replot_all.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Activation, Dense, Dropout\n",
    "from keras.layers.core import Flatten, Dense\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.metrics import categorical_crossentropy\n",
    "from keras.optimizers import Adam\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.convolutional import *\n",
    "from keras.models import load_model\n",
    "\n",
    "import keras\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import shutil\n",
    "import os\n",
    "import random\n",
    "\n",
    "from PIL import Image\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn import datasets\n",
    "\n",
    "import time\n",
    "from PIL import Image\n",
    "from numba import jit\n",
    "import pickle\n",
    "from pathlib import Path\n",
    "import skimage.measure\n",
    "\n",
    "from keras.optimizers import Adam\n",
    "\n",
    "from keras.models import load_model\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def chart_regression(pred,y,sort=True, plot_name = 'pred'):\n",
    "    pred_folder = 'pred/'\n",
    "    ex = '.png'\n",
    "    t = pd.DataFrame({'pred' : pred, 'y' : y.flatten()})\n",
    "    if sort:\n",
    "        t.sort_values(by=['y'],inplace=True)\n",
    "    \n",
    "    y_true = test_label\n",
    "    y_pred = np.squeeze(pred)\n",
    "    erro = (np.abs((y_true - y_pred) / y_true)) * 100\n",
    "    \n",
    "    \n",
    "    a = plt.plot(t['y'].tolist(),label='exact')\n",
    "    b = plt.plot(t['pred'].tolist(),label='prediction', linestyle='dashed')\n",
    "    plt.ylabel('$E_{GS}/L$')\n",
    "    plt.xlabel('samples')\n",
    "    #axes = plt.gca()\n",
    "    #axes.set_ylim([-4,0])\n",
    "    plt.legend()\n",
    "    plt.title('RMSE ='+ str(np.round(100000*score)/100000))\n",
    "    #plt.savefig(pred_folder + plot_name + ex)\n",
    "    plt.show()\n",
    "    \n",
    "from sklearn.utils import check_array\n",
    "def mean_absolute_percentage_error(y_true, y_pred): \n",
    "    y_true, y_pred = check_arrays(y_true, y_pred)\n",
    "\n",
    "    ## Note: does not handle mix 1d representation\n",
    "    #if _is_1d(y_true): \n",
    "    #    y_true, y_pred = _check_1d_array(y_true, y_pred)\n",
    "\n",
    "    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 10k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv('csv/pool_10k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_pool_more/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "X_1 = dataset.iloc[:,1].values\n",
    "Y_1 = dataset.iloc[:,3].values\n",
    "\n",
    "W = np.random.permutation(np.c_[X_1.reshape(len(X_1), -1), Y_1.reshape(len(Y_1), -1)])\n",
    "\n",
    "X = W[:, :X_1.size//len(X_1)].reshape(X_1.shape)\n",
    "Y = W[:, X_1.size//len(X_1):].reshape(Y_1.shape)\n",
    "\n",
    "x_train = np.array([X[i] for i in range((len(X)//10),len(X))])\n",
    "y_train = np.array([Y[i] for i in range((len(Y)//10),len(Y))])\n",
    "\n",
    "x_test = np.array([X[i] for i in range((len(X)//10))])\n",
    "y_test = np.array([Y[i] for i in range((len(Y)//10))])\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test[0:1000]:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test[0:1000]:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/pool_train_10k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score (RMSE): 0.014074658902991817\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool_10k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "result = pd.DataFrame(pred)\n",
    "result.to_csv(\"pred_pool_10k.csv\")\n",
    "\n",
    "label = pd.DataFrame(test_label)\n",
    "label.to_csv(\"label_pool_10k.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv('csv/rand_more.csv')\n",
    "path_folder = 'data/matrix_del_rand_more/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "X_1 = dataset.iloc[:,1].values\n",
    "Y_1 = dataset.iloc[:,2].values\n",
    "\n",
    "W = np.random.permutation(np.c_[X_1.reshape(len(X_1), -1), Y_1.reshape(len(Y_1), -1)])\n",
    "\n",
    "X = W[:, :X_1.size//len(X_1)].reshape(X_1.shape)\n",
    "Y = W[:, X_1.size//len(X_1):].reshape(Y_1.shape)\n",
    "\n",
    "x_train = np.array([X[i] for i in range((len(X)//10),len(X))])\n",
    "y_train = np.array([Y[i] for i in range((len(Y)//10),len(Y))])\n",
    "\n",
    "x_test = np.array([X[i] for i in range((len(X)//10))])\n",
    "y_test = np.array([Y[i] for i in range((len(Y)//10))])\n",
    "\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_rand_10k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score (RMSE): 0.19041727407295692\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xd4VFX6wPHvSe8dCD2h9xqagIAgooIN7Cg2UNHVde1rQ9dVf66C64oiil1RQbGgIiIgRbr03gKEllASAumZ8/vj3MlMkkkBkswkeT/Pk2dum3vPzcC8uae8R2mtEUIIIYrycncBhBBCeCYJEEIIIVySACGEEMIlCRBCCCFckgAhhBDCJQkQQgghXJIAIYQQwiUJEMJjKaUSlVKZSqnTSqkjSqmPlFIhTvs/UkpppdQVRd73hrX9NmvdTyn1ulIqyTrXXqXUpBKuY/95qwLK/5BV7jSl1AdKKf8SjvNTSs20yqGVUgOL7I9QSn2slEq2fiYU2R+nlFqglMpQSm1TSg0p4TrzrfP7WOt1lVLTlVKHrDIuVUr1Ot/7FjWHBAjh6UZorUOALkBX4Mki+3cAY+wr1pfftcBup2OeBBKAnkAoMAhY6+o6Tj/3n0+hlVKXAE8Ag4E4oBnwfClvWQKMBo642DcJCLLO0xO4RSl1u9P+6Zj7iQaeAmYqpeoUKc/NgE+R84YAq4DuQBTwMfCTcxAWtZsECFEtaK2PAL9iAoWzH4G+SqlIa30YsIHCX7Q9gFla60PaSNRaf1LJRR4DTNNab9ZanwT+Bdzm6kCtdY7W+g2t9RIg38UhI4BXtdYZWutEYBpwB4BSqhXQDXhOa52ptf4G2AiMtL9ZKRUOPAc8VuS6e7TWE7XWh7XW+VrrqYAf0Pp8blzUHBIgRLWglGoEXArsKrIrC/gBuMFavxUo+uW/HPiHUmq8UqqjUkqdRzluUkqllvLTxDq0PbDe6a3rgXpKqehzvXSR5Q5O19mjtU4vcq32TusvAe/g+unEcVKlumACRNHfsailJEAIT/edUiodOAAkY/4SLuoT4FbrL+UBwHdF9r8M/B9wM7AaOKiUGlPkmO+KfNGPdVUYrfUXWuuIUn72W4eGAGlOb7Uvh5bzvp3NAZ5QSoUqpVpgnh6CSriO/VqhAEqpBKAv8L/SLqCUCgM+BZ7XWhc9n6ilJEAIT3eV1joUGAi0AWKKHmBVzdQBngZma60zi+zP11pP1lr3BSKAfwMfKKXaFrmO8xf9e+dZ7tNAmNO6fTndxbFleQDIBHYC32PaHJJKuI79WulKKS/gbeBBrXVeSSdXSgViquqWa61fPofyiRpKAoSoFrTWfwAfAa+VcMhnwMMUr14qep5MrfVk4CTQ7mzLoZS6uUhvp6I/9iqmzUBnp7d2Bo5qrY+f7TW11ie01jdrrWO11u0x/29XOl2nmVLK+cmks7U9DNM4/5VS6gimQRogSSnV37off8wT10Hg7rMtm6jZivZqEMKTvQEkKqW6aK3XFdn3JrAYWFT0TUqpvwPrgBVALqaqKZTiPZnKpLX+HPi8HId+AnyklPocOIx5uvmopIOtL2p7O4OfUioAyNZaa6VUcyDV+hkKjMNUpaG13qGUWgc8p5R6GtNO0wnTSJ0GNHC6TGNMYOkOpCilfIGZmKeTW7XWtnLcl6hF5AlCVBta6xTMF+8zLvad0Fr/rl1PcJIJvI5ppD0G3AeM1FrvcTrmxyJPArPOs6xzgFeBBcA+66eg/UQptdnqemq33SpnQ0xvrUygqbWvO6ZnUjqmPeVmrfVmp/fegHlSOAm8AozSWqdYPbaO2H+AFOv4o1rrHOACYDgm6KQ63Xv/87l3UXMomTBICCGEK/IEIYQQwiWPCBBKqWutR26b1S1PCCGEm3lEgAA2AdfgooFRCCGEe3hELyat9VaA8xjgKoQQooJ5RIA4VzExMTouLs7dxRBCiGplzZo1x7TWdco6rsoChFJqHhDrYtdTWuvvz+I84zD9wGnSpAmrV6+uoBIKIUTtoJTaV57jqixAaK1d5qg/h/NMBaYCJCQkSB9dIYSoJJ7SSC2EEMLDeESAUEpdrZRKAvpgJiz51d1lEkKI2s4jGqm11rOA80ptYJebm0tSUhJZWVkVcToBBAQE0KhRI3x9fd1dFCFEFfKIAFGRkpKSCA0NJS4uTrrNVgCtNcePHycpKYn4+Hh3F0cIUYU8ooqpImVlZREdHS3BoYIopYiOjpYnMiFqoRoXIEAG3FU0+X0KUTvVyAAhhBDi/EmAqEYWLlzIn3/+6e5iCCHcKS+7yi4lAaIakQAhRC33yxPwYl2YcTvklzjNeIWRAFEJPvvsM3r27EmXLl24++672bdvHy1btuTYsWPYbDb69+/P3LlzAbjqqqvo3r077du3Z+rUqQXnmDNnDt26daNz584MHjyYxMREpkyZwqRJk+jSpQuLFy921+0JIdxlxTsA/LJfkasrv22wxnVzdfb8j5vZcuhUhZ6zXYMwnhvRvsT9W7du5auvvmLp0qX4+voyfvx4/vjjDx5//HHuueceevXqRbt27Rg6dCgAH3zwAVFRUWRmZtKjRw9GjhyJzWZj7NixLFq0iPj4eE6cOEFUVBT33HMPISEhPPLIIxV6T0KI6mXe8Tp0PZNLbLh3pV6nRgcId/j9999Zs2YNPXr0ACAzM5O6desyYcIEZsyYwZQpU1i3bl3B8W+++SazZpkxggcOHGDnzp2kpKRw4YUXFow7iIqKqvobEUJ4rMu6NCY2PKDSr1OjA0Rpf+lXFq01Y8aM4eWXXy60PSMjg6SkJABOnz5NaGgoCxcuZN68eSxbtoygoCAGDhxIVlYWWmvpWiqEKFGLumFVch1pg6hggwcPZubMmSQnJwNw4sQJ9u3bx+OPP87NN9/MCy+8wNixYwFIS0sjMjKSoKAgtm3bxvLlywHo06cPf/zxB3v37i04B0BoaCjp6eluuCshhCdp0qx1lVxHAkQFa9euHS+++CJDhw6lU6dOXHzxxSQmJrJq1aqCIOHn58eHH37IsGHDyMvLo1OnTjzzzDP07t0bgDp16jB16lSuueYaOnfuzPXXXw/AiBEjmDVrljRSC1FLzYi4k9dCHkU16VUl11NaV98pFRISEnTRCYO2bt1K27Zt3VSimkt+r0K433VTlqEUfHV3n/M6j1JqjdY6oazj5AlCCCGqCe/MY4T6Vd3XtgQIIYSoDrJOMT1tNMMzvq2yS0qAEEIIT6c1euErAGQF1Kuyy0qAEEIIT7fvT9TyyQAE1mtWZZet0eMghBCiJrDFtGaLXycyczWXDRteZdeVJwghhPBwiw7aGH7qCdYO+hhfn6r7u14CRDUQEhICwKFDhxg1alSpx77xxhtkZGQUrF922WWkpqZWavmEEJXIls/xtT/Qzns/t/aJq9JLS4Bwk/z8/LN+T4MGDZg5c2apxxQNED///DMRERFnfS0hhJulH4HkrfD9fYzc9jCPBPxIgG/lJucrSgJEJUhMTKRNmzaMGTOGTp06MWrUKDIyMoiLi+OFF16gX79+zJgxg927dzNs2DC6d+9O//792bZtGwB79+6lT58+9OjRg2eeeabQeTt06ACYAPPII4/QsWNHOnXqxP/+9z/efPNNDh06xKBBgxg0aBAAcXFxHDt2DICJEyfSoUMHOnTowBtvvFFwzrZt2zJ27Fjat2/P0KFDyczMrMpflxDCmS0fbDZ4vTW83RvWTwfgeGjVpNdwVvMbqT+8vPi29ldBz7GQkwGfX1t8f5eboOvNcOY4fH1r4X23/1Suy27fvp1p06bRt29f7rjjDt5++20AAgICWLJkCWDyNk2ZMoWWLVuyYsUKxo8fz/z583nwwQe59957ufXWW5k8ebLL80+dOpW9e/eydu1afHx8ClKCT5w4kQULFhATE1Po+DVr1vDhhx+yYsUKtNb06tWLAQMGEBkZyc6dO5k+fTrvvfce1113Hd988w2jR48u130KISrYC1FQp3DWgoM6mtNdxlV5UeQJopI0btyYvn37AjB69OiCoGDPq3T69Gn+/PNPrr322oKJhQ4fPgzA0qVLufHGGwG45ZZbXJ5/3rx53HPPPfhYDVZlpQRfsmQJV199NcHBwYSEhHDNNdcU5HOKj4+nS5cuAHTv3p3ExMTzuHMhxDmzpz5K2VqwKTOqHTfkPE3DOpFVXpya/wRR2l/8fkGl7w+OLvcTQ1FF03Xb14ODgwGw2WxEREQUmhuitPcXdbYpwUvLueXv71+w7O3tLVVMQrhLfm6xTat6vcmBWclVMv9DUfIEUUn279/PsmXLAJg+fTr9+vUrtD8sLIz4+HhmzJgBmC/w9evXA9C3b1++/PJLAD7//HOX5x86dChTpkwhL8/MS1tWSvALL7yQ7777joyMDM6cOcOsWbPo379/BdypEOKs2fIh44R5dZabUWj15FWfsWHuJ3T22U/TqOAqLKAhAaKStG3blo8//phOnTpx4sQJ7r333mLHfP7550ybNo3OnTvTvn17vv/+ewD++9//MnnyZHr06EFaWprL89911100adKETp060blzZ7744gsAxo0bx6WXXlrQSG3XrVs3brvtNnr27EmvXr2466676Nq1awXftRCiXFL3w6vxMHVg4e25hZ/eI78bzf35n/BSj0zCg3yrrnwWSfddCRITExk+fDibNm1yazkqkif8XoWoMfYuho+tEdEProfIOLOcmQrL3yb/yGa8t892HH/lZOhacR1HJN23EEJ4IpvNERwA/tsZJnWECeHgH4oe+CRP+z9e+D2BVd9ADbWhkdoN4uLiatTTgxCiAqXtL3nbttmsOKKYsVLzbGAggToTrv0YWl9WtWW01MgAcbY9fETpqnM1pBAeJS/HPDGU5Otb6Q309n0Ov3v/gH1LzLgtN/GIKial1H+UUtuUUhuUUrOUUuecGyIgIIDjx4/Ll1oF0Vpz/PhxAgKqvoudEDXOvOccy95+JR42+tL+eNdtDT3urIJClcxTniB+A57UWucppf4PeBJ4vIz3uNSoUSOSkpJISUmp0ALWZgEBATRq1MjdxRCi+ko9AKeTwcspl1J0CzJv+IaDf37JLzvSyT+RyF2+cwnRpxnWu7v7yurEIwKE1nqu0+pyoPSUpaXw9fUlPj7+/AslhBBnS2s4kwJ+webn4F+w7C3Y+iPk50C3MQWHZpxO5aIpWzlyqjV+Pl7c0vtaQgZPhPSj4OURlTueESCKuAP4yt2FEEKIQtKSYNpQuG02RJUwq9uf/4PfrASbz6XCd+MLpc1gmyMzQ1DGIYJDvXnlmo5c2rE+4YHWOAc39VhypcoChFJqHhDrYtdTWuvvrWOeAvIA18OHzTHjgHEATZo0qYSSCiGECxtnwqmDsOYjuPgF18dscPrbduErEBwDzrXdGcewefnhZcsB4KcH+ld5Cu+zUWUBQms9pLT9SqkxwHBgsC6lhVlrPRWYCmagXIUWUgghSuJt/YXvIl8SACnbHcn2AI5uMpmhExcXOmxJXmteC/gbjw2oTz8PDg7gIVVMSqlhmEbpAVrrjLKOF0KIKhfd0rw2sgYg2/Jh7WfQtC+c2ANfXAtNnXKuZaaaAHFiLyx6tWDz3JbP8cV1gwjx94iv31J5RksIvAWEAr8ppdYppaa4u0BCCFFIq6EwIQ06jDTr676AHx+At7pDnpVDad8SkxYDYP8y2D4H+j5ARoRjsp9/jR5SLYIDeMgThNa6hbvLIIQQpdIazhwDHz9Y8gZs/9mxL+uUY/n7+6zj82G6mf8lWccSZ43drU6DeD0iQAghhMex2Ux3013zwD8MDq2FXx6DLjfDuiL9aH64v9RTxakjZuHJg5VU2MohAUIIIYpa+5l5Eug62iw7yzldrlO8m/AzecunYPMLY3zMOrw7XQv+IZVQ2MojAUIIIYqyVxOlbC++L6fsfjRL/Ptz9+rLWBs5kCZ3v493iH+Z7/FEntJILYQQlUNrM/+CzVbyMblZsHsBLJ5o0m7bHdlY/Nis1BKrivK0FxdkvclyOgHQNX0h0dU0OIA8QQghqoOsU3BkA8T1K/vYonbOhS+uA+VtqoyueLP4MXMeNwPgisrLKr4tMxVO7nV5KR9l49OHR9I8PxGmTD77snoYeYIQQni+b+6Cjy43vYjO1hlrKLPOh78+Lr5/9Qeug0NJEm6HKY5Atc5WOO1G8zohENvBdImd4HrK4OpCAoQQwvO1GmpebXln9z7nUc+thoGXr2O0s81m0mH8/GjZ54ntCEBq/+dYtW5twebZagBtg8+YlQsfhRH/PbvyeTgJEEIIz+cbZF5zM83I5DlPmpHMzj65EpZbY2y1hqkDTbI8e4PziT1gy4WTiTD/36bb6sKXSw06Nt9gALKP7gQgYvHzBB1eWbC/z52v4x9jZY/ufCN0v+08b9SzSIAQQni+3QvMa16WqW5a/rbJdZRqTdW5fwXsWWjaEsCk1j60FjZ+7TjHsR3mdcPXJvXF+xeZ9QZdoUkfs1ynTaHLTsq8HAB/7WiLaNi2V8FydGxTaNgd/EIgunmF3KonkUZqIYTnO7zevC74NxxcbZa3zjZf9Bf8zaTZBugy2rzmltAVVXnBUatnUkCE6ZHUbQxHvepSb/8ySNlW6PDOzRvBvsKniBj5X9h/vcnN5OMHvoFmbIQtv/CEQDWABAghhOdrlADHtpuJd+zSDphXe3AA2PyteVJIWolLTybBSw3MclgDEyBm/53dtvbUK1qf4u3PkPYNHAHiuk9g8yzwC4IWTsmpfQPNa1622VeDSIAQQng+e6ptZ67SbudmlBwcbp9j2iHsWgyG5C0ApFFkhHOdNtDuKug51lQ/ZaVBXF9od2Xx8/Z/GPr9w2NmgatINe+OhBCebcv3cGClqZL56WHT6OzMlg//qgOr3ofs06a3kfMXO4BPAGyaWf5rjvkR6rQu6J66/uKvmLE5vWB3zzZO0xQPed78bJ5lyhbbwQSHkihVI4MDSIAQQlSVnDOw41f4+laYdrFpRF71Pnw7znHMG53ghSjTyPzzYzCpvWl4PnW48LmGT3IsX/m2Y7njtWD1PALALxT6PQTxF7LskKPX010/Hmd9eljBenSoVU3UbBBc8ACkHzJVWrqU0de1gFQxCSEq3rGdpmG54yjHtr8+gTlPONZ9rBQUna6DvBzIOA6pTi3COt+0EaycCnH94XQyZKeZL3zndoeWQyEoBkLqwVXvwPuD0W2v4GDH8axKPMGKPSdY8dpC9h47w48BLejILu69tDvX974eVoTB/BchIBzGLoC6bc3TwIp3zbnzcyrvd1QNSIAQQlQsWz68Zc265hwgGvcqfJz9yze8MbxYp/RzJi6Gh7ebwNOwOyxxeoIIjIToFuQpH+ZtPc5avwdYvMjGlp9N19jQAB96xkVxa5+mtOy+FE4ncUeMNQVNWCPzGtsRGnZznDOkrunR5O13ljdfs0iAEEJULPvYBICXGsI/rcR2ofVNN9N+/4DBz0DiUrP90F/lO++qaXDRU4U22ZQPL/y0nQkHluMD3LNjDSH+oQxpW5cbmkaS0DSK1rGheHs5TdIT4DQ/WWQcRDUrPoZh1IewbXaNHNtwNpR2nmS7mklISNCrV692dzGEEM5OJ8NrLUveHxkP9682CfR2/+76mLrtzJPIMad027Gd4J7FpGflcnDiANrkbAagbf5XbPW+nsOhHdh75fd0bRJJoF/NGo9Q0ZRSa7TWCWUdJ08QQoiK8d190GyAGZlcmpN74V/RpR+TdQrumANvdID2V8PmWeRnpvHt6gO8PncHR049xSt9vRgceZQtfS8B2zHqKy/q17CBau4mvZiEEKU7cxxSrUFp342H56NcH7fuM1g22fROOh/BdSCkDml+scwZtY1vsnsAcDo1hUdnbiDIz5tP7ujJDSMupU6/28wcz96+NW4UsyeQJwgharv1X5p6+IgmZu6Ebrea7YlLTIptu2dPOuZi/uQqGPKc42kh18pVdHgdfDv27K4f0QRS96O9/cGWy6qoK5msrmfZS/PIybMR6t2Iof5hJMf04aer+tE2Ngwv5zYFUWkkQAhR2817Hhr3NN1F5zwJbUeYAOCcLTUk1nQ3tduzAH7NhQGPwSdXwEVPn901mw82jdaJiwoata+K/Jbvj11OzwPv81TYFYzu1ZTLO9WnXWwogVNfILR1R2gQXsaJRUWSACFEbedvpZmw5ZmkczlnzMhl5w4sp484MqXaRTeDWfeY5fkvln2dpn1NNdDeRdD+apJbXMtPGw+TtPRLlp8MIzM7j6TQzkS1voDfhg8o/N6/rTn3+xPnTAKEELXNxplmPMHAJ0yaiGM7TM+jHGvim/QjkH2q7PP0e8gMUFsysVyXXR1/N1szwrhp71Je+z2Rd742PZjiY/rw6E2tubRDLEotOte7EpVAAoQQNcWaj02SuSa9Sj/ul8fMqOX2V5mRw2BGLO/6zSyn7nP9voFPQmismRhHeZnBa0rBbT8D2szOlri4xMv+36+7WK1bcSzgerYG9+ahrnFc3imWFnVDz/5eRZWQACFEdZCfC7Mfgv7/MA3Kdsd2mvQV134EPz5gtjnPg5yyHZa8AQMehYg4a6PVwLtjDkQ7DRqz+/2FwutB0dDpBjPAzccaWbz6A1MegAsfA98Ak64C0NEtOd5gAH8eyOaK1E+5PvsZ1tGcmy5oxceXtCbIb8T5/CZEFZIAIYSnSj8C+5ZCh5FwYAWs/dT8dX/zN6YraZNe8PvzsGsepFizpTXtCy/EwJAJcMH9Jp3F+i+g5RBY9JppEPYJMMfOmwDJ1gQ5/mGOaqWTieY1pJ6p+/cPhWO74M2ucPnr0HqYIzgALH4d0g+bkcfAd7b+PLTqIgK9NUdbdePmLsN4qUEYzesUSaktPJ6MgxCiouVmwpc3w/Hd5/b+L2+GP9+CWXfDzDsgO93Ro6jn3fDrk/DBUPOlHRABwXWhrjVV5r6lZt7luVZKil3zzOvW2aaL6qlD4O30d+GGL81r11uKl2PUhyY4AJxJgVNJZkY3MF1Tg+ug4weyMb+JCV6WZSl+jLuwGfMfG8zYMXdwRecGEhyqKXmCEKKiHd1i/ppuPqh8uXzyc2HdF6Zu38fPvNf6axwwYwwyT5rlyDiIaGqW0w6YL+bAKDOhjStHrOk1N39rXkPqmjkWiko2aSvodqvJugqOmdLAkfba2qbTj3I0uDUjkv5G0LZ9/GElZj1WpzfPXDyc0FZty75v4fHkCUKI87HyPThYpAumtv7aL6jzL8Pu+ab94PB607U0pF7h/blnTKMymMFrMa3M8vQbzWvmCVj6ZuH3BNc1r9nphbfvXwb3ODUkNxtkXvcshH9sg0tfNW0YD+8wDd6A1poTAY0B+FpdwrA3FqHys4k9tYHYsAAevn4YtphW0Ps+Yu77ldBW/ct338LjyROEEOfq8Ab4+REzmnjcQsd2+1/79jmTS5KfC14+jr/Opw0x7Q32YGCXkwH1Opjl9dOh5cVmOS/TcUzGcbh5Jnw+ylQ7NeoBJ/fB6aPFr5uXDXfNh52/mieSPSYtNrkZEFbfLIfWIys3n8+X7OW7tQfZeDAN+IKQfT50beLPgqhHadEgmh8G9zWpLjqvND2aRI3iEQFCKfUv4ErABiQDt2mtD7m3VEI4yc0ymUXrd3Zse9f6SzmuyF/MmanmNXExJNxuqn8Ob4B67SHIymNks8G/YqD3eEf6CoBN3xS/9ukj8OnVZjl1Pyz6T+H9EU3M3AV+Vj3/FW+aWdu2/2TmSgiKLhx0XqoPj+wyVVpvdinYbNs1n7Wno1iy8xibDqWxKvEEqRm5tKoXwv2DWjCgdR26NYm0UmcX6UorwaFG8ogAAfxHa/0MgFLqAeBZ4B73FkkIJ788aurmH9oC4Q0L78vNLLxuyzWv9naB5G3w8XAY/Q20GALbfzF/4QMc/Ktw+0FIrAkIXj7Qaphpi/huvGN/5oniZWvcGzZ+DR8OM+tnUpyOP2mqkeq2NW0Xcx6HsIYQEGZ6Hjl59LeTfHP6T7wUxMcEM6h1Xa7p1pD+LcuYzEfUWB4RILTWzsM2g4HqO0mFqJkOWpPaZJ4sHiBWvQeXv2aWk9bAX1aPHnsPotwM87ryfWjSB354AOIvNNs63wAHljvOdfqIebXlwZ4/zHKRL/JiNn5deH3us4XXr5xsypyfx5kmA9iWG8u21Yf4a/dpXrcOuTz73wTFduO/w5vSv2UdooJr90xqwvCIAAGglPo3cCuQBgxyc3GEKKzrLeav731/Qn62mfbSlfVfQMpWx3pOhqnzBzi4GvyCTZfUTTPNtqObSr5mTnrJ+1xpcbHptdRqKIQ3gQUmP9IHG3PYcXQDu1NOs3rfSbTeA0BMiB8/Rd9BvQaNeH/IndQPDyzt7KIWqrIAoZSaB8S62PWU1vp7rfVTwFNKqSeB+4HnSjjPOGAcQJMmTSqruEIUZm+8/eVR89r37459yhs+vsIkujuTUrjKaN4EWPmuWT6TAl+PKXzeVe/DHXPNHMs7fim7HNe8Z+ZJ3vC1aWOw8/bHVrcd+sAq9p3M46e87vzN2vXC7C1EB/sRFxPMXf3i6RkfTet6oTSMDMTb6+Kz+S2IWsbjphxVSjUFftJadyjrWJlyVJy39KOmF5E9ADjLOAHTLoYLHzWT2Pz0sJkNrbwS7oDju0z20tIMfRH2Lja9iorqfjus+dCxbqXR0Fqjno8A4OvOH3Ld+tuZmH89//D+ikM6iivzX2WVz10A7L7voAxUE4WUd8pRjxgHoZRynsD2CmCbu8oiapG1n8F7g2DjDNf7d841X/Cz7obPrjm74AAwfJKjy2tRAeEwdr5Znvu0Izh0GV34uNaXmdxLrS6FJ5PYeTSdJ7/dQNd//cazuWPYaIvj3ytMo3hcK/M3VWCHESx79oqC60hwEOfKU9ogXlFKtcZ0c92H9GASlU1r+P4+s3x0s0mLseJd6HW3Gf18fHfJX+7lNW+CYyQzmGR3Ha6BpFWmF1PSquLvWfdZodWsxf8j4MQeOLGHgW+uJvF4Bt5eiqHt6tGy+T/Irj+BJbGhEHA91wDkjyfSy9t0O22YAHVan989iFrN46qYzoZUMdUAp5NN6ujgmKq9bupHvmuGAAAgAElEQVQBeMOqxew62rQbbP3RrNdpY8YVlOSBdYXGDxS4/Rc4cwy+dpHXCMyUnV7WQ/t/WhTujgocSHiCxqtfIcMnnDMEMje3C9NyhjDf/xFO60AejP+RPs2j6d+yDq1jJUW2OHflrWLylCcIUVu9ZtUuTighl1BFObQWYjubbqRbfig8O9rawn+1lxocut0KUfGu99XvAn5Bxbdf+Kgjv5GdFRySmw6n7r7ZvKVu4K8/8/jAD4Ly0lju24VVbZ/grvhotmaG0rhNd6bFtix+biEqkQQIUfMlrYb3B8PQf5tJbc6m+2hYI7jyLVj+tmmTaNq3+MA4O3twiGkN4Y1gt5kxLbnLvRxpdT+HtyZzJC2Lw2lZPGG9ZdKuerzsC4uDLubhweEwD2zd7+CiEZO4qODE0ltPuIcECOFeddo4ks+dr8xUk1IiujlMHWjq+Xvc5ZjfwJ4C+2z4+JmsrDmnTYDYOMMEiaKa9EFrzaG0LLYP+ZGdScncbQWInq8uL3Sor7diQ+jb9K6TzSWxQeSePsmXV1+G8guC6M/xan5R8fML4QYSIETVyE43X96RcYW352U7JrA5X5+PMg2/z540VUqH1sLKqaW/57LXTMI9Z6ENIL1IKrAfrFEFtnyIaAwXPQ3zzUC09e0eZdqZfix4fi7pWXkM81pJiMoEX/OW569oT/3wABpEBBIbHkBUkB9eXs65i25wLLYdfvb3LUQlkQAhqsZHl5t01kXbGk7uNT8j3yu8fcFLJodQ0z6Ft+fnwkfDTW+jDteYnkb5uabbqL1X0KJXzXvtWUqLCgh3DGb79Sl4bK8JUt+OhRN7IHlL8ffYezSN/pbk9Cy2nqrPAGCZbs+df7UlPDyf4Z0a0L5BGKPn3GSOfXQPeHkxJjCy3L8mITyJBAhRNQ6vd729wyjzl/6bXc2gsITb4ZuxJqXFmo/hke2mzv+PV6HpBaZ76oHl0P5qkx/pg2HQc6wZyGa38OXSyzJ8kpmpDUzajMBI0y30uk/ME81LToPmOl5X6K0DXv+Dfccz6KqOMcAfNjW9jekXD6ZTo3CT9hpgjnVwcHT5fz9CeCAJEKJq1G1n/jLXunBqaG8/k8wu/TDM/xecSXaknPDxg5TtsP1nWDLR/Ng590I6sAIa9Sz52s7zLYMjONjZy+PlbRqaBz0FC/7Nn5fMZm6SL3+9tYQfrEPzbZpHL2nN5dHh8C2M7VMfGkec9a9DiOpAAoSoGl1Hm1nPDq83M67Zk92t/8JxTH6OSYVtdybFpLqwJ7srSdIq6GJV6zTtB/uWFN7vHBxc0Fqzcu8J/tqfyo6j6QzYt51h2pebvj9FWIAPzes6RiIvfmyQeVLYtt1sWDcd2o4ofML7i8wwJ0Q1JQFClE9ulvlr36uE7CxnjpkBb7kZpivp5a+Dj79jf7dbodsYM8DsTAo8vs8xwY2zwKjC6yXNtVzU7IfgomfMwLvo5ib76rQhjv2hDSA0Fg79Vehtb7T+jK9fmc+htCwA6ocH8ETeUgJULu/dmsDA1nXw9faCl8MhO81RjdS4t0nSd8H9xcsS06J8ZRbCw0mAEGWz2eCVxjDsZdNt1NmehXB0C/z6ZOHtna6HeKeZ1r69G1L3OUYPb/8Zcs4Uv1ZEE0ha6bocfqFw7Yemt5IrXW6CiW3N8uBn4cLHzLwL9dqbwFQkOAD8siWFjq068feLW3FJu1jCg3whdS6kJXFxU6e5oVsMhiMbHOvB0fCci8l7hKhBJECIsuVnm+qfLBdVNZ9c6fo9TZx6H5065EhNrbzMk8R39xZ/T3gTqNcOSpoiISfdZFgtyYeXOpaDY+Cip9Ba8/PGI1zu9CSSpoP4KehKbsqczlcjY4joUiTjQEQT8+Os51gziY8QtYhHZHMVHs7eBuBcZWTXcqhJMVGUt49Jhrf288JfrNrm6H7aYSSMnAb+4WY9Kh5+f6HwebrfZnoo1e9sch395jRbWodRcMssx7p/GAA232C+WLGfJ77ZwID/LOS+Lwo/ORwbs5ibBnYDICK0nJlOf3sOFk8s+zghahB5ghBly88xr7/+E/rcZwaLfX2rCRhnjpn1opb+1+Q4OrkPvneaU7nDKNPrCExX1rCGkG39db/3j+LnyTgOQ543xza9wDElJ5gA45Rx9WB6Hl/GvMA3SREcmrWRiCBfOjYM554BzdGb+6D2L4MJaTQHiG8OjRKgYbfy/Q5KqvYSogY7rwChlFqqtXaRd0DUGDYbbPnesZ6yHXyDYNvs4sdGNDXtDOD4S9/Lu/Ax0S0gKxXSDpiurX++WfK1/ULNpDu7fofN38JVkwvtXrj7FD8dDOA/1nrDM1v4yasz1w1pwFVdGtI0OsjRqNz9B8h1avNQqvzBQYha6nyrmBpUSCmEZ9n0Daz/yiwfXF04FcXknuZLv+O1xd9nDw7O9iwsvP7HK2YsRFHRVqbSAU7jG8YtgOYXmXkacjPYcTSdZa3M/uM6lAcXK1akhjOhzU+cDGtD9qDnmP/wQP4+pBVxMcGO4ABmTMX5jGgOa1RxKUGEqCbKfIJQSv0P2Gj9bNJaO6fCrL6TSdRGe/6AT66A0Ppw3afQuEfxY7R2DCRrfzUkby1+TEg9GPm+yau06RuTnqK8AsJNd1O7wCjIPAHHd5r15oMh+zR0vp5k77osWpPEKJuZMW3opEVAZ54Lv5nbsz/ny/sH06ZhtBUIVpS/DOfiwRJGggtRg5Wnimkj0Am4GeiglDqFI2DIrCXulJtlegX5+JV8zKnDEBRl2gs2fm22pR+Gnx6Ce5YUPz7juHltOwJedEpfUactpFjBYsUU6HWPSVgX1Ry+K8cEgBPS4J1+JtGd/S/5wEi4/lOTp8nyzzVB7EoeQdLaYxxKSwLgRPCNDAjaw8QrOpPQNIomq5bAcmgbfBpUFU005C3NdaL2KU8V00bgb1rrAVrraKA/MAU4BbiYZV1UmX/Xg7d7lbzfZoOJbRxPBJ2ud+yr19HMqubs2E44aI0CLpKDqCA4gHlqeKWJedpod0Xh8zormrY6K808QSTcAXH9IfMkWTm5hQ75cf0h8m2a3s2jeWxYa2b/rR93PfUOrR+eyzXdGtEkOgiOWv1gz+bJRQhx1srzZ9EYYLJSagcmDdkcrfXPwM+VWjJRPqV9SdoHpdkblH2dZjtb/4X5ufdP+OtTuPh5eMtpPMD6L4ufr8dYWPWeGeAWXMc09PoFw9AXYcNXxY93rp7SGtL2w8lGEN2cA70nEH70Nq74IpnErC/4OPhNOgemsOHhoYXbDlzpPR72LoJ6HUo/TghxXsp8gtBa36O17gZMACKBj5RSy5RSLymlLlRKeZd+BlFpwhsXT01xaB283gbOHIe8IjOfOY8hsPvrU1jxjqNqyc4+sM1ZttX8lLLNjE628w0sfFzfB81r+mHzWq8DKEX6zT/zQ8uXGDrpD/p/dJROJ/+POrGN+XJcby5s04AIf1V2cABofampsgqpW/axQohzVu6KVa31NmAbMEkpFQgMAq4FJgJlTn4tKkGjBDi62bF+aC18eLnpzrlnAbS/BgY9baqNjm6BfUuhYQKcOuj48l7xDvgEljyNJpjG6JOJsGueY1s7pxHUedY4icHPQf9/mOWl/wVg8+h1zN6VzYI3FrHtSDqQSseG4Tx/RXsuaB5Ni7ohJij8kQzHd53vb0QIUYHOuuVNKVVfa30YU8Uk1UzulHXKTNkJsHEmfHOnY983d8LaT2HY/5kG1sTFZvtlr8J7RdoG8jJNgr2SnEw0VVEhsbBllqlecg4QQVHmOu2vJifPxp+7jzHQ2nXl+xvIw4cecZE8fHErusdF0qdZdPEnhZu+KjtrqxCiSintqk96aW9Q6gOt9R1KqZuB8cC/rTaJKpeQkKBXr17tjku7X24W/Ke5yWs07CWYEF7ysdEtzWhoV+MUXAlvbAaytb7cUdX06J4SJ8DJzbexfM9xpi7aw4akNNIyc+nmvZtbIzdyovc/GdYhlgYRgS7fK4SoekqpNVrrMmt+zqXvXqr1OhToB7yHPElULeenhV3zYM/Q0o+3jzEor0tegq9vMQnq7AHC27dgd1ZuPnuPneG7tQf5edNhDp7MxKYhIsiXIW3rcXmnWHrGX0KIv3QNFaI6O5f/wT5KqaeBA1prrZRykbNZVIicDDOILLQ+pO439fqXvQa75zuOOba95IyqAJf+B3551CyH1DMNzbkZhY/xDTLbOow0XVgb9TBjJKId8xpk48MPqw/wx44U5m09SlauDYA+zaK5qktDmkQFMaBVHeqGyWhjIWqKcwkQjwB9gT/P4xyiJLlZplE5pB7sXQg/PQwP74BZ95i5mDvfUDwVdWki40xQWTnVfOlPbFs4QHS8Fq5828y6Zn8qyTkNsR0BSOv1COErXuOSN5eReCKLiCBfhndqQP+WMXRrEknjqKDi1xRC1Ajn8uX+NnA38I5SyguYXrFFqsXOHIf/NDPL3v5mEBrA4XUmOIDp6eNqzIFd7/tguZXUbsDj0GqoybZarwOkJZnzAtz5m5nOs1FPMxLbJwZGfgBbv2ePrs8vC3YxZ9MRog8rBnhdQkigHx/e3pGBreqUryuqEKLaO5cAcUZrnW+l3HgeeBX4vWKLVUs5p7LuciPsmGuWrZQTACQuNYPjGnQrPkPa2AVm8Jo9QLS7yrzmZcOHw8zyoKdhwYuOfEhHTI6hk2dy+GnjaRZs68jvM03a7S6NI+hzyY1c1OEhbo8Orsg7FUJUA+cSICIAtNaPACilMko/XJSL1iZvkp23HwW5EO1ZRK+cbLqcgqmCcnb1u6a94rORZj0y3szOBoUn+hnwqPkBaHkJqdHdeGv2Fj5fsZ/M3HxCA3z4+5CWXJfQWHoeCVHLnUuAmKOUmgK8DuQhKb8rxvrphafhXDnVsWyfcCd5q2Me58Ql8M9D8JL16285FFa+53hP+mETdJQqNCfD2v0nWb7nBLuST7Mr9UE2bkpFs5dBresy5oI4LmwZI1VIQgigfOm+22utC4braq2nK6X2A48CClPNJM6VfRxKopVZNSgGMo65PnbZW47lnHTw8oGrp8KscbB/mQkGdnlZpsdSQFihU1z9tulbUC/MnxZ1Q7i1TxyjujeiQ8NSxlEIIWql8jxBfAp0A1BK3aW1fl9rvRRYqpQK0lpLFVN5nDps6v3tX+IZJ0wOo8k9TdZVe3VQScHBFS9fqNPKWlEmBffJfTDkOdMt1goO2Xn52CuZesVHMen6LlJ9JIQoU3nSfTvXN4wvsm9xBZYFpdQjSimtVFUl+a8iKTtM2u3lbzu2vRoPUweZL/JTSaZb6UVPF35f3XYw4ImSz+vl5Uig5+1rAsJVk00Su0YJbEhK5fW52+n6wm/8nt+V+RGj+GJsbwkOQohyKc8ThHMujqKV0+c7ZanjxEo1Bi4G9lfUOT3C0S3w7oVmeZfV2WvvIvPqPMfC4Gcd1Ux2yVvMT2k2f2de7Y3XwOnsPN79Yzf/m2+S313Svh5ZnadzSft6eHtJ+4IQonzKEyBilVK3AespHiAqcsrRScBjwPcVeE73W/EOWFNm0mqYY1QzQIshJlVG99tNRtQZt5XvnPU6wsm9ZrntcFg9jdOxPflk4S6mr9zPgRMmM+sl7evx7Ij2NJQnBiHEOShPgJiASed9O9BIKbUZk/Z7G1AhVUFKqSuAg1rr9WX1oFFKjQPGATRpchYjit2lxRD46xOzvOwtkwnVPpGPPX123baFp/cs6kqraiovC3rcCUveMMn3AJpfROL9h/j7V+tYdyCJhKaR3NCjCW1iQxnQqg4+3hX2kCeEqGXKDBBa66nO60qpRpg5qjsCi8p7IaXUPCDWxa6ngH9ikv+VySrPVDDZXMt7fbfxdhqDUFI21V8eK/0c34+Hh7ZAeEOzvu0n8AtiQ1IqE3/bwcLtKQT6evO/G7syorP0OhZCVIyzHgehtU4CkjjLDK5a6yGutiulOgLxgP3poRHwl1Kqp9b6iKv3VCtf31Ix53Ee7Ja0EoArtiwl2M+b2y6IY9yFzaTxWQhRodyeaE9rvREomDtSKZUIJGitz6K/pwfzCXRUBzkb/Bz88X+m2siuThsznacr3n4AJB47Q5y16eGLW3FrnzjCg3xdv0cIIc6D2wNEjRfTEg66mNTo9+dh5DTYPAu2zTbb6nWApheYoNL1Zlj7GfiFwKJXWbY/nelrdjN3yxFeoxdDgnbzt8Etq/ZehBC1iscFCK11nLvLcN5s+TD3Geh1N+SXMo1mdrqZEW7bbDN/9DXvmbENdsNeJnPG3QQCN36wlhB/Xy7v2IDm/WcSUD+sxNMKIURF8LgAUSMcWmsyqh5eD8klVBmBaVcIijbVR91uKRQccvNtfLZ8H3u3xHC5rQ33DmzB2P7NiAr2q4IbEEIICRAVw5YPc5+G1peCLQ8+vdpsTz/kGAPhio8/NOwGTycXpOCw2TTvLd7DB0v3cvRUNh0bXsrIqx7l8cYRVXAjQgjhIAHiXCyZZKb9HPOjWc/LNplUwxqYQGF3Yk/h9zXu7Zj4B0xbgxUYsnLzWX8glQk/bmHr4VP0aRbNhBHtGdYhVrKrCiHcQgLEuZg3wbxmnICgKDNfgy23cHBw5eYZ8J/mjl5NkU0BWJV4gjEfrCQjx8zH8M/L2jC2fzMJDEIIt5IAcT5ejYeO17nuxupKzmnHsV6+5Ndpxw9rk3h0xgbCAn15ZWQnBrepS7C/fCxCCPeTb6Jz0e8hU80EsPFrkxvJlYgmJltrWENoM9zMAhfRBC54gLxWl3PXx6tYuD2FZjHBPDuiHQNb13V9HiGEcAMJEOWhNUwdAL3Hmx5HA56AtiPgvYvM/qMbXb8v/Qg8sR98g8Hb+lX/fSMLtyfz6OQNpKRnc0ffeJ6+vC1ekmVVCOFhJECUR85p02V11t1mPbajGQldlvtWQoBjpjabTTN18R5e+WUbDcIDmHxTNy7tECvBQQjhkSRAlIdzwj2AIxvh81GFt4U1MhP/2EXGQ1R8weqKPcd5eMZ6kk5m0rZ+GO+PSZA03EIIjyYBojx8yhicdvnrEN0CPrnSsc03qGBxyc5j3PHRKhpEBPDS1R25vkdjmbhHCOHxZLKA8jhZQpruoS+a17xsaDbQMS4C4IbPAfhpw2HGfbqahpGBfHdfX27q1USCgxCiWpAAUR77l7nebh/38Os/reNWAKDbjmB5ahjPfr+JB79cS9v6YXx6Z08igiRNhhCi+pAqprIkrYafHy39mMAo82pN6PPm3oZMWrucAF8vLmpTl9eu60xYgKTkFkJULxIgyvL+4OLbnBukg2KgQRcAVoZdwkf5D7MsuyevjmrPiE4NCPTzrsLCCiFExZEAcS6i4k1ivlXvQcYxSDtIZk4+L/2yjcOBFzB7fF/poSSEqPakDaI0+U6ZWAOjTHI9gMTFMOwV+NtfZj1lK//4eh3rk1L5+5BWEhyEEDWCPEGUxtsXBj9rRk/HD4B3+zvt88EW2YxNw2Ywc0c+v2w6wgMXteDGnk3cV14hhKhAEiBKk5cNMa1g0zfQbJBje72OJKdnMf6zv1i9zzxlDGlbl4cubuWmggohRMWTAFGaxMXw1WizHBJbsHnhoG/422t/kJ6dx5394rlnQHPqhPqXcBIhhKieJEC4kpVmZonLPu3YtuGrgsUXZv2FUl58cVcvLmgR44YCCiFE5ZNGalfmPQ9vJcCBlY5tIXXhvpV8Hv0Ae0/ZeGd0dwkOQogaTZ4gXFk/HXIzYPnkgk3ax5/pewJ56mBv7h3YnL4SHIQQNZw8QbiibcU2Hc0P45+zNtKufhi3942r+jIJIUQVkwBRlNbgF+xY7347AAvpjo+X4tvxF1A3NMBNhRNCiKojAaKoBS9BxnHHepvh6L9v5K3kzlzVtSEBvpI6QwhRO0iAKOrP/xVeT97MIeqQlOVPl8YR7imTEEK4gQSIohr3MK8jp5nXg3/x565jALSsG+KmQgkhRNWTAOEsOx32LjLLAeHgFwphDViwPZnYsAC6N410b/mEEKIKSYBwluY0p/TCVyAnHX0ykXX7U2nXIAwfb/l1CSFqD/nGc7b6Q8eylxkikrFnBYfSshjQqo6bCiWEEO4hA+WcrXzXsXx8Fz+E3chbKV3o3jSS0b2buq9cQgjhBh7xBKGUmqCUOqiUWmf9XOaWgkQ4perOOMYTqVcR27IbM+/pg7eXckuRhBDCXTzpCWKS1vo1t139+G5I3Q8dr4MGXdin6pPxXT4juzVEKQkOQojaxyOeIDzCp1eZV78g8nrey8TEeJSCHnFR7i2XEEK4iScFiPuVUhuUUh8opUrsT6qUGqeUWq2UWp2SklJxV49qZl59g/l0+T6+X3eIewY0p4FMHyqEqKWqLEAopeYppTa5+LkSeAdoDnQBDgOvl3QerfVUrXWC1jqhTp0K7Fl0Otm8tr+a79YepHPjCB4f1qbizi+EENVMlbVBaK2HlOc4pdR7wOxKLk5xZ1Kg+21k1+/G5kO/Mu7CZlVeBCGE8CQeUcWklKrvtHo1sKnKLp6dDgteNgHCN5j5W5PJs2nJuySEqPU8pRfTq0qpLoAGEoG7q+SqRzbBlL6O9c438PPCI8SE+DO4bb0qKYIQQngqjwgQWutb3HLhJRMLrR4gljmbVnN9j8Yy7kEIUet5RBWT23j7FVpdv2ENufmauy9s7qYCCSGE56jlAcK30Kr/kbV4eynp2iqEENT6AGE9QfS5H4ATOV5EB/tJ9ZIQQlDbA4SX9QSxbykAgaf2Ehsu800LIQR4SCO127QdDgFhENUcZo1j+wkbcR2C3V0qIYTwCLU7QMT1Mz82G7PX7efdra34p4x/EEIIoLYHiIwTkJcNYfV5cm8nBrWL5va+ce4ulRBCeITa3QYx/0WY0o/0rFzSs/JIaBopqb2FEMJSe58g0pJg9TQADpzIBJAGaiGEcFI7nyAOrYNJ7QtWJy/YRYCvFwky94MQQhSonQHi2I6CRd3uShbtTOHqro1oKAPkhBCiQO0MEA26mddLXuLoJVNJz8qjbf1Q95ZJCCE8TO1sgwi0urLOfZqjeY0Bb1rWlQAhhBDOaucTxOlkCK4L2kbWgfUAtKgb4uZCCSGEZ6mdASJ5C5wxU4ym5Sh8vBTRwX5lvEkIIWqX2hkgcjMLFk/lKKKC/fCSBH1CCFFI7QwQeVkFi6k5EBPi78bCCCGEZ6qdAcLpCWLv8Sxa1pP2ByGEKKp2BginJ4gfMzswtF2sGwsjhBCeqXYGiN73FiyGkMUFzaPdWBghhPBMtTNA+IfCRU8D0D9oH5HSg0kIIYqpnQFi28+w5hMA6vtnu7kwQgjhmWpngNj6A9jymBr1CIuCL3F3aYQQwiPVzlQbuZngF8w83yH41c4QKYQQZaqdX495WeAbQGZuPoG+3u4ujRBCeKTaGyB8AsnKzSdAAoQQQrhUOwNErjxBCCFEWWpnG8Qts8CWS9ZrKwnwkwAhhBCu1M4A4RdEXr6N09l58gQhhBAlqJ1VTMDKxBNk5dpIaBrp7qIIIYRHqrUBYk3iSQAuaBHj5pIIIYRn8pgAoZT6m1Jqu1Jqs1Lq1cq+3i+bjtC5cQThgb6VfSkhhKiWPKINQik1CLgS6KS1zlZK1a3M653KymXrkVM8NKRVZV5GCCGqNU95grgXeEVrnQ2gtU6uzIttSkpDa+jcOKIyLyOEENWapwSIVkB/pdQKpdQfSqkeJR2olBqnlFqtlFqdkpJyThfbdCgNgE4Nw8/p/UIIURtUWRWTUmoe4GpmnqesckQCvYEewNdKqWZaa130YK31VGAqQEJCQrH95ZGakYuPlyIiSNofhBCiJFUWILTWQ0rap5S6F/jWCggrlVI2IAY4t0eEMmTk5BPk541SqjJOL4QQNYKnVDF9B1wEoJRqBfgBxyrrYqez8wjx94j2eSGE8Fie8i35AfCBUmoTkAOMcVW9VFEycvIIkgAhhBCl8ohvSa11DjC6qq53JjufYAkQQghRKk+pYqpSZ7LzCJYkfUIIUapa+Wd0z/gogiRACCFEqWplgHhsWBt3F0EIITxeraxiEkIIUTYJEEIIIVySACGEEMIlCRBCCCFckgAhhBDCJQkQQgghXJIAIYQQwiUJEEIIIVxSlZgTr9IppVKAfef49hgqMWOsh5J7rh3knmuH87nnplrrOmUdVK0DxPlQSq3WWie4uxxVSe65dpB7rh2q4p6likkIIYRLEiCEEEK4VJsDxFR3F8AN5J5rB7nn2qHS77nWtkEIIYQoXW1+ghBCCFEKCRBCCCFcqnUBQik1TCm1XSm1Syn1hLvLU1GUUo2VUguUUluVUpuVUg9a26OUUr8ppXZar5HWdqWUetP6PWxQSnVz7x2cO6WUt1JqrVJqtrUer5RaYd3zV0opP2u7v7W+y9of585ynyulVIRSaqZSapv1efep6Z+zUuoh69/1JqXUdKVUQE37nJVSHyilkpVSm5y2nfXnqpQaYx2/Uyk15nzKVKsChFLKG5gMXAq0A25USrVzb6kqTB7wsNa6LdAbuM+6tyeA37XWLYHfrXUwv4OW1s844J2qL3KFeRDY6rT+f8Ak655PAnda2+8ETmqtWwCTrOOqo/8Cc7TWbYDOmHuvsZ+zUqoh8ACQoLXuAHgDN1DzPuePgGFFtp3V56qUigKeA3oBPYHn7EHlnGita80P0Af41Wn9SeBJd5erku71e+BiYDtQ39pWH9huLb8L3Oh0fMFx1ekHaGT9x7kImA0ozOhSn6KfOfAr0Mda9rGOU+6+h7O83zBgb9Fy1+TPGWgIHACirM9tNnBJTfycgThg07l+rsCNwLtO2wsdd7Y/teoJAsc/NLska1uNYj1SdwVWAPW01ocBrNe61mE15XfxBvAYYLPWo4FUrXWete58XwX3bO1Ps46vTpoBKcCHVrXa+0qpYGrw56y1Pgi8BuwHDmM+tzXU7M/Z7mw/1wr9vKo4vJQAAARZSURBVGtbgFAuttWofr5KqRDgG+DvWutTpR3qYlu1+l0opYYDyVrrNc6bXRyqy7GvuvABugHvaK27AmdwVDu4Uu3v2aoiuRKIBxoAwZgqlqJq0udclpLusULvvbYFiCSgsdN6I+CQm8pS4ZRSvpjg8LnW+ltr81GlVH1rf30g2dpeE34XfYErlFKJwJeYaqY3gAillI91jPN9FdyztT8cOFGVBa4ASUCS1nqFtT4TEzBq8uc8BNirtU7RWucC3wIXULM/Z7uz/Vwr9POubQFiFdDS6v3gh2no+sHNZaoQSikFTAO2aq0nOu36AbD3ZBiDaZuwb7/V6g3RG0izP8pWF1rrJ7XWjbTWcZjPcr7W+mZgATDKOqzoPdt/F6Os46vVX5Za6yPAAaVUa2vTYGALNfhzxlQt9VZKBVn/zu33XGM/Zydn+7n+CgxVSkVaT15DrW3nxt2NMm5oBLoM2AHsBp5yd3kq8L76YR4lNwDrrJ/LMHWvvwM7rdco63iF6dG1G9iI6SHi9vs4j/sfCMy2lpsBK4FdwAzA39oeYK3vsvY3c3e5z/FeuwCrrc/6OyCypn/OwPPANmAT8CngX9M+Z2A6po0lF/MkcOe5fK7AHda97wJuP58ySaoNIYQQLtW2KiYhhBDlJAFCCCGESxIghBBCuCQBQgghhEsSIIQQQrgkAUKIKqaUWqiUqtTJ5oWoCBIghBBCuCQBQghAKRWslPpJKbXemnPgeqXUs0qpVdb6VGsUr/0JYJJSapE1H0MPpdS3Vv79F61j4qz5Gj628vXPVEoFubjuUKXUMqXUX0qpGVYuLZRSryiltljvfa1qfxtCGBIghDCGAYe01p21mXNgDvCW1rqHtR4IDHc6PkdrfSEwBZP+4D6gA3CbUsqeObQ1MFVr3Qk4BYx3vqBSKgZ4Ghiite6GGR39Dyun/9VAe+u9L1bOLQtROgkQQhgbgSFKqf9T6v/bu3uWOMIoiuP/A4JiY29hGrERK4MoIUW+gUgKt9JWSKF9CjsLP4FguZBaCSmS1gVLQUSwEGELC21EWGLjtbjPknWZxfeFyPk1+8zuDDNTLJc7A+fqc0RcAV/KRLJDMghwsmP/3Y7jjiLiPCJugFP+haU1I6JR1nUyDqXTLDm4qiHpgMza+UAWk7/AtqQFoPWqd2r2SAMP72L2/kXEiaRpMr9qQ9Jvsiv4GBFNSetkxk/bTfm87Vi3t9v/q+4cm+5tAX8iotZ9PZJmyFC6ReAbWaDM+sodhBkgaRRoRUSdHE7TnvF7Wd4LfO15cG9jkubKugbsdf2+D3ySNF6uYVjSRDnfSET8AlbJcD6zvnMHYZamgE1Jt2Sa5gowTz5COiOj4p/qGFiStEWmcd6bBx0RF5KWgR+SBsvX34FrYEfSENllrD3j3GYv5jRXszdQxr7+LC+4zf5LfsRkZmaV3EGYmVkldxBmZlbJBcLMzCq5QJiZWSUXCDMzq+QCYWZmle4AESluaXHEGZcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_10k')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Just test\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Radom small mag"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_folder = 'data/matrix_rand/'\n",
    "dataset = pd.read_csv('csv/rand_small_mag.csv')\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "X_1 = dataset.iloc[:,1].values\n",
    "Y_1 = dataset.iloc[:,2].values\n",
    "\n",
    "W = np.random.permutation(np.c_[X_1.reshape(len(X_1), -1), Y_1.reshape(len(Y_1), -1)])\n",
    "\n",
    "X = W[:, :X_1.size//len(X_1)].reshape(X_1.shape)\n",
    "Y = W[:, X_1.size//len(X_1):].reshape(Y_1.shape)\n",
    "\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in X[0:100]:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in Y[0:100]:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100,)\n"
     ]
    }
   ],
   "source": [
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/my_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 10s 99ms/step\n",
      "Score (RMSE): 0.29153563767115553\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_small_mag')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## All mag"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_folder = 'data/matrix_rand/'\n",
    "dataset = pd.read_csv('csv/rand_all_mag.csv')\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "X_1 = dataset.iloc[:,1].values\n",
    "Y_1 = dataset.iloc[:,2].values\n",
    "\n",
    "W = np.random.permutation(np.c_[X_1.reshape(len(X_1), -1), Y_1.reshape(len(Y_1), -1)])\n",
    "\n",
    "X = W[:, :X_1.size//len(X_1)].reshape(X_1.shape)\n",
    "Y = W[:, X_1.size//len(X_1):].reshape(Y_1.shape)\n",
    "\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in X:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in Y:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[-2.7323301387724515 -3.949807386562641 -4.292664872991386\n",
      " -2.9770328888102267 -1.7457435968077153 -2.1743015333328657\n",
      " -1.7463395765861107 -4.48284342479733 -3.8789251049547406\n",
      " -2.9011617442591517 -4.029131327064219 -2.546208553139333\n",
      " -3.4367701587954427 -4.028309443751435 -4.233514899865487\n",
      " -1.8639447974808736 -3.608808193735397 -2.1922059629386936\n",
      " -2.668716719102872 -4.723793491647036 -2.6078269500779587\n",
      " -4.408438420299164 -2.5383586518093875 -2.9917362239575285\n",
      " -2.518857602411294 -2.8886030770821534 -2.8661795072610015\n",
      " -2.88752923795654 -3.811850125811977 -3.0791966050603468\n",
      " -2.019767097772924 -3.404261472410024 -4.813049749565659\n",
      " -3.883985019396383 -1.8893471918928475 -3.533506254305937\n",
      " -4.557988377938983 -3.087884386681636 -2.832073013802593\n",
      " -3.93141837341244 -2.9535203960424496 -3.225592771118229\n",
      " -2.638527918766078 -1.8400322853686035 -3.470560457275978\n",
      " -3.401607264531941 -4.497769743676696 -3.3046175108841496\n",
      " -3.3506899269939248 -2.611605624415814 -3.8148155921340647\n",
      " -3.3508326040403054 -1.9004157906380104 -3.6797027522001686\n",
      " -3.6394851466346068 -4.301576687942552 -2.020358786191103\n",
      " -3.773036704712641 -2.6433326223533444 -3.4661140521550067\n",
      " -3.1138412821686985 -3.475742673962105 -3.011183348512591\n",
      " -4.196716353917154 -2.8492611499794123 -3.851823287905821\n",
      " -2.7331715778175525 -2.744110147755922 -4.171161947641701\n",
      " -3.634166244360365 -3.287838013871164 -4.63797441272557\n",
      " -3.729482128803492 -5.1456303194315485 -3.309289026338901\n",
      " -2.6065788510063066 -3.4279192663194085 -2.6605630228140704\n",
      " -4.044907040226935 -2.2820090540755418 -1.9706557795886808\n",
      " -3.9838541992443233 -3.8919713009351002 -2.960108296797185\n",
      " -3.0330799990054 -3.2108417491147914 -7.474405051140968\n",
      " -4.290755049880766 -3.094201942215861 -5.341484490055611\n",
      " -2.2469668993547325 -3.297426164228344 -2.8008871132047317\n",
      " -3.5440058529547636 -3.110666205914943 -3.900361204743749\n",
      " -2.1938490570674443 -3.1233412034999057 -3.401472910088759\n",
      " -2.2145678142981806]\n"
     ]
    }
   ],
   "source": [
    "print(test_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/my_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 10s 103ms/step\n",
      "Score (RMSE): 2.262712124310257\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xd8VFX+//HXJ70nJKFDSOhFqaEJoiiy6NrL167oImvZr7q7ftddy+q6+nOLqy6uZUWxrCwWLNgLiIgUJUgXkA4hQEJC6iSZzMz5/XEmGMKkAElmkvk8H488YOa2c+feue97zr1zrhhjUEoppWoL8XcBlFJKBSYNCKWUUj5pQCillPJJA0IppZRPGhBKKaV80oBQSinlkwaEUkopnzQgVMASkZ0iUi4ipSKyX0ReFpG4GsNfFhEjIufXmu5J7/tTva8jROQfIpLtndcOEXmijuVU//3rBMt+vYisFJFi73L/JiJhdYzbV0TmiUieiBSIyGci0q/WOD1F5EMRKRGRgyLyN+/7kSLyoojs8g5bJSJn15ju6lrr5fB+NiO8wyeKyEIRKRKRnSeyzqrt0YBQge48Y0wcMBQYBvyh1vAfgeurX3gPwpcB22qM8wcgExgFxAMTgVW+llPj71cnWO4Y4E4gFRgNnAncVce4ScD7QD+gI/AdMK/GOkUAXwBfAp2AbsBr3sFhwB7gNCARuB94U0TSAYwxs2uuF3ArsB343jt9GTAL+L8TXF/VBvk8o1Eq0Bhj9ovIZ9igqOkD4BoRaWeMOQRMAdZig6DaSOBdY0yO9/VO719zlvfZGi/3ishsbDD5Gvc7bCgA4K3d3CciKcaYfGAqkGOMebzGZGu905YBD9Z4/0MR2QGMwPc6Xg+8arxdKFQvW0QmHdMKqqCgNQjVKohIN+BsYGutQRXYs+8rvK+vA16tNc5y4DcicquInCwicgLluEpECuv5S6tj0gnAhkYuZgKw3xsOAGOAnSLyibd56SsRObmO8nUE+vpaloj08M679uejlE8aECrQvSciJdhmlFzgAR/jvApcJyKJ2KaW92oNfxT4K3A1kIU9o7++1jjv1TrQ3+SrMMaY/xpjkur52117GhG5AdvE9VhDK+sNwqeB39R4uxs2AGcAXYCPgHnepqea04YDs4FXjDGbfMz+OmCxMWZHQ+VQCjQgVOC70BgTD5wO9Me26R/BGPMN0B64D/jQGFNea7jbGPO0MWYctr3/EWCWiAyotZyaB/qZTVF4EbkQ+AtwtjHmYAPjtgc+B54xxsypMagc+MYY84kxxokNmhRgQI1pQ4D/AE6grusn1wGvHO+6qOCjAaFaBWPMIuBl6j4Lfw34LQ00nxhjyo0xTwOHgIHHWg4fdwXV/kurMe4UYCb2Avi6BubbDhsO7xtjHqk1eC1QZ7fL3iazF7EXuC8xxlT5GGcctvYxt5GrqpQGhGpVngTOEpHaF6rBNr+cBXxde4CI3Ckip4tItIiEeZuX4jn6TqYG1b4ryMffbu8yz8A291zivRBcJxFJAD4Dlhhjfu9jlNeAMSIySURCsXdHHQQ2eoc/i61NnFe79lTD9cDbxpiSWssOEZEoINy+lKjaTVcqeGlAqFbDGJOHrSHc72NYgTFmQfXdObWUA/8A9mMPrLdhD9zba4zzQa2awLsnWNz7sbedflxjnp9UD/RecL7H+/Ii7J1WN/iqjRhjNgPXAM9haz4XAOcbY5zeC8+/xN7dtb/GtFfXWFYU8D/4bl6a4P18PgbSvP///ATXXbURog8MUkop5YvWIJRSSvmkAaGUUsonDQillFI+aUAopZTyqVX3xZSammrS09P9XQyllGpVVq5cedAY076h8QIiIETkMmyHYwOAUcaYrMZMl56eTlZWo0ZVSinlJSK7GjNeoDQxrQcuxsePnJRSSvlHQNQgjDEbAU6gk02llFJNLFBqEI0mItNFJEtEsvLy8vxdHKWUarNarAYhIvOxT8Oq7V5jzDwf7/tkjHkeeB4gMzNTfwaulFLNpMUCwhijT6xSSqlWpNU1MSmllGoZAREQInKRiGQDY4GPvM8eVkop5UeBchfTu8CJdq98YpwO8LggKqHhcY0Bf9xxVVEEB7dAUTYU50ByT+g35fjmlb8NQiMgoQuEhB7btO4qQCC0EbuPMVCwHSpLoIuvxzjUo7IEslfAgR+gz1nQvt+RwyuK7TYLj4awqBPbJsX7IDIOIuOPfx4eD5Tuh7hOEFLj3Mvtgr1ZkNQDEjofOU3OaijeC11HQLyvS3Te+XpcEHaMj2koO2i3bWTikeU5VjX390M7oaocUvse+37TFKoq7PcgNrXu5btdUFls99OYZAgNB48bJMQ/39uGeNz271i3bwsIiIBocS4n/PtUe1AJjwZHPuRvhTP/CON/DY4CWPUaDLkC4jocOe2SGfDlw/bg3HkwdBoMHQdCj/F2A+f9CDsXw4EN0HEQ9D4T2qUfOQ+3C7bOh7Vv2GWPmAonXWyH1RU+S/8Fn9975HvhMXD7Kt8Hlp3fwFtTISLWfpk7DIS4jjDievve2jdh0V8gJMyGBAKuSrhznV2PqnL72dRUXbYdi+DNqdDjFMg4FXpPgg41nt5ZVQEb3oXNH8Hu5VCWB31+Ble/aYd/+GsbcFUOu9zUvtB3CvSZBCUH4MuHYN9aOLAejMdO066HDYiC7bDwUXvALajxOIe7d0J0u6M/h2p7V9p1LtlnDxzuKijcDTd9aYPh83thyxcw7FrIvBFSevneDuWF8P2rsPEDmPQgpI+D7V/B/D9B3ia7TrHtoc9kOOM++9n+9zLY9iVIKPQ/BwZfAWljITbFHuxev8rOO6ErtO8PiV1hwu8gqTssexq++gs4y+z6dzzJfl6n/sYeIJ1ldj+oLqu7yh4QAd6ZDtsW2ANjTArEdrAhfeEzdvj6tyHjdFuOmjxuyP3BbrtdS+2/v1xkvwvfzYRl/4KIOPsZxaTa9b3wWRtCXz9mp4nrCPEd7UmIx2W/WwCbP4F2GXYddiyyn+XIX0D6+Lq3Hdjv5OMDweV9HlJMit2nRv4CumXC1gXw6e/tyVOV46fppn0J3UbA6v/CB7fbcrfv/9O2O7x/fA+bP4aCHXBoB1SWwpT/Z/ftw5+Lx847Mu6n99xVsOIF+OYJ+11K7G6329CroddEe+Kx9nU4uNUeY4zHln38nZA2BjZ9BG/dAO5KCIu2J6idh8DP/wFJafhbcAaEcdsd1FVh/1L6wKCLIWOCHb59IXxxPyx4CAZfDmNvsyEAMOBcyFllv5g7vrYHeQTu3WeHL3wYfpgH4bFQVWbf638uXDHb/v/Lh+H7/9gzzdj2Nmg8bjtsx2L45nE4/ym7I31yN4y7w+5IaWNg4n3Q6SRI7GanrSg+Mhw2fgCbP4Vzn7Bf3AHn2QNQ7ibYthA8VdC+r93ph19npy3cBUV77UEkLJLDT7Z841q7jik97ZfFkQ/dRsKkByDjNBhyuT0wbvkMPr8PUvvBpbNs+eZ4hyWmQa8zbdl7nGLnW1VuDyChEfbA5qmCNa/bM/c+k2wo/fi5/bxPvctO22GAXR+AtW/ZAO46wn4JI2LtPCPibICt/q/dZqFh9sDWvp8Njrwf7cEosbsNwJBw+9mXF9gv/Jjb7Hb87t+w/GmI7wxDr/rpwLb0X/bAsXqO3a5dhkFEjB0WFm3nMfx6G2TZWbDlc5jyqB0+6pe2rPvW2BOPjR/AmFvt8PTxcONndp/au9IeRPavs+sOkNwLTr4MohLtSceuJfDDezDBO/yj38L+9XYfTeoO826Di2dC91Ew+pf2BKX8kA3p0ryfwgNs8Hz9D5j6oT3TBltDnX2prSkAxHeBHmPtvgAwcpoNqZzv7ThleVCy/6caivHYzzRvE5QesPt2x5Ps5+jxwAd32n0/OtmOF5UEfX925P7bebA9iBpj988Jd0FkAoyaZrdlRJz9jDd/BL3OsAERk2KDp89kO25Ugt3Hqg+ynU6yn2llMWz8EF4+BwZeAJe8CLkbYeZEG+CJ3SA5w37eCd3stCtesMFXmus9dvSzn+vpf7DBvPhx6NDfBnzhHtjzLfScaKctPwTzH7SBmdzLfseKsu3JGNjj0OjptpZXWWRPQLZ/ZcsC9jPL32Jr0gXb7X6bPu6nY1Uza9UPDMrMzDTN1tVG3o+wYqb9Qlc57Aa/9t2jzypL8+yGSxvtnW6z3THbpdsv27Yv7QHl1N/a4U+PtsOGXWu/GDW/sOvmwvu32zMRj/exwuc+YWsy9dn0sf2CvHqBrbVM/fDos393lW1yiO/UcDXbGFj+LGTNsgeGyDj7pex3Npz2uyPHLc6xZ0EbP4DL/2O/WNsX2WEZExpXpTfGfmHCoxoet+YZcm1bF8BrF9svfkwqZL1oA/ash+z8JbThZrGivXZ9sr+zwVS93R7yBtRJl8DYW+1ZXn08Ht/NOlUVsPULezDpOrz+edTFVekNc2xgLfkn5HmfPtouHS572QZYQ7YugDlX2AP4dfPsQbUoG964BkbeZMMrKe34m2V8NY2V5sG6N20Y9jvHnjxVb/elT9kTKFfFT+On9Iabvzl6fwa7Lxhz7E0zToetBRXsgIuetfNY+4Ytj68m5h8/h43z7EE+LAp2L7PNnr/eYPen4hx7QuHrc/K4wVlqvxeNVb2PGwMvTLK1ZbAnccYDp90NE++pfx4NEJGVxpjMBsfTgGiAo8AeaAr3wM8eObE26sYo2A4f/sYelH/2qD0rrM+eFfDipJ/OfqYtgLgG++Bqu5Y9A5/9wX6ZRt8CZ9xraxknqrLUfmmrD8yBxBjblLR/nW0eO5aD0eZPbCCERcFvN9v9zl/X2MAeUKtrPJUl0HloQLbN43Y17hrciSjcbWuunYfY2ntKb+8Ac8LXfzQggsknd9vaxw0fH30hNxhtmW8vYh7rRfFgtf5tmPsLmPxnOOV//V0a1QI0IIKNyxmYZ1qqdcjfZpu9GtPMp1q9xgZEcF6kbos0HNSJSOnl7xKoABQQP5RTSikVeDQglFJK+aQBoZRSyicNCKWUUj5pQCillPJJA0IppZRPGhBKKaV80oBQSinlkwaEUkopn/SX1EopFeBcbg8b95VQ5fFgDBhjGNI9ifDQ5j3H14BQSqkAVlbpYtorWSzbnn/E+yvvm0RKXPP2LqwBoZRSAaq4ooobXlrB6j2F3PfzAfTuEIeIIEB8VB3PRWlCGhBKKRWA9hQ4uHX292zaX8y/rhzG2Sd3bniiJqYBoZRSASCnsJyl2/JZti2fb3fkk32onIiwEP597QjO6N/RL2XSgFBKqRa2cV8x327PZ19RBTlFFazLLmRnvgOA5NgIRqUnc+O4DCb270BGahM8EfE4BURAiMjfgfMAJ7ANuMEYU+jfUimlVNNyewzPLNzKkwu24PYYIsJC6JwYRZ8OcVw7Np1xvVPo2yGekBA/PfK1loAICOAL4A/GGJeI/BX4A3C3n8uklFJNJqewnDvfWM13Owo4f0gX7jlnAB0TIhF/Pf+7EQIiIIwxn9d4uRy41F9lUUqppnSwtJLnv97Of5btIkTgH5cN4eLhXQM6GKoFREDUciPwhr8LoZRSx8oYw8tLd7ItrxSPgQqnm0/W76fS5eb8IV34zVn9SEuJ8XcxG63FAkJE5gOdfAy61xgzzzvOvYALmF3PfKYD0wHS0tKaoaRKKXV83l21lz998ANJMeGEhQgiwuRBHfnfM/rQu0Ocv4t3zMQY4+8yACAi1wM3A2caYxyNmSYzM9NkZWU1b8GUUqoR8ksrmfT4IjJSY5l78ykBc6HZFxFZaYzJbGi8gGhiEpEp2IvSpzU2HJRSKpA88tFGSipcPHrx4IAOh2MRKL25/guIB74QkdUi8py/C6SUUo21eEse76zay82n9aJfp3h/F6fJBEQNwhjT299lUEqp+hhjyC2ppKLKjctjKHe62ZZXyg/7inlv1V4yUmP51Rlt61AWEAGhlFKBxhjDhpxivtyUy6rdh1iTXURBmfOo8cJDhb4d43nkopOJCg/1Q0mbjwaEUkp5GWPYuK+E99fk8PG6fewucCACfTrEMWlABwZ1SSQuMoywUCEiNIT01Fh6tY8jIixQWuublgaEUirolTvdvLx0J++t2svmAyWEhQin9E7ltom9OGtgJ5JjI/xdRL/QgFBKBb3HPt/Mi9/sYESPdvz5wpP4+cmdgzYUatKAUEoFtXKnm7ey9nDekC48deUwfxcnoLTNhjOllGqkD9bkUFzh4toxPfxdlICjAaGUClrGGF5dvpO+HeMYmd7O38UJOBoQSqmgtSa7iPV7i7l2TI9W0btqS9OAUEoFrdeW7yImIpQLh3X1d1ECkgaEUiooFTqcfLAmh4uGdSU+KtzfxQlIeheTUqpNM8ZQXOHiUJmT/DInuwvKWJddzLc78ql0ebhGL07XSQNCKdVmrcsu4r5561mz58hH3EeGhTCwSwL3njOAAZ0T/FS6wKcBoZRq1YwxON0eKpwenG4PHmOorPIwa8kOXl22k5S4SP7vZ/3onBhFu9gIuiRG06t9LGGh2sLeEA0IpVSr4PYYHE4XxRUusnYWsHjLQZZsPciB4go8Pp57JgLXjunBbyf3IzFarzEcDw0IpVTA2J3vYG9hOQVlTvJKKtiSW8rm/SVsyS2lqLzqiHETo8MZ3zuVnu1jiQoPJTIshMiwEEJDQggNgZO6JjKoS6Kf1qRt0IBQSvldocPJQx/+wDvf7z3i/fioMPp3iufcwZ1JjYskLjKMmMhQBnVJ5OSuiYS2kSe3BSoNCKVUi3I4XRwsceI2Bo/3mQsPffADhQ4nt5zei/G9U0mJiyAlNpLUuAj9AZsfaUAopZpVldvDa8t38VZWNnsLy49qKgI4qWsCr944ioFd9I6iQKIBoZRqNou35PHQBz+wJbeUET3acf6QLnRKjKJ9fCQRoSGIQGxEGKf3a693FQUgDQilVJMxxrA1t5RP1+/n0w372ZBTTFpyDDOvy2TSgA7aXNTKaEAopY5LWaXLe5dRMZv2l7DZ+5fvfW7z8LQkHjhvIFeOSmtzz2oOFhoQSqmjGGMoKq8i+1A5JRUujDG4jWFnvoNVuw+xench2w+WHR4/OjyUvp3imTSgIyd3S+SsgR3pmBDlxzVQTUEDQil12LJt+cxYsIX1e4soqXT5HCc1LoJhae24cFhX+nWKp3+neLq3iyFEbzltcwIiIETkz8AFgAfIBaYaY3L8Wyql2jZjDA6n+3BN4ZmvtvLV5jw6J0Zx8fCudE+OoVu7GBKiwwgVITRE6JgQRbd20XotIUgEREAAfzfG3A8gIrcDfwRu9m+RlGobqtwedhc42J5Xxva8UjYfKOHHAyVszS2lospzeLzE6HDuOac/141N12sGCgiQgDDGFNd4GQv46FlFKdUQl9vDpv0lrNpTyPrsItbtLeLHAyW4anRW1DEhkr4d47l6dA86xEeSGB1OUkw4Y3ulap9F6ggBERAAIvIIcB1QBEysZ7zpwHSAtLS0limcUgGsyu3h/dU5zF2ZzZrsQhxONwDtYsI5qWsi0/r2pE+HOHq2j6VnahyJMRoCqnHEmJY5WReR+UAnH4PuNcbMqzHeH4AoY8wDDc0zMzPTZGVlNWEplQp8bo/hkMPJwdJKlm/LZ+biHewtLKdX+1hO7dOeYWlJDE9rp9cKVJ1EZKUxJrOh8VqsBmGMmdTIUf8LfAQ0GBBKtRUut4eDpU5KKqqodHmoqHJzoLiSzQdK2HKghL2F5RSXV1Hk/avZvXVmj3b8+cJBTOynP0RTTSsgmphEpI8xZov35fnAJn+WR6mWUFrp4vdvr2X59nzyy5z4qsyHCPRIiSUtOYb0lFgSo8NJjA4nNS6C1PhI0lNiOamrdmmtmkdABATwFxHph73NdRd6B5Nq4/JKKrnx5RX8sK+Yi4Z1pWtSNB0SIkmICicqPJSo8BCSYyPo1T5O7yhSfhMQAWGMucTfZVCqpew4WMbUl77jQHEFM68bwRn9O/q7SEr5FBABoVRbVlRexbNfbWPV7kNsP1hGXkkl7WLCmXPTGIaltfN38ZSqkwaEUs3EGMMn6/fzwPsbKChzMrR7Eqf3bU/P9nGcO7gz3ZNj/F1EpeqlAaHUcfB4DBUuN6WVLsoq3RQ6nBQ6qjjkcJJbUklucSUb9xWzbHs+g7ok8NLUkXoxWbU6GhBKNZIxhu93H+Llpbv4bP1+nG5PnePGRITSKSGKe87pz43jMvRhOKpV0oBQyge3x/D+mr18tv4Abu/9p9mHytm4r5j4qDAuH9mdbu2iiYkMIzYilKSYcJJiImgXE0H7+EjiIvWrpVo/3YtV0HK5PTjdHpwuD1Xun36EkLWzgMe/+JEtuaV0TYomwds/UWxEKA9feBIXDetKrAaACgK6l6s2r6LKTV5JJQdLK9m0v4Rvt+fz3Y4Ccooq6pymV/tYnr5qOGef1Emfc6CClgaEalOqn3GQtesQCzYeYMHGXPYWlh8xTmpcJKN7JnN5h3giw0OICA0hPFTA201F+7gIJg3oqNcNVNDTgFCt0o8HSnhy/o/sKSinrNJFaaULh9NNmdN1uMuK6PBQxvdJ5arRabSPiyQ1PoL0lFgyUmO1zyKlGkEDQrUqReVVPDn/R15dtou4yDCGpyXRIyWGuMgwYr0XjKMjwujfKZ6xvVK0mwqlToAGhApo+4sqeG7RNnbll7GvqII9BQ4cVW6uGpXGXZP70S42wt9FVKrN0oBQAckYw1srs/nzhz9Q6fLQp0Mc3dpFMzI9mctHdtcfnSnVAjQgVEA5VOZk2fZ8Xl+xh69/zGNURjJ/v3QwPVJi/V00pYKOBoTyC5fbw5ebclm4OY/iiipKK1wcKK5g84ESjIH4yDD+dP4grh3TQ28zVcpPNCBUiyooczJ7+S7mfLebnKIKEqLCSI2PJD4yjI4JUfz85M6c0juVwd0SCdfbTJXyKw0I1SIKHU5eWLyDl5bsoMzp5tQ+qfzxvEFMGtBBf2+gVIDSgFDN6kBxBa8s3cl/lu2ipNLFzwd35s4z+9CnY7y/i6aUaoAGhGoW67KLeGnpDj5Yk4PLY5gyqBN3TOpD/04J/i6aUqqRNCDUCXG5PRxyVFHpclPp8rB8ez5zvtvN+r3FxESEcvXoHtwwLl3vQlKqFdKAUMdl/d4i5q7MZt7qvRxyVB0xrH+neB66YBAXDO1KorcnVKVU66MBoRpU3QHe97sPsWhzHot+zGNLbikRoSGcNagjYzKSiQwPJTIshIzUWE7umqh9HSnVBmhABLn80kq+2XqQfUUV5JdWkl/q5GCZk4KySg6VVVFcUUVZpQuPtwO8iNAQRma049qxPTh/SBeSYrSrC6XaqoAKCBG5C/g70N4Yc9Df5WmLyipdbNpfwuo9hXy+YT8rdhYcPvhHhoWQGhdJSlwEqXGR9O0QT0J0+OGO8Pp3jmdMRgrREdoBnlLBIGACQkS6A2cBu/1dlrbEGMMP+4p5f3UO8zceYPvBssPdYfftGMdtE3tz1sCO9GofR0xEqDYNKaUOC5iAAJ4AfgfM83dB2gKH08WbK/bw2re72ZpbSliIcErvVM4b0oVBXRIZ1CWBLknR/i6mUiqABURAiMj5wF5jzBo9gz02xRVVbM0tZX9RBR5j8BjYvL+Y15bvpqi8imFpSTx84Umcc3JnkrVrbKXUMTihgBCRJcaYcY0cdz7Qycege4F7gMmNnM90YDpAWlpaI0va9sz+dhczFmzhQHHlUcNEYPLAjkyf0JMRPZL9UDqlVFsgprpB+ngmFtlhjMk4oQKInAwsABzet7oBOcAoY8z++qbNzMw0WVlZJ7L4VmnZtnyufmE5w9PaceaAjvTuEEfXpGjCQwURISE6jA7xUf4uplIqQInISmNMZkPjNViDEJGngHXev/XGmJIag48/XapnYMw6oEON5e0EMvUuJt8OFFfwv3O+JyM1lpdvHEVcZEC0Eiql2qDGHF3WAYOBq4GTRKSYnwJDe1xrQVVuD7fN/h6H082cm8ZoOCilmlVjA2Km8bZFiUg3bGCcDHzW1AUyxqQ39TxbM7fHsCGniCVb85m/8QArdx1ixpXDtDdUpVSza0xAXA88LSI/Ap8CnxpjPgY+btaSBbmtuaXMXZnNO99nk1tiL0T37RjHA+cN5PwhXfxcOqVUMGgwIIwxNwOISH/gbOBlEUkEFmIDY4kxxt2spQwiP+QU8/BHP7B0Wz6hIcLEfu05b0gXTumVSvv4SH8XTykVRBrdiG2M2QRsAp4QkWhgInAZ8DjQ4NVwdTS3x1DpclPlNpRWunjuq23M/nYXidHh/P7s/lw8vKvejaSU8ptjvsopIp2NMfuwTUzazHQMSitd3PfuOjbuK+FgaSUFDic17zIODRGuG5vOryf1JTFGu8lWSvnX8dwG8whwo4hcDdwKPOK9JqHqUely88v/ZLF8ewET+3VgRHo7UuMiiYkIJSxECA8NYWyvFPrqxWelVIA4noAo9P47GRgPzERrEvVyewy/eWMNS7bm84/LhnDJiG7+LpJSSjUo5DimCROR+4A93ltfy5q4TG2Kx2N44P31fLRuH/f9fICGg1Kq1TieGsRdwDhg6QnMIygUOpz85s01fLkpl1+e1pNpp/b0d5GUUqrRjufg/gzwS+BZEQkB5jRtkdqG1XsKuW329+SWVPDQBYO4dkwPfxdJKaWOyfEERJkxxu3tcuNPwN+wne0FtT0FDmYt2cGPB0rYnlfGvqIKuiZFM/fmUxjSPcnfxVNKqWN2PAGRBGCMuQtARBz1j962OV0eZi7ezowFWwDo3zmBsT1T6N0xjqtGpekzm5VSrdbxBMSnIvIc8A/ABQRtvw8bcoq44/XVbM0t5eyTOvHH8wbSOVGf0qaUahsa0933IGPMhurXxpg5IrIb+D9AsM1MQeejtfu46601JEaHM2tqJmf07+jvIimlVJNqTA3iP8BwABGZZox5wRizBFgiIjHGmKBqYvJ4DE8u2MKMBVsYnpbEc9eO0O4wlFJtUmN+B1HzIdG31hq2uAnLEvAqqtzc/voqZizYwmUjujFn+hgNB6VUm9WYGkTNp8ZJrWHH80O7VqnQ4WT6qyv5bmcBfzi7P9Mn9ESk9scEpgH1AAATjUlEQVShlFJtR2MCopOITAXWcHRAnPAjR1uDPQUOpr70HXsKynnqymGcp89jUEoFgcYExIPY7rxvALqJyAZst9+bgNTmK1pg2LS/mOte/I6KKjf/+cUoRvdM8XeRlFKqRTTmgUHP13xd65GjXzdTuQJC1s4Cbnx5BdERobx18yn066Q9rSqlgscx/w7CGJMNZNPGe3Bdvj2f62d9R9ekaF79xSi6tYvxd5GUUqpFaUd7dfj3om0kx0bw1s1jSYnTR30qpYJP0NyFdCzKnW6WbsvnZ4M6aTgopYKWBoQPy7YfpNLl4cwBHfxdFKWU8puACAgReVBE9orIau/fOf4sz5ebcomJCGVURrI/i6GUUn4VSNcgnjDGPObvQhhj+HJjLuN7pxIZFurv4iillN8ERA0ikGw+UEJOUQVn9NfmJaVUcAukgPiViKwVkVki0q6ukURkuohkiUhWXl5ekxfiy025AEzUgFBKBbkWCwgRmS8i6338XQA8C/QChgL7sM+a8MkY87wxJtMYk9m+ffsmL+fCTbkM6pJAxwTthE8pFdxa7BqEMWZSY8YTkZnAh81cHJ8KHU5W7jrEbRN7+2PxSikVUAKiiUlEOtd4eRGw3h/lWPRjHh6DXn9QSikC5y6mv4nIUGzvsDuBX/qjEB+s2UdKbARDuiX5Y/FKKRVQAiIgjDHX+rsMP+QUM3/jAe6c1IeQEH3Og1JKBUQTUyCYsWAL8VFh3DAuw99FUUqpgKABga09fLphPzeOyyAxOtzfxVFKqYCgAYG39hAZxo1ae1BKqcOCPiA27rO1hxvGZ5AYo7UHpZSqFvQB8cLiHcRFhvELrT0opdQRgj4g9hY6GNg5QWsPSilVS9AHhMPpJiZSe21VSqnaNCCcbmIiNCCUUqo2DYhKFzERAfF7QaWUCihBHxBlTjexWoNQSqmjBH1AlDvdRGsNQimljhLUAeF0eXC6PVqDUEopH4I6IMqdbgBiIrUGoZRStQV1QJQ5XQBag1BKKR+COiAc3hpEtAaEUkodJcgDoroGoU1MSilVW1AHRFll9TUIrUEopVRtQR0Q5VW2BqE/lFNKqaMFdUBU1yD0IrVSSh0tqAOi+hqE3uaqlFJHC+qA0BqEUkrVLagDorxKb3NVSqm6BExAiMj/ishmEdkgIn9riWWWVboICxEiQgPmY1BKqYAREI3vIjIRuAAYbIypFJEOLbHc6mdBiEhLLE4ppVqVQDl1vgX4izGmEsAYk9sSC3U4XcTqBWqllPIpUAKiL3CqiHwrIotEZGRdI4rIdBHJEpGsvLy8E1pomdOt1x+UUqoOLXb6LCLzgU4+Bt3rLUc7YAwwEnhTRHoaY0ztkY0xzwPPA2RmZh41/Fg4Kl3azYZSStWhxY6OxphJdQ0TkVuAd7yB8J2IeIBU4MSqCA0o0+dRK6VUnQKliek94AwAEekLRAAHm3uh5RoQSilVp0AJiFlATxFZD7wOXO+reamplTld+itqpZSqQ0AcHY0xTuCall6uo9Ktv6JWSqk6BEoNwi/KnC7tyVUppeoQtAFhjNFrEEopVY+gDQin24PLY/SHckopVYegDQhH9dPktAahlFI+BW9AVGlAKKVUfYI3ICr1caNKKVWfoA2IMqf3YUGRWoNQSilfgjYgtAahlFL1C96AcOo1CKWUqk/QBkSZU2sQSilVn6ANCIdeg1BKqXoFfUDEhGsNQimlfAnegPBepNYnyimllG9Be/pc5nQTERpCRFjQZqRSAaOqqors7GwqKir8XZQ2JSoqim7duhEeHn5c0wdtQDicLmL0+oNSASE7O5v4+HjS09MREX8Xp00wxpCfn092djYZGRnHNY+gPX12ON3EhGtAKBUIKioqSElJ0XBoQiJCSkrKCdXKgjgg9GlySgUSDYemd6KfadAGRJk+TU4ppeoVtAHhcLr0DialVIv46quvWLp0qb+LccyCOCDcxOqvqJVSLaC1BkTQHiEdTrdeg1AqAP3pgw38kFPcpPMc2CWBB84b1OB4r732GjNmzMDpdDJ69GjuueceJk2axLJly0hOTua0007j/vvvZ/LkyVx44YXs2bOHiooK7rjjDqZPnw7Ap59+yj333IPb7SY1NZUXX3yR5557jtDQUF577TWeeuopTj311CZdv+YStEfIskqXXoNQSh22ceNG3njjDZYsWUJ4eDi33norixYt4u677+bmm29m9OjRDBw4kMmTJwMwa9YskpOTKS8vZ+TIkVxyySV4PB5uuukmvv76azIyMigoKCA5OZmbb76ZuLg47rrrLj+v5bEJ2oAod7r1GoRSAagxZ/rNYcGCBaxcuZKRI0cCUF5eTocOHXjwwQd56623eO6551i9evXh8WfMmMG7774LwJ49e9iyZQt5eXlMmDDh8O8OkpOTW35FmlBABISIvAH0875MAgqNMUOba3nGGMqcLr0GoZQ6zBjD9ddfz6OPPnrE+w6Hg+zsbABKS0uJj4/nq6++Yv78+SxbtoyYmBhOP/10KioqMMa0qdt1A+IitTHmcmPMUG8ovA2805zLq3R58Bj0l9RKqcPOPPNM5s6dS25uLgAFBQXs2rWLu+++m6uvvpqHHnqIm266CYCioiLatWtHTEwMmzZtYvny5QCMHTuWRYsWsWPHjsPzAIiPj6ekpMQPa3ViAiIgqomN3v8B5jTncsq8HfVpDUIpVW3gwIE8/PDDTJ48mcGDB3PWWWexc+dOVqxYcTgkIiIieOmll5gyZQoul4vBgwdz//33M2bMGADat2/P888/z8UXX8yQIUO4/PLLATjvvPN49913GTp0KIsXL/bnah4TMcb4uwyHicgE4HFjTGY940wHpgOkpaWN2LVr1zEvZ0+Bg1P/tpC/XTqY/8nsftzlVUo1jY0bNzJgwAB/F6NN8vXZisjK+o6z1VrsFFpE5gOdfAy61xgzz/v/K2mg9mCMeR54HiAzM/O40u3ww4K0BqGUUnVqsSOkMWZSfcNFJAy4GBjR3GU5/LhRvQahlFJ1CqRrEJOATcaY7OZeUPnhp8lpQCilVF0CKSCuoJkvTlc7fJFaf0mtlFJ1CpgjpDFmakst6/DzqPWHckopVadAqkG0mOprEFqDUEqpugVlQFRfg9CuNpRSzSUuLg6AnJwcLr300nrHffLJJ3E4HIdfn3POORQWFjZr+RojKAOirFIvUiuljp3b7T7mabp06cLcuXPrHad2QHz88cckJSUd87KaWlC2sTicLiLDQggLDcp8VCrwvfTzo98bdCGMugmcDph92dHDh14Fw66Gsnx487ojh93wUYOL3LlzJ1OmTGH06NGsWrWKvn378uqrrzJw4EBuvPFGPv/8c371q18xcuRIbrvtNvLy8oiJiWHmzJn079+fHTt2cNVVV+FyuZgyZcoR8z333HNZv349brebu+++m88++wwR4aabbsIYQ05ODhMnTiQ1NZWFCxeSnp5OVlYWqampPP7448yaNQuAadOmceedd7Jz507OPvtsxo8fz9KlS+natSvz5s0jOjr6mD7mhgTlEdLhdOsFaqXUUTZv3sz06dNZu3YtCQkJPPPMMwBERUXxzTffcMUVVzB9+nSeeuopVq5cyWOPPcatt94KwB133MEtt9zCihUr6NTJ12+C4fnnn2fHjh2sWrWKtWvXcvXVV3P77bfTpUsXFi5cyMKFC48Yf+XKlbz00kt8++23LF++nJkzZ7Jq1SoAtmzZwm233caGDRtISkri7bffbvLPIyhrEGVOFzH6K2qlAld9Z/wRMfUPj01pVI3Bl+7duzNu3DgArrnmGmbMmAFwuE+l0tJSli5dymWX/VSDqaysBGDJkiWHD9LXXnstd99991Hznz9/PjfffDNhYfb401B34N988w0XXXQRsbGxAFx88cUsXryY888/n4yMDIYOtZ1ejxgxgp07dx7XOtcnKI+Sjko3sforaqVULbW76q5+XX2A9ng8JCUlHfFciPqmr+1YuwOvr6+8yMjIw/8PDQ2lvLy80fNtrKBsYtIahFLKl927d7Ns2TIA5syZw/jx448YnpCQQEZGBm+99RZgD+Br1qwBYNy4cbz++usAzJ492+f8J0+ezHPPPYfLZW+1b6g78AkTJvDee+/hcDgoKyvj3XffbdHHlQZlQJTrNQillA8DBgzglVdeYfDgwRQUFHDLLbccNc7s2bN58cUXGTJkCIMGDWLePNvX6D//+U+efvppRo4cSVFRkc/5T5s2jbS0NAYPHsyQIUP473//C8D06dM5++yzmThx4hHjDx8+nKlTpzJq1ChGjx7NtGnTGDZsWBOvdd0CqrvvY5WZmWmysrKOebq/fbqJmIhQfnVGn2YolVLqWAVCd9817zZqS1pFd9+B5HdT+vu7CEopFfCCsolJKaVqS09Pb3O1hxOlAaGUCgitubk7UJ3oZ6oBoZTyu6ioKPLz8zUkmpAxhvz8fKKioo57HkF5DUIpFVi6detGdnY2eXl5/i5KmxIVFUW3bt2Oe3oNCKWU34WHh5ORkeHvYqhatIlJKaWUTxoQSimlfNKAUEop5VOr/iW1iOQBu45z8lTgYBMWpzUIxnWG4FzvYFxnCM71Pp517mGMad/QSK06IE6EiGQ15qfmbUkwrjME53oH4zpDcK53c66zNjEppZTySQNCKaWUT8EcEM/7uwB+EIzrDMG53sG4zhCc691s6xy01yCUUkrVL5hrEEoppeqhAaGUUsqnoAsIEZkiIptFZKuI/N7f5WkuItJdRBaKyEYR2SAid3jfTxaRL0Rki/ffdv4ua1MTkVARWSUiH3pfZ4jIt951fkNEIvxdxqYmIkkiMldENnm3+di2vq1F5NfefXu9iMwRkai2uK1FZJaI5IrI+hrv+dy2Ys3wHt/WisjwE1l2UAWEiIQCTwNnAwOBK0VkoH9L1WxcwG+NMQOAMcBt3nX9PbDAGNMHWOB93dbcAWys8fqvwBPedT4E/MIvpWpe/wQ+Ncb0B4Zg17/NbmsR6QrcDmQaY04CQoEraJvb+mVgSq336tq2ZwN9vH/TgWdPZMFBFRDAKGCrMWa7McYJvA5c4OcyNQtjzD5jzPfe/5dgDxhdsev7ine0V4AL/VPC5iEi3YCfAy94XwtwBjDXO0pbXOcEYALwIoAxxmmMKaSNb2tsb9TRIhIGxAD7aIPb2hjzNVBQ6+26tu0FwKvGWg4kiUjn4112sAVEV2BPjdfZ3vfaNBFJB4YB3wIdjTH7wIYI0MF/JWsWTwK/Azze1ylAoTHG5X3dFrd5TyAPeMnbtPaCiMTShre1MWYv8BiwGxsMRcBK2v62rlbXtm3SY1ywBYT4eK9N3+crInHA28Cdxphif5enOYnIuUCuMWZlzbd9jNrWtnkYMBx41hgzDCijDTUn+eJtc78AyAC6ALHY5pXa2tq2bkiT7u/BFhDZQPcar7sBOX4qS7MTkXBsOMw2xrzjfftAdZXT+2+uv8rXDMYB54vITmzz4RnYGkWStxkC2uY2zwayjTHfel/PxQZGW97Wk4Adxpg8Y0wV8A5wCm1/W1era9s26TEu2AJiBdDHe6dDBPai1vt+LlOz8La9vwhsNMY8XmPQ+8D13v9fD8xr6bI1F2PMH4wx3Ywx6dht+6Ux5mpgIXCpd7Q2tc4Axpj9wB4R6ed960zgB9rwtsY2LY0RkRjvvl69zm16W9dQ17Z9H7jOezfTGKCouinqeATdL6lF5BzsWWUoMMsY84ifi9QsRGQ8sBhYx0/t8fdgr0O8CaRhv2SXGWNqXwBr9UTkdOAuY8y5ItITW6NIBlYB1xhjKv1ZvqYmIkOxF+YjgO3ADdgTwDa7rUXkT8Dl2Dv2VgHTsO3tbWpbi8gc4HRst94HgAeA9/Cxbb1h+S/sXU8O4AZjTNZxLzvYAkIppVTjBFsTk1JKqUbSgFBKKeWTBoRSSimfNCCUUkr5pAGhlFLKJw0IpVqYiHwlIs3ykHmlmpIGhFJKKZ80IJQCRCRWRD4SkTXe5wtcLiJ/FJEV3tfPe3+EVF0DeEJEvvY+e2GkiLzj7Zv/Ye846d5nM7zi7Zd/rojE+FjuZBFZJiLfi8hb3r6zEJG/iMgP3mkfa9lPQylLA0IpawqQY4wZ4n2+wKfAv4wxI72vo4Fza4zvNMZMAJ7DdnNwG3ASMFVEUrzj9AOeN8YMBoqBW2suUERSgfuAScaY4UAW8BsRSQYuAgZ5p324eVZZqfppQChlrQMmichfReRUY0wRMNH7dLJ12I7/BtUY//0a023wPn+jEtvNRXVnaXuMMUu8/38NGF9rmWOwD65aIiKrsX3q9MCGSQXwgohcjO0yQakWF9bwKEq1fcaYH0VkBHAO8KiIfI6tFWQaY/aIyINAVI1Jqvv38dT4f/Xr6u9V7X5sar8W4AtjzJW1yyMio7Ad0F0B/AobUEq1KK1BKAWISBfAYYx5Dfsgmupn+R70Xhe4tM6J65YmImO9/78S+KbW8OXAOBHp7S1DjIj09S4v0RjzMXAnMPQ4lq3UCdMahFLWycDfRcQDVAG3YB/juA7Yie0q/lhtBK4XkX8DW6j1fGBjTJ6ITAXmiEik9+37gBJgnohEYWsZvz6OZSt1wrQ3V6Wagfcxrx96L3Ar1SppE5NSSimftAahlFLKJ61BKKWU8kkDQimllE8aEEoppXzSgFBKKeWTBoRSSimf/j8mi9lJ+Am9agAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_all_mag')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_folder = 'data/matrix_pool/'\n",
    "dataset = pd.read_csv('csv/pool.csv')\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "X_1 = dataset.iloc[:,1].values\n",
    "Y_1 = dataset.iloc[:,3].values\n",
    "\n",
    "W = np.random.permutation(np.c_[X_1.reshape(len(X_1), -1), Y_1.reshape(len(Y_1), -1)])\n",
    "\n",
    "X = W[:, :X_1.size//len(X_1)].reshape(X_1.shape)\n",
    "Y = W[:, X_1.size//len(X_1):].reshape(Y_1.shape)\n",
    "\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in X[0:100]:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in Y[0:100]:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/my_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score (RMSE): 0.5256468943250299\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Retrain 500\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_retrain_500.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_retrain_500.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_500/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_rand_500.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score (RMSE): 0.5060998222673577\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_500_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_pool_retrain_500.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_pool_retrain_500.csv')\n",
    "\n",
    "path_folder = 'data/matrix_pool_500/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_pool_500.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "50/50 [==============================] - 5s 97ms/step\n",
      "Score (RMSE): 0.10770684056968804\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128,verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool_500_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Retrain 1k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_retrain_1k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_retrain_1k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_1k//'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_rand_1k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 10s 97ms/step\n",
      "Score (RMSE): 0.411783127528642\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZEAAAEWCAYAAACnlKo3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsnXd4VFXegN+TTnpCQkshAem9g0gTBERUxC4qioioa1kb66q7uq7rfuuKsooidhTsAhakKb13AekQSAglhfSeOd8fZy5TMpNM+iSc93nuM3PvPffccydwfvf8qpBSotFoNBpNVfCo7wFoNBqNpuGihYhGo9FoqowWIhqNRqOpMlqIaDQajabKaCGi0Wg0miqjhYhGo9FoqowWIhqNRqOpMlqIaBo0QogEIUS+ECJHCHFWCPGJECLQ6vwnQggphLjO7ro3zcfvMe/7CCFeF0Ikmfs6IYR4w8l9jO3tGhj/n83jzhRCfCSE8HXhmr+bxz7K6tgtQoiNQog8IcRqB9fMFUIcEkKYjGe2OjfH7rkKhRDZVufjhBBLhBAXzGN9WwjhVb0n1zQWtBDRNAaulVIGAj2BXsCzducPA5ONHfMEeDNwzKrNs0BfoD8QBIwAdjm6j9X2p+oMWggxBvgLMBKIA9oAL1VwTVvgJuCM3al04E3g304u3QM8BOy0PyGlnG79XMAXwDdWTd4BzgMtUb/xMHNfGo0WIprGg5TyLLAMNdFZ8yMwWAgRZt4fC/wOnLVq0w9YKKVMlooEKeW8Wh7yZOBDKeV+KeUF4GXgngqueRuYARRZH5RSrpRSfg0kO7pISjlbSvkrUFBe50KIAOBG4FOrw/HA11LKAvNvvBToUsE4NZcIWohoGg1CiGjgauCo3akC4AfgNvP+3YC9gNgMPCGEeEgI0U0IIaoxjjuEEBnlbLHmpl1QKwSDPUBzIURTJ/3eDBRJKZdUdWwucCOQAqy1OjYLuE0I4S+EiEL9xktrcQyaBoQWIprGwCKzDj8RpXb5u4M284C7hRAhKHXMIrvzrwL/B0wCtgOnhRCT7dosshMG9zsajJRygZQytJztlLlpIJBpdanxPci+T7Od51/A405+g5piMjBP2ibVW4MSeFlAEur3sf/9NJcoWohoGgMTpJRBwHCgIxBh30BKuR6IBJ4HfpJS5tudLzWrfAYDocArwEdCiE5297EWBu9Xc9w5QLDVvvE920Hbl4DPpJQnqnlPpwghYlACdp7VMQ+UivB7IAD124ahBK5Go4WIpvEgpVwDfAL810mTz4EnKavKsu8nX0o5G7gAdK7sOIQQk+y8new3Q521H+hhdWkP4JyUMs1BtyOBR83eUWeBGOBrIcSMyo6vHO4GNkopj1sdCzff620pZaF5bB8D42rwvpoGjBYimsbGm8BVQgh74zrA/4CrsNX3AyCEeFwIMVwI0UQI4WVWZQVR1kOrQqSU8+28uOw3Q501D7hPCNHZbPR/HiUEHTES6IpyGuiJMqA/AMw2j99TCOEHeAEeQgg/IYS31fP5mM8LwNt83v7//93295dSpgIngAfNv0soSuW1B40GLUQ0jQwpZQpqcn7Bwbl0KeWvdvp+g3zgdZTHVirwMHCj3Vv5j3YrioXVHOtS4D/AKuCkebtozxFC7BdCTDK3TZNSnjU2oBS4IKXMMTe/y/wM7wJDzN+t1W3LzccuB+aavw+1utcgIBpb116DiSiPthSU00IJ8OfqPLum8SB0USqNRqPRVBW9EtFoNBpNldFCRKPRaDRVRgsRjUaj0VQZLUQ0Go1GU2UafSbOiIgIGRcXV9/D0Gg0mgbFjh07UqWUkRW1a/RCJC4uju3bt9f3MDQajaZBIYQ46Uo7rc7SaDQaTZVxCyEihLjZHFhlEkL0LaddghBirxBitxBCLy80Go2mnnEXddY+VFTsey60HWFOxaDRaDSaesYthIiU8gBANUo4VIri4mKSkpIoKCi3Po+mkvj5+REdHY23t3fFjTUaTaPALYRIJZDAciGEBN6TUs511EgIMQ2YBhAbG1vmfFJSEkFBQcTFxdWZ4GrsSClJS0sjKSmJ+Pj4+h6ORqOpI+rMJiKEWCmE2Odgu74S3QyWUvZGVVZ7WAgx1FEjKeVcKWVfKWXfyMiyHmoFBQU0bdpUC5AaRAhB06ZN9epOo7nEqLOViJRyVA30kWz+PG/OoNofB2m9XUELkJpH/6YazaWHW3hnuYIQIkAIEWR8B0ajDPIajUajsWP+lpMs3Xe21u/jFkJECHGDECIJGAT8LIRYZj7eSgixxNysObBeCLEH2Ar8bK7HoLFj9erVbNy4sb6HodFo6on03CL+9fMBfthzutbv5RaGdSnlQqBMgR+z+mqc+ftxbEuJapywevVqAgMDufzyy+t7KBqNph54b80x8opL+fOo9rV+L7dYiVyqfP755/Tv35+ePXvywAMPcPLkSdq1a0dqaiomk4khQ4awfPlyACZMmECfPn3o0qULc+danNKWLl1K79696dGjByNHjiQhIYE5c+bwxhtv0LNnT9atW1dfj6fRaOqB81kFfLopgQk9o2jXPKjW7+cWK5H65KUf9/NHclaN9tm5VTB/v7ZLuW0OHDjAV199xYYNG/D29uahhx5izZo1zJgxg+nTpzNgwAA6d+7M6NGjAfjoo48IDw8nPz+ffv36ceONN2Iymbj//vtZu3Yt8fHxpKenEx4ezvTp0wkMDOSpp56q0efSaDTuzzurj1FcKnlsZLs6ud8lL0Tqi19//ZUdO3bQr18/APLz82nWrBkvvvgi33zzDXPmzGH37t0X2//vf/9j4UKl8UtMTOTIkSOkpKQwdOjQi3EZ4eHhdf8gGo3GbTidkc+CLae4uU80cREBdXLPS16IVLRiqC2klEyePJlXX33V5nheXh5JSUkA5OTkEBQUxOrVq1m5ciWbNm3C39+f4cOHU1BQgJRSu9VqNBoASk2S15YeBOCROlqFgLaJ1BsjR47k22+/5fz58wCkp6dz8uRJZsyYwaRJk/jHP/7B/fffD0BmZiZhYWH4+/tz8OBBNm/eDMCgQYNYs2YNJ06cuNgHQFBQENnZ2fXwVBqNpj44kZrLLe9tYtHuZO4bEk9UaJM6u/clvxKpLzp37sw///lPRo8ejclkwtvbm5kzZ7Jt2zY2bNiAp6cn3333HR9//DF33HEHc+bMoXv37nTo0IGBAwcCEBkZydy5c5k4cSImk4lmzZqxYsUKrr32Wm666SYWL17MW2+9xZAhQ+r5aTUaTW1QUFzKvE0JzFxxGB9PD964tQcTekbV6RiElLJOb1jX9O3bV9oXpTpw4ACdOnWqpxE1bvRvq9HUPnlFJSzYcor31h4nJbuQKzs249WJ3Wge7Fdj9xBC7JBSOi3NYaBXIhqNRtMAOJ9dwOpDKaw5lMLaIylkF5QwqE1T/ndbLwa2Ca83+6gWIhqNRuPmbDiaypRPtlFYYqJZkC9ju7Tgln4x9Iurf49MLUQ0Go3GjdmTmMG0eduJaxrAzFt70LllsFt5ZWohotFoNG7K0fPZ3PPxVsIDffjsvv40q0GbR02hXXw1Go3GDVl/JJU7P9iKl6cHn983wC0FCGghotFoNG5F0oU8pn+2gzs/3IKPlwfzpvSndVOr6HMp4Zt74cjK+hukFVqd1UgIDAwkJyeH5ORkHn30Ub799lunbd98802mTZuGv78/AOPGjWPBggWEhobW1XA1mkue1JxCtidcYFtCOvtOZ3Ihr4iMvGLScovw9hQ8Nbo9U4e0wc/b0/ZCISDtCGycBe2qXeuv2mgh4saUlpbi6elZcUMrWrVqVa4AASVE7rzzzotCZMmSJeW212g0NcvKP84x7bPtmCT4eHnQtVUwbSICCfX3JiLQl9sHxJYfdd76Ctj+ERTlgk/d5MhyhhYi9URCQgJjx45lwIAB7Nq1i/bt2zNv3jw6d+7MlClTWL58OX/605/o168fDz/8MCkpKfj7+/P+++/TsWNHTpw4wR133EFJSQljx4616Xf8+PHs27eP0tJSZsyYwbJlyxBCcP/99yOlJDk5mREjRhAREcGqVauIi4tj+/btREREMHPmTD766CMApk6dyuOPP05CQgJXX301V1xxBRs3biQqKorFixfTpEndpVbQaBoTc9YcIybcn5m39KBrVAi+XpV4Wdy/CI4sh9JCOLYKOo2vvYG6gFsIESHEa8C1QBFwDLhXSpnhoN1YYBbgCXwgpfx3jQzg42vKHusyAfrfD0V5MP/msud73gG9JkFuGnx9t+25e3926baHDh3iww8/ZPDgwUyZMoV33nkHAD8/P9avXw+oHFtz5syhXbt2bNmyhYceeojffvuNxx57jAcffJC7776b2bNnO+x/7ty5nDhxgl27duHl5XUxVfzMmTNZtWoVERERNu137NjBxx9/zJYtW5BSMmDAAIYNG0ZYWBhHjhzhiy++4P333+eWW27hu+++484773TpOTUajYWDZ7PYfvICz1/TiT6tqxDncew3yE0F3xA4/Eu9CxF3MayvALpKKbsDh4Fn7RsIITyB2cDVQGfgdiFE5zodZQ0TExPD4MGDAbjzzjsvCo5bb70VUFl8N27cyM0333yxcNWZM2cA2LBhA7fffjsAd911l8P+V65cyfTp0/HyUu8KFaWKX79+PTfccAMBAQEEBgYyceLEi0Wt4uPj6dmzJwB9+vQhISGhGk+u0Vy6zN98Ch8vD27sHV21DlKPQPPO0O4qOLZaGdrrEbdYiUgpl1vtbgZuctCsP3DUXCYXIcSXwPXAH9UeQHkrBx//8s8HNHV55WGPfcCQsR8QoHScJpOJ0NBQm7oi5V1vT2VTxZeXR83X1/fid09PT/Lz813uV6PRKHILS1i46zTju7UkLMCnap2kHoKO42HYDPALUYb2esRdViLWTAF+cXA8Cki02k8yHyuDEGKaEGK7EGJ7SkpKLQyxZjh16hSbNm0C4IsvvuCKK66wOR8cHEx8fDzffPMNoCb5PXv2ADB48GC+/PJLAObPn++w/9GjRzNnzhxKSkqAilPFDx06lEWLFpGXl0dubi4LFy7UGYA1mhpk8e5kcgpLmDQwtmod5KZBXhpEtIeQKPANrNkBVoE6EyJCiJVCiH0Otuut2jwHlACOZkVH4tbhq7OUcq6Usq+Usm9kZGTNPEAt0KlTJz799FO6d+9Oeno6Dz74YJk28+fP58MPP6RHjx506dKFxYsXAzBr1ixmz55Nv379yMzMdNj/1KlTiY2NpXv37vTo0YMFCxYAMG3aNK6++mpGjBhh0753797cc8899O/fnwEDBjB16lR69epVw0+t0VyaSCmZv+UkHVsE0Ts2rGqd5KZAaGuI7Kj2934L395Xc4OsAm6TCl4IMRmYDoyUUuY5OD8IeFFKOca8/yyAlPJV+7bWuGsqeGsvqsaEO/y2Go07sjsxgwmzN/DyhK7cNbB1zXS6+V1Y+hd4ZCc0bWs5vu87yEyCgQ+Bp3eVunY1FbxbqLPMXlczgOscCRAz24B2Qoh4IYQPcBvwQ12NUaPRaKrDF1tO4e/jyYSerWqu0/Zm9/79Cy3HSkvg15dh3/fgUftmb7cQIsDbQBCwQgixWwgxB0AI0UoIsQRASlkC/AlYBhwAvpZS7q+vAVeXuLi4RrcK0WguKbZ/DDs+dalpTmEJP/6ezLXdWxHkV7WVAQA/PAq/zLDsh8fDZaNgzf9Bklnjsv97uHAChj5dJ0Z3txAiUsrLpJQxUsqe5m26+XiylHKcVbslUsr2Usq2UspXqnnP6g5bY4f+TTWXFEeWw4+PgslUYdMf9ySTV1TKrf1jqnfP46uVXcSaie9DUEuVT6s4H9b+F5p1hg7jHHZR07iFEKlr/Pz8SEtL05NeDSKlJC0tDT8/98w0qtHUOJ0nqM/kXRU2/XJbIu2bB9Irphr56YrzIeOU8syyxj8cblsAE2YrwZZ6CIY8CR51M727RZxIXRMdHU1SUhLu7P7bEPHz8yM6uooBVBpNQyLtGBSZ3eSPLIPoPk6bHjiTxZ7EDF4Y37l6xaTSjgKyrBABaNFVfSZugx63Q5cbqn6fSnJJChFvb2/i4+PrexgajaahcvBnWPEChLeFw8tgxF+dNv1qWyI+nh7c0MthWJvrpB5Wn46EiEFMP7XVIZekOkuj0WiqxdnfITgaet4O5w9AfplUfwAUFJeycNdpxnRtQbgrEerpx+GL26EgS+2XFFnOeftD68HQ9LIaeICa45JciWg0Gk21OLMHWnaHflNhwHTwDXLY7Ic9yWTmF3NbPxcM6vkZsOBWuHDS0t/ihyGsNQz/K3S4Wm1uhl6JaDQaTWUoylVJEFt0hyZhTgXIyj/O8fyifXSNCmZQm6bl91laDN/cA+kn4K7vLa65/k1h7Wvw3X0qo7gbooWIRqPRVIZz+wEJLXuo/SMrYN71NqqnH/ckM/3zHXRsEcRnUwbg4VGOQV1K+OUZOL4Krn0T4qxy6I19Fa76h4r9+FdLWP9G7TxTNdDqLI1Go6kM0f3gsT3gb67HU1oEx1eza8MvbBfdOHg2m+93JdGvdTgf3tO34uDCggwVcT74cehlV6NHCBj8GITGwsLp0Lxr7TxTNdBCRNM4MJnAVAJeVUyvrdG4ihAQFofJJNl8NJXvdoTyqvRi+4oveaXEi6YBPlzfoxWvTuxOE59yKhamHIKm7ZRK7N5fIKKD87ZdblBxKfWc9t0RWohoGgc/PgK7PocXHWc01pgpyFRbaBVTkbsz2z6EnPOqzkZtBtqt+helLbozZVMz1hxOIdjPi4eCe3GH1wlumnqVa3VCTm2GT8bDmH/BgGnQzIWkpW4oQEALEU1jYdfn6rMwxy1qLLgt/zYLj8YobH+ZAaZiiB9ia1dwxr7v4ehKiB+mhGrrQRVfU1IE62ayu9UdrDk6hmfGdmDK4Hj8Vv4Guz4nwN/FvFh/LAYPT+jmqP5ew0ILEU3jIDRWpYTQNDzWvW7O9VRN91VTsfq0zy3ljENLlFH85AZl3H5wg1NPq4ukHABTMZ+cCOb6nq14cFhbFYXeqpfy2CrOV9VQK+LUZojqo1KWNHC0d5amcRDeBmIG6FWIq5QU1vcILGx8W9XFqA5FuZbvuakVt5cSTm6EtiPghvfUC8iy5yq8LO/kTgDSgjvxzwldLWlMetymXHNdESBFeSpYMWZAxW0bAFqIaBoHQ5+BnpOUTlzjmNISy3dXJtq6Iqw1eFbTISL9uOW7KyuRjFOQdRpiL4fYgTD4Udj5Kexe4LB5Zn4x3+9MYtXqleTIJvzljqurntI9eadyAmkkQkSrszSNg6je8Mk4yH1e1VHQlMVIGAhqog2pZi6nmqAgU2XB9QupXj95acrlNi/VtReJU5vUZ+vL1eeI5+D0Tlj0IMXNuzPrd28On8smv7iUnMIS9p3OpLhUMtc/k/yIrnSPsVNDlRTB7H7Qd4pyyS2P6H4wZTk071z553RDtBDRNA4S1qvPrOT6HYc7Y0Q8t+ql6k+4A5mn1WdBNQ39bYbDM8dg9f/Zlol1hpTQqreyxQB4+cKd31N6ZAVPri7mhz0nad8sAH9fb5p4e3LP5XGM69aSHtHj8MBB/RAvH/X7ph2t+N5evhDbOFYh4CZCRAjxGnAtUAQcA+6VUpbJaCaESACygVKgxJX6v5pLACnhi9vUdy1EnBMS5X5eWXk1rFYbPqPiNqASJ/a83eaQycObv+6P4Yc9ibwxMI8b/DfB6H9aGpSWgIcAnMR+hLWGCwnl39dkgt9ehs7XQ6uero3VzXELIQKsAJ6VUpYIIf4PeBZVc90RI6SUbqTQ1dQ7xflKxwxKz61xjpQqbYdfCIRWs8peTWDYZob9pXr9fP+Aym475Am1qnHi9VRqkqzcd5r0/FIkAonEJMFkkuxJzOD7Xad55MrLuKFgNmz9Ci5/DAIj1e82ZzB0vRGGPeN4DGFxkLil/HGmHYH1M9VqSQuRmkNKudxqdzPQ8J2nNXXHRVWI0CuR8kjeDRtmqTxMV/wZRr1Y3yOyCJF+U6veh5Rw+BfoehMs+6syjj+b6LDpe2uPsXf5PP7p/RE3Fr1IgrRV600b2oYnrmoPaY/Aznmw5V0Y+TflyZVyEELKEbyhrVXsSWkxeDoxup/arD5jBlblSd0StxAidkwBvnJyTgLLhRASeE9KOddRIyHENGAaQGxsI4zM1dhiCJGBD0F0XzWpuGl0bxny0pWRt+M1tX+vCyeUAAH38c4y1FkpB9TqqCppa/LS1L+BiHZQlAOFWVBcAN62pZr/SM7ijRWHeS8ykbDcEr587FaElzdCgKcQeHt5EGx4XEW0g87XwdYPVE6r3fPBJ0gdc0bsIOh1Xq2MnQmRxC0qM68rdpsGQp0JESHESqCFg1PPSSkXm9s8B5QA8510M1hKmSyEaAasEEIclFKutW9kFi5zAfr27asLqTd2Cs0FfC67Ei4bVb9jqSzLX4Ddn8MDay1ZYWuLwhz16RfiPkJk2Azl3vvptfDn/RBShfLKhjG76WWWlWhuio26rrCklCe+3k2ovw9DfQ/jEd6fFuEVBBZe8YSKLN/4P9i/CLpOBJ8A5+3bjVJbeZzarFx7G8pLjgvUWZyIlHKUlLKrg80QIJOB8cAkKaXDiV9KmWz+PA8sBPrX1fg1bkxkB7j7B4jsCCfWQfa5+h6R6xg5rI6urP17FZpdfMPia96gXVU8PFWgKFiq+VWW1CPqs2lbCIhU3+1iRWauOMzBs9m8fm0sXuf3W1x7y6NVT+Wym7wbinPLZth1hMmkViKOKMyB/AuNJj7EwC2CDYUQY1GG9OuklA4rrwghAoQQQcZ3YDSwr+5GqXFb/EKgzTDlYvnpeDixpr5H5DrDZ0DLnnBoae3fq8i8EgmLc5+VyKbZcOBH9b0wu/y2zvDyU+66IbFWQsTyfL/sPcPctce5vX8sQ4/PUgfbjXat7/FvwPiZcNXLFU/+JpPKTbbm/xyf9w2EZ47DwAddu3cDwV1sIm8DvigVFcBmKeV0IUQr4AMp5TigObDQfN4LWCClrIP/eRq35/xBOL8f2oxQ+w3NQ6vDOFj9qpr4AiJq7z4eXmqiHfRw9eMyaoqd8ywrkMIqrkS636w2gPB4GPG8+gS+2HqK5xbupVdMKM9f0wnynlapTqJ6u95/aKyKaK8IDw/197twsuy54gLIT4fgVipOpBHhFkJESumw8rxZfTXO/P04UMtKY02D5NAS+PUleO4s+IY0LA+tn59SXj+P7KhdAQLK/XXIE7V7j8qSm6rUkNnJNSPYAiJg2NNIKXl39VH+s/QQV7YPZ07Xg/h4D1KrsLC46t/HGWFxjmNFlj0Lf/wAf9rWKJIuWuMW6iyNploUZinjrJefCqhrSEIk5aCKcalLb52cFJW91jC01xemUuVZ1aon3PKZ8m6qSh+vd4QtylEzOSOfT5dt5PbXF/KfpYeY0LMVcy/PwGfJ48oNuLZxJER+/wa2fwS9JjU6AQJushLRaKpFQSb4BiuPl+BW7ilEclIg5yy06GZ7vDALApurvE2b34VrZ7mWCbYqrHxJlXKNHQhf3Vk3HmHlkX8BkEpdVJ7rbDmcTjhCVPYZPtl2lo/WrOJUeh5rfR4jzr8bEybO4pa+MXhsmKkaxw+tubE7I6y1UlsVZIFfMKQchh8fU4ker/xb7d+/HtBCRNPwKci0JPAb8Zx7uk++M0C9ddunHSnIUiVSCzJh79eqDGrHcbUzhlOblTdUx/Fqv76N6/kZyk4TEAFHf1X5vCqRlHB7Qjoff7aY2cCevAi6RYdwW/8Ymh+IITZAQn+z59t5c5BgRbVCaoL4oerfoDSpeKWf/qxsIDd9CJ6Nc7ptnE+lubQoyLIIkcoYTOuSoFaOJ7HCbPXG2nqwWk0dWlJ7QqQoW02mhu0lL6127uMqEZfBC6lqwv1PPHS/Dcb9x6VLf9yTzJPf7OER//NQBG88dBMENlMnz7SADKuI9ZQDyu5SF0T1URsoV9/gltBlglohN1K0ENE0fK75r/J+Acg6o1x824+BJmH1Oy5rinKUG6o9MQOgeRcVqd3uKji8VOn5PZwk+asOhdngE2gRIq5WAKxNhADhqRwiXPDOKjVJ/vfrEWb9eoR+cWE8EJAIKVEW115Qz3d6h/puKlVxJG2G18rwyyCl+l2lhKDmcOMHdXPfekQb1jUNn7A4aGZ+0zy3HxY+oHTR7oLJZEk5kmv39n/7AkveqLYj1QSUdqx2xlGYo1ZDfqFKjVTf6qwjK2Dxw5ZxVRAnci6rgEkfbGbWr0eY2CuKz+4bgE/bIXD5o7YqzIBI9WwmEwgPeHgLDKjD2Iy3+sKXt8PZSyOMTa9ENA2fHZ9AZCdVo8FQG7hTrEjOWcv3rCQIaOq4XUR7ZXgvKaidcYTGqlQgQsCd39Wuq6srnN4Buz6H8W8qlZ4DF9+SUhMHz2az5UQ6s1cdJb+olP/e3IMbe0ep0rSXP1K2347jVRS8NClhWZfPKQSExapn+/lJuG9Z3d27ntBCRNPw+eUv0H+qnRBxIw+twOZw/Wz11p1vVSYn/Th8fI2KiO5wNcT0g+nra28c01ZZvrcZXnv3cZXcVLUq8vRG+gZTcCGZeWuOcTwll5ScQlJzCjl6Poe8olIAukWF8OZtPWkbGaiuP7EWovqW9WaL6m2xjR1ernJrDXyw7hwuwuLg7F646h91c796RgsRTcOmpAhK8i2Gdb8Q8A5wLyHi4QktuqvvBVZCJD9DBdnVB6e2qPxZdZE92Bl5qciACP6z9CC7j48mK6+A/UkHiQzypVmQLxGBvtzcJ5o+ceH0aR1GVGgTy7U5KTDvehjyJFz5vG2/xflKrRkWD79/BYlbYdBDdfdc/aepVO+NqHpheWghonFO8i5lW+hxa32PxDmGMdbXLEQuxoq4kTrr6K9w2KzWsF6JXBy7ldfWZxNVOvsRf63ZMWSehq/vhhHPqkzHW99Tf9/6FCK5qZwpDuTd1ccY1aknN3ZtwZB2ETQP9qv42gOLlbqqyw1lz11IgA9Gwk0fq2DOyA41PvRyiR9aNzEpboI2rGucs/5NWDitvkdRPoYe3ViJANz6OVztwFU08zQsnO48y2ptceAHZbcI4rcuAAAgAElEQVQB25WIkTPKN9hyLCtZvUXXNPkX4PR2KMpV+/4R9W5YT80rZesFf27oFcX7Y5twk+c61wQIqOJPkR0tNdKtMTy1ss8qz6xmdeTee4mihYjGOZEdAKHcJN0VY1K2FiLNOir3SnuWPw97vlButHVJxilo1gn+kqg8iQwMbyQ/KyES3LJ2VHHGvYxVT0CkWgmVFNb8vVxgf3ImQ848xictnufVid0Qh36BRdNdG0/maVVpsOuNju0cTcKUV1bSVigtVE4XmlpDCxGNY0pLlHEQ6T4ZXx3RvBs89jvEXWE5dmYPrP1vWeFXWlS3YzPISFSeUX7BtpNecEvodK1tPEtQK8g+U/NjuChEzALL8BCr44DDM5n5vLbsIJM+2EKovzdz7+6Dn7en5SXAuqbItg9UOV97Di8FJHSZ6PgmHp5qpZW0AxB6JVLLaJuIxjEXElT0NChViLsmjvPyUfmKrEncCr+9DL3usl2RdLwGDv5Ut+osKSEzUQU/bpilJvG+96pzba9UmzXBLSHnnBLiNZkmo8gsRHzMnk3+VgGH1YymllJy+FwOWQXFFBabKCotpaRUUmqSFJsk57MKOJ2Rz4nUXNYdScUkJde28+dV+QYB5yQEjbKskAqzINCsjvr5SfU54EHbsrl9pyibQ4TD5N+KgEjlJfWnreDhpFStpkbQQkTjmJSDlu/Wenx34/QOOL5GecT4mifIQLPgyDlnK0S636YMsd5NyvZTW+SlKRVNaGulSmsSZhEijmjRDdqNgeI88Ax23q6y+AQpd1hj1RM/FKZvUCVlq0JxPuz7npQ2N/DXxX+w4o/yq0n6+3gSFdqEqVfEc+fA1sSUJsLsNZB3l2pgrJAcrXpPbbR1SRZC1UAvjzGvqD7r8m99iaKFiMYxqYfU5+h/QnAV6l7XFSc3qloiRtQ3QFAL9ZljN7FlJdX9xBIQAc+fU6q1w7/YCuSfnoCE9ept2aDz9WqradqPVptBk1C1VZUtc2Dli7wjfmdN8XCeHtOBHtGh+Hp74OPpgaeHwNND4OUhiAzyJaSJN8JalZdgNuobajXDLmQdtT71N/jgSjj2m0WIHFyiVpNj/lX++NuOUL/v+T+g911Vf05NhbiFEBFCvAxcD5iA88A9Rj11u3aTAcMp/J9Syk/rbpSXGCmHlX7eUUSwO1GQpYyohpoGLIn4ss/atv38Rkg9DNfMhH731d0YjUp2fqG2Ve/y05Wban1QWgI7PlaleWP6VerSklITW/cc4HJgjPdu7pj+PO2aVzJDrlHj3VCrtewJD6yz1FsHiO4DD22xXXXs+w6Or4br3i6//9QjsP1DW0cGTa3gLob116SU3aWUPYGfgDKJ94UQ4cDfgQFAf+DvQgg3yrDXyEg9pHTOybsgM6m+R+OcgkylT/ew+qcc6GQlknNefR6qg+JEBgd+hF9mqEm7SWhZF18/O5VVXroqsrTtw5odx9r/wgejLPvCQ43LUaGmvHTlnOCArIJipny6nReS+pLp04IBpTtpF1wF7z3Dvdhwx/UNhJbdLSrJCwmw9X21kjOSUZaWwNEVyr7kUcHUtc5cQ6S+U7tcAriFEJFSWqfvDACkg2ZjgBVSynQp5QVgBTC2LsZ3SXLtLFWreu5w2L2gvkfjHOtaIgbefvDEARj8mOVYSaFlAs+rw/iIY6tgz5fKSO4XqiZCaf7nXZhVNj28X6gydldXcO/41JLJFiDjpG16dKMeuKNYkTlD4L2hKoGhFb8nZXDjOxvZeDSVaRPHEnLvN4i+91XN683LF8Lbgr9ZnVVSpLyxTu9U+4nbYMlT6kXglxmw+wtI3KL+3u3HVNy/IYzsnS40NY5bqLMAhBCvAHcDmcAIB02iAKv/BSSZjznqaxowDSA2NrZmB3qpYFS88wkyV6BzUwqzygoRKOtxZJ32vC6D7DITVdJDgJF/g1F/t5wrzFaFmKzx8FArKXtVXGWQUk28Q56w1LYozLZMrAb+EWVdfAtzlO0I1G/bJJTdiRnMWnmYVYdSCA/wYf7kHgwoWA8BQ+Dqf1dtjL3uVJuBEMoba8RzKu+VkXEgNFblyDr/h1J5eXhDG0fTgx1XvqACEduOrNr4NC5TZysRIcRKIcQ+B9v1AFLK56SUMcB84E+OunBwzNGKBSnlXCllXyll38jISEdNNOWRekS9PRdmK28edxYiN30Md/9Q9vjvX8PGtyz7hiorvG3dCpGMU8ozC8oGxnW6Fi5zMMkFt6xeTq38CyqfmLWdyEi3bk1A07IxKQd+VJ9TlnEw04Opn25jwuwN7E7M4OkxHVj7zAgGBKeqTAaJW9VqJWF99YQegKc3ePtbUsFknVapbHyDoN1o5UDh4QXdby2rAnSEn9mV2h2rXDYy6kyISClHSSm7OtgW2zVdANzooIskIMZqPxpwoyx7boKUSh+cerTqfRxZrmpylBQqPb47CxFvP8cxLEeWK/WIQXAUXPO6yhvl7Vc3sSJSWgINQQVvfjcV0k+o/Sufhz73lL0uqKUqrlVVjIj337+0HDMKUlkTe7l6YTDXOJFSUrhzAfmBMTyxyZerZ61jy4l0nh7TgXUzruThEZcR6OsFKWbPvciOaqX1yTWwe37lxvjDI7DkGdtjvkGWYMOsZMtqsv1YMJWo1fGE2ZW7j6bWcQt1lhCinZTyiHn3OuCgg2bLgH9ZGdNHA8/WxfgaFPkXlMvrry+VreftKimHoEm40pm7+0pk1avQoqt6q7cmsDlkn1MTuRAqXqTfVLXvYgnWcjnzuzIKB7d03sZQIRlCJC8d9n4DvScrg6+pRL2B23PZqOrp8o3VxZk9mAqyWXUij2a0pcDkx7rlh8gpLCUjv4jCnMvJD+nJ2Q/+IK+ohJTsQiKKJxAjUng6+z7GtL+BAbfOINTfx7b/lINqVRDeRgUBxg6CPV/BFU+UffM3fn97knaUNXr7BltWIplJEGLWVkebvccOLVGlZjVuhVsIEeDfQogOKBffk8B0ACFEX2C6lHKqlDLd7Aq8zXzNP6SU6fUzXDfG0HH3rYYLa+phS+bTYc+U37a+2fwO9JzkWIiU5FtsJhcSlEqnRdeaue+CW8AnAKatdlw7HZRK5anDFkO6EddQkKG2/4uDsf8HA6fbXtdncvXGZpXB+MG3v2NZaiSgsvWKI0cJ8PEipIk3of7eBPmF0srfkwivUvw6xNC2WScuiwyk27dz6RGRBvYCBNRLRnhbSxR591vhp8eVCtE64/Px1fDd/coQPvplS6BjznmlrovuY9uvdXXDuxdbVoueXurve4lUCmxouIUQkVI6Ul8hpdwOTLXa/wj4qK7G1SDJM8vVjuOqdr2U6k3TCHizzklVX2yYBV5+MOAB2+Mmk5p0HBnWjYDD7HPq/KZ3lJ1n+jpY8QIMeqTS8REXyc+wvO3/+Bjc+GH5unfjnF+o5XpjsvQJcHxNaYn6rETqkx0n09l4NI3CoyVcRxztSaCl6RyzbruKKzs2w8/bEy8PYRv0B/DF7Wq16d0TWlwLcYNVdb6MU45vlHLQNntu91tV3Y6F06A4V6Ul2futypgc2Fx59/kEKiP82v+qrbRQrbisufVz9XeGssGQN39afzE1mnJxCxdfTQ1irEQKslTFv6K8yl2fm6omlAjzSuTCSRVXIR36MNQNK/4GvzxTdgyFWYB0bGgNbKZiIYzfI/e8yskkS+GPxWq1VVXSzTXQ44davJ8c8fvX8OUkS2Zam5WIWW3jaOynNsM/I+HkBpeGU1hSyos/7OfGdzfx+orDLMm5jI/aKaeCvw1uwvU9owh6pwfem98qK0BA5e86tQm2vKu8oABCYpS9wxGTf7Kt2ufjD3ctUivCqD4qtui7+yCmPzy4QVVUNFa03k2g83Xw8LaykfkhUcrYn5MCK1+y2F5AxYo4Uv1p6h23WIloapCo3uqtzctXTQrNOlVOPRIQAU8eAk+zquKPRWoSf/Z0WRfRijizR1X0q66HTMseqq+UQ7YZWR3VEjGIGwIvpFoC1XJS1FuxEdxWnViRpu3gzu+Vy6mRtsOR7j9pu8rrZfyWPkGW+zsqSGXgH6Heul3I5nsqLY+HF+xk7+lM7h0cx5+vak8wZs+s/4TikZmohFjWaWWDcUTvybDxf8qY39WsFAiNVc4Jjp7LkR3I2w8mvGP5LSa8q7LsevvZrigGPez8YY7+qoRYVB9YPxPih9R9QSlNpdFCpLER1EIZH6WEZl1U1G/vu12fyIWwqILAosfOv1A5IXJsFXw2Aca/odQb1eGWz2BWd6VjtxYihdmAcCxEDOFhkHMOmndRk6unb/luvsUFavJzhl+wrWvuH4vVhHu9nedQ1mkIibb89h4e8LTZa86odOjrKMbFPEnb1RXJLyrlZHoux87nsjvxAjtOXmDf6Sz8vD14764+jOli/ru9MwzC4+HRXervZ6zGfJzYbrx84MaPVECi4ekW1Ud5kRXn2arckrZDwjrlpODMFiQE9LzD8bnyOLoSdn4G175p/h0choFp3AwtRBobZ/eqCT9+qLIh/PioUlW0vty16/d8pXI6DXxQ7VsLkdAY59fZY0RcH15ePSGSn6Eiopt3UxOaNS26wt/ScRIupALuWvZQE1rueQgcoSa4AAdBdgYZp+DNbupN2tlEeHiZWum1Ga72d86z2KKssc8ibE1orMpLFuJgovQJAN8QMs+f4qctJ9l8PJ2dJy9wOsPiluzj5UH3qBDuHRynsuKG+1uuzzoNrQdZBEJ5qx6DmH62NqKuE9Vmz5HlsPY1lZ69pvENVinrDVtMNVPUa+oGLUQaG1veU290Tx6EbjcrW8LBn10XIvu+U54zjoRIZTBWLVnVTN9x7Ff4dopKW+7Is6q8HEoHlygh1PMOuOE9y6QU2cF5Jt/dX6hPo4ysI1a/qn6XNsPVfkAzOO/AKz3nPMQMsD224u9KkI16UWVItuNCbhGLd5/myuJg/ti9j+e27aNZkC/948O5tV8McREBxDcNoH2LQHy9PMtcT1GesrkEtVSR3nu+siSbrKw6Esqqs1IOQlh8+Su1qmIIuZRDSqCUJ/Q0boMWIo2NvHQV4wHK4NlmhCU/kSvknLMkMISqC5EuNyh1U3Ur5xn5noxVkMlkERwJ65XxetSLjgMOg5pDjjmSusPVluN3LXR8L5NJBc3FD4X+9ztuIyWkHVMeSQaBzdRKx37CDY6CyPa215/ZA0U5UJhDZn4RSbkeHEvNY/epDHYlXmDf6UyKSyWF4ePo3CmKVSOGE9fU37FB3BGGHSU4Sqmjdn8O3W6ErjdZIuddobgA/tdTvUxY5yBLOaSCDGsDw8kg/bhWZTUgtBBpbOSn206od3zpvK0jcs7ZvvGHxSmf/eZViK/ofXflr7EnM0nZDbyawOwBylg7fIY6d3Yf7PxUCRFHBDZXE35uqkrsF9O//BoUJzcou8CVz6so7sJM29TkoPoqzLIt5hTYTKncCjJsS91OKZsht9A7mKzTx1j06sPcI3/kmsJ5gMDP24NuUSFMuSKe63tE0bnVNa78OmUx7CjBLZV3GqjPmyqZFdjbD0oKbFPXlxZD2lHoUEX38YowVh7XvqlWO5oGgRYijY28NOWRVRVMJqWCCbTS4/sE2FaVc5X5N6txXPGEmnRDq5gIM9OcNsTLR8UQHF9tESKGd5avk1xKgc2VYEjaBl/cpoocRfdRxtvfv4J7frJtf3avWsV1HA+z+yt7ym126TzSzIbxpm0tx4JaqtVbQaaNECkuNZFTUEJRqYmiEhPL9p8l+FAuV8oMOoVDaU4Q797Uh9im/rRvHoS3p5VqrrREraKCWlWc9tya4FYw/Fnlom0qVsfST1TtbxhqFytiuPzWlsdUh3Hwl0Tl/FCZZ9bUK1qINDby0m3VV1vmKlffR3ZW7KFVkKHaBNoZgw/8pI5VJjgvaZuKNXj/SiVM7CdjV8lItKQAaTMcNs02JxMMVMLJO8B5QF5IlErqZ7ydG8Wqcs8rD6OiPKXyMxj0kMpl5eOv8jXt+lzZRqy9k4wYEWsh0nUiSVFjWXs4lbWHd3DkfDbNs/9ghul9/lo8lf0y7mLT2c2a0jQnnyuifSAxhKu7OUmbsuNjlQr9yUO23nIV0bQtDP+L+m4qVVlvf30Jlj+v/g04M/Q7IjRWFSczCG8Dz52tvaA/L19AwKLpyo7VZnjt3EdTo2gh0ti4/Utbl1dZqnTMeemWmAZn+IfD8yll4wmWPKWii10VIkW5yoYSEq1USEdXOs+hVBGj/m4xgrcZDhveVN5mUX2UG3F56qkhT6pt7Wtq34jRMKrp5aWCj3mFVFKoJjFDqHS+Dra9D0dW2OZr6nYzRPcj0RTJ2i0n2Xkyg52nLnAiVRnio0Kb0D06hGFNS+mRcJxpwy4jN6wTXp6C1uH+DEg7C9v2qtT05WWjDTHbgE5udOwl5YzM0youJTBSuTlHtLMEEFoLTFcIbQ1H7P52tRnwl5eu0qf8sdg9MiVoXEILEXekIFOpbapSa9t+ojfqVWQnVyxEQKkRPOzyJVU2CWOmOXdTSLSa5Pd8oQSZ9du7q1gXIIodqGI8jq9WdSKCWqiaGRWRc14JVsOjKMAsRHJTLWq2T69Taprr/me+1+VqRXfgB1sh4uXLT2eCeOqb9RQUm4gI9KF/tD+f+L2Hb69baX75OGUE37oTEuD6K3pZVkAAbe5T3lIfj3McI2LQdoRya17yFLQe7PoKYukMlZn34S1q/8GNsOpfSpB6O0mx4oy4IcrWU1qk7CqfXqdWa/Z5ymqKkkIlQEAb1hsQWvHojix6CL6+WxmFK0NeuspTlGlJwHfRrdWV1OIn1sIPj5YVGH6hFvuDKxi685BoiBmovidudf16g9w0tdowUoR4N1H2kLghStjd+R10u8n59RmnYP4tsH+hcsM1uLgSMXuOpR2DxM22xnJPL+h4jYpzKVGV+0wmybqP/8q3X35E11Yh/PrkMLY9N4p3Jl9O6/SNtMg/YvGiyjmnJl5nnnG9JztOA2/g5Qs3fqBWdYsetK0yWFKoEkqe3KjqfxTmWM5lJdsWuhLCkga+snaGDmNh3GtqLCfWwqmNFmN9bWC9MtNCpMFQrX8RQgjXkvtoKochPMqLVXB43VE14RjqC7BdiVTE6Z3K28nDTmVR2ZWIT4AqJBTaWrmD+oaoSbqyJG5RUe9pRyzHhjypJjeoWD3m4QVHlkG3W1TwoEFQC2U09zAvxI/+qj7t37CveAIeXI/Jw5tNx9KY+skW+ia8z70tTzL//gG0jQxUQkMItdowCl+B+u4fUTZy/uxe+OAqpWbqfnP542/WEca8ojzU8lIhYYOKmXk1Gmb1gI+vhv2LbN2os87YTsDHflM2MWcpTyqitFjZjvYvVBHvtVkp0NtK3eYoCFPjllRXnaVDSmuDMa/A5xMr9/YPlqhpaxffoBbQbkxZY7kjcs6pN1b7oLQmYXBmt+vjiB0Ik76x7E98r6yrrCtcXNFU0bMroBkg1PNYq/nCWsMDay37x36DsDhkWBxpOYUcO59DSk4h6bmC5IwCluxdzan0PC7zzaCJKGLowIEI+0A/eyESGqtUUvaYSiFpq/Iaa9rWccoWa/reBz3vVKq4FX9T9qXek5UQDG6lVk+G44Hh0WWd28pIwOlqsKk1hdnw71jl8nzgR5UZujaCDA2sXwp0oGGDoUIhIoR4C9hr3vZJKbOtTtdjatdGjKFDr7QQMb+RWqtQPL1h0teuXZ9zzlZ/bzDsaduAs4qwN6JbB/pVhoxTKj7EsGFUFk8vde2615V9qXmXi6cKiktZdySVE+cucPfRNazzHc4zL6/gQl6xTRetRBrvBn9CxqjHGRjbChaAsFZ7GQQ0s02YOPQpx2MyHAGWP69WdyP/Vv4zCGGZuK/6h8pFZh1tX1oMJ9apZyspUJ5T1ulCws3xFr0mlX8fR/gGqW3nPOW51+WGyvdRWUJjIW5o7d9HU2O4shLZC3QHJgFdhRBZWISKfl2oDU5tVsWKOo2v3HX55pVIEwfR2654R9nHiBjYV6CriPk3KQP47QvUflGeSr1iKoYet7vupZWZaJvAsCr4hSpPqENLoHkXTmfk8/nmkwza/CDHSqJ4o+RGMvxvJ7NJZ8a2a0m7ZoFc1iyQFiF+hPn7EOZVhNdbL8GZjyHE/Pdw5CDQrBMuvVP5WXmTVfZt25GAP38APh2vkj92uhaue0s5BRgYUepGSd7KEhqr3Ky73KBSxtc2j+yqVA0VTf3jqhB5X0pVzEEIEY0SKt1QJWurjbli4fWoyobngXuklGWU+EKIUvN4AE5JKa+rifu7HXu/UcZM+4p3FZGXpuwZ9pPToofh3F5bFY4jhLC4llqTehSOrlC++xWpX0AZfa0j3D29lYfWsV9VavTxM8sWYyouUM9sLTAyEiuX9NG6u1IT64+kEp0fSDvgpVWpzF/+C0UlJjwE3B6URZdWIdxx93iC/cp7w/ZViRJX/l0JYq8mKgDQnqtesnyXUiVxHPhg2dTn1oGRzoIkK0OLbmo8h5dBrzvLZgkwVJO/vex8dVQeoa2VmuzmT6o9VJfQAqTB4cpfbDIwWwhxGFgKLJVSLgGW1OA4XpNSvgAghHgU+BvmErl25Espe9bgfd2T/AxIPaTyQnW/xfXrBv1J1YOwf3P38nVepc6ayT86Pn72d1j6FxWnUZEQkVIZgtuPtRzz9FY2krX/VckLz+yG2xbYvtF/f78yQo9/UwkYT2+4dpaKc3GBguJSfk/KZHfiBXadymDT8TQy8oq53O9mFrCLy+LbcG9kHOH+Pozv0YqoH95XOv9Tv6m6IOW50PabqupteHjBM8cr9nIqyFCrKEeFvDw8lLND9hnXBHJFCAHtR8Pe75Rrb2G2eh7rMU5839Zjq1L9e0DKgarH+WgaPRUKESmlUe+8I3A18IkQIgRYhRIqG6R08X+683tkWe0GcKnbWgoy1OeJNZUTIgERju0HwS2V/r2iOhnOqEwSxrx0pZsPibY97uGp3HNjB8A398K86+GhTWrVdOgXFY/R6Tr4bzu4ZZ6yo7TsXqb7k2m5bDmeTmFJKYUlJk5n5LPz5AX2J2dRYlL/bGLD/bmyQzPGdWvJMCR8DZNG9oUoq3QwAREqGeIXt8MVf4aRLzh/Jt9AJaB/fUmpj+xrg4Nyt136F5j4geWYI/UTqDKwH4ysOeNxuzGw4xMliM/th+fO2Z6vzL8he3pPVvE0plK9StA4xOV/FVLKg8BB4A0hRBNgBHAzMBPoW92BCCFeAe4GMs19O8JPCLEdKAH+LaVcVN37uh1SWibrvEpmzt3zlXq77TDW9rihfsk+YzG02pOVrGJEhjxR1pOnMkLESP1uL0QM2gxXCR2TtqlJtDAHljwNkZ1UuvY3uih1XnQ/OLxURcqb037sTszgrg+3kF1gcVf18/agR3Qo04a2oVdsGL1iQ4kI9LXcb+UO22cw8I+w2JAceVHZ0/9+9SYe0c7xeSmVUMo6bXEddiZEQqLhmtdtDP3Vos0wZYNK3qVsGDWZd6rdKLVpNE6o9KuFEKKllPIMSp3lskpLCLEScJQE6Dkp5WIp5XPAc0KIZ4E/AX930DZWSpkshGgD/CaE2CulLBORJ4SYBkwDiI2tontofWF42EDl06+vnwkR7csKEcPlszwhkpmk7B4DHih7rjJCxDtABdFFlpMEsmV3yyrjy9uV6mfKcpWWo8sE2POlMuQufhjuWQJBLdh56gKTP9xKWIAPXz8wiIhAX3y8PAjw8cTLs5xJc8iTKuLb3sXYuL+XH0T3r/i5fIPUisUZF/NypVhiQ5y5Va/5P/V2329qxfd1BZ8AVcf8pye0yklT51TlleUVACHEJCHEBiGES3mhpZSjpJRdHWyL7ZouAG500key+fM4sBro5aTdXCllXyll38jISBcfy03wbqJqg3e4xvKm7Cp5aY7rajS9DPrcW/Zt3JocswrE0duz4Zaan1HxGCIuU7aMCAdusPYU5Sr1S/9pSs0Fqu5FcZ4qrgUkE8HPv5/h7g+3Eh7ow5fTBtKpZTCRQb6ENPEuX4CAUkU5epPueQeEt1W1Q7x8yp6vLMbvlnNO5ejqdJ3zxInnDyr1XWkVAwAd0byLSixZVduHRlNFqqLkNGaS0cAVwPtU08guhGgnpTTCkq9Dqc3s24QBeVLKQiFEBDAY+E917uu2GHXO04+7fo2UZTP4GoTGWupWOyPbXLwp0MHE5xsMj+52rp6xpjBbRR7bR2o7widAZZb1CyE5I5/Xlh3i+PlS3hcRNEtYR4n0YMicQ5TiSVxTf76YNpCWIU4qElaWrDPK2aCmVgO+wUqllHNOCab4cmIdTm1Un8W54FkDxnVQAin9eOXdsTWaalIVIeIlhHgeSJRSSiFEJXNzOOTfQogOKBffk5g9s4QQfYHpUsqpQCfgPSGECbWC+reU8g9nHTZYzh+ETW+plBvjZ7p+XUGm8mRylqvJVKre8J0Zc3POK08cR4Z5IZyrwez54VGV2uOR7S41L/UN4ZMNCby+/BBSQv/4cL7yfZrr0z8mxJTBv8b3pHXTALpHh+DvU4OGXR9/FYBonVyxOggB7a5S9qeKPJmMOh0+VShX6wxPL7hmJsQPq7k+NRoXqMr/yqdQq4CN1ejDBimlM/XVdmCq+ftGVGxK4yb9mKpjUdk3ZCNa3VGgIcA7A5XKw5m/v3cTaNXb+Qpi5zwVI1FRvqes0xfVOFJKzmQWcOhcNsfO51BQXEqpCUpNJlJziziXWcDRlBxOpuUxvEMkL1/flZhwf6A/fLQYCOfWfrVk0/ILqXy1v4owaqbMv0UJbPuiVwZTlikjvCurtcpg1FLXaOqQqgiAd4AHgHeFEB7AFzU7pEscw+5w/qCKq7jmddeKEoXFwZOHndeMCGxefibfIU+Un1Z95zz15lyREMlMgvih/J6UwX2fbiclu9Bhs/AAH5oH+9E2MtJuzXQAAB2fSURBVJCnRndgfPeWtnXEb5mnapE3RHLO2mYNtie4lW1qEo2mAVMVIZIrpSw1pz95CWWX+LVmh3UJY3hASRMc/AmGPu2aEPHwLD9gLqhl1TLpGjQJs00w6IjSEsg+Q75/Sx78fCc+nh68fH0XOrQIpl2zQPx9PfEUAk8PYSswHBEYCTQwp4g1/4F93yvVYlVq0ms0DZCqeGeFAkgpnzInY8yr2SE1MGb1VAn+aoqCDEBYMrO66uabtB1WvWqpvWFPcEtlPHcURQ0q+G/D/5z370o6+BxVOnXBQRMpOYW8e2dv7hoUR//4cMICfPD1Uu64FQqQhoqpREV3Z59xzQlBo2kEVEWILBVCzBFCtBNCxHMpp4OXEi6cgK3v12yfwVEWA7mrbr6nNsOafzuvfx3USlWos649YWAyqVoV5QmJJmHK+6s433kbLz/WxzzAV2db8vL1XegeXU7p2saIUX4XWb46S6NpRLiSCr6LlHK/sS+l/EIIcQp4GhAoldalSaE5K/7Ah2quz5EvqC3bHLfh6kokPx2Ep/N8TK0HqboQjoy5+RdUht3yao7EDUHums/Bg/v57Kgvm46lkVtYQn5xKX4lWYxkG9+YhlNqGsZt/WJqzyDuzhi/X+srVA14jeYSwBWbyGdAbwAhxFQp5QdSyg3ABiGEv5Ty0lVnGTmumtTCG3eTMLUiES568BiBhs5URS17qM0R5QQa5heVsvPUBTYntmOL72y2zj9HE28PHm/1BzFBqcQUHOGy7C34luTQttcYRHg8dw5s7dqYGxvG73fFny3BkxpNI8cVIWI9Kz0EWGWYYx1w6b5yGZ5UPzxSNgV3VVn6rFKLDHkCnqhEGExemvMYEVBqsqxkldHXPhbkohCxrESKS0288vMB5m85SXGpxNND0DMmklcnRjPBcwNNfvyHahgcDe2vhKFPc3+LS9yYHBKjshd711BApEbTAHBFiFhbYu1fc2sw01sDpDaMp0eWqxoRlSU/o3whYipRyQ2HPQMj/mp7zstXFTIy17VOyynkwfk72Xoindv6xTCmawv6tg4jyM9ce70oAiLbqHQqAeXc81IjJEpFqn8yDmYklJ9mRqNpJLgiRFoIIe4B9lBWiFzaKduDWsCA6bC7BkNl8jMsk8+SZ1RdjTGvVHzd3T9ASTlGb09v5eabdrTsudaXw5RfADh4Nov7PtlOak4hb97akwm9osq29/HX6hpn5JxXhcH8LjGnAs0liytC5EVUqvd7gWghxH5UbquDQBWLXzcSctOUUbowU6UVqW4EspTKzmJMQCkHy/eGssbDo2y1QHvih8KRZSqew0FtiITUXCa9vwUvT8HXDwyiR4yeCCvNBnOOssbqxqzR2OFKUaq51vt25XErqLfayNn1Gfz+lfpekOk4g25lKMpVaifDUO8fDmf3VXydlPDTn1VN9svKqf3Q4Wr4/UtI2mpbM+THxylKS2ByymOYpGTB/YNoG1mDeZ00Gk2jpdIR61LKJCCJmi2P2zAxvLP6Tyv75nlyk7I1RPV2vb/iPJWe3Ejn3STMtTiRvHTY8bEqN1ueEGl7pVK1HFpiI0RKk7ZzJM3EucICFtw/UAuQ6nD/byruRqO5RND1LqtDfobypBr3WtlzH5sLQ72Y6Xp/gc3g0Z2W/Sbh5hgOk+NqdXnpsHXuxdobFaba8AuG2+YjW3Tnpz3JrDuSgveZnbyStpdviifz9qTe9I7VxuBqoeNDNJcYWohUB8N+UVIIiJopbmRNeBto1UsZzO3tHXnpMHuAKkTU/mq44nGIHVhhl3lxI3l+0T6+33mapgE+vOm9iHyPAIbd8hgjOpcTbKjRaDQOuLRddKtL/gWVj+qfzZSKyBHOclU54sRa+GQ8XEhQ+70mKfWII4O5fzg8vhce2gJ3fOmSADmWksMNb2/Ab8883uqZxNZHOjOkaD1N+k9mRI+2ro9To9FozOiVSHUY9Cc14S95ymIfMXjykDKSV4YLJyFhnSoO5QreftCsY4XNCktKeX/tcd767SgBvl78NXIDgfkHIeBuVciojS5kpNFoqoZeiVSHdldBj9vVd/v640EtICS6cq6eRp4sw8X3/AF49wpIWF+27bLnYOPbDruRUnIht4iDZ7P4+fczXD1rHf9dfpiRnZqx5NEhBHa/ViVsLM6HPpN1SVWNRlNl3EqICCGeEkJIcw11R+cnCyGOmLfJdT2+MpzaAoVZ4OGlXHwNzvyuUsR/dRec2+/8ensKMlSuLKOErYcXnNtbtpiUlLB7PqQeKtPF7sQMBr36G71eXsHYN9fx8IKdlJokn9zbj3cm9aFFiJ+yochSmH+TihnRaDSaKuI26iwhRAxwFXDKyflw4O+owEcJ7BBC/CCldDHNbQ1jMsFHY1TRKL9QW3VW1mmVIv7CCWg7QpWldYX8DBUjYqxejMh1ezffzCS1amnR3ebwrwfO8fCCnUQG+fLC+M40D/alebAf3aJC8PO2CoRs1Ut9nt5R8yVaNRrNJYXbCBHgDeAZYLGT82OAFVLKdAAhxApgLPVVnrcwE5Bq0r/8EZVHysC6ZkdWsuPrd34GyTth3OsW993AZhDdz9LGUGvZp4M/+7v6tMrK+8XWUzy3cC9dWoXw0T39iAzydT52Dw+YtkaVu9WR1RqNphq4hRARQlwHnJZS7imn6l0UkGi1n2Q+5qi/acA0gNjYWqprYdhA/EKVF5U1uanmcyGO65rv+BR+fBRaD4adn6qVQaueMPwvtu08vcA3RLnzWnNmjzK+m1c4n21K4IXF+xneIZLZd/QmwNeFP2urnhW30Wg0mgqoMyEihFgJOCoW/hzwV2B0RV04OObQf9acqmUuQN++fWsnSaSxOmgSBoU5KmWJUeM8LxW8mkDTdkq1Zc2+75QAuWwU3PQRvNYO+t/vfFJvN8pSKtfAw0sJIJ8AvtmeyAuL9zOqU3PevbM33p5uZebSaDSNnDoTIlJKh/k4hBDdgHjAWIVEAzuFEP/f3n3HSVXeexz//KhLr4vA0kRREKS5EiyJjSgBr+2l15LY7vWam6tJzDXRFJOYZoompujLiEYliTWmSCxYUa81rEpHEXSBdSlLcSnLsrvs7/7xnHGHYQaWYWdndvb7fr3mteeceebMcziwP57zlN8kd18bV7QMODFufxDwYkYq2xjxCalmfTk8YvryW+FYtwEw/MSwcu6GZbt/7p0/h0df598fhugWHQWrXg/v/elsKCqGk7/TUP7ce/b87hOugxOu4/EF5Vz/1wV8ekRfbrtoggKIiDS7rD/OcveFwCeJOcysFCh29w0JRZ8GbjKz2LocpwLfapZKJnPQGDhvJhSODIEkfojvMVeFV3UltE+YKLh5Zej3aF8Q9oceA6/+JrRkyueFWeqNsHz9Vq55aB5HDe3FnRcftXvHuYhIM8np/7qaWbGZ3Q0Qdaj/CJgbvX4Y62TPiq79YPRZYeZ4QY/QMkmcnV7QY88l168ugem3NOwPOTZMSlz9r92XgY955gaYcVLD/oo58NuJPD3nRdqYcccXjqJzh6z/X0BEWqmcCyLuPizWCnH3Ene/Iu69e9z90Oh1b/ZqCVQsgxUvhMBR0DMEgtoo3fwfToPnfwjr34UnroWP40Ytt2nTMA8EYPDRoZN81Rvg9Xvma6+rgU0rGvbXzINNK3hwaQ2njelP3657GYUlIpJhORdEWox3/gQPXBC2Y7/4Y4+01i0Ks8F3bIK5d8OG98Px92aH/pOd2xrOU9ADvr4cxl8U7ScEkU69wmOx2KTANQvY3mkgZdUFXDQpQyPPREQaSUEkXdVxEwMHfwpOuykslFhbDTXbQr7z7gND2dhckQ9fggV/gfaddz9Xlz7hPIecAr0P3v29WKKrWEf+2gUs2jWU4YVdmDz8AJNgiYgcIAWRdMXnQu83KnSkd+oZhvdCCCKx5FKxILJ+KRQetmdukI0r4Olvw5Tvw7Djd3+vS7QCzMJHYec2fOMKXt1exEWThrCXOTUiIs1CQSRdOzY3PHqqqwl9JDs2N0w07NI3ZDbsUghboyBS8S4UjtrzXB26wNJ/hqXgEx0+DU75Xlg+pbaKt/tMp8RGc87EQZm5LhGR/aAgkq7Y4yyALWVw+9Hw3lPQrgBGng69osdSPQZDTVVouWxdk3zp9m7RHMxnboBt63d/r30n+PS1UHg4lW16cdnGSygccxK9uzRxAiwRkTRobGi6zvhdWHEXGlok1ZUhSFxwf0O5/3oh9HdULAuBpd8+FmNMkoBq0/Ya7nnlQ2a+Xsr2nXVccsywJrkEEZEDpSCSrthKuBBGWMGeOUWgYYHDwsPgq/NSn++LL8OSWbt1ulfuqOXOl1Zw76ulVNftYuro/lx10qGMKerRBBcgInLgFEQayz0M2+3QGep3wYJHYFAx9B0RllPv2D20RObcFBZYvPbdEEDefy4M8z3v3vBoKpUB42DAONydtZU7eGLBGm6bs5yPq2o5Y9xAvnLKoRzar1vqz4uIZIGCSGP9/viw1Mk5d4Zg8Y//hqk/D0EEGnKK1FaFSYOxFsj2Clj2FPzxzLBO1tSfJj390jVb+MkTS1lcXsnmqloAPj2iL9dPHamWh4jkLAWRxih9NUwg3F4R9j9ZwTduYuBnfxA6yF+/PQzvjYnNFVn95l7XxfrR40tYXL6FqaP7M7qoO+MH92TsoJ4py4uI5AIFkcZY+Vr4uW0dVG9p6PuIzRMBGHNO+Pn8DxvmdgB0j0t5UphkZBbw3tqtvLZiI9dPHcmXTjykCSsuIpJZCiKNsbm0YXvNPNhVE7bjlyjZXArbN4Z5IvHpcLsPaNjul2SOCHDfax/SsV0bLjh6cJNVWUSkOSiINMbm0pBgauP78NHb0COa6Bf/OOulm8OCjOPOh95xrYn4IbtJWiKbt9fw93c+4uwJRfTS3A8RaWEURBpjc2lYjmTs+TBkcmhRXPkS9IzLONipZ+hwn3Ljnp8/bya8dV+YeJjg4ZLVVNfWc9lxwzJTdxGRDFIQ2ZdddeHxVa9hcMI3Go4nprMt6Am128Ps9A4JCyyOPiu8EtTtquePr5VyzPA+jOzfvenrLiKSYVr2ZF/atoNvvA8nXA91O6HsLXj3yZDmNl5swuFNA2DJY4069dOL11FeWa1WiIi0WDnVEjGzrwM3A4VJ0uNiZruAhdHuKnc/o9kq16YNrJoL902HrgeFJU8mfKHh/fj+kcScIEms2ljF9x5bxKH9ujJl1EEZqLCISOblTEvEzAYDnwVW7aXYDncfH72aJ4AseQweuTQkkhowDrAw1Dd+eC/A0ONCnwnsPsQ3icodtfzHzLnU1Tt3XVJM2zZa0l1EWqacCSLArcB1gO+rYLNa/S9YNjusadWxGxQeHo4nprHtUQSDJ4XtzqmDSO2ueq66/21WbtzOnRcfxcF991xwUUSkpciJIGJmZwAfufv8fRQtMLMSM3vDzPbsqW4435VRuZKKiooDq9zm0jAKK5ZIauDEqCYJQaR2B8x/KGx3Tp1x8NZnl/HK8g3cdPaRTB7eJ2U5EZGWoNn6RMzsOaB/kre+A3wbOLURpxni7uVmNhx4wcwWuvuKxELuPgOYAVBcXHxgLZvNK8PIrJiiiTD/gZACN151JZTNDR3sbdsnPVVVTR1/emMl08cO4LxiTSwUkZav2YKIu09JdtzMjgQOBuZH6V4HAW+b2SR3X5twjvLo5wdm9iIwAdgjiDRhpUNLZOixDcdGTg/9IYM/tXvZWMvk2K+kPN3j89ewtbqOS5UPRETyRNZHZ7n7QqBfbN/MSoHixNFZZtYLqHL3nWbWFzgO+EVGK1dbBX2G775cSfeBcOS5e5ZtXxB+bl2753uR+99cyYh+XTl6WK+UZUREWpKc6BNJxcyKzezuaHcUUGJm84E5wM/cfUlGK9ChS0gWVXx54z8z966khxeWVTK/rJLPf2oIZhqNJSL5IestkUTuPixuuwS4Itp+DTgyS9VqnGO/nHKl3vvfXEmn9m0556hBzVwpEZHMybkgklPenAELH4HLn0rZWb6bU3+c9PCW6loem1fOGeMG0r2gEecREWkhFET2Zt3CMDqrMQEkwZbqWhZ9VMm26jpeXb6BHbW7+PzkIRmopIhI9iiI7M3m0t2H9zbSjppdnPG7VyjdWPXJsYlDlKlQRPKPgsjebC6FwZP3+2O3z1lO6cYqbj53LCP7d6dLx7YM7Nmp6esnIpJlCiKp7KqFyjIYO2y/PrZ8/TbufHkF50wo0oRCEcl7OT3EN6t2boXDPgcDJzT6I+7Od/+xiE7t2/KtaclT4YqI5BO1RFLp3BsufGC/PvLYvHJe/2AjPzprDIXdOmaoYiIiuUMtkSayelMVP/jnYsYN6sFFkzQKS0RaBwWRJrBtZx1XzCyh3uHXF0xQfhARaTX0OOsA1dc71zw0j+UV25h5+STlBxGRVkUtkQP0q2eX8dzSdXx3+iiOH7H3jIYiIvlGQeQAVFbV8vuXVnD2hCIuPXZYtqsjItLsFEQOwHNL11FX71x67DCtzCsirZKCyAGYvXgtA3oUMLaoR7arIiKSFQoiadq+s46Xl1Vw2uj+tNFoLBFppRRE0vTSsgp21tVz2uhkaeNFRFqHnAgiZnajmX1kZvOi17QU5aaa2XtmttzMvtnc9Yw3e9Fa+nTpwKSDe2ezGiIiWZVL80RudfdbUr1pZm2B24HPAmXAXDOblfEUuUnsrNvFC++u5/SxAzSxUERatZxoiTTSJGC5u3/g7jXAQ8CZ2ajIa8s3sm1nHaeN0aMsEWndcimIXG1mC8zsHjPrleT9ImB13H5ZdGwPZnalmZWYWUlFRUWTV3T2orV069iOYw/p0+TnFhFpSZotiJjZc2a2KMnrTOAO4BBgPLAG+GWyUyQ55sm+y91nuHuxuxcXFhY22TUA1O6q59ml6zh5VD86tmvbpOcWEWlpmq1PxN2nNKacmd0FPJ7krTIgPsvTIKC8Caq2X154dz2bttfwb2MHNvdXi4jknJx4nGVmA+J2zwYWJSk2FxhhZgebWQfgAmBWc9Qv3iNzV9OvW0dOPLxpWzgiIi1RTgQR4BdmttDMFgAnAV8DMLOBZvYkgLvXAVcDTwNLgUfcfXFzVnJtZTVz3lvPecWDaNc2V/7oRESyJyeG+Lr7xSmOlwPT4vafBJ5srnol+uvbZdQ7/Ltyp4uIALnTEsl59fXOw3NXc8zwPgzto5whIiKgINJob3y4kVWbqjj/aLVCRERiFEQa6eG5q+le0I6pmmAoIvIJBZFG2L6zjqcWreWsCUUUtNfcEBGRGAWRRlhcvoWaunoN6xURSaAg0giLyysBGD1QyadEROIpiDTC4vIt9O3agX7dOma7KiIiOUVBpBGWlG9h1IDuyqMuIpJAQWQfaurqeX/9Vj3KEhFJQkFkH5at20rtLmf0wO7ZroqISM5RENmHJeVbABRERESSUBDZhyVrttC5Q1uGaakTEZE9KIjsw+LySkYN6E4b5VIXEdmDgshe1Nc7S8q36FGWiEgKCiJ7sXJTFdtrdimIiIikoCCyF7FO9SMGaHiviEgyORFEzOxGM/vIzOZFr2kpypVGGRDnmVlJpuu1uLySdm2Mw/p3zfRXiYi0SDmR2TByq7vf0ohyJ7n7hozXhrDcyaH9utKxnVbuFRFJJidaIrlqcfkWjlB/iIhISrkURK42swVmdo+Z9UpRxoFnzOwtM7sy1YnM7EozKzGzkoqKirQqs35LNRu27dRyJyIie9FsQcTMnjOzRUleZwJ3AIcA44E1wC9TnOY4d58IfA64ysw+k6yQu89w92J3Ly4sTC8HSHVtPdPHDqB4aKp4JiIizdYn4u5TGlPOzO4CHk9xjvLo53oz+zswCXi5ySoZZ0ifztx+0cRMnFpEJG/kxOMsMxsQt3s2sChJmS5m1i22DZyarJyIiDSfXBmd9QszG0/o8ygFvghgZgOBu919GnAQ8Pcop0c74AF3n52d6oqICORIEHH3i1McLwemRdsfAOOas14iIrJ3OfE4S0REWiYFERERSZuCiIiIpE1BRERE0qYgIiIiaTN3z3YdMsrMKoCVB3CKvkCzLPiYQ1rjNUPrvO7WeM3QOq97f695qLvvc8mPvA8iB8rMSty9ONv1aE6t8ZqhdV53a7xmaJ3Xnalr1uMsERFJm4KIiIikTUFk32ZkuwJZ0BqvGVrndbfGa4bWed0ZuWb1iYiISNrUEhERkbQpiIiISNoURFIws6lm9p6ZLTezb2a7PpliZoPNbI6ZLTWzxWb21eh4bzN71szej37mXYpHM2trZu+Y2ePR/sFm9mZ0zQ+bWYds17EpmVlPM3vUzN6N7vcxreQ+fy36u73IzB40s4J8vNdRavH1ZrYo7ljS+2vBb6PfbwvMLO0MfAoiSZhZW+B2QhreI4ALzeyI7NYqY+qAa919FDCZkHb4COCbwPPuPgJ4PtrPN18Flsbt/xy4NbrmzcB/ZqVWmfMbYLa7jySkVVhKnt9nMysCvgIUu/sYoC1wAfl5r+8DpiYcS3V/PweMiF5XElKUp0VBJLlJwHJ3/8Dda4CHgDOzXKeMcPc17v52tL2V8IuliHC9M6NiM4GzslPDzDCzQcB04O5o34CTgUejInl1zWbWHfgM8AcAd69x94/J8/scaQd0MrN2QGdgDXl4r939ZWBTwuFU9/dM4I8evAH0TMgw22gKIskVAavj9suiY3nNzIYBE4A3gYPcfQ2EQAP0y17NMuLXwHVAfbTfB/jY3eui/Xy758OBCuDe6BHe3VGa6by+z+7+EXALsIoQPCqBt8jvex0v1f1tst9xCiLJWZJjeT0W2sy6An8FrnH3LdmuTyaZ2enAend/K/5wkqL5dM/bAROBO9x9ArCdPHt0lUzUB3AmcDAwEOhCeJSTKJ/udWM02d93BZHkyoDBcfuDgPIs1SXjzKw9IYDc7+5/iw6vizVvo5/rs1W/DDgOOMPMSgmPKk8mtEx6Ro88IP/ueRlQ5u5vRvuPEoJKPt9ngCnAh+5e4e61wN+AY8nvex0v1f1tst9xCiLJzQVGRCM4OhA64mZluU4ZEfUF/AFY6u6/intrFnBptH0p8Fhz1y1T3P1b7j7I3YcR7u0L7v55YA5wblQs3655LbDazA6PDp0CLCGP73NkFTDZzDpHf9dj15239zpBqvs7C7gkGqU1GaiMPfbaX5qxnoKZTSP877QtcI+7/yTLVcoIMzse+D9gIQ39A98m9Is8Agwh/EM8z90TO+1aPDM7Efi6u59uZsMJLZPewDvAF9x9Zzbr15TMbDxhIEEH4APgcsJ/JPP6PpvZD4DzCSMR3wGuIDz/z6t7bWYPAicSlnxfB3wf+AdJ7m8UUG8jjOaqAi5395K0vldBRERE0qXHWSIikjYFERERSZuCiIiIpE1BRERE0qYgIiIiaVMQEclRZvaimRVnux4ie6MgIiIiaVMQEdkPZtbFzJ4ws/lRforzzex7ZjY32p8RTeSKtSRuNbOXo/wdR5vZ36LcDj+OygyL8nvMjPI6PGpmnZN876lm9rqZvW1mf4nWOsPMfmZmS6LP3tK8fxoiCiIi+2sqUO7u46L8FLOB29z96Gi/E3B6XPkad/8M8HvCkhNXAWOAy8ysT1TmcGCGu48FtgD/E/+FZtYXuAGY4u4TgRLgf82sN3A2MDr67I8zc8kiqSmIiOyfhcAUM/u5mX3a3SuBk6IseQsJizmOjis/K+5zi6P8LTsJy47EFsBb7e6vRtt/Bo5P+M7JhORor5rZPMIaSEMJAacauNvMziEsXyHSrNrtu4iIxLj7MjM7CpgG/NTMniG0LordfbWZ3QgUxH0kth5Tfdx2bD/27y9x7aHEfQOedfcLE+tjZpMIiwpeAFxNCGIizUYtEZH9YGYDgSp3/zMh2VEsN/WGqJ/i3JQfTm2ImR0TbV8IvJLw/hvAcWZ2aFSHzmZ2WPR9Pdz9SeAaYHwa3y1yQNQSEdk/RwI3m1k9UAt8iZBydCFQSkgjsL+WApea2Z3A+yTku3b3CjO7DHjQzDpGh28AtgKPmVkBobXytTS+W+SAaBVfkSyKUhI/HnXKi7Q4epwlIiJpU0tERETSppaIiIikTUFERETSpiAiIiJpUxAREZG0KYiIiEja/h+T1pRTCGcTyQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_1k_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_pool_retrain_1k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_pool_retrain_1k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_pool_1k/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_pool_1k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 10s 102ms/step\n",
      "Score (RMSE): 0.09719719362809788\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128,verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool_1k_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Retrain 5k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_retrain_5k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_retrain_5k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_5k/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_rand_5k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "500/500 [==============================] - 49s 99ms/step\n",
      "Score (RMSE): 0.34683049896530543\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_5k_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_pool_retrain_5k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_pool_retrain_5k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_pool_5k//'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/retrain_pool_5k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "500/500 [==============================] - 49s 99ms/step\n",
      "Score (RMSE): 0.10310340238569268\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128,verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool_5k_test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Train 1k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_train_1k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_train_1k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_1k_train/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/rand_train_1k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 11s 108ms/step\n",
      "Score (RMSE): 0.45117996807851257\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_1k_test_train')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_pool_train_1k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_pool_train_1k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_pool_train_1k/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_pool_1k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 11s 110ms/step\n",
      "Score (RMSE): 0.11429127339658202\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128,verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_pool_1k_test_train')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# 25k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rand"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_train_25k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_train_25k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_25k_train/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/rand_train_25k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2500/2500 [==============================] - 254s 102ms/step\n",
      "Score (RMSE): 0.7076151220692025\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_25k_test_train')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Train 50k\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rand"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_rand_train_50k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_rand_train_50k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_rand_50k_train/'\n",
    "\n",
    "#writing the columns of the csv in an list\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/rand_train_50k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1000/1000 [==============================] - 114s 114ms/step\n",
      "Score (RMSE): 0.2059529695097349\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='pred_rand_25k_test_train')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Pool L = 16\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/xtest_l16_pool_10k.csv')\n",
    "dataset_y = pd.read_csv('csv/ytest_l16_pool_10k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_16_fortran/'\n",
    "\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_l16_pool_10k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score (RMSE): 0.26912805576136206\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='train_l16')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 20k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_l16_20k.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_l16_20k.csv')\n",
    "\n",
    "path_folder = 'data/matrix_16_20k/'\n",
    "\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(200,)\n"
     ]
    }
   ],
   "source": [
    "print(x_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_l16_20k.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200/200 [==============================] - 20s 98ms/step\n",
      "Score (RMSE): 0.06380012712256235\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='train_l16_20k')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_true = np.sort(test_label)\n",
    "y_pred = np.sort(np.squeeze(pred))\n",
    "error = np.abs(np.subtract(y_pred,y_true))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.8477394067574668\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"
    }
   ],
   "source": [
    "x = y_true\n",
    "y = y_pred\n",
    "ylower = y - error\n",
    "yhigh = y + error\n",
    "\n",
    "score =np.mean(np.abs((x - y) / x)) * 100\n",
    "\n",
    "\n",
    "graph = plt.plot(x, y,'.')\n",
    "plt.errorbar(x,y, yerr=error)\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(200,)\n",
      "zero bin std data:\n",
      "[-5.48551]\n",
      "zero bin std data:\n",
      "[-5.478584]\n",
      "zero bin std data:\n",
      "[-5.4603086]\n",
      "zero bin std data:\n",
      "[-5.4578013]\n",
      "zero bin std data:\n",
      "[-5.424314]\n",
      "zero bin std data:\n",
      "[-5.423528]\n",
      "zero bin std data:\n",
      "[-5.401954]\n",
      "zero bin std data:\n",
      "[-5.3818727]\n",
      "zero bin std data:\n",
      "[-5.339108]\n",
      "zero bin std data:\n",
      "[-5.334227]\n",
      "zero bin std data:\n",
      "[-5.2593217]\n",
      "zero bin std data:\n",
      "[-5.1665525]\n",
      "zero bin std data:\n",
      "[-5.1391377]\n",
      "zero bin std data:\n",
      "[-4.991859]\n",
      "zero bin std data:\n",
      "[-4.98074]\n",
      "zero bin std data:\n",
      "[-4.9181685]\n",
      "zero bin std data:\n",
      "[-4.775209]\n",
      "zero bin std data:\n",
      "[-4.6562552]\n",
      "zero bin std data:\n",
      "[-4.6230063]\n",
      "zero bin std data:\n",
      "[-4.587458]\n",
      "zero bin std data:\n",
      "[-4.551089]\n",
      "zero bin std data:\n",
      "[-4.4231834]\n",
      "zero bin std data:\n",
      "[-4.3873734]\n",
      "zero bin std data:\n",
      "[-4.3736825]\n",
      "zero bin std data:\n",
      "[-4.3059835]\n",
      "zero bin std data:\n",
      "[-4.3009686]\n",
      "zero bin std data:\n",
      "[-4.2934027]\n",
      "zero bin std data:\n",
      "[-4.2907643]\n",
      "zero bin std data:\n",
      "[-4.2045975]\n",
      "zero bin std data:\n",
      "[-4.1825986]\n",
      "zero bin std data:\n",
      "[-4.1647487]\n",
      "zero bin std data:\n",
      "[-4.153446]\n",
      "zero bin std data:\n",
      "[-4.1503015]\n",
      "zero bin std data:\n",
      "[-4.1379633]\n",
      "zero bin std data:\n",
      "[-4.071291]\n",
      "zero bin std data:\n",
      "[-4.051655]\n",
      "zero bin std data:\n",
      "[-4.0328093]\n",
      "zero bin std data:\n",
      "[-3.9789207]\n",
      "zero bin std data:\n",
      "[-3.944572]\n",
      "zero bin std data:\n",
      "[-3.9352438]\n",
      "zero bin std data:\n",
      "[-3.8414235]\n",
      "zero bin std data:\n",
      "[-3.8304522]\n",
      "zero bin std data:\n",
      "[-3.7902286]\n",
      "zero bin std data:\n",
      "[-3.7802746]\n",
      "zero bin std data:\n",
      "[-3.7665784]\n",
      "zero bin std data:\n",
      "[-3.7306778]\n",
      "zero bin std data:\n",
      "[-3.6952462]\n",
      "zero bin std data:\n",
      "[-3.6463473]\n",
      "zero bin std data:\n",
      "[-3.641824]\n",
      "zero bin std data:\n",
      "[-3.6167958]\n",
      "zero bin std data:\n",
      "[-3.6015403]\n",
      "zero bin std data:\n",
      "[-3.5692356]\n",
      "zero bin std data:\n",
      "[-3.5429542]\n",
      "zero bin std data:\n",
      "[-3.5291283]\n",
      "zero bin std data:\n",
      "[-3.5263436]\n",
      "zero bin std data:\n",
      "[-3.4946146]\n",
      "zero bin std data:\n",
      "[-3.4881089]\n",
      "zero bin std data:\n",
      "[-3.4622533]\n",
      "zero bin std data:\n",
      "[-3.4433033]\n",
      "zero bin std data:\n",
      "[-3.415123]\n",
      "zero bin std data:\n",
      "[-3.3713543]\n",
      "zero bin std data:\n",
      "[-3.3527534]\n",
      "zero bin std data:\n",
      "[-3.3466794]\n",
      "zero bin std data:\n",
      "[-3.3284166]\n",
      "zero bin std data:\n",
      "[-3.3044493]\n",
      "zero bin std data:\n",
      "[-3.2589152]\n",
      "zero bin std data:\n",
      "[-3.1957417]\n",
      "zero bin std data:\n",
      "[-3.1955554]\n",
      "zero bin std data:\n",
      "[-3.1449487]\n",
      "zero bin std data:\n",
      "[-3.1409347]\n",
      "zero bin std data:\n",
      "[-3.1003714]\n",
      "zero bin std data:\n",
      "[-3.0874171]\n",
      "zero bin std data:\n",
      "[-3.0511258]\n",
      "zero bin std data:\n",
      "[-3.0349643]\n",
      "zero bin std data:\n",
      "[-3.0099475]\n",
      "zero bin std data:\n",
      "[-3.0013797]\n",
      "zero bin std data:\n",
      "[-2.9859996]\n",
      "zero bin std data:\n",
      "[-2.9816294]\n",
      "zero bin std data:\n",
      "[-2.975722]\n",
      "zero bin std data:\n",
      "[-2.971152]\n",
      "zero bin std data:\n",
      "[-2.9518583]\n",
      "zero bin std data:\n",
      "[-2.9414165]\n",
      "zero bin std data:\n",
      "[-2.926284]\n",
      "zero bin std data:\n",
      "[-2.873842]\n",
      "zero bin std data:\n",
      "[-2.8550742]\n",
      "zero bin std data:\n",
      "[-2.8161604]\n",
      "zero bin std data:\n",
      "[-2.8076901]\n",
      "zero bin std data:\n",
      "[-2.736564]\n",
      "zero bin std data:\n",
      "[-2.6254232]\n",
      "zero bin std data:\n",
      "[-2.6226244]\n",
      "zero bin std data:\n",
      "[-2.6052835]\n",
      "zero bin std data:\n",
      "[-2.5640042]\n",
      "zero bin std data:\n",
      "[-2.5624382]\n",
      "zero bin std data:\n",
      "[-2.5071962]\n",
      "zero bin std data:\n",
      "[-2.4859111]\n",
      "zero bin std data:\n",
      "[-2.4479172]\n",
      "zero bin std data:\n",
      "[-2.4045777]\n",
      "zero bin std data:\n",
      "[-2.396506]\n",
      "zero bin std data:\n",
      "[-2.3918345]\n",
      "zero bin std data:\n",
      "[-2.3636258]\n",
      "zero bin std data:\n",
      "[-2.36289]\n",
      "zero bin std data:\n",
      "[-2.3524675]\n",
      "zero bin std data:\n",
      "[-2.3432672]\n",
      "zero bin std data:\n",
      "[-2.2788613]\n",
      "zero bin std data:\n",
      "[-2.241847]\n",
      "zero bin std data:\n",
      "[-2.2306466]\n",
      "zero bin std data:\n",
      "[-2.1722858]\n",
      "zero bin std data:\n",
      "[-2.160454]\n",
      "zero bin std data:\n",
      "[-2.1394112]\n",
      "zero bin std data:\n",
      "[-2.1373875]\n",
      "zero bin std data:\n",
      "[-2.1300094]\n",
      "zero bin std data:\n",
      "[-2.0460758]\n",
      "zero bin std data:\n",
      "[-2.0385745]\n",
      "zero bin std data:\n",
      "[-1.9976095]\n",
      "zero bin std data:\n",
      "[-1.985228]\n",
      "zero bin std data:\n",
      "[-1.967957]\n",
      "zero bin std data:\n",
      "[-1.9630265]\n",
      "zero bin std data:\n",
      "[-1.9457662]\n",
      "zero bin std data:\n",
      "[-1.9279116]\n",
      "zero bin std data:\n",
      "[-1.8630652]\n",
      "zero bin std data:\n",
      "[-1.8603646]\n",
      "zero bin std data:\n",
      "[-1.8352827]\n",
      "zero bin std data:\n",
      "[-1.8319799]\n",
      "zero bin std data:\n",
      "[-1.807423]\n",
      "zero bin std data:\n",
      "[-1.799658]\n",
      "zero bin std data:\n",
      "[-1.7682859]\n",
      "zero bin std data:\n",
      "[-1.7664348]\n",
      "zero bin std data:\n",
      "[-1.7567958]\n",
      "zero bin std data:\n",
      "[-1.7446749]\n",
      "zero bin std data:\n",
      "[-1.7142152]\n",
      "zero bin std data:\n",
      "[-1.6959015]\n",
      "zero bin std data:\n",
      "[-1.6919636]\n",
      "zero bin std data:\n",
      "[-1.6564605]\n",
      "zero bin std data:\n",
      "[-1.6515044]\n",
      "zero bin std data:\n",
      "[-1.6384333]\n",
      "zero bin std data:\n",
      "[-1.6361692]\n",
      "zero bin std data:\n",
      "[-1.6089665]\n",
      "zero bin std data:\n",
      "[-1.6078024]\n",
      "zero bin std data:\n",
      "[-1.5740992]\n",
      "zero bin std data:\n",
      "[-1.5289499]\n",
      "zero bin std data:\n",
      "[-1.4862293]\n",
      "zero bin std data:\n",
      "[-1.4510101]\n",
      "zero bin std data:\n",
      "[-1.4104785]\n",
      "zero bin std data:\n",
      "[-1.385837]\n",
      "zero bin std data:\n",
      "[-1.3849883]\n",
      "zero bin std data:\n",
      "[-1.3498267]\n",
      "zero bin std data:\n",
      "[-1.2234023]\n",
      "zero bin std data:\n",
      "[-1.2216958]\n",
      "zero bin std data:\n",
      "[-1.2139075]\n",
      "zero bin std data:\n",
      "[-1.213291]\n",
      "zero bin std data:\n",
      "[-1.1932341]\n",
      "zero bin std data:\n",
      "[-1.1833389]\n",
      "zero bin std data:\n",
      "[-1.166296]\n",
      "zero bin std data:\n",
      "[-1.1515754]\n",
      "zero bin std data:\n",
      "[-1.1312352]\n",
      "zero bin std data:\n",
      "[-1.119417]\n",
      "zero bin std data:\n",
      "[-1.093932]\n",
      "zero bin std data:\n",
      "[-1.0910932]\n",
      "zero bin std data:\n",
      "[-1.0676836]\n",
      "zero bin std data:\n",
      "[-1.066001]\n",
      "zero bin std data:\n",
      "[-1.0655829]\n",
      "zero bin std data:\n",
      "[-1.0300136]\n",
      "zero bin std data:\n",
      "[-1.0275375]\n",
      "zero bin std data:\n",
      "[-1.0250789]\n",
      "zero bin std data:\n",
      "[-1.0011563]\n",
      "zero bin std data:\n",
      "[-0.9994409]\n",
      "zero bin std data:\n",
      "[-0.9972681]\n",
      "zero bin std data:\n",
      "[-0.9951249]\n",
      "zero bin std data:\n",
      "[-0.9924173]\n",
      "zero bin std data:\n",
      "[-0.98854685]\n",
      "zero bin std data:\n",
      "[-0.9830381]\n",
      "zero bin std data:\n",
      "[-0.9798637]\n",
      "zero bin std data:\n",
      "[-0.97741646]\n",
      "zero bin std data:\n",
      "[-0.97553504]\n",
      "zero bin std data:\n",
      "[-0.96622175]\n",
      "zero bin std data:\n",
      "[-0.9647028]\n",
      "zero bin std data:\n",
      "[-0.9588818]\n",
      "zero bin std data:\n",
      "[-0.95084494]\n",
      "zero bin std data:\n",
      "[-0.9484134, -0.94793624]\n",
      "zero bin std data:\n",
      "[-0.9484134, -0.94793624]\n",
      "zero bin std data:\n",
      "[-0.9413443]\n",
      "zero bin std data:\n",
      "[-0.9368749]\n",
      "zero bin std data:\n",
      "[-0.9360906]\n",
      "zero bin std data:\n",
      "[-0.935258]\n",
      "zero bin std data:\n",
      "[-0.9351546]\n",
      "zero bin std data:\n",
      "[-0.93178546]\n",
      "zero bin std data:\n",
      "[-0.92894685]\n",
      "zero bin std data:\n",
      "[-0.9218908]\n",
      "zero bin std data:\n",
      "[-0.9211629]\n",
      "zero bin std data:\n",
      "[-0.9128251]\n",
      "zero bin std data:\n",
      "[-0.9060021]\n",
      "zero bin std data:\n",
      "[-0.9056249]\n",
      "zero bin std data:\n",
      "[-0.9046451]\n",
      "zero bin std data:\n",
      "[-0.90349156]\n",
      "zero bin std data:\n",
      "[-0.9015921]\n",
      "zero bin std data:\n",
      "[-0.9001468]\n",
      "zero bin std data:\n",
      "[-0.8992924]\n",
      "zero bin std data:\n",
      "[-0.89608634]\n",
      "<class 'numpy.ndarray'>\n",
      "(200,)\n",
      "zero bin std data:\n",
      "[-1.5978152281017537]\n",
      "zero bin std data:\n",
      "[-1.8560784348491188]\n",
      "zero bin std data:\n",
