTbWdcc8vzOdLczQ5Tb7ec1QVTCvC4hZqGTq5dGj1haJ01lyRxDEUisGYBJcK5W0lVf0CxZlZy9XaIrVrmvVF+FHFVvVLq68DXJ2LUeLDWkdlrm18+2sYq83RxJHYpZqyGi5J8b0TNPdbyOityjybJWJQXZFMTY6WB06lt7+WiquKQYyV5Xj5+RZxJ2nFiTAkd/qJ/8XALy6uKQhvENFGxPhN9g/5hCcUWZc8py+fvx9p45VgbeV538Avbm+VibnnBmBUzVtLV6vB1KtMT6NxnTTaSqr0D/nMSuWfMbBlrMbL1gW7u6uNgUxcb5kZO1tqj+exRZBkwdMFI1TLD1TijRx+Wcx9NkrHu19yZ/pr7oD9Aw+m+4b2fKcA+a+hEaw81DZ28c0nkTkRNODley7kHaDN7L+yTPOeU5dM3GOCP+5pYMbM4pMFvYUXhqLKMUorf7qxnVXUxlQ6XshKpubtdwuLKSVROykl6Ix9k0PgBMKQWq879b0eMIVrRxqraHe1YkbsxGTKSc4+tvM6aLxOtDNKivCCHngF/6K7PNKTxdB/+gEqpc7dPhnz8dWPJxNVLYluwoBl+r/YN+ofHNI+QZcBIjI7sslxYWcjvdtfT3NUXbM6z8+qJDo619vBvl85JlvkJoyw/m+JcT4gkNRG2bl5E74D/nAyNS18PEoF826n4n99opSjXw6Io86DtDn2s7rOSXG+wMsBOvAlV3xh6r6XNt3T1p7VzH1kpkwrsM92feL2Ri2YUBZNjmrGxigzODvpp67bW4dllmeFxxSOdu/WZ29/QSfn8cOf+0Kt15GdnRZ3p4iREhCc+/haKfIlx7guSnG+ykzGyDBiRe+/AEEop/nqklfVzJkfVWOON3DtGjdzHcNpmyVMssgxEaWRKI4I17qWplWV6+oe0JDNOfCGaez+FOVkhunOpGdF63MJFM0LnyCwYpWKmq2+QJ15v5LplFVF7PpxGxSTfmPkyJ5Ier26MWKfiR1u6aersY8O86JtufPEkVPO89Ni3NpmMNt7ATp7Xjc/jHtO5O6KRKQHUtvficQtTk1zHOxr52VnUd/TaJBnt3OMhuCR7wE9rpNEbGHXbQ34V9vmZ5PMwvdgXMan66N5Gzg76uSnDqoucSEY599xsN+09A/zlDWMv4WhrzOzfxDljyDL2LlV7dj/WMbIiQnlhdkyaO5D2jUx17b1UFedGnNN9rsjLdtPdP8Tj/2hkxYyijKpBPxcEq2WGjIRqaQTN+X+3XBT18YsqIydVH9pZxwVT8lkew9RIzcTIKFnGOhX/y5FWZpbmjqr5htS5xxC5Q/h8mZ6BIXI8rtBRwFHYumkRt2+cO+p9inweslySEbJMKvV2MCqYWrr62d/YyTU6ao8bny1yjzbJszjPG7UJaWHFJI639oSUox5q6mJv3WluWlnl6C1UmUJmOXevmzNnh/j7sfYxlw/HK8sAdIxoZDKGhsV28nPp/PIxoxWXSzKikam2vTellTJgJNet8SbaucePlVDtMxOq8Y6mXVhZiFJwsGl4u9hDr9bhcQs3rJg+yiM1iSKzZBlvFq1mTe5Yzt3Svz1uGVM+sLroRm5kSkbJYnmajyA40zvImbODKXfu1t9lxYwix9dSOxGfbflNR+9A3CMjrIqZ2+/bxZyyfKqKc3mypokrF05N+uA4jUFGRe5WFO0SWDdndOduReuxjPIMau4jZJnuvqGEt7OneyNTXUfqyyBh2Lm/c5QWeE10rDPbU2f6CKj45wFVFvn44rULWTOrlLODfp4/1ExP/xDvu2Tm2A/WJITMitzNZqIl04vGXKxrRe6jDQ2zmOTzIEKwmcMiHlkmVsoKcthTdzqhv/NcYq0TTHXkPn9qAeUF2VybBrXUTsQKfupPnwWgZBwbgz6wYVbI7bD9wJqkklHO3XK0G+aOPZTHF0fknuV2McnnCYvcewaGgt2niaK8IJu2noHwna1pwnADU2qlkFXVJez4whUptSGdsRr76jsM5z45AVKKduznlvTzHqNgOff1Y+jtYLzRsrNcMUXuYCRVwzX3sRd1xEtZQTZKEZzEl27UtvdSkuelIAGLqDWpQ0TI8bg42WFF7lonTzcyyrlfsXAKX7t+MWtjHKeZ63WPWSljUZIb3qWaDFnGPoIgGn872sp3njrIoD8Q8ee7azuoN7XvRKKU4j9+s5cvb6+Jep86B5RBahKDz+Omycz/lCZ5760m8WSUcy/M8XDz2pkxn/75PLE790jDw5JSLVM4eiNTZ98gH3tgDz/801E+ev9uBoZCHfwvXz7BjXf+jU89vDehdoFRyvbbXfU8sKM2ZG6+HSeUQWoSg8/jxh9QiEDxGDksjfPIKOceLz6vO3ZZJjd0pnsgoOgd8CchoTr6fJnvPn2Y9p5+bltXzVM1p4IOXinFD559gy9uq6E0z8srx9s5aSbDEsGJ1h6+8th+phX56B8K8GezC9hO/5Cfk6fPUnUOxplqko8V+BT5PGmZ/znfOa//YlUluTG3pVuRu1JGZ0zPQHLWtlkJ2kiyzL6TZ/jlyye4Ze1MvrxpEVs3LeLp/ae4/b7dbH10P9979jA3rpjOQx+5BIDtexKzvnbIH+CTD+8hyyU8+OG1FOZk8cz+U2H3e/5AM/6AYtU5WESgST6Wc4+3gUnjDDKqWiZefnzzxcTaBV2S52HQr+juH6IgxzM8VybBzt2b5aI41xMmywQCii88so+SvGw+9Y75ANy6rhqXwH9tMzTwD22YxeevWYDLJayYUcS2PSfjnpn9VE0TX3t8P9csrmDL6hnMmpzHj144ymu1p/nBluVUleRy2YXlPHfgVFhFz+92n6SsIJu3xJDQ1jgf66xWNx2lJ+e1c49Vbwd7I9MgBTke20TIxI8CNRqZQiP3B1+tY2+d4WALbZUot1xSTXGel+6+Id6zanhmx/UXTeOL22o40NgZ1wzpe/92grbuAX76l+Pc9dIxVleXsKu2g03LKoMLrq9cNJVH9jSw680O1sw2ktdt3f28cKiZD2yYpU/hMwSrSzVRiyo05xb9KYwRaypewxlDx+6JcVHHeCgvyKGle9i5t3b3860nD3LJ7FI2LQvvuLx2aSVbVs8IGcb0ziUVZLmER/acDLt/45mzQXnJTnNnHy8fa+NDb5nNy5+9jE+/Yz6NnWepLMrhq5sXB+/31gvK8LpdIdLMtj0NDAUUN+q5IRmD1QuiI/f0RDv3GFlQUUie183t9+3mxcMtMS/qGA/ltsj9TO8gH7jnVc4O+Pnq9YtinqZXmp/NWy8oY/ueBgKBYUf+m511XPKN57l/R23YYx77RyNKwaZlFZQX5nDHxrm89OmN/OlTl4Z0/OZnZ7FubinPHDgV/JL43e56Fk8rZP7Ugon81zUOItvS3HUZZFqinXuMVEzyse2jGyjLz+bWn+/g/54/AiQnci8rNCZDnu4d4OafvcLBxi7uvHkFc8vjc5ybl1fSeKaPV463A/DqiXY+/4fXAbjrxWP4A6HR+6P/aGBBRWHI84hIRJnl7Qun8GZbL280d3OwqZOahk4dtWcYVuSuZZn0RDv3OJhbns8jd6znPSurePmYsYA7GZF7WX42A/4AN931MoeauvjxLSu4fMGUuH/PlQunkud1s23PSerae/nIr3ZRVZzLt29cSm17L0/uawret669l9dqT0eUfSJxhWnP0zVN/H73SbJcEvNjNenBsCyjI/d05LxOqI4Hn9fNt961lDWzS9i+t4GKSYlfJWc1Mp1o7eWuWy5m44Xl4/o9Pq+bdyyayuOvN7K7tgN/QPGz21YxoySXO188yt0vHeWaJVMREbbvNcomYx20NaUwh2VVRTxZ08Spzn42XliuS+YyDKtaJtIWJo3z0ZH7OLlhxXTuef/quCpuYmVxZSEzS3O5633jd+wW1180ja6+IY619HDnv6xg1uQ83C7hQ2+Zxd76M/zNWIAoAAAGMUlEQVT9mCHZPLq3gRUziuIaHXDlwinsO9lJS1e/lmQyEF9Qc9fOPR3Rzt2BzC7L58VPb2Tj/Ik5doB1c0q5evFUvv2upayz1Z/fuGI6pXle7n7pKG+c6uJgU1fcssqVCw1ppjjXw2UT/BLSOI8cr66WSWe0LJPhZLld3HnzxWHHczxubltXzf88c5jsLDcugWvinH0+tzyflTOLWTd3Mt4xloxr0o/rllaSneXWcluaIpHqnc81K1euVDt37ky1GecdHT0DrPvm85wd9LN+bin3fWhtqk3SaDRxICK7lFIrI/1Mh1vnMcV5Xt6zqgpAV7poNBmGlmXOc27fOAcRo8tVo9FkDtq5n+eUF+TwpesWpdoMjUaTYLQso9FoNBmIdu4ajUaTgWjnrtFoNBmIdu4ajUaTgSTNuYvIVSJySESOiMhnk/U8Go1GowknKc5dRNzAD4GrgYXAe0VkYTKeS6PRaDThJCtyXw0cUUodU0oNAA8Cm5P0XBqNRqMZQbKc+zSgzna73jwWREQ+LCI7RWRnS0tLkszQaDSa85NkNTFF2gUXMsRGKXU3cDeAiLSIyJsTeL7JQOsEHp8snGoXONc2p9oFzrXNqXaBc21zql0Qn20zo/0gWc69Hqiy3Z4ONES