LUHMES.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy as sp\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "sns.set_theme()\n",
    "\n",
    "from tqdm import tqdm\n",
    "\n",
    "import scanpy as sc\n",
    "\n",
    "from causarray import *\n",
    "\n",
    "import os\n",
    "path_result = 'results/LUHMES/'\n",
    "os.makedirs(path_result, exist_ok=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<KeysViewHDF5 ['cell_ids', 'counts', 'covariates', 'gene_names', 'sgRNA']>\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4881, 1944) (4881,) (4881,) (4881, 8)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "import h5py\n",
    "\n",
    "with h5py.File('data/LUHMES/LUHMES.h5', 'r') as f:\n",
    "    print(f.keys())\n",
    "    counts = np.array(f['counts'], dtype='float')\n",
    "    sgRNA = np.array(f['sgRNA'], dtype='S32').astype(str)\n",
    "    gene_names = np.array(f['gene_names'], dtype='S32').astype(str)\n",
    "    cell_ids = np.array(f['cell_ids'], dtype='S32').astype(str)\n",
    "    covariates = pd.DataFrame(np.array(f['covariates'])).astype(float).values\n",
    "    pt_state = covariates[:,-1]\n",
    "    \n",
    "    covariates = np.c_[np.ones(covariates.shape[0]), np.log(np.sum(counts, axis=1)),  covariates]    \n",
    "\n",
    "    id_cell = (~np.char.startswith(sgRNA, 'RELN')) & (pt_state>3)\n",
    "    counts = counts[id_cell] \n",
    "    sgRNA = sgRNA[id_cell]\n",
    "    cell_ids = cell_ids[id_cell]\n",
    "    covariates = covariates[id_cell]\n",
    "    pt_state = pt_state[id_cell]\n",
    "\n",
    "    covariates = np.c_[\n",
    "        covariates[:,:-2],\n",
    "        OneHotEncoder(drop='first', sparse_output=False).fit_transform(covariates[:,-2:])\n",
    "    ]\n",
    "\n",
    "    print(counts.shape, sgRNA.shape, cell_ids.shape, covariates.shape)\n",
    "\n",
    "gRNA_names = [\"PTEN\",\"CHD2\",\"ASH1L\",\"ADNP\"]    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PTEN 1944 (1014, 1444) (1014,) 458 (1014, 8)\n"
     ]
    }
   ],
   "source": [
    "gRNA_name = gRNA_names[0]\n",
    "\n",
    "id_cells = ((np.char.startswith(sgRNA, gRNA_name) | np.char.startswith(sgRNA, 'Nontargeting')))\n",
    "\n",
    "Y = counts[id_cells]\n",
    "A = np.char.startswith(sgRNA[id_cells], gRNA_name)\n",
    "W = covariates[id_cells]\n",
    "W = W / np.sqrt(np.sum(W**2, 0, keepdims=True)) * np.sqrt(W.shape[0])\n",
    "\n",
    "id_genes = np.any(Y!=0, axis=0) & (np.sum(Y>0, axis=0)>=50)\n",
    "Y = Y[:,id_genes]\n",
    "print(\n",
    "    gRNA_name, \n",
    "    len(id_genes), \n",
    "    Y.shape, A.shape, np.sum(A), W.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "colors = [sns.color_palette(\"muted\")[1], sns.color_palette(\"muted\")[0]]\n",
    "sns.set(font_scale=1.5)\n",
    "fig, axes = plt.subplots(1, 2, figsize=(10,4), sharey='row', sharex = True)\n",
    "\n",
    "for i,name in enumerate(['PPP1R1C','GNG3']):\n",
    "    j = np.where(gene_names[id_genes]==name)[0][0]\n",
    "    axes[i] = sns.histplot(pd.DataFrame({'x':Y[:,j], 'condition':np.where(A==1, 'treatment', 'control')}), \n",
    "        x = 'x', hue = 'condition', hue_order=['treatment', 'control'],\n",
    "        palette=colors, linewidth=0.1, edgecolor='k', ax=axes[i])\n",
    "    \n",
    "    for k in range(2):\n",
    "        axes[i].axvline(np.mean(Y[A==k,j]), color=colors[k], linestyle='--')\n",
    "    axes[i].set_title(gene_names[id_genes][j])\n",
    "\n",
    "    axes[i].set_xlabel('Gene expression levels')\n",
    "\n",
    "plt.savefig(path_result + 'LUHMES_gene_expression_levels.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Doubly robust estimation and inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PTEN 1944 (1014, 1444) (1014,) 458 (1014, 8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CHD2 1944 (756, 1360) (756,) 200 (756, 8)\n",
      "ASH1L 1944 (725, 1360) (725,) 169 (725, 8)\n",
      "ADNP 1944 (895, 1416) (895,) 339 (895, 8)\n"
     ]
    }
   ],
   "source": [
    "for gRNA_name in gRNA_names:\n",
    "\n",
    "    id_cells = (np.char.startswith(sgRNA, gRNA_name) | np.char.startswith(sgRNA, 'Nontargeting'))\n",
    "\n",
    "    Y = counts[id_cells]\n",
    "    A = np.char.startswith(sgRNA[id_cells], gRNA_name)\n",
    "    W = covariates[id_cells]\n",
    "    W = W / np.sqrt(np.sum(W**2, 0, keepdims=True)) * np.sqrt(W.shape[0])\n",
    "\n",
    "    id_genes = np.any(Y!=0, axis=0) & (np.sum(Y>0, axis=0)>=50)\n",
    "    Y = Y[:,id_genes]\n",
    "    print(gRNA_name, len(id_genes), Y.shape, A.shape, np.sum(A), W.shape)\n",
    "\n",
    "    # record covariate adjusted residual for plotting later\n",
    "    if gRNA_name == 'PTEN':\n",
    "        df_glm_p, df_glm_nb = glm_test(Y, W, A[:,None])\n",
    "        resid = fit_glm(Y, W, np.zeros((Y.shape[0],0)), family='nb')[-1]\n",
    "\n",
    "    for func in [ATE, SATE]:\n",
    "        df = func(Y, W, A, B=10000, alpha=0.05, c=0.01, family='nb')\n",
    "        df.to_csv(path_result+'{}_{}.csv'.format(gRNA_name, func.__name__), index=False)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Residual plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "gRNA_name = 'PTEN'\n",
    "id_cells = (np.char.startswith(sgRNA, gRNA_name) | np.char.startswith(sgRNA, 'Nontargeting'))\n",
    "\n",
    "Y = counts[id_cells]\n",
    "A = np.char.startswith(sgRNA[id_cells], gRNA_name)\n",
    "W = covariates[id_cells]\n",
    "W = W / np.sqrt(np.sum(W**2, 0, keepdims=True)) * np.sqrt(W.shape[0])\n",
    "\n",
    "id_genes = np.any(Y!=0, axis=0) & (np.sum(Y>0, axis=0)>=50)\n",
    "Y = Y[:,id_genes]\n",
    "\n",
    "df_ATE = pd.read_csv(path_result+'{}_ATE.csv'.format(gRNA_name))\n",
    "df_STE = pd.read_csv(path_result+'{}_SATE.csv'.format(gRNA_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x648 with 14 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_col = 5\n",
    "colors = [sns.color_palette(\"muted\")[1], sns.color_palette(\"muted\")[0]]\n",
    "sns.set(font_scale=1.5)\n",
    "fig, axes = plt.subplots(3, n_col, figsize=(15,9), sharey='row', sharex=True)\n",
    "\n",
    "idx = np.where((df_ATE['rej']==1)&(df_STE['rej']==1))[0]\n",
    "for ij in range(n_col):\n",
    "    j = idx[np.argsort(np.abs(df_STE['tvalues_init'].values)[idx])[ij]]\n",
    "    axes[0,ij] = sns.histplot(pd.DataFrame({'x':resid[:,j], 'condition':np.where(A==1, 'treatment', 'control')}), \n",
    "        x = 'x', hue = 'condition', hue_order=['treatment', 'control'], bins=np.linspace(-10,10,50),\n",
    "        palette=colors, linewidth=0.1, edgecolor='k', ax=axes[0,ij])\n",
    "    axes[0,ij].set_title(gene_names[id_genes][j])\n",
    "\n",
    "i = 2\n",
    "idx = np.where((df_ATE['rej']==1)&(df_STE['rej']==0))[0]\n",
    "for ij in range(n_col):\n",
    "    j = idx[np.argsort(np.abs(df_STE['tvalues_init'].values)[idx])[ij]]\n",
    "    axes[i,ij] = sns.histplot(pd.DataFrame({'x':resid[:,j], 'condition':np.where(A==1, 'treatment', 'control')}), \n",
