MARS_rasheds_data_SSTnodetrend.ipynb

{
 "cells": [
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "# MARS modelling on seasonal and monthly MLSA using SST EOFs as predictors"
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "%%javascript\nIPython.load_extensions('gist');",
   "execution_count": 71,
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "application/javascript": "IPython.load_extensions('gist');",
      "text/plain": "<IPython.core.display.Javascript object>"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "!date",
   "execution_count": 1,
   "outputs": [
    {
     "output_type": "stream",
     "text": "Thu 19 Mar 2015 10:13:57 NZDT\r\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from datetime import datetime \nfdate = datetime.utcnow().strftime(\"%Y-%m\")",
   "execution_count": 2,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "[Nicolas Fauchereau](mailto:[email protected])"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### Station Name"
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": true
   },
   "cell_type": "code",
   "source": "list_of_stations = ['Guam','Yap','Pohnpei', 'Majuro', 'Tarawa', 'PagoPago', 'Rarotonga', 'Moturiki']",
   "execution_count": 3,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "station_name='Tarawa'",
   "execution_count": 4,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### lag (0 is SYNCHRONEOUS, 1 is 1 month lead time, 2 is 2 months lead-time)"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "lag = 2",
   "execution_count": 5,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### time-scale of the predictand: monthtly (`'1M'`) or seasonal (`'3M'`)"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "tscale = '3M'",
   "execution_count": 6,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## imports and settings"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "%matplotlib inline",
   "execution_count": 7,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "import os, sys\nimport numpy as np\nimport pandas as pd\nfrom matplotlib import pyplot as plt\nfrom scipy.io.matlab import loadmat, whosmat\nfrom netcdftime import num2date\nimport seaborn as sns\nfrom sklearn.metrics import mean_squared_error\nimport warnings",
   "execution_count": 8,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from datetime import datetime, timedelta\nfrom calendar import monthrange",
   "execution_count": 9,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "warnings.filterwarnings(action='ignore',module='pandas')\nwarnings.filterwarnings(action='ignore',module='matplotlib')",
   "execution_count": 10,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "rootdir = os.path.join(os.environ['HOME'], 'research/NIWA/CLCP1407') ",
   "execution_count": 11,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictor_dir = os.path.join(os.environ['HOME'], 'drives/auck_projects/CLCP1407/RawData/predictors')\n#predictor_dir = os.path.join(os.environ['HOME'],'research/NIWA/CLCP1407/data/predictors/')",
   "execution_count": 12,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Load local functions"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "#%%writefile ../../code/library/get_rasheds_data.py\ndef get_rasheds_data(dpath = '../../data/predictands/', fname = 'IWLS_SL_data_op.xlsx', station='Guam'):\n    from dateutil import parser\n    data = pd.read_excel(os.path.join(dpath, fname), sheetname=station, parse_cols=\"B:N\", \\\n                     na_values=[9999], header=0)\n    data = data.stack()\n    dates = [parser.parse(\"-\".join(map(str, [y, m, 1]))) for y, m in zip(data.index.get_level_values(0), data.index.get_level_values(1))]\n    dates = [d + timedelta(monthrange(d.year,d.month)[1]-1) for d in dates]\n    data.index = dates\n    data = pd.DataFrame(data, columns=['msla'])\n    data = data / 10.\n    data = data.reindex(pd.date_range(start=data.index[0], end=data.index[-1]+timedelta(28), freq='1M'))\n    # detrend\n    import statsmodels.formula.api as sm\n    data['X'] = np.arange(len(data))\n    y = sm.ols('msla ~ X', data=data).fit().predict()\n    y = pd.DataFrame(y, index=dates)\n    data['trend'] = y\n    data['detrend'] = data.loc[:,'msla'] - data.loc[:,'trend']\n    data = data.loc[:,['msla','trend','detrend']]\n    data['seas_msla_detrend'] = pd.rolling_mean(data.loc[:,'detrend'], 3)\n    data['seas_msla'] = pd.rolling_mean(data.loc[:,'msla'], 3)\n    category, bins = pd.cut(data.loc[:,'msla'], [-np.inf, -10, -5, 5, 10, np.inf], retbins = True, labels = [-2,-1,0,1,2])\n    data['m_cat'] = category.copy()\n    category, bins = pd.cut(data.loc[:,'seas_msla'], [-np.inf, -10, -5, 5, 10, np.inf], retbins = True, labels = [-2,-1,0,1,2])\n    data['s_cat'] = category.copy()\n    return data    ",
   "execution_count": 13,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla = get_rasheds_data(fname='IWLS_SL_data_op.xlsx', station=station_name)",
   "execution_count": 14,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla.head()",
   "execution_count": 15,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>msla</th>\n      <th>trend</th>\n      <th>detrend</th>\n      <th>seas_msla_detrend</th>\n      <th>seas_msla</th>\n      <th>m_cat</th>\n      <th>s_cat</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1980-01-31</th>\n      <td> 0.1</td>\n      <td>-4.684616</td>\n      <td> 4.784616</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>  0</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>1980-02-29</th>\n      <td> NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>1980-03-31</th>\n      <td> NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>1980-04-30</th>\n      <td> NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>1980-05-31</th>\n      <td> NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            msla     trend   detrend  seas_msla_detrend  seas_msla m_cat s_cat\n1980-01-31   0.1 -4.684616  4.784616                NaN        NaN     0   NaN\n1980-02-29   NaN       NaN       NaN                NaN        NaN   NaN   NaN\n1980-03-31   NaN       NaN       NaN                NaN        NaN   NaN   NaN\n1980-04-30   NaN       NaN       NaN                NaN        NaN   NaN   NaN\n1980-05-31   NaN       NaN       NaN                NaN        NaN   NaN   NaN"
     },
     "metadata": {},
     "execution_count": 15
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla.tail()",
   "execution_count": 16,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>msla</th>\n      <th>trend</th>\n      <th>detrend</th>\n      <th>seas_msla_detrend</th>\n      <th>seas_msla</th>\n      <th>m_cat</th>\n      <th>s_cat</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2014-08-31</th>\n      <td> 10.2</td>\n      <td> 4.531784</td>\n      <td> 5.668216</td>\n      <td> 4.690424</td>\n      <td> 9.200000</td>\n      <td>  2</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-09-30</th>\n      <td>  9.0</td>\n      <td> 4.553992</td>\n      <td> 4.446008</td>\n      <td> 5.301549</td>\n      <td> 9.833333</td>\n      <td>  1</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-10-31</th>\n      <td>  NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td>  7.1</td>\n      <td> 4.598409</td>\n      <td> 2.501591</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>  1</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td>  4.6</td>\n      <td> 4.620617</td>\n      <td>-0.020617</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>  0</td>\n      <td>NaN</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            msla     trend   detrend  seas_msla_detrend  seas_msla m_cat s_cat\n2014-08-31  10.2  4.531784  5.668216           4.690424   9.200000     2     1\n2014-09-30   9.0  4.553992  4.446008           5.301549   9.833333     1     1\n2014-10-31   NaN       NaN       NaN                NaN        NaN   NaN   NaN\n2014-11-30   7.1  4.598409  2.501591                NaN        NaN     1   NaN\n2014-12-31   4.6  4.620617 -0.020617                NaN        NaN     0   NaN"
     },
     "metadata": {},
     "execution_count": 16
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla.ix['2012':].plot()",
   "execution_count": 17,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x10e1f2450>"
     },
     "metadata": {},
     "execution_count": 17
    },
    {
     "output_type": "display_data",
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeMAAAFgCAYAAACIUxFOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XlgnHWd+PH3M3cmM7mPtklzt9Om90EPzgJKRXF1ucR7\nXRYVQfAHuJ4IKIIoIojAgsCiLCqIiIAIcrcc9qD3NW2TtM19TjJX5n5+f0ySps1MMkkmV/N5/WOc\neY7vhKf5zPf6fBRVVRFCCCHExNFMdAOEEEKI6U6CsRBCCDHBJBgLIYQQE0yCsRBCCDHBJBgLIYQQ\nE0yCsRBCCDHBdCM5yWaz6YHHgWLACNwO1AEvAQd7DnvIbrc/k4xGCiGEEKeyEQVj4PNAq91u/6LN\nZssEdgK3Ab+02+33JK11QgghxDQw0mD8Z+DZnp81QBBYAdhsNtungEPAt+x2u3v0TRRCCCFObcpo\nMnDZbDYr8DfgEcAE7LTb7dttNtv3gUy73f7t5DRTCCGEOHWNtGeMzWabDTwHPGC32/9ks9nS7XZ7\nV8/bzwO/Hux8VVVVRVFGenshhBBiKooZ+Ea6gCsf+CfwDbvd/lbPy6/YbLbr7Hb7FuB8YOugrVEU\nWltdI7m9EAnJzbXKMybGlDxjYrhyc60xXx9pz/j7QDrwI5vN9qOe174F/MpmswWBRuCrI7y2EEII\nMa2Mas54lFT5RinGkvRaxFiTZ0wMV26uNeYwtST9EEIIISaYBGMhhBBigkkwFkIIISaYBGMhhBBi\ngkkwFkIIISaYBGMhhBBigkkwHqVLL/0kwWBwopshhBBiCpNgPEqS0lMIIcRojTg39Vh75s3DbDnQ\nktRrnjYvj8vPq4j7/ssvv8h7720gEAjQ3t7GZZd9lo0b36G6uoprr72eDRvepr6+Dr/fz2WXXcH6\n9R/vO7e6+jC/+c29hMMRuro6uemm77Jw4eKktl8IIcSpadIG44nS3e3jnnvu5403/snTT/+BRx55\ngm3btvK73z1Gc3MTDz/8vwBs3vyvvnNUVaWmpoZrr/0WZWUVvPbaK/z97y9KMBZCCJGQSRuMLz+v\nYtBe7FhQFIU5c+YCkJpqoaSkFACr1YpOp+eb37yBu+76KR6Ph/XrLzzhvJycXJ544jGMRiNer4fU\nVMu4tl0IIcTUNWmD8USJNwesqip2+37uuOMX+P1+Lrnkor5halVVue++u7nlltspLi7hsccepqmp\ncTybLYQQYgqTYHyS3mB8YlCO/tzR0c7VV/8nGo2Wz33ui2i12r731q+/kJtv/g55efnMm1dJe3vb\nOLdcCCHEVCVVm8QpSyrqiLEmz5gYLqnaJIQQQkxSEoyFEEKICSbBWAghhJhgEoyFEEKICSbBWAgh\nhJhgEoyFEEKICSbBuJ9AIMBLLz2ftOvdc89dbN/+YdKuJ4QQ4tQkwbif9vY2Xnzxb0m7nlR0EkII\nkYhJm4HrucMvsb1ld1KvuSxvERdXXBT3/d///nGOHKnm7LNXsXLlKrq7vXz3uz9iy5Z/8frr/0RR\n4PzzL+DSS6/gpz+9FYPBQGNjI+3tbfzgB7cwd+48nn/+WV544a9kZGTh83Wzbt35Sf0MQgghTj3S\nM+7ny1++kpKSMr7ylasoKSnloYceJxKJ8Oabr/PQQ4/xm9/8lo0b3+HYsaMoisKMGbO45577ufTS\nz/DCC3/F4XDwzDN/5JFHfsfdd9+HoijSOxZCCDGkSdszvrjiokF7sWOhNzWoqqrMnl0MQHV1FU1N\njVx33dcBcLtd1NXVAjB3rg2AvLx8du/eSX19LcXFpeh00V/rokVLmMB0o0IIIaYI6Rn3o9FoiEQi\nwPH53uLiEkpLy7n//oe5//6HWb/+45SXn1jasTfgFhYWUVNTjd/vQ1VV9u/fKz1jIYQQQ5q0PeOJ\nkJmZRSgUJBAI9AXRioo5rFhxGldffSWBQIAFCxaSm5sHDKzwlJGRwZe//J9cffV/kZaWhlYrv14h\nhBBDk6pN4pQlFXXEWJNnTAyXVG0SQgghJikJxkIIIcQEk2AshBBCTDAJxkIIIcQEk2AshBBCTDAJ\nxkIIIcQEk2A8iC996TNx3/vLX55O6r3+8Icn+cc/XkrqNYUQQkwNEoxH6Pe/fzyp15NMXUIIMX2N\nKEWUzWbTA48DxYARuB3YDzwBRIA9wDV2u33EGUVa//wnXFu3jPT0mKwrTyP3sivivu/z+fjxj2+m\nq6uTgoJCIpEI1dWHuffeu1FVlfT0dL73vR/x7LNP43Q6ueeeu5g/fwEvvfQ3VFXlyiu/RldXF888\n8wc0Gg2LFy/l61+/lscee5impkYcjg6ampq47robWLVqDRs2vM0TTzxKeno6iqLw0Y9+LKmfVwgh\nxNQw0p7x54FWu91+NvAx4AHgl8D3e15TgE8lp4nj5/nnn6W0tIwHHvgtX/jClwkGg9x110+54Ybv\ncP/9D7Nmzek89dTv+fKXryQtLY0bbvgOqqqSlpbOgw8+ypw5c3n88Ue4776HePDBR2ltbWHLlk0o\nioLBYODuu3/N9dffyNNP/4FwOMz99/+Ke+99kF/96gHS0zMm+uMLIYSYICNNnvxn4NmenzVAEFhu\nt9s39Lz2D+AC4PmRNiz3sisG7cWOhWPHjrJ27RkAFBWVkJ6ewdGjNfzylz8DIBQKMXt20YDzel+r\nq6uls9PBTTddB4DX66W+vg6AOXPmAtEKT4GAH4fDgcWSSlpaGgBLliwb2w8nhJhwje0emjq82GZn\nYDbpJ7o5YhIZUTC22+0eAJvNZiUamH8I3N3vEDeQPurWjbOSkjJ2797JWWeto76+jq6uToqKSvjh\nD28jP38GO3Zsw+l0AtA/pbdGEx1gmDmzgLy8fO6990G0Wi0vvfQ35s2rZMOGt4gOFhyXmZmJ2+3B\n4eggMzOLffv2sGzZivH6qEKICfDbF/dxpMmFokDZrDQWlmazoDSL0plWtBpZwjOdjbiskM1mmw08\nBzxgt9v/aLPZft7vbSvQOdQ1cnOtI739mLjqqv/ge9/7Htdd91UKCgrIzs7i1ltv5Wc/+zHhcBhF\nUbjjjjvIzbUyZ04FP//5jzn99NNJTTWSm2slN9fKVVddybe+9XUikQiFhYV89rOX8uGH75OWlkJu\nrhWn04zBoGPGjAx+8pPb+O53/x9paWmYzea+Y0TyyO9TjLXhPGOtXT6sZgOFeRbsxxxU1Tv527s1\npKboWTInh2Vz81huyyMvyzyGLRaT0YiqNtlstnzgbeAbdrv9rZ7XXgB+abfb37HZbP8DvGG32/88\nyGWkapMYU1JRR4y14Txj/kCYq+95hwWlWdz4maV4fUH2H3Wwt6aDPTUdtHX5+o7NzzKzsCSLBWVZ\nzCvKwGSQcqyninhVm0b6X/j7RIehf2Sz2X7U89r1wK9tNpsB2MfxOWUhhJj2HG4/AJkWIwBmk54V\ntjxW2PJQVZUWRzd7ajrYW9PB/mMO3thWxxvb6tBqFK44fw7nryicyOaLMTbSOePriQbfk60bVWuE\nEOIU5XBFg3GG1TjgPUVRyM8yk59l5vwVhYTCEarqu9hT08Fb2+r564Zqzlo8E4NeO97NFuNEVgwI\nIcQ46OwJxlkxgvHJdFoNtqJMLjmnnPNWFOD1h9hyoGWsmygmkARjIYQYB73D1LF6xoM5e8ksFODt\nHfVj0CoxWUgwFkKIceBwnjhnnKic9BQWlmVTVe+ktsU9Fk0Tk4AEYyGEGAd9C7iG2TMGWLdsFiC9\n41OZBOMp5NJLP0kwGBzVNQarRAVSjUqIseJw+dFqFCzm4WfeWlyeTabVyAd7mvAFQmPQOjHRJBhP\nIeNR2UmqUQkxNjrdfjKtRjQj+Deh1Wg4e8ksfIEwm/fLQq5T0aTdSf7+m1VUJ3n1YNm8PE4/rzzu\n+8eOHeXOO29Dq9Whqiq33HI7zz33Z3bt2kEkEuEzn/kc5577EbZv/5AnnniUSCRCd3c3t9xyO3l5\n+fzoR9/F4/Hg9/v46le/wWmnrYl5n5dffpH33ttAIBCgvb2Nyy77LBs3vkN1dRXXXns9Z555Dnfc\ncRv19XX4/X4uu+wK1q//eN/51dWH+c1v7iUcjtDV1clNN32XhQsXx7xXrEpUAFVVh7nvPqlGJcR4\niERUutwBygrSRnyNsxbP5IX3anhrez1nL5mVxNaJyWDSBuOJsHXrZiorF3H11d9k164dbNz4No2N\nDTz44KP4/X6+/vWvcNppazhypIabb/4JOTk5PPnk//LWW69z1lnrcDq7+OUv78fhcHDs2NFB79Xd\n7eOee+7njTf+ydNP/4FHHnmCbdu28uc//4nly09j587tPPLIEwBs3vyvvvNUVaWmpoZrr/0WZWUV\nvPbaK/z97y/GDca9laiuuupqjh07wre//S0A7rrrdn7wg1spLi7hpZee56mnfs9Xv/oNnnvuGW64\n4Tu8/PKLpKWlc+edd+N0dvGNb1zFY489idFo5Cc/+dGAalRbtmziT396ihUrTuP++3/FY489SVpa\nGrfd9sPk/McRYgrr8gSIqOqwF2/1l5VmYmlFDtsPtXGkyUnJjNiBPRIJ4m7dQmr2UrQ6Sas5VUza\nYHz6eeWD9mLHwkUXfYqnnvodN954HRZLKnPm2LDbD/DNb34NgHA4TGNjAzk5Odx77y8wm820traw\nePFSSkvL+Ld/u5hbb/0BoVCISy+NX3FKUZS+Kk6pqRZKSkoBsFqtBAIBzGYz1113I3fd9VM8Hg/r\n1194wrk5Obk88cRjGI1GvF4PqamWuPc6uRJVRkYmAEeP1nD33XcCUo1KiLHWm/BjJIu3+jtnaQHb\nD7Xx9vYG/uPC2MHY1bKJrsY3CQU6yZr98ZjHiMln0gbjibBx4zssWbKMr3zlKl577RUeeeQhVq1a\nzbe//X1CoRBPPvm/FBQUcsMN1/LMM38jJSWFn/70ViKRCNXVh/F6vfz85/fS1tbG1Vdfyemnnxn3\nXoPNpba3t2G37+eOO36B3+/nkksu6humVlWV++67m1tuuZ3i4pK+oeJ4YlWigmhgvvnmH5OXly/V\nqIQYY33Zt0bRMwZYWJpFdpqJTfua+cx5FaQYT/wTrkbCuFs3A+Dp2EXGrPPRaEd3TzE+JBj3M2/e\nfH7601vR6/VEIhF++tO7ePXVf3DNNVfR3e3l7LPPxWw2c8EFF3LNNf9FTk4uRUUltLe3UVhYxOOP\n/5a33nqdSCTCVVd9fdB79Qbjk4OyokB2dg4dHe1cffV/otFo+dznvohWq6U38K1ffyE33/wd8vLy\nmTevkvb2trj3+fSnL+HOO3/M1VdfycyZs7Bao9+mb7rpe/zkJz/qq0b1ve9FU4yXlJTyk5/czMqV\nq/valpmZyRVXfJ5rr72KcDjCzJmz+OhH1w9ov6IoaLVabrrpu9x00/VYLFbM5pRh/BcQ4tTUOYpt\nTf1pNArnLJ3Fcxuq+dfeJs5dfmK+ao9jD+GQG40ulUjIg6djF9bc00Z1TzE+RlS1KUmkapMYU1K1\nSYy1RJ+xZ9+u4uV/HeW7n1/O3NkZo7pnl9vPTQ++z8zsVG77z9P6vhCrqkrTgYcJ+lrJn/ufNB/6\nX/TGbGbM+7rsaphEkl21SSTgl7+8iyNHqge8fvfdv8ZoTO7Q0XjeSwgxPMmaMwZItxhZNieHrfZW\nqhuclBekA+BzVRP0tWDOXIgxtQBzRiVexx787qOYrCWjvq8YWxKMx9CNN37nlLyXEGJ4HK5oreLR\nzhn3OmdZAVvtrby9vb4vGLtaPgAgLS+6pdKSsxKvYw/utq0SjKcASfohhBBjzOEOYEnRo9cl50/u\n/OJM8jJS2HygBY8vSKC7GZ+rGqOlGIM5ugfZmDobvSkPb+cBQkGZrpnsJBgLIcQYUlWVTpc/KUPU\nvTSKwjnLZhEMRXh/dxOulmguAmve8URDiqJgyT0NiOBp25a0e4uxIcFYCCHGULc/jD8YTmowBjhj\n0Ux0WoXNe6vwOHajM2aTkjb3hGNSMxehaIy427ehquGk3l8klwRjIYQYQ6Op1jSYNLOBFbY8SqzV\noEZIy1szYNW0RmsgNXsJ4aCL7q6DSb2/SC4JxlNIMqo2CSHGV6drZHWME7FuSS4rZzfhDxsxZ8VO\niWvNiSbdcbVuSfr9RfJIMJ5CZK+gEFNPR+9K6iT3jAFmphwhRR9i09F8PL7YOSP0plyMlhL87iME\nfa1Jb4NIjkm7tclR/xrezn1JvaY5o5LMgo/Gff9UrNokhJhYnUncY9yfqkZwt24iomrZdHQGObub\n+NjqgTnmAay5p+F3H8HV9iFZhVJFbTKatMF4IpyKVZuEEBPL4Q4AyR+m7u6yEwo4SMlcij9s4p0d\n9VywanbMeskp6XPR6q142neSMfM8NFpDUtsiRm/SBuPMgo8O2osdC6di1SYhxMTq6xmnJTcYO3uS\nfGTNPJ1V81t4f08TB446qCzJGnCsomixZC+nq+kdPI7dffPIYvKYtMF4IpyKVZuEEBPL4fJj0Gkw\nG5P359bvqSXgqcOUNge9KYd1S/W8v6eJt3c0xAzGAKk5y+lq2oi7dSuW7OWyBmWSkWDcz6lYtUkI\nMbEcLh8ZVmNSg5+zJ8lHWt5aAMoL0ijMTWX7wVa63H7SYwyJ6/RWzBnz8Hbuw++pxWSJPb8sJoZU\nbRKnLKnaJMbaUM9YKBzhq794G9vsDL7z+eVJuWfQ30HjvgfQp8xghu2/+oL8Gx/W8dRrB7nknDI+\nsbYk5rk+1xFaDv8ec+ZCckouTkp7xPBI1aYJIJWUhJjeklXHuD9XyyZAHZDkY+2CGfz57cO8s6OB\nC9cUx1zIZbQUozfl4u3cRzh4AVq9rDeZLCQYjyGppCTE9Nbp6llJnaRgHA514+nYgVafhjmz8oT3\nzCYdq+fns3FXI3trOlhUlj3gfEVRsOSsxFH3D9zt20mfcVZS2iVGT5J+CCHEGOlNhZmshB/utg9R\nI0GsuatRFO2A99ctKwDg7e31ca+RmrUYRWOIXkuNJKVdYvQkGAshxBhxJDEVphoJ427djKIxYMlZ\nFvOYkhlWivOt7DzcTofTF/MYjdZIatZiwkEn3V2HRt0ukRwSjIUQYow4elJhJmOY2uPYQzjkxpK9\nHI3WFPMYRVE4Z+ksIqrKlgMtca9lyVkJgLtN8lVPFhKMhRBijDiSlApTVVVcLR8ACta81YMeu7As\nus/4cF1X3GMMKXkYLUX4XNUEfe2japtIDgnG46i6+jA7d26f6GYIIcZJp8uPokBa6ujST0aDZgvm\nzAXoDOmDHpudZiLdYuBwQxeDbV215JwGgLtt66jaJpJDgvE4euutN6ipGbjVSQhxanK4/aSlGtBp\no39qO155mbpf/oKOV17GX18/aLDsz9WT+jItL3bxmf4URaGiIJ0ud4D2OPPGAOb0eWh0qbg7dhKJ\nSGnWiTZptzb9o7aV3R3upF5zUZaFC2fnxn0/kWpKJ/P7fdxxx200NzcTDAb5f//vvyktLeVnP7sd\nj8dNW1srF198GWeeeQ7/+MdLGAwG5s2bz7x5lTFaIIQ4VaiqisMVoDA3te+1zrfeINTejnf/Xtqe\nfQZddjapi5aQumgx5nnz0cTIPxDobsbnqsZoKcZgnpXQvctnpfOhvZXD9V3kpKfEPEbRaLHkLMfZ\ntBGvYw+W7NiLwsT4mLTBeKIMVk0pVjB+/vm/MGtWIbfddid1dbW8//67GAx6PvKR9Zxzzrm0tbVy\n7bVf49OfvpSPf/yTZGfnSCAWYhrw+EKEwpG++WI1EiHU2YmhcDZZ6z+GZ9dOPHv30PX2m3S9/SaK\nTkfKvPmkLlpM6qIlGPLyAHD1pL60JtAr7lVREB3Krqp3sqZyRtzjLNkrcDa9i6t1C6lZSyVf9QQa\nVTC22WyrgZ/Z7fZzbTbbMuBFoHet/EN2u/2ZkV77wtm5g/Zix8JQ1ZRiqa09xpo1pwNQWDibyy//\nLK2tLTzzzB/ZsOFNzGYL4XC47/gJTD8qhBhHvVuLevcYh10uCIcx5OWRtvYM0taegRoO46uuwrN7\nF+5dO/Hu2Y13z25a//gU+hkzMC9ZiL+iBp0xi5S0uQnfu3iGBZ1Woao+/iIuAJ0hjZR0G91dBwh4\n6zGmFo78A4tRGXEwttls/w18AegdS14B3GO32+9JRsMmynC/GRYXl7J//z7OPPMc6uvrePzxh8nI\nyGLhwkV8+tOXsm3bVj744F0ANBqNBGMhpom+VJg9e4xDnQ4AdJnHqyopWi0pc+aSMmcuORdfSrCj\nA8+eXXh278K7by+uPe9iqJiFWhtEqUz8b5Nep6U438qRJhf+YBijfmCCkF7W3JV0dx3A1bpVgvEE\nGk3P+DBwMfBkz/9fAcy12WyfIto7/pbdbk/upO84GKyaUiyf+tTF3Hnnj7n22q8SiUS4/vqb8Hjc\n3HvvL9iw4W1KS8swm80Eg0Fstnk88MCvKSkpZdkyqScqxKns5G1NIUdPMM7IjHuOPiuLjLPXkXH2\nOiLBIB37X8Qb3kP39ipCa53o0tISvn95QTpVDU6ONDqxFcW/p9FSis6YjbdzL+HQBWh15oTvIZJn\nxMHYbrc/Z7PZSvq9tAl4xG63b7fZbN8HbgG+Pcr2jasLL7yo7+fVq9eyenW0PNmcOTbuvvvXMc8x\nGAzccsvtA17//e+fHvDa2rVnsnZt/BrHQohTx4Bg3NczzkjofI1ej2oJQRdEGrtxvreRrAs/kfD9\nywvSYUstVQ2DB+NovurldNa/htexD2vuyoTvIZInmQu4/mq323snKJ4HYkevfnJzrUm8/di69dZb\nqaqqGvD6o48+KhWYJrGp9IyJqSneM+YLRaekyoqyyM214g14AcgpKSA9weeycV8jOr0FTVCL+70N\nzP3C5SiaxHakrtLreOj5PdS2eob8d5BuXU1n/esE3fvJrTw3oeuL5EpmMH7FZrNdZ7fbtwDnA0Pu\nJJ9KtWavuebGmK87nQEg9uIuMbGknrEYa4M9Yw09r6vBEK2tLpx1TQC4FROBBJ7LUMBJ0N9FSroN\ny8o0nO+/y9GNm0mtXJBw+7LSjOyraaelxTnEehgNRksR7s4aGuvr0RkSHw4XwxPvi1Eykn70rkj6\nOvArm832FrAWGDh2K4QQ00Sny4/JoCXFGO3zHB+mjj9k3F/A2wCAwTyL9HPWAdD1zlvDakP5rHRc\n3iCtnd1DHmvOjAZ5r2PvsO4hkmNUPWO73X4EOL3n552ATIgKIQTROeP+OalDnQ405lQ0hsRSYwa8\n0TKIRnMBxvxSDAWFuHdsJ9TViS49sXnnioJ0thxooareSV7m4AuzzBmVOGpfwevYQ1r+2oSuL5JH\n0mEKIUSSBYJhPL7QicHY4Ui4Vwwn9owVRSHjnHUQDuN8792Er1Hek/zj8BD7jQG0OjOmtDIC3Y1S\nPGICSDAWQogkO3mPccTnI9LdjS4jsR6tqqr4vQ3ojNlodNFyidY1p6MYDHRteAc1EknoOkX5FvQ6\nzZDJP3qZM3qGqjtlqHq8STAeY++88xZtbW0T3QwhxDjq3daUMWBbU1bcc/oL+dtRw34M5oK+17Rm\nM9ZVqwm2teLdvy/uuW6Xn388u5v2Fjc6rYbiGVZqW934AqEh72vOmIei6PA49kiConEmwXiMPfvs\nn/B6p1zuEyHEKMRN+DHMxVvG1BMLQ6SfHd12NNhCrm0fHOXI4Xa2vHsEiM4bqyrUNDiHvK9Ga8SU\nPoeQr42gryWhtorkmLSFIp558zBbDiT3YThtXh6Xn1cR9/1jx45y5523odXqUFWVW265nby8/AHH\n1dYe4667bicUCmE0mrjttjvo6GjjN7+5l3A4QldXJzfd9F2cTieHDh3k9ttv5cEHH0Wnm7S/biFE\nEjnccRJ+DJJ9qz+/J7p4q3/PGMBUWopx9uzoQq7OzgHD3t3eAPZd0S1URw614Xb6KJ/VM2/c4GR+\nydA989SMBXR37sfr2IMhZeDfPzE2pGfcz9atm6msXMS99z7IlVd+Dbc7do/2gQfu5Utf+k/+538e\n57LLruDQoQPU1NRw7bXf4r77HuTzn/8Sf//7i5x++pnMmTOXH/7wNgnEQkwj8XvGic0ZB7z1oGgG\nBENFUaK940iErnc3DDhv7/YGQqEIeTOtqCrs29FIRUF0z3Ci88am9DkoGgMex14Zqh5HkzZCXH5e\nxaC92LFw0UWf4qmnfseNN16HxZLK1752TczjamuPsXDhIgDOPPNsAHbu3METTzyG0WjE6/WQmmoZ\nt3YLISaXTteJC7iCPcFYn8CcsRoJEehuxpAyA0Uz8E+0dc1aWp99mq6N75D18Yv6MnKFQmF2f1iP\nwajjwksX8cdHNrF/ZyMrzigmJ91EVX0XqqoOWQxHo9GTkj4Pr2OXVHIaR9Iz7mfjxndYsmQZ9933\nIOvWnc///d/vYh5XXFzKvn3R1YavvfYKf/nLM9x3391ceeXX+MEPbqWsrKLvG6VGoyGS4MpHIcSp\nweH2o9UoWFOje4qHM0wd6G4GNTxgiLqXNiUF66rVhNrb8e7b0/f6wT3N+LxBFiybhTnVgG3RDLye\nAEcOtVFRkI7HF6Kpw5tQ+1N7EoB4HHuGOFIkiwTjfubNm8+jj/4P119/NS+88Fcuu+yKmMddc831\nPPnkE3zzm1/jtdde4YILLmT9+gu5+ebvcNNN16GqKu3t0RXUCxcu5vbbb8HlkrSMQkwXDpefdIsB\nTU8vNORwoOh0aCxDj5j1318cT8bZ6wDofOdtILoVaufmWjQahUUrokF8wbLo+Xu2NfTtN66qH3oR\nF4AprQyNNgWvYx+qKp2J8TBph6knQkFBIQ8++GhCx91334MnvPaZz3yez3zm8wOOveqqq7nqqquT\n1kYhxOQWUVW63AFKZhzPQRzqdKDLyEyoXnpf5q3U2D1jAGNJKcaiYjw7dxDqdFDXGqazoxvbohmk\n9sxTZ2anUlCcQf3RTs5cPhOAqoYuzlw8c8g2KIoWc0Yl7vYP8buPYrKWDnmOGB0JxnEEg0FuuOHa\nAa8XFRXz7W9/fwJaJISYClyeAOGI2rd4Sw2HCXd1YaiYk9D5fk8DisaIzpgd9xhFUUg/Zx0tT/6O\nro0b2OkjFCOuAAAgAElEQVQsAmDJqhPndxcsK6D+aCeOY10Y9JqEMnH1MmcuwN3+IR7HHgnG40CC\ncRx6vZ777394opshhJhierc19SX86OoCVU1oj3Ek7CPkb8NoKR2yF522eg2tzzzNsfd30pCezuyy\nLLJzTxwGL5mTTarFwKG9zZTmWTlY34XXF8JsGvpPv9FShFZvpbtzP2rhx1E02iHPESMnc8ZCCJFE\nA7Y1DWfxVpxkH7FoTCmkrV5DtWY2AEtXzR5wjFarYf6SmQT8YWYZdKhATWNi88aKosGcUUkk7MPn\nGljLXSSXBGMhhEiik7c19e0xTiAY+z29i7fizxf3p1l+Bi2WYtI03RQUx97DPH/pLBQFIh3RMoqJ\n7jcGMGcuBGRV9XiQYCyEEEnUEa9nnMAwde/iLcMgi7f6szcBiobZjVv6gv7JLFYjpXNz8Hb5SCWx\nCk69DOZZ6AyZdHfZiUSCCZ8nhk+CsRBCJFHnKPJSB7wNaPVWdHrrkMf6fUH272zEbIA8Vw3OGBm5\nei1YFg3uRQYdVQ1OIglm1lIUBXPmAtRIkO6ugwmdI0ZGgrEQQiRR3wKuk4ephwjGoYCTcNCV8BD1\n3u0NhIIRFq8uQms00rVxQ9zSigXFGWRkpWAJRgj6QzS2J5b8A44PVXtlqHpMSTAWQogkcrj8pJp0\nGPTR1cd9w9Tpg+elTiTZR69wKNKT+lLLgpVFpK1ZQ8jRgWf3rpjHK4oS7R2rkMPw5o0NKXnoTXl0\nOw8TCfkSPk8MjwRjIYRIok63v2+IGqLBWGtNQxmiWEwiyT56HdrXjNcdYP6SWRiMOtLPGbq0om1R\nPhqtQi4Kh+o6E/kofcyZC0AN4+06MKzzROIkGAshRJJ0+0N0+8N9e4xVVSXkcCQ0X3y8bOLgGbJU\nVWVHT+rLxSujgdtUVIyxpBTP7l0EO9pjnmc06ZlbmY8JhYYjsRd7xWPuyVUtQ9VjR4KxEEIkSaf7\nxG1NEa8XNRAYMhirqkrA24DOmINGaxr02NqaDhxtXirm52FJO35sxjnrQFXp2hh/IdfCnrzVOlcA\njy/x1dF6YxYG8yx8rhrCQU/C54nESTAWQogkGWnCj5C/DTUSSCjZx45NtcDA1JfWVWvQmEw4392A\nGg7HPDd3hhW9xUA6sPdg25D36i+6kEvF27lvWOeJxEgwFkKIJBkQjBNcSZ1oso/WJhf1RzspLMkk\nJ//E7U8aoxHrmtMJORxxF3IBlMzPRUHhwM7GwT/MSWSoemxJMBZCiCTpG6YeZs+4L9nHECupd27p\n7RUPTH0JPUPVDL6Qa+Wq2YRQcTe6CIcSL4+o01sxWorxe2oJBRJfjS0SI8FYCCGSpLdnPNw9xgFv\nAyhaDCn5cY9xO30c3tdCVm4qs0tjX884uwhTWRmePbsJtscehs6wmug26VAiKof3twz5mfo7vud4\n77DOE0OTYCyEEEkykmFqNRIi0N2EISUfRRN/+9OurXWoarRXPFhFp/Sz14Gq4tqyOe4x2UXRPc/b\nt9TFPSYWc8Z8QINHgnHSSTAWQogkcbj86LQaLCl6ILFh6kB3E6iRQeeL/b4Q+3Y0YrYYmFOZN2gb\nzJXRud3uqsNxj6koy6ILFUeLm7Zm96DX60+rM2NKKyPY3UjQF3sLlRgZCcZCCJEkDrefTKuhr+ca\ncjhQjEY0KSlxzzleNjF+MN6/s4FgIMyiFQVotYP/2dZnZaPLzMJXdRg1Tg7qioJ0Woi+t3d7/aDX\nO1mqpMccExKMhRAiCcKRCE5PoG+PMdCX8GOwYeXjyT5iL94KhyPs2lqHTq9hwbKhtz4BmMrLCTud\nhNpizxvPzEnFb9AS0sDBvc0E/KGErguQkm5DUXR4HHvjBnsxfBKMhRAiCbrcAVSVvuxbkWCQsNuV\nwErqBhStEZ0xO+b7Vftb8LgCzF8yE6NJn1BbUsoqAOiujj1UrVEUygrSaYpECAUjHNzTnNB1ATRa\nI6b0OYT8bQS7Ez9PDE6CsRBCJMHJi7fCndH8z4Mt3oqEugn52zGaZ8XsPfemvlQUWLyyMMYVYjOV\nlwPgG2TeuHxWGq1Ei0js2V4/rF6uDFUnnwRjIYRIguPBOJqiMpHFW37v4Mk+nJ0+2ls8FFdkk5YR\nf975ZMaiYhSdju6qqrjHVBSmEwIMWSk42rw01ia+d9iUVoGiMeBx7CYS9id8nohPgrEQQiSB4+SE\nHwlsawoMEYxbGp0AzCoavPziyTR6PcbiEvy1x4j4YwfLspnpKIAjWumRPdsSX8il0eix5q4iHHTR\ncexFmTtOAgnGQgiRBJ2uE4tEBB0dAOgHDca9ZRNjL8xqaXABkD8zbdjtSSkrh0gE35GamO+bTTpm\n5aRyqMNLZo6ZmoNteN2J93LTZ67DmDobb+c+3G1bh90+cSIJxv1EVJVmh3eimyGEmIJ6e8YZVgMA\nod454zjD1Kqq4vc0oNWnodVbYx7T0uhEUSAn3zLs9pjKo4u4Bp03LkgnEIowsyybSESl6kBrwtdX\nFA3ZJZeg0Zlx1P+TgHd4ua7FiSQY9/P+7ia+9/C/eG5D/HkWIYSIpXOYqTDDQReRkHvQLU2tzW6y\n8yzo9Npht6c3GHdXx/97Vl4Q7XH7TFoUhWGnx9QZ0sgu/jSoYdpqniUS9g27nSIqfu61BNhsttXA\nz+x2+7k2m60CeAKIAHuAa+x2+5SaSNhqjz6IL71/lDSzgY+sjJ2MXQghTtbh8pNm1qPrScoR6nSA\nRoM2LT3m8ceHqGPPF3e0egiHIuTNjN1rHoo+MxNd1vHkH7FWa1cURNt2tN3DrKIM6o924uryYU0f\nvKZyfylpFaTln4Gz+T3aj71ITsmlcfdVf2hvod3pJ9WkI9WkJzWl539NOswmPXrd9O0fjjgY22y2\n/wa+APTmUrsH+L7dbt9gs9keAj4FPD/6Jo6PQDDMgaMOstNMBMMR/vj6IdJSDayaHz9xuxBCQHTI\nudPlZ2Z2at9rIUcHuvQMFE3sABMYItlH7+KtvBHMF/cylVXg3rqZYGsrhryBaTTzs8ykmnRU1Xdx\n5poS6o92cnh/C8vWFA3rPukzz8XvqaW7cz/uti1Yc1cNOOZYs4sH/jr4ViijXou5J1BbUqIBOi3V\nwEdXFp7wuz0VjaZnfBi4GHiy5/8vt9vtG3p+/gdwAVMoGB841kkgFOG0eXmsWZDPXX/Yxm9f3Eeq\nSc+C0qyJbp4QYhLz+kMEQpG+ldRqJEKosxNTcXHcc45vaxp88VberJH1jAFSystxb92Mr+pwzGCs\nURTKC9LZVdVOTmE6Go0yomDcO3/cdOBhHPX/xJBaiPGkz/Xq5mMAXHx2GWmpBjy+IJ7uEF5fELcv\nhKc7iNcXwuML0u70Udd6PCvYv/Y28fVPLWBxec4IfgtTw4iDsd1uf85ms5X0e6n/uIQbiD02M0nt\nroomPV9Unk1RvpVvXryYe57ZyW/+upv//uwySkfx7VQIcWrrK53Ym/DD7YZweJDFWxEC3gb0plw0\nWmPMY5obnegNWjJH0SM0lc8Bopm40taeHvOY8llp7Kpqp67DS2FpJseqOnC0e8nMNg/rXjq9lezi\nf6e16inaap5lpu2raHTR4e4Op4/N+1uYlZPKJ9YWD5oetFc4EsHrC7Grqp3fv2rnvj/v4rJzK1g/\nRNWqqWpUc8Yn6V+l2gp0DnVCbu7Iv/Elk6qq7DnSQYpRx+nLCtFpNeTmWtEadNz1+y38+i+7+Pm1\nZzErd/grGsXEmizPmDh15eZaqW3vBqAw30purhW3K7oq2TozP+Yz2O1uRo0ESMsqjvm+3xfC0e6l\nuCyb/PyRdwQiGZXUGwyEjtbE/bewonImf91YQ6Ojm+WrizlW1UHjsS7mzhvBFF3uUrRqE03Vb+Bu\nfpmyJV9CURRe+OAo4YjKZefPIS9veJ+nrDibyopcfvq/m3nmrcO0ufxcc+kSDHEWtQW6Hfi8raRl\nzx1++ydQMoPxdpvNdo7dbn8HuBB4Y6gTWltdSbz9yDW2e2ju8LLCloujw9P3+pyZVr5wgY3fv2rn\nBw+9x/e+sKJvGEpMfrm51knzjIlTU+8zdqQuunJar4n+XXNXR+eDg6bUmM+gu/0gABFNXsz36486\nQIXMHPOon2FDUTGe6iqaa1vRmAYuzMo061AU2HWolY8unYVWp2Hn1lrmL50xoh6oPm0tRsthOlv2\nUL3vDXRpK3jlgyOkWwxUzs4Y0efJTNHxgy+u4DfP7eLNrbUcbeji2osXkd6zcj26TawWV+smujsP\nACqzFlyHzjC8ZCnjId6XomQsXetdMX0jcJvNZnufaJB/NgnXHhe7eoaoF5cPTNS+blkBnz6zlLYu\nH796ZgdeX3C8myeEmOR6h6mzelNh9iT8iLet6XjZxHiLt3rmi5MwPZZSPnjyjxSjjsJcC0eaXGh0\nGorLs+hs99Le4ol5/FAURUNOycVodKl0NrzG5l078QXCfGRF4ahWS2dajXznc8tZU5lPVYOTH/9u\nK0cbHXg6dtFsf5SWQ0/Q3bkffUo+2cUXT8pAPJhR9YztdvsR4PSenw8B60bfpPHXF4zLYldN+eQZ\nJTi9Ad7cVs+vn93FDZ9ZGneIRAgx/RxP+NGzx3iIvNQBTz0oWvSm2EPBzQ3RldT5o1i81cvUW8Gp\n6jDmefNjHlNRmE5ti5udh9uomJ9Htb2Nw/tbRpRsBECrt5JT/O+0VP0fOeE3SE9Zzrpl8es1J8qg\n13LVJyspytXQXLcJz9ENKMYgoJCSPg9r3mqMqUVTck55+m7q6tHtD3GwtpPiGda+IY+TKYrC5z4y\nl5Xz8jhY18XDL+wlHInEPFYIMf04TkqFGXLEr9ikRkIEupsxpMxA0cT+Ut/S6MKcaiA1CdNiKQlk\n4vrIikK0GoVn3jrMrOIM9AYth/e3jCrntCmtDKdmOWkmH19acxSzcfSzogFvEx3HXqAy5TnOrTiG\nThPh/ZoC9ng+RXbpZZgsiS0Om4ymfTDed6SDcESN2yvupdEoXHVRJfOLM9l+qI0nX7VLcnQhBBDN\nvmXUa0kxRoPrYD3jQHcTEMEQJ9mHx+XH4/KTN9OalMCiy8hAl52Nr7o67t+smdmpnL+ikNZOH2/u\naKB0Tg6uLl/fcPlIqKrKn7fmUtORTrahHlfrphFeJ4K38wDNh35Hk/0RPB070RnSySy8kJTiq9nW\nPJ9n323job/uwR8Ij7i9E23aB+O+IeqKwYMxgF6n4dqLF1Gcb2XDzkb+urF6rJsnhJgCHG4/GVZj\nX/AMOTrQmFPRGAf2bHuTfRjjVmrq3V+cvO2UKeUVhN0ugi3NcY/5tzNKsJr1vPT+UWaURL9EHN43\nvPSY/e090kFtq4cq7xnR+eP61/F76uIer0bCBH3tdHcdwtW6GUfdq7RW/YmGvb+mreYZ/O6jmKxl\n5JZ9lpnzr8Gaexqz87O4+csrsc3O4MODrdzxfx/S1tU94jZPpGSupp5yVFVlV3U7lhQ9pTMSe/BT\njDq+dfkS7nzyQ0mbKYQgGIrg8gYpyOmffcuBLiv2F/whk300Jm++uJeprALX5k34qqow5M+IeYzZ\npOeSc8p54h8HeK+6DaNJx+EDLaw9rxyNZvg99Fc3RZN8nHfaXHJSs2k5/CRtNX8hp/QSwkEXIb+D\nkN9B0N9BKOAgHOji+Hrg4xStCUv2ciy5qzCkDExcYjUbuPGKpfzhtYO8vaOBn/xuK9devIg5hdNo\nAddUd6zZTZc7wNoFM4b1sKWnGrjhiqXc8eSHkjZTiGmus6+OcXQldcTvJ9LdjS4jdjAIeOtRtCZ0\nxtiZ/XoXb+XOSF4wTikvB6KLuNJOPyPucWcumslb2+r51/4WPl2RS/3hdhprOykojl8GMpZjzS72\nHnEwryiDkhlpQBrpM86hq+kdmg8+PuB4rc6CMbUQnTELnTETnSGz72eNNmXI4XqdVsMX19soyLXw\nx9cP8Ys/7uAX3zid9FTDsNo9kaZ1MN5V1QbE3tI0lLyMFG64fElf2sxMqzGhb2JBfwdex160egup\nWUtQlGk/UyDElNa3eMs6dLWmcKibkL8Dk7UsZoBRVZXWJhcZWSkYTfqktdE4uwhFr8dXHX8RF0TX\nxnzuo3O48/+2sbfdQwbRSk7DDca9qS8/tvp4Ws20GWehqiEiYR86Q0/Q7Qm8Gu3og6aiKJy/opCC\nnFR2HG4j1TS1wtu0jgS7qttRFFhYNrLc00X5Vq7590WoKjzw1z19/yhPFgl142r7kKaDj9O47zd0\nNb5Fx7EXaT70BEFf+2g+ghBigh3vGQ+9rSnQN0Qde764s8NLwB9Oyv7i/hSdDlNJKf66OiK+wedU\n5xRmsKYyn0MOLzqDlmp7K+Fw4rtH+qe+XNRvYayiaMiYdT5Zsz9BWv5azBnzMKTkJyUQ9zevOJMr\nzp/TVz1rqpharU0ilzdAdb2TioJ0UkfxDbSyJIvLz6vA6Qnwm+d2EQxFV/Opahhvl53Wmj9Tt+ce\nHLV/J+Cpw2QtJavo3zBnVBLw1NF04GGcLR+gqrJVSoipyDGgjnFvwo+BX/J7yyYa4iX7SEJxiHhM\n5RWgqvhqYif/6O/SdeUY9BpaIxF83aFoRrAEvb61jnBEPWVzSI+VqdWPT6I9NR2ojGyI+mQfXVlI\nbbOL9/Y08vwbH3B+pQtv514iIS8AelMuqVmLMWcuQmeIfuO1ZC/F69hHR93LdNa/hrdzP9lF/4be\ndOpWJRHiVBR/mHrgtFXA05N5K+5K6tGXTYwnpbwcBz3JP+ZXDnpsVpqJT6wp5rWNNWSi4fC+FoqG\n2P4J4PWFeHtHPekWA2sqYy8UE7FN22B8PAXm6INfOOjk4uXtnJa9gwyTB3cbaHRmrLmrSc1ajD4l\ndo5Xc2YlRksxjrpX8HbupfHAw2TMPBdr3hqZSxZiihgQjOMMU6tqBL/nGFpDBlp97MxWLY0uNFqF\nnLzkF6XpzcQ1WPKP/tavKmLjzgb8zgBV9jbO/lgYnW7wzIMbdjbgC4T5xNriUaW+nI6m5W8rElHZ\nU91OptVIYe7Iy5OpapiO2r/TsPc+XM1vkZHiw96axx+3LcBp+TKZhesxmGcOOlSj1aeSU3oJOaWX\nodGa6Gx4neaD/0vQ1zbidgkhxo/D7UejKH0rd+Nl3wp4G4iEfaRYy2JeJxyK0NbsJifPgnYMApku\nPR19Ti7d1VUJJSwy6LV85vw5dAChYJja6o5Bjw+FI7y2tRajXpuU1JfTzbQMxtUNTjy+EIvLs0c8\npxEJddNy+CncbR+iN+WSNfsiChfeSEnlFRxuz+Khv+2ntTPxzefmjPnMnH815syFBLz1NB54GGfz\ne5N+LtnrC+ILhIY+UIhTVKfLT7rF0Lc9MujoQNHp0FpOnPf1OasAMKWVx7xOW4ubSEQdkyHqXqby\nciIeD8HmpoSOXz43l/SZ0c+xZXPtoMdu3t+Mw+XnrCUzR7UOZ7qalsF45yi2NEF0e1LTwcfxu4+Q\nkm4jf+6VWHKWo9GZmFOYwecvmIvHF+L+v+weVno2rc5MTsnF5JRe3tNLfqOnl9w6onaOFVVVOVjb\nycMv7OX6X7/Lz57aJqlBxbSkqiqdbn/f4i2IDlPrMjIHfNHvdlUBCiZLacxrtfTsLx6LxVu9TOXH\ni0YkQlEULr/Qhg+V1jon3d2xq9apqsorm2rRKAoXSBKkEZmWwXh3VTs6rcL8Ye6dA/C5j9Jsf4yQ\nvx1r3uk9gfPEpfnrlhawblkBda1uHnt5/7ADlTlj3km95EcmRS+52x/irW11/OjxzfzsqW1s2teM\nokSTpxxrdk9o24SYCE5PgFBY7ZsvVsNhwl1dA4aoIyEfAU89htQCNLqBNYUhuWUT40npmzeuSvic\n2XlW0mZY0QAvvXYw5jH7jjioa3Wzcl4uORkpyWjqtDPtFnA5XH6OtbhZUJqFyTC8j+9u30lH7Yug\nQlbRJ7FkL4t77Oc+Mof6VjdbD7Twcr6FT6wtGda9envJ3oxKOmr/TmfDGwCk5cfPnjNW6lrcvLW9\nnvf3NuEPhNFqFE6bl8d5ywtwd4d44K+72bSvmeIkZgwSYipo65mK6lu85XSCqg7IvuVz1wAqKdbY\nQ9QAzY1ODEYtGVljF8yMhYUoBkPCPeNe6z8yhxf/bzvV+1vpOs8/oMLdK5uOAicm+RDDM+2C8e7q\nwWsXx6KqKl2Nb+FsfheN1kRO6WWYrLGHmnrptBq+8e+L+PETW3junWoKcy0sqRj+ym1zxjyMlmJc\nrZswpVUM+/yRCoYifHiwhbe31XOwrguI/sH5+Ooizloyq29YLhgKk2LUsvlAM5eeW45G9hWKaaTd\n6QNibWs6cY/xUPPFfl+Qro5uCksGDm8nU2/yj+5DBwl3d6NNSSzwFxamY7QYsLj9/PnNQ/zXvy3s\ne29g6ksxEtNumHrn4Z754gSqNAFEIkHajjyLs/lddMYs8udeOWQg7pWeauCblyxCp9PwyIt7aWz3\nDLu9DpefN7e38fjGDA40jP13p7aubv7yThXffvA9HnlhHwfrulhQmsU3L17Ez69eyyfPKD1hfkyv\n07J8bi4dTj+He4K2ENNFe1dPMD454Ue/bU2qqtLtqkKjNQ1SHGLskn2czFRW3pP8Y3hV55YsL0CD\nwsF9LVT3zG8DvNqzsEt6xaMzrXrGwVCEfUcd5GemkJ9pHvL4cNBNa/WfCHgbMFqKyCm9HK1u6PP6\nK5mRxn9cOI/fvriP+/+ymx9+aSXmIXKmurwBPrS3snl/M/ZjnaiAosDaMd5Ev2FnA7975QCqCqkm\nHetXzWbd0gLyswb/zKsr83lvdxOb9jczd/bUqpTS3+Mv70ejKPzHhfMmuiliimjvKdeXcfIe435z\nxiF/O+FAF+aMyrj5A8ZjvrhXSsUcHET3G6dWLkj4vDmVeWzeUEMWCn94/SDf/+IKOl1+Nu9vHpD6\nUgzftArGB+s68QfCLF489HBxoLuZ1qo/Eg46Sc1aQtbsi1A0g294j2ftghkca3bx6uZafvviXr55\n6eIBw7leX4jth1rZtL+ZfTUOIj2LvuYUprNqfj4r5+WNeQUSs1FHZUkWayrzOW1eHgZ9Yp93fnEm\nVrOerQda+NxH5qDVTL0Bl5pGJ+/uakQBLj6njDTz1Kn2IiZOR1ecYep+PWOfK9oDNcXZXwzHKzXl\nzRynnjGJr6julZaRQt4sK2qDix0NTj7Y00R9q0dSXybJtArGu/uybg3+Da676yBtR55DjQRIn3ke\naflnjPpBu3RdOXUtbnZWtfP8xmouPrscfyDMzqo2Nu1rZnd1O6FwNACXzLCyan4+q+bnkZUWe+Xl\nWFg5L4+V8wbWCx2KVqNh5bw83tpWz/6jDhaWTr1vyK/01F5Vgb3VHaxdKKn8JotOt59IRB3XfwuJ\nGjBM3TkwFWb3EPPFqqrS0ujEkmYk9aSFUWNBl5aGPjcPX3UVaiSCMowvzxXz82hpcJGj0fDs21X4\ng2FJfZkk0yoY76xqx6jXxh1KVVUVV+tmOuv/iaJoySm9DHPG/KTcW6vR8LVPLeQnv9vCS+8f5WiT\nG3utg0Awul2pIDe1LwAnMoQ+2ayen89b2+rZtK95ygXj1s5uttpbSE810OUJsLu6XYLxJKGqKr98\negdOT4CffW0tKcbJ9Serrasbs1GH0RAdRTq5Z6xGQvjdR9CZctAZ0mNew+300+0JUmYbv7z0pvJy\nXP/6gEBTE8ZZseexY6mYl8f7b1RRbjHynjM6RD9dU18erO0k02okN0lbuabNb7DZ4aW5w0tlSWbc\nB8fTsZPO+lfR6FLJm/PlpAXiXpYUPd+8ZDFGvZbd1e1kWIxcdHoJP75yFT+5cjWfPL1kSgZigIrC\ndDKtRrYdbCUYmtxZw072zy21qCpcfl4FGRYDe2o6iEQkiclk0NThpb7Vg8sb5I0P6ya6OQO0d/n6\nhqgh2jPWWtNQdNEvDX5PLWokOOiWprEsDhFPSk/yj6HqG58s1WpkVlEGAaefPIsRk2H6pr6879ld\n/PJPO5L2927aBOOhhqjVSJiupndQFB0z5n4FY+rYPGCFuRZu+cpp3PIfp3HnV9dw8dllFOYmPyn8\neNMoCqvn59PtD/dtH5sK3N1BNu5qICvNyGnz8lhYlo27O0hNk3Pok8WY23H4eI72Vzcfo9s/eVKv\n+oNhPN3BvsVbqqoScjhOWLzVt6VpkPni3sVb+bPGLxgPNxNXfxXzo1NZn1o8ix/9x2nTNvXlGYtm\n0NLZzWtbB08TmqhpE4x7qzTFW/Hn6dhBONCFJWcFOuPwM3MNx4wsM8UzrKfcgofVlfkAbNrXPMEt\nSdzb2+sJBCNcsHI2Oq2mb/9575c3MbF2HGpDAT66cjYeX4jXJ1HvuLO3WlPPPG+k24saCJyQ8KPb\nVQ2KFqO1JO51WhqcKArkzhi/L+XGgkIUo3FYmbh6ldlyUBSor+5gxhA7LU5lnz6zFEuKnhffP0Kn\n2z/q602LYOwPhDlwrJPZeZaYi0CiveJ3URQdafmnT0ALTw1F+RbyM1PYebhtShSPCIbCvP5hHSlG\nHWctic6bVZZkolEUdg9RoUaMPZc3wOH6LsoL0/n0WaWkmnT8c/MxvL7J8Wz1lk7s29bUV60pmvAj\nHHQT7G7CZClCo4nde4xEVFqaXGTmpKIfZkbA0VC0WkwlpQQaGwh7vcM6N8VsoLA0i9YmF12O4Z17\nKjGb9Fx8Thn+QJhn3x7+l5qTTYtgvP+og1A4EneI2tOxg3CwC5OujGO33I5n755xbuGpQVEUVlfm\nEwhF2HEofgnISNg3KQpLfLC3GacnwLqls/oWBplNeioK0znS6MTpDUxwC6e3XVXtqCosrcghxajj\nY6uL8PhCvPFhcoYFR6s3GGdZT0r40TNMfXxLU/z5Ykebh1AwMi5bmk6WUl4xouQfcHyo+vD+yVXE\nZrydvXgWRXkW3t/TRFXD6JIeTYtgvKs6/nxx/15xaFsXwZZmGh68H9+RmvFu5ilh1fzBh6oj4QD1\ney4EnpUAACAASURBVH5FW/XTE1r4IqKqvLr5GFqNwkdOqjKzqCwrusWpRnrHE6k3W97SnjSy5y0v\nxJKi559baidF79jhPqln3HniSuqhtjTB+Cb7OFnvfmPfCOaNS+fkoNUqHNzTRDg8tRZsJpNGo/C5\nj84F4A+vHezLDzGiayWrUZOVqqrsqmoj1aSjLMYCCXdPrzg1ewXe7XtRjEbUQID6+35FoKVlAlo8\ntc3KSWV2noU9NR24Y5RbUzR6jJZiup0H6aj9+4T1kHdVtdPY7mVNZf4Jq2Hh+LqCqbQQ7VQTDEXY\nXdNBXmYKM7Oj85IpxmhWuOjc8cT3jnvzUmcNyEudiaqq+FzVaHUW9Kb4e/ePr6Qe/56xqXxkyT8A\njCYdcxbk09nRzeYN07vjMnd2Bqvm51HT6OL93YnViY7llA/G9W0eOpx+FpZlD8gMpUZCOJs2oig6\njIFCwi4n1hUryfv8Fwm7nNT/6u5oFRYxLKsr8wlHVLYdHDiEpSgKOSWXYkiZiad9O86mDRPQQni1\nJ8nH+hj5dGfnWaJbnKo7RvVNV4yc/ZgDfyDM0oqcExY69vWON09877i22Y1Go/QtYurfMw52NxEJ\neTCllQ+6ULOlwYVWpyErN3Vc2tyfzpqGPj+/L/nHcJ1xfgXpmSns2FTL0Wm+4PHycysw6DQ8+07V\niFf8n/LBuHcVdawqTdFesRNLzkp8ew8BkLp4CRnrziProk8SbG2h/te/IuLzjWubp7pVPVm84g1V\na7QGcss/i9aQQVfTO7jbto1n86hpdGKv7WRhWVbMbWWKovRtcTrSM4woxteOk4aoe/X2jr3+EK8n\naUvJSEQiKsdaXBTlW/vSxvbvGSeSAjMYDNPe6v7/7L13eBzndfb9m5ntFYvFohONIAD23sUuSqRE\nFapZsiLbcm+xnf7Fb/zGTt44iZO4FyW2I1uybFWrUqIoiWIXewOIQvTesdjeZ74/FgCJXgg2Wfd1\n4QKwM/Pss7uzz3nOOfe5D45UE5J0fZZifV4+ciBAuLV10tdqtCq23jMHURLY+0Y5Xs+VM4pvViRa\ndNyxKhu3L8zrR+qmNMafhDEWgHl5g1uaxb3iSwxq7/lzIEkY5sRbg9nvuQ/LLesI1dXS8oufokSv\nf47qZkFSgp6ZGRbK652jUv4ltYnkmY8iqgz0NO4i4Bq5afnVQL/05bYVo3eZWfBRqPq6QVEUzlZ1\nYdCqyM8crlq1ZenluePhqZBrgbYeP+GIzMzL5hd1OhG0WkS9/lK+eAxj3NXuRVGuT764H/2h6qnk\njQEcqWbWbsknGIjw7qulyFPwsD8s2LYyC7tFxzsnGmnvmTzL/ENtjH3BCFVNLvLSLZiHCP97uy95\nxYo/RqiuFn3+LCRDPOQkCAIpj30K44KF+C+U0P7bJ28IBvDNgpWzU1CAE+Wj593VOjuOvIcRBImu\n2hcJ+a5+DWm/9GVWionZ2aPXk18qcbp5jfHNer82dnjpcYdYMNOOagSPUaeJM6v9oSjvnLw+dcf1\nbfGIycyMSzXF0V4nqgQbihwh5GtAo09DUo8efu64hs0hRkO/Eldgkkpcl2Pu4nTyCh20Nrk4eah+\nuqZ200GjlvjY5nxissKz71VO+vobS+h1mnGhNp7zG8qiVuQo7vbLvOLjZwAwzl8w6DxBkkj7wpdp\n+q9/x/3BYaSEBBz3P3jN5n8zY3lRMn94r5Ljpe1sHcJWvhxaYyZJuQ/QWfMcnTXPkjLrcdS6q6dt\n3S99uW1F1pi5PINOTX6GhcomFx5/eNhm7kZGW4+fX75eiiQJ/NXHFqGdYPetGwX9IeqF+aNrNW9e\nksHuYw3sOdHI1mWZGK6xClR9e9wY52fGjbEciRDzeNBkZBLy1oEio7OM7hXDJfLWtVTeGgpNRiaC\nVjcl8Y9+CILAxu0FdLZ5OHWknvQsK5k5icPOkyNhmv7ze0SdPUhmC5LZjKrvtzTwO/63ymxGslgQ\ntVe/ccZ0Ymmhg6KsBM5Vd1Nc0z2ptpIfas94IF88c/CXesArdixDUpvwnT8HxPPFQyFqtWT8+V+g\nTknF+dYunO+9c/Un/iGA1aSlKMtGdYubzt7AmOfqrQUkzrgTOeqns/r3xCLeqzKny6UvJ9Kdav5M\n+01X4nS8rJ1/+s0JalvdVDW5eGbPtQv/TxfOVXUhiQLz84Yv6P3QaVRsX5lF4Dp5x/VtHgQgt8+Q\nxlz9gh+2uOoWY9cXA7S3eNDpVZit168blSCK6HL7xD98vimPo9Wpue3eOYiiwLuvl+EfIT3l/uAI\nweoq5FCIcGsL/pJi3B8cxrlnN10vPU/7b35Ny09+SON3/4nav/8bqr7yBSq//Hlqv/l3uI8euZKX\nec0gCAKP3FqAIMAf3q0kOomyrw+lMZYVhZYuH8U13VhNGrJSLpF0BnnFyWtQolH8pRdQJSWhSRu5\ne4lkNpP5jb9CslrpfPb3eE4ev1Yv5aZGvzzm8bLx5TFNSUuwpK4nGnbSWf0H5Nj0C268P0T6cjzc\nTCVOkajM7/ZU8MSrF1CAz9w5m+xUM4eKWzl4ruV6T2/CcHpC1LZ6KJiRMK63u2lJxnXJHcuKQn27\nh1S7AV2fWMzl3ZqC7moEUYPWOHpEKOAP43EFSU6zXHdZ3IGmEbVXpiKVnGZh1cY8Ar4I775eNqjZ\niiLLOPfsBkki5zv/j1k//x/yf/bf5P7bfzDjm/+X9K99g5RPfYak+x/Cdts2LKvXYpg3H01aOlFX\nL22/+h+6/vjilFjf1xozkk1sXJxBW49/Us1NbvowtaIodLmC1La6qWvzUNf3OxiOAbB+Yfqgm93b\nfYZYxI05eTWS2oS/vAw5EMC8as2YXwq1w0HG1/+Spu/9K22/+h8kkxlD0fR2dfqwYWmhg6ffruBY\naQd3rs4Z93xr6gZiEQ++7jN01b6AY2Y8nzwdiERjvDdE+nI8zEg2Ye3v4qQoiDeolnhHb4BfvFJC\nfZuHTIeRL907jzS7kYIZCXznyRP87p2LZKeayUq5frnJieJc9cgs6pGg06jYviqLF96vZs+JRu5d\nN3ZYeLrQ6QwQDMfITr30fvYbY9GuJxrqRm8pQBBHv3cviX1c/8/kUr1xNcZ5C8Y5e2wsWJ5JS0Mv\ndVXdnPmgnqVrcwDwnTtLpK0Ny9p1A6IoolaLqHWgTnKMOWa4rZXmH/+QnjffINzeRuqnP3fDh693\nrsvjeGk7rx2uZdXcVKzG8dNc0+oZFxYWni4sLHy/7+fX0zl2P3q9Ic5UdvLHAzV8//mzfP3Hh/i7\nJz7giVcvsPtYAxUN8R6Ta+al8ujWAh7YeClUNOAVi2osyXENal9xPERtGiFEPRS6rGzSv/I1FEWh\n5Wc/JtR4/YUHbmQYdWrm59lp6vTS3DV+CEwQBBJn3InOkk/QU01PwxvTRkIakL5cnD7hnriCIDA/\n147HHxkg7NxoOH2xk+88eYL6Ng+3LEjj/3xiGWn2OGnIkaDnszvmEInK/PzlkutelzsRnOuTUV04\na2K9fTcvjjOr3znZiO8aecf9+eKcyzY3/TXGiiU+h7FUt+Ay8tZ1zBf3Q5/X5xlPkVF9OQRBYNOd\nRZgsWk4cqqOlIR6+73n7LQBst2+b9Jia1DSyvvkt9IVFeE+dpPHfv0ukb/MzFNFojPLiNo7uqyF2\nHVu5mvRq7l2XRyAU44/7JxZxmDbPuLCwUAdQUVGxabrGvByvH65l39mWAT3YfiRZdczOtpGbZiE3\nLb77H22xjXvFnj6vOL5g+c6fR9Bo0BcWTWgehtlzSP3M52j7nydo+tF/kfX3/4Dafu2agt9sWDEn\nmbNVXRwvbWfn+vE9F0EQScp5gI6qp/D1nENSW0hIv7JbapD05dLRQ4cjYf5MO4eKWymu7ib3Opag\nDEU0JvPivrhHqFGJfObO2aydnzbsvEWzkrhjVTZvHq3nf98s4ys75133sOhoCEVilNY7yUgykjzB\nhu1ajTTgHb9zjbzjur6N2UiecVTdC9EJGOMbyDOWTCbUKakEa2tQZBlBvDIfTadXs/XuObzyzBne\nfa2UuzYlEqyqxLhgIdr0qbWmlUwmMv/ir2l/5incBw/Q8C/fIeOr30CXkwOAuzdA6dkWys61EgxE\nUalEFizLwGC6fh70xsXp7D/bzKHzrWxakkFO6tjrx3R6xgsBQ2Fh4duFhYXvFRYWrpyugZs7vbx8\nsJZwJK7Is3NdLn/x0EJ+9LVb+N6X1vCle+exbWUWhVm2UQ3xSF5xpLOTcGsLhqLZiJqJs2UtK1bh\neOgRYr29NP/gv4h5rw7h6MOAxfkONGqRY2XtE/ZyRUmDI+8RVNpE3O0H8XSdvKI5jCV9OR7m3oAl\nTt2uIP/2zGn2nGgkzW7gHz65bERD3I+d63MpnJHA6Yud7Dlx40ZzSut6iERlFk3QK+7H5sWZmA3X\nzjvuj5LMSL5kSCNOJwgQkdtRaWyotaOTzxRFoaPVjSVBh/4GYenrZ8bFPwKV00P4S820snJDHj5v\nmL1vXkQBbLdvv6IxBZWKlE88juOhh4m5XDR877tU7PmAt14s5pknjnHmaPzeXrwqi4c/t+K6GmIA\nSRR55NYCFOD371SOu/5NpzH2Af9RUVFxO/BF4JnCwsJpGb8/Cf74HbP52gMLuGttLvPz7JMqN/F2\nnyYW8WBKWnbJK+4LURvnjx+iHgrbbbdju31bPJ/xkx8iRyZOOIpGYoPIDR9maDUSi/KT6HAGBsJ7\nE4GkNuKY+XFElQFn41v4eyumPIexpC/Hg0GnZmaGhZoW94ha29ca56q6+PaTx6lpcbNqbgrf+uSy\nYSpiiiLjajtIV+2LOJv24Os6zmc2qyhICfLm4TIqG0cO8V1v9Hf6GqukaSRoNRLbV2YTCMXYc/zq\nbjYURaGh3UOKTY9Bd2njH+11IqTqUeTwuCVN7t4gwUD0uop9DIV1/QYAOl94btpIUotWziAzw0hH\nzEJz7nr0BYVXPKYgCBjW34rrzi/wQdoO9p4OUVfVjSPNzOY7i3jsK6tZtTHvujLUL8fsbBtLCx1U\nNbs4Ok6f9+kkcF0EqgAqKioqCwsLu4E0oHm0CxyO8UM0Xn+YD0rbSbbpuXV1LpI4+RCbHIvQWnoE\nUVSTO+c21Jr44tVZcQGArE1r0E5gLkOR9MXPcDHgo+vAQdwvPUv+V788aggwGolRWdZO8elmKss6\nMBg07HhoAbP6uhx9mLF1VQ7Hyzo4X+tk+fzJhKnMWM2f5eLJJ+iufwlb4uNY7AWTem5nIEpFYy9L\nipJZPGd07xEgEvYS8nUiqrSIkgZJ0iBKWlbNS6OyyUVDl58NSzIn9fzThVhM5um3ynjp/SrUKpGv\nPriQ21ZmD7vfIiEPNed/j9c5vC3exxf1jdVxnBZPAjqDDY0uYdCPyZaLpLr2C5ksKxTX9pBg0rJi\nQcakv+cPbi1kz4lG3j3VxCPbZ1+1uvC2bh++YJQlRSkD65fDYabe7UIzO05MSsmcR8IY60l7Yzxf\nnDcraUJr4DWBYwmB9bfQdeAQSskpkrdsnpZhV6ir6IxauKjOY40vyozc0SMG46G91c3Jw3WcP9VE\nJBxD0lnJCDSQ1n6O/Mw55K9bi6i+tvXmE8GXHljEl/79PV7aX8PW1bmjnjedxvhxYAHwlcLCwnTA\nAowpeNrZOb6ntPtYA6FwjA1r0+npnlo42NN5nEjIjTl5Db0uBfAgh0L0ni9Bk5GJGx1MYC4jIeHh\nx/DUN9Lx7l5IySRh06WbWJZlmut7qSztoPZiJ+FQnOFtTdTj6Q3yh18dZ/bCNNZsnolmgqSimxFZ\ndgN6rYr9p5vYsSprkqzkBOw5cVGQylO/JjFrByb7ogld6XCYefbtcgA2L0of834LuCrpqvsjijy8\nPnIm8M0tIkr7Cc7tMyCIGgRJjShqEPqaxitKDJQYihKLt4aU+/+ODTkWQ0DE7FiBJeWWMRm3AKFw\njPp2D3/cX83FJhfJNj1fvnceWSlmuroGfx9C3ka66l4kFvGgtxaRkHErctRPLOwmGnETC7toamvD\n6+0mUfYTDQ33kEWVAWvqekz2pePObTpR3eKi1xPilgVpU/6e37Z8Bs+/X8Xv3yrjvgnwE6aCM32K\ncqkJOjo7PTgcZjraXYR7etBmZAECITl1zHutqiI+hsGindAaeK1g3rGT7qPHqf3t76BgHqJuYnn7\n0RB1uXDve4+FSfmcUK/ihadO8uDjy9DpJ2YwI+EoHleIrg4vpWdbaG2M9ws2WbQsWZ3F7IVpqCOL\naPlZHZ37DuBpbCH9K19DZblxIg4QDz/fviKLN47U8dvXS/jiAyOvX9NpAX4NPFlYWNjfhufxioqK\nK4p3yLLC3tNNaFQi6xZMrBxlKAY0qEU1luTVA4/7y8tQIpFhqluThajRkP7lP6fh/32bjmefQZOR\ngceURuWFdqrKOwj44qFNk0XL3MXpzJqTQqLDSHeHj71vlFF2rpWmOieb7igkYwx5xpsZapXI0gIH\nh4pbqWpyUTAjYfyLLoPekk9y/mN01TxHT8NrRMNOrKkbxyUitXX7xpW+VBQFT8dRelveBUHE5FgR\nfzwWRpHDyHIYJRamvc2JWoph0EWJRfwowTAwSqpBkBAEMV6WJUgIgoQgqBAkLYIgEYt4cLXtx99b\nSmLWXWiNcW87EpVp6vTGy/RaPdS2uWnp8tGfalpWlMzj24uG8SIURcHbdQJn0x5AISH9VszJq+Pv\njzYRLlNktGYo/PSlYs4e7eKetTO4c0US0bCLWMRNJNCBp+skzqbdeDqPk5C2GX3C7GtC+OoPUU+k\npGk0bFqcwe5jddTUnKU17RhGS/al92Ga0J9quZy8FfN6USQZEuKKcqI0dq6yo9WNIIAjZXiTkusJ\ndaKdxO130v3qy3TveuOK1QZ7338XJRpl5pblKJocThys4/1d5Wy7P67/Hw5F8biCfT8hPO7gwP9e\ndzyUfzkyc2zMW5JBdr4dsS9yIisqUv/qr+n8zZP4Thyn/l++g+1LX8A4IwetdHXz8ec7L9DkbUFW\nZGKKjNz3c+nvGDFFRlEUIrYoxqJO3u04zxe5ysa4oqIiCjw2XeNBvOawyxVk/cJ0TBPcTQ2Ft/s0\nsagXS/KaQTqxvuLzwMiqW5OF2m5H//HPc+7F9znyQiUBKR4Q0OlVfQY4mdRM66BFISnFxP2fWsrJ\nw3Wc+aCB1/5wjvnLMli5IQ/1TSZfOBGsnJPCoeJWjpW2T9oYA+hMWaQUPE5n9R9wtx0kGnJhz7pr\nTO/t1QPVY0pfKnKUnsZdcda2ykRS3sfQGkcOo79RWcrh4ja+9cll5KZZ4mQMJYYshxEQ4kZXlABx\n3MVfjgVxNr+Hr/sUbRf/l2b/LPZV51DXHiQau2TgNWqR/AwruWkWCrMShrUTjI8VpqfxDfzOEkSV\nkaSc+9CZRw+FiYLAZ3bM5jtPnuC1w43kZSQyPy9n4Lg5eTWutgN4u07RVfciGmMmtvStaE2TY6FP\nFuequlBJInNHkFGcCBQ5RsRTwpdvOYtO6CXihV5vJSFfA/asexGnKfTeT97KGlLWJGbqQRifRR2L\nyXS2e7Enm1BdwfdcURQUlGELvzzkZ7CRiF36W77sPC47d3Emhn0meva8RU2hjajNPKJxifX9P7Ih\nikEozMJ39qLo1byS0ExEagF7EnVV3fziJ+9ASEKIjfz6FVEmpg0RtYWIaYNEtAF8iZ3U6H3sa5WJ\ntVx6PqV/Q5yvsCJoZHVxN83/9q+8sy6Rzz74bSyaq5MGUBSFZ8pfxBuZhGqZBYQxqEI3dGy0n7h1\n69Kp5elkOTLgFZtT1gw8rigKvvPnEA2GAfWZqaKmopNTh+vp6vCCbQGSHCFdaWfhfRuZke8YszWa\nJImsXJ9HTn4Se98oo/hkMw01PWy+s4jUjOHdam5mFGUnYDGoOVHewce3zhrWW3oiUOuSSCn4NJ01\nz+J3nicWcePIfWjEhdYbiPDO8Qbso0hfxiJeOmufJ+xrQmNIJyn3IVSa0cNb8/PsHC5uo7gmXuIk\nCAIIKiRxcl+hdqefp3ZXUN1iJNU4n7vnVpFpqmRHfgMnjPPRmGaSk2YmN81Cut044AGMhEiwm67a\nF4gEO9AYM0nKeWDM19APo07Nl3fO47tPn+KXr5fyj59ajr2P8CKpjSTO2I7ZsYLe1r0Eestor3wy\nHvZO33JVdMO7egM0dfpYMNOOVjO+gVIUZWDhj0b9+HvOEug6hRL1ohMESjvtnG2z8/ByLwHXRZrK\nn0BJ2YSsSRjRcMT/V4YZtGGGTJaplWuw5Am83xpBVmR0LSqEC7XkZ8VDuvt76nE5nx/RS5IVBdml\nQopm0aZu5L9OHRt5PvJl1zHyfGTl6tXQzponcsdhme4XX2DX+qmtQwsr/KhCEY7OM3DCWQKAaoaG\nPM8apJCaiMZHWBsgogkQ04WIaoPEdGFi2hBoYkiiiChIiIKIKAhoBQm9YEUSLn9c7Ps//uPfKHIh\no4uid8q4c18P2u1hmHqKekwIgsDfLPtzugM98XmI4mVzkhAQBs21/7jTNToJ9IY1xs1dPkrrnBRl\nJZCZPLVwjq+rzytOWYukMgw8Hm5pJtrTjXn5CgRpartTRVE4ebiek4fqEEWB7Hw7s2YnozvyKoFj\nR9Ad9SEWPD6hsVLSLTz4+DKOHajl/IkmXvndGRatymL52hwk1YdDsVQSRZYVJbP3dDNldU7mTUJA\nfdA4aiPJsz5Bd93LBFzltFc+iWPmI6g0g73tt4/HuQY7b8kdJn0Z9rfSWfMcsYgbQ8JcErPvRhTH\njrzMzU1EEOLSmHevHd3zHAuRaIyfv1xCY4eXDIeRrNQU3MbFJBrLSeAkW/NOYrCFsGXcNmK3n8u9\nIX9vGe7GXShyGE3iIrQpt+CKxZD93Zc8GJRRvaGYKsbG9Sr2nm3k+3u6uGttNoKgDF74VeloLDrs\n/mpwleN3VdCtSaZVk05EkIgpsTG9pNG9M+Wy4zFc/hDa+UFajWr+75G3RvD0lEHjKyiYBIGlOjWL\ntWq0gkBYUTgXinAiGMGj8kBmPf/ZCut0GlbjJtr4Crv9IUrDVyh8kgIysLuufOCheQ0BCmfbCEdl\n9jiLR0teAGDryCIDaFM34na3XGZMpEGGRRRE1KIaSRAR+ozOoMVdEAcZJGHIY/1GYbTjlz+vKAiD\nrykQiDS+Rn5DK5/RrkXJzx5ynjTqWJIgIsgKvt3fQ1aruPPRb3KPxTJwHlsEVKKqb07x8aYVCyG4\nvA7f+XNorJOPwE0GSfpEkvSTs/aWMcTGblhjvLfPK94ySZGGfshyBHf74bhXfFmuGBizMcREEInE\neH9XOdXlnZitOrbfPw9734ZBnvUpGlubcR88gC4nl4QNExOsUKkl1m7JJ3dWEnt3lXPmgwbqq7rZ\nsqOIpJtAxnAiWDknhb2nmzlW2j5pYywPWZgNmXcgS3pCPWdoLf8V2hk7QGtHVmTO13Tx5rlKbClq\nsvKiVPXWDlwv+OrRdR9DUGL4LLPp0OdR01E8gpc0JKynyDiKmmnwBHm+vBVJYhTDExvFCMm09Xhx\n2oMkZakwWbW0KTItEZkDPTIWJG5RKyQ7i+nuKeZgSKY8Mty4CcB6vYZVOg0RRYkbGOchqD40pc9E\nOwt6gafLPxjzvAK1xAa9lqRwO+ZQG0eDYU4GI0zFtF3yGuILeDgqI6hAkERkhT5DpBpkWERBQBIk\nEgSZQiHIDCGICAQRuShYaJIsyBo1RRYRFIFjZR3o1GrISONizENeqJ67jDpWJyTTps9BGDB0l3sw\nwoABGzAy4iWDVtvs4bXD9aybn86aeelIokiizUS362UEcysaUvjWqvuGGCpp0HMc2l3FRTr4y02f\nHlgzbkQEP5FGw798h8Q9J8hetWNSTov7+FE8PU6smzZjd1z76gNddg667Jxr/rxXihvSGPuDEQ6X\ntGK3aFk0a2oelL+nmFjUizl5zSCvGPqMsSBgmDd/0uN6PSF2v1RCZ5uH1Ewr2+6bO6hwX9RoSP/K\nn9Pwz9+h4/e/Q5uRiT5/1oTHT89K4KFPL+OD96spPdvKS789zZI1WSxclYlGdXVp+70hF+c7LxCV\no8gol4XLRsgNyTF0rd3YSxrwJZpoXTRj3FxVTJExzndzGpnWo2+hoAw5d/QwoTKKv7FUq2aLXsFd\n+xyv+YJUR+KMdd08CAI/Ob9/4Ny1Og236DWEFYXXfUGqnCeAExN/g8ygMsP+luFlQxOCAJJJICiK\ntPkHexchQeI1WWKeSmGpKsZtOpG5GjUnZANBQYWIiEGEFYKHZML4UHFaSkJj1bFkiCcU92BGMS5D\nPBtZhj3Hm+j1hNmwKIP8dNtl5w7xwlCIeGvQ9BazQS+w3mRDsC9FshQhiUM9tuGe3kjekD8Y5es/\nPkhmsol//NTyEd82RVEIeetxd3xA0B3vE6vS2rEkr8aYuICCEVIFvWXnOF/ezdbNa0i06AZC+knB\nDtLVapJyH5xQSP9yVJfVILt9LMmYzSxbfF1y2M341H5igCWhiATD2DrLHc0e1BoJW9LofY5vBOhy\ncrCsXYf70AFcB/YPqhIZC4qi4Nz9FggCtq2Tl778U8YNaYwPnW8lHJHZtCYDUAjHIqOGwUbOucQQ\n2w4CIh1qBy3dFSj9x/1+DNWVxDJTOOm9iOwZK7Q2+DlDPdB7RI8SFFFnBfEs6OD5uooR52O9NZul\nr5ZQ+ePvceCBOfgNqlFzQ8PCiYqMrJPRFyaSXjOXk4fqOXDmHItuS2ZL4Zrx3r4p4/3GQ7zbsH/M\nczRhmaK6IPOqgjh6435RTIS37R2ENKOHnPq9IfQCSkyhOxBGo5IGLfqXe0ODc0PCEGNzadFHEClW\nAsyNtnK/Sc+7HUaO92hZlO8gJy2BUDCGGoWsQC3WaA9hUUuLeR4LHRYWDw3X0fc8Qw0L8d+dziC/\neauC+XlJ7Fw3c8zc0OWvwe2N8E+/OUUoHONbn1zOjHE8omiol57GN8jw1JCpCpKQthmNIb2vc2si\nSAAAIABJREFUbCmM3lpIZvY9zJamh5Q0z+zjn35zksP7YcsnZw9oW48IexFy5mbc7YfxdBxD6TiI\nOtiKPXsnkmrypTAltd3EZIXFo7CoYxEPXXWvEPLWAqA1zsCcvAa9tWBMotzcnETOV3dzobaHdQvT\nUevspBR8mp7GXfidxbRV/A9JOfePSXYbin7yVvaQSFVMF3/c6Jg75vUeV5DengDZMxPH5ALcKEja\neT/ek8fpevWPmFesRDKOv4EIlJcRaqjHtHQZmuTx25R+hEu4bsb4z3f9X8KRyIgMvXA0hm65zK4A\n7No3+bHzVBIPmvWUhCLsKnlm0LFZ9UHukBVOJHg4Xvb8hMe0dKeSWbsQQRZom1FGd0otdIxxgQEC\ni02sP+1l/pulvLo1CVTSoNzLpZyOhFbQDDdEZhE5vRG5IhVDSwLRBj1cuYjNqLg9ezO51mwEGOzd\nICA0tMDR0yjnLkAkAqKIauF8BI0aTpzmr7W3YlyzZuQc0mXeUEO7h28/eQKjRcvXH1pE+jR5CC5n\nHW2Vz7I12ccyRy4LFu8kOdlCa3MTnTXPEYn2oDVmkZH7IPkj5GMnglk2hReVXmqrFDK3Z0yoXlqW\nFZ7cVYIvEOXRrQXjGmIAlTYBx8xH8TuLcTa9jbP57b4jAta0zVhS1k5ruU6a3cinthfx369d4Ecv\nnOfvH1s6ZpcZUdKRkL4FU9IyehpeJ+iuoq3ilzhyH0RjGFtYZSjOVo2uuhX01NJV90fkqA+deSbW\n1PUTZnXP7ROXuFDXM9ClS5Q02LPvRWvMxNn0Nh1VvyMhfTPm5LE7tvWjrt2DzazFctl7I8ciKNYY\nijOK2jR2WVZjX1/sKxG+uJZQWa0k7ribrhefp/v1V0h++NFxr+nZ/SYAttvvuNrT+9DhuhljWY6N\n6A0FwzLt7gAWg440m3FQ6E3qq98cizwgCiIFgUqIedDYl7AjxXKZ8RNJLDkAuClas52CjJRhuSER\nYZBnhCJQd9pNdbULlVpg6R3ppOXOGYFEMTREKCJuFOj49S/h2FH+rrmQlE98ampv1kro7vRitV1Z\nEf54MKj1LHLMG/g/5vXiPnoE14H9hFviQmpqhwPrug1Y1t6CyppApLOT2hOnkc8UY900flgqK8XM\n/RvyeGl/Df/6u1N87YEFzMq8MqJFNCbzy7edtLYv4HNrK7BJJXTXyeikdbRV/A456sNoX0Ji5vYr\nErIQBYH5uYkcLmmjod0zrvA7wFvH6ilv6GVRfhKbl0xcfUwQBIyJC9CZZ+Js3kPI14g9awc689UR\ns1g5J4XmLh9vHKnjh8+f428/vnjc7lYqjRXHzI/jatuPu+0g7RefxDbjjgmLssRkmeLqbmxm7eCe\n44qMu+0grrb9gEhCxm2YHSsntQFJsxtItGi5UNuDLCsDnqggCJgdy9EYUumqfZHelvcI+ZqxZ98z\nZn1wrzeEyxseVgft7a1FUAnQNf7cBoxx3s1hjAEStmzFdWA/vXvfw7p+E9r00fUeQo2N+C+UoC8o\nRJ93bdpZfphw3Yzxz+76lxHVZ/7r2TM01Dn56qdXTMiLGIpwoJ228tNoTTncOmvnoGOKLFNT8wJY\nE1ix7M5xv9yXiFouzFYddzwwn0TH5LyqlE88TrilBdeBfWizc0jYsHHC1yqyTOBiBZ5TJ/GXXkB9\n191YVl29MDXEcz6Byou4DuzDe/IESjQKkoRp2QoSNmxEX1g0qKuL2uFAl5eHv6yUqNs9IfWbO1fn\nkGDS8uSb5fzns2f5wt1zWVIwdq5trPn+bs9FSuucLMrPJH/BRrrrnsffW0rlqVJAwJa5DVPS8mnx\nJufPtHO4pI3i6u5xjXF1s4uXD9RiM2v59J1TE8+Q1EaScnaOf+I0YOe6XNy+MAfOtfDTPxbzjQcX\noh6HzS8IIglpm9AaMuiqf4WehtcI+5qwZW5DGKfsq6rJhS8YZcXslIH3Jhbx0V3/MkFPDZLaSlLu\n/QOiKJOBIAjMzUnk4PlW6ts9wzpuaY0zSC38HF11LxFwldNW0UlS7oNo9COHVhv62yamDg5R97aX\nAiD6x14XZFmmqc6J2aq76pvq6YSoVuN46GFafvojOp/7PRnf+KtR7+OePf1tEq+sIcSfKm6onHFL\nl48LdU4KZyRMyRADeDqOAmBJXjXsWLC2hpjXg+WW9eMujHGiVjGdbV7SMq3cPoSoNVGIWi3pX/lz\n6v/523T8/mm0mZlj1jYrsRj+inK8p07gPX2KmCe+CIgmE6Lh6pI+/Bcr6HjqN4Tb4qIl6pQUrOs3\nYlm9dkwja16+kmBNDd5TJ0jYtGVCz7V2fhoWo4afv1zCz14u5s9uK2TT4sm3V9t9vIED51rISjHx\n+bvnoNKoSM7/M7ob3iDsq8U242704wgxTAZzcvpLnHq4a4wSJ38wyn+/dgFFUfjcjjlTFq25lhAE\ngcduL8DjD3OmsotfvlHKF++eO6H8pt5aQFrh5+isfR5v92nCgTaSch8YVnJ2OfpD1P1dmoLeerrr\n/kgs4kFnmYU9+94p5aH7MTc3boxLantGbH8pqU0k5z9Gb8t7eDo+oP3ir9GZclHrU9DoU1DrU1Bp\nExEEYaBtYtYQY+zurECJyqiFsYmmHS0ewqEY+bOTb9gWlqPBuHARhjlz8V8owVd8DtOC4ZGPSE83\nnuPH0KSnX7Gq4Z8qbihj/N7p/nKmqdHhYxEPPmcxKq0dnWU4g3miqlsdrW7eeqkEvzdM0YJU1t9e\nMKZ4x3hQJzlI+8KXaf7Bf9Ly85+S/a1vo0q4tEgp0Sj+8lI8J0/iPXsaua8lo2Q2Y92wEdPS5RgK\ni6ZcEz1RBKsqiXR1Yl6xCuv6DXEveAILh3n5CjqffxbP8WMTNsYQF9L4248v5ocvnOPptytweoLs\nXJc34cXqVEUHL75fjc2s5esPLESnid/OgqgiKedekpJMw/Sbx0Lv/n0gQML6jaOeY9KrmZlupbrF\nhTcQGdHIKorCU2+X0+UKsmNNDkU3kcypJIp84e65fP+5s5ws7+D3BjWPbh2bLNUPldZGSsGncTa+\nia/nHG3lv8Sec9+om6GzlV1o1RKFM6y42w/T27IXgIT0LRPO446FOTmJCMCF2h7uWpMz4jmCIGLL\n2DqQRw64LxJwX2ojKIhq1LpkrFE1y2ZIZFqykWNmRElLLOIhFOpCbgmito4dem64yfLFl0MQBBwf\n+zj13/kWnc/9AeOceQiqwaaj9913IBbDdtv2K+6H/KeKG8YY+4NRjhS3kWjRsrhgavq0ns6ToMiY\nk0fOL/nOnwNJwjhnzqhjVJV1sHdXOXJMZs3mmSxYnjktO1njnLkkPfAQXS88R8svfkrmX/wV/osV\neE+ewHv2DLLfD4BktWLdtAXz0mXoCwqv6Y1t234ntm13TPo5VQk29AWFBCrKifT0oE6c+IKTm2bh\nm48t5QfPneONI/X0esJ8YlvhMKGOoahtdfPL10vRqCW+/sCCEfsUT+Zzi3R30fHMUwiiiGXlakTt\n6PnD+XmJVDW7KK3rYcUIXbcOnW/leFkH+RlW7rklZ8JzuFGgUUt87YEF/NszZ9h7uhmLUTNhoRNR\nVJOYdTcaYybOpt10Vj+DNW1TvCnGZZ9Ha7ePdmeAVUUWehueJ+iuQlKbsefcj840+VaXI8GkV5OT\nZqG62UUgFB0zB25ImI0hYTaxiJdwoJ1IoP3Sb38ruRaZ3Dngb67G3wwqjQ2xz2uXGwKoFoy94Wqs\n7UEQuGn157UZGSRs3ETv3vfo3fsuttsu8UNifh+9+/chWRMwrxwekfwIE8MNs4U5VNxKKBJj0+KM\nKUklynIEb9dJREmPMXG45xvt7SXUUI+hoHDUbiQ1FZ2882opoiiw/YH5LFwxY1pDSrbbtmFesZJg\ndRVVX/8qLT/+Ie4jhxG1WhJu3cqMv/smef/xA1IefQxD0exrvsMUBGHKz2lesRIAz4ljk742xWbg\nm48tJSfVzKHiVn7yUjGhcGzU87tcAX704nkiMZkv3jN3kE7wVOF8522Q5XiUoqJ8zHPnz4yHJIur\nu4cda+328cy7F9FrVXz+7jlTupdvBBh0av7ioYUkWXW8crCWfWdG7YQ6DIIgYE5aSsqsTyGpLbha\n36er5jnkaHDgnLNVXWRa3dyavZ+guwqdOY/Uws9PmyHux9zcRGKyQnnDxHo4S2oTestMLClrSMrZ\nSdrsL5Iw6y954sgiTrYvxuxYidaUgxwLEva3gAJynR+VbXQjGwxE6Gz1kJJhRau7YfyfScN+905E\no5Hu118l6nYPPO7avw8lFMR269YbsoXhzYIbYqWQFYW9p5pQq0TWL5xadyZfz3nkWABT0tIRpQ19\nxX2qW/NHDlHLssKx/TWIosC9jy4me+b0a/AKgkDKJz+NbmY+KpsN29bbmfH3/0Duv/8XyQ8/in5W\nwU0b4jEvWQaiiOfE8SldbzFq+NuPL2ZeXiLFNd187w+ncfvDw87zB6P86MXzuH1hHtkya9KN6EdC\nzOvFdfDAQOjNX3J+zPOzUsxYDGqKa3uQlUtiJJFojCdevUA4IvOp7UUkWW8eos5IsJm1/OXHFmHS\nq3l6TwWnKsaq5RsOrTGD1MLPoTPnEnBfpK3il4T9bSiKQrDnOI+vKEaFH2vaRhwzHx1RAvRKMa+/\nxKkvTDwVNHQGaPOYwDAXW+btpMz6BBnz/5r0ud/A0pyL4oqgShjdGDfVOVEUyMq9Ob3ifkgmE/Z7\ndiIHAnS/8hIAciSC8913ELQ6rJMgp36E4bghVv7i6m46egOsnJMypabg/W3wEETMjpFVfHznx84X\nV5W209sToHB+KklXsbWZqNWS9ff/QN6//SeOjz2Cfmb+TWuAL4dkNmOYM5dQXS3h9vYpjaHTqPja\n/QtYOy+V2lYP3336FB1O/8DxmCzzxKslNHf62LI0k1uXTU8nod59e1FCIex334uo0+ErLh7zfFEQ\nmJdnx+0L09h+KSf9wr5qGju8rF+YzvIRmlPcjEhNNPAXDy1Eo5b479cuUF4/MQ+zH5LaiGPmo1hS\nbiEadtJ+8X9prfgtS1PLCMc0JOf/GdbUsQmVgeoqfKUXpjT/vHQLOo10Rca4X+zjcia1IAioNBai\nfZ//WJ7xzVjSNBoSNmxCk56B6+ABgg31eI4dJebqJWH9BqSrTDD9sOOGsALvXmF3pqC7kmioG6Nt\nPpJ6eMhSiUbxlV5AnZyCJjV12HFZljl5uB5RFFi6JntKc/gIcVY1TC1U3Q+VJPLpO2dz5+psOpwB\nvvv0Kera3PGWZe9UUlLbw4KZdh7ZMnGJ0bEgR8L07n0XUa/HumkL+tlziHR2jLuhmN+nrX2+Jh6q\nPlvVxbsnm0izG3jk1umZ242C3DQLX905H0WBn/zx/ECZz0QhCCIJ6ZtJyv0YCBLRQAO1PVbq5XvG\nVcBSZJmWn/2Ylh//gEjP5A2qShKZnW2j3Rmgszcw6ethdOUtgHB3D4JKhWQaOVWiKAqNtT3o9KoP\nhca8IEk4Hv44KAqdz/4e5563QJJI2Hrb9Z7aTY/rboxbu31cqO2hINM65dyfu6+cyexYOeLxQOVF\nlFAQ44KRKfeVFzpwOQMULUzDbJ0eicE/RZgWL0FQqaYcqu6HIAjcv2Emf3ZbAR5/hH9/5gxPvlXO\nvjPNzEg28YUJlttMBO4PjhBzu7Fu2ISk12OcF79HfOOEqi/v4uT0hPjfXWWoJJEv3jMP7YewH/Xc\n3EQ+u2MOwVCM7z9/jo4pGDZDQiFpRZ/ndOcynjoxj3n540c2grU1xNxulGiUnrd2TWXql9S4pugd\n17d7MOnVJFqGk/rCPd1ICQmjevbOLj8+T5jMnJtDAnMiMM6Zi3HRYgIXKwi3tGBesRJ14vSn9f7U\ncN2N8UDP4imGHMP+NkLeOrSmXDSG4V4vgPf86PniuFdchygJLF09veSRPzVIBgPG+QsJNzcRam66\n4vE2L8nkyzvnEZMVDp1vJcGk4esPLBhXGWqiUGQZ557dIEnYbt0KgHF+vHnIeKFqk15NXnqcqfuL\nV0vwBiJ8bHP+lOvjbwasnJPCI7fOwu0L8/1nz+LyDc/pjwdFsvJ2iRFHgoE0u2Hc833nzgIgqFS4\nD+6fknd8JcbYF4zQ2RskO8U0zOAqsRhhZy9q2+jh50sSmDd3vngoHA8+DH2llom3/WmKfCixGIo8\nfX2lr6sx9gejHC5pw2a+knKm0UU++uE7fw5Bq0VfMFzY+WJJO+7eIHMWpmGyfOQVXykGWNXHRw5V\nK4rCoXcraaiZ2MK4tDCZv354EYtnJfGNBxeSOI2fke/cWSJtbVhWrRkg4KgT7WjSMwhcLEcOj21s\n5ufZUZS4ktRk5S5vVty6bAY71mTT0RvgB8+fJRCaWBPFaEzG7QtzsryDUDjGollJE6pU8J49g6BW\nk/TQwyjRKM7dk/eOkxP0JFl1lNY7iU1y8WzoD1GPoLYWdbtBlgdpBgzFzaZHPVFoUlJI/dRncDz0\nCNoZ08PduNlQ+3/+jrZfPjFt411Xnv3h4lZC4Rg7VmdPqQQkGvHgc5ag0iahs4ysahVubyfS3oZx\n0eJhtPtYLJ4rliSBxas/yhVPB4wLFiJotXiOH8N+733DFtxoVKaiuI3SMy3c9cgi0jKt445ZMCOB\nghnT3yi85+1++b7BmtrGefNx7tlN4GIFxjHabC6YaeeVg1cmd3kzYue6vD7ZzFZ+8tJ5VsxJwReI\n4A9G8QUj+AJ9v4NR/MEI3mB0WKnaRFjw4Y4Owi3NGBcsJGH9Rnr3vI3rwH5s23egHoMwNRRCH+Fu\n35lmals95GeMf8/1o76PoJWdOjyFFnXGyWyjMamjkRgtjS4SHUaMI9TB3+ywrL660rw3OtRJDjwn\njmNZu27MdWKiuG6esSwrvHe6CZU09XImb+cJUGQso4h8wNiqWxUlbXhcQeYsSsf0IfyyXA+IWi2m\nhYuJdHYQqq8bdlytlth6z1xkWeGtF4txdvuHD3INEKiqJFhViXHBQrTpgz3afjm/8fLG2SlmPrGt\nkL98aOFNIXc5XYjLZhayeFYS5Q29PLW7gpf21/DWsQYOnGvl1MVOyht66XIFAIEUm57Z2TaWFTpY\nvzCdhzblU5Q1/ubKd+4MAMZFixFUKhLv3BH3jqeQO56bE/dMS2qG14aPhfr2fvLW8PRDtLfPGI+y\nMWhpdBGLyh86r/gjxOF46GEQBDqf+0Ncw/8Kcd0849MVHXQ4A9wyP21K5UxyLIy36xSipMeQOLoW\n6mj1xbGYzKnD9UgqkcWrPsoVTyfMK1biOX4Uz/Fj6HKGs2Wz8hLZuL2Q99+sYNfz57nvE0swjNGy\n72rgklc8PN+ly5+FoNXiKxk7bywIAhsXffhD0yNBEuNktdMXO5EVBaNOjVGnwqhXY9CpMGhV46qo\njQdvX764XwvZsnot3btex3VgH4l33Dlmbe9QzM5OQBQELtT1cO+6iXcUqmvzoNeqcCQMrxm/ZIxH\nNrb9IeqsvA9XvvgjxKHLysa6bgOuA/vo3f8+ti1br2i862aMXz9UA0xdh7pf5MOSum5EkQ8AORgk\nUFGOdkbWsLBW+fk2vO4QC5ZlDoSQ6jwBTna5WJ2cQIbxo/zxVGGYOw/RYMBz4jhJDzw0Yh110YI0\nPK4gJw/X8+YLxdzz8UWoNdeGhRxua8N39gzanNwReQSiWo2haHY8p9zZidoxtY5SH3aoVSIr5wyX\nA50OxHw+Ahcr0OXmDeRkBZUK+x130f7Uk/S89SbJj4zfX7cfBl0f4a7FhT8YwaAbP5IRCEVp7/FT\nlDUyWzraRyYbbVPQWNuDSiWS2peKickKh9qd1HuCI54/EsKhKHVlXYT8kQlfA/GNok2jwqC6ut+p\ndIeRu9bkDOjCjwV/NMa7zd30TpBn0A9JhFXJCcy0jE/4u9aw33sfnhPH6H71FSwrVyOZpk7gvH6e\ncXkHszKtI+ZixoOiKHHiliBhThpZ5APAX1aKEo0O6yISi8qc/qAelUpk8aoZRGWZd5p7ONTmRAHO\ndnvYnG5nQ5oN6U8kDzidENVqTEuW4j50kEBVJYYRDB7Aslty8LhDVBS38c6rpWy7fy7iNRBAce7Z\nDYpC4rY7Rk1vGOctwHfuLL6S85NqfvERpge+kvMgyxgXDu4QZFmzlu5dr8W94+13jkmeGop5uXFN\n8bJ6J0sLxxdlaewYPV8MEG5tAUCdPHwsrzuIs8tPVl4iKpVEeyDECzXttPhDE55vsDOAu7wHOTw1\nxm7XlK6aHM5Vd3OyvIPP7pgzZl/y8l4fL9e144mMLnM7Fi44faxOTuD2TDuaK4y4TCdUFgv2u+6h\n8/ln6Xr1ZVIefWzqY03jvCaNqZYzBdwXiYZ6MCYuQlKPvhMZLV9cdr4VrzvEwhWZ9IrwQmkj7YEw\niVo1t6QmsK/FybvN3VT0+nggNwWH/tqGUD8MMC9fifvQQTwnjo1qjAVBYMO2AnyeEPXV3Rx6p4p1\nt826qkSoqMuF+8gh1I5kTEuWjnpePyHDV1L8kTG+DvCdjeeLTYsWD3o8nju+i46nfkPP7l0kPzxx\n73hubiKvHKqlpLZnQsa4boBJPbIxDtbXobEnorIOJ4Q11sZD2Jm5Ng61OdnT1E1UUViaZOG2TDua\nMTadgVCUF9+v4khxFypJYOfGmaxdkMZkvhYdgTCv1HfQHYyQpNNwX04yaYbp5cXIisKbR+vZfbSB\nf3vmNNtWZnHvLXmDemCHYjJvNnZyotONJAhsy7SzItmKwMRfTEcgzIu17XzQ0Uul28dDualkmm6c\nyGXC5lvp3b8P1769JGzYiDZzanbtuhnjTUszWTxriuVM/SIfySOLfEB80fUVn0M0mdDlXWrhFo3G\nOP1BPZJaxJtn5RelDcQUWOmwsn1GEhpJZGGimdfrOznb4+GnpQ1sy0xiZbIV8SMvecIwFM1GMpvx\nnjxB8sOPjtr+UZJEbt85l1eeOcOFMy2YrbqrmsPvff9dlGgU2223jylDqnY4UKem4i8vQ45EPhLA\nv4ZQolF8JcWokpLQZAxPY1nX3ELPG6/j2r+PxG0T945z0swYtCpKanpQFGXcTd9YyltRVy+x3l4s\ny5eNeG1jbQ9RncQRnUxjYxdGlcTDOcnMsY0dxqxocPLrXWV0uYJkpZj47I45ZDomH/q06jV83Wrg\n7aZuPujo5X+rW9iUlsjGtESkaRQfeXBjPgtnJvHrXaW8dbSB4upuPrtjDlkpZmo9AV6sbcMZipKq\n1/BgXuqUNgQzTDq+OncGe5q6OdzeyxNljWxMT2TTNL+WqUJQqXB87GFafvxDOp/7Axl/+TdTciik\nb3/729M/uwlg9fz0bwcDk8uDAIT9rbha96Iz52FJWYOiKES6OuP9gI8dxblnN50vPEfPa68gB4OY\nlizFvPRSKPvC6RbK63vwrU6jPBDErJZ4JD+Ntam2gQ9WLYrMTTSRrNdQ6fZzwemjwRsg16xHd5Vz\nMB8WCKJIpLuLQEUF+lkFaEYI5fVDUolk5ydRXd5J7cUurDY99mkQzzAatfgvazYhB4O0/s9/I2p1\npH76s8N6sg5FpKODYOVF9IVFaBw3j9Z0uK2N5h//APcHRzCvWHnV+2BPN/zlZbgP7seyei2mERrV\nC6KIoNHgO3MaRZYnXFYiCgJ1bW6qW9ysnps6LgP+lYM1hKMyD23OH7a4Biov4jn2AY4N61DnDZY/\njcVkdp1tpHN+Ii5ZZk6CkU8VpI/JQ4lEY7y4r5qndlcQDMe4c00On7trDgmmqXuzkihQmGAk26Sn\n2h2gzOWj0u0j26THOI0qcXarjnUL0vAFIpyv6eHg+VaqPQEO+ryEYgob0mx8LC8Vq2bqG1pJECiw\nGskx66nxBCjr9VHh8pFt1mFSX/9OWOrkFIK1NfhLL6DLykaTljbquUaj9jsjPX7djDHwbf8IXXnG\nghKN0lP9BtGYE6FaR++re+h87vc4d7+F9+QJApUXiXS0Ixn06AsKMS9fSeL2HUj6OBMyEo7y3Ac1\ndMy1EZAEFiaa+WRBOqmj7NZS9FoW2S10BsNUugOc6nJjUatI1Wv+ZGpKrwSiTof78CEElYRp0ZIx\nz9VoVWTm2qgsbaemvJO0TCuWERisk8FQY9y77318Z05hu307xrnzxh9AEPAcPYLKbJnY+TcA3B8c\nofmnPyTa1Um0u5uY14NpSN71Rkfvu+8QrK0h6b4HRiXPaTNn4D5yiMDFCqzr1iPqJha29IeinKvq\nJjXRQF76cCGPfoTCMZ7dW0lemoV1I5Reek4cI1BeRsY9d6HYLkX4PJEoT5c10WxVIwH35aVyW6Yd\n7Rgbovo2D99//hxnK7tIsen5+oMLWDsvbdrkMxN1apYmWfBEolx0+TnZ6UYriWQYtdO2jqkkkYX5\nSVgSdZTU9tDW7CHWG+LPFmdxywz7tEUVE7V9ryUafy2nOt2oRYFMo+66rsmCIKDNysF1YB/Bmhqs\nGzaOugm+6Y1x+++eovWpJ1CKIii9EQIvXSDa04PKbsc4Zy6WVWuxbduO42MPY99xD5aVq+Oh0j5D\n7ApH+dW5etptGtSCwEP5qWzJsKMehzCk7QtbWzUqLrp8FDu9tAfCzDQbbigiwY0IlS0R96GDhJoa\nSbj1tnE9NINRQ0q6hYsX2qm52ElOfhL6Kyh5utwYK7EYrb96AiUaJe3zX0TUju9xqGwJON95m5jX\nc8PnjeVQiPanf0v3qy8jqtWkfOJxIl1d+M6fQ52cPOU81rWGoih0PPMUKArJH39s1FSCIIoIanU8\nt6woE/aOTTo175xsQhKFMZngdW0eDp5rZXGBY6ApyOXofWcP4bZWch7/JCElfl9fcHr5zcUWOsNR\ntD1B7nPYWZxtH9VIxGSZXUfq+OUbpbh9YbYsyeTL980fsYzqSqEWRebaTKTotX3RPi91ngB50xTt\niykKB1qd7O1xoU0zYJUFnO1+zpZ3YNSryUk1T5/hF0Xm2EykGeKvpbTXR23fa9Ffx8gVBd9OAAAg\nAElEQVSlymwm5vPiLylG1OnRzxq5Ycxoxvj6+/cTQKCmBte+vWg2pSNIAnrtbFL+v0+izZwx7o5Y\nURTO9Xh4rb6TIDL6nhBfXD0Th2XiN7wgCCx3WJlpNvBibRsXnF7qPQF25iYzO+HDq0V8pRBEEfPy\nFTj37MZ/oWQYGWckZGTb2HRnEe+9XsauF85z32NLpkW9yHPqBNGuLqybNqMyj+4RXQ5RrcFQWISv\n+DyR7m7U9htTDD/U3ETrEz8n3NqCNiubtC98GU1KCrqZ+TT88z/S/vRv0WbloE2fmrjOtUS4uYlo\nV1c8vD5OGsGydh09b74Rr/HcdseIRKqhSErQk5JooKzBSTQmj1oLPVa+GOLkLcliQZNoI9jm4vWG\nTs50e1AJAjPag3Chm6INRaPOo7Xbx6/eKKO21R1XcLtj9oCG9tXEvEQT2WYdL9d1UN7r40cXGrgr\ny8Fi+9SNZVcwzAs17TT64mm/+2elU7DGyLHSdp5+u4Kn367gTGUnj2+fjW0axZXm2ExkmXS8UtdB\naa+PH5c0sCMriSVJluvmJdvvuhf30Q/ofuM1LGvWoLJOnO1/U3jGXS8+T7ijGc22dESVjpSln0Fj\nd4z7ZfVHY7xY2877rU6QFazlTjY5LBQWTC3/p1dJLE6yoBVFKlx+znZ7cIWj5Fr0qD4EPYmvBiSj\nEdeB/YCAeenIZJehsCebkCSB2otdNNf3MmtOMpJq8u9vv2esKArt//srYm43aZ/7EpJx4n1XY34f\n/uLzaNLS0eXkTHoOVxOKouA6uJ/Wn/+EmMtFwq1bSfv8l1BZ4psNyWRCnZyM59hRAhXlWNbcMu53\n5nrDdWA/gfIyEu+4C23m2BoEgiQhqNT4zp4GmHAqob3HT2WTizk5NpKsI2/K951tpqHdy33r87AM\nic7EPB66X3kJQ0EhvQuX8Ivieuq8QTIMWv4sJ5mLb/3/7d13fFTnmejx35mukUZ91LuAA5Lo3cYG\n3LDBFeM42fQ4zcneOGWTm+zNzb1372Y3N4nTdjfFjh3HLY6xYzs2NuCOjQ3YNIEEB1DvvY5mRlPO\n/WMGLEBCBUlDeb6fDx+ko/POvAOv5pm3Pe8J0jJiKRkmh0JQ13ljXwO/fe4wHb0eVhanct+meWRO\nYJHWRFmNBuYlxhBvMXG8Z4BDXf00D3gpiI0a12ifruvsbuvhiRNNdA36mZcYw2dnZZ6a9styxrCy\nJI3GdheHqzrZeaiJpFjbhBakjcRiNDA3MYZEq5ljvaHX0hR+LdYIjFwaLJbQmej79xJwuYhZePb0\n3EXbM/Z3d1N3/AS77/4Cnf4ElIAZ5UDtmMr6gjp+XSfHbkV5txHTgI8Fdw2/UjcY1Nn2QS3bP6jD\n7x99X19Qh8FgkC16Ay+P6xVNkALXr8rl7uXnPv/1fHzY1sPW+nZ0fXzlsqJt3JaXQqL17AUa1tw8\nzM4U+g/sI+j1EjSZ+ffH95GX7uCuNYUjJgtYuCKHvh4P5Qea2P58GTdtmotxgr9c7qNH8NbWELN4\nyYgLyTo9Pl6oaUVH57bcFJJsoTfg6JJ5tBHa9xq/es2Enn8qBNxuWh97hL49uzHYo0n/8leJWRja\nqtXlDb0Wf1Dn9pIFoa0Xb7xG6xOPkfaFL0a45ufmOngAjMZTw87d4dcyGNS5PTflrG2GsavCveO3\n3iBh3U1j6h0X5yfy+t56Dld1ouYMn7CjprkPi8lA2jAnS3lqa/CbTLw3byX795zAAFybEVqpXH2s\nDV0f/mCIzl4PD798hPLqLmKizHzx5iKWzI7MwkBFUVjijKMw1s4zVS2Ud7s4drAa8zjmqYOEti5F\nGQ3cWZDKvKSzRxESHFa+9bH5vHWgkb++cZw//L2Mx7Zp49qmNeb66OALBnlLh3etFfz480tJnsIt\nUH/SGog2G9mQ7TxtQVzcVavpfvMNet97l/i11wybhXA4F3QwDuo6b3xYyjt3fJ6AyUw8vVit0SjK\n2N6UFSW0SCumtpfdXR7mXpmLbZgVlK3dbh56qZzj9T1E20xjHkrRAZcvgCcwecdojURRFBxTvN/Z\nbjISZzGjjyMa+4I6x3sH+M3hGjbkOFlyxhCRoig4li2nc8uLuEoPErVoCbqu8+a+Bg5XdoyYLEBR\nFK66YSauvkFqKjrYsfUYa9arExp+6twa+riUsG79WT/TdZ0P23vZUtvGYDD0uv+jrJb12U6WOmOx\npKRgTklloDyUQOZC6Fl6qqtp+sNv8bW1YiucQfqX78WclISu6+xt7+WlM17LjWtvIr2ygt733iVq\nlkrcqqsi/AqG5+/uxlNVSdTsORjsdva19/JibRve8O/Xf5bXsi4rmRVDthkazGYS12+g9YnH6Nr+\nSuhov1HMzonHaFAoq+rkztWFZ/3c5w/Q2O4iL80x7AE21Q3NvLjxHnpjkkmLtrIxN4Ws8Erpk/uL\nsws+Csa6rvN+WTNPvHoct9fP/MIkPnfTbOLOY6X0ZEmwmrlHzeS9lm4OdPQRHOcncWeUhQ3ZTmLP\nkYFLURTWLsykKC+Bv75+IpyzfGrohEZEMRuwTOEugtDWuFCCqIreAe7IS2V2fGjETTEaSfn4P1B/\n/09pfepJsv/7P4/pfSvy7ywj6Pb6eKaymcq4dGy6h2sNu1iYOYu4tPH1DH2Dfh5/uhyL1cj8pacP\nG+m6zjulTfzl9eN4BwMsUZ18ep06oVzZl4KihJhR90GeSdd1DnT08WJtG89Vt3Kky8Ud+Sk4hmw3\nOBmM+/bsxrF0GT/41GKef7eSrbtr+cnj+7hxxdnJAgAMBgPX31bEC08e4OihZgzGUIAeT5Yub10d\nA2WHiZqlElVwek7i3kE/z1W3oPUMYDMa2JidjHtgkDe6enm+ppXyrn425qcSXTKX7jdeC2UTmz1n\nXP8+k0nXdbpff422zU9BIEDCTRtIvu0OFJOJPp//1Dyg1aAwu9OPq9dDU3Y0L9Z3knLtx5n7yhb6\nNr/IzPRsYgvzIvY6RtJfGj67eOFinjjRRHm3C6vBwJ35qVgMCi/UtPJSbRtHu/vZmJdKfHgkJnbV\n1aHe8ZtvkLBu/alh+pHYLCZmZMZxrK6bvoHBs37f69tcBIL6Wck+AkGdN5s6eTM+G10xsDLOyqdW\nzKGn0wWE/n/qqjqx2kw4w2V7BwZ5bKvG3mNtWC1GPnfTbK6al35B7cYwKAqr0hJYlTa1ObRTE+x8\nY9PI5whcTBRF4TMzM3i3uZtXGzp49HgjS52xrM92YjUasM8pImbRYvr37aXvg93ELhv5iN+TLrhg\nrOs6+8Nv7t5AkNxgPavNH5CRuYrY1PEf2XVobwMet48lq/KwDslH293v5ZFXjlJa0YHdauLLtxSx\nvCj1gvoluRgoisLC5FjyHVH8rbqFoz0ufn24httyU5ibGHpDsmZmYcnMwnXoIIGBAcx2O3etmcGC\nGcn88aWzkwUMZbYYWX/XXLb8tZTyA024+ge5/rYizGPcJ9m5PXwgxI2nHwhxqLOP56tbcQeCzIi1\nc4XJyntPHcbr8ZNoNdA1J4FjwE8/rCA1MIv4bDsHt9YRfciP1WbCYjVhsRmxWE0YFAVdBx0d9FAb\n1s/4e+h1o8nAnHlp4zo/O9DfT/MjD+E6sB+jw0HaPV8+NZR7uLOP52vaGPAHyLKYse5pwdXhxmQ2\nkFLXR+ecBFqTbbxx4wYStG52b67Gaq0jNj4KR5yN2PgoYuNtxMbbSM+KwzyGPMNTwXVgP7W5s9gV\nn8dAt4t8RxSb8lNJCAfdPEcUz1W1crTHxW/CC48WJDlCveObNtD65ON0bXsF5113j/pcJQWJaHXd\nHKnpYtmc01dVD7d4q8Xt5ZnKFhoGvMS4XVz13qus+v73Tptj7e4YoL/XS+FsJwaDwoHj7Tyy9Si9\nrkFmZcdzz4Y5U7JSWkSGQVG4Oj2BWXF2Nlc280FbLyd6B9iUn0a+I4rku+7GVXqQ9s1PEzN/4ag7\nOC6oBVz9Pj+bK1t4u7kLo6KzKriL5aYDOJNWEZc1/qG1Qa+f7c+XYzSGelimcM/rg6Ot/Orpg9S3\nuSjOS+Dbdy9gZvbwyeDF2NhMRuYnOYg2GTnWM8DBzn46PD4KYqMwGwyhJf/lZVjS0rHlhObtk2LD\nyQI8fkorOnintAmjQWFGZtxp/xdmi5GZRSm0NfdRV9lJfU0X+TOTRw3IpoE+qv/wAJb0dJwf+wSK\nouD2B/hbVSuvNXaiKLAh20lOq4d3t2gEgzqz56WTnGAn1R3E4PHTE2OizxmFOyYaf6uf3g43HW0u\nWpv6aKrrob66i7qqLuqrP/rTUNNNQ003jbWhP011PaE/9T001/fQWNvNkYNN2KLMJKfGjLz1xe3G\nU1VJ//79ND/8R7zhIdysb38XW04ubn+A56tbebWhE12HBZhxv1ZD0O1j2dX5bPjYPBYuzmJufAwm\nX4DGQABXahQGqw9H9wADLp3ONhctDb3UVnRyvKyVIwebsFhNJKWMXK+pMDDg5oWaNvYuX4uOwk3Z\nydyam3LaQQdDFx5pp7YZeilwRBGTm0PPzndwHz9G3NWrR33js5gNvH2gEZvFxMJZp+9lfvtgIzXN\nfdx+VT6OaAs7W7p5qqKZHp+fRfFRXPnHX5CemkzcyitP2z53rKyFuqouZi/I4JUDDWx+q4JAMMim\nNTP4zI3qZXXM5uUkxmxicXIcOjpa9wD72nsZDAYpTE1G8flCJwcajadG1aZlAZeqqgbgt8A8wAt8\nUdO0irGULe/q57nqVlz+ALlRcOXgSzgMvRiPxxC3ePWE6nNobwNej59lV+VhtZlweXw8sf0Yu8pb\nsJgMfOqGWaxdmClBeJIYFIWVqfHMjLOzubKFA519VPa5uTM/hdyly+l4/m/0fbCbuCtXnSpjs5j4\nzDqVBTOS+dMrR3j27UoOnujgnpvnkJrw0eIZi9XE+rvm8tYrGscOt/DcY/vZ8LF5xCWM3NNofGlL\naDj3hptQDAaO97h4tqqVXp+f7GgbG3OclL1VxYHDLdijLazbWEzaGQfPt3sGeaayhdp0CMbq3Jab\nRn5iIoNeP16vn0GPH10HRQmNEhD+O/Q9gHLqZyf/7mjrZ9dblby99RgnjrSy5iYVu+LFW1uLp7YG\nb20N3tpafG2tH1VEUUi67Q4SN9yCYjBwomeAZ6ta6PH5yYiykFnVT3NZPVF2M9fdWkRWXsKpfzdn\nmoP1aQ6WewZ5prKZmswkvPFWbjK4KVq8hN5uN73dHjpa+zm8r4G3tx7j0N4Grry2kKy8qd9uU9E7\nwGatjt5Zc0kdHODji2aTGjV8MD258Kgg1s4zlc2Udbmo7qtlY14Kaetvpu3Jx+nathXnpo+d8zlz\nUh3ERJkpqz47NWZNcx8mo0KUw8JDWgNVfe5T6SzzmmupH/Riy8076zHrqrroRefx3TV09nnPK52l\nuLiYDAo3ZCUzOz6azZUtvNPcjdYzwKa112Pc+S5d214hbtXV59weqYxnsc5oVFXdCNysadoXVFVd\nDvxA07TbR7hdb2vrw+MP8FJdG/vaQ3v01ib5Keh+BiUYZPDlJjLu+saE5ukGvX4e/10oh/Wn7l3B\nscYe/vTyUbr6vBRmxHLPzUWkJV54R3JdKk4mAXi9sYOgDitS4ih68gGCVZUU3v9rjI6zV172u308\nvl1jz5FWLGYDd18zkzULMk57o9R1nT07qtj3fi02u5kNd80lJf3sOcLAgIuq730HxWoj88f/j20t\nPexu7cGowLUZSSyKsfPqc2W0NvWRku5g3cYSYkZYuBfQdba/u4ed5jiCRiPLnHGsD+cxnwhfRzud\nRyp4f38PTS4LBt3PjPYPyeo5eip9viE6GltOLtbsHKw5OdgKZ2BxpjAYCLKtvp33W3swKLAiNoae\nN2vp7XKTnhXH9bcVnXNfdlDXebuygdfb+ggaTSyyKdxSVHBqG4ir38ueHVUcLW0GILcwiZXXFJIw\nzKri8+ULBk/lG1Z0nXl732HDDauJKZwxpvJBXWdnSzfb6zsI6DqLEqOZ8cff0kwazekLyMhJYM1N\nKhbr8H2O379wmD1HWvm/X1xOZnJo8Y0/EORrv3ibtJmJGLNj8IbTWd6el0KM2UTn1pdpf+Zp0r/6\ndRxLluJ0Omhr68Pt8fGvv36XJl3HoCisX5nLrVfmnfeZzuLiMxgI8kp9O7vDv6NX+vrJf/g3xC1d\nRvqX78XpdAzb+5vsYHw/sFvTtKfD39drmjbSZkF9V0ULz1a10D3oJ8NuZUNiJ8bmF1EUE57n6zDr\nieT+7/87oZ7rwT11vPdGBQuvyOG4x8eb+xowGhRuW5XPTStyhl0lKSZfg8vD5soWWj2DJAQGWfn3\nJyhZdz3xq9eOWGZ3eQuPb9dwefyU5Cfy+fVnJwso29/AO9uPYzQZuOG2YnJnnP6Js/2F5+h88QUG\n7/4021IL6fD6SA0nqzd0etj2XBkDrkHUklSuvnEWplEy9ww2N7P3l79k5/q76bQ7SLSauSs/lVzH\n+OYAe3a+Q8sjD4OuowMtMQUcS1mBz2AhKcrPqsVxOOcUYEpMPKvd1/V72FzVTLvHh9NmYVnASNlr\nlQT8QRYsz2b56vwxL26rKj3EMw09dCWnkmA2sqkwnfwhr6WtuY/33qigsbYbg0GheGEGS1blDbsb\nYSLqXR42VzbT5vGRbDVzxQuP4exqo+DnvzrnAR7DaR7w8tTxRloH/RjdfhLLu4jq9qKjkJBkZ93G\nkmE/TLxb2sTDLx/h49fO5IaloQxlRxu7eeBALTZnaJ/qmQkxmh74HX17dpP37z/F4kzB6XTw4aFG\nfvu3Q7T1eoi1mvhvd8+nMGP0LVbi0jZ0JC6lu50rt26m5KtfJeeKxdMSjB8EntU0bWv4+xogX9O0\ns/b+PPDeCf39hg4UYFlKHMXGGlytOzEYbZianPS99j7Jt91BzJKRzyseSVAP8tJTpfR5fLTHWGjr\n8ZDpjOaLG4omdH6yOD++YJBXwz0ggkEW1mlctWHdOY9R6+338sxrJzhW04XNauSW1QVkn7G4q7mx\nlwO7atGDOsWLM8jJDwVkd8UJOl9+iYYZRRydEUoEsTg5lpWpcdQca2fvuzXous7CFTnMLB7boj1d\n12n45c/xeb3Uf/Ze9naGzrpdmhzL7IToMR0J52vvoO3pJ8FgxLF0KWZnCubkFPwGM4f3N9BS34ti\nVJhVnEL+zORTgfXkXNSeth4AFiTEYK3ooeZYO2azieVr8snKHf9K2I5Xt7O7y8WxOQtBUVicHEvR\nkNei6zotjb0cLW1ioN+H2WJgRlEKOQVJE97zDXCsx8Xu1h50YEGSgyW+Ptof/AMxi5eQfPvGMT+O\nrus01fWgHW6mubGX/qxoXJmhHm5BRTkJTpWaqh5MZoV5y3JIOyMXdf+Ajz9uKSc3LYY7riqkzT3I\nq/Xt+IE4xcCmmek4zljMVv/r+wm6XGT/4H+iKApH63v4y3aNQFAnBfjSnfMonOBpdOLS4/YHTp0A\naPT5WHFsP5/77lenrWe8S9O0zeHv6zRNGzYp7i3feWHynvgcFGDd8hzuuCofs5y4FFFVfW6e2q/R\nFzW2DFi6ruNucNF3vBs9OC3NRYhxS3BYKTQYMLl8fOG+KzFN4olI4tJwqLOP57Q6PCYzD65fNGww\nnuw9DDuBW4DNqqquAEpHunFuSSppMVY8PbW4+5swGK3EO4vwNnfQe/gw0QX5OGbNGncFdHSOljbj\ncnmZtyiL61fmUTxMoncx/ZxOBylH97Plgw8xLlhITGHB6IVynfSVeDlxtI2Af/iAHAgE6elyE/AH\nMOl+rH4X1uRkomKiyYqNQvcHqTzWRl+vhyi7hRmq87RtbmPlaWuje+8+ogsKcMyaiV/Xaeh1n0pK\ncS6DnZ34+/sxORxYEkbuxQaDOq4+Dx63DxSF6BgL9mgLFqORGH+QuhOdBIJBnKkOsvMSz/tkn4DX\nS8fO9/AHAnjmLcJvHnmPfTAYZKB/ELfbB7qO2WoixmEddYj/TGajQrbDjilc9/Z3d+IfGCD12mvO\neZjI4KCf1uY+2lv68PuDKIpCYnI0qRmx2IfsFfb7/WgHywnYorClp+P3B+jtchMIBLFYTcTGR50a\nDenq89DnGiQl0Y7daqKlsY++Xi/XL88966zcwc5OOvd8QHR+Hg5VBSA2xsK1CzN58Kdvk6c6Sc8Y\ney5icfm4xulgbryZl17eAQx/gt1kB+PngOtVVd0Z/v7zI934488tQ9v7F1ydBzDbnDgLP4nR7KD2\nX/4X3o568v/pYxNKzN9Y103H+3Usm5nCjTeEgnlbW99EXouYAtbZJSx69DGC+3fi/MQnSbjmutEL\npQIz0855S19bNy8+8DY9xjjS46LYsGkxGZkJlB9qZOvfDpPUO8hSNYVrNsye8D7aoDeTitcexNLY\nRO6X1o25XN+He2ja8giWrGxyvvY/MZwj4J1UfaKdHVuP4ap3kZyikJJho7y8iXyzgdUb5jCreOQT\nh8ZrIFun/hc/w/ShRu6P/g/GmHOv/u3qcPH+GxXUVHRCq4fFV+aydFXehNZ2DLa2Uv3Y60QvWEjm\nNcOfctPZ5mL/rlpOHGnFFtSZEWWmeFkGJYsysI+QxaqhbAuufQfJ/+n9mBOT8Lh9vPb3cuqquojz\nKdy4sYREZzSHKzv4xdMHmTc7jY2rC/j6L3aQlRLDx9eenZmra/tx2tp2k3bHQmKXhT5IOp0O3nn9\nGACpmbHyXiNGZrBx8803jPjjSQ3GmqbpwL1jubey9AlcnaVY7Bk4C/8Bo8nOwDENb10tMYuXTPiE\nnIN76gCYv+zcSeZFZJgcsWR99/s0/Op+2p58nEBvD0m3bTyv7WVBn4/uR37PguoTaCWbaOqx8fwT\nB1i2Kp/tL5Th9wdZdlUei67IPa/nMVitRM1UGThShr+7G1P86L0gX1sbLX/+E4rFQsZX7h1TIAbI\nm5FM+hfjeP/NSo4cbKK9tZ/4JDvr7igmMXnsB12MhX1OEUm33k7HC8/R9Iffknnft8+Z9jMhKZr1\nd82jrqqTt7ceY+/OGtqb+7n2ltnjHnFwHdwPMOyZy7quc3hvA++9WUEwoJOQZGfe0ixmFaeOOhRs\nLyrBVXqQgfIy4lZdjS3KzPq75rFnRxX7d9Xy7KN7uWbDbGYWJmEyGjhc1cnKkjQG/cER15V4aqoB\nztrWVFfVCZyeAlOI8YrYkuLullKsMTmkzPg0RlNopWP3668CkHDdyJ8ezvmYnQNUH+8gJd1BWpas\nZrxQ2XJyyf7+DzE7U+h86UVaH3sEPRCY0GPpuk7Lnx/GrR0lfuF8bv369cyel0Z7Sz8vP3sIxaBw\n450lLL5yYj23M0XPDWW9cpUdGr1ufj9ND/yOoNtNyic/jSV9fEcYWm1m1tykcsvH57N0VR6bPrto\n0gPxSYkbbiF6wUIGjpTT/Kc/ogdHH3rPzk9k0+cWk5WXQE1FB8/+eR+dba5xPW//wVAKzOh580+7\n7vX42Pa3Mt597QQWi4l1dxRz9xeXUrQgY0xzstHFxQAMlB0+dc1gUFixpoAbbi8CYPvz5ezfWc2s\nrFjqWvs5VNEBjHxsore2BoPNhtn50eEOwaBOXVUX0Q7rlGz/EpePiAXj1Lw1OAs/icEYGmbydXTQ\nv38f1pxcbDOGH64aTemH9QDMX5YtiTwucJaUFLK//z+w5uTSs+NtGn//XwR9YztSc6iOvz9P3673\nsRUUkHbPlzGajKy5SWX56nxyC5O48zOLyJ/E1a32cApK16HRg3H783/DU1WJY8VKYq9YNer9I8nK\nS2DJqrwpTVOpGAykf+mr2Apn0Ld7F+3PPD2mcrYocyjT14pserrcPPvoXiqOto5eEAi4XLiPadgK\nCk4797W5oYfND39I1fF2MnLi+dgXllCgOsf1O21OS8eUmIirvOysDxaFs1O48zOLiUuIYv+uOpJ6\nBzEC2z8MjaoN1zMOejwMNjdhzck9betVU303Xo+f7PwEec8R5yViwThr1gYMhiG5ot98HYJB4q+9\nbkKN2uP2oZU244i1UqDK1oKLgSkujqzvfp+o2XNw7d9Hwy/vJzAw9p5Vz8536HzxBczJTjL+8Zun\nUiAqisKilbl89mtXkDDJPUlLegamxCQGysvO2Zt3HT5E19aXMaekkvqpz1wUb9QGq5XM//ZNLGnp\ndG3fStf2rWMrZ1BYsabwtB7nrrcqCI6yAt51uBSCQaLnLwTCeel31fL84/vp6/WyZFUet3x8/jmT\nmIxEURTsRSUEXS68tTVn/TzRGc2dn11EbmEini4PxSgM9g+iKJCdcvacubeuDnQd6xlD1BVaGwA5\nMkQtztMFkfki6PXS887bGGMcOJYtn9BjlO1vxO8PMndJ1rhO9RGRZYyKIvO+bxOzeAnuYxp1P/0J\n/u7uUcsNHCmn5dFHMNijybzvW6Oe1DNZFEUheu5cggMuPFWVw97j7+6m+aEHUEwm0r9yLwbbxXM4\ngDEmhsxvfQdjfDxtTz9F7+5dYy57Zo9zy9OloVXhI3AdCM8XL1iIe2CQLZsPseutSqKiLdz6idDQ\n/PmsFo8uDu0xdw0Zqh7KajNz06a5LL4iFysKRSiUmEyU7qql+ng7rj7vqXs9tdUA2HJzT3uMiqOt\nKAqn0o8KMVEXRNTq272LoMtF3Oo1Y17gMlTAH+Tw3gYsViNz5qdPQQ3FVDKYzaR/5WvErb2Gwfo6\nan/yrwy2NI94v7ehnsbf/geKopDxj98Y91zs+YouCR0D5zp89s49PRik6Y9/INDXR/Kmu4fNYXyh\nMyclk3XfdzBERdH88IMMHCkfc9mhPc766i6eeWQv7S1nrzDW/X5chw9hTnbS5o/m6Yc/pK6yk+yC\nRO76/BIyJ5DE5Ez2OUWgKKfNG59JURSWXZ2PITsOH2DzBflwZw2vPHuYR//rff78H+/x8uZSDpT1\n0mbPxp+ceeq8b6/HT31tNynpsRPaKifEUBEPxrqu0/X6q2A0Erfmmgk9xvHyFq+8uJQAABGvSURB\nVAZcg8yZnz5iHlpxYVMMBlL+4dMk3XYH/vZ26n7yYzzV1Wfd5+/upuHXvyTodpP6+Xuwz1Knva72\nOXPAaBx23rjzlS24jx4hev4C4q8dw7atC5Q1O5uMr38DRVFo/K/f4BlmqHfEsuEe55Irc+nr8fDc\nY/s5VtZy2j0DxzQCbg81eVfx0lMHcbsGWbGmgA13zcUePTnniRtjYrDm5uGuOEHQc+4D7RcuyqAU\nnZlr8ll/11yWXpUXyoJmVKip6OToQBKlGdfy12eq+PN/vseWp0vZse0YelAnO196xeL8RfwIRbd2\nlK5tr+BYuoy4VeM/JlHXdd7YchSP28d1txZhtUkwvlgpioJdnY0xLo7+Dz+gd/cubPn5WMKrV4Me\nD/W/+Bm+5iaSbt846h7locfbTWo9TWYGtKN4Kk4Qt+aaU3PV7uPHaX74QUzx8WR9859GPcbvQmdO\ndmJJTaNvzy76D+zDsWgJxuixzcErikJmbgLJqTFUn2jnRHkrgx4/mbnxGAwKTa+8xh7fDOoDycTE\nWtnwsXnMnILzxP3t7aFFYoUzsaSNvFc9Izma2TkJLC5KISEpmoyceGYUpTB/aTZFJckoLz5KrMNI\n7KwZeAZ8tDX309keWt+wYm3hiIeMCHGmaTlCcSK6wtuZ4q+9fkLl66u76GxzMaMoBUfc2A9rFxeu\n+NVrMcY4aH7w9zT8+hek3/NlYpYspemB3+GtrSF21VUkbrglonWMLpmL++gRBsoOEbvySgL9/TQ9\n+DvQddK+9NVRE2dcLBxLl+Hv6aHtqSeo/9XPyfn+D4c9cWsk+TOT2fiZxWz722FKP6ynvaWPogUZ\n7KhPYtBuJW9GItfcPGfKhnntxSV0vvwSA2WHh93LfJKiKMweYWjc0NlCkquewtw5pGwMzUO7BwZp\nb+knJsZGglO2NInzF9Fhal9bG64D+7Hm5WMrODvjzVicSvKxVJJ8XEoci5eQ+c3vYLBYaHrw99T/\n7Ce4Sg9in1NM6qc+G/HVydFDtjjpuk7znx/G39lJ0q23R2TofColXHc9CTeux9fSQsN//JKg1zt6\noaHlk+xs/Mwi8mcl01jXw2svHsGnmJhrb+HGO+dO6XxrVOEMFKsVV/nI88aj8YaTfVhzPlq8FWW3\nkJ2fyKyiycuEJi5vEQ3G3W++DrpOwrXXT+jNtaOtn7qqLtKz44Y901Zc3Oyz55D13e9jdDhwHz+G\nJTOL9Hu/fs7sUNPFkpmFKSEBV/lhul9/Ddf+fUSpsyPeY58qyXfehWPlFXgqK2n6w2/HnaTFYg0l\n7lixtgBntJ8l9VuYt+L8MqKNhWIyYVdn42tuxtfRPqHH8NSE5stteXmTWDMhThexYBxwu0PbmWJj\nJ3RMIkDpno+SfIhLky0nl+wf/JCEG9eTed+3MdovjCFBRVGwl8wl2N9P21+fxBjjIP1LXxn3WbwX\nC0VRSPvsF7AXh9JMtjz2COM98U1RFBYuz2FZ73vE+ntOjS5MNXt4i9NAWdmEyntrqlHMZixpslND\nTJ2IvXO0vrWDoNtN3Oq1GMzjH6Ya6PdyrLyFuMQo8mbIqUyXMoszBeemj2FOvLASK5wKJrpO6he+\niCn+0l5Vq5hMZNz7j1hz8+h99x06XvjbuB/D392Np6oS+ywVo31qUnue6dR+4wkMVQd9PryNDViz\ns895opQQ5yti431NL70MRiPxa9ZOqPyhfQ0EAzrzl2ZFfP5QXJ6ii0uwZGTiWLqMmDNyK1+qDDYb\nmd/4FnU/+TGdL72IKS6BuDVr0QcHCXo8BD3u0N9u90ffuz+67q0PrfE4mXVrOphT08JZ08rRg8Fx\njV4MNtRDIHBW5i0hJlvEgrG7vh7HipWn5aQdK58vQNm+RmxRJmaVnPtoPSGmisEWRd6//DjS1Zh2\nprg4Mr/1T9T9+7/S+sSjtP7lcRjDwRInKSYTMQuHP9N1KiiKgr24mN53duCpriaqYAznaIedmi/O\nyR3lTiHOT0RXwsRfM7HtTNqhZrweP4uvyMU8hhNchBCTy5KSQuY3v03b00+h+/0YoqIw2GwYbFFD\nvrZhiIrCeNq1KEwJCdO+9Su6uITed3YwUH54XMH41Epq6RmLKRaxYKx+7zvo4/ilOEnXdUo/qMdg\nVChZNL1pEIUQH7Hl5pH93e9HuhpjYp/9UWrMpJtvHXM5T20NismENSNzCmsnRAQXcCVfecWEylUf\n76Cny82s4lTsMZL1RggxOmNMDLa8fNyVFaOmxjxJ9/sZrK/DkpV9QWynE5e2i24fxsEPQgtA5kmS\nDyHEONiLiyEQYODo0THd721sQPf7Zb5YTIuLKhi3NvXSVNdDdn4CSc5LI92gEGJ62IvC+43HuMVJ\n5ovFdLqogvFBSfIhhJigqIJCFKsN1xiTf5w8qerMM4yFmAoXTTDu6/FQcbSVRGe0HOQthBg3xWTC\nPmcOvpZmfO1to97vrakGoxFLpkyJial30QTjQ3vr0XUkyYcQYsKii4oBcJWfu3esBwJ46+qwZmRO\nKEOgEON1UQTjE0daOfRhA/YYCzPllBQhxAR9lKf63PPGg02N6D6fzBeLaXPBr9cv29/Ajm3HsViN\nXH9bEUbTRfH5QQhxATKnpGJKTmbgyLlTY57KvCXzxWKaXLCRTdd19r5Xw45tx4mym7n1EwvIyB5/\n6kwhhDhJURSii0oIDgzgqa4a8T5vePGW9IzFdLkgg7Gu67z3RgV7dlThiLVy+6cW4kxzRLpaQohL\ngL04NG98rqFqT001GAxYs2TnhpgeF1wwDgaDvLnlKKUf1JOQbOf2Ty8iPvHCOMNWCHHxO5Uac4RF\nXHowiLeuFkt6BgaLZZprJy5XF9Scsd8X4NUXyqk+0UFKhoMNd83DFiUrGYUQk8cYHY0tvwB3xQkC\nbjfGqKjTfj7Y3Izu9cp8sZhWF0zP2Ovxs+XpUqpPdJCVl8CtH58vgVgIMSXsxSUQDOI+euSsn3lr\nqwGw5uRNb6XEZe2CCMYDrkH+/pcDNNb1UDjbyfpNczFbLqhOuxDiEvLRfuOz540/WkmdN51VEpe5\niEe8vh4PLz51kJ4uN0UL0rnqhlkYDJLUQwgxdWz5BRhsNgaGSY3prakGRcGaLYu3xPSJaM+4s93F\nc4/vo6fLzcKVOVy9TgKxEGLqKSYTUbPn4GttYbCt9dR1PRjEW1uDJS0dg80WwRqKy03EgnFDbRfP\nP74fV98gK9cWsmJ1gaS5FEJMm+iT2biGrKr2tbUS9HiwyrGJYppFLBg/+rv3GfT6WbteZcFyGQ4S\nQkyvU0cqDtlv7AkfmyjzxWK6RWzOOBjUWXdHMfmznJGqghDiMmZOScGc7AylxgwEUIxGvDUnM29J\nz1hMr4j1jL/1o+slEAshIkZRFOzFxQTd7lOpMU/2jGWYWky3SesZq6qqAPXAsfCl9zVN++eR7rdH\nW3ANeCfr6YUQYtzsRSX0vP0WA2WHsRUU4q2twZyaelYiECGm2mQOUxcCezVNu3USH1MIIaaMfc4c\nUBRcZYdxrLyC4MAA0SVzI10tcRmazGC8GMhUVfUNwA18S9O0Y6OUEUKIiDHao7EVFOKpqsR9JJSN\nS4aoRSRMKBirqnoP8M0zLn8N+DdN055VVfVK4HFg2XnWTwghppS9qBhPxQm6Xt0GyEpqERkTCsaa\npj0EPDT0mqqqUYA//POdqqpmnH/1hBBiakUXl9D54gsMNjUC0jMWkTGZw9Q/AjqBn6mqOh+oHa2A\n0ylnFIupJW1MjEZPnE+j3U5gYABbWippeWnjKi9tTEyGyQzGPwEeV1V1PaEe8udGK9DW1jeJTy/E\n6ZxOh7QxMSZR6hz69+/FlJk9rjYjbUyM10gf3iYtGGua1gPcMlmPJ4QQ08VeUkL//r3Y8vIjXRVx\nmYr4qU1CCBFpsVesQvf5ib1yVaSrIi5TEoyFEJc9g9lMwnXXR7oa4jIW0SMUhRBCCCHBWAghhIg4\nCcZCCCFEhEkwFkIIISJMgrEQQggRYRKMhRBCiAiTYCyEEEJEmARjIYQQIsIkGAshhBARJsFYCCGE\niDAJxkIIIUSESTAWQgghIkyCsRBCCBFhEoyFEEKICJNgLIQQQkSYBGMhhBAiwiQYCyGEEBEmwVgI\nIYSIMAnGQgghRIRJMBZCCCEiTIKxEEIIEWESjIUQQogIk2AshBBCRJgEYyGEECLCJBgLIYQQESbB\nWAghhIgwCcZCCCFEhEkwFkIIISJMgrEQQggRYRKMhRBCiAiTYCyEEEJEmARjIYQQIsIkGAshhBAR\nJsFYCCGEiDDTRAuqqnoHsEnTtE+Gv18B/ArwA9s1TfuXyamiEEIIcWmbUM9YVdVfA/8GKEMu/w74\nhKZpq4DlqqoumIT6CSGEEJe8iQ5T7wTuJRyMVVWNBayaplWFf74NuO78qyeEEEJc+s45TK2q6j3A\nN8+4/DlN055WVXXNkGuxQO+Q7/uAgkmpoRBCCHGJO2cw1jTtIeChMTxOL+AY8n0s0D1KGcXpdIxy\nixDnR9qYmGrSxsRkmJTV1Jqm9QKDqqoWqKqqADcAOybjsYUQQohL3YRXUwN6+M9JXwWeAIzANk3T\nPjifigkhhBCXC0XX9dHvEkIIIcSUkaQfQgghRIRJML7Aqar6lqqqs0b4WbWqqpbprpO4tEgbE1NN\n2tjoJBhf+HROT65y5s+EOF/SxsRUkzY2CgnGF4f/rarqVwBUVZ2tquqbka6QuORIGxNTTdrYOUxb\nMA4PU6jT9Xzi8iNtTEw1aWNiqkxnz/jMrVBiBKqqxqiqenLbmcLp/24jDfUIaWNjJm1swqSNjZG0\nsfE5n33GE+FUVfXngA1IB36oadoLqqqWAm8B8wj9h90WTiRyuXoE+E9VVXcATmAroX8vgEWRqtRF\nQtrY2DyCtLGJkjY2No8gbWzMpnvOeD5wv6ZpNwBfBr4evu4AntQ0bQ3QANw0zfW60NwP/AzYDWwG\nngLWh+dYFiKfzM9F2tjYSBubOGljYyNtbBymtGesqmoM4NE0zR++9C7w/fABFPoZz78//HcdoU+c\nly1N094Hlp5xedkw9+VPT40uXNLGJkba2NhJG5sYaWPjM9U940eAVaqqGoAU4JfAo5qmfYbQcM7Q\n55dPSWIiHkHamJhajyBtTEyxqZ4zvh/4TfjrzcBR4Oeqqt4H7AISRygnDVqMlbQxMdWkjYkpJ7mp\nhRBCiAiTpB9CCCFEhEkwFkIIISJsUueMVVU1Aw8DuYAV+FfgCKEFEEHgMPB1TdP08P1OYCdQomna\noKqqccDjhLYIWIBva5q2azLrKC5uk9DGooEngXhgEPispmmN0/06xIXrfNvYkMeZTWhOOWXodSGG\nM9k9408CbZqmXQ3cCPwXocUP/xy+pgC3Aaiqug7YTmh14knfAl4N79P7XLi8EEOdbxv7IvCBpmmr\nCX3w+9401l1cHM63jaGqamy4jGca6y0uYpMdjDcDPxry2D5gkaZpO8LXXgGuC38dAK4FuoaU/yXw\nQPhrM+Ce5PqJi995tTFN034N/Fv421xOb39CwHm2MVVVFeAPwA+Q9zAxRpM6TK1pmgtAVVUHoQb9\nQ+DnQ27pB+LC974Wvndo+Z7wtTTgMeC+yayfuPidbxsLXw+qqvo6UALcMPW1FheTSWhj/wvYomla\nafi65GEWo5r0BVyqqmYDbxDaFP8XQnMsJzmA7lHKzwVeA36gado7k10/cfE73zYGoGnatcDVwLNT\nUklxUTvPNvZJ4J5w2sc0YNuUVVRcMiY1GKuqmkpo/uR7mqY9Er68X1XV1eGvbwJ2DFc2XL6I0CfR\nT2iaJg1YnGUS2tgPVFX9dPhbF+Af6V5xeTrfNqZp2kxN09ZqmrYWaEZGX8QYTHYGrn8mNHzzI1VV\nT8653Af8RlVVC1AOPHNGmaFZR/6N0Crq34SHd7o1TbtjkusoLm7n28YeAv6squoXACPw+Smur7j4\nnG8bG8t1IU4jGbiEEEKICJOkH0IIIUSESTAWQgghIkyCsRBCCBFhEoyFEEKICJNgLIQQQkSYBGMh\nhBAiwiQYCyGEEBEmwVgIIYSIsP8P7z0xBBd79DgAAAAASUVORK5CYII=\n",
      "text/plain": "<matplotlib.figure.Figure at 0x10e1f20d0>"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Load the predictors: SST EOFs"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "sst = pd.read_csv(os.path.join(predictor_dir, 'ERSST_EOFS1_9_-40_40_nodetrend.csv'), index_col=0, parse_dates=True)",
   "execution_count": 18,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla.tail()",
   "execution_count": 19,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>msla</th>\n      <th>trend</th>\n      <th>detrend</th>\n      <th>seas_msla_detrend</th>\n      <th>seas_msla</th>\n      <th>m_cat</th>\n      <th>s_cat</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2014-08-31</th>\n      <td> 10.2</td>\n      <td> 4.531784</td>\n      <td> 5.668216</td>\n      <td> 4.690424</td>\n      <td> 9.200000</td>\n      <td>  2</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-09-30</th>\n      <td>  9.0</td>\n      <td> 4.553992</td>\n      <td> 4.446008</td>\n      <td> 5.301549</td>\n      <td> 9.833333</td>\n      <td>  1</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-10-31</th>\n      <td>  NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td>  7.1</td>\n      <td> 4.598409</td>\n      <td> 2.501591</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>  1</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td>  4.6</td>\n      <td> 4.620617</td>\n      <td>-0.020617</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>  0</td>\n      <td>NaN</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            msla     trend   detrend  seas_msla_detrend  seas_msla m_cat s_cat\n2014-08-31  10.2  4.531784  5.668216           4.690424   9.200000     2     1\n2014-09-30   9.0  4.553992  4.446008           5.301549   9.833333     1     1\n2014-10-31   NaN       NaN       NaN                NaN        NaN   NaN   NaN\n2014-11-30   7.1  4.598409  2.501591                NaN        NaN     1   NaN\n2014-12-31   4.6  4.620617 -0.020617                NaN        NaN     0   NaN"
     },
     "metadata": {},
     "execution_count": 19
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "sst.tail()",
   "execution_count": 20,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2014-10-31</th>\n      <td> 0.079347</td>\n      <td> 1.454287</td>\n      <td>-1.560152</td>\n      <td>-1.145130</td>\n      <td> 0.248758</td>\n      <td> 1.713660</td>\n      <td>-0.120309</td>\n      <td> 2.025361</td>\n      <td>-0.935043</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td>-0.281475</td>\n      <td> 1.749576</td>\n      <td>-1.454797</td>\n      <td>-0.888209</td>\n      <td> 1.478064</td>\n      <td> 2.160027</td>\n      <td>-0.040456</td>\n      <td> 2.629284</td>\n      <td> 0.620553</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td>-0.286381</td>\n      <td> 1.577672</td>\n      <td>-1.288308</td>\n      <td>-0.375071</td>\n      <td> 2.845861</td>\n      <td> 1.725399</td>\n      <td>-0.417040</td>\n      <td> 2.992877</td>\n      <td> 0.162500</td>\n    </tr>\n    <tr>\n      <th>2015-01-31</th>\n      <td>-0.238318</td>\n      <td> 1.626915</td>\n      <td>-1.564301</td>\n      <td>-0.006681</td>\n      <td> 2.554465</td>\n      <td> 1.441429</td>\n      <td>-1.511443</td>\n      <td> 1.296484</td>\n      <td>-1.100238</td>\n    </tr>\n    <tr>\n      <th>2015-02-28</th>\n      <td>-0.355435</td>\n      <td> 1.346813</td>\n      <td>-2.194495</td>\n      <td> 0.065253</td>\n      <td> 2.292541</td>\n      <td> 1.085200</td>\n      <td>-1.190221</td>\n      <td> 0.226652</td>\n      <td>-0.375395</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2014-10-31  0.079347  1.454287 -1.560152 -1.145130  0.248758  1.713660   \n2014-11-30 -0.281475  1.749576 -1.454797 -0.888209  1.478064  2.160027   \n2014-12-31 -0.286381  1.577672 -1.288308 -0.375071  2.845861  1.725399   \n2015-01-31 -0.238318  1.626915 -1.564301 -0.006681  2.554465  1.441429   \n2015-02-28 -0.355435  1.346813 -2.194495  0.065253  2.292541  1.085200   \n\n                EOF7      EOF8      EOF9  \n2014-10-31 -0.120309  2.025361 -0.935043  \n2014-11-30 -0.040456  2.629284  0.620553  \n2014-12-31 -0.417040  2.992877  0.162500  \n2015-01-31 -1.511443  1.296484 -1.100238  \n2015-02-28 -1.190221  0.226652 -0.375395  "
     },
     "metadata": {},
     "execution_count": 20
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### calculates 3 months averages of the PC scores"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "sst = pd.rolling_mean(sst, 3)",
   "execution_count": 21,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "sst.head()",
