6. Evaluate Model (Figure 3).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "33158c23-41d1-44f0-9895-a8f5d57c8bd0",
   "metadata": {},
   "source": [
    "# Cross-Validation and AUCs of LRM per cell line idea (unique drug-indication pairs)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1a82511-43d1-4741-aaf5-a4225babb4af",
   "metadata": {},
   "source": [
    "## *1. Import the DRH and clinical trials datasets*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "13214de2-f5ce-450d-a0dd-00ce6449701d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import data manipulation tools\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f7b44556-1be9-409b-947b-d3960d638406",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>drug1 tas</th>\n",
       "      <th>drug2</th>\n",
       "      <th>drug2 tas</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "      <th>known ind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>midostaurin</td>\n",
       "      <td>0.799386</td>\n",
       "      <td>altrenogest</td>\n",
       "      <td>0.812819</td>\n",
       "      <td>0.908978</td>\n",
       "      <td>Estrus</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>altrenogest</td>\n",
       "      <td>0.812819</td>\n",
       "      <td>midostaurin</td>\n",
       "      <td>0.799386</td>\n",
       "      <td>0.908978</td>\n",
       "      <td>Leukemia, Myeloid, Acute</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ixazomib-citrate</td>\n",
       "      <td>0.747696</td>\n",
       "      <td>ixazomib</td>\n",
       "      <td>0.791817</td>\n",
       "      <td>0.889649</td>\n",
       "      <td>Multiple Myeloma</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ixazomib</td>\n",
       "      <td>0.791817</td>\n",
       "      <td>ixazomib-citrate</td>\n",
       "      <td>0.747696</td>\n",
       "      <td>0.889649</td>\n",
       "      <td>Multiple Myeloma</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>carfilzomib</td>\n",
       "      <td>0.763093</td>\n",
       "      <td>ixazomib-citrate</td>\n",
       "      <td>0.747696</td>\n",
       "      <td>0.839315</td>\n",
       "      <td>Multiple Myeloma</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              drug1  drug1 tas             drug2  drug2 tas  spearman corr  \\\n",
       "0       midostaurin   0.799386       altrenogest   0.812819       0.908978   \n",
       "1       altrenogest   0.812819       midostaurin   0.799386       0.908978   \n",
       "2  ixazomib-citrate   0.747696          ixazomib   0.791817       0.889649   \n",
       "3          ixazomib   0.791817  ixazomib-citrate   0.747696       0.889649   \n",
       "4       carfilzomib   0.763093  ixazomib-citrate   0.747696       0.839315   \n",
       "\n",
       "                 indication  known ind  \n",
       "0                    Estrus      False  \n",
       "1  Leukemia, Myeloid, Acute      False  \n",
       "2          Multiple Myeloma       True  \n",
       "3          Multiple Myeloma       True  \n",
       "4          Multiple Myeloma       True  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### initialize dictionaries to store the DRH and clinical datasets\n",
    "DRH_data = {}\n",
    "clin_data = {}\n",
    "### store a list of the cell lines used\n",
    "cell_lines = [\"A375\", \"MCF7\", \"PC3\"]\n",
    "\n",
    "### iterate through each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## store the DRH and clinical datasets in their respective dictionaries\n",
    "    file_path = \"./DRH_clin_data/DRH_data/\" + cell_line + \"_unique_ind.txt\"\n",
    "    DRH_data[cell_line] = pd.read_csv(file_path)\n",
    "    \n",
    "    file_path = \"./DRH_clin_data/clin_data/\" + cell_line + \"_unique_ind.txt\"\n",
    "    clin_data[cell_line] = pd.read_csv(file_path)\n",
    "\n",
    "DRH_data[\"MCF7\"].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ef146af-be5a-46d1-a29c-01f3d47efe19",
   "metadata": {},
   "source": [
    "## *2. Generate the shared data filtered by TAS cutoff*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "dc25ad15-1483-4aa4-bd14-3de9d70d0753",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to select only drug-indication pairs above TAS cutoff\n",
    "def gen_tas_datasets(cell_line_data):\n",
    "    ## initialize a dictionary to store all the filtered datasets at each tas cutoff\n",
    "    tas_data = {}\n",
    "    ## iterate through each tas cutoff\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        ## store only drug-indication pairs below tas cutoff in the dictionary\n",
    "        drug1_above_tas = cell_line_data[\"drug1 tas\"] > tas_cutoff\n",
    "        drug2_above_tas = cell_line_data[\"drug2 tas\"] > tas_cutoff\n",
    "        tas_data[tas_cutoff] = cell_line_data.loc[drug1_above_tas & drug2_above_tas]\n",
    "        \n",
    "    return tas_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "11e9894c-e6fa-46a6-9f36-4b5b3608fa16",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to get the intersection dataset for a cell line\n",
    "def get_shared_data(DRH_data, clin_data, cell_line):\n",
    "    ## get the DRH and clinical trials data for that cell line\n",
    "    DRH = DRH_data[cell_line]\n",
    "    clin = clin_data[cell_line]\n",
    "    ## create a dataframe containing the rows in both datasets\n",
    "    cols = DRH.columns.tolist()[:-1]\n",
    "    shared_data = pd.merge(DRH, clin, on=cols, suffixes=(\"_DRH\", \"_clin\"), how=\"inner\")\n",
    "    \n",
    "    return shared_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "31044e4d-b936-4980-b4db-3d64a2a675c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>drug1 tas</th>\n",
       "      <th>drug2</th>\n",
       "      <th>drug2 tas</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "      <th>known ind_DRH</th>\n",
       "      <th>known ind_clin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ixazomib</td>\n",
       "      <td>0.791817</td>\n",
       "      <td>metronidazole</td>\n",
       "      <td>0.641509</td>\n",
       "      <td>0.770292</td>\n",
       "      <td>Rosacea</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5-aminolevulinic-acid</td>\n",
       "      <td>0.618329</td>\n",
       "      <td>sitagliptin</td>\n",
       "      <td>0.587846</td>\n",
       "      <td>0.750782</td>\n",
       "      <td>Diabetes Mellitus</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sitagliptin</td>\n",
       "      <td>0.587846</td>\n",
       "      <td>5-aminolevulinic-acid</td>\n",
       "      <td>0.618329</td>\n",
       "      <td>0.750782</td>\n",
       "      <td>Glioma</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>sitagliptin</td>\n",
       "      <td>0.587846</td>\n",
       "      <td>5-aminolevulinic-acid</td>\n",
       "      <td>0.618329</td>\n",
       "      <td>0.750782</td>\n",
       "      <td>Keratosis, Actinic</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>macitentan</td>\n",
       "      <td>0.561786</td>\n",
       "      <td>phenylbutyrate</td>\n",
       "      <td>0.508335</td>\n",
       "      <td>0.741032</td>\n",
       "      <td>Urea Cycle Disorders, Inborn</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   drug1  drug1 tas                  drug2  drug2 tas  \\\n",
       "0               ixazomib   0.791817          metronidazole   0.641509   \n",
       "1  5-aminolevulinic-acid   0.618329            sitagliptin   0.587846   \n",
       "2            sitagliptin   0.587846  5-aminolevulinic-acid   0.618329   \n",
       "3            sitagliptin   0.587846  5-aminolevulinic-acid   0.618329   \n",
       "4             macitentan   0.561786         phenylbutyrate   0.508335   \n",
       "\n",
       "   spearman corr                    indication  known ind_DRH  known ind_clin  \n",
       "0       0.770292                       Rosacea          False           False  \n",
       "1       0.750782             Diabetes Mellitus          False           False  \n",
       "2       0.750782                        Glioma          False           False  \n",
       "3       0.750782            Keratosis, Actinic          False           False  \n",
       "4       0.741032  Urea Cycle Disorders, Inborn          False           False  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### initialize a dictionary to store all the shared data across all cell lines\n",
    "shared_data = {}\n",
    "\n",
    "### store a list of all tas cutoffs of interest\n",
    "tas_cutoffs = [0, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5]\n",
    "\n",
    "### iterate through each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## get the rows present in both DRH and clinical trials\n",
    "    cell_line_data = get_shared_data(DRH_data, clin_data, cell_line)\n",
    "    ## generate the store the tas datasets for this shared data\n",
    "    shared_data[cell_line] = gen_tas_datasets(cell_line_data)\n",
    "    \n",
    "shared_data[\"MCF7\"][0.5].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f54cc8e-bcbe-499e-b68f-3f1ed0573e8b",
   "metadata": {},
   "source": [
    "## *3. Generate the training and testing datasets* "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "bd88dc51-2e29-4ab7-8224-0801ed742c9e",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to split the neithers while maintaining same percentage of 0s for tas cutoff\n",
    "def split_neithers(shared_data):\n",
    "    ## store whether the drug-indication pairs are in DRH and clinical trials independently\n",
    "    in_DRH = shared_data[\"known ind_DRH\"]\n",
    "    in_clin = shared_data[\"known ind_clin\"]\n",
    "    ## get the drug-indications that are only in clinical trials\n",
    "    only_in_clin = shared_data.loc[~in_DRH & in_clin]\n",
    "    ## get the dataframe containing the other drug-indications\n",
    "    DRH = shared_data.drop(only_in_clin.index)\n",
    "    ## get the drug-indications that are not in either DRH or clinical trials\n",
    "    neither_data = shared_data.loc[~in_DRH & ~in_clin]\n",
    "    \n",
    "    ## get the relevant information about the datasets\n",
    "    DRH_true = DRH[\"known ind_DRH\"].sum()\n",
    "    clin_true = only_in_clin[\"known ind_clin\"].sum()\n",
    "    total_neithers = len(neither_data)\n",
    "    ## calculate the number of neithers to add to clinical trials\n",
    "    num2add = (clin_true*total_neithers)/(DRH_true + clin_true)\n",
    "    clin_neither = neither_data.sample(n=num2add, random_state=42)\n",
    "    \n",
    "    return only_in_clin, DRH, clin_neither"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d0bb01ea-a0ef-4fb5-9089-8654284b4e8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate training dataset for one cell line\n",
    "def gen_train_test_data(shared_data, cell_line):\n",
    "    ## initialize dictionaries to store the training and testing datasets\n",
    "    train = {}\n",
    "    test = {}\n",
    "    \n",
    "    ## iterate through each tas cutoff\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        ## store the dataset for that cell line and tas cutoff\n",
    "        shared_tas = shared_data[cell_line][tas_cutoff]\n",
    "        ## randomly split the neithers while maintaing percent trues\n",
    "        only_in_clin, DRH, clin_neither = split_neithers(shared_tas)\n",
    "        ## generate the training datasets\n",
    "        train_tas = DRH.drop(clin_neither.index)\n",
    "        train_tas = train_tas.drop([\"known ind_clin\"], axis=1)\n",
    "        train_tas.rename(columns={\"known ind_DRH\": \"known ind\"}, inplace=True)\n",
    "        ## store the training dataset in the dictionary for all tas cutoffs\n",
    "        train[tas_cutoff] = train_tas\n",
    "\n",
    "        ## generate the testing dataset\n",
    "        test_tas = pd.concat([only_in_clin, clin_neither])\n",
    "        test_tas = test_tas.drop([\"known ind_DRH\"], axis=1)\n",
    "        test_tas.rename(columns={\"known ind_clin\": \"known ind\"}, inplace=True)\n",
    "        ## generate the separate datasets for each TAS cutoff\n",
    "        test[tas_cutoff] = test_tas\n",
    "    \n",
    "    return train, test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "33777f31-ff71-4e05-a547-966b4af61b7d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>drug1 tas</th>\n",
       "      <th>drug2</th>\n",
       "      <th>drug2 tas</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "      <th>known ind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5-aminolevulinic-acid</td>\n",
       "      <td>0.618329</td>\n",
       "      <td>sitagliptin</td>\n",
       "      <td>0.587846</td>\n",
       "      <td>0.750782</td>\n",
       "      <td>Diabetes Mellitus</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>valrubicin</td>\n",
       "      <td>0.794730</td>\n",
       "      <td>doxorubicin</td>\n",
       "      <td>0.732233</td>\n",
       "      <td>0.725946</td>\n",
       "      <td>Urinary Bladder Neoplasms</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>doxorubicin</td>\n",
       "      <td>0.732233</td>\n",
       "      <td>valrubicin</td>\n",
       "      <td>0.794730</td>\n",
       "      <td>0.725946</td>\n",
       "      <td>Urinary Bladder Neoplasms</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>lurasidone</td>\n",
       "      <td>0.566841</td>\n",
       "      <td>palbociclib</td>\n",
       "      <td>0.575843</td>\n",
       "      <td>0.712300</td>\n",
       "      <td>Breast Neoplasms</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>mebendazole</td>\n",
       "      <td>0.684513</td>\n",
       "      <td>podophyllotoxin</td>\n",
       "      <td>0.704934</td>\n",
       "      <td>0.699762</td>\n",
       "      <td>Condylomata Acuminata</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    drug1  drug1 tas            drug2  drug2 tas  \\\n",
       "1   5-aminolevulinic-acid   0.618329      sitagliptin   0.587846   \n",
       "12             valrubicin   0.794730      doxorubicin   0.732233   \n",
       "14            doxorubicin   0.732233       valrubicin   0.794730   \n",
       "19             lurasidone   0.566841      palbociclib   0.575843   \n",
       "31            mebendazole   0.684513  podophyllotoxin   0.704934   \n",
       "\n",
       "    spearman corr                 indication  known ind  \n",
       "1        0.750782          Diabetes Mellitus      False  \n",
       "12       0.725946  Urinary Bladder Neoplasms       True  \n",
       "14       0.725946  Urinary Bladder Neoplasms       True  \n",
       "19       0.712300           Breast Neoplasms      False  \n",
       "31       0.699762      Condylomata Acuminata      False  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### initialize a dictionary to store the training datasets for each cell line\n",
    "train_data = {}\n",
    "test_data = {}\n",
    "\n",
    "### iterate through each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## generate and store all the tas datasets for that cell line\n",
    "    train, test = gen_train_test_data(shared_data, cell_line)\n",
    "    train_data[cell_line] = train\n",
    "    test_data[cell_line] = test\n",
    "    \n",
    "train_data[\"MCF7\"][0.5].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c6d680eb-ed8e-401d-8426-21fd0ac6a045",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>drug1 tas</th>\n",
       "      <th>drug2</th>\n",
       "      <th>drug2 tas</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "      <th>known ind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>metronidazole</td>\n",
       "      <td>0.641509</td>\n",
       "      <td>sitagliptin</td>\n",
       "      <td>0.587846</td>\n",
       "      <td>0.727970</td>\n",
       "      <td>Diabetes Mellitus</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>belinostat</td>\n",
       "      <td>0.672252</td>\n",
       "      <td>romidepsin</td>\n",
       "      <td>0.795148</td>\n",
       "      <td>0.692400</td>\n",
       "      <td>Lymphoma, T-Cell, Cutaneous</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>romidepsin</td>\n",
       "      <td>0.795148</td>\n",
       "      <td>belinostat</td>\n",
       "      <td>0.672252</td>\n",
       "      <td>0.692400</td>\n",
       "      <td>Lymphoma, T-Cell, Peripheral</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>mepacrine</td>\n",
       "      <td>0.806568</td>\n",
       "      <td>crizotinib</td>\n",
       "      <td>0.670493</td>\n",
       "      <td>0.676249</td>\n",
       "      <td>Carcinoma, Non-Small-Cell Lung</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>amsacrine</td>\n",
       "      <td>0.660436</td>\n",
       "      <td>mitoxantrone</td>\n",
       "      <td>0.681046</td>\n",
       "      <td>0.610274</td>\n",
       "      <td>Leukemia, Myeloid, Acute</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             drug1  drug1 tas         drug2  drug2 tas  spearman corr  \\\n",
       "5    metronidazole   0.641509   sitagliptin   0.587846       0.727970   \n",
       "33      belinostat   0.672252    romidepsin   0.795148       0.692400   \n",
       "34      romidepsin   0.795148    belinostat   0.672252       0.692400   \n",
       "49       mepacrine   0.806568    crizotinib   0.670493       0.676249   \n",
       "144      amsacrine   0.660436  mitoxantrone   0.681046       0.610274   \n",
       "\n",
       "                         indication  known ind  \n",
       "5                 Diabetes Mellitus       True  \n",
       "33      Lymphoma, T-Cell, Cutaneous       True  \n",
       "34     Lymphoma, T-Cell, Peripheral       True  \n",
       "49   Carcinoma, Non-Small-Cell Lung       True  \n",
       "144        Leukemia, Myeloid, Acute       True  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data[\"MCF7\"][0.5].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b8b7e58-bc11-40d2-98ac-a7b43f1ff041",
   "metadata": {},
   "source": [
    "## *4. Evaluate Model*"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58bdba4b-d476-4966-9620-61ac620f19c7",
   "metadata": {},
   "source": [
    "### a. Cross-Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "acb0d4e8-54a1-41d1-b99d-871b59a2aca9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import the data modeling and evaluation tools\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import StratifiedKFold, cross_validate\n",
