visualize_outputs.ipynb

{
 "metadata": {
  "name": ""
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pandas as pd\n",
      "import matplotlib.pyplot as plt\n",
      "import matplotlib.patches as mp\n",
      "import seaborn as sns\n",
      "sns.set()\n",
      "#sns.set_context(\"paper\")\n",
      "%matplotlib inline"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 58
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "splitnum = 3\n",
      "meltype = 2\n",
      "if meltype == 1:\n",
      "    olap = 0.4\n",
      "    nu = 'n015'\n",
      "else:\n",
      "    olap = 0.5\n",
      "    nu = 'n005'\n",
      "\n",
      "outdir = \"/Users/rachelbittner/Desktop/exper_output/ver1\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 285
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "melodia_df = (pd.io.json.read_json(\"melodia_scores/melody%s/nu_%s_melody%s.json\" % (meltype, nu, meltype))).transpose()\n",
      "clf_df = pd.DataFrame.from_csv(\"%s/CC_exper_mel%s/splitnum_%s/olap_%s/all_mel_scores.csv\" % (outdir, meltype, splitnum, olap))\n",
      "diff_df = pd.DataFrame(columns=['OA melodia', 'OA clf', 'Diff'])\n",
      "diff_df.loc[:, 'OA melodia'] = melodia_df.loc[clf_df.index, 'Overall Accuracy']\n",
      "diff_df.loc[:, 'OA clf'] = clf_df['OA']\n",
      "diff_df.loc[:, 'Diff'] = melodia_df.loc[clf_df.index, 'Overall Accuracy'] - clf_df['OA']\n",
      "print diff_df\n",
      "print diff_df.describe()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "                                       OA melodia    OA clf      Diff\n",
        "Phoenix_SeanCaughlinsTheScartaglen       0.515244  0.468565  0.046679\n",
        "MusicDelta_LatinJazz                     0.529300  0.376734  0.152566\n",
        "MusicDelta_FunkJazz                      0.607898  0.639281 -0.031382\n",
        "Phoenix_ElzicsFarewell                   0.334339  0.353304 -0.018965\n",
        "Phoenix_LarkOnTheStrandDrummondCastle    0.511190  0.645931 -0.134741\n",
        "MatthewEntwistle_ImpressionsOfSaturn     0.526184  0.490820  0.035364\n",
        "SweetLights_YouLetMeDown                 0.563910  0.508493  0.055417\n",
        "Phoenix_BrokenPledgeChicagoReel          0.501287  0.534522 -0.033235\n",
        "MatthewEntwistle_FairerHopes             0.264697  0.325380 -0.060683\n",
        "Creepoid_OldTree                         0.538993  0.598233 -0.059239\n",
        "Phoenix_ColliersDaughter                 0.594568  0.580319  0.014249\n",
        "MusicDelta_InTheHalloftheMountainKing    0.053245  0.112479 -0.059235\n",
        "MatthewEntwistle_TheArch                 0.723347  0.587746  0.135600\n",
        "MatthewEntwistle_Lontano                 0.307679  0.257675  0.050004\n",
        "MatthewEntwistle_TheFlaxenField          0.451285  0.441201  0.010084\n",
        "HeladoNegro_MitadDelMundo                0.619466  0.508826  0.110639\n",
        "MatthewEntwistle_DontYouEver             0.639344  0.554730  0.084615\n",
        "MichaelKropf_AllGoodThings               0.382473  0.453482 -0.071009\n",
        "PortStWillow_StayEven                    0.577832  0.403765  0.174067\n",
        "Phoenix_ScotchMorris                     0.624656  0.653002 -0.028346\n",
        "MusicDelta_Pachelbel                     0.311158  0.382343 -0.071185\n",
        "       OA melodia     OA clf       Diff\n",
        "count   21.000000  21.000000  21.000000\n",
        "mean     0.484671   0.470325   0.014346\n",
        "std      0.158045   0.137777   0.083236\n",
        "min      0.053245   0.112479  -0.134741\n",
        "25%      0.382473   0.382343  -0.059235\n",
        "50%      0.526184   0.490820   0.010084\n",
        "75%      0.594568   0.580319   0.055417\n",
        "max      0.723347   0.653002   0.174067\n"
       ]
      }
     ],
     "prompt_number": 286
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "trackid='MusicDelta_LatinJazz'"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 287
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_fpath = \"%s/CC_exper_mel%s/splitnum_%s/olap_%s/melody_output/%s_pred.csv\" % (outdir, meltype, splitnum, olap, trackid)\n",
      "melodia_fpath = \"Melodia_Output_nu/nu_%s/%s_MIX_vamp_mtg-melodia-nu_melodianu_melody.csv\" % (nu, trackid)\n",
      "annot_fpath = \"/Datasets/MedleyDB/Annotations/Melody_Annotations/MELODY%s/%s_MELODY%s.csv\" % (meltype, trackid, meltype)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 288
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf_dat = pd.read_csv(clf_fpath, header=None, index_col=None)\n",
      "clf_dat.columns = ['time', 'f0']\n",
      "melodia_dat = pd.read_csv(melodia_fpath, header=None, index_col=None)\n",
      "melodia_dat.columns = ['time', 'f0']\n",
      "melodia_dat.loc[melodia_dat['f0'] < 0,'f0'] = 0\n",
      "annot = pd.read_csv(annot_fpath, header=None, index_col=None)\n",
      "annot.columns = ['time', 'f0']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 289
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ymin = 256\n",
      "ymax = 1024\n",
      "xlim = [48, 66]\n",
      "c1, c2 = sns.color_palette('deep', 2)\n",
      "\n",
      "ant = mp.Patch(color='k', label='Annotation: Melody %s' % meltype)\n",
      "clf = mp.Patch(color=c1, label='Classification Method')\n",
      "mda = mp.Patch(color=c2, label='Melodia')\n",
      "\n",
      "fig = plt.figure()\n",
      "\n",
      "ax1 = plt.subplot(2,1,1)\n",
      "plt.semilogy(annot['time'], annot['f0'], 'ok', alpha=1, basey=2, markersize=5)\n",