    "        x = 'x', hue = 'condition', hue_order=['treatment', 'control'], bins=np.linspace(-10,10,50),\n",
    "        palette=colors, linewidth=0.1, edgecolor='k', ax=axes[i,ij])\n",
    "    axes[i,ij].set_title(gene_names[id_genes][j])\n",
    "\n",
    "i = 1\n",
    "idx = np.where((df_ATE['rej']==0)&(df_STE['rej']==1))[0]\n",
    "for ij in range(n_col):\n",
    "    if ij>=len(idx):\n",
    "        fig.delaxes(axes[i,ij])\n",
    "        continue\n",
    "    j = idx[np.argsort(np.abs(df_STE['tvalues_init'].values)[idx])[ij]]\n",
    "    axes[i,ij] = sns.histplot(pd.DataFrame({'x':resid[:,j], 'condition':np.where(A==1, 'treatment', 'control')}), \n",
    "        x = 'x', hue = 'condition', hue_order=['treatment', 'control'], bins=np.linspace(-10,10,50),\n",
    "        palette=colors, linewidth=0.1, edgecolor='k', ax=axes[i,ij])\n",
    "    axes[i,ij].set_title(gene_names[id_genes][j])\n",
    "\n",
    "\n",
    "legend = axes[0,0].get_legend(); handles = legend.legendHandles\n",
    "labels = [t.get_text() for t in legend.get_texts()]\n",
    "for i in range(3):\n",
    "    axes[i,-1].set_xlim([-6,6])    \n",
    "    for j in range(n_col):\n",
    "        try:\n",
    "            axes[i,j].get_legend().remove()\n",
    "        except:\n",
    "            pass\n",
    "        if i==2:\n",
    "            axes[i,j].set_xlabel('Residuals')\n",
    "\n",
    "axes[0,0].set_ylim([0,200])\n",
    "axes[1,0].set_ylim([0,200])\n",
    "axes[2,0].set_ylim([0,250])\n",
    "\n",
    "legend = fig.legend(handles=handles, labels=labels, loc=9, ncol=10, title='Condition', frameon=False, fontsize=14)\n",
    "def legend_title_left(leg):\n",
    "    c = leg.get_children()[0]\n",
    "    title = c.get_children()[0]\n",
    "    hpack = c.get_children()[1]\n",
    "    c._children = [hpack]\n",
    "    hpack._children = [title] + hpack.get_children()\n",
    "legend_title_left(legend)\n",
    "\n",
    "axes[0,0].text(axes[0,0].get_xlim()[0]-5, axes[0,0].get_ylim()[1], '(a)', fontsize=16, weight='bold')\n",
    "axes[1,0].text(axes[1,0].get_xlim()[0]-5, axes[1,0].get_ylim()[1], '(b)', fontsize=16, weight='bold')\n",
    "axes[2,0].text(axes[2,0].get_xlim()[0]-5, axes[2,0].get_ylim()[1], '(c)', fontsize=16, weight='bold')\n",
    "\n",
    "\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(top=0.9)\n",
    "plt.savefig('figures/LUHMES_{}_residuals.pdf'.format(gRNA_name), bbox_inches='tight', pad_inches=0, dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Summarized statistics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import scipy as sp\n",
    "def summarize(Y, resid, A):\n",
    "    res_diff = np.r_[\n",
    "        np.abs(np.mean(resid[A==1]) - np.mean(resid[A==0]))/2,\n",
    "        np.abs(np.mean((resid+Y)[A==1]) - np.mean((resid+Y)[A==0]))/np.std((resid+Y)[A==0]),\n",
    "        np.abs(np.median((resid+Y)[A==1]) - np.median((resid+Y)[A==0]))/np.std((resid+Y)[A==0])\n",
    "    ]\n",
    "\n",
    "    res_overall = np.r_[\n",
    "        np.mean(Y==0),         \n",
    "    ]\n",
    "\n",
    "    res = np.r_[res_diff, res_overall]\n",
    "    return res\n",
    "    \n",
    "idx_list = [\n",
    "    np.where((df_ATE['rej']==1)&(df_STE['rej']==1))[0],    \n",
    "    np.where((df_ATE['rej']==0)&(df_STE['rej']==1))[0],\n",
    "    np.where((df_ATE['rej']==1)&(df_STE['rej']==0))[0],\n",
    "\n",
    "]\n",
    "\n",
    "res_all = []\n",
    "for i, idx in enumerate(idx_list):\n",
    "    for j in idx:\n",
    "        res = np.r_[summarize(Y[:,j], resid[:,j], A), j, i]\n",
    "        \n",
    "        res_all.append(res)\n",
    "res_all = np.array(res_all)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "               Absolute difference\\nof residual mean  \\\n",
      "Test                                                   \n",
      "ATE only (10)                               0.137157   \n",
      "Common (22)                                 0.307546   \n",
      "STE only (4)                                0.231693   \n",
      "\n",
      "               Absolute difference\\nof fitted response mean  \\\n",
      "Test                                                          \n",
      "ATE only (10)                                      0.436272   \n",
      "Common (22)                                        0.250622   \n",
      "STE only (4)                                       0.151403   \n",
      "\n",
      "               Absolute difference\\nof fitted response median  \\\n",
      "Test                                                            \n",
      "ATE only (10)                                        0.032957   \n",
      "Common (22)                                          0.153083   \n",
      "STE only (4)                                         0.152631   \n",
      "\n",
      "               Proportion of\\nzero counts      j  count  \n",
      "Test                                                     \n",
      "ATE only (10)                    0.746548  275.0   10.0  \n",
      "Common (22)                      0.312623  132.5   22.0  \n",
      "STE only (4)                     0.286489  288.5    4.0  \n"
     ]
    }
   ],
   "source": [
    "col_names = ['Absolute difference\\nof residual mean',\n",
    "    'Absolute difference\\nof fitted response mean',\n",
    "    'Absolute difference\\nof fitted response median',\n",
    "    'Proportion of\\nzero counts', 'j', 'Test']\n",
    "_df = pd.DataFrame(res_all, columns=col_names)\n",
    "_df['Test'] = _df['Test'].map(\n",
    "    {0: 'Common ({})'.format(np.sum(_df['Test']==0)), 1: 'STE only ({})'.format(np.sum(_df['Test']==1)), \n",
    "    2: 'ATE only ({})'.format(np.sum(_df['Test']==2))})\n",
    "_df['count'] = _df.groupby('Test')['Test'].transform('count')\n",
    "\n",
    "_df.to_csv(path_result+'{}_summary.csv'.format(gRNA_name), index=False)\n",
    "\n",
    "print(_df.groupby('Test').median())\n",
    "_df = _df.melt(id_vars=['j', 'Test'], value_vars=col_names[:-2], value_name='Value', var_name='Metric')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9oAAAF+CAYAAAB9BWKWAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAB2zUlEQVR4nO3dd1QU19sH8C9dqVJEbFhQsFAEFREVlRgRsEVj7BVjoVixYWKPYIsR1NhbYo+910SiRmPvKWJQUEEEZQER2N15//Blfq5L1YWlfD/ncM7unblzn5ndu8uzc+eOhiAIAoiIiIiIiIhIJTTVHQARERERERFRWcJEm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIVYqJNREREREREpEJMtImIiIiIiIhUSK2J9uPHjzFjxgx069YNjRo1QufOnQtcd//+/ejUqRMcHBzg6+uLo0ePFmGkRERERERERAWjrc7G//33X5w7dw5OTk6Qy+Uo6C29jx8/jilTpmDEiBFo1aoVTp8+jQkTJsDAwABt27Yt4qiJiIiIiIiIcqchFDS7LQJyuRyamu9Oqk+dOhV3797F4cOH863n7e0NW1tbLFu2TCzz8/NDcnIyfvnllyKLl4iIiIiIiCg/ah06np1kF0ZMTAwePXoEX19fhXJfX1/cuXMHSUlJqgqPiIiIiIiIqNBK3WRojx49AgDY2NgolNerV09hOREREREREZE6lLpEOzk5GQBgbGysUG5iYqKwnIiIiIiIiEgdSl2inU1DQ0Phefal5h+WExERERERERUntc46/jHeP3NtYWEhlkskEgDKZ7rz8+pVGuRytc0HpzKJiS8xa1YIAGDWrPkwN7fIpwYRERERERF9DE1NDZiaGuS6vNQl2nXr1gXw7lrs96/TjoqKUlheUHK5UCYS7awsKeLj48XHZWGfiIiIiIiISqNSN3S8Zs2aqFu3Lo4ePapQfvjwYTg4OMDMzExNkRERERERERGp+Yx2eno6zp07BwB4+vQpUlNTcfz4