    "from sklearn.metrics import make_scorer, roc_curve, roc_auc_score, precision_score, recall_score, f1_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c4fa6461-c5d4-4581-9e63-8f13cd8de0a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialize the parameters of the logistic regression model\n",
    "lrm = LogisticRegression(class_weight=\"balanced\", solver=\"lbfgs\", random_state=42)\n",
    "\n",
    "# intialize the parameters of the Stratified KFold\n",
    "skf = StratifiedKFold(n_splits=3, random_state=42, shuffle=True)\n",
    "\n",
    "# create a dictionary to define the evaluation metrics of interest\n",
    "metrics = {\"AUC\": make_scorer(roc_auc_score),\n",
    "           \"precision\": make_scorer(precision_score),\n",
    "           \"recall\": make_scorer(recall_score),\n",
    "           \"F1 score\": make_scorer(f1_score)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2a0fb141-3f76-4af2-a68e-8803d254a199",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to perform cross-validation for one cell line\n",
    "def gen_cv_scores(train_data, cell_line):\n",
    "    ## initialize dataframe to store the evaluation metrics for training and validations datasets\n",
    "    cv_scores = pd.DataFrame()\n",
    "    ## iterate through the tas cutoffs\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        # store the tas dataset\n",
    "        tas_data = train_data[cell_line][tas_cutoff]\n",
    "        # define the inputs and outputs to train the model\n",
    "        train_cols = [\"spearman corr\"]\n",
    "        X, y = tas_data[train_cols], tas_data[\"known ind\"]\n",
    "        # perform stratified kfold cross-validation on training and validation datasets\n",
    "        tas_cv = cross_validate(lrm, X, y, cv=skf, scoring=metrics, return_train_score=True)\n",
    "        # add the results to the dataframe\n",
    "        tas_scores = pd.DataFrame(tas_cv).drop([\"fit_time\", \"score_time\"], axis=1)\n",
    "        cv_scores = cv_scores.append(tas_scores.mean(), ignore_index=True)\n",
    "    ## add the tas cutoffs\n",
    "    cv_scores.insert(0, \"TAS cutoff\", tas_cutoffs)\n",
    "    \n",
    "    ## rename columns\n",
    "    cv_scores = cv_scores.rename(columns={\n",
    "    \"test_AUC\": \"val_AUC\",\n",
    "    \"test_F1 score\": \"val_F1 score\",\n",
    "    \"test_precision\": \"val_precision\",\n",
    "    \"test_recall\": \"val_recall\"})\n",
    "    ## reorder columns\n",
    "    cols = [\"TAS cutoff\", \"train_AUC\", \"train_F1 score\", \"train_precision\", \"train_recall\", \n",
    "            \"val_AUC\", \"val_F1 score\", \"val_precision\", \"val_recall\"]\n",
    "    cv_scores = cv_scores[cols]\n",
    "    \n",
    "    return cv_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "3041d639-3d1c-4982-ac5f-ae043e5404bf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAS cutoff</th>\n",
       "      <th>train_AUC</th>\n",
       "      <th>train_F1 score</th>\n",
       "      <th>train_precision</th>\n",
       "      <th>train_recall</th>\n",
       "      <th>val_AUC</th>\n",
       "      <th>val_F1 score</th>\n",
       "      <th>val_precision</th>\n",
       "      <th>val_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.615249</td>\n",
       "      <td>0.094405</td>\n",
       "      <td>0.051072</td>\n",
       "      <td>0.623329</td>\n",
       "      <td>0.616155</td>\n",
       "      <td>0.094663</td>\n",
       "      <td>0.051227</td>\n",
       "      <td>0.622742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.20</td>\n",
       "      <td>0.625780</td>\n",
       "      <td>0.116637</td>\n",
       "      <td>0.063945</td>\n",
       "      <td>0.663082</td>\n",
       "      <td>0.627616</td>\n",
       "      <td>0.117479</td>\n",
       "      <td>0.064395</td>\n",
       "      <td>0.673611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.25</td>\n",
       "      <td>0.657541</td>\n",
       "      <td>0.143552</td>\n",
       "      <td>0.080150</td>\n",
       "      <td>0.694444</td>\n",
       "      <td>0.661763</td>\n",
       "      <td>0.144099</td>\n",
       "      <td>0.080635</td>\n",
       "      <td>0.690476</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.30</td>\n",
       "      <td>0.785755</td>\n",
       "      <td>0.276081</td>\n",
       "      <td>0.164581</td>\n",
       "      <td>0.859259</td>\n",
       "      <td>0.789437</td>\n",
       "      <td>0.276741</td>\n",
       "      <td>0.165221</td>\n",
       "      <td>0.866667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.35</td>\n",
       "      <td>0.798413</td>\n",
       "      <td>0.340721</td>\n",
       "      <td>0.209626</td>\n",
       "      <td>0.910714</td>\n",
       "      <td>0.801423</td>\n",
       "      <td>0.349206</td>\n",
       "      <td>0.216399</td>\n",
       "      <td>0.916667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.40</td>\n",
       "      <td>0.863636</td>\n",
       "      <td>0.426002</td>\n",
       "      <td>0.271193</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.853535</td>\n",
       "      <td>0.417647</td>\n",
       "      <td>0.265873</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.45</td>\n",
       "      <td>0.844309</td>\n",
       "      <td>0.456443</td>\n",
       "      <td>0.295895</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.845635</td>\n",
       "      <td>0.480556</td>\n",
       "      <td>0.326923</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.50</td>\n",
       "      <td>0.868708</td>\n",
       "      <td>0.487179</td>\n",
       "      <td>0.322222</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.861722</td>\n",
       "      <td>0.490476</td>\n",
       "      <td>0.327778</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   TAS cutoff  train_AUC  train_F1 score  train_precision  train_recall  \\\n",
       "0        0.00   0.615249        0.094405         0.051072      0.623329   \n",
       "1        0.20   0.625780        0.116637         0.063945      0.663082   \n",
       "2        0.25   0.657541        0.143552         0.080150      0.694444   \n",
       "3        0.30   0.785755        0.276081         0.164581      0.859259   \n",
       "4        0.35   0.798413        0.340721         0.209626      0.910714   \n",
       "5        0.40   0.863636        0.426002         0.271193      1.000000   \n",
       "6        0.45   0.844309        0.456443         0.295895      1.000000   \n",
       "7        0.50   0.868708        0.487179         0.322222      1.000000   \n",
       "\n",
       "    val_AUC  val_F1 score  val_precision  val_recall  \n",
       "0  0.616155      0.094663       0.051227    0.622742  \n",
       "1  0.627616      0.117479       0.064395    0.673611  \n",
       "2  0.661763      0.144099       0.080635    0.690476  \n",
       "3  0.789437      0.276741       0.165221    0.866667  \n",
       "4  0.801423      0.349206       0.216399    0.916667  \n",
       "5  0.853535      0.417647       0.265873    1.000000  \n",
       "6  0.845635      0.480556       0.326923    1.000000  \n",
       "7  0.861722      0.490476       0.327778    1.000000  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### redefine the tas cutoffs\n",
    "tas_cutoffs = [0, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5]\n",
    "\n",
    "### initialize a dictionary to store results from cross-validation for all cell lines\n",
    "cv_scores = {}\n",
    "\n",
    "### iterate through each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## generate and store results from cross-validation for that cell line in dictionary\n",
    "    cv_scores[cell_line] = gen_cv_scores(train_data, cell_line)\n",
    "cv_scores[\"MCF7\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f15d458d-7eb2-49e6-bc9d-7073cd9f52c9",
   "metadata": {},
   "source": [
    "### b. Cross-Validation ROC Curves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7208aae1-8cb9-49f9-b0d3-75413ce76f27",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import data visualization tools\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "684727da-9624-48ed-9264-834b824e0b6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# set theme of figures\n",
    "sns.set_style(\"whitegrid\")\n",
    "\n",
    "# set the font sizes of plots\n",
    "plt.rcParams[\"axes.titlesize\"] = 17 # title \n",
    "plt.rcParams[\"axes.labelsize\"] = 15 # X and y axis labels\n",
    "plt.rcParams[\"xtick.labelsize\"] = 15  # X tick labels\n",
    "plt.rcParams[\"ytick.labelsize\"] = 15  # Y tick labels\n",
    "plt.rcParams[\"legend.fontsize\"] = 15 # legend"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "1abc046a-1c71-41de-8a17-1922df81343f",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to compare training to validation scores across all cell lines and tas cutoff\n",
    "def compare_metric(metric_scores, metric1, metric2, metric):\n",
    "    ## create a dataframe with the metric\n",
    "    scores = pd.DataFrame()\n",
    "    scores[\"TAS cutoff\"] = tas_cutoffs\n",
    "    \n",
    "    ## initialize a figure\n",
    "    #plt.style.use(\"seaborn-whitegrid\")\n",
    "    fig, ax = plt.subplots(1, 2, figsize=(20,10))\n",
    "    ## iterate through each cell line\n",
    "    for model in metric_scores.keys():\n",
    "        # store the data for that cell line\n",
    "        model_scores = metric_scores[model]\n",
    "        # create a line plot showing the results on training datasets\n",
    "        ax[0].plot(model_scores[\"TAS cutoff\"], model_scores[metric1], marker=\"o\", label=model)\n",
    "        # create a line plot showing the results on validation datasets\n",
    "        ax[1].plot(model_scores[\"TAS cutoff\"], model_scores[metric2], marker=\"o\", label=model)\n",
    "        \n",
    "        # add the data to the dataframe\n",
    "        scores[model + \"_\" + metric1] = model_scores[metric1]\n",
    "        scores[model + \"_\" + metric2] = model_scores[metric2]\n",
    "        \n",
    "    ## add titles to each plot\n",
    "    ax[0].set_title(\"ROC Curves of Individual Cell Line Models \\non Training Datasets\")\n",
    "    ax[1].set_title(\"ROC Curves of Individual Cell Line Models \\non Validation Datasets\")\n",
    "    \n",
    "    ## add labels and axis limits to each plot\n",
    "    for i in range(2):\n",
    "        ax[i].set_xlabel(\"TAS cutoff\")\n",
    "        ax[i].set_ylabel(metric)\n",
    "        ax[i].set_ylim(0.4, 1)\n",
    "        ax[i].legend()\n",
    "        \n",
    "    # show plot\n",
    "    plt.show()\n",
    "    \n",
    "    return scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "1c5a82a7-3a32-4757-ac6b-86b4b64fb9ae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJ8AAAJ/CAYAAADWGXLJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi41LCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvSM8oowAAIABJREFUeJzs3XlcTfnjP/BXpVKSLGWJ7F20iZS0F4WSYSjGnmEw1jHGnhhjG8YY65cxmI+oJFsh+zaGsW+NLYbKIJK03+r8/vDrjOvedFtv9Ho+Hh7z6X3e533e57hXr8/7vM/7qAmCIICIiIiIiIiIiKgMqKu6A0RERERERERE9Oni4BMREREREREREZUZDj4REREREREREVGZ4eATERERERERERGVGQ4+ERERERERERFRmeHgExERERERERERlRkOPhFVAnFxcRg5ciRsbW0hkUiwefPmcj3++fPnIZFIcP78ebFs2rRpcHd3L1Z77u7umDZtWqH14uPjIZFIEBERUazjKCsiIgISiQTx8fFlepwPkUgkWLlypfizomteEoMGDcKgQYNKpa2PSUk+pyXZl4iIPg3MYMxgJcUMVr77UtmpouoOUMUSERGB6dOniz+rq6ujVq1a6NChA8aPH49mzZop3O/WrVv49ddfceHCBSQnJ8PAwAA2NjYYPnw4LCwsFO4jlUqxc+dO7Nu3D/fu3UN6ejpq166N9u3bw8/PDx07dlSqz6dPn8b27dtx7do1vH79GtWqVYOZmRl69OgBX19faGhoFP1CfGJmzJiBu3fvYsyYMahZsybMzc0LrOvu7g4TE5NyD0eVSXp6OoKDgxEdHY2HDx8iKysLdevWhZ2dHQYMGAAzM7Ny60t8fDw8PDwwYcIEjBkzptyOWxQrV67EqlWrAAD79+9H8+bN5er06tULMTExMDExweHDh8u7i0REJcYM9mliBqtYmMGKhhmMShMHn0ihMWPGoEmTJpBKpbh9+zbCwsJw7tw5REZGok6dOjJ1d+zYgcDAQBgZGaFPnz5o2LAhnjx5goiICPj5+WHevHno27evzD6vX7/GV199hStXrsDBwQGjRo2Cvr4+nj59isOHD2PIkCHYvn072rVrV2AfBUHAvHnzsG3bNrRs2RL9+/dHvXr1kJKSgj/++APTp0/Hs2fPMGrUqDK5Rh+L7OxsXLx4EQMHDsTQoUNV3R3R999/D0EQirXvwYMHoaamVso9Kh9PnjzB8OHD8c8//8DDwwO+vr7Q0dHB48ePcfDgQURERODEiROoV6+eqrsqY+PGjaruArS1tbFnzx588803MuX37t1DTEwMtLW1VdQzIqLSwwz26WAGq1iYwYqPGYxKAwefSCEHBwfY2NiIPzdv3hxBQUHYvXs3vvzyS7H8+vXrmDNnDiwsLPDbb79BT09P3BYQEICAgAAEBQVBIpHA0tJS3DZjxgxcvXoVP/74I3x9fWWOPXbsWERERKBKlQ9/PH///Xds27YN/fr1w5w5c6Cu/t9TpAEBAbh27RpiY2OLfQ3elZmZCS0tLZljfCxevXqFvLw86Ovrq7orMjQ1NYu9r5aWVin2pPzk5uZi7NixSEhIwG+//QZ7e3uZ7RMnTqwQAUORinDNXV1dERkZiUmTJskE3z179sDQ0BAmJiZITExUYQ+JiEqOGUwWM1jpYwZjBisqZjAqDR/fv+KkEvkh6PHjxzLlq1atgiAIWLJkiUzoAQA9PT0sXrwYeXl54nRNALhx4waOHDmCXr16yYWefL1795YJSu/LysrC2rVr0aRJE8yaNUthILGyskLv3r0BfPi58/efXc9/dvzs2bNYuHAhHB0d0bZtW8TExEAikWDr1q1ybfz7779o1aoVfv75Z7EsNTUVS5Ysgbu7O8zNzeHq6orFixcjMzNTZt9z585h0KBBsLOzg6WlJdzc3PDtt98iPT29wPPPd+nSJQwZMgTW1tawtrbGkCFDcOXKFXH7ypUr4ezsDODt35VEIoFEIim03XflX7s1a9Zg9+7d6Nq1K8zNzdGjRw/88ccfcvUfPHiA4cOHw8rKCg4ODli8eDGkUqlcvfefxe7Ro4f49/W+oUOHws3NTbxLp2i9gefPn2P8+PGwtraGra0tZsyYgdTU1EKPm0/RmgFHjx7FqFGj4OzsDHNzczg7O2POnDlISUkp4Gp92OHDh3Hr1i2MHDlSLvQAgIaGBkaOHClzx+3FixcIDAyEo6MjzM3N0aVLF/zf//0f8vLyitWH4np/vYGifi6U/T58iK+vLxISEnDx4kWxTBAEREZGwtvbW+HjHXl5ediwYQO8vLxgbm4OR0dHzJ07V+Hf4c6dO+Hl5QULCwv4+vri2LFjCvshCAK2bdsGX19fWFhYwM7ODpMmTUJCQkKh51CS7zsRVU7MYMxgzGDMYMxg/x2TGezjxZlPpJT8L3SNGjXEsszMTJw9exbt27dHkyZNFO7XtGlTtGvXDmfPnkVmZiaqVq2KI0eOAHj7fHBxXbp0Ca9evcLAgQNLdPfmQxYsWIBq1aphxIgRyMzMRPPmzdGkSRNERUVh4MCBMnWjoqIgCAJ69OgB4G0wGzJkCB49egQ/Pz80btwYd+7cwe+//4579+5hw4YNUFNTQ2xsLEaOHIkWLVpg9OjRqFatGp48eYJjx44hLS0Nurq6BfbvwoULGDZsGIyMjMRp7aGhoRg8eDA2b96M9u3bo0uXLjAwMMD8+fPRpUsXdOnSpdjX48iRI3j16hX69esHbW1t/P777xg7diyOHz8OAwMDAMDLly8xaNAgZGRkYOjQoahZsyb27NmDs2fPFtq+t7c3li9fjkePHqFx48ZieWJiIv766y8MGzaswGneWVlZGDp0KB4/fowvvvgCjRo1wqFDhzB16tRiny/w9hehhoYGBg4cCAMDA/z9998IDw/H3bt3sX379iK3V9TPflJSEvz8/JCdnQ0/Pz8YGRnh8uXL+Omnn5CQkIB58+YVuQ+lTZnPhbLfh8KYm5ujefPm2LNnDzp06ADg7aKe//77L3r27ImYmBi5fYKCghAaGgo3NzcMHjwY9+/fR0hICK5du4aQkBDxbuKuXbswY8YMmJmZ4dtvv8XLly/x3XffoX79+nJtfv/99wgJCUGPHj3Qv39/vHz5Elu3bkX//v2xe/du1KpVS2H/S/J9J6LKixmMGYwZjBlMEWYwZrCPDQefSKHU1FQkJSWJ6w0sWLAA6urq8PLyEuv8888/kEqlaNWq1Qfbat26NS5evIhHjx5BIpGI07CLevfnXffv3y9xG4XR0tLC1q1bZYKVj48PVq9ejSdPnqBBgwZieVRUFFq3bi0uwrd582bcvXsX4eHhMn2USCQIDAzE2bNn4eDggDNnziArKwu//vqrzD+WEyZMKLR/CxcuhI6ODsLCwsQ1IHr16oVu3bph4cKFCA8PR6tWrcTgI5FI0LNnz2Jfj0ePHiE6Olo8lp2dHT777DNERUVhwIABAIANGzbgxYsX2LJli7hYqb+/v1LHzQ8+kZGR+Prrr8XyAwcOIDc3VwyVioSFhSE2NhYLFizA559/DgD44osvMHjw4GKfLwAsW7YMOjo6MmVt27bFd999h0uXLqF9+/ZFai82NhbVq1eHsbGxUvV//vlnpKWlYd++fTAyMgIA9OvXD/Xq1cP69esxbNgwNG3atEh9KG3KfC6U/T4ow9fXFxs3bkRgYCC0tLSwZ88etGjRAm3atJGre/fuXYSGhsLHxwfLli0Ty1u0aIF58+Zhx44dGDBgAHJycrB06VI0a9YM27ZtQ9WqVQEAtra2GD58uMzf15UrVxAcHIzvv/8efn5+YrmXlxd69eqFLVu2YNKkSQr7XpLvOxFVHsxgzGDvYwZ7ixlMFjPYW8xgHw8+dkcKffXVV7C3t4ezszNGjhyJ9PR0LFu2TOYNHfnTad+f6v2+atWqydRXdr8PKY02CtO3b1+5O3re3t4QBAFRUVFi2YMHDxATEwNvb2+xbP/+/Wjbti0MDQ2RlJQk/unUqROAt9M+AYhrABw+fBi5ublK9y0xMRG3bt1Cz549ZRYfNTIygq+vL27cuIGXL18W/aQ/wNPTU+ZYrVu3hp6eHuLi4sSy48ePQyKRyLwlR0dHR+YXREEaNWqEtm3bYv/+/TLlUVFRaN68+QcD9okTJ2BgYCATsPLvlpVEfugRBEH8PwP5C7DeunWryO2lpqaK34fCCIKAgwcPwsXFBVWqVJH5HDk6OkIQhFJ7hW9JKPO5UPb7oIwePXrgzZs3OH78ODIzM3Ho0KECHx05ceIEgLfrj7yrb9++0NfXF7ffuHEDL168gL+/vxh6AMDR0REtWrSQ2Xf//v2oWrUq3NzcZM6ldu3aaNq06QfPpbjfdyKqXJjBmMHexwzGDKYIMxgz2MeGM59IoenTp8PU1BRpaWk4cOAAjhw5Irf4ZH7oUPRM97vS0tIA/BeA8vdLS0sr9gKM77ZRVkxMTOTKmjVrBjMzM+zfvx8jRowA8PYfQjU1Nfj4+Ij1/vnnH2RmZip8phyAGEq6d++OiIgIBAYGYunSpbC1tYWrqyu8vb0/OP0zfwq+ojsu+f9Qx8fHo3bt2kqebeEU3SmqUaMGXr9+Lf785MkThc/yK3tnyMfHB/Pnz8ft27fRqlUrxMfH4+rVq4XelUhISECjRo3kPqMlvSMVGxuLpUuX4s8//0RGRobMtuKsOfB+IPiQpKQkvH79Gnv27MGePXsU1intcFscynwulP0+KHs8Gxsb7N27Fzk5OUhLSyvwjmz+2hHvv55cS0sLJiYm4vYnT54AUPx5adq0qcxU8vxzcXR0VHjMRo0aFdj34n7fiahyYQZjBnsfMxgzmCLMYLKYwSo+Dj6RQubm5uICl126dMHo0aMxbdo0WFlZoW7dugCAxo0bQ1NTE7dv3/5gW7dv34ampqa4JkHz5s1x+PBh3LlzR3xmuKjyp1bfvn0bnTt3LrT+h55lLmj0u6BXhvr4+GDx4sV48OABmjVrhsjISLRr107mueS8vDzY2tpi9OjRCtvIn76b/3z2pUuXcPr0afzxxx+YNWsW1q5di9DQUBgaGhZ6buWloLfMFPdVvYrkT1ePiopCq1atxDtw74bKkiros/D+5yA1NRWDBg2CtrY2xo8fj8aNG0NHRwe5ubn48ssvi3XezZo1Q0xMjNwjA4rkL2bZvXt3uddk5/vQL9nyosznQtnvg7J8fX0xb948JCUloUOHDoVey9KU/9aiFStWKNz+oVcNf0zfdyJSHWYwZrD3MYMxgynCDCaLGazi42N3pJQpU6YgIyMDq1evFst0dHTQsWNHXLp0Se4NLPkePXqES5cuoVOnTuJUSg8PDwBvF5crrvbt28PAwAD79u1DTk5OofXzF+l8/05JdnZ2kV8L2r17d6irqyMqKgoxMTF4+PCh3C/mxo0bIzU1FZ06dVL4591ppGpqarCxscGkSZMQHh6O9evXIyEhAaGhoQX2If9Ox8OHD+W25a/n0LBhwyKdV2lo0KCBwj4pKlOkTp066NixozilPjIyEpaWlgrvgL7L2NgYcXFxcp8FRcfV19dXeMfs/bdknDt3Di9fvsTixYsREBAADw8PdOrUqUTXtSif/Vq1akFPTw9SqbTAz1FFCD7KKMr3QRldu3aFmpoaLl++XOB0b+C/78CDBw9kyqVSKeLi4sTt+cFJmc9u48aN8ebNG1hYWCg8l8LWoCjO952IKjdmsP8wgxWMGezDmMGYwZjBVI+DT6SUZs2aoUuXLoiIiMCzZ8/E8q+//hpqamr47rvv5KZ+p6WlYerUqVBTU5NZvDD/1ZYREREyz+2/a/fu3bh+/XqB/alatSpGjRqFf/75BwsWLFD4ytPr16+Lr/XV09NDrVq15J7P3r59e5Gf+61Xrx5sbGwQFRWFqKgoVKlSBV27dpWp0717d8TExODw4cNy+2dlZYnX6tWrV3Lb8xft+9CUYkNDQ5ibm2PPnj0y02UTExOxZ88eWFpalup0b2W5urrizp07Ms9cZ2RkICwsTOk2fHx8kJCQgJ07d+LOnTtK3XFzcXFBcnKyzNTo3Nxcha9kNjExwZs3b2Sm8aalpWH37t0y9fLvJr3/2dq4caPS5/I+Ly8vtGnTBuvXr1e4VkBubi42bNiAp0+fQkNDA127dsWxY8cUfhdSU1ORnZ1d7L6UJ2W/D8rS19dHUFAQxo4dK/fde5erqysAYNOmTTLlO3bswOvXr+Hm5gbg7SyD2rVrIzQ0VOa1w2fOnBEX1n33XARBKPCuW1JSUoH9Ke73nYgqN2aw/zCDFYwZ7MOYwZjB3scMVv742B0p7csvv0R0dDQ2btyIGTNmAACsra0xe/ZszJs3Dz4+PujduzcaNGiAf//9FxEREXj69Cnmzp0LKysrmbYWLVqEkSNH4ptvvkFERAQcHR2hr6+Pp0+f4ujRo7h16xZCQkI+2J+hQ4fi4cOHCA4OxoULF9C1a1fUq1cPKSkpOHfuHE6ePCnzxgN/f3+sXbsW06dPh5WVFW7evIlz586hZs2aRb4W3t7emDNnDrZv345OnTrJvdZz+PDhOHHiBMaPHw9fX19YWlpCKpXin3/+wYEDB/DLL7/Azs4Oa9aswblz5+Dm5gZjY2Okp6cjIiICGhoa6Nat2wf7MG3aNAwbNgx+fn7w9/cHAISEhCArKwvTpk0r8jmVhhEjRmDv3r0YM2YMBg0aJL7m9/23lXxIly5dEBQUhB9++AHq6uro3r17ofv4+flh27ZtmDNnDu7evYtGjRohOjpa4S/U/LdujB07FoMHD4ZUKsXOnTtRu3Zt/Pvvv2K9du3aoWbNmpg6dSoGDhwIHR0dHD9+/IO/2AqjoaGBlStXYvjw4Rg6dCi6dOmCDh06QEdHB3FxcTh48CAeP34s3kmaPHky/vrrLwwYMACff/45JBIJ0tPTcf/+fRw8eBD79u0rtburFy5cwJo1a+TKbWxsYGtrW6K2lf0+FEX+G3U+xNTUFP7+/ggNDUVqaiqcnJwQGxuLkJAQmJmZoU+fPgAATU1NfPPNN5g5cyYGDBgAX19fJCUlITg4WFx3JZ+NjQ0GDRqE//3vf7h79y6cnZ1RrVo1xMfH4+jRo/D29sa4ceMU9qck33ciqtyYwf7DDKYYM9iHMYMxgzGDqR4Hn0hplpaWsLW1RVhYGEaNGiX+su/fvz/MzMzw22+/ISwsDMnJyahRowZsbGzwyy+/wMLCQq4tAwMDbN26FeHh4YiMjMTq1auRmZmJ2rVrw8bGBjNmzIC1tfUH+6OmpoZ58+bBw8MD27dvR3BwMF6/fg09PT2Ym5tjyZIlMndsRo8ejVevXuHAgQM4cOAA7OzssHnz5mK9CrZr166YP38+0tLSFN4V0tbWxpYtW/Drr7+Kd+eqVasGY2NjDBw4UHzVqYeHB54+fSrePdPX10ebNm0QGBhY6Pl36NABmzdvxi+//IK1a9cCACwsLLB06VLxbSDlrU6dOti6dSvmz5+PTZs2QU9PDz179oSDgwOGDx+uVBvVq1eHi4sLDh06BHt7e6Wewa5atSq2bNmCH374AWFhYdDU1ETnzp0xePBguVcM16hRA2vWrMHChQuxdOlSGBkZYdiwYahWrRqmT58u1jMwMMCGDRuwePFirFmzBlpaWnBycsKSJUvEN4QUR8OGDREREYHg4GAcPHgQp0+fhlQqRb169dCxY0esWLFCXNOjVq1aCAsLw9q1a3H06FGEh4ejevXqaNy4Mb7++utSfT797NmzOHv2rFz5qFGjShx8lP0+lIWgoCA0atQI4eHhOHPmDAwMDODn54dJkyZBS0tLrNenTx8IgoANGzbgxx9/RNOmTbFkyRIcOnQIf/31l0ybs2bNgrm5ObZt2yY+BlOvXj3Y29t/MMCU5PtORJUbM9h/mMEUYwYrHDMYMxgzmGqpCaW5Uh0REREREREREdE7