behavioral_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Behavioral analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set some default display parameters for plots and tables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sns.set(style=\"ticks\", context=\"paper\",\n",
    "        rc={\"xtick.major.size\": 3, \"ytick.major.size\": 3,\n",
    "            \"xtick.major.width\": 1, \"ytick.major.width\": 1,\n",
    "            \"axes.linewidth\": 1, \"lines.linewidth\": 1})\n",
    "mpl.rcParams[\"savefig.dpi\"] = 150\n",
    "pd.set_option(\"display.precision\", 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load in the cached complete data file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"data/punch_data.csv\").query(\"stim\")\n",
    "data_corr = data.query(\"correct\").copy()\n",
    "data_resp = data.query(\"answered\").copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remove between-subjects differences in median RT so that error bars on plots represent within-subject error on the effects of interest."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def remove_intercept_variance(x, group_median):\n",
    "    return x - x.median() + group_median\n",
    "data_corr.loc[:, \"rt_within\"] = data_corr.groupby(\"subj\").rt.transform(remove_intercept_variance,\n",
    "                                                                       data_corr.rt.median())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def savefig(fig, name):\n",
    "    fig.savefig(\"figures/{}.pdf\".format(name), dpi=300)\n",
    "    fig.savefig(\"figures/{}.png\".format(name), dpi=120)\n",
    "    fig.savefig(\"tiffs/{}.tiff\".format(name), dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/mwaskom/anaconda/envs/punch/lib/python2.7/site-packages/rpy2/robjects/functions.py:106: UserWarning: Loading required package: Matrix\n",
      "\n",
      "  res = super(Function, self).__call__(*new_args, **new_kwargs)\n"
     ]
    }
   ],
   "source": [
    "%load_ext rpy2.ipython\n",
    "%R -n library(lme4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%R\n",
    "lr.test = function(m1, m2, name){\n",
    "    print(summary(m1))\n",
    "    out = anova(m1, m2)\n",
    "    chi2 = out$Chisq[2]\n",
    "    dof = out$\"Chi Df\"[2]\n",
    "    p = out$\"Pr(>Chisq)\"[2]\n",
    "    test_str = \"\\nLikelihood ratio test for %s:\\n  Chisq(%d) = %.2f; p = %.3g\"\n",
    "    writeLines(sprintf(test_str, name, dof, chi2, p))\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%R -i data\n",
    "data$log.cpe = log(data$pe_context)\n",
    "data.corr = data[data$correct,]\n",
    "data.corr$log.cpe = log(data.corr$pe_context)\n",
    "data.resp = data[data$answered,]\n",
    "data.resp$log.cpe = log(data.resp$pe_context)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Basic statistics about the behavioral data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Perceptual coherence of experimental stimuli"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Summarize the average coherence (based on the actual signal that was acheived during stimulus presentation) for each context."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>color_signal</th>\n",
       "      <td>15</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>motion_signal</th>\n",
       "      <td>15</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.48</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               count  mean   std  min   25%   50%   75%   max\n",
       "color_signal      15  0.18  0.06  0.1  0.15  0.16  0.18  0.36\n",
       "motion_signal     15  0.25  0.10  0.1  0.17  0.23  0.32  0.48"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.groupby(\"subj\")[[\"color_signal\", \"motion_signal\"]].mean().describe().T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Response accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>color</th>\n",
       "      <td>15</td>\n",
       "      <td>0.89</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.83</td>\n",
       "      <td>0.85</td>\n",
       "      <td>0.91</td>\n",
       "      <td>0.93</td>\n",
       "      <td>0.97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>motion</th>\n",
       "      <td>15</td>\n",
       "      <td>0.87</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.76</td>\n",
       "      <td>0.80</td>\n",
       "      <td>0.86</td>\n",
       "      <td>0.93</td>\n",
       "      <td>0.98</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        count  mean   std   min   25%   50%   75%   max\n",
       "color      15  0.89  0.05  0.83  0.85  0.91  0.93  0.97\n",
       "motion     15  0.87  0.08  0.76  0.80  0.86  0.93  0.98"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.pivot_table(data_resp, \"correct\", \"subj\", \"context\").describe().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Generalized linear mixed model fit by maximum likelihood (Laplace\n",
       "  Approximation) [glmerMod]\n",
       " Family: binomial  ( logit )\n",
       "Formula: correct ~ context + (context | subj)\n",
       "   Data: data.resp\n",
       "\n",
       "     AIC      BIC   logLik deviance df.resid \n",
       "  6180.4   6215.8  -3085.2   6170.4     8751 \n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-5.6914  0.2153  0.3084  0.4314  0.5610 \n",
       "\n",
       "Random effects:\n",
       " Groups Name          Variance Std.Dev. Corr \n",
       " subj   (Intercept)   0.2370   0.4869        \n",
       "        contextmotion 0.6681   0.8174   -0.38\n",
       "Number of obs: 8756, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "              Estimate Std. Error z value Pr(>|z|)    \n",
       "(Intercept)     2.2321     0.1367  16.333   <2e-16 ***\n",
       "contextmotion  -0.1743     0.2242  -0.778    0.437    \n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "contextmotn -0.424\n",
       "\n",
       "Likelihood ratio test for main effect of context:\n",
       "  Chisq(1) = 0.58; p = 0.445\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = glmer(correct ~ context + (context | subj), data.resp, family=binomial)\n",
    "m.null = glmer(correct ~ 1 + (context | subj), data.resp, family=binomial)\n",
    "lr.test(m, m.null, \"main effect of context\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Proportion missed trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.027111111111111086"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1 - data.answered.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Effect of context on reaction time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>color</th>\n",
       "      <td>15</td>\n",
       "      <td>1.27</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.93</td>\n",
       "      <td>1.19</td>\n",
       "      <td>1.27</td>\n",
       "      <td>1.31</td>\n",
       "      <td>1.55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>motion</th>\n",
       "      <td>15</td>\n",
       "      <td>1.20</td>\n",
       "      <td>0.21</td>\n",
       "      <td>0.74</td>\n",
       "      <td>1.10</td>\n",
       "      <td>1.17</td>\n",
       "      <td>1.36</td>\n",
       "      <td>1.57</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        count  mean   std   min   25%   50%   75%   max\n",
       "color      15  1.27  0.15  0.93  1.19  1.27  1.31  1.55\n",
       "motion     15  1.20  0.21  0.74  1.10  1.17  1.36  1.57"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.pivot_table(data_corr, \"rt\", \"subj\", \"context\").describe().T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/mwaskom/anaconda/envs/punch/lib/python2.7/site-packages/rpy2/robjects/functions.py:106: UserWarning: refitting model(s) with ML (instead of REML)\n",
      "\n",
      "  res = super(Function, self).__call__(*new_args, **new_kwargs)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ context + (context | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2149.2\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.6982 -0.7213 -0.0747  0.6567  3.5396 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name          Variance Std.Dev. Corr\n",
       " subj     (Intercept)   0.023685 0.15390      \n",
       "          contextmotion 0.008625 0.09287  0.45\n",
       " Residual               0.076159 0.27597      \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "              Estimate Std. Error t value\n",
       "(Intercept)    1.27249    0.03998   31.83\n",
       "contextmotion -0.07203    0.02479   -2.91\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "contextmotn 0.411 \n",
       "\n",
       "Likelihood ratio test for main effect of context:\n",
       "  Chisq(1) = 7.07; p = 0.00783\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ context + (context | subj), data.corr)\n",
    "m.null = lmer(rt ~ 1 + (context | subj), data.corr)\n",
    "lr.test(m, m.null, \"main effect of context\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Analyses of context prediction error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test to support using log(CPE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Data: data.corr\n",
       "Models:\n",
       "m.lin: rt ~ pe_context + (pe_context | subj)\n",
       "m.log: rt ~ log.cpe + (log.cpe | subj)\n",
       "      Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    \n",
       "m.lin  6 2340.7 2382.4 -1164.3   2328.7                             \n",
       "m.log  6 2323.8 2365.5 -1155.9   2311.8 16.928      0  < 2.2e-16 ***\n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m.lin = lmer(rt ~ pe_context + (pe_context | subj), data.corr)\n",
    "m.log = lmer(rt ~ log.cpe + (log.cpe | subj), data.corr)\n",
    "print(anova(m.lin, m.log))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2323.2\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7391 -0.7389 -0.0813  0.6501  3.7294 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name        Variance Std.Dev. Corr \n",
       " subj     (Intercept) 0.029650 0.17219       \n",
       "          log.cpe     0.001389 0.03726  -0.10\n",
       " Residual             0.078099 0.27946       \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error t value\n",
       "(Intercept)  1.32021    0.04522  29.197\n",
       "log.cpe      0.08338    0.01230   6.779\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.028 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE):\n",
       "  Chisq(1) = 21.89; p = 2.88e-06\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe + (log.cpe | subj), data.corr)\n",
    "m.null = lmer(rt ~ 1 + (log.cpe | subj), data.corr)\n",
    "lr.test(m, m.null, \"main effect of log(CPE)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe + context_freq + (log.cpe + context_freq | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2327.4\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7477 -0.7381 -0.0862  0.6513  3.7316 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name         Variance Std.Dev. Corr       \n",
       " subj     (Intercept)  0.029063 0.17048             \n",
       "          log.cpe      0.000657 0.02563   0.02      \n",
       "          context_freq 0.001133 0.03366   0.19 -0.98\n",
       " Residual              0.078089 0.27944             \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "              Estimate Std. Error t value\n",
       "(Intercept)   1.320468   0.045205  29.210\n",
       "log.cpe       0.082480   0.013498   6.111\n",
       "context_freq -0.002014   0.026630  -0.076\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) log.cp\n",
       "log.cpe      0.014       \n",
       "context_frq -0.067  0.469\n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE) controlling for context frequency:\n",
       "  Chisq(1) = 20.72; p = 5.31e-06\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe + context_freq + (log.cpe + context_freq | subj), data.corr)\n",
    "m.null = lmer(rt ~ context_freq + (log.cpe + context_freq | subj), data.corr)\n",
    "lr.test(m, m.null, \"main effect of log(CPE) controlling for context frequency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of congruency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ congruent + log.cpe + (congruent + log.cpe | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2319.1\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7372 -0.7319 -0.0811  0.6591  3.6819 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name          Variance  Std.Dev. Corr       \n",
       " subj     (Intercept)   0.0283522 0.1684              \n",
       "          congruentTRUE 0.0004795 0.0219    0.42      \n",
       "          log.cpe       0.0013177 0.0363   -0.04 -0.71\n",
       " Residual               0.0779400 0.2792              \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "               Estimate Std. Error t value\n",
       "(Intercept)    1.326064   0.044326  29.916\n",
       "congruentTRUE -0.014440   0.008527  -1.693\n",
       "log.cpe        0.082284   0.012108   6.796\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) cnTRUE\n",
       "congrntTRUE  0.226       \n",
       "log.cpe      0.073 -0.338\n",
       "\n",
       "Likelihood ratio test for main effect of stimulus congruency:\n",
       "  Chisq(1) = 2.77; p = 0.0958\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ congruent + log.cpe + (congruent + log.cpe | subj), data.corr)\n",
    "m.null = lmer(rt ~ log.cpe + (congruent + log.cpe | subj), data.corr)\n",
    "lr.test(m, m.null, \"main effect of stimulus congruency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interaction between log(CPE) and congruency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ congruent * log.cpe + (congruent * log.cpe | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2323.3\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7171 -0.7301 -0.0803  0.6573  3.6747 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name                  Variance  Std.Dev. Corr             \n",
       " subj     (Intercept)           0.0261345 0.16166                   \n",
       "          congruentTRUE         0.0020521 0.04530   0.49            \n",
       "          log.cpe               0.0019659 0.04434  -0.15 -0.80      \n",
       "          congruentTRUE:log.cpe 0.0005535 0.02353   0.59  0.99 -0.78\n",
       " Residual                       0.0779189 0.27914                   \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                       Estimate Std. Error t value\n",
       "(Intercept)            1.321662   0.043210  30.587\n",
       "congruentTRUE         -0.004866   0.020247  -0.240\n",
       "log.cpe                0.077730   0.015597   4.984\n",
       "congruentTRUE:log.cpe  0.009629   0.016491   0.584\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) cnTRUE log.cp\n",
       "congrntTRUE  0.129              \n",
       "log.cpe      0.054 -0.684       \n",
       "cngrnTRUE:.  0.059  0.911 -0.648\n",
       "\n",
       "Likelihood ratio test for interaction between log(CPE) stimulus congruency:\n",
       "  Chisq(1) = 0.34; p = 0.557\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ congruent * log.cpe + (congruent * log.cpe | subj), data.corr)\n",
    "m.null = lmer(rt ~ congruent + log.cpe + (congruent * log.cpe | subj), data.corr)\n",
    "lr.test(m, m.null, \"interaction between log(CPE) stimulus congruency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Effect of presenting context cue early"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ early + (early | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 1692.3\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-4.3710 -0.7239 -0.0881  0.6373  3.6658 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name        Variance Std.Dev. Corr \n",
       " subj     (Intercept) 0.030924 0.17585       \n",
       "          earlyTRUE   0.006932 0.08326  -0.03\n",
       " Residual             0.071727 0.26782       \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error t value\n",
       "(Intercept)  1.31694    0.04561  28.872\n",
       "earlyTRUE   -0.15738    0.02235  -7.042\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "          (Intr)\n",
       "earlyTRUE -0.044\n",
       "\n",
       "Likelihood ratio test for Main effect of early cue:\n",
       "  Chisq(1) = 22.71; p = 1.89e-06\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ early + (early | subj), data.corr)\n",
    "m.null = lmer(rt ~ 1 + (early | subj), data.corr)\n",
    "lr.test(m, m.null, \"Main effect of early cue\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interaction of log(CPE) effect with cue timing manipulation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe * early + (log.cpe * early | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 1518.6\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-4.1219 -0.7222 -0.0889  0.6326  3.6499 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name              Variance  Std.Dev. Corr             \n",
       " subj     (Intercept)       0.0226173 0.15039                   \n",
       "          log.cpe           0.0016605 0.04075  -0.55            \n",
       "          earlyTRUE         0.0108059 0.10395   0.24  0.02      \n",
       "          log.cpe:earlyTRUE 0.0009251 0.03042   0.75 -0.56  0.76\n",
       " Residual                   0.0699290 0.26444                   \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                  Estimate Std. Error t value\n",
       "(Intercept)        1.43695    0.04049   35.49\n",
       "log.cpe            0.11913    0.01490    8.00\n",
       "earlyTRUE         -0.21244    0.03106   -6.84\n",
       "log.cpe:earlyTRUE -0.05282    0.01650   -3.20\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) log.cp erTRUE\n",
       "log.cpe     -0.187              \n",
       "earlyTRUE    0.098 -0.229       \n",
       "lg.cp:rTRUE  0.173 -0.639  0.721\n",
       "\n",
       "Likelihood ratio test for interaction between log(CPE) and early cue:\n",
       "  Chisq(1) = 8.77; p = 0.00307\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe * early + (log.cpe * early | subj), data.corr)\n",
    "m.null = lmer(rt ~ log.cpe + early + (log.cpe * early | subj), data.corr)\n",
    "lr.test(m, m.null, \"interaction between log(CPE) and early cue\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE) on early trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.corr\n",
       " Subset: early\n",
       "\n",
       "REML criterion at convergence: 959.7\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-2.6291 -0.7064 -0.1222  0.6053  3.5183 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name        Variance  Std.Dev. Corr\n",
       " subj     (Intercept) 0.0405214 0.20130      \n",
       "          log.cpe     0.0007233 0.02689  0.38\n",
       " Residual             0.0732255 0.27060      \n",
       "Number of obs: 3898, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error t value\n",
       "(Intercept)  1.22452    0.05310  23.060\n",
       "log.cpe      0.06649    0.01236   5.382\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.364 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE) on early cue trials:\n",
       "  Chisq(1) = 16.86; p = 4.03e-05\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe + (log.cpe | subj), data.corr, subset=early)\n",
    "m.null = lmer(rt ~ 1 + (log.cpe | subj), data.corr, subset=early)\n",
    "lr.test(m, m.null, \"main effect of log(CPE) on early cue trials\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE) on context repeat trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.corr\n",
       " Subset: !context_switch\n",
       "\n",
       "REML criterion at convergence: 1188.3\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7846 -0.7363 -0.0927  0.6446  3.5860 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name        Variance  Std.Dev. Corr \n",