      "plt.semilogy(clf_dat['time'], clf_dat['f0'], '.', color=c1, alpha=1, basey=2, markersize=3)\n",
      "ax1.set_ylim(ymin=ymin, ymax=ymax)\n",
      "plt.xlim(xlim)\n",
      "locs,labels = plt.yticks()\n",
      "plt.yticks(locs, map(lambda x: \"%1.0f\" % x, locs))\n",
      "plt.ylabel('Frequency (Hz)')\n",
      "plt.tick_params(axis='x', which='both', bottom='off', top='off', labelbottom='off')\n",
      "\n",
      "ax2 = plt.subplot(2,1,2)\n",
      "plt.semilogy(annot['time'], annot['f0'], 'ok', alpha=1, basey=2, markersize=5)\n",
      "plt.semilogy(melodia_dat['time'], melodia_dat['f0'], '.', color=c2, alpha=1, basey=2, markersize=3)\n",
      "ax2.set_ylim(ymin=ymin, ymax=ymax)\n",
      "plt.xlim(xlim)\n",
      "locs,labels = plt.yticks()\n",
      "plt.yticks(locs, map(lambda x: \"%1.0f\" % x, locs))\n",
      "plt.xlabel('Time (sec)')\n",
      "plt.ylabel('Frequency (Hz)')\n",
      "\n",
      "fig.legend(handles=[ant, clf, mda], labels=['Melody %s Annotation' % meltype,'Proposed Method','Melodia'], \n",
      "           loc=(0.125,0.52), ncol=3, columnspacing=1)\n",
      "plt.savefig('/Users/rachelbittner/Desktop/mel%s_%s_comparison.eps' % (meltype, trackid), format='eps', bbox_inches='tight', pad_inches=0.0)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAERCAYAAABsNEDqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsnXdUlNfWh5+BAQlFBanWCMYYk9iiRk1Eg8YY1MTesGOJ\nUSwoCaKi8RJFsUTRaBRQg1iJDVu8atQbjRVLig17QVEwClbKfH/wzTgD05mBUc+zFmsxbznv7637\nnLPP3kcik8lkCAQCgUCgJ1YlLUAgEAgELxfCcAgEAoHAIIThEAgEAoFBCMMhEAgEAoMQhkMgEAgE\nBiEMh0AgEAgMQmqugrOzswkLC+PWrVs8f/6coUOH4uPjQ2hoKFZWVrz11ltMmjQJiUQCQF5eHoMH\nD6Zly5Z0796dJ0+eEBwcTGZmJjY2NkRFReHq6mouuQKBQCDQE7O1OJKSknBxcSEhIYGYmBimTJlC\nZGQkwcHBJCQkIJPJ2L17t2L7H374gczMTIUh2bhxIz4+PqxYsQJ/f39iY2PNJVUgEAgEBmA2w9G6\ndWtGjBgB5LcmpFIp//zzDw0aNADA19eXgwcPArBjxw6srKxo2rQp8nhEOzs7/v33XwBFq0MgEAgE\nJY/ZDIe9vT0ODg5kZWUxcuRIRo0aRV5enmK9g4MDmZmZnD9/nq1btzJy5EiUg9g//fRTjh8/Tps2\nbYiLi6NTp07mkioQCAQCAzCbjwMgNTWV4cOHExAQQNu2bYmKilKsy8rKonTp0mzatIk7d+7Qp08f\nbt68iY2NDRUqVGDnzp3079+frl27cu7cOYKCgti8ebM55QoEAoFAD8xmOO7du8eAAQOYNGkSjRo1\nAuCdd97hyJEjNGzYkP3799O4cWM+//xzxT7z58/Hzc2Npk2bsnHjRhwdHQFwcXEhKytL5zFlMpnC\nRyIQCAQC82A2w7Fo0SIyMzNZsGABCxYsAGD8+PF8//33ZGdn4+PjQ+vWrTXuHxwczMSJE0lISCA3\nN5fvv/9e5zElEgl372aa7BxMgZubk9CkJ5aoS2jSD6FJfyxRl5ubk0HbS1617LiWeEOEJv2wRF1C\nk34ITfpjiboMNRwiAFAgEAgEBiEMh0AgEAgMQhgOgUAgeEW5eDGF1q39KF/ehapVvRg16mtiY38i\nIyO9SOWadTiuQCAQCEqGixdTaNy4nuJ3Tk4OK1euAFYwe3YUv/9+BGdnF6PKFi0OgUAgeAWZOnWK\nxnV376Yxb94co8sWhkMgEAheQ9avX2f0vsJwCAQCwStIWFi41vVlypQ1umxhOAQCgeAVxMenGpGR\nMzSunzdvodFlC8MhEAgsnr17d1O7dg1q167B3r350zEkJx/Dz+9j/Pw+Jjn5WAkrtEwGDPiKpUtX\nqmQXd3PzYOfOfdSpU9focsWoKoFAYNHs3bubrl07KH4r/y+ndWs/ALy9fUhIWIePT7Vi02fptGnT\nlps3izb8tiB6tTgyMzP5+++/OXPmDJmZlhUqLxAIXm1Gjhym97aXLl2kSZMPuHz5khkVCbS2OPbt\n20dMTAwpKSl4enoilUpJTU3F29ubwMBAmjVrVlw6BQKBQC9kMhkREZPZvHlDSUt5ZdFoOEJDQylX\nrhzh4eG89dZbKuvOnz9PYmIiSUlJzJw50+wiBQLB68vcuQvUdk8JSg6N2XFv376Np6en1p1TU1Px\n8vIyizBjscSsk0KTfliiLqFJP8ytaejQgfzyy1qD9/Pw8GD58lXUq1ffDKqMw1LvnyFo9HHIjUZE\nRIRi7m85ISEhABZnNAQCwavJwoUxhISEGbzfnTt3aN3aj5MnT5hB1euLTuf4hg0b6NGjB2fPnlUs\nu3DhgllFCQQCQX6upQ9wdy+Nu3tpoqKmMmFCOGlpD1X+pk2L0llWcHBQMSh+fdBpOCpVqsS0adMI\nCgoiKSkJQEzPKhAIzIo8Qd/Fi6qV1IiIKcyfP6+EVAnk6DUct06dOqxYsYLly5czffp0pFIR/iEQ\nWCIZGenExv5kktTZJYm2BH3/+c9Eld8dOnTG3t5Ba3l///0nAwb0IjCwD/PmzX6pr40loNNwyH3n\nHh4eJCQkkJaWxp9//qn3AU6dOkXv3r0BuHr1Kj169CAgIIDJkycryl62bBldu3ala9euzJ8/X2X/\nixcvUr9+fZ4/f673MQWC15GMjHQaNqzDuHEhjBsXQsOGdbh/P6OkZZmcguN5XFzKceLE37Ro0VLr\nPlu2bCYpaSMREZNp2vTDV/LaFBc6DUdiYqLi/1KlSjFr1ixWrVqlV+FLlixhwoQJZGdnAzBt2jSC\ng4NJSEhAJpOxe/durl+/TlJSEmvWrGHt2rUcOHCAc+fOAZCVlcX06dMpVaqUMecmELxWhIV9w8OH\nDxS/Hz58QI0a3iQlvXzxDNoS9NWoUbPQMmdnF1atWk9a2kPOnbuis/y7d9NYsmRRUSS+1mjscxo3\nbpzWHevW1Z3npEqVKsyfP59vvvkGgH/++YcGDRoA4Ovry4EDB2jevDkxMTEKv0lOTg52dnbIZDLC\nw8MJDg7m66+/1vuEBILXlU2b1hdaJpPlERjYl4CAPbz33nt06NAZF5dyJaDOMHx8qnH48Em6dGnP\ntWtXVNYtX75S677Ozi5IJJJCLZOCzJwZycyZkQA4OjoSFxdP8+YtdGpLTj5GQEBX0tPvqSy3trZm\n9ux59OjRW2cZLzsaWxwNGjSgYcOGNGjQgEOHDin+ly/Xh1atWmFtba34rXwj7e3tyczMRCqV4uzs\njEwmY/r06dSsWVNhcJo1a0aNGjWKcHoCweuDtkErCQnLGTcuhBYtmr40XTRVq3pz7NhpDh8+Sbt2\n7WnXrj2HD5+kalVvnfuOHj3GoGNlZWXRtWsH9u/fp3W75ORjtG7tV8hoAOTm5jJy5DDWrFlt0LFf\nRjQGACrTvn17Nm7caNQBbty4wZgxY1izZg3NmjVj3778G7Nr1y7++OMPJk6cyLNnzwgLC8PR0ZHJ\nkycjkUho1aoVHh4eQL6fpHbt2sTHxxulQSB4HejVqxcJCQk6t/viiy+Ii4ujXDnLb3kUhT59+hj8\nzahQoQI3btzQuL5u3bqcPHkS/1GFW3fbfugIgJ2dHU+ePDFM7EtGsQ6Peueddzhy5AgNGzZk//79\nNG7cGJlMxtdff02jRo0YNGiQYtudO3cq/vfz8yM2NlavY1hiRKbQpB+WqOtl0jRxYgTr1q3TOZBk\n8+bNeHt7c/ToaaPnnNZXU0ny888/8+WXXenTpzuPHz/Sa5+8PJnW88jJydOrHG1lWOK1MjRyvFgM\nh7wJHRoaysSJE8nOzsbHx4fPPvuMXbt2cfToUbKzs9m/fz8AY8aMoU6dOoX2FwgEhcnISGfIkED2\n7dujcZuq9dpRruK7pN/4m8vJSTx8+JCQkGBiYpYVn9ASwNe3GVeupAIQGTmF2bO159ZLTb2Fu3tp\nQDVF+969u+nbt6eiJSFvXagjKkq/OBN1vhIPDw/c3DyYOfMHi0qTUhCNXVXKw2JXr15N9+7dVdYP\nHz7cvMqMxBItudCkH5aoy9I1ZWSkU6/eezpr1FXrteMd3/6K3/IP3yef+LFwYWyRHeaWfp3kREVF\nEhU1Ve8yJBIJ8+YtIChI9wAda2spP/zwI926dde6nZubE7/++ptiDhFNFHWyJUMwWYtDJpMpavrd\nunUrmiqByRkQ+aJ2GReq/QEsCXYeucba31LIk4HUWsLikE9KWpIKE2MOc/Pei4+tpV3Dr2fv4+nz\nXJ3XbsOGRL26YS4nJ6kYDjm//baHOnXeoXlzP2xsbAkLC3+lJ0EKCQklJCRU8bt27Rqkpt7SuL1M\nJiMkJFhrme3atSc29meDdIwdOwqAloPjsLV/Mfe3cktm4MA+HDumf8xccaLRcAwaNAg7OzutOz97\n9kzEWJQAykZD+belffzy/r8tm5MrY0DkHso42DIn6OOSFfX/KBsNsNxrmJMrY3DUb0YZXolEwp07\nDxRdNAW7V+QO3m0/dGTHjm0AbN26mT/+SNZr5JKg6CgbDci/J2mXjnBscyQ3blwvIVW60TgcNyQk\nhLVr15KVlVVoXVZWFgkJCYwePdqs4gQvL60aVi5pCXphZ2uteyMDuHgxhcDAPgQG9uHixRSjy/kx\nWPckaRcvphATs1jj+rFj82OxQkPDCQ+P0Ou4eXl5itqwJpKTj9GgQS28vJypWtWLUaOGMX369Jcy\nzcncuQu0