uOYTERERERERERGVGQ4+ERERERERERFRmeHgExERERERERERlRkOPhERERERERERUZnh4BMREREREREREZUZDj4REREREREREVGZ4eATEZWq2NhYSCQSREVFFWv/SZMmoWvXrqXcKyIiIqKPQ05ODiQSCdasWSOWnT17FhKJBBcvXix0/+XLl6NNmzal3i9nZ2fMnDmz1Nslosqhiqo7QERlZ9q0adi1a1eh9YyNjXHs2LFy6FHFFhsbi+7du4s/V6lSBXp6ejAxMUGHDh3Qr18/mJiYFLv9HTt2IDs7GwMGDCiN7pbYjRs3cOLECfj5+aFu3bqq7g4REdFH4+jRoxgzZgymT5+OoUOHKqzz6tUrODk5wc3NDStXrizfDhbTn3/+iYsXLyIgIADVqlVTdXdE/fv3x+XLl8Wfq1Wrhjp16sDc3Bw+Pj5wc3ODmppasdq+desWjh07hr59+6JevXql1eViy8vLw+rVq2Fubg43NzdVd4eo1HDwiegT5u/vD3t7e/HnBw8eYN26dejfvz+sra3F8tIMF82aNcP169ehqalZrP2XLFkCQRBKrT/F4ePjA2dnZwiCgJSUFMTExGD79u343//+h9mzZ8PPz69Y7e7YsQPp6ekVZvDp5s2bWLVqFVxdXTn4REREVATOzs4wMDDA3r17Cxx82r9/P6RSKXr27Fni43Xs2BHXr1+HlpZWidv6kHPnzmHdunXo16+fXD48cuQI1NVV9+CMoaEhpkyZAgDIzMxEXFwcjh07hqioKDg5OeGXX36Brq5ukduNiYnBqlWr4OTkVGEGn1atWgV/f38OPtEnhYNPRJ8wa2trmUGm8+fPY926dbC2tlY6CKWnpxfpF7mamhq0tbWL3Nd8xR20Kk1t2rSRuz6TJ0/GV199hTlz5qBJkyawtbVVUe+IiIhI1TQ1NeHt7Y3g4GDExsaiefPmcnX27t0LAwMDODs7l/h46urqJcpXpaGsB74Ko6enpzCfrVy5EqtXr8bcuXOxePFiFfWOiArDNZ+IVOjevXsYPXo0bGxsYGVlBT8/Pxw/flymzvnz5yGRSLB792789ttvcHNzg4WFBfz8/HDr1q1S7c+kSZNgbW2NuLg4jB49Gu3bt8ewYcMAvL0rNHXqVHTu3BkWFhaws7PD2LFj8fDhQ5k2FK35tHTpUkgkEsTFxWHGjBno0KEDrK2tMWnSJKSkpMj14d01n7KysiCRSDBv3jwcPXoUPXr0gLm5Oby8vHDgwAG5c4iPj8eoUaPQtm1b2NnZYdasWbhx40aJ1qEC3t5tW7ZsGQBg7dq1YnlmZiZ+/vln9O7dGx06dIClpSV69eqFvXv3yuzv4OCAa9eu4d69e5BIJJBIJOJ5KtsG8PZRuS+//BL29vawsLCAs7Mzxo8fjxcvXsjUO336NAYOHAhra2u0bdsWgwcPxpUrV8TtS5cuRVBQEACgT58+Yp+uXr1apOMQERFVROWRsfIHQhT9vn78+DGuXr2Kbt26iYM2cXFxCAoKQrdu3dC2bVu0b98eAQEB4u/eDylozaf8bGRhYQEvLy9EREQo3D88PBxDhw6Fg4MDzM3N4eHhgeXLlyM7O1us8+2332LdunUAAEdHRzEbPH36FIDiNZ9evXqFwMBAODo6wtzcHN26dcOmTZtkZrHnr2EVGBiIEydOwNfXFxYWFvD09CxRNgPe3vQcP348bG1tsXfvXsTFxYnbjhw5glGjRsHZ2Rnm5uZwdnbG3LlzkZqaKtZZvnw5Zs2aBeDtEwP555x/nZVpAwDS0tKwePFieHh4iDnZ398fhw8flqn3/PlzzJ49W7xenp6eWL9+PfLy8gAAjx49gpmZGQAgNDRU7E/+dVf2OEQVEWc+EanIw4cP0a9fP1SpUgVDhgyBnp4eIiIiMHr0aCxfvhzdunWTqf/7779DKpVi0KBByMnJwcaNGzF27FgcOnSoVGcL5eTkYPjw4bC0tBSnNgPAyZMn8fjxY3z22WcwMjLCkydPEBISggEDBiAyMhK1atUqtO0JEyagbt26mDhxIh4+fIjg4GBUrVoVCxcuLHTfy5cvIzo6Gv3790eNGjUQGhqKyZMnw9zcHI0aNQIApKamYvDgwXjx4gUGDhyI+vXrIzo6GrNnzy7+BXlH06ZNYWVlhQsXLiArKwva2tpITk5GeHg4unXrhs8//xxSqRTR0dGYMmUK8vLy8NlnnwEAAgMDsXTpUmRnZ+Obb74BAFSvXh0AlG7j+fPnCAgIQO3atTF8+HDUqFEDz549w+nTp/HixQvUqVMHALB7925MmzYNnTp1wsSJE5GXl4fw8HAMGjQIwcHBsLKygre3N16+fImIiAiMHz8eDRs2BACYmJgofRwiIqKKqLwylpWVFZo0aYJ9+/Zh4sSJMmsO5Q9IvTtT5/r167h06RI8PT1hbGyMly9fIjw8HIMHD8auXbsUzp76kDNnzmDs2LFo3LgxJk6ciIyMDCxZskTho/Rbt25Fy5Yt4ezsDB0dHVy+fBnr16/H06dPxdlCX3zxBdLT03H06FEEBgZCT08PAFCjRg2Fx8/MzMTgwYMRGxuL/v37o1mzZjhx4gQWLVqEJ0+eyA1UXbt2DUePHkW/fv1Qs2ZNhIaG4ttvv4WZmRmaNGlSpHN/32effYa//voLZ8+ehb+/P4C3A26ampoYOHAgatSogZiYGOzYsQP37t3D1q1bAQDdunXDixcvEB4ejnHjxomZslmzZkq3AbzNedHR0RgwYABatGiB1NRUxMTE4Nq1a+jSpQsA4OXLl/D394dUKoW/vz8MDQ1x8eJFLFu2DE+fPkVgYCDq1KmDxYsXY+rUqejYsSN69+4NAGjcuLHSxyGqsAQiUolx48YJrVu3Fu7evSuWvXnzRnB3dxccHR0FqVQqCIIgnDt3TjA1NRXc3d2FjIwMse7hw4cFU1NT4fjx40ofM7+t3bt3K9w+ceJEwdTUVFi8eLHctvT0dLmy+/fvC23atBE2btwoU2ZqaipERkaKZT/++KNgamoqTJ06VWb/2bNnC2ZmZjLnNXHiRMHLy0v8OTMzUzA1NRXMzMyEBw8eiOVPnjwR2rRpIyxfvlwsW7dunWBqaiocPHhQLJNKpYK/v79cnxTJ7/uvv/5aYJ3Zs2cLpqamYl+m64qjAAAgAElEQVSkUqmQnZ0tUycvL0/o37+/4OPjI1Pet29fwdvbW65NZduIiooSTE1Nhdu3bxfYvzdv3gjt2rUTpk+fLlOempoquLq6CoMHDxbLtm3bJpiamgrXr1+XqavMcYiIiCqq8sxYq1atEkxNTYULFy7IlHt6egpdunSRKVOUpV6+fCnY2toKgYGBYplUKhVMTU2F1atXi2V//PGH3HF69Ogh2NnZCa9evRLL7ty5I7Rq1Upo3bp1ocf+5ZdfhFatWgnPnj0Ty3766SfB1NRUeP78uVx9JycnYcaMGeLPmzZtEkxNTYWwsDCxLC8vT/j6668FiUQixMbGypxPmzZthPv374t1nz59KpiZmQlLly6VO9b7+vXrJ5MP33ft2jW5DKvonHfu3CmYmpoKV69eFcvCwsIEU1NT4cqVK3L1lW3D2tpa+P777z94DtOnTxfs7e2FxMREmfLFixcLEolEePTokSAI/12v2bNny7WhzHGIKio+dkekArm5uTh9+jScnZ3RsmVLsVxPTw/9+/fH8+fPERMTI7NPr169ULVqVfHnDh06AIDM9OLS8sUXX8iV6ejoiP87LS0Nr169Qs2aNdGwYUOlH/97v11bW1tIpVJxOveH2Nvbo2nTpuLP9evXR8OGDWXO/9SpU6hbty48PT3FsipVqig8n+LKX3wzLS1NbD//rmh2djaSk5Px6tUr2Nvb4969e8jKyiq0TWXbyJ8pdfToUUilUoVtnTx5EqmpqejRoweSkpLEP1lZWejYsSMuXbpU4L75lDkOERFRRVTeGcvX1xdqamoyj95du3YN//zzD3x9fWXqvpulMjMz8erVKwCAubl5kZdS+Pfff3Hnzh189tlnMDAwEMtNTU3RqVMnufr5x87Ly0NKSgqSkpJgY2ODvLw8/P3330U6dr4TJ06gZs2a4uwc4O1jcMOHD4cgCDh58qRMfTs7O5nZXXXr1kXjxo1LJcu+n8+A/85ZEAS8efMGSUlJaN++PYC3L11RhrJtVK9eHdeuXSsw0+bl5SE6OhouLi5QV1eXyWhOTk4QBAHnz58vtD+FHYeoIuNjd0QqkJSUhPT0dHFK77tatGgB4O3aRZaWlmK5sbGxTL38KdCvX78u1b5VqVIFDRo0kCtPTk7GTz/9hEOHDolhKZ+JiYlSbb/frr6+PgDlzkFRn2rUqCGz75MnT2BiYiL3qt2STuV+V36oefcNMDt27MCWLVsQGxsrPrOfLzU1VakFQpVpw8HBAZ6enlixYgU2bNgAGxsbuLq6wsfHR/w85K/BVdCbdwDgzZs3H3xMUpnjEBERVUTlnbEaNWqEdu3aITo6GrNmzYKWlhb27NkDAHKDT9nZ2Vi5ciX27t0rN3hQ1KySkJAAADI35vI1bdoUf/75p0zZ5cuXsXz5cly9elVmnScAcutvFqUPTZo0gYaGhkz5u9f5XYqynL6+fqlkWUX57P79+1i6dCnOnTuHjIwMmfpv3rxRql1l25g6dSqmT58OV1dXtGrVCp06dYKPjw/atGkDAEhMTERqaioiIiIKXJcrKSmp0P4UdhyiioyDT0QfiYJebSu8s6BjadDS0lJ4rHHjxuHGjRsYNmwYWrVqhWrVqkFdXR1z586VGywpSEnO4f1gU5R9S9Pdu3ehqakpBqhdu3Zh1qxZ6Ny5M4YPH47atWujSpUqOHr0KLZu3arUtVG2DXV1daxcuRLXr1/HqVOncPbsWcyfPx+rV6/G1q1b0axZM/F6LFmyBIaGhgqPl7+GQ0GUOQ4REdGnoqQZq2fPnggMDMSpU6fg6uqK/fv3w9raWu7m3Lx587Bz507xhSD6+vpQV1fH2rVry3Qmy+PHjzF06FA0btwY06ZNQ4MGDaCtrS2uy6Rsjiupssyy9+7dA/Df2kgpKSkYOHAgdHV1MWHCBJiYmEBHRwdSqRQjR45U6pyL0kb37t1ha2uL48eP4+zZs9i5cyd+++03TJ48GSNGjBDP0cfHB59//rnC4ylzM7ew4xBVZBx8IlKBWrVqQVdXFw8ePJDbFhsbCwDiAtAVwYsXL/DXX39h8uTJGDlypFguCAKSk5MrTF8bNGiAx48fQxAEmdlP77+Rr7gePnyIa9euwd7eXpzNFBUVhebNm2P16tUydd+fag5AbkZWvqK0AQCWlpawtLTE2LFjcfPmTfTt2xdbtmzB3LlzxeBSu3ZthdPulemPMschIiKqiFSRsbp164b58+dj79690NDQwKtXrzBhwgS5elFRUfj888/lFuL+6aefinzM/NlaijLO+2VHjhxBVlYWfv31V5nFyIuSVQrqw507d5Cbmytzk1AVWXb37t1QV1cXs8+ff/6JV69eYfXq1eJjcsDbmUzKKmobderUQd++fdG3b19kZGTgyy+/xMqVKxEQEIA6depAV1cXOTk5Jc5nHzpOQTdriSoCrvlEpAIaGhpwcnLC6dOnxV/QwNspw9u3b4eRkVGFmj6bf6fq/TtT4eHhSE5OVkWXFHJ2dsazZ89w6NAhsSwnJwfbt28vcdsvXrzA5MmTAQCjRo0Sy9XV1SEIgsy1efHiBXbv3i3Xho6OjsKp7cq2kZycLPd30LJlS2hqaopTv11dXaGrq4s1a9bITasHZKd0569j8P7Uc2WOQ0REVBGpImPp6+vDzc0NJ06cQHBwMDQ1NeXeqJfft/dn3Jw7dw43btwo8jHr168PU1NT7N69WyaL3b17F2fPnpWpm5/j3j12bm4uNm3aJNeurq4uAOUexXNzc0NSUpJMXhEEAb/99hvU1NTg4uJStJMqBkEQsHLlSvz111/w9fUV31ZXUHbduHGjXBsFnbOybeTk5CA1NVWmTEdHB02aNEF2djYyMzNRpUoVeHl54ciRIwrX90pNTRVzm4aGBrS0tOQylzLHIarIOPOJSEUmTZqEP/74AwMGDMDAgQPF1wDHx8dj+fLlqFKl4nw9a9WqBWtra6xduxZpaWlo0KABrl+/jmPHjsmtk6BKX3zxBUJCQjBlyhRcu3YN9evXR3R0tLhgt7J382JiYrBnzx5xccmYmBgcPHgQUqkUc+fOha2trVjXw8MDgYGBGD16NNzc3JCYmIiQkBA0aNBAbmDOzMwMf/75J5YtWwaJRILq1avDxcVF6TbCwsKwc+dOeHh4oFGjRsjJycG+ffuQlZWF7t27A3i7TsXcuXMxdepU9OzZEz4+PjA0NMTTp09x/vx56Ojo4NdffwXwdoFTAFi7di2ePXsGLS0tdOrUCTt27Cj0OERERBWVKjKWr68voqOjcfr0aXTp0kVmEfB87u7u2LVrF6pWrYpWrVrh/v372LlzJ1q2bKnUC0reN2XKFHz11Vfo168f/Pz8kJ6ejuDgYJiamoqPoQGAk5MTli5dihEjRsDPzw+5ubnYv38/cnJy5No0MzMDACxbtgyenp7Q1NSEh4eHzILs+fz9/REeHo7AwED8/fffaNq0KU6ePImTJ09i8ODBpf6YfmpqqrieVlZWFuLj43H06FHcv38fTk5OmDNnjli3ffv2MDAwwLfffouBAwdCW1sbx48fl1u3FIA4GLlu3Tq8fPkSWlpasLe3V7qNlJQUuLu7o0uXLpBIJKhRowZu3bqFiIgIuLq6iutQTZkyBRcuXED//v3Rp08fmJqaIi0tDffu3UN0dDQOHDiAevXqAXj793Dq1Cls3rwZhoaGaNSoERo2bKjUcYgqqorz/26JKpmmTZsiJCQEy5Ytw6ZNmyCVStGqVSusXbsWbm5uqu6enBUrVmDBggUIDQ1FdnY22rZti82bNyMwMFDVXRNVr14dv//+O+bPn49t27ZBW1sbXl5e8Pb2xuDBg6GlpaVUO5GRkYiMjISGhgb09PRgYmKC/v37w9/fX1xLIJ+fnx9SUlIQGhqKP/74A8bGxhg9ejTU1dURFBQkU3fEiBF4/Pgxtm/fjjdv3qBp06ZwcXFRug17e3vcuXMHBw8eRGJiInR1dWFqaor/+7//g6urq1jP19cX9evXx/r167F582ZkZmbC0NAQVlZW6Nu3r1ivefPmmDZtGn7//XfMnDkTubm5CA0NVfo4REREFZEqMpaLiwsMDAyQnJwst9B4vsDAQFStWhXR0dHYuXMnJBIJVq1ahZ07d+LatWtFPqazszNWrlyJFStW4KeffoKxsTGmTJmCR48eyQw+5T/av2LFCixbtgzVq1dH165d0adPH/Ts2VOmTQcHB4wcORK7du3CsWPHxLfW5Q+KvKtq1ar4/fffsXz5chw4cACvX7+GsbExpk2b9sEXnxRXYmIivvvuOwBvZyvVqVMHFhYWmDx5Mtzc3GRuMtaqVQvr16/H4sWLsXr1amhpacHFxQWLFy+Go6OjTLtNmzbFjBkzsGXLFjEPBQcHw8bGRqk29PT08MUXX+DPP//E8ePHIZVKUb9+fYwZMwbDhw8X69WuXRs7duzA2rVrcfToUYSFhUFfXx+NGzfG2LFjZV4GExQUhKCgIPz000/IyspCnz59MGfOHKWOQ1RRqQnlvFrvo0ePsHHjRly5cgX379+HjY0N/ve//xW635s3b7BgwQIcOXIEeXl5cHV1xaxZs1CzZs1y6DURfcwiIyMxefJk7Ny5U5ztQ0RU2TCDERERkaqU+5pP9+7dw8mTJ9G0adMivVJ04sSJOH/+PObPn49Fixbh5s2b+Prrr8uuo0T0UXr/eXepVIqtW7fCwMAAEolERb0iIlI9ZjAiIiJSlXJ/7M7d3R2dO3cGAIwfP17hc7fvu3LlCs6cOYOtW7eiQ4cOAIC6deuib9++OHv2bKFvDCCiymPMmDEwNjZG69atkZaWhgMHDuDWrVuYOXMmNDU1Vd09IiKVYQYjIiIiVSn3mU/5bw0oilOnTqFOnTpi6AHevgK8YcOGOHXqVGl2j4g+ck5OTrh69Sp+/PFH/PLLL8jNzcX8+fMxePBgVXeNiEilmMGIiIhIVT6KBccfPHig8G0JzZs3x4MHD1TQIyKqqIYNG4Zhw4apuhtERJ8EZjAiIiIqDR/F4FNKSgqqV68uV66vr4/4+PhC9798+bLSr1gvqry8vGLdSaTi4fUuX7ze5Y/XvHzxepevsrzeampqaNu2bZm0XZkxg1E+Xu/yxetdvni9yx+vefmqCBnsoxh8KikdHR20bt26TNr++++/y6xtksfrXb54vcsfr3n54vUuX2V5vf/+++8yaZdKhhns08HrXb54vcsXr3f54zUvXxUhg30UQ436+vp48+aNXHlKSgr09fVV0CMiIiKiTx8zGBEREZWGj2LwqVmzZnj48KFceUHrEBARERFRyTGDERERUWn4KAafnJ2dkZiYiIsXL4plN27cQFxcHJydnVXYMyIiIqJPFzMYERERlYZyX/MpIyMDJ0+eBAA8e/YMqampOHjwIADAxcUFOjo66NKlCzp06IAFCxYAAKytreHo6IipU6di6tSpUFdXx48//oj27dujU6dO5X0KRERERB8dZjAiIiJSlXIffHr58iUmTJggU5b/89GjR9GwYUPk5uYiLy9Pps7y5cuxcOFCzJgxA3l5eXBzc8PMmTPLrd9EREREHzNmMCIiIlKVch98atiwIe7cufPBOseOHZMr09fXx8KFC7Fw4cKy6hoRERHRJ4sZjIiIiFSl3AefiIiIPmYpKSl4/vw5pFKpqrvy0ZFKpUq/jvddmpqaMDIy4tvViIiIKjFmsOIrTgYr7fzFwSciIiIlpaSk4NmzZzA2NoaOjg7U1NRU3aWPSkZGBnR0dIq0jyAIyMjIQEJCAgBwAIqIiKgSYgYrmaJmsLLIXx/F2+6IiIgqgufPn8PY2Bi6uroMPeVETU0Nurq6MDY2xvPnz1XdHSIiIlIBZrDyVRb5i4NPRERESpJKpUWeuUOlQ0dHh9PsiYiIKilmMNUozfzFwSciIqIi4N021eB1JyIiqtyYBcpfaV5zDj4REREREREREVGZ4YLjRERElczKlSuxatUq8WdDQ0NYW1tjypQpMDExEcujo6MRHByMmJgYZGZmwtjYGK6urggICEDdunWRnZ2NKVOm4ObNm0hMTISuri7Mzc0xceJEmJubq+LUiIiIiCqkyp6/OPhERERUCVWvXh2//vorACAuLg4rVqzA0KFDERkZCV1dXSxatAhbtmxB7969MXToUOjp6eH+/fsICQlBfHw8Vq9ejby8PKipqeGrr75Co0aNkJaWhs2bN2PIkCHYvXs3GjVqpOKzJCIiIqo4KnP+4uATERFRJaShoYG2bdsCANq2bYv69etjwIABOHnyJLS1tbFp0yb88MMP6NOnj7iPra0t/P39cebMGQBA1apV8fPPP8u0a29vDzs7Oxw5cgTDhg0rvxMiIiIiquAqc/7i4BMREZEK7L6SgB+j7+BJcgYaGOhgipcEn1kbq6w/+dO0ExIScOrUKZiZmckEn3waGhpwcXEpsB1dXV1oa2vzzXRERERUIVWkDFaZ8hcXHCciIipnu68kYHrEDSQkZ0AAkJCcgekRN7D7SoLK+pSQ8PbYderUwZUrV+Dk5KT0voIgICcnB4mJiViyZAk0NDTg7e1dVl0lIiIiKpaKlsEqU/7izCciIqIS2HkpHmEX44q0z5XHycjOzZMpy5Dm4rvw69j+12Ol2/GzaYTP2zcs0rHflZOTA+DtmgNBQUGoVq0aOnXqhOzsbNSvX1/pdjZs2IBly5YBAGrVqoX169fD2Fh1s7iIiIjo0/exZrDKmr84+ERERFTO3g89hZWXheTkZJiZmYk/N2jQAMuXL4eamhoAiP9VRq9evWBvb4/ExERs27YNX331FYKDg9GiRYtS7zcRERFRcak6g1Xm/MXBJyIiohL4vH3DIt/5clh0DAnJGXLlxgY6CP3KvrS69kHVq1fHpk2boKamBkNDQxgZGUFNTQ1SqRRaWlp48uSJ0m0ZGhrC0NAQAODs7Axvb2+sX78eS5YsKavuExERUSX3MWawypy/uOYTERFROZviJYGOpoZMmY6mBqZ4ScqtDxoaGrCwsIC5uTnq1q0r3mnT1NREu3btxDeqFFWVKlUgkUgQF1e0afBEREREZU3VGawy5y8OPhEREZWzz6yNsbC3BYwNdKCGt3fbFva2UOnb7t41ZMgQ3Lx5E7t27ZLblpeXh1OnThW4b1ZWFm7duoWGDYu/FhURERFRWajIGexTz1987I6IiEgFPrM2rhBBRxF3d3cMGzYMM2fOxOXLl+Hh4QFdXV08ePAAISEhMDY2hrOzMyIjI3Hq1Ck4OTnByMhIXHMgMTERw4YNU/VpEBEREcmpqBnsU89fHHwiIiIiOdOmTYO1tTW2bt2KyZMnIysrC8bGxnB3d0dAQAAAoFmzZti7dy8WLVqE169fw8jICJaWlti5cydatmyp4jMgIiIi+rh8yvmLg09ERESVzLhx4zBu3LhC63l5ecHLy6vA7W3atMH69etLs2tEREREn6TKnr+45hMREREREREREZUZDj4REREREREREVGZ4eATERERERERERGVGQ4+ERERERERERFRmeHgExERERERERERlRkOPhERERERERERUZnh4BMREREREREREZUZDj4REREREREREVGZ4eATERERERERERGVGQ4+ERERERERERFRmami6g4QERGRari7uyMhIQGHDh1C48aNxfJ79+5h0aJFuHPnDpKTk1GnTh04ODhgwoQJMDIyEutJJBKF7WpqauLmzZsAgPj4eHh4eMjV6d69O5YvX17KZ0RERERU8VXGDMbBJyIiokroypUrSEhIAABERkbi66+/Fre9efMGDRs2xGeffQYjIyPEx8dj9erVuHXrFsLDw1Glytv4EBoaKtfuqFGj0K5dO7nyqVOnwszMDNra2gCAmjVrlsVpEREREVVolTWDcfCJiIioEoqKioKuri5atmyJqKgomeDTrl07mfBiZ2eHevXqISAgAHfu3IGZmRkAoG3btjJtXr9+Ha9evYKPj4/c8Zo2bQpLS0vo6OiU0RkRERERVXyVNYNxzSciIiJVuB4GLDcHggze/vd6WLkdOjc3FwcOHIC7uzs+//xzxMbG4vbt2x/cx8DAAAAglUoLrJMfptzc3Eq1v0RERESlhhlMJTj4REREVN6uhwH7xgOv4wAIb/+7b3y5hZ/z58/jxYsX6N69O7y8vKCpqYnIyEi5enl5ecjOzsaDBw+wbNkyWFhYwNLSUmGbgiCIYUrRnbXp06ejXbt2cHR0xMKFC5GZmVnq50VERET0QcxgKstgfOyOiIioJK5uB65sLdo+8ReA3CzZMmkGsGcscGmL8u1YDwTa9i/asfF2fQF9fX04OTlBS0sLDg4O2L9/PyZPngw1NTWx3ogRI3DmzBkAgJmZGTZs2AB1dcX3rS5evIhnz57B29tbplxLSwsDBgyAg4MDNDU1ce3aNWzYsAGPHz/G2rVri9x3IiIiIgDMYP/fx5LBOPOJiIiovL0fegorL0XZ2dk4fPgwOnfuDC0tLQBv33qSkJCAK1euyNSdPXs2wsLCsGTJEqSnp2PEiBHIylLcx8jISNSoUQOOjo4y5UZGRggMDISHhwc6dOiAcePGYdq0aTh27Fih08yJiIiIShUzmMoyGGc+ERERlUTb/kW/87Xc/P9P935PjUbAsKjS6VcBTp06hZSUFLi4uCAlJQXA28UstbS0EBUVJbPIZZMmTQAAVlZWsLGxgYeHB/bt24c+ffrItJmTk4NDhw7B09NTDFMf0rVrV8ydOxc3b95Eq1atSu/kiIiIqPJgBvuoMhhnPhEREZU3j0BA871n8jV13paXsaiot8FqwoQJ6NChAzp06AAXFxdkZ2fj4MGDyM3NVbifsbExatSogbg4+cD2559/IikpSW66d2HenV5OREREVOaYwQCoJoNx5hMREVF5s/R7+9+j84DX8UCNhm9DT355GUlPT8fx48fh4+MDPz/ZY/39999YuHAhzp07BwcHB7l9Hzx4gOTkZDRs2FBuW1RUFAwNDWFnZ6dUP6KjowFAfF0wERERUblgBgOgmgzGwSciIiJVsPQr86DzvqNHjyIjIwODBw+GlZWVzLZ27dph7dq1iIyMxJkzZ6ChoQErKytUr14dsbGx2LhxI0xMTOTurGVnZ+PIkSPo1auXwoUwV65cibS0NLRr1w6ampq4fv06Nm7cCE9PTz5yR0REROWPGUwlGYyDT0RERJVEVFQUmjRpIhd6AEBTUxPdunVDZGQkZs6cidDQUISFhSErKwv169eHp6cnRo4cCV1dXZn9Tp06hTdv3hQ43btZs2bYuHEjduzYgczMTDRo0ADDhw/H6NGjy+QciYiIiCoaZjAOPhEREVUa69at++D2oKAgBAUFAQB69eqlVJudO3fGnTt3Ctzu7e0thqKMjAzo6OgUWJeIiIjoU8QMxgXHiYiIiIiIiIioDHHwiYiIiIiIiIiIygwHn4iIiIiIiIiIqMxw8ImIiIiIiIiIiMoMB5+IiIiIiIiIiKjMcPCJiIiIiIiIiIjKDAefiIiIiIiIiIiozHDwiYiIiIiIiIiIygwHn4iIiIiIiIiIqMxUUXUHiIiIqHytXLkSq1atQuPGjXHo0CG57Z6ennj06BHGjh2LcePGieXR0dEIDg5GTEwMMjMzYWxsDFdXVwQEBKBu3boAAIlEovCYwcHBMDMzw7Rp07Br1y6FdZYtWwYfH59SOEMiIiKiikVV+cvGxgazZ8/Gvn37FNYpr/zFwSciIqJKSFtbG/Hx8bhx4wYsLCzE8uvXryMhIQHa2toy9RctWoQtW7agd+/eGDp0KPT09HD//n2EhIQgPj4eq1evFusGBATAy8tLZv8WLVoAAMaMGYN+/frJbNu+fTsiIyPRqVOn0j5NIiIiogpDVflr5MiRGDhwoMy28s5fHHwiIiKqhHR0dGBmZob9+/fLhJ/9+/ejY8eOuHnzplh27NgxbNq0CT/88AP69Okjltva2sLf3x9nzpyRadvY2Bht27aVO2ZGRgZMTExgYmIiUz5z5kzY29ujVq1apXV6RERERBWOKvIXADRq1Ag6OjoyZeWdv7jmExERkQpEPYiCZ7gnLLdYwjPcE1EPosq9D927d8eBAwcgCAIAQBAEHDhwAN27d5ept3nzZpiZmckEn3waGhpwcXEpdh9u376N+/fv83E7IiIiKheqzmCVNX9x8ImIiKicRT2IQtDZIPyb9i8ECPg37V8EnQ0q9/Dj6emJFy9e4NKlSwCAixcvIikpCZ6enmIdqVSKK1euwMnJSel28/LykJOTI/7Jzc0tsO7+/fuhra2Nzp07F/9EiIiIiJRQETJYZc1ffOyOiIioBPbG7sWue4oX0C7I9cTryM7LlinLzM1E4B+BCL8brnQ7vVr2gm9z3yId+136+vpwcnJCVFQUbGxsEBUVBScnJ1SvXl2sk5ycjOzsbNSvX1/pdn/44Qf88MMP4s/t2rXD9u3bFdbdv38/XFxcoKenV+zzICIiosrnY81glTV/cfCJiIionL0fegorL0ve3t5YsGABpk+fjujoaMyaNUthPTU1NaXbHD58OLp16yb+XK1aNYX1rl27hri4OHz77bdF6zQRERFRMVSUDFYZ8xcHn4iIiErAt7lvke98eYZ74t+0f+XK61erj01dN5VW15Ti7u6OWbNmYfny5cjIyICbm5vMdgMDA2hpaeHJkydKt9mgQQOZRTQLEhUVhWrVqsHV1bWo3SYiIqJK7mPOYJUxf3HNJyIionI2od0EVNWoKlNWVaMqJrSbUO590dXVhaurKzZv3gw3Nzfo6urKbNfU1ES7du3k3qhSUnl5eThw4AA6d+6MqlWrFr4DERERUQlVlAxWGfMXB5+IiIjKmXczbwR1CkL9avWhBjXUr1YfQZ2C4N3MWyX96d+/P9zc3NCvXz+F24cMGYKbN29i1y75dRXy8vJw6tSpIh/zwoULeP78OfaoUskAACAASURBVLy9VXPOREREVPlUpAxW2fIXH7sjIiJSAe9m3iobbHqfnZ0d7OzsCtzu7u6OYcOGYebMmbh8+TI8PDygq6uLBw8eICQkBMbGxnB2di7SMaOiolCzZk04ODiUtPtERERESqsoGayy5S8OPhEREVGhpk2bBmtra2zduhWTJ09GVlYWjI2N4e7ujoCAgCK1lZOTg0OHDsHT0xNVqjCKEBERESnyKeUvJj4iIqJKZty4cRg3btwH65w/f16uzMvLC15eXh/c786dO4Uev0qVKjh37lyh9YiIiIg+FZU9f3HNJyIiIiIiIiIiKjMcfCIiIiIiIiIiojLDwSciIiIiIiIiIiozHHwiIiIiIiIiIqIyw8EnIiIiIiIiIiIqMxx8IiIiKgJBEFTdhUqJ152IiKhyYxYof6V5zTn4REREpCRNTU1kZGSouhuVUkZGBjQ1NVXdDSIiIlIBZjDVKM38xcEnIiIiJRkZGSEhIQHp6em8+1ZOBEFAeno6EhISYGRkpOruEBERkQowg5WvsshfVUqlFSIiokpAX18fAPDkyRNIpVIV9+bjI5VKi3X3TFNTE3Xr1hWvPxEREVUuzGAlU5wMVtr5i4NPRERERaCvr89BkGL6+++/0bp1a1V3g4iIiD5CzGDFVxEyGB+7IyIiIiIiIiKiMsPBJyIiIiIiIiIiKjMcfCIiIiIiIiIiojLDwSciIiIiIiIiIioz5T74dP/+fQwZMgRWVlZwdHTEihUrkJubW+h+9+7dQ0BAAKysrGBnZ4c5c+YgLS2tHHpMRERE9PFjBiMiIiJVKde33b1+/RpDhw5FixYtsGbNGjx+/BiLFy9GXl4eJk2aVOB+b968wZAhQ9CkSRMsX74cycnJ+PHHH5GYmIg1a9aU4xkQERERfXyYwYiIiEiVynXwKSQkBFlZWVi1ahX09PTg4OCA1NRUrFq1CiNGjICenp7C/bZt24bMzEysW7dOfLWigYEBRo8ejRs3bsDCwqI8T4OIiIjoo8IMRkRERKpUro/dnTp1Co6OjjIBx9vbG5mZmf+PvfuOq7J+/zj+Ooe9tyBLRRQBFXGBA3PlzNJSS1uao7LUtGnD/Nm3Mi2TXJWrtOFKTbPU3BMXooiIA5UhKA5AkHnO+f1xlFRAQeGcA1zPx8MHcJ/73PfFHV/4fN/3574+HDhwoNT3xcbG0rhx46JBD0C7du1QKBTs2LGjUmsWQgghhKjqZAwmhBBCCH3SafgUHx+Pj4/PXdvc3d2xsLAgPj6+1Pfl5eVhYmJy1zYjIyOUSiVnz56tlFqFEEIIIaoLGYMJIYQQQp90+thdZmYmNjY2xbbb2tqSmZlZ6vvq1KnDX3/9RUFBQdEAKCYmBpVKRUZGxgPPm5eXR2xs7MMXfh+5ubmVdmxRnFxv3ZLrrXtyzXVLrrduyfXWHxmDiUcl11u35Hrrllxv3ZNrrluGcL11Gj49rAEDBrB48WI+++wzRo8eTXp6Ov/3f/+HkZERCoXige83MzPD39+/UmqLjY2ttGOL4uR665Zcb92Ta65bcr11qzKvt74HVNWVjMHEbXK9dUuut27J9dY9uea6ZQhjMJ2GT7a2tmRlZRXbnpmZeVcvgXvVr1+fyZMn8+WXX7Js2TKUSiUDBw4EwMXFpdLqFUIIIYSoDmQMJoQQQgh90mn45OPjU6yvQEpKCjk5OcX6ENyrf//+9OnTh/Pnz+Pk5ISDgwMhISEMGDCgMksWQgghhKjyZAwmhBBCCH3SacPxDh06sHv37rvuvP3999+Ym5vTunXrB77fzMwMPz8/nJ2dWbt2LWq1mp49e1ZmyUIIIYQQVZ6MwYQQQgihTzqd+fTcc8+xZMkSRo8ezYgRI0hMTGTWrFkMGTLkrqV/H3/8cVq1asUXX3wBQFZWFnPnzqVVq1YYGRmxf/9+Fi1axGeffYa9vb0uvwUhhBBCiCpHxmBCCCGE0Cedhk92dnb89NNPTJ48mddeew1bW1tefvllRo8efdd+KpUKtVpd9LVSqSQ2NpYVK1aQm5tLgwYNCA8Pp2vXrrosXwghhBCiSpIxmBBCCCH0Seer3fn6+rJ48eL77rN169a7vra0tGThwoWVWZYQQgghRLUmYzAhhBBC6ItOez4JIYQQQgghhBBCiJpFwichhBBCCCGEEEIIUWkkfBJCCCGEEEIIIYQQlUbCJyGEEDXPseXwbWMaLWsD3zbWfi2EEEIIIYSoFDpvOC6EEELo1bHlsG4MFOSgAMhI1H4N0HSgPisTQgghhBCiWpKZT0IIIWqWLZOhIOfubQU52u1CCCGEEEKICifhkxBCiJpDo4GMpJJfy0iEbV9C3D+QmaLbuoQQQgghhKjG5LE7IYQQ1Z9GA2e3wPYpgKbkfZTGsHMqaNTar63dwL0Z1G7230fb2jorWQghhBCi2jm2HLZMplFGEth5QpeJVaPtwa26qWp1GxAJn4QQQlRfGg2c2QLbv4TkQ2DrCc2eh5hVdz96Z2IBfb6DRr0hNRouHoGLUZASBac3SSAlhBBCCPGoqmrfzTvqBqpO3QZGwichhBDVj0YDZzbfCp0Og50XPDFDGzwZm4JPR9gyGU1GEop77155h2r/3Zaf/YBAyhXcgyWQEkIIIYS4n9L6bv79LuRmaL9WKABF6R/Lso+iPPvd7yPajxs/LLnuzZ+Cfx8wNv/vnKJUEj4JIYSoPjQaOP2vNnS6GAl23tAnHIIGa0On25oOhKYDORkbi7+///2PaWp1n0AqShtKlRRI1W6mDaUkkBJCCCFETZabAbHrtDOGSnw9Hf5+R7c1VYTMi/C5GyiUYGqtHTOaWt36/M6vrcDM5p7XS9vXGsysKy7QMqDHHCV8EkIIUfVpNNrwZ/uX2jDI3lv7GF3QoLtDp4ryoEAq5VYodeZfCaSEEEIIUfMU5MCpjXB8JZzaBKo8UBqBWlV8X1t3GLkT0GjHdEUfKWFbaR8p4373+1jK+Za9CNmXi9dtbg/t34K8LO04MP/2x1ufZ6Xe8fWtbbfHhQ9SLNCyAlObe76+FVQVC7Ru7XthD+z4CgpzDeIxRwmfhBBCVF0ajXZgs2PKrdCpDjw5Uxs6GZnotpZHCqSa/ffongRSQgghhKiKVIVwbjtEr4TYvyD/BljVgpZDockAuBZ/d+8k0Pbd7Pp/YO2it7IfqPvnJdfda1r5QhyNRnuMu4KqrHtCq1vbKiPQAu35t0yW8EkIIYQoE40G4v7R3s1JiQKHuvDkLAh6Tveh0/2UKZCKKj2Quj1T6nYgVVVXWjGgKd9CCCGEqEAaDSQegOgVELMabl4BMzsIfAoa94d6HbQzngA8W2o/ltZ301Ddru9Rx2AKBZhaav9RQWHb/QKtX54p+T0ZSRVz7nKS8EkIIUTVodFA3N+wfQqkHtOGTk/NhqbPGlbodD8PG0hZ1YK0WFAXardlJMLa0XAjVbtKn0KhnaJ933+39uF++yoqtmlmVV3ZRgghhBCluxSjDZyO/wHpCdoeRQ17QJP+4Ps4mJiX/L7y9N00JLfqNjj3C7TsvErus2XnqZPS7iXhkxBCCMOn0cDJ9dqZTqnHwKEePDVHOwioKqHT/ZQlkIpe8V/wdFthLvz7ifZfhXpQkFXa64rir6cnFK9bj1O+hRBCCPGQrp/XPlIXvVJ7Q0xhBPU7QaePwK8XmNvqu0Jxpy4TS35csMtEvZQj4ZMQQgjDpVZD3O3QKRocfaDvXGgyEIyq+Z+wewOpo0tL3/fpedpZUqX+05ThtfvtU5bj3PE6d+x3Lb7kmvU05VsIIYQQ5ZB1Wfs4XfQKSDqo3eYVCr2+hsB+YOWs3/pE6e54XNAQHnOs5iN3IYQQVZJaDSf/gh1T4VI0ONaHvt9rm1VW99CpNHaepUyd9jLsGUQJEawvvEq4gz2pxka4FaoYez2d3sZO+q5MCCGEECXJzdA2DI9eAed2aG8muTaGrpOg8TPaVYVF1WBAjznW0BG8EEIIg6RWw8l1t0Kn49rQqd8P2oaVNTV0us3Apk6X1frgfkw6t5pcpbaPVIqJMZOcnaBeP3rruTYhhBBC3FKQA6c3aQOnU5tAlaftrdl+vLaPU60q1J9JGKQaPpIXQghhENRqiF2rDZ0ux4CTL/T7UXt3raaHTrdV1EorOhZ+ZX9R8HRbrlJB+JX9Ej4JIYQQ+qQqhHPbIfoPiF0H+Te0C5y0HKqdbe7RomIXIRE1mozohRBC6I9aDbF/3gqdToBTA23/osbP/Lcsr/iPoa60Uoo8VR4p2SklvpaanarjaoQQQgiBRgOJB+D4Sm0vp+w0MLODwKe0M83rhsmNP1Ep5KdKCCGE7qnVcGKNNnRKiwXnhvD0fGj8tIRO1cDNgpusOLWCn2N+LnUfNys3HVYkhBBC1HCXYrSr1B1fqV2J1tgcGvbQPlLn+ziYmOu7QlHNSfgkhBBCd9SqW6HTtFuhkx88s0C7WooOQ6f18esJjwwnNTsVtxg3xjYfS28feQjsUWXkZfDbyd/4NfZXMvIyCHEL4cn6T/Jr7K/kqnKL9jM3Mmds87F6rFQIIYSoAa6fvxU4/aGdYa4wgvqdoOOH0Kg3mNvqu0JRg0j4JIQQovKpVdqp3TunQdpJvYVOoA2eJu2dVBSGpGSnMGnvJAAJoB7SlZwrLD6xmGUnl3Gz8CaPeT7GiKYjCHIJAqCBQ4P/wj4rCfuEEEKIR3Zsecl9ILMua8dc0Ssg6aB2X69Q6PU1BPQFaxf91i10ypBuuEr4JIQQovLcDp12TIUrceDSCPov1A5+9PR4XXhk+F2zcAByVbmER4ZLIFJOKVkpLIpZxKrTq8hX5dO9bneGNxmOn6PfXfv19ulNb5/exBrAMr9CCCFElXds+d0r4GYkwp9vwK7p2vGWRg2ujaHrJG0fTXtvfVYr9MTQbrhK+CSEEKLiqVVwfBXsnApXToGLP/RfdCt0Uuq1tJRSGl2nZKew6fwmOnp1xNTIVMdVVS3nM86z4PgC/jr7FwBP1H+CYY2HUdeurn4LE0IIIWqCLZP/C55uU+Vrx1ztx2v7ONWSmz01naHdcJXwSQghxMO7d8p3549BodTOdLp6GmoFwICfwP8pvYdORVQWYHSz+HaNgrd3vI29mT29fXrT17cvjRwb6b4+AxZ3LY750fPZdGETJkoTBvgNYGjgUGpb19Z3aUIIIUTNkZFU8naNGrp8ottahMFRqVVEpUUZ3IrDEj4JIYR4OCVN+V79GqCBWoEwcDE06mM4oRMQcyUGjSIHNAoUCk3Rdo3ahLyUvix8qRNrzqxhedxyfo39lUaOjejr25fe9Xpjb26vx8r162jaUeYdm8eOpB1YmVgxJHAILwa8iLOFs75LE0IIIWoeO0/tuKuk7aJGylflE5ESwdaErWxL3Ma13Gul7quvFYclfBJCCPFwSpryjQYsneC13QYVOgGk56Yzfvt4lGp7ctIew9RpOwqTdDQF9uSldceZNoR5hhHmGUZ6bjp/n/ubNWfWMOXAFL459A0dvTrS17cvbd3bYqys/n8+NRoN+1P3M//YfPan7sfOzI43mr3BoEaDsDOz03d5QgghRM0V2A/2fnf3NhMLbdNxUWNk5WexK3kXWxO2sit5F9kF2ViZWNHBowOdvTuTXZDNlANTDGbF4eo/ehZCCFE5SpvyffOawQVPKrWKD3Z9QFpOGi43x3E23ZGC9NC79sk2KyAi/iqhPk7Ym9sz2H8wg/0HE3ctjjVn1vBX/F/8e+FfalnUok/9PvT17VstexxpNBp2JO1g3rF5HLtyDGcLZ95p+Q4DGg7A0sRS3+UJIYQQNZtGAwkRYOEIppaQkXz3aneiWruSc4XtidvZkrCF/Sn7KVAX4GjuSI+6Peji3YWQ2iF39S41NzY3mBWHJXwSQgjxcKrQlO+5R+ey5+IeOjmNYm2MI08392B//DUupufgbm/BC6HerDicxOB5EbzT3Y/XOtRHqVQA4Ofox/ut32d8i/HsSNrB6jOrWRSziAXHF9DMpRl9ffvSvW53rE2t9fxdPhqVWsWmC5uYHz2fU9dP4WHtwSehn/CU71OYGZnpuzwhhBBCAJzbCUkHoNfX0HqEvqsROpB0I4ktCVvYmrCVI5ePoEGDh7UHgxoNoot3F4JcgjAqZRVpQ1pxWMInIYQQDydsPPw17u5tBjjle0fiDn449gNhbr1Yv8ObHoFufDMgCIVCcdcf4hfb1OX9P44xdUMch89fZ/rAZthZmhQdx8TIhK51utK1TlfSbqaxLn4da86sYdK+SXx18Cu6enelX4N+tHBtgVJhWDO/7qdAVcBf8X+x4PgCLmReoJ5dPT5v/zk96/XERGny4AMIIYQQQnd2TgNrNwh+Ud+VPJT18ev/m4kTo9+ZOIZKo9Fw6voptiZsZUvCFuKuxwHQ0KEhrwW9RhfvLjR0aIhCodBzpeUj4ZMQQoiHc/Oq9qO1K2RdNsgp3wmZCUzYNYGG9o2IOtIJd3tTvurftMQ/1tZmxswaFEzruo78b/0Jes/cxZznm9PUs3ijcRdLF15p/ApDA4dy7Mox1pxZwz/n/mFd/Do8rD14yvcpnqr/FO7W7rr4Nh9KbmEuf5z+g59ifiI1OxV/R3+md5xOF+8uVSo8E0IIIWqMhAg4vwu6fwEm5vquptzWx69n0t5JRT2IUrJTmLR3EkCND6BUahVH044WzXBKykpCgYLgWsG80/IdOnt3xsvGS99lPhIJn4QQQpRfQQ5EfA8NusHzK/RdTYlyCnMYt30cSoUSy/RhpN1Qs/K15thZlD6bR6FQ8HLbujT1tOPN347Qf+4+PukTwAsh3iUGVgqFgiCXIIJcgniv1XtsvrCZP8/8yZyoOcyNmktI7RD6+vali3cXzI0NY5CYlZ/FsrhlLD6xmGu51wiuFczE0Im092hf5e6gCSGEEDXKzmlg6Qwthui7kocSHhl+V/NrgFxVLlMPTqWFawtcLV1r1FgkX5XP/pT9bEnYUrRCnYnShJDaIQxrMoyOXh2r1crCEj4JIYQov6jf4OYVaKef1TIeRKPR8Nm+zzh9/TRPuk3kl20qPnkigCCv4rOYShLs7cBfo9szbnkUn6w5zuHz1/i8XxOszEr/s2lhbEGf+n3oU78PSTeSWHt2LX+e+ZMPdn2AjYkNPer1oJ9vPxo7N9bLwCo9N51fYn/ht5O/cSP/Bm3d2zKiyQhaurXUeS1CCCGEKKfkw3BmM3T5FEyt9F3NQ0nNTi1x+7Xcazy+8nEsjC2oa1uXunZ1qWdXT/vPth51bOsYzE28R5VdkM2upF1sSdhStEKdpbElYZ5hdPHuQphHWJXvI1oaCZ+EEEKUj1oFe2eCR0uo007f1ZRoedxy1sWv4+l6r/DbRkseD6jFK+3qlusYDlamLHy5FXO2n2H6v6c4fjGT719ojm8tmwe+19PGk1HNRvFa0GscTD3ImjNrWHd2HStOraC+XX36+vblifpP6ORu1uWbl1kcs5jlp5aTU5hDF+8uDG8ynMbOjSv93EIIIYSoIDu/AXN7aDVc35U8lEvZl1AqlKg0qmKvOZk7MarZKM5lnONcxjmOXj7KP+f+KXpdgQJ3a3dtKGVbryiYqmtbF2cLZ4OfLXU152rRCnURKRF3rVDX2bszIbVDasTiLhI+CSGEKJ/YtXD9HDw+GQzwj/3RtKNMOTiFNrXbs3lfE2rZKJlWSp+nB1EqFbzZuQHB3g6MXXqEJ2ft4cunm/BUM4+yvV+hJKR2CCG1Q/gw5EM2nt/I6jOr+ebwN8yInEGYZxh9ffvSwbNDhTf3TrqRxKLji1h9ZjUqjYqe9XoyvPFwfB18K/Q8QgghhKhkqcchbj10nADmtvquptyu5Fxh+KbhKFFipDQiX51f9Jq5kTnvtnq3WM+nnMIcEjITigKpc5nnOJ9xnshLkeQU5hTtZ21ifVcYdftzLxsvTI1MdfY93is5K5ktF7awJWELUWlRqDVqPKw9eK7Rc3Tx7kIzl2alrlBXXUn4JIQQouw0Gtg9A5x8oZHhNYa8mnOV8dvH42bphir1OS5lZLP8tTbYWz7a4KOdrzN/jQ5j9O+RjF0axaHz1/n4CX/MjMs+aLAxtaF/w/70b9if+PR41pzVzobanrgdR3NHevv0pq9vXxo6NHykWuPT41lwfAHr49ejVCh5yvcpXgl8BS/bqt2kUgghhKixdn0NpjYQ8qq+Kym3a7nXGL5xOJduXmJ+9/mkZKf8t9qdVemr3VkYW+Dn6Iefo99d29UaNZdvXiY+I74omDqfeZ6IlAjWnl1btJ9SocTT2rPEYMrB3KHCv8+iFeoSt7I1YSsnr50EoIFDA0Y2HUkX7y74OfgZ/CytyiThkxBCiLI7txNSoqBPOBjY3ZpCdSHv7XyPjLwMBrp/xZyDWXzYqxHNvStmgOFmZ85vI0KZtjGOH3fGczQpndmDm+PlaFnuY/nY+zC+xXjGBI9h78W9rD69mt9P/s6SE0sIcAqgn28/etbriZ2ZXZmPeeLqCeZHz2fzhc2YGZkxqNEghgQOwdXKtdz1CSGEEMJApMVBzBpoPw4sKj40qUzpuemM2DSC5Kxk5nSdQ3PX5oB2ZbvY2Fj8/f3LfUylQomblRtuVm60dW9712vZBdmczzjPucxzdwVT+y7uu2u2lb2Z/V1h1O1wytPGE2NlyRHJ+vj1xUKznvV6aleou7CFrYlbSbyRiALtYjRvt3ibLt5d5ObfHSR8EkIIUXZ7ZoC1KzR9Tt+VFDPzyEwOpB7gtYCPCF+TT5dGtRje3qdCz2FipOTDXv60qOPAOyuO8sTM3UwfGEQX/4cLeIyVxnTw7EAHzw5cz73O+vj1rDmzhs/3f860g9Po7N2Zvr59Ca0dipHSqMSBT22r2vwY/SN7kvdgbWLN8CbDeSHgBRzNHSv0exdCCCGEHuyaDiYW0OYNfVdSLhl5GYz8dyTnM84zs8tMWrm1qvRzWplYEegcSKBz4F3bVWoVF7MvasOoO8KpHUk7WH1mddF+xkpjvG28iwVTp66f4ouIKRRo8gBIyU7hw10f8tm+z8guzMZYaUxI7RCGNh5KJ69O1WqFuook4ZMQQoiySTkGZ7dqV1kxMawVRzZf2MzC4wvpW78/y7a54myt5usBQSiVlTO1uXugG43cbBj1ayTDfj7EqI71Gf94Q4yNlA99TAdzB14IeIEXAl4g9mosa86sYf259Ww4vwFXS1cCHAPYm7KXPNV/A58JuyagQYODmQNjgsfwXKPnsDF9cEN0IYQQQlQB1+IhegWEvg5WVSfQuJF/g9f+fY0z6WcI7xRebIaSrhkpjfCy8cLLxosOnh3uei0jL4Pzmef/myl1K5zambSTQk1hqcdUoyZfVchXYV8R5hkm468ykPBJCCFE2ewJ1/YbaPmKviu5y7mMc3y852MaOzfmyvkeJKdfZfmroThYVW6TyTpOVvzxelv+b10Mc7afJTLhOt8NCqaWzaMHc/5O/vg7+fN2y7fZlriNNWfWsC1pW7H9NGiwNbVlY/+NWBhbPPJ5hRBCCGFAdn8LSmNoO1rflZRZdkE2ozaP4uS1k0zvOJ0wzzB9l3RfdmZ2BLkEEeQSdNf2AnUByTeSOZdxjjFbx0AJ9zML1Hn08umlo0qrvoe/RSuEEKLmuH4eYlZDyyFgYa/vaorcLLjJuG3jMFWa0t72bf45foV3uvnRoo5uHjkzNzHiy6eb8s2AIKIS0+n93W4i4q9W2PFNjUzpXrc7c7vORVHSqAft3UUJnoQQQohqJj0Ron6H5i+BjZu+qymTmwU3eWPLG0RfiWbqY1Pp5N1J3yU9FI1GQ1zKTVZG5PH5SiXqgpLHvqVtFyWTmU9CCCEebN9sUCghdJS+Kymi0Wj4dO+nnMs8x4Tg6UxclkZHPxde7VCxfZ7K4pkWngR62DLql0gGz4vg3e6NeLWDT4U+9udm5UZKdkqJ24UQQghRzewJ135sN1a/dZRRbmEuY7aO4cjlI0wJm8LjdR7Xd0nlolJriEy4zobjqWw4nkpyeg5KBYTUc+JSRi80TitQKAuK9teoTTDK6IVGo6nRK9iVh4RPQggh7i/7KkQugabPgq27vqsp8mvsr2w4v4HXmo7m+w1KHCxN+KYS+zw9SCM3W9aObs/7fxzjqw0nOXzhGt8MaIadpUmFHH9s87FM2juJXFVu0TZzI3PGNq8ag1IhhBDV15ojyUzbGMfF9Bzc7S14t7sffYM99F1W1XUjFSIXQ7NBYG/4q6XlqfIYu20sB1IP8Hn7z+lZr6e+SyqT/EI1EfFX2RCTyqaYS1zJysPUSElYA2fGdm1AV39XHK1MWXPEiw83aVA4/oPCJB1NgT35ad0pyGzKm78d4Yt+TSpsvFedSfgkhBDi/g78CIU50G6MvispcvjSYb459A2dvToTG9uShGsp/D4iFCdrM73WZW1mzKxBwbSq48Dnf8fSe+Yu5j7fgiaedo987N4+vQGKrXZ3e7sQQgihD2uOJDNhVTQ5BSoAktNzmLAqGkACqIe1dyaoC6H9eH1X8kAFqgLGbx/P3ot7mdx2Mn3q99F3SfeVk69i5+k0Nh5PZXPsJTJzC7E0NaJTo1r0CHSjo58LNuZ3B0nan+OXmbYxtChgfbt7Q1Jv5DJ90ykiE67zzcAg2tavOk3h9UHCJyGEEKXLz9aGT369wMVP39UAkHYzjXd2vIOHjQdB5q8x+Vg873RrSIiPk75LA0ChUDCkXT2CvOx549dInpm7l0+fDGBwa+9Hnpbd26e3hE1CCCEMyrSNcUXB0205BSqmbYyT8OlhZF+BQwuhyQBwrKfvau6rQF3AOzveYWfSTj4J/YR+Dfrpu6QSZeYWsO3kZTYcT2V7XBo5BSrsLEzoFuhGj0A32jdwxtzE6L7H6BvsUeLPc3tfZ95aGsXz8/fzagft6semxtJauyQSPgkhhCjdkV8g5xq0e0vflQD/DXKyPaw/uAAAIABJREFUC7L5oNm3jFl8nrAGzozq6Kvv0ooJ9nZg/Zgw3loWxUerj3Po/HU+79cYS1P50yuEEKL6uJieU67t4gH2zYaCHAh7W9+V3FehupAPdn7A1sStfND6Awb6DdR3SXe5mpXHvycusSEmlT1nrlCg0lDLxoz+LTzp0diN1vUcMTF69JCoqac9f41pz2d/xfL9jrPsPpPGjGeD8a1lXQHfRfUiI2AhhBAlUxXC3lngFQreIfquBoBvD39L5OVIJrf5gq/WpWNrYcL0gc301ufpQRysTFk0pBWzt51h+uZTHE/OYO4LzfGtZaPv0oQQQogK4W5vQXIJQZOLjX4fha+Sbl6DA/MgsC+4NNR3NaVSqVV8tPsjNl3YxDst3+F5/+f1XRKgDTw3xmgbhh88fw21BrwcLRjarh7dA90I9rKvlDGjpakxXz7dhI5+LnzwxzGemLmLT56omFnv1YmET0IIYQDubtSZYhiNOmNWQ0YC9PxKv3XcsuHcBpacWMLgRoPZecSL81eS+WV4iMEPbpVKBaO7NCDY24GxS4/w5Kw9THmmKU8GGU7zdiGEEOJhvdvdj3dWHKVQrblr+838Qk5fukEDV7nhUmYHfoT8GxD2jr4rKZVao2bi3on8fe5vxjYfy8uBL+u1nvi0LDbEpLLxeCpHkzIA8HO14c3ODegR6IZ/bRudBUDdA91o5mXPOyuO8tHq42w7mcZXzzTRe09SQyEPIwohhJ7dbtSZnJ6Dhv8ada45kqy/ojQa7RK/zn7QsIf+6rjlzPUzTNw7keBawdRVPMvqI8mM7dKwSjV2bN/AmfVjwgiobcuY348w8c/j5BWqHvxGIYQQwoD1CXLH0tQIM2MlCsDD3oIJPf2wMDXmuR8jOJmaqe8Sq4bcTIiYA369wa2xvqspkVqjZvK+yaw9u5ZRzUYxvMlwndeg0WiIuZjB9E1xdPt2B52/2cHUDXGgUPB+j0ZsffsxNo7rwPjHGxLgbqvzmUeutub8PLQ1nzwRwM5TafQI38X2uMs6rcFQycwnIYSoJLkFKtJvFnAtO5/0m/lcv1nAtZv5pGfnaz/eem3vWe1z6HfSe6POs1vgUjQ8NRuU+r1PkZWfxbjt47A0tmRUwCSGzI+jbX0n3uxseH2eHsTNzpzfR4YydcNJ5u06x9HEdGYNbo6Xo6W+SxNCCCEeys5TaWTmFvL9C83p0bh20fbHA9wYPG8/g36M4JfhIQS6P/rKr9XawfmQmwEdDHPWk0aj4Yv9X/DH6T8Y0WQErzV9TWfnVqs1HEm8zobjqWyISSXxWg5KBbSu58ikPgF0C3TD3d5CZ/U8iFKpYFj7erSt78TYpUcYsuggQ9vV5f0ejR7Y2Lw6k/BJCFGt3P34mkWFPL6m0WjIKVBx/WYB17PzuX4z/1agVMD1m/m3tt36/GY+17O1n9/ML31Wi7WZMfaWJjhamRYLnm7Ta6POPeFgUxua6Ld5pEaj4ZM9n5B4I5HZnX/g4+WJWJuZMOO5ZhgZaJ+nBzExUvJR7wBa1HHk3RVHeWLmbr59NojOjVz1XZoQQghRbksPJuBkZVrs75iPizXLXg1l0I8RDJ63n1+GhdDEUwKoEuVnw75Z4NsVPJrru5piNBoNUw9OZVncMoYGDmV08OhKn1FUoFKzP/4aG2JS2BhzibQbeZgYKWjv68ybnXzp6u9q8I+z+de2Ze2b7Znyz0kW7TnP3jNXCR/UjEZutvouTS8kfBJCVBu3H1+7vdzv7cfXgKIASqPRkJVXWDTr6N7A6PqtGUpFgdKtffIK1aWe19bcGAcrUxwsTXGxNqOhqw0OlqY4WplqAyZLU+wtTXGw0n5uZ2mCmfF/dz3aTdlqWI06kyPh3E54/DMwNtVPDbf8FPMTmxM2807Ld1i9z5yzaVf4ZVgItWzM9VpXRejRWNuH4PVfInnlp0O80ak+47o2xLgCVl4RQgghdCHtRh5bYi8ztF3dEpeXr+NkxbJX2zBoXgSD50ewZFgIzbzs9VCpgTv8E9y8Ch3e03clxWg0GqYfns4vsb/wgv8LjGsxrtKCp9wCFbtOX2HD8VQ2x14iI6cACxMjOjVyoXugG50a1cLW3KRSzl1ZzE2MmPRkII/5ufDuimM8OWsPH/RoxJC2dQ12wZzKIuGTEKLamLYxrih4ui2nQMV7K48xd/vZW4+65Zc600ihAHsLk6IgycPegsbutkVfO1j+95qjlQn2lqbYW5g8cljwbne/u0Kz2zJu5rP5xCW6Buh4RsyecDCzgxZDdHvee+xP2c+MyBl0q9MNq5zOrDx8jDGdfWnnW3X6PD1IHScrVo1qy6S1MczedpbIC+mED2pWLcI1IYQQ1d/qI0kUqjU828qr1H28HC21AdSPEbw4fz8/vdKaFnUcdFilgSvIhT3fQd0wg1ld+DaNRsPMIzP5KeYnnvV7lvdavffIwdO9i+yM7lwfSzMTNh5PZVvcZW7mq7A1N6ZrgCs9At3o0NClWjyq1smvFhveCuP9lceY/NcJtp9K4+v+TallW3PGfBI+CSGqjdIeU8tXqanrbEmwpf2t8MjkVpiknY10+3NbCxO9PMp1e1bWnY8LDg+rx6rIZIYvPsT4xxvyZidf3dwduXoWYtdCu7Fgrr8pwanZqby38z3q2tZlSMN3GTg3ipB6joztarjLDj8scxMjpjzTlJZ1Hfl4TTRPfLebmYOCCfFx0ndpQgghRKk0Gg1LDybSoo4DvrXuv6Kdh70Fy14NZfC8/by0YD+LhramdT1HHVVq4I4sgaxUeGaevisp5vtj3zMveh7PNHiGD0M+rJDg6d6nFD5YdRwAZ2sz+gV70KOxG6E+TphUw5ngztZmzH+5Jb/sT+Dz9SfoEb6Lr55pyuO6vtGsJxI+CSGqDXd7ixIfX/Owt+CHF1vqoaKy6xvsQd9gD2JjY/H39wdgUGtvJqyKZvq/pzhxMZOvBwZhbVbJv7b3zQKlMYS8XrnnuY98VT5vb3+b3MJcpnT5hreWnMLS1IjvBgVX2T5PZdG/hSeNPWx5/ZdIBs/fz7vd/Xi1g4/OV2kRQgghyuLwhevEp2Uz9Zn6Zdq/tp0FS0eGMnheBC8vPMDCIa1oU7+G32gpzIfdM8ArRDvzyYDMj57PnKg5PFn/SSa2mYhS8ehh0NSNJ4vN9AdwtjZl/4ddqvU47zaFQsGLoXVo4+PImN+jGLH4EINDvPm4tz+WptU7nql+caIQosYa0rZusW0WJka8291P98VUAHMTI6YPDOLj3v5sOpHK03P2cOFqduWdMOsyHPkVggaBjf7uwEw9OJVjV47xWbvP+HnHTeIu3WD6s81wrQHTkhu52bL2zXb0CHRjyj8nGbH4MBk3C/RdlhBCCFHM0oOJWJka0btp7QfvfIurrTlLR7bB08GCoT8dYPfpK5VYYRVw9HfITNL2ejKgm00/x/xMeGQ4ver1YnLbyRUSPF24ms3F9NwSX7ualV8jgqc7+dayYfUbbXm1gw+/H0jgiZm7iU7K0HdZlUrCJyFEtZGWlYcCcLM1R4F2xtOXTzd55NXu9EmhUDA8zIfFr4Rw+UYeT87aw85TaZVzsv0/gCof2o6pnOOXwbqz61gWt4whgUPISQ9k6cFERnWsz2MNXfRWk67ZmJswa3Awn/YJYMepyzwxaxfHk6v3YEQIIUTVciO3gPXHUugT5I5VOWdlu9iYsXRkKHWdrBj280F2VNa4xtCpCmH3dHAPBt8u+q6myK+xv/L1oa/pVqcbn7f/HCPlo/VbUqs1LNl3nh4zdlFavORub/FI56iqzIyNmNDLn1+HhXAzT0W/OXuYu/0sKnXJ/WmrOgmfhBDVQm6BiuWHEunZxI2ID7twbkpv9nzQuUoHT3dq38CZtW+0p7adOUMWHeDHnWfRaCrwD1NeFhycB/5PgLNvxR23HOKuxTF532RaubWij9cwPlwVTau6Dox/vPr1eXoQhULB0Hb1WPZqG1QqDU/P3ctv+xNYHZlEuylbqffBetpN2cqaI8n6LlUIIUQN9NexFHIKVAy8T6Px+3GyNuO3EaHUd7FmxM+H2HryUgVXWAUcXwnXz0OHdw1m1tPyuOVMOTCFzl6dmdJhCsbKR3sMLDk9h5cWHuCTP2NoWdeBT58MwOKe5uFV+SmFitLW15kNb4XxeIArX204yfPzI0rtZVuVSfgkhKgW1h9LIf1mAS+E1NF3KZXG28mSP15vS4/Gbnzx90neWhZFTn7x5+YfSuTPkJsB7d6qmOOVU0ZeBm9tewtbU1s+azOFMb8fw9RYyXeDgh95NcGqrLm3A3+NCSPUx4kPV0fzzopjJKfnoEE7oJuwKrpKBFBrjiTTbspWev0cL6GZEEJUA0sPJtLQ1ZpgL/uHPoajlSm/jQjBz82GV5ccZlNMagVWaODUKtj1DdQKhIY99V0NAKtPr+aziM/o4NmBrx/7GhOlyUMfS6PRsPxQIj2+3UlkwnU+79eYxa+0Zkjbenz5dBM87C2qzVMKFcXe0pQ5zzdnav+mHEvKoMeMnaw/lqLvsipU9e5oJYSoMZZEXMDHxaraN660MjNm9uDmzNl+lq83xXHmchY/vNgCTwfLhz+oqgD2zYY67cFT943Z1Ro1H+3+iNTsVBb1WMTsLZc4mXqDRUNaUduuZk7DvpOjlSk/DWlF0ORN3MgtvOu1nAIV/1t/olg/rDtvoN57L/XOBub33mhV3LXfvZXcvaGs59h56jKzt50lr1AN/BeaATLYFEKIKigu9QZHE9P5uLf/Iy+KYW9pyi/DQ3h54QFG/RrJzEHB9GxS9h5SVdaJP+HKKei/CJT6v8m27uw6Pt37Ke3c2zG943RMjB4+eLqcmcuEVdFsOXmZ1vUc+bp/EN5O/41TS1pkR2gpFAoGtvSidV1Hxi6L4o3fItl60pP/eyqw8hcd0oGq/x0IIWq848kZRCWmM/GJgBqxMphCoeCNTr7417Zh7O9RPDlrD3Oeb06oz0MGb9ErITMZ+oRXbKFlNO/YPHYk7WBC6wkkptTit/1HePUxHzo1qqWXegyRUqkg657g6bYrWfkMmheh44oeTU6Bimkb4yR8EkKIKmjZwURMjBQ83dyzQo5nZ2HCkmGtGbLoIG/+foQZag19gtwr5NgGSa2GnV+DUwMIeErf1fDPuX/4eM/HtHZrzYxOMzAzMnvoY607epFP/jxOTr6KT54IYGjbuihrWCPxilDX2YqVr7Vh5pbTzNp2hoPnr/Hts81oUcdB36U9EgmfhBBV3i8RFzA3UfJMi4oZBFUVnRu5subNdoxYfIgX5u9nYp8AXgytU74ATqOBPeHaad++XSuv2FLsTd7L7KjZ9PbpTRvnJ+kzaw8t6jjwTrea/ex/SdztLUgu4fl/Z2tTZg5qXvS1hjt6gd3TFuzOL+9tGXbn+4q/duf7NKW+du8GDRpe+elQsZqBatnLQAghqru8QhWrjiTRLcANRyvTCjuujbkJP7/SmlcWHWTs0iOo1Jrqe4Pi1D9wOQb6/QCP2Mz7Uf174V8m7JpAcK1gvuv8HebGD7ey8LXsfD5Zc5z10Sk087Lnm4FB1HexruBqaxYTIyXju/kR1tCFt5ZGMfCHfYzp3IA3OtWvsi0pJHwSQlRpGTkF/Bl1kaeCPLCzePgpwlVVfRdr1rzRjnFLo5j4ZwzHkzP4rG9jzIzLOJg5vQnSYqHfjzpvdpmclcx7u97D18GX91t+zAvzjmCkVPDdoGBMqugf1cr0bnc/JqyKJqfgvz5fFiZGfNw7wKAfN/UoJTSrqSvbCCFEVfbviUuk3yx46Ebj92NtZsxPr7Ri2E+HGLc8ikK1hv7V7caiRgM7p4FDXWjcX6+lbEvYxns73qOJcxNmd5mNpcnDtXDYFJPKh6ujycgp4L0efowM86my4YghalXXkX/eCmPimuN8u/kUO0+n8e3AZnc9ylhVyE+FEKJKWxWZRE6BihfbVN9G4w9ia27CvJdaMrqzL8sPJfHsDxFcyswt25v3hIOdFzR+unKLvEeeKo/x28ejUqv4tuO3fLvpPDEXM/lmQBAeEkqUqG+wR5Vs0vludz9Z2UYIIaqJZQcT8bC3oL2vc6Uc39LUmIVDWtHe15l3Vx5l2cGESjmP3pzZAhePQPvxYKS/eSA7k3Yyfsd4/J38mdN1DlYmVuU+RkZOAeOXRzFyyWFq2Ziz9s32jOroK8FTJbA1N2HGc8GEP9eMU6k36PXdLlZFJlXsytc6IDOfhBBVlkaj4ZeICwR52dPYw07f5eiVUqng7W5+BNS25e0VR+kzczffv9iC5t73eTY88SBc2AM9psAjNJZ8GF/u/5ITV08Q3imcmAumLN53geHt69E1wFWndVQ1t5t0ViW36522MY6L6Tm421vwbne/Kvd9CCFETZd0/Sa7z1xhTOcGGFViHx8LUyPmvdSSV5cc5v0/oilQaXghtBrcZNRoYOdUsPWEoEF6K2Pvxb2M2zaOBvYNmNt1LjamNuU+xo5Taby/8hhpWXmM6ezLm50bYGosoVNle6qZB829HRi/PIrxy4+yLS6N//VtXGWe/pCfECFElbUv/ipn07J5sToMSCpIzya1WTWqLeYmRjz3Q8T97xjumQHm9hD8ou4KBFadXsUfp/9gRJMR+FqF8v7KYwR52fNej0Y6rUPoTt9gD/Z80Jm/X/ZhzwedJXgSQogqaMWhJAAGtKz8R+HMTYz48aUWdGlUi4/XHOfnvecr/ZyV7vwuSNwP7d8C44rrl1UeB1IOMGbrGOra1eXHx3/Ezqx8N2+z8gr5cHU0Ly88gLW5MatHtWV8Nz8JnnTIy9GSpSPb8G53P/6JTqHnjJ1ExF/Vd1llIj8lQogq69eIBOwsTHiiaQ1YkrccGrnZsvbNdoT4OPL+H9FM/PM4BSr13TtdOQ0n10PrkWCmu4aQMVdj+Dzic0JrhzK88Wu8+XskCgXMGhQsAxchhBDCQKnUGlYeTqK9rzOeDrrpNWNmbMTcF1rQLcCVT9fGMH9XvE7OW2l2TgNrVwh+QS+nP3zpMG9ufRNPa0/mdZuHvbl9ud4fEX+VnuE7+f1AAiM7+PDX6PY09SzfMUTFMFJqV75e+XpbTI2VDJoXwdQNJ8kvVD/4zXokI30hRJV0OTOXjTGpDGzpibmJflcKMUT2lqYsGtKKEWH1WLzvAs/P38+VrLz/dtj7HRibacMnHUnPTWf8tvE4WTgxtcNUpm08zbGkDKYNCMLLseo1TRRCCCFqit1nrpCcnsOzldBo/H5MjZXMfr45vZq48b/1sXy/46xOz19hEvbDuZ3QdgyY6L63ZdTlKEZtHoWrpSvzu8/H0dyxzO/NLVAxed0JBs2LQKlQsPzVNnzYy1/G3wagmZc968eEMbCFF3O2n+WZuXs5m5al77JKJeGTEKJKWnowkUK1hsEh8shdaYyNlHzUO4AZzzbjaGI6T87czfHkDLiRCkeXau+8WbvopBaVWsUHuz4gLSeN6R2nc+BsHov2nGdI27p0D3TTSQ1CCCGEeDjLDybiYGnC43rozWhipOS754LpE+TOlH9OMmvraZ3X8Mh2TgNLJ2g5VOenPn7lOK9vfh1nC2cWdF+As0XZm8UfSbhOr+92sXDPOV4MrcM/Y8NoVbfswZWofFZmxnzVvynfv9CcxOs3eeK73fx+IMEgm5FLw3EhRJVTqFLz2/4Ewho4U8+5/Ktz1DR9gz3wrWXNyMWHeGbuXtY12kxDdSG0eVNnNcw9Opc9F/cwsc1E7JQ+DF6xiyYedkzoJX2ehBBCCEN2LTufTSdSeTG0LmbG+pntYmyk5NuBQRgrFXy96RSFag1juzRAoai8xucVJjkSzvwLXSaCqW7HrbFXYxn570jszOxY0H0BtSxrlel9eYUqwjef5vsdZ3GzNefX4SG0q6QVDkXF6NG4Ns28HHh7RRQTVkWz7eRlpjzTFEcr/fQXK4mET0KIKmfLycukZubyf08F6ruUKqOxhx1rR7fn7SW7cDv9KzFOnfGzq6OTPwI7Enfww7Ef6OvblyfrPc2zP0ag0cDswc31NogVQgghRNmsikyiQKXR+SN39zI2UvL1gCCMlApmbD5NoUrD290aGn4AtesbMLeDViN0etq4a3GM+HcEViZWLOi+ADerss00j7mYwdvLj3Iy9QYDW3ry8RMB2JpXjdXUajo3O3OWvBLCgt3nmLrxJD1m7OTp5h6sO5pya8XhFL2uOCzhkxCiyvkl4gK17czp0qhsd2+ElrO1GQsbH8foUg7vpXTC8aeDzBwUjL1l5d0RScxMZMLuCfg7+vNRyEd8vTGOqMR05jzfHG8n6fMkhBBCGDKNRsPyQ4kEednj52aj73IwUiqY+kxTTIwUzNp2hgK1mg96NDLcACr1OJz8Cx77AMxtdXbas+lnGfnvSMyUZizsthAP6weHDQUqNXO3n+W7LadxsDJl4ZCWdG6k+8csxaNRKhWM6OBDW18nhi46yPc7/mvUn5yew4RV0QB6CaCk55MQoko5dyWbXaevMKi1N8ZG8iusXArzMNo/F3w68tLTT7I//hpPztrDydTMSjldTmEO47aPQ4GC6R2ns+d0BvN2neOlNnXo1URWKBRCCCEMXVRiOqcuZfGcnmc93UmpVPB53ya8EOrNDzvi+d/6WIPsbwNoZz2ZWkPIqzo75bmMcwzbOAylQsn87vPxsn3wf7vTl27wzNy9TP/3FL2a1GbTWx0keKriAt3tMFYWD2VzClRM2xinh4pk5pMQoor5NeICxkqFQQ2CqoxjyyErFfrN5dn63vjWsuH1Xw7z9Jy9fDMgiJ4VGAhpNBo+2/cZp66fYnaX2ShVTry9YheB7rZ82Mu/ws4jhBBCiMqz7GAiFiZGPNHUsG4aKZUKPnuqMcZKJQt2n6NQpWbSk4GGNQMq7RTErIb2b4Glbpp0J2QmMHzjcDRoWNBtAfXs6t13f5Vaw4Ld8Xy96RTWZsbMeb653CCsRlIyckvcfjE9R8eVaMm0ASFElZFboGLF4SS6B7pRy9Zc3+VULWo17AkHt6bg0wmAFnUcWDe6PQ1dbXj910i+2RSHWl0xdw6Xxy1nXfw6Xg96ndDa7Rj9+xEKVRpmD24uS/MKIYQQVUB2XiHrjl6kd9Pa2Bhgzx+FQsGnfQIYEVaPn/dd4KM1xytsHFMhdk8HY3MIfUMnp0vOSmbYpmHkq/OZ120e9e3r33f/81eyefaHfXzx90k6NnRh41sdJHiqZtztLcq1vbLpPHw6c+YML7/8MkFBQbRv357w8HBUKtUD3xcdHc0rr7xC69atad26NUOGDOHo0aM6qFgIYSjWHb1IRk4BL4TW0XcpVc+pf+DqaWg3Fu64K+hqa86yV0MZ2NKTmVvPMGLxITJzCx7pVMfSjjHl4BTCPMJ4NehVvt4Ux+EL1/ny6SbUldUJhdAbGYMJIcpjfXQK2fkqg55trlAo+LCXP693rM9v+xOYsCraMAKoa+e0M85bvgLWLpV+utTsVIZtHEZ2QTY/Pv4jDR0alrqvWq1h8b7z9AzfRdylG0wfGMQPL7bAxcas0usUuvVudz8s7rnpa2FixLvd/fRSj04fu8vIyGDIkCH4+voyZ84cEhIS+Oqrr1Cr1YwbN67U96WkpDB06FACAgKYOnUqAAsWLGDo0KGsW7cODw/9dGsXQujWL/sT8K1lTaiPbqYuVxsaDeyeAfZ1IKBvsZfNjI346pmmNPawY/K6E/SdvYd5L7Wkvot1uU91Necq47ePx9XSlS/DvmRH3BV+2BHP4BBv+gS5V8R3I4R4CDIGE0KU17KDifi4WNGijoO+S7kvhULBe939MFEq+G6rtgn5tP7aVfH0Zve3oDSGtqMr7RTr49cTHhlOanYqSoUSY4UxP/X8CX+n0tsbJKfn8N7Ko+w5c5UODV346pkm1LbTzywYUfluNxWftjHu1mp3FjVntbulS5eSl5fHrFmzsLa2pl27dmRlZTFr1ixGjBiBtXXJ/0dn+/btZGdnM3v2bGxstKssBAcHExoayo4dOxg8eLAuvw0hhB5EJ2VwNDGdSX0CDOt5/qogIQKSDkCvr8Go5F/7CoWCl9rUpaGrDaN+jaTvrD2ED2pWrmaThepC3tv5Hul56SzpuYSbuaaMX76fRm42THwioKK+GyHEQ5AxmBCiPM5cvsHhC9eZ0NOAV5K7g0KhYHw3P4yUSr7dfAqVWsM3A4L0szhNeiJE/QYtXgbbynmMbX38eibtnUSuStvTR6VRYaQw4kLmBRo7Ny62v0ajYcWhJCb/dQKNRsOXTzfhuVZeVeK/rXg0fYM96BvsQWxsLP7++u27qtP/Ne7cuZP27dvfNcDp3bs3ubm5HDhwoNT3FRYWYmRkhIXFf6mspaUlRkZGhruygRCiQv0ScQELEyOebuGp71Kqnj3hYOkEzZ5/4K6hPk6sfbMd3k6WDPv5ELO3nSnz79mZR2ZyIPUAH4d+TAN7P8b8foS8QjWzn5c+T0Lom4zBhBDlsfxQEsZKBU83r1rjrrFdG/Budz/+jLrI2GVRFKjUui9i73eARtvqoBKo1Cq+PvR1UfB0W746n/DI8GL7X87MZdjPh3jvj2MEutuy4a0ODGrtLcGT0Dmdhk/x8fH4+Pjctc3d3R0LCwvi4+NLfV+3bt2wsLBgypQpXL16latXr/Lll19iZ2dHz549K7tsIYSeZdws4M+jyfQNdsfWABteGrTLsdp+T61fBVPLMr3F08GSla+1pU9Td6ZtjOON3yLJzisscd/18evptrIbTX5uwsLjC2nt1pq+vn35dvMpDp6/zhf9mjzU43tCiIolYzAhRFnlF6r543ASXfxrVck+QG908uXDXo1YfyyF0b8dIb9QhwHUjVQ4/DMEDQJ77wo5ZJ4qj8OXDjPv2Dxe3/w6YUvDuJJzpcR9U7NTiz7XaDT8GZXM49/uZM+ZK3zaJ4DfR4Ti5Vi28aA2mivYAAAgAElEQVQQFU2nj91lZmYWTdm+k62tLZmZmaW+z9XVlcWLF/Pqq6+yZMkSAFxcXFiwYAGOjg/u/ZKXl0dsbOzDF34fubm5lXZsUZxcb90ylOu95kQGuQVq2tZSG0Q9lamir3nt/Z9ha2TOGfvHUJXzuK8FmeFi7MiiyFROJF1jYidXatv8F/7turKLH87/QL46v2hb1OUoJvw5k6X7fOjewAY/80xiY0v//a5vhvIzXlPI9dYfGYOJRyXXW7f0eb33XMjmanY+bd0UVfa/eZgLjGzlxI8HU3nphx1MeMwVU6PSZ/pU1PWuFfUdjuoCztZ+koKHPF5WYRZxN+KIzYol7kYcZ7PPUqjR3gT0tPAkxD6E/df3c6PwRrH3Opk6ERsbS3quitkRV9h9IZtGLma83c4VT7tc4uJOPtL3V5Hkd4puGcL11mn49LAuX77M2LFjCQwM5H//+x8Av/32GyNHjmTp0qW4u9+/ia2ZmVmlPd9oCM9O1iRyvXXLEK63RqPh3793EOxtT5/2zfRaiy5U6DXPSIaETdBqGA2bhT7UIQIC4LGgNEb/foTxG1KZNag57Rs4k5qdypKjS+4KnkA75fuvKyto6Po/ZrzYDgtTw37czhB+xmuSyrze+h5QVVcyBhO3yfXWLX1e76kRB3CzNef5zsH66ZlUQfz9wdP9PBP/jGHGwSzmvtCi1DYAFXK9s6/AqjXQZAC+rbuV+W0Xsy4SeTmSI5eOEHk5kjPpZwAwVhoT6BTIi3VepLlrc5q5NMPe3B4o3vMJwNzInHdC3iEhx4GP1keTmVPIez38GBnmY5D/HeV3im4ZwhhMp+GTra0tWVlZxbZnZmZia2tb6vsWLFhAYWEh3333HSYm2rvuoaGhdO/enYULF/Lxxx9XWs1CCP3ad/Yq8WnZfDMgSN+lVD0Rc0CjhtBRj3SYDg1dmD+sPuP+/IPhf/+Gs3MyGYWppe6vMbrO7OeDDT54EqImkTGYEKIsUjJy2HEqjVEdfQ0ysCivl9rUxVip5MPV0YxYfIh5L7WsvD6UEXOgIAfC3i51F7VGzenrpzlyWRs0Hbl8pOhROWsTa4JqBdGzXk+CawXTxLkJ5sbmJR6nt09vgKLV7tys3BgR+AZbDnqw6shhAt1t+WV4EI3cSv/9LoSu6TR88vHxKdZXICUlhZycnGJ9CO4UHx+Pr69v0aAHwNTUFF9fXxISEiqtXiGE/i2JuIC9pQm9m1bOaiHVVs51OPwTNH4GHOqU++3JWckcSj3EwdSDHLp0iOSsZLAGC6y4cr0OgQ5duWq8gau5xXsO2JvWwrdW8cd7hBD6I2MwIURZrDyUhFoDA1t66buUCjM4xBtjpYL3Vx1j2M8Hmf9Sq4q/QZZzHfb/CAFPgYtf0eY8VR4xV2KIvBxJ5KVIoi5HcaNA+7hcLYtaNHdtTnBgMM1dm9PAvgFGyrLXVZDRjOwzH3AjPQcTK1M+O64mO/8iY7o04M1OvpgaV/3wUFQvOg2fOnTowIIFC8jKyipabeXvv//G3Nyc1q1bl/o+d3d3du7cSX5+PqampgDk5+dz+vRpOnXqpJPahRC6dykzl00nLjGsfT1ZLa28Di2E/CxoN+aBu2o0GpKzkouCpkOph7iYfREAezN7Wri24MWAF2np2pL6dr7M3R7PN//P3p2HR1Xe/R9/Z7LvQEjISsIWdmQL+6IIooAKsmgVW21rxVa0trW1fZ5Wq/1179Pa+vjYVq0VNzQsIsENsSCyhT0gkIRA9o2Qfc/M+f0xYQkkkEAyJ8l8Xtc1V5Iz58x8Zy5I7vmc+3zvT5PpF+2Om+9bNBi15x/LxfDgp5N+0GEvS0SujcZgInI1NpvB6r2ZTBkQRN+g7tWUellcFK4WF56MP8QD/9rDqw/E4evZjh+Fd/8D6sopnbyCQ1nb2J+/n/0F+zly5gj1tnoABgQOYG6/uYwNGcuYkDFE+EVc84pz6w9k89O1SVTXWwEoqqzDBXhiziAeuzm2vV6VSLtyaPh0zz33sGrVKlauXMlDDz1EZmYmL7zwAg888ECTpX/nzJlDXFwcv/71rwFYunQp8fHxPProo9x7770YhsGbb75JYWEhd999tyNfgog40Nt7MrDaDO6b2D6rhTiN+hrY9RIMnA2hIy+72zAMsiqymsxsyq3MBaCnZ0/Gh47n68O/TlxoHAN7DMTi0vTM2cqbBzE0LIDvvemC1XchnsEf4+JeglHfA9vZW6kv7f69uUS6Go3BRORqdqYVkVVczZNzB1995y5o8bhI3FxdeGL1QR741x7+9eAE/K4zgMqtyGV/9pcc+OpV9vcfROqWhzEwcLO4MSxoGMuHLmdMyBhGh4ymp1fPNj22zWZQWl3PmYpaCitqOVNRx5nyWs5U1PLajtPng6dzDGB1YpbCJ+m0HBo+BQYG8tprr/Hss8+yYsUKAgIC+MY3vsHKlSub7Ge1WrHZLiyJOWLECF5++WVeeOEFfvzjHwMQGxvLq6++ypAhQxz5EkTEQeqtNt7ek8GM2GCig3zNLqdrOfQ2VBbA1McBe9iUWZ7J3nx72JSYl0h+VT4Avbx6Ma7POB4c8aB9ZlOPAZeFTc2ZPawPgT4eFJSNoaFsTJP7/vDxCRaOiWj/1yUi10xjMBG5mtWJmQR6uzN3eKjZpXSYO0dH4Gax8Ng7B7j/ld38+5sTCPByv/qB2Ps1nSw5eX5W04GCA+dP3vl6uzG61yDmRt3I2D5jGdF7BN5u3pc9htVmcLayjqLKWs6U13GmovZCuHTRz2cqaimqqKPBZlz2GG4Wl2a3A+SUVLfh3RBxLIevdjdw4EBef/31K+6zZcuWy7ZNnjyZyZMnd1RZItLJfHYsn/yyWn61sO39ipyazYqx43kyIm4gsS6fvV88RWJeIgVVBYA9bBrfZzxxoXHEhcbRP7D/NU/5LiyvbXa7Bj4inZPGYCLSkpKqOj46msfX4qK6fauD+aPCcLXAo28dYPbf/0yN30ZsrsVYdvRkcb+HeHrW/QDUWes4WnS0SdhUXmfv1xTsHczYPmP5xpCvMfaT/0d0z+GkjXmJMxW1pGfXse9EzvlZSmcqLoRKZyvraC438nC10NvPg97+nvQJ8GJ4eAC9/TztN39Pevt5ENz4c6C3O9N//znZzYy3wntcHniJdBYOD59ERFrjjV0ZhAd6MWtIiNmldHqGYZBelk5ifiKJJ9az17eWQrdi2PUsQV5BxIXGnQ+c+gX2u+aw6VLhPbw18BEREekG1h/Ipq7BxrK47tNo/EpuHRHGjWM3sKv8HVws9bgAhlsx76X/iaS12/DxquNI0RHqbXUA9PaIItJ9En7ug7DU9aeiJIBDWXX0KV/LfUYhSwpnsPfE9ibP4eVuOR8gRfb0YUzfHhcCJT/P82FTbz9PArzc2jQ+e3Lu4CY9nwC83V277SWT0j0ofBKRTietsILtqWf40S2xuFraJyjpTgzD4FTZKfbm2ZuDJ+Yncqbavupcb8OFOJs74yf+F3FhE4gJiGm3sOlSGviIiIh0fYZh8E5iJiMjAhkeHmh2OQ6zp/RNXNzqm2xzsVg5XrYLW35frNWTsFZFY62OodzqyynAz9PNHhr5GQwO8uDhqg/I9B3HwslL+LafJ8H+HgT52mcr+Xq4dtgY7Fx7gz98fIKckmrCe3jz5NzBansgnZrCJxHpdN7cnYGbxcVpzr4BJKQl8Pz+58mrzCP0aCiPj32c+f3nA41hU+mp883BE/MSKaopAuzTvs9dQhdX70L0e9/CZcGfYUjHNwLWwEdERKTrS8ou5XheOc8tHGF2KQ5lcy2mpWjokdi/nJ+VZA+b7N97e1x0SWLiy3CqiIA7X2V5f8e3iVg4JkJjLulSFD6JSKdSXWflvb2Z3DoilBB/L7PLcYiEtASe2fEMNdYaAHIrc3l6x9PszNlJdUM1e/P3crbmLAAhPiFMCp90/jK6vv59L5xVe2Mx+AbDDfc6rHYNfERERLq21YmZeLpZuOOGcLNLcSiLtSeGW3Gz21fePOjKBzfUwfa/QOQE6DezgyoU6V4UPolIp/LB4RzKahpYPsl5Go0/v//588HTObXWWt4/+T59fPowJXyKfWZTnzgi/SObn8KddwRSN8Osn4O7c4R2IiIicn2q66xsOJjD/JFhBHq3btW37mJxv4d4L/1PuFgutA8wbO4s6ffQ1Q8+vBpKM2HBn6GDLq0T6W4UPolIp/LGrnQGhfgxsV8vs0txmLzKvGa3u+DCp0s+bV2/gC+fBw8/iPtWO1cnIiIi3dWmpFzKaxucqtXBOU/Pup/P3nqd4roCDGxYrD1ZctFqdy2yNsAXf4Kw0TBwtmOKFekGFD6JSKdxKLOEw1ml/PKO4R3WoLEzCvUNJbcyt9ntrXofSjLgyBqY9Ah49+yACkVERKQ7Wr03k5ggH6c66XfO8bPHKa7P46mJTzGWsQwdOrR1Bx5dC8Wn4O43NetJpA0sZhcgInLOG7vS8fFwZdFY5+ohtGLUisu2ebl68fjYx1v3ADv/1z74mfTddq5MREREuqu0wgr2nDrLsrgopzrpd058cjyerp4s6L+g9QfZbLDtjxAyDAbP67jiRLohhU8i0imUVtWz4VAOd46OIMDLuXoOWCz2X8VBXkG44EKYbxjPTHnm/Gp3V1R1Fva/DiOXQaBzhXYiIiJy7d7dm4WrxYUlYyPNLsXhquqrSEhLYG7MXAI9A1t/4LENcOYEzPgRWPRRWqQtdNmdiHQK7+3LpLbBxvJJfc0uxeHik+OJCYhhw8INHD9+vPXTvgH2/BPqq2DqYx1XoIiIiHQrDVYba/ZncdPgYEICnG+hko9Pf0xFfQVLYpe0/iDDsM96ChoEwxZ2XHEi3ZTiWhExnc1m8ObuDMb27cHw8DacfeoGUopTOFR4iCWxS9o+5b2uCvb8HWJvhZA2BFYiIiLi1D4/UUhheS13xznfST+wn/gbEDiA0cGjW39Q8keQnwTTfwgW144rTqSbUvgkIqbbcbKIU2cquX9ytNmlONyalDW4W9y5Y8AdbT/44JtQVQRTv9/+hYmIiEi3tToxg2B/T24aHGx2KQ534uwJDp853LYTf4YBW38PPaJhZBtmS4nIeQqfRMR0b+xKp6ePO7eNCDO7FIeqaahhw8kNzO47m55ebVylztoAO/4GkROg76SOKVBERES6nYKyGj4/UcjisZG4uTrfx8H45Hg8LB7cPuD21h90cgvk7IfpPwBX5+pNKtJenO+3jYh0KnmlNXx6LJ9lcVF4uTvXFOZP0z+lvK6cxbGL237wV+uhJB2mfV/L/IqIiEirxe/PwmozuDsuyuxSHK66oZqNaRu5JeaW1jcaNwzY9gcIiIAbvtaxBYp0YwqfRMRUb+/JwGYY3DfB+S65i0+Op69/X+JC49p2oGHAl89D71iIva1jihMREZFuxzAM3k3MZEK/XvTr7Wt2OQ53TY3G07+EjJ32Ngdunh1XnEg3p/BJRExTb7Xx9p4MZsYG0zfIx+xyHCqtJI39BftZHLsYi0sbfxWn/QfyDsOUx7TMr4iIiLTa7lNnOV1Uxd3jnW/WE9hP/PUL7MfYkLGtP2jr78E3BMbe33GFiTgBfWoREdNs/iqfgvJalk90vllPa1LW4Obidm2Nxr/8C/iFwqhl7V+YiIiIdFvvJmbi7+nGvJHO1WcTILk42b7C8KA2NBrP3AOntsLUx8Ddu2MLFOnmFD6JiGlW7Uonooc3Nw0JMbsUh6q11rLh5AZu6nsTvb17t+3gnIP2mU+THtHUbxEREWm10up6EpJyuWN0ON4eztVnE2BN8jWsMLztD+DdC8Y92HGFiTgJhU8iYorUggp2nCzi3ol9cbU4V8Psz9I/o6S2pG39Bs758nnwDIDxGgSJiIhI6204lENtg81pG41/kPYBc6Ln0MOrR+sOyjkIKZ/A5O+Bp1/HFijiBBQ+iYgp3tydjrurC8ucsOfAmpQ1RPhFMClsUtsOPHvKvsrd+AfBq5UrtIiIiIhgv+RuaFgAIyOcbwxxboXhNp342/YH+3hrwkMdV5iIE1H4JCIOV1XXQPy+LG4dEUawv3NdOpZels6evD0sHnQNjcZ3vgAWN5j4SMcUJyIiIt3S0ZxSkrJLuXt8ZOv7HXUj8cnxxATEML7P+NYdkP8VHN8IE1fohJ9IO1H4JCIO98GhHMprGrh/knM2Gnd1cWXhwIVtO7DyDBx4A0bdDQHO1yRURERErt27iZl4uFlYOCbC7FIcLrU4lQMFB1gS24ZG41/8ETz87OGTiLQLhU8i4lCGYbBqVzqxffyIi+lpdjkOVW+t5/3U95kZOZNgn+C2Hbz779BQC1Mf75jiREREpFuqqbey7kA2tw4PpYePh9nlONyalDY2Gj+TAkfWQty3wadXxxYn4kQUPomIQx3KKuVIdhn3T4p2umnfWzK3cLbmbNsbjddWwJ5/wJD50HtQxxQnIiIi3dLHR/Moq2lwykbjNQ01bDi5gdl9Z9PTq5UnPb/4H3DzgsmPdmxxIk5G4ZOIONQbu9Lx8XB1ymnfa5LXEOYbxpTwKW078MAqqCmBqd/vmMJERESk21qdmElUL28m9w8yuxSH+zT9U8rqylp/4q/4NBxebV/cxa+Ns9RF5IoUPomIw5RU1fHBoRwWjYnA38vd7HIcKrM8k525O1k0aBGuFtfWH2ith53/C32nQFRcxxUoIiIi3U56USU7ThaxbFwUFotzzTgHe6Px6IBo4kJbOYba/mewuMKUlR1bmIgTUvgkIg4Tvy+L2gYby52w0fjalLVYXCwsGriobQceWQulmTBNs55ERESkbd7bm4XFBZaMjzS7FIc7WXKS/QX7WTKolY3GS7PhwJsw5n4ICO/4AkWcjMInEXEIm83gjV3pjI/uydCwALPLcah6Wz3rU9czPWI6ob6hrT/QMODL5yF4KAyc03EFioiISLfTYLURvy+LmbHBhAV6m12Ow8Unx+NmceOOga1sNP7l84ChE34iHcTN7AJExDl8efIMp4uq+P7sWLNLcbhtmds4U32m7Y3GUzdDwVFY+BJYdK5AREREWm9bSiF5ZTU8c8cws0txuFpr7flG4728rrBi3eF34bNnGVKaBRgQPRV69HVYnSLORJ9mRMQhVu1Mp5evB7eNbMPMn24iPiWeEJ8QpkVMa9uBXz4PAREwYnHHFCYiIiLd1urETIJ8PZg1pI/ZpThcqxqNH34XPngMSjNxwbBvy95n3y4i7U7hk4h0uNzSajYfy2fZ+Cg83drQbLsbyKnI4cvsL1k0cBFuljZMNs3aB6e/gMnfAzePjitQREREup3C8lo+O1bA4nGReLg530e++OR4+vr3vXKj8c+ehfrqptsaauzbRaTdOd9vIhFxuLd3Z2AA9010vmnMa1PWAnDXoLvaduCXfwGvQBj79Q6oSkRERLqztfuzaLAZLBsfZXYpDpdWmsa+/H0sjl2MxeUKH3dLs9q2XUSui8InEelQ9VYbbydmcmNsMFG9fMwux6EabA2sS13HlIgphPu1YdWUM6lw7AOIewg8/TuuQBEREel2DMNg9d5MxkX3ZGCIn9nlONya5DW4Wdy4c8CdV94xsIUVAFvaLiLXReGTiHSoT47mU1hey/JJ0WaX4nDbs7dTUFXA0kFL23bgzr+BqwdMfLhjChMREZFua196MWmFldwd53yzns41Gp8VNYsg76Ar7zzhO5dvc/eGm3/RMcWJODmFTyLSod7YlU5ED29uHBxidikOF58cT2/v3syImtHqY1yri+Dg2zDmPvBzvvdMRERErs87iZn4ergyf2SY2aU43Gfpn1FSW3L1FYYbaiHpPXD3Bf9wDFwgMApu/yuMWuaYYkWcTBu634qItE1qQTk704p4cu5gXC0uZpfjUHmVeXyR/QXfHPFN3C3urT6uV8q7YKuHyY92YHUiIiLSHZXX1JNwOJc7R4fj6+l8H/XiU+KJ9ItkYtjEK+/42bOQdxjueQuGzOf4sWMMHTrUMUWKOCnNfBKRDvPGrgzcXV2cctr3utR12Axb6xuNH34X/mcYQcf+Da6e9qV+RURERNpg4+FcquutTjn2Ol16msS8xKs3Gk/ZDDtfgLhvw5D5jitQxMk5XxwuIg5RVdfAmn1Z3DYijN5+nmaX41BWm5W1KWuZHDaZKP9WDP4OvwsfPAb11bgANFTbfwZN/RYREZFWeycxk9g+foyO6mF2KQ63JmUNbi5uLBy4sOWdKgpg/QoIGQa3/MpxxYmIZj6JSMfYcDCH8toG7p/sfI3Gd+TsIK8yj8Wxi1t3wGfPQn1102311fbtIiIiIq1wIq+cQ5klLBsfhYuLc7U7qLPW8X7q+9zU9yZ6e/dufiebDdY/ArXlsPgVe3NxEXEYzXwSkXZnGAardqUzuI8/46N7ml2Ow8Unx9PLqxezoma17oDSrLZtFxEREbnE6sRM3F1duGtspNmlONyWjC0U1xazZNAVGo3v/j9I3Qzz/gh9hjmuOBEBNPNJRDrAwcwSjuaUsXxytNOdeSusKmRr1lbuHHAn7q6taDRedRYsLZwHCHS+waOIiIi0XW2DlbUHsrhlWCi9fD3MLsfh4pPjifCLYFL4pOZ3yD0Enz4Ng+fbez2JiMMpfBKRdrdqVzq+Hq4sGhNhdikOtz51PVbD2rpG41Vn4fU7wbDZm4xfzN0bbv5FxxQpIiIi3cqnX+VTUlXPMidsNJ5els7uvN0sHtRCo/G6Soj/Fvj2hjv+Bk52YlSks1D4JCLtqriyjo2Hc1k0NgI/J1vi12bYWJOyhgmhE4gJjLnyztXFsGohFJ6A+96FO1+AwCgMXCAwCm7/q5qNi4iISKusTswkooc30wa20O+oG1uTvAZXF9eWG41/+BMoSoW7/gG+QY4tTkTOc65PhiLS4d7bl0ldg43lk5yv0fiu3F1kV2Tz2JjHrrxjdQm8vhAKjsE9b8HA2fbto5Zx/Ngxhg4d2vHFioiISLeQVVzF9tQzPDZrEK4W55rVU2etY33qem6MupFgn+DLdzi6Dg6sgmk/gH4zHF+giJynmU8i0m5sNoM3d2cQF9OTIaEBZpfjcPHJ8fTw7MHN0Te3vFN1CaxaBPlH4e43YNAcxxUoIiIi3c57e+0LlCwd73y9Irdk2huNL41devmdJRmw4XGIGA83/czxxYlIEwqfRKTdfJF6hvSiKqec9XSm+gyfZ3zOHQPuwPPS/k3n1JTCG4shLwnuXgWxcx1bpIiIiHQrVptB/L4spg3sTWRPH7PLcbj45HjCfcOZHD656R3WBljzkL2v5uKXoTWLwIhIh1L4JCLt5o1d6QT5enDriFCzS3G4DSc30GA0sHjQ4uZ3qCmzB0+5h2DZ6zD4NscWKCIiIt3O9tQzZJdUc7cTNhrPKMtgd+5uFsc202h82x8gcxcs+B/o1c+cAkWkCYVPItIuskuq+exYPsviovB0czW7HIcyDIM1yWsYGzKW/j36X75Dbbk9eMo5AEtfgyHzHF6jiIiIdD/vJmbS08edOcP6mF2Kw61JaaHRePoO2PZ7GHWPFm8R6UQUPolIu3hnTwYGcO+EvmaX4nCJeYlklGewJHbJ5XfWlsMbSyB7Hyz5Fwxd4PgCRUREpNs5W1nHJ1/lsWhMpNOd+Ku31rM+dT0zI2cS4hNy4Y7qYvvldj2iYf4fzStQRC6j1e5E5LrVNdh4e08mNw0OIaqXc/Yb8PfwZ070Jc3DayvgzaWQlQhLXoVhd5hToIiIiHQ7a/dnUW81nPKSu88zP+dszdmmJ/4MAz54HCry4FufgKe/eQWKyGU080lErtsnX+VxpqKW+52w0XhxTTGbMzZzx4A78HLzunBHXSW8tQwy98CSV2D4wpYfRERERKQNDMPg3b2ZjI7qweBQ5wtZ4pPjCfMNY0r4lAsb978OX70Ps/4bIsaZV5yINEvhk4hct1U704ns6c2M2GCzS3G4DSc3UG+rb9povK4S3robMnbC4n/C8EXmFSgiIiLdzsHMEpLzK5xy1lNmeSY7c3dy16C7cLU0Xm5YmAwfPQX9ZsKUx80tUESapfBJRK5LSn45u0+d5b6J0bhaXMwux6EMwyA+OZ4bgm9gUM9B9o11VfbgKf1LuOufMKKF1e9ERERErtHqxEy83V1ZMCrM7FIcbm3KWiwuFhYNbDy511ALa74Jbl6w6O9g0Udckc5I/zNF5Lq8sSsdD1cLy8ZHml2Kw+3L38fpstMX+g3UVcHbjcHTor/DyGYakIuIiIhch8raBj44lMOCUWH4e7mbXY5D1dvqWZeyjhmRM+jj27jC3+ZnIC8JFr4IAc4Xxol0FQqfROSaVdY2sHZ/NvNGhhLk52l2OQ63JmUNfu5+3BJ9C9RXwztfg1NfwML/09K+IiIi0iESDudSWWd1ykvutmZupaimiKWxS+0bUj6FXS/ChO/A4NvMLU5Erkjhk4hcs/cP5lBe28ByJ2w0XlpbyienP2F+//n4YIF37oW0rfazbjfcY3Z5IiIi0k2t3ptJ/2BfxkX3NLsUh4tPjqePTx+mhk+F8nxYtwJChsOc58wuTUSuQuGTiFwTwzB4Y1c6Q0L9nXLwszFtI3W2Opb0vx1W3wcnP4c7/xdG32t2aSIiItJNpRaUsy+9mHvionBxca5em1nlWezI2cHiQYtxxQXWr4C6CljyKrh7Xf0BRMRUCp9E5Jrszyjhq9wylk+KdrrBz7lG4yOChjHkk+cgdTPc8TcYc5/ZpYmIiEg3tjoxEzeLC3eNdb5em2tT1uLi4sKiQYtg1//CyS0w99cQMsTs0kSkFRQ+icg1eXNXOn6ebiwcE2F2KQ53qPAQqSWpLCk+C6mfwu1/hbH3m12WiIiIdGN1DTbW7s/m5qEh9HayXpv1tnrWpa5jesR0QktzYfMvYcgCGP9Ns0sTkVZS+CQibXa2so6Nh3NZNCYCP083s8txuPgT7+KDhfPlQy8AACAASURBVNvSEmHBX2DcN8wuSURERLq5LcfzKaqs4564vmaX4nDbMrdxpvoMS/rNh/hvgW+wfda5k82+F+nKnO9To4hct/f2ZlJntTllo/GyqiI+TtvIgrIyfOb9CcY/aHZJIiIi4gTeScwkNMCLGbHBZpficO+lvEeITwjTDn0AZ9PgGx+ATy+zyxKRNtDMJxFpE5vN4M3dGUyI6cXgUH+zy3Esaz0J6+6lBoMlox6CuG+ZXZGIiIg4gdzSarYlF7JkXCSuFuea7ZNdkc2O7B3cFTgUt0NvwvQfQr/pZpclIm2k8ElE2mRbSiEZZ6tYPtnJZj1Z6zHee4D4ytMM9Qxm+Mz/MrsiERERcRLxe7OwGbBsfJTZpTjc2pS1uODCXQc3QGQc3PiU2SWJyDVQ+CQibfLGrnR6+3lw6/BQs0txHGs9xH+TI2mfkOzpwZIxK8yuSERERJyEzWawem8mUwYE0TfIx+xyHKrB1sC6lLVMs7kTZjVg8cvg6m52WSJyDRQ+iUirZRVXseV4AXfHReHh5iS/PqwNsObbcGwD8UNvxNvNm3n95pldlYiIiDiJnWlFZBVXc3ec88162pa1jcLqMywpzIIFf4aeMWaXJCLXyEk+PYpIe3h7TwYG8LUJTrLKirUB1j4EX62n4uan+bAqnVtjbsXPw8/sykRERMRJrE7MJNDbnbnONOu8UfyhfxLSYGX6wIUwconZ5YjIdVD4JCKtUtdgY3ViJjcPCSGypxNM+bY2wLqH4ehamPMcm0IiqW6oZkmsBj4iIiLiGCVVdXx0NI+Fo8Pxcnc1uxyHyj1zjO1FSSxqcMNt/h/MLkdErpPCJxFplY+O5nGmoo77JjlBo3GbFdY/AkfiYfYzMPUx1qSsYVDPQYzsPdLs6kRERMRJrD+QTV2DjWXOdsmdYbD2o5UA3HXzH8HTyVZYFumGFD6JSKu8sSudqF7ezBwUbHYpHctmhfXfhaR34eZfwLQnOFp0lK+KvmLJoCW4uDjX8sYiIiJiDsMweCcxk5ERgQwPDzS7HIdq2Psqa2uzmeobRfjAW8wuR0TagcInEbmqE3nl7Dl1lvsmRmOxdOPwxWaF978Hh9+BWf8N038IwJrkNXi6erJgwAKTCxQRERFnkZRdyvG8cueb9VRwnO1bf0mBmxtL4n5odjUi0k4UPonIVb25Ox0PNwvLxnfjwY/NBhtWwqG34ab/ghlPAlBVX8WmU5uYGzOXAI8Ak4sUERERZ7E6MRNPNwt33BBudimOU18Da75FfIAvwV69mNF3ptkViUg7aTF8OnXqFHfddRdbt25t8eCtW7dy1113kZmZ2SHFiYj5KmsbWLs/m/kjw+jl62F2OR3DZoMPVsLBN+HGn8LMH5+/66PTH1FZX6lG4yLiMBqDiUhNg40NB3OYPzKMQG93s8txnM1Pk3fmGF94urNw0GLcLU702kW6uRbDp3/961/4+Pgwc2bLafPMmTPx9fXl5Zdf7pDiRMR86w9mU1HbwPLu2mjcZoONj8OBN2DGj+HGp5rcHZ8cz4DAAYwOHm1SgSLibDQGE5Ht6ZWU1zY41yV3yR/D7pdYO2QmBnDXoLvMrkhE2lGL4dP27dtZvHjxVR9g8eLFbN++vV2LEpHOwTAMVu1MZ2hYAGP79jC7nPZns0HCE7D/dZj+I7jpZ03uPnH2BElnklgcu1iNxkXEYTQGE3Fe6w9kM/W3W/jT9kJcLS7kllSbXZJjlOfB+kdo6DOCtZQxJXwKkf6RZlclIu2oxfCpoKCAqKirJ+2RkZEUFBS0a1EiYq5zA595r5/ieF45IyICul/4Yhiw6Yew7zWY9gN7g/FLXmN8cjweFg9u73+7OTWKiFPSGEzEOa0/kM1P1yaR3Rg4WW0GP1t3hPUHsk2urIPZbLDuYair4svp3yW/qkDtDkS6oRbDJz8/P4qLi6/6ACUlJfj6+rZrUSJinksHPgAfHMrpXgMfw4BNP4K9r8LU78PNv7gseKpuqCYhLYHZ0bPp4dUNZ32JSKelMZiIc/r9x8eprrc22VZdb+UPH58wqSIH2fk3SPsP3Pob4vN3EeQVxMwoNRoX6W5aDJ9Gjx7Npk2brvoACQkJjB6tXigi3cUfPj5x2cCnpt7WfQY+hgEf/hgSX4YpK2H2M5cFTwCfnP6E8vpynXkTEYfTGEzEudTUW1m1K52ckppm78/pzpfeZe+Hz56FobeTN+RWtmVvY9GgRWo0LtINubV0x4MPPsgDDzzAgAEDeOSRR3B1dW1yv81m48UXX+Tjjz/mtdde6+g6RaSDWW0Gu08VNZnxdLFuMfAxDPjoKdjzD5j8KMx5rtngCeyX3MUExDC+z3gHFykizk5jMBHnUFpVz6pdp3ltx2nOVNTh7urCbcYX/NjtXcJdzpBj9Ob3DcvYFzDH7FI7Rm05rPkW+PWB2//KupTV2AybGo2LdFMthk8TJ07kpz/9Kb/73e945513mDx5MuHh4QDk5uayc+dOiouL+elPf8qECRNa/YSpqak899xzHDx4EH9/f5YuXcqjjz562cDqYn/729944YUXmr3vBz/4AQ8//HCrn19ELrDaDPacOktCUg4fHcnnTEUtLoDRzL7hPbwdXV77Mgz4+Gew+yWY9F245VctBk+pxakcLDzID8f9sPv1uhKRTk9jMJHuLbe0mle+OMXbezKorLMyMzaYFTMH4Hb0PUbsexlvlzoAIl3O8Dv3lzkyLAaYZWrNHWLTj6H4NHxjI1avQNamrGVy2GSi/J1ohT8RJ9Ji+ATw9a9/nfHjx7Nq1Sr27t3Lxx9/DECfPn2YMWMGy5cvZ+jQoa1+stLSUh544AEGDhzIiy++SEZGBr/73e+w2Ww88cQTLR63dOlSpk+f3mTb5s2b+ec//8mMGTNa/fwiYg+cEk+fZVNSLpuS8jhTUYuXu4VZQ0KYPzKcytp6nt7wVZNL77zdXXly7mATq75OhgGf/DfsehEmroC5v24xeAJYk7IGN4sbdwy8w4FFiohcoDGYSPeTWlDOS1vT+PDgaXobxTw80IW7BroQ6ZoKyWvg8L+hMXg6x9uljriTfwW6WdCbFA+H3oIZP4aYqXyZtY28yjyeHP+k2ZWJSAe5YvgEMGzYMH7zm9+0y5O988471NbW8sILL+Dn58fUqVOpqKjghRde4KGHHsLPz6/Z40JDQwkNDW2y7cUXX6R///5tGniJOCurzWDvucDpSB6F5bV4ujUGTqPCmDUkBB+PC78OPNxc+cPHJ8gpqSa8hzdPzh3MwjERJr6C62AY8OnPYecLMOE7cOtvrxg81Vpr2XByA7P7zqaXVy8HFioi0pTGYCJdUF0VlOVAeY79a1k2BdlpZGek4VaRw1MuxfzRvdS+b0bjDcDDHxpaaHFQmgUJP4LhC6HvZLC0PFuxSyg+DRufgKiJMPMngL3dQS+vXtwUdZO5tYlIh7lq+NSetm3bxrRp05oMcObPn88f//hH9uzZw6xZrZtOWlxczI4dO1ixYkVHlSrS5dlsBnvTixtnOOVS0Bg43TT4QuDk69n8r4CFYyJYOCaCY8eOde0PF4YBm5+GHX+DuG/Dbb+/YvAE8Gn6p5TVlbE4drGDihQR6Xgag4m0g5qy84GS/Wsz39eUXHaYu+GHj0sQ7kGR+EVNh6C+EBDeeIsA/zDwCoA/j4DSzMuf180bDqyCxH+CbwgMuwOGLYToKV0viLLWw5pvAy5w1z/B1Y38yny2ZW3jgeEP4O6qRuMi3VWL4dOsWbNa7HXi5uZGr169iIuLY/ny5YSEhLTqydLS0pg0aVKTbeHh4Xh7e5OWltbqgc8nn3xCfX09CxYsaNX+Is7CZjPYn1HMxsO5fHgkl/yyWjzcLNw0OJj5o8KZNSQEvxYCp27j8Lv2VVNKs8DTz97Mcvw34bY/XDV4AvuZtyj/KCaEtr6PiohIe9IYTMTBDAOqiy8Kks59zW0aLtWVX36sb4g9ROoZA9FTsPqFsr/Yh3dTrCSe9cYSEMHy6UO4Z0JUk1nmzbr5F/DBY1B/0Qwod2+4/a8weB6kfAJfrYcDb9pX7fUNgaG322dERU/tGkHUf34LWYmw5FXoGQ3A+tT1WA0riwfpxJ9Id9bib8C5c+e2eJDNZqOwsJD33nuP9957j7feeot+/fpd9cnKysrw9/e/bHtAQABlZWWtLNm+tPDw4cOJiYlp1f61tbUcO3as1Y/fFjU1NR322HI5vd+XsxkGxwtr2Xa6gu3plRRVWXG3uDA+wpsHbghhQpQPPu4WoJTMtNI2PXZXe78D0j8iLPG3WKyNSxXXlmO4uJLjGkXZiRNXPT67Opt9+fu4N/JeThy/+v4doau9512d3m/H0vvdOhqDtY7+PTlWl32/DRuutcW4VxXiVl2Ae1VBM18LsVhrmx7mYqHBK4gG7xDqfcJp6Dmaep8QGnxCqPcOpsEnhAav3hiuHgBU19v4KKWcdftLKKy0EtPDnaVTezCjnx9ulhrST6ZcvVb3kQSM+wnBh1/CvSqfep8+FI5aQZn7SEjLBNehMHIoLkO/j1/uDgIyt+B34E0se1+hwbMn5ZE3UhY1i6rgMWDpfCcbfQr20feLP1HabwG5rsPg2DFsho3VX61mZMBIKrIrOJZtzr+xLvvvuwvTe+5YneH9bvG30k9+8pOrHlxVVcXy5cv585//zF//+td2LawlBQUFJCYm8qMf/ajVx3h6enbYpUNd/rKkLkbvt53NZnAgs4SEw/ZL6vLKavBwtTBzcDALGi+p8/e6/mnLXe79/mgpnAueGrkYViKOv0rEbT+46uEJiQm4ubjx0JSH6O3du6OqvKIu9553cXq/Hasj32+zB1TtSWOw1tH/X8fqlO+3tQEq8i+ZrXTxZXA5UJ4Ltvqmx1ncISDMfslbyMQLl78FhIO//XI4F78+uLu64Q5cab3foopa/r3jNP/emUlpdT0TYnrx+6UDuHFw8LWtmDt0KNz6xPn3OwJotuvmyLHAo1BXCSmf4vbVenomf0zPk+vApzcMXWC/NC9mOrh2giCq6ixs+n8QNIAe9/ydHp72S4C3Z2+nsK6Qn0z+CUNjzPv31Sn/fXdzes8dqzOMwa7rN5GPjw8PPvggv/rVr1q1f0BAABUVFZdtLysrIyAgoFWP8eGHH2IYBvPmzWtTrSJdnWFcCJw+TMolp9QeOM2IDeYntw3m5qF9CGiHwKlLsdmgKAUy90DWnub7JID9EryrqLPW8f7J97mp702mBU8iIq2lMZh0ew219uCoSW+lSy6Dq8gDw9b0ODevC0FS9OSmwdK57316g8VyXeVlnq3in1+k8e7eTGrqbcwZ1ocVMwcwLrrndT1um3n42i+7G77Q3uw89VM4uh4Ovwf7XgOfIBiywH5/zAxzgijDgA0robIQvrbZ3hah0blG47OiWnfpr4h0Xdf926d3795UVla2at/+/fuTlpbWZFtubi7V1dX079+/VY+xadMmxo0bR1hYWJtrFelqDMPg4LnA6Uge2SXVjYFTb340dzCzhzlZ4FRTBtn7LoRNWYlQ03gpoVcP+4Czoeby4wIjr/rQWzK2UFJbon4DItJlaAwmXVZdVWOw1NxspcbvKwsvP87D70KQNGBW00Dp3PfePVvV4/FafZVTxktbT5KQlIvFBRaNieA7M/ozMOTyy1odzsMHht1pv9VVQepme4+opHjY/2/w7nVhRlS/GeCo5t57X4XjG+GWX0H46PObC6sK+U/mf/j68K+r0biIE7ju8OnIkSOEh4e3at8ZM2bwyiuvUFFRcX61lU2bNuHl5cWECVdv7puVlcXBgwd5+umnr6tmkc7MMAwOZZWScDiHTUn2wMnd1YUZg4L5wZxYZg/rQ6C3E/yBNgwoSr0QNGUmQsFXgAG4QPAQ++AqcgJETYCgQXAkvvlGnTf/4qpPF58cT7hvOJPDJ3fYSxIRaU8ag0mndI0rwuHd80KIFD7mwipwF4dLXq2bpdfeDMNgZ1oRL21NY1tyIb4ernxzagzfmtaf0EAvU2q6Kg+fxlXx7rCPi1I322dEHVkL+1+3v9/nZkT1m9lxQVTBMfj4Z/awcNL3mtylRuMizuWaw6fa2lo+//xzXnrpJR544IFWHXPPPfewatUqVq5cyUMPPURmZiYvvPACDzzwQJOlf+fMmUNcXBy//vWvmxy/adMm3NzcuPXWW6+1bJFOyTAMDmeVsikpl42Hc88HTtMG9uaJObHMcYbAqbbCPqvpXNCUtce+8gyAZyBEjrcPoCLj7N97BV7+GKOW2b+eW+0uMNIePJ3b3oKMsgx25+1m5ZiVWFyubxq+iEhH0xhMTGEY9r49TWYr5V4eLjW7Ilxw44pw0c1cCtcYMnn4OP41XYXVZvDJ0Txe2nqSQ1ml9Pbz4Mm5g1k+MZpAny40LnP3tq+KN/T2xiDqM/uMqKPr4cCqxiBqPgxbBP3bMYiqr4H4b4GnPyx8qcmljjbDxpqUNUwMnUh0QHT7PJ+IdGothk+TJk1qsUmezWajoqICq9XKvHnzWLFiRaueLDAwkNdee41nn32WFStWEBAQwDe+8Q1WrlzZZD+r1YrNZrvs+ISEBCZNmkSvXr1a9XwinZlhGCRll5KQlEvC4Vyyiqtxs7gwbVBvvj97ELcMC+1aA5u2MAw4m3bJrKajF/o29B5sHwSdm9XUe3DrezOMWnbVsOlSa1LW4OriysKBC9v4QkRE2p/GYOJwNhtUnbnibKXBpdlwyYpwuFjAL9QeIgUPbv5SOP9QcPM053Vdo5p6K+sOZPOPbWmcOlNJdJAPv1o4giXjIvFydzW7vOvj7m2/9G7oAns4dPIzewh19H048Ia9jcHFM6LcPK79uT79uX18d188+PdpctfOnJ1kV2Tz/bHfv84XJCJdRYvh03333dfiwMfV1ZVevXoxfvx4BgwYQH19fbP7NWfgwIG8/vrrV9xny5YtzW5///33W/08Ip2RYRgcyS6zB05JOWSetQdOUwf25rGbB3HLsD708LmOP/KdVV0lZO9vOqupqsh+n2cARIyDGU/aw6bIcfYzcA5Sb61nfep6ZkTOIMQnxGHPKyLSEo3BpF1dbUW48hz7DKYrrQgXNpri4EkExYxsGiz59ekcK6m1k7Kaet7clcGrX56isLyWkRGB/O+9Y7l1RCiulo7rI2Uady/7yb4h8xuDqC32GVHHNsDBN+yzzIc09ojqf2PbgqgTH8Kef8Ck78KgOZfdHZ8cT0/Pnszqq0bjIs6ixb8Wl54Ju5RhGOzatYtXX32VTz/9lD179rR7cSLdgWEYHM0pOz/DKeNsFa6NgdPKmwZxy/BuFjgZBhSftjcDz9wDmbsh/ygYVvv9QYMg9lb75XNRE+y9myzmnUX8PPNzztacZUnsEtNqEBG5mMZg0mqXrQjXzKylFleEawyR+rZuRbiCY8cI6qbLoheU1fDKl6d4a1cG5bUNTBvYm7/cPZopA4JaDIK7HXcvGDLPfmuobQyi3odjG+Hgm/YgavB8+4yo/jddOYgqy4X134XQkTD7mcvuPtdofPmw5Xi4dqMxsIhcUZtPVRw8eJCNGzfy0UcfUVRURGBgIPPnz++I2kS6LMMw+Cq3jITDuSQk5ZJeZA+cpgwI4ns3DeCWYaH09O0mf2zrqyHngD1kOjer6dwKNR5+EDEWpv+gcVbTePDpXJdsrElZQx+fPkwNn2p2KSIiV6QxmJPpkBXhGmcydfCKcF1FWmEF/9iWxtr92TTYbNw2MowVMwYwMrKZvpLOxM0TBt9mvzXUQtp/7JfmHU+AQ2/Ze3EOmWdf+GXALPv+h9+90HPT1cMeeC5+tdlLLt8/+T4NRoMajYs4mVaFTydOnCAhIYGEhARycnJwd3envr6ep556ivvuuw83t+4z3VbkWhmGwbHcchKSckg4nMvpiwKnR2YO4JbhofTqKoFT4wBiyKVNuw0DSjIuzGrK2gN5SWBrsB/XawAMnH1hVlPIMFNnNV1NVnkWO3J28N0bvotrJ65TRJyXxmDdVEsrwpVf1MD73KIbF/PqcSFYCht9+WwlE1eE60oOZpbw0n9O8vFXebi7Wlg6PpLvzOhPdJCv2aV1Pm6eEDvXfmuoswdRX62H4xvh0Nv29gkhwyBnP1jr7MdYa+1Ny3MPQnBsk4ezGTbik+OJC40jJjDG4S9HRMzT4oglMzOTjRs3kpCQwMmTJ3F1dWXatGk8/vjjTJgwgRtvvJFhw4Zp0CNOzTAMjueVn5/hdOpMJRYXmDKgNw/PHMDcrhQ4nXP4XfjgMaivxgWgNNM+dXrXi43T9/Pt+7n72Hs1TX28cVZTHPgGmVl5m61NWYvFxcKiQYvMLkVE5DyNwboww7CHRi007W7VinA9+kLfSV1mRbiuwjAMtiYX8tLWk+xKO0uAlxvfu3Eg35gSQ7B/12qIbho3D4i9xX5r+Auc2mqfEXXwTcBouq+13j4T6pJFYHbl7iK7IpvHxjzmuLpFpFNocdQyZ84cXFxcuOGGG3j22We55ZZbCAy0T0EtL2/mD6aIkzAMgxP5FwKntEJ74DR5QBAPTe/P3OF9CPJrh0GMzWqf6myttX89f6uxn1lqqLmwrdl9LtnW2scpybi8N4StHvIOw4glF81qGt6lm4zW2+yNxqdFTCPUN9TsckREztMYrJNqxYpwlOXY/65erNkV4cKazlryD+tyK8J1FQ1WGwlJuby0NY1juWWEBnjx3/OHcs+Evvh5dt1xjOncPOyNxAfNaQyfmlGaddmm+OR4enj24Obomzu4QBHpbFr8jRseHk5OTg7Jycns3r2b4OBgpk2bprNs0v3ZbE2DGmstRn0N6QXF7DiezZ7UXM6UluPlUscdfbyJm+DDqDBv/F1zoCER9rcQ9lgvCn1aCoQu3ufcpWzXxcXeVNTNo/GrJ7h6Nt3m4WdvKurmab8Vn275fbnrH+1QU+ewLWsbhdWF/HzQz80uRUSkCY3BTNDSinAXXwbX7IpwbuAffuEyuMHzms5W6oYrwnUV1XVW3t2byT+/SCOruJoBwb78fskoFo6OwMPNcvUHkNYLjLTPlG9u+0XOVJ/h84zPuXfovXi6KmwVcTYt/iXcsmVLk8aWGzduJDAwkDlz5jBjxgznWflBHMcwrhDI2AMbn7wUsKRfsk8zM3gue5yaq4Q9F4c+ly9b7QLENN7uBTh3Jd3ZxtvhZl6Pm1dj0HPxzcvehNHNyz513q3XhZ+b2+dcQNSafZoLlixubW8