      "[-1.8714283248496018]\n",
      "zero bin std data:\n",
      "[-2.0585679802905466]\n",
      "zero bin std data:\n",
      "[-1.562843109291484]\n",
      "zero bin std data:\n",
      "[-1.6570986318118264]\n",
      "zero bin std data:\n",
      "[-1.5360069694349088]\n",
      "zero bin std data:\n",
      "[-1.7938740686341321]\n",
      "zero bin std data:\n",
      "[-0.9999148061427581]\n",
      "zero bin std data:\n",
      "[-1.531802829769025]\n",
      "zero bin std data:\n",
      "[-1.1851676489404477]\n",
      "zero bin std data:\n",
      "[-0.6879912246688497]\n",
      "zero bin std data:\n",
      "[-2.1353192424552767]\n",
      "zero bin std data:\n",
      "[-0.21167031809211598]\n",
      "zero bin std data:\n",
      "[-0.21609812224064767]\n",
      "zero bin std data:\n",
      "[-2.756198519755762]\n",
      "zero bin std data:\n",
      "[-2.8032068180520135]\n",
      "zero bin std data:\n",
      "[-1.8874234480833167]\n",
      "zero bin std data:\n",
      "[-2.1094721283332074]\n",
      "zero bin std data:\n",
      "[-2.0639496968728337]\n",
      "zero bin std data:\n",
      "[-2.8785265169295156]\n",
      "zero bin std data:\n",
      "[-1.619070598431329]\n",
      "zero bin std data:\n",
      "[-2.0853575561365667]\n",
      "zero bin std data:\n",
      "[-2.093190856631097]\n",
      "zero bin std data:\n",
      "[-0.8131940385554229]\n",
      "zero bin std data:\n",
      "[-0.9024904066135332]\n",
      "zero bin std data:\n",
      "[-2.682276328963492]\n",
      "zero bin std data:\n",
      "[-2.710680106828125]\n",
      "zero bin std data:\n",
      "[-0.8901614872410633]\n",
      "zero bin std data:\n",
      "[-1.0073594867714033]\n",
      "zero bin std data:\n",
      "[-0.8551696638385761]\n",
      "zero bin std data:\n",
      "[-0.6591021953193592]\n",
      "zero bin std data:\n",
      "[-0.888519324275863]\n",
      "zero bin std data:\n",
      "[-1.888908638469012]\n",
      "zero bin std data:\n",
      "[-1.1152926703077155]\n",
      "zero bin std data:\n",
      "[-2.522667809366434]\n",
      "zero bin std data:\n",
      "[-3.0308881114217674]\n",
      "zero bin std data:\n",
      "[-1.7952211195730667]\n",
      "zero bin std data:\n",
      "[-2.157901693870056]\n",
      "zero bin std data:\n",
      "[-2.814992876430591]\n",
      "zero bin std data:\n",
      "[-1.0074260034876505]\n",
      "zero bin std data:\n",
      "[-2.4527505433310766]\n",
      "zero bin std data:\n",
      "[-1.716970491455616]\n",
      "zero bin std data:\n",
      "[-2.8362135907112656]\n",
      "zero bin std data:\n",
      "[-2.719590595713332]\n",
      "zero bin std data:\n",
      "[-1.757322901845803]\n",
      "zero bin std data:\n",
      "[-2.1457816793588096]\n",
      "zero bin std data:\n",
      "[-0.9218633614135594]\n",
      "zero bin std data:\n",
      "[-1.5043553857019942]\n",
      "zero bin std data:\n",
      "[-1.2399100362862372]\n",
      "zero bin std data:\n",
      "[-1.7384275318607127]\n",
      "zero bin std data:\n",
      "[-0.968474427781558]\n",
      "zero bin std data:\n",
      "[-1.236374535988038]\n",
      "zero bin std data:\n",
      "[-1.850745189129116]\n",
      "zero bin std data:\n",
      "[-1.9626797483001286]\n",
      "zero bin std data:\n",
      "[-1.772540497481746]\n",
      "zero bin std data:\n",
      "[-1.62006936676208]\n",
      "zero bin std data:\n",
      "[-1.3095344542168674]\n",
      "zero bin std data:\n",
      "[-1.7168442639912103]\n",
      "zero bin std data:\n",
      "[-2.2109388367592895]\n",
      "zero bin std data:\n",
      "[-1.052209809060551]\n",
      "zero bin std data:\n",
      "[-1.093579770159069]\n",
      "zero bin std data:\n",
      "[-1.4709530078991835]\n",
      "zero bin std data:\n",
      "[-1.6583568848689163]\n",
      "zero bin std data:\n",
      "[-1.7447068774254029]\n",
      "zero bin std data:\n",
      "[-2.9673047306613425]\n",
      "zero bin std data:\n",
      "[-1.2533061949506188]\n",
      "zero bin std data:\n",
      "[-2.750979331080024]\n",
      "zero bin std data:\n",
      "[-1.635033133382146]\n",
      "zero bin std data:\n",
      "[-3.1505653469242034]\n",
      "zero bin std data:\n",
      "[-3.2166914618624247]\n",
      "zero bin std data:\n",
      "[-2.8587036576307914]\n",
      "zero bin std data:\n",
      "[-2.0779679543933285]\n",
      "zero bin std data:\n",
      "[-2.717713589713691]\n",
      "zero bin std data:\n",
      "[-1.9025830749218793]\n",
      "zero bin std data:\n",
      "[-1.936574936213253]\n",
      "zero bin std data:\n",
      "[-1.5722126648789307]\n",
      "zero bin std data:\n",
      "[-2.329622689329222]\n",
      "zero bin std data:\n",
      "[-2.258139448041824]\n",
      "zero bin std data:\n",
      "[-2.262043031916898]\n",
      "zero bin std data:\n",
      "[-2.273132657646943]\n",
      "zero bin std data:\n",
      "[-2.9424748854434872]\n",
      "zero bin std data:\n",
      "[-2.9475491463436523]\n",
      "zero bin std data:\n",
      "[-1.4595590619184715]\n",
      "zero bin std data:\n",
      "[-1.1093112716142488]\n",
      "zero bin std data:\n",
      "[-1.1188671623504378]\n",
      "zero bin std data:\n",
      "[-3.3188649072855014]\n",
      "zero bin std data:\n",
      "[-3.3152970936215747]\n",
      "zero bin std data:\n",
      "[-1.2564655334587413]\n",
      "zero bin std data:\n",
      "[-1.822997841467147]\n",
      "zero bin std data:\n",
      "[-2.519764289218662]\n",
      "zero bin std data:\n",
      "[-1.2118238468901543]\n",
      "zero bin std data:\n",
      "[-3.250016592482377]\n",
      "zero bin std data:\n",
      "[-3.3119967822535767]\n",
      "zero bin std data:\n",
      "[-3.3105929832689096]\n",
      "zero bin std data:\n",
      "[-2.958137602234801]\n",
      "zero bin std data:\n",
      "[-1.4773880916083182]\n",
      "zero bin std data:\n",
      "[-1.3321811393674179]\n",
      "zero bin std data:\n",
      "[-2.3741800059514766]\n",
      "zero bin std data:\n",
      "[-1.8804214781813378]\n",
      "zero bin std data:\n",
      "[-1.9849807420321965]\n",
      "zero bin std data:\n",
      "[-2.8052813018539116]\n",
      "zero bin std data:\n",
      "[-2.887693367129336]\n",
      "zero bin std data:\n",
      "[-2.830204726082053]\n",
      "zero bin std data:\n",
      "[-1.6134641768625353]\n",
      "zero bin std data:\n",
      "[-1.16311225452269]\n",
      "zero bin std data:\n",
      "[-3.1964756342969665]\n",
      "zero bin std data:\n",
      "[-2.7425082476702953]\n",
      "zero bin std data:\n",
      "[-2.3029875973376086]\n",
      "zero bin std data:\n",
      "[-3.219284654137475]\n",
      "zero bin std data:\n",
      "[-2.9611809393750925]\n",
      "zero bin std data:\n",
      "[-1.2282419175362722]\n",
      "zero bin std data:\n",
      "[-3.3497834248681224]\n",
      "zero bin std data:\n",
      "[-1.2855597651974773]\n",
      "zero bin std data:\n",
      "[-2.2004270007779505]\n",
      "zero bin std data:\n",
      "[-2.497762639268399]\n",
      "zero bin std data:\n",
      "[-2.2630697128456037]\n",
      "zero bin std data:\n",
      "[-2.5436750824084013]\n",
      "zero bin std data:\n",
      "[-3.5120357264754656]\n",
      "zero bin std data:\n",
      "[-3.081279380897153]\n",
      "zero bin std data:\n",
      "[-3.4973345022222913]\n",
      "zero bin std data:\n",
      "[-2.256695234874884]\n",
      "zero bin std data:\n",
      "[-3.4587278649983557]\n",
      "zero bin std data:\n",
      "[-2.33660296624517]\n",
      "zero bin std data:\n",
      "[-3.3543706269345703]\n",
      "zero bin std data:\n",
      "[-1.8451207328687251]\n",
      "zero bin std data:\n",
      "[-3.0901835134298836]\n",
      "zero bin std data:\n",
      "[-3.26499978272567]\n",
      "zero bin std data:\n",
      "[-3.2694246868111203]\n",
      "zero bin std data:\n",
      "[-1.9501439700153593]\n",
      "zero bin std data:\n",
      "[-2.042261383084906]\n",
      "zero bin std data:\n",
      "[-4.32081349697258]\n",
      "zero bin std data:\n",
      "[-2.2762231292605706]\n",
      "zero bin std data:\n",
      "[-3.380562539436143]\n",
      "zero bin std data:\n",
      "[-3.780424199178651]\n",
      "zero bin std data:\n",
      "[-4.495480013650715]\n",
      "zero bin std data:\n",
      "[-2.8333510289379853]\n",
      "zero bin std data:\n",
      "[-4.914263395129862]\n",
      "zero bin std data:\n",
      "[-3.0525364885018504]\n",
      "zero bin std data:\n",
      "[-3.443698805096022]\n",
      "zero bin std data:\n",
      "[-3.3980166962137917]\n",
      "zero bin std data:\n",
      "[-3.3104260217407173]\n",
      "zero bin std data:\n",
      "[-4.576710300872363]\n",
      "zero bin std data:\n",
      "[-4.382204983443845]\n",
      "zero bin std data:\n",
      "[-4.376509985512933]\n",
      "zero bin std data:\n",
      "[-2.556289409731634]\n",
      "zero bin std data:\n",
      "[-1.4205418426155145]\n",
      "zero bin std data:\n",
      "[-1.4156945453996566]\n",
      "zero bin std data:\n",
      "[-1.4004837533573056]\n",
      "zero bin std data:\n",
      "[-1.4004486077406635]\n",
      "zero bin std data:\n",
      "[-2.4572755123136973]\n",
      "zero bin std data:\n",
      "[-2.2645315179644623]\n",
      "zero bin std data:\n",
      "[-1.0783016004915273]\n",
      "zero bin std data:\n",
      "[-0.8878157622960438]\n",
      "zero bin std data:\n",
      "[-0.7352574702246109]\n",
      "zero bin std data:\n",
      "[-0.69718363012482]\n",
      "zero bin std data:\n",
      "[-0.8052570137610113]\n",
      "zero bin std data:\n",
      "[-0.8277965589948233]\n",
      "zero bin std data:\n",
      "[-1.7203056467777658]\n",
      "zero bin std data:\n",
      "[-1.678637977510567]\n",
      "zero bin std data:\n",
      "[-1.6687425096693045]\n",
      "zero bin std data:\n",
      "[-0.5833964956794181]\n",
      "zero bin std data:\n",
      "[-0.5564025024474909]\n",
      "zero bin std data:\n",
      "[-0.5310113578143183]\n",
      "zero bin std data:\n",
      "[-0.42538274746709975]\n",
      "zero bin std data:\n",
      "[-0.7651384553109357]\n",
      "zero bin std data:\n",
      "[-0.7683125004083285]\n",
      "zero bin std data:\n",
      "[-0.7745005042759522]\n",
      "zero bin std data:\n",
      "[-0.7873664477578554]\n",
      "zero bin std data:\n",
      "[-0.4899724346168831]\n",
      "zero bin std data:\n",
      "[-0.8591460912537364]\n",
      "zero bin std data:\n",
      "[-0.8969368787741653]\n",
      "zero bin std data:\n",
      "[-0.8644792004541858]\n",
      "zero bin std data:\n",
      "[-0.7352975453580988]\n",
      "zero bin std data:\n",
      "[-1.105581597961459]\n",
      "zero bin std data:\n",
      "[-0.991303197360719]\n",
      "zero bin std data:\n",
      "[-0.8980542589881788]\n",
      "zero bin std data:\n",
      "[-0.6570456274605109]\n",
      "zero bin std data:\n",
      "[-0.41319009390294065]\n",
      "zero bin std data:\n",
      "[-0.3500946518553615]\n",
      "zero bin std data:\n",
      "[-0.33990524818822393]\n",
      "zero bin std data:\n",
      "[-0.2691295600553949]\n",
      "zero bin std data:\n",
      "[-0.3196227588869599]\n",
      "zero bin std data:\n",
      "[-0.2983984796363268]\n",
      "zero bin std data:\n",
      "[-0.28498067653838627]\n",
      "zero bin std data:\n",
      "[-0.18889393875620913]\n",
      "zero bin std data:\n",
      "[-0.2189082602841934]\n",
      "zero bin std data:\n",
      "[-0.4711459530977624]\n",
      "zero bin std data:\n",
      "[-0.56764674257779]\n",
      "zero bin std data:\n",
      "[-0.820533633839722]\n",
      "zero bin std data:\n",
      "[-0.7490257920909048]\n",
      "zero bin std data:\n",
      "[-1.0851951798292847]\n",
      "zero bin std data:\n",
      "[-0.9862436094167868]\n",
      "zero bin std data:\n",
      "[-0.8243584698573815]\n",
      "zero bin std data:\n",
      "[-0.7286680561675231]\n",
      "zero bin std data:\n",
      "[-0.36956275089684626]\n"
     ]
    }
   ],
   "source": [
    "y_true = np.sort(test_label)\n",
    "y_pred = np.sort(np.squeeze(pred))\n",
    "error = np.divide(np.abs(np.subtract(y_pred,y_true)), y_true)*100\n",
    "\n",
    "bin_edges = [(y_true[i]+y_true[i+1])/2 for i in range((len(y_true)-1))]\n",
    "\n",
    "results = make_bin(y_true, y_pred,bin_edges,method='percentile')\n",
    "\n",
    "results_residual = make_bin(y_true, error, bin_edges, method='percentile')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_regression_results(results, results_residual,truth_values=np.unique(y_true),individual_truth_values=y_true,individual_medians=y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Negativity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_neg.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_neg.csv')\n",
    "\n",
    "path_folder = 'data/matrix_neg/'\n",
    "\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_neg.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 25s 246ms/step\n",
      "Score (RMSE): 0.009894175362264807\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='train_neg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000,)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAFjRJREFUeJzt3X+QXWd52PHvo5Xl1I1DhKxAY1mWlQiK+FGwhCzKxENLADlm7ISQqewm4zSmHpiYSabNtMK0oSM6jBN3GpiJOkUFd/LDxKHGppoAFe7g8BerSCsgRrgiYmvZi0gjpK2TqRmvVvv0j3vXrDb3as/dvXfPr+9nxqM9955z7/OOn/Psu+85530jM5EktcOasgOQJK0ei74ktYhFX5JaxKIvSS1i0ZekFrHoS1KLWPQlqUUs+pLUIhZ9SWqRtWUHsNg111yTW7ZsKTsMSaqViYmJ72XmxqX2q1zR37JlC8eOHSs7DEmqlYg4XWQ/h3ckqUUs+pLUIhZ9SWoRi74ktYhFX5JaxKIvSS1SuVs2pUFt2fe5skMYmafvv7XsENQw9vRVa00u+ND89mn1WfRVWxOnp8sOQaodi75q69HjU2WHINWORV+1lWUHsArevv1lZYeghil0ITci9gAfA8aAT2Tm/Yve/xfAe4BZ4CzwK5l5uvveXcC/6e767zPz94YUu1ruNT/+kku233vzVvb9zKtKimY4PnXkGe577MkXt9/yyh8rMRo10ZI9/YgYAw4AtwDbgTsiYvui3b4K7MzM1wGPAL/dPfalwIeAm4BdwIciYv3wwlebTT8/Q3R/XgNc/XeuKDOcoZh+foY13Uatic62NExFhnd2AacyczIzZ4CHgdsX7pCZT2Tm893NcWBT9+d3AI9n5vnMnAYeB/YMJ3S13e6tG7hi7RoCWLt2Dbu3big7pBXbvXUDa9dEp01rohFtUrUUKfrXAs8u2J7qvtbP3cAXBjk2Iu6JiGMRcezs2bMFQpK6Mi/9twkiLv1XGqIiRb9X5vU8wyLiF4GdwAODHJuZBzNzZ2bu3LhxyTUAJADGJ88xO5ckcHEuGZ88V3ZIKzY+eY7Zi3OdNl2ca0SbVC1Fiv4UcN2C7U3AmcU7RcRPAx8EbsvMFwY5VlqO3Vs3sG7tGsYCrmjQ8E7T2qRqiVziz+KIWAt8C3gr8B3gKHBnZp5YsM8b6FzA3ZOZf7Hg9ZcCE8CN3ZeOAzsy83y/79u5c2e6cpaKmjg9zfjkOXZv3cCO65txj0AT26TRi4iJzNy51H5L3rKZmbMRcS9wmM4tmw9m5omI2A8cy8xDdIZzfhj4b9EZh3wmM2/LzPMR8WE6vygA9l+u4EuD2nH9+sYVxia2SdWxZE9/tdnT1yCa2CtuYps0ekPr6UtVNXF6mn/6iXFmZudYt3YND71nd+2LZBPbpGpxGgbV1vjkOWZm55hLuDDbjDtdmtgmVYtFX7XVxDtdmtgmVYtj+qq1Jo5/N7FNGj3H9NUKTbzTpYltUnU4vCNJLWLRl6QWsehLUotY9FVrE6enOfDEqUatl9vENqk6vJCr2mrig0xNbJOqxZ6+aquJDzI1sU2qFou+aquJDzI1sU2qFh/OUq018UGmJrZJo+fDWWqFJj7I1MQ2qToc3pGkFrHoq9aaeHtjE9uk6nB4R7XVxNsbm9gmVYs9fdVWE29vbGKbVC0WfdVWE29vbGKbVC3esqlaa+LtjU1sk0av6C2b9vQlqUW8kKvaauJFzya2SdViT1+1tfCi50xDLnqOT57jhQvdNl1oRptULRZ91db6q9Yx170kNZed7bpbf9U65q+yzdGMNqlaHN7RZW3Z97myQyjsxJnnyg5hxb6xqA33PfYk9z32ZEnRLG0NMHn/rWWHoQHY01dfdSr4AA8deabsEFYsyg5gQHPA1prlSdtZ9KUKedeNm8oOYWBzZQeggVj0pQrxTh2NmkVffb19+8vKDmEgTzdkbLlu7ahbnrSdF3LV1zVXX3nJ9p03beYjP/fakqJplyoX/g8+9uQl1082LsoTVZs9ffX18zduYu1Y59Li2rHg52s43qzhe9eivKjjdYg2s+jrstbQuaPERNFCayI6eRF1u99Insvqa3zyHLNzSQIX59KnQwV08+LiXCcvLvrUcN1Y9NWX0/yqF/Oi3pxaWZflNL/qxbyonqJTK3v3ji5rx/XrPan1t5gX9eXwjiS1iEVflzVxepoDT5xi4vR02aGoQsyL+nJ4R325oId6MS/qrVBPPyL2RMTJiDgVEft6vH9zRByPiNmIePei9y5GxNe6/x0aVuAavYWLlFxoyCIlWjnzot6W7OlHxBhwAHgbMAUcjYhDmfnNBbs9A/wy8Bs9PuL7mfn6IcSqVTZ/a96F2TlvzdOLzIt6KzK8sws4lZmTABHxMHA78GLRz8ynu+85y2qD7Lh+PQ+9Z7e35ukS5kW9FSn61wLPLtieAm4a4Dt+KCKOAbPA/Zn52cU7RMQ9wD0AmzdvHuCjNWremqdezIv6KjKm32tyjUGe6NrcfWDgTuCjEfETf+vDMg9m5s7M3Llx48YBPlqSNIgiRX8KuG7B9ibgTNEvyMwz3X8ngT8F3jBAfJKkISpS9I8C2yLihohYB+wFCt2FExHrI+LK7s/XAG9mwbUASdLqWrLoZ+YscC9wGHgK+HRmnoiI/RFxG0BEvDEipoBfAD4eESe6h78KOBYRXweeoDOmb9GvER/CUS/mRX054Zr68iEc9WJeVFPRCdechkF9+RCOejEv6s2ir76cN129mBf15vCOLst509WLeVE9zqevofAhHPViXtSXwzuS1CIWfUlqEYu+JLWIRV+SWsSiL0ktYtGXpBax6EtSi1j0JalFLPqS1CIWfUlqEYu+JLWIRV+SWsSiL0ktYtHXZbksnnoxL+rLqZXVl8viqRfzot7s6asvl8VTL+ZFvVn01ZfL4qkX86LeXC5Rl+WyeOrFvKgel0vUULgsnnoxL+rL4R1JahGLviS1iEVfklrEoi9JLWLRl6QWsehLUotY9CWpRSz6ktQiFn1JahGLvvpy+lz1Y27Ul9MwqCenz1U/5ka92dNXT06fq37MjXqz6Ksnp89VP+ZGvTm1svpy+lz1Y25Uj1Mra8WcPlf9mBv1VWh4JyL2RMTJiDgVEft6vH9zRByPiNmIePei9+6KiL/o/nfXsAKXJA1uyaIfEWPAAeAWYDtwR0RsX7TbM8AvA59adOxLgQ8BNwG7gA9FhN0DSSpJkZ7+LuBUZk5m5gzwMHD7wh0y8+nM/HNgbtGx7wAez8zzmTkNPA7sGULckqRlKFL0rwWeXbA91X2tiJUcK0kasiJFP3q8VvSWn0LHRsQ9EXEsIo6dPXu24EdLkgZVpOhPAdct2N4EnCn4+YWOzcyDmbkzM3du3Lix4Edr1HzUXv2YG/VV5JbNo8C2iLgB+A6wF7iz4OcfBj6y4OLt24EPDBylVp2P2qsfc6PeluzpZ+YscC+dAv4U8OnMPBER+yPiNoCIeGNETAG/AHw8Ik50jz0PfJjOL46jwP7ua6o4H7VXP+ZGvRV6OCszPw98ftFrv7ng56N0hm56Hfsg8OAKYlQJ5h+1vzA756P2uoS5UW9Ow6C+fNRe/Zgb1eM0DFoxH7VXP+ZGfTnLpvryDg31Y27Ulz199eQdGurH3Kg3e/rqaeEdGjPeoaEFxifP8cKFbm5cMDfqxqKvntZftY657jX+uexsS9DJhfnbP+YwN+rGoq+epp+feXEOjehuS2Bu1J1j+jW3Zd/nRv4dib05/cDCnn4CDxw+yQOHT47s+56+/9aRfXYb2dOvsdUo+PPue+zJVfsuVdvHv/ztVf2+1czzNrDo15S3yqks333u+2WHoBWw6NeUd0yoLE67UG8W/ZravXUDa8d6LVcwGo6rat7v330TN2+7ZtW+7yM/99pV+6428EJujc3/xh4L+PDPvpY7b9pcajxqj9+/+6aRffbE6Wn2HvwKFy4mV4wFr3z51SP7rjayp19T45PnmJ37wWR53janphifPMfFbm7PzaVDmUNm0a+p+eltxwKnt1WjmNuj5dTKNeb0tmoqc3twRadWtqcvSS3ihdyacqZDNZW5PVr29GvKdUrVVOb2aFn0a8qLXWoqc3u0vJBbY17sUlOZ24NzjdwWcJ1SNZW5PToO70hSi1j0JalFLPqS1CIWfUlqEYu+JLWIRb/GJk5Pc+CJU66ipcYxt0fHWzZrykfV1VTm9mjZ068pH1VXU5nbo2XRrykfVVdTmduj5TQMNeaj6moqc3twTsPQAj6qrqYyt0fH4R1JahGLviS1iEVfklrEoi9JLWLRl6QWsehLUosUKvoRsSciTkbEqYjY1+P9KyPij7vvH4mILd3Xt0TE9yPia93//vNww5ckDWLJ+/QjYgw4ALwNmAKORsShzPzmgt3uBqYz8ycjYi/wW8A/6b737cx8/ZDjliQtQ5Ge/i7gVGZOZuYM8DBw+6J9bgd+r/vzI8BbIyKGF6YkaRiKFP1rgWcXbE91X+u5T2bOAs8B8xNm3BARX42IL0fET60wXknSChSZhqFXj33xhD399vkusDkzz0XEDuCzEfHqzPzrSw6OuAe4B2Dz5s0FQhI4P4may9wenSJFfwq4bsH2JuBMn32mImIt8BLgfHZmc3sBIDMnIuLbwCuAS2ZUy8yDwEHoTLi2jHa0jnOOq6nM7dEqMrxzFNgWETdExDpgL3Bo0T6HgLu6P78b+FJmZkRs7F4IJiK2AtuAyeGE3m7OOa6mMrdHa8mefmbORsS9wGFgDHgwM09ExH7gWGYeAj4J/EFEnALO0/nFAHAzsD8iZoGLwHsz8/woGtI283OOX5idc85xNYq5PVrOp19jjnuqqcztwTmffgs457iaytweHadhkKQWsehLUotY9CWpRSz6ktQiFn1JahGLviS1iEW/xiZOT3PgiVNMnJ4uOxRpqMzt0fE+/ZpyfhI1lbk9Wvb0a8r5SdRU5vZoWfRran5+krHA+UnUKOb2aDn3To05P4maytwenHPvtIDzk6ipzO3RcXhHklrEoi9JLWLRrzHvZVZTmduj45h+TXkvs5rK3B4te/o15b3Maipze7Qs+hVU5E/b3Vs3sHZsDQGMjXkvs5pj99YNrF0TndxeE5fNbYeBBufwTsVMnJ7mjoNf4cLF5Iqx4I/ueVP/P23nn7Go2LMW0opFANn9t7eJ09Pc8V/GX1xA/Y/+ucNARdjTr5jPHJ9i5mKSwMzF5DPHp3ruNz55jtm5zn6zc+mfwGqM8clzXJid6+T2ZYZ3Hj0+xUx3v5nZOR7tc67oUvb0K2Zxv+ZTR57hU0eeuewxcwnrr1o3uqCkVbT+qnXM/+06Bzxw+CQPHD655HH+vVuMPf2K+X8vzC7ruPsee3LIkUjl+MaZ55Z13FKdI3VY9Cvmiyf+suwQpFL1H8XXMFj0K+aif6Oq5d5146ayQ2g0i37F3PKaly/ruKfvv3XIkUjl2HH9ej7zvn848HE/+/ofH0E0zeOF3Ir56N43APCn3zrLW16x8cVtqU12XL++UEfm1x/+qufKgCz6FfRLb9rCtpdd7QNX0hI8VwZn0a8Y5x2RivFcWR7H9CvGeUekYjxXlseiXzGuDyoV47myPK6RW0GuDyoV47nyA66RK6nx5gv9/NBO2wt/ERb9ivHilFSc58vgHNOvGC9OScV5vgzOol8xXpySivN8GZzDOxWz4/r1/OY7X80XvvFdbnnN3/NPVekyPF8GZ9GvmInT0+z/kxPMzM5x9OnzvPLlV5vIUh+eL4NzeKdiFo5RzjhGKV3W+OQ5XrjQPV8ueL4UUajoR8SeiDgZEaciYl+P96+MiD/uvn8kIrYseO8D3ddPRsQ7hhd6M62/ah1z3UcnXBFLurzFq2x5vixtyaIfEWPAAeAWYDtwR0RsX7Tb3cB0Zv4k8DvAb3WP3Q7sBV4N7AH+U/fz1MeJRasGLd6W9AOLV9la7qpbbVJkTH8XcCozJwEi4mHgduCbC/a5Hfh33Z8fAX43IqL7+sOZ+QLwvyPiVPfzvjKc8C+1Zd/nRvGxpTr7Ny+UHYJUWd9bdH4UWVO6Dka5PkaR4Z1rgWcXbE91X+u5T2bOAs8BGwoeOxRNLPgAX/zm/yk7BKmymnp+jLKeFSn6vZasXDxhT799ihxLRNwTEcci4tjZs2cLhCRJWo4iRX8KuG7B9ibgTL99ImIt8BLgfMFjycyDmbkzM3du3LixePSSpIEUKfpHgW0RcUNErKNzYfbQon0OAXd1f3438KXsTN95CNjbvbvnBmAb8GfDCf1STV0jtqntkoahqefHKNu15IXczJyNiHuBw8AY8GBmnoiI/cCxzDwEfBL4g+6F2vN0fjHQ3e/TdC76zgK/mpkXR9SWxiaApP487wfjfPqS1ABF59P3iVxJahGLviS1iEVfklrEoi9JLWLRl6QWqdzdOxFxFji9go+4BvjekMIpU1PaAbalqprSlqa0A1bWluszc8mnWytX9FcqIo4VuW2p6prSDrAtVdWUtjSlHbA6bXF4R5JaxKIvSS3SxKJ/sOwAhqQp7QDbUlVNaUtT2gGr0JbGjelLkvprYk9fktRHI4t+RLy/uxD7iYj47bLjWamI+I2IyIi4puxYlisiHoiI/xURfx4Rj0XEj5Yd0yAiYk83p05FxL6y41muiLguIp6IiKe658evlR3TSkXEWER8NSL+pOxYViIifjQiHumeJ09FxJtG8T2NK/oR8Y/orM37usx8NfAfSg5pRSLiOuBtQN0X/nwceE1mvg74FvCBkuMpLCLGgAPALcB24I6I2F5uVMs2C/zLzHwVsBv41Rq3Zd6vAU+VHcQQfAz4H5n594F/wIja1LiiD7wPuL+7GDuZ+Vclx7NSvwP8K3osM1knmfnF7vrJAON0VlGri13AqcyczMwZ4GE6HYvayczvZubx7s9/Q6ewjGTd6tUQEZuAW4FPlB3LSkTEjwA301mbhMycycz/O4rvamLRfwXwUxFxJCK+HBFvLDug5YqI24DvZObXy45lyH4F+ELZQQzgWuDZBdtT1LhQzouILcAbgCPlRrIiH6XTKZorO5AV2gqcBf5rd6jqExHxd0fxRUuunFVFEfE/gZf3eOuDdNq0ns6frm8EPh0RW7Oitykt0Zb7gLevbkTLd7m2ZOZ/7+7zQTpDDA+tZmwrFD1eq2Q+FRURPwx8Bvj1zPzrsuNZjoh4J/BXmTkREW8pO54VWgvcCLw/M49ExMeAfcC/HcUX1U5m/nS/9yLifcCj3SL/ZxExR2c+i7OrFd8g+rUlIl4L3AB8PSKgMxxyPCJ2ZeZfrmKIhV3u/wtARNwFvBN4a1V/CfcxBVy3YHsTcKakWFYsIq6gU/AfysxHy45nBd4M3BYRPwP8EPAjEfGHmfmLJce1HFPAVGbO/9X1CJ2iP3RNHN75LPCPASLiFcA6ajgZU2Y+mZk/lplbMnMLnaS4saoFfykRsQf418Btmfl82fEM6CiwLSJuiIh1dNaAPlRyTMsSnR7EJ4GnMvM/lh3PSmTmBzJzU/f82At8qaYFn+55/WxEvLL70lvprC0+dLXs6S/hQeDBiPgGMAPcVbNeZVP9LnAl8Hj3L5fxzHxvuSEVk5mzEXEvcBgYAx7MzBMlh7VcbwZ+CXgyIr7Wfe2+zPx8iTGp4/3AQ92OxSTwz0bxJT6RK0kt0sThHUlSHxZ9SWoRi74ktYhFX5JaxKIvSS1i0ZekFrHoS1KLWPQlqUX+P8tlvnzFhK1TAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataset = np.loadtxt('neg-delta=-1.dat')\n",