   "execution_count": 22,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1980-01-31</th>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n    </tr>\n    <tr>\n      <th>1980-02-29</th>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n    </tr>\n    <tr>\n      <th>1980-03-31</th>\n      <td>-1.107874</td>\n      <td>-0.591140</td>\n      <td>-1.028791</td>\n      <td>-0.779700</td>\n      <td>-0.126471</td>\n      <td>-2.484766</td>\n      <td> 0.272488</td>\n      <td>-2.005085</td>\n      <td>-1.027807</td>\n    </tr>\n    <tr>\n      <th>1980-04-30</th>\n      <td>-1.049593</td>\n      <td>-0.788299</td>\n      <td>-0.866360</td>\n      <td>-0.810652</td>\n      <td>-0.511990</td>\n      <td>-2.234189</td>\n      <td> 0.175152</td>\n      <td>-2.154401</td>\n      <td>-0.420754</td>\n    </tr>\n    <tr>\n      <th>1980-05-31</th>\n      <td>-0.982191</td>\n      <td>-0.935345</td>\n      <td>-0.449225</td>\n      <td>-0.666945</td>\n      <td>-0.667239</td>\n      <td>-1.542010</td>\n      <td> 0.283610</td>\n      <td>-1.834780</td>\n      <td> 0.074049</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n1980-01-31       NaN       NaN       NaN       NaN       NaN       NaN   \n1980-02-29       NaN       NaN       NaN       NaN       NaN       NaN   \n1980-03-31 -1.107874 -0.591140 -1.028791 -0.779700 -0.126471 -2.484766   \n1980-04-30 -1.049593 -0.788299 -0.866360 -0.810652 -0.511990 -2.234189   \n1980-05-31 -0.982191 -0.935345 -0.449225 -0.666945 -0.667239 -1.542010   \n\n                EOF7      EOF8      EOF9  \n1980-01-31       NaN       NaN       NaN  \n1980-02-29       NaN       NaN       NaN  \n1980-03-31  0.272488 -2.005085 -1.027807  \n1980-04-30  0.175152 -2.154401 -0.420754  \n1980-05-31  0.283610 -1.834780  0.074049  "
     },
     "metadata": {},
     "execution_count": 22
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### calculates the seasonal cycle and include it in the predictors"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "#seas_cyc = msla.groupby(msla.index.month).mean()\n#seas_cyc_arr = seas_cyc.values.copy()\n#seas_cyc_arr = np.resize(seas_cyc_arr, (len(sst),4))",
   "execution_count": 23,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors = sst.copy()",
   "execution_count": 24,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "#predictors['seas_cyc'] = seas_cyc_arr[:,0]",
   "execution_count": 25,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors.head()",
   "execution_count": 26,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>1980-01-31</th>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n    </tr>\n    <tr>\n      <th>1980-02-29</th>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n      <td>      NaN</td>\n    </tr>\n    <tr>\n      <th>1980-03-31</th>\n      <td>-1.107874</td>\n      <td>-0.591140</td>\n      <td>-1.028791</td>\n      <td>-0.779700</td>\n      <td>-0.126471</td>\n      <td>-2.484766</td>\n      <td> 0.272488</td>\n      <td>-2.005085</td>\n      <td>-1.027807</td>\n    </tr>\n    <tr>\n      <th>1980-04-30</th>\n      <td>-1.049593</td>\n      <td>-0.788299</td>\n      <td>-0.866360</td>\n      <td>-0.810652</td>\n      <td>-0.511990</td>\n      <td>-2.234189</td>\n      <td> 0.175152</td>\n      <td>-2.154401</td>\n      <td>-0.420754</td>\n    </tr>\n    <tr>\n      <th>1980-05-31</th>\n      <td>-0.982191</td>\n      <td>-0.935345</td>\n      <td>-0.449225</td>\n      <td>-0.666945</td>\n      <td>-0.667239</td>\n      <td>-1.542010</td>\n      <td> 0.283610</td>\n      <td>-1.834780</td>\n      <td> 0.074049</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n1980-01-31       NaN       NaN       NaN       NaN       NaN       NaN   \n1980-02-29       NaN       NaN       NaN       NaN       NaN       NaN   \n1980-03-31 -1.107874 -0.591140 -1.028791 -0.779700 -0.126471 -2.484766   \n1980-04-30 -1.049593 -0.788299 -0.866360 -0.810652 -0.511990 -2.234189   \n1980-05-31 -0.982191 -0.935345 -0.449225 -0.666945 -0.667239 -1.542010   \n\n                EOF7      EOF8      EOF9  \n1980-01-31       NaN       NaN       NaN  \n1980-02-29       NaN       NaN       NaN  \n1980-03-31  0.272488 -2.005085 -1.027807  \n1980-04-30  0.175152 -2.154401 -0.420754  \n1980-05-31  0.283610 -1.834780  0.074049  "
     },
     "metadata": {},
     "execution_count": 26
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### Truncate the predictors data frame so that it matches the predictand's period"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train = predictors.truncate(before=msla.index[0], after='2012-12-31')",
   "execution_count": 27,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from datetime import datetime, timedelta",
   "execution_count": 28,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_test = predictors.truncate(before='2013-01-01')",
   "execution_count": 29,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_test.head()",
   "execution_count": 30,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2013-01-31</th>\n      <td> 0.441183</td>\n      <td> 0.553970</td>\n      <td>-0.649460</td>\n      <td> 0.460341</td>\n      <td>-0.784978</td>\n      <td> 0.345687</td>\n      <td>-0.913241</td>\n      <td> 1.160401</td>\n      <td>-0.560570</td>\n    </tr>\n    <tr>\n      <th>2013-02-28</th>\n      <td> 0.626748</td>\n      <td> 0.450770</td>\n      <td>-0.048795</td>\n      <td> 0.119587</td>\n      <td>-0.799325</td>\n      <td>-0.161482</td>\n      <td>-0.302423</td>\n      <td> 1.323378</td>\n      <td>-1.129079</td>\n    </tr>\n    <tr>\n      <th>2013-03-31</th>\n      <td> 0.706321</td>\n      <td> 0.313734</td>\n      <td> 0.670206</td>\n      <td>-0.134879</td>\n      <td>-1.158049</td>\n      <td>-0.516534</td>\n      <td> 0.290282</td>\n      <td> 1.463259</td>\n      <td>-0.860440</td>\n    </tr>\n    <tr>\n      <th>2013-04-30</th>\n      <td> 0.710833</td>\n      <td> 0.419355</td>\n      <td> 0.888320</td>\n      <td>-0.468886</td>\n      <td>-1.076923</td>\n      <td>-0.920763</td>\n      <td> 0.239916</td>\n      <td> 1.627752</td>\n      <td>-0.182579</td>\n    </tr>\n    <tr>\n      <th>2013-05-31</th>\n      <td> 0.688718</td>\n      <td> 0.479600</td>\n      <td> 0.368737</td>\n      <td>-0.810292</td>\n      <td>-0.821610</td>\n      <td>-1.052582</td>\n      <td> 0.039362</td>\n      <td> 1.949067</td>\n      <td> 0.367747</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2013-01-31  0.441183  0.553970 -0.649460  0.460341 -0.784978  0.345687   \n2013-02-28  0.626748  0.450770 -0.048795  0.119587 -0.799325 -0.161482   \n2013-03-31  0.706321  0.313734  0.670206 -0.134879 -1.158049 -0.516534   \n2013-04-30  0.710833  0.419355  0.888320 -0.468886 -1.076923 -0.920763   \n2013-05-31  0.688718  0.479600  0.368737 -0.810292 -0.821610 -1.052582   \n\n                EOF7      EOF8      EOF9  \n2013-01-31 -0.913241  1.160401 -0.560570  \n2013-02-28 -0.302423  1.323378 -1.129079  \n2013-03-31  0.290282  1.463259 -0.860440  \n2013-04-30  0.239916  1.627752 -0.182579  \n2013-05-31  0.039362  1.949067  0.367747  "
     },
     "metadata": {},
     "execution_count": 30
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train.tail()",
   "execution_count": 31,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2012-08-31</th>\n      <td> 0.194940</td>\n      <td> 0.578025</td>\n      <td> 0.959596</td>\n      <td> 0.633007</td>\n      <td>-0.776405</td>\n      <td> 1.194343</td>\n      <td> 1.056317</td>\n      <td>-0.154364</td>\n      <td>-0.770291</td>\n    </tr>\n    <tr>\n      <th>2012-09-30</th>\n      <td> 0.157989</td>\n      <td> 0.741021</td>\n      <td> 0.834498</td>\n      <td> 1.444838</td>\n      <td>-0.525554</td>\n      <td> 0.981528</td>\n      <td> 1.195169</td>\n      <td>-0.053278</td>\n      <td>-0.574374</td>\n    </tr>\n    <tr>\n      <th>2012-10-31</th>\n      <td> 0.121739</td>\n      <td> 0.823729</td>\n      <td> 0.383890</td>\n      <td> 1.785692</td>\n      <td>-0.543366</td>\n      <td> 0.994475</td>\n      <td> 0.932145</td>\n      <td>-0.068096</td>\n      <td>-0.538249</td>\n    </tr>\n    <tr>\n      <th>2012-11-30</th>\n      <td> 0.168521</td>\n      <td> 0.894391</td>\n      <td>-0.212990</td>\n      <td> 1.442968</td>\n      <td>-0.744425</td>\n      <td> 0.872075</td>\n      <td> 0.142523</td>\n      <td> 0.265779</td>\n      <td>-0.024213</td>\n    </tr>\n    <tr>\n      <th>2012-12-31</th>\n      <td> 0.277662</td>\n      <td> 0.772436</td>\n      <td>-0.691938</td>\n      <td> 0.853209</td>\n      <td>-0.719597</td>\n      <td> 0.598226</td>\n      <td>-0.656387</td>\n      <td> 0.728071</td>\n      <td>-0.169520</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2012-08-31  0.194940  0.578025  0.959596  0.633007 -0.776405  1.194343   \n2012-09-30  0.157989  0.741021  0.834498  1.444838 -0.525554  0.981528   \n2012-10-31  0.121739  0.823729  0.383890  1.785692 -0.543366  0.994475   \n2012-11-30  0.168521  0.894391 -0.212990  1.442968 -0.744425  0.872075   \n2012-12-31  0.277662  0.772436 -0.691938  0.853209 -0.719597  0.598226   \n\n                EOF7      EOF8      EOF9  \n2012-08-31  1.056317 -0.154364 -0.770291  \n2012-09-30  1.195169 -0.053278 -0.574374  \n2012-10-31  0.932145 -0.068096 -0.538249  \n2012-11-30  0.142523  0.265779 -0.024213  \n2012-12-31 -0.656387  0.728071 -0.169520  "
     },
     "metadata": {},
     "execution_count": 31
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### scale the predictors"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from sklearn.preprocessing import StandardScaler as scaler",
   "execution_count": 32,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "scale = scaler()",
   "execution_count": 33,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "#### the following because the scaler doesnt deal with missing values"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "pred_vals = predictors_train.dropna()",
   "execution_count": 34,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "scale.fit(pred_vals)",
   "execution_count": 35,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "StandardScaler(copy=True, with_mean=True, with_std=True)"
     },
     "metadata": {},
     "execution_count": 35
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train_std = pd.DataFrame(scale.transform(pred_vals), \\\n                                    index=pred_vals.index, \\\n                                    columns=pred_vals.columns)",
   "execution_count": 36,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_test_std = pd.DataFrame(scale.transform(predictors_test), \\\n                                    index=predictors_test.index, \\\n                                    columns=predictors_test.columns)",
   "execution_count": 37,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### reindex to have complete time-series"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train_std = predictors_train_std.reindex_like(predictors_train)",
   "execution_count": 38,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train_std.tail()",
   "execution_count": 39,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2012-08-31</th>\n      <td> 0.227956</td>\n      <td> 0.664142</td>\n      <td> 0.952708</td>\n      <td> 0.646773</td>\n      <td>-0.829453</td>\n      <td> 1.334988</td>\n      <td> 1.175453</td>\n      <td>-0.087977</td>\n      <td>-0.911161</td>\n    </tr>\n    <tr>\n      <th>2012-09-30</th>\n      <td> 0.190992</td>\n      <td> 0.832512</td>\n      <td> 0.820478</td>\n      <td> 1.511263</td>\n      <td>-0.556788</td>\n      <td> 1.100605</td>\n      <td> 1.328511</td>\n      <td> 0.030118</td>\n      <td>-0.685482</td>\n    </tr>\n    <tr>\n      <th>2012-10-31</th>\n      <td> 0.154729</td>\n      <td> 0.917947</td>\n      <td> 0.344181</td>\n      <td> 1.874227</td>\n      <td>-0.576149</td>\n      <td> 1.114864</td>\n      <td> 1.038577</td>\n      <td> 0.012807</td>\n      <td>-0.643869</td>\n    </tr>\n    <tr>\n      <th>2012-11-30</th>\n      <td> 0.201528</td>\n      <td> 0.990939</td>\n      <td>-0.286726</td>\n      <td> 1.509272</td>\n      <td>-0.794692</td>\n      <td> 0.980059</td>\n      <td> 0.168169</td>\n      <td> 0.402864</td>\n      <td>-0.051746</td>\n    </tr>\n    <tr>\n      <th>2012-12-31</th>\n      <td> 0.310707</td>\n      <td> 0.864963</td>\n      <td>-0.792979</td>\n      <td> 0.881259</td>\n      <td>-0.767705</td>\n      <td> 0.678456</td>\n      <td>-0.712478</td>\n      <td> 0.942946</td>\n      <td>-0.219126</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2012-08-31  0.227956  0.664142  0.952708  0.646773 -0.829453  1.334988   \n2012-09-30  0.190992  0.832512  0.820478  1.511263 -0.556788  1.100605   \n2012-10-31  0.154729  0.917947  0.344181  1.874227 -0.576149  1.114864   \n2012-11-30  0.201528  0.990939 -0.286726  1.509272 -0.794692  0.980059   \n2012-12-31  0.310707  0.864963 -0.792979  0.881259 -0.767705  0.678456   \n\n                EOF7      EOF8      EOF9  \n2012-08-31  1.175453 -0.087977 -0.911161  \n2012-09-30  1.328511  0.030118 -0.685482  \n2012-10-31  1.038577  0.012807 -0.643869  \n2012-11-30  0.168169  0.402864 -0.051746  \n2012-12-31 -0.712478  0.942946 -0.219126  "
     },
     "metadata": {},
     "execution_count": 39
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_train_std.tail()",
   "execution_count": 40,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2012-08-31</th>\n      <td> 0.227956</td>\n      <td> 0.664142</td>\n      <td> 0.952708</td>\n      <td> 0.646773</td>\n      <td>-0.829453</td>\n      <td> 1.334988</td>\n      <td> 1.175453</td>\n      <td>-0.087977</td>\n      <td>-0.911161</td>\n    </tr>\n    <tr>\n      <th>2012-09-30</th>\n      <td> 0.190992</td>\n      <td> 0.832512</td>\n      <td> 0.820478</td>\n      <td> 1.511263</td>\n      <td>-0.556788</td>\n      <td> 1.100605</td>\n      <td> 1.328511</td>\n      <td> 0.030118</td>\n      <td>-0.685482</td>\n    </tr>\n    <tr>\n      <th>2012-10-31</th>\n      <td> 0.154729</td>\n      <td> 0.917947</td>\n      <td> 0.344181</td>\n      <td> 1.874227</td>\n      <td>-0.576149</td>\n      <td> 1.114864</td>\n      <td> 1.038577</td>\n      <td> 0.012807</td>\n      <td>-0.643869</td>\n    </tr>\n    <tr>\n      <th>2012-11-30</th>\n      <td> 0.201528</td>\n      <td> 0.990939</td>\n      <td>-0.286726</td>\n      <td> 1.509272</td>\n      <td>-0.794692</td>\n      <td> 0.980059</td>\n      <td> 0.168169</td>\n      <td> 0.402864</td>\n      <td>-0.051746</td>\n    </tr>\n    <tr>\n      <th>2012-12-31</th>\n      <td> 0.310707</td>\n      <td> 0.864963</td>\n      <td>-0.792979</td>\n      <td> 0.881259</td>\n      <td>-0.767705</td>\n      <td> 0.678456</td>\n      <td>-0.712478</td>\n      <td> 0.942946</td>\n      <td>-0.219126</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2012-08-31  0.227956  0.664142  0.952708  0.646773 -0.829453  1.334988   \n2012-09-30  0.190992  0.832512  0.820478  1.511263 -0.556788  1.100605   \n2012-10-31  0.154729  0.917947  0.344181  1.874227 -0.576149  1.114864   \n2012-11-30  0.201528  0.990939 -0.286726  1.509272 -0.794692  0.980059   \n2012-12-31  0.310707  0.864963 -0.792979  0.881259 -0.767705  0.678456   \n\n                EOF7      EOF8      EOF9  \n2012-08-31  1.175453 -0.087977 -0.911161  \n2012-09-30  1.328511  0.030118 -0.685482  \n2012-10-31  1.038577  0.012807 -0.643869  \n2012-11-30  0.168169  0.402864 -0.051746  \n2012-12-31 -0.712478  0.942946 -0.219126  "
     },
     "metadata": {},
     "execution_count": 40
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "msla.tail(10)",
   "execution_count": 41,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>msla</th>\n      <th>trend</th>\n      <th>detrend</th>\n      <th>seas_msla_detrend</th>\n      <th>seas_msla</th>\n      <th>m_cat</th>\n      <th>s_cat</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2014-03-31</th>\n      <td> 19.7</td>\n      <td> 4.420743</td>\n      <td> 15.279257</td>\n      <td> 11.001465</td>\n      <td> 15.400000</td>\n      <td>  2</td>\n      <td>  2</td>\n    </tr>\n    <tr>\n      <th>2014-04-30</th>\n      <td> 16.1</td>\n      <td> 4.442951</td>\n      <td> 11.657049</td>\n      <td> 14.212590</td>\n      <td> 18.633333</td>\n      <td>  2</td>\n      <td>  2</td>\n    </tr>\n    <tr>\n      <th>2014-05-31</th>\n      <td>  9.6</td>\n      <td> 4.465160</td>\n      <td>  5.134840</td>\n      <td> 10.690382</td>\n      <td> 15.133333</td>\n      <td>  1</td>\n      <td>  2</td>\n    </tr>\n    <tr>\n      <th>2014-06-30</th>\n      <td>  7.1</td>\n      <td> 4.487368</td>\n      <td>  2.612632</td>\n      <td>  6.468174</td>\n      <td> 10.933333</td>\n      <td>  1</td>\n      <td>  2</td>\n    </tr>\n    <tr>\n      <th>2014-07-31</th>\n      <td> 10.3</td>\n      <td> 4.509576</td>\n      <td>  5.790424</td>\n      <td>  4.512632</td>\n      <td>  9.000000</td>\n      <td>  2</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-08-31</th>\n      <td> 10.2</td>\n      <td> 4.531784</td>\n      <td>  5.668216</td>\n      <td>  4.690424</td>\n      <td>  9.200000</td>\n      <td>  2</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-09-30</th>\n      <td>  9.0</td>\n      <td> 4.553992</td>\n      <td>  4.446008</td>\n      <td>  5.301549</td>\n      <td>  9.833333</td>\n      <td>  1</td>\n      <td>  1</td>\n    </tr>\n    <tr>\n      <th>2014-10-31</th>\n      <td>  NaN</td>\n      <td>      NaN</td>\n      <td>       NaN</td>\n      <td>       NaN</td>\n      <td>       NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td>  7.1</td>\n      <td> 4.598409</td>\n      <td>  2.501591</td>\n      <td>       NaN</td>\n      <td>       NaN</td>\n      <td>  1</td>\n      <td>NaN</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td>  4.6</td>\n      <td> 4.620617</td>\n      <td> -0.020617</td>\n      <td>       NaN</td>\n      <td>       NaN</td>\n      <td>  0</td>\n      <td>NaN</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            msla     trend    detrend  seas_msla_detrend  seas_msla m_cat  \\\n2014-03-31  19.7  4.420743  15.279257          11.001465  15.400000     2   \n2014-04-30  16.1  4.442951  11.657049          14.212590  18.633333     2   \n2014-05-31   9.6  4.465160   5.134840          10.690382  15.133333     1   \n2014-06-30   7.1  4.487368   2.612632           6.468174  10.933333     1   \n2014-07-31  10.3  4.509576   5.790424           4.512632   9.000000     2   \n2014-08-31  10.2  4.531784   5.668216           4.690424   9.200000     2   \n2014-09-30   9.0  4.553992   4.446008           5.301549   9.833333     1   \n2014-10-31   NaN       NaN        NaN                NaN        NaN   NaN   \n2014-11-30   7.1  4.598409   2.501591                NaN        NaN     1   \n2014-12-31   4.6  4.620617  -0.020617                NaN        NaN     0   \n\n           s_cat  \n2014-03-31     2  \n2014-04-30     2  \n2014-05-31     2  \n2014-06-30     2  \n2014-07-31     1  \n2014-08-31     1  \n2014-09-30     1  \n2014-10-31   NaN  \n2014-11-30   NaN  \n2014-12-31   NaN  "
     },
     "metadata": {},
     "execution_count": 41
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### shift the MSLA to construct the predictand"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictand = msla.loc[:,['msla','seas_msla']].tshift(-(lag + 2))",
   "execution_count": 42,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Data munging"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "indata = predictors_train_std.copy()",
   "execution_count": 43,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "if tscale == '3M': \n    indata['predictand'] = predictand['seas_msla']\nelse: \n    indata['predictand'] = predictand['msla']",