       " subj     (Intercept) 0.0300518 0.17335       \n",
       "          log.cpe     0.0005138 0.02267  -0.46\n",
       " Residual             0.0747662 0.27343       \n",
       "Number of obs: 4507, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error t value\n",
       "(Intercept)  1.28630    0.04658  27.618\n",
       "log.cpe      0.06905    0.01202   5.746\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.015 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE) on context repeat trials:\n",
       "  Chisq(1) = 18.21; p = 1.98e-05\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe + (log.cpe | subj), data.corr, subset=!context_switch)\n",
    "m.null = lmer(rt ~ 1 + (log.cpe | subj), data.corr, subset=!context_switch)\n",
    "lr.test(m, m.null, \"main effect of log(CPE) on context repeat trials\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE) on context switch trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.corr\n",
       " Subset: context_switch\n",
       "\n",
       "REML criterion at convergence: 1117.3\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-2.7475 -0.7378 -0.0615  0.6737  3.5047 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name        Variance Std.Dev. Corr\n",
       " subj     (Intercept) 0.036993 0.19233      \n",
       "          log.cpe     0.002425 0.04924  0.64\n",
       " Residual             0.080976 0.28456      \n",
       "Number of obs: 3199, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error t value\n",
       "(Intercept)  1.29422    0.05127  25.246\n",
       "log.cpe      0.01655    0.02010   0.823\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.568 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE) on context switch trials:\n",
       "  Chisq(1) = 0.71; p = 0.399\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe + (log.cpe | subj), data.corr, subset=context_switch)\n",
    "m.null = lmer(rt ~ 1 + (log.cpe | subj), data.corr, subset=context_switch)\n",
    "lr.test(m, m.null, \"main effect of log(CPE) on context switch trials\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interaction with context switch events"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ log.cpe * context_switch + (log.cpe * context_switch | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2259.6\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7349 -0.7371 -0.0800  0.6576  3.6092 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name                       Variance Std.Dev. Corr             \n",
       " subj     (Intercept)                0.030488 0.17461                   \n",
       "          log.cpe                    0.000699 0.02644  -0.33            \n",
       "          context_switchTRUE         0.003813 0.06175   0.14 -0.06      \n",
       "          log.cpe:context_switchTRUE 0.003751 0.06125   0.44 -0.36  0.93\n",
       " Residual                            0.077259 0.27796                   \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                            Estimate Std. Error t value\n",
       "(Intercept)                 1.286064   0.046944  27.396\n",
       "log.cpe                     0.068924   0.012665   5.442\n",
       "context_switchTRUE          0.007775   0.024081   0.323\n",
       "log.cpe:context_switchTRUE -0.053132   0.024398  -2.178\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) log.cp c_TRUE\n",
       "log.cpe      0.052              \n",
       "cntxt_sTRUE -0.065 -0.456       \n",
       "lg.cp:_TRUE  0.156 -0.493  0.919\n",
       "\n",
       "Likelihood ratio test for interaction between log(CPE) and context switches:\n",
       "  Chisq(1) = 4.36; p = 0.0367\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ log.cpe * context_switch + (log.cpe * context_switch | subj), data.corr)\n",
    "m.null = lmer(rt ~ log.cpe + context_switch + (log.cpe * context_switch | subj), data.corr)\n",
    "lr.test(m, m.null, \"interaction between log(CPE) and context switches\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main behavioral figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhYAAAKuCAYAAAD93lXLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAXEQAAFxEByibzPwAAIABJREFUeJzsvXeYlPW5//+aXrf3CgvSm5SliYAKliM2jC2aoEejJzkG\nc3kS45WcaIzJdZKcnOPPBL+px5w0TYwKih6jKEV6XWCBXWBp2/vOTi9P+f0x7MCws8sCM9v4vK6L\nK5mnzOcz487zvJ/7c9/vW6OqqopAIBAIBAJBHNAO9AQEAoFAIBAMH4SwEAgEAoFAEDeEsBAIBAKB\nQBA3hLAQCAQCgUAQN4SwEAgEAoFAEDeEsBAIBAKBQBA3hLAQCAQCgUAQN4SwEAgEAoFAEDeEsBAI\nBAKBQBA3hLAQCAQCgUAQN4SwEAgEAoFAEDeEsBAIBAKBQBA3hLAQCAQCgUAQN/QDPYGrAVVV8fl8\nAFgsFjQazQDPSCC4OhG/RYEg8SRMWHzwwQd885vfBOCBBx7gpZdeStRQgx6fz8f06dMBKCsrw2q1\nDvCMBILBQW1tLUuWLOlxf0pKCnl5eSxatIgVK1aQnp5+ReOJ36JAkHgSJizeeeedyP9fu3Ytzz33\nHDabLVHDCQSCIc7YsWNJSkqKvJZlmc7OTqqqqqisrOStt97iD3/4A2PHjh3AWQoEgouREGFRX1/P\njh07SElJoaSkhP3797N27VoefPDBRAwnEAiGOBqNhu9973uUlpZ22+dwOHj++efZuHEjK1eu5KOP\nPhJLGALBICYhyZvvvvsuqqoyY8YMbrzxRgD+9re/JWIogUAwzElNTeXHP/4xBoOB06dPs2XLloGe\nkkAg6IW4CwtVVVm9ejUAixYt4rbbbgOgsrKSAwcOxHs4gUBwFZCamsqYMWMAqKqqGuDZCASC3oi7\nsNixYwd1dXXodDpuuukmioqKmDp1Kqqq8uabb8Z7OIFAcJUgSRKAyNUSCAY5cRcWXUmbs2fPJisr\nC4Bly5YB8NFHH+F0OuM9pEAgGOZUV1dz/PhxdDod119//UBPRyAQ9EJchYXL5WLdunUA3HnnnZHt\ny5YtQ6fTEQgEePfdd+M5pEAgGCaoqhr1WpZlOjo6WL9+PV/5ylcAePLJJ8nLyxuI6QkEgj4S16qQ\nDz/8kEAggMVi4eabb45sT09PZ/78+WzevJm//e1vPProo/EcViAQDHFUVeXLX/5yr8c8+eSTPPPM\nM/00I4FAcLnEVVh0LYPccMMN3dZB77zzTjZv3sypU6fYsWMHc+fOjefQAoFgiBPLx8Lj8VBdXU0g\nEOD111/H6/Xyne98B61WdCMQCAYrcRMWx48fp7y8HI1Gwx133NFt/5IlS7BYLPh8Pv76178KYSEQ\nCCL05mMRCoVYvXo1P/jBD/jzn/+MLMu8+OKLAzBLgUDQF+ImLLqiFaqq8tWvfrXXYz/99FPa2trI\nyMiI1/ACgWCYYjAYuP/++2lubmbVqlW89dZbPPXUU+Tm5g701AQCQQziEk8MhUK8//77QNjbPzc3\nN+a/nJwcIFw29vbbb8djaIFAcJXQ1VNEURSOHDkywLMRCIYWoZBMKCT3y1hxiVhs2rSJ9vZ2NBoN\nr7/+OpMmTerx2DvuuIPjx4/z1ltv8eSTTwprXoFAcElcWD0iEAhio6oqAV8If0BCllVMZh0Ggy7h\n48YlYtEVfRg3blyvogKI9Aupq6vj888/j8fwAoHgKmDTpk0AaLXai15nBIKrGVlScDn9dLR58fkl\nQINO138Jz1c8UktLC5s3bwZg+fLlFz3+rrvuwmKxAAgnToFAEKGnSISqqnzyySf86le/AuCWW26J\nLKsKBIIwqqri9wZxdHhxOHxIkoJWpx2QVYErXgpZs2YNsixjNBqjTLF6wm63c8cdd/DWW2+xefNm\nGhsbRRKWQHCVo6oqL7/8Mna7PWq7JEnU1dXR3t4OwOTJk/n+978/ADMUCAYnsqTg9QYJBmW0Wg0a\njQadbmBTDK5YWKxevRqNRsMNN9xAampqn8754he/yFtvvYWiKLz11lusXLnySqchEAiGMBqNJmZz\nMZPJRHp6OjfccAO33HILd955p/CwEFz1qKpKICDh94WQZRWttn+XOi7GFQuL//u//7vkc8aPH09l\nZeWVDi0QCIY4hYWFQ/Ja8Obv95BflMKiJWMGeipDFrcrwJYNJ7hu8SiSks0DPZ0hgSKfjU4EZNCE\n840GOjoRi7g6bwoEAsHVwM3LJmAwJj67fjhTX+Og5nSHqAzsA6GghNcbCudNaDVoB1F0IhZCWAgE\nAsElkpElWrfHC1E+HBtVVQn4Q/h8EoqiotNpB9VyR28IYSEQCASXyJu/30N+YQoz5hTx1//dy023\njePksVZqqx1odRpKRmcwb2EJ+vM8A8rL6qg81ITLFcBmMzJhSi5TZxRE9tdWO9i3q4aOVg8arYbC\nEanMnj8Se5IJgGNHmvh8/Qnu/MIUtn9+irYWDxargUnT8qLex+sJsv3zU9TVOECFUWMzMZn0VB1r\n4aFHZ0XmP3J0Bu2tHpoaXVwzLovcvCQ2fVrFQ4/Niox5/mddtDS87KOqKgf21nH0cBMedwB7kolJ\n0/KZNO1c19kP3iknOdVCcoqZIwcb8PskMrNtzFtYQlZOEseONLHp03BOzV//dy9jJ2aLZaWzdHlP\neP0hUAfvckdvCGEhEAgEl4rm7L+zbNlwgnETc7h52Xiam9zs2X4Gs8VA6fwRAOzccory/Q1MnZ5P\nQXEqLU1udm09jaKoXDurkOMVzWxcd5xrxmUxfVYhfl+IvTuref/vB7nnwWuxWA3hgVSVzz46ypQZ\n+ZTOH8HRw03s3HKa9AwrhSPSkCWFD989hCQpzF9YgsGoY/+eOtpaPFhthqiPcORgA1NnFHDtrEL0\nBh2dHd4+f9ZjR5qZXlpITl4y9XWdbP/8JIGAxIzZRZHjTlW1kpZu5brFo1FVlZ1bTrPuw0oeemwW\nxSXpTJ9dSNmuWpbePp70TBEBUhQVvzeEPxBCo9Gg1WijvvehhBAWAoEgoSiqEr5IDmOKR6YxZ8FI\nAPKLUqmrdlB9qp3S+SMIBCTK9zcweVoes68LH1NQlIrPG6Sx3hm+6W49TdGING64ZWzkPXPyk3n7\nz/s4WFbHnLPnqSrMmFPEuIlhH4+cvCROnWij+nQHhSPSqDragqPDxz0PTiMzO1y6m1+Yyl//sLfb\nnO3JpojwAXoWFufh6PBx9HATpfNHMG1mYfizFKei0cD+PbVMmpqHyRy+raiKym13TYrkooSCMhvX\nHaetxUNmtp3kswmbmdn2qAjJ1YYsK3g9Z8tFNZphUfU0ZD9BY2Mjs2bNYvfu3b0eFwgE+O///m9u\nuOEGrr32Wh588EG2bNnS7biFCxcyfvz4bv8cDkeiPoJAMGxRVRWn301tZyN1nY0DPZ2Ek52XFPXa\najciSQoAzQ0uVEVl5OjopovzFo7i1jsn4ujw4fOGGD02M2p/coqZ7NwkGmo7o7bn5J4bS6fTYrEY\nkM72gKiv7SQ5xRwRFQAGo47ikjQuzGTIuIwoQX2tA1WF4pJ0FEWN/BtRko4sKTTWn5trWro1KsHV\najcCRL6Xq51QSKLT4cPRETaz0um0aLRDNERxAUMyYtHQ0MDjjz+Ox+O56LHf/e532bhxI//2b//G\nyJEjWb16NU899RR/+MMfmDUrvN7Y3t5Oc3Mz3/72t5k5c2bU+UlJSbHeViAQxCAoh+jwduIJ+dBq\nNOi0V0flhF4f/Tk1Gk0kKTHgDwGcW864gIBfCu+3Gbvts1iNtLVGX+f0MXo9dIkGny8UcxyL5YJt\nGvpe1XKeIuma69t/Lut2mEYDHk8w8lpn6P6dgEjWDAYlvO4giqyg1euGTELmpTCkhIWqqqxevZqf\n/OQnkde9UVtbywcffMALL7zAQw89BMDcuXPZt28fb775ZkRYdNXRL1myhKKioh7fTyAQdEdRFVwB\nN06/h6AcwqDTY9ANqUtLQjGawt+F3xciJdUS2e52BXA6fJjPCgHveTflLryeIGZz379Lm91IQ62z\n23afL3Txk8/e+BUl+rp6fkdMozE8l2XLJ3cTJipgt1+9Sxq9oaoqPl+IgD+EooQjTVr98BXdQ0oq\nVVZW8v3vf5/ly5fz05/+9KLHZ2dn884770RZjYftTnUEg+d+xBUVFdhsNiEqBIJLwB8K0OBs5nRH\nHR0+J2jAqDcIX4KzdN2es3OT0Go1nDnZHrX/4L461n98jLR0K1abgRPHWqL2Ozv9NDe6yM1P7n2g\n877u/IIUXE4/bS3nohySJIf9Ii4y3y6h4HEHItsc7d5IlAIgryA8F58vRGa2PfLP75fYu6OaQKAP\nAqZr2lfB34miqHjcATrawt+jRjN0SkavhCH1WJGfn8+6devIyclh586dFz3eaDRGuiCqqkpjYyOv\nv/46NTU1vPDCC5HjKioqSElJYeXKlWzbtg1Zllm8eDHf+c53yMrKStjnEQiGGrKi4PS7cQXcSKqE\nQWe4OqMTlxDNN1sMTJ6eT3lZPTqdltyCZJoaXBwpb2Tu9SPRaDSUzh/JpnXHWf/xUcaMyz5bFVKD\n2aJnyvT8Ps9l9Lgs9u+t5ZMPKyidW4zBpKe8rB6/LxTtbhlj/gWFKej1WnZsPs2sucUEgxJ7d9aE\nkzHPHp+eaeOa8VlsXl+F2xUgM8uGw+Fjz7ZqklLMUREZLhJR7orknKpqpWhEGqnp1t4/5xBCkRU8\nniChoIxmEBlaKUr/5LcMqStCSkoKKSkpl3Xub37zG1555RUA7r//fubNmxfZd/ToUZqbm5k8eTIr\nVqzgxIkT/PznP+eRRx5hzZo1kW6sfcXrjc6u9vl8lzVngWCw4A8F6PB14pMCkdwJA7FzBgYTCfst\nXuRh+4IKTeZcNxKLxUBFeSMH9taRlGJiweJRjJ8cbsA4dkI2BoOO/XtqWfdhBQajnqIRqZTOH4HF\nei73IuZD/nnbtFoNt901ie2fn2LLhpNodRquGZeF2ayn0+GLeU4XRpOeJbePZ/fWM6z7sIKkZDMz\n5xRxrKI56vhFS8awf08tFeWNuF0BrDYjo8dlMmveiOgoRIzJnr8pvyiFgqJUdm87Q11NJ7feOTH2\nlzmEkGUFjztAKCij1WkHhaBQFAW304PX48WeZCE5OfECTqMO0UyanTt3smLFCv70pz9RWlp60eOP\nHz9OZ2cne/fu5bXXXuPWW2+NLKccOHAAo9HIhAkTIsfv27ePL37xi7z44ouR/Iy+Mm7cuB73lZWV\nYbUOH2UuGL4oqkKnLzo6cVmoKkWpF3nqThBX22+xo82Lo8NLyTXRFSZr/nYAu93EktvHD9DMhjeS\nJON1BwmFFLQ6zaBY5pFCEp0dTgKBYNgGXKvFYjVRNDIn4WMPqYjFlTBmTNjVbdasWUiSxC9+8Que\nffZZcnNzmTZtWrfjZ8yYQVJSEkePHu3vqQoEA0pXZYc35AvnJA2R6IQgnGj52UdHmTDFycjR6aiK\nyonjrbQ2uyNeGIL4EQpJeD2hSLmoTj/wEQqf14+700MoFEKn13arWOoPhrWwqK+vZ+vWrdx1110Y\njefCiRMnhkNuzc3N2O12/vGPfzBt2rSI+IBw+CgUCpGenn7J45aVRZdi+Xw+5s+ff5mfQiBIPKqq\n4g56cPhckcoO/TDInbjafovZuUncdNs4Du6t43hlM6iQmW3jtrsmkVd4ecvIgu6EQhIedxBFGhwl\no4qi4HF58Lh9KLKM3qCPWZbcXwz9K0cv1NbW8r3vfQ+r1crtt98e2b5161aMRiMlJSUYDAZefvll\nli5dys9+9rPIMevXr8fv9zNnzpxLHne4hVcFwxdZkenwdeIO+lBVFb1Oh1E/fKITV+NvseSazG5L\nIYL4EApJeN0hZElGq9cNeMmoLMl0Olz4vf5I11OtduBv6wM/gzjidrupqqqiuLiY9PR0SktLmTt3\nLi+//DJut5uioiI2bNjAG2+8wcqVKyPmV08++SS/+MUvyMzMZOHChRw7doxVq1axZMmSyxIWAsFg\nJxAK0O7rxBcKhEO4w8BGWCBIFKFQOIdCkpVB4UERCARxOdwEAkH0eu2ARidiMaSFxYUJMocPH2bF\nihX8+Mc/5u6770aj0bBq1SpWrVrFr3/9a1paWigpKeGHP/why5cvj5z3ta99jfT0dN544w3efPNN\n0tLSeOihh/j617/e3x9JIEgonqCPdp+DkCyFjaz0Q/oSMKxob/OyZX0Vd943NbLttz/fysw5RcyY\nU5zQsV1OP3/9370sWjqGsROyEzrWUCIUkvF6gudyKAZ4ycPj9uJ2epAlCb1Bj2GQCYouhmxVyFDC\n6/Uyffp0YHhmogsGN6qq4vC5cAZcKKjo+9tmewCrQi5kMP8W9+6oZt+uGr6y8rrItuZGFza7EVuC\nHS2FsIima8mjK0IxkKiqisflxe1yoyoquiuIloiqEIFAcEXIioLD14krGPZy0Gm1Q8tqV0B2ruhV\n1J/IsoLbFRgUEQpFUXA6XPg84eosrU47ZLyyhbAQCIYZkiLR5nHgDfnQakX+xGCipdnNri2naW12\no6qQnWtn1rwR1JzuYN+uGiB6+eO3P9/KjDlFzJxTTH1tJx++e4h/umcSZbtqaG5yY7EYmD67iOKR\naWzdeIK6agdGk56pMwqYfG04SnTsSBObPq3iocdmRbUnf/P3e8gvTGHR0jHd5hkretI1t675AFQd\nbeHA3lo6HX4MBh2FxanMWTASa4yGaoOZLuvtYFAecEEhSRKd7U4C/uDZEtbBudzRGwkXFqdPn6am\npob29nZ0Oh2ZmZkUFBSIvhwCQZwJyiHaPB14Q/5hUy46nAgGJP6x5jAFxaksuX08sqxQtquWj9Yc\n5t6Hp+PxBDl6uIm77p+KzX6e2+YF77PhH8eYNquQGbOL2L+nji0bTpCSYmbUmEwmTcvnyIEGtn9+\nipy8JLJyeol4XGgPeuHuHvZ1bW6sd7Jx3XFmzC4iryAZjyvAzq2nWf+Poyy7d0pfvpIBR1VVvJ4Q\nfn8IrVYzoILC7/PjdLiRQiF0et2gS8i8FBJy5amsrORPf/oTmzZtorW1NeYxRUVFLFq0iPvvv5+x\nY8cmYhoCwVWBPxSg3evALwcw6AzDqlx0ONHR7sPvl5g0LY+cvHAzr9Q0K5WHGtFowHb2Kf9iyx/j\nJuVE+ofoDTree+sgWblJzJwbjiKkZ1o5daKNpgZX78LiIlws+66x3oler2XazILIDdlkNtDS7L7s\nMfsLVVUJ+EJ4fSFg4ARFWNiEEzIVWTkrKIb+A0FcP8Hx48f50Y9+xI4dO8jMzGT27NmMHTuW4uJi\n7HZ7OInM4aChoYH9+/ezdu1a/vznP7Nw4UKeffZZxo8XdrMCQV/xBH20eTuQlK5mYEJQDGbSM61Y\nLAY+XlvBqDGZFBanUlCcyuxLdMTMzjsnFixnW65n59gj28zm8LZgQCKR5BWksGf7Gd7+Sxmjrsmk\ncEQqhSNSKRqZltBxrxS/X8LrDYAK2gFaJlRVFbfTjdvlRaOCVj80lzx6Im7C4r//+7/54x//yK23\n3spf/vIXZs6cedFzJEliy5YtvPvuu9x3332sWLGCb37zm/GakkAwLHEHvLT7HMiqgl6rE4JiiGAw\n6Fj2hSmU7a7h5LFWKsob0eu1jJmQzbyFJX1+H6Ox+w1oIMLmOXlJ3HrnRMrL6ikvq2P/nlosVgPT\nS4uYNC2v3+dzMULBs26ZqhoWFAPQzuPChMyBrjhJFHETFmfOnOG9995jxIgRfR9cr2fx4sUsXryY\nEydO8Oqrr8ZrOgLBsEJVVVwBDw6fE5mwoNBrhs8TztVCapqFG24ei6qqNDe6qaps5kh5I8kp5ouf\nfLmcTZZQlOi1jVBQ7uWc8P+oqhrxC4p1fOGINApHpCFJCvU1Dg7tb2DbppNk59qvaBkmnkiSjMcd\nRA6ddcscgAZhYYdMJ35fAJ12eEUnYhE3ufTqq69ekqi4kNGjR/Pzn/88XtMRCIYFqqrS4XVyxlFH\nu8+BRqvpfx8KQVw4c7KNP/5mJ15vEI1GQ05eEtfdMBqTSY/bFUCjTcwNz3A2wuFxByLbHO1eAv6e\nl0qMkXOCkW2N9c6oY3ZtPc2avx0AQK/XUlySzpwFIwFwu4IMNIqi4uz00dnhR1UZELdMKSTR1tRO\nY10zoWAIvV6XsP/Og4l+zxI5dOgQOp0uqkW5QCCIRlEVOrydOAOeSIdRwdAmJy8ZjUbDug8qmTaz\nAINRx8njrQSDEiWjM2hvC/uNVB1tITs3KW5RjPzCFPR6LTs2n2bW3GKCQYm9O2swmfXQQ4JmcUk6\nOzafZvNnVUydUYDbHWDfrpqISIFwtOLgvjo2rjvONeOyUGSFA3vrMJv15BcNXMOzbZtOcuRgA6PH\nZnJtadGAdBwNBoI4OpxIwXCFh8E49BMyL4WEfeOqqvLzn/+cb33rWwDIssyKFSv4whe+wD333MMj\njzyC2z34s4cFgv5EURXavA7OdNTjCnrQ63TCh2KYYLYY+Ke7J2Iy6dn8WRUfv19Ba7OHJf80nrzC\nFEquySArx86mdccp31cX8z0uJ4pvMulZcvt4VEVl3YcV7NtZw8w5RWTl2HvMM0hJtbD45jG4XAE+\nXnuEwwcaWHjTNZHKFQgLlptuG0dHm5dPP6xk/cfHMBh13H7vZEymgbmRut0BDh9sQFXhxLFWZLl/\njaWDgSDNja20NLaDqqI36Lu1nhgoLlwKSyQJs/T+3e9+x89+9jMWLFjA7373O95//32ee+45brnl\nFsaMGcNvf/tbvvjFL/Ltb387EcMPKgazjbBgcBAdoWB4RSiEpbcgwYRCEh5XEH8gxNq/H4psv+uB\nqRj7IVoQCATpbHcihaRB5T/h80m0t/roaPPT0RHg5ttHD21L7/fee48lS5awatUqAP7xj39gMpn4\nj//4D6xWKx6Ph08++eSqEBYCQU909fFw+J1oNKDXDZ6L0pUiKzJVbdXUORtZMWP5xU8QCC4RRVFx\nu/yEQme7jvZzdO9CQTHQokJRVDo7ArS1+Whv9ePzJrbkuCcSJiyqq6v58pe/DEAwGGTHjh2UlpZG\nnhBGjRpFc3NzooYXCAY1XVUe7T4HoBk2gsIX8nOo6Rj7Gyo40FCB+2yfEiEsBPFEVVV8vhA+78A4\nZp7vkqk36AdUUAQCMu2tPtpa/XS0+1H6efknFgkTFjabDZfLBcCuXbvwer0sWLAgsr+2tpb09PRE\nDS8QDEpUVaXTH45QwPBY8ujwdVJWf4Sy+iNUtFQhKb2UMQoEwPbPT3L4QAOTpuUxb+GoSzo3FJRw\nuwIwAD4QPq8fl8OJJMlnBUX/55KoqorbFaKtxUdbqw+3K9Tr8VabnrQMM+kZln6aYQKFxaRJk/jL\nX/5CYWEhv/rVr9BoNNx8881IksTGjRt54403ooSGQDDccQU8tHk7gKEvKOqdzeyrP8y+ukOc7Kjp\n8TidRsf4rL6bPwmGP1JI5vCBcILl4QMNlM4b0acn/q7Oo11+FP2J3+fH2TFwgkKWFRztAdpafbS1\n+AgGlR6P1Wo1pKabyMi0kJ5hxmzRI4UkTldVU9BP1ToJ+3aef/55HnvsMVauXAnAY489Rn5+Ptu3\nb+fpp58mMzOTr3/964kaXiAYNHRZb3c5ZQ5FVFXljKOOPXWH2Ft3iAZXz8uYNoOFqXnjmZ43kcm5\nY7HqE2j+JBhyyIoa6UOiquHXvd2IIo3CfEG0Om2/ioquCIU8AH08QkGZtlY/rS3h5MveqjrMZh3p\nmRYyMs2kppnR6jQE/AGqjlRx9FAVJypOEQqGmDKzf2weEvYtjR49mvfff5/t27eTl5fHtddeC8C4\nceP41re+xT333COWQgTDGm/IR6unA1mR0ev0Q84pU1EVqtrOsKe2nL31h2jzOno8NsuWzvS8iUzP\nn8TYzJHREZnEFJ4JrgJCIQm3Mwga+tWt8sIIRX+N7fNJtDX7aG3x0ekI9HpscoqRjCwLGZkWrLZw\nWavX7ePgnkMcLa/i9LFqZHlgliUTKr9SU1O57bbboralp6fz+OOPJ3JYgaDfeG3Hn1lzZB13T1zK\nv859BIBAKECLt4OgHMSgMwyp9uWKqnC05SS768rZW3eITr+rx2NHpOYzPX8SM/MnU5iSO2jq9QWJ\n5UryI/qKqqq43QGCAblf8yjCSZmus1UeiV/y6IrGtDT7aG324XH3nC+h02lIyzCTcTYy0WVW5nZ6\n2Lf9CJUHjlF9sha1h8iGwdh/PYXi9q3dc889PPvss1x//fWXdf769et59dVXee+99+I1JYEgofil\nAGuOfIKiqqw5so4vT78HV8CDX/IPqW6jETFRW86eunKcgdjGdRo0jMkYwcyCycwomEyWTUQcrzYu\nNz/iUvB7g3i9ITT9WO0RDARxtDvPq/JInKBQVRWXM0jr2chEbyWhRpOOjEwzmdmW8BLHWTtwt9PN\ngV3HqThwjJpTdT06qFqsZsZMGs24KdcwcmxxIj5OTOL27X3pS1/iW9/6FsXFxdx1110sWbKEnJze\njThaWlr48MMPeffdd2lsbOS5556L13QEgoQjyRLK2TC/oipUt9eRZLEPCUGhqArHW8+wq/YAu2sP\n9igmtBotE7JGR8REqnlwNJYSDAyXmh9xKYSrPYKoqGj7SVDIkoyjrZNAIHjWhyIxgqJLTLQ0+Whp\n9hLw97xEYbXpycyykJFlISnZGIkEup0eKg8ep+LA0V7FhD3Zxrgp1zBu6hiKSwr77bs8n7h9i8uX\nL2fx4sX84he/4Gc/+xk//OEPmTBhAuPGjaOoqIikpKSwGZDDQUNDA/v27ePMmTNYLBbuvfdennrq\nKbKysuI1HYEg4VxoWqvTDx773lioqsqpjlp21uxnV+1BOnydMY/TabRMyhnDrIKpzMifiN1k6+eZ\nCq4mFEXF6fAhSQpanRZNP/QzVxSFzg4nPrcPXYKMrc6JCS8tTT4CgZ7FRFKykcwsC5nZFqy2cw8m\nXo+PowePc2T/UapP1Ha75nSRkpbMuKljGD9lDAUj8ga80Vlc5Vl6ejovvvgizzzzDG+//TYbN27k\ngw8+IBRIr0HVAAAgAElEQVSKXjcyGo2Ulpby8MMPc8cdd5CWlhbPaQgECccVcFPT2TjQ0+gTdc4m\ndlSXsaNmPy2e9pjHhMXEWEoLw2LCZhRW14L+wdHhxWw29NuTtdvpwdXpCncKjrPdd5fHRHOTl5am\n3iMTyalGsrKtZGZbMJvPzSPgD3Ds0AmOlFVy6lg1ihK7tDQlLZnx08YwYdpY8ooGV45TQuI+qamp\nPPHEEzzxxBMEg0Gampro6OhAo9GQlZVFRkYGBsPgDxcLhi6xkirjQUAK0uxpQ1LkfrcPvhTavQ52\n1Oxne3UZNZ0NMY/RarRMzL6G2YXTmFkwSYgJwYCg02n75abo8/pxdnSiKvGvMPG4z4qJRi8+X885\nEylpJrKyLWRmWzGZzs1BkiROVJzmcFklVYdPIEmxBUlyWhITpo5lwrXjyCvKGVRi4nwSnq5uNBop\nKiqiqKgo0UMJBED3pMrHZ92HWW+6oveUFYVWTzueoBeD3oBeq6P3YrD+xxfys6eunG1n9lHZchI1\nxiKsBg1jM0cyp+haZhVOIdlkH4CZCq5WVFXF6w7265hdiZmyJKHT69DE6XnA75dobvTS3OjttZoj\nNc1EVo6VzCwLxvPEhKqoVJ+q5fDeCioPHsfvi31FsSfbmDAtLCYKigd+maMvDJ06OIGgj1yYVCnJ\nElymsDjXJKwTrVaLQT+4Im2yInOo6Rhbz+yjrP4wISX209LItALmFF3LnMJppFtT+3mWAsE5K+5g\nqP+8FdqbO5BlGYNRF5coRSik0NIUFhO9+UwkpxrJzrGSlW2NEhMALY1tHNp7hMP7KnE6Ypdzmy0m\nxk8by6Tp4ykaVTCoo6OxGLLCorGxkWXLlvHLX/6S0tLSHo8LBAK89tprrF27lo6ODsaPH8/TTz/d\nzU68vLycn/70pxw6dAi73c7y5ct5+umnxZLNVYw74KPV20a4Sdjg+qnUdDaw9fRetteU9eg1kWVL\nZ17xdOYVTycvKbufZyi42vjLr97ls7VbuOmOBTz8L+eazqmqitsZFhThZY+e7ajjTZeouBIURaW9\n1UdTg5e2Vl+Pfm/2JAPZuVaycqxRORMAXrePw2WVHNp7hIaappjnG4x6xkwazaTp4xk1bmS/GoLF\nm8F1tewjDQ0NPP7443g8nose+93vfpeNGzfyb//2b4wcOZLVq1fz1FNP8Yc//IFZs2YBUFNTw2OP\nPcaMGTN49dVXOXHiBK+88goOh4OXXnop0R9HMMiQFIkmVytBOTSoBIU76GVHdRmbT+/hjKMu5jE2\no5U5RdOYXzyD0enFg3YNVjC8CPiDfLp2M6qi8tnaLXzh0WWYzEYCfgmPJ4CmnxqGeVzeqNeX+/ff\nVdHR1OClucmLFIothixWPdm5VrJzrFHVHBAuZa2qOEX57sNUVZyKmYSp0WgoGVvM5JkTGDv5Gowm\n42XNd7AxeK6afUBVVVavXs1PfvKTyOveqK2t5YMPPuCFF17goYceAmDu3Lns27ePN998MyIsfvvb\n35KUlMT/+3//D71ez8KFCzGbzbz88sv8y7/8C3l5eYn9YIJBgaqqtPs66fS70Gt1g0JUKKrCkaYq\nPj+9i331h2N2DtVpdFybP4H5xTOYljcevXbg5y24upAkKeL4qCgKUkgiGFAIBaV+6e0RDARxtDkI\nBHpOnOwLgYBMU4OHpgYPXk/s9zIYtGExkWuN8pnoorm+hQO7D3NobwU+jy/me+QWZjN55kQmXjsO\ne/LwK+ceUlegyspKvv/97/Pwww8zd+5cnnrqqV6Pz87O5p133mHEiBGRbWHlrCMYPJdAtGXLFhYt\nWoRef+7ruOWWW3jppZfYsmUL9913X/w/jGBQ4Q35aHaHlz0Mg0BQtHo62Hx6N5tP76HdF7tHR0la\nEQtGzmRO0bXYRUWHoA/0hx03hEtIrTZrwkXFO39Yy5Z125kxbxq3LL8Rne7Sx1MUlbZWH431Htrb\n/DGNp7RaDRlZFnLyrKSln3PA7MLv83N4XyUHdh2msTb2Uoc92cbkGROYUjqRrNzMS57nUCLhV9AN\nGzawYcMG6uvrefbZZ7FYLOzYsYPly5djMl1aQl1+fj7r1q0jJyeHnTt3XvR4o9HIpEmTgPDTaGNj\nI6+//jo1NTW88MILAPj9furr6xk5cmTUuenp6djtdk6dOnVJcxQMLSRFosndRkAKDLhjpqTIlNUf\nYdOpnRxuOh6zqiPZZGf+iBlcP2IWBSm5AzBLwVClP+y4u9BotQmvXmhraWfLJ9tRVZV92w9w47Lr\n0VxCoz+PO0RjvZumBi+hHpY6UlJN5OSF8yb0+uilHFVVqT5Zy4Gdh6g8cCxmiahOr2Pc5GuYUjqR\nkrEjhlwS5uWSMGEhSRLPPPMMn332WSRU9Pjjj1NdXc1LL73Eu+++y//8z/+QnJzc5/dMSUkhJeXy\n+sn/5je/4ZVXXgHg/vvvZ968eQC4XOHEN7u9e9mdzWbD7Y5tddwbXm/0Op/PFzscJhg4FFWh3evA\n6feg1+kGVFS0eNrZdHInn5/eHdNaW6vRMi13PNeXlDI1d/yQbb0+EIjf4jkSacctS/2XkBkMBOlo\ndeDz+CLL4aqiIssy+otESGRZoaXJR0OdG2dn7LJXk1lHTp6N3DwbFmv3b8jj8lK+5zD7d5TT3ho7\nmphfnMvU0klMnD4Os8V8iZ9w6JMwYfGb3/yG9evX893vfpfFixezdOlSAG666Saef/55fvazn/HL\nX/6Sb3/724maQhQ33ngjM2fOZO/evbz22mv4/X5++tOf9uhq1sXlKMzp06df7jQF/UBXtYdGo8Gg\nH5hlD1mROdhYyYaTOyhvPBYzOpFjz+D6kbNZMGImqZa+C3DBOcRvMbF0+VI4HYkXbIqi4Ghz4PeF\n+3po9X2/NnvcQeprPTQ1epCl7r81rVZDZraF3DwbqemmbnkTqqpypqqGsu0HOXqoCkXuft+w2CxM\nmTmBqbMnk503vJc6LkbCrqpr1qzhrrvu4ktf+hLt7edshA0GA48++ihnzpzh008/7TdhMWbMGABm\nzZqFJEn84he/4NlnnyUpKdxUKVaFidvtjuwX9C+JcM6UFYVmdys+KTBgeRSdfhebTu1i48mdMXMn\n9FodswqmsKhkDuOzRomqDsGgxunwYTDqE55L4XF5cDpcaLXaS1q+aW7y0tTg6TE6YbMbyCuwkZ1r\nw2DoLlR8Xh8Hdx+hbPtB2ls6ur+BBkaNHcG0OVMYO2n0kC4RjScJu7o2NDT0+rQwfvx43n777UQN\nD0B9fT1bt27lrrvuwmg8V8YzceJEAJqbm8nNzSUnJ4czZ85EndvW1obH42H06NGXPG5ZWVnUa5/P\nx/z58y/jE1ydJMI5s9Pvos3rQKfV9ruoUFWVE+3VfFq1ld215chq97XYXHsmi0fN5boRM0kSTb/i\nhvgtJhiNJqHiNxQK0dHqQJbky7ppH6/sLgZ0Og3ZuVbyCuzYkwwx519f3ci+bQc4UlYZM3fCnmxj\n2pzJTJs9mdT0y1ueH84k7Aqbnp5OdXV1j/srKipIT09P1PBAuNz0e9/7Hlarldtvvz2yfevWrRiN\nRkpKSgC47rrr2LhxI88//3xEgHz88cfodDrmzp17yeNarSJD/0qIp3MmQJ2zGaPe0O+CIiiH2FVz\ngE9PbOV0R3ffCZ1Gy/T8Sdw4eh4TskaL6EQCEL/F+HKxEv94jtPZ4cTr8qI36vskKjraA7Q0xTaL\ng7CBVX6hnewcK7oYyyhSSKLiwDH2bCmLbWKlgdHjRjJ93lSumTBqQNqRDxUSdqW9+eabefPNN7n1\n1lspLCyM2vfRRx/x9ttvx72M0+12U1VVRXFxMenp6ZSWljJ37lxefvll3G43RUVFbNiwgTfeeIOV\nK1dGljmeeOIJPvzwQ77yla/w6KOPcvr0aV555RUeeOABcnNF5v1QR0WNe8KjO+DhzQMfRG3zBL3Y\njBYcPifrT25nw8kduALdl9jSLCksLpnNopI5IndCMGQIhWQc7d6LH3iF+Lw+HO1ONHBJ3UePHGxF\nb4g2mNJqw9GJ/EI7ScmxzaecDhf7th2gbEd5TN8Jq93CtNmTmT53KqkZIjrRFxImLFauXMmuXbt4\n8MEHGTUqXC/96quv8tJLL3H69GmKi4v5+te/fkVjXPiEd/jwYVasWMGPf/xj7r77bjQaDatWrWLV\nqlX8+te/pqWlhZKSEn74wx+yfPk5y9lRo0bxP//zP/znf/4nzzzzDOnp6Tz22GOsXLnyiuYnGBh8\nUmLbg7kDHn686dfUOqPbpv/n57+hJL2IvXWHYy53jMscxU3XzGdG/iRR2SEYUvi9QTyeYEKf0mVJ\npr3NgRQMXTRC4fNJnKqKXZFxPqVzczHHqOxQVZW60w3s3ryPyvLjEXOv8ykcmc/M66YxbuqYKI8j\nwcVJ2LeVlJTEm2++yeuvv87HH3+MyWTi0KFDFBYW8vjjj/PUU09dUqnphcyZM4eKiopu2yorK6O2\n2e12nn/+eZ5//vle32/WrFn87W9/u+z5CAYeVVVp9XTQ5G5J6Dh/P/SPbqICoMXbQYs3ek3XoNUz\nb8QMlo6+jqLUq8vBVVEVZEUZFIZjgssjbG3tJxRS0Ol1yMErc7bsaQynw4Xb5UGn0/YoKlRVxdkZ\npPaMi9YWH1Koe0KmwQCh0PnnRFdvyLJM5YFj7Pp8X8zlDr1Bz+QZ45m5YDo5+VlX9sGuYhL6i7da\nrTz99NM8/fTTiRxGICAQCtDobgVAn2BPipoe+nScT7olhRtHz2dxyWzsV0kypqzIKKqKBg1GvQGr\n3ozVaMGkHx79D64muhqKXbd0Lsu/tCyhfT5aG1rR6bU9elCoikpLi4/aMy5czp5brms1IZyOABbb\nuUq+PTtqmLtgBJIssX/HQfZs2Y+rs7tXTEp6MjPnX8u0OZOwWC1X/qGuchL+KOH1emlpaenRL6Ir\ngVIguBxUVaXN24Ez4OmXJ2NP0EeHz9njfpPOwGOz7mNWwZRhv9whKeH+EDqdDqPOSLLJhtVgFRGK\nIc75DcW2rtvBHQ/ektDmWBqtNuYSiywrNNZ5qK1x4fd1X1o0W6L/zjyeAHp99ENFe2sH7/yxgpqT\npwgFQ1xI8ehCSq+fzphJo4e9K+bFPJviScKuADU1NTz33HPdyr3OR6PRdFvOEAj6iifoo8XTDqgJ\nv5l1+Dr55PgWNpzcgb+XHI45RdOZW3RtQucyEKiqGmmAZtTpMWgNpJqTsRrM6Ia5gLqaUBSVjjZ3\ndEMxScZ4ZdXeUfi9vedAhYIydbVu6mrcMbuKpqSZKCpOwmqDTWtjv4e7s4O6U0dpbajlwuYfWp2W\nSdPHU3r9DHILsy/3Ywx6ZElGUUGn1aA3GrBYzCSn9U/0NGFX45deeony8nKWLVtGYWFhzOYworxO\ncDn0p9FVs7uN/zu6kS1n9sTsLHo++UnZ3D/ltoTOp7+IEhJ6Axa9CZvRilnf3ZVQMDwIhWRcnX6U\nGImM8UBRFDpaHXjcsV06A36JmjMuGuo83eeggewcK4XFSZHqDr/PH3WIqqo4O9o4ebgMR1v3/AmL\n1cz0+dOYdd007MndWzgMZboszUGDwaBDZzCQlGLHaDJG3Xt1uv757SbsqlxWVsY///M/8+yzzyZq\nCMFViDvgpcXTjlab2C6k9c4m1lZuYGfNfpQLEsBsBgvXjyylzetgd93ByPZvXPfYkM2nUFUVWVVA\nDUd/zAYTSUY7ZkMcH1UFgxavN4jPGwrnUiRAOPq8PhxtncghheNHopOrXU4/zU1+mho8XGiTodVp\nyMu3UVic1G3p40Kqyvfgc3dfprQl2VmwdDZTZk3CaBrYRoPxQlEUFFkJtyUwGjBajFisZvQG/aAQ\n/gm7MlutVnJychL19oKrDEVVaHa34Q35EtowrKazgbUVn7G7trxb/45UczK3jl3IopLZWAxmPEFf\nlLCwDbHW5ZIso6Ji1OmxGMwiInEVoqoqrk4/kqxcNEEz6A+xf09t1LaQX8LYi9fE+SWkiqyya3s1\nbld0guXBslYg+m/OYNBSUGQnv8iOoQcL7wv7dVwoKuwpaRSOuobb752DyTK0E4gVWUGWVbRaMBgN\nWKxmLFYLesPgzGdK2KweeOAB3n77be69917M5quvu5sgfvikwNkohTZhoqLG0cB7FZ+yp668275s\nWwa3j1/M/OKZQzoxMexiKqPX6jDpjaRbUrAYzGg1wztpbajSVZlx0x0LePhfll/8hEtEkmS87gBo\nNBdNXAz6Q2xYdxynI3r5YfOGE9x4y9iY57g6Xbic50pIK8vrcbsCaLr9vZ0TFSaTjqIRSeQW2HoU\nOrIkU77nCFs/3Rlzf2pGDgWjx5OSnsWc+blDUlTIsnw2MVqL3mDAarNitpquyE8jvLw0xJdCvva1\nr1FWVsbSpUtZsGABmZmZMZ+ExFKJ4GI0uloS5lBZ19nI6iPrYgqK/KRslk24kTmF04ZkgmIkT0JV\nMeoN2A1WkpOShrQ4ulo4vzLjs7Vb+MKjyzCZ43uDdHX6MffxplteVt9NVAC4nAHK99czYUp0EmRL\nQwsmkymqhNTR6cdgMqOL+fenMnZCOjl5NrTa2Dc/WZI5sPsw2z7bibOju3X3qHEjWbB0HidOnPPa\nMFyCc+dAosgKiqKEH56MBmxJVsxm82U3NVNVNZKnotdp0Oq0mE06jOb+WQpK2Le+evVqtm3bhqqq\nrF69usfjhLAQXIxERCkaXS2sObKOnTUHui15FKbkcueEJcwqmDzknuYlJdxnxag1YBXLG0MWSZIu\nqMyQMBFfYXEpLpodHT23Re9o775Po9VE2pr7vBJnTnUiSXr0+u5/h0G/D6tdT15B7ITKiKD4dCdO\nR8+9QDo7nKSkJ8OJ9h6PGSx0JVtqNBoMJiNWmxWL9fKERJeI0BC2MNfqNOh1eoxmPTpdYpvE9UTC\nhMUvf/lLCgsL+fd//3dGjhwZsypEIIiFM9DdwCZetHk7eO/Ip2w5s7dbUmZhSi53T1zKjPxJQ0ZQ\nhKMSEjpNeHkjXAJqQTfMa/IFl04oFH/XzC5kqft7B/wSZ045aazvSsrsIRIhS6SmdhcVsixTvvsI\nWz7dETNCcSFtze1sXrcDW2rspZmBRg7JqITdPc1WEza79ZJzJFRFPWtCFy0iDKawwdhgeYBImLBo\naWnhueeeY9GiRYkaQjDMUFSFJlcrjl4MqC5GT83BZFXmw8oNfHZiW7ey0fzkHO6ZuJSZQyRCISky\nqqKg1+qxGs0km+wYhbuloBe87iCuzsvvoZOWZqGjrXsDMlVVMRtVWhujowSnTnTS2hyKUbqqEvT7\nMZrPuVva7CbGTji3lKIoCkfKKtn88XY62jqjztbqtJjNJrwxmoUBtNS3DBphoaoqsiSj1Woxmo0k\npSRhtvQtenj+UoaGcJmoRqfBYNJjMA5cJKKvJExYjB49mpaWxPZsEAwfvCEfze42NBoN2svMZ+ip\nOdjL61cRUkL4pWg74GxbBndPWsrcomsHvaCQFDl8EdcZSbbYsBttQzLvQ3D5bP/8JIcPNDBpWh7z\nFo7q0zlRVR8xWoX3lSnT82lt9XTLs9BoghSPygWio3/1Ne6oTqNms44Ro1JITzdy9Egz7e3njp81\ntwiDSY+qqhw7VMWmj7bR2tQW9X5anZZpsycz/6bZvPuHtT0Ki4FGDsmoGjDodRjN4aiEwdj7Um5E\nRKgqWp0WrU6DQafDYNCh18d2JR3sJExYfOMb3+DZZ59l/Pjx3HTTTRgMw6N+WNA3XtvxZ9YcWcfd\nE5fyr3Mf6fG4rsZhrqD7inMpemoO5gpGty5Ps6Rw14QlLBg5a9DabndVcOi0Oix6E5nWNJErcRUj\nhxQOH2hAVeHwgQZK541A30MZZuQcWcHp8J1X9XH5ls5Gs4Eblo5h/95azpw812hv9rxiTGYD1ac6\nYp9n1FJckkxegT2SlDlhSh5bN53rt2Mw6jldVc3GD7dQXx39+9VoNUydNYnrls4hNT3csjynIDtm\nAzGArH5uHCbLMooCBr0WvTEclTCZjb1W2SiKiqqqaDUatPqwH4/RqEVvGDxLGVdKwoTF73//e8xm\nM9/4xjfQ6/UkJSXFzLPYsmVLoqYgGCD8UoA1Rz5BUVXWHFnH47Puw6zvbrQUlEI0uFvOmjJdufC8\nWHMwm8HCsvE3ctM18zEmuFHZ5SArMooSNqiyG20km+yigkMAgKyoEfMoVQ2/7u0vIxiQcLkCkZLN\ny/GguBCj2cCEKTlRwsLjkTh0oBVnZ3QEQafTUDI6hYJie6/+GB6ng7+/votTx85E79DApOnjuf6W\neaRnpkXtuuGfFlB3uoGWxtao7RnZ6Vy/dC77dic2eVMKSWeNqYzh6g2LudccQkVWUFX1bNmtFrPF\ngNEwNCMRfSVhVy1JkigpKRFNxq5CJDlcmQBdT94SXCAsOv1OWr2O8A0+Tio9KHdvMtRFqjmJH938\nTWzGwdW5UJJlQMWoM2I32Ug2iyUOwZXh9Qbxd7locnEPir6UICqKQmeHE1dndJ7F4YNtMY8/fnAb\nc667t0dREfB5qT5+iOa6M932jZk4ikX/dB3ZebGjDxabhUe+dh/r3t/EoT1HItvvf/zuhHUm7RIT\nRrOJlLRkTObYjrQXVmiEhYQJo1HXYxntcCRhwuJPf/pTot5aMITpStD0y8GYUYOeki97EwTOgJvV\nhz+h1hk7PAowLW/ioBEVYW8JMOuNpNlSsBqFSZXgylFVFbczQDAkR93QL+ZBMXPuiF7fNxAI4Ozo\nPPsA0LebY0tDMxv+bwv/dN/S6PfyB9j66S72fb63W7fNwpICbrz9egpL8i/6/habhZvvXhwlLCy2\n+BoxdokJk6VnMaEqKqqigCYsIgwGHUajblgta1wOIs4q6DcCoQD17ha0Gk3M3Iaeki//v62/5zuL\nv9qtD4ekyKw/sZ3VRz7BF+p+4ewiPymb+ybfGp8PcZlIiowGDRa9iQxrKha9+aq+8AjiiyIrdHaG\nfwMXRgku1YOi2zEtndiSLDTWezhV1dltf2d7C6cqunexbqprPjc/ReHArkNs+mgbXnd01CMjO50b\nl13PNRNHDehvIuItodViMhtjiglFVlABrUaDTq9Bb9JjHAJVGv1N3ITFAw88wMqVK7nuuusir/vy\nRf/1r3+N1xQEg5gOrxOHr7NXS9qeki/rXc38/dA/eGzmvZFtR5qr+Mv+96i7IEqRbcsgw5pKRcuJ\nyLaBag7W1YvDojeRbksdNBETwfAiGJDCVtnaxNzcfD6JY5XNuJzBbvuaayo5fqi7a+35nK6q5tM1\nG2luiM6JMJjMFI+ZxD0PXYfRPDDPuIoS7sFhMOgwW81R3hKKoiJLChpNOGdEe5Uua1wOcfuv2dLS\ngs/ni3otEAA0uFox6PQX9bnvLfmya1+718FfD37ArtqDUfstBjN3T1zKTaPnE5CC/Ov7L0b29Wdz\nsPPLQtNsKdiMFvEkI0gYUV1Je6AnDwqAtPRosauqKp3t0VGJg/taokpHz8eW1PO4qRmpvPO/73O0\nvCpqu96gJ2/EWApKxqHT6/s9ibFLTOj1Oqw2K7YkKzqdLlytcfafVqfBatFjNPX//IYDcRMW69ev\n7/W1YGjR13LRviApIcw9XJj6iqqq/OPY56w58kmUH4UGDQtLSrl30q0km8PufZdvA3RlaNCQYUnF\nbrKKnAlBwnE7/QSC8kW7kvbkQZGUbGLKtefyGYKBIO0t7fj9Pf+CMrMstLace4C8fulc6k7VdavQ\nsFjNHDtUhSxFm9FNnjmB62+ez8H9l2+CdzmoqooUUtAbtBExodFowyJCo0Gr5Wy1hk4IiTiQsG9w\n1apVHDt2rMf9Bw8e5MUXX+xxv2DgOFcuqrDmyDr80pXdqrt3M4xNUWpBj/uavR389eAHUaJiVHox\nL9z4NI/N/EJEVPQXqqoiXVCFUpCcTbLZLkSFIOG4On2EpIu3OodzHhQjRkWXbV5/w+hIRUhnh5PW\n5nYcHV62rj/Z7T2sNj3TZmQxbmJ61HaLNVyhMXnWxKjtPq8/SlQUjMjj0Wce4s4v3kZSShL9hSRJ\nyJKM0WQkJz+TrNxMbEk29HodZouetHQLaRlWklMsmM0GISriREKFxdGjR3vcv3v3bt55551EDS+4\nAmKWi/YRVVVpcscuQbsY902+lcLk3Jj7PMFzoVy70cpjM+/l32/4GiXpRZc11uUiK+Hvw6QzUpSS\nh/bsModWo0UvPCcE/YQsq5e0xGY0G7i2NPq3YjDrURSF5sZWPE4vp084OLCvHZ0henmkteEkEyen\nkJoeu+rCYrMwZ+GMmPtsSTbueOhWvvz0g+QX5/V5vvEiOS2F7Lws0jJSMFsM2JNMpGdYSUm1YLUa\nhZBIEHG7ElZXV/OlL30JWT6nUn/wgx/wk5/8pNuxqqrS3t7OqFF9s6UVDA0kRaLO2UxQ7p7k1Rfs\nJhvPL3qKNw98wNbqvTGPWVgym/sn39avyZiKqiDLCia9kWSLnSSTLRKVuHvizZElo1gmYAJBItD0\nkjz47h/XsuWTHSy4eS7Lv3xHj8f5PD48Lg8ej8Txig48nhBarZYLWulw4nA5m5LoVjoK4c6jOzft\nZsu6HVHbtVotpQuns2Dp3B49H+KNFJTRX2Bbnp5mw55iEcmW/UzchEVxcTEPPvgg27ZtA6C1tZWs\nrCwyMjK6HavVasnIyOCJJ56I1/CCAcYd8NLiaUOv0/d56SMWkqrgCXUvgStIzuHRGfcyJnPkFczy\n0gjJElqNhiSjldTklJjGVf8695ErzkERCC5GINC3qGEwEGTzx9tRVZUtn+xg2QO3YDTFzm9ytDqp\nb/BSeya6c2jA3710+/zS0S6c7a38788/o625u9Pll56+n4IRF/ejiCdZuRlY7BagJrLNZDUKUTEA\nxDV2+9WvfpWvfvWrANx44408++yzLFmyJJ5DCAYhbZ52OgOeK7KfVlWVbdX7eGP/+92ExR3jb+Su\niYURPKsAACAASURBVEvQaxO/1BCOTsiY9CaybRmiqkMwKPB5+5bnJEnhqiQIVz9Ikoyxh4DBwQNt\n+H3RgqXhTBWnjx6MfULXGKEQp48epKmmey5GFxnZ6T3uu1IURUGWFLigcWp6lr3H6hVB/5KwK3Wi\nq0IaGxtZtmwZv/zlLyktLe3xuGAwyOuvv86aNWtoamoiJyeHO+64gyeffDKqMdrChQtpbu6uynfs\n2EFqampCPsNQR1EVGlwthOTQFYkKh8/J/+57h/0NFTH33zp2UcJFhazIqCrYTRbSklP6RcQIBH2l\nt6ZWfUFVVTo7oisxzhcVFoue9uYTnDzS3egKwo2/AI4fPkHZ5nUEA+dFNTQwbfZkDuw8dEVz7I1w\nVYeERqvDbreQlpGMql5+UzVBYhmSV8+GhgYef/xxPB7PRY/94Q9/yNq1a/na177GlClTKC8v57XX\nXqO+vp4f/ehHALS3t9Pc3My3v/1tZs6cGXV+UlL/ZTAPJUKyRJ2zEY1Gc0W9LXbU7OdP+1ZHRSlM\nOiOBy8zTuFRCsoRBqyfdkkqSySaiE4JBQTzvmZIk0drYjtMRO+pRWGxn5OgUgv5UaqpOxmzuNWfR\nTFb/6UMq9kcn5GfmZHD7A0vJyE5PiLCQJRlZUbFYjeTkZZCUYo1Uwnjcsb05BAPPkBIWqqqyevXq\nSEJoV8ivJzo6Ovj73//Ot771Lf75n/8ZgLlz5wLwX//1X3zzm98kLS2NyspKAJYsWUJRUf9WGQxF\n3AEfLZ7WmFUQfe314Q56+eO+d7sZXU3KHsODU5fxvU9fif/Ez9KVjGkxmMhOSsdkEEmXgsGDJMl0\ndsTnpul1e+loC9JY7+1mx2226Bg/KYOU1PDff0/NvWbOn8YfV/0Nn+ec+NdotBRdM5H7VtyAyWLA\n7+vZUv9SURUIBkMYTXrSM1PJyE5GrxeN+YYSQ0pYVFZW8v3vf5+HH36YuXPn8tRTT/V6vMfj4aGH\nHuLGG2+M2t7VcbWmpoa0tDQqKiqw2WxCVPSBdm8nDn9nzDbnfe31cajpGL/b/RYO/7nQrFlv5MGp\ny1hUMgdvL30/rgRZCbcvTjJZSeshGVMgGEhCQQmnM4AmTmWQrc2d1Jzx4ejoHq24dmY2pgustGM1\n9/pkzYaoY5LSMrhm8iys9mR0cbrhq2q4D4eiKNjtZopLsuPeVEzQfwwpYZGfn8+6devIyclh586d\nFz2+sLCQF154odv2zz77DIPBEBEYFRUVpKSksHLlSrZt24YsyyxevJjvfOc7ZGXFbt17taGqKo2u\nFvxyMKaogIv3+nj42jt5u/wjPqnaErV/XGYJT5Q+QJYtMQlfsiwjKwoppiRSLUliuUMwKPH7Qng9\nwT6ZXvWV8v2t9HSZv9RxDEY9C2+5Dl8oK66/IUmS0elU0jOTyMhOE9GJYcCQEhYpKSmkpKRc0Xus\nW7eONWvW8Mgjj0TyJ44ePUpzczOTJ09mxYoVnDjx/7N33/FN1fv/wF8neyctnbRllAKlUKC07I0g\nVUG5BS9UEC4OxIle8Ot1XScKLryAgx8qyhRFwAEyBAEBQZBNCxQKBVqgdDdp0ibnnN8foUnTJG3S\nJi1t38/HQx+ckzM+UU7yzme83xewcOFCTJkyBRs3boRc7l3xqLIyx27MqjVUmiKWY3G5+BoA3mVV\n0ko11fq4kH8Jb+1cjCvF12z7RAIhxndNxuhOg/2arTJQqUW4OsRv1ye3r6byLJYZKmA01lzzozau\n6nxY01hb/9wqSIb8vJp7AyvKK7Btw+9O+6OiIzBm4miotRrs2+3+OfcExzkOYYdGBCA4WOd1siqR\nSARGwFjTcgsEtdYiIg3Hr/8nWJbFwYMHkZeXB45zPRtp3Lhx/myCg23btmH27NlISkrC888/b9v/\n9ttvQyKRoEuXLgCAxMRExMTE4IEHHsDGjRuRmprq1X0SEhJ82u7GdrX4OtQyVb3yU2SX5IKvsj4s\nQhOKmX0eQJTO99n42GoZftSShq9sSm4PTeFZ1JeaUFFee82PmrAWFjev56PYRYl0gYBBTGcdgoLl\n2L8nx+01srNy8OOqX1GU7xicDL1rIAaM6ANGwMBirtusUo6z/ksoFkAkcfxxogvQ1CkDplQmwcix\ng7Hj5724Y+wgSGUSj/N9EP/yW2Bx5swZzJgxw+USzkoMwzRYYPH1119j/vz56NevHz755BNIJPb1\nzj169HA6vlevXlCr1TWmJW8phEKRR12fUboIZBZedfla1aBiZMxA/DP+bkjcDKnUlYVjIQCDQEX9\nerUIaQg8z6O0xASLhatXammDvgzFBcUoKqzAqWOOn7cKpRg9EkOhUIrdBgUcx2H/jr/wx7Y/wXPO\nE+ITB/aoMdOnOzwPW6VQqUwIkVAMgYCBTOGYxKo+Js9MweSZKT65FvEdvwUW7733HoqKivDUU08h\nNjbW4Yu8IfE8j7lz52LlypUYM2YM5s2b59BlptfrsWXLFvTo0QMdO3a07ec4DmazGYGB3o/7Hz3q\nuBbcaDRiwIABdX8TTcT93ZJxIT/L5TwLwFrj45Gkf6Jn6ziXr9cVy3FgAAQpdFBLVdCX174MmbQM\nt+uzyPM8SopN4G5141dlKDVg48pNDvvK9M49ETzPo6igGEa9EZez9LiSVQqWdQwMevQKhkLpPoAv\nKSzFj6s340qmfXiDYZhaV9zVhrNwEAoAmapyxRUPjU4NtUaJigq2xnNJ0+e3wOLIkSN46KGH8NRT\nT/nrFh756KOPsHLlSkyfPh0vvPCC0+tisRhvvfUWRo0ahQ8++MC2f+fOnTCZTOjbt6/X91QoFPVq\nc2PLLyus03mVtT7+t+8bZBRccnitU1B7PN73AQTIfdebYB3yYBAo1zZ4dVPSNNyOzyLP8ygqMgKc\nc80PQ6kBi9/+AteuOAbnyxd/i+hug2zbLMuhMK8Q5SYzzqYVulz1AQACofuehvNpmfh13TYYy+zz\nLrSBGiSPvwNrl26oy1uzUevkkMkkt/6sglrTMDlihAIGDGPtLWEY6zZpeH4LLKRSKUJDQ/11eZf0\nej3Onz+PNm3aIDAwEOnp6Vi6dCni4+ORnJyMY8eOORwfExMDlUqFGTNmYNGiRQgKCsKQIUNw7tw5\nLF68GCNHjqxTYNFUWSuT5sFQUbflnmbWgnWntzgFFckdh+D++Lt8tryzMktmgFwDrYxWeZCmg+N4\n6zwIxnUhsV/WbnUKKgDgRs5NRHezb9+8loeKCg5pJ/JRXm7vAZArav9I5zgOWWdPYN+vGQ774xI6\nI3n8SDjlyq6BdZkoj+qPIGdhodYoodapGvT5FImF6NojHKePX0PXHuEQiWmFSWPwW2AxYsQIbNu2\nzeuJj96o/hf29OnTmDZtGubNm4dx48Zh27ZtAIBTp05h4sSJTucuX74cvXv3xhNPPIHAwECsXr0a\na9asQUBAAFJTU/H000/7re23G47nkF1yAxzH1Sl9cKGxGIv+XIHMgstOr43tcodPgorKOh5amRoB\nCq1fV5IQ4msWC4uSIhMYAeP2yzY765rL/dUV5pfj/LkihxUWIaEKtGkvx66f3J9XbizDmaN/Ql9s\nLxwmlogw+h93IL53HBiG8SrZlUgkgFojQbnJ8f2ER4VCpWmcSdP9h0Sj/xCqnN2Y/BZYTJ48Gc88\n8wwee+wx3HXXXQgMDHT5hTVo0CAXZ9eub9++SE9Pd9pXmUUTAGbNmoVZs2bVei2GYZCamurXIOh2\nxnIsrhZfB5i61STIyLuERX8uR0m53g+tszJbzFBI5AjWBFJiK9LkVCa+8lWOinNnqgxXMkCHGB0i\n2qhQbnJfrOziuUs4tm87LGZ7uvzg8CD848ExCAr1bC4ZV20Oh0IhhkAggNnsuBqjLpM9SfPht8Bi\nwoQJAICcnBzs3r3b5TEMwzgFB6RhmVkLrhZfg1AgrFOX5e6LB7H8yEawvL07dmDbROzL+ttnbRSA\nQZSudb0KnRHSWMpNFuj11qDi2KErOH/2JmI6B6Nnb+dMvxFtw3H5guuVVa6IxALExbdCQKD7LJU8\nx2PfjoPYs3W/wyhH997dMHr8cIdijO5wHA8Bw0Cpsk/CZxjrsArHCxAUGuBxm2keRPPnt0/qd955\nx1+XJj5SbqlAdsmNOn1hsxyLb09swvYqWTSFjBAPJtyH3pE9fBpYtNaEUFBBmiSTyYKyW0GFxcIh\n4+xNgAfOn72JbgkREIkcezDGTByNSxlXnOZZaFu1crq2QilCtx7BNc6rKDeV4+c1W5Fx+oJtn0Ag\nRIeuvZA8fiBEYg96UHhAIZdAIrX2FLbvoMOlzCK0aadBQLAWGq3Kq4JgNA+i+fPbp3VKCq0tvp2V\nm8uRXZpbpy9so9mETw+swskb9hwfGqkKT/efio5B7WCoqHt2Q47nwHH1W44mEoogYBhwPA8BI3BZ\nLI0QfzOVVcBQZs+myXGcrceA529tw/GLXalW4qlXHsGGFZtweK99qWxMt97Vrs6hW/dWtU7WXP35\nDyjMK7JtyxQqxCb0h1Kj8/h9qDUyW1ABALFxAeieEIpWoQEQCusWFNA8iObN75+4Z86cwW+//Ybs\n7GyIxWKEh4dj+PDhiI2N9fetiRtGswnXSnIhFnmfoCq/rBAL9i5zyFXRVheBWQOmIVDh+YeVKxaL\nBWqZEgEyTb2uIxNJMS7uTmxM245xcaMgE1H1UtKwysoqYCqrW4pumUKGwaP7OwQWgmrziowGAzIz\n8tC1Z+sar1U1qOgQ2x5Bkd0hErvPKcRZ3GfWZFnra61CAyFXUIEw4p5fA4v33nsPX331ldP+hQsX\n4sEHH8RLL73kz9sTF8rMRlwrvQmJm6CiprLnlwqvYsG+ZSg2ldpeS4qIx6O9J0IqqnsCNJZjIWAE\naK0NhVQk8UmCqyf7TcGT/abU+zqEeKtMXwGTyVynbJoV5RXIyy2AJ0s+i4udV2+4S2w1aFQ/9B/e\nF/v/cJ3S2zaHQi2F1OzcbrPZAoVSjsBgHS3vJrXyW2Cxbt06fPXVVxg2bBgef/xxREdHg+M4ZGZm\nYsmSJVi+fDni4uIatFZIS1dmNuJ6aZ7bVNo1lT2/r8tIfPX3OpSz9hnlYzoPR0q30fVa9mm2WKCT\naxzScNNQBmmqDPoKlNcxqDDoy1CcXwyRRIQrFx2Da08yYeqL9Vjx6XcO+0RiEe6bfBc6x3e09TgA\nuJVHw552W64QQyq1PmcWs9nhGizLITwqEDI59VIQz/gtEcDKlSvRu3dvfPbZZ+jRowfUajW0Wi0S\nEhLw6aefonfv3li1apW/bk+q0Zcbcb0kr8Y5FTWVPf/sr9W2oELICDA9cQImxN9V56DCmuSKR6Q2\nzKm2R+VQhoAR0FAGaTLqE1SUFpWiuKAEQrEQlzKLkXnesRBYucl53pJWa/+iL8wrwmfzljkMfQCA\nUq1Am+hIANYy6RFRKoABIiJVYMBAJGKg0clsQYUrYRHBFFQQr/gtsMjMzMTo0aNddpsJBALceeed\nyMjIcHEm8TV9uQG5hjyIaykrXFPZ80oykQTPDXwIQ9v3qXN7zBYLNFI12uhaux2SebLfFGx/6Bsa\nziBNQlk9gorC/CLoSw0QigS4kFGErMwSp2N43nHug1IlRacuIQCAovxifLVgFcwVZqfzigtK8Ptm\n+8qtmM4BGDwsEh1jA6DWSqFUSiGo9hldvXeEclIQb/mtj1kul6OoqMjt60VFRY1WmKwlKTaVIr+s\nEGIfVBLVSFWYPehhtA2IqNP5leXMI27NpSCkOSgrq9ucCp7nkZdbAEu59dyMM4W4lu16flHrCA1K\nq+SfS+oXBbFUhGtXruO7LzfWmBjrRnburftZqwwrFBKHVR5VWViO5lCQevNbj0Xfvn2xatUqXLni\nXB73ypUrWLVqFXr3rr6EivhSkbEE+WVFHgcVUTr3AYNcJMUrw5+sc1BhZi1QiOVoq4ugoII0G6Zb\nqz/q0lORdy0flgozGCGDs2kFDkGFUuX4zHaKc6y7JJaIcOHMRaz89HsYSmvPIcFZOIjFQmirLR2t\nqsJshkqtQEjrIK/fi0gksvVsCAQChwrSpOXx2//9WbNmYcKECRg7dizuvvtuREdHg+d5ZGZmYvPm\nzRAIBB6l2yZ1U2wqQYGx2Ks8Ffd3S8b5vEvILr3hsF8kEOHVEU8hROWcpMcTFtaCSE04VFJ5nc4n\n5HZkMjrmqfDarUJkZ04XIPe6PTjQaCXoGKvGzo3uTz19NB2/fr/9Vi4M67XcLSQJjQiBWieD0E26\nfo7jwPNAWOtgSKQSlJsqwAgY8LdKunsSJEhlEowcOxg7ft6LO8YOglRGPx5aMr8FFh06dMDKlSvx\n1ltvYf369Q6v9ejRA6+88go6derkr9u3aNbhD++CCgCQiqQIVOicAotXhz+F1pq6V6qN1IZRUEGa\nFZPRDIOhHkHFLWerBRXaACniewTBXKWeR3U5F89h35njtm2xRIyxk0bjj20HcPN6nsOxQWGBGDtx\npNugwmJmIVNIERiss9UJqmuQMHlmCibPpMSIxM95LLp27Ypvv/0WeXl5yM7Otq4CiIxEUJD3XW3E\nM/pyQ53mVFSwZiz+c4VDNs1KQUrP6wAAgEggBAMGPKzLRWViWtVBmo/ycgvKDBVeBxVGg+PKjvNn\ni5B7wx5U6AKk6NYzCEKhAGbneZjgeR6XM07j6gV7fSWFSo6Jj6QgPCoUbWOisP2n3Th1OM32+vRn\nJkKpch3Umy0WBARpoVI7VyGlIIHUR4MMhAUFBVEw0QDKzEbk6gu8zqhZwZqxaP9yl0FFXQgYAcbE\njsCms7/TclHSrJjNLAyl5V7PqSgt0aO4sNRhn7ugwhWe53Ex/RiuZZ237dMGapA6YzwCg62Bv0wh\nx8ixQx0CC4XSOajgOA48GIRHhEAkprkQxPd89rdq0KBBeO211zBq1Cjbtif27t1b+0GkVuXmclwv\nrX1JaXVm1uLUU9EzPA7HrqXVcJZrleO9bXTheHbgv/DswH95fQ1Cblccy6Gk2Oh1fYzigmKUGYwQ\niVyfp9XVHFQAwI6fdzsEFUGhrZD62HiotSpr2zgeQgEDtbbmfBMWloVEIkZQaKBt6IMQX/NZYNG+\nfXsolUqHbdIwKizmOhUUs3AsPj2wEieun7Ht6x3ZHQ/2HIdnfnnTu2uxFsjFMoSqgmi5Gml2OI5H\nUZH3QUVhfhFMRhOEIiEsZuc6HBqtBPG1BBUAkHbU/oyqtAFInTEBaq3185ZjeUilIsgVYhjLnNN8\nVzKbLdBo1dAGqr16D4R4y2eBxYoVK2rcJv7BciyyS657HVRwPIf/99e3OFqlZyIpIh6P9UlFucX9\nxDFXLBYLApU6aGX0gUWaH57nUVxk9DpgLrhZgAqTGcJbPRXXcxxzVChVYsT3DIZQ5HnPgTqgFeIS\nB0N+a4iDY3koVRKIayk9bjZbEBQSYDuPEH/yW1/Yiy++iOPHj7t9/cCBA3j00Uf9dfsWgeM5XC25\nAaHAu19RPM9j2d8/4K+r9v8/Ca27YmbfByDy8lpm1oIwTTAFFaTZKi02ATy8CizycwtQUW6G4FbQ\nkJdrxIUMx4SBcfGtIBK7XwJanTYwGF2ThkAkFlvLrrM8VBppjUEFz/OwWFiERQRTUEEajM96LDiO\ng8VisW1v2LABffr0QZcuXVweu2/fPhw4cMBXt29xeJ7HtZJcADwYN/U6Vh37CTsu7McdHQZgcs97\nbeetPbEJf1w6ZDuuW2gnPNF3sldBBcdz4HkebXThEAloAhhpnvSlJrAc73Faa57nkX+zEJYKe9Ks\n4qJypJ/KdzpWInH9vHEch63rdzrsi4qOQERMHwgreyZ5QKOTOaXjdtWe8KgQr4dwCKkPn30j5OTk\nYMyYMSgvL7flmn/xxRfx4osvuj0nPj7eV7dvcW7o82DhWLcTsMotFfjt/D7w4LHjwn5M6JYMqUiC\nzWd3YUvGHttxnYLa4+n+U70aSuF4a9rfSE242/XxhDR1JmMFKspZj1eA8DyP/NwCWMwW2znGMjNO\nHc8Dx9VenRSwVhrdsu43pB8757B/zKRknDhqryGi1khqDSoAILR1MAUV9VRwswhfL/oOU5+cgKDQ\nwMZuTpPgs8AiMjISr776Kg4dsv4S3rhxI5KSkhAZGel0rEAgQKtWrZCamuqr27coeYYCmCwVNX6p\nWzgW/K1UfBzPwcKxOHTpML4/9avtmDa61nh24L+8SrHNchwkQjHC1cE0SZM0W2azxasEWJVBBWu2\nByLmChYnj+XZJm0yjLVeR03X2P7TLhw7eMrpNbHYcQm5p88eFRCrv9PHzuLk4XT6vPOCT/uwx48f\nj/HjxwMAsrOz8fjjj2PAgAG+vEWLV2Qsgb6izOt5FadzM/DV3+ts26GqVpg96GEoxJ6Pu3I8B5lI\ngjB1sFf3JqQp4TgeJcUmr37pVwYVjJCxXeP0iXwYy+zDwwoFA4PBHlmYKywQie1B/Z6t+3H4j6Ou\n28R61g7W4uGBxGvVq74S9/w2OL5ixQpUVFRg69atGDJkCORy6xfY5s2bAQDJycm0jtpL+vIyFBg9\nLypW1ZeHvgN3q/SytUrpI15NuLRwLJRieZ3rhRDSFFSuAPHms6n68AfP8zh/thDFRfaKowxjQV6u\nEXKl/Zk7fOAK+g1qC7FUhL/2HMG+7Qdtr4VGBONG9k3btkxR+0e1xcxCoVZ43O6abN2wC7t+3Y/8\n3EIEtNJi2N0Dcdf44QCAU0fO4qfVW3HlUg6EQiG69YrFPx8ai8BgHQDgj+0H8fX/1uLlD2dh1efr\ncTkzBxqdCiPvHWK7BgAUFRRj9ZKNOH30LMADfYb0hFKtxIFdf+ODr/8LAJg97Q0kDeyOy5k5uHDm\nEvoNT0THuPb4asG3+ODr/6JViD0r8Oxpb6BLj4545N8PALDOVdm8bif2bDmAgrwiBIUEYOS9QzDy\n3sG2c979v0UIjQhGaHgQdvyyF6XFBrSLiUTqjH8gunMb/LH9IL5a8C0A4Pnpb2HQyN54+Nb1iXt+\nCywKCgrwyCOPIC0tDT/88AO6du0KANi6dSu2bt2K1atXY8mSJQ65L4h75ZYK5Bry61z+vJy1LiGV\niST496CHvAoQWI6FWqJAkJLGF0nzpi+1BgOednsX5hfBXGX1BwDkXNU7VCqVShgUFRqdzjXoy3Eu\nPRc8W4Tfftxl2x8U1gop08bis3e+su2T1LKc1GKxQKNTQ1jLcZ749osfsX3jbiSPH46uCZ2Ree4y\nvv/qZ7Asi4BWWnzx4Wr0G9YLYyaNQmmxHhtXbsHb//4Yry+aA43uVsIunscn73yN5JThmDB9DPZs\nOYDvvvwJUe3C0S0xFuYKC+b/5xNUlJsxeWYK5AoZflm7HZczc6AL1Di057ef9+Ku8cMxZuJISGVS\nXLt6w1WzwTAMqv5fW774e+zd/hfGThqFmLj2OHPiPFYv2YAygxH3pt5pO+fw3uNo3SYMDz4xARzH\nYe0XP2Hx3K/wwdf/Rc8+XTF20ij8/O12PP3qQ4hq37re/31bAr8FFh9//DEuXryId955Bx07drTt\n/+CDD3DnnXfilVdeweLFi/HCCy/4qwnNBstxyCm54XWuiuoEjABP9J2CdgHO815qopbIKaggzZ7J\nWAFzheeTNUuKSmEqM9nyVABAUWE5LpyzLytVayQwlhlcnQ4AuJiRhVN/2XsqKtN0iyWeP+uVQYVG\np4ZBX3sJ9ZoY9GXYvnE3Rt43BPdPHwsAiOvZCSWFpTh38gIuZ2YjPjEWj/3fg7ZzOsa1x0uPzcOW\nH3binw9bV5+BB8ZNTsbgO/taj+nSHn/vP4Hjh9LQLTEWf/5+GNev3sTrC2ejbYz186hLj454fvpb\nTm0KCg3EhH+NsW27Cyyqun41F7u3HsD908fi7gkjAABdEzqDYRj8snY77hgzCEq1AjzPg2M5zHl7\nJmRya+kBk7EcX3y4Glcyc9A2JhLB4dYfYW07RDr0kBD3/DYWsWfPHkydOhUpKSmQSOzjiGKxGPfc\ncw8mT56M7du31/n6169fR1JSkm2yqDsVFRX4/PPPkZycjISEBCQnJ+OTTz6BuVqVn5MnT+LBBx9E\nQkICBg8ejAULFjgd01hySr3PVVE57FHVlJ73oXt4rNf3D1TQw0SaN4uFRZnB7HFQYSg1QF9icAgq\nyk0WpJ3Ms03QlEiF6NojCO46P8pKi3H60CFbzgqFSo7UGdY03Rzr2Xh+1aDCFy6cyQLH8Uga0N1h\n/wOP/QOTZoxDcWEp+g7r5fBaSHgQYrq0w5mT5x32d+jSzvZnkVgEtVaFcpO15zT9eAZCwlvZggoA\nkMml6NknzmkuQ5voCK/fR9rxDIAHevaJA8uytn969u0Gc4UF505n2o5t3SbMFlQAQEArLQDY2kq8\n57cei6KiIoSEhLh9PTw8HLm5uXW69rVr1/Dwww/DYHD/S6DS22+/jZ9//hlPPPEE4uPjcfLkSXzy\nySfIycnB3LlzAQBXrlzB9OnT0atXL/zvf//DhQsXsGDBAhQVFeGNN96oUxt9Jc9QAJbjvF7Wufns\nLoft4dH9MKJDfx+2jJDmged5lBabPA4qTEYTigpLHRJTcRyPtJP5MFfYV4B07d4KUqkQWo0MJUWm\natcw4PzpI2BZ6+ROsUSMiY+kIDA4ABzLQ6GqvXCfr4MKANCXWD9T1S6uabj1mjZA4/SaRqfG5cxs\nh31SqfNqM/5WEFVarHd5j+r7GIaBzMOy7VUDksr38fLM+c4HMkBhfrG9ndWuX7mShiZr1p3fAot2\n7dph586dmDx5ssvX//jjD0RFRXl1TZ7nsWHDBsyfP9+2XZPCwkJ8//33eP755/HQQw8BAPr16wcA\n+PDDDzFnzhwEBARg6dKlUKvV+PTTTyESiTBkyBDIZDK89dZbmDlzJsLDw71qp6/oy40oKTd4PQRy\nOPskfj2322FfStfRHp/vKusfIc2VvqQcgGdzKiwWCwpuFjllu8zMKEJJsf0XbsfYAGi01uCgTj4U\n8gAAIABJREFUU1wICguNtvkbFrMZ6Yf3wlJxaz6HgEHK1DEIjwq1pei21NJbyrIsAnRanwYVgLXX\nBLB+8YdF2Fd/5d8sxJWLOQCsRdWqKy4ogVrj+Xy5gCAdzpw477S/tKjUxdGOKmdSVP+cqtrDUPk+\nXpj3JGSKakEaD1tFWOIffhsKmThxIvbt24f/+7//w6lTp6DX62EwGJCWloZXX30Vu3btwsSJE726\n5pkzZ/D6668jJSUF7733Xq3HGwwGpKamYsSIEQ77KwukXblyBYC1wurQoUMhqlIZdPTo0eA4rtGq\nr7Ich1xDntdBRXbxdSz9a63TfoGb7JzO92UhE1OZc9IylJssMJtZj/I9cByHm9fyIKoWVNzMLUP2\nFb1tO6y1EuER1kmM2zf+jg9eWQRDYRYiorTgOQ5nj/2JMr092dVdE0aiQ5f21p4KRe11PwBArpT7\npZhYh85tIRQKcPSAYy6NX9ftxIaVW6ANUOPA7iMOr+Vey8P5M5fQsWt0jdeuOiE2tnsM8m4UOPRy\nVJRX4MThM7VOnJUrrBVcC27a57LkXLkBQ6l9fknnbh0AWAOkdjFRtn/0JWVYv+JXGPS193ZXotWL\n3vNbj0VqaiouXLiAlStX4qeffnJ6ffLkyZg6dapX12zdujW2b9+O0NBQHDx4sNbjIyMj8d///tdp\n/44dOyAWi9G+fXuYTCbk5OSgXbt2DscEBgZCpVLh4sWLXrXRV66X5npdt8NoNmHhn8ttK0C8xXEc\npEIJ1FJaqUOaP47jYSzzLF8Fz/O4eS3PabjEZLTgXFqBbVupEiOms3XZpbnCjMN7j4HneRw5cBxD\nkgfi8N6tKMqzTz7sP6I3evaNB8fykMlFkEitbREKhWAYBjzPQyBgHOZyAIDu1jwAX1NrVbhz3FBs\n3bALYrEInbt1wPkzl/D7pn2Y9Og4yBUyfLlgDT6fvwIDRiSitMSAjSu3QK1RIjllWI3XrtrD3H9Y\nIjZ9twML3/wSKVPvhkIpx9YNu1BarHfIbumqV7pLz44QS8T4dulG/GPq3TAajNiwcot1MuatY6La\nt0b/EYlYtnAt8m4UoF3HKFy7mosfvt6EkPBWCIsIqfEeVSlu1Vg5vO844pO6oHVUaC3/FYlfizy8\n8sormDRpEnbt2oXs7GywLIuIiAgMHz4cnTp18vp6Wq0WWm39Hqjt27dj48aNmDJlCtRqNW7etK4V\nV6lUTscqlUro9Xqn/bUpK3OcmW00Oi81q0lhWQnMHOvVvAqe5/HF4e9wQ5/n1b0qcTwHkVCEMHUw\nDBX1m1lOyO2ipmfRoC+HRCLz6DqFeYW3vuTtzyTP8Ug/lQ+LxfrFJBQyiOveypatk2VZ25cWz/E4\nfugUrmXZu/87x3fEsLsGgeN4SKRCyGT2peQSqRgDRiThz98Po//wJEgkdVtmXhf/fPheqHVq7Nq8\nD5vX7URwWCs8+OT9GHaXdY6WTCHFprW/YeFbX0GukKF7UhdM+Nc9jsMyLjodqvZECIQCzHl7JlYt\nWY/li7+HUCTEgBFJUGmUuH411+U5lRRKOZ5+dTq+/+oXLHzzSwSHBWLc5GTs2/GXw20fee4B/PLd\nb/h9834U3CyENlCDfsN7IWXq3bbrMgzjuoekyq64nh0R17MTvl/2C9KOncNzb8zw7D9kC+b36lEx\nMTGIiYnx9208sm3bNsyePRtJSUl4/vnnAdQ+n6Au3WAJCQl1ah9gzVdRaCyGWOTd/5ptGX/g72x7\n9+VdnYY6zbNwp7L2R2t1CKWtJc1KTc+ipYKFVFr733dDqQHlpgqnXoNLF0uc5lUoFO4DgF2b/rD9\nWanR4e777wTAQCQEFArnCYpjJ43C2EmjAAAV5Ran1/3prvHDHZJZVZU0sAeSBvZwe+7gUX0xeFRf\np/2VSa8AIDvrOnKu3MDTrzzkcMwbsz5CaOtgl+dUFZ/YBfGJjgUu+1VbrSIQCnBv6p22nBWu/Gf+\nU077unTviGWbFti2pTIpnn/ncbfXIM78GliUl5cjPT0dZWVlDl/gLMtCr9fjzz//xNtvv+3PJth8\n/fXXmD9/Pvr164dPPvnEtgS2sqfC1QoTvV4PtbrhyoHzPI/rpTe9DioyCy7ju5ObbdvdQjvh7s7D\nPAoseJ4HeCBSG2YLKkRCEQQMA47nIWAEENUzfwYhtyOBqPYhEHOFGcUFJRBVyytRXFSOyxft8yRC\nwxUIDa95CLHyM1AskaJLr4EQicUAAyg9WAHCejihWiQSgREw4Dlr74rIy8+ShmIymvDpu19jxD2D\nkDggHhzL4eCeo8g6fxUTK3NhkCbLb3/rzp07h+nTpyM/37lcsO3mIpHfAwue5zF37lysXLkSY8aM\nwbx58xweNqVSidDQUGRlZTmcl5+fD4PBgA4dOnh9z6NHHfP9G41Gj2qm3DQU1HpMdYYKIz49sArs\nrbwVAXItHuuT6vFkTQvHoo0u3OF4mUiKcXF3YmPadoyLGwWZiCZzkqaprs8iYA0E8nPznYIK1sLh\nzGn7syqTCxHT2bNVBgzDILbXAEjl1uRMao201l5Cs9mC4LBWALJqPA6wLp0cOXYwdvy8F3eMHeS0\nlPJ20SG2HZ548V/4dd1O7N9xCDzPo21MJGa//Rhiu98ePdyk7vwWWHz00UcoKSnBjBkzwDAMlixZ\ngv/+978oKSnB+vXrIZPJsHat8+oFf7Rj5cqVmD59utssnwMHDsSuXbvwn//8x9aTsXXrVgiFQtvy\nVG8oFN7n6y8zG6GvMHiVspvneXxz5AfklRUCsK78mNnnATAAVh9znDBrqCiDUuJYcMxssaC1Jhgi\ngfNfgyf7TcGT/aZ4/T4IuZ3U5VmsVJhXBFfZrTLPF8FktA9NxHZtBZHIs0A+Oi4BmoAga9tUtZc+\nN5stCAoJgFgisVVHZRhAWMMqlskzUzB5ZopH7WlMvQf1QO9B7odUSNPlt3U0R48exf33349///vf\nmDlzJoRCIdq1a4eZM2di3bp10Ov12Lhxo0/vqdfrcezYMRQUWH9NpKenY+nSpYiPj0dycjKOHTvm\n8E/lxMxHHnkE+fn5ePTRR/H7779j2bJlmDdvHiZOnIiwsDCfttEVjudwozTP6zoge7P+xl9XT9i2\nx8WNQoQmBPN2L8G+y387HPvxvmXQl9uHeyysBa2UOsjEnk1eI6QlMZQaUFFe4TTHqrDAhJyr9ueo\nTTs1tDrXPXpnT2Y4bMclxCKsjb0HVFxLQi6LxQKtTg25Ug6RWIiuPcKtibd6hDsteSXkduK3wMJg\nMKBLF+vkGrlcjoiICKSlpQEANBoNUlJS8O2339brHtW7EE+fPo1JkyZhz549AKyTNQHg1KlTmDhx\nIiZNmmT7JzU1Fenp6QCA6OhofPnllzCZTJg1axa++eYbTJ8+HS+//HK92uepXH2+15NEc/X5WHXM\nHpjFBkdjTOxwfH9qC66WXHc6Pqc0F9+f2gLAOl6rlCi8qm5KSEthsVhQXFTqNFnTYuFwttrS0rbR\nrlepFdwsxLYNuxz2DbtrkMdtYFkOcoUMmgD7M9p/SDQeeXog+g+pOV9EY8nOuoa3Z//PYd/0u5/D\nj6u2+P3eN2/kY/rdz2Hv9r/8fi9SO78Nheh0OoelmpGRkcjIsEfwoaGh9coR0bdvX1tgUHXfmTNn\nbNuzZs3CrFmzPLpeUlJSgwzNVFdmNqLMbPIqERbHc1h6aC1MFuuMdKVYjhm9J0HACHClKNvteVeK\nsm3r4oOpqBghTnieR/6NAohcTOy8dKEY5SYWgHU4onNcIAQuhiQsZgs2LP8F5grH7JlCDydSchwP\noVDQ5LJD/vXHMVxIv+Sw79UFzyIgSNdgbaBFbbcHv/VY9O7dG+vWrUNenjWvQmxsLP7880/buvIj\nR47UOydFU8fzPG7o873Orrnl3B5k5F+ybU9LHI9AhWcPr4VjEaEOpWWlhLhQUljiMmFSSXG5Q3bN\nqHYaqDWuJ0bu+HkPbuTcdNrPc57VnuA4DiHhQc3iGY3u3NZW1Iu0HH7rsXjssccwadIkjBgxArt3\n78bEiRPxzTffYNy4cQgNDcWhQ4fwz3/+01+3bxIKygoh8LBGQaWrxdex/vRW23a/qJ7oE2mvRBil\ni0Bm4VWX50ZowxCuDva6UiohLUFFeQUMBqNTbwXH8TiXVmjbVihFaNveuRAXAGSkZeLvfcdcvqZQ\n1j6HymJmERIR5HFBtIZ2KeMK1n75Ey6dvwKe49Ehti1Spt6NE4fT8dNq69Dz9Lufw7jJo3Hf5GRM\nv/s53Dd5NMZNTkb6iQy8959P8fw7j+OnNduQeTYLaq0K9z4wGj16d8GKT37A6aNnoVDJkZwyHHeO\nGwoA+GP7QXy14Ft88PV/HcqWz572Brr06IhH/v2AUzs3rPwVP63ehmWbFzjsr9oeADiw629s+m4H\nbuTchFQmRbfEWEx8eCx0gRQM1Yff/vbGxsZi3bp1GD9+PAICAtC2bVt8+OGHYFkWaWlpuOeee2xJ\nqloiC2dBcbnBq7kVLMfiy8Pfw8JZu2N1Mg0eTBjncMz93ZIRqXGecBquCsb0XhMgp8mahDjheR6F\neYUuh0CuXi6FwWAf1ujUxfUQiL7EgE1r7UG/dYmonatrV2Uxs9AFaRo0y6Y3jAYTPnzlc2h0Kjz9\nykN4/D/TUG6qwAevfI4ho/th8GhrUqxXFzyLIcn2SsrV/0t9Pn85Evp1w3NvzEBYRAiWL/oO8//z\nCSLbh2PW648gulNbrPl/G5F59nKN7WEYpuafZW5erNydcToTSz9Yhd6De2L2WzOROmMc0o+dw2fz\nltd4X1I7v2ZPiYmJwWuvvWbbTk5ORnJysj9v2WTc0Od7XQtkW8ZeXCy8YtuenjgeSonjcjqVVIn/\nDH0Ma47/4rAyZM7gR9A2IKJ+jSakmSopKrUu5ay232SyICvTnggrPELpchUIz/PY9N02lOmtKcNF\nIiHu/ucofLPQswnqLMdDrpRBpb596/RkX74OfWkZRt03BDFdrIUcw6NCsHvLnxAwjG3II7pz2xqv\nM2R0P4z+xzAA1rwbbz33MaI7t8U/ptwFwFrn4+/9J3DhzEVEd25T9wbXMvJ09nQmJFIJ7p4wAiKx\n9atQpVHiUsaVmk8ktfJ7WrazZ89i165dyMnJwdSpU6FQKJCRkYEhQ4b4+9a3rTKzEeWWcq+Wl97Q\n5zkMgQxsm4ge4V1cHquSKvFAz3sdAovoQO9K1BPSUrAWFmWlZRC6WMJ54VwRuFtzI8RiAdrHuO4i\nP37wFC6k2yej33HvULQK8WyCtHVCNRAY3HCTHOsiqn041FolFry2FH0G90S3xFh069UZ908f69V1\nKoMSAFDrrJmPO1QJRiqDq8ogzV9i42Ow/pvNePnx+eg9qAfik7qgW6/O6J7k+nOVeM6vgcXcuXOx\nYsUKANZuq+TkZOj1ejz99NMYNmwYFi5caEtI1ZLc1Bd4nQhr+ZENMHPWpDwaqQqpPbx7mD3NxElI\nS5OfV+gyqCjMNyEv1/7l1j5G67KkeVF+MX77aZdtOzq2HXoN6AFjmcmj+1vMLFq3vf0nVEtlUrz0\n/jP46dtt+GvPMez69U+IJWIMvCPJq4RclWXPHa/d8N8DMV3a4d9vzsCWDbuwZf0ubPpuBzQ6NcZO\nGoWR9w5u8PY0J34LLFatWoUVK1Zg2rRpGDVqFKZMsWZx7NOnDx588EGsWLECX3zxBZ544gl/NeG2\nVGwqBc94Nju80p9XjuJ0rn2p7pSe90ElqXtGQUKIlbHMCNZsccpZwXM8Lpwrsm2rtRKEtXYepuA5\n6xBIRbl1DoZMLsU9/7QWF6spO2Yli5lFYIjOo9Ltt4OwyBDMmDMFPM/jwpks7N95GL9v2oeQ8CC/\n3bNyJkX1gpEmY7n7c24FaTzP2/7s6vhuibHolhiLigoz0o6dw/aNe7Dq8/XoENsW7TvVYximhfPb\nz9hvv/0Wd9xxB1588UVER9sTumi1Wrz88ssYM2YMfv75Z3/d/rZVZCzxalWGoaIMa47b/zt1D4tF\n7yqrQNxysWSOEGLH8zyKC4qdggoAuJZjcJiw2bFzgMsehaMHTyLrvH1MfvT4O6DWqsBzPBTKmn+F\ncxwHmUIKpapp/Eg4euAUnpr4MooLS8EwDGK6tMPUJydAoZQj/2ah34Kjyh6Ogpv2QC/nyg0YSsu8\nOufc6UyHY75f9jPemPURAEAiEaNnn66Y+Ii1AFr+zUKQuvNbYHHp0iUMGuQ+01zv3r2Rne0+mVNz\nxXjwK6aq9ae3ofRWKm6JUIwHE8Z51GVqYRu2zDIhTU1piR6ulg5YLBwuXSi2bYeGK1zmrCguLMHO\nn/fYtjvHxyCuZ2dwHA+pTFTrii8et/+8iqo6xrUHwzBY+OYXOPLnSaQdO4evF66FscyEpIE9oFBa\naxEd2PU3bl53X3zSW7E9YiCWiPHt0o04cTgdB3cfwcI3v4RSrXA7P7NHn64AgGUL1+L00bP4Y9tB\nLF/8PWRy+8TbbomxyDp/FUs/XIVTR87i2F+nseqz9VCpFejSo6PP2t8S+S2wUKvVNVY2zcrKatCS\n5LcLb+Y6XC7Kwc4Lf9q27+0y0qOMmRaORaCy6XxgEdIYDKVlLvNFXLlUArPZ2u0uEDJoH+P8LPE8\nj60/7EBFuTX7rUwuxeiUO8AwDAQMA7m89jlUQSEBXqfyb0wqjRJz5s6EUqXAso+/xYLXliLr/FU8\n+fK/ENs9BkmDuqN9pyh88eFq/PrD764vUodpJEqVAk+/Oh0sy2Hhm19i46otGDc5Ge07Rbm9XFhE\nMB6dMxl5Nwqw4LWl+O2nP/DQrEnQtbLnH+nSvSMef3EacrKuY/HbX2HJ/BWQyqV4Yf5TTaYX6Xbl\ntzkWQ4cOxZo1a5CSkgK53LGq5vHjx7FmzRqMGjXKX7dv8niex8qjG8HfisnDVEEY3bH2CUU8z0Mi\nEEEjVfm7iYQ0aUKh89dSeTmLq5erZNhsq4ZU6tzFf+ZEBs5XWQUy8r5hUGmU4Cw8lBrXRcmqE0ub\n3sT1th0i8e+3HnP5mi5Qi/9+/G+HfVUTVHXp3hHLNjkmrAoObeWUxKr6eQAQn9gF8YmOqzX6DetV\n43UGjEjCgBFJDvve/X8vOWxThVX/8Ftg8eyzz2L//v247777kJCQAABYuXIlvvzyS+zfvx9arRbP\nPPOMv27f5B26egLnqqTtntzzPo9Sf1s4Fq11ITCZ3U9sIoQ4FzEEgKzMYvvyUokAUW2ce1VNxnJs\n22D/Rd6uYxvEJ8WB563niN2UUK86DFpb6XNCmjK/9cOFhoZi3bp1GDlyJI4cOQIA2LFjB/7++2/c\neeed+O677xAZGemv2zdpFawZ353cbNtOCI9DfFjnWs9jORYBMi1EAr+nJyGk2SkrM+Najr0ketv2\nGghdBAm7f90HQ6n1OKFIiOQJ1iGQqhM2DXojflq7zeG8ivJytG2vodLnpNnz6zdQcHAw3n33Xcyd\nOxeFhYVgWRaBgYEQeVjlr6XanrEXeWXWWclCRoCJ3e/x6DwBI0CAwnUNA0JIzbIyS2zZGmVyEcIj\nnIcTb2Tn4sj+47btQaP6ITAoABzHQ64QgWEYGPRG/L8PVuJGtmMhsmUL1+K51x/FnWPj/fo+CGls\nDTJzSCAQoFWrVggJCbEFFZcvX8a0adMa4vZNSmm5AT+f2WnbvqPDAISpg2s9r4I1I1TZqtbjCCHO\nygxm5F63L19sF61xqgfC8zy2rt9pq34aGKRD32GJAAABw0AqtU7Y3LLhd6egAgDyrhdg28bd/noL\nhNw2fNp1cOLECXz66ac4evQoACAuLg5PP/00evWyT7LhOA7Lli3DokWLYDJ5lpmuJfk5fQdMFuv8\nCKVYjnvjRtZ6DsdzUIkVkIo9mzRGCHFUtR6IXCFCSKjzqoBTf6fj6qUc2/ad/xgBkUgEjuWhVNuf\nvWuXb7i9z+XMHLevEdJc+KzH4tChQ3jggQewe/duBAQEICIiAn/99RemTp2Kw4cPAwCys7ORmpqK\n999/HyKRCG+88Yavbt8s5BkKsTPTvrx0TOwIjzJsshyHYJVndQkIIY7KDGbk3nDsraieb6ai3Izf\nN/1h2+4UH4Po2HbgeUAkYtxO2CSkJfLZ0/D5559DLBZjxYoV2LJlC9avX4/ffvsN0dHReP/993Hx\n4kVMnDgRx48fx9ChQ7Fp0yZMnDjRV7dvFjakbbOVRA+Ua3FHzIBaz2E5Fq0UWqoFQkgdXb5k761Q\nKEQIdtFbceD3Q9CX2Cdsjhw7FIA1pbdC5dhTGN4m1O292nagCsOk+fPZt1FaWhomTZqEpCT7uuHw\n8HDMmTMHJ06cwHPPPQeDwYC3334bS5YsQWio+4evJcopycX+rCO27XFxd0LiQaEyBgy0MpqwSUhd\nmEwWh7kVUe00TstQSwpLcWDXYdt236GJ0LXSguMAiVQEQbXjk/8xHKERzvOiWkeFYsK/xvj4HRBy\n+/FZYFFaWoqOHZ3ToHbs2BE8z+PatWtYs2YNJkyY4KtbNis/pf9mT4alDsbAtr1qOQMws2YaAiGk\nHq5mldrK6khlQoSEOfdW7Nm6HxazNUW+Uq1A/xF9AFiTSMrlztPUlCo5ZsyZgl4Dujnsf/aNR6HS\nOBcyI6S58VlgYbFYIBY7/8KWSq3dhI888ghiY2N9dbtmJbv4Og5esS9hG9dlpEeFyqRCCRRiucvX\nREL7LykBI4DIg+RahLQk5goO17LteSui2qqdVoLkXruJE4dP27aHJg+EVCYBx/GQyYRu6/YoVXIk\np4xw3KemNNGkZWiwgfmuXbs21K2anB+r9Fa0VoegT1TtKWbNFjNCVO5LFctEUoyLuxMCRoBxcaMg\nE9GKEUKqupZjsGfZFAtclkXftWmvLbdFq9BAdO9t/RxjGEAqcz9UyXE85HKZ7xtNSBPQYD9jPanI\n2RJdK83Foasnbdv3xY2sdSImx3NQShS1pvh+st8UPNlvik/aSUhzcy1bD8AaHLSOVEFYrSDZ5cyr\nDvVAht89GAKh4FbQUEtJdJZFQCutz9tMSFPg08DiwoULOHTokMO+khLrjOuzZ8+6zLjZu3dvXzah\nydl8dpettyJcHYLekd1rPYdlWQRpAvzdNEKatYoKDhIxwAisgUVVPM9j96/7bNuR7VqjY9doAJXJ\nsNwPVbIsB41ODcZFkTNCWgKfBhaff/45Pv/8c5evzZs3z2kfwzBIT0/3ZROalPyyQoeVIPd0HlZr\nbwXLsdDKVB7NwSCE1C40TAlJtUDhUsZlXMnMtm0PvXsgGIaxJsNS1V6VVK1Tocxg9HlbCWkKfBZY\nPPnkk16fU5/hkevXr2PMmDH47LPPPO71SEtLw/3334/t27ejdevWDq8NGTIEubm5TuccOHAAOp2u\nzu2syZZze8DyHACglUKHfm0Saj2H54EAhX/aQ0hLFNHGubdiz5b9tu12HdugbYcoAIBAwEBcQ/Ew\ni8WCwKAAGvolLZrPAounn37aV5eq1bVr1/Dwww/DYDDUfvAt586dw4wZM8BxnNNrBQUFyM3NxQsv\nvIDExESH19Rq57LJvmCoKMPui3/Ztu/qNBSiWnohWI6DTqahZFiE+IhWJ4WqWg9E5tlLyM66Ztse\nkmxNVOdJb4VQLIJC5XqlFiEtRZNag8jzPDZs2ID58+fbtmtjNpuxYsUKLFq0CBKJxOU5Z86cAQCM\nHDkSUVFRvm20G7syD6KCNQMAVBIFBrfzpNeFh07un0CHkJaodZRzb8W+7Qdt29Gx7RDZztq7WVtv\nRUWFGWEuEmMR0tI0qZ++Z86cweuvv46UlBS89957Hp2ze/dufPLJJ5g5cybmzJnj8pj09HQolcoG\nCyosnAXbz9snhg2P7g+pqOZfQizHQSt1zgpICKkbsUSAoGDH3oXLmVcdCo0NGtUPAG7lrXC/vJTn\necjkUkiktc+/IKS5a1I9Fq1bt8b27dsRGhqKgwcP1n4CgPj4ePz+++/QaDRYv369y2PS09Oh1Wrx\nzDPPYP/+/WBZFsOGDcNLL72E4GDf/wL56+oJFJmsq2VEAqFHNUF46q0gxKfCwpVOCbGq9la0jYmy\n9VaAgdMEz6rMFguCw1r5pZ2ENDVNKrDQarXQar1bG+5JTZKzZ88iNzcX3bp1w7Rp03DhwgUsXLgQ\nU6ZMwcaNGyGXezdmWlZW5rBtNDrODt+eYe+t6BeVAJ2s5oCB5TjoqLeCEK/V9CxWT4iVc/k6LmVc\ntm0PHNUXgHXCtFTq/qOS563JsETiJvVxSojf0JMA4O2334ZEIkGXLl0AAImJiYiJicEDDzyAjRs3\nIjU11avrJSS4X91xqfAqLhZesW2P6jjIo2tSbwUh3qvpWZTJHHsgqhYai2gbblsJwnM8ZDUEFhYz\ni+Bw91lwCWlpKLAA0KOHcwrtXr16Qa1W4+zZsz691+5M+0qQjq3aoa2udQ1HAxzHQSNVUW8FIX5U\nmF+EsycybNv9hieBYRjwvHUIxN3zx/M85CoZRCLKK0NIpRYfWOj1emzZsgU9evRwqM7KcRzMZjMC\nA72vHnr06FGHbaPRiAEDrPMoDl09YfuvPqJD/1qvxfE8AuRUFp2QuqjpWazqr91HbCvGAoJ06Ni1\nAwCAZznINO7r7FBvBSHOmtSqEH8Qi8V46623sGTJEof9O3fuhMlkQt++fb2+pkKhcPin6hwNM2ct\nv6yRqpAUEV/jdTieg0aqpN4KQuqopmexkrHMhBOHTtm2+w5NhEBg/WgUigW2KsHV8TwPuZJ6Kwip\nrln1WOj1epw/fx5t2rTxuKdBKpVixowZWLRoEYKCgjBkyBCcO3cOixcvxsiRI+sUWHhiSLvetRYR\nYzkOOjkVMiLEn44fPAVzhTXglyvliO8dB6D2hFhmi4V6KwhxoUkHFtV/yZ8+fRrTpk3DvHnzMG7c\nOI/OAYAnnngCgYGBWL16NdasWYOAgACkpqb6NZvo0OiaAxae5yEXSSEUtPhOJUL8huN6Ej/DAAAg\nAElEQVQ4/L3/mG07oV88xOJb+SoEcJsQy7YShHorCHHSZAOLvn37OhUw69u3ry2LpispKSlISUlx\n2s8wDFJTU71e/VFXcSExCFbW3KNiZi0IU1MWP0L86XxaJooLrDllGAGDXgOsE7l5DpBKalgJYmEp\nbwUhbtDP4UbgSfpuqUgCidB9pj9CSP0d3mvvrejcLQYanXVZN8/zkMrdBxYSqZjyVhDiBgUWDUwh\nliMxoluNx5hZC3QyWglCiD8V5BU5JMRKGmTPeSESMW4nbVrMFugC6fkkxB0KLBpYn6getfZECBgG\nKqmigVpESMt04i/7SpDgsFaIio4AYJ20Ka2hLohILIJU5n4JKiEtHQUWDWxM5+E1vs7xHNQSZY3H\nEELq79QR+3ysnv262yZ2MzVM2rSwrG24pDYikQjMrVokAoEAIhENnZCWgQKLBqaW1Rw0WDiWlpgS\n0gCMBmvdEJFIiG6J1nT+PAeIa5g7wYCBQuVZ7SCpTIKRYwdDIBDgjrGDIJVR5VPSMlAIfZtRiGS0\nxJSQBtSlZ2fIFTIAAM9xkMldD3OwLAel2rshyskzUzB5pvNKNEKaM/oGu42YWQt0lL6bkAbVs689\nA25NmTY5noNGq2qoZhHSZFFgcRsRMgLIxbLGbgYhLUZgkA6R7a2FADmOh8RN7gqe5yGTSSEQ0kcm\nIbWhp+Q2wfEcVFKatElIQ4rv3dUhG69E4nrSptlsgS6Q5j4R4gkKLG4TFo6l3BWENLBuvbrY/iwS\nMm4L/kkkYohryMRJCLGjwOI2IadJm4Q0qDbRkdDeSnRVU+4Ki4WFSku9iYR4ir7JbgNm1gKtjCaF\nEdKQuvaKtf2ZYRi3uSsYBlCqKGEdIZ6iwOI2wICBQuzZ2nhCiG90ju9o+7NI5GYlCMdDoZS7HSIh\nhDijwKKR8TwPhVhGH1yENLDKORMcy0Micz1/wsJaoPYw0yYhxIoCi0Zm4SzQyeiDi5BGwwBiketh\nEIlEDJGb1wghrlFg0ciEjBBSMRU0IqQx8DwgFrv+GLRYWKg1NPeJEG9RYNGIeJ6HUkJzKwhpLDzL\nQSJ1v4zU07oghBA7CiwakZm1QEPDIIQ0HgHjchiE53nIFVKa+0RIHVBg0YhEAiEkQtdr5wkh/sXz\ngFjkfhhEo6OEdYTUBQUWjYTjORoGIaQR8Zz71SBCoZAybRJSRxRYNBILy0JLKbwJaTQM43oYhOM4\nmltBSD1QYNFIxAIRxEL6RURIY3H3+LEWDmpK4U1InVFg0Qg4noNCQuXRCWlMEonr+U0SqRhCIeWu\nIKSuKLBoBDQMQkjjE7mYuMmyHJQaqgtCSH1QYNEIRAIhDYMQ0shcLSXleA4KJc2vIKQ+KLBoYJQU\ni5Dbl0QihkBAH4uE1AfD8zzf2I1o7nieh9FoBADI5VQpkZDGQs8iIf5HgQUhhBBCfIb6/AghhBDi\nMxRYEEIIIcRnKLAghBBCiM9QYEEIIYQQn6HAghBCCCE+Q4EFIYQQQnyGAgtCCCGE+AwFFoQQQgjx\nGQosCCGEEOIzFFgQQgghxGcosCCEEEKIz1BgQQghhBCfocCCEEIIIT5DgQUhhBBCfEbU2A1oCXie\nh9FoBADI5XIwDNPILSKkZaJnkRD/a1Y9Fk899RRGjBjhsC8rKwszZ85E79690a9fP7z++uvQ6/UO\nxxgMBrzxxhsYNGgQEhISMGPGDFy8eNFn7TIajUhISEBCQoLtQ40Q0vDoWSTE/5pNYPHjjz/it99+\nc/gFUlJSgmnTpqGgoADz58/H7NmzsXnzZjz77LMO586ePRvbtm3DnDlz8N577+HGjRuYOnUqSkpK\nGvptEEIIIU1asxgKuXHjBubOnYuwsDCH/WvWrEFxcTE2btwInU4HAAgNDcWMGTNw5MgR9OrVC0eP\nHsWuXbuwdOlSDB48GACQmJiIO+64A6tXr8bMmTMb/P0QQgghTVWz6LF45ZVXMHjwYPTv3x88z9v2\n7927F0lJSbagAgAGDhwIpVKJPXv22I5RKBQYNGiQ7ZjAwED06dMHu3fvbrg3QQghhDQDTb7H4vvv\nv0d6ejp++eUXzJs3z+G1Cxcu4J577nHYJxQKERkZaZtDceHCBURFRTlN4oqKisIvv/zi38b72Mvb\nPnTaN/fO2R4f64q78wkh7m39Kc1p3+h74zw+1hV35xNyu2nSgUV2djbmzZuHefPmOfRKVNLr9VCp\nVE77FQqFbQJnaWmpy2OUSqXTJE9PlZWVOWzTJDFCGgc9i4Q0vCYbWPA8j5deegnDhg3DqFGjXB7D\ncZzb8wUCge06tR3jrYSEhDqdRwjxLXoWCWl4TTawWLVqFc6dO4cPP/wQFosFgD1IYFkWDMNArVbD\nYDA4navX6xEeHg4AUKlUyM/PdzrGYDBAo9H48R00bfUddqEhFkJ8o77DLjTEQnytyQYWW7duRWFh\nocOky0pdu3bFU089hfbt2yMrK8vhNZZlcfXqVYwePRoA0L59e+zbt8/pGllZWYiOjq5T244ePeqw\nbTQaMWDAgDpdi7ifD0LBCakNPYu+RYEJ8USTDSzefPNNh/FTnuexePFinD59Gp9//jmCg4PBMAy+\n+OILFBQUIDAwEIB1FYjRaLQFJIMHD8aSJUuwZ88eDBkyBABQUFCAw4cP4/HHH69T2xQKRT3fHSHE\nF+hZJKThNVhgUVRUhPPnzyMpKQkAsH//fnzzzTcQCoWYMmWK178i2rdv77RPp9NBLBaja9euAIDU\n1FSsXLkSDz30EJ566ikUFhbi/fffx9ChQ9GzZ08AQFJSEvr06YPnn38ec+bMgU6nw6JFi6DVapGa\nmlrPd03c8XRVCiHEv9z1Qni6WoWQ6hoksMjIyMCUKVMQEhKCn3/+GZcuXcKMGTMAAGKxGLt378bS\npUvr3UXJMIzDstHAwEAsX74c77zzDubMmQOlUom77roLL7zwgsN5ixcvxrvvvov3338fHMchMTER\nCxcuhFqtrld7CCGEkJamQQKLhQsXQigUYs6cOQCAdevWwWKxYNWqVYiLi8PUqVOxZMmSegcW7777\nrtO+jh07YtmyZTWep9FoXJ57u/Lm135z7hmgSaGksdHESavm/N6I9xok8+ahQ4cwbdo0DB06FACw\na9cutGnTBomJiZDL5bjvvvtw8uTJhmgKIYQQQvyoQQILo9FomzyZnZ2N8+fPO6zmqD6EQQghhJCm\nqUGGQlq3bo0jR47g/vvvx4YNGwAAw4YNA2BNYrVt2zZERkY2RFNIHTTn4RRCmgqaTEmaigYJLMaN\nG4cFCxbg1KlTyMzMREREBAYMGICMjAy88MILSEtLw6uvvtoQTSF+1thBCM27IOT2mPNwO7SBNI4G\nCSwee+wxMAyDH3/8Ef369cOLL74IkUgEjuNw9epVPPPMM5g8eXJDNIUQQgghftRgeSxmzJhhW2Ja\nqVOnTti3bx/EYnFDNYMQQgghftQgkzfdYRiGggpCCCGkGfFLj4Wr+h2e2Lt3r49bQgghhJCG5JfA\nwlW6bUIIIYQ0f34JLFasWOGPyxJCCCHkNteocyyqunbtWmM3gRBCCCH11CCrQniexzfffIM//vgD\nZWVl4Hne9prFYoFer8fly5eRlkYJYIh/UH4LQhqfuyRflN+ieWmQwGLp0qX46KOPIJFIoFQqUVhY\niLCwMBQVFcFkMiEiIsJpKSohhBBCmp4GGQrZsGEDYmNjsX//fqxduxaAdR7G4cOH8eqrr6KoqAgp\nKSkN0RRCCCGE+FGDBBbZ2dm47777oFKp0KZNG2g0Ghw+fBgikQiTJ0/GoEGD8L///a8hmkIIIYQQ\nP2qQoRChUAiVSmXbjoqKwrlz52zb/fv3x8cff9wQTSGEEL9L+3uf0764xIGN0BJCGl6DBBbt2rXD\n6dOncf/999u209PTba+bTCaUl5c3RFMIsaEJneR25U1g4q8gpiGDIypY1rw0yFDImDFjsHbtWsyd\nOxdGoxFDhgzBwYMHsXr1ahw8eBDLly9HdHR0QzSFEEIIIX7UID0W06ZNQ0ZGBlatWoXZs2fjnnvu\nwapVq/Dmm29aGyES4bXXXmuIphBCCKnCVc/E/2fvvsPjqM7Fj39nZruqVS0XyQ0XbIrBBtsQMGA6\nBFNCCZcQMMHXxEBiQskvCZeEFAjcBC4tQEKH3EvoYDoYE5oxtig27sYqtiRLVt2+M3N+f4y01qrL\n2iLJ5/M8euSdnXLW9u68e8573iN7C6SBSEpgYbPZuO222/jFL36By+UC4KmnnuL111+nsbGRo48+\nmkmTJiWjKZLUq/4MkcjhFGl/kswgpD/DI3IoZXBJ2rLpAHl5edE/OxwOFi5cmMzLD0ld3bgkSUq+\n7oo79ZXsGZD2FwkJLK677jouvvhiDjvsMACWLVuGoii9Hvff/y1vopIkSZI0lCUksFi+fDnz58+P\nBhavv/56n46TgYU0HMjhEWmw6qrXZDiTQySpkZDAYuPGjT0+liRJkiRpeEpKjsXixYs57bTTOOus\ns5JxOUmKu4H2QnSXKyN7Moa2eORNDIdehFTXvBjo8bIXI76SElh8+umnzJ8/PxmXkiRJGpaGQwAi\n7R+SUiBr0qRJrF+/PhmXkiRJkiQphZLSY/HDH/6QP/zhD3z33XfMmTOH3NxcNE3rtN8FF1yQjOZI\nkiRJkpQgSQksfv3rXwOwZs0a1qxZ0+U+iqLIwEKSJEmShriEBBaXXHIJV111FXPnzgXg8ccfT8Rl\nJEmSpEFKrvC6/0pIYLF69Wrq6uqij4888shEXEaShjxZ82LokMmTw5ucLRI/SS3pLXXvy6oNvLs1\n9oNrTNbIFLVGGipkYBJ/677aRX2tL2bbuEm5KWrNvpHlw5NPBiZ7ycBikAjpIRoCTTHbMp3pKWqN\nlEoyWEitcFAn4I/EbNMjZopaI3WU6poZ+2uw0B8JCyzeeecdysrK+nXM0qVLE9QaSRr65IJ00lDX\n3XDSUMq9GGiBrv1BwgKLt99+m7fffrtfx8jAQpIkSZKGtoQFFosXL2bevHmJOr0kSZIk9UjOTEmN\nhAUWkyZNkrNBJClFZJ6GNNQNhwTU7oZNhtrr6C+ZvClJkiQNaYlaDE72buwbGVhI0n5C9mJI0sDF\nIwF1uM82SUhgsXDhQsaOHZuIUyfURx99xF//+le2bdtGbm4uF198MZdffnmqmyVJkjSsyeJjw0tC\nAovbbrstEadNqC+//JL//M//5PTTT+fnP/85X3zxBXfccQe6rnPllVemunmSJCWJoQdRlFDMNj0S\nTFFrpHjrz7BJfwIeOZSylxwKaXXPPfcwffp0br/9dgCOPvpodF3nwQcf5NJLL8XpdKa4hZIkJcOu\nHdvZvrkiZlt29uwUtUYajoZDYmpPZGABhMNhPv/8c6655pqY7SeddBJ///vfWbNmjZw6Kw1bMvdC\nklJvOOVdyMACqKioIBKJMH78+JjtJSUlAHz33XcysJAkSZL6ZX/NHZGBBdDS0gJAWlpazPa2xz6f\nr9MxPfH7/d0+DgQCXR4TCgTRQ7HrE0SC4U7bks3v9/e5Df3ZN5kS9RqS+XqT/Ro6/h+OB7fbjaIo\ncT9vT/blvRgOh4jo4ZhtwVCw07aerpmIfQcqUe2SryE+vlq1otO27y0YF/frJON9qAghREKvMASs\nXbuWH/7whzz66KPMnTs3ul3XdWbMmMF1113HT37ykz6fb8qUKYlopiQNaaWlpXg8nqReU74XJSlW\nMt6HakLPPkRkZGQAnXsmvF5vzPOSJO277noIJElKnmS8D+VQCFBcXIymaZSXl8dsb3s8ceLEfp2v\ntLQ05rHP5+Poo48G4OOPP076t7aBCgQC0RyTTz75BLfbneIW9d1QbjsMr/a7XK6kX384vReH0/8F\n2f7kSvb7UAYWgNPpZNasWbz99tsxBbHeeustMjMzOfjgg/t1vp4+rDwez5D6MOvI7XYP2fYP5bbD\n0G9/svMrYPi+F4f6/wXZ/tRJxvtQDoW0WrJkCV999RXXXnstK1eu5K677uKRRx5h8eLFsoaFJEmS\nJPWRDCxazZkzh3vuuYfvvvuOpUuXsnz5cm688UYWLVqU6qZJkiRJ0pAhh0LaWbBgAQsWLEh1MyRJ\nkiRpyJLTTSVJkiRJihs5FCJJkiRJUtzIwEKSJEmSpLiRgYUkSZIkSXEjAwtJkiRJkuJGBhaSJEmS\nJMWNDCwkSZIkSYobGVhIkiRJkhQ3MrCQJEmSJCluZGAhSZIkSVLcyMBCkiRJkqS4kWuFJIEQgkAg\nAFjL7aZi+WhJkuR7UZKSQfZYJEEgEGDmzJnMnDkz+qEmSVLyyfeiJCWeDCwkSZIkSYqbuA6FeL1e\nXnrpJd5//302bdpEY2MjDoeDsWPHMm/ePC688ELGjRsXz0tKkjSEVVZWsmDBgi6fU1UVp9NJfn4+\n06dP5/vf/z7HHXdcklsoSVJ/xS2wWLFiBb/85S9pbGwEIDs7mylTptDU1MS2bdvYvHkzTz31FFdf\nfTWLFy+O12UlSRomJk+eTEZGRvSxYRh4vV4qKiooLy/njTfe4KijjuLuu+8mPT09hS2VJKkncQks\nHnnkEf785z8DcNppp/HTn/6UiRMnRp+vra3lgQce4JlnnuGvf/0rwWCQa6+9Nh6XliRpGFAUhd/8\n5jfMnj2703PhcJhXXnmFP/3pT3z88ccsWbKEf/zjHzgcjhS0VJKk3gw4x+KLL77gzjvvBGDp0qX8\n5S9/iQkqAPLz87n55pu56qqrAHjwwQdZv379QC8tSdJ+wOFwcN555/Hggw+iaRqrV6/miSeeSHWz\nJEnqxoACCyEEN998M6Zpcuihh7J06dIe91+yZAlFRUUIIXjssccGcmlJkvYzs2bN4oILLgDg4Ycf\nJhQKpbhFkiR1ZUCBxZo1a9i+fTsAP/nJT3rd326388c//pFHHnmEW2+9dSCX7pPq6mpmzZrF6tWr\ne9yvrKyMqVOndvo588wzE95GSZL6ri2waG5u5osvvkhxayRJ6sqAciw++eQT6yQ2G3PmzOnTMXPn\nzh3IJfusqqqKRYsW4fP5et13w4YNADz++OO43e7odpfLlbD2SZLUf1OmTCEtLQ2fz8fnn3/OUUcd\nleomSZLUwYACi7beitGjR5OWlhaXBg2UEIIXX3yR22+/Pfq4Nxs2bKCoqIgjjzwy0c2TJGmARo8e\nzebNm6mqqkp1UyRJ6sKAhkKampoAGDFiRFwaEw8bN27klltu4ZxzzonOVOnLMVOnTk1wyyRJioe2\nLzFtU9slSRpcBtRj0TZsoOt6XBoTD6NGjeKdd96hsLCQVatW9emYDRs2UFJSwoUXXsi3335LZmYm\nZ599Ntdeey02m1xORZIGk0gkAiDX+ZCkQWpAd838/HwAGhoa4tKYeMjKyiIrK6vP+9fX17N7925M\n0+T6669n9OjRfPLJJzz88MNUVVVFp9L2h9/vj3ks1ySQpPhpaWkBIDMzs9d95XtRkpJvQIHFhAkT\nAKipqcHr9fapGl59fT1er5fi4uKBXDpu0tPTeeyxxygpKaGoqAiwprU5HA7uuusulixZ0qkuR29m\nzpyZiKZK0n4vHA6zc+dOgD69L+V7UZKSb0A5FieccAJgld797LPP+nTMs88+y0knncTJJ588KIZQ\nHA4Hc+bMiQYVbY499lgANm3alIpmSZLUha+//hpd11EUhcMOOyzVzZEkqQsD6rEYM2YMhxxyCF99\n9RV///vfu11MqE04HOZf//oXAJMmTRoU+Qs7duzgs88+4/TTT49ZpyAYDAKQk5PT73OWlpbGPA4E\nAsybN29gDZUkKfr5UVBQ0GX5747ke1GSkm/AJb3/3//7fyiKwpdffskDDzzQ47533nknO3fuRNO0\naHnvVKutreWWW27hzTffjNn++uuvk5GRwYwZM/p9To/HE/PTvjaGJEn75vPPP+eVV14B4Morr+xT\n8qZ8L0pS8g24y+CQQw7hyiuv5MEHH+Tuu+9m69atLFmyhEmTJkX3qays5K677uK1115DURR++tOf\nMn369IFeep94vV62bt1KcXExOTk5zJo1i7lz53LbbbcRDAaZOHEiH3zwAU899RS//OUv5SqKkpQk\n3dWc8fl8vPTSS9x5550IIZg3bx4XXXRRklsnSVJfxWUs4uc//zkjRozgjjvuYPny5Sxfvpy8vDxG\njhxJc3Mz5eXlgJXPcO2117Jo0aJ4XLZPOn6rWb9+PZdeeim33XYbCxcuRFEU7rnnHu69914ee+wx\namtrKSkp4dZbb+W8885LWjslaX8mhODWW2+NCeR1Xae5uZmKigpM00RRFE444QTuuOMOVHXAna2S\nJCWIIvpSmrKPysrKePbZZ1m9ejVlZWX4fD5cLhfFxcXMnTuXCy+8kLFjx8brckOG3++PZqeXlpbi\n8XhS3CJJGhwqKytZsGBBl8MaqqridrspKiri4IMP5swzz+zz0gHdke9FSUq8uGZPlpSUcP3118fz\nlJIkDWNjxoxh48aNqW6GJElxJPsTJUlKGMMUhCJGqpshSVISpX6+pyRJw0rEMPGHDYK6gWEINFWh\n0K6lulmSJCWJDCwkSRqwsG7gCxuEdBPTFNhUBVVRUDWFuCVxSZI0JMjAQpKkfRIxTLwhnZBuIARo\nimL9aHJxMEnan8nAQpKkPhFCENZN/BG9tWcCbKqCpqggYwlJklrJwEKSpG4JIQhEDPxhg4ghAIFN\nVdEUFU2mTUiS1AUZWEiSFMMUgmBbMKELFAU0VcGmKsiuCUmSeiMDC0mSMIXAHzYIhA0ipomKYgUT\nMl9CkqR+koGFJO2nDFPgD+kEdAPdEGiKgqoq2GW5bEmSBkAGFpK0H4nmTIQMwoaJ1jot1C57JiRJ\nihMZWEjSfiAUMfCGdMLtEjDtmuyZkCQp/mRgIUnDlFW0SicUEYjWYEImYEqSlGgysJCkYcQwBd6Q\nTiBiRCtgajKYkCQpiWRgIUlDXNuMDn9YRzcFNlkBU5KkFJKBhSQNQaYQBMKttSZME422GR0ymJAk\nKbVkYCFJQ0gwYuAL6YTbFa6S00MlSRpMZGAhSYOcbpp4gwZBXUcIrCRMOcwhSdIgJQMLSRqEus6b\nkIt9SZI0+A2KwGLHjh1UVFRQX1+Ppmnk5eUxevRoxo4dm+qmSVLSCCEI6ga+oEFY5k1IkjREpSyw\n2LhxI08++SQrV66krq6uy33Gjh3Lsccey/nnn8/kyZOT3EJpMFu3q5lvdjZz0ewxqW7KgEWniIZb\nhzo0ddjkTZimwBcxKMhwpropkiQlSdIDiy1btvCHP/yBzz77jLy8PI444ggmT55McXEx6enpCCFo\nbGykqqqKL7/8kldffZWnnnqKY445hmXLljF16tRkN1kahD7cUoeiDO1v8mHdoDmoE9bN1noTQz+Y\nEEJQ6w2zo95P2Z4AlY0BQrrJf587I9VNkyQpSZIaWPzlL3/hiSee4JRTTuHpp5/m8MMP7/UYXdf5\n6KOPeOGFF/jBD37ApZdeyi9+8YsktFYa7ESqG7AP2nInfGEdwxDYtaFdWlsIQWMgwo49Acrq/ZTV\nBwhEjFQ3S5KkFEpqYFFWVsbLL79MSUlJn4+x2WzMnz+f+fPns23bNu6+++4EtnDwWLmljs++a6DB\nHybLZWfuhBzmT84DYFONl3c27KaqOYiqKEwpTOeMGSPJ9tgB+HxHA/9au5Or50/gpa+q2NkYJMNl\n4+iJudFzgDV18fX1Nazb1UwgbFKY6eTkaQVMK8oA4BfPr+PEaQWcfGBB9Ji3vq3hnQ213Nn6DfSf\nX1TS5I+Ql+5kbUUjIzx2rlswiRteWM9JBxawflczNS0hTpiSz4nTCmjwh3ntmxo27/aiGyYlOR7O\nPHgko7PdANT7wvzxzc1ccuRYvqxsYlONF01VOHhUJmcdUoTDpnL/yu1sr/NH27jkmPFMzE9L/D9K\nq5e+quLjbXs4amIuCw8p6tMxumnSEowQiljhkKYqqEN0Zoc/bFBW72fHHuunKaj3uL9MEZGk/UtS\nA4uBBgUTJ07kf/7nf+LUmsHr1a+r+XBrHfMn5zG5IJ3yhgDL11VjCEGWy8b/frGTmWOzWDA1H29I\n561vd/M/H2xj2QmTSHda/6RCwBOrKjj2gDxOm+Fk1XcNvPZNNUVZTqYUZmAKwYMf7aDOG+aUAwso\nyHDyRVkjj35azn8eM44JedaNui+jDdvr/Ng1lcvnFhM2TNTWg97bWMvpMwrJz3CS47HjDenc88F2\nHJrKOYcWYddU/r11D/et/I5rj5tIYebecfjnSndxxLgRXD63mLKGAG+sqyHNqXHajJGcO3MUz6yu\nBODcmaOSOn4f1k0+2roHAXy8bQ+nTS/EYeu+xyEYMWgJ6kRM05rZMQTvsrppUtkQZEe9n+/q/NS0\nhHrcXwFGZjopyfVQMsLNqNagUZKk/cOgmBXS0bp169A0jWnTpqW6KUkXCBt8uLWO703K5fQZIwE4\noCAdb1Bne62PXU1BphSmc/ERe2fMjM9N48/vbOGDzXWccdDI6PaTphVwxLgRAIzL9fDNrmY2VHmZ\nUpjBxuoWKuoDXDa3mOmjMqPXqfOF2VbriwYWfWEKwXmHjSLLbY/ZPiHPwzEH7O0heX1dDYGwwdUn\nTWCExwHAtJEZ3P72Ft76toYfzSmO7nvgyAzObH0tkwrS2Vzj5duqFk6bMZLCTBdOm4qiKBTnePrc\nzngwTBEdgjGF9bij9sMd1