omr6jVQOIrm5N8hpCvEOYHjnd7FJERJrQGOwqWupL2JKGuktCpHZaEc4/3H6pnEVBRmdSXFnH6zvT+ffO05ytrGNcdE+evn04Nw8JwWJx8v87HeXmX5xv2XCeu7d9+0XeT21sNB6rRuMizuiKp2FGjx7N6NGj+dnPfsauXbvYuHEjn3zyCfHx8bi4uPDuu+/i5eXFyJEjHVVv59HWgU9nZhiN4U1bQ5sa+4DuqvtcIey5eKbPuSaFVxDdmtdzceDTXPjj5mVv2Hk+2Ln4Pg+Kalw4VljL4bwacips1OJO3+CejIwJYXS/UHr4+109EHJ177qryLRyANGV5Vbksj17Ow+Negg3i85Gi0jnozFYC5rrS7hhJRQeh96xbVwRrjFEGnDT5U27tSJcl5NdUs3LX6Txzp5MquutzBoSwiM3DiAupnOtststnfsMdG61u2Y+G9kMG2tS1jCuzzj6B/Y3qVARMVOrPnVZLBamTJnClClT+OUvf8nWrVvZtGkTmzdvZuPGjcTExPDhhx92dK2dR3MDnw8am+a1JYAyDHszvmuZxdPiPhcFQld9nIt6A7UHV4/Lw55Lf/YKuDy0uST8aTbYadzndHYeMQNirzyr6BoGiqkF5SQcziMhKYfk/ApcXCAuphcLbgzj1hGhhPh7tc971BVcNIAwSrNw6erhajPWpq4F4K5Bd5lciYjIlWkMdonPnm16cgTsY5kv/nThZ60I51RO5JXz960n2XAoB4A7bgjn4ZkDGBzqb3JlTmbUsiuOFffk7SGzPJPvjv6uA4sSkc6kzaf83d3dmT17NrNnz6a6uprNmzeTkJDQEbV1Xs0NfOqr4YPH7cuOtmo2UOM+l64McS0sbpdcdtVM+OPje/V9mg2ELv7Zq4XjGr93wLTz6rpjEDG0XR4rtaCCTUm5JBzO5UR+uT1wiu7FL+8Yzm0jQgkJcKLA6VKNA4jjx44xdGj7vN+dRYOtgXUp65gSPoUIvwizyxERaTWNwWi2gbGdCzy6197I28PXoSWJ4xmGQeLpYl7aepItxwvwdnfl/snRfHt6fyJ6eJtdnjQjPjmeQM9A5kTPMbsUETHJdV1v4u3tze23387tt9/eXvV0DS0NfOqroOB405k47j1a6MvThmCn2Uu8LtrH4urY19+FpRVWnF+l7niePXAaH92TZ24fxm0jw+jjzIGTk/gy+0vyq/J5asJTZpciInLNnHYMdqXGxr0HOr4ecSibzWDzsXxe2nqS/Rkl9PL14AdzYrl/UjQ9fT2u/gBiiqLqIj7L+IyvDfmaGo2LODE1O7kWLQ58ouDRPY6vR64orbBxhlNSHsdyywB74PT07cO4bUQYoYEKnJxJfEo8QV5BzIyaaXYpIiLSVk7Ql1AuV9dgY/3BbP6xLY3Uggoie3rz7J3DWTouCm8PnYTt7Dac3ECDrYElg5aYXYqImEjh07XQwKfTO3Wm8vwldV81Bk7jonvyiwXDuG1kKGGBmpLtjPIr89mWtY0Hhz+Iu8Xd7HJERKStnKAvoVxQUdvA27szeGX7KfLKahgaFsDz94xm/sgw3Fy1ymBXYDNsxCfHMzZkLP17qNG4iDNT+HQtNPDplE6fqSQhKZdNSbkczbEHTmP79uDnC4Zx24hQwtUDwOmtS12HzbCxeJCW+BUR6bK6cV9CsSssr+W1HadYtTOdspoGJvcP4reLRzIzNhgXrUDYCLto+wAAIABJREFUpSTmJZJRnsGKG1aYXYqImEzh07XSwKdTSC+6EDgdybYHTmP69uC/5w9l3sgwBU5yntVmZV3KOiaGTSQqIMrsckREROQS6UWV/GNbGu/ty6LeauPW4aE8PHMAo6N6mF2aXKP45HgCPALUaFxEFD5J15N5toqExkvqkrJLARgd1YP/mjeU20aGEtnTx+QKpTPambuTnMocnhj/hNmliIiIyEWOZJfyf1tP8mFSLm4WC4vHRfDQ9P70D/YzuzS5DmdrzrI5YzP3DL4HLzf1WBVxdgqfpEvIPFvFpqRc4vdkk1KUBsANUT342bwh3DYijKheCpzkyuKT4+nl1Yubo242uxQRERGnZxgGX6YW8dLWk2xPPYO/pxsPzejPt6b2I0SrD3cLG1LtjcbV7kBEQOGTdGJZxVXnm4YfyrLPcIoN8uSntw1h3kgFTtJ6hVWFbM3cyvJhy3F3VaNxERERs1htBh8eyeXvW9NIyi4l2N+Tp24bwr0T+xLgpb/R3YVhGMSnxDMmZAwDew40uxwR6QQUPkmnklVcxYdJeWxMyuVQZgkAoyIDeeq2IcwfGUZFfjpDhw4wuUrpat4/+T4Nhs68iYiImKWm3kr8viz++UUa6UVV9O/ty2/vGsmisRF4urmaXZ60s735e0kvS+c7o75jdiki0kkofBLTZZdU82FSLhsP53KwMXAaERHAT261B059gy7McDqWb1aV0lWdW+J3fJ/xxATGmF2OiIiIUymtqueN3en868tTnKmo44aoHvz0tiHMGRaKq0Ur13VX7yW/h7+HP7dE32J2KSLSSSh8ElPklFTbL6lLyuVAhj1wGh4ewI9vHcz8kWFEB/maXKF0F7tzd5Ndkc3KMSvNLkVERMRp5JXW8Mr2NN7anUFlnZWZscE8PLM/k/sH4eKi0Kk7K64pZnP6ZpbGLlWjcRE5T+GTOExuaTWbkvLYlJTLvvRiAIaFBfDkXHvgFNNbgZO0v/jkeAI9A5kdPdvsUkRERLq91IJy/r41jfUHs7EZsGBUGA/PGMCw8ACzSxMH2XByA/W2ehbHqt2BiFyg8Ek6VF5pDZuSctmUlMvexsBpaGPgNG9kGP0UOEkHKqouYkvmFu4ZfA+erp5mlyMiItJt7Usv5qWtJ/n0q3y83C3cO6Ev357eXwvEOBnDMIhPjueG4BuI7Rlrdjki0okofJJ2l192IXBKPG0PnIaE+vOjW2KZNzKM/sF+JlcozmLDSfsSv0til5hdioiISLdjGAafnyjgpf+ksef0WQK93Xls1kC+MSWGID+d9HFG+/L3cbrsNM9Nfc7sUkSkk1H4JO2ioKyGD4/kkXA4l8T0sxiGPXD64ZxY5o0KY4ACJ3EwwzBYk7KGsSFjGdBDKySKiIi0l3qrjQ8O5fD3rWmcyC8nPNCLXywYxt1xUfh66uOFM0pIS+D5/c+TW5mLCy4YhmF2SSLSyeivg1yzgvIaPjqSx8bDuSSetgdOg/v488Rs+wyngSEKnMQ8WuJXRESkfVXVNfDOnkxe2X6K7JJqBvfx53+W3cDtN4Tj7moxuzwxSUJaAs/seIYaaw0ABga/3v1rPFw9mN9/vsnViUhnofBJ2qSgvIaPGwOnPY2BU2wfPx6/eRDzR4YxqI+/2SWKAFriV0REpL2craxj1cGzbHpvCyVV9UyI6cVzC4dz0+AQrVwnPL//+fPB0zk11hqe3/+8wicROU/hk1xVYXktHx3NI+FwDrtP2QOngSF+PDZrEPNHhRGrwEk6kYS0BP6878/kV+Xj4+bDZxmfaeAjIiJyDTLPVvHyF2ms3ptJTb2N2UP78MiN/RkX3cvs0qSTMAyD3MrcZu/Lq8xzcDUi0pkpfJJmnamo5aPGHk67TxVhM2BAsC8rZw1igQIn6aQunfZd1VDFMzueAVAAJSIi0kpf5ZTx920n2Xg4F4sLLBwdwexImDv5BrNLk04kvzKfX+78ZYv3h/qGOrAaEensFD7JeUUV52Y45bIrzR449Q/25dGbBjJ/VDixffw0tVo6NU37FhERuTaGYbAr7SwvbT3J1uRCfD1c+ebUGL45rR9hgd4cO3bM7BKlkzAMg41pG/nNnt9Qb63n9v6382n6p03GYF6uXjw+9nETqxSRzkbhk5Mrqqjl46P5JCTlsPNkY+DU25fv3TSQ+aPCGNzHX4GTdBktTe/WtG8REZHm2WwGn3yVx/9tTeNQZgm9/Tx4cu5glk+MJtDH3ezypJMprCrk2Z3P8p+s/zAmZAzPTX2O6IBopkZM5fn9z5NXmUeobyiPj31cJ/5EpAmFT07obGUdHx/NY1NSLjtOFmG1GfTr7ct3bxzIvJFhDA1T4CRdS521jleSXsGg+WV9Ne1bRESkqdoGK+v2Z/OPbWmknamkby8fnls4gqXjIvFydzW7POlkDMNg06lN/Hr3r6m11vKj8T9i+dDluFrs/1bm95+vsElErkjhk5MobgycEi4KnGKCfFgxsz/zRoYxLCxAgZN0SYl5iTy781lOl51mVO9RnCg+Qa219v+3d+9hUdb5/8dfIGdwwFAT8IgEmgKKB1JJa0trF6/ya1vfWmuxlNJaU8vyMjdLM1072JqUptmatptm2e7XMKufmaamlmWah1ZFwwMeMOU8HGbm94dJTuBp5L6HGZ+P6+JSPjP3zHvey+K719z3Z6pv57RvAAB+VWit1L825uqttft0rKhcHWMsyvpTZ/2+Y5Qa+DILoqb8snxN3jBZK3NXKqlJkib3mqw24W3cXRYAD0P45MVOlVbo0+1H9dG2PK3bky+b3aFWkSF6qPfpwKlDNIETPNdJ60m99M1L+r+9/6eYsBjNunmW0mLSlJ2TzWnfAAD8xrFCq95at1//3PCTisqr1CsuUtPv6qRecZHMg6iVw+HQJ/s/0fMbn1dpZake7/K47rv2vuqznQDgUhA+eZmC0kp9suP0puHr9uSryu5Qy6tC9GDvWKUTOMELOBwO/Wfvf/TyNy+ruKJYQxOH6sGkBxXsFyyJ074BADjbvvwSzVmzVx9sPqQqu12/7xilYX3aKrF5uLtLQz32s/VnTd4wWZ/99JkSGydqcq/Jio2IdXdZADwY4ZMXKCit1Kc7Tl9St3b36cCpxVXBGnJ9G/VPjFbHGAIneIecghw999Vz+uboN+rUpJMm9Jigaxpd4+6yAACod74/cEqzV+/Viu1H5N/AV3d2ba7M62PVunGou0tDPffp/k/1/MbnVVRRpJEpIzW4w2D5+fKfjQAuD79FPFRBWaU+23FU2VsPa+2efFXaHGreKFhD0tooPSlKiTHhBE7wGuW2cr257U3N2zZPQX5BeqbHMxp4zUD5+vi6uzQAAOoNh8OhNbvzNfuLvfoq54QsQX56+Ia2GtyzjZo0DHR3eajnTlpPasrGKVqxf4WujbxWb/Z7kzf5ANQZwicPUlBWqf+346iyt+Xpy93HVWlzKCYiWPf3aqP0xCglNSdwgvfZmLdRz214Tj8V/qQ/tPmDnuj2hBoHN3Z3WQAA1BtVNruyt+XpjdU52pFXqKstgRr/h/a6J7WlwgIZ93FhK39aqUkbJqmwolAjOo/Q/R3vl7+vv7vLAuBF+Neoniu0/hI4bc3Tl7vzVWGzKyYiWIN7tlZ6UrSSCZzgpX62/qyXvn5Jy3KWqUXDFnrj5jfUM6anu8sCAKDeKKuwacnmA5r7ZY4O/Fymtk1C9cIdSbq9c7QC/dgUGhd2ynpKUzdN1fJ9y9Xuqnaa03eOEq5KcHdZALwQ4VM9VGSt1P/beTpwWvPf04FTdHiQ/tyjldKTotSpRQSBE7yWw+HQv/f8Wy9vflkllSXKTMzUg0kPKsgvyN2lAQBQL5wqrdCCr37S/PX79XNJhVJaRujp9Gt1c/ur5evLjIiLsyp3lSZtmKRT1lN6uNPDGpo4lLOdABiG8KmeKLJWauXOY/poa57W7D6uiiq7osKDdN+ZwKl5BMMEvF7OqRxN2jBJm49uVkrTFD193dOKaxTn7rIAAKgXDp8q05tf7tOir3NVWmHT79o11bA+bdWtdSPemMRFKygv0LRN07QsZ5niG8Vr1s2z1O6qdu4uC4CXI3xyo+LyKq385QynL/57OnBqZgnSvamnA6fOLQiccGWwVlk1d9tcvfXDWwrxC9HEnhM1IG4AG4oDACDpv0eLNHv1Xv3flsNySLo9OVoP9olVu2YWd5cGD7Pm4BpNXD9RJ6wn9FDSQ3oo6SH5N+BsJwDGI3wy2ZnAafm2PK368XTgdLUlUINSWyo9MUopLRsROOGK8tXhrzR5w2TlFuWqf2x/jek6RpHBke4uCwAAt/t6/8+a/cVerdx1TMH+DXRfj1YaktZGzRuFuLs0eJjCikK9+PWL+veefysuIk6v3vSqOkR2cHdZAK4ghE8mKCmv0ue7jil7a55W/XhM5VV2NW0YqD91b6n0pCh1IXDCFehE2Qm9+M2Lys7JVsuGLTWn7xz1iO7h7rIAAHAru92hlbuOafbqvdr800k1CvHX6Jvj9ecerdQoNMDd5cEDrTu0Ts+sf0bHy44rMzFTw5KHKaABP0sAzEX4ZJDSCufAyVp5OnC6p3tL/SExSl1bETjhymR32PXh7g81ffN0lVaV6qGkh5SZlKnABoHuLg0AALepqLLrP1sOac6aHO0+VqzmjYI18bYOuqtrCwUH8Ml1uHTFFcV66ZuX9MHuD9Q2vK3+fuPf1bFxR3eXBeAKRfjkon9/d0gvfvKjDp8qU3REnp64JUH9OlytVbuOK3vbYX2+63Tg1KRhoP63a4vTgVPrq9SAwAlXsD0n9+i5Dc/p22PfqsvVXTThugmKjYh1d1kAALhNcXmVFm3K1by1+5RXYFW7Zg014+5OSk+Mkl8D9j6Ea9YfXq9n1j+jY6XH9EDHB/Rwp4d5ow+AWxE+ueDf3x3SuKXbVFZpkyQdOlWmx9/bIl8fH1XaHWocFqg7u7RQelKUuhE4AbJWWTVn6xz944d/KDQgVJN6TtKAuAF8Mg8A4IqVX1yu+ev2a8FX+1VordJ1sVdp6sBE9Ylvwr+PcFlJZYle/uZlLfnvErW2tNaC3y9QcpNkd5cFAIRPrnjxkx+rg6czbA4pyN9XCzK6qXsbAifgjPWH1uu5Dc/pYPFB3db2Nj3e9XFdFXSVu8sCAMAtck+Uas6Xe7Xkm4OqsNl1y7XN9FCfWHVu2cjdpcHDbczbqAnrJiivJE+DOwzWI50eUZBfkLvLAgBJhE8uOXyqrNb10gqberTlU7oAScovy9eLX7+o5fuWq7Wltd7s96ZSo1LdXRYAAG7xw6ECzV69V8u35cnP11cDU2KU2TtWbZuEubs0eLjSylK9svkVLfpxkVpZWunt37+tzk07u7ssAHBC+OSC6IhgHaolgIqOCHZDNUD9YnfY9cHuD/TK5ldkrbJqePJwDUkcwj4DAIArjsPh0Pq9JzR79V59uTtfYYF+yuwdqwd6tdHVFs5IweX7+sjXenrd0zpcfFj3tr9Xj6Y8qmA//psEQP1D+OSCJ25JcNrzSZKC/RvoiVsS3FgV4H67T+7WpK8macvxLerWrJuevu5ptQlv4+6yAAAwlc3u0Iofjmj26r3adqhATRoGauyt7TToupayBPm7uzx4gdLKUr363av6585/qkXDFvrHrf9Ql6u7uLssADgnwicXDOgcI0lnfdpdsJ64JaF6HbjSlFWV6Y3v39Db299WWECYJvearNva3saGqQCAK4q10qYPvj2ouWtytP9Eqdo0DtXUgYn6n84xCvJv4O7y4CW+Pfqt/rrurzpQdEB/avcnjUwZqRD/EHeXBQDnRfjkogGdYzSgc4x27typ9u3bu7scwG3WHlqryRsm61DxIQ2IG6DHujymRkFsmgoAuHIUlFXqnQ0/6R/r9iu/uFzJzcM1a1CK+nVoxofQoM6UVZVp5ncz9c6OdxQdFq23bnlL3Zp1c3dZAHBRCJ8AuOR46XG98PULWrF/hVpbWjMAAQCuOEcKrHpr3T79a2Ouisur1Du+iYb1iVWP2EjO/kWd2nJsi55e97T2F+7X/yb8rx7r8hhnOwHwKIRPAC6J3WHX+/99X3/f/HeV28r1cKeHNaTjEAU0CHB3aQAAmGLPsWLNWbNXH353SDa7Q/2TovVQn1h1iA53d2nwMtYqq17b8poW7Figq0Ou1tx+c3Vd1HXuLgsALhnhE4CL9t+T/9XEryZq6/GtSm2Wqr9e91e1Dm/t7rIAADDFt7knNfuLvfps51EFNPDVPd1bKvP6WLW4ijNQUDeyc7I149sZOlJyRJFbIyUfKb8sX3+M/6Me7/K4wgLC3F0iALiE8AlADWcPPs22N9OwpGHaX7RfC7YvkCXAoilpU9Q/tj+XFAAAvJ7D4dAXPx7XrNV7tWnfzwoP9teIG+OU0bO1IsMC3V0evMh/9vxHz214TuW2cklSvjVfkvRAhwc0uutod5YGAJeN8AmAk+ycbD27/llZbVZJUl5Jnp796lk55NDAawZqdMpoRQRFuLlKAACMVWmz66Oth/XG6hztOlKk6PAgPd3/Wt3drYVCAxmh8Sub3aaSqhIVVxSruLK4xp8llSUqqihSSWVJ9XpJZYmKKoucbjsTOv3Wx/s/JnwC4PH4lxO4AjkcDpVVlVUPQaWVpSqpLFFJZYn+tulv1cFT9f3lUGRQpCb2nOimigEAMEdpRZUWf31Ab365T4dOlSn+6jC9fGeybusULf8Gvu4uD3XI7rCrrKqsRjB00cFRxek/y6rKLvhcvj6+CvMPU5h/mEIDQtXQv6EigyLVytKqev0f2/9R67FHSo7U9UsHANMRPgEeospepZLKkl+DoqoSlVT88mdlza/SylLnYKnK+TaHHJf0/D9bfzbolQEA4H4/l1To7fX7teCr/TpZWqlurRtp0u0ddGNCU/n6cpl5feJwOGS1WWsNhs4VHJ0Ji4orf7ntl79fzDwU6h/qHBwFNFSz0GZqGNDw19sCwqr/DPU/HS6FBvx6XLBf8AW3K1ixf4XySvJqrDcLbeZyrwCgviB8AgzicDhUYa+oEQadHSDVCIfOCpN+e//fno10Ln6+fqeHI/9QhfiHKNQvVOFB4Yr2i1aof2iNrxD/EKf7j1g5QsfLjtd4XAYfAIA3OniyVG9+uU+Lvz6gskqbbm5/tYb1iVXX1le5uzSvVGmrdDprqKTyrMvVzhUcnR0w/RIcVTmqLvhcwX7BNYKjxsGNT4dDvwRHtQZIZ4VIof6h8vUx54y3kSkjnbY+kKSgBkEamTLSlOcHACMRPgFnOd/laNXf13KmUY1g6Zcg6WIGI+n0cBTiF+IUCjUNaarW/q2rB6IzQVKof6hCA0J//ftvvgIaBFxWDx7v+jiDDwDA6+3MK9Qbq/dq2dY8+Uga0DlGD/WO1TVXN3R3afWSzWFTQXlBjYCotuDofJesVdgrLvhc/r7+NUKh6LDo6jfLzhsc/fJniH+I/H39TehM3UmPTZekXz/0JbSZRqaMrF4HAE9G+ASP547L0XzkU+tZQ5FBkQoLCKsRJNV2ltGZ70P8QuTnW3/+r8jgAwDwVg6HQxv3/azZq/fqix+PKzSgge7v2VpDrm+jqPBgd5dnCLvDXj37OAVFZ+1fdDHBUVlVmfT1+Z+rgU+D6vDnTCjUJKSJ2vi3cdrr6ELB0eW+kebJ0mPTlR6brp07d6p9+/buLgcA6kz9+S9eXDEu53K04wXHpT2qu8vRAsMVHRZdazh05j5hAb8568g/9KKu2/dkDD4AAG9itzv06Y6jmr16r7YcOKXI0ACN6Rev+65rrfCQ+nl2zNlnY/92v6KzQ6Ta1s8OjkoqSy74xtqZN9XODo4sgZbqs43C/MNUVlCmNtFtzhscBTUI8ur5CADgOsInXJT6cjmaj3zUJKSJWy5HAwAAnqW8yqZ/f3dIb6zJUc7xErW8KkTPDeioO7s0V5B/A8Oet8JWcc6Nrn8bHP1236OzgyObw3bB5wr2C3YKgc5cul/bPkbnOvMoxD/kgvsa8YYUAOByED55MXdfjnb2QBMZFHnBja7P3P98l6Mx+AAAcGXLzsn+9dLw7bVfGl5krdS/NubqrXX7dLSwXB2iLZp5T2f9vmMz+TU4d8hyZnaqdaPripqXq9XYMPuXPyvtlRd8HQG+ATUuN4sJiznvBti/3fMo1D+0Xl26DwDAufCvlYsuZvC5VO78dLTfnkXE5WgAAKC+yc7JdvpQjLySPD27/lnZHXbd0OIG/XTyhP759X/10bZ9Kq0qUbtofw3oHaZmEcf0U+X3emnz+S9Zu5jZ6ex9jc6EQk1DmqqNf5vaN8I+x6eocUY2AOBKQvjkgtoGn2fWP6NjpcfUvVl3Uy5HC2oQVOMsorM/HS3U7/yXoJ0dJjH8AAAATzDj2xk1AiKrzaqn1j7lfMdoKURSrqR/7T295COf6kvPzgRBEUERat6wea1nFTX0b+h03zP3CWwQyBttAABcIsInF9Q2+JTbyjV98/RzHnO5l6OdfaZRfft0NAAAADPklRypdd3hkGz5/dWlRZRuvba1Yq9qXCM4CvYLvuC+RgAAwBgkGC44co7BR5JevfHVWje+5nI0AACAy+NTFSGH38maN1RFaPVDz6hpwyDziwIAABdE+OSCZqHNlFeSV2M9KjRKN7a80Q0VAQAAeL+yo/0UGLVUPr6/bujtsPur/NgtBE8AANRjnHvsgpEpIxXUwHnACWoQpJEpI91UEQAAgPdr6ttT1ryBsldEyOGQ7BURsuYNVFPfnu4uDQAAnAdnPrngzKfaVX/aXWjdfNodAAAAzu2JWxI0bmmFSvZ2rl4L9m+gJwYmuLEqAABwIYRPLkqPTVd6bLp27typ9u3bu7scAAAArzegc4wk6cVPftThU2WKjgjWE7ckVK8DAID6ifAJAAAAHmNA5xgN6BzDG4AAAHgQ9nwCAAAAAACAYQifAAAAAAAAYBjTw6c9e/YoIyNDycnJSktL04wZM2Sz2S7q2E8//VR33HGHkpKSlJqaqiFDhqi0tNTgigEAADwfMxgAAHAXU/d8Kigo0ODBgxUXF6fXX39dubm5mjZtmux2u0aPHn3eY5csWaJJkyZp6NChevLJJ1VYWKgNGzZc9NAEAABwpWIGAwAA7mRq+LRo0SKVl5crKytLYWFh6tWrl4qLi5WVlaXMzEyFhYXVetzPP/+sKVOm6Omnn9Zdd91Vvd63b1+zSgcAAPBYzGAAAMCdTL3sbs2aNUpLS3MacNLT02W1WrVp06ZzHvfxxx9LkgYMGGB4jQAAAN6GGQwAALiTqeFTTk6OYmNjndaio6MVHBysnJyccx63detWtWnTRu+//7569+6tDh066M4779S3335rdMkAAAAejxkMAAC4k6nhU2FhoRo2bFhj3WKxqLCw8JzH5efna9++fZo1a5bGjBmjWbNmKTg4WEOHDlV+fr6RJQMAAHg8ZjAAAOBOpu755CqHw6HS0lLNmDFDvXv3liSlpKToxhtv1DvvvKNRo0ad9/jy8nLt3LnTkNqsVqthj42a6Le56Lf56Lm56Le56LfnYQbDGfTbXPTbXPTbfPTcXPWh36aGTxaLRcXFxTXWCwsLZbFYznucj4+PUlNTq9fCwsLUoUMH7d2794LPGxgYqPbt27tW9AXs3LnTsMdGTfTbXPTbfPTcXPTbXEb2290DVX3HDIbLRb/NRb/NRb/NR8/NVR9mMFMvu4uNja2xr0BeXp7Kyspq7ENwtrZt28rhcMjhcDitOxwO+fj4GFIrAACAt2AGAwAA7mRq+NS7d2+tXbvW6Z235cuXKygoSN27dz/ncTfccIMkaePGjdVrRUVF2r59u9q1a2dYvQAAAN6AGQwAALiTqeHT3XffrYCAAI0YMULr16/X4sWLlZWVpcGDBzt99G/fvn311FNPVX+fmJiom266SePHj9eHH36oL774QsOHD5efn58GDRpk5ksAAADwOMxgAADAnUwNn8LDwzV//nzZbDYNGzZMM2fOVEZGhh599FGn+9lsNtntdqe1F198UTfddJP+9re/6dFHH5Wfn5/efvtthYeHm/kSAAAAPA4zGAAAcCfTP+0uLi5OCxYsOO99Pv/88xproaGhmjhxoiZOnGhUaQAAAF6LGQwAALiLqWc+AQAAAAAA4MpC+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADJGGmCAAAZFUlEQVTDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDED4BAAAAAADAMIRPAAAAAAAAMAzhEwAAAAAAAAxD+AQAAAAAAADDmB4+7dmzRxkZGUpOTlZaWppmzJghm8123mMOHjyohISEGl+jR482qWoAAADPxgwGAADcxc/MJysoKNDgwYMVFxen119/Xbm5uZo2bZrsdvtFDTFjx45VSkpK9feNGjUyslwAAACvwAwGAADcydTwadGiRSovL1dWVpbCwsLUq1cvFRcXKysrS5mZmQoLCzvv8W3atFGnTp1MqhYAAMA7MIMBAAB3MvWyuzVr1igtLc1pwElPT5fVatWmTZvMLAUAAOCKwQwGAADcydTwKScnR7GxsU5r0dHRCg4OVk5OzgWPHzdunNq3b6+0tDRNnTpVVqvVqFIBAAC8BjMYAABwJ1MvuyssLFTDhg1rrFssFhUWFp7zuICAAA0aNEi9evVSWFiYNm3apLlz5yo3N1ezZs264POWl5dr586dl1X7uVitVsMeGzXRb3PRb/PRc3PRb3PRb/dhBsPlot/mot/mot/mo+fmqg/9NjV8clXTpk01YcKE6u9TU1MVGRmpiRMnateuXWrXrt15jw8MDFT79u0NqW3nzp2GPTZqot/mot/mo+fmot/mMrLf7h6ovBUzGM6g3+ai3+ai3+aj5+aqDzOYqZfdWSwWFRcX11gvLCyUxWK5pMe69dZbJUk//PBDndQGAADgrZjBAACAO5kaPsXGxtbYVyAvL09lZWU19iG4WD4+PnVRGgAAgNdiBgMAAO5kavjUu3dvrV271umdt+XLlysoKEjdu3e/pMf65JNPJEkdOnSo0xoBAAC8DTMYAABwJ1P3fLr77ru1cOFCjRgxQpmZmTpw4ICysrI0ePBgp4/+7du3r7p166YpU6ZIkmbOnKmSkhKlpKQoLCxMX3/9tebNm6d+/fpdcK8BAACAKx0zGAAAcCdTw6fw8HDNnz9fkyZN0rBhw2SxWJSRkaERI0Y43c9ms8lut1d/Hxsbq3nz5mnJkiUqLy9XVFSUhgwZouHDh5tZPgAAgEdiBgMAAO5k+qfdxcXFacGCBee9z+eff+70fXp6utLT040sCwAAwKsxgwEAAHcxdc8nAAAAAAAAXFkInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAYwicAAAAAAAAYhvAJAAAAAAAAhiF8AgAAAAAAgGEInwAAAAAAAGAY08OnPXv2KCMjQ8nJyUpLS9OMGTNks9ku+ni73a6BAwcqISFBq1atMrBSAAAA78EMBgAA3MXPzCcrKCjQ4MGDFRcXp9dff125ubmaNm2a7Ha7Ro8efVGPsWTJEh09etTgSgEAALwHMxgAAHAnU898WrRokcrLy5WVlaVevXrpnnvu0SOPPKL58+eruLj4gscXFBTolVde0ahRo0yoFgAAwDswgwEAAHcyNXxas2aN0tLSFBYWVr2Wnp4uq9WqTZs2XfD4GTNmKCUlRT169DCyTAAAAK/CDAYAANzJ1PApJydHsbGxTmvR0dEKDg5WTk7OeY/dtWuXPvjgA40dO9bIEgEAALwOMxgAAHAnU/d8KiwsVMOGDWusWywWFRYWnvfYyZMna9CgQWrVqpUOHjx4Sc9bXl6unTt3XtIxl8LIx0ZN9Ntc9Nt89Nxc9NtcRvW7vLzckMf1FsxgqAv021z021z023z03FzunsFMDZ9clZ2drX379mn27NkuHd+pU6c6rggAAMD7MYMBAIC6YOpldxaLpdZNLQsLC2WxWGo9prKyUi+88IIyMzNlt9tVWFhY/RhlZWUXtUkmAADAlYwZDAAAuJOpZz7FxsbW2FcgLy9PZWVlNfYhOKOsrExHjhzR1KlTNXXqVKfbRo8erZYtW+qzzz4zrGYAAABPxwwGAADcydTwqXfv3po3b56Ki4urP21l+fLlCgoKUvfu3Ws9JiQkRAsWLHBay8/P12OPPabHHntM1113neF1AwAAeDJmMAAA4E6mhk933323Fi5cqBEjRigzM1MHDhxQVlaWBg8e7PTRv3379lW3bt00ZcoU+fn5KTU11elxzmx2GR8fr+TkZDNfAgAAgMdhBgMAAO5k6p5P4eHhmj9/vmw2m4YNG6aZM2cqIyNDjz76qNP9bDab7Ha7maUBAAB4LWYwAADgTj4Oh8Ph7iIAAAAAAADgnUw988mT7NmzRxkZGUpOTlZaWppmzJghm812weOKioo0btw4devWTV26dNHjjz+ukydPmlCxZ3Ol3xUVFZo2bZr+9Kc/KSkpSQkJCSZV6x1c6fnWrVs1btw49e3bV8nJybrllluUlZWl8vJyk6r2XK70e/fu3RoyZIjS0tLUsWNH3XDDDRo/fryOHTtmUtWey9Xf4WfY7XYNHDhQCQkJWrVqlYGVegdX+n3w4EElJCTU+Bo9erRJVaO+YgYzFzOYuZi/zMcMZi5mMHN50gxm6p5PnqKgoECDBw9WXFycXn/9deXm5mratGmy2+0X/B9k1KhR2rdvnyZPnixfX1+99NJLeuSRR/Svf/3LpOo9j6v9tlqtev/995WUlKTOnTtrw4YNJlbt2Vzt+ccff6zc3FxlZmaqVatW+vHHHzVjxgz9+OOPmjlzpomvwLO42u+ioiI1b95cAwYMUNOmTXXw4EG99tpr2r59u95//335+fErvDaX8zv8jCVLlujo0aMGV+odLrffY8eOVUpKSvX3jRo1MrJc1HPMYOZiBjMX85f5mMHMxQxmLo+bwRyoYfbs2Y6uXbs6ioqKqtfmzJnjSEpKclr7rW+//dYRHx/v2LRpU/Xa999/74iPj3esW7fO0Jo9mav9djgcDrvd7nA4HI6FCxc64uPjDa3Tm7ja8xMnTtRYW7RokSM+Pt5x8OBBQ2r1BpfzM/5ba9eudcTHxzt++OGHui7Ta1xuv0+dOuVITU11vPfee474+HjH559/bmS5Hs/Vfh84cID+ogZmMHMxg5mL+ct8zGDmYgYzl6fNYFx2V4s1a9YoLS3N6dNf0tPTZbVatWnTpvMe17hxY3Xr1q16LSkpSc2bN9eaNWsMrdmTudpvSfLx8TG6PK/kas+vuuqqGmvt27eXJE5DPo/L+Rn/rYiICElSZWVlndboTS633zNmzFBKSop69OhhZJleoy5/vgFmMHMxg5mL+ct8zGDmYgYzl6fNYIRPtcjJyVFsbKzTWnR0tIKDg5WTk3NJx0lS27Ztz3vclc7VfsN1ddnzLVu2yNfXVy1btqzLEr3K5fbbbreroqJCOTk5evnll5WYmKikpCSjyvV4l9PvXbt26YMPPtDYsWONLNGrXO7P97hx49S+fXulpaVp6tSpslqtRpUKD8AMZi5mMHMxf5mPGcxczGDm8rQZjItVa1FYWKiGDRvWWLdYLCosLHTpuIMHD9Zpjd7E1X7DdXXV8+PHj2vWrFm6/fbbFRkZWZclepXL7XdmZqbWrl0rSerQoYPmzp0rX1/eOziXy+n35MmTNWjQILVq1Yrf2xfJ1X4HBARo0KBB6tWrl8LCwrRp0ybNnTtXubm5mjVrlpElox5jBjMXM5i5mL/MxwxmLmYwc3naDEb4BMAlFRUVGjVqlEJCQjRu3Dh3l+PVnn76aRUUFGj//v2aNWuWMjMz9e677yowMNDdpXmV7Oxs7du3T7Nnz3Z3KVeEpk2basKECdXfp6amKjIyUhMnTtSuXbvUrl07N1YHAPUT85e5mMHMwQxmLnfNYMS2tbBYLCouLq6xXlhYKIvFct7jioqKLvm4K52r/YbrLrfnDodDY8eO1Z49ezRnzhyFh4cbUabXuNx+t27dWsnJybr99ts1b9487dixQ8uWLTOiVK/gSr8rKyv1wgsvKDMzU3a7XYWFhdWPUVZWVuvj4bS6/B1+6623SpJ++OGHOqkNnocZzFzMYOZi/jIfM5i5mMHM5WkzGGc+1SI2NrbGNZJ5eXkqKyurdT+Bs4/bvHlzjfWcnBzdfPPNdV6nt3C133Dd5fb8+eef18qVK/XWW2+pbdu2RpXpNeryZzwmJkbh4eE6cOBAXZboVVzpd1lZmY4cOaKpU6dq6tSpTreNHj1aLVu21GeffWZYzZ7MiN/hbGR85WIGMxczmLmYv8zHDGYuZjBzedoMxplPtejdu7fWrl3rlCIuX75cQUFB6t69+3mPO378uL755pvqtW3btunAgQPq3bu3oTV7Mlf7DdddTs/feOMN/fOf/9SLL76orl27Gl2qV6jLn/GcnBydOnVKzZs3r+syvYYr/Q4JCdGCBQucvqZPny5Jeuyxx/TSSy+ZUrsnqsuf708++UTS6X01cGViBjMXM5i5mL/MxwxmLmYwc3naDNbg2WeffdawR/dQ11xzjRYvXqyNGzeqadOmWr9+vaZPn66MjAz16dOn+n59+/bVrl27dNNNN0mSoqKitGXLFr3//vuKiorSvn379Oyzz6pt27YaNWqUu15OvedqvyVp9erV2rVrlzZv3qwdO3bommuu0Z49exQcHMzp4ufhas+XLVumiRMn6n/+53+UmpqqI0eOVH8FBAQoODjYXS+pXnO139OmTdOGDRtUUlKi48ePa/Xq1Zo4caIaNWqkCRMmyN/f310vqV5zpd++vr5q3ry509eZYSgjI0PXXXedG19R/ebqz/fMmTO1atUqlZaW6ujRo1q6dKleffVV3XTTTfrzn//srpcDN2MGMxczmLmYv8zHDGYuZjBzedoMxmV3tQgPD9f8+fM1adIkDRs2TBaLRRkZGRoxYoTT/Ww2m+x2u9PaK6+8oqlTp+qpp56S3W7XjTfeqPHjx5tZvse5nH5PnDhRhw4dqv5+5MiRkqSpU6dq4MCBxhfvoVzt+bp16yRJS5cu1dKlS53uS8/PzdV+d+zYUQsXLtR7772n8vJyRUVFqV+/fnrwwQcVEhJi9svwGJfzOwWXztV+x8bGat68eVqyZEn1z/eQIUM0fPhws18C6hFmMHMxg5mL+ct8zGDmYgYzl6fNYD4Oh8Nh6DMAAAAAAADgisWeTwAAAAAAADAM4RMAAAAAAAAMQ/gEAAAAAAAAwxA+AQAAAAAAwDCETwAAAAAAADAM4RMAAAAAAAAMQ/gEwGUJCQkX/Nq4caPTMWPGjFFCQoKWLl1a62MePHhQY8aMUZ8+fZSYmKg+ffro4Ycf1ubNmw1/PcePH9fMmTN1+PBhl44/evSohg4dqi5duighIUHffPONJOnVV19VWlqa2rVrp/Hjx9dlyQAA4ArEDOaMGQyo//zcXQAAz7V48eLqv1utVmVkZGj48OG64YYbqtfj4uKc7rNy5UpJUnZ2tgYOHOj0eCdPntRdd92lZs2a6YknnlDjxo116NAhrVy5Ulu2bFGXLl0MfT35+fnKyspSjx49FB0dfcnHv/baa9q9e7emT5+u8PBwxcXFacuWLXrttdc0ZswYde3aVY0bNzagcgAAcCVhBnPGDAbUf4RPAFzWqVOn6r+XlJRIklq2bOm0frbPP/9cpaWl6tGjh7766iudOHFCkZGR1bd//PHHOnnypLKzs9WoUaPq9TvuuEMOh8OgV1F39u3bp06dOqlPnz7Vazk5OZKke++9V8HBwe4qDQAAeBFmMGfMYED9x2V3AEyTnZ2tmJgYjR8/XjabTStWrHC6vaioSAEBAbJYLDWO9fHxueDjL1q0SP3791diYqJ69uypkSNHqri4WJJ0zz33aPTo0U73X79+vRISErR371799NNPGjBggCRp0KBBSkhI0LXXXlt939zcXA0fPlydO3dW586dNXz4cB04cECSVFVVpYSEBG3atEkrVqxQQkKC+vbtqzFjxmjcuHGSTg+JZ58GDgAAYBZmMGYwwN048wmAKYqKirRmzRrdd999uuaaa5SQkKCPPvpIgwYNqr7PtddeK6vVqrFjx+r+++9X+/bt5et7cRn5zJkz9dprr+nee+/Vk08+KavVqlWrVslqtSosLOyCx0dFRWnatGkaO3asJk2apISEhOphq7y8XIMHD1ZgYKCef/55+fr66tVXX9W9996rZcuWyWKxaPHixZowYYIiIyM1cuRIBQYGKjQ0VNHR0XrjjTf0zjvvyN/f3+kUeAAAAKMxgzGDAfUB4RMAU3z66aeqqKhQenq6JKl///6aPn26Dh8+XH1t//XXX6/77rtPCxcu1LJlyxQaGqq0tDTdc8896tGjxzkf++TJk5ozZ44eeOABPfnkk9Xr/fr1u+j6AgIClJCQIElq27at02nrS5Ys0dGjR/Xpp58qJiZGkpSYmKh+/frpvffe09ChQ9WpUyeFhoYqIiLC6dgWLVpIkpKSkhQYGHjR9QAAANQFZjBmMKA+4LI7AKbIzs5W69at1aFDB0nSH/7wBzkcDmVnZzvd769//as++eQTPfHEE+revbvWrFmj+++/X0uWLDnnY3/33XeqqKjQHXfcYUjtW7duVWJiYvXQI0kxMTFKTk425RNgAAAAXMUMBqA+IHwCYLj8/Hxt2LBBv/vd71RYWKjCwkJZLBZ16NChxuAjSa1bt9bQoUM1e/ZsrVy5UvHx8Xr55ZfP+finTp2SJDVp0sSQ+o8fP+60KecZjRs3VkFBgSHPCQAAcLmYwQDUF4RPAAy3YsUK2Ww2vfXWW+rWrVv11/bt27Vz507t3bv3nMdGRkZq4MCBOnnypE6ePFnrfSIiIiSdHlDOJTAwUJWVlU5rFzu0NGnSRCdOnKixnp+fr/Dw8It6DAAAALMxgwGoLwifABjuo48+Unx8vBYsWOD0NXfuXPn5+VW/81bbcCFJ+/fvV1BQ0Dk3rUxJSVFgYKA+/PDDc9bQrFmz6o/cPWPdunVO3/v7+0uSKioqnNaTk5O1bds2HT58uHrt8OHD+v7779WlS5dzPicAAIA7MYMBqC/YcByAoQ4dOqQtW7boySefVGpqao3be/bsqezsbD366KP64IMP9PHHH+v2229XQkKCqqqqtG7dOi1evFj33Xdf9WDyWxEREXrooYeUlZWl8vJy9e7dW+Xl5Vq1apVGjx6txo0b6+abb9aHH36oadOm6frrr9dXX32l9evXOz1OTEyMAgICtHTpUgUHB8vf318dO3bUH//4R7355pvKzMzUX/7yF/n4+GjmzJlq3Lix7rrrLkP6BgAAcDmYwQDUJ4RPAAyVnZ0tX19f9e/fv9bbb7vtNo0ZM0Y//PCDbrjhBh0+fFjvvfee8vLy5OfnpxYtWuiZZ57RnXfeed7neeSRR9SoUSMtXLhQ7777rsLDw9WtWzcFBwdLkm666SaNGjVK7777rhYtWqR+/fpp3Lhx+stf/lL9GMHBwXruuef0+uuva8WKFbLb7dqxY4cCAwP19ttva8qUKXrqqackSampqXrqqadksVjqqFMAAAB1hxkMQH3i43A4HO4uAgAAAAAAAN6JPZ8AAAAAAABgGMInAAAAAAAAGIbwCQAAAAAAAIYhfAIAAAAAAIBhCJ8AAAAAAABgGMInAAAAAAAAGIbwCQAAAAAAAIYhfAIAAAAAAIBh/j84wh0Bpuf0HQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x720 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAS cutoff</th>\n",
