2.mutation-classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create a logistic regression model to predict mutation from gene expression data in TCGA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\rvelazquez\\AppData\\Local\\Continuum\\Anaconda3\\envs\\cognoma-machine-learning\\lib\\site-packages\\statsmodels\\compat\\pandas.py:56: FutureWarning: The pandas.core.datetools module is deprecated and will be removed in a future version. Please use the pandas.tseries module instead.\n",
      "  from pandas.core import datetools\n"
     ]
    }
   ],
   "source": [
    "import datetime\n",
    "import os\n",
    "import time\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.linear_model import SGDClassifier\n",
    "from sklearn.metrics import roc_auc_score, roc_curve\n",
    "from sklearn.model_selection import train_test_split, StratifiedKFold\n",
    "from dask_searchcv import GridSearchCV\n",
    "from sklearn.pipeline import Pipeline, FeatureUnion\n",
    "from sklearn.preprocessing import StandardScaler, FunctionTransformer\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import plotnine as gg\n",
    "\n",
    "from utils import get_model_coefficients, get_genes_coefficients, theme_cognoma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Specify model configuration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Genes: ['7157']\n",
      "Diseases: None\n"
     ]
    }
   ],
   "source": [
    "# We're going to be building a classifier with multiple genes filtered by two diseases \n",
    "# Example:\n",
    "# gene_ids = ['7157', '7158', '7159', '7161']\n",
    "\n",
    "# Information passed into the notebook is stored in environment variables\n",
    "gene_ids = os.environ.get('gene_ids')\n",
    "if not gene_ids:\n",
    "    gene_ids = ['7157'] # TP53 is the default\n",
    "else:\n",
    "    gene_ids = gene_ids.split('-')\n",
    "    \n",
    "disease_acronyms = os.environ.get('disease_acronyms')\n",
    "\n",
    "if not disease_acronyms:\n",
    "    disease_acronyms = None # use all of the diseases as default\n",
    "    # disease_acronyms = ['LUAD', 'BLCA']\n",
    "else:\n",
    "    disease_acronyms = disease_acronyms.split('-')\n",
    "    \n",
    "print(\"Genes: \" + str(gene_ids))\n",
    "print(\"Diseases: \" + str(disease_acronyms))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Here is some [documentation](http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDClassifier.html) regarding the classifier and hyperparameters*\n",
    "\n",
    "*Here is some [information](https://ghr.nlm.nih.gov/gene) about the genes*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "path = os.path.join('download', 'expression-matrix.tsv.bz2')\n",
    "expression_df = pd.read_table(path, index_col=0)\n",
    "\n",
    "path = os.path.join('download', 'mutation-matrix.tsv.bz2')\n",
    "mutation_df = pd.read_table(path, index_col=0)\n",
    "\n",
    "path = os.path.join('download', 'covariates.tsv')\n",
    "covariate_df = pd.read_table(path, index_col=0)\n",
    "\n",
    "path = os.path.join('download', 'expression-genes.tsv')\n",
    "expression_genes_df = pd.read_table(path, index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Select acronym_x and n_mutations_log1p covariates only\n",
    "disease_cols = [col for col in covariate_df.columns if col.startswith('acronym_')]\n",
    "\n",
    "# Filter covariate columns by disease if a list was provided\n",
    "if disease_acronyms:\n",
    "    disease_cols = [col for col in disease_cols if col.endswith(tuple(disease_acronyms))]\n",
    "    \n",
    "selected_cols = disease_cols + ['n_mutations_log1p']\n",
    "covariate_df = covariate_df[selected_cols]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Filter the rows by disease type\n",
    "# subsection of columns with row-wise max\n",
    "has_disease = covariate_df[disease_cols].max(axis=1) > 0\n",
    "covariate_df = covariate_df[has_disease]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# filter by sample_id\n",
    "expression_df = expression_df[expression_df.index.isin(covariate_df.index)]\n",
    "\n",
    "# filter by sample_id\n",
    "mutation_df = mutation_df[mutation_df.index.isin(covariate_df.index)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "sample_id\n",
       "TCGA-02-0047-01    0\n",
       "TCGA-02-0055-01    1\n",
       "TCGA-02-2483-01    1\n",
       "TCGA-02-2485-01    1\n",
       "TCGA-02-2486-01    0\n",
       "TCGA-04-1348-01    1\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The series holds Gene Mutation Status for each sample\n",
    "# Take max of mutation status, meaning if any of the genes mutated the value should be 1\n",
    "y = mutation_df[gene_ids].max(axis=1)\n",
    "y.head(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Gene expression matrix shape: (7306, 20468)\n",
      "Covariates matrix shape: (7306, 33)\n"
     ]
    }
   ],
   "source": [
    "print('Gene expression matrix shape: {}'.format(expression_df.shape))\n",
    "print('Covariates matrix shape: {}'.format(covariate_df.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Set aside 10% of the data for testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.645907\n",
       "1    0.354093\n",
       "dtype: float64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Typically, this type of split can only be done \n",
    "# for genes where the number of mutations is large enough\n",
    "X = pd.concat([covariate_df, expression_df], axis='columns')\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y, test_size=0.1, random_state=0)\n",
    "\n",
    "# Here are the percentage of tumors with TP53\n",
    "y.value_counts(True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Feature selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def select_feature_set_columns(X, feature_set):\n",
    "    \"\"\"\n",
    "    Select the feature set for the different models within the pipeline\n",
    "    \"\"\"\n",
    "    n_covariates = len(covariate_df.columns)\n",
    "    if feature_set == 'covariates':\n",
    "        return X[:, :n_covariates]\n",
    "    if feature_set == 'expressions':\n",
    "        return X[:, n_covariates:]\n",
    "    raise ValueError('feature_set not supported: {}'.format(feature_set))\n",
    "\n",
    "# Creates the expression features by standarizing them and running PCA\n",
    "# Because the expressions matrix is so large, we preprocess with PCA\n",
    "# The amount of variance in the data captured by ~100 components is high\n",
    "expression_features = Pipeline([\n",
    "    ('select_features', FunctionTransformer(select_feature_set_columns,\n",
    "        kw_args={'feature_set': 'expressions'})),\n",
    "    ('standardize', StandardScaler()),\n",
    "    ('pca', PCA())\n",
    "])\n",
    "\n",
    "# Creates the covariate features by selecting and standardizing them\n",
    "covariate_features = Pipeline([\n",
    "    ('select_features', FunctionTransformer(select_feature_set_columns,\n",
    "        kw_args={'feature_set': 'covariates'})),\n",
    "    ('standardize', StandardScaler())\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Elastic net classifier and model paraemeters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Parameter Sweep for Hyperparameters\n",
    "regularization_alpha_list = [10 ** x for x in range(-10, 10)]\n",
    "# Chose n_components based on number of positives (or negatives, if that is less)\n",
    "min_class_size = min(y.sum(), len(y) - y.sum())\n",
    "if min_class_size > 500:\n",
    "    n_components_list = [100]\n",
    "elif min_class_size > 250:\n",
    "    n_components_list = [50]\n",
    "else:\n",
    "    n_components_list = [30]\n",
    "\n",
    "param_grids = {\n",
    "    'full': {\n",
    "        'features__expressions__pca__n_components' : n_components_list,\n",
    "        'classify__alpha': regularization_alpha_list\n",
    "    },\n",
    "    'expressions': {\n",
    "        'features__expressions__pca__n_components' : n_components_list,\n",
    "        'classify__alpha': regularization_alpha_list\n",
    "    },\n",
    "    'covariates': {\n",
    "        'classify__alpha': regularization_alpha_list\n",
    "    }\n",
    "}\n",
    "\n",
    "# Classifier: Elastic Net\n",
    "classifier = SGDClassifier(penalty='elasticnet',\n",
    "                           l1_ratio=0,\n",
    "                           loss='log', \n",
    "                           class_weight='balanced',\n",
    "                           random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define pipeline and cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Full model pipelines\n",
    "pipeline_definitions = {\n",
    "    'full': Pipeline([\n",
    "        ('features', FeatureUnion([\n",
    "            ('expressions', expression_features),\n",
    "            ('covariates', covariate_features)\n",
    "        ])),\n",
    "        ('classify', classifier)\n",
    "    ]),\n",
    "    'expressions': Pipeline([\n",
    "        ('features', FeatureUnion([('expressions', expression_features)])),\n",
    "        ('classify', classifier)\n",
    "    ]),\n",
    "    'covariates': Pipeline([\n",
    "        ('features', FeatureUnion([('covariates', covariate_features)])),\n",
    "        ('classify', classifier)\n",
    "    ])\n",
    "}\n",
    "\n",
    "# Construct cross-validated grid searches\n",
    "cv_pipelines = dict()\n",
    "for model, pipeline in pipeline_definitions.items():\n",
    "    cv = StratifiedKFold(n_splits=3, random_state=0)\n",
    "    grid_search = GridSearchCV(\n",
    "        estimator=pipeline,\n",
    "        param_grid=param_grids[model],\n",
    "        cv=cv,\n",
    "        n_jobs=1, \n",
    "        scoring='roc_auc',\n",
    "    )\n",
    "    cv_pipelines[model] = grid_search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting CV for model: expressions\n",
      "\truntime: 0:06:26.144749\n",
      "Fitting CV for model: full\n",
      "\truntime: 0:06:05.821173\n",
      "Fitting CV for model: covariates\n",
      "\truntime: 0:01:04.474832\n"
     ]
    }
   ],
   "source": [
    "# Fit the models\n",
    "for model, pipeline in cv_pipelines.items():\n",
    "    print('Fitting CV for model: {0}'.format(model))\n",
    "    start_time = time.perf_counter()\n",
    "    pipeline.fit(X=X_train, y=y_train)\n",
    "    end_time = time.perf_counter()\n",
    "    elapsed = datetime.timedelta(seconds=end_time - start_time)\n",
    "    print('\\truntime: {}'.format(elapsed))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# expressions\n",
      "{'features__expressions__pca__n_components': 100, 'classify__alpha': 1}\n",
      "cv_auroc = 91.512%\n",
      "# full\n",
      "{'features__expressions__pca__n_components': 100, 'classify__alpha': 1}\n",
      "cv_auroc = 91.607%\n",
      "# covariates\n",
      "{'classify__alpha': 0.01}\n",
      "cv_auroc = 84.097%\n"
     ]
    }
   ],
   "source": [
    "# Best Parameters\n",
    "for model, pipeline in cv_pipelines.items():\n",
    "    print('#', model)\n",
    "    print(pipeline.best_params_)\n",
    "    print('cv_auroc = {:.3%}'.format(pipeline.best_score_))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize hyperparameters performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "cv_results_df = pd.DataFrame()\n",
    "for model, pipeline in cv_pipelines.items():\n",
    "    df = pd.concat([\n",
    "        pd.DataFrame(pipeline.cv_results_),\n",
    "        pd.DataFrame.from_records(pipeline.cv_results_['params'])\n",
    "    ], axis='columns')\n",
    "    df['feature_set'] = model\n",
    "    cv_results_df = cv_results_df.append(df)\n",
    "    \n",
    "cv_results_summary = (cv_results_df\n",
    "    .groupby(['classify__alpha', 'feature_set'])['mean_test_score']\n",
    "    .max()\n",
    "    .reset_index())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAusAAAG9CAYAAACsxjhKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1120bcc0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (29159709)>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(gg.ggplot(cv_results_summary, gg.aes(x='classify__alpha',\n",