nodQysZUwhBvT/C9jof3+3xU14fQO/idbaXl+agJNdNSY6H4hFuXHZt\n7/kS3WBJkgaVlAYWQgjuueceKioquOOOOzAMg8svv5xVq1YBMGvWLP72t7+Rnp6eymYmVVm9HyHg\noNabfZuzDimipjnEHe9sYeaMrJjnctMdlOR62Fbri9le0u6ma1NV0p02QoYJWL0MmqpEg4o2V8+f\n0O82pzm1TkEF0Omb6pZaL6Oy3WS57TE35KmF6aytaIxte25swJDlttPgD/e7baPYoJ0AACAASURB\nVMlkDXfohCLW37GmWtNFh4KQblLW2iOxvc7X6/BGulNjXI6HcbkeSnI8ZLgG5XcUSZJSIKWfBv/4\nxz+4//77OfroowFYvnw5q1at4uSTT+aAAw7g4Ycf5r777uPGG29MZTOTyhe2Et/ahjTa84etD/sM\nV+ebeIbTxs5AIGabw9b5piaEaD2XgcehdXp+Xzi6ST50dhgi8IcM9vjC3PDC+i73j7QGPdY5Y9uu\nKNbwzmA0FIc7hBDU+cJsr/Wzrc5HZWOAnjolbKpCcY6bcbkexud6yEtzDPlZOZIkJUZKA4uXX36Z\nBQsWcO+99wLw5ptv4nQ6+dOf/oTH48Hn8/H222/vV4GFu7UL2Rs2yG+3vcEfpqopCEBLMNLpuOZg\nhLQugpH22t8G3HYtGsS0V9loBSdjWnsbzA5385Budjqmr9wOjYmetOgQR5u2KwyFG3JHNS0BXDYN\nbQgMd0QMk7L6ANtqfWyr89HcS69EQYaD8blpjM/1MGaEC9sgf32SJA0OKQ0sysvL+dGPfgRAOBzm\ns88+Y/bs2Xg8Vjf4hAkT2L17dyqbmHTFOW5UVWH9rmbGtxsO+GBzHV9WNpHhslFa0cSskhHR5/Z4\nw5TtCXDMAbl9vs6EPA8rt9SxsbqFqSOtWSCmEPzfFzvJT3fwoznFOO0qjf7YIGbHHv8+v7aJeWms\nrWgkL90RMwb/ytdVRAzBuTNH9flciqKkZOxeN2IDK01RB3X9ieZghG21PrbW+ijrJVfCZVMZn+dh\nQm4a4/M8XfaaSZIk9SalnxxpaWm0tLQA8Pnnn+P3+6PDIgCVlZXk5OSkqnkpke60ccykXD7csgeb\npjAhL42yPX4+2V7P9w8uwmVX+b8vdvLU5xUcXpyNL6Tz9obdpDk1jm2XKNmV9reUaUUZlOR6+OcX\nlZw6vZAcj4MvyhvY3RLi/MNHA1YC5ZeVTZTkeMhNd/BFWQN7fPue53DsAbmsKW/kwX/vYP7kPNwO\nja8qmli1o4Gz+jhts43HobFjj5+tu628jXgN63SnbbijJdTzt/xUE0JQ0xJiy24rmOhtBkdhhpOJ\n+WlMyPMwKssVndEjSZK0r1IaWEyfPp2nn36aMWPG8Le//Q1FUTjppJPQdZ0PPviAZ555JibQ2F+c\ncdBI0p02Pt1ez4pNdeSmOTh35ijmjLeCLKdN5f1NdTz2aTkuu8rUwgxOm1HYawJd+1uGqij85KgS\nlq+r4c31uwkbJqOyXCz+3jjGjrCGQb5/cBGGELz6TTWqojBzbBanzRjJv9bs7PKcvcl027l6/gSW\nr6/mudJd6IZJfoaT8w8fHZ290pe2Axw1MYeKhgAPf1zGhbNGM3Nsdj9a0jdCCHxhA29IR4i22R2D\n78ZrmILyej+bW4OJnoIfh6YyPtfDxHwPE/LSZK+EJElxpwiRupS4bdu2cdlll0WHOy677DJuvPFG\nPv30Uy677DLy8vJ44oknmDCh/zMVBhO/38/MmTMBKC0tjQ71SIOTKQQtQZ1Aaw5K+9yPYMTgrhXb\no49/dtyEmGGdZAnpJtvrfGzZ7WVbnb/H3Jdst51J+WlMyk9j7Ah30nNZBAyatULke1GSEi+lX1cm\nTpzIK6+8wqeffkpRURGHHnooAFOmTOH666/n7LPPHtBQSHV1NWeccQYPPPAAs2fP7tMxuq5z4YUX\n4na7efLJJ/f52tLQYwpBc8AKKFR18CWTBsIGm2u9bK7xsqM+0GUNjTZjsl2twUQ6uWl2OYNDkqSk\nSXk/aHZ2NqeeemrMtpycHBYtWjSg81ZVVbFo0SJ8Pl/vO7fz0EMPsW7dOo444ogBXV8aOgxT0ByM\nEAgb2FQF2yAqte0L6Wze7WVjjZfyhkC3U25tqsK4XA+TC6yeCY8j5W9tSZL2U0n99Dn77LNZtmwZ\n3/ve9/bp+Pfff5+7776bl19+udt9hBC8+OKL3H777dHHfbVx40Yeeugh8vJ6ToKUhgfdNGnyRwgb\nJpqiDJrFwLwhnU01VjBR0RDodj+XTWVSQRqTC9IZn+sZNO2XJGn/ltTA4pJLLuH666+nuLiYs846\niwULFlBYWNjjMbW1tSxfvpwXXniB6upqbrjhhh7337hxI7fccgsXX3wxc+bMYfHixX1qWzgc5oYb\nbuBHP/oRpaWlfX5N0tDTcbnywVCfwRfS2bTby8Zqq2eiO+lOjQPy05lSmJ6SfAlJkqTeJDWwOOec\nc5g/fz733HMPd955J7///e+ZNm0aU6ZMYezYsWRkZFjLMDc2UlVVxdq1aykrK8PtdnPuueeyePFi\n8vPze7zGqFGjeOeddygsLIyWBu+L++67D9M0Wbp06YCHYaTBKawbNAd0wobArqW+hyIYMdi028uG\nai9le/zd1uXIdNmYUpjOlIJ0Rme7ZL6EJEmDWtIHYnNycviv//ovrr32Wp577jk++OADXnvtNSKR\n2EJMDoeD2bNnc/HFF3PmmWcyYkTP0xHbZGVlkZWV1fuO7Xz99dc8+uijPP300zgcjn4d2xW/P7aI\nVCDQ/TdQKX66W848ohs0BVsDClXBnsIciohhsrXWx7dVLWyv82N0M1SX6bIxtTCdqSMzKMp0ymBi\nH8n3oiQlX8oyvLKzs7niiiu44oorCIfD1NTU0NDQgKIo5Ofnk5ubi93eeU2MeAuFQtx00038+Mc/\n5qCDDorLOdums0n7prsAoSddLWeuqQqN/jAhw8QehxoU/rDB+5trY7YFwkav001NISivD7CuqpnN\nNV7CRtfBRIbTxtSR6UwrzKAoSwYT8SDfi5KUfIMiddzhcDB27FjGjh3b+85xdtdddyGEYMmSJei6\nVVioLeHTMAw0Lfk1CpJp5ZY6PvuugQZ/mCyXnbkTcpg/OY/7V24HReGqY8ZH991a6+VvH+5gyTHj\nmZhvLave4A/z2jc1bN7tRTdMSnI8nHnwSEZ3WNm0r7oKEBy23ocsOi5nXucNAgo2VYnLGh7+sME/\nV1dQ64vtWXuudCcXH1HcZeXP3S0h1u1q5tvqFryhzuuyAHjsmhVMjMxgjBzmkCRpGBgUgUUqvfXW\nW+zatavLbzbTp0/ntttuY+HChf06Z8fkz0AgwLx58wbUzkR49etqPtxax/zJeUwuSKe8IcDyddVW\n97yi9FpV0xvSueeD7Tg0lXMOLcKuqfx76x7uW/kd1x43kcLM/hdF6hgg9FSroSe6IXDHscz3yq11\nnYIKgD1+nZVb6zj1QCsJ2RfSWV/dwrpdzexu6br8uUNTmVyQxoFFGYzL8QyZpdWHoqHyXpSk4WS/\nDyz+9re/xeR3CCG4+eabURSF3/3ud4wePbrf5xwK1fwCYYMPt9bxvUm5nD7DWm30gIJ0vEGd7bVW\n7Y/ebukfbtlDIGxw9UkTGOGxclOmjczg9re38Na3NfxoTnEiX0KUbpjU+2LXxIj3N/+a5u7X3Khu\nCrKpxss3u5rZXufrcvlxVYEJeWlML8pgUn5ayhNH9xdD4b24L4KvP49it+M88ftxPW/kqy+IlH6G\n58dL43renoQ//YDAkw+Q8fv7UHPyMHZVEHj6IdKvv7XP5/D+5RZQFNJ//l+Ja6jUZ/tdYOH1etm6\ndSvFxcXk5OQwefLkTvukpVnd/NOnT09285KmrN6PEHDQqMyY7W2Lgd2/cnuvN+cttdYCYFlue0zP\nwtTCdNZWNMa/0R0YpqApECaom/vcsxEPtd4wL35V1eVzRZlOZozKZNrIjIQvlDZYCCGiPU+aquKy\nyyAq3kKvPYvz9B/E/7zvvQZJHo6zHXQYaTf8ASXTWu8nsvZTjO829+8kcghxUBn2gUXHm+P69eu5\n9NJLex3iGO5j3b7WdTAGsgiVP2SwxxfmhhfWd/l8xDAT8s28bS0PX0iP5lAYSmIDi8JMJ9Xd9Fp0\njGkynDZmjMpgxqhMctMGPstoMBNCYIjWIEJRsalgs2k4bSoOmypXS02k1C3zFFdqeiZqembvO/ak\ndfhWGhyGdWBx5JFHsmHDhk7bNm7c2ONx+8MaIe7WmQzesEH7yiAN/jB7vGEURcHs8MEV7rDQlduh\nMdGTxpkHjYzZ3nZUvIs3tQUU/pC1lkcyhxOmFWawqdpLsJvFvjRVYUpBGgeNzqQkxzNsb6i6aSKE\n1RNh08CuargdGpqqDNvX3B+h914j/NF7mHtqUbNzcBxzIs4FZ0afj2z4mtDy5zB2laOoGrYDD8F1\n9sWoI3KB1mGBpx4k7fpbCT77KEblDpSMbJzHnRI9T9NVF1jXev05Qq8/R9b9/weAsbOc4EvPoG+1\nPvNsUw/Cfe6PUPMKAPA9eCf6+lLSf3UHWuEoAIKvPUvojRdI+9nNBF99FqP12KarLiDt5/+F7YAD\nO71Gs6WZ4L8eQ9+0DhHwoRaOxrngDBxHHkPkqy/wP3gH6b+8HW3sOOs1rfqQwOP34f6PxTjmHW+1\ntWIH3j/dSNr1v8es3mkNhdx6L+FPVxB6/floG5yn/wDX6echdJ3Q688R/vwjRHMjan4hzhO/j2PO\nsXsbJgSht18m9MGbCG8L2thxuM6/DFvJxIH/w0r9MigCixUrVrBixQp27drFsmXLcLvdfPbZZ5xz\nzjk4nYNjVcThpjjHjaoqrN/VzPjcvePQH2yu46vKZsbmuKluCsYcs70utibAxLw01lY0kpfuiJly\n+crXVUQMwbkzR8Wtvb6QTkswgkLy1vII6yYbqlsorWzqtrciL83BYWOzOLAoIyWrnCaSaVq9EYqi\nRNdQSXfZcWiarPjZhcDzTxJ+/3WcJ56JberB6GVbCb74NMIwcJ28MHqDtc8+Cuep5yC8zQRffRbv\nHb8m/Ze3o2a0fmsXJv6H/4pzwRm4Fv6Q8MfvEXzhKdRRxdgPPIS063+P745fYz/qeBxHnQCAUbML\n752/QRs5xsqPMAyCbzyP987fkP6rO1AzMnFf9BO8ty4j8MzDpP/8vzDKthN680WcJ34f2wEH4r7o\nCvyP3oOiKLguugJtZNf5ZYFH78H0NuP+4ZUobg/hzz4g8Ph9qCPysE07CGw29I1fRwMLfeM31u8t\nG6KBRWR9KUpGJrbxBxCu3mmdWFFwHL0As7GeyCcrSLvhD6jZ1iKU/kf/B31dKc7TzsU2bhKRdWsJ\nPHE/aDYcs4+y/g62bQRDx33RFaDrBJ57Av/9t5Pxp7+hDILquvuTlAYWuq5z7bXX8t5770WHHhYt\nWkR5eTm//e1veeGFF/jHP/5BZuYAu8mkTtKdNo6ZlMuHW/Zg0xQm5KVRtsfPJ9vr+f7BRdg1hQ1V\nLbzydRUHFmXwXZ2fNeWxeRPHHpDLmvJGHvz3DuZPzsPt0PiqoolVOxqiuRoDFTEMvC0RDFOgJenD\nYY8vzNqKJtbtau5xOXKA/zhizLAJKDr2RrgcGm6bhk0mmvZK+H2E338dx3Gn4lr4QwBsU2cgmhsx\ntm5AnHQWwReewnbgoXguuyZ6nG3CFFp+t4zQu6/gPvs/ottdp5+HY95xAGgTphAp/Rx93VrsBx6C\nbfwBAKjZudjGTQIgtPw5FKeLtGt/g+JyWeeeMoPm31xN6J1XcJ/zH6iZWbguXETgH3cT/vh9Qu8t\nRxtdgvNMqwdEKxqD4nKDokTP2xV967c4T/8B9kNmWccdMA0lIxNsNhSHE9sB09E3fhNNLNU3rUMr\nnoC+5du95/j2S2zTO8/EU7NzosFEWxuMneXopatw/eDHOI87NfrazPo6jC3roTWwwG4n7ae/RPFY\nOXIi4CPw1IOYVZVoo5OTSC5ZUhpYPPTQQ7z//vv86le/Yv78+Zx44okAnHDCCdx0003ceeedPPDA\nA9x4442pbOawdcZBI0l32vh0ez0rNtWRm+bg3JmjmDM+B1MI9vjCrC5r5NPt9UzMT+PSOcXc+8H2\n6PGZbjtXz5/A8vXVPFe6C90wyc9wcv7hozliXN8qpfZmjy9MutOe8G/IphBsrfWxpryRsvquqzOO\nznIxvSiDtzfWdvn8UNKWG4GweiLsmkq6S8Np0+SQxj7Qv9sCwsQ+M3ZVZPcPfgyAUVWJaGnC3nYT\nbKXmF6JNmIyx+duY7dqEvUnlis2GkpGJCHU/M0nf9A22yTPA4UAYrTVTXG5sk6agb/g6up/j8Hno\naz8j8PSD4HCQftPtKP2s1WObPIPQq/+HWb4d2/RDsU2fGRMU2Q46jOBLTyMMHbNuN6KxHuf5l+F/\n6L8x6+tQ3B6M7ZtxHndan66nb7OGru0zj4zZnvaTZTGPtaKx0aACQM2xhoBEILanVUq8lAYWL730\nEmeddRaXXHIJ9fX10e12u50f//jHlJWV8e6778rAIoHmT85j/uTOq7mqisLpM0ZGp6K2ufPcGTGP\nc9Md/OjI+H0b6LgabaIXCAtEDL6ubGZNRSPNQb3T8w5NZcaoDA4dk0VBhpNgxBiSgYVhCkxhTXu1\nawo2m4bHoWFTlWGfqJwMwtcCgJLR9XICwu+1nm+d+dCempGFUbkjZpvi6DAErCgguu89E14vkTWf\nEFnzSafnlIzYHl/7nGOJlK5CLRiFWtj/nkXPomsJvfki4TWfECldZfVwTD0Y98VXoubkYZ8+08oP\n2bYJo2onauEobAcdDg4n+pZvrdemqNgOPKRP1xM+b5evo5OOf2dtX0Z6+HuTEiOlgUVVVVWPJXen\nTp3Kc889l8QWSakUjBjsbun+W1k81XnDrClv5JtdzehdTFXNT7dyJ6YXZfap8udgY5oCA4GmqNg1\n8DhsuOwyNyJRFHdr93tLMxTsvVmb9XWYtdUorbMeRFNDp2PN5gaU9IyBXd+Thm3awTgXnBH7hDXn\nd+/DcIjgvx5HHVWMubOM8Luv9rsWhuL24Dr7YlxnX4yxuwr9y9UEX3+OwD//TtpPb0LNL0QtHIW+\n8RvMml3YJk9H0TRsE6egb15vBSIHHIjidPX5emD93SqtwyQARvVOhM+LbeKUfrVfSryUfmLm5ORQ\nXl7e7fMbNmwgJyen2+el4cEwBXXeEPW+cEJvfEIIvtvj49k1O/n7J2WUVjbFBBWqYtXguHj2GC6f\nW8zMsdlDJqgwTEHENDGEQFMhzWWjMMNFYaaTnDQnaU6bDCoSSBs/CTQN/esvYraH3nkF/yN3o44a\ni5KZTWT1xzHPm7U1GNs3Y5s4tX8X7NDLpE0+EKOqAnXMOLTiCWjFE1DHjie04nUiX66O7hd88WnM\npgY8i6/DMf8UayZIW/IkQC89hGZTA82//E8iaz+zrltQhPOk72ObehBmQ110P9uMw9A3rUffuhFt\nsjWzxDZlhrXt26+wHXx49xfp0AbbpGkARDr83QZfeIrgvx7rsb1SaqS0x+Kkk07in//8J6eccgpj\nxoyJee6NN97gueee4wc/iH8RGGlwEELQHNTxhXVsijXOb5hdr6kxEIYp2FDdwqodDdR6O5fZdts1\nZo7JZObYbDJcg2KiVK+soQ0roVX2SKSemp6J4/jTrAJTNju2A6ahb99M+MN3cJ33I2umxVkXEXjy\nAfyP/A/2I76H8LUQeu1fKOmZOE44o+cLdOhUU9xpGNs2om/+FtvkA3Gddh7eO36F/77bcBxzEtht\nhP/9LvpXq/FceR0A+uZvCa98C9dZF6Hlj8T1/QuJfPk5gcfvJ+36W1FUFcWThrF9s5VwOWYcSlp6\n7OvMGoFaUETg2UcRwQBqXiFG+Tb09V/iPOXs6H72gw4j/N5rANgOsAoNapOnI156Jvp8d9p6f8Kr\nP8I2fjLamBLsh80h+MJTEA6jjSkhsm4t+rq1eBZf3/s/jpR0Kf0Uveaaa/j888+58MILmTBhAgB3\n3303v/3tb9mxYwfFxcVcffXVqWyilCDBiEGjP4ICcVkkrCsh3eTLyia+KGukJdQ5fyI/3cGskmym\nj8wY9DMfhBDopkBVFOw2RQYSg5D77P9Azcgi/OE7hN55GTWvEPdFi3AcvQAAx9z5KC43obdewv/g\nnSguN7bph+Ja+EPUzK5zM6I6/DM7Tz2b0OvP47v/NjJu/gva6GLSr/sdwZf/F/9j9wCgjSrGs+QG\n7AcdjggFCTz5AOroEhytQx+K04X7wkX4H/gz4XdewXnyQpzzT8Ffth3ffX/CfclV0amc7Xl+8nOC\nLz1D8NX/Q3hbUEfk4jztXFynnhPdR5s4FVwe1Kzs6GvTiieAy42anYuaW9DtS7UfNofI5/8m8MT9\nOOYdj/uiK3BfdjXqa/8i9P5yq0ZF0Rg8V14XnZkii2MNLoromC2XZH6/n0ceeYS33nqL8vJyDMNg\nzJgxnHDCCSxevHhYTDX1+/3RXJLS0tJhu35BXximoNEfJqyLTvUo2pYlX7erJbpt8VEljOilemVX\nx80ck8m31d4up4tOyPVwxLgRlOS4+524GIwY3LVi78yYnx03IWHTTdt6JWyagrM12VKuMTIw8r0o\nSYmX8n5fj8fD0qVLWbo0eYveSMknWqtmeluHPboKKvq7LHlPx5VWNsc8VhWYXpTJEeOyyU8fnEXX\nuhrecDvk9E9JkoaWlAcWYH2LqK2txTS7nhY0fvz4JLdI2hcvfVXFx9v2cNTEXBa2K5AV0q1hDyG6\nH/bo67LkfT2ujdOmcuiYLGYVD778iejwhqrg0BTSnDacNjm8IUnS0JbST9qKigpuuOEGSktLu91H\nUZRO631Ig09YN/lo6x4E8PG2PZw2vRCbptDgazfs0cP9sqdlybt7rsEfZnONt9vj0p0aV8wrGVSV\nMQ1TYCKwqQpOu0aOQ1a2lCRpeElpYPHb3/6Wb775hjPOOIMxY8agdVEBThbvGRralskGa7VPX1gn\nFDFQuxj2GKhGf4SPt9ezrqq5xwUe0522QRFU6KYZXeMkwyWHNyRJGt5SGliUlpZy+eWXs2zZst53\nloaUBl+YDJe9z/v3tCx5YaaVE9EcjPDJ9nq+3tncaanyno5LBd00AQW7ppDtseOyaTJIlqQOAs8+\nRnjlmziOPQX3+T9OdXOkOElpYOHxeCgs7HrsXBra+jt74dhJeexqCHTKl8j12JhdPIJ3N9ZSWtmE\n0SGiyHLZME1BS9jodNyxkzqXKk8UIQSGCYpq5XVkuhw4bKoMJiSpGyIcIvzBGyAE4ZVv4lp4UedS\n5tKQlNLB3QsuuIDnnnuOYDDY+87SoGUKQb1vYKW4PQ6Ni2aPZcao2NLGJblpPL6qnC/KG2OCiiyX\njdOmF7D46HFcNq+k03HnzRzd7UySeIoYJiYCp10jP8PByEwXIzwOnHbZQyFJPdJ1omOZpmk9loaF\nlPZYXHXVVZSWlnLiiSdy9NFHk5eX1+WHsRwqGbwCEZ1Gf6RTT8K+8Dg0FkzJj6lHsbaiKWafDKeN\neRNyOHh0ZnT2RFfHuRMUVCitOaii9c9FWe6EXUtKLDMcBlnHQtoPCNMEQwdVRdESf9tPaWDx4osv\n8sknnyCE4MUXX+x2PxlYDD6msApdBSMmdk1FVwYeWIjWpcu74rarzB2fw8yxWUkvEiWEQBcCTVXI\ndNk5amIOn2yv56iJuTKoGMJEbTWGHkYdkdvvpcMlaTCxAgcDEYlAJIzQI2AY1o9pgLCS65W0dLQR\nuQlvT0oDiwceeIAxY8bw61//mnHjxnU5K0QafKxy3GGrvHScbvJVTUHe21RLZWPssJhNVThi3AiO\nKMlO6gwPq8YEaBq47Bo5Ti26hPvZh47i7ENHJa0tUoLY7Qg9gllVCU4XSmYWah9X3JSGt8GWVCqE\nANNA6DqEwwg9DEbr8JGhRwMHFMXqlWirF6QqoFq3+WQOzKY0sKitreWGG27g2GOPTWUzpA66K3TV\n1ksRioi41V5oCeqs3FLHuqqWLp+/bM5YcpNYKVM3TRQFXDabFUzIGhPDmqIoYLeDaWDW7cZUVVRP\nOkpG5t4PZ2m/koqk0pjAIRJGRFoDB0Nv7XWwikeKjoGDAtiSHzj0JqWBxcSJE6mtrU1lE6QOuip0\n5bCphHSDBn84Wo9hoHTTZHVZI59srydidD+MkuZM/H9R0xQYCByaygiPY1DUvpCST2n9gDZ9LSje\nZnA4UdxuFLcnKePS0iDRVVLpAAOLthwH0drDICLthirahitoDRwUJXZoTtOsHwZX8NCTlL5bfvaz\nn7Fs2TKmTp3KCSecgN3e97oHUmJ0LHRlmIKmQARf2IjbKqTban28u6mWBn/s1NL8dAffm5jLC19V\nxeU6PWmbHqpq4HFqpDltsmiVBLD3Q900MJubUJoaEHaH1ZPhSRuUPRmDret+fxPT4xAM7s1xaEua\nbD9U0SlwUK0fhk7g0JuUBhaPPvooLpeLn/3sZ9hsNjIyMrrMs/joo49S0DoJoNYbxKFp2OOwfkVT\nIMK7m2rZsjs2QdNtVzlmUi6HjMki3MVqpPFkmgJTgMuhkOO0y6EOqUfWDUBDEQKzudEKMmw2VJcb\nJS0j2suRSvt7PYhkBFXR5Eg90prjEGkNGgwrcAAreBgiQxWJltJ3ha7rjB8/Xi4yNpgJBrwolmEK\nvihr5KNte4i0m5aqADPHZnHMpNyEDj/I3gkpHpR23dGm34fS0mwFGU63lZORqiAjAV33Q0U8gyrR\nms9g+mLzvYyqShSnq+vkyHaBQ9tDKcWBxZNPPpnKy0tdEB0W3xhokaddjUHe+LaGWm84ZvuYbBcn\nTSugICNxH4Cyd0JKFEVVrRsMYAb9KL4WK3J1OFBcLpmXkSz9CKqEELG9DpHw3h4H02ztcQAR6lDs\nT9PAbpdBQz/I//lSlGEKar0Dq6DZJqyb/HvbHtaUxxa48tg1jpucx4xRGQmtTGkIIXsnpKRoCzIA\nMHQrL6OxwQo07HYUm92azupyDcr8jHgYLDkewjCsoCEchIiOMI0u6zmgae2mZO7991MAIpFuzj40\nCSEwwmHCfj9aegae4VbH4oILLuCaa67hqKOOij7uy83lf//3fxPdtP1eIKLT6NPjdhN+8vNKWkKx\nJXoPGZ3J/Ml5uBMw7NGxp6UwwyWLV0kpobTL4sc0MEM6BHwoZmthFM0GwzNmvwAAIABJREFUNg3F\n4UJxusBmGzTl3/clQEhGjocwTatXwTT3Tsk0IghfbL6WuXsXuNygqLEJkimq55BsQgiMSISwz0ck\nFMKIRDANK6BSNA2HqpGMWrNJDSxqa2sJBAIxj6XUawpE8IcMbJrCvuRO+sMG72+O/bdsH1Tkptk5\n5cBCxo5wD7SpnViFrMSA80AkKVEURYmZMogwIWJiBoNgmtaNTrNZz9s0FM0OdgeK3Wb1eqjJWcxu\nnwOEOOV4mAE/imGA3q6nwTSjPQ1dFoEyOqwvotmsHqL9hKHrRAIBwoEARjiM3trbomoaauv/Gy0F\nuT9JveL777/f42MpuUwhaPCFiRhin2tT+MMG/1xd0WlVUrC+JMwZn8O8CSOiVSvjZW/+hEquy57w\n2SSSFG8xPRuwN+AIhVrH/EFRhLUwjapGgwxU1UoUdbhQ7DbrG3w8xClA2FsdMmIFCe1yGBAiGiwI\nf4fehsb6ztN5FWL+jvbnrw9tQxpBnw89FEIPhxCmQFUU1NbgwTZISjakNMfi3nvv5aSTTmLy5Mld\nPv/111/z/PPP89vf/jbJLRv+DFNQ5w1Z79sBfNt/f3Ntl0EFwKT8NI6ZFN/xPMM0QVFIk/kT0jAV\nk7PRkTDBMDH1CPi81o06GIjZxajZheJOi9ZMiJ6r7b3S9pZpfaxgXU+EY8vpm81NKJGw1XvQNpur\nLTgQwmqLKRCB2ADB2F2F4nJ3nkHRnqrGBlVYgdZwzUHpr56GNNTWvydNs8EgHe1NeWBRUlLSbWCx\nevVqGVgkQEg3qPdFsA1w+GBXU5AN1d5un28Oxm8ZZN00URWFTLcdj0PmHEv7N0VRotMchR7p+GRr\n8NAaABg99+YJrBwG0SFAMQM+FETPwzCq0ikIUmw2lEHyzXmoMA2DsN9vDWlEIhjhMEIIVJstpUMa\n+yqpLS0vL+eSSy7BMIzott/97nfcfvvtnfYVQlBfX8+ECROS2cRhzxfSaQpEBrR4mCkEn33XwL+3\n7aFDzmTc6aYJyFLbkpRIiqru7dFo26YoCcvtEH4fwbdeit0W8KO4949l7PVwONoboYdCmKaBoqjR\n4EEb4oFZUgOL4uJiLrzwQj755BMA6urqyM/PJze3c3e5qqrk5uZyxRVXDOia1dXVnHHGGTzwwAPM\nnj272/1CoRD33Xcfr776Kg0NDUydOpWlS5dy9NFHD+j6g0ljIEwgbAwoqGgORnj1mxoqGgK97luY\n2f/xWU1VULC+RSkK5KQ5yHI7+nWcqgy8qJckSYkh/D78j92L2B1but//zMOkXX4NiictRS1LDGGa\nREIhwj4fejiMHgpZn1OtCZaqpqEOs5W9k963smTJEpYsWQLA8ccfz7Jly1iwYEFCrlVVVcWiRYvw\ndZiS1JVf/epXfPDBB1x33XWMGzeOF198kcWLF/P4448za9ashLQvWYQQ1LclaQ5gDHNrrZfX1tUQ\njOztWs1w2lAVQVPQiNk312Pj2El5/b6GqigcVpxFaUUTR03M7VNQAeCwqRw9KTe6KqvDJsdqJWkw\nCr33WqegAoC6GkLvvYbrzAuS36g4Mg2DSDBIyOfDCIcxdB0hBFrrtOKh3hvRFykdtEnUrBAhBC++\n+GJ0iKVjjYOOKisree2117j55pu56KKLAJgzZw5r167ln//855AOLEwhqPNa43X7+i3eMAUrt9Tx\neVljzPZpI9M5eVoBprCSONft2lsK97yZo/H0o46EEAJDQLpL46JZY/jh7LH9bufCQ4pilnmXJGnw\nMaoq9+m5ZNiXIZpofoTfjx4OY0QiVnJl27BGInMjhAA9Yk271XXrz7qO0vbYaL/NgCO+l7i2tDN0\nskH6YePGjdxyyy1cfPHFzJkzh8WLF/e4f0FBAc8//zwlJSXRbYqioGka4XC4hyMHt7ZKmirs8+wJ\nb0jnzW93U9m4N2PcriqcOC2fg0ZlRsdgF0zJjwks+lqcqq0Ohduhkee2y1kekiSlRF+HaAxdJ+Tz\nEQkEOuVHKIqCzdHay2qarTf1iFXNU9dR2h5Hf9pv06PbFb1doNAaHCjtAofoNsPo+DJ6fo0ysNh3\no0aN4p133qGwsJBVq1b1ur/D4WD69OmAdaOrrq7mkUceoaKigptvvjnRzU0IwxTUtoRQlb6v99FV\noaunV1cSaDf0kZfmYOEhI8lLH3hlvYhh4rCpFKQ74l7nQpKkwUkrGoO+q6Lb55LNKgMeIvjG810O\n0Zi11TQ/fj9KyUTMgB8iERQ9gmLo2CN7gwSl3Z+JRFDM/t30h5NhGVhkZWWRlZW1T8c+9NBD/PWv\nfwXg/PPPZ+7cuf0+h9/vj3ncvtpoMkQMgzpvuF836+4KXbUPKmaMyuDkaQUDSv4EK+hpS8yUMz2k\nREr1ezHRhuLsCucJZ2BU7Oh8E88rxHnCGb0eL4SwCm+Fgojm2OFZfX2pVaEzFLQWEwuHrD+HQxAK\ntf62Hrdt61S9swNVUaBmJ9TsZLB//RGq2loy3gY2O2g2hM0W3Zas/uBhGVgMxPHHH8/hhx/OmjVr\nuO+++wgGg/z5z3/u1zlmzpyZoNb1LqIb1Pki/e4BWLm1rttCVwpw8oEFHDI6c0DTz9qWL093aWS4\nhn8Ck5R6qXwvJtqQnV1hd+A671JC772GuWnd3s0HHUZk7WdWPY1QEBEMWEFB2+OQFSQQClplvrsQ\neu1fyXoV/SIA7HbrZt/6I9pu/tHfdkTMY+u30No/tu09Vmu3re3PvXzuJ+uGLwOLDg444AAAZs2a\nha7r3HPPPSxbtoyRI0emuGW9C+sGdd4I9g7lud/dWMuaikYOH5vNgqn5XR5b09z9qqY5aXYOHbNv\nPUBtdEPgsCnkZThkHoUkxUEqZ1cIXUcE/IiAD9FYH/Nc+MN3QI9YzwcDiKAfEQhAMGAFCd30EERW\nvJGw9vaVaF2PpP1nlG6YkJGJWjTWuqm3BQitv63HjpjAQbR7nrYeg/3ocy/lgYVhGKxatYq6ujpM\ns+sKcQsXLkxoG3bt2sXHH3/MWWedhcOxd3rjgQceCMDu3bv7FViUlpbGPA4EAsybNy8+je1GSDeo\n7yKoiBgma8obEcCaikaOPSC3y6GMSA/V+QZUTMu03qgj0uxy2ENKulS8F5MlXrMrhK5j7tlN6L3l\nMdtDK95AEQLh91kBhL81kPD7INJ9Unvk0xV9vnYiKJlZqGPGoTic1uqxDifYHRiKgi7AQMG02VpX\nl7Wew+5AGDqBl/4Je2pjprn70rNxn3MJwhX/RRSHq5QGFhs3buTKK69k9+7d3e6jKErCA4vKykp+\n85vf4PF4OP3006PbP/74YxwOB+PHj+/X+Tye5I5vhnSDPb5wlwGA0Xpjh7bqvoKO9/fNu700+Lse\nBoF9K3QFEDFN0hw2Ml2DZ1loaf+S7PfiYCFME7OuBuFtwfQ2I7wtCJ+39afdn/1ea2ihC/rn/05s\nI1UtZkhDLZmI4klHcblQXG4rKHB28WeHE/9zj0PNrq7Pm56J+/zL0EMhgi0tRIJB9HA4ZgpoV1+V\nFMB9zn8Q/uhd2LEput195vnW2idDjGkYePfsobmmJvpz9H8uScq1UxpY/PnPf6axsZGlS5cyderU\nmN6CRPJ6vWzdupXi4mJycnKYPXs2c+bM4dZbb8Xr9TJ27FhWrFjBM888wzXXXENGRkZS2rUvokHF\nPs6q+KKskXc3db98/b4WujJMQWGmE4dN9lJIUjyI1gXHzOYmREsjit1BtxV6qnfiv/dPiW+UZkNx\nuxHevVPNbTMOQ0nPsBJIXW4Ut8cKCFzWb9xuFKcboUfw//lX0ePcFy7qc9KpomldvvagrhPx+Qjs\n2BFTlMrWx3uL4nLj/N6JMYGF4nT16dhUCba00FRTTXN1DU3VVTS1BhEtu3dbC5e1s18EFmvXruXy\nyy9n6dKlCb1Ox2/L69ev59JLL+W2225j4cKFKIrCvffey7333suDDz5IbW0t48eP5/e//z3nnHNO\nQts2EGHdYI+3656K3gghWLE5tuiVQ1MoynJRVr83c76/ha7aFGTIoEKS+kOEw4imesymBkRTY+vv\nBkRzI2ZzI6K5qcchiIRIS8dxxDEoHg+KJx3cHmtpc7fHWkFVjxB86yWMr1ZHD3EcfxpqTh++jPQy\nG6MnbVNWTVOgtiv8FwgZ2PNGDqkFu/rCNE28dXU0VVfTVFVlBRDV1TRVVxPydr8QZKqk9G/f6XRS\nWFiY0GsceeSRbNiwodO2jRs3xmxLT0/npptu4qabbkpoe+IlEk3U7H9QYZiC5etq+LZ677eMTJeN\nH8wcRYbLxl0rtke396fQVXty6EOSYgnDQDQ1YDbsQdTXYTbusf7cWI/ZWA/+3pce6BdVRUnLsH7S\n01t/Z1jDDW2/09IJvPwM1HSRBAooWSNwHHtS168nRbNS9HCY8KFzaf52PWbdbnLS9vYoqHn5OOYe\nl5DrJoOh6zTX1NBUVUXjrl00Vle1BhLVmPo+BGKKQlpODpkFBWQlcQJCSgOL448/nrfffjtaRlvq\nm4hhUNdNTkXvx5q88nU12/fsnd9fkOHk/MNGke60EYz0v6iLbppy0S9JonVWQUsTZt1uRN1uzD27\nMffUYtbXIRr3WNUYB0rTUDKyUTIyUTOywO1GX/Np9Gn35deg5hZYvQt9GCK1jRmH3k1g0VPBqmTN\nShFCEAkECHpbiASCmIaBZrfjOfeSIZsP0RZANO7a1fqzk8Zdu2jevRuxD/9H7G43WSOLyBpZSFbh\nSDJbf2cUFPR5GCieUhpYXHzxxVxzzTUsXryYU089lZycHNQu3gjDaYXRgTJM0e/iV+298GUVVe2m\nlpbkuDnn0CKc+zhsoRuCTFmKW9rPCCGsxMiaXZi7q6yf2hrM2moIdz91u1eqipI5AiUrGzVrBErW\nCNTMbJTMbGtbRhZ40mJ6BEXAHxNYqPkj+1Uga18LViVyzQ9hmoR8PoJer7UaqBDY7PbYlUCHQD6E\nME1a6upo2LmTxp07adhl/W6qqbEqfvaHopCek0tW0UgriCgaSVbhSLKKRuLKGFiNoXhLaWBx3nnn\nAdZ0z5UrV3a5j6IonYYy9leGKdjdEhxQ+ev2QcXUwnTOOKhwn85nmgKhQEGmE01VCIT33/K10vAm\nhEDU12FUVWBWVWJW7cSs3mnNqNgHSmY2yohc1NYfJTsXNTsHJTsHJSOzT70M8aR40vD8eGk0V+LT\nshrWVzdwyCklHJfEIluGYRBuaCDk92NEItC6XtNQyZcI+bzUV1bSUFlJw86dNFRW0rhrF3qof4Gm\nomlkFhSQXVREVtGo1t9FZBUWYnMOfCmFZEjpv9gf//jHVF5+SDGFtfaHFqeo9NAxmZw0rWCfehoi\nhonboZHttg+qKFmS4kEEgxiV32FU7MCsLMPYWQ5Bf+8HtqOkZaDkFaDmFaDmFqDm5KHm5KOMyEGx\nJ79rujeKJw3XKWfTtHYV66sbEMBXb7/D0YuXYHd13QsQ7zU/GnZWorrTUNtNCx2MTNOkZfdu6ivK\nqa+opKGygvrKSvwNDf06j6KqZBYWkj1qlPVTZP3OLCwc1K+/L1La+sE842IwEaL/C4q18YY6J/wc\nUZLNcZPz9ikoiBgm2R47HsfQ/o8vSW1EMIhRthXjuy0YO7Zh1uy0ir70gZKeiVpQhFpYhJo/EjW/\nEDWvcFCv1dETQ7Sre2OaGLpOd8X393kIRdcJNDcTrt9D+xDLZrP3WpI62XTTpK6sjD3V1dRXVFBf\nXk5DZSV6P1e9zsjPJ3v0aEaMGk32qP/P3nnHR1Hn//85M9vTO6mE3kGkowgqcgionGIXFAtYTj09\ny+HZy52eX/3d6amIYgPEU+zogdiwIAhIk94JaaSXzbYpvz822WRTNyHJJmGej0cIO20/u9mdec37\n836/3klEpaR0CQHREB3iVe3Zs4evv/6azMxMjEYjiYmJnH322fTv3z/YQ+sQFNq9H+KWiIoPt/h/\n6cekRzKpT/NFhVp5oo0PN+udSHUaRVVVPLKCw+nG4XYjiiIJ0SdnCd8WeH7+hopjh1AzjoDWdMKc\nEBOHlJiKmJiC2C0ZMSEJMbTjety0NbWnUKqwXXWTryLk+4UL2fr5ZwybMYNRl1+B225HkWVvrkQH\ni3Z6nE4KMzIoOHSQooNZFNhdFDlcaBufDfgYJpuN6NRUopKTiUpJISo5hcikpAajPl2VoAuLf/7z\nn7zxxht1lr/wwgvMnj2bBx54IAij6jiUODx4FK3ZVRd2l8zyTZkUOfwdNcf1iG62qFA0FaMoEhWi\n9/nQ8UfTNNweGbvTicPpwS3LKLIKAoiCiCSJoLXcr6Atca9djSQ1kLRssSGlpiOlpiMmd0dKSu20\nUYi2pGoKxV5DWFS9Ty67na2ffYqmaWxbuZKB552HyWLFYGy6AaFWU+g1FjyqlQCpKUrdyhvB/4Eg\nCMguF4UZGeQfOULB0aPkHz1CSU5OwJEqBIHwhASiU1KITk0jKiWF6JQUbFFR+vQwQRYWK1as4I03\n3mDSpEnccsst9OzZE1VVOXToEK+++irvvPMOAwcObHNL746K3SVjd8nNLit1eBTe25xJgf3kzXSq\nqj4irI2fDCRRQIDKBj7o5addFFlWqHC5sTuceDwybllBAyRRRBJFREFENHbCiJbVhpTeB6lHb6Tu\nvRHjEto9ibKrYC8qxJWbS0Vxic/fRlNVNLnyoi8IlT+13l+DAYxGBEFEEESvIBAE6u/1LXh/apVS\niuERYKqKDmiAhuqRKcrIIP/QIfIOHiD/8CGKMzMDLuuUjEavcEjrTnRKKjFpqUQmJ2OsnUipVf6j\nqtUCRdOqm49VvRdV/wpi9TpB6FKCJKjCYunSpYwaNYpXXnnF700dPnw4L7/8Mtdeey3Lli07JYWF\nS1YocTTfAMslq7y/OZO88tZx6Au3GZoUFQAmg8iZvWP4+WABZ/SKwWTQT8qdnSoR4XC6cHtkPIqC\noqoIiBgNEiBg7MRzxGJiGqaBQ5D6DERMTNGFRCNoqlp5wQShKoQgCCDWtdZ2qwJSTBwmq39FiRgX\njxgSWr3A6fBfHxEFzfWgqH0DY7Zgt9vJ27uXvP37yNu3j/yDB1ACzIkwWKzEpHcnrjiXWJuFmBAL\nYfc/5T/uZqKpNUSGpiHgfR9RlMplCigqmlr1mBriRKsWLDXFStVvAdCq/iJqLbGCV6AFgaCeFQ4d\nOsR9991Xr1ITRZEpU6bw3HPPBWFkwUVRNQpbYIAlqyofbc3yKymNDTWRX0NkONxKs7qM2oyBf0Rm\nDktk5rDEgLfX6Tgoqorb7cHucOFwe/DIci0RAZIoIYldx6bdOvdPmKwd30yppVTlN5x2wYVMuvnm\nJrev7Z4LeBuFGY2IkhGMJgSjwZtgKYh4nA4qCovwlNv9kjCl0FDvtoYmLuaShCYI3g6qgggNTUs1\ngux2k19WQW65g9wyByduuyXg6gyD2Ux0j57E9elDTM9exPbuQ0RSEoLbhfSvR3zbKS0YV00En/ip\n+l15vFpTQs2JV2hVAsRPrGigqDXEijd6oqFUBW/qRHjaiqAKC6vVSnFxcYPri4uL260xWUdB0zTy\ny5tfVqpqGp/vyPXr82ExiH6iAmDFlkyuHp3WYP+Pek8uOl0KTdNweWTsDidOtwe3R0ZWFEDAKEkI\ngtDlRMSphsfpZOvnn6GpKls//4wzrrvOl0CoqSooqvfGVhC9UxCShGire1cuxcYhhYX7HiseDxX5\n+bjtdlRFRTIZkVpaPms0oY04AzavQxsx3tu+vAkcxcXk7tnNiT27ObFnD/kHDwRkdS0aDER1Tyeu\nTx9ie/chrncfIlJSqs22OhmCb4qotlipZ9sa/xfbKYk0qMJizJgxLFu2jIsuuojU1FS/dRkZGSxb\ntoxRo0YFaXTBoajCXTn9Friw0DSNr/fksTe32rAn0makuJ5W6AUVMmsP5HP+wLo9WlRV61LzfDpe\nFFXF7nBR4XThdntwy96EN7FGXoQ+ddW1UGTZl0OgqSqKw4HRZPZGHywmBJsNDP4+NBXZdduQO0tK\nMIeE4igpxllaiux2ezuGSlLDia/NQJt8IdrkC+tfp2mU5eSQs2snubt3kbt7N6VZmQEdNzQhgbg+\nfYnv24+4Pn2J7tkzKNbWpypBFRZ33nkns2bN4oILLmDatGn07NkTTdM4dOgQX375JaIocueddwZz\niO2K3SXjktVml3P+erSY3zJKfI97xtqw1+NfUUVuaV0nOEXVMBoEwiz6l6+zI8sKdqcTu8ON2+PB\noyiIgohB6vx5ETpNo8myNyReAzExGSkissF9HMXFfHzX7ZxXq3r2/dtuZtqjj2MJj0A0GNr04qxp\nGsXHM8jZ+Ts