cQCAg4MDqlevjpCQEOzfvx/3798X640ZMwbjx4+HtbU13N3dcebMGVy4cAGrV69Wy34QERERERERZVNrop2YmIixY8cqlGU/Dw0NRY8ePSCXyyGTyRTW8fb2xtu3b7Fq1SqsX78e1tbWWLJkCdq2bVtssRMRERERERHlREPInq67nEpMTC0T1zO/eBGPMWNGAgDCw1fD0rKKmiMiIiIiIiIqmzQ1NWBubpjr8lI3GRoREVFBSaVZSEuTICMjHXK5LP8KRCqgqakFPb2KMDAwhra2jrrDISIiNWCiTUREZZJUmoWkpHjo6xvBzMwKWlpa0NDQUHdYVMYJggCZTIa3b9OQlBQPM7MqTLaJiMqhUjfrOBERUUGkpUmgr28EQ0MTaGtrM8mmYqGhoQFtbW0YGppAX98IaWkSdYdERERqwESbiIjKpIyMdFSoYKDuMKgcq1DBABkZ6eoOg4iI1ICJNhERlUlyuQxaWlrqDoPKMS0tLc4NQERUTjHRJiKiMovDxUmd+P4jIiq/mGgTERERERERqRATbSIioo8UGDgCgYEj1NJOeno6Fi2aj27dOqF162b47rtZAIDWrZth/frVRR4TERER5Y639yIiIvpAVNRDbNq0Dg8e3ENSUiKMjIxRs6Y1nJ2bws9vpLrDAwBs3/4TDhzYi4EDh6J27TqoXr2GukNS8OjRQ/z66xn4+HRB1arV8lz3yy+7IC7ueb7bHDr0a5Uc/xMnjiI5+TW++qrfJ2+LiIgoJ0y0iYiI3nP79k2MHTsa5uYW8PbuDEvLKnj5MgH37t3Bli0bFBK9pUtXFEtMObVz/fpV1Ktni5EjAxTKz5y5UCImgXv0KAobN66Fs3PTfBPtMWMmIj39jfj83LlfERn5K8aPnwxDQ0Ox3MamvkpiO3HiGJ48iWaiTURERYaJNhER0Xt++mkjKlSoiHXrfkKlSpUUlr18maDwXEdHp1hiyqmd169foVIlU6VyPT294ghJpTw82ik8j42NQWTkr2jXzhPm5hbqCYqIiOgTMNEmIiJ6z9OnsahTp65Skg0AFhaVFZ5nXze9fPkasUwikWD58qWIjPwVcrkAV9cWGDduMrp376Qw9Hn9+tXYuHEttm/fix07fsavv55BVlYmWrRwx6RJ02BiUinHdq5fv4oxY0b9/5L/0Lp1MwBAePgquLg0Q+vWzZSGWKempmLz5vX47bezePnyBUxMKqFJExcEBIxF5cqWyMrKwubN6/HHHxfw9GksMjMzYWNTD4MHD0Pr1m0V9rl162bo1q0HWrduizVrVuDx42hYWlaBn99IfP55JwDA0aOHMH/+bAB4L1YgJGQmfHy6FPSlUHLlyiVs2bIRf/31AIIgR8OGjTFihD8cHJzEdd68eYMNG9bg3LmzSEx8CQMDQ9SpUxdDhgyHi0szBAaOwM2b18V9yXb+/FUAwJkzp7B9+0948uQxNDSAKlWs0L59Bwwd+vVHx01EROUPE20iIqL3WFlVw507t/Dvv/+gfn3bQtWVy+WYOnUC7ty5hS5duqNePVtcu/YnJk8em2udWbOmw9zcAsOHj0Js7BPs2bML2tramDXruxzXr127Dr79dg5WrgyHoaEhBg0aJpbnJD09HYGBI/Do0UN06uSLhg0bQyJJxqVLFxAbG4PKlS2RlpaG/fv34LPPPoevb1dkZGTg1KljmDYtGIsXh6NFi5YK27x37y5+//0cunXrgc6du+HQoQOYO3cG6te3Q+3adeDk5IwePXph797d4jXkAGBv71io4/m+U6eOY+7cGXB2boavvx4NQZDjyJGDGDt2NCIi1qBxY3sAwJIloTh79jR69OiFOnXqQiJJwf37d/Hw4T9wcWmGwYOHISUlBYmJCQgKmqDQxpUrlzFrVghcXJpj5MgAaGlp4smTx7h588ZHx01EROUTE20iIqL39O8/COPHB8DPbwDs7BrCyckZLi7N0LRp83yHZf/++zncvn0To0YFYsCAIQCAHj16Ye7cGfjnn79zrFOrVm3MmDFXfC4IwN69uxAcPE3h+uRsZmbm8PLywebN62FqagYvL588Y9q2bQsePvwHM2bMRceO3mL54MF+EAQBAGBkZIS9e49AV1dXXP7ll70xdGg/bN/+k1Ki/d9/Udi8eYeYQLdv3wE9e3bGkSMHERAwFtWr14CjYxPs3bsbzZu3gItLM3yK9PR0fP/9Qnh5+WD69FliebduPTFw4FdYs2YFli37EQBw4cJ5dOnSXSmJzta8uRvMzbciLS1V6dj98cd56Ovr4/vvI0rEde5ERFR68fZeRERE72natDlWrFiLVq3aIDr6P+zY8TMmTx6Hrl074siRg3nWvXTpIjQ1NdGjRy+F8l69+uRa54svvlR43qSJC2QyGeLj4z5+J97z229nULt2XYUkO5uGhgYAQEtLS0yys7KyIJEkIy0tDU2auODvv/9Squfi0kzhDLqZmTmsrWvj2bNYlcT8oStXLiMlRYKOHTvh9evX4t/bt2/RrJkrbt26AalUCgAwMDDAgwf3kJDwotDtGBgY4u3bt/jzz0uq3gUiIipneEabiIjoAw4OTggNXQKZTIaoqH9x8eJ57NjxM0JD58DKqiqaNm2eY734+OcwMzOHvr6BQnnNmta5tmVlVVXhuZGREQBAIkn+xL145+nTp2jd2iPf9Q4d2o+dO7fh8eP/xDPdwP+S8fd9GDPwLm6JRPJpweYiJuYxAGD8+MBc10lJSYGpqSlGjw7C/Plz0LNnZ9Svb4cWLVqiY0fvXIfWv6979544e/YUJk0aCwuLymjWzBVt27ZH69ZtczwOREREuWGiTURElAstLS3Y2jaArW0DODg4YezY0Thx4miuifbH0NTMeXDZ+8nup8ovRzx16jgWLJiH1q090L//IJiamkFLSwtHjx7CqVPHldYvjpjfJ5e/2+706bOUJqTLlj3MvkMHLzRp0hTnz5/DlSuX8MsvO7F162aEhMzMd5i9ubkFNm3ajitXLuPy5Yu4fPkPHD9+BG5u7li0aBmTbSIiKjAm2kRERAXQsGFjAMDLly9zXadKlaq4evVPvHmTpnBWOybmSZHHl5vq1asjKioqz3XOnj2FatWqIzR0iUIyefTooY9uV5VJafXqNQAAlSpVQvPmLfJd38LCAt2790T37j2RkpKCkSOHYOPGtWKinVdsOjo6cHdvDXf31hAEAatWLcfWrZtx584tODo2Ucn+EBFR2cdrtImIiN5z7doVyOVypfI//rgA4N3kZblxc3OHXC7H3r27Fcp3796h0hgLo127zxAd/QinT59QWpZ9Bjr7DPX7Z6SfPo1FZOSvH91uhQoVAbwb0v2p3NxawtDQCJs3b0BmZqbS8levXgEAZDIZUlNTFZYZGRmhatXqSEn537D2ihUrIDVVOa7k5NcKzzU0NGBrawdANftBRETlB89oExERveeHHxYhPT0dHh7tUKtWHQiCHH///TdOnDgKExMTfPVV31zrtmnTFvb2jli9egWeP38GG5v6uHbtTzx79hSAas/yFlS/foNw7txZzJnzLa5cuYyGDRshJSUVly9fhJ/fSDg7N0Xr1m1x7tyvmDp1Alq18sDLlwnYt283atashYcP//modu3sGkBTUxM//7wRqakp0NPTQ6NG9qhWrXqht6Wvb4DJk6dj9uzpGDKkLzp29IaZmTkSEl7g+vWrqFChIpYsCcebN2/wxRfeaNvWE/Xq2cLAwAB37tzC5csXFSaos7NrhN9+O4sffliMRo3soampgQ4dvBAWNg/Jya/RtGlzVKlSBQkJCdizZxfMzS3g7OzyUceBiIjKJybaRERE7wkIGIdz587izz8v4fDhg8jKyoSFRWV07NgJgwYNQ9Wq1XKtq6mpiYULlyI8/HucPn0Cp06dgKtrC8yZE4a+fXso3D6ruFSsWBHLl6/Fhg1rEBn5K06cOApTUzM0aeIiTtLm7d0Zr14lYd++Pbh69U9Ur14DQUETEBsb89GJtoVFZUyeHIKff96MhQu/g0wmQ0jIzI9KtAHA07MDKleujJ9+2oidO7chI+MtzM0t0KhRY3Tu3B0AUKFCBfTo0QtXrlzG+fPnIJPJULVqNQQEjFOY+f3LL3sjOvoRTp48hj17dkIQBHTo4AUvL28cPLgfBw7sRUqKBKamZmjZshWGDv1aaYI7IiKivGgIRTVzSSmRmJgqTrJSmr14EY8xY0YCAMLDV8PSsoqaIyIiUq+4uMewsqql7jAAAP/88xeGDRugdC9rKvtK0vuQiIhUR1NTA+bmhrkvL8ZYiIiIyryMjLdKZTt3boWmpiaaNOHwYyIiovKAQ8eJiIhUaOnSRUhNTYWjoxMADfzxx3lcuXIZ3bt/ydFGRERE5QQTbSIiIhVycWmOXbu24erVy3j79i2srKphxAh/9O8/WN2hERERUTFhok1ERKRCHTt2QseOndQdBhEREakRr9EmIiIiIiIiUiEm2kREREREREQqxESbiIiIiIiISIWYaBMRERERERGpEBNtIiIiIiIiIhViok1ERERERESkQky0iYiIiIiIiFSI99EmIiIqJS5f/gO7d2/Hgwf3kJaWBjMzczRv3gJ9+w5E7dp11B1eifHjjxGIjn6EBQuWAgCuXLmMw4cP4P79u0hKSkTVqtXQqZMvevfuDx0dHbHe/v17EBn5Kx4+/Bdv375FrVq1MWDAELRt215cRyaToV+/nvDzG8X7pRMRUa6YaBcRI+MKqKCnk/+KKiKTpYmPzc0NUbmyUbG1ne1tRhZSJG+LvV0iosIq7s/obJ/yObl+/Wps3LgWbdq0xcSJ02BmZob4+DicPHkco0YNw/Hjv6o42tIpMfEl9uzZiR9++FEsO3BgLzIy3mLYsBGoUsUK9+7dxcaNaxEd/R+++Wa2uN6WLRvQokVLdOvWE/r6FXH27BlMnz4J06bNgK9vVwCAlpYW+vcfjPXrV8HTswO0tfmvFBERKeO3QxGpoKeDfpO3Flt7sowU8fHY0P3Q0iv+RHvbwv5IARNtIir5ivszOtvHfk7++eclbNy4FgMHDsXIkQEKy7y8fHD+/DlVhVjqHTiwF1WrVoO9vYNYFhw8DZUqVRKfu7g0gyDIsXbtjwgMHC8u27Bhq8J6zZu7IS7uGXbu3Com2gDQoYMXli1bjIsXz8PDo10R7xEREZVGvEabiIiohNu2bQtMTc3g5zcyx+WtW7cVH8vlcmzZsgG9enVD+/Yt0bt3d+zcqfijwvr1q/H5523w8OG/GD3aD56erTBoUG/cuXMLUqkUERHfo3PnDujWzQvr16+GIAgqqQsAt27dgL//cHh6toK3tydmz/4GiYkvxeXPnz9D69bNcPLkMfzww2J4e3uiS5eOWLw4DBkZGfkeq2PHDuOzzzoqlL2fPGerX98OgiDg9etX+a736tUrhTJ9fX24ubnj2LHD+cZDRETlExNtIiKiEkwqleL27Vto1sy1QMOUV6xYhnXrVuHzz72wYMFSeHi0R0TEUmzatE5pu3PmfANf3y747ruF0NHRRUjIJCxaNB8ZGRmYOXMefHy6YuPGtYiM/FUldf/66wHGjfOHjo4O5syZjzFjJuD69SsYO9ZfKYles2YlpFIpZs+ej969++HQoX3Yvv2nPPc9NjYGz58/g4ODU77H6fbtm9DR0UH16jXyXS+n698dHZvg+vUrkMlk+bZFRETlD4eOExERlWASSTIyMzNQpYpVvuu+fv0ae/bsRJ8+/TFihD8AwNXVDWlpqdi6dTO++qof9PX1AQBZWVkYNSoI7u6tAQB6enoYM2YUEhNfYvHicADvhk6fP38OZ8+eRtu2nmI7H1t3y5YNMDU1w+LF4eIkZDVrWmPUqGE4c+YkfHy6iG00aNAQwcFTxX24d+8OfvvtLIYMGZ7r/j94cA8AYGNTP8/jFBPzBL/8sgNdunRXmAztQydPHsfdu7cxd26Y0rL69e2QlpaGx4+jUbeuTZ7tERFR+cMz2kRERCVY9tBrDQ2NfNe9f/8upFIpPD0Vh0536OCF9PR0/Pvv32KZpqYmmjVzFZ9bW9cC8C6pfV/NmrXw4kW8QtnH1r19+wbatGmrkNza2zvCyqoqbt26oVDX1bWlwvPatesiIUExjg8lJr6EpqYmjI2Nc10nLS0V06dPQpUqVTFqVGCu6z18+C8WLw6Fl5cP2rfvoLTcxKSS2CYREdGHeEabiIioBDMxqQRdXT3Ex8flu25KigQAYG5urlBubm4B4N3Z8Wx6enrQ1dUVn2trv0t+DQ0VJ9PU0dFGZmamQtnH1k1JSYGZmWJsAGBmZg6JRKJQZmSkuC1tbeU4PpSZmQktLS1oauZ8HiErKwshIZOQkpKCH3/cAH19gxzXi4t7juDgMWjYsDGmTv02x3Wy978g140TEVH5wzPaREREJZi2tjacnJrg6tU/IZVK81w3+0xuUlKSQnn2WVdjY5OiCbKAjIyM8epVklJ5UlJinmehC8rY2BhZWVk5Jr9yuRyzZ3+Dv/9+gMWLw2FllfNQ/NevX2PChECYmppi/vyFuQ4tz/5Rw8REvceUiIhKJibaREREJVzfvgORlJSIjRvX5rj8woXfAQANG9pDW1sbZ8+eUlh+9uwpVKxYEba2DYo81rw4OjZBZORvCj8Y3Lt3F3Fxz+Hk5PzJ27e2rg3g3czlH/r++wW4ePF3zJ+/GDY29XKs/+bNGwQHj0FWVhYWLQqHgYFhrm1lt1GzZq1PjpuIiMoeDh0nIiIq4Vxd3TB06NfYuHEt/vvvET7/3AtmZuaIj4/H6dMncOfOLRw7dhaVKlXCl1/2wY4dP0NXVxcODk64evVPHDiwF35+I1GxYkW17segQcMwevQwBAePQa9efSGRJGP16hWoXbuu0i25PkbDho2hpaWFv/66rzBT+JYtG7B//x706zcQurp6uHv3jrisevUaMDU1BQBMnz4J//77N6ZNm4G4uOeIi3surmdra6cwXP7Bg3dt5HRLMCIiIibaRERU7rzNyMK2hf3V0u7H8vMbicaNHbB79w4sWhSKtLRUmJtbwNm5KZYtWymu5+8/BkZGRjh0aD