rpdIJERFzda6fsKEyQAkJW2gWrVKVKtWSWfQ5cyZP6gukBV2ur/xxhsa9zfVc2YsGn0c\nubm5rFq1ihUrVuDk5ISnpyfW1tbcunWL+/fv06dPH3r06GFxeatehn7WksYSNYFmXfPmzSQiIj93\nUY0a77B0aUKxdacYcq3kifmUKVvWmdWrf+FejjtgGoPq5ubEoUMnCh1LmZCQMJUuGYCYmIWEhX2r\n+P3ZsASsbV58nJRbJEOHDuO776YVKlcex6CLCRPCGTFirM7t1HHxYooiV1VRus70vXf79++jZ8/O\nPH/+zKDyq1evQXz8aqpWzY/ODwzsq3a7goGBcl0nT56ge/eOWo2tra0tbdt+wfr1LzJ4WFtLyc3N\n0VunPn4sQ30cOuM4ZDIZZ8+e5cqVK1hbW1O5cmXefvttix3ppOlBGTg9vysi5lvDuyJ2HrkGGPfS\n6/PwRiee5tTFe9jaWPNjcDOiE08DENS5lsHHM5UmeHHeckzx0Rsc9RuA2q4Xdbo0jYTRNxCsqBhi\nOJo3b8I///xF1XrtgHy/QtV67ajeuDs2pRyws7XWqyWhj6amTZvx++/71a5/6623OXDgqNp1cXGx\nhIa+6ClQjkco2JU1fPhowsO/U1n20Uf1uXDhfKFylc/Zt3d+bbp51SyGDx+hxxm94OLFFD7+uAG5\nublA/kf34MHjRt1rQytI+/fvIyhoCA8fPuDRI81+o8jIOQwYEKiyrEoVT548eay1/OjoxXTr1r2Q\nrvv3Mxg0qK9K8GH9L/KN/rHNkXrr14R8YISVlZXGrliTG46XDXUPys4j11i9R7U5Z0hf9uCo38jJ\nVX+ZdJWj6+EdHf07Dx6pH3dvrv52bZqUjVZBX4ocYz6Amq6hBIj9//MsqCsjI50aNaqqLe+zz/yJ\njzd/hK4hH59GrXtRw7efxvVWVqo9w8rnbqimMmXKquSlUqZ06TKkpGjvHx86NJBfflmn+F2w9aHM\niNZlFUNDq1WrpPa4vr1/wLGcasVix7zO3L79b6FtQXNlLDCwD7k+fRS/t/3Q0SjnMxjfsvb0LEte\nXp5Gx7VUKuXWLdXoew+PMjTtNafQNVA2xnZ2dly7lqZRV+XK7jx9+rRQcKG2ob/qKKhbTv7jl/8M\nFvQ3mmwGQMELpNbFf5kssz1XvGzYkKhx3c2bmqN7SwqJVfHdNQcHe43ratfWPYRz4cJYhg/Xz0ep\nPINe7dq11W7z+MFtvcp6VSlTpvCHuiA5OTnExv5EevrL5QdSx2vR4ihJLNGfYImaoLCu2NifGDcu\nRO22Y8eG8s03qlOJxsQsZPz4UGQyGV5eXixdmlDkICpDrpU6HwLArwsCANPETGRkpNOlyxcapzaw\nsrIyaFSUOr+MHOXuJzlDhw5j4ULtDmU54eERL1VXlRxln5o61HVV7d27m65dOxh8LHgRaHjkyB+M\nHDnMqDJ04eRUmmPHNGcKMHmLY9CgQWzfvl2RGl0gKC46dOiMp2fhee1tbW3p3Fk1tkj+0ZbXg1JT\nU4t9rumBA4cSGTlH8fvXBQEKowH5MRPvv1/d6CSD8mA/TUajceOPDB5KK89CW61a9cLr6nfAp77q\nx3DhwgUMHz5apVUjldrg7++vsp0xRkOu5+DB44qkhtqMxs4j1wr54UzBiBFj1Y5Ck0is1BoNgObN\nW5CYmKRhQqn8lmjVeu0UxliZS5cu0qTJBzRq9BHR0aqj5GxtS9GlS88Cy2zp1atwTI4mWrRopdVo\nGIPOFseRI0fYsGEDhw8fplmzZnTo0IFatczjtDUFd+9msmpVPGPGjCQnJ4cyZcqyZMlStemSAyP3\noO7kTelbsMTavSGakpOPMXBgX41jyqVSKa2Gr82vgRTowzf0Osp1FfKt5OWRB5zZv1RR+3V2dubQ\noROKl6F8eRdycgqPNClfvgInT54xSIc6TepYtSq+UA1RIpHw7rvv8ddfmgO3Pv/cn+XLDffPaGuB\ngWogmib/lLZ70qlTW/73vxcO989HrEFiZaN2W7m/Jq6AfyojI50NGxI5evQImzdvIDc3Fw8PTzp2\n7EylSpVNmtZdfo5WEvWDXizl3fPz+5jKLSdrXC9/ro315ZgCk7c4GjZsyLRp09i2bRu1a9cmKCiI\nNm3asGzZMouclW/SpHGMHDlM8RF58OBfvdIlCwojH3qpLRBJfp3zyI8BUMZUtUF5qeUqvqtYdv/+\nfdavX6d+ByW0RQUXBXVGA/JHIWozGgA7dmw3i6aiMmOGamxB9lP9EgPKychIp0WLpowbF8L69evI\nyclBJpNx+3YqP/4YzbhxIdSoUZUGDWoVKfZgcNRvKoYxT5ZvRDQZy5KmQoWKWtdXelf3HCCWhl4+\njkOHDrFp0yYOHjyIr68v/v7+HDhwgDNnzuidtbY4+O9//0urVq2AF8MMc7Of8OuCALy8ynPq1Nli\n12QptR5l9NXk5/cxf/112qCymzZtzpBv57P98DUc37DhPwM/LLKuypXd8ar5KaDa3z5tWhSBgUOA\nwv4FOba2tty4UXjuBFNoevr0qVFlWllZaRxtpA1teakM9W1o4vLlS3z99SCOHz+q1scB6ruh3Nyc\niIycqbVFpIxEIuHQoRNG6dVmIJRbVJby7sn9SOqG2MqXHU+abvT1MAUmz477ySefULFiRTp16sSk\nSZMUaUgaNmxIp06djFNpJvr3z+/3azk4TrFM0xBDgXk4deoErRpWZvWeFI3DjA1l+vRZhWr3zs7O\ndOzYRfF74MChHD+ezC+/rFHZbtYs1TnsiwsrKyu8vauRklI45mHCBM2OV23I59YeNWoo27fnt1qs\nraU0bPghP/ywwCQfnapVvdm+fTegLiGghPDw/xjluyiITCYjImKyUV0zZRxsASwmfY0u5H6k7t07\ncfnyRZV1xzZH4uPzVokaDWPQ2VW1bNky5syZQ/v27ZHJZFy9ehXIH+1g7OROxUt+g0rbHMavO6tW\nxVOxoiuenmX59NNmREZOoXx5F/7+2/AEa3l5uQQG9tG9oQH06NGb6OjFlCpVCqlUSufO3VX8G3IW\nLlxCdPRi7OzssLOzUwRcmYPp02dpXT916iwOHjxGWtpDwsMjsLKywsrKyminsRxnZxe2bdtGWtpD\n0tIekpqawaZN283y0QkJCVUcJ//vgVbtHTp0NrkGdcwJ+vilMRpJSRsoX74cH35Yh8uXL+Lo6MjO\nnTtVrusffxg3aqwk0dlV9fPPP7N+/Xo2btzIjRs3GDhwIP369SuUZt0SUO6qAvAfmQgSCWf2L+Ny\nchIDBgSqjHopDiyluayMsiZNffVFxZiuE0u/VgVZs2Y1QUGDVZZJJBKmTZutduRNUZEHZ04Z+pHF\nXqdJk8bpPVwXoFw5VxIS1ppl7omSfp60pSFJTEzC17foWQRMhckjx9u0acO6deuwt88POHry5Ald\nunRhy5Ytxqs0I7/8soUhQ/qTnn5PZVRIVvo19sePUpvDxxg0RXwXHLXi5ubEF2M2qR29JcdUqSj0\nRfmFqlChXKGh1ur6tl1dXbGxsSU6+id8fZvpld+nadPm/PLLZqN0FUQfx6c5Iu1L+uOjTHTiaU6k\nvPDXWFImX+XrpNzF5ejoRL9+A5k/X3uFzRxzT5T0vdMUZQ+UmM9VEyYfVZWTk4ONzYsheTY2Nhab\npwryZ/w6c+YSUqmU7fNejPWXpwKYObNw4jZzY8kRlgWNxucj1vCOb/9Cczd4euY/6PJakq9vM51z\nVp86VXwxFK8D5spdZmqUu7guXbpJePh3pKU9LJRyRZng4KBiVCgoKjqd4y1btqRv3774+/sjk8nY\nuXMnfn5Fq+l06NABR0dHACpWrMi0afkf86lTp+Lt7a3oBlu2bBnbtuXPE+Dr66tz1kF1Bm3bDx3V\nTixfVOR9rIH/XxPWlnNIrkq+jTx+pG411xL7GMinrTQnHh6eJiuru181/nc6ldT0R4pkkHLkXTjF\nQcG4FhsbG2bO/EGR+dTcxIX6KRJ2CiybOXPmaeyqio7+qZjVmBa9suPu2LGDY8eOIZVKadCgAS1b\ntjT6gM+ePaN79+5s2PAiX31GRgbffPMNV69eZeDAgXTr1o3r168zatQoEhMTkUgk9OjRg8mTJ/P2\n229rPhk9W0JvvVWdn39eXSypuUu6uayOy5fP8OGH+g+TBfVdCbrSLBja/WCJ10pZk7aU4uZ0xGvT\nZCnoo0lbKg/lZyU5+ZhiThBD5t4uOCumJVynrVu3MHhwX0XL3tHRiU2bNvL++w1KVFdBTN5VJZFI\n8PHxoXXr1rRo0YLSpUtz9Kj6lM36cPbsWZ48eUJgYCB9+/bl1KlTPH78mBEjRvDll18qUkZ4eXkR\nGxurMAY5OTk6ZyTUlwsXztOkyQdcvnzJJOW9bOg7FbC1tRQfn2oaDYCmNAtubh4m6bOOjJyCu3tp\n3N1L4+XlTGTkFGrXrqFY5uFRhnnzCqdcNxejRmlu8Y4dK7padKEulYerq1sho9G6tR9//XWav/46\nXexpY0xNmzZtuXkzXaXrrqg9NpaAzhbHd999x2+//UalSpVUlsfHxxt1wPPnz3Pq1Cm6dOnClStX\nGDRoEL/++itWVlbMnz8fV1dXlRFbMpmMGTNm8PjxY7777jstJatvcZQuXUajgwogOHgsoaHmG6rr\n5ubEmDEhijklCk7qUpxcvJjCl19+TlraHY3byKcbLQmUa4j9+wewdWuSjj3yKeoQV301aZufWl2q\nbXNhCTXpgphKk69vI86e/Udl2Tvv1GTfvkNa91NOH1QwDYqlYYm6DG1xINPBp59+Knvy5ImuzfTm\n2bNnsqdPnyp+d+7cWXb79m2ZTCaTRUdHy1atWqVY9/TpU1lwcLAsPDxclpeXp7Ns8v3QKn8tWrRQ\nu1z577vvvjPZ+RVk4sSJao/5888/m+2Y6jh//rzO6wDIXFxcilWXOhITE/XSqvxnZ2cnS0xMNKuu\nMWPGaDx+r169zHrs14UKFSoUurYVKlTQuV+74I2ytsEbZZ1Dk4pBpUBniyMwMJDo6GjFcNyisnr1\nas6dO8ekSZO4c+cO/fr1Y+vWrYVaHDKZjIEDB9KoUSMGDRqkV9nKLQ5nZ2fWrNmIk5OT1ik25Tg6\nOjJ37gLatTMuNbImPDzKoO4SF2cNFfInyElK0h2wWZLjy+U1MW3DGHWxdOlK2rRpa3JNkJ/yo379\n98nKylLZxsHBgeTkv02afVRfTZaCqTRNmjSehQujVZYNGzaCSZP0mzPdHJpMjSXqMrmPo3Tp0rRp\n04bg4GDGjRun+DOWzp07k5WVRUBAAMHBwUybNk1lmJ78479r1y6OHj3K//73P3r37k3v3r05efKk\n1rLt7OwUffLnzl2lTp26WtNGK5OVlUVgYF+2bjVdfMrevbvVGg1AbSZXc6It3kIikeDlVd7igpKM\noX//nkRGGpfSQxsZGekEBQ1VMRo2NjYEBPQtVqPxqjNyZDBubu6K325u7owYEVyCigTq0NniWL8+\nfyir/IMuk8mQSCR06GDamrkpkEgkWid+qVWrOrdva5+pTJ9pN/VBU9I9OWXKlGXNmvVGjR4xFF0T\n01iKwZDXxLRF3BqLsb4sNzcnzp27wrvvvkVubmFjb47ANX00WWKN1VhNBeeVt7GxoUuXbrz/fm06\nduxitFG2xOsElqnLLHOOX79+nZSUFD7++GNu375dyFFuKciNm6a89tqGUypja2vLwoVLjO620uc4\nY8Z8y6xZ0wss+4Zvv51g1DE1oWzA/EfmT8W6bW5+TiF7e3t+/nmNRRiNiTGHsba2YnL//GGKW7du\nYcSIIWRnZ9OpUzfCw7/j6tWrfPVVIJcuGZeS25isAW5uTnze41usPBupLJfPA12jRk3279fuuDU1\nRfnwKM8pb0qC5x9AJpMZnEPqq68GsH69+imCi9rtaMh1kg/l7e6XP0R/9Z78Z0xqLWFxyCeK7eQZ\nIwouN5eu4sLkhmPr1q0sWrSIJ0+esGrVKjp06MDYsWNp3759kYSaA12GA+DkyRP07dtD73kajIn5\n0JaO3MurPGFh4QQFfaV2vanjAcqXd6FJj5mKyHl48dEryYljoHAKDdAvjUbBGqohODo6EhcXr3Zi\nL3UMiNxTaJ4ReHENzZ06omCqle5+1Qho867RHx515bVqWFnD1oaXp+7+KW9TxsFWYVx0xQEVtfVv\njOEoSMFJovQ5X1PqKi5M7uNYsmQJq1atwtHRETc3N9avX8/ixYt17VZiWFtbM2HCZI3r69Spy4gR\no/Uu78KF8zRuXE8RSxATs9BobVKplFOnzmqsYQGEhJhmWOnFiym0bu1XyJeSm/3EJOWbAys9M9mE\nhoYTEhKme0M1ZGVlFXliL7nRAFRSu78MWFKyIHPNr21Kun5i/iDhlxGdLY6OHTuyfv162rdvr0ij\n3q5dO5KS9BtjX5x06dKFkJAJOjOyZmSk06hRPf79975Rx9E07zDkf7B79OjMlSuFgwvlrYkaNaqS\nkZGudn87OzuuXUszSpeyBl0jyYoykY65MLQmpm/Xoyb08Xu4uTkxceIktT4iV1dXDhw4VuyOcUut\nsRqqSVtcDBRvV1VxYom6TN7ieOutt4iPjyc7O5szZ84wceJEatSoYbRAc7Ju3Tq9PoQuLuU4fPgE\nvXoZN2/E+PFj1S6Xf7ALGg1ra2uioxdTv359atV6W6PRAIiKmqd2eUZGOiNHfk3lyu74+FQgKWmD\n2u0Apk7V7AivWq8dH7Udwg/L93Lhrs5UZRZNvXr12blzH+XLV0AikWBra0t09GKWLl2JVKr73GbP\nnklUVKTO7eQRz/Ku0HfeeZeJE79j6sJtHL2QxcSYw0yMOVzk83ndmDt3AVXrtaPl4Dh8e/9A/S9C\nqf9FKLblzqrxAAAgAElEQVS2tiYfVl0cFPf0tTuPXDPZ9MyGorPF8ejRIxYuXMjBgweRyWQ0atSI\nYcOGKZIUWhqGWnJjaq2aoqt9fT/k7NkzhZZLpVL+978jWlsBUqmUOXN+VOvf0NRC0vRyaYvZ8B+1\nXmuWUsjvt5W/AMWZutvUNTFNMTQF0TYnREFNO49cUzhONWHua2apNVZDNKnMGf7/PiQrgP9/Novi\nfDZWU1EomCdLeZm6qRZMoUubkTI0garJWxwODg6MHTuW9evXs2HDBr799luLNRrGIK+1vvee/he5\nZs33Ci3LyEhXazTkaGsFODo6cutWhkan+IYNiWq71UaOHKp2ezHbYT7jx0/Ua7v09HsvfU6klxn5\n7IjoqNC8DMintX3V0dmeV9ct5e7uzv79+80iqCSoU6cue/b8rjP2Qk5cXOE8XRs2aHZ4R0REceCA\nZmfsRx81pUoVD548yXdcF6wB79y5XacmObpGHI3yL8enn/py926mopmrblSNJU0SZCwjRozFysqW\nKVP0G+LcqlUzlVxie/fupl+/AB4/fqzYZujQYcR9lz8NgLbrJ9DOq/B8KSO1zu/GVB6ObO5zlJdf\nEs+hXnEccrKzs9m1axcnTpwgLMy4US3mpqhNwLi4WMaPH0tubm6hdS4u5di+fbdaP0ps7E+MGxdS\naHm3bt2Ijl6i0WFtY2NLdnbhmQThRXCZu3tptesLdlVpMxoeHp7Ex6+hTp26FtnVAebtWrh8+RIB\nAV1JSTmv1/Zffz2SH3+cq3bd8OGjCQ/XnnDTnFji/ROa9McSdZm8q0oZGxsbPv/8cw4dKt6Ap+Jk\nwIBAUlPvc+jQCZV04eXKuWqdVL5Dh85UqFBRZVl4eASrV68GUJv6pEWLVri4aB6RM2xY/nzWmuYZ\nUTYae/fu1trSsLKyKvYIZ0uialVvDh48xo4d+jkvNRkNQOc0qIJXg8DIPYqJ2gSq6OyqUp5wSSaT\nceHCBWxtX/1+PG9vH06c+Jv169cB6Ex94OJSjj17fte6vfzjpUzt2ppHqN25k58epX79hhw9qjpq\nx97enr17d9O8eQudgVSCF8h9Wt27d9Q6uk0gEGhGZ1dVaGhooayzPXr0sNi0I5bYBNSmSdtHv2nT\n5jRoUF9rSyIxMYl+/XoUytiqbjt5ahFLbCpD8evSFUegCdFVVRihSX8sUZdZclW9TFjiDdGlaf/+\nfQQEdOHZs6eKZVZWVvzxRzKNGtXVOqS0TJmyPHjwr8b1Dg6OLF++SiUflSU+uFD8uowZil3SRgMs\n8/4JTfpjiboMNRw6u6r8/PyQSCRqP14SiYTdu3cbdEBBYXx9m3H9ehqXL18iImIyABMmTNYrmFHb\nvBVubu78/bdxCQFfB+TdVsHB+dO+zp4dTefOX2i8pi9jUJpAYA50Go527dphb29Pt27dkEqlbNmy\nhePHj/Ptt9/qFVwl0J+qVb0LJR0cPXqM1q4qbfdg4cJYk2l7VZEPxdYHYTQEgnx0jqrav38/Q4YM\noWzZsjg6OtK9e3cuXryIq6srbm5uxaHxtePixRQCA/sQGNiHLl16EhISpvAzlS3rrFcZS5eutIh0\n6foyMeYwHb7ZXOKpO+bMUZ/yJTrachN7CgTFjV7DcX///UWNbNeuXTg4OGjZWlAUkpI20LhxPZKS\nNpKUtJHGjetRs+Y73LnzgNjY5XolZpw48buXrnZ8894jcnJl3Lz3qFjz/RSkXbsOLF26Ejs7OyB/\nCLqpU90LBC87OruqIiIiCAkJIT09HZlMhre3NzNmzDD6gNnZ2YSFhXHr1i2eP3/O0KFD8fT0ZMiQ\nIbz55psA9OzZk88//5x9+/bx448/AvD+++8zYYJpJzmyNJKTj6md+a5//97s3LmP0aP1S7luqvnh\nixOptYScXMvo+mzTpq0iQ7ElOjIFgpJGp+F499132bZtGxkZGdja2hY5T1VSUhIuLi5ERUXx4MED\nvvzyS4YNG8aAAQPo37+/YrusrCxmzpxJfHw8ZcuWZfHixWRkZGgNmHvZGTVquMZ1cgeuLkqXLvPS\nzREBsDjkE/GRFgheEnR2Vd24cYP+/fvTrVs3Hj16RO/evbl+3fhZuVq3bs2IEfk157y8PKRSKX//\n/Td79+6lV69ejB8/nkePHnHixAmqV69OZGQkAQEBuLu7v9JGA9A6rBY097/LqV69BkePnir2+SEE\nAsHrhU7DMWnSJAYMGICDgwNubm588cUXhIYaNm+zMvb29jg4OJCVlcXIkSMZPXo0tWrV4ttvv2XF\nihVUqlSJ+fPn8++//3L48GFCQkJYsmQJy5cv58qVK0Yf92WgfftOGtfNnh1Nu3YdiI7+qdA6iURC\neHgEv/9+RBgNgUBgdnQajvv379O0adP8ja2s6NKlC5mZRetOSE1NpW/fvrRv3542bdrw6aefUrNm\nTQA+/fRTzpw5Q9myZXnvvfcoV64c9vb21K9fnzNnNKctfxUYOTIYNzd3lWUeHp6KZIcA3br14Ny5\nK0ybFsW0aVGcO3eFO3ceMHy4aaacFQgEAl3o9HHY2dlx+/Ztxe9jx45RqlQpow947949BgwYwKRJ\nk2jUqBEAAwcOZPz48dSqVYuDBw/y3nvv8e6773LhwgXu37+Pk5MTp06dolu3bjrLNzQCsjjQV5Ob\nmxNnz55h5cqVQP4gAXXdc25uToSGqp+F0NSaihtL1CU06YfQpD+WqktfdKYcOX36NBMmTOD69etU\nqlSJBw8eMHfuXOrUqWPUASMiItixYwdVq1ZVLBs7diyRkZFIpVLc3d2ZMmUKDg4ObNu2jZiYGAD8\n/f0ZOHCgzvItzblqiQ5fS9QElqlLaNIPoUl/LFGXyXNV7d27lyZNmnDlyhXy8vLw9va26Oy4lnhD\nhCb9sERdQpN+CE36Y4m6TD4fx4wZM7C1taV69erUqFHDoo2GQCB4Ndl55JpipjtByaPTx1G5cmXG\njRtH7dq1Fb4NiURC+/btzS5OIBAIAFbvyU/WKabptQx0Go6yZcsCcOrUKZXlwnAIBILiZueRa8J4\nWAAaDcedO3fw8PAgMjKyOPUIBAJBISq4OnDn/mP+dzpVGA4LQKOPY8iQIYr/4+LiikWMQCAQqMO9\n7BtIrfXKySooBvS6E5s3bza3DoFAINBIUOdaPH2eq8ieHFiCGZQFehoOgUAgEAjkCMMhEAheCrr7\nVVP8Hxtq2FzxAtOi0TmekpKCn1/+zUlLS1P8D2KucYFAUPy0alhZOMYtBI2GY8eOHcWpQyAQCAQv\nCRoNR8WKFYtTh0AgEAheEoSPQyAQCAQGIQyHQCAQCAxCGA6BQCAQGIQwHAKBQCAwCGE4BAKBQGAQ\nwnAIBAKBwCCE4RAIBAKBQZSY4Th16hS9e/cG4MyZMwQEBNC7d28CAwNJT08HYOXKlXTq1InOnTuz\na9eukpIqEAgEAiV0TuRkDpYsWcLmzZtxcHAAYOrUqUycOJEaNWqwZs0alixZwogRI4iNjeXXX3/l\n8ePHtG/fnpYtW5aEXIFAIBAoUSItjipVqjB//nxkMhkAs2fPpkaNGgDk5ORQqlQpJBIJAI8fP+bR\no0dYWYleNYFAILAESqTF0apVK27cuKH47ebmBkBycjIJCQkkJCTwxhtv0LZtW/z9/cnLy1OZWEog\nEAgEJUeJGA51bNu2jUWLFrF48WKcnZ1JTk7mxIkT7NmTP2FLYGAgdevWpVatWlrLcXNzKg65BiE0\n6Y8l6hKa9ENo0h9L1aUvFmE4Nm3axNq1a4mPj6dMmTIAPHnyBDs7O2xtbQFwcnIiKyurJGUKBAKB\ngBI2HBKJhLy8PKZOnUr58uUZPnw4AB9++CHDhw/nwIEDdOnSBWtraz744AOaNGlSknIFAoFAAEhk\ncg+1QCAQCAR6YBFdVabm+fPnXLlyxaRlvvnmm4puM4Flk52TR9r9xyYt093ZHhupGNlXUuTk5pD2\nON2kZbrbl0Nq/Up+As3OK3nVrly5wttvv23SMs+dO0f16tU1rj98+DB