Targets.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "dd4092d7-ce3d-4d78-aac6-4b100e13027d",
   "metadata": {
    "tags": []
   },
   "source": [
    "# packages and imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "77ad7a12-b6f3-4dcb-8e35-4e948e18243f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-08-23 13:19:25.313288: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-08-23 13:19:25.474248: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
      "2023-08-23 13:19:25.481119: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory\n",
      "2023-08-23 13:19:25.481141: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.\n",
      "2023-08-23 13:19:25.514136: E tensorflow/stream_executor/cuda/cuda_blas.cc:2981] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
      "2023-08-23 13:19:35.111025: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory\n",
      "2023-08-23 13:19:35.111151: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory\n",
      "2023-08-23 13:19:35.111164: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n",
      "2023-08-23 13:19:48.760906: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory\n",
      "2023-08-23 13:19:48.760949: W tensorflow/stream_executor/cuda/cuda_driver.cc:263] failed call to cuInit: UNKNOWN ERROR (303)\n",
      "2023-08-23 13:19:48.760977: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (sdx48): /proc/driver/nvidia/version does not exist\n"
     ]
    }
   ],
   "source": [
    "#imports\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from statistics import mean\n",
    "\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "from alibi.explainers import IntegratedGradients\n",
    "\n",
    "\n",
    "from data_makers import *\n",
    "import utils\n",
    "from utils import *\n",
    "import mean_model\n",
    "from mean_model import meanModel\n",
    "import ML_models\n",
    "from ML_models import *\n",
    "\n",
    "import sklearn\n",
    "\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "\n",
    "from Nikhil_code import data_preprocessing as dp_nb\n",
    "from Nikhil_code import model_selection as ms_nb\n",
    "from Nikhil_code import testing as t_nb\n",
    "import Nikhil_code.Data_imports as di_nb\n",
    "import Nikhil_code.pairs_train_test_split as tts_nb\n",
    "import Nikhil_code.Learning_curve as lc_nb\n",
    "\n",
    "import shap"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "addb9ebb-ecbc-4166-a31a-7cc41694d115",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Feature selection (FS) and data creating for each drug"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "54f7fa21-d750-4a39-a907-1d0526bda60e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Falling back to the 'python' engine because the 'c' engine does not support skipfooter; you can avoid this warning by specifying engine='python'.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of missing prot values 0.386335609896865\n",
      "num non overlapping prot and target cls: 10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Falling back to the 'python' engine because the 'c' engine does not support skipfooter; you can avoid this warning by specifying engine='python'.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "num non overlapping cls: 930\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((38, 8457), (38, 17417), (38, 22804), (38, 38), (345, 345))"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#input data\n",
    "prot, rna, phospho_ls, one_hot_cls, one_hot_drugs, ic50_df1 = di_nb.read_input_data(transformation='norm')\n",
    "_all_cls = prot.index\n",
    "_all_drugs = ic50_df1.columns\n",
    "assert prot.shape[0] == rna.shape[0] == phospho_ls.shape[0]\n",
    "assert phospho_ls.shape[0]  == one_hot_cls.shape[0]\n",
    "prot.shape, rna.shape, phospho_ls.shape, one_hot_cls.shape, one_hot_drugs.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "48da9d79-f104-4bed-833e-984954088463",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(11583, 22804)\n",
      "(11583, 23149)\n"
     ]
    }
   ],
   "source": [
    "##PHOSPHO\n",
    "#create prot data for all drugs\n",
    "x_all_phos, x_drug, y_main_phos = dp_nb.create_all_drugs(\n",
    "    phospho_ls, one_hot_drugs, ic50_df1, _all_cls)\n",
    "\n",
    "#fmt index to include drug cell line paris\n",
    "cls_drugs_index = x_all_phos.index + '::' + x_drug.index \n",
    "x_all_phos.index = cls_drugs_index\n",
    "y_main_phos.index = cls_drugs_index\n",
    "x_drug.index = cls_drugs_index\n",
    "\n",
    "x_all_phos = x_all_phos.astype(np.float32)\n",
    "print(x_all_phos.shape)\n",
    "X_main_phos = X_main_maker(x_all_phos, x_drug, short = False)\n",
    "print(X_main_phos.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1fbdee40-ca26-4f30-9294-1b7050f40a12",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(11583, 8457)\n",
      "(11583, 8802)\n"
     ]
    }
   ],
   "source": [
    "##PROTEOMIC\n",
    "#create prot data for all drugs\n",
    "x_all_prot, x_drug, y_list = dp_nb.create_all_drugs(\n",
    "    prot, one_hot_drugs, ic50_df1, _all_cls)\n",
    "\n",
    "#fmt index to include drug cell line paris\n",
    "cls_drugs_index = x_all_prot.index + '::' + x_drug.index \n",
    "x_all_prot.index = cls_drugs_index\n",
    "y_list.index = cls_drugs_index\n",
    "x_drug.index = cls_drugs_index\n",
    "\n",
    "x_all_prot = x_all_prot.astype(np.float32)\n",
    "print(x_all_prot.shape)\n",
    "X_main_prot = X_main_maker(x_all_prot, x_drug, short = False)\n",
    "print(X_main_prot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "a081fc6f-5d46-4989-a5b4-eae5d5ea1c12",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(11583, 38)\n",
      "(11583, 383)\n"
     ]
    }
   ],
   "source": [
    "#one hot data creation for all drugs\n",
    "x_hot, x_drug_hot, y_main_onehot = dp_nb.create_all_drugs(\n",
    "    one_hot_cls, one_hot_drugs, ic50_df1, _all_cls)\n",
    "\n",
    "cls_drugs_index_hot = x_hot.index + '::' + x_drug_hot.index \n",
    "\n",
    "x_hot.index = cls_drugs_index_hot\n",
    "x_hot.columns = np.arange(len(x_drug.columns), len(x_drug.columns) + len(x_hot.columns))\n",
    "print(x_hot.shape)\n",
    "X_main_onehot = X_main_maker(x_hot, x_drug, short = False)\n",
    "print(X_main_onehot.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1684003-3ddb-4662-9eda-4635d9b0a541",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Early stopping and parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a8ac959f-8308-4ba8-b836-bb1f0880bb02",
   "metadata": {},
   "outputs": [],
   "source": [
    "# set early stopping and learning rate scheduler\n",
    "es = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=20, min_delta=0, restore_best_weights=True)\n",
    "lr_scheduler = keras.callbacks.LearningRateScheduler(scheduler)\n",
    "callbacks_list =  [es,lr_scheduler]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "f3cf2f52-0aa8-4c56-8575-523b1ab3b411",
   "metadata": {},
   "outputs": [],
   "source": [
    "# set parameters and train\n",
    "num_epochs = 200 # max epoch for training\n",
    "learning_rate = 1e-2 # initial learning rate, decays via learning rate scheduler\n",
    "momentum = 0.5\n",
    "initializer_seed = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3bae2f5f-bcb9-4037-b195-37a1098c0335",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Filter for drugs, features or all"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df067fc7-593e-4708-979f-86e123d86b9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "model_type = 'rfr'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0a1bdc43-dfa8-488e-8de6-0fbd2ec017ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "# function that filters the X and y dataframe according to the Anti-Cancer drug database \n",
    "test_type = 'drugs_features'\n",
    "\n",
    "if dtype == 'phospho':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='all')\n",
    "elif dtype == 'proteomic':\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='all')\n",
    "elif dtype == 'full_network':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='all')\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='all')\n",
    "    dd_X = pd.concat([dd_X_phos, dd_X_prot], axis=1)\n",
    "\n",
    "\n",
    "#create prot data for all drugs\n",
    "dd_x_all, dd_x_drug, dd_y_main = dp_nb.create_all_drugs(\n",
    "dd_X_prot, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "#fmt index to include drug cell line paris\n",
    "cls_drugs_index = dd_x_all.index + '::' + dd_x_drug.index \n",
    "dd_x_all.index = cls_drugs_index\n",
    "dd_y_main.index = cls_drugs_index\n",
    "dd_x_drug.index = cls_drugs_index\n",
    "\n",
    "dd_x_all = dd_x_all.astype(np.float32)\n",
    "print(dd_x_all.shape)\n",
    "dd_X_main = X_main_maker(dd_x_all, dd_x_drug, short = False)\n",
    "print(dd_X_main.shape)\n",
    "\n",
    "## ONE HOT\n",
    "#one hot data creation for all drugs\n",
    "x_hot, x_drug_hot, y_main_onehot = dp_nb.create_all_drugs(\n",
    "    one_hot_cls, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "cls_drugs_index_hot = x_hot.index + '::' + x_drug_hot.index \n",
    "\n",
    "x_hot.index = cls_drugs_index_hot\n",
    "x_hot.columns = np.arange(len(dd_x_drug.columns), len(dd_x_drug.columns) + len(x_hot.columns))\n",
    "print(x_hot.shape)\n",
    "X_main_onehot = X_main_maker(x_hot, dd_x_drug, short = False)\n",
    "print(X_main_onehot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "34ec045f-f68c-4393-9d73-3a30b774673d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#CHANGE JUST CL:DRUG\n",
    "\n",
    "test_type = 'drugs'\n",
    "\n",
    "# function that filters the X and y dataframe according to the Anti-Cancer drug database \n",
    "if dtype == 'phospho':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='drug')\n",
    "elif dtype == 'proteomic':\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='drug')\n",
    "elif dtype == 'full_network':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='drug')\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='drug')\n",
    "    dd_X = pd.concat([dd_X_phos, dd_X_prot], axis=1)\n",
    "\n",
    "\n",
    "#create prot data for all drugs\n",
    "dd_x_all, dd_x_drug, dd_y_main = dp_nb.create_all_drugs(\n",
    "dd_X_prot, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "#fmt index to include drug cell line paris\n",
    "cls_drugs_index = dd_x_all.index + '::' + dd_x_drug.index \n",
    "dd_x_all.index = cls_drugs_index\n",
    "dd_y_main.index = cls_drugs_index\n",
    "dd_x_drug.index = cls_drugs_index\n",
    "\n",
    "dd_x_all = dd_x_all.astype(np.float32)\n",
    "print(dd_x_all.shape)\n",
    "dd_X_main = X_main_maker(dd_x_all, dd_x_drug, short = False)\n",
    "print(dd_X_main.shape)\n",
    "\n",
    "## ONE HOT\n",
    "#one hot data creation for all drugs\n",
    "x_hot, x_drug_hot, y_main_onehot = dp_nb.create_all_drugs(\n",
    "    one_hot_cls, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "cls_drugs_index_hot = x_hot.index + '::' + x_drug_hot.index \n",
    "\n",
    "x_hot.index = cls_drugs_index_hot\n",
    "x_hot.columns = np.arange(len(x_drug.columns), len(x_drug.columns) + len(x_hot.columns))\n",
    "print(x_hot.shape)\n",
    "X_main_onehot = X_main_maker(x_hot, dd_x_drug, short = False)\n",
    "print(X_main_onehot.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cf10b410-f185-4cc7-8090-0a233740e045",
   "metadata": {},
   "outputs": [],
   "source": [
    "#CHANGE JUST FEATURES\n",
    "\n",
    "test_type = 'features'\n",
    "\n",
    "# function that filters the X and y dataframe according to the Anti-Cancer drug database \n",
    "if dtype == 'phospho':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='feature')\n",
    "elif dtype == 'proteomic':\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='feature')\n",
    "elif dtype == 'full_network':\n",
    "    dd_X_phos, dd_ic50_df1, dd_one_hot_drugs = phospho_target_maker(X_main_phos, phospho_ls, ic50_df1, one_hot_drugs, ctype='feature')\n",
    "    dd_X_prot, dd_ic50_df1, dd_one_hot_drugs = proteomic_target_maker(X_main_prot, prot, ic50_df1, one_hot_drugs, ctype='feature')\n",
    "    dd_X = pd.concat([dd_X_phos, dd_X_prot], axis=1)\n",
    "\n",
    "\n",
    "#create prot data for all drugs\n",
    "dd_x_all, dd_x_drug, dd_y_main = dp_nb.create_all_drugs(\n",
    "dd_X_phos, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "#fmt index to include drug cell line paris\n",
    "cls_drugs_index = dd_x_all.index + '::' + dd_x_drug.index \n",
    "dd_x_all.index = cls_drugs_index\n",
    "dd_y_main.index = cls_drugs_index\n",
    "dd_x_drug.index = cls_drugs_index\n",
    "\n",
    "dd_x_all = dd_x_all.astype(np.float32)\n",
    "print(dd_x_all.shape)\n",
    "dd_X_main = X_main_maker(dd_x_all, dd_x_drug, short = False)\n",
    "print(dd_X_main.shape)\n",
    "\n",
    "## ONE HOT\n",
    "#one hot data creation for all drugs\n",
    "x_hot, x_drug_hot, y_main_onehot = dp_nb.create_all_drugs(\n",
    "    one_hot_cls, dd_one_hot_drugs, dd_ic50_df1, _all_cls)\n",
    "\n",
    "cls_drugs_index_hot = x_hot.index + '::' + x_drug_hot.index \n",
    "\n",
    "x_hot.index = cls_drugs_index_hot\n",
    "x_hot.columns = np.arange(len(x_drug_hot.columns), len(x_drug.columns) + len(x_hot.columns))\n",
    "print(x_hot.shape)\n",
    "X_main_onehot = X_main_maker(x_hot, dd_x_drug, short = False)\n",
    "print(X_main_onehot.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f579fb29-553d-4899-b016-02b82a36fa3a",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Test models for drug targets"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d8464ca-38b7-4075-9efd-2b146c50da92",
   "metadata": {},
   "source": [
    "reform dataframes to only contain drugs and targets that have known interactions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf06e625-33b7-4402-94ac-ab5141265f2f",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## Random forest"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dfda4ac4-09eb-4c37-bcf5-efd91af6a397",
   "metadata": {},
   "source": [
    "drug and features filtered"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e95d141d-7133-48ea-850a-cfa1041dd397",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "model_r2 = []\n",
    "model_standard_error = []\n",
    "mixed_set_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    print(f'Seed: {seed}')\n",
    "    #tts\n",
    "    dd_X_train, dd_X_test, dd_y_train, dd_y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(dd_y_train, noRepeats=True)\n",
    "    mm = meanModel(dd_y_train, dl)\n",
    "    prediction = mm.predict(dd_y_test.index)\n",
    "    mean_r2 = r2_score(dd_y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #MIXED SET\n",
    "    #MODEL\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    #random forests\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    model_r2.append(r2)\n",
    "    model_standard_error.append(standard_error)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "799e528c-137d-47d7-8b25-561b20f1c320",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'drugs_features_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_standard_error, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_standard_error', 'model_MSE', 'model_pearson'], \n",
    "           file = f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "           "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d462eedc-fefa-4f29-a574-711522a9eb39",
   "metadata": {},
   "source": [
    "drugs filtered only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0fc09fbc-8854-4dd0-a98b-442272082dad",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "model_r2 = []\n",
    "model_standard_error = []\n",
    "mixed_set_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    print(f'Seed: {seed}')\n",
    "    #tts\n",
    "    dd_X_train, dd_X_test, dd_y_train, dd_y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(dd_y_train, noRepeats=True)\n",
    "    mm = meanModel(dd_y_train, dl)\n",
    "    prediction = mm.predict(dd_y_test.index)\n",
    "    mean_r2 = r2_score(dd_y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #MIXED SET\n",
    "    #MODEL\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    #random forests\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    model_r2.append(r2)\n",