9++mkjLC2rICBgLPr0GaCKXfgkDRo0xNKlK7B69XJ8++1U6Orqws3NHUFB46Gnp/fJ269YsSJatmyFP/64gE6dfMXyP/+8BADYtu0nbNumeIuwkJCZ4mznV65cBgDMmzdTadu7dx9E1arVxOeXLl1Au3affXLMRERUNmkI2dOZllOJiamQy1V/CCpXNkK/yVtVvt3cyDJS8PLubgCAhX0vaOkZ5VND9bYt7I+EhJRib5eIKCdxcY9hZcVhveXNhQu/Y9asEBw4cEK8lZmq/fvvPxg+fCC2b9+LatWq57ku34dERGWTpqYGzM1zv8SI12gTERFRmdGqVRvUqlUH+/fvKbI2duz4GV5ePvkm2UREVH6pNdGOjo6Gn58fnJ2d4ebmhrlz5yI9PT3fem/evMHixYvRoUMHODk5oWPHjli+fHm+t/0gIiKisi84eBr09YvmenSZTAZr61oYPnxUkWyfiIjKBrVdoy2RSDBo0CBUq1YNy5YtQ1JSEkJDQ5GUlISlS5fmWXfWrFk4ffo0xo8fj/r16+P27dsIDw+HRCJBSEhIMe0BERERlUQNGjREgwYNi2TbWlpaGDzYr0i2TUREZYfaEu0dO3ZAIpFg//79MDMzA/Duyys4OBj+/v6oX79+jvWkUimOHz+O4cOHY+DAgQAANzc3PHv2DEeOHGGiTURERERERGqltqHjkZGRcHNzE5NsAPDy8oKuri4iIyNzrScIAmQyGYyMFCf7MjY2hlwuL7J4iYiIiIiIiApCbYl2VFQU6tWrp1Cmq6sLa2trPHr0KNd6Ojo66NGjB3766SfcunULaWlpuHTpEnbt2oX+/Yv/Vi1ERERERERE71PrNdrGxsZK5cbGxkhOTs6z7qxZszBz5kx89dVXYtmQIUMQGBhY6DjympKdCq9y5eK/rRgRUU5evNCEtjZvrkHqpampye9GIqJySG2Jdm4EQYCGhkae6yxevBi//fYb5s2bh9q1a+PmzZtYsWIFLCws8PXXXxeqvaK8j3Z5xPtoE1FJIZfLIZXykiJSL7lczu9GIqIyKL/7aKst0TY2NoZEIlEqT0lJgY2NTa71/vnnH2zYsAErV67EZ599BgBo3rw5pFIpwsPD0bdvXxga8iw1ERERERERqYfaEm0bGxtERUUplGVmZuLJkyfo0aNHrvUePnwIAGjYUPG2HY0aNUJmZibi4+OZaBMRUZlz8uRx7N69HTExjyGXC6hcuTIcHJwwcmQA/vvvEcaMyf++zuHhqwAg13UbNbLHmjWbVBl2rr77bhb++us+fvppl0q2t3//L9i3bw82b96e4/IffliEX37ZiR49emHChCli+fr1q3H37m0sXbpCJXEQEREBaky0PTw88OOPP+LVq1cwNTUFAJw6dQqZmZlo27ZtrvWqV68OALh37x6qVasmlt+9excaGhoKZURERDkxNdGFtq5esbcrzczAq+TMQtfbtm0LfvwxAl991Rd+fiMBAI8eReHUqWN4+TIBdnYNsGrVRnH9f/75C99/vwAhITNhbV1bLK9Tpw7+/vsvAFBaBgD6+vqF36kSICMjA5s2rUdg4Lgcl//77z84cuQQDAwMlJb16tUHO3ZsxfXrV+Hi0qyIIyUiKn+kUikAQFu7xF21XKTUtrd9+vTBzz//DH9/f/j7+yMxMRFhYWHw8fFRmI08JCQE+/fvx/379wEA9vb2cHR0xMyZM5GYmIhatWrh9u3bWLNmDXr27ImKFSuqa5eIiKiU0NbVw7WFw4u93aaT1wEofKK9e/cOeHt3RlDQBLHMzc0d/foNhFwuh6amJuztHcRlmZkZAIC6dW3QoEGjHLeZ17LS5uzZU8jIyEC7dp8pLRMEAd9/H4a+fQfg6NFDSsuNjU3g4dEOu3dvZ6JNRKRiUqkUEye+m7B6yZLl5SrZVtt0rMbGxti8eTP09fURFBSE0NBQ+Pj4YP78+QrryeVyyGQy8bmWlhZWrVqFDh06YM2aNRg5ciT27t2LYcOG4dtvvy3u3SAiIipyKSkSmJtb5LhMU7P4vsrlcjm2bNmAXr26oX37lujduzt27tyqsM769avx+edt8OhRFAICvsZnn7VC//5f4rffzuS63devX6N9+5Y4eHCf0rLx4wMwduzoPOM6evQQPDza5fgP3JEjB/Hy5Uv07z8o1/rt23fAxYvn8fr16zzbISKiwklKSkR8fBzi4+OQlJSo7nCKlVp/UqhTpw7Wr1+f5zphYWEICwtTKDM3N8ecOXOKMjQiIqISw86uIfbt+wVWVlXRurVHrkl3YchkcnE4XzZNTc08E/cVK5Zh9+7tGDBgCJo0ccGVK5cREbEU6enpGDLkfyMEpFIpZs0KQY8eX2HwYD/88ssOzJwZgh079qFqVeVLvCpVqgQPj3Y4fPgAunb9QiyPi4vDtWtX8M03uX/nZ2S8xZ07t+Dl5aO0TCJJxqpVEZg0aTr09Crkug1HxyaQyWS4ceMq2rfvkOt6REREBcUbjBIREZVwEydOQaVKlbBo0Xx069YJvXp1xQ8/LMKzZ08/epsjRw5Bu3ZuCn8rV4bnuv7r16+xZ89O9OnTHyNG+MPV1Q0BAWPRtesX2Lp1M968eSOum5WVhZEjA9G9e0+4urph+vRZAIDff/8t1+137doD9+/fxaNHD8Wyo0cPwsDAEG3bts+13r///gOpVIp69eorLVu9egXq1bPNsz7wbpRdlSpWuH//Xp7rERERFVT5GSRPRERUStWtWw8//bQLV69exp9/XsbNm9fwyy87cfToYaxYsQb169sVepvffDMbtWvXUSgzMzPPdf379+9CKpXC07OjQnmHDl44eHAf/v33bzg5OQN4d2a8efMW4jomJpVQqZIpXrx4kev2XVyaoUaNmjh8+ADGjJkIQRBw9OgheHl5Q08v94nrEhNfAgAqVTJVKH/w4B6OHTuMDRu25lRNiYlJJXFbREREn4qJNhERUSmgo6ODli1bo2XL1gCAy5f/wOTJ47Bx4zrMn7+o0NurXbtOoSZDS0mRAHh3+db7soexSyTJYpmenh50dXWV4s+epC0nGhoa6Ny5G7Zv/wmjR4/BzZvXEBf3HJ07d88zroyMTHH771u6dBE6dfKFubkFUlJSALy7xjwrS4qUlBTo6+tDS0vrvZh1kZHxNs+2iIiICopDx4mIiEqhFi1awsamPh4//q9Y2jM2NgYAJCUlKZRnnwU2Njb55DZ8fLogLS0NkZG/4fDhA2jQoFGOQ8Jziis1NVWh/MmTaBw8uA/e3u3Fvxcv4nHo0Luy94eoA0BKSgqMjSt98j4QEREBPKNNRERU4iUlJSoN687IeIsXL+JRp07dYomhYUN7aGtr4+zZU7CzayCWnz17ChUrVoStbYM8aheMmZk5Wrf2wM6dW/Hw4T8YM2ZivnWsrWsBAJ49e4patWqL5QsW/ACZTHGyt1mzpqNxYwf06tUH1avXFMvlcjni4+Pg61vrk/eBiIgIYKJNRERU4g0a1AetWrWBq2tLWFhY4OXLBPzyy04kJ79Gr159P2qbjx5FQSqVKZRpa2ujQYOGOa5fqVIlfPllH+zY8TN0dXXh4OCEq1f/xIEDe+HnNxIVK1b8qDg+1KXLF5g4MQgVKlTA55975bt+tWrVYW5ugb//foCWLVuJ5U5OTZTW1dXVReXKlZXulx0d/Qjp6ek51iEiIvoYTLSJiIhKuGHDRuDChd+xfPlSvHqVhAoVKsDWtgEWLFgKd/fWH7XN+fNnK5WZmJjgyJHc73ft7z8GRkZGOHRoP376aSMsLasgIGAs+vQZ8FEx5KR58xbQ1zdAu3aeMDAwLFCd9u074I8/LijcYqww/vjjAqysqqJhw8YfVZ+IiOhDGoIgCOoOQp0SE1Mhl6v+EFSubIR+kws206kqyDJS8PLubgCAhX0vaOkZFVvb2bYt7I+EhJRib5eIKCdxcY9hZZXzUGBTE11o6+Y+k3VRkWZm4FVyZrG3W5rcvn0T/v7DsXLlOjg6NilQnaiohxgypC927NiH6tVrFLrNoUP7wcOjPYYO/brQdfOT1/uQiKise/EiHmPGjAQAhIevhqVlFTVHpDqamhowN8/9B2Ge0SYionLnXbLLhLckefkyAU+fxmL58qWws2tY4CQbAGxs6qFt2/bYuXMrJkyYUqh2r127gri4OHz5ZZ9CRkxERJQ7zjpOREREanfgwF4EBY1EVpYUISEzC11/9OgxsLSsgsIO1EtPf4NvvpkNI6PiHwlGRERlF89oExERkdr5+Y2En9/Ij65fvXoNDBgwpND1Wrdu+9FtEhER5YZntImIiIiIiIhUiIk2ERERERERkQox0SYiIiIiIiJSISbaRERERERERCrERJuIiIiIiIhIhZhoExEREREREakQE20iIiIiIiIiFeJ9tImIiEqRUaOG4e7d2wgL+x6tW3sAAAIDR+Dmzet51mvSxAXLl6/Jc93Fi8Ph5uau8pg/9Pz5M/Tq1RVz54ahffsOn7y9rKws9O3bA0FB49G2rScA4MqVSzhy5BDu37+LZ8+eokePXpgwYUqO9bdt+wn79u1GYmIi6tSpi9Gjg9Csmau4/Pr1q/j22ynYtesADAwMPzleIiIq+5hoExFRuWNsogc9Xd1ibzcjMxOS5IyPrv/s2VPcvXsbAHDq1DEx0Z44cSrS0tLE9b7/Pgx6ehUQEDBOLDMwMBAfOzg4KSzLVrt2nY+OTZ3279+DChUqwMOjvVh26dJFPHz4D5o0cYFEIsm17rZtP2HNmhUYMSIAdnYNcPDgPkyePA6rV29C/fq2AAAXl2awtq6FHTu2ws9vZJHvDxERlX5MtImIqNzR09XFkI1ji73dTUOXAfj4RPvkyWMAgGbNXHH+fCTevEmDvr4B6tSpq7Cevr4B9PX1YW/vkON2jIyMcl1W2giCgN27t+OLL3pBQ0NDLA8IGIegoAkA3p2RzklmZia2bFmPr77qi379BgJ4d+Z/8OA+2LJlA+bODRPX9fXthlWrlmPwYD9oa/PfJyIiyhuv0SYiIiolTp06Dnt7Rwwe7IeMjAycO/drsbb/6NFDTJw4Bp9/3gYdO7bF5MnjERPzRGGd1q2bYdu2Ldi4cS26dfOCt7cnZs6cludZ5YiI79Gjhy/kcrlC+Z07t9C6dTPcunUj17o3b17Hs2dP4empOARdUzP/f3Hu3r2N1NRUfPaZl1impaWF9u074NKlixAEQSz38GiPlBQJ/vjjfL7bJSIiYqJNRERUCvz11wM8fhyNzz/vhCZNXGBpWUU8w11YgiBAKpUq/eUlPj4O/v5fIzHxJUJCZmLKlG/w9GkMAgK+xqtXrxTW3bNnF/777xGmTp2BkSMDcPHieaxcuSzXbXft2gMvXsTjzz8vKZQfOXIQNWtaw8nJOde6V65chqVlFVSpYlWAPVcUHf0fAKB27doK5XXq2CA9/Q0SEl6IZcbGxqhduy6uXLlc6HaIiKj84dgnIiKiUuDkyaPQ0tKCp+fn0NDQQIcOHbFz5zYkJr6EublFobb1xx8X0K6dWw5tREJfXz/HOrt2bYNUmoWlS1fA1NQUANC4sT369PkCe/fuUrh22dTUDHPmhIrPHz+OxrFjhzB16rc5brtWrdpwcnLG4cP7xcnY0tPTcebMKQwZ4pfnvvz1133Uq1c/7x3ORUqKBLq6utDTq6BQbmRkBACQSCSwtKwiltevb4v79+99VFtERFS+MNEmIiIq4WQyGc6cOYnmzVuISe7nn3tj27afcPr0CfTu3b9Q23N0bIIxYyYolVeoUCGHtd+5desmXFyai+0DgJVVVdjbO+L27ZsK6zZv3kLhee3adZCamoo3b97kmsh36dIdYWFz8erVK5iamuLMmZPIzMyAt3fnPPclMfEl7Owa5rlO3jRyKHs3ZPz9a74BwMSkEhITX35CW0REVF5w6DgREVEJd/36FSQmJqJ1aw+kpKQgJSUFVlZVUbOmNU6ePF7o7RkaGqJBg0ZKf3ld15ySIoG5ublSubm5OSSSZIUyIyNjhec6OjoA3k0+lpv27TugYkV9HD9+BABw5MgBtGrlATMz5Tbfl5GRKW6/sIyMjJGZmYGMDMUJ6lJSUv9/uZFCua6urtK6REREOeEZbSIiohLuxIl312IvXhyGxYvDlJY/fhyNWrVqF2kMxsbGSEpKVCpPTEyEsbHJJ29fT08PXl7eOHLkANzdW+POndtYuPCHAsWVmpryUW1m387s8eP/YGvbQCyPjn4EfX0DVK5sqbB+SkqKSvaViIjKPibaREREJVhGxltERv4Gd/fW6NNngMKyN2/SMG1aME6ePIavvx5dpHE4OjbBgQN7kZz8GiYmlQC8myDt7t3bGDhwqEra6Nr1C/zyy04sWjQflStbokWLlvnWsbauhcePoz+qPXt7RxgaGuLMmVNioi2TyXD27Gm0aNFSaej48+fPYG1d66PaIiKi8oWJNhERUQmWfb/snj17w8WlmdJyF5dmOHXqeKES7ZSUFNy9e0epvFq1arkO1f7qq344cuQQxo8PxODBwyCTybFhw2oYG5ugR4+vCr5Deahbtx4aN3bAzZvXMXiwH7S0tPKt4+DghLNnT0MqlSrc3zou7jkePHg3cdnbt2/x9OlT/PrraQBAy5atUaFCBejq6mLQID+sWbMClSqZws6uAQ4d2o9nz2Ixc+Y8pbb+/vs++vYdqJJ9JSKiso2JNhERUQl28uQxWFhUVppgLJuPTxfMnTsDd+7cgoODU4G2eefOLYwapXwWevz4SejZs3eOdapUscKKFWuxYsUyzJs3E4AGnJ1dMH/+YoUJ0j5V27btcf/+Xfj6di3Q+h4e7bB06UJcu3ZF4Qz49etXMX/+bPH55csXcfnyRQDA7t0HUbVqNQBAv37vEudfftmBV6+SULt2XSxc+APq17dVaOf+/btITk5Gu3affdL+ERFR+aAhCIKg7iDUKTExFXK56g9B5cpG6Dd5q8q3mxtZRgpe3t0NALCw7wUtPaN8aqjetoX9kZDwcdfJERGpWlzcY1hZ5TzM19hED3q6usUcEZCRmQlJMifTysu4cf4AgB9+WFngOt98MxkVKlTEN9/Mzn/ljxQevgQPH/6L8PBVhaqX1/uQiKise/EiHmPGvLv9Y3j4aoVbJpZ2mpoaMDc3zHU5z2gTEVG58y7ZZcJbkvz1133cuXMbV6/+iQULlhaq7pAhX2PEiCEYOTJAaQIzVUhJScGRIwcLHRcREZVfTLSJiIhI7YYPHwQDAwP07z8YrVq1KVTdevXqY8KESYiPjy+SRDs+Pg4jRvijSRMXlW+biIjKJibaREREpHbnz1/9pPqdO3dXTSA5qFevPurVq19k2yciorJHU90BEBEREREREZUlPKNNRERERERUDhT3ZKAyWZr42NzcEJUrF/+EzeqaiJSJNhERlVmCIEBDQ0PdYVA5Vc5v7EJEJZCeri6GbBxbbO1JUzPFx8G7Z0PbsPjv+LFp6DKoYwJUDh0nIqIySUtLB1lZnFmc1CcrKwPa2jrqDoOIiNSAiTYREZVJhoYmeP36JdLSUiCTSXl2kYqFIAiQyaRIS0vB69cvYWBgou6QiIhIDTh0nIiIyqSKFQ2gra2D1NTXSEtLhlwuU3dIVE5oampBR0cXpqaW0NEp/mGSRESkfky0iYiozMpOdoiIiIiKE4eOExEREREREakQE20iIiIiIiIiFWKiTURERERERKRCTLSJiIiIiIiIVIiJNhEREREREZEKqTXRjo6Ohp+fH5ydneHm5oa5c+ciPT29QHVTU1Mxf/58eHh4wN7eHp6enli2bFkRR0xERERERESUN7Xd3ksikWDQoEGoVq0ali1bhqSkJISGhiIpKQlLly7Ns+7bt28xaNAgpKWlYfz48ahevTpiY2Px7NmzYoqeiIiIiIiIKGdqS7R37NgBiUSC/fv3w8zMDACgpaWF4OBg+Pv7o379+rnWXbNmDWJiYnDs2DFYWFgAAFxdXYslbiIiIiIiIqK8qG3oeGRkJNzc3MQkGwC8vLygq6uLyMjIPOvu2rUL3t7eYpJNREREREREVFKoLdGOiopCvXr1FMp0dXVhbW2NR48e5VovNjYWCQkJsLKywqRJk+Dk5ARnZ2dMmDABSUlJRR02ERERERERUZ7UlmhLJBIYGxsrlRsbGyM5OTnXei9fvgQArF27FikpKVixYgVmzJiBP/74A+PGjSuqcImIiIiIiIgKRG3XaOdGEARoaGjkulwmkwF4l5CHh4dDV1cXAGBgYICgoCDcunULTk5OBW7P3Nzw0wImBZUrG6k7BCIiIiIiIpE6chS1JdrGxsaQSCRK5SkpKbCxscm1XqVKlQAALi4uYpINAG5ubgCAhw8fFirRTkxMhVwuFHj9giqvCWdCQoq6QyAiIiIiohwwR1EdTU2NPE/aqm3ouI2NDaKiohTKMjMz8eTJE9StWzfXejVr1lRIsD+UkZGhshiJiIiIiIiICkttibaHhwcuXbqEV69eiWWnTp1CZmYm2rZtm2s9XV1dtGrVCteuXUNmZqZYfvHiRQCAvb190QVNRERERERElA+1Jdp9+vSBkZER/P398fvvv2P//v2YO3cufHx8FGYjDwkJQaNGjRTqBgYGIikpCQEBATh37hx2796NWbNmoXXr1nB0dCzuXSEiIiIiIiISqfUa7c2bN2PevHkICgqCnp4efH19MWnSJIX15HK5OAFaNnt7e6xbtw5LlixBYGAgDA0N4e3tjeDg4OLcBSIiIiIiIiIlap11vE6dOli/fn2e64SFhSEsLEyp3M3NDbt37y6q0IiIiIiIiIg+itqGjhMRERERERGVRUy0iYiIiIiIioFUKoVUKlV3GFQMCj10/O3bt7h79y4SExPRvHlzmJmZFUVcREREREREZYZUKsXEiYEAgCVLlkNbW61X8VIRK9QZ7S1btqB169YYOHAgxo0bh7///hsAkJSUhGbNmvGaaSIiIiIiohwkJSUiPj4O8fFxSEpKVHc4VMQKnGjv3bsX8+fPR5s2bfDdd99BEARxmZmZGdzd3XHs2LEiCZKIiIiIiIiotChwor1x40a0a9cOS5cuRfv27ZWW29vb4+HDhyoNjoiIiIiIiKi0KXCiHR0djXbt2uW63NTUFK9evVJFTERERERERESlVoETbUNDQ0gkklyXR0dHc2I0IiIiIiIiKvcKnGi7ublh7969yMzMVFoWFxeH3bt3w8PDQ6XBEREREREREZU2BU60x40bh1evXqFHjx7YunUrNDQ0EBkZicWLF6Nr167Q1tZGQEBAUcZKREREREREVOIVONGuVasWtm3bBktLS6xYsQKCIGDjxo1Yt24dGjVqhG3btsHKyqooYyUiIiIiIiIq8Qp1l3QbGxts2LABEokE0dHREAQBNWvW5LXZRERERERERP+vUIl2NmNjYzg6Oqo6FiIiolJJKpUCALS1P+prlYiIiMqYAv9HsH///gKt1717948MhYiIqPSRSqWYODEQALBkyXIm20RERFTwRHvq1Km5LtPQ0BAfM9EmIqLyJCkpEfHxceJjS8sqao6IiIiI1K3AifaZM2eUyuRyOWJjY7F161bExcVhwYIFKg2OiIiIiIiIqLQpcKJdvXr1HMtr1qyJli1bws/PD9u3b8c333yjsuCIiIiIiIiISpsC394rP56enjhy5IiqNkdERERERERUKqks0X758iXevn2rqs0RERERERERlUoFHjr+7NmzHMslEgkuX76MjRs3wtXVVWWBEREREREREZVGBU60PT09FWYXf58gCGjatClmz56tssCIiIiIiIiISqMCJ9rz589XSrQ1NDRgbGyMWrVqwcbGRuXBEREREREREZU2BU60e/ToUZRxEBEREREREZUJKpsMjYiIiIiIiIjyOKM9bdq0Qm9MQ0MD8+fP/6SAiIiIiIiIiEqzXBPty5cvF3pjuU2WRkREREREROWLlr4OtA11xcflSa6J9tmzZ4szDiIiIiIiIipDNDQ1YNW1vvi4PCnwZGhEREREREREhVHeEuxsnAyNiIiIiIiISIUKdUb7woUL2LBhA+7evYuUlBQIgqC0zoMHD1QWHBEREREREVFpU+Az2mfPnsXXX3+NuLg4eHt7Qy6Xw9fXFz4+PtDT00ODBg0QEBBQlLESERERERERlXgFPqO9atUqNGjQALt27YJEIsGOHTvQs2dPtGzZEjExMfjqq69Qq1atooyViIiIiIiIqMQr8Bntv//+G127doW2tja0tLQAADKZDABQs2ZN9O3bF2vWrCmaKImIiIiIiIhKiQIn2np6etDT0wMA6OvrQ0NDA4mJieJyKysrPHnyRPUREhEREREREZUiBU60a9asiaioKACAjo4ObGxscOrUKXH52bNnYWlpqfoIiYiIiIiIiEqRPK/R/vfff1G//rsbjHt4eGDnzp2YMmUKdHR0MHjwYHz77bfo2LEjAODJkyeYPHly0UdMRESUC2MTPejp6hZrmzJZmvjY3NwQlSsbFWv7GZmZkCRnFGubRERElLc8E+0uXbqgYcOG6Nq1K7766isMGjQI2trvqvTq1QsVK1bEsWPHoKWlBX9/f3Tv3r04YiYiIsqRnq4uhmwcW6xtSlMzxcfBu2dD27B4E/1NQ5cBYKJNRERUkuSZaA8ePBhHjx7FggULsHjxYrRo0QJdu3bF559/DgMDA3Tu3BmdO3curliJiIiIiIiISrw8r9GeNm0aIiMjsWHDBnTt2hW3b9/G1KlT0apVK0ycOBHnzp0TZx4nIiJFUqkUUqlU3WEQERERUTHL9z7aGhoacHd3h7u7O2bPno0zZ87g0KFDOHnyJI4ePYpKlSrB19cXXbp0gZOTU3HETKQgO5HJvqyBqCSQSqWYODEQALBkyXK+P4mIiD7A/+GoLCvUu1pXVxfe3t7w9vaGRCLBsWPHcOjQIWzduhVbt26FtbU1Tpw4UVSxEilhMkMlVVJSIuLj48THlpZV1BwRERFRycH/4aisK/DtvT5kbGyM3r17Izg4GO3bt4cgCLyPNhW77GQmPj4OSUmJ+VcgIiIiIrXj/3BU1n3UT0ePHj3C4cOHcfjwYcTExIjDy7t27arq+KiANHUNoKVnJD4mIiIiIiIi9Shwov3ixQscPXoUBw8exIMHDyAIAho0aIDJkyejc+fOqFy5clHGSfnQ0NCEeeOe4mMiIiIiIiJSjzwT7dTUVJw4cQKHDh3ClStXIJPJULVqVQwfPhzdunVDvXr1iitOKgAm2EREREREROqXZ6Lt7u6OrKwsGBgYoHv37ujatStatGhRXLERERERERERlTp5JtqtWrVC165d8dlnn0FXV7e4YiIiIiIiIiIqtfIca/zjjz/C29u7yJLs6Oho+Pn5wdnZGW5ubpg7dy7S09MLtY27d++iYcOGcHZ2LpIYiYiIiIiIiApDbTesk0gkGDRoEKpVq4Zly5YhKSkJoaGhSEpKwtKlSwu0DblcjlmzZsHMzAxv3rwp4oiJiIiIiIiI8qe2RHvHjh2QSCTYv38/zMzMAABaWloIDg6Gv78/6tevn+82du3ahZSUFPTs2RM//fRTUYdMRERERERElC+1TVMdGRkJNzc3MckGAC8vL+jq6iIyMjLf+tlnvkNCQqCjo1OUoRIREREREREVmNoS7aioKKXbg+nq6sLa2hqPHj3Kt/7ixYvh4uKCtm3bFlWIRERE+dLS14G2oS60DXWhpc8ffomIiEjN12gbGxsrlRsbGyM5OTnPujdu3MCRI0dw+PDhogqPiIioQDQ0NWDVtb74mIiIiEhtiXZuBEGAhkbu/6jIZDLMnj0bQ4YMQc2aNT+5PXNzw0/eBv1P5cpGxdqeTJYmPjY3Nyz29olyw/dm+aLuBJvvLyIqbcrj92R53OeSQh3HWm2JtrGxMSQSiVJ5SkoKbGxscq23a9cuJCQkoF+/fmL9jIwMAO/Okuvq6qJChQoFjiMxMRVyuVDI6PNXXjtOQkJKsbaXmJiq8FhLq3jbJ8oN35vqwc9eIqLSoTx+T5aEfeb3pOpoamrkedJWbYm2jY0NoqKiFMoyMzPx5MkT9OjRI9d6jx49wsuXL+Hh4aG0rHnz5hg0aBCmT5+u8niJiIiIiIiICkJtibaHhwd+/PFHvHr1CqampgCAU6dOITMzM88JzgYMGIAOHToolO3btw9Hjx7F2rVrYWVlVaRxExEREREREeVFbYl2nz598PPPP8Pf3x/+/v5ITExEWFgYfHx8FGYjDwkJwf79+3H//n0AQK1atVCrVi2Fbf3555/Q0tJCixYtinUfiIiIiIiIiD6k1mu0N2/ejHnz5iEoKAh6enrw9fXFpEmTFNaTy+WQyWRqipKIiIiIiIiocNQ663idOnWwfv36PNcJCwtDWFhYnusEBQUhKChIlaERERERERERfRRNdQdAREREREREVJYw0SYiIiIiIiJSISbaRERERERERCrERJuIiIiIiIhIhZhoExEREREREamQWmcdJyIqLkbGFVBBT6fY2pPJ0sTH5uaGqFzZqNjazvY2IwspkrfF3i4RERFRecdEm4jKhQp6Oug3eWuxtSfLSBEfjw3dDy294k+0ty3sjxQw0SYiIiIqbhw6TkRERERERKRCTLSJiIiIiIiIVIiJNlEpJJVKIZVK1R0GERERERHlgIk2USkjlUoxcWIgJk4MZLJNRERERFQCcTI0olImKSkR8fFx4mNLyypqjoiIiIiIiN7HM9pEREREREREKsREm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIV4mRoRERERERU7pia6EJbV6/Y2pPJ0sTH5uaGqFzZqNjapuLHRJuIiIiIqJwzMq6ACno6xdZeSUk6ry0cXmxtJaVniY/vrJ6CpxWL73hnazp5XbG3WV4x0SYiIiIiKucq6Omg3+StxdaeLCNFfDw2dD+09Io/0d62sH+xt0nlB6/RJiIiIiIiIlIhJtpEREREREREKsREm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIVYqJNREREREREpEJMtImIiIiIiIhUiIk2ERERERERkQox0SYiIiIiIiJSISbaRERERERERCrERJuIiIiIiIhIhZhoExEREREREamQtroDICIqizR1DaClZyQ+JiIiIqLyg4k2EVER0NDQhHnjnuJjIiIiIio/mGiTysilWahc2ahY25TJ0sTH5uaGxd6+NDMDr5Izi7VNKj2YYBMRERGVT0y0SWU0tXVwbeHwYm0zKT1LfHxn9RQ8rahTrO03nbwOABNtIiIiIiL6H55uISIiIiIiIlIhJtpEREREREREKsREm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIVYqJNRERE5YJUKoVUKlV3GEREVA4w0SYiIqIyTyqVYuLEQEycGMhkm4iIihzvo01ERERlXlJSIuLj48THlpZV1BwRERGVZTyjTURERFRGcbg8EZF6MNEmIiIiKoM4XJ6ISH04dJyIiIioDCrPw+Wzf1jQ1ua/ukSkHjyjTURERERlBs/kE1FJoNaf+aKjozF37lxcv34denp68PX1RXBwMCpWrJhrndTUVGzcuBGRkZH477//oKOjAwcHB0yYMAENGjQoxuiJiIiIqKQpz2fyiajkUNsZbYlEgkGDBiEtLQ3Lli3D1KlTcfjwYYSEhORZ79mzZ9i5cyfc3d3xww8/YN68eUhLS0Pv3r3x77//FlP0RERERERERDlT2xntHTt2QCKRYP/+/TAzMwMAaGlpITg4GP7+/qhfv36O9WrUqIFTp04pnPV2c3ODp6cnduzYgW+//bZY4iciIiIiIiLKidrOaEdGRsLNzU1MsgHAy8sLurq6iIyMzLWevr6+0tByAwMDWFtbIzExscjiJSIiIiIi1dDUNYCWnhG09IygqWug7nCIVE5tZ7SjoqLQs2dPhTJdXV1YW1vj0aNHhdqWRCLBv//+i7Zt26oyRKJ8ZUqzULmyUbG2KZOliY/NzQ2Lvf2MzExIkjOKtU0iIiIqWzQ0NGHeuKf4mKisUVuiLZFIYGxsrFRubGyM5OTkQm1r0aJFAIAvvvii0HGYmxsWug5RNl1tHQzZOLZY25SmZoqPg3fPhrahbrG2v2noMlSuXLxt0scr7h9iSD34OudP3T9SqkN53Geg/O53acQEm4qLOj4HStzNBQVBgIaGRoHX37NnD3bt2oXvvvsO1atXL3R7iYmpkMuFQtfLDz/UqSxLSEhRdwiFVl77ZGl8rT4FX2fKTWJiqsJjLa2yf8zK4z4D5Xe/P1V5/fyk8qEovic1NTXyPGmrtp+RjI2NIZFIlMpTUlJyPNOdk3PnzmHGjBnw9/fHl19+qeoQiYiIiIiIiApNbYm2jY0NoqKiFMoyMzPx5MkT1K1bN9/6N2/exNixY9GtWzeMHVu8Q3eJiIiIiIiIcqO2RNvDwwOXLl3Cq1evxLJTp04hMzMz30nNHj58iJEjR8LNzQ1z5swp6lCJiIiIiIiICkxtiXafPn1gZGQEf39//P7779i/fz/mzp0LHx8f1KtXT1wvJCQEjRo1Ep8nJibCz88Penp6GDx4MO7evYubN2/i5s2buH//vjp2hYiIiIiIiEiktsnQjI2NsXnzZsybNw9BQUHQ09ODr68vJk2apLCeXC6HTCYTnz98+BBxcXEAgCFDhiisW716dZw9e7bIYyciotxJpVIAgLZ2iZtvk4iIiKhYqPW/oDp16mD9+vV5rhMWFoawsDDxeYsWLfD3338XdWhERPQRpFIpJk4MBAAsWbKcyTblysi4Airo6RRbeyXhlk9vM7KQInlb7O0SEVHx439ARESkMklJiYiPjxMfW1pWUXNEVFJV0NNBv8lbi609Wcb/bu0yNnQ/tPSKP9HetrA/UsBEm4ioPOBd4omIiIiIiIhUiIk2ERERERERkQox0SYiIiIiIiJSISbaRERERERERCrERJuIiIiIiIhIhTjrOBFRGSWXZhX7LYxKwi2UiIiIiNSNiTYRURmlqa2DawuHF2ubSelZ4uM7q6fgacXiu08yADSdvK5Y2yMiIiLKCYeOExEREREREakQz2gTERERFYPivpyjJFzKIc3MwKvkzGJvl4hI3ZhoExERERWD4r6cQ92XcgDZl3Mw0Sai8odDx4mIiIiIiIhUiGe0iYhIZUz0tGFeUVt8TERERFQe8b8gIiJSGS1NDQS3riE+JiIiIiqPmGhTqcazZ0QlDxNsIiIiKu+YmVCpxrNnRERUEJq6BtDSMxIfExERFSUm2lTqMcEmIqL8aGhowrxxT/ExFY/MYr6lGaD+25plZGZCkpxRrG0SUcnDRJuIiIjKBSbYxU9XWwdDNo4t1jalqf+7nVjw7tnQNtQt1vY3DV0GgIk2UXnHRJuIiIiIiKiIcW6h8oWvMBERERERURHj3ELlCxNtIiIiIiKiYsAEu/xgok1Uymjp64jXm2np66g5GiIiIiIi+hATbaJSRkNTA1Zd64uPiYiIiIioZGGiTVQKMcEmIiIiIiq5eJ8LIiIiIiIiIhViok1ERERERESkQky0iYiIiIiIiFSIiTYRERERERGRCjHRJiIiIiIiIlIhJtpEREREREREKsREm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIV0lZ3AERERESkeiZ62jCvqC0+JiKi4sNPXSIiIqIySEtTA8Gta4iPiYio+DDRJiIiIiqjmGATEakHr9EmIiIiIiIiUiEm2kREREREREQqxKHjRERERFRmaOnrQNtQV3xMRKQOTLSJiIiIqMzQ0NSAVdf64mMiInVgok1EREREZQoTbCJSN16jTURERERERKRCTLSJiIiIiIiIVIiJNhEREREREZEKMdEmIiIiIiIiUiEm2kREREREREQqpNZEOzo6Gn5+fnB2doabmxvmzp2L9PT0AtXdv38/OnXqBAcHB/j6+uLo0aNFHC0RERERERFR/tR2ey+JRIJBgwahWrVqWLZsGZKSkhAaGoqkpCQsXbo0z7rHjx/HlClTMGLECLRq1QqnT5/GhAkTYGBggLZt2xbTHhAREREREREpU1uivWPHDkgkEuzfvx9mZmYAAC0tLQQHB8Pf3x/169fPte6yZcvQqVMnTJw4EQDg5uaG//77DxEREUy0iYiIiIiISK3UNnQ8MjISbm5uYpINAF5eXtDV1UVkZGSu9WJiYvDo0SP4+voqlPv6+uLOnTtISkoqspiJiIiIiIiI8qO2RDsqKgr16tVTKNPV1YW1tTUePXqUa73sZTY2Ngrl2dvKqy4RERERERFRUVPrNdrGxsZK5cbGxkhOTs61XvayD+uamJgoLC8oTU2NQq1fGBamBkW27ZJK19hc3SEUOwtDs/xXKmOKst8UJfbJ8oF9svRgnywf2CdLD/bJ8oF9sni2qSEIgqDyVgugcePGGDt2LEaMGKFQ3qdPH1SuXBkRERE51jt48CAmTZqECxcuwMLCQiyPjo6Gl5cXfvzxR3h6ehZp7ERERERERES5UdvQcWNjY0gkEqXylJSUHM90Z8vtzHX2tvKqS0RERERERFTU1JZo29jYICoqSqEsMzMTT548Qd26dXOtl73sw2uxs7eVV10iIiIiIiKioqa2RNvDwwOXLl3Cq1evxLJTp04hMzMzz1t01axZE3Xr1sXRo0cVyg8fPgwHBweFWcyJiIiIiIiIipvaEu0+ffrAyMgI/v7++P3337F//37MnTsXPj4+CrORh4SEoFGjRgp1x4wZg2PHjmHp0qW4fPky5s+fjwsXLiAoKKi4d4OIiIiIiIhIgdpmHTc2NsbmzZsxb948BAUFQU9PD76+vpg0aZLCenK5HDKZTKHM29sbb9++xapVq7B+/XpYW1tjyZIleZ4JJyIiIiIiIioOapt1nIiIiIiIiKgsUtvQcSIiIiIiIqKyiIk2ERERERERkQox0SYiIiIiIiJSISbapUSfPn1gZ2eHs2fPKi2LjY2FnZ0djh8/XiyxeHp6Ys6cOYWut2nTJpw7d64IIlL0YXwRERFwdnZWWOfevXvo3bs3nJycYGdnh9jYWADAokWL0Lp1azRo0ABTp04t8lip9GKfLDj2SSoO7JMFxz5JpV1ERATs7OzEPzc3NwwaNAhXr15Vd2iFsnfvXhw6dEipfOrUqejcubMaIsrb69evERgYCFdXV9jZ2WHv3r3qDqlEU9us41RwMTExuHHjBgDg0KFD8PT0VHNEH2fLli1o165dsc8O36tXL6U258yZA6lUilWrVqFixYqwtLTE77//jnXr1mHatGlwcnKCubl5scZJpQf75KdhnyRVY5/8NOyTVBpVqFABmzdvBgDEx8dj5cqVGDJkCPbs2QM7Ozs1R1cw+/btg76+Prp06aJQ7u/vjzdv3qgpqtytX78ely5dQlhYGCwsLGBtba3ukEo0JtqlwMGDBwEA7u7uOHv2LFJTU2FoaKjmqEoPKysrWFlZKZRFRUVh4MCBaNmypUIZAAwaNAiamp822EMmk0Emk0FXV/eTtkMlE/vkp2GfJFVjn/w07JNUGmlqaqJJkybicwcHB3h6emLnzp2YMWOG0vqCICAzMxN6enrFGGXO3r59iwoVKuS6vKQmsFFRUbCzs0OHDh3UHUqpwKHjpcDhw4fh7OyM0aNH4+3btzh16lSO66Wnp+Obb75Bs2bN4Orqivnz5yMrK0tcnpKSghkzZqBNmzawt7dHmzZtEBQUBKlUKq7zzz//YPjw4XB2doaLiwtGjRqF6OjoPOMbOHAgRo4cqVD24TA9T09PPH36FFu3bhWH+bw/3OTAgQPo1q0bHBwc0KpVK4SGhiIzMzPfY/Pbb7/Bx8cHDg4O+OKLL3DlyhWldd4fEnf58mXY2dkhJSUFK1euhJ2dHQYOHIiBAwciNDQUANCwYUPY2dnh8uXLAIDU1FTMmzdPPG5dunTB6dOnczwGBw8ehLe3NxwdHXH79m0AQGRkJPr06QMnJye4urpi2rRpkEgkYt3smC5evIjg4GA4Ozujbdu2WLZsGeRyuUI7UVFRGDNmDFq0aAEnJyd07twZu3fvFpcLgoBNmzahU6dOsLe3R7t27fDjjz+Cd/FTLfbJ3LFPsk+qA/tk7tgn2SfLi2rVqsHMzEy8zCF7+PX58+fxxRdfwMHBAceOHQMAnD59Wixzd3fHrFmzkJaWJm4r+z137tw5jBkzBs7OzmjVqhVWrVql1O7Vq1fRr18/ODo6wtXVFRMnTkRCQoK4PLuv79u3DzNnzkSLFi3QpUsXDBw4EH/++Sd+++03sc9HREQoxP6+gnz22NnZYd26dVi+fDlat24NV1dXjB8/HsnJyfkev/y2b2dnhzNnzuDq1ativJQ3ntEu4e7evYtHjx5hxowZaN68OapWrYpDhw7hiy++UFp36dKlcHd3x9KlS3H//n2Eh4dDV1cXwcHBAIDQ0FBERkZi4sSJqFGjBhISEnDu3DnxS+r58+fo378/qlWrhtDQUMjlckRERKB///44dOgQzMzMPno/li9fjhEjRsDFxQXDhg0D8L9f67Zs2YKwsDAMHDgQkyZNQkxMDJYuXYr09PQ8r3H7+++/ERAQAHd3d0yePBnx8fGYPHmywpfzhxo3boydO3di8ODB6Ny5M3r16iWe9dixYwd++ukn7Ny5EwBQr149ZGVlYdiwYYiLi0NQUBCqV6+OEydOICgoCD///DOaNm0qbvvevXuIiYlBQEAATE1NUaNGDZw+fRpBQUHo3r07Ro8ejVevXuGHH37A+PHjsX79eoXYZsyYAV9fX6xYsQLnz5/HypUrUbt2bXTr1g0AEB0djd69e8PS0hLTpk1DlSpVEBUVhefPn4vbCAsLw/bt28Vjfe/ePUREREBTU1Ppnzz6OOyT7JPskyUL+yT7JPskAe9+8Hn9+jUsLS3FshcvXmDmzJkYPXo0atSogcqVK+PMmTMIDAyEl5cXxo0bh9jYWCxZsgTR0dHYtGmTwjZnzJgBHx8fhIeH4+LFi1i6dClMTEzQt29fAO8+f4YMGYKmTZti6dKlkEgkWLJkCYYMGYK9e/cqnD1fsmQJ2rRpg8WLF0Mmk6FGjRqYNGkSKlSogClTpgCA0siSbIX57Nm6dSucnJzw3Xff4fnz51iwYAEWLlyI7777LtdjV5Dt79y5EwsWLMDbt28xc+bMQr8+5ZJAJdp3330nNGrUSEhMTBQEQRAWLFggNGzYUHjx4oW4TkxMjGBrayv069dPoe7SpUsFJycn4fXr14IgCIKvr68QGhqaa1vz588XnJycxLYEQRBiY2OFxo0bC+Hh4WJZ+/bthdmzZ4vPBwwYIIwYMUJhW9kxHTt2LNd6giAIKSkpgrOzs7Bw4UKF8sOHDwsNGzYUYmJico13/PjxQvv27YWsrCyx7MSJE4Ktra1CO+Hh4UKTJk0U6jZp0kRhnwRBENatWyfY2toqlO3Zs0do2LCh8NdffymUDxs2TBg8eLD4fMCAAULjxo2F2NhYsUwulwuenp7C2LFjFereuHFDsLW1Fa5cuSIIgiBcunRJsLW1FcLCwhTW69q1qzB69Gjx+YQJEwRXV1chOTk5x+Px5MkToUGDBsLWrVsVylevXi24uLgIaWlpOdajwmGfZJ/Mxj5ZMrBPsk9mY58sP7Lfs1lZWUJWVpYQGxsrBAUFCba2tkJkZKQgCIIwZcoUwdbWVrh27ZpC3e7duwtffvmlIJfLxbJDhw4Jtra2wqVLlwRB+N97btKkSQp1g4ODBQ8PD0EmkwmCIAgBAQGCh4eHkJGRIa5z/fp1wdbWVtizZ48gCP/r60OHDlXaj5w+G7Jj9/X1FZ8X9LPH1tZW6NGjh8K25s2bJzRt2jSnw1jo7Y8YMUIYMGBAntui/+HQ8RJMJpPh6NGjcHd3F3+p6tq1K2QyGY4cOaK0/ueff67wvGPHjkhPT8c///wDAGjUqBH27duHdevW4a+//lIaJnX16lW4ubkp/CpWvXp1ODs7F9ksjjdv3kRaWhp8fHwglUrFv5YtW0Imk+H+/ft51vX09IS29v8GZnz22WfQ0dFRWXwXLlyAra0tbGxsFOJr1aqVOOQtm62tLapXry4+j46ORmxsLHx9fRXq2tvbw9DQEHfu3FGo37p1a4Xn9erVQ1xcnPj80qVL6NSpE4yNjXOM9eLFixAEAZ06dVI6lqmpqfjvv/8+9XCUe+yT7JPskyUL+yT7JPtk+fXmzRs0btwYjRs3hqenJ/744w/x0o9slSpVgouLi/g8LS0NDx48gLe3NzQ0NMTyTp06QVtbW6kf5/SZERcXJ77vrl69ig4dOijMNeDs7Izq1asrbetTJmkszGdPTv0kJSVFYWj8p2yfCo5Dx0uwS5cuISEhAQEBAeIwr2rVqqF27do4ePAghgwZorD+h0PWsmcDzb5O5Ntvv0WlSpWwadMmLFq0CFWqVIGfnx8GDx4MAJBIJGjYsKFSHBYWFkX25ZOUlAQA6NGjR47Lnz17lmvdhIQEpRlPtbS0UKlSJZXG9+DBAzRu3DjH5RKJRPxCt7CwUKoLAIGBgTnW/XDfPvzHQEdHR+H6u9evX6NKlSp5xioIgsLENe97/vx5rvtBBcM+yT7JPlmysE+yT7JPll8VKlTAzz//DA0NDZiamqJq1apKk/R9+J5LSUmBIAhK5dra2qhUqZLStcwffmZkP09ISEC1atUgkUiUtpXd7ofb+pRZ+gvz2ZNTPwGAjIwMGBgYfPL2qeCYaJdg2bOozpo1C7NmzVJaHhUVBRsbG/F59hdWtsTERABA5cqVAQBGRkYICQlBSEgIHj58iG3btmH+/PmoU6cOPDw8YGJigpcvXyq18/LlS5iYmOQap66ursJkMgAKNOkCAHG7ERERqFq1qtLynMqyVa5cWdzHbDKZDK9fvy5Q2wWNz87OLtfrWvT19cXH7/8yCkD8R2bGjBlwdHRUqlvYD1xTU1PEx8fnGauGhga2bduW49mKkjqDZWnCPsk++T72SfVjn2SffB/7ZPmiqakJBweHPNf58D1nZGQEDQ0NpX4hlUrx+vVrpX784WdG9vPszwwTExOlbQHvPhPq1auXZyyF8bGfPSVl++UVE+0SKnvW1Hbt2omTomRLTU1FQEAADh06hHHjxonlp06