9+/Zl9uzZ+Pv7K5a3a9eO\n9957j2nTpqndLzo6Gjc3N7p3765Tw59//snKlSs1lqXM9OnTOXHiBDk5OXTr1o0uXboU2mbJkiUs\nX76cPXv2mNQoHjt2DCcnJ4334Pnz52zatIkuXbqwYcMGypQpozI1cVFJu/+YryJNO7XxotAWVHBz\n1Lj+8OHDjBo1imrVqiGRSHj27Bnt2rWjV69eJtVRFKZMmULr1q1p2LChYll0dDQLFy5k7969uLu7\nA5Ceno6vry8RERF06NBBbVnnz5/n4cOH1K9fHz8/P3bs2GHwM7Rjxw5SUlIUXdTaSHuczqhtkw0q\nXxc/+E+mvJOHxvXGvNPGvs9BQUFER0cbdyIlgKhCmRBvb2+2bt2q+H3u3DmePn2qdR95vIopOXTo\nEDdu3GD16tWsXLmSJUuWkJmZWWi7zZs307ZtWxXNpiAxMZG0tDSN69PS0khMTASgQ4cOJjUaJYVE\nIqFJkybEx8fz888/Ex8fT1xcnEUN6ND0rL355pts375d8Xvbtm2UL19e67P566+/kpKSYnKNloah\n77Sx7/PLZDTgFW1xlAQSiYQaNWpw5coVsrKycHR0ZPPmzbRr147U1FQAtm/fzvLly7GysuKDDz5g\nzJgxKmVERkaSnJwMQNu2benTpw+XLl0iLCwMOzs7ypYtyxtvvMGBAwdYu3Ytc+fOBaB79+6Kmg5A\nvXr1qFmzpqLc3NxcpFLVW3348GHefPNNunXrRkhIiKJm2bt3b9555x0uXLhAVlYWc+fOJS8vjzFj\nxuDl5cW1a9eoVasWkydP5uHDh4SEhPDo0SNycnIYNWoUTk5O/P7775w5c4Zq1aqxe/du/vvf//Lk\nyROcnZ2ZP38+ixYtIiUlhQULFiCTyXB1daV79+5qzz80NBRbW1tu3rzJ3bt3iYyMVDk3S0Emk6Hs\nLszKykIqlWJlZUXv3r0pV64cDx8+ZNGiRYSFhXHjxg3y8vLo168f/v7+iut+5swZJBIJc+bMoVy5\ncmqvyc6dO4mJiUEqleLu7s6cOXPIyspi/Pjx/PvvvwBMmDCB6tWrs2rVKtauXYuLiwtPnjzhs88+\nU9EtkUjw9/dn+/bt9O3bF4C9e/fyySefKM5n1qxZHD9+XKG3bt26bNiwAVtbW8W9mDRpkiI2a8GC\nBbzxxhuMGzeu0HkmJyczdepUSpcuja2tLe+99555b0wR0PVOm+p9Bvjoo484cOAAR44cYcGCBeTl\n5fH48WNmzZrFm2++WdynrhPR4jAxrVq1YufOnUB+U7Ru3boAPHjwgPnz57N8+XJWrlzJnTt3OHjw\noGK/3377jZs3b7J27VpWrlzJli1bOH/+PDNmzGDkyJEsW7ZMMarso48+UnQVXLhwARcXF4XRALC1\ntaV06dJkZ2cTGhpKt27dFA+onHXr1tG5c2eqVq2Kra0tp0+fVqyrXbs2S5cupUmTJmzZsgWJRMKV\nK1eYOnUqiYmJ7N+/n3v37rFw4UI+/vhjVqxYwdy5cxk/fjzvvvsuTZs2JSQkBE9PT/7991+WLVvG\n2rVrycnJ4c8//2To0KFUq1aNYcOG6Tx/iURCxYoViY2NpXfv3qxZs8b0N81EHDp0iN69e9O3b19C\nQkKYMGEC9vb2QP6HIy4ujrVr1+Lq6srq1atZunQpc+fO5f79+wCKFkurVq0U3UfqrsnWrVsZOHAg\nK1eupHnz5mRlZbFo0SIaN27Mzz//zJQpU5g8eTIZGRksX76ctWvXsmTJEiQSidoasaurK/b29ly/\nfp2rV6/i6elJqVKlANi3bx83b95k5cqVLF++nEWLFmFvb0/Hjh3p37+/Iq6qS5cuxMfHU7FiRQ4c\nOMCaNWvUnufkyZOZOXMmcXFxJu9ONhfq3un79++b7H2GFy2VlJQUoqKiFM/Bjh07ivdk9US0OEyE\nvHbWpk0bJk+eTKVKlahfv75i/dWrV8nIyGDgwIEAPHr0iGvXrinWX7p0iQ8++AAAqVRK7dq1SUlJ\n4fLly7z//vsA1K9fnxMnTgDwxRdfsGXLFq5fv67Wf/HgwQNGjhzJhx9+yODBgwut+9///sf9+/eJ\nj48nMzOTFStWMGPGDADeeecdALy8vLh37x6QnyZG/hF0c3Pj2bNnXLp0iS+//BIADw8PHB0dFQkq\nIf9lsLGxITg4GHt7e+7cuUNubi7qBvJpOn9lPR4eHooanCXSqFEjZs+erXZd1apVgfzzlH8wHBwc\n8PHx4fr16wCK5R988AF79+7Fy8tL7TUZN24cP/30E/Hx8Xh7e9OyZUvOnz/P4cOH2bZtGwAPHz7k\n6tWr+Pj4YGNjA+S3RDUNomzTpg1bt24lJyeHdu3aceDAAQAuXLjA33//rUhImpuby82bNwvtL285\nuLq68vTpU43nee/ePUUNun79+pw6dUrfy1vsaHun8/Lyivw+f/DBB4r3WY67uzsRERE4ODhw584d\n6tWrZ/bzNAbR4jAxlSpV4smTJ8THx/Pll18qHr6KFSvi5eXFsmXLiI+Pp2fPntSpU0exn4+Pj+Kj\nmJ2dzYkTJ3jzzTepVq0ax48fB+DkyZOK7Tt27Mj27ds5fvw4zZo1U9Hw9OlT+vXrR+fOnRk6dGgh\njZs3b6Zz587ExsYSExPDunXrOHDgABkZGYD6flp1y7y9vTl69CgAd+7c4eHDh5QtWxaJREJubi5n\nz55l9+7dzJkzhwkTJpCXl4dMJsPKyoq8vDyVsjSd/6uCPNeaj48Px44dA/K7s86fP0/FihUBFK2+\n5ORkqlevrvGarFmzhqCgIOLj45HJZOzatQsfHx/69etHfHw8M2fOpH379lSpUoWUlBSePn2KTCbj\n9OnTGvvgP/vsM3bv3k1ycjIffvihYrm3tzcffvgh8fHxLF26lM8++4xKlSop7rEmNJ2nh4cHFy5c\nAFSfZ0um4DsN+e+Dl5cXS5cuNfp9Vmc0w8PDiYyMZNq0abi7u2s09CWNaHGYCOVuAH9/fzZv3kyV\nKlUUtRAXFxf69etHQEAAeXl5VKxYkbZt2yr2bd68OYcPH6Z79+48f/4cf39/atasybhx4xg3bhxx\ncXF4enoqPkDyGn69evUKJYBcvXo1N27cYM2aNYqunWnTpik+UImJiURFRSm2t7Ozo1WrVqxbt07j\nh6XgcolEwldffUVYWBi//vorT58+5T//+Q/W1tbUrl2b2bNnM2vWLN544w0CAgJwdnamZs2apKWl\nUadOHbKzs5k5cyZ2dnZaz1/52OYYSGAqNHUDFaRr165MnDiRnj178vTpU4YPH46LiwsACQkJzJ07\nFwcHB6KionByclJ7Te7cucOQIUNwcHDAwcGBTz75hObNmzN+/HjWrFnDo0ePCAoKwsXFhaFDh9Kz\nZ0/KlClTyM+lrN3R0REvLy8qV66scr39/Pw4cuQIAQEBPH78mE8//RQHBwfee+89ZsyYgY+Pj9pn\nQ9N5RkREMH78eOzt7SlbtizVqlUr4pU3H9reaRcXF9q0aUOvXr1M8j7L+eKLLwgICMDd3R1vb2/u\n3r1b7OetD69kAODrEscxdOhQwsLCqFSpUklLsShexjiO3r17Ex0dTdmyZc12jJcZEcdhWbySV83W\n1lZrzMXLztOnT+nZsyeNGzcWRkMNNlIrrTEXgpcPqbVUa8yFoHh5JVscAoFAIDAfwjkuEAgEAoMw\nW1dVdnY2YWFh3Lp1i+fPnzN06FB8fHwIDQ3FysqKt956i0mTJimcT3l5eQwePJiWLVvSvXt3njx5\nQnBwMJmZmdjY2BAVFYWrq6u55AoEAoFAT8zW4khKSsLFxYWEhARiYmKYMmUKkZGRBAcHk5CQgEwm\nY/fuF/mEfvjhBzIzMxWGZOPGjfj4+LBixQr8/f2JjY01l1SBQCAQGIDZDEfr1q0ZMWIEkN+akEql\n/PPPPzRo0AAAX19fRaTljh07sLKyomnTpopxy3Z2dor0CfJWh0AgEAhKHrMZDnt7exwcHMjKymLk\nyJGMGjVKJejLwcGBzMxMRQqFkSNHqgS7fPrppxw/fpw2bdoQFxdHp06dzCVVIBAIBAZg1uG4qamp\nDB8+nICAANq2basSdJaVlUXp0qXZtGkTd+7coU+fPty8eRMbGxsqVKjAzp076d+/P127duXcuXME\nBQWxefNmrceTyWQWHSQmEAgErwJmMxz37t1jwIABTJo0iUaNGgH5OYeOHDlCw4YN2b9/P40bN+bz\nzz9X7DN//nzc3Nxo2rQpGzduxNExfyy+i4uLXumpJRIJd+8WTh9ekri5OQlNemKJuoQm/RCa9McS\ndbm5ORm0vdkMx6JFi8jMzGTBggUsWLAAgPHjx/P999+TnZ2Nj48PrVu31rh/cHAwEydOJCEhgdzc\nXL7//ntzSRUIBAKBAbxyAYCWaMmFJv2wRF1Ck34ITfpjiboMbXGIAECBQCAQGIQwHAKBQCAwCGE4\nBAKB4BXl4sUUGjasg7t7adzdS+Pp6cy771Zj797dunfWgjAcAoFA8Apy8WIKjRvX48qVS4pleXm5\n3L2bRteuHdi/f5/RZQvDIRAIBK8gU6dO0bo+KGiI0WULwyEQCAQCgxCGQyAQCF5BwsLCta6Pjv7J\n6LKF4RAIBIJXEB+fahw+fJKyZV0KrRs4cCi+vs2MLlsYDoFAYPHs3bsbH5+KuLuXpkKFcqxaFU9M\nzELKl3ehfHkX5s2bSWzsT8TG/kRGhmnnJn+ZqVrVm/PnrxAZOQepVIpUKiUycg5Tp04vUrkictzM\nWGqUqKVpAsvUJTTphzk17d27m65dO+i9vZdXefbuPUj16lUs7jqB5d4/Q9CrxZGZmcnff//NmTNn\nyMy0rBMWCASvNiNHDgOg08rBij85nRIG0ylhsMr2qam3WLFiebFqfN3QmuRw3759xMTEkJKSgqen\nJ1KplNTUVLy9vQkMDKRZM+P7yAQCgcBYlI2H8u9fei4G4OTJE8Wu6XVCo+EIDQ2lXLlyhIeH89Zb\nb6msO3/+PImJiSQlJTFz5kyzixQIBK8vc+cuoGvXDvzSczEdfh6IlfRFR8nDG/cpXdG50D516tQt\nTomvHRp9HLdv38bT01PrzqmpqXh5eZlFmLFYYt+h0KQflqhLaNIPc2uKi1tMaOhYg/dzc3MjPn4N\n9erVN4Mq47DU+2cIGn0ccqMRERGhmPtbTkhICIDFGQ2BQPBqMmDAYHbu3Ierq5vKcjs7O6373b17\nl9at/UTXlYnR6RzfsGEDPXr04OzZs4plFy5cMKsogUAgyM+19IEiQV+rVs3o0qUraWkPFX/XrqUx\nbVqUzrKCg4OKQfHrg07DUalSJaZNm0ZQUBBJSUkAYl5vgUBgVuQJ+i5eVK2kLly4gClTJpWQKoEc\nvYbj1qlThxUrVrB8+XKmT5+OVKr/jLOnTp2id+/eAFy9epUePXoQEBDA5MmTkbtXli1bRteuXena\ntSvz589X2f/ixYvUr1+f58+f631MgeB1JSMj/ZUIhNOWoG/+/Dkqvzt06Iy9vYPW8v7667Si5VK7\n9tskJx8zic7XFZ2GQ/5x9/DwICEhgbS0NP7880+9Cl+yZAkTJkwgOzsbgGnTphEcHExCQgIymYzd\nu3dz/fp1kpKSWLNmDWvXruXAgQOcO3cOgKysLKZPn06pUqWMPT+B4LUhIyOdhg3rMG5cCOPGhVCz\npg/NmjV65T+SLi7lOHHib1q0aKnX9qmpqcLvUUR0Go7ExETF/6VKlWLWrFmsWrVKr8KrVKnC/Pnz\nFcbnn3/+oUGDBgD4+vpy8OBBvLy8iImJUXR/5eTkYGdnh0wmIzw8nODgYGE4BAI9CAv7hocPHyh+\n5+XlcebMPy/lR1Jbgr7hw0cXWubs7MKqVesVvg/590Q5aLDjikEq+wi/h/Fo7HMaN26c1h3r1tU9\nTrpVq1bcuHFD8Vt55K+9vT2ZmZlIpVKcnZ2RyWTMmDGDmjVrUqVKFaKjo2nWrBk1atTQ5zwEgtee\nTZvWA/BlXH+sbaXcPnGNg7N+BaBVq2Z06tSF77+fgYtLuZKUqRfyBH0BAV1JSTmvWD58+GjCw78z\nqkyJlapvVt59BeDo6EhcXDzNm7fQWU5y8jECArqSnn5PZbm1tTWzZ8+jR4/eRul7mdBoOBo0aIBE\nIkEmkxEdHc2IESMUH35jneNWVi8aOI8ePaJ06fyb9uzZM8LCwnB0dGTy5MkAJCUl4eHhQWJiIvfu\n3SMwMJD4+HijjisQvA4ov5cSKwku1dxV1v/yyzr27NnFoUMncHYunDHV0qha1ZuDB43rZhs9egyz\nZ89URJIXTEtSkKysLLp27UBiYpLWrLHJycdo3dpP7brc3FxGjhyGlZUN3bp1N0r3y4JeSQ7bt2/P\nxo0bjTrAjRs3GDNmDGvWrOGrr75iwIABNGzYkPDwcBo3bkzr1q0ZOHAgjRo1YtCgQWrL8PPzY8eO\nHdja2hqlQSB4HejVqxcJCQkANBnzGYCixVGQyMhIvv3222LTVhJMmTKFSZMKj8DquGKQSusj9fhV\nxXWqUKGCSi9JQerWrUuVEXWR2tkUWic3UnZ2djx58qSo8i0a/YdHFQF5TSg0NJSJEyeSnZ2Nj48P\nn332Gbt27eLo0aNkZ2ezf/9+AMaMGUOdOnUK7a8PlhiRKTTphyXqepk0TZwYwbp163j+/LlGgyEn\nNDSUx49zGD58hFk1lSTh4eEMHTqauLhYwsLGkJeXp3OfvDyZ2vPIyEhnw4ZE7txJo4oex9Z2LSzx\nWhkaOW72FkdxY4k3RGjSD0vU9TJoyshIZ8iQQPbt26Nxn7f8a1GrV6P8HzL4JSC/dpyW9tAsmiyB\ngpoiI6cwe7b+ufW8vX1ISFiHs7MzTZt+yN27aXrtFx29WGtXlVyXJl+JjY0NM2f+UKy+EkMNh8YW\nh3I8xb179wrFVwwfPtxAaQKBwNRkZKRTr957PH78SP+dlBrwDRvWYtWq9fj4VDO9OAsjNDQca2tb\noqCv27EAACAASURBVKKm6rX9pUsXadLkA3r27K2X0bC2lvLDDz/q5d/Q5ivJzs62eF+JxhZHdHS0\nootIJpMV6i6yVMNh6bUeUzBszzdqly/wm6HX/pZYO4Ti1aXvNSypaxW8bwLPcl8EvSrrUtYUG/sT\n48aFaCynSpWqXL16GVBNRZ7zNJt/Eo8DcGHbaRo3boKrqzthYeFGGRFLfKb00VS7dg1SU29p3cbO\nzo6nT59qXN+uXXtiY382SNf779fmr79OF0oPDy98Lra2tty4cU9NCabHZC2OQYMG6Uwg9uzZMxFj\nIRBYKBKJhKNHT7Fly0YGDOjDLz0XK7qspHY2iq6rC9tO88cfBwHYunUzf/yRTNWq3iUp3aKwsbHR\najjMlcI9JyfHLOWaAo0tjqCgIJo2bYq/vz+Ojo4q67Kysti0aRMHDhzgxx9/LBah+vIy1nqKQsTh\nWQBM+HCM3vtYYu0QLFOXoZoiI6cwZ07+PRk9egyhoZoD2YqqKSlpA4MGDSAvL1ftdiEhYYSEhAKw\nf/8+OnduBxSeBEl5VBFA06bN+eWXzRqPn5S0gcGDA8nNffFha9euHWlp9/jnn7+wt3cgOnqhXjER\n6pA7oiE/nYixcSf63DtDp6UtiKOjE8eP/4mzs4tGH0rB+A43Nyd+/fW3Ql1VX8b1B2DTgKVK5TvS\nrt2XrFqVoFhmY2NLq1afsXVrkk595cq5kpCwVmdaeZM5x3Nzc1m1ahUrVqzAyckJT09PrK2tuXXr\nFvfv36dPnz706NHDoLxVxcHL/uEpDixRE2jWNW/eTCIi8nMXeXh4sHz5qmKbX8GQa6XuwyGRSAhZ\nPBmnqs68VdYbv8q+JtEUF/czgYF9NW6jbDTkJCVtUNnn0xldVCZBkg8nBRg6dBjffTetULkFy9CG\nrpgIdWRkpNO8eRNu304FwNPTi337/jAq7kTfe7d//z769OlumJ8IaNGiFT/+uFir0VBG7jSX6zp5\n8gQ9e3bm3r27Bh3XGBITN+Pr21zjepOPqpLJZJw9e5YrV65gbW1N5cqVefvtty02Q27BB6Wo/gCA\nEb+NI1emvlanqxxdD++ea/v5JWVLkTUagpubE13XDNW4vpZrTYbU6qfx2pWytmV2swiz6Cp4rTS9\nkDt37lN0ERT0ByhT1GuoTpOme1ZwuGdeTp5itjorrNQm+DFGn5ubE2XKlMXj4yovRkrxouVQunQZ\nUlKuq91369YtDBzYm9zc/OdZXR+7HOtsK+Z/rqqvWrVKKmlNCqJc3p5hiZw7d0XtdpqeLfIgj/zr\nKMuTsb7XEiZO/I6goMJpRnRhbAXJ07Os1qG7UqmUW7cyVJZ5eJSh2ufvq9wPUDXGdnZ2XLuWplbX\n1q1b6N+/J1D4niiXoQ9tF/amVJk3Ci1XDsB2snUk8uMXrWGTTeQkRyKR8M477/D555/TqlUratSo\nYbFG42Vk57W9JS3BYsnISNdYi3sZ8gwpT3Eq/xiakrfb1TZ4nzZt2pKaep+vv84f3PJLz8X80nMx\nOU+zC22ba5PH0F1j2bt3t1H6Hjz4V/dGBVC+TvIgvZclz9abzd82et82bdrq9CkXGRM+gpbVz2QG\nFvjNUNRuJEiQIaNTtbYGlTHvk/wme8ThWaQ+uqNYbmg56nC0ccDRxoEmXvnJH03RlaEP+tR05dsM\n35MfYTzfb7pZNRVE3s+tC3O0frThV9lX7X1Sbh0p1xp/6bmYtgt74+7hAUDW80d4Orgb5JdS5sKF\nC9jZ2fEs86miZqlcK507d6HOMiZPnoqrqydTpkwA4J/E47zZ/G0e3XmIZ93KKpHVyqk45syZp7Wr\nSlmHNt+Epudv3rzZRERMpu3CFzEMxT1/eFjYBEXXqDoiIgpPHDV69BgO3fkTp/JlyX2eg7Vt4U+r\nrW0p3N1LY21tTfv2HRk7dpzCKHfo0Jnp02cxcuQwxTXs8PNAo/RvGZqfmqlTwmDF0Oucp9k0ulOT\nwMAhRpVZEL0CAF8mLK3v3hL9CZaoCQrr0jbUVLmrSk5MzELCwvKNnFQqZdasuUUOojL0WrVu7ac1\njXlw8NgiOcy1jf+3sbFh8eLltGljWIUmNDQ/ulofJkwIx8enBoMG9dE56sdYH8cnn3ykGCLr5VWe\nvXsPmtXHoY758+cpjKoyXbr0ZMGCRWr3iYqK1BAjIgG0f2bl57lz506CglS7qt544w2VFCa2trZ8\n/nk7Nm36Red5yHF2dtaao8zkPo5BgwbRsWNHWrZsiY1N4fwsloYhD4pyX7W89VDUGr+8dSPvQ7TE\nj3RRNcXELCQ8fDwAU6Z8zwnvy4p1RfEpyHXtuZafeuaXlC3k5ebx7GH+SyOvSQHs3LmXOnXqqWiS\nGw1l5A5J5T51J9v8UYLKfby6NGkiIyOdFSuWc+DA76Sm3qJKlSpIJBJ27NimcR91jmt9yMhI5913\n38L7s5oqfenyGqq6eIKfTi8DYEitflrL1vzRK0x4eEShVCULF85R5IVydHRi2bKVBhsNOffvZ7B+\n/ToAOnbsYnRCRlO8e2vWrCYkJP9co6LmGRWQ5+f3MW+Pa6QSeKmM/P4p+3IMvQaXL18iPDyMmzdv\n0Lq1P66uHkyaFEpubi7t23fm++8jtZZhcsNx5MgRNmzYwOHDh2nWrBkdOnSgVq1aBh2kOLl7N1Nl\nFI6TkxOxsT+rHRpoCsd5cZRpagx5oTSlRVBG3i2j7HwDw89ZrqvgNZQ7KpWHjdra2vLnn+cVL0P5\n8i5qa8A2NjZ0WzlUrfNcH33arpUmY6ULiUTCnTuaHcyakLfAVNKHoN5wGPMc6jtiysrKitu3Vf0X\n8uuUkZHOd99NZNWqFYp1yjVmLy8vli5NKJZRcZZSafPz+5i3wxppXH96xSEubDttcCChKTG5c7xh\nw4ZMmzaNbdu2Ubt2bYKCgmjTpg3Lli2zyOlcR44cqtI/mZmZSdeuHdi/f1+hbSWaqgBFoJZrTZOX\nWVLIu0W0GQ3IHz0EhUcVyVsO5uD58+eKGpk25MZEgoRS1qbLrrxqVbxRRsMUXNh2Or/nQ/bCaEgk\nEiZMmFykctu160B09E9G75+RkU6jRvVUjAag0s0in31v61bNcSKvGjNn/qDaU1Wgqi53qhe3L6co\n6OXjOHToEJs2beLgwYP4+vri7+/PgQMHOHPmDLGx+vWNFgf//e9/adWqFVDYOenlVZ5Tp84WuyZL\nqfUoo68mP7+P+euv0waV7eHhydytMYouQENaHZp0vftuNbW5gqZNi1I4+zTV/itUqMiJE//orUFf\nTZUru2uNJtZGUbqq6tevRVaWqh43Nw+2bPnVZNHe9+9nMHjwAI1JE9V1Vbm5OREZOVPhk1J+/woG\nGMo5fPikWSPULendi4v7idDQwv46efr7U4sOkJz8V4nNk2LyFscnn3zCggUL+PDDD/n111/5z3/+\nQ+PGjRk9ejTp6elGCzUH/fvnR14qj8gQFC9Pnjwx+ciwBQsK14JdXV3p2LGL4vfAgUP5+utRhbZb\nujSh0LLiwN7enmrVqhdabqzRgPxRSseP/0mXLl2wsrJCKpUSGTmHv/++YNIPsLOzC+vWbSQkJKzA\nGolao2EsERGTTVLOy8CAAUNITEzCwUE1C8fBWb8iPfq8RI2GMehscVy9ehUHBwdcXV158uQJaWlp\nVKmiT0b64qdixYrcvHmTtgt7Y+uUPyb6eeZThVO1qCNajMGSaj1yCmqKiVnI+PGhyGQynJycyMw0\nXq+bmzuNGjUxKlne/7V353FR1esfwD9nWGUdUCDFFcW4aopbaXbRyFuUKwViiZEMiIpiiggoiAsp\ngkqEaGjigqglIol6zVwAkwtGKnlNUvhFYQre6+DCKArO+f3BnYmBMxsOM4M879eL10vODGceD1/m\nmfNdnq+ia5Wfn4fg4EA8fvwY