   "execution_count": 44,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "indata.tail()",
   "execution_count": 45,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n      <th>predictand</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2012-08-31</th>\n      <td> 0.227956</td>\n      <td> 0.664142</td>\n      <td> 0.952708</td>\n      <td> 0.646773</td>\n      <td>-0.829453</td>\n      <td> 1.334988</td>\n      <td> 1.175453</td>\n      <td>-0.087977</td>\n      <td>-0.911161</td>\n      <td> 4.866667</td>\n    </tr>\n    <tr>\n      <th>2012-09-30</th>\n      <td> 0.190992</td>\n      <td> 0.832512</td>\n      <td> 0.820478</td>\n      <td> 1.511263</td>\n      <td>-0.556788</td>\n      <td> 1.100605</td>\n      <td> 1.328511</td>\n      <td> 0.030118</td>\n      <td>-0.685482</td>\n      <td> 3.600000</td>\n    </tr>\n    <tr>\n      <th>2012-10-31</th>\n      <td> 0.154729</td>\n      <td> 0.917947</td>\n      <td> 0.344181</td>\n      <td> 1.874227</td>\n      <td>-0.576149</td>\n      <td> 1.114864</td>\n      <td> 1.038577</td>\n      <td> 0.012807</td>\n      <td>-0.643869</td>\n      <td> 4.800000</td>\n    </tr>\n    <tr>\n      <th>2012-11-30</th>\n      <td> 0.201528</td>\n      <td> 0.990939</td>\n      <td>-0.286726</td>\n      <td> 1.509272</td>\n      <td>-0.794692</td>\n      <td> 0.980059</td>\n      <td> 0.168169</td>\n      <td> 0.402864</td>\n      <td>-0.051746</td>\n      <td> 4.733333</td>\n    </tr>\n    <tr>\n      <th>2012-12-31</th>\n      <td> 0.310707</td>\n      <td> 0.864963</td>\n      <td>-0.792979</td>\n      <td> 0.881259</td>\n      <td>-0.767705</td>\n      <td> 0.678456</td>\n      <td>-0.712478</td>\n      <td> 0.942946</td>\n      <td>-0.219126</td>\n      <td> 4.766667</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2012-08-31  0.227956  0.664142  0.952708  0.646773 -0.829453  1.334988   \n2012-09-30  0.190992  0.832512  0.820478  1.511263 -0.556788  1.100605   \n2012-10-31  0.154729  0.917947  0.344181  1.874227 -0.576149  1.114864   \n2012-11-30  0.201528  0.990939 -0.286726  1.509272 -0.794692  0.980059   \n2012-12-31  0.310707  0.864963 -0.792979  0.881259 -0.767705  0.678456   \n\n                EOF7      EOF8      EOF9  predictand  \n2012-08-31  1.175453 -0.087977 -0.911161    4.866667  \n2012-09-30  1.328511  0.030118 -0.685482    3.600000  \n2012-10-31  1.038577  0.012807 -0.643869    4.800000  \n2012-11-30  0.168169  0.402864 -0.051746    4.733333  \n2012-12-31 -0.712478  0.942946 -0.219126    4.766667  "
     },
     "metadata": {},
     "execution_count": 45
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "indata = indata.dropna()",
   "execution_count": 46,
   "outputs": []
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "+ y = target \n+ X = features"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "y = indata.pop('predictand').values\nX = indata.values",
   "execution_count": 47,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "print(y.shape, X.shape)",
   "execution_count": 48,
   "outputs": [
    {
     "output_type": "stream",
     "text": "((351,), (351, 9))\n",
     "name": "stdout"
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## <hr size=3>  MARS "
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from pyearth import Earth ",
   "execution_count": 49,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "regr = Earth(max_degree=2)",
   "execution_count": 50,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "regr.fit(X, y)",
   "execution_count": 51,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "Earth(penalty=None, min_search_points=None, endspan_alpha=None, check_every=None, max_terms=None, max_degree=2, minspan_alpha=None, thresh=None, minspan=None, endspan=None, allow_linear=None)"
     },
     "metadata": {},
     "execution_count": 51
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "yhat = regr.predict(X)",
   "execution_count": 52,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "f, ax = plt.subplots(figsize=(8,4))\nax.axhline(c='.8')\nax.plot(indata.index.to_datetime(), y, 'b',lw=1.5, label='observed')\nax.plot(indata.index.to_datetime(),yhat,'r', lw=1.5, label='predicted')\nax.legend()\nax.set_title('MARS modelling for %s, NO CV, R = %4.2f' % ( station_name, np.corrcoef(y,yhat)[0,1]) )\nf.savefig('../../figures/rasheds/MARS_time-series_{}_{}lag_no_detrend.png'.format(station_name, lag))",
   "execution_count": 53,
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeMAAAEKCAYAAAAhPD1yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": "<matplotlib.figure.Figure at 0x10e9a0810>"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "res = pd.DataFrame(np.c_[y,yhat], columns=['observed','predicted'], index=indata.index.to_datetime())",
   "execution_count": 54,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "mean_squared_error(res['observed'], res['predicted'])",
   "execution_count": 55,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "8.1748887464045694"
     },
     "metadata": {},
     "execution_count": 55
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "sns.lmplot('observed','predicted',res, size=6 )\nplt.title('MARS modelling for %s' % ( station_name) )\nplt.savefig('../../figures/rasheds/MARS_snsplot_{}_{}lag.png'.format(station_name, lag), dpi=400)",
   "execution_count": 56,
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAakAAAGyCAYAAABePTsDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqjfVAAAAAElFTkSuQmCC\n",
      "text/plain": "<matplotlib.figure.Figure at 0x10e96de50>"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "## Cross Validation procedure now"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "All the scores are calculated over the test instances, **NOT** the training instances "
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "from sklearn import cross_validation",
   "execution_count": 57,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "#cv = cross_validation.ShuffleSplit(len(y), n_iter=5, test_size=0.2)\ncv = cross_validation.LeaveOneOut(len(y))",
   "execution_count": 58,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false
   },
   "cell_type": "markdown",
   "source": "y_arr = []\nyhat_arr = []\nMSE_arr = []\nregr = Earth(max_degree=2)\nfor cv_index, (train_index, test_index) in enumerate(cv): \n    res = regr.fit(X[train_index,:], y[train_index])\n    yhat = res.predict(X[test_index])\n    yhat_arr.append(yhat)\n    y_arr.append(y[test_index])\n    mse = mean_squared_error(yhat, y[test_index])\n    MSE_arr.append(mse)\nMSE_arr = np.array(MSE_arr).flatten()\nyhat_arr = np.array(yhat_arr).flatten()\ny_arr = np.array(y_arr).flatten()"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### correlation coefficient between ``cross-validated`` predicted values and observations"
  },
  {
   "metadata": {
    "collapsed": false
   },
   "cell_type": "markdown",
   "source": "CV_R = np.corrcoef(y_arr, yhat_arr)[0,1]; print(CV_R)"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### histogram of the Mean Squared Error values "
  },
  {
   "metadata": {
    "collapsed": false
   },
   "cell_type": "markdown",
   "source": "f, ax = plt.subplots(figsize=(6,6), dpi=200)\nax.hist(MSE_arr, color='k', alpha=.7)\nax.set_title('CV MSE for %s\\nmean = %6.2f'% ( station_name, MSE_arr.mean() ) )\nax.text(0.8, 0.9, 'N = %s' % (len(MSE_arr)), transform=ax.transAxes)\nf.savefig('../../figures/new_data/MARS_CV_MSE_hist_{}_{}lag.png'.format(station_name, lag))"
  },
  {
   "metadata": {
    "collapsed": false
   },
   "cell_type": "markdown",
   "source": "f, ax = plt.subplots(figsize=(8,4))\nax.axhline(c='.8')\nax.plot(indata.index.to_datetime(), y_arr, 'b',lw=1.5, label='observed')\nax.plot(indata.index.to_datetime(),yhat_arr,'r', lw=1.5, label='predicted')\nax.legend()\nax.set_title('MARS CV predicted / observed TS for %s, CV R = %4.2f' % ( station_name, CV_R ) )\nf.savefig('../../figures/new_data/MARS_time-series_CV_{}_{}lag.png'.format(station_name, lag))"
  },
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": "### and now predicts"
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_test_std.head()",
   "execution_count": 59,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2013-01-31</th>\n      <td> 0.474286</td>\n      <td> 0.639294</td>\n      <td>-0.748079</td>\n      <td> 0.462908</td>\n      <td>-0.838772</td>\n      <td> 0.400324</td>\n      <td>-0.995610</td>\n      <td> 1.448025</td>\n      <td>-0.669581</td>\n    </tr>\n    <tr>\n      <th>2013-02-28</th>\n      <td> 0.659916</td>\n      <td> 0.532691</td>\n      <td>-0.113171</td>\n      <td> 0.100051</td>\n      <td>-0.854366</td>\n      <td>-0.158245</td>\n      <td>-0.322300</td>\n      <td> 1.638427</td>\n      <td>-1.324453</td>\n    </tr>\n    <tr>\n      <th>2013-03-31</th>\n      <td> 0.739517</td>\n      <td> 0.391137</td>\n      <td> 0.646820</td>\n      <td>-0.170921</td>\n      <td>-1.244285</td>\n      <td>-0.549280</td>\n      <td> 0.331045</td>\n      <td> 1.801846</td>\n      <td>-1.015004</td>\n    </tr>\n    <tr>\n      <th>2013-04-30</th>\n      <td> 0.744031</td>\n      <td> 0.500240</td>\n      <td> 0.877368</td>\n      <td>-0.526593</td>\n      <td>-1.156105</td>\n      <td>-0.994476</td>\n      <td> 0.275526</td>\n      <td> 1.994019</td>\n      <td>-0.234169</td>\n    </tr>\n    <tr>\n      <th>2013-05-31</th>\n      <td> 0.721908</td>\n      <td> 0.562472</td>\n      <td> 0.328164</td>\n      <td>-0.890143</td>\n      <td>-0.878589</td>\n      <td>-1.139655</td>\n      <td> 0.054453</td>\n      <td> 2.369402</td>\n      <td> 0.399757</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2013-01-31  0.474286  0.639294 -0.748079  0.462908 -0.838772  0.400324   \n2013-02-28  0.659916  0.532691 -0.113171  0.100051 -0.854366 -0.158245   \n2013-03-31  0.739517  0.391137  0.646820 -0.170921 -1.244285 -0.549280   \n2013-04-30  0.744031  0.500240  0.877368 -0.526593 -1.156105 -0.994476   \n2013-05-31  0.721908  0.562472  0.328164 -0.890143 -0.878589 -1.139655   \n\n                EOF7      EOF8      EOF9  \n2013-01-31 -0.995610  1.448025 -0.669581  \n2013-02-28 -0.322300  1.638427 -1.324453  \n2013-03-31  0.331045  1.801846 -1.015004  \n2013-04-30  0.275526  1.994019 -0.234169  \n2013-05-31  0.054453  2.369402  0.399757  "
     },
     "metadata": {},
     "execution_count": 59
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "predictors_test_std.tail()",
   "execution_count": 60,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>EOF1</th>\n      <th>EOF2</th>\n      <th>EOF3</th>\n      <th>EOF4</th>\n      <th>EOF5</th>\n      <th>EOF6</th>\n      <th>EOF7</th>\n      <th>EOF8</th>\n      <th>EOF9</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2014-10-31</th>\n      <td> 0.325926</td>\n      <td> 1.774335</td>\n      <td>-1.518708</td>\n      <td>-1.333059</td>\n      <td> 0.103720</td>\n      <td> 1.315195</td>\n      <td> 0.167349</td>\n      <td> 2.069282</td>\n      <td>-0.537560</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td> 0.100170</td>\n      <td> 1.730264</td>\n      <td>-1.601095</td>\n      <td>-1.183057</td>\n      <td> 0.651315</td>\n      <td> 1.802366</td>\n      <td>-0.052539</td>\n      <td> 2.625778</td>\n      <td>-0.127494</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td>-0.129946</td>\n      <td> 1.713454</td>\n      <td>-1.577788</td>\n      <td>-0.882169</td>\n      <td> 1.671244</td>\n      <td> 2.075114</td>\n      <td>-0.201242</td>\n      <td> 3.070491</td>\n      <td>-0.082215</td>\n    </tr>\n    <tr>\n      <th>2015-01-31</th>\n      <td>-0.235872</td>\n      <td> 1.772894</td>\n      <td>-1.579250</td>\n      <td>-0.478071</td>\n      <td> 2.506648</td>\n      <td> 1.975174</td>\n      <td>-0.712396</td>\n      <td> 2.786648</td>\n      <td>-0.145645</td>\n    </tr>\n    <tr>\n      <th>2015-02-28</th>\n      <td>-0.260534</td>\n      <td> 1.634213</td>\n      <td>-1.839873</td>\n      <td>-0.139635</td>\n      <td> 2.801749</td>\n      <td> 1.580588</td>\n      <td>-1.134861</td>\n      <td> 1.851006</td>\n      <td>-0.528059</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "                EOF1      EOF2      EOF3      EOF4      EOF5      EOF6  \\\n2014-10-31  0.325926  1.774335 -1.518708 -1.333059  0.103720  1.315195   \n2014-11-30  0.100170  1.730264 -1.601095 -1.183057  0.651315  1.802366   \n2014-12-31 -0.129946  1.713454 -1.577788 -0.882169  1.671244  2.075114   \n2015-01-31 -0.235872  1.772894 -1.579250 -0.478071  2.506648  1.975174   \n2015-02-28 -0.260534  1.634213 -1.839873 -0.139635  2.801749  1.580588   \n\n                EOF7      EOF8      EOF9  \n2014-10-31  0.167349  2.069282 -0.537560  \n2014-11-30 -0.052539  2.625778 -0.127494  \n2014-12-31 -0.201242  3.070491 -0.082215  \n2015-01-31 -0.712396  2.786648 -0.145645  \n2015-02-28 -1.134861  1.851006 -0.528059  "
     },
     "metadata": {},
     "execution_count": 60
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "X_pred = predictors_test_std.values",
   "execution_count": 61,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast = regr.predict(X_pred)",
   "execution_count": 62,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast.shape",
   "execution_count": 63,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "(26,)"
     },
     "metadata": {},
     "execution_count": 63
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast_df = pd.DataFrame(forecast, index=predictors_test_std.index, columns=[\"{} forecast\".format(station_name)])",
   "execution_count": 64,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast_df",
   "execution_count": 65,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Tarawa forecast</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2013-01-31</th>\n      <td> 2.235417</td>\n    </tr>\n    <tr>\n      <th>2013-02-28</th>\n      <td> 0.098058</td>\n    </tr>\n    <tr>\n      <th>2013-03-31</th>\n      <td> 2.046161</td>\n    </tr>\n    <tr>\n      <th>2013-04-30</th>\n      <td> 2.124065</td>\n    </tr>\n    <tr>\n      <th>2013-05-31</th>\n      <td> 2.774192</td>\n    </tr>\n    <tr>\n      <th>2013-06-30</th>\n      <td> 3.102988</td>\n    </tr>\n    <tr>\n      <th>2013-07-31</th>\n      <td> 5.714778</td>\n    </tr>\n    <tr>\n      <th>2013-08-31</th>\n      <td> 1.983968</td>\n    </tr>\n    <tr>\n      <th>2013-09-30</th>\n      <td> 3.266503</td>\n    </tr>\n    <tr>\n      <th>2013-10-31</th>\n      <td> 3.373323</td>\n    </tr>\n    <tr>\n      <th>2013-11-30</th>\n      <td> 3.670890</td>\n    </tr>\n    <tr>\n      <th>2013-12-31</th>\n      <td> 4.256796</td>\n    </tr>\n    <tr>\n      <th>2014-01-31</th>\n      <td> 3.865849</td>\n    </tr>\n    <tr>\n      <th>2014-02-28</th>\n      <td> 4.965997</td>\n    </tr>\n    <tr>\n      <th>2014-03-31</th>\n      <td> 6.647030</td>\n    </tr>\n    <tr>\n      <th>2014-04-30</th>\n      <td> 5.107783</td>\n    </tr>\n    <tr>\n      <th>2014-05-31</th>\n      <td> 2.017040</td>\n    </tr>\n    <tr>\n      <th>2014-06-30</th>\n      <td> 1.009004</td>\n    </tr>\n    <tr>\n      <th>2014-07-31</th>\n      <td> 2.113653</td>\n    </tr>\n    <tr>\n      <th>2014-08-31</th>\n      <td> 3.762782</td>\n    </tr>\n    <tr>\n      <th>2014-09-30</th>\n      <td> 4.349747</td>\n    </tr>\n    <tr>\n      <th>2014-10-31</th>\n      <td> 6.083026</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td> 5.254073</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td> 2.778925</td>\n    </tr>\n    <tr>\n      <th>2015-01-31</th>\n      <td>-0.538166</td>\n    </tr>\n    <tr>\n      <th>2015-02-28</th>\n      <td> 1.283390</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            Tarawa forecast\n2013-01-31         2.235417\n2013-02-28         0.098058\n2013-03-31         2.046161\n2013-04-30         2.124065\n2013-05-31         2.774192\n2013-06-30         3.102988\n2013-07-31         5.714778\n2013-08-31         1.983968\n2013-09-30         3.266503\n2013-10-31         3.373323\n2013-11-30         3.670890\n2013-12-31         4.256796\n2014-01-31         3.865849\n2014-02-28         4.965997\n2014-03-31         6.647030\n2014-04-30         5.107783\n2014-05-31         2.017040\n2014-06-30         1.009004\n2014-07-31         2.113653\n2014-08-31         3.762782\n2014-09-30         4.349747\n2014-10-31         6.083026\n2014-11-30         5.254073\n2014-12-31         2.778925\n2015-01-31        -0.538166\n2015-02-28         1.283390"
     },
     "metadata": {},
     "execution_count": 65
    }
   ]
  },
  {
   "metadata": {
    "trusted": true,
    "collapsed": false
   },
   "cell_type": "code",
   "source": "forecast_df = forecast_df.tshift(lag+2, freq='1M')",
   "execution_count": 66,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast_df.iloc[-1,:]",
   "execution_count": 67,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": "Tarawa forecast    1.28339\nName: 2015-06-30 00:00:00, dtype: float64"
     },
     "metadata": {},
     "execution_count": 67
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "f, ax = plt.subplots(figsize=(6,4))\n\nforecast_df['2013':].plot(ax=ax, lw=3, color='r')\nmsla.ix['2013':,'msla'].plot(ax=ax)\nmsla.ix['2013':,'seas_msla'].plot(ax=ax)\nax.legend(loc=3)\nax.set_ylabel('cm')\nax.set_title(\"{}, {}: {} = {:4.2f} cm\".format(station_name, \\\n                                         forecast_df.index[-1].strftime(\"%Y%m\"), \\\n                                         tscale, \\\n                                         forecast_df.values[-1][0]), fontsize=14)\nax.grid('on')\n#ax.set_ylim(-15, 15)\nax.axhline(0,zorder=1, color='0.8')\nf.savefig('../../figures/forecasts/MARS_Forecast_{}_{}_{}.png'.format(station_name, \\\n                                                                 forecast_df.index[-1].strftime(\"%Y%m\"), \\\n                                                                 tscale))",
   "execution_count": 68,