dUvj1/uTJk6hYsSJsbW2Vtl+vXj2EhIRg+/btiIqKgoeHB5o2bYqdO3fi1atXMDU1BfDu190bN27kOUFI1apVcf78ecjlcvGXxPPnzyutp6Ojg4yMDIUyFxcX6Ovr4/nz5+jYsWO+x+V9Tk5OOHv2LKZOnSoOiztz5ozSPzOfolWrVjh37hwsLS3z/JU8J3Xr1kXVqlXx+PFj9O/f/5NjadmyJU6cOIFJkybleNua7F/ok5KSeNuFIsA+mT/2SeXlAPtkUWGfzB/7pPJygH2yPDMwMEDDhg1x7NgxDB06VCw/efIkpFIpmjVrprD+qVOnFIaPnzx5EpaWluKkZU2bNsXp06cxZcoU8cebW7du4enTp0rbyklOfT4nH/vZU1BFvf3yiol2CXX27FmkpaVhwIABaNGihdLyFi1a4PDhwwr/QMTExGDatGnw8fHB/fv3sW7dOgwaNEj8JapPnz7o0KED6tevDx0dHRw+fBiamppwdXUFAHGGRD8/P4waNQoymQwREREwMTHJ8wvQ29sbu3fvxuzZs+Hl5YU7d+4o3JIkW926dfHHH3/g/PnzMDExQY0aNWBqaoqxY8di8eLFiIuLg5ubG3R0dBAbG4tff/0VM2fOzHUGxpEjR6Jnz54YPXo0BgwYgPj4ePz4448qvXdqt27dsGvXLgwYMADDhg1D3bp1kZKSgn/++QcJCQl5zrqooaGBkJAQjB8/Hunp6WjXrh0MDAzw/PlznD9/HoMGDYKTk1OBYwkKCsJvv/2GPn36YPjw4ahSpQr+++8/JCYmIigoCHXq1MHAgQMxZcoUDB06FM7OzpDJZIiJicHp06exceNGVRyScot9kn3yQ+yT6sU+yT75IfZJKojAwEAEBARgwoQJ6N69O2JiYvD999+jZcuWSp8lly9fxoIFC9CqVStcuHABBw8exIwZM8QfzEaNGoU+ffpgxIgRGDRoEJKTk7FkyRLUq1cPvr6++cZSt25d7Nu3D2fOnIGlpWWuP1h97GdPQRX19ssrJtol1MGDB2FpaYlWrVrluLxHjx6YPHkyrl+/Lt7GYPz48fjzzz8xbtw4aGpqok+fPhg/frxYx8XFBQcPHkRMTAw0NTVRv359rFy5UrweqWrVqvj555+xcOFC8TYDrq6uWLFiRZ63LGnVqhWmTZuGLVu24MCBA3Bzc0NYWBj69OmjsN6ECRMwa9YsjBkzBmlpaQgNDUWPHj0wZMgQWFlZYePGjdi2bRu0tLRQvXp1eHh45DqhCQA0aNAAERERWLRoEQICAmBjY4OwsDBMmzatYAe5AHR1dbFp0yasWLECa9euxYsXL2BiYgJbW1v06tUr3/odO3bEunXrsGrVKgQHB0MQBFStWhXu7u4KE8IUhLW1NXbs2IElS5Zg3rx5yMrKgrW1tXjtIABMnz4ddevWxY4dO7B69WpUqFAB1tbWaN++faH3nRSxT7JPfoh9Ur3YJ9knP8Q+SQXx2WefISIiAitWrIC/vz+MjIzQpUsX8TZ/75szZw527dqF7du3Q19fH2PHjlVIPO3t7bFx40YsWbIEY8eOhZ6eHjw8PDB16lSFW3vl5uuvv8aTJ08wdepUSCQSBAYGIigoSGm9j/3sKaii3n55pSF8OKUmERERERFROXX58mUMGjQIv/zyS77XgRPlhrf3IiIiIiIiIlIhJtpEREREREREKsSh40REREREREQqxDPaRERERERERCrERLucuXfvHnr37g0nJyfY2dkhNja22GO4fPky7OzscOfOnTzX27t3L+zs7JCUlFQkcXh6emLOnDlFsm0qvQraRzIzMzFt2jS0bNkSdnZ2iIiIyPG9HRERgevXryvVz638U/A9TaUV+x2p24evY0REBJydndUYEVHRuXz5MlatWqXuMMo83t6rnJkzZw6kUilWrVqFihUrirc8KU6NGzfGzp07YWNjU+xtE+WnoH1k7969OHDgAEJDQ1GrVi1YWVnB0NBQ6b29fPly6Ovrw8XFRaF+buVE5RH7HZU0vXr1Qtu2bdUdBlGR+PPPP7FhwwaMGjVK3aGUaUy0y5moqCgMHDgQLVu2LFS9rKwsaGpqQktL65NjMDQ0RJMmTT55O0RFoaB9JCoqCpaWlujWrZtCeWl7b8tkMshkMujq6qo7FCrH2O+opLGysoKVlZW6w6AS4O3bt6hQoYK6w6BSiEPHy5DTp0/jiy++gIODA9zd3TFr1iykpaUB+N9w7ZSUFKxcuRJ2dnYYOHBgrtvKHkK1ceNGeHp6wtHRES9evAAAHDhwAN26dYODgwNatWqF0NBQZGZminVTUlIwY8YMtGnTBvb29mjTpg2CgoIglUoVYnl/mF9qaiqmTZsGFxcXtGjRAvPmzVPYZm71AGDq1Kno3Lmz+Pzly5cICQnBZ599BkdHR3z++edYsGAB3r59W+hjOnDgQIwcORLHjh2Dl5cXnJycMGzYMCQkJODFixcYPXo0nJ2d4eXlhVOnTinVj4yMRJ8+feDk5ARXV1dMmzYNEolEXP727VvMnTsXnTp1gpOTE9q3b4+QkBC8fv1aYTvZr8f27dvh6ekJFxcXDB8+HM+fPy/0PpVnquojnp6e2LJlC54/fw47OztxqOuH71E7OzsAwMKFC8X1stfJqRwABEHApk2b0KlTJ9jb26Ndu3b48ccf8eG8lb/99ht8fHzg4OCAL774AleuXCnQMch+Tx88eBDe3t5wdHTE7du3AeT/fs3KysKiRYvQvn172Nvbo1WrVvj666+RmJiocAzPnTuHMWPGwNnZGa1atcpxeNrVq1fRr18/ODo6wtXVFRMnTkRCQoK4PDY2FnZ2djh48CDmzZsHV1dX8TXLyMgQ18vv8wYAXrx4gSlTpsDNzQ0ODg746quvcPXq1XyPlZ2dHdauXYvw8HC0atUKTZs2xezZsyGTyXD9+nV8+eWXaNKkCXr37o2oqCiFugV5HR89eoQJEyagXbt2cHJygre3N1avXq0Qe0GPQ0nGfsd+V5L6HVCw1/HDoeP8vi7ZsvtBTn979+4V17t9+zaGDRsGZ2dnuLi4ICgoCHFxceLy7D6wb98+zJw5Ey1atECXLl0AAK9fv8b06dPRsmVLODo6omfPnvj9998LFN/+/fvRo0cPODo6okWLFhg2bJjC+/eff/7B8OHDxbhGjRqF6OhopbiOHz+usN0P36fZx+HixYsIDg6Gs7Mz2rZti2XLlkEul4t1li9fjjdv3ojHKPtzNz4+HuPHj4e7uzscHBzg6emJGTNmFPBVoA/xjHYZcebMGQQGBsLLywvjxo1DbGwslixZgujoaGzatEkcrj148GB07twZvXr1gqGhYZ7bPHnyJGrUqIGpU6dCV1cXRkZG2LJlC8LCwjBw4EBMmjQJMTExWLp0KdLT08Vrm0JDQxEZGYmJEyeiRo0aSEhIwLlz58QOnpNvvvkG586dw/jx41GrVi3s2bMnx8S1IF6/fg1jY2NMmzYNxsbGePz4MVauXIlnz55h2bJlhd7egwcPkJiYiEmTJiE9PR3z5s1DSEgIUlJS0KFDBwwYMABbt27F+PHjcfbsWXHI4+nTpxEUFITu3btj9OjRePXqFX744QeMHz8e69evB/DuizsrKwtjx46Fubk54uPjsWbNGgwfPhy//PKLQhy//vorHj16hG+++QZpaWkIDQ3FtGnTsGnTpo86TuWNKvvI8uXLsWrVKly/fh3Lly8HAFhaWuLp06cK6+3cuRO9e/fGwIEDxR+D6tWrl2s5AISFhWH79u0YMWIEXFxccO/ePUREREBTUxMjR44EAPz9998ICAiAu7s7Jk+ejPj4eEyePFnhn/O83Lt3DzExMQgICICpqSlq1KhRoPfrmjVrsG3bNgQHB6N+/fp4/fo1/vjjD6UfsWbMmAEfHx+Eh4fj4sWLWLp0KUxMTNC3b18AwN27dzFkyBA0bdoUS5cuhUQiwZIlSzBkyBDs3bsXenp64rZ++OEHeHh44Pvvv8f9+/fxww8/wNLSEv7+/gDy/7yRSCTo27cv9PT0EBISgkqVKmH37t0YNmwYjhw5gpo1a+Z5rLZu3YqmTZsiLCwMDx48wPfffw9NTU1cunQJI0aMgLGxMRYsWIDx48fj4MGDYr2CvI4JCQmoVasWfH19YWhoiH/++QcRERFITk7G5MmTFeLI7ziUVOx3/8N+VzL63ce+jvy+LtmyP0vet2vXLuzbtw+1atUC8C7J7t+/P1q1aoXFixdDKpVi+fLl8PPzw8GDBxVGbS5ZsgRt2rTB4sWLxREoX3/9NZ48eYIJEybAysoK27dvx8iRI7Fhwwa4ubnlGtu6deuwaNEifPHFFwgKCoIgCLhy5QoSEhJgY2OD58+fo3///qhWrRpCQ0Mhl8sRERGB/v3749ChQzAzMyv08ZgxYwZ8fX2xYsUKnD9/HitXrkTt2rXRrVs39OrVC3FxcTh8+DA2b94MAOLnbnaf+Oabb2BhYYHnz5/j2rVrhW6f/p9AZUL37t2FL7/8UpDL5WLZoUOHBFtbW+HSpUtiWZMmTYTw8PB8t9e+fXvB1dVVSE1NFctSUlIEZ2dnYeHChQrrHj58WGjYsKEQExMjCIIg+Pr6CqGhoblu+9KlS4Ktra1w+/ZtQRAE4eHDh4KdnZ2wc+dOcR25XC507dpVsLW1FRITE3Osl23KlCmCr69vru1lZWUJp06dEuzs7ISkpCSFfZw9e3aex2HAgAGCk5OT8PLlS7EsPDxcsLW1FdasWSOWJSYmCra2tsL27dvF+D09PYWxY8cqbO/GjRuCra2tcOXKlVxjvX//vmBrayvcvXtXIdY2bdoIb9++Fcs2btwo2NraCsnJyXnuA72j6j4ye/ZsoX379gplOb1HbW1thXXr1inVz6n8yZMnQoMGDYStW7cqlK9evVpwcXER0tLSBEEQhPHjxwvt27cXsrKyxHVOnDgh2NraFug93bhxYyE2NlYsK+j7dcSIEUJgYGCu287e/0mTJimUBwcHCx4eHoJMJhMEQRACAgIEDw8PISMjQ1zn+vXrgq2trbBnzx5BEAQhJiZGsLW1FYKCghS2NXr0aKFr167i8/w+b8LDwwUXFxfhxYsXYplMJhO8vb2FkJCQXOsJwrvXqEePHgplAwYMEGxtbYVbt26JZceOHRNsbW2FqKgoQRAK/jq+Ty6XC1lZWcKmTZuEpk2biu/Tgh6Hkor97h32u5LT7wr6OoaHhwtNmjTJNU5+X5dsly9fFho3bqzwv9qAAQOEXr16KXwexcbGCo0bNxb2798vCML/+sDQoUMVtnf69GnB1tZW+O2338Sy7Pf0gAEDco1DIpEITk5OwvTp03NdZ/78+YKTk5P4/+77cWV/LmbHdezYMYW6H75Psz8PwsLCFNbr2rWrMHr06FzrZWvSpImwZcuWXGOlwuHQ8TIgLS0NDx48gLe3NzQ0NMTyTp06QVtbu0BDtXLSokULGBgYiM9v3ryJtLQ0+Pj4QCqVin8tW7aETCbD/fv3AQCNGjXCvn37sG7dOvz1119KQ7Y+dPv2bQiCAC8vL7FMQ0MDn3/++UfFLfz/0DFfX184OjqicePGCAgIgCAIePz4caG316BBA5ibm4vPa9euDQBo3bq1WGZmZgYTExNxaFh0dDRiY2Ph6+urcKzs7e1haGioMPz9wIED+OKLL+Ds7IzGjRuje/fu4jbe17x5c4UzDtlnYt4f8kQ5K6o+omoXL16EIAjo1KmTUh9LTU3Ff//9B+BdX/T09IS29v8GJX322WfQ0dEpUDu2traoXr26+Lyg79dGjRrh3LlzCA8Px+3btyGTyXLc/od9t2PHjoiLixPfq1evXkWHDh0Urk91dnZG9erVlV6L9/sZ8O59//57Pr/PmwsXLqBFixYwNTUV90sul6Nly5bi0N28fNh+7dq1YWRkBEdHR4Uy4H99saCvY0ZGBiIiItCxY0c4ODigcePGmD9/PlJSUvDy5ctCHYeSiP1OEftdyeh3n/I68vu6dHj69CnGjh2LTp064euvvwbwbkTCtWvX4OPjA5lMJr4/qlSpgjp16ihdlujp6anw/OrVqzAwMFCYIE9TUxPe3t64ceNGrv3yxo0bSE9PR69evXKN9+rVq3Bzc1M4c129enU4Ozt/9Ofkx35nNGrUCBs2bMDWrVuV3tdUeBw6XgakpKRAEARYWFgolGtra6NSpUpITk7+qO2+n1wCEG+z1aNHjxzXf/bsGQDg22+/RaVKlbBp0yYsWrQIVapUgZ+fHwYPHpxjvYSEBGhra8PExESh/MP9KajNmzcjLCwMfn5+cHNzg4mJCR4+fIhp06Z91DWNH8aV/WVsZGSkVJ59XXn2sQoMDMxxm9nH6tSpU5g8eTK+/PJLjB07FqamppBIJBg+fLhSrLnFUVqu01SnouojqpaUlARBEHKdEOr58+do3LgxEhISlPqnlpYWKlWqVKB2PjwOBX2/jh49GlpaWti3bx9WrFiBSpUqoW/fvggKClIYcvfhMLfs5wkJCahWrRokEkmO/dvCwkLptTA2NlZ4/n4/A/L/vElKSsKNGzfQuHFjpfY+7MM5yan9nMqA//XFgr6OixYtwq5du+Dv7w8HBwcYGRmJQ34/7Nf5HYeSiP1OEfvdO+rudx/7OvL7unRIT09HQEAAatSoge+++04sT05OhkwmQ2hoKEJDQ5XqfXg5w4fvkbz6T1ZWFt68eZPjezv7Gv4qVarkGrNEIkHDhg1z3Hb2D0SF9bHfGUuXLsUPP/yA8PBwzJkzB7Vq1cLYsWPh6+v7UXGUd0y0ywAjIyNoaGiIE6Nkk0qleP36tdIHfkG9fwYC+N8XR0REBKpWraq0fnaZkZERQkJCEBISgocPH2Lbtm2YP38+6tSpAw8PD6V6lStXhlQqRXJyskKsH57Ryf51OCsrS6H8w4lIjh8/Dk9PT0yaNEksK+5JSLK/sGfMmKHwC3y27A/w48ePo0GDBgpfBnfv3i2WGMuTouojqmZiYgINDQ1s27Ytx7Mr1tbWAN71mQ/3RSaTKfWF3HzYtwv6ftXV1UVgYCACAwMRExODAwcOYPny5