    "    model_standard_error.append(standard_error)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7e90d48f-d499-43ee-b090-975599039ee7",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'drugs_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_standard_error, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_standard_error', 'model_MSE', 'model_pearson'],\n",
    "           f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bdbdb1aa-bfc1-4d2d-898d-850ac3f6f77a",
   "metadata": {},
   "source": [
    "features filtered only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "071cffc5-3b37-4ada-894c-e6df6211d619",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "model_r2 = []\n",
    "model_standard_error = []\n",
    "mixed_set_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    print(f'Seed: {seed}')\n",
    "    #tts\n",
    "    dd_X_train, dd_X_test, dd_y_train, dd_y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(dd_y_train, noRepeats=True)\n",
    "    mm = meanModel(dd_y_train, dl)\n",
    "    prediction = mm.predict(dd_y_test.index)\n",
    "    mean_r2 = r2_score(dd_y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #MIXED SET\n",
    "    #MODEL\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(X_main_onehot, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    #random forests\n",
    "    if model_type == 'rfr':\n",
    "        r2, MSE, pearson, standard_error = rfr(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    elif model_type == 'xgb':\n",
    "        r2, MSE, pearson, standard_error = xgb(dd_X_main, dd_y_main, test_size=0.2, random_state=seed, iterations=3)\n",
    "    model_r2.append(r2)\n",
    "    model_standard_error.append(standard_error)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d201ae2b-aa97-4343-b640-7bbb04932389",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'features_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_standard_error, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_standard_error', 'model_MSE', 'model_pearson'],\n",
    "           f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a239464-615f-4448-a1e8-865cf234741b",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Neural Networks"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0db6b26a-8f8a-4d1c-a36a-c51ebd2cfed1",
   "metadata": {},
   "source": [
    "drugs and features filtered"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ffa697a4-1469-4d7e-8d44-a7652fd46071",
   "metadata": {},
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "mixed_set_r2 = []\n",
    "model_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    \n",
    "    print(f'Seed: {seed}')\n",
    "    \n",
    "    print(f'Mean Model')\n",
    "    #MEAN MODEL\n",
    "    #tts\n",
    "    X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, y=dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(y_train, noRepeats=True)\n",
    "    mm = meanModel(y_train, dl)\n",
    "    prediction = mm.predict(y_test.index)\n",
    "    mean_r2 = r2_score(y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(x_hot, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "        \n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    print(f'Model: Deep_NN')\n",
    "    \n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "    \n",
    "    ## Code for model building and training\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "    MSE = mean_squared_error(y_test, y_pred)\n",
    "    pearson = pearsonr(y_test, y_pred)\n",
    "    model_r2.append(r2)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a17334f0-e31e-424e-9be0-322db437360e",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'drugs_features_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_MSE', 'model_pearson'], \n",
    "           file = f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "           "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "408971fe-dbc4-428c-b5ed-4d82dd3269f5",
   "metadata": {},
   "source": [
    "drugs filtered only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3cd44696-ab40-405d-acd4-14868cbf57fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "mixed_set_r2 = []\n",
    "model_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    \n",
    "    print(f'Seed: {seed}')\n",
    "    \n",
    "    print(f'Mean Model')\n",
    "    #MEAN MODEL\n",
    "    #tts\n",
    "    X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, y=dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(y_train, noRepeats=True)\n",
    "    mm = meanModel(y_train, dl)\n",
    "    prediction = mm.predict(y_test.index)\n",
    "    mean_r2 = r2_score(y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(x_hot, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "        \n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    print(f'Model: Deep_NN')\n",
    "    \n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "    \n",
    "    ## Code for model building and training\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "    MSE = mean_squared_error(y_test, y_pred)\n",
    "    pearson = pearsonr(y_test, y_pred)\n",
    "    model_r2.append(r2)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36e84a5b-7806-4e6a-9526-3b44a8f2921b",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'drugs_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_MSE', 'model_pearson'],\n",
    "           f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1cb51715-889a-4a48-93ea-c3dc13b14db5",
   "metadata": {},
   "source": [
    "features filtered only"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50171e2b-00e8-4862-99c4-a2b8816e06d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "rs = [0, 8, 23, 42, 69, 88]\n",
    "mm_r2 = []\n",
    "mixed_set_r2 = []\n",
    "model_r2 = []\n",
    "model_MSE = []\n",
    "model_pearson = []\n",
    "\n",
    "for seed in rs:\n",
    "    \n",
    "    print(f'Seed: {seed}')\n",
    "    \n",
    "    print(f'Mean Model')\n",
    "    #MEAN MODEL\n",
    "    #tts\n",
    "    X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, y=dd_y_main, test_size=0.2, random_state = seed)\n",
    "    #run mean model\n",
    "    dl = dlMaker(y_train, noRepeats=True)\n",
    "    mm = meanModel(y_train, dl)\n",
    "    prediction = mm.predict(y_test.index)\n",
    "    mean_r2 = r2_score(y_test, list(prediction.values()))\n",
    "    mm_r2.append(mean_r2)\n",
    "    \n",
    "    \n",
    "    print(f'Mixed-Set Model')\n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(x_hot, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "        \n",
    "    mixed_set_r2.append(r2)\n",
    "\n",
    "    print(f'Model: Deep_NN')\n",
    "    \n",
    "    #phospho\n",
    "    xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = seed)\n",
    "    xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = seed)\n",
    "    y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = seed)\n",
    "    \n",
    "    ## Code for model building and training\n",
    "\n",
    "    model = build_CNN(xo_train, xd_train, learning_rate, momentum, initializer_seed, mtype = 'regression')\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "\n",
    "    y_pred = model.predict([xo_test, xd_test])  \n",
    "    r2 = r2_score(y_test, y_pred)\n",
    "    MSE = mean_squared_error(y_test, y_pred)\n",
    "    pearson = pearsonr(y_test, y_pred)\n",
    "    model_r2.append(r2)\n",
    "    model_MSE.append(MSE)\n",
    "    model_pearson.append(pearson)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d537eac8-d464-4df7-a67e-1dcb6a0e30b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "table_make(f'features_{dtype}',\n",
    "           [rs, mm_r2, mixed_set_r2, model_r2, model_MSE, model_pearson],\n",
    "           ['rs', 'mm_r2', 'mixed_set_r2', 'model_r2', 'model_MSE', 'model_pearson'],\n",
    "           f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "96829e69-35b4-4900-864a-d0c42b4fc9bb",
   "metadata": {
    "tags": []
   },
   "source": [
    "# performance plots"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83742c56-4585-4527-8125-463ccd5fbf86",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## individual model metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b42998fb-01a6-4658-95f0-75e54956370b",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "model_type = 'xgb'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8a378bb8-0bb1-4547-b424-cc75c8a66251",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0179cb5c-8dd4-4716-9383-34131f31ca8a",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='proteomic'\n",
    "model_type = 'xgb'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "55c2ad85-18ca-4be4-811a-c91d46df2bed",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "71f82ede-c216-429a-a589-2b4a97017dfe",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='full_network'\n",
    "model_type = 'xgb'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2a910bc1-35fe-4ad3-8798-fc67a87ecc3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e44599dd-94b5-4084-826d-b23469704164",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "model_type = 'rfr'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58c8d2b6-b7f5-4383-97da-42f32c7b9112",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 1, figsize=(5, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes.plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes.plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes.plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes.set_title('drugs/features', style='italic')\n",
    "axes.set_ylabel('r-squared')\n",
    "\n",
    "\n",
    "# Plot 2\n",
    "#axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "#axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "#axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "#axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "#axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "#axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "#axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "#axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes.set_ylim(0.45, 0.85)\n",
    "#axes[1].set_ylim(0.5, 0.85)\n",
    "#axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "plt.savefig('plots/poor_result')\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e179e6d6-d252-446e-bb45-25efec0e6839",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='proteomic'\n",
    "model_type = 'rfr'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "048afe28-ba22-4f9d-a35e-0ab7bb167d59",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0ed31567-363c-42eb-9ed4-9b3884ee25c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='full_network'\n",
    "model_type = 'rfr'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2a19407d-f1a9-47b9-9f41-d041d11c9fd8",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}_NORMX.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}_NORMX.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}_NORMX.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "49027a90-a9fc-4c74-9ea5-51a2ea8929af",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "model_type = 'Deep_NN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8868eae9-87d7-47ab-98bf-59d962dfc974",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f290d408-2495-45e1-8d94-d7780325cd21",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='proteomic'\n",
    "model_type = 'Deep_NN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "02075c5a-0ec4-47ed-a56a-181d4b81791f",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "434b5a5c-ddfd-47a8-873f-c1bcd861db1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='full_network'\n",
    "model_type = 'Deep_NN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8de8ac23-746f-4a33-bb98-4a93d2effcac",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "27446bf2-0c1e-42b8-8d03-506a15b64c89",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "model_type = 'CNN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "270f574a-4e58-485b-a47c-3119e6373405",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fdcbd357-1a8b-416c-b3ea-154ea57c1159",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='proteomic'\n",
    "model_type = 'CNN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1a938621-3a0d-4fcb-b604-614325076183",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e08f2cc8-eaba-48ce-948b-98ec3c8bf051",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='full_network'\n",
    "model_type = 'CNN'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8fc3b7f7-c0d3-413c-be0e-d7ba77303f67",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_df_full = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "shap_df_drugs = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "shap_df_features = pd.read_csv(f'Targets/{model_type}/model performance/{dtype}/features_{dtype}.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full['rs'].astype(str), shap_df_full['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Full Network')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs['rs'].astype(str), shap_df_drugs['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Drugs Filtered')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Features Filtered')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6174045-a111-4a6c-83ce-e7148dbaf6d2",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## features filtered NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f43773e4-8fc7-4191-b4a4-18db2f48e5ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "# extract pearson info from df\n",
    "shap_df_features1[\"model_pearson\"] = [float(f'{x[8]}{x[9]}{x[10]}{x[11]}{x[12]}') for x in shap_df_features1[\"model_pearson\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4ad553f-b72d-4bf9-ae09-bd7cb8cfd465",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'r2: {mean(shap_df_features1[\"model_r2\"])}\\n MSE: {mean(shap_df_features1[\"model_MSE\"])}\\n pearson: {mean(shap_df_features1[\"model_pearson\"])}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "d0b0bece-5da4-4eed-a965-6ce458d13df9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_type = 'Deep_NN'\n",
    "\n",
    "shap_df_full1 = pd.read_csv(f'Targets/{model_type}/model performance/phospho/features_phospho.csv')\n",
    "shap_df_drugs1 = pd.read_csv(f'Targets/{model_type}/model performance/proteomic/features_proteomic.csv')\n",
    "shap_df_features1 = pd.read_csv(f'Targets/{model_type}/model performance/full_network/features_full_network.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 3, figsize=(15, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes[0].plot(shap_df_full1['rs'].astype(str), shap_df_full1['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0].plot(shap_df_full1['rs'].astype(str), shap_df_full1['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0].plot(shap_df_full1['rs'].astype(str), shap_df_full1['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0].set_title('Phosphoproteomic')\n",
    "\n",
    "# Plot 2\n",
    "axes[1].plot(shap_df_drugs1['rs'].astype(str), shap_df_drugs1['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs1['rs'].astype(str), shap_df_drugs1['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1].plot(shap_df_drugs1['rs'].astype(str), shap_df_drugs1['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1].set_title('Proteomic')\n",
    "\n",
    "# Plot 3\n",
    "axes[2].plot(shap_df_features1['rs'].astype(str), shap_df_features1['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[2].plot(shap_df_features1['rs'].astype(str), shap_df_features1['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[2].plot(shap_df_features1['rs'].astype(str), shap_df_features1['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[2].set_title('Combined dataset')\n",
    "\n",
    "axes[2].legend([\"mean model\", \"mixed-set model\", \"model\"], loc=0)\n",
    "\n",
    "axes[0].set_ylabel('r-squared')\n",
    "\n",
    "axes[0].set_ylim(0.5, 0.85)\n",
    "axes[1].set_ylim(0.5, 0.85)\n",
    "axes[2].set_ylim(0.5, 0.85)\n",
    "\n",
    "axes[1].yaxis.set_ticks([])\n",
    "axes[2].yaxis.set_ticks([])\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.savefig('Targets/DNN_features_lines')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "82d533ce-bff4-413f-a4a8-b4e75712d3ae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "shap_df_features = pd.read_csv(f'Targets/CNN/model performance/full_network/features_full_network.csv')\n",
    "\n",
    "fig, axes = plt.subplots(1, 1, figsize=(5, 5))\n",
    "\n",
    "# Plot 1\n",