    "                                      y='mean_test_score',\n",
    "                                      color='feature_set'))\n",
    " + gg.geom_jitter(size=4, alpha=0.8, height=0, width=0.05)\n",
    " + gg.scale_x_log10()\n",
    " + gg.labs(x='Regularization strength multiplier (log alpha)',\n",
    "           y='CV AUROC')\n",
    " + gg.guides(fill=gg.guide_legend(title=\"Feature Set\"))\n",
    " + gg.aes(ymin=min([0.5, cv_results_summary['mean_test_score'].min()]), ymax=1)\n",
    " + theme_cognoma()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use optimal hyperparameters to output ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred_dict = {\n",
    "    model: {\n",
    "        'train': pipeline.decision_function(X_train),\n",
    "        'test':  pipeline.decision_function(X_test)\n",
    "    } for model, pipeline in cv_pipelines.items()\n",
    "}\n",
    "\n",
    "def get_threshold_metrics(y_true, y_pred):\n",
    "    roc_columns = ['fpr', 'tpr', 'threshold']\n",
    "    roc_items = zip(roc_columns, roc_curve(y_true, y_pred))\n",
    "    roc_df = pd.DataFrame.from_items(roc_items)\n",
    "    auroc = roc_auc_score(y_true, y_pred)\n",
    "    return {'auroc': auroc, 'roc_df': roc_df}\n",
    "\n",
    "metrics_dict = {    \n",
    "    model: {\n",
    "        'train': get_threshold_metrics(y_train, y_pred_dict[model]['train']),\n",
    "        'test':  get_threshold_metrics(y_test, y_pred_dict[model]['test'])\n",
    "    } for model in y_pred_dict.keys()\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvYAAAG9CAYAAAB+ljISAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzs3Xl0XPV9///n3WbVSKNd8r7bGLAxZscYA2UnbFlKSimQ\nhpZfTpOQX39J2p5wIGnTHDhpTpLmmzQhIZCUhibp1yQswaQBG68YYzC28SrbsmRb+4w0+5177+f3\nx0gjC3kRtmXJ8vtxjo91P3eZ95Vt+TWf+dzPR1NKKYQQQgghhBBnNH2kCxBCCCGEEEKcPAn2Qggh\nhBBCjAES7IUQQgghhBgDJNgLIYQQQggxBkiwF0IIIYQQYgyQYC+EEEIIIcQYIMFeCCGEEEKIMUCC\nvRBCCCGEEGOABHshhBBCCCHGAAn2QgghhBBCjAES7IUQQgghhBgDzJEuYCzI5XJs3ryZcDiMz+cb\n6XKEEEIIMUSlpaVUV1ePdBlCnBIS7E+BzZs3841vfGOkyxBCCCHER+T3+/nhD38o4V6MCRLsT4Fk\nMgnAvffey0UXXTTC1QghhBBiKJqamvjOd75DT0+PBHsxJkiwP4Vqa2uZPn36SJchhBBCCCHOQvLw\nrBBCCCGEEGOABHshhBBCCCHGAAn2QgghhBBCjAES7IUQQgghhBgDJNgLIYQQQggxBkiwF0IIIYQQ\nYgw466e7zGQyLF26lB07drBr1y6SySRf/OIXue6660a6NCGEEEIIIYbsrO+x7+np4fnnn6e5uZmp\nU6eOdDlCCCGEEEKckLO+x76iooJnn32W8vJydu3axd///d+PdElCCCGEEEJ8ZGd9j71lWZSXl490\nGUIIIYQQQpyUsz7YCyGEEEIIMRZIsBdCCCGEEGIMOOvH2AshhBB9vEyG5NvrcLrjQzpeeR4qn+9v\n0ED3+Qcek8+jPK93QxW283bhd6WKx5nl5ej+wGG1pHG6u49bgx4MYpZF+19PKfKtLUOq31dXP2A7\n39qCl8kcXn2hRtdDeS54Ho6no2onY5SUoFD0eC7JWIZIah9BPPxuoc+wNVPKIeWnI2KTMxS24ZEq\n0fHsatxMADsdwBfKUj/uLeZ3WdT3FL5vDT01bM9HOVRqF6v43ANLqB43fkj3JMTZTIL9Cejq6iIW\nixW3W1tbAfA8j/zhP+CFEEIck1IKp7UF5bqD9rndcZyOdgCMaDmaaaIUZHLgpTNkcnH8fo+AZ6Ll\nXdx4jKytkcXDBlI5E8bVYgcMlIKeHhPNyeIcamZiVQfRnI+gawKKzY0lbAr4iCUtcukQgXAezWcB\nkLY8QOHzEtSfu4o/e8skmNMAxQvdN3OwzE8m0BfQNTTDGHiPno/qK15kyRaHSS2F0LsxM4/N+hQ6\nov3BXtMN0LTDTvRTMm0fc6w9XLI+SLcdIOaVs96+iJZxDn1neppOXh/433k4lCaycAXXvFyC4RWu\n+VZqMW1lPrLB/tfM7fENOM/yphC9ZBnn73AZ11S4/725mezWpxKvcMEr1J/vtPDQoLdcXUWIXvwe\nUzpTzN1UeHOyPVvPzvB4skmveH1P03G0w2qNW8QmZbA6k8R7X293rpoGvYzuw/5OvLj+A+67tYZT\nzT3C3zshzmQS7E/Aq6++yvPPPz+off/+/SxbtmwEKhJCiOGjuS6ad+QA5LoargsJX5r9VozSoMvM\ntkpCXZ1orksmabDb8NGZL8H05+ioMvC7fpRrEOsOUOZ2oGmKqjlvs3hDCJ9TSIp/aPszegIO6QCg\nge1rQVM6yi2EZqV5GFaGyqte5GOvQSRVqOeNzLXsCVeRLClsq125wfejVVA9+w2uXA+TDxTaPsjc\nQnN5CbYP8CtSTgm4vam1twM751f4wi5Z28XIFtryrobr6Xj92ZXDOuF7GzQcFHbewe4tJ5/3cMyB\n57megYNJX2LX8DA9k3jeT0s2iGNAErAx8DxVDNWe0vFU/5sJDYWHgatrOK6O6g32jjLwlDHgUwLP\n/dCbEOXhahqOKpwL4HgGrlZ4c9T3mgoKwb5YK7g95XheunieqzT8KX3AGwmltMJ5H+J6/a/neRpK\n67tqQVPzIfn/VYghkGB/Am666SYuvfTS4vaGDRt47rnnmDRpElddddUIViaEEMfmOIq2Vpumnk7y\nNrQdcjhk7idj58kfihKMOgCMuyDB/PETmLzPofEPq9kXL2e3XcKhSCHRpjUTNx1CQ0OhQAVBlZNb\nspSb1oFRSGb8pvsO2kp10r092lqn0ZeT0YGENw6lFIGQTSTj4c/3hjkHbM1HVqcQ7F0TpWnFJ8MM\n5YKmk9UsDmlhuvXCjoQWJI+BSyExewzsIdaUwodHVrNo1wMYeqGXOIkfBx2XQlZVmoYTVIdnS/yW\nR84tY1tpLX5/4ZouGlYoTT5YfAXSVpDD+d0cTqaGHaEyDlYFsQ0/relaUvjxSpPF41wPlHZYD77y\nyBPloB4k1TtkJp+3yGUttNJkb20ahqajNGfgH3QA8vZEPpg0E00Vvjfp9ggqnEUL9L/ZMXVv4HnK\nI6cmsqu6mmarAoBUT5iMbeKWdQMaSmloKROrpqvwPfbl0LvKsUM17LWmcuicusIfYdbAUS56RRdG\nLghKw58Loikd3fGh+1x000Hb9zG2hTW2n1P4ZitV+HMqO+x7Meu887jxyvM51fbu3XvKrynESJJg\nfwIqKiqoqKgobjc1NQGg6zqWZY1UWUKIMc5THh4eXl7ng40x3nttN/l8Blsr9KbbhlscntEaSJPo\nLqF2fBv187dze/N0DE9n5Z463okH6PblGBjpdKAHGj1wXeLGGkJv9VCyBfJ2CQG7hy5nPgktXDjc\nMCj0DQNohTSmIN9ZR8bfSiRbCGmTrWbamPShPtrDtjQNDQg5FkZ5BaZTCKG1boauoIle6uBqCqVb\nKMNDc0zoLkUr60ZFYrjxenb6J2DoFt1mKd2+SuxgDsI5VM4PJUkI5zASkcIwEt3D9eeIN13BWrMK\no7pwP17UwNPA0EGzXDRDJzBgWIyG5pWQ3vhJ3rKAvh/1JYPuiPCHsjIapHbeyE6AviH0ZYXffPSP\nx/cdNp2FGekN35kZ5IGOmkJd0YUHmDMtjk+zWBS5nOmBaRzT7cfefbYzPjRsSogznQR74KWXXiKV\nStHVVeh9WL9+PR0dHQDcdttthMPhkSxPCDGGuK5HTzpHOGzQZDezKrGWdNqjo0mjxe4gd6AM1/Ug\nFoXuEggUHiBUugvT91KSKMfXNBFlZ0geGofn93BL0r1XPyykpCPoyuNALkdHV4bSt/agOzpWbjcq\nexX5agelK/KaVegJ72Uqha4gbStaelx2J/1Alia7jiwmGIXXyhp6Ybh13kLzNHzj9hBOatQ2+mhN\nRXgzOpuYr4x0uARb+dB9eXQrDz4b3fGh9Q4B0XSFrmm0b7qLp4zDbmEC5DNBAloAV3cIh3Nomoal\nWVAKUF341TqL9rr+2/YBPkJAqLelpNAFX/KhP4h0LQMOG+BoE8Z9eIzNqadbHtUXtROZnBy0rxDm\nrz5+mBdCnLUk2AMvvPACbW1txe21a9eydu1aAJYsWSLBXoizkOOowjjmI3CVg6cUbd0p7GCculAV\nnY0+1q5K0N6aJ+1lSakE8Q5Fzt+Nmy/0FCsF6C7UdFJa6aB3VgLVuBmLVGsYNzSO3sHF/TL9Dzd6\ndRYZN4ubzUHej+caGMpBMzKFhxIPC/YWDqZy0eIRQlmH0lwKw9XIKT9lWgwv2oFl68SDQVIBEw2F\n2VbNzPxeZmb3Yax2QYXIKBM0jUggxvm+LfgyOWK+CnYGZ5DXfYUkDeg9M8ij06DNouGwoN0XnDX8\nhVhsg1toKFD0DdgezAJwsKJp6u/6gBqrmvur7x3KH58QQpyVJNgDP/3pT0e6BCHEaZBJexw4YBem\nHFRw6GCepv054l0Ot95dRs0E8Ok+TM3kT691s21rmh43iascMl4GHZ2UlxpwzZqbd/CJcxexcVUF\n7W159h9M051P4JTEwVDgAJrb39nrgXIVMasNfdZughvmQ8JCS5u9wzT630zovb3aumGjoQhk8pQG\nWpmXbMdDo9ofoyfq0jFO0RWq4kDpRDwjj5UqY1yiiYnxA9S2+GC1H9dn4Gk6PuAi7x1S3WF6/GXs\ns8aTMHvHhYyDLm82b6uZ5HVr4AwtH1IYjXL4G5++Xu5T26td6MHuodKqZlHk8lN6bSGEGGsk2Ash\nxgSlFPGYSyLhUlpmUFqmo2s6a1cl2LA+ie24tORb8Wm9D0t6KUwMbFUY6tK2by8hFeOhCfdRYZYX\nr9vtdGOrwTOr9Mm1RHgvvo6GnbV4qTCJnB+l93ZBHz6UvO+LSBItWYLpaPib6wgpRbmywUqiUjoR\nO4FuaMys7mZ8OE65P43bHcPt6oSthUuYVVW9851H0Xx+cDQCVdMpXXIdZmkhpOf2N5Jvb8NXPw6z\nvAItEEA7RlAXQghx5pNgL4Q4Y3ieYulvumjanyMY0uj0OjA0g2SPR17lySsHRzmUXb6XyNw2vjzu\nEQ4dtJk02c+WXV1kvQxZ+hffsemfwrHn3XGEpnbR5cTwaT5q6yycfJCeVIJEPoHKZQfMY2gpnUCm\nhEi+ncbtYayuEK0lWQLlnUQWvkPO8Ag4Ov5wggVqAgvC8yjzQlB8WPLCwm/zwD50EKdrD7mmRgAC\nM2YSnDEbmAxAdm8DmR3bANADAfRQiMPfNQTnzKXkokswIqXFNv+kyfgnTT5133whhBCjngR7IcSo\nkEl7vPz7GBUVJsmUy759aXK2y/V3hpg03aRED2Ni0bQ/h6NctnQ1HvVa+e4AyU6T/zB+irHEJNMW\n4NDGaSTNNNgWTDoI0Z7CwZqC8h4Ip0m/X8H3ulYRbOtf4TIfSuDmEoVDXQPdttBsCxsNmzw9mybg\nZIKoSA5HV1RcuJ1AbZq7D0xhcqoUX2QCpVdfW+wtd9Npsju24cS6SH+wGeX0v7kwKyoLv0crCC9Y\niB4sTJsYmDWb8PwFoGnooRCaz4du+dDDYfRQWHrihRBCABLshRAjqLXF5le/7CDn2Ry0D2FpFjmV\nHXDM0gONBK0uPjHuVqb5pqLrGsZhw7jNaBZHOWTdHF5PGGNSC1mjm3ysna61CzDTkcKBoQzB0gya\n7aL11EJPLZpjonmFH4PK9fCyWVwgyeGLMR116pR+GqAr/PUu0Qt8TG2fxuz6BeTb2zA+FLydWCfd\nK14/5uVU3i6GegCrqhqrqvrYNQghhDjrSbAXQnxkexqyrF2VIG8f/UFJ11XEY4VFczK+bjr0FtJd\nJpWXHGJ897mUZGo4eMCmLZYiEW7tPSs/6Dq590rxBZt5aX0X0UQJpaUGMbrILVhNPpAi2D4BvbMc\nvwZ6wEBLVeI1TylOF+4ApmYVHvbsLkzX4iu1D3uFwj14mSzeYcN0jkQz9AHDXQqnK6oy7VTmO6nV\nW6n+VRcVZjkJfXXh+9DVRXjhxWjHmS87dN48Sq9agu73H/M4IYQQ4mgk2AshhsS2PdasTNByMM+u\nnRkcR1FeYQ7Y37w/j2fkcXwp8v4khuPHsfp64AvHdmlt+HPjceIV6Dp4WQtLi2AHCsNdjFn7Mca3\ng6bQTY2yzmnQPIdcR5B4pvBGoTmXxt9yJR+OwAowNAu9t4NcR2dyVTmlRiGMWz6NyxdFmDY9wIcl\n16+jZ80mAEouvAgjGh2wXzMtAtNn9D60WpDvaKf9P5/p79B3QuAb2Luf7+rE6ezAqqktfBfKyone\ncDNmeQV6KIRREjlu6BdCCCGGQoK9EOKoXFfx3sYUy16O09XloDwARcqK44xvou3SbVSnpzBt/zVk\nMh6W5WBbOfLBQkh3jSwahRlrvN6JynOrz6PrnENk/ArP1DAnmXiT9mBP2Ul050Ks5GTY0f/QZ9/0\n5l10gb+wiJzy5TGV6l+wCDA0nTqrdkgh/sOUUmQb+5eWD8yYhW/ceJTnkW85RL6tFbtpP9k9u6m4\n7c7+8xxn0LV0nw89EMQ/bRpWTR1mtLwY6gGMkhJCc887bk1CCCHERyXBXgiB4yiSSZdotPAjYc/u\nLOvXJWk5ZGN7eQ4lEmT8CTyjf6iMDvj8GjPmeHz6hnKcjMGvftnB7o5uclWHUJqCnA+9vgOvpg1D\ngyqjkujELJoOmnbgQ1VEYXrDR6r7VCxY5CZ6iP/xVewDzf1tPT3EP9hCrmEXbqZ/eM6Hn1HVDAOz\nvByzvBKjJEzJZYvQg0F5mFUIIcSIkGAvxFmouSnHpo1pWlvy9PQUep3rzk2RvGg115ddy/7GMkIh\nnXDYoKGtiZRVmMddK8lg1ccwZzfhGjZ2JMGGVDu5vnne74KSfDs+z0YBQT2Ajg70PUCaI3rYHPEn\nw6dZJ71gkZvoof25Z/GyWZRSKNtGZTPEXvk96Ppxz7eqa6i5/7MnVYMQQghxqkiwF+Is0LQ/x//8\nd+cR93nKY2+ukcZdWUpDJplLs1x97UTiMZcXX4jhuQq9tgtz3m78EY8JE3yARdxNU6rXoGs6cSde\nvJ6lWVhGYXhM1Owfp94XxKcHpg3rvX4UqU3v4mULzwBY0XJyzfsxSiIDQn3o3PPxT5qMVVM7+MFZ\nIYQQYhSRYC/EGGTbHokel8oqC6VUf6hXhRVXO5xOXDVwfLjT4yc4oZtX4q8xp34W5RUmi66O0KgU\ndiBIpm0qNeUZSg0bXSus6uopD13Ti+Pa+4x0iHeTCTLbtuLZ9qB9ynHI7NyO09aKWVkFgG5ZVN17\nP21P/wdetn+V2apP34evtu601S2EEEKcDAn2QowR2azH+rVJNm5IAnDdDWUEojYJleLv/r8akt3w\n/R810O60F8+pvnEnvqoUut8tjh+/uewmnm1/Dlvl2fPaVLKdASCKVq6h1TXR4w583Sqz8qTHuZ8M\nJx7HS6cGtHX97n/wcoWArlwXNx4rbLsOKn/YlJq6jh4MEZg1G93no+Tiy9B0A6Msin/KVLQhDMcR\nQgghRgsJ9kKcwVoO2jz/XEd/gwJb5bGVzU9+vwd+D+M+9T7/z/S/wNAjfO2rs4h7Nfypezn7cgNX\nbp0bnEN7voPfvL+W+DvjUXkDJ2mAKgTh6oX9D7v2DbE5FePcT5SXydC9/H/J7Nh+xP1Odxw3HgPX\nPeJ+ADeVwiiJEL5gIQAlCy8ZllqFEEKI00GCvRCjXDLhkkq5xGMutXUW0fL+f7bPP9dBzIkTc+KY\nmkFeDVzgqfyy/ZgRm5SXZnxpIYyXahHmBuewuPRKqsxK9uUaWZVYyweZ7WT2l9H+x5nF803NAg0C\nlVnqp2n4tOqRHWKT6CH28u9w4nGUk0c5Lsrz8DIZNMtC9/mKx3o93aBp+MZP7L+ABmZlFVZNLWZF\nJYEZszDLouiB40+JKYQQQox2EuyFGIXWr0uyZmVPf4OCqlqDKbc00d7Wzh3VN6JrOlooSyaeQeGR\nL0wyT9mFByiZ3YER6g/5FYfNRGNqJueGzilur0qspS1fGJ4Tf2c8AD7NT4VZzvjK0v754GuuHs5b\nPialFKmNG+hZubyw7XmovI3T1YnK5zEipQSmTCEwY1bxnKRponJZNMtHYOZswgsW4qurH6E7EEII\nIYafBHshRolc1uO3z3fS3pbHwaEt34GjHPKq8ABow3546z/AX6tx82fzBDQ/V1xczf52k82RFYRn\ndKJrBp5ymRaYSplRxvmhudRaNQNepyG7h1WJtdi9vfs9buENRLapHC1WTp1VTtgIcfvdFUNa3OlE\nKKWwm5vwspnjHwzYB5pIvfcuSimc9lZUNguGgR4IYNbWgW4QmDGLyKVXFM8JzJiFpusYkVI0U37U\nCSGEGPvkfzshRokDzTZdehsNudiAdscFx8xCZRdOSTeJiQf5+6UHKTlQGDLj+jKka3KUvHchGoUn\nYPtG3f+RHcCOAdezPRsYPNOLnixlor/QY19dYw1bqAdIrl1FYv26IR/v2TnceBwvmQDDwDduPJrV\nP+xGtyz8UwYOD7J6Z7wRQgghzhYS7IUYAb97sZW3N7eT8lJMvaKHTyw6j2kz6pkQc2ltLiPrZUk5\nObjhTTJ5G4VCKdAcE+vgeLR900i7faubhjDa5zC0vm8Af/ErUyvMN6+jDxiuc/miyCm5zyOxWw5+\npFAPoPIOXjKBUVGBGS0HNPRAgNLF1+CrH4dZXjE8xQohhBBnEAn2QpwGsS6H5X/q5t3drcScrgH7\nmptydNlxxvnquXzCHM7/yzRLzWeI74wS/2A2kUO1xWPNSGEKR1+JAagTrsfQdOqs2kHzzxfH0w9D\nb72b6CG+7GVyzc0D2qPX3zRg28vncbs6MWtqe1erBfvQAZJvv4URiWBWVlFy4cUEZ5+DEEIIIfpJ\nsBdiGPX0uDz941biTjedTieOC3lboQ4L5ZmGMC+8sYcNLZWFc9weYvmLySZMdNcBTeGryDD+4m4q\np+RG3eqtx6M8j1zjXmK/X4pS/fdtRsspXXItgSnTUEqR2baV1LsbcDraUQqq//IBrKpqAIKz5hC5\nfBF6KFwM+0IIIYQYSIK9EMMob3vUXdxJw8puHLfwgGwfvbKbCVVlGB11xDYl6WQvAE5xysrC71Z5\nmnmfbB/RRaBOlJfLkdq0kcSaVQPay5ZcR2jeBWi6jptO0/PGH8ns2jngGPvQwWKw1ywLw7JOW91C\nCCHEmUiCvRAnqSG7h1e2vE/zmkpyMT8OOVJJKJ+cxfIC5LpNMu0h8jUHYe4BmHiI6NvXEPBHySTA\ntg7CYZn18H+UAb/JuIvTI7YI1MmwDzTTufQ3KMcZ0G5WVBCav4Dszu2k3t1AvrWFwzryMaPlBGfN\nkakphRBCiI9Igr0QJ0EpxQ+/3UnPwSnYkRjQG2J1SGrdGBkPQ4XxTT2EN3kvoW3z0BrPpfL6XQQn\ndQ+63odXdB3NQ2482ya5bjW5fXtRyhu8P5kcFOrLb76NwPSZJN9aQ2LdmkHn+CdOouLuT8lwGyGE\nEOIESLAX4gRs7Wzg6R91Yac1EiqBivQHUU0v/G43VxJYsIdwyyyglEjzBegVHtUXtRCZrAHR4jmj\nLch7tk1m6/u4icRRj8nu3oXTM/jNyYf5xo3DqqnDP2UagSlTUY5Desv7A/YHps3EP3UaZkWlhHoh\nhBDiBEmwF2KIGrJ7WNm9lryWZ++qCN0t08iH49A7d7yGhi+ao+b8bsLlik/Ou5wZkWtGtughctNp\nnM6O4nZy/VpyTfuHdK5uWeih0BH3GWVRjJII9sFmShcXvheaaVJz/2dJvvM2Km9TetWSk65fCCGE\nEBLsxVnmw6uufliisYT2DdV4eX1Aez5lktgfxl8/Gas0h5PwQzCDng3iBTJos/cSmnuIq6su4c9n\nfOp03Mop48S6aP/lz1He4OE0xxOYMZOya2/AOCzYK6XIt7WS2vg2mR3bi+1eJo0RLgEKD8NGLrti\n0PWEEEIIceIk2IuzRkN2D/+36/cD2jL7y4i/Mx6VNwAKgR1QCryciZPw42Ys0DzcUIrcjE342uuJ\naOMpCZgY5x7AmbudSyvm8xczPnPa7+lUyDU3HTPUV997P+j6oHbd78co6V/IystkSG58m+Tbbx3x\nOsq2IXzy9QohhBDiyCTYi7PGqsTaAdtRM0rXu1NRsf7FmLyEhd27rZke6C5eNIYyCw+B6pvnUnF+\nnBllNYS1Ei64cgqzp9+CqY2Nf0qhc87FrKwqbGgagWnTj7uqq3IcUu++Q3LDW3i53BGPqXngIcxo\n9Ij7hBBCCHFqjI00IsQQHD785rL4x9j/din1aQfPr9B1DZ9PY1dLlmAQyio02owmAPxogIUecKi5\neQe+igz31d1EUA+O0J0Mn8D0mQRmzPzI56U3vzcg1BvhMIHpMwnOPQ+rtk4eiBVCCCFOAwn2Ykw7\nfEx9j9sDgHFgAu8sD9K3AJTjKKqrTe77TDWZjMc3v7eFNrcwG0xwcpzQ1C5CU2JohuKC8DyujizC\np/tG6pZGjPI8Mts/IN/WilkWJbxgIVB4GDZyxVXE//gq4QsupOTiy9ADgeNcTQghhBCnmgR7MWYd\nbUx99/9OZqIf0mmPloN5zKDD+CtT6HoNuulRM93myj/LsCffQNbLArAgfAFXl16JpZ2dq596uRzx\nP/6B7O5dAITnzUc5DppZ+BESmH0ONRMnFR+OFUIIIcTpJ8FejFlHGlPf/d50KsxyWlpsYm4MvT6F\nMqDMGEdLvIeqshCP/PnF7Ms1sr1rO7eX38rs4EcfmnImOXyaS8038JMIL5ul+/U/ktm5fUC7yudJ\nb9lE+ILeXntNk1AvhBBCjDAJ9mLM6htTn9lfRtXWq7C8UiIdNjt7GlGWiz+iYxYmw2F15zts7j7E\n5yKfIWKUMMU/mb+v//wIVn96KM8ju2MbUJjlxjduPABOPEZ82SvYhw4OOic4cxaheQuwautOa61C\nCCGEODYJ9mJM6RtT37nPz4G3x+PZEyFZQqDEY2PrDjylQAdd14qhvurPdhOaHOeB6r8kYpxdvc52\ncxNuJgNAcPYcNNPEy2WJvfgC+cN68gHCF1yIb8JEAtNnysOwQgghxCgkwV6MKasSa2nLt3No/Vzs\nziAo8Bs6LeM3ooUV+p4JoIHlKwTTC/66kb+qv/esHTt/+DAcq64eAN0fwKysIt/ZgR4MolsWFZ+4\nB7O0bKTKFEIIIcQQSLAXY0ZDdg9t+XbSjWUkttTh4WFdsZlZqcupTF/HxDkGv2n5AICFf9nGA1Pv\nBhaNbNEjSDkOyQ3rcTrbUQqMyMDgHpgxk/Kbbis+ICuEEEKI0U3+xxZjxqrEWuLvjKPtD7NxrSxa\nXSdq60zOua6c66+vpD2e5nPV8zlnTuT4FxvjPNum+0+vkd6yCZTCN2EivgkTi/ujN98mw22EEEKI\nM4wEe3HG8zzFj59t4IPmybjKwy7tRMeAnjC+AMRnbQBupDoaovosX/w039FOZsc20ps24sTjoBQA\nxoeG2UioF0IIIc48EuzFGe/Zn7azs62DlJdGOTqaMtE0Dd3nMulv1nFF1adGusQR56bTdPzqF7iJ\nxBH3B889X8K8EEIIcYaTYC/OWA3ZPazofItt1dB9qBLl6Hi2gR608V39Hv4w/L/1f4+hGSNd6oiL\n/+HFQqjcYV6wAAAgAElEQVRXHvnWFrxcDiMSwRo3Ac3vxyw7yz/KEEIIIcYACfbijLNyeQ95W9G0\nYC2dWjvMTqO/X0Ned9BuWUVADzBx6w3cc/3cszbU2wcPkHpvI8p1UErhxLrItx5COQ7oBr4Jk9CM\n/u+NHgyOYLVCCCGEOBUk2IszRi7r8aN/bwEgOD5BU6yZshI/NaEyKq6JsTezn/r3b6U8O57b765g\n2vTACFd8einHIbtnN14uR2LlcjzbLu5zEwm8VAqrrh49FEb3+YoLTFk1tQSmTh+psoUQQghxikiw\nF6OeUoqnfthGOu2SctO05FugAZQ2gdLFbZSHwjy0+BM881Qb8axDNGqOyVCvlMLrXUzqCDvp/t9X\nye7dU9j0PAA0XQdADwSwJk5Ct3zoPh/RW24nMGXqaalbCCGEEKeHBHsx6v3k/7SSyXi02K2kvBQA\ntuOSnbiN3W/XU75/Os8E2+jpcUe40uHj5XJ0/OoXhZlsjkMpRb7lIGZlJfVf+odiuC8yDHlQVggh\nhBiDJNiLUS+T9mjI7SluVyzey35rF8rW8W2bi+MGiOec4v6+VWXHguT6dWT37MJuaRnS8Zquo2ka\nZrQc3fKT3b2D0Nzzh7lKIYQQQowGEuzFqNbaYtOQ21vcnvTXG0ilPUpevAqVDKBpGpOryik1Cn+V\nLZ/G5YvGxgJUTnecnjUrB7VrhkFw9pzBJ2g62YZdeNkseigMgB4MDXeZQgghhBglJNiLUW37ziRQ\nWEQpML4HgGlr76E71YRDnpJKly88PGMEKxw+XjZb/FozDTTLh1VdQ/ltd6L7fIOOt1tbSG/dXNyO\nXHaFPBQrhBBCnEUk2ItRZ9vWNM37ba6/OcpFV/rJzXX4w+97yE7dS/XLf0H8sLH01Re1j2Clw0cp\nRcevflncDi+4iNIrFx/1+MzuncRe+l1x26yoIHLZlcNaoxBCCCFGFwn2YlRZ+ptOGvflqKv3kct5\nhP1hbqhaQot5kK4NF3HA7aHLieEoB6siTWRycqRLPqVS724g9e47OD09A9rN8vJjnPMO3Ste7z+2\nooLqex8YrhKFEEIIMUpJsBejRstBm8Z9OZRSrN67nZ0/DfLo5y5E0zS8nAk4xLwuvJIEluUSXXgA\nnzZ4SMqZxunqJL15E55tDxhK08csLSU4e+4Rz1VK0bN6RXHbqq6h4o67Byw+JYQQQoizgwR7MSo4\njuL55zoA2G83AzDxzt18+9BK7q64AwiTctO4oQTjPrUZXdOpMitZFLl8BKs+eZ5t0/aLp4+4z6qu\nwSgro+zqa48a1DVNI7zgIpJvv4VZWkrVp+8bPL2lEEIIIc4KEuzFiGs5aBdDvac8HJWn/uNbaMyk\niXU5LP+DRVush5Z8C2bvulNVZiX3V987glWfHDeVJPXeRtLvvzdon2YYRG+4meDsc4Z0rdC587Cq\nqglMnymhXgghhDiLSbAXI64v1APsze0jenEzti9N6yGb2ev+igTQYccA0KzCg7Nnck99tmEXsVd+\nj3K9Ae2+8RMov/k2NJ//iLPe9FGuC73z1QOY0ShmNDqsNQshhBBi9JNgL0ZUPO4QDhukki4NuT2U\nnNNG6bwWmpsdavf0h3ezNIsyMkQXHuDuituZHpg2glWfnPS2rcVQr/ssguech69+HP5pM44Z6KEw\npr7zf/4bs6yM0PwF+OrGnY6ShRBCCHEGkGAvRlQ0anLOuUFeX3uI4JQYFVfsB8DzFNG2mQDcfncF\nf4q8RtyJEzWjZ3SoB8Dr76mv+czD6IHAcU9xEz0k33mbXONenFgM++ABcs1N1PzVX6NZ1nBWK4QQ\nQogzhAR7cdp5nmLHtgxz5gbRNI35C8Kcu3ASP+35A6m0R6zLYcbqv0QBeiTNytI36HF6jnvdM5F2\nnB56ADeVov2XP8ez7f7zNIhef7OEeiGEEEIUSbAXp933/+0QrvL45cYtLLlwAjecex62ZxBoqaBl\nG0zZfh0KjZSbptPdQ32+fxEqn3Z2BVk3maT1pz8a0GZEIkT/7Eb8kyaPUFVCCCGEGI1kCg1xWj39\n4zYA9uX2kdhXgkMepRQ+3UfF6puZuP3P0Cg8FJqJHiK68EDx3Bqr+ox+aBYg39lBdk/DkI5VShF/\n9cUBbdEbb6Hmwb/BP3nqcJQnhBDiLLdv3z40TeOBBx74SOdpmsaSJUuGpSYxdNJjL06bV16M0dPj\nkHGzxbat5SuY1DCJTas9OjscUm6aLifG+BsbKZ9wEE8VxqOf6Q/M9km/t7H4taZphTE1R6FpGrnm\n5uJ29PqbCJ1z7rDWJ4QQYuybMmUKUAjxQ7VkyRJWrFiBUmp4ihKnhAR7cVrYtsfO7Rkcz+Fg/iCO\ncjE/tYxGO8dzf9xCtrPwAKmj8lgVadzxzdD7s6PGqh4ToR7Ay+WKX5dcdkVxykoAL5Mhs2s74XkL\nim3hefNJvb8J/8RJBOeed1prFUIIcfYZP34827Zto6ys7COdt23bNkKh0DBVJYZKgr04LX70vVZQ\n0Gjvx1Eu7l8sxTM8JpkT6Mg5OCoPGlgV6eLwm6gZxadZZ/zwm6MJX7Cw+LV98ACd//e/0UxrQLAP\nnTcf/+Sp+KfNGPAmQAghhBgOlmUxZ86cj3zeiZwjTj0ZYy+G3bvvpFAoGnJ7ANAWv41meIT1MG5z\nNSRLMDWLUKninE8eZPJ0H3dX3M5DNQ9wf/W9Y6a3/miU4xB/9SWUU1h86/DZb6ya2sKKshLqhRBi\nzFi+fDmapvH444+zatUqlixZQiQSIRqN8vGPf5zdu3cPOP6NN97gM5/5DLNnz6akpISSkhIuuugi\nfvKTnxzx+n3j3Q8cOMBf/dVfUVdXh67rPPPMM2iaRmNjI42NjWiaVvz1+OOPA0ceY69pGitWrCh+\n3ffrw8ccaYx9R0cHjzzyCFOnTsXv91NTU8OnPvUptmzZMujYBx54AE3T2Lt3L9///veZM2cOfr+f\nyZMn8/Wvfx3P8wadIwaSHnsx7FxXUX9emoYNhW19yiFM/NRbtYS3XkWpv/DRXXWZxb01V41gpSMj\nsWYlTk9hOk8zGkXZNgxhGkwhhBBntnXr1vGtb32Lm266ic9//vNs3bqVpUuXsnLlStatW8e0aYWO\nrSeeeILdu3dz2WWXcddddxGPx3n11Vf527/9W3bs2MG//du/Dbp2Z2cnl19+ORUVFdxzzz1ks1nm\nzZvHY489xne/+10AHnnkkeLxx3rw9bHHHuOZZ56hsbGRxx57rNh+wQUXHPP+2tvbufzyy2loaGDJ\nkiXcc8897N27l9/+9re8/PLLLFu2jEWLFg0678tf/jIrVqzgtttu48Ybb+SFF17g8ccfx7ZtvvnN\nbx7zNc92EuzFsCuf18aBpjfwHarHXbKKrHIY13oBXe9Ow58O0zeY/vJFkZEt9DRQ+f7eeDSNXONe\nkhsL73h0y6L81jswSkpGqDohhBCn07Jly/iP//gP/vZv/7bY9uMf/5iHH36YL37xi7z4YmFmtB/9\n6EdMnTpwNjTHcbjlllv43ve+xxe/+EUmTZo0YP+WLVt48MEHeeqppzAMo9h+4YUX8swzzwAUe+mP\n5/HHH2f58uU0NjYO+RyAr371qzQ0NPCP//iP/Ou//mux/ZVXXuHWW2/lwQcfZMeOHej6wAEkGzdu\n5P3336e+vh6ARx99lJkzZ/Lv//7vPPbYY/ik8+uoZCiOGHarEmtxwylqbtyJaehM8U0i/s547M4Q\nnlcI9dU1FtOmH38F1jOZl8uS298IgFlWhmZZJNatKe4vueQyjEjpSJUnhBDiNJs1axYPPfTQgLaH\nHnqImTNn8vLLL9PeXljH5cOhHsA0TR5++GFc1+WNN94YtN/n8/Hkk08OCPWnk23b/OpXv6KyspKv\nfe1rA/bdcsstXH/99ezevZvVq1cPOvfRRx8thnqAqqoq7rjjDhKJBDt27Bj22s9kEuzFsFFK8fPn\n9tGdKPRSm7rJ7OAsfAcno8XKqTDL0XWN6hrrrOitz+zcgXIL4+iD55yH3dyEfehgcX944SUjVZoQ\nQogRcOWVVw7qrdZ1nSuvvBKlFJs2bQIgkUjw2GOPMX/+fEpKSopj3D/+8Y8DcPDgwUHXnjp1KlVV\nVcN/E0exfft2stksl1xyyRFny7nmmmsAeO+99wbtW7hw4aC2CRMmABCPx09xpWOLDMURw+YPL8XZ\nuKeZ9L5xmON8XHRrnodqHuC5P7Sj+/MAVFaZ3Ht/9QhXenrkW1uKX2t+P53/89/F7dA556Lp8j5b\nCCHOJrW1tcds7+7uxrZtlixZwsaNG1mwYAH33XcflZWVmKbJvn37ePbZZ8kdNpXy8a59uvT0Pjt2\ntDr6euT7jjtcaengT69NsxBZ3d4OMnFkEuzFsFBKsfz9vYUNx8DRnOK+vN2/uMVY76lXSpHdvRMv\nmcDpaCu2W9U1aBooVVijquy6G0awSiGEECOhtbX1mO1lZWX87ne/Y+PGjfz1X/81P/3pTwcc9/zz\nz/Pss88e8RojPZtaXzg/2j22tLQMOE6cGhLsxbB4+60EGS9d3K66djeJxvk894d2enoK77ajUXPM\nj6tPvrVmwDj6Pma0nOCcc/FyOcquuwHNlH+KQghxtlm9ejWe5w0YjuN5HmvWrEHTNObPn8/PfvYz\nAO64445B569cufKEXtcwDOzDplYe6jlQ6DEfyrj9OXPmEAgEePvtt0mn04OG4yxfvhw4/sw64qOR\nz/7FKRfrcnjljQPFbb22i0qjgvYN1bS35YsPzFq+sT03e65pP8m3CqHes3Oo3vl3rYpK9HCYshtu\npuL2uzDC4ZEsUwghxAjZuXMnTz311IC2p556ip07d3LrrbdSXV3N5MmTAVi1atWA41asWDHo3KGq\nqKigo6ODbDb7kc4BaGpqGtLxPp+PT3/603R0dPCtb31rwL5XX32VZcuWMWPGDK688sqhFy6Oa9R2\nE+bzeZ577jmWL19OMplkypQp3HvvvSxYsOCY5/3TP/3TERc9gMK7zaVLlx732AULFvD1r3/95G7g\nLNXRnuc/f95Op9NZbKu7fi+a5sfLF95H6rpGZZU5pofheLlsYdEpBV4uh7KzaJqGb/xkym+/e8Q/\nIhVCCDHybrzxRr7whS/wyiuvcO6557J161ZefPFFqqqq+N73vgfAxz72MaZMmcKTTz7Jli1bOO+8\n89ixYwcvvfQSd911F7/97W8/8utee+21bNiwgZtvvpmrrroKn8/H4sWLWbx48THP+e1vf8vHP/5x\nbr75ZgKBAPPnz+djH/vYUc954oknWLFiBf/yL//CmjVruPTSS9m3bx+/+c1vCIVC/PznPx/08LA4\nOaM22H/3u99lzZo13H777dTX1/P666/zjW98g29+85vMnTv3qOd98pOf5Prrrx/Qlsvl+OEPf3jE\nNwVVVVXcd999A9r63pWKj0YpxX8+015cYRag9LaNlAQsMvvLsHt84IfSUmNMPzCrlCL2yos4ySRu\nTzfk85hVVYCGVVuHERw8O4AQQoizz2WXXcbXvvY1vva1r/H9738fwzC48847efLJJ4uLU5WUlPD6\n66/z5S9/mTfffJPly5dz7rnn8txzz1FbW3tCwf7RRx8lFovx0ksvsXLlSlzX5bHHHjtmsH/ooYfY\nt28fzz//PE888QSO43D//fcfM9hXV1fz1ltv8c///M/87ne/Y+XKlZSVlXHnnXfy2GOPcd55533k\n2sWxjcpgv3PnTlauXMmDDz7IXXfdBRTeKf7d3/0dzzzzDE8++eRRzz1SeO+b3/Xqq68etC8UChWn\nXBInZ3d6Dz0XbMTu6sHZPhltXBvZFof0ivk4CT8+rfCufKwPwUm9+w65xn2obBa3swOrphbQCEyd\nRunia0e6PCGEEKPIokWLiuPNj2bq1KlHDfBKqSG1Ha6kpISf/OQnR9w3ZcqUI55vmiZPPPEETzzx\nxJDrAIqfPvR9AnEszzzzTHHxrA97/PHHP9LiWGerUfn5x+rVq9F1nRtvvLHY5vP5uP7669m+fXtx\nwYahevPNNwkEAlx66aVH3O+6LplM5qRqFrAmvQ7/3CZqp9tYN67Gnrkd/4aLcRJ+ACrMcmBsz4Rj\nt7aQWLUC5XmFWXAMAz0YwjduPOW33C5TWgohhBBi2IzKHvs9e/Ywfvz4QU9Qz5o1C4C9e/dSXT20\noRzd3d289957LFq0iEBg8AwsBw8e5JOf/CSO4xCNRrnhhhu45557ivOliqGzVR5d0ykdZ5P60zwC\njfVomoaOzuSqciqDZVy+KDJmZ8Lxcjnir/wez3Vx2lrQgyHMykqsqmoqP3GPhHohhBBCDKtRmV5j\nsRjl5eWD2vvaurq6hnytvrFjS5YsGbSvrq6O888/n8mTJ5PL5Vi9ejW//vWvOXjwIF/5yldOuP6z\n0cu/i7EvW0V2fA+zZ5RSlb2SuL8wd/3td1eM2TDfRylF959ew+nuxu3sAMCsqkL3+YnedJuEeiGE\nEEIMu1EZ7G3bxrKsQe0+nw/giCusHc2bb75JWVnZEedJ/cIXvjBg+5prruEHP/gBr732Grfffjtz\n5sz5iJWfnZJJl107MhxKBvE2zmPnBMUk++yZqx4gs/V9Mju3o1wXL5PGN24Cus9P5Sc/jVVdM9Ll\nCSGEGEWWLFly3HHwQpyIURnsfT4f+Xx+UHvfYgp+v39I12lpaWH79u3ceuutQ1pMAeDOO+/ktdde\nY9OmTUcN9l1dXcRiseJ236pqnucdse6x7qn/08bO3G7cvAaeRufsdYx/tzBsyjDVmP6eqLyNZ9vE\n/vgqUJjm0j/7HIxIBP+M2RAtH9P3L4QQZzLXdU/r66XT6Y80d/zJCoVCRxyGLMauURnsy8vLjzjc\npi9MD3U6yhUrVgBHng3naPrG7ieTyaMe8+qrr/L8888Pat+/fz/Lli0b8muNBcqDps5q0r4cyjVx\nL1uPLxaiJ3EQw1RUaj0sW3b6foidNsqjcvtWAl0dA5ozldV0jSssJkJLG5xlfx+EEEIcXUdHx0ee\nAORkTJw4UYL9WWZUBvtp06axefPmQUsQ79ixAyhM+zQUb775JnV1dR9pSE1LSwsApaWlRz3mpptu\nGjDDzoYNG3juueeYNGkSV1111ZBfayxobrL5/bvrIGuiaRpGZYILd9/Bw48N7c/oTJV6ay0pU0dV\nV+MlelCuixEtJ7LoakILLhrp8oQQQgzB3r17R7oEIU6pURnsr7jiCpYuXcqyZcuK89jn83n+9Kc/\nMWvWrGKveldXF+l0mrq6ukGz2DQ0NNDU1MSf//mfH/E10uk0lmUNGMuvlOLXv/41ABdeeOFR66uo\nqBjwqUHf8sq6rh/x2YCxzM45eChQGmp8C1O23MqtN08a89+H7JZN4LjkmxrRS0oACE6aQsn5F2CM\n8XsXQoixYqjDdE+1hQsXDvtrvPPOO8P+GmL0GZXBfvbs2Vx55ZX84he/IB6PF1eebWtr4/Of/3zx\nuF/84he8/vrrPPXUU9TW1g64xvGG4TQ0NPDtb3+bxYsXU19fTy6XY926dWzbto0bb7yR6dOnD98N\njiEN6UYUHqCjVXcxLXndmH5Y1unpxkulcNNp7EMH0AJBrJo6AOyWg3Cax2sKIYQQQvQZlcEe4Etf\n+hLPPfccy5cvJ5lMMmXKFB599NEhLT/seR4rV65k+vTpTJgw4YjHVFdXM3fuXNauXUs8HkfTNCZO\nnMjnPve5AQtjiWN789BGlFMDSsOwgyNdzrDxbJv4qy+R3dMAgBvrAsdBj0TQTIOShZcQXngJeu/M\nTUIIIYQQp9tJB/t8Pk9PTw+lpaWndPiFz+fjwQcf5MEHHzzqMY888giPPPLIoHZd1/n5z39+zOvX\n1dXx1a9+9aTrPNsl4xqeU5ijvSI2DatcG+GKTj0vk6brd/+D3fv8hZtO4XbH0Xw+fDV1VD/wWYxw\nyQhXKYQQQoiz3QkH+4aGBp5++mm2bduG53l8/etfZ/78+cTjcb797W/ziU984ohzx4uxZc6lLh80\ne2TTGnXZWVy+KDLSJZ0yXiZDatNGUu9txOudnky5DiqbwSivwIyUUnr1NRLqhRBCCDEqnNBymHv2\n7OEf/uEfaGlp4ZprrhmwLxqNYts2r7/++ikpUIxenueRjfkx8gHKfCVUVfrHzPh6pRRdL71AYt2a\nYqjXQ0FQYFZUYZZXoPl9BM+dN8KVCiGEEB/NCy+8wA9+8INhufZ3v/tdXnrppWG5tji+Ewr2//Vf\n/0VFRQU/+MEPuP/++wetnjZv3jx27dp1SgoUo1M8neTJH31A85ryYpvlGzvDcJJrV2MfaC5uW7W1\nVP35XxKcWVh4S/f5qP2bv0PTxs49CyGEODtIsB+7TmgoztatW/nEJz5BMBg84qqa1dXVR1xgSowd\n33pyL625DvRz9uO36/HZVWNmGE7uQBOJ9WuL25ppUHXPfWiaRvnH7iK9eROh8+ahjdA0aUIIIYQQ\nR3JCPfb5fJ5wOHzU/el0+oQLEqNfZ0eeTjuO8sDdW4dyDabXl5/Rw3A826Z7xet0vbiUzt8UVhUu\nfBKlKL/l9mLPvKbrhOcvkFAvhBDijPTAAw/w7LPPsmPHDjRNQ9M0HnjgAQDWr1/PDTfcQCQSoaSk\nhDvuuGPQIl7PPvss8+bNIxQKEY1Gufjii4s99FOmTKGxsZEf//jHxWs/88wzp/kOz24n1GNfV1fH\n7t27j7p/8+bNTJw48YSLEqPb0z87hKJ3+FU6SLRG46YlR55WdLRw02lyjXvB8464P9uwqziVJYBy\nXZyOdqzKSpJvv4V/6nQZdiOEEOKM9+ijj9Le3s7WrVt5/vlCR1Z1dTXr169n8eLFXHvttfznf/5n\ncWKU6667ju3bt+Pz+XjzzTd54IEH+NKXvsR3v/tdcrkcmzZtKo7SWLp0KbfccguXX345X/nKVwBk\nXaDT7ISC/eLFi/n1r3/NokWLmDZtGkAx9CxdupSNGzfy2c9+9tRVKUaNXTsz7Ijtx1W9AfnaNUxf\nMJ5pNUtGtK5jUUrR8cuncTOZoR2fz2P3rShrWuTb23BjXZgVlcNcqRBCCDG8pk+fTnV1NYFAgMsu\nu6zY/tnPfpbzzz+fl156CV0vDOi47LLLmDZtGk8//TQPP/ww69evp7y8nO985zvF826++ebi1wsW\nLMDv91NTUzPg2uL0OaFgf9ddd7Fp0yYef/xxxo8fj6Zp/OxnP6O7u5t4PM78+fO55ZZbTnWtYhR4\n8YUuHOUUe+zLDs7Bd2F2hKs6tviyV4YU6jVdp/yOjxP/w4v4p06H3jerFR+7S0K9EEKIk/Zf//Vf\nvPLKK5SVlQ37a3V3d/PQQw9x3333HffYTCbDypUr+da3voXneXi9n25XV1dz/vnns379eh5++GEu\nuugiYrEY9913H/feey+LFi2ipESmfB5NTijYW5bFN77xDV566SVWrFiBZVkcOHCAcePGcccdd3D7\n7bcX3+2JsaXxQBIVKIR6bfYeZp5fyaLI5SNc1UC5A00k161B2TYAdmtLcZ8RiVC6+JojnmdWVJLZ\ntrUwvWVvqK/61F/gGzd++IsWQgghRkhXVxeu6/KVr3ylOITmcKWlpfz/7N13eFRV/vjx953JTCaT\nXikBQkCqQKhKlSQIgSBFRPgKIiDCqrti+bkoIiu64rKuiIDu6lpAAXFFlKICIgkCUoI0pUMooZgQ\n0pNJmXJ/f0QmjklIMpkU4PN6Hp/nnnPvPeczk0Q+c+cUgMjISD777DMWL17MPffcg1arJTY2loUL\nF9KsWbPaDluUwekNqrRaLSNGjGDEiBGujEfUY2lXzZg8roIKKCpNon/lieZT6jqsUnJ27nBYqvL3\ngh54CK3RWOY5S1YmuT8llFwrSb0QQohbgJ+fHxqNhr/+9a+MHj261Hlv75JV78aOHcvYsWPJyspi\n06ZNPPPMMzz00ENs3bq1FiMW5XEqsV+4cCGDBw+mTZs2ZZ4/efIkGzZs4Mknn6xWcKJ+ycuzYdNY\nwQp4FNLQEFzXIZViKyrCnPKrvazx8ABA62HEZ8DAcpN6gNy9exzKktQLIYRwpXHjxjFw4EC6detW\n433t27ev3IVM9Ho9BQUlw2g9PT3p3bs3R44cYd68eZVq39fXlzFjxrB7926WL19ebtuidjmV2MfF\nxdG5c+dyE/uUlBTi4+Mlsb/JNGqso9XtWhILLlDUPBF3Tf372i13z05UixUAj9Zt8Y8dVqn7bPn5\n2PJy7eWQh6fVSHxCCCFEXWvXrh0ffvghK1asoE2bNgQFBTF//nwiIyO59957efDBBwkKCuLXX39l\n69atREdHM2bMGF566SWuXLlCVFSUfYXE5cuXM2jQIIe24+Pj+e677wgICCA8PJzAQJmnVltqZCB8\nQUEBWlnn+6aj02kw5jRAnx6M37EedR1OKYWXLpC7b6+97NO/7LH0ZdF4eODTfwCKouAbOQA3n5qf\n2CSEEELUhSlTpjBmzBiefPJJevTowZw5c7jjjjvYtWsXqqoydepUYmJimDVrFoWFhXTs2BGAO+64\ng9OnT/PEE08wcOBAXnnlFSZMmMC7775rb/u1114jLCyM0aNH06NHD9avX19XL/OWVOkn9qmpqaSk\npNjLFy9e5PDhw6Wuy83NZcOGDTRq1Mg1EYp6Q1VVks0paAqMaHSF6BVdXYfkIO93Sb2bjw9az+vP\n1FdVFdPB/RgjuqBoNLj5+eEXOwzDba1rOlQhhBCizvj4+LBy5cpS9REREaxZs6bc+4YOHcrQoUOv\n2/btt9/Otm3bqh2jcE6lE/vvv/+ezz77zL6T2KpVq1i1alWp61RVRVEUGYZzk/l80yn25P1EoakB\nqq04oa9vq+GoZrP92LuclW9+z3T4Z7J+iMMY0QXVZkPRaPBoVfbwMiGEEEKI+q7SiX3Pnj0JCQkB\nYNGiRcTExJQaY68oCgaDgVatWhEcXP8mVgrnxe+5TIE5GFuRgtvtJ/HNaktLQ4u6DstOVVX7saLR\n4JiWjwYAACAASURBVFHBU/f8E8fI2vIdAKafD6ALboA+tH7vniuEEEIIcT2VTuzDw8MJDw8H4MqV\nK/Tu3ZuwsLAaC0zUH+lpFopUMzZL8ZQMd5snTYwN6jiqEoWXLpD57XqseXmVuj5n53ZyEnbby1lb\n42j4uHzDJIQQQogbm1Or4jzwwAOujkPUY+u+TC8+UAGfXFql9mXwqKA6jen38o8edkjqNeWtU5+R\nTub3mxzWuNcFBuEXE4tGr6/xOIUQQgghapLTG1QBZGRkcPr0aXJzcx2GQlwTHR1dneZFPZCbY+XX\ntJJlIOl8FP/Ld9KipaHugvqNqqoUnj+H6UjJJG73Jk3x6tGz1LXmtKukLlviUOcWEEDwhMk1HqcQ\nQgghRG1wKrG32Wy8++67bN68ucyE/hpJ7G98K5ddJcuaZS/rTf51GA2oNpt9vXnT0cPk7PrR4bz/\nPSPRGBw/dJivppK6fKlDnaF5OH5DKrfGvRBCCCHEjcCpxP6rr75i06ZNREZG0rlzZ9566y0mTpyI\nh4cH69atw9PTkwkTJrg6VlEH8vKs5JoLsBZqUTsfxv3X28CjbmKx5uVy9dNPyh1LrwsMQnF3L31f\ndhZe3Xpgzcsj//hRgidMRhdYf4YSCSGEEEK4gtM7z3bt2pWnn36a7OxsAFq2bElERARRUVFMnz6d\nxMREIiIiXBqsqF0Z6RZQoaDIgmoDGqbidroHOl+lTuIpTDpfblLvP+QeDC1uQ1FKx2ZocRv6Js3I\n2bWDBtP+jLacMfhCCCFETTIajfj7+3PmzJka78vf3x/3Mh52iZubU4l9SkoKgwcPBkCjKV4pxWq1\nAmAwGBgwYACbN29m1KhRLgpT1IWsLAvGAAtqpgohaaCAR2Ahvfp612ochReSKEw6h+Vqqr3OvWkz\n3Pz8QQH3Fq0wNA93uEdVVXJ37cCrV18URUGj1+PbX4aGCSGEqDtBQUEEBck3xqLmOJXY6/V6tFot\nUJzIK4pCVlbJOGx/f39SU1PLu13cIJqHG5g8pRHp72RzsvEh9MlhtLjvLC1C+tdaDNacbNK//B9/\nnMph7NS53M2kbEVFpC77CGtODhqjJx7tbkcjTy2EEEIIcZNzKrEPDg4mOTm5uAE3Nxo1asT+/fuJ\niire7fPQoUP4+fm5LkpR63JzrShASooVa64B45HO6LyLgAu1GoclO6tUUq9xd0cf2rTM61WbjeR/\nL7SXs7ZuQd+kKRp32TBNCCFE3crIyCA3N7fiC13Ez88Pb+/a/ZZd1C2nEvtOnTqxe/duHn74YQAi\nIyP59NNPSU9PR1VVjh49ysiRI10aqKg9RUU2PvhPCsPvDWDHzgzSrGkoKCg6K3pFV2dxeXbugrF9\nR9z8A1B0ZcdRmHTeoex39yB0QZLUCyGEqHs5OTm1OqLB3d1dEvtbjFOJ/b333kuXLl0wm83odDpG\njx5NVlYWW7duRaPRMGjQIMaNG+fqWEUt2fp98YTotLxcDmt+wqYWJ8Z+3S7R17v2huEAFJw6aT/W\nenqhC7n+jrdZWzbZjz3atsfYQSZwCyGEqF+6detW433s27evxvsQ9Y9TiX1AQAABAQH2slarZdq0\naUybNs1lgYm6k5FhAeCjr45T1PYyljt/xiOtEY1aaGhpaFFrcVhzc8k7uL+kQtFc9/qiSxex5uQU\nX6oo+N0dU5PhCSGEEELUK9fPlMqQn5/PrFmz+O6772oiHlEP/Hq5CKtqw1ykYm34K6quEPc8/1of\nhmMzOY5DNLS8rfxri4rI3LzRXvaJHojiVq2NlYUQQgghbihVTuw9PDw4ffp0TcQi6gGLpXim6lXz\nVTAUoirF5cDbM+nr3avO4vLqfgdu/gHlnjf9chBLZgYA+oaNMHboVFuhCSGEEELUC1VO7AHCw8O5\nePGiq2MR9cDZxAIAcm25KO5mADwvt6BRC6VWh+EAmH752X6saLTXvdb8u8lIvgMHl7lRlRBCCCHE\nzcypxP6BBx7gu+++4+eff674YnFDyc21UWgrBEDjVzwUxv1q41qPw2oykffLoZIK7fV/Vf0HDyV4\nwmT8Bg1BFyibfwghhBDi1uPUIOStW7cSHBzM3/72N5o3b05oaCh6vd7hGkVRmD59ukuCFLXHYlbJ\ntxU/tdeGXgVAsV3/aXlNUAsLHMoebdqVfZ3NhvLb7se6wCBJ6oUQQogKrFmzhosXL/KXv/zFZW1O\nmjSJ3bt3c/z4cZe1KarOqcQ+Li7Ofnz27FnOnj1b6hpJ7G9MF5IKybOaALBqivA42rHWY1BVlbQv\nPrOXPbt0w83P3+GavJ8PkBX3PQ0eeRStl6zRK4QQQlTWmjVr2L17t0sT+9mzZ9fq5luibE4l9mvX\nrnV1HKKesFlBZ1ApyIcitzw804PAs3Y3psr5cRvWvDx72c3H1+G8arORFfc9ilbLlQ/fQ3FzI2TK\no2gMhlqLUQghhLjZFRYW4u7uXqlrW7ZsWcPRiMpwaoy9uHmNfiCQv/ypBbF3N8YnpRWKrfizX02v\niGPJSCdn5w6y4jaT+1OCvV5RKLXCTd6BnwBQrVZUVcUvZqgk9UIIIUQlTJo0iY8//pgTJ06gKAqK\nojBp0iQmTZpE27Zt+f777+nWrRvu7u4sWbKE/Px8pk+fTrt27TAajTRp0oRx48Zx+fLlUu22bdvW\nXl66dCmKorB//35GjBiBp6cn4eHhLFiwoLZf8i1FFvoWpTT2DqAgxMh3W1KxYUav0dfYijiqqmK+\nksLVlcvKPB/80BQUXcm3BQVnTpO9/Qd7Oej/xqNvWPuTe4UQQogb0ezZs0lNTeXIkSN89lnxsNfg\n4GD+/ve/k5KSwrRp05g1axYtWrQgJCSE/Px88vPzeemll2jUqBEpKSnMnz+fu+66i2PHjqHTXf8b\n/fHjxzN58mSeeOIJPv/8c5555hluv/12Bg0aVBsv95Yjib2wy8m2En/oLIP6NGHXjhx7vUZnq7E+\ns3+Ic9xd9jeKRoPfkHsc1q4vTDpP+rqv7GX3Jk3QNWhUY7EJIYQQrvTpp5/y7bff4uvrW+G169ev\ndygPGzas0v388d7fa9myJcHBwRgMBnr27OlwLjMzk7Vr13LXXXc51L///vv2Y6vVSu/evWnatCmb\nN28mNjb2urH86U9/4qmnngIgKiqKb775hlWrVkliX0NkKI6wW7o8ic1bk7mQl0xBkdleH9w99Tp3\nVU9R0nmHsr5hI4InTKbB1MfwaNXGXl94IYm0Lz93uDZgxGhZr14IIYRwEV9f31JJPcDKlSvp3r07\nPj4+uLm50bRpUwBOnjxZYZsxMTH2Y61WS5s2bbhw4YLrghYO5Im9AIp3nD2ZdgmNh4V3Pz2Btf1x\nlJSeuHkX4h3m+lnulsxM8vbvxZyeZq/zu3sQhlZt0ZQxUSdr8waHcsjD0xyG6AghhBCieho0aFCq\nbu3atYwbN46JEycyZ84cgoKCUBSFnj17UlBQUEYrjvz9HVe10+v1lbpPOEcSewHAxg3FCbYt3w2T\nScHX1ok81VJjvyA5u7aTf6JkrVs3Pz+MHSLKvFZVVYydu2HNTCfv50MEj3uo1Eo5QgghRH03btw4\nBg4cSLdu3ap87/WG17hKWd+Cf/7553To0IGlS5fa68pa5lzUD5LYCwDyNFklhS6HKUxsC6gANbLU\npe13y1kqGg3Gjp3LvVZRFLy6dgfAJ2qgDL8RQgghqqEqT81NJlOpCbKffPJJTYQlXKBaif3hw4c5\nePAgmZmZjBw5kiZNmpCfn09iYiLNmzfHy8vLVXGKGnb4SMlk2UbtC8lIBFDw0BpqfKnLho8/ieJW\n+lex6PIlzFeS8exc8mRDknohhBCietq1a8eHH37IihUraNOmDUFB5e/aHhMTw2OPPcYLL7xAdHQ0\n27Zts6+mI+ofpxJ7q9XK/Pnz2blzJ6qqoigK/fr1o0mTJmi1Wl577TVGjhzJmDFjXB2vqCGZhSXj\n6L093MkA3BQ3WhvCaWkIqbF+FTdtmUm9JTuLq59/iuKmxaNNezQeHjUWgxBCCHErmTJlCgkJCTz5\n5JOkpaUxceLEcq+dOnUq586dY8mSJSxatIi+ffvy7bff0qpVq1qMWFSWU4n96tWr2blzJw8//DBd\nu3blz3/+s/2cXq+nZ8+e7Nu3TxL7G0RKhuPkWE+tsY4iKaaqKlc++m/xscVKyof/odFfnqnTmIQQ\nQoibhY+PDytXrqzUtVqtlnnz5jFv3jyHelVVHcq/H4MP2De9+qONGzdWKVZRNU4tdxkfH09UVBTD\nhw/Hx8en1PmmTZuSnJxc7eBE7Xjjg5/tx4YmWXU+3CV3148O5YCR99dRJEIIIYQQNw6nEvsrV644\nbBv8R56enuTmun6JRFEzQny88XYrng/Rp1cAv55RseSUXnLSlVSrtcz6grOJ5CTsspcDRozCvUnT\nGo1FCCGEEOJm4NRQHA8Pj+sm7r/++muldlUT9cNf/9SRpOQs1mw/ySmfH8nc0h4ADRp0etc9vTen\nXSUrbjO2nGws2dkAaL2Kv/FRVZWrKz7GfLVkMyytpyeG8JYu618IIYQQ4mbm1BP79u3bs3Xr1lLj\nqwByc3P5/vvv6dixY7WDE7WnWUNffCOLd5BTzVoAAtz86dXX22V9mH45RNGli/akHsC7b/EOd5kb\n1jsk9QBB4ye5rG8hhBBCiJudU4n9/fffz+XLl3nxxRfZu3cvAOfOnWPjxo089dRTFBQUcN9997k0\nUFEzEjMvkZ5VvJZtkWomP8kXS447DXUNCQ30oUVLg8v6Ml8pmXeh8fDAp19/PG5rDYBXj54YWpQ8\nnW/0xDNojXU7iVcIIYQQ4kbi1FCcVq1a8fzzz/P222+zaNEiAJYsWYKqqvj6+vLCCy/QrFkzlwYq\nXC/bksPiJSfoOrSAwHwDmbZMMve1x03R4ak1unQYTv7pkxRdvmwvBz84CcVNR/6pE3i0aoMuOARF\nq0Uf2oSg+x9wWb9CCCGEELcKpzeo6tGjBx988AEHDx7kwoULqKpK48aN6dq1K+7uNTvxUrjGgr2r\nsebext7/6Sm0mWkS44s53YheKf4ix5XDcIouJjmUVRVS/rMIfaPGeLRqA4D/0BEu608IIYSor/bt\n21fXIYibVLV2ntXpdPTo0YMePXq4Kh5RS1LMV8j5paG9HNgin8x9oUDx2PrgEJ3LhuHYCgvJO3jA\nXvYbOIQrH/wHAEWrxZqbi1Z2KRZCCHGT8/f3r9WHn17yb+stx6nE/qmnnmLAgAH079+/zHXsRf33\ni+kIhame9rJfqJmL53zsw3Bc+bQ+/+hhh3LBmVP248KLF7CkpUpiL4QQ4qbn7e2Nt7fr/n0V4o+c\nSuyzsrL44IMPWLJkCd26dSMqKoo77rgDN7dqfQEgatGpgkSwFe9FEKoPJeN4ybryrnxaD2A15dmP\nNUYPChJPAcXj9/1iYnEPC3dZX0IIIYQQtyqnMvGPPvqIgwcPEh8fz549e9i7dy+enp7069ePAQMG\n0KpVK1fHKVwsI90MgMUKVy+BraEBKK5z5dP6P7KZ8rmW1GsMBoztbq+xvoQQQgghbiVOJfaKotCl\nSxe6dOlCQUEBP/74I/Hx8WzcuJGNGzfSuHFjoqOjGT16tKvjFS6QXJSC6WwAAOYiFVBQbcXnDIEF\nLn1a/3uqufiDg6LTARAwfFSN9COEEEIIcSuq9tgZg8HAgAEDGDBgAFevXiU+Pp4vv/yS5cuXS2Jf\nT63N/IacCw3Iz1fB7EZAgJarGtD5mgjunl1xA5WgqiqFZxMxX7lC4dkzABT9egl9o8YAeLRqjb5x\nqEv6EkIIIYQQLkjsr0lOTiYuLo6tW7diMpnQarWualq42AOB9/PPC8nkBR7FcLod+gAN7n4FBIw4\nirebn0v6KDx/lvR1X9nLqtmMotWi6HS4BQTgN/gel/QjhBBCCCGKVSuxz8vLY/v27cTHx3PixAlU\nVaV58+Y8/PDD9O/f31UxChfz0XrTLCwd68mO5NhsZPqfwdjlnMvat+blkr5mtUOdxsuLxhNnYTOZ\nUC3FSb4QQgghhHAdpxL7vXv3EhcXx969ezGbzfj6+jJs2DCio6MJD5cVTm4E7njg5VGINegSvsNL\nNsrQK7pqt1108YJD2e/uGAy3tUZjMGDNy0Pr6VnOnUIIIYQQwllOJfavvvqqfXOq6OhounbtKkNv\nbhBmq4UtOy8DBry8teS4F09o1SgagtwC6evdy6l2C84mUpB4GgBrZoa93qN1WzzatqfgbCIerdpI\nUi+EEEIIUUOcSuwfffRR+vXrJzua3YBWH0pg53YrHr1OoDvSniJLEQA+Wh8mBo93qk1bvomM9WtQ\nbbZS5wwtWpK9LQ7T0SN4tGqDarOhaDTVeg1CCCGEEKI0pzKsIUOGSFJ/gzrwSwbmIoX0swYysovs\n9dUZgmM1mcpM6rXe3igGD/J+PoTW05Oc3T+iFhY43Y8QQgghhChfpZ7Yp6amAhAcHOxQrsi160X9\ncCaxgPQkd6wFNjjZDI27O4pvFoDTQ3AAMjd9Yz/27BiB152/taXTkfKfxQBYsrLI/SkB7559nH8B\nQgghhBCiXJVK7B955BEURWHVqlXodDp7uSJr1qypdoDCdTYcPIo1T28va93Ar9sl/Nz8aGlo4VSb\neQf2Yb5ypaRNXz+0Xt6oNhu/LprvcG3wuIecC1wIIYQQQlSoUon92LFjURTFPkH2WlncWI6aDwPN\nUVHRGswEDjqLoWkW4Nza9arZTNYPcQ51xts7ApB/9BfH+g6dcAsIdKofIYQQQghRsUol9uPGjbtu\nWdwYrFnFK9Koig2vICuGpsWr1zg7vt6SneVQDnpgAhoPDwAKz52z12v0evzujnGqDyGEEEIIUTlO\nTZ6Ni4sjJSWl3PMpKSnExcWVe17UDdulIPtxwyGn0CgaQnTBTo2vz/x+I6nLltjLxna3o2/Q0F7O\nP33SfhwkQ3CEEEIIIWqcU4n9okWLOH78eLnnT548yaJFi5wOStQMq6V4+JRicUOjt9qXuHRmfH3+\n0SMOZa23j0NZ37Akydf6OjfURwghhBBCVJ5T69irqnrd8xaLpdpj8M1mMytWrGDr1q3k5ubSvHlz\nxo8fT5cuXa5735YtW1i4cGGZ5z7++GP8/f0d6o4dO8bSpUtJTEzEaDTSt29fJkyYgMdvQ0puFueS\ncgFQUVFVBUV7/Z9hhdSS5S29e/XBM6Krw+mg/5uAOe0qNpNJ5mMIIYQQQtQCpxL768nNzeWnn34q\nlUBX1VtvvcXOnTsZPnw4jRo1Ii4ujldeeYW5c+fSvn37Cu8fN24cDRo0cKjz/MOup2fOnGH27Nk0\nadKEKVOmkJaWxldffcXly5eZM2dOteKvb8KaetIupDEnLqWiGAsB58fWq6rKtc927mHN8b6zt/1c\nUfJlFI0GXUhDdIFBIPNlhRBCCCFqRaUT+5UrV/K///0PAEVRWLBgAQsWLCj3+hEjRjgd1MmTJ9m+\nfTuTJ0/m3nvvBSA6Opq//OUvLF26lNdff73CNrp160arVq2ue82yZcvw8vLitddew2g0AhASEsLb\nb7/NgQMHKvx24EaiKAppmhR0ei16fyvg/Nr1mRu+LrO+8OIF0r74DL+Bg3ELCEJxc/nnRiGEEOKm\nYDabmTt3LitWrCApKYmioiK++uorRo4cWdehiRtYpcfYh4eHExUVRWRkJKqq0q5dO6Kiokr9Fxsb\ny9NPP82kSZOcDurHH39Eo9EQE1Oykoper2fgwIEcP3680htkmUwmrFZruecOHjxI//797Uk9QFRU\nFB4eHuzYscPp+OsrGza0BguKAqMChju9dn3huTP2Yzf/APtx2hefAWC+kkJW/ObqBSuEEELcxObP\nn8/LL79M48aNefbZZ3nppZdo27Ztjfa5dOlSFEVh6dKlNdrPjepmeH8q/Ui1Z8+e9OzZEyjeeXbM\nmDFERETUSFBnzpwhNDTUIeEGaN26NQBnz56tcFfbF198kfz8fNzc3OjatSsPP/wwjRs3tp8/d+4c\nVqu11FN9nU5HeHg4Z86c+WOTN7SzBeco8k7HanbHqPF0Oqn/I+9efYHiITjX5B06QPCEyS5pXwgh\nhLgZff3113h5ebF582b0en3FNwhRCU6NlZg7d66r43CQkZFR5hj9a3Xp6enl3uvu7s6AAQPo2LEj\nRqOR06dPs3btWmbMmMGCBQvsHwgyMjIc2vxjP0ePHnXFS6kXcnMsvPnWRXLQofjk4I1nxTdVgr5x\nYzTu7ljzcrn62Qp7vSG8RfH4eiGEEEKU6fLlywQGBkpSL1zKqeUua1pRURE6XemJndd++QsLC8u9\nt2/fvjz55JNER0fTs2dPHnzwQV5++WVycnJYtWqVQx9Auf1cO38zOHLyt42kTB5grt64d9VmQ7Va\nHOqy47c4lH3vHlytPoQQQoib1Zw5c1AUhbNnz3L+/HkURUFRFJo3b26/Ztu2bQwbNoygoCDc3d1p\n1aoVL774IiaTyaGtoqIiFi9eTExMDE2bNsXd3Z2QkBBGjRrFgQMHHK6dNGkSkycXf5s+efJke7+/\nX7muefPmDnH8XmRkZKlV7q69lq1bt7J06VK6du2K0WgkMjLS4bq1a9cyYMAA/P39MRgMdOjQgTfe\neKPc4dIVKSgoYP78+URERODr64unpyfNmzdnzJgxHDp0qNT1lem/Mu/PjaBSWd6IESNQFIVVq1ah\n0+ns5YqsWbPGqaD0ej1ms7lU/bVk293dvUrttW/fntatW3Pw4EGHPoBy+7neJ+j09HT7E3/AvlmX\nzWYrs726ti15PxaLAdUGisaGqpb9uiti+vkgeT9usyf2NquNrIRdmE6V7GngN/J+bHo9tnr4Pggh\nhBC/52xiWR3Xkt633noLgKeeegoAP7/iPV/+85//8Oc//xk/Pz+GDRtGSEgIP/30E3PnziU+Pp74\n+Hh7jpKens5TTz1Fv379iI2Nxd/fnzNnzrBu3To2bNjAtm3b6NGjBwAjR44kMzOTtWvXMmLECDp3\n7uyy1/Svf/2L+Ph4RowYwaBBg9BqtfZzM2fOZN68eYSGhjJq1Ch8fX3Zvn07f/3rX9mzZ4/DQ9fK\nmjhxIp9//jmdOnVi8uTJuLu7c+HCBeLj49m7d6/DUPHK9l+T709tqlRiHxUVhaIoaDQah3JN8ff3\nL3O4zbVkOiAgoNS5igQFBXHp0iWHPn7f5h/7uV4fGzdu5LPPPitVn5SUxKZNm6ocW007c0SloEBX\nvERlvg6TLsupOBscSMDNlFdSkZICP1/7sKRwtX1HCo8chSM3zzAmIYQQwpUiIyOJjIy0T9D8/fLa\nR48eZfr06XTq1IktW7YQGFiyZvS8efOYOXMmixcv5v/9v/8HFOcySUlJhIaGOvRx5MgRevbsyQsv\nvMDmzcWLWfw+cR05cmS1Fjn5ox9++IE9e/bQsWNHh/rNmzczb948YmJiWL16tX3ZcVVVefzxx3n3\n3XdZvXo19913X6X7ysrKYtWqVXTr1o09e/Y4fIiwWq3k5OQ41X9Nvj+1qVKJ/bVPk+WVXa1Fixb8\n8ssvmEwmhwm0J06cAIpX6KmqlJQUfH197eWwsDC0Wi2nTp2ib9++9nqz2czZs2fp06dPuW0NHjyY\nO++8017+6aefWLFiBc2aNaNfv35Vjq0mqarK1iubKcxUUGwqRksA4QMvE9Pl3sq3YS6i4PQpck4d\nAa/iPwr3dh2w5WSj8fCg8NQJPHv0osOdzi2fKYQQQtSFs2fP1nUIDt577z0sFguLFy92SOoBZsyY\nwZtvvsnKlSvtib27u3uppB7g9ttvJyoqik2bNmE2m8scduxK06ZNK5XUA7z99tsA/Pe//3XYS0hR\nFObNm8d7773HypUrq5TYK4qCqqoYDAb7A+drtFqt/ZuPmuq/vquXC4337t2br776ik2bNtnXsTeb\nzWzZsoXWrVvbJ8Cmp6djMplo2LAhbr+tmZ6VleWQwENx4n369GmGDRtmr/P09CQiIoIffviBsWPH\n2j9AxMfHk5+ff93EPiAgwOGJ/oULFwDQaDQ1/sdTVclFKRRe8sWGGVQFvW8hQeFV+yPP2hZP3i+H\nUJTiPyA3Pz+CBg8FoOjyJcyhTTB27nbDjUMTQghxa/v90976YPfu3QBs2rSJLVu2lDqv0+k4fvy4\nQ93Bgwd5/fXX2bFjB8nJyaWG2l69epVGjRrVXNDAHXfcUWb97t278fT05KOPPirzvIeHR6nXUxEf\nHx9iY2P59ttv6dq1K/fffz+RkZH06NGjVG5TE/3Xd04l9tnZ2WRlZdG0aVN7XXJyMmvXriU3N5eo\nqCi6du3qdFBt2rShT58+fPLJJ2RmZtp3nr1y5QpPPPGE/bpPPvmEuLg43n//ffsuszNmzKBFixbc\ndttteHp6kpiYyPfff09QUBCjR4926GfChAnMmDGDF154gZiYGNLS0lizZg1dunShW7duTsdfn+w/\n9av9WDHrgfwqb0xlTnPcN8C9aZj9WN84FH3j0k8LhBBCCFE114YhV3b1wZ07dxIdHQ3AoEGDaNWq\nFV5eXiiKwpo1azh06NB1FxxxlWs52B+lp6djsVh4+eWXy703Ly+v3HPlWbVqFa+99hqffvops2bN\nAooT/smTJztsOlpT/ddnTiX277//PpcvX2b+/PkA5OfnM3PmTPsv5Pbt23n11Vfp0KGD04E9/fTT\nrFixgq1bt5Kbm0vz5s2ZPXt2hW3269ePvXv3cvDgQQoLC/H392fQoEH83//9X6mlLVu2bMnf//53\nli5dyocffoiHhwd33303Dz30kNNx1zc//5JrP9ZY9Og11mqtYR80djy6hjX7yV8IIYS4Ffn4+ADF\nD1C9vb0rvH7u3LkUFhayfft2h2HFUPy0uqwVYq5Ho9GUuypgVlZWufeV9429j48PiqJw9erVKsVR\nEaPRyKuvvsqrr77K2bNniY+P591332XhwoXk5+fz3nvv1Wj/9ZlTif2JEyccljLavn076enpjc3l\nZQAAIABJREFU/O1vf6NFixb87W9/48svv6xWYq/X65k8ebJ96aGyPPXUU6XG+z/44IM8+OCDle6n\nffv2vP76607HWd8Fehu5bDWj2sy4h+RQndFXGr0OfaPGmNOukrpsCcbbO2Jo0RJDy1YV3yyEEEKI\n67rzzjvZv38/u3fvZuDAgRVen5iYSEBAQKmk3mQysX///lLXXxt6VN5qQP7+/vzyyy9YLBb7EGco\nfqp96tSpqrwUoPj1bNiwgVOnTpXaENRVwsPDCQ8P54EHHiAkJIR169bZE/uq9l/R+3MjcGod+8zM\nTIKCSjYg2r9/P7fddhvdunXD39+f6Ojom27n1hvV+KgIOjRsgjHAitamQ6OzVbkNS1qaQ7kgsfiP\n23TkF4ouXyrrFiGEEEJU0eOPP46bmxtPPPEESUlJpc5nZmY6rE8fFhZGRkYGR44csddZrVaeffZZ\nUlNTS91/bX7gtbmBf9SjRw/MZjMrVpRsOqmqKjNnznRqyMr06dMBePjhh0n7Qy4BxcO4jx07VqU2\nU1NTOXz4cKn6jIwMCgsLMRgMTvdf0ftzI3Dq8a1Wq3X4qubw4cMMGDDAXvb09HRYbkjUHaNRSwH5\nuBUZQAPB3Uv/oV9P7r692H43Pk9VVUyHSv6n4t6suatCFUIIIW5pHTp04N///jePPfYYbdq0ITY2\nlpYtW5KTk8OZM2f44YcfmDRpEu+++y4ATzzxBN999x19+/ZlzJgxGAwGtm7dyqVLl4iMjGTr1q0O\n7ffq1QsPDw/eeustMjIy7IuRvPjiiwD85S9/YcmSJTzyyCNs3ryZ4OBgtm/fTlZWFhEREVUe2jN4\n8GBmz57N3//+d2677TYGDx5MWFgYaWlpnD592j50u127dpVu89KlS3Tp0oWIiAg6depEaGgoaWlp\nrF27FrPZzLPPPut0/xW9PzcCpxL70NBQdu3axdChQ0lISCA3N5dOnTrZz1+9erVSY8NE7Ug2F2+g\n5eZdSGDzqk2iKUw6Zz92CwjEmpWJ9bdP7VovL/TNwsq5UwghhBBVNXXqVDp37sybb77Jtm3bWL9+\nPb6+vjRr1oynn36aiRMn2q+95557+OKLL3jttddYvnw5RqOR6OhovvrqK1555ZVSbQcEBPDFF18w\nZ84c3n//ffLz84GSxLVDhw5s3LiRmTNn8sUXX+Dl5UVsbCxvvPEGY8aMcer1vPLKK9x1110sWrSI\nLVu2kJmZSWBgIOHh4cyZM4fx48dXqb3mzZszZ84c4uLi+P7770lLSyMoKIiuXbvy5JNPMnjwYKf7\nr+j9uREoqqqqVb0pLi6OhQsXYjQaKSwsJDg4mHfeecc+HuvFF19Eq9VedxbyzWTr1q28+eabPPPM\nM6W2Ua5LV81pnC48w+f/S6YoxRud0cpfHg2v0uTZtK9WUXj+HAANpj5G0aWLZHy7HgCvHnfi0+eu\nmghdCCGEqHGJiYk8/fTTLFiwgJYtW9Z1OEJUm1NP7KOjo1EUhd27d2M0Grn//vvtSX12djZ5eXkM\nGTLEpYGKqvvizBYOrWyMyeaOXgVPL32VknpVVSm6UDzGT1FA6+lF3v699vMag4fLYxZCCCGEEM5x\neomUqKgooqKiStX7+PiwYMGCagUlXOP8YTfMRSpYdWhsRjS6/Crdn/fTHlRbyWRbS3YWRcnJ9rJH\n2/Yui1UIIYQQQlRPtXeezc7OJiWleAx3gwYN7GuwirqlqipWkw6bWpyYa93UKk+cLUopSeK1/oFc\n+ei/9rKbvz/a323PLIQQQgjhjHPnzrF06dIKr/Pz8yu1zLlw5HRif/bsWf773/+WWqaoffv2TJ06\nlfDw8GoHJ5x3uvAM+ef8sVG8tbQuwFSlibOWrEwKTpesWesZ0RnT4V/QuLtTePECAcNHuTxmIYQQ\nQtx6zp07V6l5mWFhYZLYV8CpxP78+fM8//zzFBUVcccdd9CsWTMAkpKS2Lt3LzNnzuT111+314va\nty/jF6B4rwGl0IDiZqKvd69K35930HFjC2OHCIydupC3/yd87x6Mm5+fK8MVQgghxC0qMjISJ9Zy\nEWVwKrFfsWIFWq2WN998s9ST+fPnzzNz5kxWrFjBzJkzXRKkqLqCFC8sVlBt4JZvxF1rqdrE2YIC\n+7FnRBcUrRZVVfHs2r3craOFEEIIIUTdcSqxP3LkCLGxsWUOtwkLCyM2NpYNGzZUOzjhvMwrSvHE\nWUDnU+jUjrPXePfsDSAJvRBCCFENly5dKnMH1JrSuHFjgoKCaq0/UfecSuwLCwvx9/cv97y/vz+F\nhVXbCEm41lXrVVQaoaCgcWLi7O+ZfjmE1j8AXXADGYIjhBBCOMlqtWI2m2utP5vN+Yd64sbkVGLf\noEED9u7dy9ChQ8s8v3fvXho0aFCtwET1+BNIik2LBStuPgV4h+U61Y5qNpO9/QfQanFv0oTA0Q+4\nOFIhhBDi1tKtW7ca72Pfvn013oeofzTO3BQdHc2BAwd44403SEpKwmq1YrVaOX/+PPPnz+fgwYMM\nGDDA1bGKKnhiQDRdmofiazSgcbc43Y756hX7sUe7Dq4ITQghhBBC1ACnntiPHDmSxMREtm/fzo4d\nO+xjr1VVRVVV+vbty8iRI10aqKgavV6DUWtEq9U4+fGt+Gm9ml+yqZU+tKmLohNCCCGEEK7mVGKv\n1WqZMWMGBw4cYM+ePQ4bVPXs2ZPOnTu7NEhRdbm51mq3YTXlOZS1vr7VblMIIYQQQtSMau0826VL\nF7p06eKqWISLHDyRSkpWDqYiHTabWqUH9raCAnITdmE6dgR+N+kmYNhIWRVHCCGEEKIec3KQRonC\nwkIuXLjAhQsXZCWcemJV/DG+3nSZfZfOYqFqs+9NR34hd/9PxYVrm0UooPX2cXGUQgghhBDlmzNn\nDgaDocr3LV26lOXLl9dARPWf00/sk5KSWLJkCYcOHbIvp6TRaIiIiGDy5Mmy62wdyk4t+bzmprgB\nRZW+1/bb8BtVVbHmZBcn9Botik7n6jCFEEIIIcr1yCOPEBsbW+X7li5disFg4MEHH6yBqOo3pxL7\nxMREZs2aRUFBAREREfYkPikpiQMHDnDs2DFee+01WrSo/E6nwjWyi0z2Y0XVoFGc/FLGakHfsDGK\nuzsAWl9Zv14IIYQQNa+oqAi9Xk+TJk1o0qRJXYdzQ3Eq61u6dCmKojB//nxefvllpkyZwpQpU3j5\n5ZeZP38+iqKwZMkSV8cqKiE576r9WPHMv86VFVA0GDt1RhcYiPedvVA01R61JYQQQoh6Yvfu3cTE\nxODr64uXlxfdu3dn7dq1AKSnp/PII48QHByMwWCga9eu9nMAK1asQKPRkJSU5NCmxWIhODiYGTNm\nAHDy5EnGjx9PWFgYHh4etG7dmhdeeIGCggKH+xRFYe7cucyePZvQ0FAMBgMFBQWlhuLk5+czffp0\n2rVrh9FopEmTJowbN47Lly/br4mMjOSHH35g06ZNKIqCoijMmTPHfn7jxo306dMHo9GIv78/EyZM\n4OrVq78Ph3/961+0bt0ag8FAYGAgffv2Zc+ePdV7w2uJU0/sT5w4wfDhw2nZsmWpcy1btiQ2Npb1\n69dXOzhRdTl5JWvWB7j7UOTktAdFq8Xrjp7oAmUraiGEEOJmsmPHDgYMGEDXrl157733CAwM5MCB\nA/a9iYYMGcLJkyf5xz/+QbNmzfjggw+49957Wbt2LcOGDWPkyJF4eHiwcuVKnnvuOXu7mzZt4urV\nq4wfPx6AS5cuERYWxujRowkICODYsWO88sorXLx4kU8++cQhpnfeeYcuXbrw3nvvYTab0ZUxBDg/\nP5/8/HxeeuklGjVqREpKCvPnz+euu+7i2LFj6HQ6/v3vf/Pggw/i7u7OggULAOxP/desWcN9993H\n+PHjmTVrFpmZmcyaNYuRI0eyY8cOAJYtW8bMmTN5+eWX6dOnDzk5Oezdu5eMjIwa+Vm4mlOJvU6n\nw9/fv9zzAQEB6PV6p4MSzkvNLtlh1lwAFszVW/pICCGEEDeV5557jrCwMLZt22ZPoAcOHAjAunXr\nSEhIYO3atQwfPhyAIUOG0LVrV+bMmcOwYcPw9PRk+PDhfPrppw6J/cqVK2nfvj0REREAREVFERUV\nBRTP3evTpw/e3t5MnDiRt99+Gx+fkoU5vLy8WLduHVqttty4AwICeP/99+1lq9VK7969adq0KZs3\nbyY2Npb27dvj4+ODwWCgZ8+eDvc/88wzjBw50uFDRbt27ejatSsbN25k8ODB7Nmzh06dOjFr1iz7\nNcOGDavaG1yHnMr5unfvTkJCAkOHDi3zfEJCQq1slyxKS00tXgXHolop8sxAbyr+AKZXKjf51WYx\nYzOZUDw8aixGIYQQ4lZk/Wk3xp8PcXnblhrvy5iVhTkyGkIGONSbTCZ2797NK6+8UuZT8e3bt2M0\nGh2SWUVRGDt2LC+88AJ5eXl4enoybtw4hg8fztGjR2nfvj0mk4m1a9cyc+ZM+32FhYX885//ZPny\n5SQlJTmsnnjmzBmHfY9iY2Ovm9Rfs3LlSubPn8/JkyfJycmx1588efK6E21PnTrF2bNn+ec//4nF\nUjK6oWPHjoSEhJCQkMDgwYPp3r07//73v5k+fTqjRo2iV69euP823/BG4NTA6SlTppCTk8O8efM4\ndeoUJpMJk8nEyZMnmTdvHjk5OUyZMsXVsYpKuHymeIlKs2pGYy35Rezr3eu691lNJq6uWknu7p2Y\nky9jTjpPUdL5Go1VCCGEELUrIyMDm81GaGhouedDQkJK7V3TsGFDVFUlMzMTgMGDBxMYGMiKFSsA\nWL9+PXl5eYwbN85+z/PPP8/cuXOZMGEC69evJyEhgXfeeQeg1Dj7Bg0aVBj72rVrGTduHB06dODT\nTz9l165d7N69u8z2/ujKlSsAjBkzBp1O5/DflStX7PMFrn2bsGPHDqKjowkMDGTSpEmkp6dXGF99\n4NQT+wkTJqAoComJifY39Br1t7XPH3rooVL3rVmzxpnuRBX49bxA+kENppOeuBnNUAjhhjBaGspf\noUi12Uhf/RnmtDTMV4p3Edb4eKNaq797rRBCCCHqD39/fzQaDZcuXSr3/JUrV1BV1SG5T05ORlEU\n/PyKV8nT6XSMHj2alStXMnfuXD799FN69epF8+bN7fd8/vnnTJ06ldmzZ9vrDh8+XGa/ldkE8/PP\nP6dDhw4sXbrUXnf27NkK74PiYTwACxcupHfv3qXOBwUF2eN4/PHHefzxx0lNTWX9+vU89dRTAA79\n1ldOJfZRUVGyC2k9dX+rSL49lsGPnqfQWtxRFB0+2utvLmU6fAhzWpq9rOj1uPkH4NG+Q02HK4QQ\nQtwytN17YgprSbtaGK786759BDZtWqreaDTSu3dvPvnkE5577jnc3BxTwX79+vHGG2/wzTffcM89\n9wDFD20///xzunTpgqenp/3acePG8d577/Htt9+yceNG+2TVa0wmU6k5l8uWLXP6NZlMplLDh/44\nCRdAr9eXeoLftm1bmjVrxqlTp5g+fXql+gsODubhhx/m66+/5ujRo07HXZucSuyvfXIR9U+Amz/D\nB3phMm3j2MrSf9B/ZMlIJyvue4DiJ/RWK/pmYWi9fdAajTUdrhBCCCFq2euvv05kZCT9+/dn+vTp\nBAUFcejQIfR6PY899hh33HEHEydO5B//+AdNmzblww8/5ODBgw5LXkLxh4CmTZsydepUbDYbY8aM\ncTgfExPDBx98QLt27WjSpAkff/wx5887P8w3JiaGxx57jBdeeIHo6Gi2bdvGZ599Vuq6du3asWTJ\nEtatW0fjxo3t/y1atIjRo0djMpkYPnw4Pj4+XLx4ke+++45p06bRr18/pk2bhq+vL7169SIwMJCf\nf/6ZTZs28fTTTzsdd22SxclvMjnZVny8dVT2C5Wi5F/tx9bcHPTNwkDR4NG6bQ1FKIQQQoi61KtX\nL3744Qc8PDyYMmUKI0eOZOXKlYSFhaHVatmwYQP33nuvfSnIM2fO8OWXX5ZaHUZRFB544AEuX77M\noEGD7MNZrlm8eDExMTE8++yzTJgwAT8/PxYuXOh03FOnTuW5555jyZIljBw5koSEBL799ttS182Y\nMYN+/foxceJEevTowX//+18ARowYwebNm0lKSmLChAkMHTqUuXPn4uPjY99UtU+fPuzZs4dp06YR\nExPD4sWLef7553n55Zedjrs2Keq1QfHCaVu3buXNN9/kmWeeITIyss7iOJNYgLlIpVFjHStNKzj+\nvyaQ60VEg3AmTQ0p8x7TsSNkbvoW1WLBmp2FW0AgAEFjxqFvXPbEGiGEEOJmkJiYyNNPP82CBQvK\n3JvH1ZKSkkhNTa2VlQP37dtH06ZNCQkp+99/cXOSJc5vIjt2pvHTufO4a9xx66FgyXHHrYIn95bf\nJssWXb6I7rekHpCkXgghhBDiBiOJ/U3EZM1HxUaBLR9bltU+zkqnLzu7t2Skk3tgHwCKpmRUVtCY\ncWVeL4QQQggh6i8ZY38TufJryYYLqrZkqcpefb3LvN6cdtV+rPEwomvcGJCn9UIIIYQQNyJ5Yn8T\nMWEqPtCo6HP9sAJ6nyJatDRUeK//0OF49+xTswEKIYQQQogaI0/sbxKqqqLYin+cGp0NtzzfKt2v\n9fKqibCEEEIIIUQtcTqxT01NZeHChUyePJlRo0Zx6NAhALKysli4cCGnTp1yWZCiYgUFKlaKh9+4\nN8xGY9FXcIcQQgghhLiZOJXYJycn88wzz7Br1y6aNWuGzWazn/P19eX06dN89913LgtSVCwn20q+\nLR8AS647qJXfGVi1WrFmZ2PJzkK1WCq+QQghhBBC1DtOjbFfvnw5Go2GxYsX4+7uzoQJExzOd+/e\nnYSEBJcEKConI70kITc2z8B8unG516qqSuamb8k/Xrw9svlKCplbNpGTsFvWrxdCCCFq2L59++o6\nBHGTciqxP3ToEEOHDiU4OJjs7OxS54ODg0lLS6t2cKLy8oqK7Mde+JBxnWstqVfsST2AWlSIovEB\nQGP0rKkQhRBCiFta48aNadiwYa31p9Vqa60vUT84ldibTCYCAgLKPW+xWLBareWeF64X2MRCj6FW\nLl8sonH7QjJOlz8Z9vfDbVSbDa2nJxqjEa23N25+frURrhBCCHHLcXOTxQhFzXLqNywoKIikpKRy\nz584cYJGjRo5HZSourDAQB4K7M/Z9ufIs+VxwJYMuFd4n0fbdhScPAGAm2/VVtIRQgghROVZLBaH\neYk1TavVylP7W4xTiX2vXr3YuHEjAwcOxN/fHwBFKZ6suXPnTn788UfGjZPdS+tCuKE5H6euAIq/\n6tMq158fXXD8KGiK/+jdm4XXdHhCCCHELevy5cukpqbWWn9NmzYlJCSk1voTdc+pxH7MmDHs3buX\nZ599lttvvx1FUVi9ejXLli3j1KlThIeHM3LkSFfHKipgVs3oFB1Fqtle11DX4Po3KSWr57i3aFlT\noQkhhBBCiBrmVGJvNBr517/+xYoVK/jhhx9QVZWDBw/i6enJkCFDmDBhAnq9rKNemzb+cpj8NHca\newVi/W0UlJuiw0frU+49qqpizcrEza94voQuKLg2QhVCCCFuad26davxPmTlnVuT07M4jEYjU6dO\nZerUqWRlZaGqKr6+vvYhOaJ2xa8vwmTNBFJoN7ly4/dsubn2Y11gUA1FJoQQQgghaoNLpmf7yqTL\nOpVYcIYMj0tYsz0AMGlyK7jjN6oNrbcvGg8PvHr2rsEIhRBCCCFETXMqsT98+HClruvQoYMzzYsq\n2pa1E0t2CGADFFSKn9hrytlYOGfnNgC0Pr549+6L9x29ailSIYQQQghRU5xK7GfNmlWpITdr1qxx\npnlRRSZLof1YAfzc/MjQGvB08y91rSUzk8KLF+1ljU7mQgghhBBC3AycSuynT59eqs5qtZKcnExc\nXBwhISHExMRUOzhRscSCMyRnZ6JawlAtWrwCrejWxtKwyIpNq5a6XjWX7FCrqiqGNu1qM1whhBBC\niFo1adIkdu/ezfHjx+s6lBrnVGI/YMCAcs+NGjWKp556yumARNXsyNlF+obbsVk0KDYFrcWdzMyS\nnWV1+rK/WbHmZKNv2IiiSxdwb9IUjYextkIWQgghhKg1s2fPJje3kvMPb3DX373ICV5eXgwaNIgv\nv/zS1U2LMhSpZgrTDWBTUBQFb50nfn5u+Pm5ERyio1df7zLvs+bmUnghiYxv1mG9RX7ZhRBCCFH3\nioqKKr7IhVq2bElERESt9llXXJ7YQ3Fyn5KSUhNNiz9QVbDlF4+TV1Bo1sSDSVNDmDQ1hPETg2nR\n0lD2fYUF9mOtr1+txCqEEEKI+iEhIYFBgwbh7e2Nl5cXI0aM4OzZswB88803KIrCypUr7dcXFBTQ\noUMH+vTpg9VqBaB58+Y8+uijzJ8/n6ZNm+Lh4cHAgQM5ffq0Q1+KojB37lxmz55NaGgoBoOBgoKC\nCuO45uOPP6ZTp04YjUb8/Pzo0aMHX3/9tf38N998w5133om3tzc+Pj506tSJjz76yH5+0qRJtG3b\n1qHNw4cPM2TIELy8vPDy8mLIkCGlFoe59vo++OADWrRogbe3N4MGDeLcuXNViq82uTyxLyoqIj4+\nHj8/SRZrg2pTUIIyISALbasL9OpWyU2mNCU/eo1sJiaEEELcMhISErjrrrtwc3Nj+fLlLFu2jPPn\nzzNgwACKiooYOnQof/rTn3j88ce5cOECAM8//zxJSUksW7YMrVZrb+vrr79m5cqVLFq0iA8//JCT\nJ08yePBgzGazQ5/vvPMO+/fv57333mP16tXodLoK4wDYtm0bkyZN4u6777b3dd9995Geng5AYmIi\n9957L61bt2b16tWsXr2aKVOmkJmZWe7rv3DhAv369ePSpUssXbqUpUuXcvHiRe666y77671mw4YN\nrFixgrfeeov333+fw4cPM378ePv5iuKrbU6NsV+4cGGZ9bm5uRw/fpzs7GwmTZpUnbhEBRILzrAj\nZxdXksCg0aMxaGnt0YJ2LSv+QKWazWApHoevCwys6VCFEEIIUY8899xzdOzYka+//hrNbw/6evbs\nSYsWLfjoo4/sT+Hj4uKYOHEizz//vD1xb9GihUNbGRkZHDx4kKCg4o0u27ZtS7du3VixYoVDLujl\n5cW6descPhRUJo6EhAT8/f1588037fcNGTLEfnzgwAHMZjPvvPMOPj4+AAwcOPC6r3/BggUUFBSw\nefNmGjRoAECfPn0IDw/nrbfeYv78+fZrzWYz3377LR4exXsFZWZm8thjj3Hp0iVCQ0MrjK+2OZXY\nx8XFlVnv5eVFaGgojzzyCP37969WYOL6duTs4nxiEambW6EB9GZ3Avwa4aaU/yNVVZX0dV9iyUiz\n1+lCGtZCtEIIIYTYbf2JQ8af2XJ5W433lWXMItocyQBCHOrz8/PZvn07//jHP7DZbNhsxXvfBAcH\n07FjRxISEnj00Ufx9PRk2bJl9O3bl507dzJq1CgmT55cqp/IyEh7Ug/QtWtXwsPDSUhIcEjsY2Nj\nHZL6ysbRvXt3MjIymDBhAuPHj6dv3754eXnZ24mIiMDNzY0HHniAqVOnctdddxEQEHDd92b79u1E\nRkbak3qARo0aERkZyfbt2x2u7devnz2pB2jfvj1Q/NQ/NDS0wvhqm1OJ/dq1a10dh6iiItVM5r5Q\nQEGv6Alw8y93ouw11pxsrDk52PKLx7Upblp87oqs+WCFEEIIUS+kp6djtVqZMWMGM2bMKHX+2lNv\ngO7du9OmTRuOHDnCE088UWZ7ISEhpeoaNGjAr7/+WqrOmTgiIyP57LPPWLx4Mffccw9arZbY2FgW\nLlxIs2bNaNWqFRs2bOCf//wnY8eOxWq10r9/f9566y06duxYZswZGRllbqLasGFDEhMTHer8/R33\nBNL/Nnz52hyBiuKrbVVO7AsLC/nxxx8JDQ2lTZs2NRGTqCTVrEWxafH0UxjcuwFNml5/rLwl7Sq2\noiKwFg/DcW/eUpa5FEIIIW4hfn5+aDQa/vrXvzJ69OhS5729Sx4Szps3j1OnTtGhQwemT5/O3r17\n7YntNVeuXCnVRkpKCl26dHGo++PGplWJY+zYsYwdO5asrCw2bdrEM888w0MPPcTWrVsBuPvuu7n7\n7rsxmUzEx8czY8YMhg0bVmqS6zX+/v5lLvKSnJxcKpGvjIriq01VTux1Oh1vv/02U6dOlcS+Hsi3\n/f/27j0+qure//9rrrknBBJIuCYEAo3coXILIAEBq1KtxVqRtqjtOcfihWPFSosoFfWntQXr6c/z\n4KBiDUXxiHJERYEEEUEBJXKVS6AEISQhCblnbvv7R2RkTAIhJJlJ8n4+Hj6yZ+219/7MLDePz6xZ\ne61KKk8H8fqHh6iMjiS1T9866xluN4XvvIVRVQmAJSqK0AGDWjJUERGRdm2UZQRJFb0Y/oPhzX6t\nXad30aNTj1rlYWFhjBkzhn379vH000/Xf/yuXTz++OM88cQT3HLLLQwZMoQ//OEPPPvssz71MjMz\nOXv2LJ2+fWbviy++4NixY4wcOfKi8TU0jgtFRUVx6623sn37dl577bVa+0NDQ7n++us5evQo999/\nP+Xl5YSFhdWqN27cOF588UXy8vK8vzjk5uayefNm7rnnngbF0pj4WsJlJ/Zms5mYmBgqKiqaIx5p\noNJ/heMqDcJwejDMLio9HtzmkHrre75N6DGZwGrFEhmFLbaBM+iIiIhIm/Hcc89xzTXXcPPNN3PH\nHXcQExPD6dOnyczMJC0tjRtvvJFZs2YxevRofve732E2m/nrX//Kv/3bv3H99ddzzTXXeM8VHR3N\n1KlT+eMf/0hFRQXz58+nd+/e3H777Vccx6233srChQvJy8tj4sSJxMXFceTIEV577TWmTJkCwH//\n93+zZcsWfvSjH9GtWzdOnTrF3/72N8aOHVtnUg8wd+5cXn75ZaZMmcKCBQswDINFixZ7eASRAAAg\nAElEQVQRHBx82YusXiq+ltaoMfZpaWlkZmYyffp0bDZbU8ckDZC/syYp95QHYw6rBqBrzKUf1rBE\nRBKaMoBOt/ysWeMTERGRwHT11Vezbds2Fi5cyK9//WsqKiro1q0b48ePZ+DAgcybN49vvvmG999/\n3ztbzd13383//d//8ctf/pKsrCzvtOY33HADffv2Zc6cORQUFJCamsqLL77YoPzwUnGcr7NkyRLe\neustiouLiY+PZ9asWTz++OMADBo0iHfffZeHHnqIgoICYmNjmTZtGosXL673uj169ODjjz/moYce\n4pe//CVQ04u/cuVKevSo/SvHpd7DxeJraSbDMIzLPSgrK4uXXnoJp9PJj370I+Lj4wkKCqpVr64H\nE9qizMxM/vKXv/Cf//mfPt9im9PDz2+h4pyJsiIzwZEerBaYNy+BHkHd66zvLi/jzLL/H4DQqwbQ\n4Vr/TcUkIiISCI4ePcrcuXP561//SlJSUrNf78SJE+Tn5zN8eAsMxdm1ix49etT5cGtTSUhIYNq0\nabz44ovNdg25PI3qsX/00Ue928uWLav1QIRhGJhMJt5+++0ri04uyuOw1Iys+Xb2qA5WLQomIiIi\n0l41KrG/7777mjoOaQSP0wymmpXdrFFVBJkuPiuOu6wUT2UlnuoqDLcb0wXzyYqIiIhI69aoxH7S\npElNHYc0gqfKhiekArAROTD3ootTAbhLSjCqKinbtpXQlIEE9+7TMoGKiIhIm1PfdJLiP+bGHLR0\n6VK+/vrrevcfOnSIpUuXNjooaRinyeHdtncuw2y6eHMaLpd3Oyih90VqioiIiEhr06jEftOmTeTm\n5ta7/8yZM2RkZDQ6KGkYI6IMzDVLMHfvdPEFFVxFheCqGbZj6RCNydyophcRERGRANWooTiXUlVV\nhUXjt5tdSKQbW3QFw4ZFcWvn6RetW7l/n3fbHKrVZkVERETamgYn9vn5+T7L7548eZK9e/fWqldW\nVsb7779PfHx800QotWQfrcJRYsdimAi2h/LLoeMueUzZFzu826EpA5szPBERERHxgwYn9hs2bGDV\nqlWYTCZMJhOrV69m9erVteqdn+ry/vvvb9JA5TvbPinFwMATVInZdvGZcAA8DgfOU98AYLLbCf/h\nxZd5FhERkaZntVqx2Wx89dVXzX4tm83mXVxK2o8GJ/ajRo3yLnLw/PPPM3XqVPr16+dTx2QyERwc\nTN++fYmNjW3aSAWo6a3Pz3NSFXcCT7WFTsM83i9T9XGdK/Zum+x2LOERLRGqiIiIXKBr16507drV\n32FIG9bgxD4xMZHExEQA8vLyGDNmDL169Wq2wJxOJ+np6WRmZlJWVkZCQgIzZ85k6NChFz0uKyuL\nzMxMDhw4QEFBAdHR0QwaNIiZM2fSsWNHn7rz58+vczjR0KFD/bYU8KVs+6QUgErrOTiaxIECD4dj\nS0juHVXvMVWHD3m3zSEaXy8iIiLSFjXq4dmf//znTR1HLUuWLOHTTz9l+vTpxMfHs2nTJhYtWsTi\nxYtJSUmp97gVK1ZQWlrK2LFj6dq1K7m5uaxbt44dO3awdOlSoqN9Z4+JiYlh1qxZPmXf/wIQSJwO\nA7elGjxmzNaaGXHCbMEXPaYi6wvMYeF4yss0d72IiIifVFRUUFVV1WLXCw0NJTj44jmCtC3NMivO\nlTp06BBbtmxh9uzZ3HzzzQCkpaUxZ84cXnnlFZ555pl6j73zzjtJSUnxGVc2bNgw5s+fz7p167jj\njjt86oeGhjJx4sTmeSNNLPtoFcXFLqoiCzH9qzsmqwcLViyW+ofhGIaBp7oaa2xn3EHBBCUktmDE\nIiIicl5BQQH5+fktdr0ePXoosW9nAvKpiq1bt2I2m5k6daq3zG63c+2113Lw4MGL3hQDBgyo9bDI\ngAEDiIiIICcnp85j3G43lZWVTRN8Mzo/DKcqIh8cNQ/NWk1Wojte+vuZyWzG2qED4T8c1awxioiI\niIh/BGSPfXZ2Nt26dSP0e/OtJycnA3Ds2LHLeji3srKSyspKIiMja+07deoUM2bMwOVy0aFDB6ZM\nmcJtt92G1Rp4H43TYQBQHpqP2Vbz8I3ZZCYouGHfz4ITe2O2X3oWHREREWk+w4cPb/Zr7Nq1q9mv\nIYEn8LJXoKioqNZYeMBbVlhYeFnnW7t2LS6Xi9TUVJ/yuLg4Bg4cSK9evaiurmbr1q288cYbnDp1\ninnz5jX+DTQzw20CixuAoFDjonXzX12O4/QpzCEhGMbF64qIiIhI6xWQib3D4cBms9Uqt3/b21xd\nXd3gc+3du5dVq1aRmprK4MGDffbdd999Pq8nTpzICy+8wIcffsj06dPp379/I6JvXgYGxUYhJmqm\nHu1yVVm9dd0VFbiKCjEqK3BXt9zDOiIiIiLS8gIysbfb7TidzlrlDocDgKCgoAad5+TJkzz11FP0\n7NmTOXPmNOiYm266iQ8//JCsrKx6E/vCwkKKioq8r8+vyOvxeOqMu6l4PB6c1gqMvE7wbed7SnhS\nvdcseudNDI+BgYHZbm/2+ERERFoTt9vt7xBEmlRAJvbR0dF1Drc5n0w3ZDrK/Px8Hn30UUJDQ1m4\ncGGt8fr1OT92v6ys/p7wDz74gFWrVtUqP3HiBOvXr2/QdRojJ6cL5YYTI8YCHg8Ww0LpiWOsLz5U\nZ/3O2cewlZUQUu2g2h7E6SNHONuM8YmIiIiI/wRkYt+7d2/27NlDRUWFT0L+9ddfA3gXyqpPSUkJ\nCxcuxOl08sQTT1zWvPS5ubkAdT5oe960adMYOXKk9/XOnTtJT0+nZ8+ejBs3rsHXulx535zlXLEb\ntxPCpxwi1B3JTVffgNVae7pLwzDIP7wPI8iGo7yUiG7dsfdKZMQFMw2JiIi0Z8eOHfN3CH7z5z//\nmeeff55vvvmGcePGkZmZecljjh8/TmJiIv/85z+57bbbALjmmmsIDg7mgw8+aOaIpSECMrEfM2YM\na9asYf369d557J1OJxs3biQ5Odnbq15YWEhFRQVxcXHeWWyqqqpYtGgRZ8+eZfHixfUu3VxRUYHN\nZvMZy28YBm+88QZQM/d9fTp27OjzZeH8NJpms7nOZwOaitlsxmT2YA1yE967mA5WCAmpe5abyq8P\n4C4qxF1cjDk4FJPJjMViadb4REREWhOLxeLvEPxi7969PPTQQzz88MNMnz6dDh06+DskaSIBmdj3\n69ePsWPH8uqrr1JcXOxdeTYvL497773XW+/VV19l06ZNLFu2jC5dugDw3HPPcejQISZPnkxOTo7P\n3PUhISGMGlUzj/vRo0f585//zPjx44mPj6e6uprt27dz4MABpk6dSlJSUsu+6SbmLi/HXVoz773p\n22TeElH/rxAiIiLSPhw8eBCA3/zmN/Tu3dvP0UhTCsjEHmDu3Lmkp6eTmZlJWVkZCQkJLFiwgAED\nBlz0uPM/q23YsIENGzb47OvcubM3sY+NjSUlJYVt27ZRXFyMyWSiR48e3HPPPT4LYwUah6UCl7US\np8d10ekr3aUl8O1DQZawcCKuHk3o4KEtFaaIiIgEoF/96lesWLECwNuJ+fLLLzN79myfITYATz/9\nNI888oimy25FAjaxt9vtzJ49m9mzZ9db54EHHuCBBx7wKfuf//mfBp0/Li6Ohx9++Ipi9Ifs8hNU\n5lkpemkgITefhS511yvZvMm7besUQ8SY1LorioiISLuxYMEC+vTpw4IFC3jrrbeIj49vt0OS2qKA\nTeylbtVBpVAeAYaJ0OD6b0R7lzhMQcG4i84SctXAFoxQREREAlVSUhJ9+vQBYOjQoSQkJHD8+HH/\nBiVNRol9K+Op/q7J0rqOqLeeyWbDZLNh69qNyNTxLRGaiIiIXMTO7W6+ygrl442nmv1a586Fck2a\nk86Tmv1SEkDM/g5ALo+pJKLmr8lE37CLP+BrMpuxdWj4VJ8iIiIi0nopsW9FDMPAMH/7QKxR/zCc\nsl07cBbk464ob6nQREREpJUKDg4GwOFw+JTXtVioBDYNxWlFKlxV3m0jpLrOOobHQ8mWTJy5p8Fq\nhe49Wyo8ERERuYgRoyz0Sqpg+PAfNPu1du06TY8enRpUt3Pnztjtdvbv3+8tMwyDjz76qLnCk2ai\nxL4VceHybof0OFdnnarsI9+9MOkHGREREbk4s9nMT3/6U/7+97/Tp08funfvzvLly9Vj3wopsW9F\nqq1l3m17uLPOOhVf7fZum4OCmj0mERERaf2ef/55HA4HDz30EEFBQfz2t79l2LBhzJ8/39+hyWVQ\nYt+KGCbPdy9MdS8W4cj5l3dbK82KiIjI9912220+C1EBdOrUidWrV9eq+8gjj3i3ExISai1WlZmZ\n2SwxSuMosW9FggnF1iEXw+okunPt/e7yci6830w2W8sFJyIiIiJ+pUHYrUgU0US4OxJpC6NzgqfW\nflfhWe+2OTyiJUMTERERET9TYt+GVB7c593W+HoRERGR9kWJfStS4i7BZdT90CxA9Ylvx9e7nJoR\nR0RERKSdUfbXChytymZFfjr7/pVHVZVBtcOFldoLVIUNGUZwUh+cBQWYQkJqCk0tHKyIiIiI+IUe\nnm0FPindRp4zn2pzHIbLjKPQxpCwQbXqhQ+/GmOoB3dxMc6zBQAE9+3X0uGKiIiIiB8osW8FHN7h\nNyZMmDCbzPQLSa6zrsn83Y8wJhNEjBrbAhGKiIiIiL8psW9FTB4zJpMZa6cSQs0hddZxFRV6e+s1\nM46IiEjgCA0NJTo6muzs7Ga/VnR0NEGaSKPdUWLfSlSeiMKDBwtmzN9rNsPloui9tZgsFqq/Oekt\nN+kBWhERkYARExNDTEyMv8OQNkyJfStRvKsbJrcFzGBYfWfGMZxOqrKPAuCprMAcEgpA6KAhLR6n\niIiIiPiHunRbCY/DAqaaZWU7hob57Ks+ecK7bTLXTINjsloIH3F1ywUoIiIiIn6lxL6V8DhNGBiY\nzGCtCvfZV75rh3fbZK8ZT2ey2lo0PhERERHxLyX2rUDpv8JxeL4bfhPavdS7bRgGrm8flgWwxXZu\n0dhEREREJDAosW8F8nfG4gmuAmomvAy/4Cl3T2UlHmdN0h/cOwmTVY9NiIiIiLRHSuxbAY/TjLU8\ngpAIGDTOYNKA7+awdxcXebctkVH+CE9EREREAoC6d1sJi8dORJTBfdeN8yn3VFR4tw1HdUuHJSIi\nIiIBQj32rZzH6fBuB/VM8F8gIiIiIuJXSuxbOWvHToQOGIgtNhZTaBjOvDx/hyQiIiIifqChOK2E\nI7wQh9nDmzu+YHyf/nSOrlmEyt4lDnuXaQCUfvYphsfjzzBFRERExE/UY99KOG2VVH8TxeYPKvjm\nVFWddRwnc7zbNi1ZLSIiItKuKLFvJVwRxd5tu+XSzRZ9/Y+bMxwRERERCTAaitNKmL9KAbsLgBBr\nzTz21Sf+havwLOaQEOw9E6jOOQGAyWLGHBLqt1hFREREpOUpsW8lDKsLEwAmwsItABR/9D7u0ppV\naMMGD7mgcouHJyIiIiJ+pqE4rYDHMPg2q8dqstIlzo7jTK43qbdGRnm3AWxx8f4IU0RERET8SIl9\nK2BYnd5e+KAeZwFwfDvsBiCoV4JP/Y43/bSlQhMRERGRAKHEvhVwW79bUTbMFgKAq7jIWxY6cLBP\nfZNVI6xERERE2hsl9q2AccHU9LFBHQGo2PuVt8zasVNLhyQiIiIiAUaJfSvgcZvAVDMWp/JczYOz\n1qgo73710IuIiIiIMsJWIJhgXJ2rsAfDXbN6YXg8uM6dA8ASFubn6EREREQkECixbwVMHiuW8giC\nrAaRtgiqL1hh1hoT68fIRERERCRQaChOK2S4nFhCah6iDerew8/RiIiIiEggUI99K+AIL8JlMsCo\neXA2OKE3wf82B8OoGXdvGAYYWpVKREREpD1TYt8KVDiqcZyOwRFSwcnT5XSPrxlXbzKZcOSeoujd\nd3CXlfk5ShERERHxJyX2rYDrXAgeNxgVdkxum8++ygP7fZJ6c3AwmEwtHaKIiIiI+JkS+9bA6qr5\na5gICjJhGAYmkwnD46Eq+4i3mj0ujvCRYzApsRcRERFpd5TYtwIeq6Nmw4DIKAtnV72Gq+gs7vJy\nsNq8iXz0DTdhCY/wY6QiIiIi4i9K7FsBI/f8lJYmLLhxnMn9dofh0ztvDgtv+eBEREREJCBousvW\nIKi65q/ZQ+nmjTXbbhfO8wk+EHv7LA3BEREREWnHlNgHOMMwoCoYAPMFreUuLcUcEup9feG2iIiI\niLQ/SuwDnNNpAN/OUe+04i45B4Dh8WDp2AmAoB49sURE+ilCEREREQkESuwDnNPhAasbAFNMMc4z\npwEw2+2Yvu3Cjxw3wW/xiYiIiEhgUGIf4AwTMOpLCKnCZHPjcThrdlj03LOIiIiIfEeJfYALC7UQ\nkzMUW0EX4sss3nJPZYUfoxIRERGRQKPEPsCZTCYsziCCYyuINZV4F5XVmHoRERERuZDGc7QSJjN8\nE9mTuHtvp2LPblyFhZTv/sLfYYmIiIhIgFBiH+CKihy4qswYnpqZcTzlZZR/sQtXcZGfIxMRERGR\nQKLEPsD9f3/fx9kyBwYGIVaDquwjPkm9yaQ57EVEREREiX1A83gMKjyVAJg8kFB4DFexCcPjwWQ2\nY7JaiUq7VuPtRURERESJfSArL/Ng4AHAZqqmX/5hzn10CJPNjiUyiugbfkxwQm8/RykiIiIigUCz\n4gQwp8vt3Q6KOw5OF+5z5zBZbf4LSkREREQCkhL7AFbsKAXA4nETVumGmudnMYWE1JRrCI6IiIiI\nfEtDcQLY7mOnATAbbqzfdt5bOkRjjYgg8ppJ2DrF+DE6EREREQkkSuwDWGkRGB4Dk+EhvKymu94S\n1YHOv/o1JquaTkRERES+o6E4AcxjcoNR8/BsF3dBzd+7/k1JvYiIiIjUosQ+gOVUnPKOq7cZTkqD\no7B2iPZvUCIiIiISkJTYB7DJV/VnZFIl0TGHiax047RoNhwRERERqZsS+wA2pn8iN4zthd0VTL65\nC8ciEvwdkoiIiIgEKA3WDnChQeFUOLuQFdqd4BD12IuIiIhI3dRj30q4TfoOJiIiIiL1U7YYwL7I\nPsm5EyfAUgqE+DscEREREQlgSuwD2JvrTlB0sgy3JYrh5fs5Gv0Df4ckIiIiIgFKQ3ECWHWxBY8H\nrIaLBMc3mGyGv0MSERERkQDV7nvsnU4n6enpZGZmUlZWRkJCAjNnzmTo0KH+Dg2HqQIw8FSHUGCN\nptPAHH+HJCIiIiIBqt332C9ZsoR33nmHCRMmcPfdd2M2m1m0aBH79+/3d2gYbg8GBtbIIipDrHRK\nqPZ3SCIiIiISoNp1Yn/o0CG2bNnCL37xC2bPns20adN44okniI2N5ZVXXvF3eJg8JgA8didn7R2Y\n0O06P0ckIiIiIoGqXSf2W7duxWw2M3XqVG+Z3W7n2muv5eDBg+Tn5/sxOvC4a/6GVJiw2ELoE5rk\n13hEREREJHC168Q+Ozubbt26ERoa6lOenJwMwLFjx/wRFgBulwu3wwGYMGHgMbXrphIRERGRS2jX\n2WJRURHR0dG1ys+XFRYWtnRIXo7CQjxGzVAcEx4qwsP9FouIiIiIBL52PSuOw+HAZrPVKrfb7QBU\nV/vvYVVPSCihQZU4PR5MJogcrwdnRURERKR+7Tqxt9vtOJ3OWuUOhwOAoKCgOo8rLCykqKjI+/rM\nmTMAeDyeOs/XKBYLc+8dy3PLNmINriasZxOeW0RERHC73f4OQaRJtevEPjo6us7hNueT9o4dO9Z5\n3AcffMCqVatqlZ84cYL169c3aYyhQSXYEnI4fSKY9Tub9twiIiIi0na068S+d+/e7Nmzh4qKCp8H\naL/++msAEhMT6zxu2rRpjBw50vt6586dpKen07NnT8aNG9ekMcaOy2J/VQhjwkbSe2jd8YiIiMjl\n8+ckGSLNoV0n9mPGjGHNmjWsX7+em2++GahZiXbjxo0kJycTGxtb53EdO3b06c3PyalZEdZsNtc5\nZv9KjIwawcioEU16ThEREQGLxeLvEESaVLtO7Pv168fYsWN59dVXKS4uJj4+nk2bNpGXl8e9997r\n7/BERERERBqsXSf2AHPnziU9PZ3MzEzKyspISEhgwYIFDBgwwN+hiYiIiIg0WLtP7O12O7Nnz2b2\n7Nn+DkVEREREpNHa9QJVIiIiIiJthRJ7EREREZE2QIm9iIiIiEgboMReRERERKQNUGIvIiIiItIG\nKLEXEREREWkDlNiLiIiIiLQBSuxFRERERNoAJfYiIiIiIm1Au195tik4nU4ADhw44OdIREREpKHO\nnDkDgMPh8HMkIk1DiX0TqKqqAuD999/n/fff93M0IiIicjnKy8v9HYJIk1Bi3wQmTJgAQEhICFZr\n032kZ86cIT09nZkzZ9KlS5cmO6+0HLVh66c2bN3Ufq1fc7dheHg4AwcObPLziviDEvsmEBkZyY03\n3tjk5z169Cjp6emMGDGCpKSkJj+/ND+1YeunNmzd1H6tn9pQpOH08KyIiIiISBugxF5EREREpA2w\nPPbYY4/5OwipX3BwMAMHDiQkJMTfoUgjqQ1bP7Vh66b2a/3UhiINYzIMw/B3ECIiIiIicmU0FEdE\nREREpA1QYi8iIiIi0gYosRcRERERaQOU2IuIiIiItAFaoCoAOZ1O0tPTyczMpKysjISEBGbOnMnQ\noUP9HVq709i2yMrKIjMzkwMHDlBQUEB0dDSDBg1i5syZdOzY0afu/Pnz2bt3b61zDB06lMcff7xJ\n30971Ng23LhxI0uXLq1z34oVK4iOjvYpO3DgAK+88gpHjx4lNDSU1NRUZs2apVk8mkBj27C+ewvA\nYrGwZs2aS9bVfeg/lZWVrFmzhq+//prDhw9TVlbG/fffz6RJk/wdmkjAUmIfgJYsWcKnn37K9OnT\niY+PZ9OmTSxatIjFixeTkpLi7/Dalca2xYoVKygtLWXs2LF07dqV3Nxc1q1bx44dO1i6dGmtpDAm\nJoZZs2b5lH3/C4A0zpXeT7fffnutZezDwsJ8XmdnZ7NgwQK6d+/OXXfdxdmzZ1mzZg2nTp1CMwpf\nuca24YwZM7j22mt9yqqrq/n73/9e55cC3YeBpaSkhFWrVhEbG0tiYiJ79uzxd0giAU+JfYA5dOgQ\nW7ZsYfbs2dx8880ApKWlMWfOHF555RWeeeYZP0fYflxJW9x5552kpKRgNn832m3YsGHMnz+fdevW\ncccdd/jUDw0NZeLEic3zRtqxprifhg8fTt++fS9a5x//+Afh4eE8+eSThIaGAtC5c2deeOEFvvzy\nS/3adgWupA3r+twzMjIAmDBhQq19ug8DS8eOHb2/jh0+fJgHH3zQ3yGJBDyNsQ8wW7duxWw2M3Xq\nVG+Z3W7n2muv5eDBg+Tn5/sxuvblStpiwIABPkn9+bKIiAhycnLqPMbtdlNZWdk0wQvQdPdTRUUF\nbre73n27d+9mwoQJ3qQeYOLEiYSEhPDJJ59c2Zto55r638SPP/6Y4OBgRo4cWed+3YeBw2az1fp1\nU0QuTj32ASY7O5tu3br5JAgAycnJABw7dozY2Fh/hNbuNHVbVFZWUllZSWRkZK19p06dYsaMGbhc\nLjp06MCUKVO47bbbsFp1i16JpmjDP/7xj1RWVmK1Whk2bBh33nknXbt29e4/fvw4bre7Vq++zWYj\nMTGR7OzsJno37VNT3ofnzp1j9+7dpKamEhwcXGu/7kMRae30r1WAKSoqqrOH4nxZYWFhS4fUbjV1\nW6xduxaXy0VqaqpPeVxcHAMHDqRXr15UV1ezdetW3njjDU6dOsW8efMa/wbkitowKCiISZMmMXDg\nQEJDQzly5AjvvPMO8+bN469//as3mSwqKvI55/evs3///qZ4K+1WU96HW7Zswe12c80119Tap/tQ\nRNoCJfYBxuFwYLPZapXb7Xag5sEvaRlN2RZ79+5l1apVpKamMnjwYJ999913n8/riRMn8sILL/Dh\nhx8yffp0+vfv34joBa6sDVNTU32+hI0aNYphw4bxyCOPsHr1au655x7vNYB6r3N+vzROU96HH3/8\nMVFRUQwZMqTWPt2HItIWaIx9gLHb7Tidzlrl55ODoKCglg6p3Wqqtjh58iRPPfUUPXv2ZM6cOQ06\n5qabbgJqps2Uxmvq+yklJYXk5GR2797tcw2g3uuc3y+N01RtmJuby8GDB0lNTcVisTToGN2HItLa\nKLEPMNHR0d6f9i90vkxTr7WcpmiL/Px8Hn30UUJDQ1m4cGGtccL1OT/Mo6ys7DIilu9rjvspJibG\np13ODwmp7zq6Z69MU7Xh5s2bgbpnw6mP7kMRaW2U2AeY3r17880331BRUeFT/vXXXwOQmJjoj7Da\npStti5KSEhYuXIjT6eTxxx+/rAQvNzcXoM4HbaXhmuN+OnPmDFFRUd7XvXr1wmKxcPjwYZ96TqeT\nY8eO6Z69Qk3Vhh9//DFxcXGXNaRG96GItDZK7APMmDFj8Hg8rF+/3lvmdDrZuHEjycnJmhGnBTW0\nLQoLCzl58iQul8tbr6qqikWLFnH27FkWLlzoM4vKhSoqKmoNMzAMgzfeeAOomfteGu9K2vDcuXO1\nzrdz506OHDni0y5hYWEMHjyYzZs3+ySfGRkZVFZWMnbs2OZ4a+3GlbTheUePHiUnJ6fe3nrdhyLS\nVujh2QDTr18/xo4dy6uvvkpxcbF3lcW8vDzuvfdef4fXrjS0LV599VU2bdrEsmXLvCuUPvfccxw6\ndIjJkyeTk5PjM3d9SEgIo0aNAmoSjj//+c+MHz+e+Ph4qqur2b59OwcOHGDq1KkkJSW17JtuY66k\nDefNm0fv3r3p06cPYWFhHD16lA0bNhATE8NPf/pTn+vMmjWLefPmMX/+fKZOncrZs2d5++23GTp0\nKMOHD2/R99zWXEkbnnepYTi6DwPXu+++S3l5uXf2o88//5yCggIAbrjhhlqrQA/pnWUAAA4rSURB\nVIu0d0rsA9DcuXNJT08nMzOTsrIyEhISWLBgAQMGDPB3aO1OY9vi2LFjAGzYsIENGzb47OvcubM3\nsY+NjSUlJYVt27ZRXFyMyWSiR48e3HPPPT4L8kjjNbYNx40bx44dO9i9ezfV1dVER0d75zX//vSL\nSUlJ/OlPf+KVV15h+fLlhISEMHnyZH7xi18051trN67k30SPx8OWLVtISkqie/fuddbRfRi43n77\nbfLy8ryvt23bxrZt2wC45pprlNiLfI/JMAzD30GIiIiIiMiV0Rh7EREREZE2QIm9iIiIiEgboMRe\nRERERKQNUGIvIiIiItIGKLEXEREREWkDlNiLiIiIiLQBSuxFRERERNoAJfYiIiIiIm2AEnsRERER\nkTZAib2IiIiISBugxF5EREREpA1QYi/SRuTm5rJ48WLuuOMOpk+fzpIlSy77HHfffTfz589vhugC\n35IlS5g+fXqD669cuZLp06dz5syZZoxKRESk4az+DkBEmsbSpUs5fvw4t956Kx06dCA+Pt7fIbV6\n27dvJzs7m9tvv93foTSr7Oxstm/fzqRJk+jSpYu/wxERkUZSj71IG+B0Otm/fz8TJ07k5ptvZuLE\nifTv39/fYbUqc+bM4c033/Qp2759O6tWraqz/s9+9jPefPNNOnfu3BLhNatjx46xatUq8vLy/B2K\niIhcASX2Im1AcXExhmEQHh7u71BaLavVit1ub3B9i8WC3W7HZDI1Y1SXr6Kiwt8hiIiIn5gMwzD8\nHYSINN6SJUvYtGlTrfLFixczcOBA3nvvPbZv386JEycoKSkhIiKCQYMGcccdd9QadnH33XfTuXNn\nnnzySW/ZgQMHeP3118nOzqa8vJyIiAgSExP52c9+5vOrQHl5OatXr+bTTz+loKCA0NBQBg8ezKxZ\ns4iLi2vw+/jHP/7BSy+9xM6dO3E4HPTr14/Zs2eTlJTkU9/tdvP222+zadMmcnNzCQ4O5qqrruL2\n228nISHBp+6mTZtYt24dp06dwuVy0aFDB/r378/dd99NVFSUz/XXrl0LwPz589m7d2+tOO+//34m\nTZrEypUrWbVqFcuWLaNLly689957vPjii/zhD39g5MiRPsd4PB7uuusuIiMjWbp0qbf88OHDrF69\nmn379lFZWUmXLl2YOHEit9xyCxaL5ZKf2fTp00lLS2PixImsXLmSY8eO0adPH5588knOnj3L22+/\nzVdffUVeXh4Oh4O4uDjS0tK46aabvOc//z6+Ly0tjQceeACo+UVozZo1bN68mdzcXOx2OykpKdx+\n++212kVERPxHY+xFWrlp06aRmJjI8uXLGTVqFKNHjwage/fuAKxZs4Z+/fpx4403Eh4ezokTJ/jw\nww/Zs2cPzz//PJGRkfWe++TJkzz66KNER0dz44030qFDB4qLizlw4ADHjx/3Jvbl5eXMmzePgoIC\nJk+eTI8ePSgqKuK9997jd7/7HX/5y18aPGTlscceIzw8nJ///OcUFRWxbt065s+fzzPPPEOvXr28\n9Z577jk++eQThgwZwnXXXee93rx583jqqae8CWdGRgZLlizxJqJ2u52CggJ27drFuXPnvIn9982Y\nMQOPx8P+/fuZO3eut/wHP/hBnfXHjx/P8uXLycjIqJXYZ2VlcfbsWW666SZv2Y4dO3jqqaeIj4/n\npptuIiIigoMHD7Jy5Uqys7P5/e9/36DP68iRI2zbto0pU6aQlpbmLT9+/Djbtm1j9OjRxMXF4XK5\n+OKLL1ixYgW5ubn89re/BWD06NEUFRWxfv16ZsyY4f3/5vwzGi6Xi8cee4wDBw4wceJErr/+eioq\nKvjwww95+OGHeeqpp+jbt2+DYhURkealxF6klevfvz/R0dEsX76chIQEJk6c6LP/b3/7G8HBwT5l\nV199NQsWLOCjjz7illtuqffcX375JdXV1fzud78jOTm53norV67kzJkzPPvssyQmJnrLJ02axL33\n3svKlSu9vb+XEhsbyyOPPOId4jJ69GgefPBBXnrpJR5//HFvXJ988gmpqak89NBD3rqpqanMnTuX\nZcuW8fTTTwOwbds2QkJCWLx4sU8v+MyZMy8ax9ChQ9m8ebP32YVLCQ8P54c//CE7duygrKzMZ1hU\nRkYGFouFCRMmAOBwOPjb3/5GcnKyT1wXfknbs2cPAwcOvOR1T5w4waJFixgyZIhP+YABA1i2bJnP\nUKEf//jH/OUvf+Gjjz7i5z//OR07diQxMZH+/fuzfv16hgwZUuua69atY8+ePTz22GMMGzbMW/6j\nH/2IOXPm8PLLL/v8wiMiIv6jMfYibdz5pN7j8VBeXk5JSQmJiYmEhYVx6NChix4bGhoKwGeffYbD\n4aizjmEYZGZmctVVV9GpUydKSkq8/wUHB9OvXz92797d4Hh/8pOf+CSjffr0YciQIWRlZVFZWQnU\nPNQKNb3qF9ZNTEzk6quvZv/+/Zw7dw6AsLAwqqur2bFjB8098nDSpEk4nU62bNniLausrGT79u0M\nGzaMDh06ALB7926Ki4uZPHmyt03O/zd8+HCg5stLQyQmJtZK6gGCgoK8n43T6aS0tJSSkhKGDh2K\nx+PhyJEjDTp/ZmYm3bt3p0+fPj5xulwuhgwZwv79+6murm7QuUREpHmpx16kjcvKyuL111/n0KFD\ntZLzsrKyix47fvx4MjMzWb16Ne+88w79+vVj2LBhjBs3zju05ty5c5SWlvLll19yxx131Hkes7nh\nfQg9evSos+zLL78kPz+fnj17cubMGcxmc511e/bsyfbt2zlz5gxRUVHMmDGDffv28eSTTxIREcGA\nAQMYPnw4qamp3i8uTWXYsGFERUWRkZHBddddB8Cnn35KVVWVT69/Tk4OAM8//3y95youLm7QNbt2\n7Vpnudvt5s033yQjI4PTp0/X+lJzqba/MFaHw1Fv2wKUlJQQGxvboPOJiEjzUWIv0oYdPnyYhQsX\nEh8fzy9+8Qu6dOnincnl2WefvWQPts1m409/+hOHDh3iiy++YN++faSnp/PPf/6TBx98kNGjR3vP\nMXjw4IsO6/GXrl278l//9V9kZWWRlZXFvn37eOGFF1i5cqV3jHtTOT/cZu3atZw6dYquXbuSkZFB\neHh4rXH3ALNnz/YZunShjh07NuiaQUFBdZYvX76cd999l3HjxjFjxgw6dOiAxWLh6NGjrFix4rJ+\nvejVqxd33XVXvfvre05BRERalhJ7kTZs8+bNeDweFi5c6DMzTVVVFeXl5Q0+T3JysneMfX5+Pg88\n8ACvvfYao0ePJioqirCwMCorK+scEnK5cnJyas3Bn5OTg9ls9vYKx8XF4fF4yMnJqZUYnzhxAsBn\nxh+bzcaIESMYMWIEADt37mTRokW88847/Pu//3u9sTRmKsu0tDTWrl1LRkYGU6ZMYe/evUyZMgWb\nzeatc/7LRFBQUJN8ZnXJyMjgqquu4qGHHvIpP3369GWdJz4+npKSEgYNGnRZv7yIiEjL07/SIm1Y\nfYnY6tWr8Xg8lzy+pKSkVllMTAxRUVHeoRxms5kJEyZw6NAhtm7dWud5GjqsBOCtt97y6U0+evQo\nWVlZDB48mJCQEABv7/ebb77pU/df//oXn3/+OSkpKd5e5Lrew/kZc0pLSy8ay/nnEy5V70K9e/cm\nISGBzMxMMjIy8Hg8PrPVwHdDdv73f/+3znNXV1df8Xz0ZrO5Vq98VVWVdzrPC53/XOuKJS0tjaKi\nIt555506r1NUVHRFcYqISNNRj71IGzZq1CjWrl3L448/ztSpU7FarezevZvjx49fdJrL815//XV2\n797NiBEjvD3gn3/+OSdPnuQnP/mJt96sWbM4cOAAzzzzDKmpqSQnJ2Oz2cjLy2PXrl0kJSU1eFac\n/Px8Fi5cyNVXX+2d7tJut/OrX/3KW2fo0KGkpqayZcsWysrK+OEPf0hxcbG37q9//Wtv3UcffZSw\nsDCuuuoqYmJiKC8vZ+PGjZhMpkvOdtOvXz/WrVvHiy++yIgRI7BYLCQnJ19yXv60tDReeukl3nrr\nLbp161brF4jg4GDmzp3Lk08+yX/8x38wefJk4uPjKS8v5+TJk2zbto358+c3aFac+owdO5YPPviA\nZ555hsGDB1NcXMyGDRuIiIioVbdv376YzWZWr15NWVkZwcHBdOnSxTtN6u7du3n55Zf56quvGDRo\nECEhIRQUFJCVlYXdbmfx4sWNjlNERJqOEnuRNiwlJYXf//73vP7666Snp2O32xkyZAhPPfUUjzzy\nyCWPHzVqFEVFRWzdupXi4mLsdjtdu3Zlzpw5XHvttd56YWFhPPPMM6xZs4atW7fy2WefYbFY6NSp\nEykpKT51L+Wxxx5j+fLl/POf/6S6utq7QNX3h9w8+OCDJCUlsXHjRl566SWCg4MZMGAAM2fO9Fmg\n6rrrruOTTz7hgw8+oKysjIiICHr37s1vfvMbBg0adNFYxo8fT3Z2Nlu2bGHr1q14PB7uv//+Syb2\nEyZMYMWKFVRUVPh8AbrQsGHDeO6553jzzTfJzMykpKSEsLAw4uPj+fGPf1xrka3LdddddxESEsIn\nn3zCZ599RkxMDFOnTqVv374sWLDAp25sbCz33nsvb731Fi+++CIul4u0tDT69euH1Wrl0Ucf5b33\n3iMjI4OVK1cCNc8AJCcn1/o1QkRE/Ecrz4pIQPj+yq8iIiJyeTTGXkRERESkDVBiLyIiIiLSBiix\nFxERERFpAzTGXkRERESkDVCPvYiIiIhIG6DEXkRERESkDVBiLyIiIiLSBiixFxERERFpA5TYi4iI\niIi0AUrsRURERETaACX2IiIiIiJtgBJ7EREREZE24P8BerQfz+VeG98AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x63e3fef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (8634916)>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Assemble the data for ROC curves\n",