       "      <th>PC3_train_AUC</th>\n",
       "      <th>PC3_val_AUC</th>\n",
       "      <th>A375_train_AUC</th>\n",
       "      <th>A375_val_AUC</th>\n",
       "      <th>MCF7_train_AUC</th>\n",
       "      <th>MCF7_val_AUC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.638228</td>\n",
       "      <td>0.613410</td>\n",
       "      <td>0.679892</td>\n",
       "      <td>0.681330</td>\n",
       "      <td>0.615249</td>\n",
       "      <td>0.616155</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.20</td>\n",
       "      <td>0.706388</td>\n",
       "      <td>0.708054</td>\n",
       "      <td>0.685906</td>\n",
       "      <td>0.703547</td>\n",
       "      <td>0.625780</td>\n",
       "      <td>0.627616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.25</td>\n",
       "      <td>0.829911</td>\n",
       "      <td>0.839214</td>\n",
       "      <td>0.697265</td>\n",
       "      <td>0.686654</td>\n",
       "      <td>0.657541</td>\n",
       "      <td>0.661763</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.30</td>\n",
       "      <td>0.833776</td>\n",
       "      <td>0.789295</td>\n",
       "      <td>0.748609</td>\n",
       "      <td>0.757091</td>\n",
       "      <td>0.785755</td>\n",
       "      <td>0.789437</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.35</td>\n",
       "      <td>0.792710</td>\n",
       "      <td>0.793863</td>\n",
       "      <td>0.858586</td>\n",
       "      <td>0.861746</td>\n",
       "      <td>0.798413</td>\n",
       "      <td>0.801423</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.40</td>\n",
       "      <td>0.860905</td>\n",
       "      <td>0.828395</td>\n",
       "      <td>0.883934</td>\n",
       "      <td>0.879564</td>\n",
       "      <td>0.863636</td>\n",
       "      <td>0.853535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.45</td>\n",
       "      <td>0.821970</td>\n",
       "      <td>0.844697</td>\n",
       "      <td>0.873760</td>\n",
       "      <td>0.879368</td>\n",
       "      <td>0.844309</td>\n",
       "      <td>0.845635</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.50</td>\n",
       "      <td>0.821911</td>\n",
       "      <td>0.808559</td>\n",
       "      <td>0.873370</td>\n",
       "      <td>0.876944</td>\n",
       "      <td>0.868708</td>\n",
       "      <td>0.861722</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   TAS cutoff  PC3_train_AUC  PC3_val_AUC  A375_train_AUC  A375_val_AUC  \\\n",
       "0        0.00       0.638228     0.613410        0.679892      0.681330   \n",
       "1        0.20       0.706388     0.708054        0.685906      0.703547   \n",
       "2        0.25       0.829911     0.839214        0.697265      0.686654   \n",
       "3        0.30       0.833776     0.789295        0.748609      0.757091   \n",
       "4        0.35       0.792710     0.793863        0.858586      0.861746   \n",
       "5        0.40       0.860905     0.828395        0.883934      0.879564   \n",
       "6        0.45       0.821970     0.844697        0.873760      0.879368   \n",
       "7        0.50       0.821911     0.808559        0.873370      0.876944   \n",
       "\n",
       "   MCF7_train_AUC  MCF7_val_AUC  \n",
       "0        0.615249      0.616155  \n",
       "1        0.625780      0.627616  \n",
       "2        0.657541      0.661763  \n",
       "3        0.785755      0.789437  \n",
       "4        0.798413      0.801423  \n",
       "5        0.863636      0.853535  \n",
       "6        0.844309      0.845635  \n",
       "7        0.868708      0.861722  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# compare cross-validation average AUCs in training and validation datasets across all folds\n",
    "compare_metric(cv_scores, \"train_AUC\", \"val_AUC\", \"AUC\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b507509-6fda-4ce2-ab1f-0ba2f905a7fc",
   "metadata": {},
   "source": [
    "### c. Generate predictions on Clinical Trials on model trained on with and without TAS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "7a59bec3-4dff-463c-82dd-6c3eb34d473e",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate predictions on training and testing datasets for one cell line\n",
    "def gen_pred(train_data, test_data, cols, cell_line):\n",
    "    ## initialize dictionaries to store the predictions\n",
    "    train_pred = {}\n",
    "    test_pred = {}\n",
    "    ## iterate through all the tas cutoffs\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        # store the training and testing datasets at that tas cutoff\n",
    "        train_tas = train_data[cell_line][tas_cutoff]\n",
    "        test_tas = test_data[cell_line][tas_cutoff]\n",
    "        \n",
    "        # train the model on all the training data\n",
    "        X_train, y_train = train_tas[cols], train_tas[\"known ind\"]\n",
    "        lrm.fit(X_train, y_train)\n",
    "        # add the predicted probabilities to the training dataset\n",
    "        train_tas[\"predicted prob\"] = lrm.predict_proba(X_train)[:, 1]\n",
    "        # add the predicted probabilities to the testing dataset\n",
    "        X_test = test_tas[cols]\n",
    "        test_tas[\"predicted prob\"] = lrm.predict_proba(X_test)[:, 1]\n",
    "        \n",
    "        # store the predictions in their respective dictionaries\n",
    "        train_pred[tas_cutoff] = train_tas\n",
    "        test_pred[tas_cutoff] = test_tas\n",
    "        \n",
    "    return train_pred, test_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "57add3d4-c624-46b7-849c-b9a8021ca532",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>drug1 tas</th>\n",
       "      <th>drug2</th>\n",
       "      <th>drug2 tas</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "      <th>known ind</th>\n",
       "      <th>predicted prob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>macitentan</td>\n",
       "      <td>0.561786</td>\n",
       "      <td>phenylbutyrate</td>\n",
       "      <td>0.508335</td>\n",
       "      <td>0.741032</td>\n",
       "      <td>Urea Cycle Disorders, Inborn</td>\n",
       "      <td>False</td>\n",
       "      <td>0.759802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>valrubicin</td>\n",
       "      <td>0.794730</td>\n",
       "      <td>doxorubicin</td>\n",
       "      <td>0.732233</td>\n",
       "      <td>0.725946</td>\n",
       "      <td>Urinary Bladder Neoplasms</td>\n",
       "      <td>True</td>\n",
       "      <td>0.746831</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>doxorubicin</td>\n",
       "      <td>0.732233</td>\n",
       "      <td>valrubicin</td>\n",
       "      <td>0.794730</td>\n",
       "      <td>0.725946</td>\n",
       "      <td>Urinary Bladder Neoplasms</td>\n",
       "      <td>True</td>\n",
       "      <td>0.746831</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>valrubicin</td>\n",
       "      <td>0.794730</td>\n",
       "      <td>doxorubicin</td>\n",
       "      <td>0.732233</td>\n",
       "      <td>0.725946</td>\n",
       "      <td>Precursor Cell Lymphoblastic Leukemia-Lymphoma</td>\n",
       "      <td>False</td>\n",
       "      <td>0.746831</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>lurasidone</td>\n",
       "      <td>0.566841</td>\n",
       "      <td>palbociclib</td>\n",
       "      <td>0.575843</td>\n",
       "      <td>0.712300</td>\n",
       "      <td>Breast Neoplasms</td>\n",
       "      <td>False</td>\n",
       "      <td>0.734706</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          drug1  drug1 tas           drug2  drug2 tas  spearman corr  \\\n",
       "4    macitentan   0.561786  phenylbutyrate   0.508335       0.741032   \n",
       "12   valrubicin   0.794730     doxorubicin   0.732233       0.725946   \n",
       "14  doxorubicin   0.732233      valrubicin   0.794730       0.725946   \n",
       "16   valrubicin   0.794730     doxorubicin   0.732233       0.725946   \n",
       "19   lurasidone   0.566841     palbociclib   0.575843       0.712300   \n",
       "\n",
       "                                        indication  known ind  predicted prob  \n",
       "4                     Urea Cycle Disorders, Inborn      False        0.759802  \n",
       "12                       Urinary Bladder Neoplasms       True        0.746831  \n",
       "14                       Urinary Bladder Neoplasms       True        0.746831  \n",
       "16  Precursor Cell Lymphoblastic Leukemia-Lymphoma      False        0.746831  \n",
       "19                                Breast Neoplasms      False        0.734706  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### initialize dictionaries to store training and testing datasets with predictions\n",
    "train_pred = {}\n",
    "test_pred = {}\n",
    "\n",
    "### iterate through each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## generate the predictions for that cell line\n",
    "    cols = [\"spearman corr\"]\n",
    "    train_tas, test_tas = gen_pred(train_data, test_data, cols, cell_line)\n",
    "    ## store the datasets to their respective dictionaries\n",
    "    train_pred[cell_line] = train_tas\n",
    "    test_pred[cell_line] = test_tas\n",
    "\n",
    "train_pred[\"MCF7\"][0.3].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "9657dd64-5c5f-4e05-a30a-9d94f0c48a71",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to determine the cell line weights\n",
    "def gen_weights(train_pred, tas_cutoff):\n",
    "    ## store the predictions for each cell line\n",
    "    cols = [\"drug1\", \"indication\", \"predicted prob\", \"known ind\"]\n",
    "    A375 = train_pred[\"A375\"][tas_cutoff][cols]\n",
    "    MCF7 = train_pred[\"MCF7\"][tas_cutoff][cols]\n",
    "    PC3 = train_pred[\"PC3\"][tas_cutoff][cols]\n",
    "        \n",
    "    ## 1. generate the intersection dataset across all cell lines\n",
    "    shared_data = pd.merge(A375, MCF7, on=[\"drug1\", \"indication\"], how=\"inner\", suffixes=(\"_A375\", \"_MCF7\"))\n",
    "    shared_data = pd.merge(shared_data, PC3, on=[\"drug1\", \"indication\"], how=\"inner\")\n",
    "    shared_data.rename(columns={\"predicted prob\": \"predicted prob_PC3\"}, inplace=True)\n",
    "    ##\n",
    "        \n",
    "    ## 2. calculate the cell line weights using another LRM trained on their predictions\n",
    "    # define the inputs and outputs to the LRM\n",
    "    X, y = shared_data.drop(columns=[\"drug1\", \"indication\", \"known ind\"], axis=1), shared_data[\"known ind\"].astype(int)\n",
    "    # train the LRM of this data\n",
    "    lrm.fit(X, y)\n",
    "    # create a dictionary to store the weights\n",
    "    tas_weights = {\"MCF7\": lrm.coef_[0][0],\n",
    "                   \"A375\": lrm.coef_[0][1],\n",
    "                   \"PC3\": lrm.coef_[0][2]}\n",
    "    ##\n",
    "    \n",
    "    ## 3. generate the weighted average dataset (union across all cell lines)\n",
    "    weighted_train = pd.merge(A375, MCF7, on=[\"drug1\", \"indication\"], how=\"outer\", suffixes=(\"_A375\", \"_MCF7\"))\n",
    "    weighted_train = pd.merge(weighted_train, PC3, on=[\"drug1\", \"indication\"], how=\"outer\")\n",
    "    weighted_train.rename(columns={\"predicted prob\": \"predicted prob_PC3\",\n",
    "                                   \"known ind\": \"known ind_PC3\"}, inplace=True)\n",
    "    # combine the labelled outputs across all cell lines into one column\n",
    "    cols = [\"known ind_A375\", \"known ind_MCF7\", \"known ind_PC3\"]\n",
    "    weighted_train[\"known ind\"] = weighted_train[cols].any(axis=1)\n",
    "    weighted_train.drop(cols, axis=1, inplace=True)\n",
    "    weighted_train.rename(columns={\"predicted prob_A375\": \"A375\",\n",
    "                                   \"predicted prob_MCF7\": \"MCF7\",\n",
    "                                   \"predicted prob_PC3\": \"PC3\",}, inplace=True)\n",
    "    ##\n",
    "        \n",
    "    return weighted_train, tas_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "9b8f273b-877d-4751-92ae-f560cf317760",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to calculate the weighted average probability for one row\n",
    "def weighted_avg(row, tas_weights):\n",
    "    ## initialize variables to store the weighted sum and total weight\n",
    "    total_weight = 0\n",
    "    weighted_sum = 0\n",
    "    ## iterate through each weight\n",
    "    for col, weight in tas_weights.items():\n",
    "        # if drug-indication is in cell line dataset\n",
    "        if not pd.isnull(row[col]):\n",
    "            # add the weights and values to their respectives variables\n",
    "            weighted_sum += row[col] * weight\n",
    "            total_weight += weight\n",
    "    \n",
    "    ## calculate the weighted probability\n",
    "    return weighted_sum / total_weight"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "96337f13-f75f-408f-9f50-f0310aae4e5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate the predictions for the weighted average model\n",
    "def gen_weighted_avg(train_pred, test_pred):\n",
    "    ## intialize dictionaries to store the data for the weighted average models\n",
    "    weighted_train = {}\n",
    "    weighted_test = {}\n",
    "    cell_line_weights = {}\n",
    "    \n",
    "    ## iterate through each tas cutoff\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        ## 1. get the weighted average predictions across the training datasets\n",
    "        # generate the weights for that tas cutoff\n",
    "        tas_weighted_train, tas_weights = gen_weights(train_pred, tas_cutoff)\n",
    "        # add the weighted probabilities\n",
    "        tas_weighted_train[\"predicted prob\"] = tas_weighted_train.apply(weighted_avg, axis=1, tas_weights=tas_weights)\n",
    "        # store the data in the dictionary\n",
    "        weighted_train[tas_cutoff] = tas_weighted_train\n",
    "        cell_line_weights[tas_cutoff] = tas_weights\n",
    "        ##\n",
    "\n",
    "\n",
    "        ## 2. get the weighted average predictions across the testing datasets\n",
    "        # store the predictions for each cell line in the testing dataset\n",
    "        cols = [\"drug1\", \"indication\", \"predicted prob\", \"known ind\"]\n",
    "        A375 = test_pred[\"A375\"][tas_cutoff][cols]\n",
    "        MCF7 = test_pred[\"MCF7\"][tas_cutoff][cols]\n",
    "        PC3 = test_pred[\"PC3\"][tas_cutoff][cols]\n",
    "        # generate the weighted average dataset (union across all cell lines)\n",
    "        tas_weighted_test = pd.merge(A375, MCF7, on=[\"drug1\", \"indication\"], how=\"outer\", suffixes=(\"_A375\", \"_MCF7\"))\n",
    "        tas_weighted_test = pd.merge(tas_weighted_test, PC3, on=[\"drug1\", \"indication\"], how=\"outer\")\n",
    "        tas_weighted_test.rename(columns={\"predicted prob\": \"predicted prob_PC3\",\n",
    "                                       \"known ind\": \"known ind_PC3\"}, inplace=True)\n",
    "        # combine the labelled outputs across all cell lines into one column\n",
    "        cols = [\"known ind_A375\", \"known ind_MCF7\", \"known ind_PC3\"]\n",
    "        tas_weighted_test[\"known ind\"] = tas_weighted_test[cols].any(axis=1)\n",
    "        tas_weighted_test.drop(cols, axis=1, inplace=True)\n",
    "        tas_weighted_test.rename(columns={\"predicted prob_A375\": \"A375\",\n",
    "                                       \"predicted prob_MCF7\": \"MCF7\",\n",
    "                                       \"predicted prob_PC3\": \"PC3\",}, inplace=True)\n",
    "        # add the weighted probabilities\n",
    "        tas_weighted_test[\"predicted prob\"] = tas_weighted_test.apply(weighted_avg, axis=1, tas_weights=tas_weights)\n",
    "        # store the data in the dictionary\n",
    "        weighted_test[tas_cutoff] = tas_weighted_test\n",
    "        ##\n",
    "    \n",
    "    return weighted_train, weighted_test, cell_line_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "cc15136c-5850-4f28-bf27-83c2bb706f1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>indication</th>\n",
       "      <th>A375</th>\n",
       "      <th>MCF7</th>\n",
       "      <th>PC3</th>\n",
       "      <th>known ind</th>\n",
       "      <th>predicted prob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>pitavastatin</td>\n",
       "      <td>Myocardial Infarction</td>\n",
       "      <td>0.896607</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0.896607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>clofarabine</td>\n",
       "      <td>Leukemia, Hairy Cell</td>\n",
       "      <td>0.886319</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0.886319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>clofarabine</td>\n",
       "      <td>Pancreatic Neoplasms</td>\n",
       "      <td>0.882300</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0.882300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>rebamipide</td>\n",
       "      <td>Attention Deficit Disorder with Hyperactivity</td>\n",
       "      <td>0.875802</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0.875802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>atomoxetine</td>\n",
       "      <td>Gastritis</td>\n",
       "      <td>0.875802</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>0.875802</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          drug1                                     indication      A375  \\\n",
       "0  pitavastatin                          Myocardial Infarction  0.896607   \n",