    "axes.plot(shap_df_features['rs'].astype(str), shap_df_features['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes.plot(shap_df_features['rs'].astype(str), shap_df_features['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes.plot(shap_df_features['rs'].astype(str), shap_df_features['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes.set_title('Combined dataset')\n",
    "axes.set_ylim(0.5, 0.85)\n",
    "\n",
    "axes.set_ylabel('r-squared')\n",
    "\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.savefig('Targets/CNN_features_combined dataset')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "162a64b8-a4a1-4f69-b689-4682ff418ac6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# extract pearson info from df\n",
    "shap_df_features[\"model_pearson\"] = [float(f'{x[8]}{x[9]}{x[10]}{x[11]}{x[12]}') for x in shap_df_features[\"model_pearson\"]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aa10df4f-9e91-4284-8558-69f2b6928be4",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'r2: {mean(shap_df_features[\"model_r2\"])}\\n MSE: {mean(shap_df_features[\"model_MSE\"])}\\n sd: {mean(shap_df_features[\"model_pearson\"])}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ade3fcfb-fbd7-40c4-8b31-719f1f5a140d",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.std(shap_df_features[\"model_r2\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "616a99b0-f809-43a0-91cf-854481269660",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## Proteomic and full_network for drugs/feats and drugs in Deep_NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ab8e441-2bb1-4cc3-98b1-9a53ba69cb93",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'r2: {mean(metrics40[\"model_r2\"])}\\n MSE: {mean(metrics40[\"model_MSE\"])}\\n sd: {np.std(metrics40[\"model_r2\"])}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "903ce479-3a24-4d4a-9c1b-1fed57d0d4b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics10 = pd.read_csv(f'Targets/Deep_NN/model performance/proteomic/drugs_features_proteomic.csv')\n",
    "metrics20 = pd.read_csv(f'Targets/Deep_NN/model performance/proteomic/drugs_proteomic.csv')\n",
    "metrics30 = pd.read_csv(f'Targets/Deep_NN/model performance/full_network/drugs_features_full_network.csv')\n",
    "metrics40 = pd.read_csv(f'Targets/Deep_NN/model performance/full_network/drugs_full_network.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "035add3c-abed-496e-bd6c-96c9ed4202d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(2, 2, figsize=(15, 10))\n",
    "\n",
    "# Plot 1\n",
    "axes[0,0].plot(metrics10['rs'].astype(str), metrics10['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0,0].plot(metrics10['rs'].astype(str), metrics10['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0,0].plot(metrics10['rs'].astype(str), metrics10['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0,0].set_title('Proteomic')\n",
    "axes[0,0].set_ylabel('r-squared')\n",
    "\n",
    "# Plot 2\n",
    "axes[1,0].plot(metrics20['rs'].astype(str), metrics20['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1,0].plot(metrics20['rs'].astype(str), metrics20['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1,0].plot(metrics20['rs'].astype(str), metrics20['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1,0].set_title('Proteomic')\n",
    "axes[1,0].set_ylabel('r-squared')\n",
    "\n",
    "\n",
    "# Plot 3\n",
    "axes[0,1].plot(metrics30['rs'].astype(str), metrics30['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0,1].plot(metrics30['rs'].astype(str), metrics30['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0,1].plot(metrics30['rs'].astype(str), metrics30['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0,1].set_title('Combined dataset')\n",
    "\n",
    "# Plot 4\n",
    "axes[1,1].plot(metrics40['rs'].astype(str), metrics40['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1,1].plot(metrics40['rs'].astype(str), metrics40['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1,1].plot(metrics40['rs'].astype(str), metrics40['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1,1].set_title('Combined dataset')\n",
    "\n",
    "axes[0,0].set_ylim(0.5, 0.85)\n",
    "axes[0,1].set_ylim(0.5, 0.85)\n",
    "axes[1,0].set_ylim(0.5, 0.85)\n",
    "axes[1,1].set_ylim(0.5, 0.85)\n",
    "\n",
    "axes[0,1].yaxis.set_ticks([])\n",
    "axes[1,1].yaxis.set_ticks([])\n",
    "\n",
    "fig.savefig('Targets/DNN_supplementary.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "88a6e11c-e376-4192-90c1-cfb025c1f8ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics1 = pd.read_csv(f'Targets/rfr/model performance/proteomic/drugs_proteomic.csv')\n",
    "metrics2 = pd.read_csv(f'Targets/rfr/model performance/proteomic/features_proteomic.csv')\n",
    "metrics3 = pd.read_csv(f'Targets/rfr/model performance/full_network/drugs_full_network.csv')\n",
    "metrics4 = pd.read_csv(f'Targets/rfr/model performance/full_network/features_full_network.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e7319002-742e-4827-ba91-617a93495984",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, axes = plt.subplots(2, 2, figsize=(15, 10))\n",
    "\n",
    "\n",
    "\n",
    "# Plot 1\n",
    "axes[0,0].plot(metrics1['rs'].astype(str), metrics1['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0,0].plot(metrics1['rs'].astype(str), metrics1['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0,0].plot(metrics1['rs'].astype(str), metrics1['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0,0].set_title('Proteomic')\n",
    "axes[0,0].set_ylabel('r-squared')\n",
    "\n",
    "# Plot 2\n",
    "axes[1,0].plot(metrics2['rs'].astype(str), metrics2['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1,0].plot(metrics2['rs'].astype(str), metrics2['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1,0].plot(metrics2['rs'].astype(str), metrics2['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1,0].set_title('Proteomic')\n",
    "axes[1,0].set_ylabel('r-squared')\n",
    "\n",
    "# Plot 3\n",
    "axes[0,1].plot(metrics3['rs'].astype(str), metrics3['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[0,1].plot(metrics3['rs'].astype(str), metrics3['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[0,1].plot(metrics3['rs'].astype(str), metrics3['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[0,1].set_title('Combined dataset')\n",
    "\n",
    "# Plot 4\n",
    "axes[1,1].plot(metrics4['rs'].astype(str), metrics4['mm_r2'], linestyle='-', marker='.', color='orange', linewidth=1)\n",
    "axes[1,1].plot(metrics4['rs'].astype(str), metrics4['mixed_set_r2'], linestyle='-', marker='.', color='yellow', linewidth=1)\n",
    "axes[1,1].plot(metrics4['rs'].astype(str), metrics4['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "axes[1,1].set_title('Combined dataset')\n",
    "\n",
    "axes[0,0].set_ylim(0.5, 0.85)\n",
    "axes[0,1].set_ylim(0.5, 0.85)\n",
    "axes[1,0].set_ylim(0.5, 0.85)\n",
    "axes[1,1].set_ylim(0.5, 0.85)\n",
    "\n",
    "fig.savefig('Targets/rfr_supplementary.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aeccb3ef-2cf4-4665-99d4-8c745907058d",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## overall model comparison across networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "223c1bbd-89b7-41c6-b80a-c72d727f9e0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='phospho'\n",
    "\n",
    "#CNN\n",
    "CNN_full = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "CNN_phos = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "CNN_prot = pd.read_csv(f'Targets/CNN/model performance/{dtype}/features_{dtype}.csv')\n",
    "CNN = [mean(CNN_full['model_r2']), mean(CNN_phos['model_r2']), mean(CNN_prot['model_r2'])]\n",
    "CNN_sd = [np.std(CNN_full['model_r2']), np.std(CNN_phos['model_r2']), np.std(CNN_prot['model_r2'])]\n",
    "\n",
    "#Deep_NN\n",
    "Deep_NN_full = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "Deep_NN_phos = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "Deep_NN_prot = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/features_{dtype}.csv')\n",
    "DNN = [mean(Deep_NN_full['model_r2']), mean(Deep_NN_phos['model_r2']), mean(Deep_NN_prot['model_r2'])]\n",
    "DNN_sd = [np.std(Deep_NN_full['model_r2']), np.std(Deep_NN_phos['model_r2']), np.std(Deep_NN_prot['model_r2'])]\n",
    "\n",
    "\n",
    "#xgb\n",
    "xgb_full = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "xgb_phos = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "xgb_prot = pd.read_csv(f'Targets/xgb/model performance/{dtype}/features_{dtype}.csv')\n",
    "xgb = [mean(xgb_full['model_r2']), mean(xgb_phos['model_r2']), mean(xgb_prot['model_r2'])]\n",
    "xgb_sd = [np.std(xgb_full['model_r2']), np.std(xgb_phos['model_r2']), np.std(xgb_prot['model_r2'])]\n",
    "\n",
    "#rfr\n",
    "rfr_full = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "rfr_phos = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "rfr_prot = pd.read_csv(f'Targets/rfr/model performance/{dtype}/features_{dtype}.csv')\n",
    "rfr = [mean(rfr_full['model_r2']), mean(rfr_phos['model_r2']), mean(rfr_prot['model_r2'])]\n",
    "rfr_sd = [np.std(rfr_full['model_r2']), np.std(rfr_phos['model_r2']), np.std(rfr_prot['model_r2'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "56c2d455-f1e8-4e9c-b834-a65aa5f3ab13",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [10, 5]\n",
    "\n",
    "Network = (\"drugs/features\", \"drugs\", \"features\")\n",
    "r_squared_means = {\n",
    "    'CNN': CNN,\n",
    "    'DNN': DNN,\n",
    "    'rfr': rfr,\n",
    "    'xgb': xgb\n",
    "}\n",
    "\n",
    "# Calculate standard deviations for each model\n",
    "r_squared_std = {\n",
    "    'CNN_sd': CNN_sd,\n",
    "    'DNN_sd': DNN_sd,\n",
    "    'rfr_sd': rfr_sd,\n",
    "    'xgb_sd': xgb_sd\n",
    "}\n",
    "\n",
    "x = np.arange(len(Network))  # the label locations\n",
    "width = 0.10  # the width of the bars\n",
    "multiplier = 0\n",
    "\n",
    "fig, ax = plt.subplots(constrained_layout=True)\n",
    "\n",
    "# Define custom colors for each model's bars\n",
    "colors = ['orange', 'yellow', 'green', 'red']\n",
    "\n",
    "for model, r in r_squared_means.items():\n",
    "    offset = width * multiplier\n",
    "    rects = ax.bar(x + offset, r, width, label=model, color=colors[multiplier])\n",
    "    #ax.bar_label(rects, padding=10)\n",
    "    ax.errorbar(x + offset, r, yerr=r_squared_std[model+'_sd'], fmt='none', color='black', capsize=5)\n",
    "    multiplier += 1\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('r-squared')\n",
    "ax.set_title('Phosphoproteomic')\n",
    "ax.set_xticks(x + width * (multiplier - 2.5))  # Adjusted the tick positions\n",
    "ax.set_xticklabels(Network, style='italic')\n",
    "ax.legend(loc='upper right')\n",
    "ax.set_ylim(0.5, 0.9)\n",
    "\n",
    "fig.savefig('Targets/phospho_targets_comparison.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "35dda7c3-1a79-44fc-b9b4-5bd91f79634c",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='proteomic'\n",
    "\n",
    "#CNN\n",
    "CNN_full = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "CNN_phos = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "CNN_prot = pd.read_csv(f'Targets/CNN/model performance/{dtype}/features_{dtype}.csv')\n",
    "CNN = [mean(CNN_full['model_r2']), mean(CNN_phos['model_r2']), mean(CNN_prot['model_r2'])]\n",
    "CNN_sd = [np.std(CNN_full['model_r2']), np.std(CNN_phos['model_r2']), np.std(CNN_prot['model_r2'])]\n",
    "\n",
    "#Deep_NN\n",
    "Deep_NN_full = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "Deep_NN_phos = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "Deep_NN_prot = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/features_{dtype}.csv')\n",
    "DNN = [mean(Deep_NN_full['model_r2']), mean(Deep_NN_phos['model_r2']), mean(Deep_NN_prot['model_r2'])]\n",
    "DNN_sd = [np.std(Deep_NN_full['model_r2']), np.std(Deep_NN_phos['model_r2']), np.std(Deep_NN_prot['model_r2'])]\n",
    "\n",
    "\n",
    "#xgb\n",
    "xgb_full = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "xgb_phos = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "xgb_prot = pd.read_csv(f'Targets/xgb/model performance/{dtype}/features_{dtype}.csv')\n",
    "xgb = [mean(xgb_full['model_r2']), mean(xgb_phos['model_r2']), mean(xgb_prot['model_r2'])]\n",
    "xgb_sd = [np.std(xgb_full['model_r2']), np.std(xgb_phos['model_r2']), np.std(xgb_prot['model_r2'])]\n",
    "\n",
    "#rfr\n",
    "rfr_full = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "rfr_phos = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "rfr_prot = pd.read_csv(f'Targets/rfr/model performance/{dtype}/features_{dtype}.csv')\n",
    "rfr = [mean(rfr_full['model_r2']), mean(rfr_phos['model_r2']), mean(rfr_prot['model_r2'])]\n",
    "rfr_sd = [np.std(rfr_full['model_r2']), np.std(rfr_phos['model_r2']), np.std(rfr_prot['model_r2'])]\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "daad7d3b-1579-4464-a230-b2ab3f890f0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [10, 5]\n",
    "\n",
    "Network = (\"drugs/features\", \"drugs\", \"features\")\n",
    "r_squared_means = {\n",
    "    'CNN': CNN,\n",
    "    'DNN': DNN,\n",
    "    'rfr': rfr,\n",
    "    'xgb': xgb\n",
    "}\n",
    "\n",
    "# Calculate standard deviations for each model\n",
    "r_squared_std = {\n",
    "    'CNN_sd': CNN_sd,\n",
    "    'DNN_sd': DNN_sd,\n",
    "    'rfr_sd': rfr_sd,\n",
    "    'xgb_sd': xgb_sd\n",
    "}\n",
    "\n",
    "x = np.arange(len(Network))  # the label locations\n",
    "width = 0.10  # the width of the bars\n",
    "multiplier = 0\n",
    "\n",
    "fig, ax = plt.subplots(constrained_layout=True)\n",
    "\n",
    "# Define custom colors for each model's bars\n",
    "colors = ['orange', 'yellow', 'green', 'red']\n",
    "\n",
    "for model, r in r_squared_means.items():\n",
    "    offset = width * multiplier\n",
    "    rects = ax.bar(x + offset, r, width, label=model, color=colors[multiplier])\n",
    "    #ax.bar_label(rects, padding=10)\n",
    "    ax.errorbar(x + offset, r, yerr=r_squared_std[model+'_sd'], fmt='none', color='black', capsize=5)\n",
    "    multiplier += 1\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('r-squared')\n",
    "ax.set_title('Proteomic')\n",
    "ax.set_xticks(x + width * (multiplier - 2.5))  # Adjusted the tick positions\n",
    "ax.set_xticklabels(Network, style='italic')\n",
    "ax.legend(loc='upper right')\n",
    "ax.set_ylim(0.5, 0.9)\n",
    "\n",
    "fig.savefig('Targets/proteomic_targets_comparison.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5a16dc9-4e2e-4ffd-9bd6-bed9a4b957b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype='full_network'\n",
    "\n",
    "#CNN\n",
    "CNN_full = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "CNN_phos = pd.read_csv(f'Targets/CNN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "CNN_prot = pd.read_csv(f'Targets/CNN/model performance/{dtype}/features_{dtype}.csv')\n",
    "CNN = [mean(CNN_full['model_r2']), mean(CNN_phos['model_r2']), mean(CNN_prot['model_r2'])]\n",
    "CNN_sd = [np.std(CNN_full['model_r2']), np.std(CNN_phos['model_r2']), np.std(CNN_prot['model_r2'])]\n",
    "\n",
    "#Deep_NN\n",
    "Deep_NN_full = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "Deep_NN_phos = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "Deep_NN_prot = pd.read_csv(f'Targets/Deep_NN/model performance/{dtype}/features_{dtype}.csv')\n",
    "DNN = [mean(Deep_NN_full['model_r2']), mean(Deep_NN_phos['model_r2']), mean(Deep_NN_prot['model_r2'])]\n",
    "DNN_sd = [np.std(Deep_NN_full['model_r2']), np.std(Deep_NN_phos['model_r2']), np.std(Deep_NN_prot['model_r2'])]\n",
    "\n",
    "\n",
    "#xgb\n",
    "xgb_full = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "xgb_phos = pd.read_csv(f'Targets/xgb/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "xgb_prot = pd.read_csv(f'Targets/xgb/model performance/{dtype}/features_{dtype}.csv')\n",
    "xgb = [mean(xgb_full['model_r2']), mean(xgb_phos['model_r2']), mean(xgb_prot['model_r2'])]\n",
    "xgb_sd = [np.std(xgb_full['model_r2']), np.std(xgb_phos['model_r2']), np.std(xgb_prot['model_r2'])]\n",
    "\n",
    "#rfr\n",
    "rfr_full = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_features_{dtype}.csv')\n",
    "rfr_phos = pd.read_csv(f'Targets/rfr/model performance/{dtype}/drugs_{dtype}.csv')\n",
    "rfr_prot = pd.read_csv(f'Targets/rfr/model performance/{dtype}/features_{dtype}.csv')\n",
    "rfr = [mean(rfr_full['model_r2']), mean(rfr_phos['model_r2']), mean(rfr_prot['model_r2'])]\n",