7s1KqbCJPJiI7ow3PSSVOtQuca5tT7QLn2uZUu8C5tjnVLkicbcmSZV4F5onILBHxAluA7Ul6Lo1Go9GMICmRu1JqSEQ+CjwFuIGfK6VqkvFcGo1GowknaYPDlFJPAE8k6/eP4O5z9Dzx4lS7wLm2OdUucK5tTrULnGubU+2CBNnmiGUdGo1Go0ksevyARqPRZCDauWs0Gk0GktbO3Unza0Tk5yLSLCL7bMdKROQZEXnDvCxOgV1VIvInETkgIjUi8nEn2CYiOSKyQ0T2mnZtdYJdI2x0i8hrIvKYk2wTkRMi8rqI7BGRnU6xTUSKROS3InLQfL9d4hC75puvlfWvU0Q+4RDbPmm+//eJyAPm5yIhdqWtc3fg/Jp7gKtGHPss8JxSah7wnHn7XDMEfEoptQBYC9xhvk6ptq0fuEwptQxYDlwlImsdYJedjwMHbLedZNtGpdRyWz20E2z7AfCkUupCYBnGa5dyu5RSh8zXajlwMdAL/CHVtonINOBjwEql1GKMysItCbNLKZWW/4BLgKdstz8HfC7FNlUD+2y3DwEV5vUK4JADXrdtwNudZBuQC+wG1jjFLozGu+eAy4DHnPT3BE4Ak0ccS6ltQCFwHLNIwyl2RbDzSuCvTrCN4TEtJRiVi4+Z9iXErrSN3Ilhfo0DmKKUagQwL8tTaYyIVAMXAa/gANtM2WMP0Aw8o5RyhF0m3wf+EwjYjjnFNgU8LSK7ROTDDrFtNtAC/MKUsn4qInkOsGskW4AHzOsptU0pdRL4b6AWaATOKKWeTpRd6ezcx5xfoxlGRPKB3wGfUEp1ptoeAKWUXxmnytOB1SKyONU2AYjItUCzUmpXqm2Jwnql1AoMSfIOEXlrqg3CiDxXAHcqpS4CekitbBWG2S2/CfhNqm0BMLX0zcAsoBLIE5GbE/X709m5xzW/JkWcEpEKAPOyORVGiIgHw7Hfp5T6vZNsA1BKnQZewMhZOMGu9cAmETmBMa76MhH5tUNsQynVYF42Y2jHqx1gWz1Qb559AfwWw9mn2i47VwO7lVKnzNuptu0K4LhSqkUpNQj8HliXKLvS2bmnw/ya7cCt5vVbMfTuc4qICPAz4IBS6rtOsU1EykSkyLzuw3ijH0y1XQBKqc8ppaYrpaox3lfPK6VudoJtIpInIgXWdQyNdl+qbVNKNQF1IjLfPHQ5sD/Vdo3gvQxLMpB622qBtSKSa35OL8dIQifGrlQmNxKQkLgGOAwcBb6QYlsewNDNBjGimA8CpRhJuTfMy5IU2LUBQ676B7DH/HdNqm0DlgKvmXbtA75oHk/5azbCzksZTqim3DYMbXuv+a/Get87xLblwE7zb/oIUOwEu0zbcoE2YJLtWMptA7ZiBDX7gF8B2YmyS48f0Gg0mgwknWUZjUaj0URBO3eNRqPJQLRz12g0mgxEO3eNRqPJQLRz12g0mgxEO3eNRqPJQLRz12g0mgzk/wGpBwT7O1bW6gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "trainPredictPlot = numpy.empty_like(dataset)\n",
    "trainPredictPlot[:, :] = numpy.nan\n",
    "trainPredictPlot[no_seq:len(trainPredict)+no_seq, :] = trainPredict\n",
    "testPredictPlot = numpy.empty_like(dataset)\n",
    "testPredictPlot[:, :] = numpy.nan\n",
    "testPredictPlot[len(trainPredict)+(no_seq*2)+1:len(dataset)-1, :] = testPredict\n",
    "plt.plot(scaler.inverse_transform(dataset))\n",
    "plt.plot(trainPredictPlot)\n",
    "plt.plot(testPredictPlot)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction for future dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "4Ifu--Nw4bk2",
    "outputId": "75e3cd51-98a9-46f8-c55f-e6f02b30b5a7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "40\n"
     ]
    }
   ],
   "source": [