       "1   clofarabine                           Leukemia, Hairy Cell  0.886319   \n",
       "2   clofarabine                           Pancreatic Neoplasms  0.882300   \n",
       "3    rebamipide  Attention Deficit Disorder with Hyperactivity  0.875802   \n",
       "4   atomoxetine                                      Gastritis  0.875802   \n",
       "\n",
       "   MCF7  PC3  known ind  predicted prob  \n",
       "0   NaN  NaN      False        0.896607  \n",
       "1   NaN  NaN      False        0.886319  \n",
       "2   NaN  NaN      False        0.882300  \n",
       "3   NaN  NaN      False        0.875802  \n",
       "4   NaN  NaN      False        0.875802  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## redefine tas cutoffs\n",
    "tas_cutoffs = [0, 0.2, 0.25]\n",
    "\n",
    "# calculate and store the weighted probabilities for training and testing datasets\n",
    "train_pred[\"weighted avg\"], test_pred[\"weighted avg\"], cell_line_weights = gen_weighted_avg(train_pred, test_pred)\n",
    "train_pred[\"weighted avg\"][0.2].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "60aa077e-b828-4869-a095-2f041fe9ee78",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'A375': 1.2854715050256023,\n",
       " 'MCF7': 0.38506050766427624,\n",
       " 'PC3': 0.09197353144750729}"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cell_line_weights[0.2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "d8b31a04-92c7-414e-97f3-5ef3aaa5252e",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate AUCs for training and testing datasets for one cell line\n",
    "def gen_train_test_scores(train_pred, test_pred, model):\n",
    "    ## store all predictions into one dictionary\n",
    "    pred = {\"train\": train_pred,\n",
    "            \"test\": test_pred}\n",
    "    \n",
    "    ## initialize a dataframe to store the AUCs\n",
    "    train_test_scores = pd.DataFrame()\n",
    "    ## iterate through each tas cutoff\n",
    "    for tas_cutoff in tas_cutoffs:\n",
    "        ## initialize a dictionary to store the evaluation metric scores at tas cutoff\n",
    "        tas_aucs = {}\n",
    "        tas_aucs[\"TAS cutoff\"] = tas_cutoff\n",
    "        ## iterate through each dataset\n",
    "        for dataset in pred:\n",
    "            # get the model predictions and actual labels\n",
    "            tas_data = pred[dataset][model][tas_cutoff]\n",
    "            tas_pred = tas_data[\"predicted prob\"]\n",
    "            tas_actual = tas_data[\"known ind\"]\n",
    "            tas_pred_classes = (tas_pred > 0.5).astype(int)\n",
    "            \n",
    "            # store the evaluation metric scores\n",
    "            tas_aucs[dataset + \"_AUC\"] = roc_auc_score(tas_actual, tas_pred)\n",
    "            tas_aucs[dataset + \"_F1 score\"] = f1_score(tas_actual, tas_pred_classes)\n",
    "            tas_aucs[dataset + \"_precision\"] = precision_score(tas_actual, tas_pred_classes)\n",
    "            tas_aucs[dataset + \"_recall\"] = recall_score(tas_actual, tas_pred_classes)\n",
    "            \n",
    "        ## add the aucs at that tas cutoff to the dataframe\n",
    "        train_test_scores = train_test_scores.append(tas_aucs, ignore_index=True)\n",
    "        \n",
    "    return train_test_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "838be5ad-9ae7-487b-8cf4-899dd2505c20",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TAS cutoff</th>\n",
       "      <th>test_AUC</th>\n",
       "      <th>test_F1 score</th>\n",
       "      <th>test_precision</th>\n",
       "      <th>test_recall</th>\n",
       "      <th>train_AUC</th>\n",
       "      <th>train_F1 score</th>\n",
       "      <th>train_precision</th>\n",
       "      <th>train_recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.711766</td>\n",
       "      <td>0.101834</td>\n",
       "      <td>0.054919</td>\n",
       "      <td>0.698675</td>\n",
       "      <td>0.664884</td>\n",
       "      <td>0.094306</td>\n",
       "      <td>0.051011</td>\n",
       "      <td>0.623529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.20</td>\n",
       "      <td>0.675630</td>\n",
       "      <td>0.116583</td>\n",
       "      <td>0.063596</td>\n",
       "      <td>0.698795</td>\n",
       "      <td>0.685574</td>\n",
       "      <td>0.117202</td>\n",
       "      <td>0.064182</td>\n",
       "      <td>0.673913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.25</td>\n",
       "      <td>0.688780</td>\n",
       "      <td>0.138389</td>\n",
       "      <td>0.076761</td>\n",
       "      <td>0.701923</td>\n",
       "      <td>0.741159</td>\n",
       "      <td>0.142222</td>\n",
       "      <td>0.079208</td>\n",
       "      <td>0.695652</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   TAS cutoff  test_AUC  test_F1 score  test_precision  test_recall  \\\n",
       "0        0.00  0.711766       0.101834        0.054919     0.698675   \n",
       "1        0.20  0.675630       0.116583        0.063596     0.698795   \n",
       "2        0.25  0.688780       0.138389        0.076761     0.701923   \n",
       "\n",
       "   train_AUC  train_F1 score  train_precision  train_recall  \n",
       "0   0.664884        0.094306         0.051011      0.623529  \n",
       "1   0.685574        0.117202         0.064182      0.673913  \n",
       "2   0.741159        0.142222         0.079208      0.695652  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### initialize a dictionary to store all of the evaluation metrics for all cell lines\n",
    "train_test_scores = {}\n",
    "\n",
    "### iterate through each cell line\n",
    "for model in train_pred.keys():\n",
    "    ## generate and store the evaluation metric scores for cell line in dictionary\n",
    "    train_test_scores[model] = gen_train_test_scores(train_pred, test_pred, model)\n",
    "train_test_scores[\"MCF7\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "631b3244-0a5d-49a2-97a3-cf302fe46639",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>indication</th>\n",
       "      <th>A375</th>\n",
       "      <th>MCF7</th>\n",
       "      <th>PC3</th>\n",
       "      <th>known ind</th>\n",
       "      <th>predicted prob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>atorvastatin</td>\n",
       "      <td>Hyperlipidemias</td>\n",
       "      <td>0.928433</td>\n",
       "      <td>0.786987</td>\n",
       "      <td>0.868162</td>\n",
       "      <td>True</td>\n",
       "      <td>0.899704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cladribine</td>\n",
       "      <td>Precursor Cell Lymphoblastic Leukemia-Lymphoma</td>\n",
       "      <td>0.920844</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.724302</td>\n",
       "      <td>True</td>\n",
       "      <td>0.905612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>gemcitabine</td>\n",
       "      <td>Precursor Cell Lymphoblastic Leukemia-Lymphoma</td>\n",
       "      <td>0.917859</td>\n",
       "      <td>0.738818</td>\n",
       "      <td>0.728660</td>\n",
       "      <td>True</td>\n",
       "      <td>0.874283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>belinostat</td>\n",
       "      <td>Lymphoma, T-Cell, Cutaneous</td>\n",
       "      <td>0.900137</td>\n",
       "      <td>0.806237</td>\n",
       "      <td>0.857158</td>\n",
       "      <td>True</td>\n",
       "      <td>0.880876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>vorinostat</td>\n",
       "      <td>Lymphoma, T-Cell, Peripheral</td>\n",
       "      <td>0.900137</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.857158</td>\n",
       "      <td>True</td>\n",
       "      <td>0.896806</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          drug1                                      indication      A375  \\\n",
       "0  atorvastatin                                 Hyperlipidemias  0.928433   \n",
       "1    cladribine  Precursor Cell Lymphoblastic Leukemia-Lymphoma  0.920844   \n",
       "2   gemcitabine  Precursor Cell Lymphoblastic Leukemia-Lymphoma  0.917859   \n",
       "3    belinostat                     Lymphoma, T-Cell, Cutaneous  0.900137   \n",
       "4    vorinostat                    Lymphoma, T-Cell, Peripheral  0.900137   \n",
       "\n",
       "       MCF7       PC3  known ind  predicted prob  \n",
       "0  0.786987  0.868162       True        0.899704  \n",
       "1       NaN  0.724302       True        0.905612  \n",
       "2  0.738818  0.728660       True        0.874283  \n",
       "3  0.806237  0.857158       True        0.880876  \n",
       "4       NaN  0.857158       True        0.896806  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_pred[\"weighted avg\"][0].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f71fc0d8-8917-4371-b9fe-e9d72d2500a0",
   "metadata": {},
   "source": [
    "### d. Save predictions for future access"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ac3e6723-7a74-43aa-ad4b-bc63e4949fad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drug1</th>\n",
       "      <th>most correlated drug</th>\n",
       "      <th>spearman corr</th>\n",
       "      <th>indication</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>metronidazole</td>\n",
       "      <td>lacidipine</td>\n",
       "      <td>0.928048</td>\n",
       "      <td>Hypertension</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>atorvastatin</td>\n",
       "      <td>pitavastatin</td>\n",
       "      <td>0.860484</td>\n",
       "      <td>Hyperlipidemias</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>pitavastatin</td>\n",
       "      <td>atorvastatin</td>\n",
       "      <td>0.860484</td>\n",
       "      <td>Myocardial Infarction</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>pitavastatin</td>\n",
       "      <td>atorvastatin</td>\n",
       "      <td>0.860484</td>\n",
       "      <td>Stroke</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cladribine</td>\n",
       "      <td>clofarabine</td>\n",
       "      <td>0.836042</td>\n",
       "      <td>Precursor Cell Lymphoblastic Leukemia-Lymphoma</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           drug1 most correlated drug  spearman corr  \\\n",
       "0  metronidazole           lacidipine       0.928048   \n",
       "4   atorvastatin         pitavastatin       0.860484   \n",
       "3   pitavastatin         atorvastatin       0.860484   \n",
       "2   pitavastatin         atorvastatin       0.860484   \n",
       "5     cladribine          clofarabine       0.836042   \n",
       "\n",
       "                                       indication  \n",
       "0                                    Hypertension  \n",
       "4                                 Hyperlipidemias  \n",
       "3                           Myocardial Infarction  \n",
       "2                                          Stroke  \n",
       "5  Precursor Cell Lymphoblastic Leukemia-Lymphoma  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### create a dataframe\n",
    "all_pred = pd.DataFrame()\n",
    "### iterate through predictions for each cell line\n",
    "for cell_line in cell_lines:\n",
    "    ## add cell line predictions to dataframe\n",
    "    all_pred = pd.concat([all_pred, test_pred[cell_line][0]])\n",
    "### remove unnecessary columns\n",
    "all_pred.drop(columns=[\"drug1 tas\", \"drug2 tas\", \"known ind\", \"predicted prob\"], axis=1, inplace=True)\n",
    "### sort data from highest to lowest spearman correlation\n",
    "all_pred.sort_values(\"spearman corr\", ascending=False, inplace=True)\n",
    "### keep only the highest correlated drug 2 for each drug-indication\n",
    "all_pred.drop_duplicates(subset=[\"drug1\", \"indication\"], keep=\"first\", inplace=True)\n",
    "### rename columns\n",
    "all_pred.rename({\"drug2\": \"most correlated drug\"}, axis=1, inplace=True)\n",
    "all_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "4048ce6a-df44-498a-9514-e20bca8c5a73",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a new dataframe containing the weighted average ensemble model's predictions\n",
    "cols = [\"drug1\", \"indication\", \"known ind\", \"predicted prob\"]\n",
    "model_pred = test_pred[\"weighted avg\"][0][cols]\n",
    "# add information about the most correlated drug\n",
    "model_pred = pd.merge(model_pred, all_pred, on=[\"drug1\", \"indication\"], how=\"left\")\n",
    "# rename column\n",
    "model_pred.rename({\"drug1\": \"drug\"}, axis=1, inplace=True)\n",
    "# model_pred.to_csv(\"model_pred.txt\", sep=\"\\t\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "28d62284-9536-4bd7-80d4-9fd12d78c7ca",
   "metadata": {},
   "source": [
    "### e. ROC Curves of Models on Clinical Trials"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "68c71c78-396e-4bab-8852-6b5c0bee3454",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate the weighted ensemble model trained only on spearman correlation\n",
    "def gen_tas_model(tas_cutoffs):\n",
    "    ### 1. generate the cell line model predictions\n",
    "    train_pred = {}\n",
    "    test_pred = {}\n",
    "    ## iterate through each cell line\n",
    "    for cell_line in cell_lines:\n",
    "        # generate the predictions for that cell line\n",
    "        cols = [\"drug1 tas\", \"drug2 tas\", \"spearman corr\"]\n",
    "        train_tas, test_tas = gen_pred(train_data, test_data, cols, cell_line)\n",
    "        # store the datasets to their respective dictionaries\n",
    "        train_pred[cell_line] = train_tas\n",
    "        test_pred[cell_line] = test_tas\n",
    "    ###\n",
    "        \n",
    "    ### 2. calculate and store the weighted probabilities for training and testing datasets\n",
    "    train_pred[\"weighted avg\"], test_pred[\"weighted avg\"], cell_line_weights = gen_weighted_avg(train_pred, test_pred)\n",
    "    ###\n",
    "    \n",
    "    return train_pred, test_pred, cell_line_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "83e43cd7-3a20-4b5a-b0a7-a3b895af444c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: {'A375': 2.3050960060655266,\n",
       "  'MCF7': 0.5342571428875991,\n",
       "  'PC3': 0.16752145108912309},\n",
       " 0.2: {'A375': 1.2679337404141477,\n",
       "  'MCF7': 0.3805524433422714,\n",
       "  'PC3': 0.1799294882196162},\n",
       " 0.25: {'A375': 0.7885868453276291,\n",
       "  'MCF7': 0.31556450629191574,\n",
       "  'PC3': 0.31472110947314735}}"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# generate the models trained only spearman correlation\n",
    "train_tas, test_tas, tas_weights = gen_tas_model(tas_cutoffs)\n",
    "train_pred[\"weighted avg w/ TAS\"] = train_tas[\"weighted avg\"]\n",
    "test_pred[\"weighted avg w/ TAS\"] = test_tas[\"weighted avg\"]\n",
    "tas_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "4b472a87-47eb-491e-a004-1fabf6af44da",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate ROC curve with AUCs for all model at TAS > 0.2\n",
    "def gen_intersection_rocs(tas_cutoff, ax):\n",
    "    ## store the intersection of the testing datasets at that tas cutoff\n",
    "    test = test_pred[\"weighted avg\"][tas_cutoff]\n",
    "    test = test.rename({\"predicted prob\": \"weighted avg\"}, axis=1)\n",
    "    test.dropna(inplace=True)\n",
    "    ## store a list with all models of interest\n",
    "    models = [\"A375\", \"MCF7\", \"PC3\", \"weighted avg\"]\n",
    "    ## iterate through the data for each model\n",
    "    for model in models:\n",
    "        # generate the ROC curve\n",
    "        y_actual = test[\"known ind\"]\n",
    "        y_pred = test[model]\n",
    "        fpr, tpr, _ = roc_curve(y_actual, y_pred)\n",
    "        auc = roc_auc_score(y_actual, y_pred)\n",
    "        ax.plot(fpr, tpr, label = model + \" (AUC = \" + str(round(auc, 3)) + \")\")\n",
    "    \n",
    "    ## add labels and legend\n",
    "    ax.set_title(\"ROC Curves on Clinical Trials Across Shared Drug Indications\")\n",
    "    ax.set_xlabel(\"False Positive Rate (FPR)\")\n",
    "    ax.set_ylabel(\"True Positive Rate\")\n",
    "    ax.legend(loc=\"lower right\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "880eddcb-b02b-43dd-8c6f-229ac7a5b48c",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate ROC curve with AUCs for all models\n",
    "def compare_tas(tas_cutoff, ax):\n",
    "    ## store a list with all models of interest\n",
    "    models = [\"weighted avg\", \"weighted avg w/ TAS\"]\n",
    "    ## iterate through the predicted probabilties for each model\n",
    "    for model in models:\n",
    "        ## store the intersection of the testing datasets at that tas cutoff\n",
    "        test = test_pred[model][tas_cutoff]\n",
    "        # generate the ROC curve\n",
    "        y_actual = test[\"known ind\"]\n",
    "        y_pred = test[\"predicted prob\"]\n",
    "        fpr, tpr, _ = roc_curve(y_actual, y_pred)\n",
    "        auc = roc_auc_score(y_actual, y_pred)\n",
    "        plt.plot(fpr, tpr, label = model + \" (AUC = \" + str(round(auc, 3)) + \")\")\n",
    "    \n",
    "    ## add labels and legend\n",
    "    ax.set_title(\"ROC Curves on Clinical Trials\")\n",
    "    ax.set_xlabel(\"False Positive Rate (FPR)\")\n",
    "    ax.set_ylabel(\"True Positive Rate\")\n",
    "    ax.legend(loc=\"lower right\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "b70148d9-a0ac-44ff-bbf7-3042c02fee8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "### function to generate figure comparing AUCs on Clinical Trials\n",
    "def compare_aucs(tas_cutoff):\n",
    "    ## create an empty figure to place 2 subplots side by side\n",
    "    fig, (ax1, ax2) = plt.subplots(nrows=1, ncols=2, figsize=(20,10))\n",
    "    \n",
    "    ## generate the left plot comparing the ROC curves on shared drug indication of clinical trials\n",
    "    gen_intersection_rocs(tas_cutoff, ax1)\n",
    "    ## generate the right plot comparing the ROC curves of model with and without TAS\n",
    "    compare_tas(tas_cutoff, ax2)\n",
    "    \n",
    "    ## show figure\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "3ceccefa-18df-4c79-a083-59aaf050908d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "compare_aucs(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "c174d0f2-c541-4751-8e65-f166f2f243ca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "compare_aucs(0.2)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "CMapPy_env",
   "language": "python",
   "name": "cmappy_env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}