3t7TsXRpJOcf2uXLl6RrPDZtSn7u2395Me3fn46FC4Pl/tzOnfvU\nnt30vDHpkp2dJX79tQIuLn0AcJfPaI7H4yExMbnNyofrw3XKyTmMoKAANDQ0rpOxsLBAVlYWXF3l\nj3nogsYHx/fs2YOsrCxkZ2fj5s2bCAgIwCeffILp0/Wv3G/Trirgfzt9MQx+zmgcfAKe7xNfa+hD\n422uaUytHeBVhsfjqV0sT9+vVXNff32gxdRJY2MTpKbubJOkIZkhFfXWAr29TpKCiuoIDY1EeHik\nxmLRl+ukaLJBa6YqtyWNJ44JEybg4MGDMDMzA4D/feLzxtGj3GUydO3QoaMIDg5EdXUVPPcEgOEx\nuLKvSJo4WjujpbWUlfcA2q6EhzxN3wy7drWRlp8AGvtcuw7/645SMvg6aNBgnDnzg/T4b7/9H8aM\nGSn9JMVFWbE8RXHpC32KKfXnXfj5v3+N1+jrTL38/DwsWNA49pScnAo3t7FIS9uBiAj5ZUO41uW0\nlr5cJ0XlWXQ15iqPxsc4GhoaZNZvGBkZ6W3JkQsXLiAqKhw8Hg8Mw+Dwx18hy3e7NGmQlpomDQBw\nGNJDpZ/r08cJdnZ2Cp9TUtI+SkW0F8rWYugLN7exKCkpRUlJqfRTtb+/oMV07aY0WUKm+XWSVJAm\nmqO05Mj48ePh5+eH9957DyzL4uTJk3B35165qipPT09pqfbu3btj3brGkh5r166Fk5OTtBts165d\nOH68cSGVm5ub0s2jXnvtNaWvvWSJ5m6JVaVPnwwlJOszmqsXPYGRuQmqSyplZsJs2pTc4rlJSSkK\nS1I7OLykmWD1SG7uafj5fSSdYsrjGSAx8QutbfOZ4h6vV3dB6uDxeCrt+60JKe7xWHg2Eg1yipNq\ng6LyLM8z7VkfqFQd98SJEyguLoahoSFGjhyJ8ePHt/oFnzx5gunTp+Pw4cPSY0KhEEuXLsXvv/+O\ngIAA+Pj4oLKyEp9++ikyMzPBMAw+/PBDrFy5Ei+/LL+QmKp3Qs7O/bFnzwGtbJepj3/kv/12TaUk\nCwB8Ph/ffPOt3G6E5ORErFkTw/mYut0PXNdKU3sztFbTmBTt3aBsn+m2iklfqBKTovG0pm0lN/e0\ndOJBUlJKq/f10IfrdOzYUcye7Yf6esnguCW+/TYbr7wyUqdxNafxMQ4AuH79Ou7fvw/JUxmGwciR\nrfuPl5SUIDw8HI6OjmhoaMDixYvRuXNnCIVC5Ofno0uXLpg+fToaGhpQW1sLPp8PAPD29saGDRsU\nzujiShzOzv1x48Z1uc9fvjwaISFLWvV/UYU+NN6mLl4sxsSJbyusM2RmZobu3Xtg8+ZtKr3xN9/P\nwM7OARkZ36jdZ938WoWHL8LOnX+tE2peswdorJ2krd+fom1GTUxMUFnZ9vsqNI9JX6gaU9OxDh6P\nh549e2Pbtp0ySaN5cm7tQLI+XidAP+PS2NaxEqtWrcLZs2fRo4ds33d6erqcn1CsU6dOEAgE8Pb2\nRkVFBQIDA/Hdd9+he/fuyM//a6WxoaEh+Hw+WJZFfHw8BgwY0KppwA8ePJD7GMuyiI1dDR7PWGNz\n07k0rZrKMAw++ywOAQGKB8zbgiolJRiGQUVFlVrndXMbi4qK288TWgszZnjj++9lF401TxoAtPL7\nU0XzsSLCzd9fAH9/gdzHuaY4L1gQpFcDyQQAq8Q//vEP9vHjx8qeprInT56wdXV10u+9vLzYqqoq\nlmVZNjk5md2/f7/0sbq6Onbx4sXsihUrWLFYrPTckBZiaPxiGIb18PBocbz5F4/H09j/r7no6GjO\n19yyZUubvSaX69evK70OANgpU6ZoNS4umZmZCmP8YN9s6dfroe+wAFhTU1M2MzOzTeM6efKk3Jh8\nfX3b9LU7CkdHxxbX1tHRUddhkWaU3nH06NFDowNaWVlZ+PXXXxETE4Pq6mrU1tZyzs5hWRbz5s3D\nqFGjEBgYqOarMHB1HYrU1DSIxWKcOHFC4bPFYjEsLCzwxRdbMGlS6/cf5hIbyz3NNjg4GF5evhp9\nLUVCQ+XsuNaEpaUl4uOTdHYbLbmFV/SJtDlzBysAQF1dHby8vDBjhh+io1dqbCykabeCq+soZGbm\nwNd3Gurq/rr7sbbmIzo6VmvXTV+7OjQRU2Li5hZdVUlJX7bq3Pp4nQD9jEvjXVVWVlaYMGEChg4d\nChMTE+lxyUwodXl5eSEyMhIzZsyQnqfpND3JOMWpU6fw448/or6+XtqFFRoaCldXV4XnDg1d1mLQ\nu6joMmbMmIayMu6xDgAQiUQQCPywc6eRxhZv7d+fLh0Xak7e8bZw8WIxTp78p9zH//a3gfj440+e\nq4S1NknWlkj2zPbcE4DDH38FAMjI2I3s7EwsXrwUV6/+G8eOHQHDMFi/fuNzzXzKzT2NTz6ZgUeP\nHgEAzM3N4ePzEZyd+7eb66avysvL4Ovrg/LyGwCAIUNcUVVVBR6PJ10HQvSL0sHxrKysxif+7w2d\nZVkwDANPT81+MtcEhmFgYGCAgoKfOFcsx8REYuvWFKXn+PrrrFbP5JBQtnl9166OWLAgRDrLxNTU\nFCkpqRq/41G08Q/QOOZ0+fI1vXjjk3wS4xqLMTQ0xCefBOKrr/7a3a55IUtlWjPzyc7OEgcPZsud\nTaWLFcD6+om1NTGVl5dh9OhhLY5rol3q43UC9DOuNplVVVlZibKyMrzxxhuoqqpqMVCuLyTJTVFd\n+9WrY7B5c6LSc4WGhiM8fHmr4lD2Zg0AEydOxtGjLVdVa7rG0ciRg/H77xWcj3l4TERS0ma9SBqA\n7B/UsWNHsXBh4wSCpKStMtfEwcEaLMtKE0dDXb3CmkhNzZu3ACEhi1XuyrKzs4SNjS3u3avhfFwX\nK4D16Y2Ha98PeeuWJAvxmm6ZO3z4IFRW/sH5fIFgNtatk//hS5m2uE5N/7+tXZ+lT78/CY2vHD92\n7BjmzZuHzz77DPfu3cOHH36I7OzsVgeoaytWrMK6dS33DG5u48b1sLe3wpgxI1BeXqbWayxZ0rJK\nq8SgQYMRF5fAmTQAICREM3sCS/z55025jxkZGepN0mhuwoSJKCurRFlZZYtEumhR4xvPoY+24dBH\n21ROGgCwZUsyRo8ehpoaoco/c//+PeVPIlKSWlESwWeWIvjMUtwWVaO2XiQ9npt7Wm7SAICDB79u\nqxDJc1KaOLZv3479+/fDwsICdnZ2yMrKwrZtyrsFdMXAwOC5N7Rp6saN6xg9ehjs7a3g4GAt01Wi\nLktLK5w58wOSkuTf8UgWCj2vxi6A4Ur3hW6PIiJWcJT8Vl1NTQ0GDnRGTs5h5U9GY7eJPO19BfDz\nMmAMZL5XtlnW2z3HSf+tqLowoPi660qKe7z0qyNTOjjO4/Gk5UEAwN7eHgYGBgp+Qne8vLwQFhal\ntCKrp6cXEhLiIBSqt58Iy7JYtiwcPJ6x3Jk/ubmnpYN8zUn2huBajyDxwQc+nMclW3IePpwJIyMj\nfP75ZrnjIfL6jZvi8XgaTbDaFhYWgbfeGq+0S1CehoZ6lSdDJCdvbTHmYmpqir17D3b4gdsv3myc\nJKOo+6W1b7IpKdtbHRdpW0rvOJydnZGeno76+npcu3YN0dHRcHFx0UZsajt48KBKZbxtbTvjX//6\nCQLBbIWF1+RZvpx7pbJk1WvzxGBuboHMzBw4Ojpi0CBnuf3l1tZ8rFixqsXxplty1tXV4eHDhxAI\n/HDsGHeF4rVrV3Mel/Dw8FC75Lk+GjZsBE6ezEOvXn1gaGiILl26IDMzB5mZOejUyUylcwgEvhAI\nPlbYHTlpkid27twHKytrWFlZY+fOffjjjzsdPmk8r6SklhNVOnXqhK5du+ld2XF9lPrzrhbdgtqi\ndHBcJBJh69atKCgoAMuyGDVqFIKDg2XuQvSJuoNO5eVleP314WpNj5VXmt3ZuSdnf7ihoSHOnbsg\n9y6AYRiEhoYjMHAO55jDjh2p0i05m7KyskZZWWWL4wLBx8jJ4R6H4vF4ePbsmd4NzgGaHzTs3fsl\n6fRZZXg8A/zrXy1n4+nrQOaLEhNXCXZdx9TWNBUX18QEicFdBqhVTVnjg+Pm5uZYsmQJsrKycPjw\nYYSHh+tt0miNvn37obDwEiZNmqryzwwYMKjFMaHwrsJBVEV3AZaWVli6dJncgWp5tba4XLxYjPz8\nXLmPT5o0ReVztXfJyaqPR4nFzzB5sofMnUdu7mlYWlrC3t4K9vZW6NnTXuVxEaIarhLsRP8pHePg\n6payt7eXqSvV3vXp44QdO/aovBteWlrLOl2SCq5cYmMTcP58ntzHAwPnolcvB2kXl5mZOXbt2itd\nS5KZ+Q3nzyUlyb4xKpsGbGlphfh45VORXxSNXUxGmD9/NkSiWqXPr66uwujRwxAdHYNXXnFtsXaj\nrq5O44tECWktXQ7QK73jKC0tlX5duXIFiYmJ8PDw0EZsWhcQMBdxcYlyB////vdxKCq6rNbYwOzZ\ns+HvL8CyZdx7nY8ePQYbN8bJjIs8eiTCtGmeyM9vTDYPH3IXamz+5iUQyN+yUyAIQnHxz3o7/bat\nTJgwEb/9dgtFRZcxZIhq1XrXrFmFwMBZch+fN0/1kiiEvIjUGhk2MjLCu+++i8LCwraKR+f8/QW4\nfbsGv/5ageDgEHTt2g1/+9sAnDyZh0OHjshNGp6eXnB07C79nmEYJCdvQ2pq43TNvn37oajoMvr1\n6y99TljYMlRU/CY3luDgxhpdpqYtpyVKtvKVyMk5rHDNxvHjOR0uaTTVp48Tvv8+DydOnFHp+Yq6\nHRXNiiOkI1DaVdV0wyWWZXHjxg0YGyueq/0isLGxRUxMLGJiVNsL3Na2M86c+QFZWQcBgLN+UZ8+\nTigoKJY5tnfvLrnnlLxBubu/hWPHcmQeMzQ0Qm7uaYwb95ZK5dJJI8lMrOnT31d7OraEPq4vIESb\nlCaOoqIimQ2SbGxskJjYcfrJ1WFjYwuBQL2V34q2X/X2no64uNUtkgYAPHhwH9OmeSIzMwdz5yqv\nHtzRF6o15eo6FKWlv2HGjGn4/nvFlZO5bNmyQ/mTCHmBqVSrqj3Rt+l3qky9y8/Pa1Gqu0uXLjh/\nvhguLn0UThW2tuYr7FZxcHgJKSnbZWasvOjTFFUlFN7FiBGDUVur+mtqupZYa+jj749iUp0+xqXx\nsuru7u5gGIbzzYthGJw+fVqtFyQtubmNxR9/VKOmRqiwq4uLoqQxatQYHDkiv5x6R2dr2xk//XRF\n5pqPHDkEDx60XKMD6EfSIEQfKE0ckyZNgpmZGXx8fGBoaIijR4/ip59+Qnh4uFb3lOgIuLq6Fi0K\nVVieXRGulblEljrdi5Q0CGmkdFZVfn4+goKCwOfzYWFhgenTp6O8vBxdunTh3LmPPB+h8C4WLpyH\nnj3t0bevIwYOfAVhYctkxpmUsbOzV3vaMGmUmPgF5/HkZP0t