5O3eSu2snFYWFTe4niQJxIRYSQq3EXnU98YOHYI2MarNx6jRNUM8wvXr1YunSpTzx\nxBN89NFHfuuGDRvGgw8+SN++fYM0uvZFUbUWJWvuyS3ju335vseJ4WZmDk3k3U2Be/WrmobJIBId\nYtKtuTsZVdMa5RVOnC63t9xTVX1CQtCjEV2e2j0npLiEOtNYTSWm/viff1Nw8CCc1stveWlGBlve\ne48zbr2tdQZbA03TKMnMJPv37WTv2EHOzt9xlpQ0uZ8tOpqEAQOJ79efhJ49ifv0bV/rdGXkqOYn\ngNamFXI/TnWCfusyaNAg3nvvPfLz88nMzETTNFJSUoiNjQ320NqV/HJXs0VFdomTlTtyfY8jrAZm\nDU/CZBBJCDeTU09kAiAhvLpMStFUzJJEVIj3TkQvG+3YqKqKy+2hrMKJ0+X1jQBvYqUo6rkRpwKa\npoEse8WCyYwYFe23XjCa6vg7NHW87J07G1xfcPhQi8dam/K8PLK2byN7+zayduzAUdR0RCIiOYVu\ngwaRMGAgCQMHEhoXXz2F43T4REWr0YLcj46GqijeEl/VWzkiaCBIApIxvOmdW4GgC4sqYmNjTzkx\nUUWpw+OtIGrG96PMKfPh1izkylIms0HkstOTCTF7/6QTe8eSVeQgz+6fZxFjMzCxt/d9VjQVs0Ei\nylb9xdHLRjsWmqbhdLkrhYQbj6yggi/JUp/WODXQFAVBU9EkA6LZihAT6us5orld3hJDTfNWbBgM\nAQkLxeOhPD8ft708oATI5uAqLWPDm6/7Lfv07j9TfuJE4zsKAtHd0+k2aLBXTAwchDWi/V1bG8v9\nCCaqoqAqCr7bv8qSUlGSECSD19HUICEZjRhMZkSDAUGsXN+O5dTtelY688wzeeSRRzjvvPN8jwPh\np59+asthBRWPrFDeTBMsWfGWlZa7vCcPQYA/DkskJqRaINhMEleOSuXbfXn8nlXmWz5reDI2k4Si\napiN/qKiCr1sNLi43B7K7A4cLjcujwwCGESvkDDoQuKUwSsmNG/CZXgEgjUEoZ6wvGAyY5p0Pu61\nqzBNnIpgMkMDHTU1TcNZUoKjpNibiGk0IhmNxPbsSfHhgz7r/prE9OgZ8JhVTSNz+3bWvvyfOtMa\nDYmKqPR0EgcPIXHwELoNHIQ5rBk26V1k2kLTNLSqKIPmrfoQRKF+wWA2IxqMiKKI2EHPB+06qh49\nehASEuL3+FRG0zQKKpqXV6FpGqt3n/DzqpjSP470mLp2wzaTxOR+cX7CwmqSUFUNoyTUKyp02h+P\nLFNW4aDC4RUSmqb5pjb0iMSphb+YiESwhQR0p2m97Dqsl13X4HrZ7aEkOxt3RTkCQp1EzJHXXMuJ\nfXvZc8K//D8iNZWR18xp9LmdpaVk5ZeQUWzneIkd169PNbp9eGIiiUOGkTR0KN0GDzm5iEQnmLao\nOS0hVBqGiqKEIEoIBglRMnjbw5tMlaLB4I00dOIKvXY9ay1ZsqTRx6capU7Zm8zQjM/PluMl7Kgh\nFIanRDA8teFs79qomoYkCUSHdLwv4KlCTSHhlmVURUOUvKWfhk56x6XTcjRNQ5BlNKMBMTQcITSs\n1cPWxceOYY2KRDLU76JrDg9j2uN/Z8Obi+H4t77lUxY8hDncP4KgaRpFR4+QsWkjGZs2cWLf3oB7\nbER1T+eP/3qh5S+kHoI9baEqijfaQJUZqcH7YzD4Rxl8kYeuf7MQ1Fe4YMECrrjiCoYNG1bv+vXr\n17N48WJee+21dh5Z2+ORlWb3AckqcfL1njzf45RIC5P7xzXreQUBYkJMnVoNd1SeX/oFH3y9nksn\nj+Xua6b7lsuyQnmFE7vLhcstoyqqT0h4ky2DOGidoFEVnRCsVsSYeIR6mnN9++wzbHn/PYZfdgXn\n3Ht/QMdVFQV7ob/7pGQ2NvmdN4eHMebGGznwfbWwMIV7TbMUj4ecnb9z7NcNHNv4K/b8/IYO0yhS\nAA3IOho+4VB5EyhJ3mhDVWTBYDZhtFi7RKShtWhXYaGqKnKNJKGPP/6Y0aNHM2DAgHq3/fnnn1m/\nfn17DrHdKGrmFIjDrfDJtmyqbOdDTBIzhyU2u2ojJsR8Sn7wG7rotxZOl5v31/yCqml88PV6Zs+Y\ngCwruNxeHwlJkJB8YkJXEqcqmqYhKDLfL36DrZ9/1qhg8DgcbPnvcjRVZcv77zHhT3dgbMSCvCoZ\nszQ7i03vvOW3zl1ajrkF/S6Orl9P1ratZGzehKeiosHtbEYDqZEhpEaGkJHcl73ff1/vds3J12gv\nakYcwCscRMmAYPBGGAxmEwazBclo1IVDgLSrsMjKymLGjBm4XC6fdfSCBQtYsGBBg/sMGTKkvYbX\nbpQ5PagaSAF+PjVN44udud6pE7xRh4uGJhJqbv6f71Rse177on/rpedhMbfeVJCmaZQ5nL7EN0VV\nyc4rIiI0RPeR0AG87peCqiKYLSih4d524k0IBkX2VLtnKgqK7MFI3e1c5eXYC/KR3W5kh5OvnniU\noqNH/bb56h9PMOPJp+tMa9TGbbf7Pf7pPw1MWwgCcX36kDpiJKlDhhL3yZu+C27CVbM5cfhwnTEE\nkq/RVlSJhyobbO+UhBHRaMBiNmO02nzTFLpwOHnaVVikpKTw0EOu87YnAAAgAElEQVQPsXHjRgA+\n+eQTRo4cSUpKSp1tRVEkJiaGK6+8sj2H2OYoqkaZu3lTIJszSjiQV/2Fn9g7hrTorts8qbXxyIrf\nRd8jK1jMTezU5DFlSu0OKpxuXJXNu2qi50roQLWgEG2hCBGRCKKIp6w0IMHQGKqiUFFUhLPUeyzJ\n6E3G3PD6ojoXdICSjAw2LX2nXqMrt93O0Q3rObLuZzK3bmnwOSWTieRhp5E2ejSpI0dVu1t63N4y\n18rKDHNkhC9fo+a0Sn35Gq2NKnunGgHvdKPRhGg0YjabMVosGEymUyLHIdi0+zt8ySWXcMkllwCQ\nmZnJLbfcwvjx49t7GEGjxOHGZA78JHKizMV3e6vnM3vG2BiTHrhdrayqTW+k0yRVfhKldgcVLjeK\nXJ0nYTQYMBpa1wegK3EqNrarcsMUQ8MRwsJb7S64arrDVV7mvcOWJGom6RQcPtzgvjWNrmSXi2Mb\nf+XQjz9w/LfNDfpYmGw2UkePofuYsSSfNtzXyMyPeiozzEZTg/kaJ4umaaiy7K2yECuTJY1GJIMR\nY4QFo9WKZDS2q2+Djj9BlW5LlizB7XazevVqzjrrLKyV4cAvv/wSgKlTpyJ2sQ+HW9YwWwI7yciK\nyuc7clAqT8whJonpgxMCPknJqkqIRVfnLUVVVcoqHJTZnbjcHrRKPwlREBGNXetzebI4XG5y8ovJ\nzi8ip6Ck8ncxOQXF2B0ufn7j8WAPsV3QFAUBEMMivNUdrRhWL844jmg0VJofNX8qT9M0jm/5jYM/\nrOXo+l+Qnc4m97nk5YVYG+kx4jt2G1Rm+KYvoNLTwYhoNGIwGjHabL7STH3qouMR1KtOYWEhN954\nI7t27eLDDz9k0KBBAKxevZrVq1fz7rvv8uqrr/p5X3R2DIEmVgBrDxSQV6Pt+YzBCT5nzaZQVA2L\nUcJi0EPyzUFWFMrtTsocTtxuDwgCBknSjakAt0cmt6CY7IJisvOKyKoUEtn5RRSXNZzYdyrgm/II\nDUMIi2iVu+XakR4NtUlBEdOjB/kH9te7rjgjg68ef7TedZbwcLqPHU/KiBF8849qH4r2mjZQFQVV\nVhBEkAxGJJMZc5gFU2X0QZ++6FwE9a/1r3/9i8OHD/P3v/+dPn36+Jb/3//9H1OmTOHBBx/kP//5\nD/ffH1iZVVfiWGEFG49Wm9WMTIukR2xgAkvTNAyi1wCrszcVa+tqDvBeMIvL7Dhdblyy7KvgOBXF\nhKZpFJbaycorJCuviMzK39n5ReQVlQZqV3BqIcvektHImFYRFKosYy8soDQnu9n7Vhld1Zdnobj9\n84CMNhvpY8fR48wJJA0dhihJuOzlLR53oKiKgiYrUENEWMItmGwhSCa9FL4rENQz5w8//MCcOXO4\n+OKL/ZYbjUamT5/O7t27WbVq1SknLNyyypc7q+1vY0NMTOwTE/D+qgaxYZ3fAKstqzlcbg/FZXbs\nTjeKomCQDIiigKkBA6GuhkdWyCkoJvNEIZl5hWSeqBQSJwpxuj1NH6AejAaJbrGRdIuJJNH3O4rE\nmPbv9dAeeBMwBcTYBETzSWYDA+6KCuwF+XgcTiSjodl36ZqmUXjsKBFJKRRnZPgSRGsiSBKpp4+g\n16SzSR0xEkMrjDtQBASviLBYMIWEePMgdBHRJQmqsCguLiY+Pr7B9YmJiZxoqmlNF2TtgXyKHd6T\nuyDA9MEJAVeReBSVuFBTlygrbYtqDoAj2flYzUYMktTlfSVcbg+ZeYUczy3k+IkCjp/wioicgmJU\ntfnhB0kUiY+OICmuUjTEVv2OIjo8tP5Ok10xzCF7vImZ4ZEnfXF0lpRgzy9AlWUkkxFDM8Wzq6yM\n/d9+w56vVlOalVnvNtE9etJ38mR6nnkWlvC26XCpaRqqxwMaSCYjksH/dUSmpmIJa5/umjrBJajC\nIj09nW+//Zarr7663vU//vgjqamp7Tyq4JJZ7GDzsermPWPTo0iMqCcTux5kVSXCasSo51UA3uTL\ncoeLnIJaLoRdsAeHyyOTdaKQY7n5ZOQWcDy3gIzcghZPX0SE2kiOjyYpNoqkuOqfuKjwU7qU1uuF\nICLGdUM0tc7dvr2wEGtEBJKpedGy/IMH2P3lFxz66UcUt7vRbc9/4smADLIkg7eaQlNVBFFs0AIc\nvBEbVfZ6QxhMZoxWM+bQMIwWC4Io4iwrbdbr0ek6BPXsevnll/PYY49x3333MWfOHNLT0xEEgaNH\nj7J8+XK+//57HnjggWAOsV1RVI1Vu6ojNDEhJs7oFR3QvqqmYTaIASd3dlVUVaW03EGZw4nL5UEQ\nBWS565TcKqpKTkExx7K9AuJYTj7HcgrIKShudlmnKAp0i4kkOT6a5LhokuOjSIr1/g6xBiZmTyU0\nRUYwWRFjYlscpVA8Hkqz/HMnxGbcCKiKwpFf1rFr5eec2Lun7gaCQPJpp9Fzwln8+MK/mz0+g9nM\ngGnT2f3lFwyYNt1vqsRbpaF6EyxNFkwhoVjCwtp1OkWncxDUq9CVV17JwYMHWbp0KZ999lmd9Vdf\nfTVz5gTHqS0Y/HqkyK8K5PyB8RgCDNNrUG+3UkkUECrXiwLNtgDvDGiaRoXTRXF5BQ6HC6GymZfR\n6P14u2hZzkCwKbU7OJqdx9HsfI5m53EsxysmPHLzEnKNBomkuGhSEqJJiYsmJSGG5PhousVE6tGt\nQJFlxIhIxNCWhfLdDgf2/Dw8Tiey3Hh0oSF2rfycfV+vqbdPhyU8gr6TJ9Nvyh8IS+h2UkmYY2+4\nibE33ISmqsguN6IkIJktWMPCMYeEBCwk/KIfktRo9EOnaxH029sHH3yQK664gu+//57MzEwURSE5\nOZmzzz6bvn37nvTxc3JymDFjBq+88gqjRo1qcDu3280bb7zBJ598Qm5uLgkJCVxwwQXMmzcPYzs0\nzil2ePj5UKHv8Wkp4aREBWak5VFV4sPq7wFiMoic2TuGnw8WcEavmC5lL+1wuSkus1Ph9J6ovWKi\n8528qqIQR7LyOJqdx5GsPI5k51FUam965xpUCYjUhBjfT0pCDPHR4V06j6Qt0VQVNK1y6qP5icPu\nCjvleXkoLjeS2YTBZELxtExYbHlveZ1lcX37MXDadNLHn9EqDb6qjKcMZjOm0BAs4eEtjkgYrVaG\nX36lr4laY31OdLoWQRcWAL1796Z3796tftzs7GxuuOEG7PamT9BPPvkkn3/+ObfeeitDhgxhx44d\nvPTSS2RlZfHUU081uf/J8u3ePGS12ghrUp/YgPZTNJUws6HRyMbMYYnMHJbYKuNsKa1VNirLCkWl\n5ZQ5XGiqisFgCPqcf5ndwdsrf/BbVl7hrHc6wSPLZOQWcDjzBIez8jiceYKj2Xm4PIE7dwqCdwoj\nrVsM3bvFktotltRuMXSLjkTSW6W2GpqqehtRxcQ3u4zUba8UFB6PN5GxmQmZJVlZbH3/vXrXCZJE\nzzMnMGjGBcT27lPvNoGiqSqKR/Zaglus2KJjMIeEtJpr5Tn33h9wV1adrkPQhYXL5WL37t1UVFSg\n1iiPUhSF8vJyfvnlF5588slmHVPTND7++GOeeeYZ3+PGKCoq4oMPPuDee+/l+uuvB2Ds2LEAPPfc\nc9xzzz1ERQVuo91cDufb2XeiWvyc3TcWi7Hpi6WmaUiCSJilY9+lt1bZaEZuAUajhEE0eO/AO8Bd\neJndwaOLPuBYToHf8qff+oSHbryEglI7h47ncijzBIczT3AsNx9FCTznI8xmoXtinPenWyxpibGk\nJMRgNgb9q9ul0WQFwWZFigpM4FfhrrBTfqKGoKiVkOkqLWPDm6/771Oj82hRxjG2ffA+h3/6sc55\ny2izMeD8aQycNh1bdODl57XRVBVF9ooJc2gY1sjITtnOXKfjEtSz0759+5g7dy4FBQUNbmMwGJot\nLPbs2cOjjz7K1VdfzdixY5k/f36j29vtdq688krOOeccv+U9evQAICMjo82EhaJqfL03z/c4JdLC\noMTAGvXImkZCaMdPnGpJ2WhV3kRtOprPxLurfq4jKgCOnyji5n+8HnBFhiBAYmwUPZLi6Z4YS3pi\nHN2T4ogKC9Fr/dsbWUYMDUcMwMq6Ct+Uh7t+QQFeUfHlww/U23l0wq23s/OLzzn8808Nludao6IY\ncuEfW9TIS9M0FJcb0aSLCZ22J6jC4vnnn6e0tJR58+YhCAKvvvoqDz/8MKWlpXz00UdYLBb++9//\nNvu4SUlJrFmzhoSEBDZs2NDk9ikpKTz88MN1ln/zzTcYjUafwGgLth4vocBe6VkBTO4fF9CFRFFV\nwi3GLpeM6XC5KSq143C5Ka9oupdBMNA0jdzCEg5k5LBx18FGtqt/uUGSSOsWQ4/keHokeX/SEmOx\nNLPcUKcNkD0IEVEBJ2nWF6FYv/g1X1XF2Btu8m27aenbDXYeXbngviafqzQzs8EOpU1hMFuISk3T\nKzh02oWgCostW7Zw6aWXcvfdd+NwOHj99ddJT09n/PjxXHXVVcycOZNPPvmk2a3TIyIiiIg4Obe/\nNWvW8Mknn3DNNdcQFta8O4SKCv++CQ6Ho97tnB6Fnw5W3+0OTQ6nW3jTZX5ey26R0C5SWqqqKsXl\ndkrKHSiKirEyb6KjVCxUOF0cPJ7LvmPZ7D+Wzb5jOZTZ6/+b1sZokEhPiqdncjy9khNIT44jNSEm\n6HkhpwqBfhcBrzV3dByi1dbkcT1OJ2W5Oagej7ezZqUolF0udn/5BZqqsvvLLxh5zRzfxbyxzqM1\nkUymBn0panYobQhVllHd/jk74QkJHUJU/PL6IgwmE6PmXNeqxz2w9jv2ffM10x5v+3y49mD7xx9S\neOQwk+66J9hDaRFBvTLZ7XYGDBgAgNVqJTk5mV27djF+/HjCw8O5+OKLee+995otLE6Wr776ir/8\n5S+MHDmSe++9t9n7Dx8+PKDtfjlchMPjnW83SgITegc2b6qoEBPeuS27NU3D7nBSUu7A6XIjCCKS\nJCIGuWqlKhqx92gWe49ms+9oFsdyClrU+nvM4N7cddV0PaEyiAT6XcTjQYhNQKyvLXgNFI+Hstxc\n3BUVGMwmxFrTCYrs8Vlpe3MZPAFf0COSkxl++ZXs+PQTCg4eCGzcNZ/b7UY0GrBFRxMS27zckPbi\n51deYvz8W1r9uJuWLulSU4brX3+NtFGjgz2MFhNUYREZGUl5eXW9dUpKCvv3V3fmS0hI4HCAKr+1\neOutt3jmmWcYO3YsL730EqYWlJgFQqnTw6Zj1U3GxvaIDigCoagq4db2mwJpiyZgeUWl5BWXoqqa\nNzrRhi6YTVVsKIrKkewT7Dmcxe4jmew9mhVwp87E2CjSE2PZcySTolr7pMRHMf/iybqo6AzIMkJ8\n406amqpSduIEztLSFtlua5qGsRHTsW6DBjP1sScQJYnsHdsbFBYxPXrWGZfqkTFarYQmJ2OyeRsV\nehyOjush0RUt3tuAltzMdBSCKixGjRrFihUruOCCC4iNjaV///6sXLmSiooKbDYbv/3220lPaQSK\npmk89dRTLF26lBkzZvD000+3+IK3ZcsWv8cOh4Px48f7LVt3qBClsrw01CwxunvTiWJVUyDt5a7Z\nWtUcaq1mSBVON2EhVtr6mttQxcZjiz7gzNMGcCjzBHuPZgXUdMtmMdMnrRt90xLpk5ZI79RuhNks\nvud5e+UPrP1tl2/7v143k7AQvW4/2DT5XZRlhJj4RkVFRVERFYX5CKLUbEEBUHjkMOtfW0TOrp31\nrg9PSeHc+xYgVk6PNdShNCI1lZHXeA0DNUVB01TMYeGERMfUaVjWXA+JTUuXsP2jFZTmZBMaF8+w\nWZcyava1vvVH1v/CL4sWkn/gAKJBIn3ceM6648+EJXQD4PfPPuWrJx/jyjfe5ttnnyFv315s0TEM\nv+JK33H+b8QwANYtWsi6RQu5Z/M2APIO7OfHF/7N8a2/AdB91Bgm3v0XIpNTAPj0L3dxeN1PzHnv\nA6K7pwPw88KXWb/4NS5b+Bo/L3yZ479t9j3H5YsWkzpiZJ3X+L9HHqQsN5eotDR2r/ofYfEJXPfB\nR6Bp/Pr2m+z4+CPKTuQSnpjI8Muv4vQrqiPl7910PeEJ3YhKT2fLe+8iu92kjRrNOffcT3hidSl/\nU68FIG/fPta9+grHt/6Gq6wMW3Q0fc+dzFl33IXBbGbR9KmU5mSzc+Vn7Fz5GfNWrvJ7js5AUIXF\n/PnzueKKKzjnnHNYu3Ytl19+OW+//TYzZ84kISGBjRs3ctlll7XLWJ5//nmWLl3K3LlzT7qbqs3W\n+BxtUYWb7ZnVPvrje0YH1GSsvatATrYJmMPporDMTn5Rmd/yehtVtQENVWzkFZfz8fcbG903KS6K\nft2T6Nc9ib7dE0mOi25w3GEhVuZeONFPWITadEvsjkCj30XZg5CQ2OD0h+LxUJKViSoriC2849+8\ndAl7v1rt32lUEPzu2v/wwMN+lR7m8DCmPf53Nry5mAPff+tbPmXBQxhtNlRFxRYVhTUyqtHwf6Ae\nEt//v+f4bfkyRs6+lvQxY8ne+Ts/vPAvVFlmzNwb2Lnyc/73yIMMmDqNsTfcREVREetefZl3r5vN\n7Hf/iy3K23ZAU1U+v/9eRs6ew4Tb72THxx+x9l/PE9e7D+njxnP1W0tYdt1shs68mCF/9Ha0Ljx6\nhOVz5xDToyfTHn8KRfaw/vXXWD73Wq797wfYoqKZ/LcHeWvWxax56gkuX7SYnF272PDG64yacx2p\nI0Zy3oIH+fLBBSAITF7wtzpRnZoc/20zBrOZmc/9yxvVEQS+euoJdn7+KWNuuInkYaeRsWkj3z33\nT1zlZYy7cV7ln0zg4I9rse3cwbl/fQBNUfjhhX/z33nXc90HH2O0WAJ6LeV5eSy/8TqShg7j/Mef\nxGA0cejnH9m0dAkhcfGMue56Zj7/Lz68/TYSBgxg3E3zO+y0VmMEVVj079+fFStWsGzZMqKiooiK\niuK5557j2WefZdeuXUyfPr1FOQ5NUV5ezoEDB0hLSyM6Oprdu3fz2muvMWTIEKZOncrWrVv9tu/d\nuzehoU038AmUnw4WUtVYMtJqZFhy01EZRdUINRk6fBWIoqoUldoptTtQVW8iZkuSMJtjOuX3/IrK\nocxcdhzI4OdtewN6LkkS6ZWcQL/0JAb0SKZf9yTC9WhDl0eIiKo3UVPTNOx5eVQUF3vzKE4iiXjP\nqv/5PU4dNZrTLr+cz+/5i2+ZKbzuucUcHsaYG2/0ExbGUBuhsbFYWjGK6ywr5bflyxh+xVWcdfud\nAKSNHkNFQQHHt/zG6Ouu54cX/h89xp/B9Kf+4dsv+bThvDlrJhvfeZuJd94FeN+3cfNuZshFM73b\nDDuN/d99w6GffiB93HgShwwFIDQhgcTBQwD4ZdFCjFYbly58DVOlCOw+eiyvXXA+G99+i4l/vpuQ\n6BjO/esDrFxwHzs++YhNy5YQ16cvZ97irY6J6dkTY4i3LLvquA2hKgpTHnyY0DhvV+3Co0fY8clH\nTLj9TkZfO9f7/GPGgiiy4Y3XGX7ZFVjCw9E0Ddnp5JL/vOKLPkSn9+Cdqy5n18rPGTbr0oBeS/6B\n/ST0H8CFzz6HqfKzlzZ6DEfXrydj00bGXHc98f36I5mM2KKim3w9HZWglxX07t2bRx55xPd46tSp\nTJ06tVWfo7aq37lzJ9deey1PP/00M2fO5KuvvgLg999/5/LLL6+z7zvvvNOoHXhzKLJ72JVdfQd/\nZq/ogMSCIECYJeh/rgZxOF0UltpxON2IknhS7cgbM516/ObL/aYYNE0jO7+YbfuPsmP/MX4/mIHD\n1bRlstVs4oIJpzOwZwq90xJ1w6lTEDGkbrWXx+mkNDsLNK1F0x6OkpJ6l4cnJjLmhnmkjhjR4j4e\n0d3TW73tePb27aiKQt9zzvVbfvY93vLXgkOHsBcU0P8P5/utj0xJIWnoMDI2+0f+koYO8/1fMhqx\nRkbhaaQS5+ivG0gbNRqD2YwqeytZTDYbyacN5+iG9b7t+k/5A/u+WcNXTz6OwWxm9rL36kz/BII1\nMtInKgCObfwVTdPoNWGi7/kBep01kfWvL+L4ls30nng2AIlDh/lNacT3609kcgrHt2xm2KxLA3ot\n6ePGkz5uPIrHQ/6hgxRnHCN//34qigqxRgbum9LR6RBn07179/L999+TlZXFnDlzsNls7N+/n7PO\nOuukjz1mzBh2795dZ9mePdWdAe+8807uvPPOk36uQNhwtJCqIGi0zcjAAMywPIpGTKixw2U9V3US\nLS6vQFYUb3SiFS7QjZlOvbvqZ66Zdibb9x9j276jbNt/lPzisnqO0jhnDOvHpeeNO+mx6jSMoqqo\nqoqhEySwapqGPT8fR3ERUgsTtg/9+APrFi30WyYaDAy9ZBZDL56FoRnH9bYkb/uuvFVCyBZdfxdl\nZ6l3fX3heFt0DHn7/DusGmtNKwmigKY2nIToLClhz+pV7Fm9qp7j+49p0PQL2Pf1GqK6dyeqMtei\nuZhqTY05i70J9G9e+sc62wqCQHletXlhWHx8nW2sUVG+9zCQ16KpKj/+5wW2vP8eHoeDsIRuJA4e\njMFs7tTJmrUJurB46qmnWLJkCeD9Q06dOpXy8nJuv/12Jk2axAsvvNBmlRnBYFdOOZLR++Ub1yMa\nsQmxoGkaZoOAuYN4OoC3X0dBSRllFU5EUUASJYytWNlxOOtEg+t+2rqHbzf9jtrIyUqSRPqmJdI3\nLZFfduzjRGGp3/qU+CiumnpGq433VEJVNRRVRUPz5gkIICAgiiIGUUCojFQZJBGjwetFYu7gxl+y\n201JZiZoaotEhbO0lF8WLfS6ZtZiyt8eJum00wI+lqZpKB4PtsgobDEtt+0OFHOlR09FURFRad19\ny0tzcijOOIa10nHYnp9XZ197fh7WyJNzJDaHhZE+dpwvKbUKTfNvJ+9xOPjuuWeJ692HvP372bTk\n7RZ5YdS+eJsrI0CXL1pcR3RoGoQndvM9rigqqnO8isICorp3D/i1bHhzMZuXLeG8Bx+m7zmTMYV4\nq3iWzr6q2a+lIxNUYbFs2TKWLFnCtddey3nnncc111wDwOjRo5k9ezZLlizh9ddf59Zbbw3mMFuV\nqs91hNUQULRC0SA2pGMIK4fbQ0l5BQ6Xp/Ki0f4fn4YqONK6xTK0TxpD+3RnQI9kn4vlRRNH6hUb\nAaBpGqqqoWoqVSE1QRQQRQGDKCJKIgZRRJIkTEYDJoPk9R05iSmvjkBFURH2/PxKg6vmi/es7dv4\n4d//j4rCwnrX//LGa8x48umAbLhltxuzLYSolFREgwFnWWmT+5wsiYOHIBkMHFz7PcnDqgXQxnfe\nYu+a1dzy1beExMSye9X/GDTjQt/64uPHydq+rc5FtClqNzdLHTGS/IMHievbz1cVo2ka/3v4b0Sm\npBLftx8AP7z4b8rzTnDdfz9ky3+X8/PCl+k5YSIxla7IoigGdMdfO+qbMmIEABVFhX6VJEfW/8Lm\nd5dy9t33+JJTs7Zvo6Ko0Pc4Z9cuSrKyfL4cgbyWzK1biO3Vm8EXXOR7rrITueQf2E+3GvkUrdUE\nLlgEVVi89957nHvuuSxYsIDCGl/MiIgI/va3v1FUVOTrONrVGNej6dwKRdUINUtNRjXai+y8IiLD\nQjA1Y7qjOUmYhaXlbN51KCAfifAQK8P6dmdYn+4M7ZNGVD0JcKBXbADIiuIt+RVA0PDmwAgiouiN\nNFSJB5PJgMlgwGCQkESvaOjKlGZnY9C0evt6NIWqKPy2/F22f7SiUV+GkoyMJm24VY+CIAhEpXXH\n2M7umLaoKE6/6ho2LX0HyWgk5fQRZO3YzrYV7zPp7nsQBIEJt9/Bqkcf5ou//ZUB06bjKCpm3aJX\nsEZGMeKa2Y0/Qa23xhIWRubWLWRs3kTqiJGMu2k+7143m4/u/BOnXXoZktHE9g9XcGDtd1z47PMA\nZGzexNb33+PMP91BZGoqZ9z2J/Z9+w2rHnmQq95agiCKmMPDyd6+nWMbfyW+X38s4fXnotQWH3G9\n+zBw2nS+euIxSrOySBgwkMKjR/jpPy8QkZLqN+UiO518+KdbGTdvPu5yOz++9AJxffoyYOo0gIBe\nS+LgIfzy+iI2vPUGSUOGUpxxjA1vLMZos+Gp4RJrCQsnd89uMjZvotugwXWmmDo6QRUWR44c4aqr\nGg4BjRo1ypdY2ZUIMUkMTgrMJjxYtt2yrHCiyD8RrbnVHYEkYWblFbHh9/1s3HmQ/Rk5jR7PZJCY\ndubpjB3Smx5JCe1WttrRURQVRVPxnsUFDJXCQTKIGCUJs8mK2WSqjDiIHS5XJ1h4XE5fKLo5VBQW\n8N1zz5K7a1fTG9O0DXdoXBwRSUnNHkdrMfHOu7BFR7NtxQdsfOctIpJTmPzXBxh68SwABl9wESZb\nCBveXMynf7kLU0gIPc44kwl/uoOQGl1W6/1c1Vo05vqbWP/6q3x055+Yu+IT4vr05YrFb/HTSy/y\n5UN/A00jtncfZj7/b3qdNRG3o4JVjz1MXJ++jK6c+jBZbUz+6wN8fNcd/PrOW4y57nqGX34lubt2\n8dEdtzH10cfrJJtWDaa+MU599Ak2vLmYbSs+oCw3B1tMLP2nTuPM2/7kt33S0GGkjxvP6se8xQa9\nJp7NpLv+4ksibeq1eF//jTiKi/nt3WX8UvYKcX37MenueyjJPM6PL72Iq7wcc2goo+Zcy7f/908+\nvP1WLn1lkV80qTMQVGERFhbWaGfTo0ePNrtPR2dgRFokhibuBmVVJdLW/gmbFQ6nt7rD5cHtado4\nqjEaS8J84vWPkBWFjNyG//6SJPq1GH/+7jkkxHSdzOlAqSkchCrhIEm+XAaTwYDFYsIgiXoPkmYg\nis1/r3J2/s53//dPHMXVrrkRyckIokhxRkaLxmEOa71S9pYyava1foZYtel77mT6nju5wfWDL7yI\nwRdeVGf5vJX+iYwjr5nNyFpRjoT+A7jkxZfrPa7JauOmz7nLP4AAACAASURBVL6ss7zXWRN9BlsA\naSNHMe+LukmTNTn/sSfqXS5KEuNunOfzrGgIQRAYd9N8xt3UcLfsxl4LeCtlzr1/Aefev6DOuhFX\nV78v/f9wfgPiqHMQVGExceJEli9fzsUXX4y1ljPctm3bWL58Oeedd16QRtc2GCWB4alN16EbRBFr\nO5VAappGaXkFRWXV1R0mowGPLDe9cyM0loTZ0Lr46HBGDezFyIG9SE2I5sYnFvnWddUpDJ9w0Lxl\nxVJVxEHyCgez0YjZZKi0P9eFQzDQNI09q/7H+tcX+Zld9Zo4ifHzb+HXNxc3KCxqGjZpqoqqtH21\nh07r05WqNtqaoAqLP//5z6xbt46LLrrI1yxo6dKlLF68mHXr1hEREcEdd9wRzCG2OoMTw7AaG784\nyJXlpW2NrCgUlpRT7nCiaVR2FG29j4RHVgLarntiHGMG9WL04N6kdYv1RWnsjo7ZNr25qKqGrCiV\nUWHNLxHSIIlYTEbMZj3i0FFRZZlfXlvE3q+q74hFg4GxN82j33l/QBCEgGy4FbcHc1hYu1R76LQu\ngiB4Vb9OQARVWCQkJLBixQqef/551qxZA8A333yD1WplypQp/OUvfyElJaWJo3QuzujV9EnFKAmY\n2vDO1OX2kF9ShsPpRhIlpBaEhBuirMLJum17+XHLbo7l5De4XXiIlQvPGsGYIX3o1gWmN2qXvyqq\nhijgq6KwWUwYDRIGSdJzHDoRbrudb599hqxt1W68tuhozrl/ga9iARq34TaFhaK43YR1S8QSFtYu\n1R46rcvlixYHewidiqD7WMTFxfGPf/yDp556iqKiIhRFITo6uk07XgYTcxNtwT2qRlxI20Qryioc\nFJaU4/EoGI2GVotOyIrClr1HWLt5F5t3H0ZWGo9UdIsJ5++3XdUpSz5lxds/paq6wiCKGAwSUREh\niIKAqmlIokif1IQWNWzT6TjYCwr46olH/aIQcX37cu79D9RrKNWQDTea1/5ZMnZsPw8dndaiw1y9\nRVEkplaI8NixYzz00EO8/fbbQRpV+2OShBb11miMsgoHBcVlyKqKUQrMHTOQMtHME4V8u+l31m7e\nTUl5/SWiibGRmI1GjmRXG+w8eMMlHV5UVJVoCoBkkDBKXrMni9mKxeSNPtQux7zsvHG+FvO6qOjc\nlGQeZ9Vjj2Cv4bzY48wJTLj9zmY5aBotNqLT0/Uolc4pRbsLi+3bt/Pyyy/72hkPHDiQ22+/ndNP\nP923jaqqvPnmm7z44os4nV1jnj0QWjNaoWkapfYKikorUFTVmz8RoLVyY2WiD95wCTsPHWfNhu3s\nOZJV7/4RoTbOPK0/Z53enx5J8VQ4XVz36Cu+9R0pCVNRVK8pFPhyHExGA2aTDavZKyACvSjcfc10\n7r5melsOV6cdKDh8iNWPPoyztHrKYujFlzDi6tnNNi4KS4iv8/mRDEYEUURTVQRJQmqgc2qg2+no\ndDTaVVhs3LiRuXPnoigK3bt3x2az8euvvzJnzhzeeustRo4cSWZmJnfffTfbtm0jNDSUxx57rD2H\nGFRM4slHKzRNo6CknFJ7hS8hs7kJgY2Vid7+zzfx1DPVYZAkRg3sycQRAzmtbzpSB+4PoagqogBG\nowGryYjVYm6WgNDpuuTt38/qxx/BXV7dKGzMDTcxaMYFAe0fSOWA0Wpl+OVXsuX99xh+2RUYrfVH\n7wLdTkeno9GuwmLhwoUYjUafiADIzs5m/vz5PPvsszz99NPMnj2b/Px8Jk6cyOOPP05CQkJ7DjFo\neBSN2JOoBFFUlaKSckrsDt5Z+QOr129j6rjTmHvhpGYfq7Ey0dqiontiLOeOGsyZwwcQ1kEiEVXN\nr9BAMoh1HE7TE+M6/FSMTvuTf2A/qx59yOeAKIgiZ93xZ3pNnBTQ/t4y1MDE6Tn33s85997fatvp\n6HQk2lVY7Nq1iyuuuMInKgASExO55557mD9/PnfddRd2u50nn3ySWbNmtefQgk5LK0FkRSG/uAx7\nhRNB9BpKrfplG5qmseqXrVw19YxmN4FqrMEXgNlo4MzT+jN5zBB6pSQE9U5fUVQU1WuJbJQkDEaJ\nMLMZm8WCqTIPoszecNtmHR2AgsOHWfXYIz5RIRoMTLr7HtLHjQ9of1VWMJjNRKXqpaQ6Ou0qLMrK\nyujTp0+d5X369EHTNLKzs1m+fDn9+/dvz2EFHVlVibI1L9lPVVXyikopszu9fgiVFR5Oxe0Lx1b5\nJ5gJTFgUldn5389bGnXD7Nc9iQVzZxJibd+eBlV4ZAVN0zAYRExGAxEhVmxWfSpDp+WUZGWx+tGH\nfdMfgiRxzn1/JW3U6ID2Vz0yprBQwhO66aWkOjq0s7CQZRljPSVX5srGOzfeeOMpJyoAJEHA0oRp\nVhWaplFUVk5hiR1RFAOq8GiK7PxiPl27kbWbdzdaKpocF8n9114YNFFhNErERYURYjXrRlI6rUJF\nYQGrH3sYZ6m3L44gipz9l3ubISo8WCKjCI2Nbcth6uh0KjpMuSnAoEGDgj2EdkdRVSJsgUUUSu0O\n8ovLAK1VPCgycvL58LtfWbdtX52kM7PJSGxkGJknqrvOLpj7xzbPTZAVGVXTvOWdtcRWt+hIPTdC\np9XwOCv45onHKD9RnVM04fY7A57+UDwytugYQmp4WuiVHDo6HUxYnJKhbEFosieI0+Ump6AEWVUw\nSq0jKD74ej2/7NhfZ11UeAjTzxjO5DFDAa3Ny0S9fhEaJoOEyWQgJiIEq8WMpOdG6LQxP77wb4qO\nHPE9HnP9DfSedHZA+3pFRbSfqAC9kkNHB4IgLA4ePMjGjRv9lpVW1ovv3bu3XsfNUaNGtcvY2htF\n1QgxNxzSlxWFE4WlVDhdGA2GkxYV2flFvL9mPT9v20PtqrhuMZHMnDSKs07v74uGtFWvDo+soKgK\nZqORyDArIVZLm05tGA2SnytmaxuQ6XROjm/5DWPlzcyQmX9k0AV1u3PWh+qRsUZF1REVVeiVHDqn\nOu0uLBYuXMjChQvrXff000/XWSYIArt3727rYQUFDQgx1/0T1EzMNBhOvjFYYWk5H6xZz7ebfq9T\n8ZESH80l545h3NC+SM00/wkUuVY3x+T4KGIjw9vkuerDYjbprpg6DZI2ZiwjG2kZXhPFI2MNDydU\nbySmo9Mg7Sosbrvttmbv01WnRzRNw2r03knXpKi0nIKSciRRapXEzI++/ZVVv2zD7fFvgZ4SH82l\nk8cydkhfRLH13+OqlusWk5GYiBC/deYg9EzQXTF16iO6Rw8m/vnugBw1VVnGHBpKaHx8O4xMR6fz\n0q7C4vbbb2/Pp+vQyKpGjKX67ffIMll5RciK0qqtyz/7YbPf44ToCC6fMo7xw/q1aoRC0zQ8ioIk\nClhMJmIjQ7FZzAiCoOdK6HRIzKFhnPvXBzBams4d0lQVg8VKeLdu7TAyHZ3OTYdK3jyVMEnVjpCF\npWUUltgxGgwYTiKPoszu4MX/rq53XXiIlVmTxzJ59JBWyzHQNK9PhiiK2CwmEkIjsLbiNIOeG6HT\nlky4407C4pt29tU0DUGUiEhKaodR6eh0fnRhEQQUVSXSZkKWFbLyi/DIJx+lyM4v4oH/LKfc4aqz\nbvqZw7nsvHHYLK3jP+GRZURRwGo2kRDWtJhoqUDQcyN02pKkIcMC2k6TFaJ6pHXZaVkdndZGFxbB\nQACXy8XxotIWNQmriaZp/LR1LwtXrMEty/Vu43LLrSIq3LJMuCgQHRNBqNUS8In2ZASCnhuhE0wU\nt4fItDRE3ZBNRydgdGHRzqiqRmlpBXbNc9JRivziMhZ9/A1b9hxudLvGmoo1h/TEOCLDQpresB50\ngaDT2VDcHsK6dcNoDo7TrI5OZ0UXFu1MRl4h3SLD6vXrCBRN0/h+0y7e/Px7HC53K46u+vgeRcZQ\nq/V5W5Wj6uh0NFRZxhoZiSUsLNhD0dHpdOjCop0xiVKLuphWUVJewcIP17Bp1yG/5dHhoRSWlte7\nT4+kwMrjFEVF1VRCbRZSIqLbRLTo6HR0NFXFYDIRGhcX7KHo6HRKdGHRzlgMLU8A27r3CP95fzUl\n5RW+ZTaLmetmTGTEgB489toKjuX4dyZNiY/iqqlnNHpcRfU2HgsPtREdHoJYFZmomwd6ynLR3c8y\nYkBPHr7pkmAPRaeN0RSV8LTkYA9DR6fToguLdsZsaP50gqwovLd6HZ+u3eS3fFif7twy6zxiIr3h\n2kfnXcrbK39g7W+7fNv89bqZDTbuqupkGhVuIyosVM96bxQB/d3p+shuN5HJKXqypo7OSaALi/am\nmRfvwtJynl/2BXuPZPmWGQ0Ss6dNYOr40/zEQFiIlbkXTvQTFvU1DpMVBUEQiAkPJSLM1qCg0H0k\ndE4lNFXFGhGByWYL9lB0dDo1urDowOw+fJznln7hN/WRHB/NXVdNp3tibLOPp6gqmqYFHKHobD4S\nn3y/keWr15F5opDo8FAuOOt0brjo7OqpncptPvr2V45m56NqGt27xTL3wkmcO3owACt//I2n3viY\nv153Ea+sWIMsK7z24Dy8nV28XPvIyxgNEq8/NN/v+W97+g1EUeDF++bWO75yh5NXPljD2s27KKtw\n0iM5jpv+eC5nDOsHwJhrH+TGmWdz0x/P9e2z6KNvWPzpd2x4+0nfsi17j7BwxRp2H87EZDQwYXh/\n7rzy/BZX7OhUExqn23Xr6JwsurDooKzZsJ3Fn3yHolY38Jo4YiA3zjwHi6l5vTY0TUOWFcJDrcRG\nhvldaJuis5SJvvX5WhauWMNlU8Zx5rB+7D2azaKPvyG3sIQHb7gYgA++Xs/zS79g3sXnMqxvd0rK\nK1jyxY889Mr7DOmdSnx0BOAtCX531c88fOPFFJdX0CPZ/2Jz0cSRPPP2ZxzPLSAlwduMKregmN/2\nHOaxmy+td3yKqnLHP9/iWE4+N18yme5JcXzx4xbu/dcyXv7r9ZzWLx2ovzdOzSW/7TnMn555kzGD\ne/GP26+kpLyCVz/8mlv+sZi3Hr0FczM/GzpeFI+byBTdBEtHpzU4ZeoHc3JyGDlyZJ2W7Y2xa9cu\nBg0aRFZWVtMbtxKKqvLGZ9+x6KNvfKLCIEnMu/hcbrt0SrNFhUdWMBol0pPjiI+OaJao6CyUVzhZ\n/Ol3XHzOaO6+ejqjB/dm9vQJ3Dt7Bp//8BuHM70+Hll5RcyePoG5F07i9P49OHvkIO679gIUVWXb\n/mN+x5x7wSTGD+vHtDOGVy6pvuBMGTsUs9HAlz9v9S378uethFjNTBoxsN4xrtu2j52HjvPIvEuY\nNXksowb24tH5sxjYM5nNuw/Vu08VNfvRvvzBV6QnxfH83XM4o3J8L943lyPZeXxeqy+MTmBoqool\nPCKgniE6OjpNc0pELLKzs7nhhhuw2+0B77Nv3z7mzZuHqqpNb9xKuNwe/r38f2zcddC3LDLMxr2z\nL6Rv98QWHbNbTAQJMZGtNcQOyY4Dx3B7ZCYM7+9LSAU4c3h/ADb8foAeyfHceeX5gLenypHsPI7n\nFrKp8qLuqeVa2tj7HWqzcPaoQaxat5V5F3unLb74aQuTxwzB1EBH2m37jmI0SEwYPsBvee3plMZw\nutzsPHic2dMn+EWykuKiSE+MY8POA8yaPDbg4+l40VRNnwLR0WlFurSw0DSNjz/+mGeeecb3uCk8\nHg9LlizhxRdfxGQyBbRPa1BW4eQfb37C/mPZvmU9kuO5f86FvqqPltBa/UE6MlU5KH9+7p066wS8\nDqUAx3ML+Mebn7Jp9yGMBon0xDh6p3q7Vdb+M9uayCe5aOJIVq3bxrZ9RxEEgYzcAh6bP6vRMYY3\nUJ0TKKV2B6qm8fbKH3h75Q911rdmA7hTBcXtISI5SZ8C0dFpRbq0sNizZw+PPvooV199NWPHjmX+\n/KbvDteuXctLL73EzTffTHR0NA899FCbj7OgpJwnX/+Q4ycKfctGDOjJn6+a1qypD1VVUdX2EUId\