    "model_order = ['full', 'expressions', 'covariates']\n",
    "\n",
    "auc_output = pd.DataFrame()\n",
    "roc_output = pd.DataFrame()\n",
    "\n",
    "for model in model_order:\n",
    "    metrics_partition = metrics_dict[model]\n",
    "    for partition, metrics in metrics_partition.items():\n",
    "        auc_output = auc_output.append(pd.DataFrame({\n",
    "            'partition': [partition],\n",
    "            'feature_set': [model],\n",
    "            'auc': metrics['auroc'].round(3)\n",
    "        }))\n",
    "        roc_df = metrics['roc_df']\n",
    "        roc_output = roc_output.append(pd.DataFrame({\n",
    "            'false_positive_rate': roc_df.fpr,\n",
    "            'true_positive_rate': roc_df.tpr,\n",
    "            'partition': partition,\n",
    "            'feature_set': model\n",
    "        }))\n",
    "\n",
    "(gg.ggplot(roc_output, gg.aes(x='false_positive_rate',\n",
    "                              y='true_positive_rate',\n",
    "                              color='feature_set',\n",
    "                              linetype='partition'))\n",
    " + gg.geom_line(size=1.1, alpha=0.7)\n",
    " + gg.labs(x='false positive rate', y='true positive rate')\n",
    " + theme_cognoma()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AUROC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>partition</th>\n",
       "      <th>test</th>\n",
       "      <th>train</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>feature_set</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>covariates</th>\n",
       "      <td>0.853</td>\n",
       "      <td>0.844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>expressions</th>\n",
       "      <td>0.911</td>\n",
       "      <td>0.914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>full</th>\n",
       "      <td>0.911</td>\n",
       "      <td>0.915</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "partition     test  train\n",
       "feature_set              \n",
       "covariates   0.853  0.844\n",
       "expressions  0.911  0.914\n",
       "full         0.911  0.915"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.pivot_table(auc_output,\n",
    "               values='auc',\n",
    "               index='feature_set',\n",
    "               columns='partition')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## What are the classifier coefficients?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "final_pipelines = {\n",
    "    model: pipeline.best_estimator_\n",
    "    for model, pipeline in cv_pipelines.items()\n",
    "}\n",
    "final_classifiers = {\n",
    "    model: pipeline.named_steps['classify']\n",
    "    for model, pipeline in final_pipelines.items()\n",
    "}\n",
    "\n",
    "coef_df = pd.concat([\n",
    "    get_model_coefficients(classifier, model, covariate_df.columns)\n",
    "    for model, classifier in final_classifiers.items()\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>coefficient_sign</th>\n",
       "      <th>-1.0</th>\n",
       "      <th>1.0</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>feature_set</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>covariates</th>\n",
       "      <td>17</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>expressions</th>\n",
       "      <td>49</td>\n",
       "      <td>51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>full</th>\n",
       "      <td>62</td>\n",
       "      <td>71</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "coefficient_sign  -1.0   1.0\n",
       "feature_set                 \n",
       "covariates          17    16\n",
       "expressions         49    51\n",
       "full                62    71"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Signs of the coefficients by model\n",
    "pd.crosstab(coef_df.feature_set, np.sign(coef_df.weight).rename('coefficient_sign'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Top coefficients for covariates model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>weight</th>\n",
       "      <th>abs</th>\n",
       "      <th>feature_set</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>PCA_34</td>\n",
       "      <td>0.060960</td>\n",
       "      <td>0.060960</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>PCA_29</td>\n",
       "      <td>0.056451</td>\n",
       "      <td>0.056451</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>PCA_87</td>\n",
       "      <td>-0.051578</td>\n",
       "      <td>0.051578</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>PCA_37</td>\n",
       "      <td>-0.042979</td>\n",
       "      <td>0.042979</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>PCA_24</td>\n",
       "      <td>0.038007</td>\n",
       "      <td>0.038007</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>PCA_11</td>\n",
       "      <td>-0.037749</td>\n",
       "      <td>0.037749</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PCA_2</td>\n",
       "      <td>-0.035994</td>\n",
       "      <td>0.035994</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>PCA_22</td>\n",
       "      <td>-0.035715</td>\n",
       "      <td>0.035715</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>PCA_62</td>\n",
       "      <td>-0.034043</td>\n",
       "      <td>0.034043</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>PCA_57</td>\n",
       "      <td>-0.033934</td>\n",
       "      <td>0.033934</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   feature    weight       abs feature_set\n",
       "34  PCA_34  0.060960  0.060960        full\n",
       "29  PCA_29  0.056451  0.056451        full\n",
       "87  PCA_87 -0.051578  0.051578        full\n",
       "37  PCA_37 -0.042979  0.042979        full\n",
       "24  PCA_24  0.038007  0.038007        full\n",
       "11  PCA_11 -0.037749  0.037749        full\n",
       "2    PCA_2 -0.035994  0.035994        full\n",
       "22  PCA_22 -0.035715  0.035715        full\n",
       "62  PCA_62 -0.034043  0.034043        full\n",
       "57  PCA_57 -0.033934  0.033934        full"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_df.query(\"feature_set == 'full'\").head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Top coefficients for individual _genes_ for full model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>symbol</th>\n",
       "      <th>weight</th>\n",
       "      <th>abs</th>\n",
       "      <th>description</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>entrez_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4193</th>\n",
       "      <td>MDM2</td>\n",
       "      <td>-0.008231</td>\n",
       "      <td>0.008231</td>\n",
       "      <td>MDM2 proto-oncogene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64782</th>\n",
       "      <td>AEN</td>\n",
       "      <td>-0.007349</td>\n",
       "      <td>0.007349</td>\n",
       "      <td>apoptosis enhancing nuclease</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23729</th>\n",
       "      <td>SHPK</td>\n",
       "      <td>-0.006798</td>\n",
       "      <td>0.006798</td>\n",
       "      <td>sedoheptulokinase</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81626</th>\n",
       "      <td>SHCBP1L</td>\n",
       "      <td>-0.006789</td>\n",
       "      <td>0.006789</td>\n",
       "      <td>SHC binding and spindle associated 1 like</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23612</th>\n",
       "      <td>PHLDA3</td>\n",
       "      <td>-0.006680</td>\n",
       "      <td>0.006680</td>\n",
       "      <td>pleckstrin homology like domain family A member 3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1643</th>\n",
       "      <td>DDB2</td>\n",
       "      <td>-0.006492</td>\n",
       "      <td>0.006492</td>\n",
       "      <td>damage specific DNA binding protein 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1819</th>\n",
       "      <td>DRG2</td>\n",
       "      <td>-0.006370</td>\n",
       "      <td>0.006370</td>\n",
       "      <td>developmentally regulated GTP binding protein 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132671</th>\n",
       "      <td>SPATA18</td>\n",
       "      <td>-0.006295</td>\n",
       "      <td>0.006295</td>\n",
       "      <td>spermatogenesis associated 18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60401</th>\n",
       "      <td>EDA2R</td>\n",
       "      <td>-0.006162</td>\n",
       "      <td>0.006162</td>\n",
       "      <td>ectodysplasin A2 receptor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51499</th>\n",
       "      <td>TRIAP1</td>\n",
       "      <td>-0.005990</td>\n",
       "      <td>0.005990</td>\n",
       "      <td>TP53 regulated inhibitor of apoptosis 1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            symbol    weight       abs  \\\n",
       "entrez_id                                \n",
       "4193          MDM2 -0.008231  0.008231   \n",
       "64782          AEN -0.007349  0.007349   \n",
       "23729         SHPK -0.006798  0.006798   \n",
       "81626      SHCBP1L -0.006789  0.006789   \n",
       "23612       PHLDA3 -0.006680  0.006680   \n",
       "1643          DDB2 -0.006492  0.006492   \n",
       "1819          DRG2 -0.006370  0.006370   \n",
       "132671     SPATA18 -0.006295  0.006295   \n",
       "60401        EDA2R -0.006162  0.006162   \n",
       "51499       TRIAP1 -0.005990  0.005990   \n",
       "\n",
       "                                                 description  \n",
       "entrez_id                                                     \n",
       "4193                                     MDM2 proto-oncogene  \n",
       "64782                           apoptosis enhancing nuclease  \n",
       "23729                                      sedoheptulokinase  \n",
       "81626              SHC binding and spindle associated 1 like  \n",
       "23612      pleckstrin homology like domain family A member 3  \n",
       "1643                   damage specific DNA binding protein 2  \n",
       "1819         developmentally regulated GTP binding protein 2  \n",
       "132671                         spermatogenesis associated 18  \n",
       "60401                              ectodysplasin A2 receptor  \n",