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEYCAYAAACwQCa4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsnXd4XMXVh99tqqverGbLchn3im2wTTHF9ACGUGwggZBG\nEkLKlwRCIAQSIAkQICQQQkLvxYDBmAA2xr3bchu5qNqSrF53te1+f9xdWV2rspIlzfs8erS6d8rZ\n1d3fPffMmRmDpmkoFAqFYnhgHGgDFAqFQtF/KNFXKBSKYYQSfYVCoRhGKNFXKBSKYYQSfYVCoRhG\nKNFXKBSKYYQSfYVCoRhGmAfaAEXvEEK8ANzcSZFvSylf6idz+gwhxCzgMWAWUAb8C3hESql5z08H\nngGmAQeAH0gpt7XTzm+BCVLKm5odOwNY36roLinlLO/5WOBZYDFQAdzX/DMUQkwA/gHMAwqBu6SU\n7/n5vlKAvwPnAjbgReC3Ukq3ECIDOArYgTgppa1V3YeBXwHflVI+709//tLe59ROmTHA34AFQD3w\nptf2Ru/5ed7zU9E/lz9IKV/rSzsVvUd5+oOfO4AR3p9rvcdGNPt5a4Ds6jFe0V0J7EYX/Z8AvwB+\n5D0f7j2/wXv+a+BjIYS1VTs3AL8HWs9AnAxspeXndF6z8y8AMcB84A/As94bBd4+Pgfy0W84fwde\nF0JM9PPtvQVY0G8Y1wJLgV+3KmMELmin7pWAp5330ys6+ZyalwkCPkK/UZ0BLPPa80fv+XBgBbAZ\nXfQfAV4QQsztS1sVvUd5+oMcKWUNUAMghKjyHjsxoEb1nkuARinlT71/HxZCPIYuNH8HrvOe/4X3\n/M+EEJd6jz8vhDB5y30LONJO+5OAfe19Tl5v9jJgrJTyKLBPCDEfuB3YiP5U1Qh8R0rpBp4SQixG\nv0Ec6OxNCSEigDzg11LKQkAKId4Bzgb+1KzoWuAbwIfN6k4ArOgedJ/gx+fUnLlAJnCalLLBa/vv\n0J/Gfon+mcYB93qvyRwhxI/Q39uWvrJZ0XuU6A9xvGLxOPojuQXYBnxfSrlfCHEO8ArwPnAT+hf4\nT8BDwPVAEnAceFhK+YwQ4m/AKCnlVd627wQeRQ9FVAkh4oESIB2o9fZ7ORAN5AB3+xkGWePtvzVR\n3t+n0zY8sx7dA30eiED3NueiPyG0vs4nevtoj3lAkVfwm7d9j/f1ucCHXsEHQEp5ue+1EOLbwH+k\nlG2eoqWUteg3Ll/Zyeifz79aFf0Q+G2rY1cCy4FLO7AbIUQuMLKdU2uklOe2c7yrz6k5B4FLvILf\nnGjv70NANXCb9zo5HZgA7OjAVhNwH3Cr147V6CG6YiHEGuAT9KedBcAe9Ceiu9Cvi+PArVLK1teA\nwg9UeGcII4QwoAtIDjAd3Rs1AX9pViwF/Us3Ez2+/Bt0IboaGI8e6nhSCJEMfAqc1azu2eghgQXe\nv88D9kopj6MLvkD/4k5C916fE0JYurJbSlkopdzQ7H2EAt9FD6sAJKN/8ZtzAkjz1q+SUi6UUu4F\nDO10MRmYK4TIEkLkCSGeEUJEdtJ2ia9tdG+3VAjxDyHEcSHEdu9Tho830MNFnSKEWA9kAZXA061O\nfwmEe2PkPr6BLvqdMZuWISvfz5L2CvvxOTUvWyal/LKZ/Ubgx8D/fG2hXzN/RH8SWgf8VUr5RQdN\n/h74DnAb+k0nFGg+9nQP+s1wNhCL7qzkA6eh32Ce6MxeRccoT39oE4o+IPmMlLIeQAjxIrrH1Jw/\nSylzvOez0EMXW7x/PwTci34DWAuECSGmoQvWmeix9QXAx+gCv9Lb5tfA41LKfd52HkX/gqeghzj8\nwusRvuJ9Lw82e1+NrYo2AsF+tGdFF/Ct6KGaePQnnNfQwzphHbQd5H0dCfwfulBfDFwILBdCzJNS\n7pBS2tEHYrviB96+nwJeB65o1d+n6DffzUKIEeif/5rOGpRSlvvRb1/xGLojMQfA6xS8gu4k+MT6\nMSHEbinl+80rep2R7wO/kVJ+6j32Q+BW7zmAT6SUb3vPfQhcLaV8wPv388DLgX17Qxcl+kMYKWWD\nEOJZ4GYhxGnonrcvG6Y5uc3qfCCEuMAr0r7yACZve18Di9A9/GrgXfRHdIDzgW97X78EXCWE+L63\nndneOiZ/7fcOHr7mbff8ZjF4OxDSqngw0Dr00AYpZZ03tm6TUnq8/XwL2CaESPe23frmEYw+gAng\nAvZIKX3hnt1CiDOB76ELuV9IKbO8fX8H2CiEaB6W0YAP0Ad470G/IXwspXQJITpsUwixj/bDO2ul\nlB2GhbqDV5T/BvwQXYh94xi3ANVSyh96/94phEhDHwh/v1Uz8d6f7b4D3nDaPd4+NPQsJh92dC/f\nhw0/bvCK9lGiP4TxerVb0UV+OfAqejz7N62K2pvVeRBdwJ5HF+7baXZTAFYB56CL31p0j/4fQogp\n6I/h67zlXkYPJ72Ent5YjD4Q6q/toehiMQ+4SEq5tdnpY7QNoYygbVimXXxPPc046P2dgj5Q2lnb\nx4DDrc5no3+unSKEiAEWSynfbHbYJ5rx6OmhPj4B/utN47yCtnH/9rgIfdymNbZ2jnUbb0jnefT4\n+rVSyo+anU4D9raqsgP9qag1Dj+6c7X62+OvnYrOUaI/tDkH/cs4xTfwKIS4iM7jt98HfuwTJiHE\nJO9xX51VwN2AE/hUSnlECFHpPfaF1xuNBG4A5kspN3vbuaRVO13xKnro4Hwp5fZW5zZxcmDV530u\nQB+Abk2LNETvE88aYJyUssh7eCbgRo8VlwKpQohRUkpfGGqht0/Qb1wXtepjEi09046IQ0/vzJZS\n7vQem+3tOxtd+AGQUlZ6n6qWog+KthuXb46UssAPGzrCnzTQR9EHUq+SUn7S6txhWo73gH4jbH2D\nREpZLYQoRX+K3A0ghBiH7kD4m/qq6CFK9Ic25egx6quFEFvQwyTfpa0X1brO5UKIreie7+Pe8iEA\nUsq9Qgg7+sCiL8NkHXq6pO/R3o4+eedqIUQJejz6Me+5EAAhRDRgkFJWtjZACHEderbKjcAxb0wb\nwC2lLAXeAR4WQjwF/NP7nsLRB1Fb0/omswt9TOF5IcQv0YX2WeB5KWUFUCGEWAW8JIT4MfrA4VL0\nGyjesnd4J0o9hx53Pw/v05MQIgSIllIWtzZESnlYCPEpet7/d9EzX/4FPOkNO8W3qvIB+jjG596x\ngkDS5mYshIhCD+tVCCFOB36K/j53NPuf4H2vLwH3eFNrn0aP9/8a+HkH/T0B3C+EKEB/enoC2O69\n2Rnas0fRN6jsnaFHk8cmpdwI3I8+WJiFPtB6CRDrjbe2KO/lVvQ0vn3eevehi/qMZmVWAZVSykPe\nv7/y/v7U268DXbCvQg9f3Ise881t1s4LQEfpm9d47XoFPazi+9nubb8WPXVxvvfYGejphK3DNr73\n1/wzcXk/Awd6Kua7Xrt/0qzOzUAV+kSje9AHtjd76xegf46L0MMZtwFLpJS7vXV9KYUdscz7mXwB\nvM3J2H1ze318iJ6b31XWTl/Q4nPy8gT6DRb0/wnAw7T8nxwTQhillGXo2Vwz0MM6f0KfqfxCB/09\ngn6Tfg396akCfb5Ae7a0Z5va8q+HGAK1XaI3Ne8/wCj0QZcH0eOlK9AfZQH+KaUcdDNGFb3H60W+\nIaW8eKBtUSiGE4EM7ywDSqWUN3kHsHaje52PSikf67yqYhjwa1TanULR7wRS9N/m5KOhEX3gbzYg\nhBBXoA+a3SmlrAugDYpTl3u9oRaFQtGPBCy848ObE/0B+oBVCLBbSrlTCHE3ECOlbC+lS6FQKBQB\nIKDZO97JLu8BT0sp3xBCREkpq72nlwNPdtWGpmmawaAG8hUKhaKbtCucARN9IUQS8Blwu5Rytffw\np0KIO7wTbc5DX0+jUwwGA6WltYEyUzFESEiIUNeJwi+Gy7WSkBDR7vFAevp3o6+KeK8Q4l7vsTuB\nx4UQTqAIfeanQqFQKPqJgMf0+wBtONyVFb1juHhvit4zXK6VhISIdsM7anKWQqFQDCOU6CsUCsUw\nQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsU\nCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU\n6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQ\nDCOU6CsUCsUwQom+QqFQDCOU6CsUCsUwQom+QqFQDCPMgWpYCGEB/gOMAoKBB4EDwAuAB9gL/EhK\nqQXKBoVCoVC0JJCe/jKgVEp5FnAR8DTwKHC395gBuCKA/SsUCoWiFYEU/beBe5v14wRmSSnXeo+t\nBM4PYP8KRbtU1zVSZ3MOtBkKxYAQsPCOlLIeQAgRgX4DuAf4a7MidUBUoPpXKNqjwe7iN89uwuF0\nk55kZeKoGCaOimFcWjShwQH7OigUpwwBvcqFEOnAe8DTUsrXhRB/bnY6Aqjyp52EhIhAmKcYYvhz\nnRwurKLR6SbaGkxReQP5JXWs2lKA0WhgfHo008YlMG1sPBMyYgm2mPrBasVAMJw1JZADuUnAZ8Dt\nUsrV3sM7hRBnSym/Ai4GvvCnrdLS2gBZqRgqJCRE+HWdHM2rAGDxnHTOnZXKkWPVHMiv5EBeJdn5\nVRzMq+Stz7Mxm4yMTY1kwqgYZoyNZ2TS8BWJoYa/18pgp6MbWyA9/bvRwzf3CiF8sf2fAk8KIYKA\n/cA7AexfoWhDRW0jALGRwQRZTEzMiGViRiwAtkYXhwqrOJCn3wRkfhUH86v4cF0ud900izEpKhqp\nGPwEMqb/U3SRb805gepToeiKiho7ALERIW3OhQabmTYmnmlj4gGosznZvL+EV/+Xzf+2FjDmCiX6\nisGPmpylGFY09/S7whpq4dxZqaQlhLPtYGnTDUOhGMwo0VcMKypq7BgMEGUN8qu8wWDg/NPS8Wga\na3YdC7B1CkXgUaKvGFZU1DQSbQ3GZPT/0j99UhLhIWbW7DyO0+UOoHUKReBRoq8YNng8GlV1jX6F\ndpoTZDFx9oxU6mxONu0vCZB1CkX/oERfMWyornfg9mjtDuJ2xbmzUjEaDHyxrRBNU8tFKQYvSvQV\nw4aKWm/mTjc9fb1OCLNEAvkn6sgu6HhO4YmGUmoddT22UaEINEr0FcOGyhpv5k4PPH2A82enAfD5\n9sJ2z1c1VvOnLX/jn3v+q54GFKcsSvQVw4amHP0eePoA49KiGJlkZUd2KeXVbdM3v8hfi9PjJK+m\ngMNVOb2yVaEIFEr0FcOGkzn6PfP0DQYDF5yWjqbBlztaevt1jnrWHdtEiEm/oawu+Lp3xioUAUKJ\nvmLYcHI2bs88fYC5ExOJCLOwdvdxGp0n0zdXF67D4XFyWeaFjIpIZ0/Zfk40lPXaZoWir1Girxg2\nVNQ2YjIaiAj3b2JWe1jMevpmvd3Fpn3FANhcNr4qXI/VEs6ClLmcO/JMNDTWFK7vK9MVij5Dib5i\n2FBRYycmIhijwdCrdhbNTMVkNPC5N33z68JN2Fx2FqWfSZApiJkJU4kOjmJj0VYanLY+sl6h6BuU\n6CuGBS63h+o6R4/j+c2JiQjmtAmJHCurJyvnBF8UrCXUHMLZaWcAYDKaOCdtAQ63g/XHN/e6P4Wi\nL1GirxgWVNU1otHzzJ3W+NI339//FXXOes5OnU+oObTp/IKUeQSZglhTuB63Ry3doDh1UKKvGBZU\n9DJHvzVjUqPISA6nxLQXi9HCovQzW5wPs4RyRvIcqhqr2Vma1Sd9KhR9gRJ9xbCgN7NxOyJjUg2G\nYDuJboE1KLzN+UVpCzFg4Mv8r9VkLcUpgxJ9xbCgt7NxW+P2uDni3AEeI8f2J2F3uNqUSQiLY1r8\nJPJqCzhandcn/SoUvUWJvmJY0BTe6SNPf+eJPZTZy0kzT8BWb2Hj3uJ2y/nCPl8WrO2TfhWK3qJE\nXzEsOBne6b2n79E8rMpbjdFg5PqpF+npm9sL8bQTwhkbPZqREansLt1Hma28130rFL1Fib5iWFBR\n00iQ2Uh4SO+3hc4qO8Dx+mJmJ85gdNwI5k5Moqi8gf25FW3KGgwGFqV7J2sVqMlaioFHib5iWFBR\naycmMgRDLydmaZrGqtwvAbgwYxEAF8zxrr65rf3VN2clTiM6OIoNRVuwudRkLcXAokRfMeRxutzU\nNjh7teaOj4OVh8irLWBGwhSSw5MAyBgRydjUKPYcKaekoqFNHbPRzNmp82l0O1h/fEuvbVAoeoMS\nfcWQ5+Tqmr0X/SYvf9S5LY6ff5ru7X/RwVr7C1LnEWS0sKZATdZSDCxK9BVDnr6amHW0OpdDVUeZ\nGDuekZFpLc7NGp9ATEQw67KKsDW2Td8Mt4RxevJpVDZWsat0b6/sUCh6gxJ9xZCnt5un+PjU6+Vf\nlHFem3Nmk5FFM1OxO9ysyypqt/456fpkLbXWvmIgUaKvGPL0dvMUgILaY+wrP8iYqNGMjR7dbpmz\nZqRgMhqallxuTVJYAlPiJ5BTk68maykGDCX6iiFPZR9snrKqycs/t8MykWFBZCRHkFdc1+4MXYBz\n088C4Evl7SsGCCX6iiFPbz394voSdpXuZWREKhNjx3dadnx6NB5N48ixmnbPj4vOJM2awq4TWZTb\n2ub1KxSBRom+YshTUWMnNNhEaHDPJmZ9lrcGDY0LM87rMs9/fFo0ANkFVe2eNxgMnJuudtZSDBwB\nF30hxDwhxGrv65lCiEIhxGrvz7WB7l+hqKhp7HHmTrmtgq0lOxkRnsS0+Eldlh+XFoUBOFTYvugD\nzE6aTlRQBBuOb8HmsvfILoWipwRU9IUQvwKeA3zB1NnAY1LKRd6ftwLZv0Jhd7hoaHQR08PMnXXH\nN+PRPFw4ahFGQ9dfl7AQC6kJVo4cr8Hl9rRbxmw0c1baAuzuRjYWbe2RXQpFTwm0p38YWAL4noln\nA5cKIb4SQvxbCGENcP+KYU5vc/QPVx3FaDAyLX6y33XGp0fhdHnILa7tsMzC1HlYjBbWFKzDo7V/\nc1AoAkFARV9K+R7QPI1hM/BLKeXZwFHgvkD2r1D0ZvMUp8dFfk0hqdZkQsz+1x+frsf1D3UQ1wew\nWsKZlzybcnslWWUHum2bQtFT+nsg930p5U7v6+XAzH7uXzHM6I2nX1h7DJfmZnTkqG7VG+cdzJWd\niD7AwpR5AGwp3t5t2xSKntL7dWa7x6dCiDuklFuB84Bt/lRKSIgIrFWKIUF710mjW1/jPjM9ptvX\n0eYKfZLVjHTRrboJCRGMiAvjyPEa4uKsGI3tZ/zExwtGZqeSVX6AkEgDEcEq2tlfDGdN6S/R9+0u\n8QPgaSGEEygCvudP5dLSjmOjCgXoX+L2rpOCYj1f3qh5un0dZR3LBiDOkNTtumOSI1m/t5hdB4pJ\nT+xYzGcnzCC/+mNW7V/P2Wnzu9WHomd0dK0MNTq6sQVc9KWUucB87+vdwMJA96lQ+PDNxo3p5mxc\nTdM4Wp1HZFAEcSEx3e53fHo06/cWk11Q1anoz0mayfLDn7C5eLsSfUW/oCZnKYY0FbWNWEMtBFtM\n3atnr6LaUUNm1KgebbzSNJjbSb4+QFRwJBPjxpNXU0Bx/Ylu96NQdBcl+oohi6Zp3olZ3c/cyanR\nF0QbHdW9QVwfiTGhRIYHkV1QhdbO3rnNmTdiNgCb1YCuoh9Qoq8YsjQ0umh0unu05o5vFczMHoq+\nwWBgfFoUVXUOSqs7n3U7LX4yIaYQthTvUDn7ioCjRF8xZPGla/ZkNm5OdR4mg4l0a2qP+x/nR74+\nQJDJwqzEaVQ1VpNdeaTH/SkU/qBEXzFkqejhksoOt4PCuuOMjEjFYrL0uP+uFl9rzrxkFeJR9A9K\n9BVDlp4uqZxXU4hH8/Q4nu8jPdFKaLDJL9EfE5VBfEgsu05kYVeLsCkCiBJ9xZClp55+TnXvBnF9\nGI0GxqRGUVJpo7qusdOyBoOBucmzcXicag9dRUBRoq8YsjQtwdBNT/9oTS7Q80Hc5vhCPIcKq7ss\nO2/ELAA2F6kQjyJwKNFXDFkqa+0Y6N7ELE3TyKnOJyY4mujgqF7b4MvX9yfEEx8ax5io0WRXHaHc\nVtnrvhWK9lCirxiyVNQ0EhkehNnk/2VeaiujzlnfJ14+wOjkSMwmI9ldTNLycbp3QHdryY4+6V+h\naI0SfcWQRNM0Kmobu72kck51PtD7eL4Pi9lIZnIEBSfqsDW2v1l6c2YmTsNiNLO5aHuXk7oUip6g\nRF8xJKltcOJye7q9pPLR6lygb+L5PsalR6NpcPhY13H9UHMI0xOmcMJWRk5Nfp/ZoFD4UKKvGJL4\nNk/p7sSso9V5WIwW0qwpfWZLd+L6oJZlUAQWJfqKIUlPNk+xuewU1ZcwKjINk7F7C7R1xtjUKAyG\nrmfm+pgQO46ooAi2l+zG6Xb2mR0KBSjRVwxRmnL0u+Hp59bko6GRGZXRp7aEBptJT7RytKgGp8vd\nZXmjwcicEbOwuWxklautFBV9ixJ9xZCkJ7NxmyZlRY7sc3vGp0XjcmvkFPm3eUdTiEfl7Cv6GCX6\niiFJT2bjHu2jmbjt0d24fop1BOkRqeyvkNQ4hv4uT4r+Q4m+YkhSUduI0WAg2uqf6Hs0D7k1+SSE\nxhER1Pd71fpW3PQ3Xx90b9+jedhWsqvP7VEMX5ToK4YklTV2oiOCOtyUvDXF9Sewuex9Hs/3ERUe\nRFJsGEeOVePx+Jd/f1rSDIwGowrxKPoUJfqKIYfHo1FZ6+hW5k5fLbLWGePTorA1uik4UedX+Ygg\nK5PjJlBYd5xjdUUBs0sxvFCirxhyVNc78GhatzJ3jtb0bqcsfxjfwxAPqAFdRd+hRF8x5Dg5iNs9\nTz/EFExyeFKgzPJ7J63mTImfSJg5lC0lO3B7uk73VCi6Qom+YsjhS9f0dzZunbOekoZSMiJHYjQE\n7iuREBVCtDWI7MJqv9fVsRjNzE6aQa2jjoOVhwJmm2L4oERfMeTorqef28eLrHWEwWBgfHo0NfUO\nSiptftdTIR5FX6JEXzHkOLl5in+evm8QN5DxfB/jurFvro+MyHSSwhLYXbaPBqf/NwuFoj2U6CuG\nHL7F1vydjeublJURgJm4rRnfg7i+wWBg7ojZuDwudp7YEyjTFMMEJfqKIUdFTSNmk4GIMEuXZd0e\nN7k1+SSHJxFmCQ24bakJ4YQFm7uVwQMwd8RMDBj46tgGnJ6u1+VXKDpCib5iyFFRaycmIhijoeuJ\nWcfri3F4nIyODHxoB8BoMDAuLYrSKjuVtZ1vlt6c2JAY5o2YzbG6It6Sy9UGK4oeo0RfMaRwuT3U\n1Pk/MetoP8bzfTSFeLrp7V8nriTdmsKGoi2sO74pEKYphgHmrgoIIe4E7gWimx3WpJR+LTguhJgH\nPCylXCSEGAu8AHiAvcCPpJTKZVH0GVW1jWh0fxA30Jk7zRnXbPG1uRP9nxcQZAriu1O/xZ+3Pcnb\n2R+SEp7MmOiMAFmpGKr44+n/DJghpTQ2+/FX8H8FPAf4voGPAXdLKc8CDMAVPTFaoeiI7i6pfLQ6\nj3BzGIlh8YE0qwUZIyIIMhvJLuh6+8TWxIXG8J0py9DQeG7vS1Q1tt+GR9PYn1uBzK8kv6SW8mo7\ntkaXCgspuvb0gf3AiR62fxhYArzs/XuWlHKt9/VKYDGwvIdtKxRt6M6SytWNtZTbK5gSNyGgk7Ja\nYzYZyUyJROZXUW93Eh7S9YBzc8bHjGXJ2Mt459CH/CvrJX428wdYTC3b+HJ7Ia993nYyl8EAYcFm\nwkLMhIVYCAs2Ex6i/z0mNYqFU5Mx+DEWohi8+CP6TwBZQohNgC9tQJNS3tpVRSnle0KIjGaHml9N\ndUCUv4YqFP5wcjZu155+Tk3/h3Z8jEuL5mB+FYcKq5kxtvtPGeekLaCg9hibi7fzZvZylk24pkms\nnS4Pn2zKI9hi4oI56djsLhoandTbXTQ0umiwu2iwOykqq8fh8jS1uXZ3EXuOlHPrJRMJDfZHGhSD\nEX/+s0+he+r5zY719BnR0+x1BNC9kSyFogu64+kfrc4F+ncQ10fzfP2eiL7BYOB6sYSi+mI2Fm0l\nPSKVs9PmA7Auq4iqOgcXzR3JkrMyO23H6fLQ0Oiiuq6R1z4/xHZZSmFpPT+6agppCX2/r4Bi4PFH\n9G1Syj/0UX87hRBnSym/Ai4GvvCnUkJCRB91rxjKJCREUN+oL0o2PjOeiLCgTssX7jmGwWBg9uiJ\nhFj8X5ytL5gXGYrx7d3kFNf26vr+zTm385vPHuLdQx8yKTWT8XFjWLUlH4vZyNKLJ/r1xONjxsQR\nvPTJAd5bc5g/vrydH18znXNmp7coU2OvZXXORnYU7eXmGVczJrb/b5h9wXDWFH9E/3MhxKPoMXiH\n72Cz2Lw/+J4MfgE8J4QIQh8reMefyqWlars4ReckJERQWlpLUVkdQRYjtjo79vqO8+BdHhdHKvJI\nC0+mtspJLc5+tFZnVJKVQwVVHDteRZDFr9yIdrBw66RlPLnrOf667lnOtd7AiUob581Kw9XopLS0\ne+/rstNHkhwTwvMfH+DR13aw42AJ1y0aS359Pl8f28iuE1m4NDcWo4UT5dVEugffd9N3rQx1Orqx\n+SP6s4Ak7+9wIBWQwLn+dCylzAXme18fAs7xp55C0RMqahqJjQjpcjCyoPY4Lo9rQOL5PsalRZNT\nVMvenApmjU/oeTsxY7h63OW8nf0BK4rewWSay8Wn93xJidkikdQEK08t38Hawg1s++IdXEF6llBS\nWCJnpp7OvBGzCLOE9bgPxcDhT8rC+4BdSrkIuBZ9APbNgFqlUPQAh9NNnc3pV45+jjeeP5Civ2Bq\nMkaDgTe/PITD2bu18s9Onc+Y0Ml4QqpInnGUmG5sCN+a/NpCvij5hIbMlQRlHMBprsFQlcIVI27g\nd/N+waL0hUrwBzH+iP73gIXQ5LXPBH4SQJsUih7hW9bAn9m4J2fiZgTSpE5JT7RywZw0SqvsrNiY\n26u2NKC9Ew4WAAAgAElEQVTiwDg8dVGUmw6zpnB9t+o3OBvYeHwrf976FI9sfZINRVuwBlm5PPMi\nLo/5Do4j03nzw0o+2pCLR+X6D2r8Ce+YaRbL9772dFBWoRgwmjJ3uvD0NU3jaHUeEUFW4kJi+sO0\nDrli4Wi2HjzByk35nD5pBCnx4T1qZ9ehMo6fsDM7cTEFlpW8d3gFqdYRjI8ZC4DD7aTCXkGZrYJy\neyXltgrK7RWU2yoos1dic+lLNhswMDV+IgtTTmdSnNDnL2SASB7BP5dnsfzrHI4cq+G7l0/CGtq9\n+QWKUwN/RH858KUQ4k30PPslwIcBtUqh6AH+zsatbKyi2lHD9IQpAz4RKSTIzLLzx/PUe1m8tEry\n66Uzu22Tpml8tD4XA3DV6ZNoMMfzxM5neS7rZZLCEim3V1DjaH/g0mK0EBcSQ2bUKEZGpDE/ZQ6x\n7dwIM1Miuffbc/jXR/vJOlrO/f/dyq+WziQhOvArkyr6li5FX0r5ayHEN4GzACfwhJRSzaJVnHL4\nm6PftN5OP6yf7w8zxycwc1w8Ow+VsS6riDOnpXSrftbRCvJKajltQqL3SWE0146/kjfke+TVFhAT\nHM34mLHEh8QQFxpLXEhs0+/IIKvfN5mIsCB+9s3pfLg+hxUb8sgrrlWiPwjxa9qdlPJt4O0A26JQ\n9Ap/Z+OeCvH81iy7YDz7cyt5e/URZozteo6BD03T+GhDDgCXnXFyUPrM1NOZlTiNEFMwJmNP00Hb\nYjQauPLMTC45fVQv0kwVA4laWlkxZGjaJrFLTz8fk8HEyIjU/jDLL2IjQ7jyzNHU2Zy8tfqw3/UO\n5ldx5FgNM8bGMzKpZV52uCWsTwW/OUrwBy9K9BVDhopaO6HB5k7XjXG4HRTUHSM9IrXNImUDzfmn\npTEy0cr6rGJkfqVfdVZsyAXgsvkZgTNMMaRQoq8YMlTUNHaZuZNdeQSP5mFcdOdr0gwEJqORmy+a\ngAF4aZXE5e48Se5wYTUH8iqZPDqWzJTI/jFSMehRoq8YEjTYndgaXV3m6GeVHwBgSvzE/jCr22Sm\nRHLOrFSKyhtYuTm/07K+3P7LlZev6AZK9BVDgrIqPc+8M09f0zT2lR0kzBx6ymTutMfVZ40hKjyI\nFRtyOVHZ0G6Z3OIa9hwpZ3x6dNOKnQqFPyjRVwwJyqq6Ttc8Xl9MZWMVk+JEwAY4+4KwEDM3nD8O\np8vDy59lt7vb1YoNegaS8vIV3UWJvmJIUNrk6Xcc3tlbpod2JsdN6BebesOcCYlMGR3LvpwKthxo\nuXFdYWkdO7JLGZ0cyaSMgZ1RrBh8KNFXDAmawjudePp7yw9iwMCkONFfZvUYg8HAjRcKLGYjr39x\niAb7ySWSP97o9fIXZAz4jGLF4GPIiH5ZtY1tB3u6la9isFPWhadf56wnpzqP0VGjsFp6tr5Nf5MY\nHcrl8zOoqXfw7ldHASiuaGDLgRLSE61MHxM3wBYqBiNDQvQ1TePZD/bxj+V7yS8Z+psjKNriE/2O\nlhTeXy7R0JgyCEI7zblo3khS4sNZs/MYR45X88nGPDRNj+UrL1/RE4aE6B/Iq+TI8RpAX21Q0T8U\nVzScMsvsllbZsIZaOpwp6ovnn6qpmh1hNhm5+UKBBjy/4gAb9xWTHBfGLNHzTVcUw5shIfofrc8F\nwGCAnYeV6PcHMr+Su/+1iQ++zhloU9A0jbJqW4fpmm6Pm/0V2cQER5MSPqKfres949OjWTgtmeKK\nBtwejcvOyMA4RL18TdP4es/xpsXzFH3PoBf97IIqZEEVUzJjmTQqhrziWnXB9AMb95UAsGpLftPm\nJQNFvd1Fo8Pd4cSsnJp8bC4bk+MnDNqQyLWLxhIZZiEpNoy5kxIH2pyAUV3v4L+fHOTp9/fi8Zwa\nT5FDjUEv+r61Ry6fn8H0sfEA7D5SPoAWDX08Ho2dh0oxAA6Xh4/WD6y339XmKb7QztS4wRXaaY41\n1ML9t87l7htnYTIO+q9th0Rbg5k3KYmcohrW7Do20OYMSQb11ZNTVMPenAomjIxmXFo0M7yir+L6\ngeVQYRW1DU4WTksmOS6MtbuLKK5of+Zof9DV5il7yw9gMZoZHzOmP83qc6KswX4vuRy08mMib7qO\noI8+CLBVfc/1544lNNjMu18doapuYJ8ihyKDWvSbe/kA8dGhpCVYOZBXid3hGjjDhjjbZCkAcyYm\nsuSsTDyaxntrjw6YPZWdbJ5SbqugqL4EETOWIJN/gjnYCVrxIZG3LCN41UqivnMTIS/9d6BN6hZR\n1mCuOTsTW6ObN744NNDmDDkGregXnKhj56EyxqRGMmHUyVmJM8bF43J72Jfj39K0iu7h0TR2ZJcS\nFmxmwsgYZo1PIDMlkm0HT5BTVDMgNnXm6e8tPwjA5EEc2ukOlrVriPzBrRg8J1fojPjlTwl58T8D\naFX3OXtmKpkpkWw5cIK9R1W4ti8ZtKL/8cZcoG2+8sxx3hDP4dIBsGrok3O8hsraRmaMi8dsMmIw\nGLjmbD1s8s6aIwNiU1l1x57+yVTNwZWf3xPMO7cT+a2lGBwOALRm34uI/7uTkP/+e6BM6zZGg4Gb\nLxQYDQZe/kzicLoH2qQhw6AU/aLyerYeOMHIJCtTM1vOShw1IoIoaxC7D5er0f8AsD1bv5nObpYn\nPmFUDFMyYzmQV8m+nIp+tym3uJbwUAuxUS09/Ua3g+yqI6SEj2h3s++hhClbEnXD1Rjr6wBwp6RS\n+eV6nDNnNZWJ+PXPCXn+XwNlYrcZmRTBBXPSKK2yNy0j3Z+Yt2wm9MnHMR31fyezwcCgFP2PN+ah\n0f6sRKPBwIyx8dTZnBw5Xj0wBg5RNE1juzxBcJCJKaNjW5xr7u3354SteruTkooGxqVHt8ldz648\njMvjGnQTsrqLsbCAqGuvxFih33A9sbFUv7Uc9+QpVL/9Ac7ZpzWVjbjrl4Q+98+BMrXbXLFwNLGR\nwazclM/xsvp+6zfokxVEf+NCrA/eR8yCOVh/eSfG4qJ+6z+QDDrRP1FlY9O+ElLjw5k5vv1ZiSqL\nJzAUnKijtMrO9DFxWMwtZ76OTIrg9ElJ5JXU9usaSLlF+rIb49pZUz7LF9oZwvF8Q1kZUddeiem4\nnt6ohYVT/do7uMfri8ppkVFUv/k+ztlzmupYf/trQp99ekDs7S4hQWaWXTAet0fjpVWy3WWm+xrz\n5k0txkUMbjehL/2H2HkzCH/w9xiqqwJuQyAZdKK/clMeHk3j0vmjOpyVOHFUDEEWI7tO8dm5mqax\nPquIR9/Y2a9eTE/xZe3MFu1PDrryrExMRgPvrT3acqs/h4Pw3/2GyFtu7HNvyTd4PH5ky/CNpmns\nKz9IuDmM0VGn7oYpvcFQW0PUDVdjPqxnuGhBQVS/+BquWae1KKdFRlH91vs4T5vbdMz6u7sIfebv\n/WpvT5k5LoGZ4+LJLqhiXVZgvW2TPEjUTddisOvjRFpYWNM5g81G2JOPETtnGqFP/Q1stoDaEigG\nlehX1NhZt6eIpJhQ5k5I6rBckMXE5IxYisobKBnA/PHOqLc7+ecH+3j+4wPsy63kyXf3UN9s+dxT\nkR3ZpVjMRqZmxrZ7PjE6lHNmpHKi0sbXu483HQ978jHCnv0HwR9/iPWXP+1TmzoS/cK6Iqoaq5kU\nJzAaBtVl7h92O5HfWopl904ANKORmn8+j/PsRe0W1yIideGfe3rTMeu9dxP6j6f6xdzesuyC8QQH\nmXh79RFqGxwB6cNYdJyo65dgrNI9eU98AhVrNlL15vs4p04/Wa6qCusD9xJ7+kxCXnkRXIMrPXxA\nvg1CiB1CiNXen+f9rffp5nzcHo1LzhiF0dj5dPqmEM8p6O0fyKvk3ue3sO3gCcalRbFopi6Uzyzf\ni9vT+WbYA0VReT3Hy+qZMjqWkCBzh+UuW5BBsMXEh+tzaXS4MebnEfbkY03ng/63CmN+Xp/ZdbSo\nhpiI4Dbpmvt8e+EOslU1/cLlIvL7txK0bm3Tobq/PoHj8is6raZZI6h+412c885oOmb9/W8J/fsT\nATO1r4iNDOGqhaOpszl5e3XfZ4kZqquIun4JpmOFAHjCrVS//g6ejNE4F51H1f++oubZ/+DOGN1U\nx1R0nIif/4SYs+bpk+BOkcUHu6LfRV8IEQIgpVzk/fmOP/Wq6x18tfs4cZHBnDG560Wzpo2Nx8Cp\nFdd3uT28tfowf319J9V1Dq46K5NfL53FssXjmT4mjn25lQG5oPuCk6Gdzld3jAoP4sK56VTXO/hs\nWwHWe+9uelQGMGgaoX2UM15Z20h1nYPRyZFtzu0tO4ABAxMHwYYp3ULTsP7iDoJXrmg6VHfP/dhv\n/JZ/1a0RVL3+Lo7T5zcds/7hd4Q++Xifm9rXnHdaGiMTrazLKkLm9+E8HO9Tk/nAfgA0s5ma/7yM\na/rMk2WMRhqvuoaK9duofeQxPAknQ5zmw4eI+s5NRF+0CMumDX1nV4AYCE9/OhAmhFglhPhCCDHP\nn0qrtuTjdHm45PRRmE1dmx0VHkRmaiSHCqupsw182OR4WT0PvrSNTzfnkxATyt03zeby+RkYjQaM\nBgPf+8ZkkuPC+GxrAesDHLfsCdvlCUxGQ9MTVGdcOHck1lALn64/QuOXa9qcD3ntJbD3flG8o97l\ntEcnR7Q4XuuoI7emgMyoDMItYe1VHbSE/+FeQl9/penvhtvvwPaTO7vXiNVK9Wvv4Ji/8OShB+8j\n9IlH+8rMgGAyGrn5ogkYgJdWyZbjRj3F4yHix98naMO6pkO1f3sa56Lz2i9vsWC/5TbKt+ym/q7f\n4Yk46XBYdu4g6qpLsTR7AjsVGQjRrwf+IqW8EPgB8KoQolM76mxOVu84RpQ1iIXTkv3uaMbYeDya\nRtYALsCmaRqrdxTyhxe2kl9Sx5nTkvn9LXPITGnpnYYGm7nj6mmEBZt58dODHDl26qSbllbZyC+p\nY+KoGMJCLF2WDw02c/m8NGxuA2/PvQYA+zXX4U5LB8BYXk7wh+/32i5fPL+1p9+0YcoQm5AV+uTj\nhD19MhRjW3oT9fc9oK8p3l2sVqpffRvHgjNPHvrj/QS/93ZfmBowMlMiOWdWKkXlDazcnN+7xjSN\n8N/9hpBm12LdPffTeO0NXdcND6fhZ/9HxdbdNPzwJ2jB+sRAg9utj1v1gVMTKAZC9LOBVwGklIeA\ncqBTJf/f1gIanW4unjuyTapgZ8wYp4ciBmqN/Zp6B0++s4eXP8vGYjbyo6umcMslE1vExAtqj/FW\n9nI+PPIpsmE3F14QhCekiqc+2Ep5zamRHbDdz9BOcy7e8iGJ1SWsmHEJJSNGU/f7P2K/+Zam86F9\nMDvUJ/oZI1qK/t7yIZaqabcT/ru7sD54X9Ohxosvo+6vT/RM8H2Eh+vCf+bZJw899AA4B/7JuDOu\nPmsMUeFBrNiQy4nKnidqhP79CcKee6bp74bbvt/tpyYtNo76+/9IxdrNTV6/+egRwv721x7bFWg6\nHpELHLcA04AfCSFSgEigw3hGvc3JlzsKibIGcfX5gpDg9k2usFWRVXwQi8lMsDmYEHMw4YkWEke4\n2VdYiClsEtbgUCwmS4drqmuahkfz4PK4cXlcuD1u/bXmJsQURGRIRLv12mPbgRKeeGMnVXWNzBiX\nwJ03zCQuKrRFXysPreaV3e/j8rQc/Q+eDE7g3q2fEx8eTVxoNLFhMcSGRhMXFs2E+LGMjcvw25be\nsudoOUYDXHDGaKKsHW883sSxY/D4wywbOYfHL76T17/7B+6cPAbuuB3+8hA4nVi2byWh4BDMmtV1\ne+3g8WjkldSSlmhlVLqeuZOQEIHL4+Zg5SESwmKZljF20K6f38S+fbB0KezZc/LYokUEv/c2CSHt\nryraPSJgxYcwejRUVGDKyyXh0+Vw66190Hbg+P5V0/jzK9t4c/UR7v/eGd3+Pyd8uhweuPfkgW9+\nk7BnnibM5L9T2bLB6fDIw3D77QCEP/U44bd9Gyaeeo7HQIj+88B/hRC+wNctUsoOg3Mfr8+h3u7i\n6rMzqa2x0d4OuPvKJS/ue516Vzt3fW+K9g9WfAmAAQNBJou+4qIGbk0XdY9X3DvCgIGFqadzeeaF\n7caJnS43haX15JfUIvOr2LS/BLPJwHXnjuWCOel4HC5KS3Xrax11vHLgLfaWHyTCYuX6SVcRbgmn\nurGaysZqqhqr2ZVXSIWtimrNToUtF095y1UsJ8dN4LLMxYyMSOvQ5r6gsraRg3mVTBgZjcPmoNTW\ndbpcxB13ElJfz9kH1/Leguv40pDMwr0FJMYFk3j5lYR4Qwi2x56g7vGe5YoXldfTYHcxfYyV0tJa\nEhIiKC2tJbvyCA1OG6clzqSsrK5HbZ8SaBohzz+L9f7fYWg8ubxw4wUXUvvM82i1TqjtK4/cQNgP\nfkz4n/4AgPv+B6i46EqwdB3KGyhEagRTRseyM7uUj9ceYd6kjlO4W5OwcyParbfiu0045i+k+tGn\nobfp3UuWEv38f7Fs3wpOJ45bb6N6+ScwQPsfJCS076T2u+hLKV3ATf6WX/7VEcKCzZw7q624eTQP\nK3O/YGXO55gMRi4bvZgwSxgOt4NGtwOH20FJdS27jpaQGBtEUnwQDu/xRo8DAwZMBhMmowmzwYTR\nYMJsNDUdMxlMmAxGzEYzuTX5fH1sIztP7OGSURcyAkH+iTryS2rJK66jqLwed7O1flLiw/ne5ZMY\nmdTyg8+uPMwL+16n2lHLhJhx3DzpeqKC2/5zrsz08OfXdnBkfw3XnDOa+TNjqW6socxWwdfHNrKv\n/CD7yg8yI2EKl45eTIo1MNsA7sjufEJWayzrvybkvXcAMGkeLpofz/OV2TyatRqTRWPpjYu5+L23\nMQAh771N/X0PoEV3f10c3yBu67GRptDOII7nG0pKiPzpDwn68vOmY1pIiB4iu+W23oV0OsB22/cJ\nfebvGCsqMOXnEvLW69iX3dzn/fQVBoOBGy8U/O7fm3nji0PMmZDYZRo3gHn3TliyBIM3t941cTI1\nL74GffHUZDRS++iTxJx/JgaXi6BNGwh5/ZVT7nMcCE+/W9Q2OPjGggxCW4V16pz1vLjvDfZXSGJD\nYrhtyo2MikxvU9/l9vCzdeuwl5n40Xnze/S4X2dz8tXuAnbYt1Ds2c1bh9/DUxeFI3cSWkMUQWYj\nGckRjEyKYJT3Jy0xvMUOR26Pm09y/seqvNUYDAauGHMx5488u8OJQxazkR8tmcoDL27j3TU5pCVE\nMm1MOqMi05mVOA1ZeZgVR1exq3Qvu0v3MTtpOpeOvoDEsL7dMHu71JdUmNXBkhctcDqx3v1/aIAc\nP4KPb7uIreGrsYR5cLksmDQjLzi2c+i3S/juYx8RZrMR8sar2H7w427b5Vt+ofUg7r6yg1iMFsZF\nD84NU4JWrSTiztsxlp9MPnBOmUbtP/+NWwTuRqZZI2i4/Q6sD/4egLDH/4L92htOaW8/MTqUGxeP\nZ2d2GR5Nw0gn3223m6DVnxNxx+1Qr89+d6emUf3Gu2hRbZfw6CnuSZOx3X5H09yU8PvvoXHxxWgJ\np85G9qe86IcGmzj/tJZinldTwHNZL1PZWMWkOMG3J93QYWqe2WRkamYcm/aXUHCiro3n3RVOl4e/\nvb3b61kmEmo9h7DR2TRYCwidsolZcadx7YRLsAaHd9hGua2SF/a/xtHqPOJCYrhl8jK/lgaItgbz\n4yVTeeiVHTz74T7uuXk2yXHhGAwGJsSOQ8SMZV/5QVYcXcW2kl3sOLGHeSNmc3HGecSFtj9rtjvU\nNDiQBVWMSYkkpp1li1tjfuFZ1iW4WHnT1eRk6hd5angSk8Jn8eEKBykjgwgat4f1U+HwQ9dw55Of\nM+qF57F97/ZuPwIfLarBZDSQnmhtOlZmK6e44QRT4ycSZOpfsTLtzSLkzVcJXvkJmtmE84wFOOcv\nxLnwLDzJKV030NCA9b7fEvpiy7mKDbffQf1dv4NgP8ZSeon91u8S9o8nvd5+HiFvvuZ3/v9Acea0\nFM6c1vHnazp8iJDXXyH47TcwNVsCxBMdTfUb7/n3v+km9T//FcHL38OUn6vP3r33Lmr/eeosa236\n/e9/P9A2dIpH034v0vQ7saZprDu+mef3voLNZeey0Yu5XiwhuIsdkdwejW2ylOiIYMTI7oUSXvs8\nm53ZZcydmMgd10zjmjMFi8fPY0xUBrk1+RyqOcSm4m1Yg6ykWke0eZLYeSKLf+z5D6W2MmYnTueH\n028hIazrXHcfMRHBxEeHsHl/CftyK5k/Oakpg8lgMJAYlsD8lLmkWpM5Vl/MwYps1h7bSI2jlrSI\nFELMPX9s3by/hF2HyrjgtHTGpkV1WK6qsZrPD3zC843bWbdwHNVRocyuCeW6M77DFWMuZmJiBkeO\n1yJzG7jzvEsJMbrZwwnWnC0ILT7BqKhRaKMz/bbL5fbw+ufZpCdam8J+4eHBrDmyif0VknPTz2JU\nZGDHOgAMFeWEvPoS1v/7Gda//AnL9q0Yq6swVlZiydpN8CcfEfbM3wl+9y3MB/ZjqK9Di41Ds1pb\ntGPes4uoa68kePXJcI47OYWaF17F/q1bwdxPvllQMHg0gtau0e06sB/bt2+Dng5uDhCGmmpC3noD\n612/xPrAfVi2bMJY12x8JzSUqtfexT1jZseN9AaLBdfYcYS88yYA5gP7cM6Zh6fZbN7+IDw8+P72\njhv6Y9W63qBpmlZWVofD7eAN+T6bi7cTbgnjlklLmRg33q82Guwufvrk16QnWrn323O6ruBl495i\nnluxn9SEcO656TSCg1pe/C6Piy8LvmZlzuc4PE4yo0Zx7firSI9IweF28u6hD1l3fDMWo4Vrx1/B\nGclzepxN8tbqw3y6OZ8pmbHcec30duOXHs3DtpJdfJzzP8ps5ViMZs5MPYNLRp9PqDm0nVY75/G3\ndpN1tJyHf3AGidFt6+dU57G6YB07S7PwaB7C6+yc9+UBzpP1mJevbuGdrt19nBdWHuT688axeE46\nOY/+iudGN1AdHca0QhtLr3+IiCBrmz7aI6eohgde3MaiWanctFifcZuQEMF9/3ucAxXZPDj/bmJC\n+u6RvQUuF0GrPyfk9VcJWvUJhh6kN7rGjsM5/0ycC8/EmJ9H+CN/bNFO42VXUPvoE2gxvX9a6zZ1\ndcTNmdoUXqp99EnsN327/+3oLh4PlnVrda/+k48wtLMYmic+Afs3ryfsFz+lNNK/MareEPGDW5vG\nt9yjMqhYuxlCu/897CkJCRHtis0pH94xGAycaCjluayXOV5fzKjIdG6bcmO3NsUICzEjRkazP7eS\nytpGv0IVhSfqePHTg4QGm/jxVVPbCD6A2Whm8ahFzEmaybuHPmJnaRaPbH2CBanzOFKVQ1F9CanW\nZG6dvJQR4f5nF7THNWeP4VhpPVlHy3lr9WGuO7dtOqLRYGTuiFnMTpzO5uLtfJLzOV8WfM3u0n3c\nMnlpt1abbLA72Z9bwchEaxvBt7saeffQR2wo2gJAijGSy579iDPXHSLY4aLqjfdwtgpHTBujb3az\n50gZi+ekM/aK7/KXixfy9A8XsXvGSHI3/pVvTV3GhNhxXdrWNCmrWX6+3WnnUOURUq3JARF8U7Y8\nGSY4UdLmvBYSQuMll2G/bhlaaBhBG77Gsv5rLFs3t1iGAvRp++bDhwh9qeVyFFpYOLUP/YXG65cF\nZLDWL6xWGm7/KVZvOmPY43/Bft1SCDo19xc2Hj9GyMsvEPLW65gK2k7W0sxmHIsvxn7DjTjOPR8s\nFsISIqC0vTzAvqXuDw8T9MXnGKurMOXlEv7Yn6n/7X1dVwwwp7zobyncxd+3vojdbees1DNYMu5y\nLMbumz1jbDz7cyvZfbiMc2amdlq2we7i6fezcLg8/PgbU0mK7Xwqf0xINLdNvYkD5dm8dWg5645t\nAuCs1PksGXsplj6ILxuNBr7/jUn88eXtfLa1gNBgM1csbP9x0WQ0MT9lLnNGzOLTnM9Zlbeax3b8\ng8szL+x08Lg5uw+X4/ZobSZk5dbk88K+1ym1lZNqTebqzEuZe/1tBO3Rs2YaL74M57nnt2kv2hrM\nqKQIZH4VtkYXoWPHYZ02j7se+YQVl07n9aVn8Pdd/+aCUedw2ejFmIwdhxRyfMsvNMvc2VNyEJfm\n7rMJWcaSYsxZuzFn7SHos5VYtm9rt5xz9mnYr1tG41VXtxgQdJ1+Bvz8V9DYiGXHNizr1mLZsA7L\nti0tUjCbt1Pz9HN4Mgd+ANp2y22E/eMJjOXlmAoLCHnj1RYT604VzNu2EHXtVRjr2gq4a/JU7Dcs\nw77kWrR4/8OpfYmWmEj9vX8g4hd3ABD69BPYl3wT98RJA2KPj1Ne9P+6/lksRgvfmnQ9c0f0bCIP\n6KL/2ueH2NWF6Guaxn8+OUBJpY2L5430L2vFy8S48dwd83M2Ht9CXGgsk/t4hcewEAu/uG4GD7+6\ngw/W5WA2Gbj0jIwOy1uMZi4fcxHjY8by4v7X+eDISmTF4Q7TRJuzzZu140vV9GgePstbzcc5/8Oj\neThv5FlcnnkRES+/RNCeXYA3rfCBhzpsc9qYOPJKatmfW8lskYDt1u8S9fUavrFiNxOPN/D4vdfw\nWd5qDlUe4duTlxLfwWD00aIaQoJMJDe7Ge8o2gvQ/V2yPB6MuTmY9+7BkrWnSeiNpR1vBONOTKLx\nm9djv35Z1xk1wcH6oO4ZC/S/7XYs27c23QRMebnYl91Mw52/PHUyZaxWGn50J9Y//A6AsL/9Ffv1\ny04pb9+0N4uoG65pIfiemBjsV19L4w034mq2FPJAYl92MyFvvY5l80YMLhcRv/wpVR+tGrDcfRgE\nop9sTeSWSctItfq/5k57xEeHkpZgZX9uJY0Od7vhGtCXb96RXcqEkdEsOdv/wUUfFqOZs9Lmd12w\nh8RGhvCrG2by8Gs7ePero1jMJhbPaZuq2hwRO5a75v6saULYQ1se5+ZJ1zGpgxUoGx1u9uZUkBwX\nRpfPC68AACAASURBVEp8OOW2Cl7c/wZHqnOJDo7iponXMiF2HIaKcsL/dHKsqOGOn+MZOapDO6aN\njeOjDbnsOVLGbJGA48KLcaekYjp+jHE7DnFfUQYvThjFtpJdPLTlb5yWNJ0UazIp4Ukkh4/AGhSO\nrdFFcXkDYmR007iGpmnsPL4XqyWcjOZpu5oGNhvGqkoMVVUnf5eXYTq4H3PWHsx7s9r1FFujWSw4\nLrwE+w3LcCw6v+eDqyEhOBecibPZmjenIk3eflmZ7u2//oo+qHwKYDpyiOhrr8To3cHKExdH7SOP\n4bjwkn7JcuoWRiO1f32CmHMXYHA6sWzdTMjLLwzoZ3nKD+TaXY1abWXfbJrw3tqjrNiQy4+XTG3X\ngz+YV8lf3thJVHgQ990yl6jwU8ezaU1JZQOPvLqDqjoHNy0ez6J2Jq+1RtM0VheuY/nhT3Brbs4b\neRbfyLwIc6tw2baDJ/jH8r1cNj+DtPFVvCmXY3fbmZEwlaUTrm5Kj7X+8s6muLR7ZAYVX3c+UOXR\nNH721DqMRgOP/WgBBoOBsMf+TPjDDwLgnDOPyhWfsal4O+9kf4Dd3TIMEhkUQZQpjpwcjckjRnH5\n7GkkhyVSueptHgzLYsHhOm7/6AiGqkoMlZUYq6vaDaX4gxYWjmvKVFxTp+GcNgPH4ovR4uJ61NZg\nJfTpJ7Hefw+g57RXbNo54KJqLCwg+vILT657HxlF9fsruuXZ+2Zv9ydhDz9A+GN/AXSbK9ZvQ0vq\n3ThfV3Q0kHvKiz6g9dU/yJf1sXBqMrde2jIMUFnbyP3/3UK93cWvl87qNEXxVKGovJ5HXt1BTYOT\nWy6ewJnT/cs5zq8t5L97X+OErYxREencMnkpCWEnBe2ZD/ayJfsYsxaVcKBmL0GmIK4ddwWnJ5+m\nDx47nVi+XkPUDddg8F4/1S+9geOiS7rs+98r9rNhbzH3fXsOo0ZEYCgpIW7WpKbslYov1uGeOg2n\n20lxQylF9cUcryvWf9eXUGFvu456aEMjtrBg7nzif8zf2P39CDxxcbimTMM1dTquadNxTZ2Ge/SY\nAX0EPyWo///27js8qqJt4PBvs6mbQkISRHofQIqKIEgRROqLwosgIqh0QZpgBaWoFAERJIBUDSBF\nsWEvIAii6AcGAipDh5dQE0JCNptNtnx/nGUhkkAg2ZLs3NfFlZyzp0w2w5PZOTPPGIlu3AC/ZG1W\n9qWZc8jqV6DlL1xCd+4ckV074n/4EAD2kBAufrQBy71Nb3Bmbp4I+phMRLVuhv9RLZ1KVrfuXFoS\n79JbqqCP1tJ8bv52bHY7c0a0cHYPWKw2Zq5J4FBSGr0frEm7e67fXeJNTp7PYOaaBIymHAY9VLdA\nC8wAZFmy+OjABn4/s4tgfRC9xH9pUvZucixWRi3/An2V3dgDTFSJqEj/yFaU33cI/107CUjYhf/e\nPbmGxJnbtiN9zccFGnHyxz9nWbThL7q1rMrDzbUH0eFP9yf4s08AMD3Rj4zZ8/I9f95nu9hz8ji9\n7g/F/MM6kgLMnKhYmsAcK7NeXI/hX7mB7EFB2CKjsEdFYS8Vic3x1Vq5ihbk6zfQJugU98RsLhKy\nMI6wya8AYC1Xngu/7/ZIa193MZXI/3bB/6+9gNbdlrbqwzwHDdyIR4I+ELB1C5E9HnZup639mOy2\n7V12PxX0HeK/3c/WPacY37eRszW/ZuMBNu48SZM6ZXj64TuKXWbG42cuMWttAqZsC0O71qNx7YKP\nQf7jzJ+sk59itmZzb9lGmC/aSTD9iQ7olpBKj/c3EXg+/9TU9sBAUn/+DWv1Gw+1BG0o6Kh3fqHK\n7eG8+qS2gHfAjl+JfLijdj2DgZQ9+/OdGv/8wu1YzdnErxqB/8n/XSlH376kP9gZe1SUFuQjI7FF\nRrl1XHSJlJlJ9D31r7T2Z7yt5f9xp4wMIh/tRsBObYiw3c+P9KUrbrg8ZH48FfQBwocPIXj9OgCs\nlSpz4ecdEJr/bP7CyC/o+9zn1ztrasO3Eg5plfj3v8+ycedJysWE0q9T7WIX8AEqlw1nbK87CQrQ\ns+SLv0g4eL7A5zYpezfj7hhMZXMQv5/Zxe6sP7FnhzAwbhePz/ww34BvrVAR80PdSFvzcYEDPmgj\nkGpUKMXRU+mkOxa4zrm3GZY6dwCgy8wk+KO1eZ6blmHmQroZIXc6A77dz4+MyVPRrVxJducu5DRr\njrVOXa31rgJ+4RkMZI4cc2Xzndlwi89JbklWFqX69XEGfNBWtrrVgO9pGa9NwxalzTHSnzhO6OwZ\nbi+DzwX9upWjCPT3Y/fBZJKSjcR/u5+gQD3D/1vvugt+e7tq5SIY82hD/PV+vPv5PvYeufFqYXq5\nn7CXxlK7cTOmD4ij+ye70CXdjmFHHdr9emVcui2iFNn3t8E45nnSVq4jZe8BLvz5F+nLV5LTqvVN\nl7Vh9WjscGVFM50O01Wtx+D3l127yLTdzqlF8QDUTJJaucIjSP/gQ0zPjFTdMy5kemqAc01Y/akk\nglevdM+NLy8Av3Wzc9elaTO1yWvFlD0mhozJU53bIUvfRXcu/+HBruBzQT8wQM8dVUtzOiWTOR/t\nxpxjZWDnOtwe7ZqPWO5Us0Iko3o0QKfTMf/Tvfx97MK1B1ksBH65gVLdu1C6ZRNC3l+GnzEDf6uN\n2r+byUxqSNOUY5gHDiZ9/mIu/LqLlAPHSVu/gcxxE8nu2BnbbYVL43xldu6VP0xZPXphC9PmDvgf\nOkjAtp+vnJCZSfjQAZz4JQGAWmcOYK1SlYvfbCT7wQ6FKotSAAYDmVetKGV4Z7brlwO02Qgf/Uyu\nBeCNL79K1qChrr2vG5gf60OOI++PzmwmZNmiG5xRtHwu6APOxb0vpJtp37gi99xEH7i3q1M5ipHd\n62O325n3SSIH/qeNZdadO0fI7JmE3NuI7NFjSNp/nH3l67KjehM21W3DZ+0HsLb3ywDUf+FpMqa/\nhfnR3lhr1CzyUSzlYkKJjghm39ELWG2O9XPCwjD3urI2ach7SwHwO32KyK6dCP7sEw6U1bqRKlct\nQ+p3P7k01bCSm+mpgVjLaEMM9adPuba1b7cTNv4FZ983aNlGM8e84Lp7upNOR+bIsc7NkPeWoruU\n7rbbF9/+jEJoWCOGwAA/qpSNoEfrop327nf8GBHDh2Arc5vHkmbVqxbNM13vYMHn+3h77Z+UNqdj\ntIAx6C5sPe/J/8QcLatnrYouSlbmoNPpaFAjms1/JnHoZJoz86mp/2BCli8BIPC7rwn86gvCxj2P\n/uwZ7MDBsjW4zWrEtnqt98xe9RUhIZhGPkvYhHGA1trP6vNk0Sw+8i+G6W84/+iDNqLrlheA91LZ\nnbtgqV4D/8OH8EtPI3jF+5hGjHbLvX0y6EeEBjJ9SDPCQgLw1xdtKzZ49UoC/tBy72C1kh6/uugq\nq9VK8JpV6A/sR5eRgc6YoX11/PPLuOTc385kIqzGvSx6YAjZ6Ig0GymfeoqwrAwMWAmuWpGg+vUI\niS2NIdif0GB/DEH+lIsNK9AKRIXVsLoW9BMPpziDvrWWILtFKwJ/2YrOZqPUgL7O40+VLk9GcDh3\n1L1NBXwPMT05gJC4uejPnUV/5jTBq1eQNfDportBdjah09/AsOAd566sbt3JmDmnRAV8APR6TCOe\nJXyMtoBQyOIFmAYPdctwWJ8M+kCBMm3eipxWrWHuWwAEffsVwSveK7IJLaETXsawbHGBj2926Hea\nHfr9StnuaYJp4BDMXbp6fGZl7UraA/XEwyn0bFPDud/UfzCBv2zNdawtKoo9096FA7ZrVspS3Cgk\nBNOoMYS9qnUDhk6fgj08AnPPxwodlPWHDxI+bBABuxOc+8ztOnBpwdJil8+/oLJ69MIwYyr6M6fR\nnz1D8Pp1blm0xif79F0pp0UrTAOHOLfDJo5D/8/fhb5u0LrVNxXwAWyhYVjLlcfU50lSN23j4jcb\nMT/yqMcDPmgP1GtXjiIp2Uhy2pWJXtkdO2MteyXPkqWWIPW7zRyM0GYbV7395lY+U4qW6Yn+zt+P\nX3oaESOeJuLxHvhdNWfiptjtBH+wgqi2LXMH/LbtSF+2smR/qgsKwvT0cOdmyPy5YLW6/LY+NznL\nLbKyiOrQBv9//gLAUrsOqd9vueVx4/57Eojs0t6ZRya7TVvMnR/CHhaGPSzc8dXxL9Tx1RDq9WkE\nNv95klU/HKBv+1q5Fr4P3Pg9oa++jOWeJmRMn4U9PIJpq3Zx5FQ6C8a2Iijg2pafJyfc+Br/hF1E\nDO6H/sRx5z5baBjGia9ricQKWO90F1IIf240QV9/4dxnDwzE+OrkW1pCs6C8qa7oLqVT+u56zuRx\nactXkv1QtyK5tpqc5U7BwaQveR+7I8j77/+HsEnjb+lSuuRkIvr1cQZ8S+06pC1fRdZTAzA/8ijZ\nHTqR07wlloZ3Ya1eE1vZ27GHhXt9wAdoUF0bRXX10E2A7Ac7kLojgUvzF2MPj8BitXH87CUqxIbm\nGfAV97Lc1YgLW34jc/BQ7I5uHT9jBuEvjaVUt87oDx+84TUCtm4hqvV9uQK+pZYg9dufMA0dUSzq\nb1Gwh0dgGnBljoohbs61c1SKmG+8sx5gFbXJeONN53ZI/HICv/7y5i5isRAx+KlcGQXT41dDWMGW\nFfR20aWCKR8byj/HUzHn5P+xNum8kRyLjSqqP997hIVhnDqTi1/+gKXmlWVLA3f8SlSb5oTEzQWL\n5drzzGZCJ79KZI+Hcy1UbhowmNQffsZav4E7Su9VTIOGYXeMggrYnZB7jooLqKDvQllP9NMemjqE\njxmOnyOAF0To6xMJ3L4NALtOx6V3l2KtVuMGZxUvDapFk2Oxsf/4tdkzLzt6RhvDXK2cCvrextLk\nXlI3/YJxzPPYHWsM6LKyCHtjIpEdH0C/b6/zWP0BSWSnthgWXkmoZ4uJIe2DD8l4czYYrr9CXUll\nj40lq/eVkWqGeXMKdb2gjz+k1H//k+/rKui7kk7HpbfnYS2v9Vf7XbxI+LBBBXpYE/TJRxgWzXdu\nZ770CtntOrqsqJ6S1+zcf3Muj6ha+t4pOJjMcRNJ/X4LOQ3udO4OSNxNVPv7Mbz5BsHvLSWqXSsC\n9iU6X89+4EEubP6N7PadPFFqr5L5zCjsjlFKgVs3478n4QZn5C3o0/VEPDPY2VjMiwr6t2D+/LmM\nHPk0ffr04JFHujBy5NNMmPBynsfaI6NIf3c5dkcfZeCOXzHMmXXd6+v3JhI+dqRz29ypi7acXj4W\nLpzHU0/1ZvfuP2/hpykaP/+8meTk/LNx5qdGhVIYgvxJPJxMfoMKjp5OJ9Dfj3IxvtkSLC6s9Rtw\n8bufyHj1NeyOEWI6i4XQt2cR/vJzznTc9qAgMqbOIG3tJy5fSKS4sFWugrlrd+d2SNzcm76G/+87\nCB817IbHqaB/C0aMeJa4uMX07duP9u07ERe3mDeu6r//N0vTZmQ+f+WPguGtN/Hf8Vuex+oupFCq\nfx/nfxBLzVpcmr/oug+2tmzZxKJF73Hnnbe+hnBhffzxOjIzM276PL2fH/WqlSYl3UxSsvGa183Z\nVpKSjVQuG47eRx7uFWv+/phGjSF1y69kN7122VBLnbqkfr8F0+BhJW/CVSFljriS3yjoy8/RHzlU\n4HP9jh2lVL/e6LK1zLWWWnkvhQo+PDmrqFxunRqNGcyYMZWMjEskJ5+ne/eedOvWgxEjhlC6dDSX\n0tOYd28zXjtxjAw/P86OHU7XEc9Sr0kzlixZyMyZc9j4/besmf4aX544zq7gYD6PjqbP3AXMnDKJ\n7OxsUlKSGTx4GC1btnbe//33l5KcnMwLL4xm9uw4lixZyN69ewBo164jPXs+xtSpk0lPTyM9PZ2Z\nM+eyevUKEhN3Y7PZ6NXrcdq0eZC//tpHXNzb2Gw2YmNjmThxCn//vY/4+GXYbDZMJhOTJk2hTJnb\nmDjxZYxGI2ZzFkOGPIPFYuHgwQNMmTKZhQuX4X+T68c2qB7NH/+cI/FwChVicz+kPn72Ena76top\nbqzVa5L2+TcExy8ndMpkdMYMTIOexjjhdZekbigJrPXqY27bjqBNP6Kz2wlZMO+6Cwpdpku7SKk+\nPfFL0bpIbTExpK1eT36Lexb7oB+yMA7DrOn4GW++lZkfW2gYmS+M01L2FlBS0knatm3P/fe3ITn5\nPCNGPE23bj3Q6XS0a9eBli1bs/+XrXQeOoAOp09zVq/niaWL6Nb7Cc6ePU1OTg47ly3CPyODFL2e\nTWFhtBg0lKM6HY891pe77mrEvn2JLF++OFfQ799/MN988yVz5izgjz92cObMKZYsicdisfDMM4No\n1Ehb4rBRoyY8+mhvfvttO6dPn2LhwmWYzWaGDu1P48ZNmTVrGq+/Po1Klarw1VcbOH78KMeOHWXC\nhDeIiYlh1ar32bx5Iy1btiY9PY3Zs+NITU3lxInjNGvWgpo1a/HCC+NvOuCDlitIByQeSqZz09wL\nqx9R/fnFl58fWQMGk9W7L7qMDOyx165LreRmGjWWoE0/AhD84RoyXxx//ay2OTlEDHgS/4MHAK3r\nLC1+LbbKVfI9pfgH/XfjijTggzbmOOTduJsK+lFRpfnoo7Vs3foTBkMY1qse1laqVAWASFGb9S3u\nZ+Ov2wiz2bAaMwj+YAVNmjQjccFckg8f5CGjke0GA/9XvQZPDh3B8ePHWLnyPb76agM6nQ5LXsPg\n0D5xHD9+jIYNtZSt/v7+3HFHfY4ePeoogxZMjxw5hJT7GTlSy5litVo5ffoUqakpznJ2cYw4Onv2\nDHPnzsJgMHD+/DkaNLiTqlWr8fDD3Zk8+RUsFgs9ejxW8Dc2HxGGQKqVi+BQUjrGrBxCg6/Mwjx6\n2hH01cid4iskxDlnRbm+nKb3kXNPEwJ2/oEuO5uQxQsxTnw974PtdsJeGkvgti3OXZfeWYilyb3X\nvYfbO0mFEH5CiEVCiF+FEJuFEIVKc2kaNhJbaNGOW7eFhmEaduOAf/WDx3XrPqBevfpMmPAGbdq0\nxW63OV+7vBrXunWrqduxE693+A8dLl3CrtMR9upLtAU+WL2S2mYzzY1GVlauQrkmTdHr9SxfvoiO\nHf/DhAmvc9ddjfJ92KnT6ahSpSqJibsBsFgs7Nu3h4oVK+YqQ+XKVbn77kbExS1mzpwFtGnzIOXL\nVyA6OpaTjqn0a9asZOvWLcycOY1XXpnM+PGTiImJxWazceTIITIzM5k5cy7jx09mjuOhtJ+fHzab\nLY+SFUyD6tHY7Hb+Opp7DYCjp9MJCwkgtpTqElB8gE5H5qgraZeD45ejc8zW/beQhXGEfLDCuW18\n6RXM3Xve8BaeaOl3AwKllPcJIe4FZjv23RLTMyNvqkVelHQ6nTOYNm/eirlzZ7F16xaqVq2GwWAg\nJycn1/HNm7fUjokqTe2KlQg1Z2ExmWj+2gTGVavGEKOR6hUqcjo8jMdbPwBAmzYPsmDBXNavX8cd\nd9TjUp55t7Uy3HdfCxISdjF06ABycnJo27YdtWrVdpYVoEWLViQk7GL48MGYTJm0atUGg8HAiy+O\nZ/r019HpdMTExNKzZ2/at+/E8OGDiImJpVKlKqSkJFOhQiXee28pmzdvxGazMXiwtqhFvXoNmDJl\nEnPmLCA8/Obz4zSoHsNn246y51AKTepoIzrSM7NJTsuiXrXSxXIZS0W5FdntO2IRtfGX+/HLuERw\n/HJMo5/LdUzg118S+voE53ZWj15kjn2xQNd3e+4dIcRs4Hcp5UeO7ZNSygrXOaX45d4pAP0/fxPV\noTW6q1YgsoWGcdFHFwex2+08t2A7FquduSNb4OenI/FwMnPXJ/Jw8yp0a1ntuud7Uz4VxbsVh7oS\ntG41EY7hl7aYWFJ27XPm7vLf/SeRXTs5R/hlN72PtPUbrkmkmF/uHU+09COAq5urViGEn5Qyz76B\nvXv3YrXeereB14oIxjhmLBWnT3PuOv7aa6SFBULSEQ8WzHOqlw1k16EM/kj8h4qxweyR2sfaiMBM\nkm7wnpw541cy64lS5IpDXdE1vYc6ZcsSeOYMfsnnyVo8j5SePQk4c4ZaT/R1BnxzxYocmD4Na3LS\nNdeIjW2Y57U9EfTTgas//+cb8C/TF/FCJ97i4mO9CE46SeR333G+f38y2rfDl9OJ1akYyq5DGRw8\nlUWVsgZOXdCSzFUqE1KgOlBS64lS9Ly+ruiDSH7yScrNnAlAmZUrSO/ciWqjRxHgmARpiYjg2Pz5\nEF36puKGJ4L+duAhYL0QoimQeL2D69ev7/UfxQrlrflcfAsCgMItN178RZa2sPbncxw+a6HvbZVJ\nSvkf0RHB1Kh642f9xeEju+Idik1dGTYG29Kl+KWmEnTyJLX79EV//BgAdn9/MlaspXSzljd9WU/8\nufsMyBJCbEd7iDvGA2VQvFBwoD+iUiQnzmZwKCmNDFOOGqqp+K7QUExXLUd5OeADXJo9j5zmNx/w\nwQMtfSmlHbhxggjFJzWoHsPfx1L54hdtfoFaKUvxZaaBT2NYOA9dZqZzX+bo5zBflZXzZnl5x5bi\naxo6sm7+dUxLtVxNzcRVfJg9OhrTE/2c2+aHumEcNyH/Ewqg2M/ILS569HiItWs/JaAkr/lZBG4r\nbeC2qBDOpprQ6aByWdXSV3ybcfwk8A/AHhRE5ujnCr2qmAr6bqImFxVcg+ox/Ljzf5SLDiU4UFVR\nxceFhGCc9EaRXU79j7oF33zzJdu3b3VmvuzZszfbtv3MkSOHGTFiNFu3biEp6SRms5mePR+jQ4fO\nznOPHDnE/PlzsVptpKVd5PnnX6ZePd9bIu56GtaI5sed/1MrZSmKCxT7oP/RT4f4v/3nivSajWuX\n4dEHrr8socmUxdtvx7Fp0w98+OEaliyJ588/d7JixXLOnj3D4sXvA/DHHzuc59jtdo4ePcqIEc9S\nrVoNfvzxO77++ksV9P+lTuUoBv6nDnWrlPZ0URSlxCn2Qd8TdDodNR2LQYeGhlGlSlUAwsPD8fcP\nYOTIscyYMRWj0UiHDp1ynRcTE0t8/HKCgoLIzDQSWsTJ4koCnU5H8/q3e7oYilIiFfug/+gDNW7Y\nKneF/Pro7XY7Uv7DtGmzMJvNPPJIF2f3jt1u55133mLSpClUrlyF5csXc+bMaXcWW1EUH1fsg76n\nXA76uYO/9v2FCykMGzYAPz89jz/+BHq93vlahw6dmDDhJcqUuY3ateuSknLz68oqiqLcKrdn2bwF\nJTLLplK0is3UesXjfKWu5JdlU03OUhRF8SEq6CuKovgQFfQVRVF8iAr6iqIoPkQFfUVRFB+igr6i\nKIoPUUG/BOrR4yFycnI8XQxFUbyQCvolkMroqShKftSM3Ftw4sRxpk9/Db3eH7vdzqRJU/j00/Uk\nJu7GZrPRq9fjtGnzIAkJu4iPX4bNZsNkMjFp0hTKlLmNiRNfxmg0YjZnMWTIMzRu3DTP+9wom2eL\nFvczbdprKqOnoigFVuyD/qeHviLh3N4iveZdZerTvUaXfF/fufMP6tatz7BhI0lM3M22bVs4ffoU\nCxcuw2w2M3Rofxo3bsqxY0eZMOENYmJiWLXqfTZv3kjLlq1JT09j9uw4UlNTOXHi+HXLkl82z/Xr\n13H33Y3ZsyeBJUviAZXRU1GUGyv2Qd8TunTpyurVK3juuVGEhYVSs6ZAyv2MHKktYmy1Wjl9+hQx\nMTHMnTsLg8HA+fPnaNDgTqpWrcbDD3dn8uRXsFgs9OjxWL73uV42z+zsbAwGA6NGPacyeiqKUmDF\nPuh3r9Hluq1yV9i27WcaNryL/v0H8+OP37Fkybs0aXIvL7wwHovFwqpV71O+fAXGjh3BRx9tICQk\nhKlTJ2Oz2Thy5BCZmZnMnDmX5ORkhg0byH33tcj3Xtfrn09JSVYZPRVFuSnFPuh7Qu3adZg6dTIB\nAQHYbDamTp3B999/y/DhgzGZMmnVqg0Gg4H27TsxfPggYmJiqVSpCikpyVSoUIn33lvK5s0bsdls\nDB489Lr3yjubJ+h0EB0dozJ6KopyU1SWTaVE8JXMiUrh+UpdyS/Lpmrpe4HZs2dw7NiRa/a/9dY8\ngoKCPFAiRVFKKtXSV0oEX2m9KYXnK3VF5dNXFEVRVNBXFEXxJSroK4qi+BAV9BVFUXyIW0fvCCF0\nwEnggGPXb1LK8e4sg6Ioii9z95DN6sAuKeXDbr6voiiKgvuDfiOgvBDiJ8AEjJFSHrjBOYqiKEoR\ncVnQF0IMBJ791+5ngGlSyk+EEM2BD4AmriqDoiiKkptbJ2cJIUIAi5Qyx7F9UkpZwW0FUBRF8XHu\nHr0zEUfrXwjREDjh5vsriqL4NHf36b8JfCCE6AxYgH5uvr+iKIpPKw65dxRFUZQiorJs5kMI8SJa\nV1RVKaXZC8qzBRiS12gnIcQxoJaUMttF964KvAWUBgKAPcBLUsqMPI6tCDSUUn7lirJ4G2+rJ6Dq\nirfylrqiZuTmry+wFsh/PUP3snN5dZS8X3MJx8P3DcCbUso2UsoWwO9o701e2gLNXVUeL+Rt9QRU\nXfFWXlFXVEs/D0KI1sBBYDHasNIVjtZTAnAXYEP7xdUFZgBmYImU8gMXF22yEGKLlHKxEKI28K6U\nso2L7/kfYIuU8v8u75BSrhRCDBNC1ACWo7XoMoHHgZeBECHE9pLegvPiegKqrngVb6orXtXSF0Js\nEUIIT5cDGAQsd3w8NgshmqC1kDZKKVsDnwKvOPYFSSlbuek/sidUBa5d4QWOATuBqVLK+4B3gIbA\ndGC1q/8Te0ldUfUkN6+rK15ST8CL6opXBX20H9ijT5aFEFFAJ2C0EOJbIAIY6Xj5R8fX7UAtx/fS\nhWUJE0Jc/jSmI/d7k/+K6UUrCaiSx/4aQDDwG4CU8ksp5eX3xx1l82hd8aZ64iiPqit5UzHlX7yx\neydWCPEWWiW5HXhVSrlBCJEIbAEaoP0Su0op011w/77AMinlS+DspzwGnAfuBbYB9wF7Hcfb9Bhs\nKwAAA9FJREFUXFCGy+KB+UKIrUAs8B3aewJwtwvve7UNwCtCiMaXP7YLIQahvR9fo82o3iSE6I32\n8C4d9zUmPFlXvKmegKor16NiylW8raUP2se+2VLK9sAQYLhjfziwxvFRKAntL6crDARWXd6QUpqA\nj4GawHBHP1wHYKrjEFe2ImYDs9Aehq0H1gGdhRCb0foBXd6CkVIagYeAV4UQvwghdgCN0fofXwTG\nOcrTB1iNVnG7CiEedXXZ8Gxd8aZ6AqquXI+KKVfx+Dh9IUQYkCWltDgqxGi0BzwWtB++opTyASHE\nUaC2lNIshJgO7JdSrnBjOTcDj0gpL7jrnkpuxaGuqHriecWhnjjK6ZG64g0t/XighRDCDygDzAFW\nSimfRPvodXUZ1Uwy3xaPqivKjcWj6km+vKFPfzYwz/H9emA/8JYQYjSwA63vLy9u/WW5YbibcmNe\nX1dUPfEKXl9PwHN1xePdO4qiKIr7eEP3jqIoiuImKugriqL4EBX0FUVRfIhHHuQKIQKA94DKQBAw\nBfgH7am7DdgHDJdS2h3Hx6LNWKsnpcwWQoQCa4BIIBt4Skp5yt0/h+Jaha0nV12nNtoDvDKuyi6p\neFYRxBQdcBK4nJn0NynleLf+EG7iqZZ+H+C8lLIV0BFYgPbEfbxjnw7oCiCE6AD8gDb06rJBwP9J\nKe9HS170ohvLrrhPYesJQogIxzlZbiy34n6FrSvVgV2O7KBtSmrAB88F/fVoSydeLkMOcLeUcqtj\n37fAg47vrWgpWFMvnyylfAeY5tisfPVrSolSqHriaL0tBsYBJncUWPGYQtUVoBFQXgjxkxDiayFE\nLUoojwR9KaVRSpkhhAhH+2W9+q+yZAClHMduzGvGmpTSJoTYhDal+nM3FFtxsyKoJ5OAr6WUiY5t\ndyUeU9ysCOrKKWCalPIBtAZlic2G6rEHuY5Vc35Cmym3ltxJhsKBize6hpSyLdAK+MQlhVQ8rpD1\npA8w0DHdvSzwvcsKqnhcIevKTuALACnldqCcq8rpaR4J+kKI29D61F6UUsY7dicIIe53fN8J2JrX\nuY7zxwkhnnBsGtFyaiglTGHriZSy5uU+WuAM0N6V5VU8p7B1Ba1r6FnHtRoCJ1xUVI/zVBqG8Wgf\ntSYKIS73w40G5gkhAoG/0bLQXe3qqcPL0VaeGQDogf4uLq/iGYWtJwXZr5QMha0rbwIfCCE6ozUi\n+7m2uJ6j0jAoiqL4EDU5S1EUxYeooK8oiuJDVNBXFEXxISroK4qi+BAV9BVFUXyICvqKoig+RAV9\nRVEUH6KCvqIoig/5f+E3gtXuOurwAAAAAElFTkSuQmCC\n",
      "text/plain": "<matplotlib.figure.Figure at 0x11288c5d0>"
     },
     "metadata": {}
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast_df.to_csv('../../outputs/forecasts/MARS_Forecast_{}_{}_{}.csv'.format(station_name, \\\n                                                                          forecast_df.index[-1].strftime(\"%Y%m\"), \\\n                                                                          tscale))",
   "execution_count": 69,
   "outputs": []
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "forecast_df",
   "execution_count": 70,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/html": "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Tarawa forecast</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>2013-05-31</th>\n      <td> 2.235417</td>\n    </tr>\n    <tr>\n      <th>2013-06-30</th>\n      <td> 0.098058</td>\n    </tr>\n    <tr>\n      <th>2013-07-31</th>\n      <td> 2.046161</td>\n    </tr>\n    <tr>\n      <th>2013-08-31</th>\n      <td> 2.124065</td>\n    </tr>\n    <tr>\n      <th>2013-09-30</th>\n      <td> 2.774192</td>\n    </tr>\n    <tr>\n      <th>2013-10-31</th>\n      <td> 3.102988</td>\n    </tr>\n    <tr>\n      <th>2013-11-30</th>\n      <td> 5.714778</td>\n    </tr>\n    <tr>\n      <th>2013-12-31</th>\n      <td> 1.983968</td>\n    </tr>\n    <tr>\n      <th>2014-01-31</th>\n      <td> 3.266503</td>\n    </tr>\n    <tr>\n      <th>2014-02-28</th>\n      <td> 3.373323</td>\n    </tr>\n    <tr>\n      <th>2014-03-31</th>\n      <td> 3.670890</td>\n    </tr>\n    <tr>\n      <th>2014-04-30</th>\n      <td> 4.256796</td>\n    </tr>\n    <tr>\n      <th>2014-05-31</th>\n      <td> 3.865849</td>\n    </tr>\n    <tr>\n      <th>2014-06-30</th>\n      <td> 4.965997</td>\n    </tr>\n    <tr>\n      <th>2014-07-31</th>\n      <td> 6.647030</td>\n    </tr>\n    <tr>\n      <th>2014-08-31</th>\n      <td> 5.107783</td>\n    </tr>\n    <tr>\n      <th>2014-09-30</th>\n      <td> 2.017040</td>\n    </tr>\n    <tr>\n      <th>2014-10-31</th>\n      <td> 1.009004</td>\n    </tr>\n    <tr>\n      <th>2014-11-30</th>\n      <td> 2.113653</td>\n    </tr>\n    <tr>\n      <th>2014-12-31</th>\n      <td> 3.762782</td>\n    </tr>\n    <tr>\n      <th>2015-01-31</th>\n      <td> 4.349747</td>\n    </tr>\n    <tr>\n      <th>2015-02-28</th>\n      <td> 6.083026</td>\n    </tr>\n    <tr>\n      <th>2015-03-31</th>\n      <td> 5.254073</td>\n    </tr>\n    <tr>\n      <th>2015-04-30</th>\n      <td> 2.778925</td>\n    </tr>\n    <tr>\n      <th>2015-05-31</th>\n      <td>-0.538166</td>\n    </tr>\n    <tr>\n      <th>2015-06-30</th>\n      <td> 1.283390</td>\n    </tr>\n  </tbody>\n</table>\n</div>",
      "text/plain": "            Tarawa forecast\n2013-05-31         2.235417\n2013-06-30         0.098058\n2013-07-31         2.046161\n2013-08-31         2.124065\n2013-09-30         2.774192\n2013-10-31         3.102988\n2013-11-30         5.714778\n2013-12-31         1.983968\n2014-01-31         3.266503\n2014-02-28         3.373323\n2014-03-31         3.670890\n2014-04-30         4.256796\n2014-05-31         3.865849\n2014-06-30         4.965997\n2014-07-31         6.647030\n2014-08-31         5.107783\n2014-09-30         2.017040\n2014-10-31         1.009004\n2014-11-30         2.113653\n2014-12-31         3.762782\n2015-01-31         4.349747\n2015-02-28         6.083026\n2015-03-31         5.254073\n2015-04-30         2.778925\n2015-05-31        -0.538166\n2015-06-30         1.283390"
     },
     "metadata": {},
     "execution_count": 70
    }
   ]
  },
  {
   "metadata": {
    "collapsed": false,
    "trusted": true
   },
   "cell_type": "code",
   "source": "",
   "execution_count": null,
   "outputs": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "name": "python2",
   "display_name": "Python 2",
   "language": "python"
  },
  "language_info": {
   "mimetype": "text/x-python",
   "nbconvert_exporter": "python",
   "name": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.9",
   "file_extension": ".py",
   "codemirror_mode": {
    "version": 2,
    "name": "ipython"
   }
  }
 },
 "nbformat": 3,
 "nbformat_minor": 0
}