niarKlyduuYy0bjF+6zQNoiNCUVWVu55/B5PRwNuP3ULftEREUeRQ5gn+t25rfYdtlNP79yAlPpqv\nf92BIAikJ8UxqFdqg9uH2Sx+SbhVVFX79Ev3ds6sHSGraVIWYrUgAFdNPYMp44bWeZ3NnSY71dE0\nFVNIOCabnvSqo9OadL0J9xokJSWxZs0a7r//fiwBzp8OGTKE7777jvnz5yO1Qy17XlEpDy/8r5+o\nOG/sUO6dc0GzLhSyLGO1mOpcWE8FBvdKw2iQOFFYQv/0ZN+PQZJ4+YOvyM4rori8gmM5BVx41kj6\npyf7ck1+2b4PCCyaVZsZE05n7ebd/LRlDzPOPL3RbYf3T0dWVN/zgVdEPP76h7y1ci0AIVYzuQUl\nfvtt23fUl90RYjXTLz2Jw1l5fq8zPSmexZ98y297G+8Zo+OPpiqEd+sW7GHo6HQ5unTEIiIigoiI\niGbtk5CQcNLPW1Hhf2fqcDjq3e5EYQmPLlpBXlGpb9msc8dw2XnjmhWa9Xhk4qLDiQi1UWav/7m6\nMpFhNmZPm8CrH36N3eFieP8e5BWVsuijrxEEgT5piYRYzSTGRvL+ml+IjwonNMTCL9v38+XPWxAF\ngQpnUzajdYXHjAmns+ijbxAEgfPPOK3Rvc8Y1o8hvVN5bNEKbr7kPJLiovjipy0cy87nb9f/EYAz\nT+vHVxt2MLh3KsnxMXzx429knij0e+ZbL53Cn597m4cXvs8fxg1DVVWWfvkTuw5nMvfCSc17404B\nGvsu2qKiEbpgMrOOTrDp0sIiWAwfPrzJbQqKy+qIijnTz+KCs0YE/DyapqEoKqndYk75MsP5l0wm\nJjKMFV+vZ8mXPxJmszB6cG9uvXQKIVZvnsmzd17D88tW8thrKzBIEmcM68ebj9zMPf9ayrZ9R7ns\nvHEADThs1l0aFxVOn7RuxEaGEdtEHowoivzrnmt56b+rWfjR1zhdbvqkJfLCfXMZ2DMFgD9fNQ1Z\nVnlh+SokSWTK2KHcdtkU/v7GJ77jjBncmxfuvY7XP/6WBS8ux2iQGNAjmZfun9voVMypSmPfRWtk\nVDuOREfn1EEXFkGgpLyCx1//0E9UzL1wUo3SxqZRVRVBFEhPjvPrOnoqu2XOOncMs84d0+D6Pmnd\neGXBjXWWL3vydt//Z0w4nRkT6k5rfPr8PXWW5RaWsD8jh5v+eE5A4wu1Wrj/uou4/7qL6l0fHR7K\n3/90RZ3lF04c6fd41MBejBrYK6Dn1NHR0WlvdGHRBmzZssXvscPhYPz48QA43W7+8fZKsvKKfOuv\nnTGxWaJCURXMRiNJcVF1pkw6m1tmZ2TfsWzWbt7Ftxt30r1bbJ0SUp2OQ2PfRR0dnbZBFxZtgK2R\nXgMvLF/lM2wCuGLK+HrvkBtClmXCQqw+l8j66CxumZ0Vt0dm+ep1xEeF89RtdSMMOh2Hxr6LOjo6\nbcMpLSzKy8s5cOAAaWlpREdHt8tz7jhwDEHyvu3TzhjOxeeMDnhfjyITHRFCdHjLfS10Tp7BvVL5\ndmHblyHr6OjodEZOqZTo2tMGO3fu5IorruCHH+qaDTW0T2sxbkgfrp0xMeDjuz0ycZHhuqjQ0dHR\n0enQCFp7WUuewlRUVPiy08NHzqRvj1QenTerQfvn2rhlD4mxUYRa9V4GOjonQ83v4pYtW/SpEh2d\nNuCUngoJBtHhodw754KARYVHlkmOjcKmiwodHR0dnU7AKTUV0hG4+5rpRIUFZiHslj0kxUbqokJH\nR0dHp9OgC4t2pntiXEDbeWSZRD1SoaOjo6PTydCFRQfEI8vERYXrORU6Ojo6Op0OXVh0MGRZIS4y\njIhQPalMR0dHR6fzoQuLDoSiKoSHWIgIMAdDR0dHR0eno6ELiw6Coqr/v707j4uy2v8A/hm2AAEJ\nszRc0RoSlV0lMASVMtcs3HDLCpeIuoKBpmJ5TQslhWs/CdNcEJFcupq5dc2Nm1fcKiQ1MxVcCkFZ\nZT2/P7wz14dhEIZZGPy8Xy9eL+fM85znO49zZr5zznnOAwtzM7SuY0VNIiKipo6JRRMhA/D0E7zb\nIhERGTcmFk1ARUUF2j3lABMT/ncQEZFx4zeZgVVWVqK1Q0uYm3GtMiIiMn5MLAyoqroa1laWvAKE\niIiaDSYWBtamFSdrEhFR88HEwkAqKivh2Ppxnd09lYiIyBB4d1M9EEKgtLQUAGBlZcVkgshA2BaJ\ndI+JBREREWkNh0KIiIhIa5hYEBERkdYwsSAiIiKtYWJBREREWsPEgoiIiLSGiQURERFpDRMLIiIi\n0homFkRERKQ1TCyIiIhIa5hYEBERkdYwsSAiIiKtYWJBREREWsPEgoiIiLSGiQURERFpjZmhA3gU\nCCFQWloKALCysoJMJjNwRESPJrZFIt1rVj0WYWFhCAwMlJRduXIF06ZNg7e3N/r06YMFCxagqKhI\nsk1xcTE+/PBD+Pn5wd3dHaGhobh8+bLW4iotLYW7uzvc3d2VH2pEpH9si0S612wSi2+++QYHDhyQ\n/AIpKCjApEmTkJeXh08++QQRERHYvXs33nvvPcm+ERER2LdvHyIjI/Hpp5/i1q1bmDhxIgoKCvT9\nMoiIiIxasxgKuXXrFhYtWoQ2bdpIylNSUnD37l3s2LED9vb2AICnnnoKoaGhOHXqFDw8PHD69Gn8\n8MMPSEpKQt++fQEAnp6e6N+/PzZt2oRp06bp/fUQEREZq2bRYzF37lz07dsXPj4+EEIoy48ePQov\nLy9lUgEAvr6+aNGiBQ4fPqzcxtraGn5+fsptHBwc0KtXLxw6dEh/L4KIiKgZMPoei7S0NGRlZWHX\nrl1YsmSJ5LlLly5h8ODBkjJTU1O0a9dOOYfi0qVLaN++vcokrvbt22PXrl26DV7Lvjx2RaXsDd+O\n9d62Nur2JyL1ij9folLWYkZ0vbetjbr9iZoao04scnJysGTJEixZskTSK6FQVFQEGxsblXJra2vl\nBM7CwsJat2nRooXKJM/6KikpkTzmJDEiw2BbJNI/o00shBCYM2cO+vXrh4EDB9a6TXV1tdr9TUxM\nlPU8bJuGcnd312g/ItIutkUi/TPaxCI5ORkXLlzAsmXLUFlZCeB/SUJVVRVkMhlsbW1RXFyssm9R\nURHatm0LALCxscHt27dVtikuLoadnZ0OX4Fxa+ywC4dYiLSjscMuHGIhbTPaxGLv3r3Iz8+XTLpU\ncHFxQVhYGDp37owrV6RfalVVVcjOzsaLL74IAOjcuTOOHTumUseVK1fg5OSkUWynT5+WPC4tLcXz\nzz+vUV2kfj4IkxN6GLZF7WJiQvVhtInFRx99JBk/FULgH//4BzIzM7Fq1Sq0bt0aMpkMq1evRl5e\nHhwcHADcvwqktLRUmZD07dsXiYmJOHz4MF544QUAQF5eHjIyMjB9+nSNYrO2tm7kqyMibWBbJNI/\nvSUWd+7cwW+//QYvLy8AQHp6OtatWwdTU1OMHz++wb8iOnfurFJmb28Pc3NzuLi4AADGjh2LjRs3\nYsqUKQgLC0N+fj5iY2Ph7+8PNzc3AICXlxd69eqFWbNmITIyEvb29khISEDLli0xduzYRr5qUqe+\nV6UQGaO957NVykY2YNsX5e20HJF66noh6nu1ClFNekksLl68iPHjx+PJJ5/Ezp078ccffyA0NBQA\nYG5ujkOHDiEpKanRXZQymUxy2aiDgwPWr1+Pjz/+GJGRkWjRogUGDRqEqKgoyX7/+Mc/sHjxYsTG\nxqK6uhqenp6Ij4+Hra1to+IhIiJ61OglsYiPj4epqSkiIyMBAF9//TUqKyuRnJyMbt26YeLEiUhM\nTGx0YrF48WKVsmeeeQZr166tcz87O7ta922qGvJrvzn3DHBSKBkaJ07e15xfGzWcXlbePHHiBCZN\nmgR/f38AwA8//IAOHTrA09MTVlZWGD58OH7++Wd9hEJEREQ6pJcei9LSUuXkyZycHPz2228YN26c\n8vmaQxhERM1NQ+ZdEBkzvSQWTz/9NE6dOoXg4GBs374dANCvXz8A9xex2rdvH9q1099kJWqY5jyc\nQmQsOJmSjIVeEosRI0bgs88+wy+//ILff/8djo6OeP7553Hx4kVERUXh3LlzmDdvnj5CIR0zdBLC\neRdETWPOQ1OIgQxDL4nF1KlTIZPJ8M0336BPnz6YPXs2zMzMUF1djezsbISHhyMkJEQfoRAREZEO\n6W0di9DQUOUlpgrPPvssjh07BnNzc32FQUSkNZw3QaTKoCtvymQyJhVERDUYetEsosbQSWJR2/07\n6uPo0aNajoSIiIj0SSeJRW3LbRMREVHzp5PEYsOGDbqoloiIiJo4vay8WR83btwwdAhERETUSHqZ\nvCmEwLp163DkyBGUlJRACKF8rrKyEkVFRbh69SrOnTunj3DoEcT1LYgMT90iX1zfonnRS2KRlJSE\nuLg4WFhYoEWLFsjPz0ebNm1w584d3Lt3D46OjiqXohIREZHx0ctQyPbt2+Hs7Iz09HSkpqYCuD8P\nIyMjA/PmzcOdO3cwciSv/iYiIjJ2ekkscnJyMHz4cNjY2KBDhw6ws7NDRkYGzMzMEBISAj8/P6xY\nsUIfoRARNSt7z2er/BEZkl6GQkxNTWFjY6N83L59e1y4cEH52MfHB8uXL9dHKGQAmccPqZS59PY3\nQCREjwZ1K4LWN+lQty0X6aL60Eti0alTJ2RmZiI4OFj5OCsrS/n8vXv3UFZWpo9QiJQ4oZOoaeAN\ny5oXvSQWQ4YMwdKlS2Fubo6ZM2fihRdeQHR0NDZt2oQuXbpg/fr1cHJy0kcozUJtPQAAewGImjMO\ncZCx0EtiMWnSJFy8eBHJycmIiIjA4MGDkZycjI8++uh+EGZmiImJ0UcozRqHHIiIyND0kliYmZlh\nyZIliIyMhKWlJQBg48aN2L17N+7cuQM/Pz907dpVH6GQEdJ3D01Dhkg4nEKkGw0ZHuFQStOi17ub\nPvHEE8p/W1hYYMSIEfo8vFGq7YuLiPRv27thhg7BoyZ6pQAAH25JREFU4HibeKoPnSQWERERCAkJ\ngYeHBwBg5syZkMlkD91v2bJlugiHiIiI9EQnicW3336Lfv36KROL3bt312s/JhaGVeuQg+9E/R0L\nzWNOCIdHiJoGDpEYhk4Si19//bXOx6Q/xvYFri5eIiIyDnqZYzF16lS8/PLLGD58uD4OR1pW7y97\n34kGTwx01evS2F4IdXNl2JNB1DDqbmTWmP3Zi6Fdekks/v3vf6Nfv376OBQ9QhqSxPBSXCIi/dBL\nYtG1a1dkZmbq41BERKRnXP6bHqSXxGLcuHFYtGgRLl++jD59+qBVq1YwNTVV2W706NH6CIeIiIh0\nRC+Jxdy5cwEAJ0+exMmTJ2vdRiaTMbHQI0PPhWgKmsJVMLo6HhGRoegksZgwYQJmzJgBHx8fAMC6\ndet0cRgiImqiODzy6NJJYnHixAnk5uYqH/fu3VsXhyEyuMZOCuWaF0RNA68W0R69LulN6v32ZxEy\nrt6RlLW2fcxA0ZCxYGKifSc3bcRfFy5Iyp4dMNBA0TQv6u7Q2hx6MpiY/A8TiyaiokqguKxKUmbz\nWBXnQjyCmCwYVllREUry8yRlFaWlBoqGDInJgmZ0lljs378fV6407AZaYWG8yQ+ROrwhHZHhNXaB\nrkeBzhKLffv2Yd++fQ3ah4kFNQVcTIuISHM6SyymTp2K559/XlfVE+lVQy5N5fAVET3KdJZYdO3a\nlVeDEBkI52nQo6S2SaEj1ZTrc6KoumGT5j5Pg5M3iQyIwy5E1NwwsSAyEg1JQmrb9kuobsteDDI2\n6i5ZNSbN/WoTnSQWI0aMQPv27XVRtU4dPXoUn332GS5duoRWrVohJCQEU6ZM0cuxBaphIhOSMhmE\nmq2pOdPnUuNETUF9k4WROo6DtEMnicWSJcZ3Oc6ZM2cwbdo0DB48GH/729+QkZGB2NhYVFZWIjQ0\nVOfHz//zJn7+8bikrLPjkzo/LhERkTZxKOS/EhIS4OLigk8++QQA4Ofnh8rKSiQmJmLSpEl47DGu\ngklND3s3iLSjvr0mzWGVUF1jYgGgvLwc//nPfxAeHi4pDwoKwurVq3Hy5EleOktGT10SwitIiAyv\nOc27YGIB4Nq1a6ioqEDnzp0l5R073v9wvXz5MhMLIqJmQN2lqaQ9TCwAFBYWAgBatGghKVc8Li4u\nblB9JSUlah+XqrnnQNm9e6isKJeUVZSXqZTpW0lJSb1jaMi2+qSr16DP16vL11BepvqerPke1gYr\nKyvIZDKt11sXjdpiRQXKK6X37blXVqZSVtcxdbFtY+kqrkftNZRWVNRrO22QGWk7lAkhHvlLD06d\nOoVx48Zh7dq18PHxUZZXVlaie/fuiIiIwFtvvVXv+uRyuS7CJDJqp0+fhrW1tV6PybZIJKWPdmii\n09qNhK2tLQDVnomioiLJ80SkOXU9BESkP/pohxwKAdChQweYmpri6tWrknLF4y5dujSovtOnT0se\nFxcXw8/PDwBw7Ngxvf9qa6zS0lLlHJP09HRYWVkZOKL6M+bYgeYVv6Wlpd6P35zaYnN6LzB+/dJ3\nO2RiAeCxxx6Dl5cX9u3bJ1kQa+/evbCzs0PPnj0bVF9dH1bW1tZG9WFWk5WVldHGb8yxA8Yfv77n\nVwDNty0a+3uB8RuOPtohh0L+a/r06Th79izeffddHDp0CMuXL8eaNWswdepUrmFBRERUT0ws/qtP\nnz5ISEjA5cuXERYWhm+//RZRUVF44403DB0aERGR0eBQyAMGDBiAAQMGGDoMIiIio8XLTYmIiEhr\nOBRCREREWsPEgoiIiLSGiQURERFpDRMLIiIi0homFkRERKQ1TCyIiIhIa5hYEBERkdYwsSAiIiKt\nYWJBREREWsPEgoiIiLSGiQURERFpDRMLLTp69CheffVVuLm5oX///lizZk299z137hxcXFxw/fp1\nHUaoXkNjLysrQ1xcHAICAuDm5oYxY8bg6NGjeopWqqGxX7lyBc7Ozip/Q4cO1VPEUg2Jf9u2bbXG\nrvjbsWOHHiO/r6Hnv7y8HMuWLYO/vz9cXV0xcuRI7N6926AxPYhtUXNsi4Zri02qHQrSitOnTwsX\nFxfx/vvviyNHjojPPvtMODs7i8TExIfue/78eeHr6yucnZ1FTk6OHqKV0iT2iIgI4enpKTZt2iTS\n09PFrFmzRLdu3cSJEyf0GLlmsX/33XdCLpeLH3/8UZw9e1b5d/78eT1Gfl9D4799+7Yk5rNnz4oz\nZ86IIUOGiICAAJGXl9ek4xdCiBkzZggXFxexfPlykZ6eLhITE4Wrq6vYsGGDwWJSYFvUHNui4dpi\nU2uHTCy0ZMqUKWLUqFGSstjYWOHh4SHu3btX6z7l5eXiyy+/FG5ubqJXr15CLpcb5MOsobFfu3ZN\nyOVykZycrCyrrq4W/fv3FzNnztR5vA/S5LzHxcWJfv366SO8h9Ik/prWrVsnnnvuOXH27FldhFin\nhsafmZkp5HK5ygfe2rVrhZubmygoKNB7TEKwLWoD26Lh2mJTa4ccCtGC8vJy/Oc//8GAAQMk5UFB\nQSguLsbJkydr3e/QoUNYuXIlpk2bhsjISH2EqkKT2J988kls3boVw4YNU5bJZDKYmpqivLxc5zEr\naHref/31Vzg7O+sjxDppGv+DcnNzsXz5cowbNw49e/bUVai10iT+S5cuAQACAgIk5V5eXigtLcXx\n48f1HhPAtthYbIuGa4tNsR0ysdCCa9euoaKiAp07d5aUd+zYEQBw+fLlWvfr0aMHDh48iKlTp8LU\n1FTncdZGk9gtLCzg4uICGxsbCCFw48YNLFq0CNeuXcOYMWP0Ejeg+XnPyspCUVERxowZg549e8LP\nzw/Lli1DZWWlzmN+kKbxPyg+Ph5mZmZ47733dBJjXTSJ//HHHwcAZGdnS8qvXr1aa7k+YgLYFhuL\nbdFwbbEptkOzRu1NAIDCwkIAQIsWLSTlisfFxcW17vfUU0/pNrB60DR2hS+++AKfffYZAGDUqFHw\n8fHRQZS10yT2vLw8/Pnnn6iursasWbPg6OiI9PR0JCUl4caNG1i6dKnuA/+vxp7727dvY8eOHXjj\njTdgY2OjmyDroEn8vXr1Qrt27bBw4UJYWVmhe/fuyMzMxIoVKwAApaWleo8JYFtsLLZFw7XFptgO\nmVhoQXV1dZ3Py2QyPUXScI2NPTAwEJ6enjh58iRWrlyJe/fu4dNPP9VmiGppEruNjQ2++uordOzY\nEW3btgVwv/vPwsICy5cvx/Tp09GlSxedxFtTY899WloaqqurMXHiRG2GVW+axG9hYYHVq1dj9uzZ\nmDx5MgCgTZs2iIqKwt/+9jdYWVnpPaamgm2RbVETTbEdMrHQAltbWwCqmWFRUZHk+aaosbE/88wz\nAO5/IFRWViIhIQEzZ85EmzZtdBCtlCaxW1hYoE+fPirl/v7+WL58Oc6fP6+3D7PGnvu9e/fCz89P\n2a2pb5rG36lTJ6SkpCAvLw93795Fp06dlGO+LVu2NEhMTQHb4n1siw3TFNsh51hoQYcOHWBqaqoc\nn1JQPNZX49CEJrFfv34daWlpKpPDunXrBgD4888/dRStlCax//HHH9i8ebOy+1Dh3r17AAAHBwcd\nRauqMe+bW7duISsrC4MGDdJpjHXRJP6ysjLs3LkT2dnZcHBwQOfOnSGTyZCZmQngf+8hfcbUVLAt\n3se22DBNsR0ysdCCxx57DF5eXti3b5+kfO/evbCzs9P7bP2G0CT27OxszJs3D/v375eUHzt2DBYW\nFiqTiHRFk9j/+usvLFiwAHv27JGU7969G7a2tujevbtOY35QY943Z8+eBQB4eHjoNMa6aBK/mZkZ\nPvzwQ6SlpSnLKioqkJycjI4dO0Iul+s9pqaCbfE+tsWGaYrt0HTBggULGlUDAQDatm2LxMREXLx4\nEdbW1tixYwdWr16N8PBweHt7o6ioCOfOnYOFhUWt41dZWVn4/vvvMWnSJL131zY09qeffhonT57E\ntm3bYGtri4KCAqxfvx6bNm1CWFgYfH19m3Tsp06dwtatW2FtbY3i4mJs2LABGzduREREBLy9vfUW\nuybxK3z33Xc4c+YMoqOj9RpvTQ2N38TEBCUlJVi/fj3s7OxQWFiIxYsX49SpU1iyZAk6deqk95hq\nYlvUX+xsi4aJXeftsFGrYJDE/v37xdChQ0X37t3FgAEDxNq1a5XP/fjjj0Iul4vt27fXuu/WrVsN\nttqfEA2PvbCwUCxevFgEBASI7t27i6FDh4qtW7caIHLNYw8MDBQ9evQQQ4YMEWlpaQaI/D5N3jcL\nFiwQvr6+eo60dg2Nv6KiQsTHx4t+/foJNzc3MW7cOHHs2DGDxvQgtkXNsS0aTlNqhzIhhNBKykRE\nRESPPM6xICIiIq1hYkFERERaw8SCiIiItIaJBREREWkNEwsiIiLSGiYWREREpDVMLIiIiEhrmFgQ\nERGR1jCxIKNTUlKCDRs2YNSoUejduzdcXV0xdOhQfPHFFyo3Y9KVa9eu6bT+4uJi3L59W6fH0Iax\nY8ciMDBQ+TghIQHOzs64fPlyg+uqeU6dnZ0RERHR6BgfFcePH0ffvn2VN/FSyM/Px6pVqzBixAh4\neXnB3d0dwcHBSElJUbnldnR0NJydnSV/zz33HNzd3fHKK69g/fr1eHBNxW3btqlsX9vf9evXAQCH\nDh3CwIEDUVZWpvsTQgbD26aTUbl69SpmzJiBP/74A4MHD8awYcMghMCxY8cQFxeHgwcPYs2aNbXe\nA0Jb5s+fjwsXLmDz5s06qf+XX37BjBkz8PHHH8PPz08nx9AmmUym/HdQUBA6deqE1q1bN6iO2s5p\nbGwsHB0dtRZnc1ZeXo6YmBi8/fbbsLS0VJb//PPPCAsLQ0FBAYYPH47Ro0ejrKwMBw4cwIcffoh/\n//vfWL58OUxMpL8x58yZo7wFuBACJSUl2L9/Pz7++GPk5ORg9uzZku1Hjx4NLy8vtfEp6vL398fq\n1auxcuVKzJw5U1svn5oarSwMTqQHZWVlYsiQIaJXr14iMzNT5fmvvvpKyOVy8cEHH+g0joCAADF6\n9Gid1b9161Yhl8vFkSNHdHYMbRkzZowIDAxsdD26PqfNXVJSkvD39xeVlZXKsry8POHr6ysCAgJE\ndna2yj6LFi0ScrlcfP7558qyqKgoIZfLa71PSlVVlQgODhbdunUTf/75pxDif+9Vdfddqc2PP/4o\nXFxcxNWrVxvyEsmIcCiEjMbmzZtx8eJFREdHo1u3birPT5o0Cc7Ozvj2229RWFhogAi1S/A2PlQP\nlZWV2LBhAwYNGgRTU1Nl+apVq5Cbm4slS5bU2vMTGRkJBwcHpKWl1eu9ZmJigpdeeglVVVXKW4Vr\nonfv3nB0dMTGjRs1roOaNiYWZDR27dqFFi1aYNiwYWq3+fzzz5Geni653fWlS5cQHh6O3r17o2fP\nnhgxYgS+/vpryX6KseJffvkFs2fPRu/eveHm5oYpU6bg119/VW6nGC8+c+YMnJ2dsWPHDuVz//zn\nPzFy5Ei4urqid+/eeO+99yTzBlJTU+Hs7Ix58+ZJjh0TEwNnZ2fs378fCQkJmDNnDgDgrbfeksxf\nqEkR8+nTpxEeHg53d3c8//zz+OCDD5Cfn6/cLjs7G87OzlizZg0mTpyIHj16KIeQAODw4cMYN24c\n3N3d4enpidDQUJw7d07leAcOHMDIkSPh5uaGF198Edu3b1fZprY5FiUlJYiNjUX//v3h6uqKl156\nCV988QWqqqrqPKfOzs4q3eWHDh1CSEgI3N3d4eHhgddffx0ZGRmSbaKjoxEYGIhff/0VkydPhru7\nO3r37o3Zs2fjzp07as+nQnl5ORISEjBw4ED06NEDAQEB+OSTT1BcXPzQc1pdXY3AwEBER0fjww8/\nhKurK/z8/PD7778DaNh7cc+ePXjxxRfRs2dPzJ8/X228+/fvx61btxAUFKQsq66uxrfffgsnJyf0\n6tWr1v0sLCyQmpqKPXv2SIaz6qIYMqmsrKzX9uoEBQVh69atKC0tbVQ91DRxjgUZBSEEMjMz4enp\nKflVVtPTTz8teZyZmYnx48fD0tISEyZMgL29Pfbu3Yu5c+fi999/x/vvvy/ZPjw8HB07dsR7772H\nW7duYc2aNQgNDcXBgwdhamqKTz/9FIsXL4adnR3CwsLg7u4O4H5CEx8fj8DAQAQHB+P27dtISUnB\na6+9hrS0NHTo0AGjR4/Gvn378PXXX2PkyJFwd3fHkSNHkJqaipEjR2LgwIHo0KED/vrrL2zZsgVv\nvvmmsv66REREwMHBATNnzsSVK1eQkpKCs2fPYtu2bbCwsFBul5CQAD8/P8ybNw/l5eWQyWTYsWMH\noqOj4e3tjYiICBQXF2Pr1q0YO3YsvvrqK+Xxd+3ahcjISLi4uCAiIgK3bt3CggULYGJiAgcHB7Wx\nVVRUYPz48cjKysKrr76Knj174vTp04iLi8ONGzcQExOj9pwC0vkbqampiImJgVwuR3h4OMrLy5GW\nlobJkydjxYoV6N+/v3KfgoICTJo0CQMGDMDgwYORkZGB7du3o7S0FMuXL1cbb3V1NaZPn47jx49j\n1KhRkMvluHDhAjZs2ICMjAwkJyfXeU4VX7x79+5F+/bt8cEHH+DatWtwcnJq8Hvxgw8+wJgxY9Cm\nTRt06NBBbcw//PAD7O3tJeft1q1byM3NRd++fdXuBwDt27ev8/ma0tPTIZPJ4OLiIikvLi5GXl5e\nrftYWlrC2tpaUubt7Y2kpCRkZGQ8NEYyQoYdiSGqn9u3bwu5XC5mzpzZoP1Gjx4tXF1dJeO51dXV\nYurUqUIul4usrCwhxP/GikNDQyX7JyQkCLlcLo4dO6Ysqzkf4OrVq+K5554TixYtkux78+ZN4eHh\nId555x1l2Y0bN4Snp6cYMmSIyMvLEy+88ILo37+/KC4uVm5T3zkWiu1GjBghysrKlOUpKSlCLpeL\ndevWCSGEuHbtmpDL5SIwMFBUV1crtyssLBQeHh5ixowZknoLCwtFQECAeOWVV4QQ98fWfX19xdCh\nQyXHSU9PV9arEB8fL+Ryufj999+FEEJs2rRJyOVysWXLFskxoqKihIuLi8jNza31nAohJP/f+fn5\nws3NTQwbNkwSw927d8ULL7wgfH19RXl5ubJuuVwukpKSJPWNHz9euLi4iHv37qk9p9u3bxdyuVzs\n379fUv79998LuVwuNmzYUOc5VbwWZ2dnlTkEDX0vzpo1S22cNY83YcIESdnZs2eFXC4XcXFx9apD\nQXHuzp07J27fvi1u374t/vrrL/HTTz+JmJgYIZfLRXh4uHJ7Rax1/UVHR6sc5+bNmxrFR8aBPRZk\nFBS/BBXd5/WRm5uLM2fOYOTIkZJfZjKZDNOnT8cPP/yAAwcOwNnZWfncoEGDJHUonsvNzVV7nAMH\nDqC6uhr9+/eX/GozNzeHt7c3Dh8+jOrqapiYmKBNmzaIjo7G3LlzERwcjNzcXKxfv17lF11DTJky\nRfIrOjg4GHFxcfjXv/6FiRMnKsu9vb0lPQDp6ekoLi7GwIEDVX5t+vv7IyUlBX/99Rdu3ryJ3Nxc\nvPnmm5Lj+Pj44Nlnn5UMEdR08OBB2Nra4tVXX5WUz5o1C2+99Rbs7e3r9RrT09NRWlqK119/XRKD\nnZ0dQkJCEBcXhzNnzsDb21v53Msvvyyp47nnnsOJEydw584dPPXUU7UeZ8+ePbCxsYGnp6fknLi5\nucHOzg4HDx7E+PHjleU1z6lC27ZtJe85Td6Lffr0eeh5qaysxPXr1+Hj4yMpV/TqaTpk8corr6iU\nmZmZ4ZVXXlEZygOAN998E76+vrXW9eSTT6qUPfXUU7C0tNT5ZdtkGEwsyCjY29vDwsKiQWs75OTk\nAACcnJxUnlOUKbZRqNmtr/gSq3m9/4OuXLkC4P7k0drIZDLk5eXhiSeeAAC89tpr2LlzJ44fP46x\nY8fC09OzPi9HrWeeeUby2NTUFI6Ojsq1AxRatWpVa9zR0dFq475+/bqyno4dO6ps4+TkhJ9++klt\nbDk5OWjfvr3K5YytWrVSiacu2dnZyuPVFoPiWA8mFjXrV/xf1pWcXr16FUVFRSpf1AoPO6fqyjV5\nL9bn/Ny9excAJHOKACgv99V0LZSlS5cqj29iYgJbW1s4OTlJLmV9UNeuXdWeM3VatGihdviEjBsT\nCzIaHh4eOHv2LCoqKmBubl7rNl9//TUOHDiA8PBw5eREUcuMd0WiULOeml+A9aGoKyEhATY2NrVu\nY2dnp/x3fn4+fvvtNwD3J06WlJQ0qsfCzEy1GVdVVanMRan5WBF3TEwMOnXqVGvdnTt3xo0bNwBA\nZeElRR11TfxrSA+TphT/vzX/L+s7IfFBVVVVcHR0xN///vdan6/5xapuvk/Nck3ei3XNJVJQvMaa\n9T755JNo3749Tp8+Xef+n3/+OS5cuIDIyEi0a9dOWe7h4aEyX0nbhBD1eo1kfHhVCBmNoKAglJaW\nYufOnbU+L4TAli1bcPjwYdja2io/KBUz8h+kKGvbtm2j41Jcyte6dWv4+PhI/kxNTWFiYiLpvl+4\ncCHu3LmD6Oho5OTkIDY2tlHHV/Q8KFRUVCAnJ6fWHoYHKc6PnZ2dStw2NjYQQsDS0lLZdV/bapoP\n68p2dHREdna2yhdfVlYWIiIiJFfc1CfWS5cuqTyn+L9s06ZNvep62HHy8/Ph7e2tck4KCws1XnhN\nV+/Fxx9/HGZmZrVe7RIUFIQrV67g+PHjte5bXl6ubC91TcDVlbt37yp78ah5YWJBRiM4OBgdO3ZE\nbGxsrZdDrly5Ej/99BMGDx6M9u3b44knnoCrqyt2794t+QKsrq5GYmIiZDIZAgICGhyHiYmJZGhE\ncTXCqlWrJF+gN27cwLRp07Bs2TJl2f79+7F7925MmjQJkydPxvDhw7F582bJh7+i16Su4ZcHbdy4\nUXLc1NRUFBcX46WXXqpzP19fX1haWuLLL7+ULIVeWFiId999F7Nnz4aZmRm6desGR0dHbN68GUVF\nRcrtMjIykJWVVecxAgICUFBQoJIMpqSkYPfu3covtJrntLZYrayssHbtWpVYk5OT4eDgADc3tzpj\nqY/+/fujpKQEX331laRc0Qu2a9cujerV1XtRJpOhbdu2KkM0wP3LlVu2bIm5c+cqh5IUqqqq8NFH\nH+HmzZuYOHFio3rMNHHr1i1UVVXpvFeEDINDIWQ0zM3NsXLlSkyZMgWjR4/Gyy+/DFdXVxQXF+PQ\noUPIyMhAjx49EBMTo9xn7ty5mDhxIoKDgxESEgJ7e3vs27cPJ06cwMSJEyWT5eqrVatWuHDhApKT\nk9GnTx907doVr7/+OtauXYuQkBC89NJLuHfvHlJSUlBVVYWoqCgA94dAFixYgHbt2uHdd98FcH9+\nw+HDhzFnzhzs3LkT1tbWyrHt1NRUFBQUYMiQIXXGc+rUKUyePBlBQUG4ePEitmzZAm9vb4wYMaLO\n/ezt7REREYFFixbhtddew4gRI2Bqaoq0tDTcvHkTy5YtUyY58+fPx9tvv43g4GCMGjUKBQUFWL9+\nPR5//PE6F1caNWoUtm/fjtmzZ+PMmTN49tlncfLkSezcuROhoaHKiX01z2mXLl0k9bRs2RKRkZFY\nuHChMtbKykps2bIFeXl5iIuL00q3enBwMP75z39i6dKlOH/+PDw9PXH16lVs2rQJ7dq1w5QpUzSu\nWxfvReD+JNqdO3cqJwgr2NvbY+XKlZg2bRqGDRuG4cOHQy6XIz8/H3v27MH58+fRr18/hIWFafya\nFE6dOlXn0FOPHj0k80vOnDkDAGonfJJxY2JBRqVr167YsWMHNm7ciAMHDuD7779HRUUFnJyc8P77\n72PChAmSseoePXogNTUVK1aswIYNG1BeXo5nnnkGS5YsUfniVffBWLP8nXfewfz58/HJJ5/g7bff\nRpcuXRAVFQUnJyekpKRg2bJlsLa2Ro8ePRAWFoaePXsCuD8EkpeXh6SkJDz22GMA7ndlR0dHIyoq\nCkuXLsX8+fPh4+ODQYMG4eDBg8jIyEBQUJBkKKWmhQsX4rvvvkNsbCzs7e3xxhtv4J133qnX+Zww\nYQLatm2LL7/8EgkJCTA3N4dcLkdiYqJkfQHFPR7i4+OxYsUK5boZJ0+elKzCKJPJJOfLwsIC69at\nQ3x8PPbu3YutW7eiQ4cOiImJwZgxY+o8pzWFhISgTZs2SEpKQnx8PCwsLODm5oZPP/0UHh4eamN4\nWPmDzM3NsXbtWvzf//0fvvvuO+zZswetW7fGkCFD8M477zRqyEAb78Xa+Pv7Y8uWLfj555/h6uoq\nec7LywvffPMN1q9fjyNHjigXH5PL5Vi4cCGCg4NVjtuQYyu2TUtLw5YtW9RuM3v2bElikZGRgZYt\nW0r+36j5kIm6fm4QUZO1bds2zJkzB6tXrzaKm5WRblRUVGDAgAEYOHAg5s6da+hwHqqqqgr+/v4Y\nOnSosjePmhfOsSAiMmLm5uaYPHkydu3aJZl/0lQdOXIEd+7cweTJkw0dCukIEwsiIiM3duxY2NjY\nIDU11dChPNSqVaswYcIEtYuUkfFjYkFkxDRZq4GaH0tLSyxcuBCJiYm1rjfSVPzrX/9Cbm6ucvIy\nNU+cY0FERERawx4LIiIi0homFkRERKQ1TCyIiIhIa5hYEBERkdYwsSAiIiKtYWJBREREWsPEgoiI\niLSGiQURERFpDRMLIiIi0homFkRERKQ1TCyIiIhIa5hYEBERkdYwsSAiIiKt+X9a65+fRo4G9AAA\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11578f210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f = plt.figure(figsize=(3.5, 4.5))\n",
    "\n",
    "reg_axes = []\n",
    "hist_axes = []\n",
    "gs = plt.GridSpec(8, 2)\n",
    "\n",
    "# Make it easy to switch to black and white if necessary\n",
    "black_and_white = False\n",
    "if black_and_white:\n",
    "    pe_color = \".25\"\n",
    "    congr_pal = [\".1\", \".4\"]\n",
    "    early_pal = [\".1\", \".4\"]\n",
    "    switch_pal = [\".1\", \".4\"]\n",
    "else:\n",
    "    pe_color = \"#429c5c\"\n",
    "    congr_pal = [\"#5f486f\", \"#9a91c4\"]\n",
    "    early_pal = [\"#355d7a\", \"#72acd3\"]\n",
    "    switch_pal = [\"#8b312d\", \"#f06855\"]\n",
    "\n",
    "# Initialize the subplot axes\n",
    "for row in range(2):\n",
    "    for col in range(2):\n",
    "\n",
    "        rr = slice(row * 4, row * 4 + 3)\n",
    "        hr = row * 4 + 3\n",
    "        reg_axes.append(f.add_subplot(gs[rr, col]))\n",
    "        hist_axes.append(f.add_subplot(gs[hr, col]))\n",
    "\n",
    "        if col:\n",
    "            reg_axes[-1].set(yticklabels=[])\n",
    "            hist_axes[-1].set(yticklabels=[])\n",
    "\n",
    "        if row:\n",
    "            reg_axes[-1].set(xlim=(.1, .9), ylim=(1.06, 1.53))\n",
    "        else:\n",
    "            reg_axes[-1].set(xticklabels=[])\n",
    "            hist_axes[-1].set(xticklabels=[])\n",
    "            reg_axes[-1].set(xlim=(.1, .9), ylim=(1.11, 1.39))\n",
    "        hist_axes[-1].set(xlim=(.1, .9), ylim=(0, 450))\n",
    "\n",
    "# Set some keyword arguments that will be shared across subplots\n",
    "bins = np.percentile(data_corr.pe_context, [10, 30, 50, 70, 90])\n",
    "bins = 6\n",
    "reg_kws = dict(x_bins=bins, units=\"subj\", n_boot=10000, logx=True,\n",
    "               scatter_kws={\"s\": 12}, line_kws={\"lw\": 1.5})\n",
    "\n",
    "hist_kws = dict(bins=30, rwidth=.95, alpha=.7, linewidth=0)\n",
    "text_kws = dict(ha=\"center\", size=8)\n",
    "\n",
    "# ===================================================================================\n",
    "# Plot showing the relationship between log(CPE) and RT\n",
    "\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr, ax=reg_axes[0], color=pe_color, **reg_kws)\n",
    "hist_axes[0].hist(data_corr.pe_context, color=pe_color, **hist_kws)\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Plot showing the (non) effect of incongruent stimuli\n",
    "\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"not congruent\"),\n",
    "            ax=reg_axes[1], color=congr_pal[1], **reg_kws)\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"congruent\"),\n",
    "            ax=reg_axes[1], color=congr_pal[0], **reg_kws)\n",
    "hist_axes[1].hist([data_corr.query(\"congruent\").pe_context,\n",
    "                   data_corr.query(\"not congruent\").pe_context],\n",
    "                  histtype=\"barstacked\", color=congr_pal, **hist_kws)\n",
    "reg_axes[1].text(.55, 1.18, \"congruent\\nstimulus\", color=congr_pal[0], **text_kws)\n",
    "reg_axes[1].text(.35, 1.33, \"incongruent\\nstimulus\", color=congr_pal[1], **text_kws)\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Plot showing the interaction between log(CPE) and early cue presentation\n",
    "\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"early\"),\n",
    "            ax=reg_axes[2], color=early_pal[0], **reg_kws)\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"not early\"),\n",
    "            ax=reg_axes[2], color=early_pal[1], **reg_kws)\n",
    "hist_axes[2].hist([data_corr.query(\"early\").pe_context,\n",
    "                   data_corr.query(\"not early\").pe_context],\n",
    "                  histtype=\"barstacked\", color=early_pal, **hist_kws)\n",
    "reg_axes[2].text(.6, 1.10, \"early cue\", color=early_pal[0], **text_kws)\n",