    "rfr_sd = [np.std(rfr_full['model_r2']), np.std(rfr_phos['model_r2']), np.std(rfr_prot['model_r2'])]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1be178ce-8bc2-4c74-aa52-132f0a95b1c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [10, 5]\n",
    "\n",
    "Network = (\"drugs/features\", \"drugs\", \"features\")\n",
    "r_squared_means = {\n",
    "    'CNN': CNN,\n",
    "    'DNN': DNN,\n",
    "    'rfr': rfr,\n",
    "    'xgb': xgb\n",
    "}\n",
    "\n",
    "# Calculate standard deviations for each model\n",
    "r_squared_std = {\n",
    "    'CNN_sd': CNN_sd,\n",
    "    'DNN_sd': DNN_sd,\n",
    "    'rfr_sd': rfr_sd,\n",
    "    'xgb_sd': xgb_sd\n",
    "}\n",
    "\n",
    "x = np.arange(len(Network))  # the label locations\n",
    "width = 0.10  # the width of the bars\n",
    "multiplier = 0\n",
    "\n",
    "fig, ax = plt.subplots(constrained_layout=True)\n",
    "\n",
    "# Define custom colors for each model's bars\n",
    "colors = ['orange', 'yellow', 'green', 'red']\n",
    "\n",
    "for model, r in r_squared_means.items():\n",
    "    offset = width * multiplier\n",
    "    rects = ax.bar(x + offset, r, width, label=model, color=colors[multiplier])\n",
    "    #ax.bar_label(rects, padding=10)\n",
    "    ax.errorbar(x + offset, r, yerr=r_squared_std[model+'_sd'], fmt='none', color='black', capsize=5)\n",
    "    multiplier += 1\n",
    "\n",
    "# Add some text for labels, title and custom x-axis tick labels, etc.\n",
    "ax.set_ylabel('r-squared')\n",
    "ax.set_title('Combined dataset')\n",
    "ax.set_xticks(x + width * (multiplier - 2.5))  # Adjusted the tick positions\n",
    "ax.set_xticklabels(Network, style='italic')\n",
    "ax.legend(loc='upper right')\n",
    "ax.set_ylim(0.5, 0.9)\n",
    "\n",
    "fig.savefig('Targets/full_targets_comparison.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40b062d6-13f3-45f1-93e2-0a97b9014d60",
   "metadata": {
    "tags": []
   },
   "source": [
    "# SHAP for drug targets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6e79dbeb-f5a6-4aa8-8eeb-5a4afbbe3538",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = 23)\n",
    "print(f'dd_X_main.shape: {dd_X_main.shape}')\n",
    "print(f'X_train.shape: {X_train.shape}, X_test.shape: {X_test.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a470dadf-ea47-46a6-bd2b-dccd3570980e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = 23)\n",
    "xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = 23)\n",
    "y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = 23)\n",
    "\n",
    "print(f'dd_x_all.shape: {dd_x_all.shape}')\n",
    "print(f'xo_train.shape: {xo_train.shape}, X_test.shape: {xo_test.shape}')\n",
    "\n",
    "print(f'dd_x_drug.shape: {dd_x_drug.shape}')\n",
    "print(f'xd_train.shape: {xd_train.shape}, X_test.shape: {xd_test.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7396aa3b-1064-4b5c-b781-cad82748a008",
   "metadata": {
    "tags": []
   },
   "source": [
    "## build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "437ed5b1-aab0-4ed7-afb9-62752910a534",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "model_type == 'Deep_NN'\n",
    "if model_type == 'Deep_NN':\n",
    "    ## fit the model\n",
    "\n",
    "    model = build_Deep_NN(xo_train, xd_train, learning_rate, momentum, initializer_seed, complexity = 4)\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "    shap.explainers._deep.deep_tf.op_handlers[\"AddV2\"] = shap.explainers._deep.deep_tf.passthrough\n",
    "\n",
    "    # DeepExplainer to explain predictions of the model\n",
    "    explainer = shap.DeepExplainer(model, [np.array(xo_train), np.array(xd_train)])\n",
    "\n",
    "elif model_type == 'rfr':\n",
    "\n",
    "    #run rfr\n",
    "    classify = RandomForestRegressor(n_jobs=-1, max_depth=300, n_estimators=200)\n",
    "\n",
    "    #fit model\n",
    "    classify.fit(X_train, y_train)\n",
    "\n",
    "    #creates explainer instance\n",
    "    explainer = shap.explainers.Tree(classify, X_train)\n",
    "\n",
    "elif model_type == 'xgb':\n",
    "    #run xgb\n",
    "    xgb = XGBRegressor(max_depth = 75, \n",
    "                       n_estimators = 300, \n",
    "                       seed = 42, \n",
    "                       min_child_weight = 3, \n",
    "                       gamma = 0, \n",
    "                       colsample_bytree = 0.3, \n",
    "                       reg_alpha = 0.1,\n",
    "                       n_jobs=-1)\n",
    "\n",
    "    #fit model\n",
    "    xgb.fit(X_train, y_train)\n",
    "\n",
    "    #creates explainer instance\n",
    "    explainer = shap.explainers.Tree(xgb, X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d062367c-38ab-41ef-90e3-366c12363d75",
   "metadata": {
    "tags": []
   },
   "source": [
    "## significant SHAP finder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1cad1cc5-7f8c-4d35-856a-eedf2ea670bd",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "indexes, features, targets = SHAP_targets(dd_X_main, explainer, X_test, y_test, dtype = 'phospho')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6666b6f4-e5ca-4a90-b2a5-b55341be5a50",
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_targs = [len(x) for x in targets[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "481a65fa-bad8-42b3-9c93-fabd228a81cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "SHAP_df = table_make('SHAPs',\n",
    "           [indexes, features[0], features[1], features[2], features[3], targets[0], number_of_targs, targets[1], targets[2], targets[3]],\n",
    "           ['cl:drug',' total number of features', 'target percentage of total', 'total number of significant features', 'target percentage of significant', 'drug targets', 'number of targets', 'percent significant', 'significant targets', 'insignificant targets'], file = False)\n",
    "SHAP_df = SHAP_df.sort_values('percent significant', ascending=False)\n",
    "SHAP_df.to_csv(f'Targets/{model_type}/explainability/SHAP/{dtype}/{test_type}_{dtype}2.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49e88a79-cacb-4b44-ab74-1b14042f8fe7",
   "metadata": {
    "tags": []
   },
   "source": [
    "## SHAP plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "07157ab5-62b4-422e-9fd9-c216392c01fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "index = index_finder('KYSE-150::Bosutinib', X_test)\n",
    "index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ffde7a80-7e52-4769-a9a9-d17358e80487",
   "metadata": {},
   "outputs": [],
   "source": [
    "#specify the cell line and drugs that are tested\n",
    "cell_line, drug = list(X_test.index)[index].split('::')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "91fc92eb-b2db-48a7-a945-a940dbb297a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(f'cell line: {cell_line}\\ndrug: {drug}\\nIC50: {y_test[f\"{cell_line}::{drug}\"]}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "188d192f-58ee-4f0a-86d3-7c8ac8b49884",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "if model_type == 'rfr':    \n",
    "    # make a shap value array for the prediction\n",
    "    shaps = explainer.shap_values(X_test.loc[f'{cell_line}::{drug}'], y_test[f'{cell_line}::{drug}'], check_additivity=False)\n",
    "    \n",
    "elif model_type == 'Deep_NN':\n",
    "    shaps = explainer.shap_values([np.array([xo_test.loc[f'{cell_line}::{drug}']]), np.array([xd_test.loc[f'{cell_line}::{drug}']])])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "de0ae372-e6b1-4e08-9a01-b4c35137404c",
   "metadata": {},
   "outputs": [],
   "source": [
    "dtype = 'full_network'\n",
    "\n",
    "# drug list and their associated proteins\n",
    "if dtype == 'phospho':\n",
    "    dd = drugData(X_main_phos, isPhospho = True)\n",
    "    target_lists = [i.split(', ') for i in dd['Targets']]\n",
    "    \n",
    "    drug_dict = {dd.index[i]:target_lists[i] for i in range(len(target_lists))}\n",
    "\n",
    "elif dtype == 'proteomic':\n",
    "    dd = drugData(X_main_prot)\n",
    "    target_lists = [i.split('; ') for i in dd['Targets']]\n",
    "    \n",
    "    drug_dict = {dd.index[i]:target_lists[i] for i in range(len(target_lists))}\n",
    "\n",
    "if dtype == 'full_network':\n",
    "    from data_makers import drugData\n",
    "    dd_phos = drugData(dd_X_main, isPhospho = True)\n",
    "    target_lists_phos = [i.split(', ') for i in dd_phos['Targets']]\n",
    "\n",
    "    #prot\n",
    "    from data_makers import drugData\n",
    "    dd_prot = drugData(dd_X_main)\n",
    "    target_lists_prot = [i.split('; ') for i in dd_prot['Targets']]\n",
    "        \n",
    "    # account for disparity between phos and prot drugs available\n",
    "    overlap_target_lists = [target_lists_phos[i] + target_lists_prot[i] for i in range(len(target_lists_phos))]\n",
    "    missingIndexes = len(target_lists_prot) - len(overlap_target_lists) # the added protein indexes not covered by phos indexes\n",
    "    prot_to_add = target_lists_prot[-missingIndexes:] # the section of array that these indexes cover\n",
    "    target_lists = overlap_target_lists + prot_to_add\n",
    "                \n",
    "    drug_dict = {dd_prot.index[i]:target_lists[i] for i in range(len(target_lists))}\n",
    "\n",
    "#targets for the specific drug\n",
    "shap_targets = drug_dict[drug]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7c146fa6-4e50-460d-ba61-2ab2cbf6b3d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "if model_type == 'rfr':\n",
    "    #find where the targets are situated in the shap list\n",
    "    target_and_index = [[i, v] for i, v in enumerate(X_test.loc[f'{cell_line}::{drug}'].index) if v in shap_targets]\n",
    "    #separate them for the plot\n",
    "    protein_targs = [l[1] for l in target_and_index]\n",
    "    indexes = [l[0] for l in target_and_index]\n",
    "elif model_type == 'Deep_NN':\n",
    "    #find where the targets are situated in the shap list\n",
    "    target_and_index = [[i, v] for i, v in enumerate(xo_test.loc[f'{cell_line}::{drug}'].index) if v in shap_targets]\n",
    "    #separate them for the plot\n",
    "    protein_targs = [l[1] for l in target_and_index]\n",
    "    indexes = [l[0] for l in target_and_index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "094fa7e9-c6e8-4846-a056-9797eb77dae1",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [20, 20]\n",
    "rot = plt.xticks(rotation=45)\n",
    "\n",
    "# Sample data\n",
    "if model_type == 'rfr':\n",
    "    x = list(range(0, len(X_test.loc[f'{cell_line}::{drug}'])))\n",
    "elif model_type == 'Deep_NN':\n",
    "    x = list(range(0, len(xo_test.loc[f'{cell_line}::{drug}'])))\n",
    "y = shaps[0][0][0]\n",
    "\n",
    "# List of indices for points to be crosses\n",
    "cross_indices = indexes\n",
    "\n",
    "# Add axis labels\n",
    "plt.xlabel('index')\n",
    "plt.ylabel('shapley scores')\n",
    "\n",
    "# Add a title\n",
    "plt.title('shapley scores for a DNN model trained on features_full_network predicting SNU-387::Pemetrexed')\n",
    "\n",
    "# Plotting\n",
    "plt.scatter(x, y, marker='o')  # Plot circles for all points\n",
    "plt.scatter([x[i] for i in cross_indices], [y[i] for i in cross_indices], marker='x', color='red')  # Plot crosses for specific points\n",
    "\n",
    "plt.savefig('Targets/example_scatter_SNU-387::Pemetrexed')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23016cc3-2964-49d9-8efa-dc3ebc1f6f0d",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## significance plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "b84364e5-3a1e-4d73-a9c3-96ec23c2c6bd",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_type1 = 'drugs_features'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "0ec9791b-36c8-41e2-88a3-cac2fd1c55f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df1 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label1 = f'RF_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "abd49d5e-5293-4191-90b5-2b248a766db3",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df2 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label2 = f'RF_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "e9c1403a-693c-44dc-866e-3b4e5900d50c",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df3 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label3 = f'DNN_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "27f87ecf-c1d2-47fe-b148-9b24d0787e4b",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df4 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label4 = f'DNN_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "7ac36f93-f203-461d-bd03-a1be5513129d",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'IG'\n",
    "SHAP_df5 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label5 = f'DNN_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "e476d517-d530-4164-9bed-091e5592620b",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'IG'\n",
    "SHAP_df6 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label6 = f'DNN_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "0fec5b85-8943-4f9b-b1b5-93f83e34614d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAALICAYAAABiqwZ2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAEAAElEQVR4nOzdeVjU1f7A8fdhR0AWt9xxScm0BMlbruBuWrnkklqYG2pWtug1U1HLUvPnRbMyU8CbllkuLTf3QNPI1FKz1FLBJbcUcFcQzu+PYSaGmcFhX/y8nmcemPM928z44IfD53uO0lojhBBCCCGEMHAo7gkIIYQQQghRkkiALIQQQgghRBYSIAshhBBCCJGFBMhCCCGEEEJkIQGyEEIIIYQQWUiALIQQQgghRBYSIAshyjylVC2llFZKrS7uuRQWpVRXpdQ2pdSlzNd6TSklP+OFECIP5IenEOJuEJT59edinUUBUUqtU0pdVkpVyHzeHPgaaADEANOAV7XWGUUwl5czA/IBhT2WEEIUFafinoAQQhQBY4C8p1hnUQCUUi2ALsB0rfXFzOKXAUegh9b6xyKeUpl5b4UQwkjJSXpCiLJOKfU/4FHgHq31ueKeT34opb4DmgJ1tNaXMsvOA2la6+rFMJ+DQA3AuyhWrIUQoihIioUQokxQSjkppV5USu1XSt1QSh1XSo1XSikMq5x/GYNjpdS0zLSAbkqpp5RSP2Tm7P6deX1i5vWeVsapbSufWSl1r1LqY6XUOaXU9cx+2yil+mS2eTlb/SpKqVlKqX1KqRSl1BWl1B9KqWVKqQAr/bcHQoHZWutLSqlflVIaqARUyxxDK6VWZGlTTSk1Ryl1MPN9uaSU+kYp9aCV/j2VUmOVUt8qpRIz659RSm1QSrXJVndW5tgBgCeQnmX8QZl1FmU+D7QyVtvMa3OzlW9VSmUopSoppV7PnHeqUurzLHXcM1M7dmammtxQSv2klOqVfZzM+q2VUquVUkeVUjeVUucz679lrb4QQkiKhRCi1FNKuQDfAB2BvcACoAKGXNx7gXsw5OgaGdMCRgMtgdXANuB2tuvW0gaCM7/+km0ObTPn4AasAo4DDwObgLXZ2yilGmeO6Qasz3y4A/WBPsBsK2O/CZwH3s18/imG1eQ+wGZgR2b595ljtAa+BLyA/2W+B/5ALyBEKfWw1vpAlv7bAm9lzutb4HLmfB4D2iul2mitf8jy3iwFwoAfMl+n0dbMr0FAKpB1DCOL9zHzl5mmwIXM+VbMnPNtYF9mnSqZr7Vx5utdjCFAfxJYpZQaorWOztLnRGAGcALYkNl3lczxuwATrcxNCHG301rLQx7ykEepfgAfARqYTGbqWGZ5m8xyDUzNUv5XZtmvGNIusveXAPxtY6y3M9t2z1JWFUPgeh4Izlb/v1nm4JOl/GsgHWhsZQxfwDFbWffMPsZmK389s7xXtvJ7gWQMgeF92a51y2yzMlt5VcDDynyMYy/MVj4is3yElTbOwC1gl4338dPMto2zzdn4Xi0HXLK1ccUQmN+y8nprAknAWf5JH6yCIbj+PntfmdcrFve/XXnIQx4l8yEpFkKIUi1zB4dhwDat9Rtaa9ONFVrrbcDBzKd7MutXBqphCLL6aa3PZuvPD8Mqq62bzowrn1l3xIjAkOYwQmu9O1t942rmMa11SpbyxsAVIDH7AFrrZK11epY5KWA6hsB+YbbqtnbomAd4Y7hx72DWC1rr/wFngH9lKz+jtb6WfT7881742Tk2QBPAhZzfxxv88/lk7e8wMExrnZqtzYuZdf6ttTZLcdFan8SwCl8Fw+cHhvQPR+APK32htb5gY25CiLucpFgIIUq75zO/TrFx3bjTgzGIa5b59Rut9e9W6huDtOyBbtbr57XWpwGUUh4Y0gx+01qvtWN8ox+AAcBepdSHmfM5iHW9gUBgpNb6ppX5JGutE40FSql6QFcMQfDjSqnHrfTpwj8pJcZ2AcAYIARDkOmRrc0pK2OnYViJz87m+6iU8gbqYVhdTrfSZoHW+oaVPsdgWF2uopSaauV6A+MQmV9/Ay4BQ5RSlTCsSm/UWidbaSuEECYSIAshSrtOGILQbTau1wXOaa3/ynxuvGHsKxv1jQG0xcqnUqouhlXU9VmK22DII95go7+qmV+zB8gjgKPA0xjyjWcrpY4A/9Fav59lTAcMudQJQFS2+fhiCGS/y9Z35yxjR9iYF8ChLH0NxLDarYFYDHm+l4EMDDcGtiEzDzizvhOGVeLftda3rPRt833MvKawfE9sfjZKqUYY0igAJuTwmtIxrLSjtb6glGqF4T14FEMudbpSahPwuta6TOyLLYQoeBIgCyFKLaWUG1AZ+CVrakWW6y0wpFOsy1JsXKXcaaNbY5BmLXh6NPNr1hv0amd+PWGjv1Br/WWmMkwBpmSu3PYBXgHeU0qd11p/kVl1INAIGKy1TsvWt60Uh1qZX9tprWNtzMtEKVUVw81uxzPbnMx2vV3mt1lXgxth+MXAVpAZiO0b9Ky9j8Y2Z7XW1t5L42uarrXOKeg3ow03IfbJvJGzDYZfTPoADymlqtsI7oUQdznJQRZClGbpmY/KNq5Py/yaNYgLwpD7+4eNNgEY9hQ+nrVQKeUKhFvpzxhg+WbvKDOfeaiVNma01oe01m9gCJDBEHwaV2kjMOTkLrPS1FaAbEwxqGhrzGw6YAh2l1gJjh8CWgHXybLijGG3CbAMco0CgBPZg/rM92Rg9nkrpWplztdWaktuX5MZrXWq1nqz1rovsB3DLidV8tKXEKLskwBZCFFqZQZffwLVlVKPZb2mlPo3hsAP/rlBzweoA/xsbcU5UyrgrJQy5rMa84yjMNxYB+ZBoXGF9CmllHuWNuWBlRgCulNaa+Meyx5KqU6ZqRNZ5+uIYbcI+Ge7tmcx5OpOyZara2QMkLMHqcbT9F5WSpXL3kgpVUEp1SRLkTGvOSBbvcbAFxiC01+yzaFC5tfLVuYFhvexamYaiLG/ihjek3sw5D9nXV2+04l8ezLbPJWZY539NbkqpR7J8jzQRr36GD7HE1jmVAshBCApFkKI0u9tDPvxrso8IOMshpvMmgAnMeStGlcq73QDHhhyiR8Ctiql1mDYY7c9sB/DTW/lgGPGylrrXUqpzRiC8f1KqW8wbEfWk39WqbOOF5w5ximlVDyGXSw8MOzJWxdYpLXekrliPRlD3u/nWBcEXMVyNfxLDHnEocAxZThJ8DyGnTb8gdbAKP65ue67zOthSqkKwO9AQww3+n2FIb0he+BqfD4jM5C+huFGReNcNwCDgF1KqXWZY3fEsL9xGnAw2w2HOX42WuvzmQd7TAEOKKW+wpCX7YXhJL/WGPahjs9s8kLm6/kJw8165zH8cmS8YXGIlpP/hBC2FPc+c/KQhzzkkd8Hhu2/jmEIvM5hOKgjCENO7YUs9V7BcBPaUzn05YZhi7TTGNIKdmNIrfDBcMNanJU2FTGkQKRkaTMS6Js53gtZ6tYD3sGQA30+c85JGA7a6JOl3guZbR+3MU/PzPlst3HdFcPNbHsz53Qj8/3YBIwD/LLVb4ohUL6C4abHr4FHgJcy5/G0lTHGYEj/uJlZZ0aWa34YfnH5O7PPbRhyf5tk1o3J1tfXmeVV7/BZ9wPiMOxOkYbhl5adGH5RCshSrwfwceb8LmNY0U7EkGt9b3H/m5WHPORRsh/GzdSFEEIUoMy9izdhOJ2ultb6TC7alsOww8UJrfW/7lRfCCFEwZIUCyGEyIfM/XWTtPnBHq7ALAypGf/JTXCcaQyGPN2wApuoEEIIuxX5CrJS6kVgOIabPj7SWkdm3tX8GYbcuESgr5aN3IUQpYBS6v8wpA58h+HP/VUxBMY1MGwv10NbOcVNCCFEyVWku1hk3sgxHGgOPAh0V0rdiyFPbovW+l5gCzlvAi+EECWJ8ea9x4BXMdwE9icwGOgmwbEQQpQ+RbqCrJTqA3TWWg/LfD4Zwx6iQ4EQrfWZzA3r47TWDYtsYkIIIYQQQmQq6hzkAxi2BKqA4Y7qRzHc7V3FmKOXGSRb3fRfKTUCwylIeHh4NAsICLBWTQghhBBCiDvas2fPBa11pezlRRoga60PKqVmYbiz+yqG/T1v56L9ImARQHBwsN69O6etTIUQQgghhLBNKXXcWnmRn6SntV6itQ7SWrfBsPfnn8C5zNQKMr+eL+p5CSGEEEIIAcUQIBvTJ5RStYBewKcYTmoybmcUhuEUKCGEEEIIIYpcceyDvCozBzkNeE5rnayUmgmsVEoNBU5g2DJJCCGEEEKIIlfkAbLWurWVsosY9g0VQgghhBCiWBV5ioUQQgghhBAlmQTIQgghhBBCZCEBshBCCCGEEFlIgCyEEEIIIUQWxbGLhRBC3JVu3bpFUlISV65cIT09vbinI4QQZYqjoyNeXl74+fnh6uqar74kQBZCiCJw69YtTpw4ga+vL/7+/jg7O6OUKu5pCSFEmaC1Ji0tjcuXL3PixAlq1aqVryBZUiyEEKIIJCUl4evrS8WKFXFxcZHgWAghCpBSChcXFypWrIivry9JSUn56k8CZCGEKAJXrlyhfPnyxT0NIYQo88qXL8+VK1fy1YcEyEIIUQTS09NxdnYu7mkIIUSZ5+zsnO/7PCRAFkKIIiJpFUIIUfgK4metBMhCCCGEEEJkIQGyEEIIIYQQWUiALIQQQgghRBYSIAshhBCiVEhMTEQpxeDBg+1uExMTg1KKmJiYQptXXkydOhWlFHFxcQXe9+DBg1FKkZiYWOB93y0kQBZCCFFklFJmD0dHR/z8/AgJCSEmJgattUUbY1CU0yOvQcbhw4cZPnw49evXx93dHQ8PD+rUqUOnTp2YPn06586dM6sfEhJyx/GMwUlOAdnJkydxdHREKcXEiRNznKO196xixYq0a9eO5cuX5+bl2pzLhAkTaNasGb6+vjg7O1O5cmU6dOjAvHnzuHTpUr7HEKWT8d/73UhO0hNCCFHkIiIiAEhLS+PIkSOsWbOGrVu3snv3bhYsWGC1jbe3N2PHjrV6zd/fP9dz+O677+jWrRs3b97kkUceoUuXLpQrV47ExET27t3Lpk2baNGiBVWqVMl133eyePFiMjIyUEoRHR3N9OnTcXLK+b/krO/Z4cOHWbt2LbGxsezZs4e5c+fmeR5jxozh1q1bPPjggzz11FP4+vpy8eJFtm/fztixY3njjTe4cOFCnvovCXr27MnDDz9M1apVi3sqohSRAFkIIUSRmzp1qtnzHTt20KZNG95//31eeeUV6tSpY9HGx8fHol1+hIeHc/PmTWJiYggLC7O4vn//fnx9fQtsPKP09HSioqIoX748AwcO5IMPPuCrr76iV69eObbL/tq3bNlCx44diYyM5IUXXsj1LwmffPIJw4cPx9fXl1WrVtGtWzeLOjt27OC5557LVb8ljbe3N97e3sU9DVHKSIqFEEKIYteyZUsCAgLQWrNnz55CH+/8+fMcOXIEb29vq8ExwAMPPEDNmjULfOx169Zx6tQp+vXrx+jRowH46KOPct1P+/btTe/Zrl27ctX2ypUrPP/88wCsWLHCanAMhs9l586dFuVbtmyhS5cu+Pn54ebmRoMGDZgwYYLVdAzjn+nT0tKYPn069erVw83NjYCAALPXvXDhQpo0aYK7uzs1atQgIiKCjIwMm6/h0KFD9OjRAz8/Pzw8PGjVqhUbN260qGcrB9nf3x9/f3+uX7/OuHHjqFWrFq6urtSvX59Zs2ZZTfcB2LlzJ08++ST33HMPLi4u1KxZk/DwcE6fPm21/p49e+jSpQteXl6UL1+eDh06EB8fb/N15cbmzZtp3bo1Hh4e+Pn50aNHDw4dOmSzfkxMDL1796Zu3bq4u7tTvnx5WrZsybJly8zqGdOatm7dCpin+YSEhJjqxcbGMmLECBo1akT58uVxd3encePGTJs2jZs3bxbIaywusoIshBCiRDAGJEVx4qC3tzdOTk5cvXqVM2fOFOmf3xctWgQYcpUbN25MUFAQGzdu5Pjx49SuXTtXfRnfs9zmiX7xxRckJSXx8MMP06lTpxzrurq6mj3/8MMPGTVqFB4eHvTp04fKlSsTFxfHrFmz+Prrr9mxYwc+Pj4W/fTv35+dO3fy6KOP4uzszBdffMGIESNwdnZm//79LF26lO7du9O+fXu++uorpk+fTrly5fj3v/9t0VdCQgKPPPIIjRs3Jjw8nDNnzvDZZ5/RtWtXPvnkE/r162fX+5CWlkanTp04ffo0Xbt2xcnJibVr1zJhwgRu3rxpSmsxio6OZvjw4bi6uvL4449Ts2ZN/vzzTxYvXszXX3/Njz/+SK1atUz1f/jhBzp06EBqaiq9evWifv367N27l5CQENq1a2fXHG354osv6NevHy4uLvTr14+qVauyfft2HnnkER544AGrbUaNGkWjRo1o06YNVatW5eLFi3z77bc8/fTTHD58mDfeeAMw/LUmIiKCmJgYjh8/bvY+ZP1LxaxZszh06BAtWrQwpSvt2LGDqVOnEhcXx+bNm3F0dMzX6ywuEiALIUQxU9NKz00wOsL6qlp+bdu2jcOHD+Pi4kLz5s2t1klJSbGaYtG0aVN69OiRq/FcXV154oknWLVqFa1atWLUqFG0bt2aJk2aUK5cuTu2j4mJsXmj3t69e222++uvv/j2229p0KABLVq0AAyB8gsvvMDixYtNAYo9Nm/ezOHDh1FK8dBDD9ndDmD79u2AYRU6N44fP84LL7yAp6cnP/30EwEBAaZro0eP5oMPPmD8+PGmXwKyOnHiBAcOHDAFz6+88goBAQG89NJL+Pj4sH//fqpXrw4Y0knq16/PnDlzeOWVVyzys7dt28arr77KO++8YyobM2YMjzzyCCNHjqRr166UL1/+jq/n9OnTPPjgg2zatAl3d3fAkOvdoEED/vOf/zBx4kTTL2x//PEH4eHh+Pv7s3XrVtNcwZDP3rFjR1588UXWrFkDGH55GTJkCDdu3GDt2rU88cQTpvrz5s2zmU9vj6tXrxIeHo6DgwPff/89wcHBpmsvvfQSkZGRVtsdOHCAevXqmZWlpqbStWtXZs6cyciRI6levbopnSkuLo7jx4/bTG16//33qVOnjsUvaJMnT+bNN980BfGlkQTIQgghipzxP9ysN+lprZkzZ47N1dxLly4xbdo0i/KwsLBcB8hgSGvQWrNmzRrGjRsHgIODA40bN+axxx7j+eeft3mD3tKlS3M9HsCSJUtIT08326ZswIABvPrqq0RFRTF16lSbK25Z3zPjTXpaa1566aVcrzyfOXMGgBo1auSq3bJly0hNTTUFt1nNmDGDZcuW8fHHH/Puu+9arDzPnDnTbGW5bt26tGrVitjYWP7v//7PLOD08fHhscceIyYmhr/++svi9Xl7ezNlyhSzsuDgYAYOHMjSpUtZs2aNzdSZ7ObPn28KjgEqV67ME088wX//+18OHz5M48aNAfjggw9IS0tj3rx5ZnMFaNeuHY8//jhff/01V65cwcvLix9++IHDhw/Tpk0bs+AYDMH8u+++y9GjR+2aY3ZffvklSUlJPPPMM2bBMRj+nURHR1tNd8keHAO4uLjw3HPP8d1337FlyxaeeeYZu+dRt25dq+Vjx47lzTffZMOGDRIgCyGEEPbKHugqpViyZAnPPvuszTa1a9cu0H1djTenJSYmsmHDBnbv3s2uXbvYv38/+/fv54MPPmD9+vVWV2djY2PNcjGzGjx4sNUAOiMjg6ioKBwcHMyCkAoVKtC9e3dWr17N//73Px5//HGr/RrfM6UUPj4+tG7dmqFDhzJo0KBcv/a8pmb8/PPPAFbTA3x9fQkMDGTbtm0cOnSIBx980Ox69kAOoFq1agA0a9bM4poxCD116pRFgBwUFISXl5dFm5CQEJYuXcovv/xiV4Ds7e1N/fr1LcqNuefJycmmMmPe8NatW63mfJ8/f5709HT++OMPmjVrZnqv2rZta1HX0dGRVq1a5TlAzqlvb29vmjZtasofzurEiRPMmjWLLVu2cOLECW7cuGF2/a+//srVPK5du8a8efNYs2YNf/zxB1euXDHL3c5tfyWJBMhCCFHMCittoSQz/id67do14uPjGTp0KCNHjqR27dr5zs3MLX9/f8LDwwkPDwcMAdno0aP5+uuvGT58eI4pE7mxYcMGjh8/TufOnS1WIJ999llWr17NokWLbAbItm4aywtjYHrq1KlctTOuStpa5TeWp6SkWFyztpOEMXUip2tpaWkW12yt7N9zzz1m87wTa7nSWcdOT083lV28eBHALK3DmqtXr5rN4U5zzYu89H3s2DGaN29OcnIyrVu3plOnTnh7e+Po6EhiYiJLly7l1q1bds8hLS2Ndu3a8dNPP9G4cWP69etHpUqVTCkp06ZNy1V/JY0EyEIIIYqNh4cHHTp04OuvvyYoKIiwsDAOHz5sVx5wYalRowYrVqzA19eXffv2cfHiRSpUqJDvfo15uRs2bLC5crt+/XpOnjxZKLtnZNWqVSuioqLYsmVLrvKejYHs2bNnuf/++y2uG1M3CntbtewHuBidPXu20MY39nnp0iW78puN9e801/zMJTd9z507l4sXLxIdHW1xEuGnn36a67ShL7/8kp9++omwsDCLHULOnDljNR2qNJFt3oQQQhS7Bx54gOHDh3Pq1Cn+85//FPd0cHV1xcXFpcD6O3v2LN988w3ly5dn6NChVh8tW7Y07ZFc2J588kn8/PyIj49n8+bNOdbNugoYGBgIYPUGxZSUFPbu3Yubmxv33Xdfgc43u59//pkrV65YlBvnZZxnQXr44YcB+P777+2qHxQUBGA11SE9Pd10o2Re5NT3pUuXrP7V48iRIwD07t3b4pq1fgBTPnzWlfT89FeaSIAshBCiRJg0aRJubm7MmTPHLPezMFy7do033njD5gpcZGQkV69epVGjRgWyehwVFcXt27cZOHAgixcvtvow7te7ZMmSHPf/LQheXl7Mnz8fgH79+rFhwwar9X788UceeeQR0/NBgwbh7OzMu+++awqQjCZPnszly5cZNGiQxQ16Be3SpUtMnz7drGz37t0sX74cb29vevbsWeBjjhkzBmdnZ1566SX++OMPi+upqalmwXOLFi1o2LAh27Zt48svvzSru2DBgjznHwM88cQT+Pr68sknn7B7926za1OnTrWaYmLcni37LzcbNmxg8eLFVscx/ts/ceKE3f0dO3bM6tZ8pY2kWAghhCgRqlevTnh4OPPmzWP27Nm8/fbbhTZWWloaU6ZMYdq0aTRv3pymTZvi6+tLUlISO3bs4Ndff8XDw4OFCxfmeyytNUuWLAFg2LBhNuvVr1+ftm3bEhcXx7p162we3lFQBg4cyI0bNxgzZgxdunShadOmtGjRwnTUdHx8PPv27aNixYqmNv7+/kRGRvLcc88RFBRE3759qVSpElu3biU+Pp6AgABmzZpVqPMGaNOmDYsXL2bnzp20bNnStA9yRkYGH374oV0pELkVEBBAVFQUQ4YM4f7776dLly40aNCAtLQ0Tpw4wffff0+lSpVMB3UYf9np2LEjvXv3Nu2DvG/fPjZv3kyXLl1Yv359nubi6enJokWL6NevH61btzbbB/nAgQO0adOGbdu2mbUZPXo00dHR9OnTh969e1O9enUOHDjA+vXr6du3L5999pnFOO3bt+fzzz+nV69ePProo7i7u1O7dm2efvppHnvsMerXr8/cuXP59ddfCQwM5MSJE3zzzTd069bNalBdqmitS+WjWbNmWgghSovff/+9uKdQIgDa8F+PdWfPntXlypXT5cqV02fPntVaa52QkKABXbt27QKbR3p6ul63bp1+6aWXdPPmzXXVqlW1k5OT9vT01E2aNNEvvviiTkhIsGjXtm1bDejY2FibfYeFhWlAR0dHa6213rhxowZ0YGDgHee1fPlyDejHH3/cVHan9yy/Tpw4ocePH68DAwO1t7e3dnJy0hUrVtQhISH6P//5j7506ZJFmw0bNuiOHTtqHx8f7eLiouvVq6fHjRunk5OTLeoa3zNrjO+Vtfc6IiLC4r02/lsICwvTv//+u3788ce1j4+Pdnd31y1atNDr16+36Cc6Otrs8zCqXbu2zX9T1sY22r9/vw4LC9O1atXSLi4u2tfXV99///16xIgResuWLRb1d+/erTt37qw9PT21p6enbt++vf7hhx9yHMNeGzdu1C1bttTu7u7ax8dHP/744/rgwYM239cdO3bo0NBQ7ePjoz09PXXLli31mjVrdGxsrAZ0RESEWf3bt2/r1157TdepU0c7OTlpQLdt29Z0/cSJE3rAgAG6WrVq2s3NTTdq1EjPmjVLp6WlWdQtavb+zAV2aytxptIFeFdsUQoODtbZ/6wghBAl1cGDBws9L1MIIYSBvT9zlVJ7tNYWexBKDrIQQgghhBBZSIAshBBCCCFEFnKTnhBCiDIhJibGrpP2mjZtmqejqUuylJQUIiMj7ao7ePBg0w4EouQwHiV+Jz169KBp06aFOhchAbIQQogyIiYmxq79V8PCwspkgGzvwQwhISESIJdA9n5+/v7+EiAXAQmQhRBClAnWDq+4W/j7+xfoUdSi6MnnV7JIDrIQQgghhBBZSIAshBBCCCFEFhIgCyGEEEIIkYUEyEIIIYQQQmQhAbIQQgghhBBZSIAshBBCCCFEFhIgCyGEEEIIkYUEyEIIIYQQQmQhAbIQQgghhBBZSIAshBBCCJEPiYmJKKUYPHhwcU9FFBAJkIUQQhQZpZTZw9HRET8/P0JCQoiJibF63K4x+Mjpkddjpg8fPszw4cOpX78+7u7ueHh4UKdOHTp16sT06dM5d+6cWf2QkJA7jjd48GCUUsTExNisc/LkSRwdHVFKMXHixBznaO09q1ixIu3atWP58uW5eblm/P39zfp1cHDAx8eHFi1a8N5773H79u08930ncXFxKKWYOnVqoY0hRH44FfcEhBBC3H0iIiIASEtL48iRI6xZs4atW7eye/duFixYYLWNt7c3Y8eOtXrN398/13P47rvv6NatGzdv3uSRRx6hS5culCtXjsTERPbu3cumTZto0aIFVapUyXXfd7J48WIyMjJQShEdHc306dNxcsr5v+Ss79nhw4dZu3YtsbGx7Nmzh7lz5+Z5Li+++CI+Pj6kp6eTkJDA6tWriY+PZ8uWLaxevTrP/d5NqlevzsGDB/H29i7uqYiCorUulY9mzZppIYQoLX7//ffinkKJAGjDfz3mtm/frh0cHLRSSh87dszsWkJCggZ07dq1C3Qu9evX14COiYmxen3fvn36xIkTZmVt27bVgI6NjbXZb1hYmAZ0dHS01eu3b9/WNWrU0OXLl9ejRo3SgF61apXN/my9Z5s3b9ZKKa2U0gkJCTbb21K7dm0NWLQ9cOCAdnd314COi4vLdb/2iI2N1YCOiIgolP6FsPdnLrBbW4kzJcVCCCHKiNk7ZhObEGtWFpsQy+wds4tpRvZr2bIlAQEBaK3Zs2dPoY93/vx5jhw5gre3N2FhYVbrPPDAA9SsWbPAx163bh2nTp2iX79+jB49GoCPPvoo1/20b9/e9J7t2rWrwOZ3//33ExISAsBPP/0EQExMjCltZP369YSEhODt7Y1SytTu0qVLvPbaazRs2BA3Nzd8fX3p3LkzmzdvNut/8ODBhIaGAjBt2rQcU2U+/fRTQkND8fX1xc3Njfvuu48333yTW7duWZ37li1b6NKlC35+fri5udGgQQMmTJjApUuXLOoa02XS0tKYPn069erVw83NjYCAALPPY+HChTRp0gR3d3dq1KhBREQEGRkZZn3llIN8/fp1Zs2aRXBwMF5eXnh6enLffffxwgsvWKTwiJJDUiyEEKKMeKjaQ/T9oi8rn1xJaJ1QYhNiTc9LA52Zf+zs7FzoY3l7e+Pk5MTVq1c5c+YMVatWLfQxjRYtWgQYAsXGjRsTFBTExo0bOX78OLVr185VX8b3LGugWhBs9fvFF1+wfv16unbtysiRI0lMTAQgJSWFli1b8vvvv/PQQw8xduxYLly4wMqVK+nUqRMffPAB4eHhAPTo0QOApUuX0rZtW1MwDuapMkOHDiUqKooaNWrQq1cvfHx8+PHHH5k8eTJbtmxh06ZNZmkpH374IaNGjcLDw4M+ffpQuXJl4uLimDVrFl9//TU7duzAx8fH4rX279+fnTt38uijj+Ls7MwXX3zBiBEjcHZ2Zv/+/SxdupTu3bvTvn17vvrqK6ZPn065cuX497//fcf3MTk5mdDQUPbt20fDhg0ZMmQILi4uHD16lKioKHr16lUoKTyiAFhbVi4ND0mxEEKUJkWVYvHdse90xdkV9eTvJuuKsyvq7459VyTj2gsb6QJbt27VDg4O2sXFRZ8+fdrsmjHFwtvbW0dERFg81qxZk6e59O7dWwO6bt26+p133tE//vijvnbtWo5tjCkWYWFhVucSERGhH3zwQZspFqdOndKOjo66QYMGprL58+drQE+aNMnqmLbes02bNplSLBITE3P34rV9KRbbtm3TWmsdHR2tAa2U0uvWrbPoa8SIERrQI0aM0BkZGabyP/74Q5cvX167uLiYjXOnFAvjeD179tTXr183uxYREaEBHRkZaSpLTEzULi4u2svLSx88eNCsvjGNZfjw4Wblxs8yODhYJycnm8qPHj2qnZ2dtY+Pj/b399enTp0yXUtOTtYVKlTQFStW1GlpaaZy47/RsLAwszGeeuopDeiRI0fq9PR0s2uXL1/WKSkpVl+/yL/8pljICrIQQpQAalrBrgC+se0NANr9t12B9akjLHeYyCvj7gVZb9LTWjNnzhybq7mXLl1i2rRpFuVhYWGmVcnc+Oijj9Bas2bNGsaNGweAg4MDjRs35rHHHuP555+3ubq3dOnSXI8HsGTJEtLT083+FD9gwABeffVVoqKimDp1Ko6OjlbbZn3PjDfpaa156aWXcr3ynFVkZKTFTXo3btygZ8+etG7d2qzuE088QZcuXczK0tLSWLZsGZ6enrz99ttmq8733nsvL7zwAm+++Sb//e9/mTJlil1zmjdvHk5OTkRFReHu7m52bfLkySxYsIDly5fz4osvArBs2TJSU1N55ZVXCAgIMKs/Y8YMli1bxscff8y7776Lq6ur2fWZM2earSzXrVuXVq1aERsby//93/9RvXp10zUfHx8ee+wxYmJi+Ouvv3J838+fP89nn31G1apVmTNnDg4O5lmtXl5edr0XonhIgCyEECVAQQWfxrSKUcGj+GD3B6Z0i5Ime6CrlGLJkiU8++yzNtvUrl3b9Cf9guDr68uqVatITExkw4YN7N69m127drF//37279/PBx98wPr163nooYcs2sbGxpqlBmQ1ePBgqwF0RkYGUVFRODg48Mwzz5jKK1SoQPfu3Vm9ejX/+9//ePzxx632a3zPlFL4+PjQunVrhg4dyqBBg/Lw6v8xb948U7+enp488MADDBo0iJEjR1rUbd68uUXZoUOHuH79Oi1btsTPz8/iert27XjzzTf55Zdf7JrP9evX2bdvHxUrViQyMtJqHVdXVw4ePGh6/vPPP5vGys7X15fAwEC2bdvGoUOHePDBB82uBwcHW7SpVq0aAM2aNbO4ZgyYT506lWOAvGvXLjIyMmjTpg0eHh4264mSSQJkIYQoI7LmHIfWCSXUP9TseUmiM3Ncr127Rnx8PEOHDmXkyJHUrl3bapBTmPz9/QkPDzflyJ46dYrRo0fz9ddfM3z4cPbu3Vsg42zYsIHjx4/TuXNns1VJgGeffZbVq1ezaNEimwGy8T0raAkJCXZvk3fPPfdYlBlvgLO18m8sT0lJsWuM5ORktNb8/fffVv9iYE1+5mBtazZjbnNO19LS0nKck3Gs7J+1KB1kFwshhCgjdp3eZRYMh9YJZeWTK9l1uuB2OChoHh4edOjQga+//pr09HTCwsK4fv16sc6pRo0arFixAhcXF/bt28fFixcLpF/jzXkbNmywOPzjscceA2D9+vWcPHmyQMYrDNZuBjQGkWfPnrXa5syZM2b17sRYLzAw8I73IxXWHAqCMW3jr7/+KrIxRcGRFWQhhCgjxrccb1EWWie0xK0eW/PAAw8wfPhwFi5cyH/+8x9ef/31Yp2Pq6srLi4upKamFkh/Z8+e5ZtvvqF8+fL06dPHap1Dhw6xY8cOoqKiTIeClAYNGzakXLly7N27l+TkZHx9fc2ux8Yath4MCgoylRnzrNPT0y368/T05P777+e3334jKSnJatpGdoGBgaxevZq4uDjat29vdi0lJYW9e/eatokrKs2bN8fBwYFt27Zx7do1SbMoZWQFWQghRIkwadIk3NzcmDNnDsnJyYU61rVr13jjjTds7kMbGRnJ1atXadSoERUqVMj3eFFRUdy+fZuBAweyePFiqw/jXsNLliyx2Ge3JHNxcWHgwIFcvXrV4ia8o0ePMn/+fJydnXn66adN5cb39MSJE1b7fPnll0lNTWXIkCFW0yKSk5NNeccAgwYNwtnZmXfffZcjR46Y1Z08eTKXL19m0KBBFjfoFaZKlSrRv39/zpw5w6uvvmrxmV69etXq/syiZJAVZCGEECVC9erVCQ8PZ968ecyePZu333670MZKS0tjypQpTJs2jebNm9O0aVN8fX1JSkpix44d/Prrr3h4eLBw4cJ8j6W1ZsmSJQAMGzbMZr369evTtm1b4uLiWLduHd26dcv32EVl5syZfP/99yxYsIBdu3YRGhpq2gf5ypUrLFiwgDp16pjqN2zYkOrVq5tSWWrVqoVSiqeffpratWszZMgQ9uzZw/vvv0+9evXo3LkztWrVIikpiYSEBLZt28azzz5r+nz8/f2JjIzkueeeIygoiL59+1KpUiW2bt1KfHw8AQEBzJo1q8jflwULFnDgwAEWLlxIXFwcnTt3xsXFhYSEBDZs2MBXX31l82ZPUbwkQBZCCFFivPbaa3z00UfMnz+fsWPHFtohCuXLl2fdunVs3LiRHTt2sHbtWv7++2/c3NyoU6cOL774ImPHjrX75rWcbN68mWPHjhEYGGiWZmDN8OHDiYuLY9GiRaUqQPbz8yM+Pp63336b1atXM3fuXNzd3WnevDnjxo2jU6dOZvUdHR1Zs2YNEyZMMAXRWmtatWpl2hnivffeo2vXrixcuJDNmzeTkpKCn58ftWrVYty4cRa7d4wePZr69eszZ84cVq1axfXr16lZsybjxo1j4sSJVg8JKWy+vr788MMPREZG8tlnn7Fo0SIcHR2pWbMmQ4YMoVGjRkU+J2EfVVh3xRa24OBgvXv37uKehhBC2OXgwYNFmv8ohBB3M3t/5iql9mitLfb6kxxkIYQQQgghspAAWQghhBBCiCwkB1kIIUSZEBMTY9dJe02bNs3T0dQlWUpKis1T57IbPHhwgeRWC1GWFXmArJR6CRgGaOBX4FmgHPAZ4A8kAn211oW7x48QQogyJSYmhq1bt96xXlhYWJkMkO09dS4kJEQCZCHuoEgDZKVUdeAFoJHW+oZSaiXQH2gEbNFaz1RKTQAmAP8uyrkJIYQo3eLi4op7CsXG39+/0I6iFuJuVBw5yE6Au1LKCcPK8WngCWBp5vWlQI9imJcQQgghhBBFGyBrrf8C5gAngDPAJa31RqCK1vpMZp0zQOWinJcQQgghhBBGRRogK6V8MawW1wGqAR5KqUE5tzJrP0IptVsptfvvv/8urGkKIYQQQoi7WFGnWHQAErTWf2ut04DVQAvgnFKqKkDm1/PWGmutF2mtg7XWwZUqVSqySQshhBBCiLtHUQfIJ4CHlVLllFIKaA8cBL4CwjLrhAFfFvG8hBBCCCGEAIp4Fwut9U6l1BfAz8Bt4BdgEeAJrFRKDcUQRPcpynkJIYQQQghhVOT7IGutI4CIbMW3MKwmCyGEEEIIUazkqGkhhBBCCCGykABZCCGEEEKILCRAFkIIIcRdKTExEaUUgwcPtrtNTEwMSiliYmIKbV55MXXqVJRShXKi5ODBg1FKkZiYWOB9l1QSIAshhCgySimzh6urK5UqVSIoKIhhw4axbt060tPTrbY1/ietlOLbb7+1WscYJCxevLjA2ubG4cOHGT58OPXr18fd3R0PDw/q1KlDp06dmD59OufOnTOrHxIScsegxjj3nAKykydP4ujoiFKKiRMn5jjH7J+Bo6MjFStWpF27dixfvjw3L9fmXCZMmECzZs3w9fXF2dmZypUr06FDB+bNm8elS5fyPYYonYz/3kuDIr9JTwghhIiIMNyrnZ6eTkpKCr/99hsff/wxS5YsITg4mOXLl9OgQQOb7cePH0/nzp1xdHTM9dj5aZuT7777jm7dunHz5k0eeeQRunTpQrly5UhMTGTv3r1s2rSJFi1aUKVKlQIdF2Dx4sVkZGSglCI6Oprp06fj5JTzf/HGzyAtLY3Dhw+zdu1aYmNj2bNnD3Pnzs3zPMaMGcOtW7d48MEHeeqpp/D19eXixYts376dsWPH8sYbb3DhwoU89V8S9OzZk4cffpiqVasW91REIZIAWQghRJGbOnWqRdm5c+d4/vnn+fzzz+nQoQO7d++mcuXKFvXq16/Pb7/9RlRUFMOHD8/VuPlpeyfh4eHcvHmTmJgYwsLCLK7v378fX1/fAh0TDL9kREVFUb58eQYOHMgHH3zAV199Ra9evXJsl/0z2LJlCx07diQyMpIXXngBf3//XM3jk08+Yfjw4fj6+rJq1Sq6detmUWfHjh0899xzueq3pPH29sbb27u4pyEKmaRYCCGEKBGqVKnCihUrCAkJ4eTJk7z11ltW602ePJly5coxZcoUrl27lqsx8tM2J+fPn+fIkSN4e3tbDY4BHnjgAWrWrFlgYxqtW7eOU6dO0a9fP0aPHg3ARx99lOt+2rdvT0BAAFprdu3alau2V65c4fnnnwdgxYoVVoNjgJYtW7Jz506L8i1bttClSxf8/Pxwc3OjQYMGTJgwwWo6hvHP9GlpaUyfPp169erh5uZGQECA2eteuHAhTZo0wd3dnRo1ahAREUFGRobN13Do0CF69OiBn58fHh4etGrVio0bN1rUs5WD7O/vj7+/P9evX2fcuHHUqlULV1dX6tevz6xZs9BaWx13586dPPnkk9xzzz24uLhQs2ZNwsPDOX36tNX6e/bsoUuXLnh5eVG+fHk6dOhAfHy8zdeVG5s3b6Z169Z4eHjg5+dHjx49OHTokM36MTEx9O7dm7p16+Lu7k758uVp2bIly5YtM6tnzPXeunUrYJ7mExISYqoXGxvLiBEjaNSoEeXLl8fd3Z3GjRszbdo0bt68WSCv0V4SIAshhCgxHBwcmDRpEgCffvqp1aCiWrVqvPLKK5w9e5bZs2fnqv/8tM2Jt7c3Tk5OXL16lTNnzhRYv/ZYtGgRYMhVbty4MUFBQWzcuJHjx4/nui/j+53bPNEvvviCpKQkHn74YTp16pRjXVdXV7PnH374IR07dmTHjh306NGDsWPH4ufnx6xZs2jRogUpKSlW++nfvz+LFi2iffv2DB06lOTkZEaMGEFMTAwvv/wyr7/+OkFBQYSHh+Pi4sL06dN55513rPaVkJDAI488wsWLFwkPD6dPnz7s2bOHrl278tlnn9n9PqSlpdGpUydWrVpF165dGTZsGDdu3GDChAlMnz7don50dDQtW7Zk3bp1hIaGMnbsWIKDg1m8eDHBwcGcOHHCrP4PP/xA69at2bx5M127dmXMmDG4uLgQEhJi9ReP3Pjiiy/o3Lkzu3fvpk+fPoSHh3Px4kUeeeQREhISrLYZNWoUiYmJtGnThrFjx9K/f3+OHz/O008/zeTJk031fHx8iIiIoHbt2oAhvcf4yHqD5KxZs9i4cSNNmzYlPDycYcOG4eLiwtSpU+natavN+xMKhda6VD6aNWumhRCitPj9999tXoPS88gvQHOHjm7evKmdnJw0oI8dO2YqDwsL04DetGmTvnLliq5SpYr28PDQp0+fNtWJiIjQgP7oo4/M+sxPW3v17t1bA7pu3br6nXfe0T/++KO+du1ajm3atm2rAR0WFqYjIiKsPh588EEN6OjoaIv2p06d0o6OjrpBgwamsvnz52tAT5o0yeqYtj6DTZs2aaWUVkrpxMTEXL32IUOGaEC//vrruWqXmJioXVxctJeXlz548KDZtVGjRmlADx8+3Kzc+J4FBwfr5ORkU/nRo0e1s7Oz9vHx0f7+/vrUqVOma8nJybpChQq6YsWKOi0tzVSekJBgej9effVVs3F27dqlnZyctI+Pj7506ZKpPDo62urnUbt2bQ3orl276uvXr5vKz507p729vbW3t7dOTU01lR8+fFg7OzvrevXqmc1Va623bNmiHRwcdI8ePUxlGRkZumHDhhrQa9euNasfGRlpeh2xsbE6t65cuaL9/Py0k5OT3rVrl9m1sWPHmvpOSEgwu3bkyBGLvm7duqXbtWunnZycLF6X8bOz5ejRozojI8OifNKkSRrQK1assPs15fQzNytgt7YSZxZ7oJvXhwTIQojSRAJk42u9c4CstdZVqlTRgN65c6epLGuQq7XWCxcu1IAeOnSoqY49AXJu29orKSlJ9+rVSyulTK/TwcFBP/DAA/r111/XZ8+etWhjDBjseVgLkKdNm6YB/dZbb5nKLly4oF1cXHS1atX07du3LdoY+zMG4BMnTtS9e/fWjo6OGtAvvfRSrl97165dNaA/+OCDXLV78803NaBfe+01i2tJSUnay8tLu7m56Zs3b5rKje/Z5s2bLdqEhoZqQC9ZssTi2uDBgzVgFvwbA2Rvb299+fJlizbGfzcxMTGmsjsFyH/++adFP88884wG9K+//moqMwae33zzjUV9rbXu0aOHdnR0NM1r+/btGtBt2rSxqHv79m1dr169PAfIy5Yt04B+5plnLK6lpKRob29vqwGyLatWrdKAXrp0qVn5nQJkWy5cuKAB/eyzz9rdJr8BsqRYCCGEKHEM/2/l/Kf+YcOG0ahRI6Kjo/n1119z1X9+2tpivDnt2LFjLFy4kGHDhtGkSRN+/fVXZsyYQaNGjWzm9sbGxtpcELKV05yRkUFUVBQODg4888wzpvIKFSrQvXt3Tp8+zf/+9z+b8502bRrTpk3j7bff5rvvvqN169Z8/PHHedrBwp7Py5qff/4ZgHbt2llc8/X1JTAwkJs3b1rNgw0ODrYoq1atGgDNmjWzuFa9enUATp06ZXEtKCgILy8vi3Jjfuwvv/ySw6v4h7e3N/Xr17coN+aeJycnm8qMecNbt25l6tSpFo/z58+Tnp7OH3/8AfzzXrVt29aif0dHR1q1amXXHK3JqW9vb2+aNm1qtd2JEyd47rnnCAgIoFy5cqa84t69ewPw119/5Woe165d46233uKhhx7C29sbBwcHlFJUrFgxT/3lh+xiIYQQxSwzthCZbt68SVJSEgCVKlWyWc/R0ZHZs2fTvXt3xo8fz7p16+weIz9t78Tf35/w8HDCw8MBQ0A2evRovv76a4YPH87evXsLZJwNGzZw/PhxOnfubAr+jJ599llWr17NokWLePzxx6221wX4D88YmFoLPnNivAnP1pZpxnJrecjWdpIwbm2X07W0tDSLa7a23rvnnnvM5nknPj4+VsuNY2fNob148SKAzbxoo6tXr5rN4U5zzYu89H3s2DGaN29OcnIyrVu3plOnTnh7e+Po6EhiYiJLly7l1q1bds8hLS2Ndu3a8dNPP9G4cWP69etHpUqVcHZ2Bgy/0OWmv/ySAFkIIUSJsn37dm7fvk2VKlXuuNVYt27daNeuHevXr2fTpk25Gic/bXOjRo0arFixAl9fX/bt28fFixepUKFCvvs13py3YcMGmyu369ev5+TJk4Wye0ZWrVq1Iioqii1btvDGG2/Y3c4YyJ49e5b777/f4rrxhsfC3lYt+wEuRmfPni208Y19Xrp0ifLly9td/05zzc9cctP33LlzuXjxItHR0RYnEX766acsXbo0V3P48ssv+emnnwgLC7PYIeTMmTNMmzYtV/3ll6RYCCGEKDEyMjKYMWMGAAMGDLCrzZw5c1BKMW7cuBy38