    "\n",
    "x_dat, y_dat = dataset[:,0],dataset[:,1]\n",
    "\n",
    "print(y_dat.shape)\n",
    "\n",
    "gr = plt.plot(x_dat,y_dat,'.')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "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": [
    "# example data\n",
    "x =(y_true)\n",
    "y = (y_pred)\n",
    "\n",
    "# example variable error bar values\n",
    "\n",
    "\n",
    "graph = plt.plot(x,y,'.')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_regression_results(\n",
    "        results,results_residuals,\n",
    "        truth_values,individual_truth_values,\n",
    "        individual_medians,plot_std=True,\n",
    "        bin_edges=10,real_data=None,\n",
    "        plot_dropout_err=True, plot_name='regression.png',\n",
    "        xlabel='$\\\\theta_{E}$',ylabel='$\\\\theta_{E}$',\n",
    "        xmin=None, xmax=None,\n",
    "        ymin=None, ymax=None,\n",
    "        color_std='blue', color_err='green',\n",
    "        color_pop='magenta',plot_label=None,\n",
    "        fill_label_std=None, fill_label_err=None,\n",
    "        figsize=(6, 8),alpha_hist=0.4, \n",
    "        pop_error=None,pop_error_residuals=None,\n",
    "        plot_real_residuals=False):\n",
    "   \n",
    "    pop_err=pop_error\n",
    "    pop_err_r=pop_error_residuals\n",
    "    x_bin,y_bin,y_bin_std,y_bin_percentile1sig,y_bin_percentile2sig,y_bin_percentile3sig=results\n",
    "    x_bin,yr_bin,yr_bin_std,yr_bin_percentile1sig,yr_bin_percentile2sig,yr_bin_percentile3sig=results_residuals\n",
    "\n",
    "\n",
    "    if type(xmin)==type(None):\n",
    "        if type(bin_edges)!=type(10):\n",
    "            xmin=min(bin_edges)\n",
    "            xmax=max(bin_edges)\n",
    "        else:\n",
    "            xmin=min(x_bin)\n",
    "            xmax=max(x_bin)\n",
    "\n",
    "\n",
    "    imed=individual_medians\n",
    "    itruth=individual_truth_values \n",
    "\n",
    "    ty_bin_percentile1sig=np.transpose(y_bin_percentile1sig)\n",
    "    ty_bin_percentile2sig=np.transpose(y_bin_percentile2sig)\n",
    "    ty_bin_percentile3sig=np.transpose(y_bin_percentile3sig)\n",
    "\n",
    "    tyr_bin_percentile1sig=np.transpose(yr_bin_percentile1sig)\n",
    "    tyr_bin_percentile2sig=np.transpose(yr_bin_percentile2sig)\n",
    "    tyr_bin_percentile3sig=np.transpose(yr_bin_percentile3sig)\n",
    "    \n",
    "    # ================================\n",
    "    # Initialize figure canvas\n",
    "    # ================================\n",
    "    ymax = xmax\n",
    "    ymin = xmin\n",
    "    # set up figure\n",
    "    fig = plt.figure(figsize=figsize)         \n",
    "    \n",
    "    # set up grid of panels in figure\n",
    "    grid = plt.GridSpec(3, 2, hspace=0.0, wspace=0.0, \n",
    "                        left=0.1, right=1.0, bottom=0.1, top=1.0,\n",
    "                        height_ratios=[1, 4, 1], width_ratios=[4,1])\n",
    "         \n",
    "    # main axis: predicted statistic vs. true statistic\n",
    "    ax_main = fig.add_subplot(grid[1,0])\n",
    "    \n",
    "    # hist axis [x]: histogram of data on x axis on top\n",
    "    ax_hist_x = fig.add_subplot(grid[0, 0], sharex=ax_main)\n",
    "    \n",
    "    # hist axis [y]: histogram of data on y axis on right\n",
    "    ax_hist_y = fig.add_subplot(grid[1, 1], sharey=ax_main)\n",
    "    \n",
    "    # res axis: residual of statistic vs. true stratistic\n",
    "    ax_diff = fig.add_subplot(grid[2,0], sharex=ax_main)\n",
    "    \n",
    "    \n",
    "    # calculate upper and lower bounds [for systematic error bars]\n",
    "    if plot_std==True:\n",
    "        lo_err, up_err = measure_uplow_std_limits(y_bin, y_bin_std, nsigma=1)#(ys_bin, ys_err_bin)\n",
    "\n",
    "    # calculate upper and lower bounds on diff [for systematic error bars]\n",
    "    if plot_std==True:\n",
    "        lo_err_diff, up_err_diff = measure_uplow_std_limits(yr_bin, yr_bin_std, nsigma=1)\n",
    "        \n",
    "\n",
    "    # ================================\n",
    "    # Plot data [sim]\n",
    "    # ================================\n",
    "    # plot: dropout medians [main axis]\n",
    "    ax_main.plot(x_bin,y_bin, color=color_err, alpha=0.9, label=plot_label)\n",
    "    \n",
    "    # plot: dropouts percentile errors [main axis]\n",
    "    if plot_dropout_err==True:\n",
    "        ax_main.fill_between(x_bin, ty_bin_percentile1sig[0], ty_bin_percentile1sig[1], color=color_err, alpha=0.5, label=fill_label_err)\n",
    "        ax_main.fill_between(x_bin, ty_bin_percentile2sig[0], ty_bin_percentile2sig[1], color=color_err, alpha=0.4)\n",
    "        ax_main.fill_between(x_bin, ty_bin_percentile3sig[0], ty_bin_percentile3sig[1], color=color_err, alpha=0.3)\n",
    "    \n",
    "    # plot: one-to-one line [main axis]\n",
    "    ax_main.plot([xmin, xmax], [ymin, ymax], color='black', ls='dashed', alpha=0.9)\n",
    "    \n",
    "    # plot: errors from standard deviation [main axis]\n",
    "    if plot_std==True:\n",
    "        ax_main.fill_between(x_bin, lo_err, up_err, color=color_std, alpha=0.4, label=fill_label_std)\n",
    "    \n",
    "    # plot: deltas [diff axis]\n",
    "    ax_diff.plot(x_bin,yr_bin, color=color_err, alpha=0.9)\n",
    "    \n",
    "    # plot: statistical errors [diff axis]\n",
    "    if plot_dropout_err==True:\n",
    "        ax_diff.fill_between(x_bin, tyr_bin_percentile1sig[0], tyr_bin_percentile1sig[1], color=color_err, alpha=0.5, label=fill_label_err)\n",
    "        ax_diff.fill_between(x_bin, tyr_bin_percentile2sig[0], tyr_bin_percentile2sig[1], color=color_err, alpha=0.4)\n",
    "        ax_diff.fill_between(x_bin, tyr_bin_percentile3sig[0], tyr_bin_percentile3sig[1], color=color_err, alpha=0.3)\n",
    "    \n",
    "    # plot: systematic errors [main axis]\n",
    "    if plot_std==True:\n",
    "        ax_diff.fill_between(x_bin, lo_err_diff, up_err_diff, color=color_err, alpha=0.9, label=fill_label_std)\n",
    "    \n",
    "    # plot: errors from medians\n",
    "    if pop_error!=None:\n",
    "        xp_bin,yp_bin,yp_bin_std,yp_bin_percentile1sig,yp_bin_percentile2sig,yp_bin_percentile3sig=pop_err\n",
    "        xp_bin,ypr_bin,ypr_bin_std,ypr_bin_percentile1sig,ypr_bin_percentile2sig,ypr_bin_percentile3sig=results_residuals=pop_err_r\n",
    "        typ_bin_percentile1sig=np.transpose(yp_bin_percentile1sig)\n",
    "        typr_bin_percentile1sig=np.transpose(ypr_bin_percentile1sig)\n",
    "        ax_main.fill_between(xp_bin, typ_bin_percentile1sig[0], typ_bin_percentile1sig[1], color=color_pop, alpha=0.4)\n",
    "        ax_diff.fill_between(xp_bin, typr_bin_percentile1sig[0], typr_bin_percentile1sig[1], color=color_pop, alpha=0.4)\n",
    "    # plot: zero line [res axis]\n",
    "    xmin_line, xmax_line = ax_diff.get_xlim()\n",
    "    ax_diff.axhline(y=0.2, ls='dashed', color='magenta', lw=0.5)\n",
    "    ax_diff.axhline(y=0.1, ls='dashed', color='red', lw=0.5)\n",
    "    ax_diff.axhline(y=0, ls='dashed', color='black', lw=0.5)\n",
    "    ax_diff.axhline(y=-0.1, ls='dashed', color='red', lw=0.5)\n",
    "    ax_diff.axhline(y=-0.2, ls='dashed', color='magenta', lw=0.5)\n",
    "\n",
    "        \n",
    "    # ================================\n",
    "    # Plot data [simulation]\n",
    "    # ================================\n",
    "    # plot: histogram on upper x axis\n",
    "    ax_hist_x.hist(itruth, bin_edges, \n",
    "                   color=color_err, alpha=alpha_hist, \n",
    "                   histtype='stepfilled')\n",
    "    \n",
    "    # plot: histogram on right y axis\n",
    "    blank = ax_hist_y.hist(imed, bin_edges, \n",
    "                   color=color_err, alpha=alpha_hist,\n",
    "                   histtype='stepfilled',\n",
    "                   orientation='horizontal')\n",
    "        \n",
    "    # plot: sky data main_axis\n",
    "    if type(real_data)!=type(None):\n",
    "        print('real data')\n",
    "        print(np.array(real_data).shape)\n",
    "        if plot_real_residuals==False:\n",
    "            real_tr,real_me,real_lo,real_hi=np.transpose(real_data)\n",
    "        else:\n",
    "            real_tr,real_me,real_lo,real_hi,realr_me,realr_lo,realr_hi=np.transpose(real_data)\n",
    "        yerr_lo=real_me - real_lo\n",
    "        yerr_hi= real_hi - real_me\n",
    "\n",
    "        yerr_lor=realr_me - realr_lo\n",
    "        yerr_hir= realr_hi - realr_me\n",
    "\n",
    "        #res_me=np.divide(real_me-real_tr,real_tr)\n",
    "        ax_main.errorbar(real_tr, real_me, yerr=[yerr_lo,yerr_hi], fmt='ko', markersize=5,label='DES')\n",
    "        ax_diff.errorbar(real_tr, realr_me, yerr=[yerr_lor,yerr_hir], fmt='ko', markersize=5,label='DES')\n",
    "\n",
    "\n",
    "    \n",
    "    # ================================\n",
    "    # Set plot options\n",
    "    # ================================\n",
    "    # set plot options [main axis]\n",
    "    plt.setp(ax_main.get_xticklabels(), visible=False)\n",
    "    plt.setp(ax_main.get_yticklabels(), visible=True)\n",
    "    plt.setp(ax_diff.get_yticklabels(), visible=False)\n",
    "    plt.setp(ax_hist_x.get_xticklabels(), visible=False)\n",
    "    plt.setp(ax_hist_x.get_yticklabels(), visible=False)\n",
    "    plt.setp(ax_hist_y.get_yticklabels(), visible=False)\n",
    "    plt.setp(ax_hist_y.get_xticklabels(), visible=False)\n",
    "    ax_main.set_ylabel(ylabel)\n",
    "    ax_main.autoscale(enable=True,axis='both')\n",
    "    #ax_main.axis('equal')\n",
    "    ax_main.set_xlim((xmin, xmax))   \n",
    "    ax_main.set_ylim((xmin, xmax))\n",
    "    ax_main.set_ylabel(ylabel)   \n",
    "    ax_main.grid(True, alpha=0.2) \n",
    "    ax_hist_x.grid(True,alpha=0.2)\n",
    "    ax_hist_y.grid(True,alpha=0.2)\n",
    "    ax_main.legend()\n",
    "    #ax_main.set_aspect('equal')\n",
    "    \n",
    "    # set plot options [residual axis]\n",
    "    ax_diff.set_ylim((-0.8,0.8))\n",
    "    ax_diff.set_xlabel(xlabel)\n",
    "    ylabel_plain=ylabel.replace('$','')\n",
    "    ax_diff.set_ylabel('$\\\\frac{\\\\delta '+ylabel_plain+'}{'+ylabel_plain+'}$')\n",
    "    ax_diff.grid(True, alpha=0.2)\n",
    "\n",
    "\n",
    "    plt.savefig(plot_name)\n",
    "\n",
    "\n",
    "def make_bin(col1_dat,col2_dat,bin_edges,return_vals=False,method=\"mean\"):\n",
    "    \"\"\"\n",
    "    col1= array like of the x values\n",
    "    col2= array like of the y values\n",
    "    Funtion that bins in x of a sample of (x,y) pairs.\n",
    "    USe the return_vals if you want a list of y values in each bin\n",
    "    \"\"\"\n",
    "    print (type(col2_dat))\n",
    "    print (np.array(col2_dat).shape)\n",
    "    bin_vals = [[] for x in range(0,len(bin_edges)-1)]\n",
    "    bin_valsx = [[] for x in range(0,len(bin_edges)-1)]\n",
    "    bin_median=np.zeros(len(bin_edges)-1)\n",
    "    bin_mean=np.zeros(len(bin_edges)-1)\n",
    "    bin_std=np.zeros(len(bin_edges)-1)\n",
    "    bin_center=np.zeros(len(bin_edges)-1)\n",
    "\n",
    "    bin_percentile1sig=[[] for x in range(0,len(bin_edges)-1)]\n",
    "    bin_percentile2sig=[[] for x in range(0,len(bin_edges)-1)]\n",
    "    bin_percentile3sig=[[] for x in range(0,len(bin_edges)-1)]\n",
    "\n",
    "    for i in range(0,len(col1_dat)):\n",
    "        for j in range(0,len(bin_edges)-1):\n",
    "            if bin_edges[j] <= col1_dat[i] <= bin_edges[j+1]:\n",
    "                bin_vals[j].append(col2_dat[i])\n",
    "                bin_valsx[j].append(col1_dat[i])            \n",
    "\n",
    "    for i in range(0,len(bin_edges)-1):\n",
    "        try:\n",
    "            bin_mean[i]=np.mean(bin_vals[i])\n",
    "        except:\n",
    "            print(type(bin_vals[i]))\n",
    "            print(\"Error!\")\n",
    "        bin_median[i]=np.median(bin_vals[i])\n",
    "        bin_std[i]=np.std(bin_vals[i])\n",
    "        if method==\"percentile\":\n",
    "            if len(bin_vals[i])!=0:\n",
    "                \n",
    "                lowlim=np.percentile(bin_vals[i],15.87)\n",
    "                highlim=np.percentile(bin_vals[i],84.13)\n",
    "                lowlim2sig=np.percentile(bin_vals[i],2.25)\n",
    "                highlim2sig=np.percentile(bin_vals[i],97.6)\n",
    "                lowlim3sig=np.percentile(bin_vals[i],0.15)\n",
    "                highlim3sig=np.percentile(bin_vals[i],99.7)\n",
    "            else:\n",
    "                lowlim=0\n",
    "                highlim=0\n",
    "                lowlim2sig=0\n",
    "                highlim2sig=0\n",
    "                lowlim3sig=0\n",
    "                highlim3sig=0\n",
    "            bin_percentile1sig[i]=[lowlim,highlim]\n",
    "            bin_percentile2sig[i]=[lowlim2sig,highlim2sig]\n",
    "            bin_percentile3sig[i]=[lowlim3sig,highlim3sig]\n",
    "            \n",
    "        if bin_std[i]<0.001:\n",
    "            print(\"zero bin std data:\")\n",
    "            print(bin_vals[i])\n",
    "        bin_center[i]= (bin_edges[i]+bin_edges[i+1])/2\n",
    "\n",
    "    if return_vals==False:\n",
    "        if method=='median':\n",
    "            return bin_center,bin_median,bin_std\n",
    "        elif method=='percentile':\n",
    "            return bin_center,bin_median,bin_std,bin_percentile1sig,bin_percentile2sig,bin_percentile3sig,       \n",
    "        else:\n",
    "            return bin_center,bin_mean,bin_std\n",
    "    else:\n",
    "        # if you want the a list of values inside each bin,\n",
    "        # so you can perform your own bin metrics.\n",
    "        return bin_center,bin_valsx,bin_vals\n",
    "    \n",
    "def measure_uplow_std_limits(ybin, bin_stdevs, nsigma=1):\n",
    "    lo = [ybin[i] - nsigma*bin_stdevs[i] for i in range(0,len(ybin))] \n",
    "    hi = [ybin[i] + nsigma*bin_stdevs[i] for i in range(0,len(ybin))]\n",
    "        \n",
    "    return lo, hi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "(200,)\n",
      "zero bin std data:\n",
      "[-5.48551]\n",
      "zero bin std data:\n",
      "[-5.478584]\n",
      "zero bin std data:\n",
      "[-5.4603086]\n",
      "zero bin std data:\n",
      "[-5.4578013]\n",
      "zero bin std data:\n",
      "[-5.424314]\n",
      "zero bin std data:\n",
      "[-5.423528]\n",
      "zero bin std data:\n",
      "[-5.401954]\n",
      "zero bin std data:\n",
      "[-5.3818727]\n",
      "zero bin std data:\n",
      "[-5.339108]\n",
      "zero bin std data:\n",
      "[-5.334227]\n",
      "zero bin std data:\n",
      "[-5.2593217]\n",
      "zero bin std data:\n",
      "[-5.1665525]\n",
      "zero bin std data:\n",
      "[-5.1391377]\n",
      "zero bin std data:\n",
      "[-4.991859]\n",
      "zero bin std data:\n",
      "[-4.98074]\n",
      "zero bin std data:\n",
      "[-4.9181685]\n",
      "zero bin std data:\n",
      "[-4.775209]\n",
      "zero bin std data:\n",
      "[-4.6562552]\n",
      "zero bin std data:\n",
      "[-4.6230063]\n",
      "zero bin std data:\n",
      "[-4.587458]\n",
      "zero bin std data:\n",
      "[-4.551089]\n",
      "zero bin std data:\n",
      "[-4.4231834]\n",
      "zero bin std data:\n",
      "[-4.3873734]\n",
      "zero bin std data:\n",
      "[-4.3736825]\n",
      "zero bin std data:\n",
      "[-4.3059835]\n",
      "zero bin std data:\n",
      "[-4.3009686]\n",
      "zero bin std data:\n",
      "[-4.2934027]\n",
      "zero bin std data:\n",
      "[-4.2907643]\n",
      "zero bin std data:\n",
      "[-4.2045975]\n",
      "zero bin std data:\n",
      "[-4.1825986]\n",
      "zero bin std data:\n",
      "[-4.1647487]\n",
      "zero bin std data:\n",
      "[-4.153446]\n",
      "zero bin std data:\n",
      "[-4.1503015]\n",
      "zero bin std data:\n",
      "[-4.1379633]\n",
      "zero bin std data:\n",
      "[-4.071291]\n",
      "zero bin std data:\n",
      "[-4.051655]\n",
      "zero bin std data:\n",
      "[-4.0328093]\n",
      "zero bin std data:\n",
      "[-3.9789207]\n",
      "zero bin std data:\n",
      "[-3.944572]\n",
      "zero bin std data:\n",
      "[-3.9352438]\n",
      "zero bin std data:\n",
      "[-3.8414235]\n",
      "zero bin std data:\n",
      "[-3.8304522]\n",
      "zero bin std data:\n",
      "[-3.7902286]\n",
      "zero bin std data:\n",
      "[-3.7802746]\n",
      "zero bin std data:\n",
      "[-3.7665784]\n",
      "zero bin std data:\n",
      "[-3.7306778]\n",
      "zero bin std data:\n",
      "[-3.6952462]\n",
      "zero bin std data:\n",
      "[-3.6463473]\n",
      "zero bin std data:\n",
      "[-3.641824]\n",
      "zero bin std data:\n",
      "[-3.6167958]\n",
      "zero bin std data:\n",
      "[-3.6015403]\n",
      "zero bin std data:\n",
      "[-3.5692356]\n",
      "zero bin std data:\n",
      "[-3.5429542]\n",
      "zero bin std data:\n",
      "[-3.5291283]\n",
      "zero bin std data:\n",
      "[-3.5263436]\n",
      "zero bin std data:\n",
      "[-3.4946146]\n",
      "zero bin std data:\n",
      "[-3.4881089]\n",
      "zero bin std data:\n",
      "[-3.4622533]\n",
      "zero bin std data:\n",
      "[-3.4433033]\n",
      "zero bin std data:\n",
      "[-3.415123]\n",
      "zero bin std data:\n",
      "[-3.3713543]\n",
      "zero bin std data:\n",
      "[-3.3527534]\n",
      "zero bin std data:\n",
      "[-3.3466794]\n",
      "zero bin std data:\n",
      "[-3.3284166]\n",
      "zero bin std data:\n",
      "[-3.3044493]\n",
      "zero bin std data:\n",
      "[-3.2589152]\n",
      "zero bin std data:\n",
      "[-3.1957417]\n",
      "zero bin std data:\n",
      "[-3.1955554]\n",
      "zero bin std data:\n",
      "[-3.1449487]\n",
      "zero bin std data:\n",
      "[-3.1409347]\n",
      "zero bin std data:\n",
      "[-3.1003714]\n",
      "zero bin std data:\n",
      "[-3.0874171]\n",
      "zero bin std data:\n",
      "[-3.0511258]\n",
      "zero bin std data:\n",
      "[-3.0349643]\n",
      "zero bin std data:\n",
      "[-3.0099475]\n",
      "zero bin std data:\n",
      "[-3.0013797]\n",
      "zero bin std data:\n",
      "[-2.9859996]\n",
      "zero bin std data:\n",
      "[-2.9816294]\n",
      "zero bin std data:\n",
      "[-2.975722]\n",
      "zero bin std data:\n",
      "[-2.971152]\n",
      "zero bin std data:\n",
      "[-2.9518583]\n",
      "zero bin std data:\n",
      "[-2.9414165]\n",
      "zero bin std data:\n",
      "[-2.926284]\n",
      "zero bin std data:\n",
      "[-2.873842]\n",
      "zero bin std data:\n",
      "[-2.8550742]\n",
      "zero bin std data:\n",
      "[-2.8161604]\n",
      "zero bin std data:\n",
      "[-2.8076901]\n",
      "zero bin std data:\n",
      "[-2.736564]\n",
      "zero bin std data:\n",
      "[-2.6254232]\n",
      "zero bin std data:\n",
      "[-2.6226244]\n",
      "zero bin std data:\n",
      "[-2.6052835]\n",
      "zero bin std data:\n",
      "[-2.5640042]\n",
      "zero bin std data:\n",
      "[-2.5624382]\n",
      "zero bin std data:\n",
      "[-2.5071962]\n",
      "zero bin std data:\n",
      "[-2.4859111]\n",
      "zero bin std data:\n",
      "[-2.4479172]\n",
      "zero bin std data:\n",
      "[-2.4045777]\n",
      "zero bin std data:\n",
      "[-2.396506]\n",
      "zero bin std data:\n",