    "inference_data = pd.read_csv('test_csv.csv')\n",
    "\n",
    "No_test_data = len(inference_data)\n",
    "\n",
    "# Number of features to be found.\n",
    "print(No_test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 714
    },
    "colab_type": "code",
    "id": "806G0Jv8u1vE",
    "outputId": "aaecc443-1862-4a76-dbf4-ffc38b2eab3a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[131.32393]]\n",
      "[[134.24103]]\n",
      "[[135.0666]]\n",
      "[[135.29958]]\n",
      "[[135.36526]]\n",
      "[[135.38379]]\n",
      "[[135.389]]\n",
      "[[135.39049]]\n",
      "[[135.3909]]\n",
      "[[135.39102]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n",
      "[[135.39104]]\n"
     ]
    }
   ],
   "source": [
    "# Prediction taking the last dataset of train and predicting the rest 39 features.\n",
    "\n",
    "for i in range(79, 79+No_test_data):\n",
    "    \n",
    "    # last day of the training dataset.\n",
    "    if i == 79:\n",
    "        last_day = dataframe[i:i+1].values\n",
    "        last_day = scaler.transform(last_day)\n",
    "        X_test = []\n",
    "        X_test.append(last_day)\n",
    "        X_test = np.array(X_test)\n",
    "        X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))\n",
    "        pred_price = model.predict(X_test)\n",
    "        pred_price = scaler.inverse_transform(pred_price)\n",
    "        print(pred_price)\n",
    "        \n",
    "    # rest 39 features.   \n",
    "    else:\n",
    "        last_day = pred_value\n",
    "        last_day = scaler.transform(last_day)\n",
    "        X_test = []\n",
    "        X_test.append(last_day)\n",
    "        X_test = np.array(X_test)\n",
    "        X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))\n",
    "        pred_price = model.predict(X_test)\n",
    "        pred_price = scaler.inverse_transform(pred_price)\n",
    "        print(pred_price)\n",
    "        \n",
    "    # replacing the old prediction value with new for next feature prediction.\n",
    "    pred_value = pred_price"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "TerraBlue_assignment.ipynb",
   "provenance": []
  },
  "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"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}