    "reg_axes[2].text(.3, 1.38, \"concurrent\\ncue\", color=early_pal[1], **text_kws)\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Plot showing the interaction between log(CPE) and context switch trials\n",
    "\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"context_switch\"),\n",
    "            ax=reg_axes[3], color=switch_pal[1], **reg_kws)\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr.query(\"not context_switch\"),\n",
    "            ax=reg_axes[3], color=switch_pal[0], **reg_kws)\n",
    "hist_axes[3].hist([data_corr.query(\"not context_switch\").pe_context,\n",
    "                   data_corr.query(\"context_switch\").pe_context],\n",
    "                  histtype=\"barstacked\", color=switch_pal, **hist_kws)\n",
    "reg_axes[3].text(.6, 1.18, \"context repeat\", color=switch_pal[0], **text_kws)\n",
    "reg_axes[3].text(.4, 1.37, \"context switch\", color=switch_pal[1], **text_kws)\n",
    "\n",
    "# ===================================================================================\n",
    "# Tweak the labels and positioning of the figure\n",
    "\n",
    "for i in range(4):\n",
    "    reg_axes[i].set(xlabel=\"\")\n",
    "    if i % 2:\n",
    "        reg_axes[i].set(ylabel=\"\")\n",
    "        hist_axes[i].set(ylabel=\"\")\n",
    "    else:\n",
    "        reg_axes[i].set_ylabel(\"Reaction time (s)\", labelpad=3)\n",
    "        hist_axes[i].set_ylabel(\"Trials\", labelpad=3)\n",
    "\n",
    "[ax.set(xticks=[.1, .3, .5, .7, .9], yticks=[0, 400]) for ax in hist_axes]\n",
    "sns.despine()\n",
    "[sns.despine(ax=ax, bottom=True) for ax in reg_axes]\n",
    "[ax.set(xticks=[]) for ax in reg_axes]\n",
    "\n",
    "# Errorbar width isn't configurable in seaborn but default is too wide\n",
    "for ax in reg_axes:\n",
    "    for line in ax.lines:\n",
    "        line.set_linewidth(1.5)\n",
    "\n",
    "# Label the panels\n",
    "f.text(.55, .02, \"Context prediction error (CPE)\", ha=\"center\", size=8.8)\n",
    "f.text(.02, .96, \"A\", size=12)\n",
    "f.text(.54, .96, \"B\", size=12)\n",
    "f.text(.02, .50, \"C\", size=12)\n",
    "f.text(.54, .50, \"D\", size=12)\n",
    "\n",
    "f.tight_layout(pad=.5, w_pad=.75, rect=[0.01, .03, 1, 1])\n",
    "savefig(f, \"behavioral_main\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Tests of other variables that might influence RT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ context_switch + (context_switch | subj)\n",
       "   Data: data.corr\n",
       "\n",
       "REML criterion at convergence: 2301\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.7801 -0.7348 -0.0760  0.6609  3.6474 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name               Variance  Std.Dev. Corr \n",
       " subj     (Intercept)        0.0349269 0.18689       \n",
       "          context_switchTRUE 0.0009234 0.03039  -0.59\n",
       " Residual                    0.0778957 0.27910       \n",
       "Number of obs: 7706, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                   Estimate Std. Error t value\n",
       "(Intercept)         1.20586    0.04843  24.897\n",
       "context_switchTRUE  0.07594    0.01016   7.471\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "cntxt_sTRUE -0.488\n",
       "\n",
       "Likelihood ratio test for main effect of context switch:\n",
       "  Chisq(1) = 24.09; p = 9.18e-07\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ context_switch + (context_switch | subj), data.corr)\n",
    "m.null = lmer(rt ~ 1 + (context_switch | subj), data.corr)\n",
    "lr.test(m, m.null, \"main effect of context switch\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ frame_switch + (frame_switch | subj)\n",
       "   Data: data.corr\n",
       " Subset: !context_switch\n",
       "\n",
       "REML criterion at convergence: 1229.6\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.8383 -0.7399 -0.0929  0.6545  3.6685 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name             Variance  Std.Dev. Corr \n",
       " subj     (Intercept)      3.548e-02 0.188357      \n",
       "          frame_switchTRUE 7.151e-06 0.002674 -1.00\n",
       " Residual                  7.549e-02 0.274761      \n",
       "Number of obs: 4507, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                 Estimate Std. Error t value\n",
       "(Intercept)      1.197253   0.049004  24.432\n",
       "frame_switchTRUE 0.016215   0.008237   1.968\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "frm_swtTRUE -0.173\n",
       "\n",
       "Likelihood ratio test for main effect of cue switch on repeat trials:\n",
       "  Chisq(1) = 3.85; p = 0.0497\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ frame_switch + (frame_switch | subj), data.corr, subset=!context_switch)\n",
    "m.null = lmer(rt ~ 1 + (frame_switch | subj), data.corr, subset=!context_switch)\n",
    "lr.test(m, m.null, \"main effect of cue switch on repeat trials\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ evidence_switch + (evidence_switch | subj)\n",
       "   Data: data.corr\n",
       " Subset: !context_switch\n",
       "\n",
       "REML criterion at convergence: 1224.9\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.8847 -0.7327 -0.0846  0.6539  3.6780 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name                Variance Std.Dev. Corr \n",
       " subj     (Intercept)         0.038959 0.19738       \n",
       "          evidence_switchTrue 0.000915 0.03025  -0.64\n",
       " Residual                     0.075305 0.27442       \n",
       "Number of obs: 4507, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                    Estimate Std. Error t value\n",
       "(Intercept)          1.21384    0.05133  23.647\n",
       "evidence_switchTrue -0.01402    0.01135  -1.235\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "evdnc_swtcT -0.504\n",
       "\n",
       "Likelihood ratio test for main effect of evidence switch on repeat trials:\n",
       "  Chisq(1) = 1.56; p = 0.212\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ evidence_switch + (evidence_switch | subj), data.corr, subset=!context_switch)\n",
    "m.null = lmer(rt ~ 1 + (evidence_switch | subj), data.corr, subset=!context_switch)\n",
    "lr.test(m, m.null, \"main effect of evidence switch on repeat trials\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Linear mixed model fit by REML ['lmerMod']\n",
       "Formula: rt ~ response_switch + (response_switch | subj)\n",
       "   Data: data.corr\n",
       " Subset: response_switch != \"nan\"\n",
       "\n",
       "REML criterion at convergence: 2382.6\n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-3.8901 -0.7400 -0.0821  0.6671  3.7838 \n",
       "\n",
       "Random effects:\n",
       " Groups   Name                Variance  Std.Dev. Corr \n",
       " subj     (Intercept)         0.0330810 0.1819        \n",
       "          response_switchTrue 0.0001848 0.0136   -0.28\n",
       " Residual                     0.0791966 0.2814        \n",
       "Number of obs: 7577, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                     Estimate Std. Error t value\n",
       "(Intercept)          1.247506   0.047188  26.437\n",
       "response_switchTrue -0.018109   0.007367  -2.458\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr)\n",
       "rspns_swtcT -0.195\n",
       "\n",
       "Likelihood ratio test for main effect of response switch:\n",
       "  Chisq(1) = 5.36; p = 0.0207\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = lmer(rt ~ response_switch + (response_switch | subj), data.corr, subset=response_switch != \"nan\")\n",
    "m.null = lmer(rt ~ 1 + (response_switch | subj), data.corr, subset=response_switch != \"nan\")\n",
    "lr.test(m, m.null, \"main effect of response switch\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Supplemental behavioral figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABAwAAAEmCAYAAADm7AKBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10c4150d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "with sns.plotting_context({\"lines.linewidth\": .3}):\n",
    "    f, axes = plt.subplots(1, 5, figsize=(7, 2), sharey=True)\n",
    "sns.despine()\n",
    "\n",
    "reg_kws = dict(units=\"subj\", logx=True, scatter_kws={\"s\": 12}, line_kws={\"lw\": 1.5})\n",
    "point_kws = dict(units=\"subj\", n_boot=10000, order=[False, True])\n",
    "\n",
    "# ===================================================================================\n",
    "# Reproduce panel a from the main figure\n",
    "\n",
    "axes[0].set(xlim=(.1, .9))\n",
    "sns.regplot(\"pe_context\", \"rt_within\", data_corr, color=\"#429c5c\", x_bins=6, ax=axes[0], **reg_kws)\n",
    "axes[0].set(xlabel=\"Context prediction error\", xticks=[.1, .3, .5, .7, .9])\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Show the main effect of context switches\n",
    "\n",
    "sns.pointplot(\"context_switch\", \"rt_within\", data=data_corr,\n",
    "              color=sns.xkcd_rgb[\"dark rose\"], ax=axes[1], **point_kws)\n",
    "axes[1].set(xlabel=\"Context switch\")\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Show the main effect of cue switches, only on repeat trials\n",
    "\n",
    "# Remove across subject intercept variance for just repeat trials \n",
    "data_repeat = data_corr[~data_corr.context_switch].copy()\n",
    "data_repeat[\"rt_within\"] = data_repeat.groupby(\"subj\").rt.transform(remove_intercept_variance,\n",
    "                                                                    data_repeat.rt.median())\n",
    "sns.pointplot(\"frame_switch\", \"rt_within\", data=data_repeat,\n",
    "              color=sns.xkcd_rgb[\"dull blue\"], ax=axes[2], **point_kws)\n",
    "axes[2].set(xlabel=\"Cue switch\")\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Show the main effect of stimulus congruency\n",
    "\n",
    "sns.pointplot(\"evidence_switch\", \"rt_within\", data=data_repeat,\n",
    "              color=sns.xkcd_rgb[\"bronze\"], ax=axes[3], **point_kws)\n",
    "axes[3].set(xlabel=\"Evidence switch\")\n",
    "\n",
    "# ===================================================================================\n",
    "\n",
    "# -----------------------------------------------------------------------------------\n",
    "# Show the main effect of response switches\n",
    "sns.pointplot(\"response_switch\", \"rt_within\", data=data_corr[~data_corr.response_switch.isnull()],\n",
    "              color=sns.xkcd_rgb[\"dusty purple\"], ax=axes[4], **point_kws)\n",
    "axes[4].set(xlabel=\"Response switch\")\n",
    "\n",
    "axes[0].set_ylim(1.1, 1.4001)\n",
    "\n",
    "start = .02\n",
    "panels = list(\"ABCDE\")\n",
    "for i, ax in enumerate(axes):\n",
    "    ax.set_ylabel(\"\" if i else \"Reaction time (s)\")\n",
    "    f.text(start + (bool(i) * .05) + (i * .18), .9, panels[i], size=12)\n",
    "        \n",
    "f.tight_layout(w_pad=.3)\n",
    "savefig(f, \"behavioral_supplement\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Relationship between response accuracy and CPE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Generalized linear mixed model fit by maximum likelihood (Laplace\n",
       "  Approximation) [glmerMod]\n",
       " Family: binomial  ( logit )\n",
       "Formula: correct ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.resp\n",
       "\n",
       "     AIC      BIC   logLik deviance df.resid \n",
       "  6266.1   6301.5  -3128.0   6256.1     8751 \n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-6.5789  0.2603  0.3424  0.4222  0.5109 \n",
       "\n",
       "Random effects:\n",
       " Groups Name        Variance Std.Dev. Corr \n",
       " subj   (Intercept) 0.07904  0.2811        \n",
       "        log.cpe     0.05183  0.2277   -1.00\n",
       "Number of obs: 8756, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error z value Pr(>|z|)    \n",
       "(Intercept)   1.7592     0.1148  15.329  < 2e-16 ***\n",
       "log.cpe      -0.3380     0.1059  -3.193  0.00141 ** \n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.238 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE):\n",
       "  Chisq(1) = 9.75; p = 0.00179\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = glmer(correct ~ log.cpe + (log.cpe | subj), data.resp, family=binomial)\n",
    "m.null = glmer(correct ~ 1 + (log.cpe | subj), data.resp, family=binomial)\n",
    "lr.test(m, m.null, \"main effect of log(CPE)\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "pe_context\n",
       "(0.132, 0.251]    0.89\n",
       "(0.251, 0.343]    0.89\n",
       "(0.343, 0.436]    0.89\n",
       "(0.436, 0.547]    0.86\n",
       "(0.547, 0.817]    0.87\n",
       "Name: correct, dtype: float64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "q = data_resp.pe_context.quantile(np.linspace(0, 1, 6))\n",
    "data_resp.groupby(pd.cut(data_resp.pe_context, q)).correct.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interaction between congruency and log(CPE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Generalized linear mixed model fit by maximum likelihood (Laplace\n",
       "  Approximation) [glmerMod]\n",
       " Family: binomial  ( logit )\n",
       "Formula: correct ~ log.cpe * congruent + (log.cpe | subj) + (congruent |  \n",
       "    subj)\n",
       "   Data: data.resp\n",
       "Control: glmerControl(optimizer = \"bobyqa\")\n",
       "\n",
       "     AIC      BIC   logLik deviance df.resid \n",
       "  6235.8   6306.6  -3107.9   6215.8     8746 \n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-7.3426  0.2603  0.3307  0.4161  0.6162 \n",
       "\n",
       "Random effects:\n",
       " Groups Name          Variance Std.Dev. Corr \n",
       " subj   (Intercept)   0.07324  0.2706        \n",
       "        log.cpe       0.05756  0.2399   -1.00\n",
       " subj.1 (Intercept)   0.03066  0.1751        \n",
       "        congruentTRUE 0.12096  0.3478   -1.00\n",
       "Number of obs: 8756, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                      Estimate Std. Error z value Pr(>|z|)    \n",
       "(Intercept)             1.5226     0.1383  11.007  < 2e-16 ***\n",
       "log.cpe                -0.4527     0.1280  -3.536 0.000406 ***\n",
       "congruentTRUE           0.6555     0.1980   3.311 0.000931 ***\n",
       "log.cpe:congruentTRUE   0.3372     0.1655   2.037 0.041658 *  \n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) log.cp cnTRUE\n",
       "log.cpe      0.389              \n",
       "congrntTRUE -0.568 -0.406       \n",
       "lg.cp:cTRUE -0.451 -0.536  0.810\n",
       "\n",
       "Likelihood ratio test for interaction between log(CPE) and congruency:\n",
       "  Chisq(1) = 4.12; p = 0.0424\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = glmer(correct ~ log.cpe * congruent + (log.cpe | subj) + (congruent | subj),\n",
    "          data.resp, family=binomial,\n",
    "          control=glmerControl(optimizer=\"bobyqa\"))\n",
    "m.null = glmer(correct ~ log.cpe + congruent + (log.cpe | subj) + (congruent | subj),\n",
    "               data.resp, family=binomial,\n",
    "               control=glmerControl(optimizer=\"bobyqa\"))\n",
    "lr.test(m, m.null, \"interaction between log(CPE) and congruency\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Main effect of log(CPE) on congruent trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Generalized linear mixed model fit by maximum likelihood (Laplace\n",
       "  Approximation) [glmerMod]\n",
       " Family: binomial  ( logit )\n",
       "Formula: correct ~ log.cpe + (log.cpe | subj)\n",
       "   Data: data.resp\n",
       "Control: glmerControl(optimizer = \"bobyqa\")\n",
       " Subset: congruent\n",
       "\n",
       "     AIC      BIC   logLik deviance df.resid \n",
       "  2632.2   2663.8  -1311.1   2622.2     4083 \n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-5.7659  0.2885  0.3173  0.3561  0.4423 \n",
       "\n",
       "Random effects:\n",
       " Groups Name        Variance  Std.Dev. Corr \n",
       " subj   (Intercept) 2.526e-01 0.502545      \n",
       "        log.cpe     9.379e-05 0.009684 -1.00\n",
       "Number of obs: 4088, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "            Estimate Std. Error z value Pr(>|z|)    \n",
       "(Intercept)  2.24793    0.19917  11.286   <2e-16 ***\n",
       "log.cpe     -0.03642    0.13794  -0.264    0.792    \n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "        (Intr)\n",
       "log.cpe 0.689 \n",
       "\n",
       "Likelihood ratio test for main effect of log(CPE) on congruent trials:\n",
       "  Chisq(1) = 0.07; p = 0.793\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = glmer(correct ~ log.cpe + (log.cpe | subj),\n",
    "          data.resp, subset=congruent, family=binomial, \n",
    "          control=glmerControl(optimizer=\"bobyqa\"))\n",
    "m.null = glmer(correct ~ 1 + (log.cpe | subj),\n",
    "               data.resp, subset=congruent, family=binomial,\n",
    "               control=glmerControl(optimizer=\"bobyqa\"))\n",
    "lr.test(m, m.null, \"main effect of log(CPE) on congruent trials\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Interaction of early cue and congruent trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Generalized linear mixed model fit by maximum likelihood (Laplace\n",
       "  Approximation) [glmerMod]\n",
       " Family: binomial  ( logit )\n",
       "Formula: correct ~ early * congruent + (early | subj) + (congruent | subj)\n",
       "   Data: data.resp\n",
       "Control: glmerControl(optimizer = \"bobyqa\", optCtrl = list(maxfun = 50000))\n",
       "\n",
       "     AIC      BIC   logLik deviance df.resid \n",
       "  6241.6   6312.4  -3110.8   6221.6     8746 \n",
       "\n",
       "Scaled residuals: \n",
       "    Min      1Q  Median      3Q     Max \n",
       "-5.5018  0.2526  0.3237  0.4106  0.6292 \n",
       "\n",
       "Random effects:\n",
       " Groups Name          Variance Std.Dev. Corr \n",
       " subj   (Intercept)   0.10044  0.3169        \n",
       "        earlyTRUE     0.08103  0.2847   -0.02\n",
       " subj.1 (Intercept)   0.18411  0.4291        \n",
       "        congruentTRUE 0.10630  0.3260   -0.63\n",
       "Number of obs: 8756, groups:  subj, 15\n",
       "\n",
       "Fixed effects:\n",
       "                        Estimate Std. Error z value Pr(>|z|)    \n",
       "(Intercept)               1.8465     0.1512  12.209  < 2e-16 ***\n",
       "earlyTRUE                 0.2161     0.1172   1.844 0.065132 .  \n",
       "congruentTRUE             0.4734     0.1315   3.601 0.000317 ***\n",
       "earlyTRUE:congruentTRUE  -0.2818     0.1365  -2.065 0.038942 *  \n",
       "---\n",
       "Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1\n",
       "\n",
       "Correlation of Fixed Effects:\n",
       "            (Intr) erTRUE cnTRUE\n",
       "earlyTRUE   -0.219              \n",
       "congrntTRUE -0.481  0.213       \n",
       "erTRUE:TRUE  0.171 -0.470 -0.518\n",
       "\n",
       "Likelihood ratio test for interaction of early cue and congruent trials:\n",
       "  Chisq(1) = 4.23; p = 0.0398\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%%R\n",
    "m = glmer(correct ~ early * congruent + (early | subj) + (congruent | subj),\n",
    "          data.resp, family=binomial,\n",
    "          control=glmerControl(optimizer=\"bobyqa\", optCtrl=list(maxfun=50000)))\n",
    "m.null = glmer(correct ~ early + congruent + (early | subj) + (congruent | subj),\n",
    "               data.resp, family=binomial,\n",
    "               control=glmerControl(optimizer=\"bobyqa\", optCtrl=list(maxfun=50000)))\n",
    "lr.test(m, m.null, \"interaction of early cue and congruent trials\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}