      "[-2.3918345]\n",
      "zero bin std data:\n",
      "[-2.3636258]\n",
      "zero bin std data:\n",
      "[-2.36289]\n",
      "zero bin std data:\n",
      "[-2.3524675]\n",
      "zero bin std data:\n",
      "[-2.3432672]\n",
      "zero bin std data:\n",
      "[-2.2788613]\n",
      "zero bin std data:\n",
      "[-2.241847]\n",
      "zero bin std data:\n",
      "[-2.2306466]\n",
      "zero bin std data:\n",
      "[-2.1722858]\n",
      "zero bin std data:\n",
      "[-2.160454]\n",
      "zero bin std data:\n",
      "[-2.1394112]\n",
      "zero bin std data:\n",
      "[-2.1373875]\n",
      "zero bin std data:\n",
      "[-2.1300094]\n",
      "zero bin std data:\n",
      "[-2.0460758]\n",
      "zero bin std data:\n",
      "[-2.0385745]\n",
      "zero bin std data:\n",
      "[-1.9976095]\n",
      "zero bin std data:\n",
      "[-1.985228]\n",
      "zero bin std data:\n",
      "[-1.967957]\n",
      "zero bin std data:\n",
      "[-1.9630265]\n",
      "zero bin std data:\n",
      "[-1.9457662]\n",
      "zero bin std data:\n",
      "[-1.9279116]\n",
      "zero bin std data:\n",
      "[-1.8630652]\n",
      "zero bin std data:\n",
      "[-1.8603646]\n",
      "zero bin std data:\n",
      "[-1.8352827]\n",
      "zero bin std data:\n",
      "[-1.8319799]\n",
      "zero bin std data:\n",
      "[-1.807423]\n",
      "zero bin std data:\n",
      "[-1.799658]\n",
      "zero bin std data:\n",
      "[-1.7682859]\n",
      "zero bin std data:\n",
      "[-1.7664348]\n",
      "zero bin std data:\n",
      "[-1.7567958]\n",
      "zero bin std data:\n",
      "[-1.7446749]\n",
      "zero bin std data:\n",
      "[-1.7142152]\n",
      "zero bin std data:\n",
      "[-1.6959015]\n",
      "zero bin std data:\n",
      "[-1.6919636]\n",
      "zero bin std data:\n",
      "[-1.6564605]\n",
      "zero bin std data:\n",
      "[-1.6515044]\n",
      "zero bin std data:\n",
      "[-1.6384333]\n",
      "zero bin std data:\n",
      "[-1.6361692]\n",
      "zero bin std data:\n",
      "[-1.6089665]\n",
      "zero bin std data:\n",
      "[-1.6078024]\n",
      "zero bin std data:\n",
      "[-1.5740992]\n",
      "zero bin std data:\n",
      "[-1.5289499]\n",
      "zero bin std data:\n",
      "[-1.4862293]\n",
      "zero bin std data:\n",
      "[-1.4510101]\n",
      "zero bin std data:\n",
      "[-1.4104785]\n",
      "zero bin std data:\n",
      "[-1.385837]\n",
      "zero bin std data:\n",
      "[-1.3849883]\n",
      "zero bin std data:\n",
      "[-1.3498267]\n",
      "zero bin std data:\n",
      "[-1.2234023]\n",
      "zero bin std data:\n",
      "[-1.2216958]\n",
      "zero bin std data:\n",
      "[-1.2139075]\n",
      "zero bin std data:\n",
      "[-1.213291]\n",
      "zero bin std data:\n",
      "[-1.1932341]\n",
      "zero bin std data:\n",
      "[-1.1833389]\n",
      "zero bin std data:\n",
      "[-1.166296]\n",
      "zero bin std data:\n",
      "[-1.1515754]\n",
      "zero bin std data:\n",
      "[-1.1312352]\n",
      "zero bin std data:\n",
      "[-1.119417]\n",
      "zero bin std data:\n",
      "[-1.093932]\n",
      "zero bin std data:\n",
      "[-1.0910932]\n",
      "zero bin std data:\n",
      "[-1.0676836]\n",
      "zero bin std data:\n",
      "[-1.066001]\n",
      "zero bin std data:\n",
      "[-1.0655829]\n",
      "zero bin std data:\n",
      "[-1.0300136]\n",
      "zero bin std data:\n",
      "[-1.0275375]\n",
      "zero bin std data:\n",
      "[-1.0250789]\n",
      "zero bin std data:\n",
      "[-1.0011563]\n",
      "zero bin std data:\n",
      "[-0.9994409]\n",
      "zero bin std data:\n",
      "[-0.9972681]\n",
      "zero bin std data:\n",
      "[-0.9951249]\n",
      "zero bin std data:\n",
      "[-0.9924173]\n",
      "zero bin std data:\n",
      "[-0.98854685]\n",
      "zero bin std data:\n",
      "[-0.9830381]\n",
      "zero bin std data:\n",
      "[-0.9798637]\n",
      "zero bin std data:\n",
      "[-0.97741646]\n",
      "zero bin std data:\n",
      "[-0.97553504]\n",
      "zero bin std data:\n",
      "[-0.96622175]\n",
      "zero bin std data:\n",
      "[-0.9647028]\n",
      "zero bin std data:\n",
      "[-0.9588818]\n",
      "zero bin std data:\n",
      "[-0.95084494]\n",
      "zero bin std data:\n",
      "[-0.9484134, -0.94793624]\n",
      "zero bin std data:\n",
      "[-0.9484134, -0.94793624]\n",
      "zero bin std data:\n",
      "[-0.9413443]\n",
      "zero bin std data:\n",
      "[-0.9368749]\n",
      "zero bin std data:\n",
      "[-0.9360906]\n",
      "zero bin std data:\n",
      "[-0.935258]\n",
      "zero bin std data:\n",
      "[-0.9351546]\n",
      "zero bin std data:\n",
      "[-0.93178546]\n",
      "zero bin std data:\n",
      "[-0.92894685]\n",
      "zero bin std data:\n",
      "[-0.9218908]\n",
      "zero bin std data:\n",
      "[-0.9211629]\n",
      "zero bin std data:\n",
      "[-0.9128251]\n",
      "zero bin std data:\n",
      "[-0.9060021]\n",
      "zero bin std data:\n",
      "[-0.9056249]\n",
      "zero bin std data:\n",
      "[-0.9046451]\n",
      "zero bin std data:\n",
      "[-0.90349156]\n",
      "zero bin std data:\n",
      "[-0.9015921]\n",
      "zero bin std data:\n",
      "[-0.9001468]\n",
      "zero bin std data:\n",
      "[-0.8992924]\n",
      "zero bin std data:\n",
      "[-0.89608634]\n",
      "<class 'numpy.ndarray'>\n",
      "(200,)\n",
      "zero bin std data:\n",
      "[-1.5978152281017537]\n",
      "zero bin std data:\n",
      "[-1.8560784348491188]\n",
      "zero bin std data:\n",
      "[-1.8714283248496018]\n",
      "zero bin std data:\n",
      "[-2.0585679802905466]\n",
      "zero bin std data:\n",
      "[-1.562843109291484]\n",
      "zero bin std data:\n",
      "[-1.6570986318118264]\n",
      "zero bin std data:\n",
      "[-1.5360069694349088]\n",
      "zero bin std data:\n",
      "[-1.7938740686341321]\n",
      "zero bin std data:\n",
      "[-0.9999148061427581]\n",
      "zero bin std data:\n",
      "[-1.531802829769025]\n",
      "zero bin std data:\n",
      "[-1.1851676489404477]\n",
      "zero bin std data:\n",
      "[-0.6879912246688497]\n",
      "zero bin std data:\n",
      "[-2.1353192424552767]\n",
      "zero bin std data:\n",
      "[-0.21167031809211598]\n",
      "zero bin std data:\n",
      "[-0.21609812224064767]\n",
      "zero bin std data:\n",
      "[-2.756198519755762]\n",
      "zero bin std data:\n",
      "[-2.8032068180520135]\n",
      "zero bin std data:\n",
      "[-1.8874234480833167]\n",
      "zero bin std data:\n",
      "[-2.1094721283332074]\n",
      "zero bin std data:\n",
      "[-2.0639496968728337]\n",
      "zero bin std data:\n",
      "[-2.8785265169295156]\n",
      "zero bin std data:\n",
      "[-1.619070598431329]\n",
      "zero bin std data:\n",
      "[-2.0853575561365667]\n",
      "zero bin std data:\n",
      "[-2.093190856631097]\n",
      "zero bin std data:\n",
      "[-0.8131940385554229]\n",
      "zero bin std data:\n",
      "[-0.9024904066135332]\n",
      "zero bin std data:\n",
      "[-2.682276328963492]\n",
      "zero bin std data:\n",
      "[-2.710680106828125]\n",
      "zero bin std data:\n",
      "[-0.8901614872410633]\n",
      "zero bin std data:\n",
      "[-1.0073594867714033]\n",
      "zero bin std data:\n",
      "[-0.8551696638385761]\n",
      "zero bin std data:\n",
      "[-0.6591021953193592]\n",
      "zero bin std data:\n",
      "[-0.888519324275863]\n",
      "zero bin std data:\n",
      "[-1.888908638469012]\n",
      "zero bin std data:\n",
      "[-1.1152926703077155]\n",
      "zero bin std data:\n",
      "[-2.522667809366434]\n",
      "zero bin std data:\n",
      "[-3.0308881114217674]\n",
      "zero bin std data:\n",
      "[-1.7952211195730667]\n",
      "zero bin std data:\n",
      "[-2.157901693870056]\n",
      "zero bin std data:\n",
      "[-2.814992876430591]\n",
      "zero bin std data:\n",
      "[-1.0074260034876505]\n",
      "zero bin std data:\n",
      "[-2.4527505433310766]\n",
      "zero bin std data:\n",
      "[-1.716970491455616]\n",
      "zero bin std data:\n",
      "[-2.8362135907112656]\n",
      "zero bin std data:\n",
      "[-2.719590595713332]\n",
      "zero bin std data:\n",
      "[-1.757322901845803]\n",
      "zero bin std data:\n",
      "[-2.1457816793588096]\n",
      "zero bin std data:\n",
      "[-0.9218633614135594]\n",
      "zero bin std data:\n",
      "[-1.5043553857019942]\n",
      "zero bin std data:\n",
      "[-1.2399100362862372]\n",
      "zero bin std data:\n",
      "[-1.7384275318607127]\n",
      "zero bin std data:\n",
      "[-0.968474427781558]\n",
      "zero bin std data:\n",
      "[-1.236374535988038]\n",
      "zero bin std data:\n",
      "[-1.850745189129116]\n",
      "zero bin std data:\n",
      "[-1.9626797483001286]\n",
      "zero bin std data:\n",
      "[-1.772540497481746]\n",
      "zero bin std data:\n",
      "[-1.62006936676208]\n",
      "zero bin std data:\n",
      "[-1.3095344542168674]\n",
      "zero bin std data:\n",
      "[-1.7168442639912103]\n",
      "zero bin std data:\n",
      "[-2.2109388367592895]\n",
      "zero bin std data:\n",
      "[-1.052209809060551]\n",
      "zero bin std data:\n",
      "[-1.093579770159069]\n",
      "zero bin std data:\n",
      "[-1.4709530078991835]\n",
      "zero bin std data:\n",
      "[-1.6583568848689163]\n",
      "zero bin std data:\n",
      "[-1.7447068774254029]\n",
      "zero bin std data:\n",
      "[-2.9673047306613425]\n",
      "zero bin std data:\n",
      "[-1.2533061949506188]\n",
      "zero bin std data:\n",
      "[-2.750979331080024]\n",
      "zero bin std data:\n",
      "[-1.635033133382146]\n",
      "zero bin std data:\n",
      "[-3.1505653469242034]\n",
      "zero bin std data:\n",
      "[-3.2166914618624247]\n",
      "zero bin std data:\n",
      "[-2.8587036576307914]\n",
      "zero bin std data:\n",
      "[-2.0779679543933285]\n",
      "zero bin std data:\n",
      "[-2.717713589713691]\n",
      "zero bin std data:\n",
      "[-1.9025830749218793]\n",
      "zero bin std data:\n",
      "[-1.936574936213253]\n",
      "zero bin std data:\n",
      "[-1.5722126648789307]\n",
      "zero bin std data:\n",
      "[-2.329622689329222]\n",
      "zero bin std data:\n",
      "[-2.258139448041824]\n",
      "zero bin std data:\n",
      "[-2.262043031916898]\n",
      "zero bin std data:\n",
      "[-2.273132657646943]\n",
      "zero bin std data:\n",
      "[-2.9424748854434872]\n",
      "zero bin std data:\n",
      "[-2.9475491463436523]\n",
      "zero bin std data:\n",
      "[-1.4595590619184715]\n",
      "zero bin std data:\n",
      "[-1.1093112716142488]\n",
      "zero bin std data:\n",
      "[-1.1188671623504378]\n",
      "zero bin std data:\n",
      "[-3.3188649072855014]\n",
      "zero bin std data:\n",
      "[-3.3152970936215747]\n",
      "zero bin std data:\n",
      "[-1.2564655334587413]\n",
      "zero bin std data:\n",
      "[-1.822997841467147]\n",
      "zero bin std data:\n",
      "[-2.519764289218662]\n",
      "zero bin std data:\n",
      "[-1.2118238468901543]\n",
      "zero bin std data:\n",
      "[-3.250016592482377]\n",
      "zero bin std data:\n",
      "[-3.3119967822535767]\n",
      "zero bin std data:\n",
      "[-3.3105929832689096]\n",
      "zero bin std data:\n",
      "[-2.958137602234801]\n",
      "zero bin std data:\n",
      "[-1.4773880916083182]\n",
      "zero bin std data:\n",
      "[-1.3321811393674179]\n",
      "zero bin std data:\n",
      "[-2.3741800059514766]\n",
      "zero bin std data:\n",
      "[-1.8804214781813378]\n",
      "zero bin std data:\n",
      "[-1.9849807420321965]\n",
      "zero bin std data:\n",
      "[-2.8052813018539116]\n",
      "zero bin std data:\n",
      "[-2.887693367129336]\n",
      "zero bin std data:\n",
      "[-2.830204726082053]\n",
      "zero bin std data:\n",
      "[-1.6134641768625353]\n",
      "zero bin std data:\n",
      "[-1.16311225452269]\n",
      "zero bin std data:\n",
      "[-3.1964756342969665]\n",
      "zero bin std data:\n",
      "[-2.7425082476702953]\n",
      "zero bin std data:\n",
      "[-2.3029875973376086]\n",
      "zero bin std data:\n",
      "[-3.219284654137475]\n",
      "zero bin std data:\n",
      "[-2.9611809393750925]\n",
      "zero bin std data:\n",
      "[-1.2282419175362722]\n",
      "zero bin std data:\n",
      "[-3.3497834248681224]\n",
      "zero bin std data:\n",
      "[-1.2855597651974773]\n",
      "zero bin std data:\n",
      "[-2.2004270007779505]\n",
      "zero bin std data:\n",
      "[-2.497762639268399]\n",
      "zero bin std data:\n",
      "[-2.2630697128456037]\n",
      "zero bin std data:\n",
      "[-2.5436750824084013]\n",
      "zero bin std data:\n",
      "[-3.5120357264754656]\n",
      "zero bin std data:\n",
      "[-3.081279380897153]\n",
      "zero bin std data:\n",
      "[-3.4973345022222913]\n",
      "zero bin std data:\n",
      "[-2.256695234874884]\n",
      "zero bin std data:\n",
      "[-3.4587278649983557]\n",
      "zero bin std data:\n",
      "[-2.33660296624517]\n",
      "zero bin std data:\n",
      "[-3.3543706269345703]\n",
      "zero bin std data:\n",
      "[-1.8451207328687251]\n",
      "zero bin std data:\n",
      "[-3.0901835134298836]\n",
      "zero bin std data:\n",
      "[-3.26499978272567]\n",
      "zero bin std data:\n",
      "[-3.2694246868111203]\n",
      "zero bin std data:\n",
      "[-1.9501439700153593]\n",
      "zero bin std data:\n",
      "[-2.042261383084906]\n",
      "zero bin std data:\n",
      "[-4.32081349697258]\n",
      "zero bin std data:\n",
      "[-2.2762231292605706]\n",
      "zero bin std data:\n",
      "[-3.380562539436143]\n",
      "zero bin std data:\n",
      "[-3.780424199178651]\n",
      "zero bin std data:\n",
      "[-4.495480013650715]\n",
      "zero bin std data:\n",
      "[-2.8333510289379853]\n",
      "zero bin std data:\n",
      "[-4.914263395129862]\n",
      "zero bin std data:\n",
      "[-3.0525364885018504]\n",
      "zero bin std data:\n",
      "[-3.443698805096022]\n",
      "zero bin std data:\n",
      "[-3.3980166962137917]\n",
      "zero bin std data:\n",
      "[-3.3104260217407173]\n",
      "zero bin std data:\n",
      "[-4.576710300872363]\n",
      "zero bin std data:\n",
      "[-4.382204983443845]\n",
      "zero bin std data:\n",
      "[-4.376509985512933]\n",
      "zero bin std data:\n",
      "[-2.556289409731634]\n",
      "zero bin std data:\n",
      "[-1.4205418426155145]\n",
      "zero bin std data:\n",
      "[-1.4156945453996566]\n",
      "zero bin std data:\n",
      "[-1.4004837533573056]\n",
      "zero bin std data:\n",
      "[-1.4004486077406635]\n",
      "zero bin std data:\n",
      "[-2.4572755123136973]\n",
      "zero bin std data:\n",
      "[-2.2645315179644623]\n",
      "zero bin std data:\n",
      "[-1.0783016004915273]\n",
      "zero bin std data:\n",
      "[-0.8878157622960438]\n",
      "zero bin std data:\n",
      "[-0.7352574702246109]\n",
      "zero bin std data:\n",
      "[-0.69718363012482]\n",
      "zero bin std data:\n",
      "[-0.8052570137610113]\n",
      "zero bin std data:\n",
      "[-0.8277965589948233]\n",
      "zero bin std data:\n",
      "[-1.7203056467777658]\n",
      "zero bin std data:\n",
      "[-1.678637977510567]\n",
      "zero bin std data:\n",
      "[-1.6687425096693045]\n",
      "zero bin std data:\n",
      "[-0.5833964956794181]\n",
      "zero bin std data:\n",
      "[-0.5564025024474909]\n",
      "zero bin std data:\n",
      "[-0.5310113578143183]\n",
      "zero bin std data:\n",
      "[-0.42538274746709975]\n",
      "zero bin std data:\n",
      "[-0.7651384553109357]\n",
      "zero bin std data:\n",
      "[-0.7683125004083285]\n",
      "zero bin std data:\n",
      "[-0.7745005042759522]\n",
      "zero bin std data:\n",
      "[-0.7873664477578554]\n",
      "zero bin std data:\n",
      "[-0.4899724346168831]\n",
      "zero bin std data:\n",
      "[-0.8591460912537364]\n",
      "zero bin std data:\n",
      "[-0.8969368787741653]\n",
      "zero bin std data:\n",
      "[-0.8644792004541858]\n",
      "zero bin std data:\n",
      "[-0.7352975453580988]\n",
      "zero bin std data:\n",
      "[-1.105581597961459]\n",
      "zero bin std data:\n",
      "[-0.991303197360719]\n",
      "zero bin std data:\n",
      "[-0.8980542589881788]\n",
      "zero bin std data:\n",
      "[-0.6570456274605109]\n",
      "zero bin std data:\n",
      "[-0.41319009390294065]\n",
      "zero bin std data:\n",
      "[-0.3500946518553615]\n",
      "zero bin std data:\n",
      "[-0.33990524818822393]\n",
      "zero bin std data:\n",
      "[-0.2691295600553949]\n",
      "zero bin std data:\n",
      "[-0.3196227588869599]\n",
      "zero bin std data:\n",
      "[-0.2983984796363268]\n",
      "zero bin std data:\n",
      "[-0.28498067653838627]\n",
      "zero bin std data:\n",
      "[-0.18889393875620913]\n",
      "zero bin std data:\n",
      "[-0.2189082602841934]\n",
      "zero bin std data:\n",
      "[-0.4711459530977624]\n",
      "zero bin std data:\n",
      "[-0.56764674257779]\n",
      "zero bin std data:\n",
      "[-0.820533633839722]\n",
      "zero bin std data:\n",
      "[-0.7490257920909048]\n",
      "zero bin std data:\n",
      "[-1.0851951798292847]\n",
      "zero bin std data:\n",
      "[-0.9862436094167868]\n",
      "zero bin std data:\n",
      "[-0.8243584698573815]\n",
      "zero bin std data:\n",
      "[-0.7286680561675231]\n",
      "zero bin std data:\n",
      "[-0.36956275089684626]\n"
     ]
    }
   ],
   "source": [
    "y_true = np.sort(test_label)\n",
    "y_pred = np.sort(np.squeeze(pred))\n",
    "error = np.divide(np.abs(np.subtract(y_pred,y_true)), y_true)*100\n",
    "\n",
    "bin_edges = [(y_true[i]+y_true[i+1])/2 for i in range((len(y_true)-1))]\n",
    "\n",
    "results = make_bin(y_true, y_pred,bin_edges,method='percentile')\n",
    "\n",
    "results_residual = make_bin(y_true, error, bin_edges, method='percentile')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_regression_results(results, results_residual,truth_values=np.unique(y_true),individual_truth_values=y_true,individual_medians=y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Neg del 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_x = pd.read_csv('csv/x_test_neg_del0.csv')\n",
    "dataset_y = pd.read_csv('csv/y_test_neg_del0.csv')\n",
    "\n",
    "path_folder = 'data/matrix_neg_del_0//'\n",
    "\n",
    "x_test = dataset_x.iloc[:,1].values\n",
    "y_test = dataset_y.iloc[:,1].values\n",
    "\n",
    "test_data = []\n",
    "load_test_data = []\n",
    "for file in x_test:\n",
    "        path = os.path.abspath(path_folder+ file)\n",
    "        H = np.load(path)\n",
    "        load_test_data.append(H)\n",
    "load_test_data = np.array(load_test_data)\n",
    "test_data = load_test_data[:,:,:,np.newaxis]\n",
    "test_data -= np.mean(test_data)\n",
    "test_data /= np.std(test_data)\n",
    "\n",
    "test_label = []\n",
    "for file in y_test:\n",
    "    test_label.append(file)\n",
    "test_label = np.array(test_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = load_model('h5/train_neg_del0.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 10s 97ms/step\n",
      "Score (RMSE): 0.005405196906876172\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"
    }
   ],
   "source": [
    "pred = model.predict(test_data, batch_size=128, verbose=1)\n",
    "score = np.sqrt(metrics.mean_squared_error(pred,test_label))\n",
    "print(\"Score (RMSE): {}\".format(score))\n",
    "\n",
    "chart_regression(pred.flatten(),test_label, plot_name='train_neg_del0')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000,)\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"
    }
   ],
   "source": [
    "dataset = np.loadtxt('neg_del_0.dat')\n",
    "\n",
    "x_dat, y_dat = dataset[:,0],dataset[:,1]\n",
    "\n",
    "print(y_dat.shape)\n",
    "\n",
    "gr = plt.plot(x_dat,y_dat,'.')\n"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}