7EmItqk0HfeHH36Q/vvUqVMwNzdv\ns4A6ssY+91da1KQyNTVFdfV9zJoVoNJ5Fi8Oo6TRSpK6VKampgAap6C3ZgMoQl5kSruqYmNjERYW\nhrt374JlWTg5OSE+vvUrFuvr67Fs2TLcunULT58+xdy5c/HSSy8hKCgIvXv3BgB89NFHePfdd5GX\nl4ctW7YAAF555RVERUW1+nXbgy++2ITa2pYrnNesWYG//30sDh06qPQcDMPg/fe92yK8DmPChIn4\n4487APRzIJMQXVOaOAYOHIjjx49DKBTC2Nj4uetU5eTkwNbWFgkJCbh//z6mTJmC4OBg+Pv7Y9as\nv1br1tbWYsOGDUhPTwefz8e2bdsgFApha/viLmI7fPiQ3McWL16ATp06yR24lUhLy+jQC/0IIW1P\naVfVzZs3MWvWLPj4+EAkEmHmzJmorGxZkVVVHh4eCAkJAQCIxWIYGhri6tWryM3Nha+vL5YvXw6R\nSIRLly6hf//+iIuLw4wZM2Bvb/9CJw2gcWqtIsqK9s2fv4gGcAkhbU5p4oiJiYG/vz/Mzc1hZ2eH\nyZMnIyIiotUvaGZmBnNzc9TW1mLhwoVYtGgRBg8ejPDwcOzduxc9evTA5s2bce/ePRQVFSEsLAzb\nt2/H7t27UVFR0erXbQ8+/3yz3Mc2bUrGuHFvITMzB2ZmsmNMxsbGSE7exrmXByGEaByrhKenJ8uy\nLDtlyhTpsUmTJin7MYVu3brFvv/+++yhQ4dYlmXZBw8eSB8rKytj/fz82Pz8fDYoKEh6PDY2lj1+\n/PhzvW57UFxczDo5ObEMw7AGBgZs37592eLiYl2HRQghUkrHOExNTVFVVSX9vri4GCYmJq1OVP/9\n73/h7++PmJgYjBo1CgAQEBCA5cuXY/DgwSgoKMCgQYMwcOBA3LhxAzU1NbC0tERJSQl8fLi3VW1K\n3wYy1R1c7dmzPwoLL7c4rsn/l74O+OpjXBSTaigm1eljXBpfOR4REYHZs2ejsrISkydPxv3795GU\nlNTqAL/88ks8fPgQKSkpSElpXKC2bNkyrFu3DoaGhrC3t8fq1athbm6O0NBQCASNJazfe+899OvX\nr9WvSwghRDOU1qrKzc3F66+/joqKCojFYjg5Oel1dVx9zOQUk2r0MS6KSTUUk+r0MS6Nbx0bHx8P\nY2Nj9O/fHy4uLnqdNAghhLQ9pV1VPXv2RGRkJIYMGSId22AYBlOnqr5HNyGEkBeH0sTB5zeuLSgp\nKZE5TomDEEI6JrmJo7q6Gg4ODoiLi9NmPIQQQvSc3DGOoKC/Sk2npaVpJRhCCCH6T6XquEeOHGnr\nOAghhLQTKiUOQgghRIISByGEELXIHRwvKyuDu7s7AODOnTvSfwO01zghhHRkchPHiRMntBkHIYSQ\ndkJu4ujevbs24yCEENJO0BgHIYQQtVDiIIQQohZKHIQQQtRCiYMQQohaKHEQQghRCyUOQgghatFZ\n4igpKcHMmTMBANeuXcOMGTMwc+ZMCAQC3L17FwCwb98+fPDBB/Dy8sKpU6d0FSohhJAmlO7H0Ra2\nb9+OI0eOwNzcHACwdu1aREdHw8XFBV9//TW2b9+OkJAQ7NixA9999x0ePXqEqVOnYvz48boIPCfc\nAAAACmBJREFUlxBCSBM6uePo1asXNm/eDMl255s2bYKLiwsAoKGhASYmJmAYBgDw6NEjiEQi8HjU\nq0YIIfpAJ3ccb7/9Nm7evCn93s7ODgBw8eJFZGRkICMjA506dcLEiRPx3nvvQSwWy+wPQgghRHd0\nkji4HD9+HF9++SW2bdsGGxsbXLx4EZcuXcKZM2cAAAKBAEOHDsXgwYN1HCkhhHRsepE4vv32W3zz\nzTdIT0+HtbU1AODx48cwNTWFsbExAMDS0hK1tbVKz2VnZ9mmsbYGxaQ6fYyLYlINxaQ6fY1LVTpN\nHAzDQCwWY+3atejWrRvmz58PAHjttdcwf/58nD9/Ht7e3jAwMMDw4cPx+uuv6zJcQgghABhWMkJN\nCCGEqICmKhFCCFELJQ5CCCFqocRBCCFELZQ4CCGEqEUvpuM+j7t37+L999/Hrl27IBaLERUVBYZh\n0Lt3b3z22WfSFeja4unpCQsLCwBAjx49EBQUhIiICPB4PDg7OyMmJkanMXXv3h0ff/wx1qxZAwMD\nAxgbGyM+Ph6dO3fWWUw9evTA2rVrAQA5OTnIyMjAgQMHtBqPvLiWLFmC5cuX4+HDh2BZFuvXr9f6\ntsrNYxIIBDpv56mpqTh79izq6+vh6+uLYcOG6bydN4/JxcUFsbGx4PF4OmvnzWPy9PQEoPt23jwu\nNzc3REVFqd7O2Xbs6dOn7Lx589h33nmHLS8vZz/99FM2Ly+PZVmWDQ0NZc+cOaPVeOrq6tipU6fK\nHAsKCmIvXLjAsizLrlixgv3+++91HpOvry977do1lmVZ9sCBA+y6det0HhPLsuzVq1dZPz8/1sfH\nR6vxSHDFFR4ezv7zn/9kWZZlCwsL9aJN6bqdFxYWskFBQSzLsqxIJGKTkpLYOXPm6LSdc8Wk63bO\nFRPL6r6dc8UVERGhVjtv111V8fHx+PDDD6UlS0xNTXHv3j2wLAuRSAQjIyOtxlNaWorHjx9DIBDA\nz88Ply9fxi+//IKRI0cCANzc3FBQUKDTmEpKSpCYmNiiNpiuY6qpqUFiYiKWLVsmrWGmbVy/v0uX\nLqGqqgqzZs1CTk4ORo0apfOYdN3Oz58/j5dffhnz5s3DnDlz4O7ujqtXr+q0nXPFpOt2zhWTPrRz\nrrguXryoVjtvt11VWVlZsLW1xRtvvIHU1FQAgK+vL/z9/bF161ZYWVnh1Vdf1WpMnTp1gkAggLe3\nNyoqKhAQECDzuJmZGR4+fKjTmAIDA/Hdd98BkK0NpsuY/P390b9/f0RERGj9j1tRXAEBAbh58yas\nra2xc+dOpKSkSCs36yqmwMBAbNy4EYGBgTpr50KhELdv30ZqaioqKysxZ84cmTdBXbTz5jHNnTsX\nJ06cAKC7dt48pqCgIPTr10/n7Zzr9/fnn3+q1c7bdeJgGAYFBQUoLS1FeHg4ampqsG/fPvTt2xcZ\nGRmIi4vDihUrtBZT79690atXL+m/+Xw+rl27Jn1cJBLByspKa/HIi+nOnTu4ePGiTG0wXcZ069Yt\nGBoaYuXKlXj69CnKysqwbt06REZG6jQuPp+P27dvw93dHQCkn2J1GZO1tTUWLVqk03ZuY2ODvn37\nwtDQEH369IGJiQnu3LkjfVwX7ZwrJqFQiMLCQp218+YxVVVVwcjISOftnOtaicVitdp5u+2q2rt3\nL9LT05Geng4XFxesX78eYrFYuseHvb09Hjx4oNWYsrKyEBcXBwCorq6GSCTCmDFjcOHCBQBAfn4+\nRowYodOYamtrUVRUhIyMDKSnp2t9oJcrJicnJ5w4cQLp6enYtGkT+vXrp/U/Jq64RCIRxo8fj9zc\nXADAhQsX4OzsrPOYGhoadNrOhw8fjnPnzkljqqurw6hRo3TazpvH9PjxY+Tl5em0nTeP6aWXXsLR\no0d13s65fn9vvfWWWu38hSg5MnPmTKxatQpVVVX4/PPPYWJiAmNjY6xZswbdunXTWhwNDQ2IjIzE\nrVu3AABhYWHg8/mIjo5GfX09+vbti9jYWK3ONmkaE8MwWLx4MebOnYtu3bpJZ+q8+uqrWLBggU5i\nAhqvk6urKwDg5s2bWLJkiU5mm3DFZW9vj6ioKDx69AhWVlbYuHEjLC21V6COK6ZHjx7ptJ0DQEJC\nAoqKiiAWixEaGgpHR0edtvOmMbEsi0WLFiE0NFSn7bxpTJLrNGbMGAC6bedccfXp00etdv5CJA5C\nCCHa0267qgghhOgGJQ5CCCFqocRBCCFELZQ4CCGEqIUSByGEELVQ4iCEEKIWShykw1u9ejWmTp2K\nCRMmYNCgQZg6dSqmTp2KrKwszJ49G//5z380/prr16+XqSrwPK5evYqEhASNnIsQVdA6DkL+588/\n/8TMmTNx5syZNn2d0tJS7NixQ6Nv9qGhoQgKCkL//v01dk5C5Gm3taoI0TSuz1Du7u5IT09HUVER\ncnNzcefOHVRXV8PPzw+3bt1CYWEh+Hw+vvrqKxgbGyM7Oxt79uyBWCzGwIEDERMTA2NjY5lzpqWl\nYcqUKQAak0hMTIy0euu6devQq1cv5OfnIzk5GQ0NDejevTvWrFkDPp+PgoICaXkdR0dHbNiwARYW\nFpg0aRLS0tKk5UkIaUvUVUWIEpLSGf/+97+xY8cOaWHBsWPH4siRIwCAc+fO4caNGzh48CAOHDiA\n7Oxs2NraYseOHTLnYlkWeXl50lpOu3fvxqxZs3Do0CH4+vqipKQEQqEQmzZtQlpaGg4fPowxY8Zg\nw4YNePr0KcLCwrB+/Xrk5OTg5ZdfRnZ2NgBgxIgROHv2rBavCunI6I6DECUkdyJDhw6Fubm5tMDg\n6NGjAQCOjo548OABioqK8Pvvv2PatGkAgPr6egwcOFDmXDU1NQAa944BgHHjxmH16tU4d+4c3nzz\nTXh4eCAvLw+3b9/GzJkzAQDPnj0Dn8/H9evX4eDgIN1jYtGiRdLzWlhYgGVZ3Lt3D3w+v60uBSEA\nKHEQorLmXU48nuwNu1gshoeHB6KiogA0lhd/9uyZzHMYhoGBgYH0+3feeQeurq7Izc3F7t27kZeX\nh3HjxmHYsGHYunUrAODJkycQiUQypcsBoLa2FiKRCA4ODgAAQ0PDFjER0haolRGiIa+++ipOnToF\noVAIlmWxcuVK7NmzR+Y5NjY2EIvFePz4MYDGQe0rV67Ax8cHISEh+OWXXzBkyBBcvnwZFRUVAIAt\nW7YgISEBTk5OEAqFKC8vBwBs374d+/fvB9CYRFiW1fo+GKRjojsOQppoXgqcYRjpl7Lnubi4IDg4\nGH5+fhCLxRgwYABmz57d4jXc3Nxw4cIFjB07FoGBgYiKisKWLVtgYGCAyMhIdOnSBWvXrsWnn36K\nZ8+eoWvXrkhISICxsTESEhKwdOlS1NfXo1evXoiPjwcA/Pjjj3jzzTc1fDUI4UbTcQnRstLSUmzd\nuhVJSUkaO2dISAgWLFig9Y2mSMdEXVWEaJmLiwu6du2qsQWAV65cgaOjIyUNojV0x0EIIUQtdMdB\nCCFELZQ4CCGEqIUSByGEELVQ4iCEEKIWShyEEELUQomDEEKIWv4fOQTWPHXFKikAAAAASUVORK5C\nYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x13471ced0>"
       ]
      }
     ],
     "prompt_number": 290
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "0"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}