       "51499                TP53 regulated inhibitor of apoptosis 1  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca_for_full = (\n",
    "    final_pipelines['full']\n",
    "    .named_steps['features']\n",
    "    .get_params()['expressions__pca']\n",
    "    )\n",
    "classifier_for_full = (\n",
    "    final_pipelines['full']\n",
    "    .named_steps['classify']\n",
    "    )\n",
    "gene_coefficients_for_full = get_genes_coefficients(\n",
    "    pca_object=pca_for_full,\n",
    "    classifier_object=classifier_for_full,\n",
    "    expression_df=expression_df,\n",
    "    expression_genes_df=expression_genes_df,\n",
    "    num_covariates=len(covariate_df.columns)\n",
    "    )\n",
    "gene_coefficients_for_full.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Top coefficients for individual _genes_ for expressions model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>symbol</th>\n",
       "      <th>weight</th>\n",
       "      <th>abs</th>\n",
       "      <th>description</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>entrez_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4193</th>\n",
       "      <td>MDM2</td>\n",
       "      <td>-0.008032</td>\n",
       "      <td>0.008032</td>\n",
       "      <td>MDM2 proto-oncogene</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64782</th>\n",
       "      <td>AEN</td>\n",
       "      <td>-0.007234</td>\n",
       "      <td>0.007234</td>\n",
       "      <td>apoptosis enhancing nuclease</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81626</th>\n",
       "      <td>SHCBP1L</td>\n",
       "      <td>-0.006931</td>\n",
       "      <td>0.006931</td>\n",
       "      <td>SHC binding and spindle associated 1 like</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23729</th>\n",
       "      <td>SHPK</td>\n",
       "      <td>-0.006899</td>\n",
       "      <td>0.006899</td>\n",
       "      <td>sedoheptulokinase</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23612</th>\n",
       "      <td>PHLDA3</td>\n",
       "      <td>-0.006629</td>\n",
       "      <td>0.006629</td>\n",
       "      <td>pleckstrin homology like domain family A member 3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1643</th>\n",
       "      <td>DDB2</td>\n",
       "      <td>-0.006350</td>\n",
       "      <td>0.006350</td>\n",
       "      <td>damage specific DNA binding protein 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132671</th>\n",
       "      <td>SPATA18</td>\n",
       "      <td>-0.006340</td>\n",
       "      <td>0.006340</td>\n",
       "      <td>spermatogenesis associated 18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1819</th>\n",
       "      <td>DRG2</td>\n",
       "      <td>-0.006139</td>\n",
       "      <td>0.006139</td>\n",
       "      <td>developmentally regulated GTP binding protein 2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60401</th>\n",
       "      <td>EDA2R</td>\n",
       "      <td>-0.006048</td>\n",
       "      <td>0.006048</td>\n",
       "      <td>ectodysplasin A2 receptor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51499</th>\n",
       "      <td>TRIAP1</td>\n",
       "      <td>-0.006038</td>\n",
       "      <td>0.006038</td>\n",
       "      <td>TP53 regulated inhibitor of apoptosis 1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            symbol    weight       abs  \\\n",
       "entrez_id                                \n",
       "4193          MDM2 -0.008032  0.008032   \n",
       "64782          AEN -0.007234  0.007234   \n",
       "81626      SHCBP1L -0.006931  0.006931   \n",
       "23729         SHPK -0.006899  0.006899   \n",
       "23612       PHLDA3 -0.006629  0.006629   \n",
       "1643          DDB2 -0.006350  0.006350   \n",
       "132671     SPATA18 -0.006340  0.006340   \n",
       "1819          DRG2 -0.006139  0.006139   \n",
       "60401        EDA2R -0.006048  0.006048   \n",
       "51499       TRIAP1 -0.006038  0.006038   \n",
       "\n",
       "                                                 description  \n",
       "entrez_id                                                     \n",
       "4193                                     MDM2 proto-oncogene  \n",
       "64782                           apoptosis enhancing nuclease  \n",
       "81626              SHC binding and spindle associated 1 like  \n",
       "23729                                      sedoheptulokinase  \n",
       "23612      pleckstrin homology like domain family A member 3  \n",
       "1643                   damage specific DNA binding protein 2  \n",
       "132671                         spermatogenesis associated 18  \n",
       "1819         developmentally regulated GTP binding protein 2  \n",
       "60401                              ectodysplasin A2 receptor  \n",
       "51499                TP53 regulated inhibitor of apoptosis 1  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca_for_expression = (\n",
    "    final_pipelines['expressions']\n",
    "    .named_steps['features']\n",
    "    .get_params()['expressions__pca']\n",
    "    )\n",
    "classifier_for_expression = (\n",
    "    final_pipelines['expressions']\n",
    "    .named_steps['classify']\n",
    "    )\n",
    "gene_coefficients_for_expression = get_genes_coefficients(\n",
    "    pca_object=pca_for_expression,\n",
    "    classifier_object=classifier_for_expression,\n",
    "    expression_df=expression_df,\n",
    "    expression_genes_df=expression_genes_df\n",
    "    )\n",
    "gene_coefficients_for_expression.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Investigate the predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "predict_df = pd.DataFrame()\n",
    "for model, pipeline in final_pipelines.items():\n",
    "    df = pd.DataFrame.from_items([\n",
    "        ('feature_set', model),\n",
    "        ('sample_id', X.index),\n",
    "        ('test_set', X.index.isin(X_test.index).astype(int)),\n",
    "        ('status', y),\n",
    "        ('decision_function', pipeline.decision_function(X)),\n",
    "        ('probability', pipeline.predict_proba(X)[:, 1])\n",
    "    ])    \n",
    "    predict_df = predict_df.append(df)\n",
    "\n",
    "predict_df['probability_str'] = predict_df['probability'].apply('{:.1%}'.format)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature_set</th>\n",
       "      <th>sample_id</th>\n",
       "      <th>test_set</th>\n",
       "      <th>status</th>\n",
       "      <th>decision_function</th>\n",
       "      <th>probability</th>\n",
       "      <th>probability_str</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sample_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TCGA-L5-A4OH-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-L5-A4OH-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6.500973</td>\n",
       "      <td>0.998500</td>\n",
       "      <td>99.9%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-22-4591-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-22-4591-01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>6.113978</td>\n",
       "      <td>0.997793</td>\n",
       "      <td>99.8%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-21-1081-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-21-1081-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.424201</td>\n",
       "      <td>0.995611</td>\n",
       "      <td>99.6%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-EI-6513-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-EI-6513-01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>5.327085</td>\n",
       "      <td>0.995165</td>\n",
       "      <td>99.5%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-46-3765-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-46-3765-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5.156173</td>\n",
       "      <td>0.994269</td>\n",
       "      <td>99.4%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-L5-A8NR-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-L5-A8NR-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.983415</td>\n",
       "      <td>0.993196</td>\n",
       "      <td>99.3%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-CV-7261-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-CV-7261-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.839534</td>\n",
       "      <td>0.992151</td>\n",
       "      <td>99.2%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-B6-A0I6-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-B6-A0I6-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.754566</td>\n",
       "      <td>0.991461</td>\n",
       "      <td>99.1%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-BR-8373-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-BR-8373-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.615660</td>\n",
       "      <td>0.990201</td>\n",
       "      <td>99.0%</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TCGA-20-1684-01</th>\n",
       "      <td>full</td>\n",
       "      <td>TCGA-20-1684-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4.593257</td>\n",
       "      <td>0.989982</td>\n",
       "      <td>99.0%</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                feature_set        sample_id  test_set  status  \\\n",
       "sample_id                                                        \n",
       "TCGA-L5-A4OH-01        full  TCGA-L5-A4OH-01         0       0   \n",
       "TCGA-22-4591-01        full  TCGA-22-4591-01         1       0   \n",
       "TCGA-21-1081-01        full  TCGA-21-1081-01         0       0   \n",
       "TCGA-EI-6513-01        full  TCGA-EI-6513-01         1       0   \n",
       "TCGA-46-3765-01        full  TCGA-46-3765-01         0       0   \n",
       "TCGA-L5-A8NR-01        full  TCGA-L5-A8NR-01         0       0   \n",
       "TCGA-CV-7261-01        full  TCGA-CV-7261-01         0       0   \n",
       "TCGA-B6-A0I6-01        full  TCGA-B6-A0I6-01         0       0   \n",
       "TCGA-BR-8373-01        full  TCGA-BR-8373-01         0       0   \n",
       "TCGA-20-1684-01        full  TCGA-20-1684-01         0       0   \n",
       "\n",
       "                 decision_function  probability probability_str  \n",
       "sample_id                                                        \n",
       "TCGA-L5-A4OH-01           6.500973     0.998500           99.9%  \n",
       "TCGA-22-4591-01           6.113978     0.997793           99.8%  \n",
       "TCGA-21-1081-01           5.424201     0.995611           99.6%  \n",
       "TCGA-EI-6513-01           5.327085     0.995165           99.5%  \n",
       "TCGA-46-3765-01           5.156173     0.994269           99.4%  \n",
       "TCGA-L5-A8NR-01           4.983415     0.993196           99.3%  \n",
       "TCGA-CV-7261-01           4.839534     0.992151           99.2%  \n",
       "TCGA-B6-A0I6-01           4.754566     0.991461           99.1%  \n",
       "TCGA-BR-8373-01           4.615660     0.990201           99.0%  \n",
       "TCGA-20-1684-01           4.593257     0.989982           99.0%  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Top predictions amongst negatives (potential hidden responders to a targeted cancer therapy)\n",
    "(predict_df\n",
    "    .sort_values('decision_function', ascending=False)\n",
    "    .query(\"status == 0 and feature_set == 'full'\")\n",
    "    .head(10)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsIAAAHRCAYAAABguN9xAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x34d23cf8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (3720817)>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_df['status_'] = predict_df['status'].map(\n",
    "    lambda x: 'negative' if x == 0 else 'positive')\n",
    "\n",
    "(gg.ggplot(predict_df, gg.aes(x='probability', \n",
    "                              fill='status_'))\n",
    " + gg.geom_density(alpha=0.6)\n",
    " + gg.facet_wrap('~feature_set', ncol=1)\n",
    " + gg.labs(x='probability', y='density')\n",
    " + gg.guides(fill=gg.guide_legend(title=\"\"))\n",
    " + theme_cognoma())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}