ahSpYp4HSjwvwTiw+eVK1cW9/PD+IF3fT77zE9B5fd5Y2JigtatW2PcuHFKdVVxF4WcFPR1PH78OHr37q1we5WScpZXFdjvFLHfvaPufvexryO/r0uHkJAQJCYm4pdfflEYVZD9eTRy5Eh06NBBqd6HiWlO/wN/+H8p8K7/6OjoQF9fP8d4TE1NAbybZCy3Wezz2nb252Ru/wOr+gdLS0tLzJ8/H4Ig4N69e1i7di2Cg4NhZ2dX6M8J4qzjZYKBgQEaNmyIY8eOKZSfPHkSUqkUzZo1U0k7Li4u0NfXx/Pnz+Hg4KD0l9MvffXq1UNISAg0NTWVZgfN5ujoCA0NDZw4cUIsEwRBaTK07ET+/e2kpqbi5s2bCuu9fftW6ZYphw4dKtS+fqq6deuiatWqePz4cY7Hqlq1aiUm1vKguPpITnR0dHI8i5FTefaZmKSkpBzfN9lfuE5OTjh79qzCDL9nzpxR+gIuqIK+X99Xs2ZNBAYGwsrKSqlvf9h3T548CUtLS/GfjKZNm+L06dMK8d66dQtPnz79pNcip88bd3d3REVFoW7dukr71ahRo49uKy8FfR0zMjIU+r8gCDh8+HCRxKQO7Hd5Y79TraJ+Hfl9XfKtWrUKZ86cwYoVK5TOIOvr68PZ2RkPHz7M8f2RPWFabpo2bYq0tDRERkaKZXK5HMePH4ezs3OuPyA5OztDX19facK8D7d96dIlvHr1Six7/vw5bty4IfZNCwsL6Ojo4OHDh+I6MpkMFy9ezDPu3GSf4c7t8k4NDQ3Y29sjODgYcrkcjx49+qh2yjue0S4jAgMDERAQgAkTJqB79+6IiYnB999/j5YtW6JFixYqacPIyAhjx47F4sWLERcXBzc3N+jo6CA2Nha//vorZs6cCSsrK/Tp0wcdOnRA/fr1oaOjg8OHD0NTUxOurq45btfGxgZeXl4IDQ1FRkYGatWqhV9++UXhAwd4N+zG2dkZERERMDQ0hI6ODjZs2KB0b0N3d3ds2bIFW7ZsQd26dXH8+HE8ePBAJcegoDQ0NBASEoLx48cjPT0d7dq1g4GBAZ4/f47z589j0KBBcHJygru7O+bMmYOIiAg0bdoUFy9exNmzZ4s11vKiOPpITurWrYvTp0+jWbNmqFixIurUqQNDQ8Mcy+vUqYOBAwdiypQpGDp0KJydnSGTyRATE4PTp09j48aNAICRI0eiZ8+eGD16NAYMGID4+Hj8+OOP+d5JIDcFfb/6+/ujUaNGaNSoEQwMDPD777/j+fPnSkM1L1++jAULFqBVq1a4cOECDh48iBkzZkBT891vu6NGjUKfPn0wYsQIDBo0CMnJyViyZAnq1atX6OFp+X3eDB06FEeOHMGAAQMwaNAgVK9eHa9fv8bdu3fFM4WqVtDX0d3dHTt37kTdunVhYWGBXbt2lZjh1KrCfpc79jvVKurXkd/XJdu1a9ewbNky9O/fH3K5XOEkjLW1NczMzDBlyhQMGjQIY8aMQefOnWFiYoIXL17g8uXLaNeuXY5nurO1a9cOjo6OmDx5MiZMmIAqVapgx44d+O+///K8/ZWhoSHGjBmDsLAwyGQydOzYERoaGrhy5QratGmDFi1aiDP/+/n5YdSoUZDJZIiIiICJiQn69+8P4N3nhZeXF7Zu3Qpra2tYWFhg586dSE9P/6jjZWNjA6lUis2bN8PFxQWGhoaoXLkyhg0bhm7duqFOnTqQyWTYvn07DAwM4OTk9FHtlHdMtMuIzz77DBEREVixYgX8/f1hZGSELl26IDg4WKXtDBkyBFZWVti4cSO2bdsGLS0tVK9eHR4eHuKwGxcXFxw8eBAxMTHQ1NRE/fr1sXLlyhyv1cr23XffYe7cufj++++hra2Nzp07w9/fHzNnzlRYb/HixZgxYwZCQkJgamqK0aNH49q1awrDtwICAvD69WusWLECcrkcnp6emDt3LoYOHarSY5Gfjh07Yt26dVi1ahWCg4MhCAKqVq0Kd3d3cUKcPn36IDY2Fjt27BBvDxEeHi5OsEKqU1x95EMzZszA/Pnz8fXXX+Pt27fYsmULWrRokWv59OnTUbduXezYsQOrV69GhQoVYG1tjfbt24vbbNCgASIiIrBo0SIEBATAxsYGYWFhmDZt2kfHWZD3a9OmTXHs2DFs2rQJMpkMtWrVwoIFC/DZZ58pbGvOnDnYtWsXtm/fDn19fYwdO1b8ZwEA7O3tsXHjRixZsgRjx46Fnp4ePDw8MHXqVIWhfgWR3+dNpUqVsHPnTixbtgzff/89Xr16BVNTUzRu3DjX+zWrQkFexxkzZmDmzJmYP38+dHV10aVLF3h5eSlc9lLasd/ljf1OtYrydeT3dcn2+PFjyOVy/PTTT/jpp58UloWGhqJHjx5o0qQJtm/fjoiICEyfPh1v375FlSpV4Orqmu+waC0tLaxduxYLFy7EkiVL8ObNG9ja2mLVqlX5/mg4dOhQmJqaYuPGjTh48CAMDAxgb28vznlUtWpV/Pzzz1i4cCGmTJkCAHB1dcWKFSsU5l6YPn06srKyEBoaigoVKmDAgAFo2LAhNmzYUOjj1b59e/Tr1w9r165FYmIimjdvjvXr16NBgwbYunUrnj17Bj09PTRu3Bjr16/P8xpzyp2GkN+U0ERERAVw+fJlDBo0CL/88gscHBzUHQ5RucB+R0RUMvEabSIiIiIiIiIVYqJNREREREREpEIcOk5ERERERESkQjyjTURERERERKRCTLSJiIiIiIiIVIiJNhEREREREZEKMdEmIiIilfH09MTUqVPVHQYREZFaMdEmIiIqhfbu3Qs7OzvY2dnh0qVLOa7z5Zdfws7ODp06dSr09n/99VdERER8aphERETlEhNtIiKiUkxPTw+HDh1SKo+OjsadO3egp6f3Udv99ddfsXz58kLXO378OObOnftRbRIREZUVTLSJiIhKsXbt2uHEiRPIzMxUKD948CAsLCxgb29f5DEIgoCMjAwAgK6uLnR0dIq8TSIiopKMiTYREVEp5uvri7S0NPz6668K5YcPH4aPjw+0tLSU6hw5cgS9evWCk5MTmjZtilGjRuHhw4fi8qlTp2Lnzp0AIA5Pt7OzQ2xsrFg2Y8YMHDt2DF26dIGDgwOOHDkCIOdrtDMzM7Fy5Up06tQJ9vb2cHd3x+jRo/Hvv/+q9FgQERGVFNrqDoCIiIg+XpUqVdC8eXMcOnQIXl5eAIBbt27h8ePH6NKlC/766y+F9desWYMlS5agY8eO6N69O9LS0rB9+3b07dsXe/fuRc2aNdG7d288f/4cly5dwsKFC8W6ZmZm4uNr167hxIkTGDBgACwsLFC3bt0c45PL5Rg9ejTOnz8PLy8vDBgwAOnp6bh8+TLu3buH+vXrF8FRISIiUi8m2kRERKVcly5dMHv2bCQnJ8PExAQHDx5E7dq14ejoqLDes2fPsGzZMgQGBiIoKEgs7969O3x8fPDjjz9i/vz5cHZ2Rq1atXDp0iV069YtxzajoqKwb98+NGzYMM/Y9u/fj/Pnz2PcuHEYPXq0WP71119DEIRP2GsiIqKSi0PHiYiISrlOnTpBQ0MDJ06cgFQqFYd0f+jkyZOQSqXw8fFBUlKS+KetrQ0nJ6dcZy/PibOzc75JNgCcOHECxsbG8PPzU1qmoaFR4PaIiIhKE57RJiIiKuWMjIzQrl07HDp0CFWqVEFiYiI6d+6stF50dDQAwMfHJ8ftVKxYscBtWltbF2i9J0+eoHbt2tDV1S3wtomIiEo7JtpERERlQJcuXTBmzBgAgKOjI2rXrq20jlwuBwCsXbsW2trK/wJoahZ8oFtBbxsmCALPXBMRUbnDRJuIiKgMaNeuHYyMjPDnn39i+vTpOa6TfRa6WrVqqFevXp7bU1VyXKtWLdy4cQOZmZk8q01EROUGr9EmIiIqA3R1dTFz5kwEBgbmOGwcALy8vKCtrY3w8HDx7Pb7kpKSxMf6+voAgOTk5E+Ky8vLC8nJydi0aZPSMk6GRkREZRXPaBMREZURuSXY2WrWrIlJkyYhNDQUX331FTp27AgTExM8e/YM586dg6OjI+bMmQMAsLe3BwDMmTMHHh4e0NbWRvv27cUEvKC6deuGgwcPYsmSJbh37x5cXV2RkZGBy5cvw9vbG927d/+ofSUiIirJmGgTERGVI0OGDEHt2rWxfv16rFq1CjKZDFWqVEHTpk3x5Zdfiut16tQJN27cwLFjx3DkyBEIgoAzZ84UOtHW0tLC6tWrsWrVKhw+fBhnzpyBiYkJnJycxGSeiIiorNEQOG6LiIiIiIiISGV4jTYRERERERGRCjHRJiIiIiIiIlIhJtpEREREREREKsREm4iIiIiIiEiFmGgTERERERERqRATbSIiIiIiIiIVYqJNREREREREpEJMtImIiIiIiIhUiIk2ERERERERkQox0SYiIiIiIiJSof8DZmauSet3lLYAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1008x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(font_scale=1.4)\n",
    "fig, ax = plt.subplots(1, 1, figsize=(14,6))\n",
    "\n",
    "sns.barplot(_df, y='Value', x='Metric', hue='Test', ax=ax)\n",
    "plt.ylim([0,1])\n",
    "\n",
    "ax.legend(title='Significant Tests')\n",
    "\n",
    "fig.tight_layout()\n",
    "fig.subplots_adjust(top=0.88)\n",
    "plt.savefig('figures/LUHMES_{}_metric.pdf'.format(gRNA_name), bbox_inches='tight', pad_inches=0, dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Discoveries on all gRNAs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_sig = []\n",
    "for gRNA_name in gRNA_names:    \n",
    "    id_cells = (np.char.startswith(sgRNA, gRNA_name) | np.char.startswith(sgRNA, 'Nontargeting'))\n",
    "\n",
    "    Y = counts[id_cells]\n",
    "    A = np.char.startswith(sgRNA[id_cells], gRNA_name)\n",
    "    W = covariates[id_cells]\n",
    "    W = W / np.sqrt(np.sum(W**2, 0, keepdims=True)) * np.sqrt(W.shape[0])\n",
    "\n",
    "    id_genes = np.any(Y!=0, axis=0) & (np.sum(Y>0, axis=0)>=50)\n",
    "\n",
    "    _df_ATE = pd.read_csv(path_result+'{}_ATE.csv'.format(gRNA_name))\n",
    "    _df_STE = pd.read_csv(path_result+'{}_SATE.csv'.format(gRNA_name))\n",
    "\n",
    "    df_sig.append(\n",
    "        [gRNA_name,\n",
    "        ', '.join(gene_names[id_genes][(_df_ATE['rej']==1)&(_df_STE['rej']==1)]),\n",
    "        ', '.join(gene_names[id_genes][(_df_ATE['rej']==1)&(_df_STE['rej']==0)]),\n",
    "        ', '.join(gene_names[id_genes][(_df_ATE['rej']==0)&(_df_STE['rej']==1)]),\n",
    "    ])\n",
    "\n",
    "df_sig = pd.DataFrame(df_sig, columns=['gRNA', 'Common', 'ATE', 'STE'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{llll}\n",
      "\\toprule\n",
      "gRNA & Common & ATE & STE \\\\\n",
      "\\midrule\n",
      "PTEN & PTH2, PTGDS, NEFM, EEF1A1, C21orf59, MFAP4, ALCAM, NEFL, ITM2C, CRABP2, SLC25A6, EIF3L, WLS, PPP1R1C, CCER2, GNB2L1, SVIP, RGS10, H3F3A, DRAXIN, TCP10L, EIF3K & PCP4, MYL1, RAI14, DNER, MAP7, SNCA, TSC22D1, NRP2, GNG3, SKIDA1 & EIF3E, PRDX1, GNG5, TCF12 \\\\\n",
      "CHD2 & MLLT11, EEF1A1, MARCKSL1, NEFL, HN1, GNG3, ID4, EEF2, STMN2 & PRDX1, TUBB4B & PCP4, CALY \\\\\n",
      "ASH1L & MT-CO1, MT-CYB, FXYD7 & MT-ND1 & PKP4 \\\\\n",
      "ADNP & C21orf59, MAP1B & MEG3, PTGDS, LHX2 & NEFM, RTN1, C1orf54 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(df_sig.to_latex(index=False))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "vitae",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}