Srotrnh6uqKi4tLgfV39uxZvvnmG8qXL8/QoUOtPlq2bGnaI7mwPfnkk/j5+REfH8/mzZtzrJt1FTAwMBDA6kmCKSkp7N27Fzc3N+67774CnW92P//8M1euXLEoN87LOM+C9PDDDwPw/fff21U/KCgIwLRVWlbp6els3749z3PJqe9Lly5Z/avHkSNHAEzpFFlZ6wcwHc5jbTeKvPRXmCRAFkIIUSKcP3+e/v37ExcXR61ate54ZLJRYGAggwYNYt++fXz66ae5GjM/bbO6du0ab7zxhs0VuMjISK5evUqjRo0KZPU4KiqK27dvM3DgQBYvXmz1Ydyvd8mSJYUa/AN4eXkxf/58APr168eGDRus1vvxxx955JFHTM8HDRqEs7Mz7777rilAMpo8eTKXL19m0KBBFlvDFbRLly5ZbMO2e/duli9fjre3Nz179izwMceMGYOzszMvvfSSKc84q9TUVLPguUWLFjRs2JBt27bx5ZdfmtVdsGABR48ezfNcnnjiCXx9ffnkk0/YvXu32bWpU6daTTEx/nUn+y83GzZssHlcu/Hffvbt63Lq79ixY/z73/+241UULEmxEEIIUeSMp7hlZGSYjprevn07qampNG/enOXLl5tuzLHHjBkz+Pzzzy2CrMJua5SWlsaUKVOYNm0azZs3p2nTpvj6+pKUlMSOHTv49ddf8fDwYOHChXkew0hrzZIlSwDDzYa21K9fn7Zt2xIXF8e6detsHt5RUAYOHMiNGzcYM2YMXbp0oWnTprRo0cJ01HR8fDz79u0z+1z9/f2JjIzkueeeIygoiL59+1KpUiW2bt1KfHw8AQEBzJo1q1DnDdCmTRsWL17Mzp07admyJWfOnOGzzz4jIyODDz/80K4UiNwKCAggKiqKIUOGcP/999OlSxcaNGhAWloaJ06c4Pvvv6dSpUqmGxSNv+x07NiR3r1706tXL+rXr8++ffvYvHkzXbp0Yf369Xmai6enJ4sWLaJfv360bt2afv36UbVqVbZv386BAwdo06YN27ZtM2szevRooqOj6dOnD71796Z69eocOHCA9evX07dvX6v7R7dv357PP/+cXr168eijj+Lu7k7t2rV5+umneeyxx6hfvz5z587l119/JTAwkBMnTvDNN9/QrVs3q0F1obJ112xJf8g2b0KI0sTeLYfKOrJtXebi4qIrVKigg4KC9LBhw/S6det0enq61bbZt2rLbsKECaZ+77TNW27a2iM9PV2vW7dOv/TSS7p58+a6atWq2snJSXt6euomTZroF1980eoWWcZtr3Lamss4d+O2Yhs3btSADgwMvOO8li9frgH9+OOPm8qMr7OwnDhxQo8fP14HBgZqb29v7eTkpCtWrKhDQkL0f/7zH7M9hY02bNigO3bsqH18fLSLi4uuV6+eHjdunNk+x0Y5bRVmfK+svdfGbfyyvtfGbd7CwsL077//rh9//HHt4+Oj3d3ddYsWLfT69est+slpm7fatWtbnZe1sY3279+vw8LCdK1atbSLi4v29fXV999/vx4xYoTesmWLRf3du3frzp07a09PT+3p6anbt2+vf/jhhxzHsNfGjRt1y5Yttbu7u/bx8dGPP/64PnjwoM33dceOHTo0NFT7+PhoT09P3bJlS71mzRodGxtr2kowq9u3b+vXXntN16lTx7TXedu2bU3XT5w4oQcMGKCrVaum3dzcdKNGjfSsWbN0WlqaRd07ye82b0qX0tung4ODdfY/AwghREl18ODBQs+jFEIIYWDvz1yl1B6ttcWegZKDLIQQQgghRBYSIAshhBBCCJGF3KQnhBBCWBETE0NiYuId6zVt2pQePXoU+nyKUkpKCpGRkXbVHTx48B33qxZFz3gj7J306NHD5kl5dzMJkIUQQggrYmJi7Np/NSwsrEwGyPYezBASEiIBcglk7+fn7+8vAbIVEiALIYQQVlg7vOJu4e/vX6BHUYuiJ59f/kgOshBCCCGEEFlIgCyEEEIIIUQWEiALIYQQQgiRhQTIQgghhBBCZCEBshBCCCGEEFlIgCyEEEIIIUQWEiALIYQQQgiRhQTIQgghhBBCZCEBshBCCCGEEFlIgCyEEEIIUYIkJiailGLw4MHFPZW7lgTIQgghioxSyuzh6upKpUqVCAoKYtiwYaxbt4709HSrbQcPHmxq9+2331qtM3XqVJRSLF68uMDa5sbhw4cZPnw49evXx93dHQ8PD+rUqUOnTp2YPn06586dM6sfEhKCUirHY62Nc4+JibFZ5+TJkzg6OqKUYuLEiTnOMftn4OjoSMWKFWnXrh3Lly/Pzcs14+/vb9avg4MDPj4+tGjRgvfee4/bt2/nue87iYuLQynF1KlTC20McXdxKu4JCCGEuPtEREQAkJ6eTkpKCr/99hsff/wxS5YsITg4mOXLl9OgQQOb7cePH0/nzp1xdHTM9dj5aZuT7777jm7dunHz5k0eeeQRunTpQrly5UhMTGTv3r1s2rSJFi1aUKVKlQIdF2Dx4sVkZGSglCI6Oprp06fj5JTzf/HGzyAtLY3Dhw+zdu1aYmNj2bNnD3Pnzs3zXF588UV8fHxIT08nISGB1atXEx8fz5YtW1i9enWe+72bVK9enYMHD+Lt7V3cU7lrSYAshBCiyFlb6Tt37hzPP/88n3/+OR06dGD37t1UrlzZol79+vX57bffiIqKYvjw4bkaNz9t7yQ8PJybN28SExNDWFiYxfX9+/fj6+tboGOC4ZeMqKgoypcvz8CBA/nggw/46quv6NWrV47tsn8GW7ZsoWPHjkRGRvLCCy/g7++fp/mMHTvWrO1vv/3GQw89xJo1a9i6dStt27bNU793E2dnZwICAop7Gnc1SbEQQogyYvZsiI01L4uNNZSXBlWqVGHFihWEhIRw8uRJ3nrrLav1Jk+eTLly5ZgyZQrXrl3L1Rj5aZuT8+fPc+TIEby9va0GxwAPPPAANWvWLLAxjdatW8epU6fo168fo0ePBuCjjz7KdT/t27cnICAArTW7du0qsPndf//9hISEAPDTTz8BEBMTY0obWb9+PSEhIXh7e6OUMrW7dOkSr732Gg0bNsTNzQ1fX186d+7M5s2bzfofPHgwoaGhAEybNs0szSN76sqnn35KaGgovr6+uLm5cd999/Hmm29y69Ytq3PfsmULXbp0wc/PDzc3Nxo0aMCECRO4dOmSRV1jukxaWhrTp0+nXr16uLm5ERAQYPZ5LFy4kCZNmuDu7k6NGjWIiIggIyPDrK+ccpCvX7/OrFmzCA4OxsvLC09PT+677z5eeOEFixQekXcSIAshRBnx0EPQt+8/QXJsrOH5Qw8V77xyw8HBgUmTJgGGYEZrbVGnWrVqvPLKK5w9e5bZuYz+89M2J97e3jg5OXH16lXOnDlTYP3aY9GiRYAhUGzcuDFBQUFs3LiR48eP57ov4/udNVAtCLb6/eKLL+jevTteXl6MHDmSvn37ApCSkkKLFi2YOXMm3t7ejB07lt69exMfH0+nTp348MMPTX306NHD9EtJ27ZtiYiIMD2yrmQPHTqUAQMGcOTIEXr16sVzzz2Hn58fkydPpkuXLhY50h9++CEdO3Zkx44d9OjRg7Fjx+Ln58esWbNo0aIFKSkpVl9r//79WbRoEe3bt2fo0KEkJyczYsQIYmJiePnll3n99dcJCgoiPDwcFxcXpk+fzjvvvGPX+5icnEyLFi2YMGECV69eZciQIYwaNYr77ruPqKgoDh48aFc/wg5a61L5aNasmRZCiNLi999/L5JxvvtO64oVtZ482fD1u++KZFi7AdrwX49tN2/e1E5OThrQx44dM5WHhYVpQG/atElfuXJFV6lSRXt4eOjTp0+b6kRERGhAf/TRR2Z95qetvXr37q0BXbduXf3OO+/oH3/8UV+7di3HNm3bttWADgsL0xEREVYfDz74oAZ0dHS0RftTp05pR0dH3aBBA1PZ/PnzNaAnTZpkdUxbn8GmTZu0UkorpXRiYmLuXrzWunbt2hrQCQkJZuUHDhzQ7u7uGtDbtm3TWmsdHR2tAa2U0uvWrbPoa8SIERrQI0aM0BkZGabyP/74Q5cvX167uLiYjRMbG6sBHRERYXVuxvF69uypr1+/bnbN+LlHRkaayhITE7WLi4v28vLSBw8eNKs/atQoDejhw4eblRs/y+DgYJ2cnGwqP3r0qHZ2dtY+Pj7a399fnzp1ynQtOTlZV6hQQVesWFGnpaWZyhMSEkz/LrJ66qmnNKBHjhyp09PTza5dvnxZp6SkWH39dyN7f+YCu7WVOLPYA928PiRAFkKUJnf6YQ0l/1EQ7AmQtda6SpUqGtA7d+40lWUNcrXWeuHChRrQQ4cONdWxJ0DObVt7JSUl6V69emmllOl1Ojg46AceeEC//vrr+uzZsxZtjEGVPQ9rAfK0adM0oN966y1T2YULF7SLi4uuVq2avn37tkUbY3/GAHzixIm6d+/e2tHRUQP6pZdeytPrNwbIL774oo6IiNCTJk3SAwcONAXHPXv2NNU1Bqw9evSw6Cc1NVWXK1dOe3p66osXL1pcnzRpkgb0tGnTTGV3CpCbNm2qnZyczAJXo9u3b+sKFSrohx56yFT25ptvakC/9tprFvWTkpK0l5eXdnNz0zdv3jSVGz/LzZs3W7QJDQ3VgF6yZInFtcGDB2vA7JcSawHyuXPntIODg65ataq+evWq1dcp/pHfAFlu0hNCiBJAW2YS5IkxrWLUKPjgA1i5EjLTM0sVre/8p/5hw4Yxf/58oqOjefHFF2nSpInd/eenrS2+vr6sWrWKxMRENmzYwO7du9m1axf79+9n//79fPDBB6xfv56HrOS8xMbGmvJ0sxs8eDBLly61KM/IyCAqKgoHBweeeeYZU3mFChXo3r07q1ev5n//+x+PP/641X6nTZsGGN5jHx8fWrduzdChQxk0aFAeXv0/5s2bZ+rX09OTBx54gEGDBjFy5EiLus2bN7coO3ToENevX6dly5b4+flZXG/Xrh1vvvkmv/zyi13zuX79Ovv27aNixYpERkZarePq6mqWnvDzzz+bxsrO19eXwMBAtm3bxqFDh3jwwQfNrgcHB1u0qVatGgDNmjWzuFa9enUATp06Re3atW2+jl27dpGRkUGbNm3w8PCwWU8UDAmQhRCijDAGx8agODTU/HlpcfPmTZKSkgCoVKmSzXqOjo7Mnj2b7t27M378eNatW2f3GPlpeyf+/v6Eh4cTHh4OGAKf0aNH8/XXXzN8+HD27t1bIONs2LCB48eP07lzZ1OQZfTss8+yevVqFi1aZDNA1gX1W1k2CQkJdu+Acc8991iUGW+Aq1q1qtU2xnJbOcDZJScno7Xm77//Nv1ScCf5mYO1rdmMW+7ldC0tLS3HORnHyv5Zi8IhN+kJIUQZsWuXeTAcGmp4XoAbEhSJ7du3c/v2bapUqXLHQKtbt260a9eO9evXs2nTplyNk5+2uVGjRg1WrFiBi4sL+/bt4+LFiwXSr/HmvA0bNlgc/vHYY48BsH79ek6ePFkg4xUGa38hMAaRZ8+etdrGeBOkvXsEG+sFBgbeMX2zsOZQEHx8fAD466+/imzMu5kEyEIIUUaMH2+5UhwaaigvLTIyMpgxYwYAAwYMsKvNnDlzUEoxbtw4i+2yCrNtbri6uuLi4lJg/Z09e5ZvvvmG8uXLM3ToUKuPli1bmvZILk0aNmxIuXLl2Lt3L8nJyRbXYzO3aQkKCjKVGQ99sXYKo6enJ/fffz+//fab6S8TdxIYGAhg9YTDlJQU9u7da9omrqg0b94cBwcHtm3bVqBbFArrJEAWQghRIpw/f57+/fsTFxdHrVq17nhkslFgYCCDBg1i3759fPrpp7kaMz9ts7p27RpvvPGGzX1oIyMjuXr1Ko0aNaJChQp5HscoKiqK27dvM3DgQBYvXmz1YdxreMmSJYUa/Bc0FxcXBg4cyNWrV5kyZYrZtaNHjzJ//nycnZ15+umnTeXG9/TEiRNW+3z55ZdJTU1lyJAhVtMikpOTTXnHAIMGDcLZ2Zl3332XI0eOmNWdPHkyly9fZtCgQbi6uub1ZeZapUqV6N+/P2fOnOHVV1+1+EyvXr1qdX9mkTeSgyyEEKLIGU9xy8jIMB01vX37dlJTU2nevDnLly+nYsWKdvc3Y8YMPv/8c4tgprDbGqWlpTFlyhSmTZtG8+bNadq0Kb6+viQlJbFjxw5+/fVXPDw8WLhwYZ7HMNJas2TJEsBws6Et9evXp23btsTFxbFu3Tq6deuW77GLysyZM/n+++9ZsGABu3btIjQ0lAsXLrBy5UquXLnCggULqFOnjql+w4YNqV69uimVpVatWiilePrpp6lduzZDhgxhz549vP/++9SrV4/OnTtTq1YtkpKSSEhIYNu2bTz77LOmz8ff35/IyEiee+45goKC6Nu3L5UqVWLr1q3Ex8cTEBDArFmzivx9WbBgAQcOHGDhwoXExcXRuXNnXFxcSEhIYMOGDXz11Vc2b/YUuSMBshBCiCJnvFnKxcUFLy8vateuzTPPPEPv3r3p1KkTDg65+wNnzZo1GTt2LDNnzsz1XPLT1qh8+fKsW7eOjRs3smPHDtauXcvff/+Nm5sbderU4cUXX7Q4gjmvNm/ezLFjxwgMDDRLM7Bm+PDhxMXFsWjRolIVIPv5+REfH8/bb7/N6tWrmTt3Lu7u7jRv3pxx48bRqVMns/qOjo6sWbOGCRMmmIJorTWtWrUy7Qzx3nvv0bVrVxYuXMjmzZtJSUnBz8+PWrVqMW7cOIvdO0aPHk39+vWZM2cOq1at4vr169SsWZNx48YxceJEU05wUfL19eWHH34gMjKSzz77jEWLFuHo6EjNmjUZMmQIjRo1KvI5lVWqsO5iLWzBwcF69+7dxT0NIYSwy8GDB4s0X1EIIe5m9v7MVUrt0Vpb7M0nOchCCCGEEEJkIQGyEEIIIYQQWUgOshBCCGFFTEwMiYmJd6zXtGlTevToUejzKUopKSk2T53LbvDgwQWSWy1ESSIBshBCCGFFTEwMW7duvWO9sLCwMhkg23vqXEhIiATIosyRAFkIIYSwwtohEXcLf3//QjuKWojSQHKQhRBCCCGEyEICZDvN3zmf0KWhtIluw+qDq4t7OkIIIYQQopBIioWd/rz4J3GJcQD0adSneCcjhBBCCCEKjawg28nJ4Z/fJW5n3C7GmQghhBBCiMIkAbKdJEAWQgghhLg7SIBsJwmQhRBCCCHuDhIg28nRwdH0vQTIQgghhBBllwTIdpIVZCGEEEKIu4MEyHbKGiCn6/RinIkQQgghhChMEiDbSVaQhRBCiIKVmJiIUorBgwfb3SYmJgalFDExMYU2r7yYOnUqSqlCOYFx8ODBKKVITEws8L6FdUUaICulGiql9mZ5XFZKjVVK+SmlNiml/sz86luU87KHBMhCCJF/Simzh6urK5UqVSIoKIhhw4axbt060tOt/5XOGCQopfj222+t1jEGKYsXLy6wtvYytp86darV62lpafz3v/+lR48e1KxZE3d3d8qVK0edOnXo2bMnixcv5tq1a3kaO6uTJ08yYcIEmjVrhq+vL87OzlSuXJkOHTowb948Ll26lO8xROkUEhKCUqq4p1EqFOlBIVrrw0BTAKWUI/AXsAaYAGzRWs9USk3IfP7vopzbnUiALIQQBSciIgKA9PR0UlJS+O233/j4449ZsmQJwcHBLF++nAYNGthsP378eDp37oyjo6PNOoXRNq8OHTpE7969+f333/Hx8aFdu3bUqVMHJycn/vrrL7Zt28batWt57bXX+Pvvv/M8zuLFixkzZgy3bt3iwQcf5KmnnsLX15eLFy+yfft2xo4dyxtvvMGFCxcK8NUVrZ49e/Lwww9TtWrV4p6KKMOK8yS99sBRrfVxpdQTQEhm+VIgjhIWIDsq2cVCCCEKirVV1nPnzvH888/z+eef06FDB3bv3k3lypUt6tWvX5/ffvuNqKgohg8fnqtx89M2r86cOUP79u05ffo0zz//PG+99Raenp4W9TZt2sS4cePyPM4nn3zC8OHD8fX1ZdWqVXTr1s2izo4dO3juuefyPEZJ4O3tjbe3d3FPQ5RxxZmD3B/4NPP7KlrrMwCZXy1/IgJKqRFKqd1Kqd35+Q07L2QFWQghCleVKlVYsWIFISEhnDx5krfeestqvcmTJ1OuXDmmTJmS65SE/LTNq4kTJ3L69GkGDBjA/PnzrQbHAB07dmT37t15GuPKlSs8//zzAKxYscJqcAzQsmVLdu7caVG+ZcsWunTpgp+fH25ubjRo0IAJEyZYTccw/pk+LS2N6dOnU69ePdzc3AgICOCjjz4y1Vu4cCFNmjTB3d2dGjVqEBERQUZGhs3XcOjQIXr06IGfnx8eHh60atWKjRs3WtSzlYPs7++Pv78/169fZ9y4cdSqVQtXV1fq16/PrFmz0FpbHXfnzp08+eST3HPPPbi4uFCzZk3Cw8M5ffq01fp79uyhS5cueHl5Ub58eTp06EB8fLzN15UbmzdvpnXr1nh4eODn50ePHj04dOiQzfoxMTH07t2bunXr4u7uTvny5WnZsiXLli0zq2fM9d66dStgnuoUEhJiqhcbG8uIESNo1KgR5cuXx93dncaNGzNt2jRu3rxZIK+xtCiWAFkp5QI8Dnyem3Za60Va62CtdXClSpUKZ3I2mO1ikSG7WAghRGFwcHBg0qRJAHz66adWg5pq1arxyiuvcPbsWWbPnp2r/vPTNi+uX7/Op58a1oKMaSU5cXLK2x92v/jiC5KSknj44Yfp1KlTjnVdXV3Nnn/44Yd07NiRHTt20KNHD8aOHYufnx+zZs2iRYsWpKSkWO2nf//+LFq0iPbt2zN06FCSk5MZMWIEMTExvPzyy7z++usEBQURHh6Oi4sL06dP55133rHaV0JCAo888ggXL14kPDycPn36sGfPHrp27cpnn31m9/uQlpZGp06dWLVqFV27dmXYsGHcuHGDCRMmMH36dIv60dHRtGzZknXr1hEaGsrYsWMJDg5m8eLFBAcHc+LECbP6P/zwA61bt2bz5s107dqVMWPG4OLiQkhIiNVfPHLjiy++oHPnzuzevZs+ffoQHh7OxYsXeeSRR0hISLDaZtSoUSQmJtKmTRvGjh1L//79OX78OE8//TSTJ0821fPx8SEiIoLatWsDhn+LxkfWGyRnzZrFxo0badq0KeHh4QwbNgwXFxemTp1K165dbd4fUCZprYv8ATwBbMzy/DBQNfP7qsDhO/XRrFkzXZSifo7STEUzFT147eAiHVsIUfr9/vvvti9C6XnkE6C5Qz83b97UTk5OGtDHjh0zlYeFhWlAb9q0SV+5ckVXqVJFe3h46NOnT5vqREREaEB/9NFHZn3mp629jO0jIiJMZVu3btWArlGjRp76tNeQIUM0oF9//fVctUtMTNQuLi7ay8tLHzx40OzaqFGjNKCHDx9uVt62bVsN6ODgYJ2cnGwqP3r0qHZ2dtY+Pj7a399fnzp1ynQtOTlZV6hQQVesWFGnpaWZyhMSEkz/Jl599VWzcXbt2qWdnJy0j4+PvnTpkqk8OjpaAzo6Otqsfu3atTWgu3btqq9fv24qP3funPb29tbe3t46NTXVVH748GHt7Oys69WrZzZXrbXesmWLdnBw0D169DCVZWRk6IYNG2pAr1271qx+ZGSk6XXExsbq3Lpy5Yr28/PTTk5OeteuXWbXxo4da+o7ISHB7NqRI0cs+rp165Zu166ddnJysnhdxs/OlqNHj+qMjAyL8kmTJmlAr1ixIhevqnjl+DM3C2C3thJnFleKxVP8k14B8BUQlvl9GPBlkc/oDiTFQgghioarqysVKlQAsHnDmqenJ9OmTePatWtmK2X2yE/b3Dp79iwA1atXt3o9JiaGqVOnmj327t2b63HOnDkDQI0aNXLVbtmyZaSmpjJmzBgCAgLMrs2YMQMvLy8+/vhjbt26ZdF25syZ+Pj4mJ7XrVuXVq1akZKSwuTJk81es4+PD4899hgXLlzgr7/+sujL29ubKVOmmJUFBwczcOBAUlJSWLNmjd2vaf78+bi7u5ueV65cmSeeeIJLly5x+PBhU/kHH3xAWloa8+bNs/h82rVrx+OPP87XX3/NlStXAMPq8eHDh2nTpg1PPPGEWf0xY8ZQr149u+eY3ZdffklSUhIDBgwgODjY7NrUqVNt5lxbG9PFxYXnnnuO27dvs2XLllzNo27dulZ3uRg7diwAGzZsyFV/pVmR36SnlCoHdATCsxTPBFYqpYYCJ4A+RT2vO5GjpoUQoujozNSKnLakGjZsGPPnzyc6OpoXX3yRJk2a2N1/ftrmxp1eR0xMjCkv1Mjf35+mTZsW6Di2/Pzzz4AhIMzO19eXwMBAtm3bxqFDh3jwwQfNrmcP5MCQwgLQrFkzi2vGIPTUqVOmP/UbBQUF4eXlZdEmJCSEpUuX8ssvvxAWFmZxPTtvb2/q169vUV6zZk0AkpOTTWXGvOGtW7eya9cuizbnz58nPT2dP/74g2bNmpneq7Zt21rUdXR0pFWrVhw9evSOc7Qmp769vb1p2rSpxb8TgBMnTjBr1iy2bNnCiRMnuHHjhtl1a7+M5OTatWvMmzePNWvW8Mcff3DlyhWzNKfc9leaFXmArLW+DlTIVnYRw64WJZasIAshCo2Nm4fuVjdv3iQpKQmAnO43cXR0ZPbs2XTv3p3x48ezbt06u8fIT9vcMG5FZiuwyHqoxKRJk5gxY0aexjEGpqdOncpVO+NNeLa2TDOWW8tDtraqacyhzulaWlqaxbUqVapYHf+ee+4xm+edZF3RtjZ21hzaixcvAtjMiza6evWq2RzuNNe8yEvfx44do3nz5iQnJ9O6dWs6deqEt7c3jo6OJCYmsnTpUqsr/7akpaXRrl07fvrpJxo3bky/fv2oVKkSzs7OAEybNi1X/ZV2xbnNW6kiN+kJIUTR2L59O7dv36ZKlSr4+/vnWLdbt260a9eO9evXs2nTplyNk5+29goODsbV1ZWTJ0/y559/cu+99xbKOK1atSIqKootW7bwxhtv2N3OGMiePXuW+++/3+K6MXWjsLdVO3funNVyY4pKYYxv7PPSpUuUL1/e7vp3mmt+5pKbvufOncvFixeJjo62OInw008/ZenSpbmaw5dffslPP/1EWFiYxQ4hZ86cYdq0abnqr7STo6btJCvIQghR+DIyMkyrqAMGDLCrzZw5c1BKMW7cuBy3ESvotvYoV64cTz31FIDVXRQKypNPPomfnx/x8fFs3rw5x7pZVwEDAwMBrB6PnJKSwt69e3Fzc+O+++4r0Plm9/PPP5tyfbMyzss4z4L08MMPA/D999/bVT8oKAjAaqpDeno627dvz/Nccur70qVLVvPSjxw5AkDv3r0trlnrBzAdjmNtN4q89FeWSYBsJwmQhRCicJ0/f57+/fsTFxdHrVq1mDhxol3tAgMDGTRoEPv27TNtqWav/LS114wZM6hWrRrLli3jpZdesrn/cn6OgPby8mL+/PkA9OvXz+bNVD/++COPPPKI6fmgQYNwdnbm3XffNQVIRpMnT+by5csMGjTIYmu4gnbp0iWLXyB2797N8uXL8fb2pmfPngU+5pgxY3B2duall17ijz/+sLiemppqFjy3aNGChg0bsm3bNr780nwvgQULFuQ5/xjgiSeewNfXl08++cRiL+ypU6da/bdh/OtK9l9uNmzYYPO4dOPNr9m3r8upv2PHjvHvf5eos9uKhKRY2EkCZCGEKDjGk/QyMjJMR01v376d1NRUmjdvzvLly6lYsaLd/c2YMYPPP//cIsgr7Lb2qFatGlu2bKFXr15ERkaydOlS2rVrR926dXFwcODcuXPs2LGDP//8k8qVK1vsJmGvgQMHcuPGDcaMGUOXLl1o2rQpLVq0MB01HR8fz759+8zeV39/fyIjI3nuuecICgqib9++VKpUia1btxIfH09AQACzZs0qqLfCpjZt2rB48WJ27txJy5YtOXPmDJ999hkZGRl8+OGHdqVA5FZAQABRUVEMGTKE+++/ny5dutCgQQPS0tI4ceIE33//PZUqVTId1KGUYsmSJXTs2JHevXvTq1cv6tevz759+9i8eTNdunRh/fr1eZqLp6cnixYtol+/frRu3Zp+/fpRtWpVtm/fzoEDB2jTpg3btm0zazN69Giio6Pp06cPvXv3pnr16hw4cID169fTt29fq/tHt2/fns8//5xevXrx6KOP4u7uTu3atXn66ad57LHHqF+/PnPnzuXXX38lMDCQEydO8M0339CtWzerQXVZJivIdpKjpoUQouBMmzaNadOmMWvWLD755BNSUlJ45plnWLduHfHx8VZ3IshJzZo1TVtR5VZ+2torICCAvXv3EhMTQ6tWrYiPj2f+/PnMmzeP7777jvvuu4/Fixdz9OhR05/+82LYsGH8+eefjB8/HqUUy5cvZ9asWXzxxRf4+vryn//8x2Klc/To0WzYsIGHH36YVatWMXfuXM6fP8+4ceOIj4/Hz88vvy//jurUqcMPP/yAr68vCxcuZOXKlQQFBfHtt9/Sr1+/Qht30KBB7Nmzh4EDB7J//34WLFjAsmXLOHLkCE8++STvv/++Wf2WLVvy/fff06FDB9atW8e7777LrVu3iIuL41//+le+5vLkk0+yfv16mjVrxsqVK1m4cKEpbaZOnToW9R944AFiY2Np0aIF3377LR988AGXL19m9erVjBw50uoYw4YN47XXXuPSpUvMnj2byZMns2TJEgA8PDz47rvvGDBgAL/99hvz589n//79TJ482eJkvruB0qX07ung4GCd1yM582Jr4lZCloYA0LpWa7Y9uy3H+kIIkdXBgwcLPY9TCCGEgb0/c5VSe7TWFnsWygqyncx2sdCyi4UQQgghRFklAbKdJAdZCCGEEOLuIDfp2UkCZCGEuLvExMSQmJh4x3pNmzalR48eBTp2SkoKkZGRdtUdPHjwHfeLFkXPeCPqnfTo0SPXJyeKwicBsp3kqGkhhLi7WDsG2pqwsLBCCZDtPZghJCREAuQSyN7PLy9Hi4vCJwGynWQFWQgh7i7WDs8oKv7+/pTWm+iFgXx+pZvkINtJAmQhhBBCiLuDBMh2MtvFIkN2sRBCCCGEKKskQLaTrCALIYQQQtwdJEC2kwTIQgghhBB3BwmQ7SRHTQshhBBC3B0kQLaTrCALIYQQQtwdJEC2kxw1LYQQQghxd5AA2U6ygiyEEEIIcXeQANlOEiALIYQQQtwdJEC2kwTIQgghxN0hMTERpRSDBw8u7qmIYiIBsp0cHcx3sZAjJIUQIveUUmYPV1dXKlWqRFBQEMOGDWPdunWkp1u/z2Pw4MGmdt9++63VOlOnTkUpxeLFiwusrb2M7adOnWr1elpaGv/973/p0aMHNWvWxN3dnXLlylGnTh169uzJ4sWLuXbtWp7GBsPx1FnfWwcHB3x8fGjRogXvvfcet28X3uJOXFxcjq9diNLG6c5VBICDckCh0BgC4wydYbb1mxBCCPtFREQAkJ6eTkpKCr/99hsff/wxS5YsITg4mOXLl9OgQQOb7cePH0/nzp1xdMz9z+H8tM2rQ4cO0bt3b37//Xd8fHxo164dderUwcnJib/++ott27axdu1aXnvtNf7+++98jfXiiy/i4+NDeno6CQkJrF69mvj4eLZs2cLq1asL6BWVbdWrV+fgwYN4e3sX91REMZEAORecHJxIy0gDDDtZOCIBshBC5IW1lcZz587x/PPP8/nnn9OhQwd2795N5cqVLerVr1+f3377jaioKIYPH56rcfPTNq/OnDlD+/btOX36NM8//zxvvfUWnp6eFvU2bdrEuHHj8j3e2LFj8ff3Nz3/7bffeOihh1izZg1bt26lbdu2+R6jrHN2diYgIKC4pyGKkaRY5ILkIQshSrTZsyE21rwsNtZQXgpUqVKFFStWEBISwsmTJ3nrrbes1ps8eTLlypVjypQpuU5JyE/bvJo4cSKnT59mwIABzJ8/32pwDNCxY0d2795d4OPff//9hISEAPDTTz8BEBMTg1KKmJgY1q9fT0hICN7e3iilTO0uXbrEa6+9RsOGDXFzc8PX15fOnTuzefNms/4HDx5MaGgoANOmTTNL84iLizOr++mnnxIaGoqvry9ubm7cd999vPnmm9y6dcvq3Lds2UKXLl3w8/PDzc2NBg0aMGHCBC5dumRRNyQkBKUUaWlpTJ8+nXr16uHm5kZAQAAfffSRqd7ChQtp0qQJ7u7u1KhRg4iICDIyMsz6yikH+fr168yaNYvg4GC8vLzw9PTkvvvu44UXXuDcuXPWPwRR6kiAnAsSIAshSrSHHoK+ff8JkmNjDc8feqh455ULDg4OTJo0CTAEU9bu96hWrRqvvPIKZ8+eZXYug//8tM2L69ev8+mnnwL/pJXkxMmpcP6wa3wfswbAAF988QXdu3fHy8uLkSNH0rdvXwBSUlJo0aIFM2fOxNvbm7Fjx9K7d2/i4+Pp1KkTH374oamPHj16EBYWBkDbtm2JiIgwPbKuZA8dOpQBAwZw5MgRevXqxXPPPYefnx+TJ0+mS5cuFjnSH374IR07dmTHjh306NGDsWPH4ufnx6xZs2jRogUpKSlWX2v//v1ZtGgR7du3Z+jQoSQnJzNixAhiYmJ4+eWXef311wkKCiI8PBwXFxemT5/OO++8Y9f7mJycTIsWLZgwYQJXr15lyJAhjBo1ivvuu4+oqCgOHjxoVz+iFNBal8pHs2bNdFHzmemjmYpmKvri9YtFPr4QovT6/fffi2ag777TumJFrSdPNnz97ruiGddOgDb812PbzZs3tZOTkwb0sWPHTOVhYWEa0Js2bdJXrlzRVapU0R4eHvr06dOmOhERERrQH330kVmf+WlrL2P7iIgIU9nWrVs1oGvUqJGnPnOjdu3aGtAJCQlm5QcOHNDu7u4a0Nu2bdNaax0dHa0BrZTS69ats+hrxIgRGtAjRozQGRkZpvI//vhDly9fXru4uJiNExsba/HaszKO17NnT339+nWza8b3LTIy0lSWmJioXVxctJeXlz548KBZ/VGjRmlADx8+3Ky8bdu2GtDBwcE6OTnZVH706FHt7OysfXx8tL+/vz516pTpWnJysq5QoYKuWLGiTktLM5UnJCRoQIeFhZmN8dRTT2lAjxw5Uqenp5tdu3z5sk5JSbH6+kXRs/dnLrBbW4kzZQU5F2QFWQhRaJQqmEe7dnDhArzxhuFru3YF13cRcXV1pUKFCgA2b1jz9PRk2rRpXLt2jcmTJ+eq//y0za2zZ88Chpu+rImJiWHq1Klmj7179+ZrzMjISKZOncrkyZMZNGgQDz30EDdu3KBnz560bt3arO4TTzxBly5dzMrS0tJYtmwZnp6evP3222arzvfeey8vvPACqamp/Pe//7V7TvPmzcPJyYmoqCjc3d3Nrk2ePJkKFSqwfPlyU9myZctITU1lzJgxFrnAM2bMwMvLi48//thqasbMmTPx8fExPa9bty6tWrUiJSWFyZMnm30WPj4+PPbYY1y4cIG//vorx9dw/vx5PvvsM6pWrcqcOXNwcDAPoby8vOSmvjJEbtLLBQmQhRCFpqC2jjSmVYwaBR98ACtXQmZ+aGmibaQEZDVs2DDmz59PdHQ0L774Ik2aNLG7//y0zY07vY6YmBi2bt1qVubv70/Tpk3zPOa8efNMY3p6evLAAw8waNAgRo4caVG3efPmFmWHDh3i+vXrtGzZEj8/P4vr7dq148033+SXX36xaz7Xr19n3759VKxYkcjISKt1XF1dzdITfv75Z9NY2fn6+hIYGMi2bds4dOgQDz74oNn14OBgizbVqlUDoFmzZhbXjAHzqVOnqF27ts3XsWvXLjIyMmjTpg0eHh4264myQQLkXMgaIKdnWN+nUwghio0xODYGxaGh5s9LiZs3b5KUlARApUqVbNZzdHRk9uzZdO/enfHjx7Nu3Tq7x8hP29yoWrUqgM3Vyaw3sU2aNIkZM2bke8yEhASz3N+c3HPPPRZlxhvgjHPPzlhuKwc4u+TkZLTW/P3330ybNs2uNvmZg7VVXGNud07X0tLScpyTcSxbfw0QZYukWOSCrCALIUq0XbvMg+HQUMPzXbuKd165tH37dm7fvk2VKlXuGOh169aNdu3asX79ejZt2pSrcfLT1l7BwcG4urpy8uRJ/vzzz0IZIz+srWwbg0hjekh2Z86cMat3J8Z6gYGBd7y/qLDmUBCMaRt3SsUQZYMEyLkgAbIQokQbP95ypTg01FBeSmRkZJhWUQcMGGBXmzlz5qCUYty4cRbbdRVmW3uUK1eOp556CoDp06cXeP+FoWHDhpQrV469e/eSnJxscT02c5eUoKAgU5nx0BVrpyB6enpy//3389tvv5n+MnAngYGBABbbxIFhJXfv3r2mbeKKSvPmzXFwcGDbtm1FtkWgKD4SIOdC1pPzJEAWQoiCdf78efr3709cXBy1atVi4sSJdrULDAxk0KBB7Nu3z7Slmr3y09ZeM2bMoFq1aixbtoyXXnrJZnBlbW/f4uDi4sLAgQO5evUqU6ZMMbt29OhR5s+fj7OzM08//bSp3HhT5YkTJ6z2+fLLL5OamsqQIUOspkUkJyeb8o4BBg0ahLOzM++++y5Hjhwxqzt58mQuX77MoEGDcHV1zevLzLVKlSrRv39/zpw5w6uvvmrxC9XVq1dLzGco8k9ykHNBVpCFEKJgGE/Sy8jIMB01vX37dlJTU2nevDnLly+nYsWKdvc3Y8YMPv/8c4tgqrDb2qNatWps2bKFXr16ERkZydKlS2nXrh1169bFwcGBc+fOsWPHDv78808qV65cIk5wmzlzJt9//z0LFixg165dhIaGcuHCBVauXMmVK1dYsGABderUMdVv2LAh1atXZ8WKFbi4uFCrVi2UUjz99NPUrl2bIUOGsGfPHt5//33q1atH586dqVWrFklJSSQkJLBt2zaeffZZFi5cCBhuVIyMjOS5554jKCiIvn37UqlSJbZu3Up8fDwBAQHMmjWryN+XBQsWcODAARYuXEhcXBydO3fGxcWFhIQENmzYwFdffWU6lEWUbhIg54LZTXpabtITQoi8Mt6s5eLigpeXF7Vr1+aZZ56hd+/edOrUyWILrTupWbMmY8eOZebMmbmeS37a2isgIIC9e/fy6aefsmrVKuLj4/nmm29QSlG5cmWaNm3Kv//9b/r162fzpL2i5OfnR3x8PG+//TarV69m7ty5uLu707x5c8aNG0enTp3M6js6OrJmzRomTJhgCqK11rRq1cq0M8R7771H165dWbhwIZs3byYlJQU/Pz9q1arFuHHjGDRokFmfo0ePpn79+syZM4dVq1Zx/fp1atasybhx45g4caLZVm5FxdfXlx9++IHIyEg+++wzFi1ahKOjIzVr1mTIkCE0atSoyOckCofSBbW1UBELDg7WhXEkZ45jLgpmz5k9AOwavovgapZbyQghhDUHDx4s0nxJIYS4m9n7M1cptUdrbRHQSQ5yLkiKhRBCCCFE2ScBci5IgCyEEEIIUfZJDnIuODrILhZCCHG3iImJITEx8Y71mjZtSo8ePQp07JSUFJunzmU3ePBguw8GEULYRwLkXJAVZCGEuHtYOwbamrCwsEIJkO09dS4kJEQCZCEKmATIuSBHTQshxN3D2iEVRcXf35/SehO9EGWB5CDngqwgCyGEEEKUfRIg54IEyEIIIYQQZZ8EyLkgR00LIYQQQpR9EiDngqwgCyGEEEKUfRIg54IEyEIIIYQQZZ8EyLlgtouFll0shBBCCCHKIgmQc0FWkIUQQgghyj4JkHNBAmQhhBBCiLJPAuRckF0shBBCCCHKPgmQc0FWkIUQQgghyj4JkHNBjpoWQoiy7+DBg0RERPDEE09Qq1YtlFIopbh9O+eFkaSkJMaOHYu/vz+urq5Uq1aNIUOGcOrUqSKauRCioDjduYowkhVkIYQo+zZs2MD06dNxdHTk3nvvxc3NjZs3b+bY5uLFi7Ro0YI//viDdu3a0b9/fw4dOkR0dDT/+9//iI+Pp27dukX0CoQQ+SUryLkgAbIQQpR9Xbt25ccff+TKlSscPHiQKlWq3LHNxIkT+eOPP3jppZfYsmULM2fOZO3atcybN4/z588zevToIpi5EKKgSICcC44OcpOeEELkxaFDh1BK0a5dO5t1mjRpgrOzM2fPni3CmVlq2LAh//rXv3B3d7er/rVr1/j444/x8PBg2rRpZtfGjBmDv78/GzZs4NixY4UxXSFEIZAAORdkBVkIIfImICCA0NBQYmNj+eOPPyyu//DDDxw4cIAnnniCe+65pxhmmHfx8fHcuHGDli1b4uXlZXbNwcGBTp06ARAbG1sc0xNC5IHkIOeCBMhCiMISEhJiUda3b19Gjx7N9evXefTRRy2uDx48mMGDB3PhwgWefPJJi+ujRo2iX79+nDx5kqefftri+iuvvMJjjz3G4cOHCQ8Pt7g+adIkOnTowN69e2natGmeXldWo0ePJjY2lkWLFjFnzhyza4sWLQKwOg9r1q5dy969e+0e28fHh7Fjx9pdPzcOHz4MQIMGDaxev/feewGs/mIghCiZJEDOBTlqWggh8q5Hjx5Uq1aNmJgYZsyYgaurKwApKSmsXLmSevXq0aFDB7v6Wrt2LUuXLrV77Nq1axdagHzp0iUAvL29rV43lqekpBTK+EKIgicBci7ICrIQorDExcXZvFauXLkcr1esWDHH6zVr1szxesOGDXO8XhCrxwBOTk4MGzaM6dOns2rVKgYMGADAxx9/zI0bNxgxYgRKKbv6iomJISYmpkDmVdi01gB2vzYhRPGTHORckABZCCHyZ8SIETg5OfHhhx+ayhYtWoSLiwvPPvtsMc4s74wrxMaV5OwuX75sVk8IUfLJCnIuyFHTQgiRP9WrV+exxx5jzZo1HDx4kOTkZA4cOEC/fv2oVKmS3f2UpBzkhg0bArZzjP/880/Ado6yEKLkkQA5F2QFWQgh8m/06NGsWbOGRYsWkZycDNh/c55RScpBfvjhh3F3d2fHjh1cuXLFbCeLjIwMNm7cCEBoaGihjC+EKHhFnmKhlPJRSn2hlDqklDqolHpEKeWnlNqklPoz86tvUc8rJ7N3zCY2IdYsQD5x6QSzd8wuxlkJIUTp1L59exo0aMDSpUtZuXIlDRo0yHXwGBMTg9ba7kdiYmLhvBjA09OTp59+mmvXrjF16lSzawsWLCAxMZHOnTvLSXpClCLFsYI8D1ivtX5SKeUClAMmAlu01jOVUhOACcC/i2FuVj1U7SH6ftGXZ5v+kx/3XcJ3jGsxrhhnJYQQpZNSipEjR/Lyyy8DuV89LmwXLlzg1VdfNXsOMHToUNONdhMmTCAgIMBU56233iIuLo65c+eyd+9emjdvzsGDB/nyyy+pXLky7733XtG+CCFEvijj3bVFMphS5YF9QF2dZWCl1GEgRGt9RilVFYjTWjfMqa/g4GC9e/fuwp1wFrEJsXT/tDvX064D0KFuBzY9vanIxhdClG4HDx7kvvvuK+5plBjJyclUrFgRZ2dn/vrrLypUqFDcUzJJTEykTp06OdaJjY212Ls6KSmJadOmsXbtWs6cOUOFChXo2rUr06dPp0aNGoU4YyFEdvb+zFVK7dFaB2cvL+oV5LrA30C0UupBYA/wIlBFa30GIDNIrmytsVJqBDACoFatWkUz40yhdUKp6lmVo8lHAahYrmKRji+EEGXJvn37yMjIoE+fPiUqOAbw9/cnL4tHfn5+zJs3j3nz5hXCrIQQRamoc5CdgCDgA611IHANQzqFXbTWi7TWwVrr4Nzc7VwQYhNiOXn5pOn56Suni3R8IYQoS2bPNtzDMWbMmGKeiRBCWCrqAPkUcEprvTPz+RcYAuZzmakVZH49X8TzylFsQix9v+jLv6r9y1T246kfiU2ILcZZCSFE6fLrr7/y9ttv8+STT7Ju3Tq6d+/Ov/71rzs3FEKIIlakAbLW+ixwUillzC9uD/wOfAWEZZaFAV8W5bzuZNfpXax8ciXVy1c3lQVVDWLX6V3FOCshhChd9uzZw8SJE9m0aRN9+vQhOjq6uKckhBBWFccuFs8DyzN3sDgGPIshUF+plBoKnAD6FMO8bBrfcjwA0Xv/+WHu6+ZrKhdCCHFngwcPZvDgwcU9DSGEuKMiD5C11nsBi7sFMawml2iujq6m7+WgECGEEEKIsqnIDwopzZwdnU3fS4AshBBCCFE2SYCcC7KCLITIj6Lcd14IIe5WBfGzVgLkXHBxdDF9LwGyECI3HB0dSUtLK+5pCCFEmZeWloajo2O++pAAORdcnf5ZQU7X6cU4EyFEaePl5cXly5eLexpCCFHmXb58GS8vr3z1IQFyLkiKhRAir/z8/EhOTubChQukpqZKuoUQQhQgrTWpqalcuHCB5ORk/Pz88tVfcWzzVmplXUGWAFkIkRuurq7UqlWLpKQkEhMTSU+Xv0IJIURBcnR0xMvLi1q1auHq6nrnBjmQADkXJAdZCJEfrq6uVK1alapVqxb3VIQQQuRAUixywc3JzfS9BMhCCCGEEGXTHVeQlVIOQGegA9AcuAdwA5KAP4AdwGqt9YlCnGeJIDnIQgghhBBln80VZKWUl1JqCnAKWAuEAkcyv18KfA+4AxOAY0qpjUqp1oU94eJktotFhuQPCiGEEEKURTmtICcAvwDjgK+01ldsVVRKBQJ9gdVKqWla6wUFO82Swd3J3fS9rCALIYQQQpRNOQXInbXWe+zpRGv9C/CLUuoNoFaBzKwEkl0shBBCCCHKPpspFvYGx9naXNdaH8rflEqg2bMhNtbsJr2H/7xhKBdCCCGEEGWK7GJhj4cegr59qfzTbwCEJMDi5VcM5UIIIYQQokyxax9kpVQGYOvYJw1cBvYB87XWawpobiVHaCisXEnAo135pgY0Pw1PP+XCutDQ4p6ZEEIIIYQoYPauIL8M/AUcBGZjuHHvHeAQcBqYB6QDXyilBhXCPItfaCipTRrR7Qh8EAwbaqaSoTOKe1ZCCCGEEKKA2XuSXjVgh9b6qWzlE5RSKwBfrXUHpdR/gfHAsoKcZIkQG4vrgUPE1oZRuyG2juZG2g08XDyKe2ZCCCGEEKIA2buC/CwQbeNaNPB05vefAffmd1IlTmws9O3LrSd7El8L+vaBlZ/Drc3ri3tmQgghhBCigNkbIDsBATau3Zeln1TgZn4nVeLs2gUrV0KjRrjdhrg6hiA546edxT0zIYQQQghRwOxNsVgBvK2UcgK+Bv4GKgFPANP5Z3U5CENectkyfjwAjvv34pq5/XFcHTgVPoCKxTgtIYQQQghR8OwNkF/EsDr8Joab84xuAR9huGkPYCewpcBmV8I4upUzBcgAV1OvFt9khBBCCCFEobArQNZapwIvKqWmAU2Ae4CzwK9a66Qs9eIKY5IlhUO5crhJgCyEEEIIUabZu4IMQGYwvLWQ5lLyubnhlv7PUwmQhRBCCCHKHpsBslLqUWC71vpy5vc50lp/W6AzK4lcXXG/rTCemSIBshBCCCFE2ZPTCvI3wMPAT5nfa0DZqKsBx4KdWgkkK8hCCCGEEGVeTgFyHeBMlu+Fmxtu6bKCLIQQQghRltkMkLXWx619f1dzdcXt9j+L6BIgCyGEEEKUPXYdFKKUuk8p9XCW5+5KqbeUUmuVUs8X3vRKGDc3XCXFQgghhBCiTLP3JL33gceyPJ+DYW9kN2CWUmqc1VZljZubbPMmhBBCCFHG2RsgNwbiAZRSzsAgYKzWugswERhSONMrYVxdJUAWQgghhCjj7A2QPYDLmd8/nPl8debzn4HaBTyvksnNzewkvSupV4pvLkIIIYQQolDYGyAfwxAYA/QEftFaX8x8XhG4OyJFNzdcb2vTU1lBFkIIIYQoe+w9Se8/wAdKqT5AIPBslmshwP4CnlfJ5OqKa5oEyEIIIYQQZZldAbLWeolS6k/gIWCC1npLlstJQGQhzK3kcXXFOV2bjkyRAFkIIYQQouyxdwUZrfU2YJuV8qkFOaESzcGBdAeFS7om1UkCZCGEEEKIssjuANlIKeUEjABCMRw9HQt8qLW+nWPDMiLN2QG32+kSIAshhBBClFG5DpCBeUAz4GMMu1mMBx4AwgtwXiVWmosTrrcNp4VcTb2K1hql1B1aCSGEEEKI0sJmgKyUCtRa/2Ll0pPAvVrry5n1fsMQLN8VAfJtZ0c80x35m3QydAY3b9/E3dm9uKclhBBCCCEKSE7bvK1XSn2klKqUrfws0BFAGZZOQ4EzhTS/EifNxREvXEzPJc1CCCGEEKJsySlAbghcBw4qpV7NPEEPYAzwrlLqPJAMDABGF+40S47bLk54aWfTcwmQhRBCCCHKFpsBstY6RWv9ItAaaA/8rpR6XGv9PeAPdADaAHW01luLYrIlQbqzE14ZEiALIYQQQpRVdzxJT2t9UGvdFRgLzFJKbQLqa633Zz5uFfYkS5J0Fyc89D+p2xIgCyGEEEKULXcMkJVS7kopb631/4AmwDpgq1JqgVLKr9BnWMKkuzrjkeFoei4BshBCCCFE2WIzQFZK1VdK7QCuAUmZJ+m11VrPBe4DnIFDSqnnlVKOtvopazKcnSl3+5+3TQJkIYQQQoiyJacV5P8CicA9gA+wFFitlHLTWl/QWodjyEPuCfxayPMsMdJdXXDP+Odtu5J6pRhnI4QQQgghClpOAXIjIEZrfV5rfQV4H/ACahgrZOYgtwNeL9xplhwZri643/7nYBBZQRZCCCGEKFtyOklvI4ab8soDNzAcL30UOJa9otZ6TeFMr+TRri64ZTlUWwJkIYQQQoiyJacAeQgwCZgAuAB7gA5a64yimFhJpV1dcE3/5/mNtBvFNxkhhBBCCFHgbAbIWuurGIJjkYV2ccUl7Z/fEW7evlmMsxFCCCGEEAUtp10slK1rOclru9JCu7nimqZNz2/clhVkIYQQQoiyJKeb9P5QSg1TSnnY05FSqplS6r+U9VVnN1ecb/8TIMsKshBCCCFE2ZJTDvIEYCowTym1EfgBOABcAG5h2PqtDtAM6ALUBBYDUYU33RLA1Q3XLCkWsoIshBBCCFG25JSDvEoptRrDXsfPYDhquipgXD5VQCqGm/c+BD7WWp8v1NmWBG5uON+WHGQhhBBCiLIqpxVktNYa2JT5QClVFcPBIW5AEpCgtU4t7EmWJMrdTW7SE0IIIYQow3IMkLPTWp8BzhTSXEoF5eqGc9YUC9nmTQghhBCiTMnpJj1hhYNbOUmxEEIIIYQowyRAziXl7o5L6j8nhchNekIIIYQQZYsEyLnk6F5OcpCFEEIIIcowCZBzyZBiIfsgCyGEEEKUVXYFyEqpZ5RSFWxc81NKPVOw0yq5HMt54JKWJcVCbtITQgghhChT7F1Bjgbq2bhWJ/P6XcHBvRwuabKCLIQQQghRVtkbIKscrlUALhfAXEoFJ3cPs10s5CY9IYQQQoiyxeY+yEqpJ4AnshRNVkr9na2aG9Aa2GXvgEqpROAKkA7c1loHK6X8gM8AfyAR6Ku1Tra3z8I2eza4x4zlnp6NqHw0isCkNEJ+eJD+e7zA9TL/p7tT49vtON3bkMR7K+Py8z7SXh3LZwc+o+k9TenfuD8rDqygnl89Hqr2ELtO72J8y/HF/bKEEEIIIYQVynBYnpULSg0HRmQ+bQYcAq5lq5aaWf6m1jrBrgENAXKw1vpClrLZQJLWeqZSagLgq7X+d079BAcH6927d9szZL7FxsI7j25k6c0BxLd9hG5bv+EG5QwXnW6CyuC2A7zVzpFxW9MZ1M+ZjbXTcHN0w9HBEY3G2cGZKW2n8Pb2t1n55EpC64QWydyFEEIIIYR1Sqk9Wutgi3JbAXK2xrHAKK31oQKYSCKWAfJhIERrfSbzOOs4rXXDnPopygAZYEOdkXRO/BCANBxxJJ1UXNAOqbhmwOZHqhC47xyD+jkTV0eRmpGKq6MrWmvSMtIY9MAg1h1ZJ8GxEEIIIUQJYStAtisHWWsdWhDBsbE7YKNSao9SyrhCXSXzGGvjcdaVrTVUSo1QSu1WSu3+++/s2R6Fy83bxfS9M+loFG6k4p5hSNDuFH+OhcGwsXYaqRmpKBS30m+RmpGKRvPx/o8ZFTxKgmMhhBBCiBLOZg5ydkqpakB3oAaG3OOs9J1SIrJoqbU+rZSqDGxSStkdeGutFwGLwLCCbG+7gnDphqvp+z0EEsgv3MQVrdJw0xmcvMedkbtvsL2e5QpyakYqjzd4nA92f0Cof6gEyUIIIYQQJZi9+yD3BI4B7wFDgT5WHnbRWp/O/HoeWAM0B85lplaQ+fW8/S+h8MXGwvVjp0zPG/E71/HgNo6kK2duOkCV8zd4p60jyz5LIyTBELsrFM6OzigU3yV+x2utXqPvF32JTYgtrpcihBBCCCHuwN5t3t4CNmJIhaiuta6T7VHXnk6UUh5KKS/j90An4ADwFRCWWS0M+DJXr6KQ7doFdXz+uQfxdxrxGF+znEEsL9eez/vfT5KXI/+q3Iz/TuxO90v38H+d/o8HqjzAwCYDaXpPU9rUbsPtjNusfHIlu07bvemHEEIIIYQoYvbepHcV6KG13pyvwZSqi2HVGAzpHZ9orWdkntK3EqgFnAD6aK2TcuqrqG/SY+ZMeO01AGYznn8zy1Deegb7/12NJhHvgY359PqsF4MeGESv+3oV1WyFEEIIIcQd2LpJz94c5B+AhkC+AmSt9THgQSvlF4H2+em70Ln+k4Pspm4ZbjUEyHDkmpcrXLxos2l51/JcvnXXnKUihBBCCFGq2Rsgvwwsz1xJ3gSkZK+gtb5egPMqeVz+2cXCVaVmCZCduOLpAkm2F7zLu5bn0s1LhTxBIYQQQghREOwNkPdnfo3mn9AwO8f8T6cEy76CbKQduermANevQ1oaODtbNJUVZCGEEEKI0sPeAHkItgPju0OWFWQXlfpPeYYjN9Jvgq8vJCdDZcstnMu7lufva0W7b7MQQgghhMgbuwJkrXVMIc+j5MthBfnm7Zvg52fIQ7YRIB9NOloUsxRCCCGEEPlk7zZvImsOMllXkJ24kXYDKlSwmYfs7erN5VRJsRBCCCGEKA1yc5JeP2A40ADLk/TQWls9HrrMyLKC7EqWFeQMR27evmoIkG3sZCE5yEIIIYQQpYe9J+kNAJYCRzAcNf0V8E1m+8vAgsKaYImRNUDOKcXCCtnFQgghhBCi9LA3xWIc8AbwXObz97XWQ4A6wAWgbG/xBuY36WVPsbidc4qFrCALIYQQQpQe9gbI9wI7tNbpQDpQHkBrfQWYBYwpnOmVIDmmWNx5BVkCZCGEEEKI0sHeAPkSYIwQ/wLuy3JNARUKclIlkq0VZO1455v03LwlQBZCCCGEKCXsvUlvN/AAsAFD/vEUpdRtIBWYAuwsnOmVIFlXkHXuVpC9XLy4fOsyWmuUUoU9UyGEEEIIkQ/2BshvA7Uzv5+S+f37GE7P2wWMKPiplTBZVpCds+Ug30y/meMKsrOjMy6OLlxPu46Hi0dhz1QIIYQQQuSDvQeF/Aj8mPl9CvCEUsoVcNVa3x25A1lWkF20+S4WnVftg5vvw/Hj/5THxsKKFVCvHowfb8pDlgBZCCGEEKJksytAVkp5AZ5a6zPGMq31LeCWUqoqcEVrfbWQ5lgy2NzFwpHf/T3g3U1w9aohMAbo0QOUgjVrgH9u1KvqVbUIJy2EEEIIIXLL3pv0lgDTbVybCiwukNmUZFlWkJ0zsuYgO7GzQTnDarHW0K2b4WEMjkNDgf9n777Do6q2Bg7/zpQkJPTepWmUEkGIWACNgA1QUAhRxH5VvF71WtBYULCgUfxsV8WugEJAQRELICMIIoYmTVAg1NA7gSRT9vfHzpmSTJIJZFLXe588OXPKzJ7kXFnZs/ZaUslCCCGEEKKiCDVA7gXMKuDY97nHKzf/HGQVWMUiy5UFV12l85BPntRf99/vDY5BKlkIIYQQQlQUoQbItSi4GUgWUKdkhlOO5atiofQDT26Zt19+AadTn2cY8OabvnQLZAZZCCGEEKKiCDVA/gfoV8Cxq4FNJTOccsxiAZsvZduOU28oK+euO+jLOf72Wx0kKwWDBnmD5JqRNTmSLe2mhRBCCCHKu1DLvL0FvGcYRg7wKbALaALcgm4/PSIsoytvIiLA5dKb5OAkAjw2zk4/BklJ+ishAbp3h6uvhk2bIC0NEhKoGSEzyEIIIYQQFUFIM8hKqQ+AZ4B7gVXAvtzv/waeyj1e+QVrN+2x8volETB+vA6OU1KgeXM4ehTGjyflYljxxWt0/Ox7Vu5eCYAj3UHKohTvdyGEEEIIUX6EmmKBUup5oCk61eLm3O9NlVIvhWls5U+wUm/mIj1TfDx89x3MnAlA3602Wtz1CFvaNWDymsm8tvg1EqclYrPYSJyWSHzT+NJ8B0IIIYQQogghB8gASqkjSqkflVKTcr9XraTaAmaQTzhP4PLo1AsSEmDiRFi1Cp58ki4PjGX7+6/ydvW1eJSHh2c/TEKrBMYuHEvq4FQAmUUWQgghhChHCgyQDcO42jCMmn7bhX6V3pDLULAZZI8Nl8fFb9t/853Xvz9UqwYvvggjRtDlxoe4u+vdZLuzMTCYum4qI7rptG2ZRRZCCCGEKF8KW6T3HXAB8EfudmEUYC2pQZVbwWaQlX7b36z/hl5n5JaDdjggOxvuuw/efZcVZ9fms52fMTxuOBNWTcDA4OVFL/PmkjeZPnQ6Ca0T8r6SEEIIIYQoI4UFyK2BDL9tEXQGOTdA3vANr17+KsYvv0BiIrRoAXfdxYoLW9Pirkd456V7uHfjVMZdPo6Rc0aS484BBZPXTGbymsnE1o9lw/4NtK3blvim8UxeM5m2ddsy8uKRZfBGhRBCCCGqrsIC5E/QVSvWA5cAs5RSB0plVOVVkBlkQ9lRwKZDm+j1aS8mbexMy9RUeOwxyMxkzhku+r7/Kg0WziX1UZ1zHGWLItOZSY4nh1V7VrF672oynZnE2GMYkzCGgVMGYmAwfej0sniXQgghhBBVWmGL9HoCtXO3PwHahn005Z1fgGzOIEdZq3v3Ldy2kIvqTcd9SS+IiYHMTEZePJIuNz7Epe98D+ic45k3zGTCoAkA/L7zd044dZPCk66TPD73cW9wLKkXQgghhBClr7AZ5O3AEMMwjgMG0Dp3Oyil1LqSHly545diYc4gx1hrcNLvlJ3HdpJxLIMWuQGyv7SMNFIHp3oD35/Tf+bTlZ+iUMTWi2XDgQ14lIf7u98vwbEQQgghRBkpLEAeC7wDPIhehPdFAecZVOFFejYjiuV3Lee898/zHttyeEvQANk/n9iR7uDrv74m2h6Ny+1iw4ENRFgjyHHn8OaSN0lolSBBshBCCCFEGSgwxSK3O15zdP6xAdwHXBbkKyH3e+UXZJGeywVdmnQhsUOi99iWw1u8KRbBONId3jzj5xKew261A2C32LFb7CilGDRlEI50R/jeixBCCCGECKqwGWSUUnuAPYZhjAa+UUplFHZ+pRdkBtnt1o9b1WrlPVZUgJyWkUZShySSOiaRlpHGsE7DvFUsvtnwDc8nPE9aRhppGWkyiyyEEEIIUcoKDZBNSqnR4R5IhRBkBtkbINdu5T2mA+QGBQbI/qkWeQPgjOMZNIhpwPgB40tmzEIIIYQQolgKDJANw0gFkpVSm3K3C6WUSizqnAqvsBlk/wD5yBaIaVVggFyYVrVa6QBbCCGEEEKUicJmkBsA9tzthuiFeFVbATnIEGwGOQYOHiz2S7SqLQGyEEIIIURZKjBAVkol+G1fWiqjKe8KmUE+o/YZ3mPbjmzDU78allOZQa7dioXbF57eOIUQQgghxCkrrFGIyKuQHORoezQNYxoC4PK4OGTJPqUUizNqnyEzyEIIIYQQZSikRXoAhmE0BfqjS79F5T2ulBqZ76LKJsgMslLg8YDFomd/92buBWAPx6lXzAA5ZVEKZ9U7KyBAdqQ7SMtIC1jYJ4QQQgghwiekANkwjEHAl+hmIHshd/rURwGVP4IL0moa9CyyGSD/sfMPADI8R2hfzAA5vmk8idMSyXJmcTjrMCt2rSBxWiKpg4tcIymEEEIIIUpIqCkWLwKzgUZKqWZKqdZ5vtqEcYzlh1+KRVTuDDIEr4W8zXWg2CkWCa0TuO7s68hx5/Do7EcDguOURSmnPm4hhBBCCBGyUAPkFsCbSqnil2WoTAqZQQY4u/7Z3n1/Ht94SjnISR2TsFqsfLjiQ+7uejcAidMSiW8af4qDFkIIIYQQxRFqgPwbEBvOgVQIfjPIkYZvBtks9dajZQ/vvkUHVqJOIUAGqGavhsWw8PLClxk0ZRCpg1Olo54QQgghRCkJNUB+CLjLMIxbDMNoahhGdN6vcA6y3AiySA98M8jt6rajUUwjAHapY7iOHSnW0zvSHSROS2TG0Bncdd5duJSLHHfedG8hhBBCCBFOoQbIq4BOwCfAduBYkK/KL0iZN/AFyIZh0POMngBk2kEdP16sp0/LSPPmHE/7axpWw0qENYLJayaf5sCFEEIIIUSoQg2Qbwduy/26vYCvyq+IGWSAni1zA+QIsJ7MIv3gZtIPpZN+KJ0TzhOFPr1Zys1cnNe2bltev/J1vl7/NY50Rwm+ESGEEEIIUZCQyrwppT4N8zgqhgJmkM0cZPAFyC4ruA0457W2ZOc27K4eUZ3vb/zeO8scjDmLnNA6gdh6sdSKrEXq4FTSMtIkD1kIIYQQohRIJ73iCGEGOa5RHHWr1QX0LHKM03fseM5x3vrjrUJfYuTFI72B8Nn1z2b9/vUktE6QRiFCCCGEEKUkpADZMIx0wzA2F/C10TCM5YZhfGIYRtdwD7hMFZGDDGC1WPlgwAfENYojK9LK2dWa06xGM+/x1XtXh/xyZ9c/mw0HNpzemIUQQgghRLGEOoP8FTodowawBPgu93tNwA4sBS4AfjcM44owjLN8KGAG2T/FAuC6c67jz3v+pEmjtiwaOpsN923AwADg7wN/c9J5MqSXi60Xy/r9609/3EIIIYQQImShBsh7gb+BFkqpG5RS9yulbkA3EPkH2AZ0RHfbGx2WkZYH/o1CVPAZ5AAxMZCZSUxEDO3qtgPAozys27euyJdKWZTCvhP7WL9/PUopQJeBk456QgghhBDhFWqAfD/wmlIqy3+nUuok8H/Av5VSbuADdDm4yikgxSJ4DnKA3AAZdG6yadWeVUW+VHzTeP41818oFHsz93prJEtHPSGEEEKI8Ao1QK4NNCrgWCOgeu72EaCgcLHiK6LVdD6nESAntE4gdXAqJ5wneHTOo97Sb1LJQgghhBAivEINkL8DUgzDuM4wjAgAwzAiDMMYDKTkHgc9e7yp5IdZTvi3mi4kB9nLL0Du1NA3sR7qQr2E1gl0bdKVCasmMKLbCAmOhRBCCCFKQUh1kIF7gM+AaYAyDOMYesGeAcwERuSelwE8UdKDLDdCKPMWICYGTujmIP4zyL/v+J1hXw8LekmLmi146MKHaBjTEEe6g3X71nFm3TN5d+m7JLRKkCBZCCGEECLMQm0Uchi41jCMDkA3oDGwG1iqlFrrd960cAyy3ChgBjmUFIvWdVoTY48h05lJpjOTL1Z/UeDLHDhxgBs73UjitEQ+uuYjRswaQergVEmzEEIIIYQoBaHOIAOQGwyvLfLEyqpGDbBawe2mJseIJItsovIHyCkpEB8fECBbfpnPxxvOZmibZUW+zJKdSziz3plcd/Z11I+uj0d5GL9sPMk9kpm8ZjJpGWnEN41n8prJtK3bVpqICCGEEEKUoAIDZMMw2gOblFLZuduFUkoVXbusorNaoXFj2LkTgCbsYgut8+cgx8dDYiJcdRVUrw4OByQmMujLSXx/hpuDJw/me+qTrpP8a+a/AEg/nM6jFz3KL1t+IXFaIq1qt+Lbv79l1j+zsFqsPHPJMwycMhADg+lDp4f7XQshhBBCVCmGWWM33wHD8AAXKKX+yN0OfqLOQ1ZKKWuYxhhUt27d1NKlS0vzJbXu3eGPPwDowa8sogdz50Lv3nnOczhgwADo1Ak2boTUVEgoODVCKUXNl2pyPOc4APsf3U+96Ho40h1cM/ka6lWrx9YjW4mwRmA1rERYI5g+dLqkWwghhBBCnCLDMJYppbrl3V9YikUCsM5vWwA087WNboaeSQ6ag5yQAJdeCrNmwdNPFxocAxiGQavarVizdw2gZ5HrRdcjoXUCt5x7C/9L+x9RtiiyXLoU9SMXPSLBsRBCCCFEGBRY5k0pNV8pddxvu9Cv0htyGfMLkJuSARRQ5s3hgEWLoEkTePdd/bgIrWu39m6nH0rXT5PuYMraKQyPG06WK4tIq66k8caSN3CkF/2cQgghhBCieEKqg2wYRkPDMFr7PTYMw7jLMIzXDcMYEL7hlUNNm3o3C5xBzs05ZuJEOHwYvvxSPy4iSG5Vu5V3e8vhLd7ueck9kvlmwzfE2GOItEXSomYL3B43g6YMkiBZCCGEEKKEhdoo5FPgv36PRwPvAFcC0w3DuLVkh1WOhZJikZamc4779dMzyM2b68dpaYU+dcAM8uF00jLSSB2cisvjIqlDEjNvmMmMoTNoXL0xd553J0M7DCUto/DnFEIIIYQQxRNqmbfzgPcBDMOwoBuDPKGUSjEMYzTwIDqIrvyCpFjkC5BH+pVdO/dc+PNPGDq0yDzk1nUCA+R3+r0DkC/XeHjccNbtW8f4AeNP4Q0IIYQQQojChDqDXAs4kLvdFagLTMp9PA9oV8LjKr+CzCAX2GoaIC4OVq0K6anzplgU5NzG5/Lnnj9Dek4hhBBCCFE8oc4g7wDaA78C/YD1SqmducdqAVnFeVHDMKzAUmCnUqq/YRh1gSlAK2ALkKiUOlSc5yw1+XKQFW63EfzclBSwWAIDZIdDp1qMzN/cwz/FYsvhLSilMIzA505ZlMK6vetYsXsFLy98mc2HNhNbP5Ypa6YAMLTjUDbs30Dbum2lmYgQQgghxCkINUD+GEgxDKMPOkBO9jt2AfBXMV/3gdxrauY+fhz4WSn1kmEYj+c+fqyYz1k6atbUzT+OH6caWdTmMG53neDnxsfD9deDLffHbC7eS00NenqtqFrUiarDoaxDZLmyiP8gnk+u/YROjTr5nrJpPC/8+gLZrmyOZB9h0upJZDozibJGAfBHxh/E2GMYkzBGmokIIYQQQpyCkFIslFJjgf8Au3O/v+l3uC7wYagvaBhGc3SQ7X/NtcBnudufAQNDfb4ykSfNImgdZNA5x1Onwr59esbYDI4LyUX2z0NetmsZ106+1ts8BHQ+8oyhM7BarLy08CUynbqVdbY7myy3nsg/4TzBo3Me9QbHUi9ZCCGEECJ0oeYgo5T6XCn1H6XUR8qv/Z5S6h6l1GeFXZvH68BIwOO3r5FSalfu8+0CGga7MLe03FLDMJbu27evGC9ZwvIEyIXmIPfuDfXqwSuvwIgRRS7Uu63zbQGP0w+n8/jcxwP2JbRO4L74+1C5zQ3jGsXl2/YoDzfF3STBsRBCCCFEMYUcIJcEwzD6A3uVUstO5Xql1PtKqW5KqW4NGjQo4dEVg18eclMyCp5BBp1WceQIDB8eUsOQ+86/j43/2cg9Xe/x7vtf2v9YuXul7ynTHXy88mOi7dFEWCJYtWcVkdbIgG2rYeXjFR9LnWQhhBBCiGIq1QAZuBi4xjCMLcBk4DLDMCYCewzDaAKQ+31vKY+reM44w7t5GfMKDpDNnOMLL4T+/XV6RQgNQ9rWbcs7/d6h35n9vPs+WPaBfsp0hze3+LmE57Bb7QAYGFgM/eu0WWwMOnsQLo9LmokIIYQQQhRTqQbISqlkpVRzpVQrIAmYp5S6CfgWuCX3tFuAb0pzXMU2cKB3cyhTiDiwK/h5ZsOQjh1h716dXhFCwxAAwzB45KJHvI+/WPMFJ50nSctII6lDEtOHTsflcTGs0zDGXT6OuEZxxDWKY9zl4+jQoAMWw0LtqNoktk8kLSMNR7qDlEUpp/vOhRBCCCEqPcMvnbh0X9gwLgUeyS3zVg9IBVoC24AhSqmDhV3frVs3tXTp0rCPsyBpERcT7/wNgN8vH8UFP40u+ORnnwWlYHQh5wShlOLMt85k06FNADSIbsDLfV7mxk43EmmLLPA6s0W12+Nm+d3LST+UTuK0RFIHp0pOshBCCCFELsMwlimluuXdH9IMsmEYN+cGscGO1TUM4+biDkgp9YtSqn/u9gGlVG+l1Jm53wsNjsuDT2o+4N3u9Nv4IO30/DRsqGeQi8kwjIBFe/tO7OP2b2+n3Vvt+GL1FxT0x01C6wRSB6dywnmC+3+4X4JjIYQQQohiCDXF4hOgbQHHWucer1JmV7+OXTQGIOb4Hli0qOCTTzFABri9y+3UiqwVsG/H0R0M+3oYg6YMItuVHfS6tIw0Lmh+ATP/nsk9Xe8hoXWCpFkIIYQQQoQg1AC5gFZxANQDjpbAWCoWm40Z/uWav/664HMbNDjlALlJjSYsuXMJb1/1No9d/BgNY3wV8L7Z8A1v//F28OFZbCzYuoA6UXV48483eW3xayROSyS+afwpjUMIIYQQoqooMAfZMIxr0Q08AG4FZgF5iw9HAT2Bv5RSl4dpjEGVdQ5ybCyc8fdsZnOF3tGyJWzZAkaQvyX++gsGDYL160/7dY9lH+PBHx/k45UfA9AwpiHpD6QTbY/2nmPmICf3SOaJn5/AZrGR6cxkRLcRvNPvHRzpDtIy0ohvGk9aRpq0oRZCCCFElVRQDnJhraYbAp38HreF3JwCnxxgNvD8aY+wgrFa4Rcu5RC1qcNh2LYNli+Hrl3zn3waKRZ51YiswTv93mHO5jlsP7qdvZl76fxeZ2pE1vCNzbAycdBErmh3BSt2rWDi6onUjarLh8t188Kp66aS3CPZG0SnLEqRIFkIIYQQIleBAbJS6gPgAwDDMBzACKXU6U+BVhI2GziJ4Dv6M5yJeuePPwYPkOvUgWPHICcHIiJO+7UjbZE80fMJRswaAcA/B//Jd86fe/4kwhrBj5t+JKlDEpPXTibKGsW7S9+lR4sejF04luQeyYxdOJbUwamnPSYhhBBCiMoipBxkpVSCBMeBrFb9fREX+3Zu3Bj8ZIsF6teH/ftL7PVv63wb5zU5r8Djb/z+hrd6xZeDvySpYxJZ7izqRNVh4faFtK3T1hscS3ULIYQQQgifwlIsAhiG0RToDzRH5x77U0qpx0pyYOWdGSBvpo1v5+bNBV/QoAHs2xfQpvp0RNoiWXLnEtbtW4fT7QTArdxcMfEKDmcdJuN4Bq/2fdVbvWLu5rn0bdOXOZvnUD+6Pkt2LmF43HAJjoUQQggh8ggpQDYMYxDwJWBFt4HOyXOKAqpUgGzXHZ5DD5BLMA/ZZLPYiGsUF7BveNxw3vrjLQB+3/l7wIK9sQvHMqLbCN5b+h4RlggmrppI58adeejCh0p0XEIIIYQQFVmoZd5eRC/Ga6SUaqaUap3nq01RT1DZmDPI22iJx6yCt3MnZAevSxyOADmYu7ve7d2e/td0ftj4A6mDU3F5XCT3SGbquqm8evmr1IiswX8v+C+jHKNwpDvCPi4hhBBCiIoi1AC5BfBmRehwF24pKeBw6EV6oBfq7a/WQj9QCrZuDX6Bf4DscOj9YdChYQf6tukL6JSLfSf2kdA6gZEXj8TlcXmD5cbVG3M46zDDOg3j8bmP89ri1+j+QXfunnk3jnQHd8+8m5RFKdJcRAghhBBVTqg5yL8BscDcMI6lQoiPh8REaNHCt++v7DY0ZJt+sHkznHVW/guuvVYHyA6HfpwavsoRz1zyDHM2zwHg05WfMm3dtIDjLo+LbFc26/evJ6VvCpNWT+Lh2Q8TZY1i7b61TFw9EbvFzqhLRnkX+gkhhBBCVBUFNgoJOMkwOgKTgNeAOcDhvOcopU6U9OAKU5aNQhwOuOIKcOq1cfzU/A4u36Ebd/C//8G99+a/4JproF072LFDB8cJ4V0c1/vz3sxLn3fK119z1jX8tuM3qXIhhBBCiEqroEYhoaZYrEI3DfkE2A4cC/JVZSQkQBu/rOvjDVr7HqSnB7/giitg5UoYMSLswTHA61e8TuPqefu6BIq2+brvXdj8Qu92g+gGfPv3t4zoNsJbBUPSLIQQQghRVYSaYnE7ulKFQE8Ib9nie7zBWUQlC4cD5s6F1q3h3Xd1gBzmILlTo07sfGgnmTmZAfs9ykPs27HsydzDCdcJIq2RKKVYvGMxkdZI3B6dt2xg8Prvr1M7qra3XrLZolq67gkhhBCiMgu1UcinSqnPCvsK90DLCzOFuHdv374fNhQSIJsXvPgi1Kun0ysSE/X+MLMYFmpE1gj4qhVVi/hm8d5z2tdvj8XQt4FHebBZbdgtdhSKTGcmD89+mOQeyQAkTkskvml80NcSQgghhKgsQk2xAMAwjPaGYQw3DOMJwzAa5+5rZxhGjfAMr/xJS9Mx7pln+vbVi88TIPvndZsX9OkDR47omePUVL2/jDSO8aVerNizgix3FgBOj5MsVxZOjxO7xY5HebAaVmb9Pcu7WE/ykYUQQghR2YUUIBuGUd0wjFRgDfAh8BxgtoR7EXgmPMMrf0aO1DFuTIxvn7VxA6hWTT84elQHwnkvqFULDh/W+xIS9P4y8k6/d2gU06jQc5weJ02qN8Gt3MzbMo+EVgn0aNmjlEYohBBCCFF2Qs1Bfg24COgNLAKy/I59DzyS+1Vl+AfIbo8BzZrBxo16R0YG1K4deEGtWjpwVgoMo9TGGYzdaid1SCrjFo/jeM7xgGMHTxxk5Z6VAOw6vsu7f+q6qWTmZNK8ZnNW7l7J0I5D2bB/A23rtiW+aTyT10ymbd22kp8shBBCiAov1AD5OuABpZTDMAxrnmNbgTNKdljlX/Xqvm23G2jaNDBAbt8+8IKoKB0YZ2X5ZpvLUK8zetHrjF759qcsSuGC5hfw3rL38h37fuP3RFmjAPgj4w9i7DGMSRjDwCkDMTCYPnR62McthBBCCBFuoQbI1YADBRyrAbhLZjgVh/8MssuFnkE27dwZ/CJzFrkcBMgFGXnxSJRSHM46TLY7G6fHyXd/f+c9buYrA5xwnmDknJFUj6jO9KHTJT9ZCCGEEJVCqIv00oCbCzg2GN1pr0oJOoNsysgIflHt2oH5yeWUYRh8OfhLvh76Nd8mfUuEJSLfOd2adkOhcCs3wzoN8wbHUjNZCCGEEBVdqDPITwFzDcOYC0xF10S+2jCM/6ID5Pyf1Vdy+QJk/xnkggJk/4V6FcQvW37Bg8f72G6xY2CwNGMpdosdp8fJe8vew+VxEVs/lrELx5LcI5mURSnEN42XuslCCCGEqHBCrYO8EL1ALxJ4GzCA0UAboI9SquxqlpWRaF8TuvwzyAWlWFSQGWSTI91B4rREzmt8XsB+s25yhDWC3q1741EePlzxIY/MfoQh7YcwduFYbBab1E0WQgghRIUU6gwySqlFQE/DMKoBdYDDSqkTYRtZOWf1W6qYLwe5kswgp2WkkTpYV7swta3blpoRNQOqWKzbt45dx3dRJ6oO7y19jxs73ejtvid5yUIIIYSoaEIKkHMbgVRXSu1SSp0ETvodawIcU0odL/AJKiH/ALmyziCbqRHZ7mxm/TMLgAbRDVhw2wLvOY50B06Pk7PqncXfB/7GYliYtHqSd9b57pl3A5DUMclbCk5SL4QQQghRnoW6SO8jYEwBx55FNw+pUmx+f1rkC5B37QKPJ981FW0G2dS+ga9k3a/bfuXvA38DvhSM5B7JHDx5kK5NuuJRHgwMlu9eztVfXM3E1ROZtHoS/b/sz5S1UyT1QgghhBDlXqgBci9gVgHHvqcKLtLLN4McFQV16/p27N2b/yKzzFsF06JmC6pH+FYlxr4dy2uLXyMtI43kHsnehXlbj2zlmrOuQaFoEN2ALFcWOe4cctw5nHCeoP9Z/SX1QgghhBDlXqgBci2goHzjLHROcpWSLwcZis5DrmApFibDMDi/2fkB+56a9xQ3dLwBl8dF6uBU7/dvbviGQbGD2HdiH9G2aFweF06Pk0hrJJNWT2JEtxESHAshhBCiXAs1QP4H6FfAsauBTSUznIoj3wwyFF0LuYKmWAC8fsXrXBt7rffxSddJWr7ekmxXNmfVO4uRF48koXUCjnQHv27/leFxwznhOkGkNRKrYSXbnU2UNYp3l76LI90h9ZKFEEIIUW6FWsXiLeA9wzBygE+BXUAT4Bbg38CIsIyuHMuXgwxFd9OroDPIAJ0adWJG0gzmb5nPpZ9d6t0/6pdRvLr4VYZ2GMqe43uYvXk2Q9oP4ZsN3xBjj0GhMDBwu91ku7O5uNHFAa2pHekOWbAnhBBCiHIlpABZKfWBYRiNgGTgIb9DWcBTSqkPwjG48ixoioX/DPK6dfkvqsAzyKZLWl3irUhhOpp9lA+W+26BCasm0Lp2a96++m2+Wf8NALH1Yxk5ZyTz0udRzV6NmzrdBEDitERSB6eW7psQQgghhChEkQGyYRgW9Gzx6+iZ5AuBesABYLFSqmJOiZ6moCkW8X6VGT76CJ54Aho18u2rwDPI/j659hOubnc16YfT+fzPz9l0KH+GTfrhdMYuHMuc4XOIskUBcPDkQV749QVOOk9SzV7NGxxLTrIQQgghyhNDKVX4CYZhQ9c9HqCU+rFURhWCbt26qaVLl5bZ66enQ5s2evuMM2DLFnRpty5dYNUq34mPPQZPPaV7U2/ZApdcAlu3lsGIw+Ok8ySz/pnF3kxdtWNZxjI+Xvmx9/j151xP6pBU5m+ZT+K0RO7peg8vLXwJl3LxdK+nGZNQUPVAIYQQQojwMgxjmVKqW979Rc4gK6VchmFsBaKLOrcqCZqDbLHA88/DNdf4Dr78su5LPWpUhS3zVphq9moMbj84YN/Z9c9m5FydU/zVX1/x3ILnePuPt0kdnMrkNZOxW+24XC5eW/waCa307HFaRpo0EBFCCCFEuRBqFYuXgScNw2gQzsFUJEFzkAH694c+fQJPnjcPUlJg2TI4ftzXRMTh0PsrmUcueoS7zrvL+/iL1V9484wnr51MhDWCBtENcLqd9P+yP4OmDJIGIkIIIYQoN0INkC9H5yFvMQxjgWEYUw3DSPX7mhLGMZZLQXOQAQwDvv0WPvRrLrhsGXTtCjfcABERcOyYDo4TEwPzlisJwzBI7pnsffzPgX9oV7cdaRlpzBg6g+lDp3PCeYIWtVpw0nmStnXaSgMRIYQQQpQboQbI9YENwB+AO/dxA7+vhmEZXTkWNMXCVK0a3H67b4He8eO6BFxqKmRn65zkxET9OKFyBoStareid+veACgUn6781FsrOaF1AneedyebDm1CoVi+e7k0EBFCCCFEuRFSgKyUSijqK9wDLW8KTLEwGQZ088v5XrpUB8PnnANvvw0jRlTa4Nh0R5c7vNsfrvgQl0f/oBzpDiatnsTwuOEYGNSNquttICKEEEIIUdZCbRTiZRiGgU632KuUChYaVnopKdChg+/xyZM69di/T0hSEhzfHc8AZgGw9L00uv36qy5/Ubu2fpLatWHDBli5EoYOhSm5mSqFbW/YAG3b6tSMyZP19sjyuaht0DmDqFetHgdOHmDbkW18s/4b6larS+K0RJJ7JPO042nu6noX45eNZ1inYfT/sj9xDeOIaxQHwMrdKxnacShT1uj3H+r2hv0baFu3LfFN45m8ZjJt67aVhX9CCCGECFmRZd68JxrG1cAzQGd0YB2vlFpuGMYHwHyl1MSwjTKIsizzZqYP79+vHxsGKKXTiy0W/QXQj1mknugPwImmbYk+slsfcDr1RdnZYLf79vlvR0bqc7KyIErXESYrC2JiYMwYGD1aH58+vVzPRD/585O8uPBFAC5ofgFxDeOIrR/Lhv0bsFqsfLH6C2pG1mTX8V2c3/R8ftvxGxHWCCxYUChvi2qALHdWSNsx9hjGJIxh9PzR3o59kr4hhBBCiLwKKvMWUoBsGMbNwMfAJGAe8AnQLTdAfhS4urTTLMq6DvLs2XDFFYWf8+/Be3h7WuNTfxGbTedv5A2irVYdQJfz4Bhg59GdtHqjlTe94lTYLfr9Oz3OkLetFiuR1kgJjoUQQghRoIIC5FAX6T0JvKKUugXIO1O8Fmh/muOrcPr21b0/CvO/aY3YVy/21F/E5dK1k51O/WVuZ2XB/feX++AYoFnNZiR2SDzl6+0WO06PE6fHSa3IWiFvZ7myuL/7/RIcCyGEEKLYQs1BPgOYU8CxLKBmyQyn4vjlFz2R26wZ7NwJderAoUOB5wwcCDc5PmPixS/RwH5Y7zx0CNatg7p1Yc8enSYBOkfDzNUwWSy6sUhkpN5vbmdnw5tv6gC5AgTJr1/xOjaLjW1HtuU7dujkIdbtW0fdanXZk7mHOlF1OJTl+0E6PU4irZEopTiSfSTk7Wx3Nm8seYOEVgkSJAshhBCiWEKdQd4OdCngWDdgY8kMp2Iwc5BHjdKx6ogROu41MyFMc+fCFaO6037DdByjHPqCnTvhpZf0yr6YGF0SLjJSXxAZGfgkZrBsGL7EZpvNFzAPGqQHU841iGnAZwM/w3GLI+BrVK9R7Dy2k5f6vIRbuRnRbQSHsg4RZYsKuF4phcXQ79/AKHLbZrHRsmZL3B43g6YMkuoYQgghhCiWUGeQPwKeMQxjDzAjd59hGEZvYCQwJgxjK7fS0nQJY//v/frpOPa77/Q5MTHQq5fOkjDPScDvgqQk/TV5cmAVC6cTVqzwvdjLL8NXX+lts4rFjBk6yP7999wnrpgzpGkZaaQOTg343u/MfjSr0Yxv//6W3cf1osaWtVpSt1rdYlWxWH9gPZe1vowIawRpGWkyiyyEEEKIkIW6SM8A3gbuQTcKsQFOwAqMV0r9O5yDDKasF+kFs2IFnHee3q5VCz74AIYMOYUnatNGl4MDHQB3y5M73qcPPPaYToSupJ6e9zTP//o8APfF38dbV78V8rUpi1LIceewYOsCZg+fDejay2kZaVLuTQghhBBeBS3SC2kGWeko+t+GYfwf0BuoBxwE5iml/i7RkVZgder4tl0uOHz4FJ+oe3dfgLxkSf4AuWlTnapRifVo2cO7/Xba2/Rt25eakTWJbxpPTERModfGN42n/xf9cXlcvLTwJRZuW8iv237lmUueIWVRitRHFkIIIUShQs1BBkAptVEpNV4p9aJS6j0JjgPVru3bzsnJv2gvZBdc4NuePj1/q76mTSEj4xSfvGK4sMWF3pxigGsnX0vCZwmc/+H5HMk6Uui1Ca0TeO6y58jx5PD5n58z659Z5LhzGD1/NDaLjYFTBjJl7RTim8aH+20IIYQQogIKOUA2DCPCMIy7DMP40DCMWbnf/2UYRkQ4B1iR1KzpK0rhdMLBg6f4RJde6tv++WcYMEAnN3s8el8VCJBrRtbkts635du/bt86hk8fzqy/ZzHr71ms3L0y6PUPXfgQ/c/qz1/7/6JmRE2yXFlk5mTy+NzHpXmIEEIIIQoVUoBsGMY5wD/A/4CO6DzkjrmPNxqGUeXqIAdjsejcY9Pevaf4ROeeC/fd53v84486SP7vf/XjKhAgA3ww4AMctzj413n/okXNFt79M/+eSf8v+9P/y/50Gd+F//3xv3zXOtId/L7jdy5odgFHc47SrEYz3MqN0+NkRLcREhwLIYQQokChziC/DxwB2iqlLlBKXaOUugBol7v/vXANsKLxT7PYt+80nujNN+GBBwL3vfeejrqbNasSAbJhGFza6lLeH/A+2/67jfvi7wt6XvLPyd6KF6CD48RpiST3SGbjoY30bdOXncd2YrfYsWDhzT/elNJvQgghhChQqFUsTgI3KKVmBDk2CPhCKVWt5IdXsPJYxQKgSxddtQ3gootg0aLTfMKlS/WiPTO94vnn4aaboGdP2Ja/8UZlluPOYcz8MazYrcvgrdi1gl3Hd3mP3xd/H//p/h9mrJ+BzWJj7MKxJPdIZvT80eS4c3C5XTSr2YwdR3dgt9jp3aY3l7W+jA37N9C2bltZvCeEEEJUMQVVsQg1QP4LGK2Umhzk2A3As0qp0+ipXHzlNUC+7DJf745zztFN807bpEk6KDb17atfJDvb10CkCvpp409cOenKgH0R1gi6N+vOjqM7qBFZg2PZx/AoD6MvHc2BkwdInptMjieH7k27s3z3cpweJzH2GMYkjGH0/NGSnyyEEEJUIacbIF8LjAOGKaWW+O2/AJgAPBpsdjmcymuAfN11uvAElGA1tpwcOOMM2O1LI8Aw4OGH4eqr9ePJk6FtW4iP17WTR1aNGdCh04aSuja1yPNi68Wy8p6VLN6+mD4T+gC6Q59C3/8GBtH2aIZ1GqbPrx9baCMSmXEWQgghKr7TDZDTgDPQ9Y/35n41zP06AGzxP18pdf7pD7lw5TVAvv12+OQTvV2zJhwpvCJZ6N57T/e09hcVpQNlw9AtqkeNgrFjdbe+Ctpdr7hy3Dk40h1sO7KNj1Z8xJKdSwo8d1inYVzc4mKmrZvGvC3zAOjYoCNr9q0BdIvql/u8zCjHKDKdmURZdcvrLHdWwLbdYufO8+5k0upJ3hlnQBqRCCGEEBXM6QbInxTnxZRS+etzlbDyGiA//DC89pretlp1uTez9Ntp+/tveOcdeOON/MeGDNFpF1UoOM7Lozys3rOaw1mHvft+2fILz85/ttDrIq2ReJQHp8eJzWJDKYVbuQvcthgWPMpDjD2GmTfMBCBxWiKpg1MlNUMIIYSoQE63k17YA97Kwr+KhWFAVhZUK6nli2edBa+/Djt2wFdfBR6bOhWefrrKBscAFsPCuY3PDdjX84yezNsyjwVbFxR63Yu9X+SxOY/h8uimLE2qN/EuACxo+4TzBFPWTuGrv76S4FgIIYSoREIKkEXo/APkiAjdbrrEAmTT7bfrxiHZ2YH7X35ZD6BLF52HHB/vy02uIjnJeVkMC18nfs1ri1/j5/SfOZp9lHZ12/HHzj/Yk7kHgNa1W7Nh/wYibZHgAoVi1/FdRFojUSr4tsWwkO3KZvyy8dzY8UbSMtKYvGZyobnLoW53btyZpI5J3vzm+Kbxkr4hhBBClKKQUizKo/KaYjFxIgwfrrdr1oTFi6F9SbZRcThg4EA9PX3rrcHTLSwWeOopPdtsGHrVYBWeWQ7m580/exfr+TMwAhbuRVgjyHZnB+Qgx9hjuPncm3l36bve60Z0G8Hnf35eaO5yqNtWixWFwm6xM+qSUYxdOFZmqIUQQogwKCjFourWCAsT/xlki0XPIJeotDRIStJBb9Om0KFD/nM8HhgzRidADx2q9zkccPfdkJKit1NSSnhgFctlrS8jrlFcvv1mcGxuR9ujGXf5OOIaxRHXKI5xl48joVUCn//5OeMuH0fvVr0BeHfpu2Q6MwEd6Ga5s055O9OZyQnnCTo27OgNjgFSFlXt35kQQghRWiRALmH+ATKEIUAeORLGj9czwiNHwqefQpMmwc89eRJWrYIrr9Tl4KZMAZsNEhN1+kUVZhgGb131Fg2iGxR63qGsQyzNWErnxp25sMWFdG/WnZ5n9GTmDTN56MKHmHvLXJ7s+aT3/ItbXFwi21bDyqLti7g57mZALwKMb1q1f2dCCCFEaZEUixK2di107Ki3a9bURSeGDQvzi86bp2eKL78cvvii4PN694Y//6zSlS7yUkp5F+b5u+e7e/h45cdBrxnaYSgpfVNoWasljnQHA6cMxOVx4XK7yPHkePOVT3fbyP1fjcga3uYljnSH5CMLIYQQJeS0qliU4CCigAVAZO5rT1NKPWMYRl1gCtAKXVM5USl1qDTHVlL8Z5BdLt8MssMRpv4dDocOjpOTYfRoiI7WM8fB/vD5+ecqX+kiL8MwsFvt+fa/cvkr/LTpJ3Yey9/pZcraKXy74Vs6NOzA6j2ribRF8lzCc4xyjPIFtrm1/U5l22a1YbPadHtsj4uj2UeZvGYy4xaP49dtv9KzZU++WvdVSIv9ADKOZdDzjJ7exX6y6E8IIYQoXKnOIBs6CohRSh03DMMOLAQeAK4DDiqlXjIM43GgjlLqscKeq7zOIGdmQvXqvsdDhsBdd8ENN4Rp4jYlxdc9b9MmnZ984AC8/bZeIeh0BgbL9evLDHKIMo5l8NPGn8hx5wAwb8u8oF37mtdsTmy9WA6dPET96Pqs2asbj3Ru3Jkdx3YQZY0qdhWL2PqxPDnvSRrHNGbLkS3UjKzJ0eyjRNmidB1mjyekxX4uj4scdw73dLuHqeumktwjWRb9CSGEELlOq1FIOBiGEY0OkEcAnwOXKqV2GYbRBPhFKRVb2PXlNUBWCiIjdVxqql4dvv22FGNS/0oXTzwBjz/uC5KvukoH0xIkn5Jft/7Kgz89yPJdy0M6P8oWxbdJ39K3bd+QX8OR7ghoPNLj4x4s2r6IGHsMmc5MrIaucuFRnpC260bV5WDWQYZ1GsZPm37yLvqTWWQhhBBVXbmpYmEYhtUwjJXodtVzlFJLgEZKqV0Aud8bFnDtXYZhLDUMY+m+fftKbczFYRhw8cWB+1q0KOVY1L/SBcC//+07NncuvP++PkcUW88zerL0X0tZe+9axl0+jhh7TKHnZ7myGPDlAAanDubnzT+H9BppGWne4NiR7mDDgQ2cU/8cMp2ZNIxuiFu58SgPLWu2DGn7YNZBDAwmrZ7EiG66Xbks+hNCCCEKVpYzyLWB6cB/gIVKqdp+xw4ppeoUdn15nUEG2LNHp1Q4HPqx1Qpz5pThhG1Oji4Jd+CAfnzPPfDuu3qAVbyRyOnacngL09ZN43jO8YD9Sik+WfkJ249u9+4zMEjpm8LDFz7szTkujDmTbKZFdGnchTmb52C32DEwQlrs5/Q46dOmD3M2zwHAgoVq9mrMvGGmNwA3G5LIbLIQQoiqplws0vOnlDpsGMYvwJXAHsMwmvilWOwtq3GVhHXrdHU1k9WqK6uVWVZDRAQ89BA8mVuO7P33oVcvHSibjUTEKWlVuxWPXPRI0GO3dL6FQVMGsWqPvhkUikfnPEqOO4cnej5R5HOnZaR5g+PkHsmMnj+aKFsUbo8bi6E//ClqsZ9hGMzdPJcR3Ubw4fIPcXqcZDozWbF7BQADpwzEwGD6ULkHhBBCCFNpL9JrADhzg+NqwGzgZeAS4IDfIr26SqlCp7PK8wyyuW5u4EA4elTv+/JL2LatDCdqldIt/dav14+tVp0cLV32wkopxZq9a7j3+3tZuG0hoIPX2cNn06dN/k5+eaUsSvFWndh0cJO3BfXK3SuLVcWiaY2mTFk7hShbFHsy9+gA2mIj2h4tJeSEEEJUWeVikZ5hGHHAZ4AVnf+cqpQaYxhGPSAVaAlsA4YopQ4W9lzlOUA2deoEa3RBA+bMgT5Fx0Ph9cMPumGI6eGH4dVXy248VUi2K5srJl7B/K3zAYi0RvKv8/7Fq5e/SqQtMqyv7b/o7+KWF9PxnY78c/AfAHq26MmC2xcEpHO4PC4JkoUQQlQJ5WKRnlJqlVKqi1IqTinVUSk1Jnf/AaVUb6XUmbnfCw2OK4o2bXzbf/9dduPwiozU/a9N//ufL1FahFWkLZLJgyfTuHpjALLd2byd9jYP/fRQ2F/bf9Hfom2L2JO5h0irDsp/3f4rPT/uGZDrLIv3hBBCVHXSajqMzjzTt13mAbLDAYMG6SDZlJ2t90mQXCoaV2/MTzf9RLemvj9UP1/1OZk5mWF93ZEXj/SmUJg5xz8M+4Gfh/+MBQsLty/kRM4JRs8fHVA94+6Zd5OyKCWsYxNCCCHKozJbpFcVxPpVct6woezGAfhKv117rU6ONhuItGih6yQPHQpTdA4rQ4fqAbdtq5OppdJFiYlrFMcfd/5B+3fas37/eo7nHOfbDd9yQ6cbwv7aaRlpJHVIIqljkjcIrhFZA4XiaPZRrG6rt2OfY4sDj8dD7za9sVlsBeY6b9i/gbZ12xLfNF6qYQghhKg0yqzM2+mqCDnIv/0WWBP5zz8hLq7sxuPVp49uOw3QuTP89ZeeTbbntlx2OvVMc1ISTJ0KNhvMmCGL+UrQCwte4CnHUwBc3OJiHroweKpFzciaJLRKwGqxlujr5y0h17Z2W5ZkLAk4Z0S3EXz+5+dkOjML7NgXY49hTMIYRs8f7a2GIR36hBBCVBTlYpFeSaoIAfLRo1CrVuC+L77QNZLLVN7FekWpUwdWrIDNm/VMtMwkn7b0Q+m0ebNN0ScCQ9oPYcrgKSHVTg5VyqIUbBZbQNvpSz+91LuI0IIFm9WGR3lweVzYLXYUKt+2xbBgNaxSDUMIIUSFVC4W6VU1NWvqMsP+zCyGMtW3LzRpEvr5hw7pa665RndBmTUrsJe2KLbWdVpzaatLQzp36rqp9PykJ1//9TUnnSdL5PVHXjwSl8cVkHO8YvcKIq2R2Cw2MCDHnYPL46JetXo4Pc6g2x7lwelxcmHzC73PI136hBBCVHQygxxmNWtC69a+xiFnnlkOFuyB7qT3wAPg8egvpXSFC6V82x5Pwdc/+yw880ypDbcySj+UztiFY9l/Yn/Q47uO7+L3Hb8H7GvfoD3fJn1L27ptS2wc/ov3pg+dzordK3h49sMA2C12nB5noR37nB4nHuWhU8NObD2ylRlDZwA659ms4SyzyUIIIcojSbEoI82b6yIRsbG+uDMzE6KiynBQDodeqGcYMGqU/vIfVFYWxMTApZfq2eJgOnTwFXkWYeF0O+n9eW9+3fZrwP760fX5484/aF2ndYm8TsqiFG8TEtDd9dweN50admLZrmU4Pc5Cc5BvPvdm3l36LgBWw8pdXe9i6rqpAV0ApbayEEKI8kgC5DISGwvffAP9+ukUXigHi/VSUmDTJr0ILy1Nb8fG5q9ikZGhA2SldOe9bt1gSe5CLosFjhzR3fhE2BzOOsyrv73K6r2rmblhJgr9/9erz7ya7274rkTzkiEwWDa798XWjy20igXApNWTyHT6ytX1atmLdfvXeYNkM5VDCCGEKE8kQC4j550HH3ygMxK++07v+/JLHZuWaw4HJCZCcjI8/TQ0baoD6UaNYPdufc78+dCrV9mOswpxpDvo/Xlvb5Dc78x+RNujiYmI4YaON3B528vLZExml75pf03jnbR3vMfqRNXBarFKcCyEEKLckkV6ZSQmRmcvtG/v2/fXX2U3npClpUFqKjz0kI7mN26Em27SOSP+54hSk9A6gbu73u19POufWUxdN5VPV37KFROvIOGzBD5b+RlOd+ktoDS79AGkrk1leNxwDAwaxzTmUNYhLmp+kQTHQgghKhyZQQ6zK66A//5XF3+49Va9b/BgXV64QnA4YMgQOH4catSAYcPgjTf0saFDdRMRUWoOnTxE5/Gd2XZkW4Hn9G7dm59u+qnEaycXJG9N5SHth/Du0neJtkdzwnmCfmf2o1mNZkFTNTo37gzAgq0L6HWG/jRi5e6VAceTOiZ5m5DIoj8hhBAlSVIsysh11+mYsmVLOP98ve/ss2HxYp3WW6NG2Y6vUGaaRWqqfiOffAK33QaHD+vjdevCrl0QEVGmw6xq9p/Yz4KtC3B5XADM3TyXj1Z8hEf5qo48n/A8/z7/397HdoudmIiYsIwnZVGKN3A1aysPaT+ED5d/iMvjwmaxYTWs+Rb4RVmjsFqsuDwust3ZRNmisBgWPB5PwHGFwm6xM+qSUZLPLIQQokRJgFxGhg/XJYQHDdIl3/JKTNSTsCW81qpkpKToVtMJCfr7W2/p7idXXw1ut++8r7/Wb1CUmW1HtvHiry8yftn4As+5tfOtfHzNxyW+sM+fGSwntE5g7ua59J3QlwhrBC6PC4/yBC0X5/Q4uaTVJfyy5Zd8x10eF27lZnjccH7Y+IM3nUNmkYUQQpQECZDLyD33wLnnwogR0K6dXueW19q1gTnK5dINN0D//no6/KKL9BS46dxzYeXKMhua0FweFxd9dBFpGQXnhr9+xes8cMEDpTamiz+6mN92/AboEnBu5S5wu6DjpvMan8ewuGHeWWTAm3ohwbIQQohTIYv0yoi5SA/g1VehTRvdftrqlx66Y0fZjK1Y2rXTC/UAxo6FZs18x1atgoMHy2ZcwstmsfHl9V/StUlXakXW8n7F2H2pFY/OeZTlu5aXyngc6Q7W7FtDtD2aCEsEbuUm0hqZb9ujPPRt07fA46blu5fz8OyHSe6RDOh6zVPWTpGufUIIIUqcBMhh5h8gDxyoZ5APH9br20xm1bRyrW1bX4B8ySU6qu/eXT9WCn79teBrRalpW7ctS+9ayuHHD3u/Dow8QNcmXQFwepzcPP1msl3ZYR2Hf3e+5xKew261A2BgYDEs3m271U6ENYI5m+cQZdM5x3mPR9ujsVvs3ud+ZPYjXD7hcm/nP7PFdcqilLC+JyGEEFWHBMhh5h8g+2vc2LddIQLkYPkhl1zi2/7ll1IdjghdpC2SyYMnE22PBmDtvrVEvRDFzdNvZs/xPWF5zbSMNJI6JDF96HRcHhfDOg1j3OXjiGsUR1yjOO/2sE7DuOXcWzi73tncHHczN3W6Kd/x5xKew2qxcukZl2IzbCgULuXivCbneYPjxGmJMpMshBCixNjKegCVXUwMbAtSkatCBcgpKYEzyKArXOzb53s8f37pj0uErF3ddqT0SeG+H+7z7puwagJ/7f+L+bfO9wbPJcU/J9i/4sRDFz4UdDuYhy58yBv8fn/j9wAM+HIALqcLAwPHFgft/9eezYc282LvFwHfIkHJTRZCCHE6JEAOs0oxgxwfr8ttHD+uq1gsW6Yff/IJfP65rmixYgVER8O//637alv8PpyIjtYtBS3ygUVZGhE/gu83fs/3/3zv3bc0YylXT7qaRy96lBqRhdccrBNVh06NOoV7mAH8G5EMnDIQm8XGuMvH4Uh38N0/3/HX/r+wWWyMcozCarHyzCXPeFM7pg+dXqpjFUIIUXlIFYswmzIFpk3L3xhkzhy4PLczcEICzJtX+mMrlrvvhg8/1E1Dfv5Z10YGuOUW2L696OsHDdI/DLu96HNF2DjdTpbvWs74ZeP5ZOUnxb7+we4P8toVr4W1VFwwKYtS2HRwE0kdk7xpFQO+HMAJ5wlv620zd7marRrDOg0jqaPu5y5NRoQQQhSkoCoWMoMcZpViBhl0u+nPPtNBbpIOPEhM1K2oX3gh+Jv0N326Lgjdtq1vX40aur1g587hGrXIw2610715d7o3706zGs14/tfni3X960teZ+OhjTSMbohhGFzR9gqGdBgSptH6+Ae1ZtrFzBtmMi99nvc9mI1SnB4nsfVjuXrS1SgUBga3dbmNV357heQeyQGNTSRYFkIIEYzMIIeZwwHPPps/RXffPmjYUG/XqVMBqqQ5HLoMx/Hj4PHoyH/mTH3MTLdwuXQQfeCA77rjx3X6RUHsdrj/fmjQQD+Oi4MrryynnVMqn3X71vHmkjdZt29doeftP7Gfv/b/FfTYb7f/xoUtLgzH8IIyA1yAxGmJXNXuKiasmoDN0H/vu5QLi2FBKYVCcXa9s1l/YD3XnHUNc9PnEtcwjnX719GzZU+a1WgmrayFEKIKk0YhZeSPP+DeeyHvUD0e3aHZbEiXlQWRkaU/vpD4t5z+4Qd45RW9f+hQX7pFQgGtf5XS3VLefz/015sxA6699rSHLUpOjjuHaydfy48bf8x37NbOt/LJtcVP1zgd5ixyco9kRs8fjdvj9uYgPz73cZweJwCNYxqzO3M3NSNqcjTnKO3rt2fd/nXettYGhrSyFkKIKkwC5DKydq1O210XZIKuWTPIyNDbW7dCy5alO7aQmS2nQQfK99yj0yqUgqefhjFjCr9eKR1Y79oVuG/8+Px/OYDuz/355yU3flEinG4n3274lsNZh9l3Yh/JP+uGHdH2aHY/vLvIRX4lyT9NwsxNBp1vPGn1JLJd2SgUHuXhrHpnseHABm+QbH63GBZQ4MHDDR1vYM7mORIcCyFEFSMBchnZskWXC966Nf+xrl1heW5TsyVL4PzzS3VoxeM/iwxw9dXgdEL16jq/uKAZ5MJkZ8OkSbBhg07L+Ogjvb9Nm+A9uUW5oZQi7r041uxd4933fv/3+VfXf5XZmPybk4y6ZBRj5o8hy5VFjjuH85qcx7Jdyzij1hlsPbKVZjWasfPYTgCq26tz3Hmcp3s9zZiEIv7YE0IIUalIq+kyUtAiPahgC/XS0nzBcWIijBsHrVrpNIvERB1AF1dkJNx+O7z8MrzzDkRF6f2bN1eAH0jVZhgGt3e+PWDfXd/dxU8bfyqjEeVvTjLqklFE2aLo3qw76/frHOStR7bStUlXdh7bid1iJ8ISwXHncRrGNOTdpe/iSD+F+1gIIUSlIwFymFWaAHnkSD1LbAbKd96pB/3KK/pxWtrpPX9EhC+NA2DRotN7PhF2N597M41iGgXse2fpO2U0Gl3pYvyA8SS0TmDkxSNxeVxMHzqdQecMYkzCGH7b8Rsjuo1g7d61RNn0H2N2q51oWzR7M/fyYPcHSZyWKEGyEEIIKfMWTikpuvlcVha89BKkp0NsrM4oaNtW7zeV+wDZNHKkfmOgy3AkJcGjj+qUiO7d9YzylCn6eHG2O3fWTUhMkyfr54yP19tt2+rXFuVGveh6rLxnJR+v+Jgn5z0JwHd/f8eOoztoXrN5GY/OVxouoXUCKYtSSB2cSlpGGjefe7O3csXK3SsZ2nEoo38ZzWd/fkZyj2Qen/s4AEM7DmXKmil0btwZgAVbF9DrjF4A3uumrJkScG4o250bd5bKGUIIUc5JDnIYmZXRjh6FF1/0lQuOidHr2p580hckjxihswwqBDMf+dxzYeFCsNn0ojuPR78hM1WiONtWK+Tk6LxmU506cPiw3m7UCKpV09tt2uiyci1ahP2titD0/rw389J1t5uGMQ2Jscec8nO1q9uOTwd+StMaTUtqeAUyq2Fc1Pwiftj4AxHWCL24z+Mhy51FlDUKq8WKy+Mi253trX7hfxwo1rbVYpXKGUIIUU7IIr0y4nBA79562/9HbbPprIITJ/TjgQP1WrcKw4z+MzN9tepK06236iBZlAtT1kwh6aukEnu+AWcN4Nsbvi2x5yuMI93B9anXcyTrCB48BZ4XZY0iy51V4PHi6ndmP5bsXCLBsRBClCFZpFdGEhJ0szkzODbTbF0uPQlr2rGj9Md2WhISdIMPtzswd7iktovyzTf6hyjKhUHnDOLiFheX2PPN/HsmD/zwQECVjHBJaJ3AfeffhwePtwEJkG87y51V6PHibDeKacSsf2ZxVburvK2zUxaleL8LIYQoWzKDHGbmRKvLpb9ycvTMcU6O7rR87Jg+r3HjwDLB5Z6ZZnHVVTBhgq5IoZR+Y6e67XRCnz4wZ46eYjcDYLOE3OLFgYnb/fpBr16Sp1xOKKXYfnQ7Ls+p/+Eyev5oPv8zsAb2Za0vo0eLHgH7mtVsxo2dbqR6RPVTfi2TmWZhduSLtEailCLHk+Pddnqc9GnThzmb5wQ9Xpxtt8etu/1hwYPH2+HvlnNvYeq6qST3SMblcUlushBClAJJsSgDZnBsGDBqlP4yc5BdLh0omwGyYejYLyKiTIccGjM4Tk6G0aP1LHJJ5CC7XLo2clSU/kFcdpnuquevQwfdfQV0EP3yy3oMhnHq9ZhFubE3cy9nvXUWR7KPFHluq9qt+Piaj08rPSFYR77SyEE2n6+atRon3SdpXqM5O47tkHbYQghRyiRALgMpKboQQ1KSroK2aZOvisWsWTq+fPRROHlSn5+erksLl3tmZz3zTSUl6RnclStPr4oFwIIFelY4KUn36Z4+XXdRAR0QW606iDZVq6aDaQmOK40FWxfwxpI32H18N0t2LMGtCs5xt1vsOG5xcHHLU0vvCNaRz7+6RTirWDz444Os2ruKWpG1OJJ9BJthw6VcnNf4PJbvXi7tsIUQohRIgFzO3HCDbkb39ts6DgT49Vfo0aPw66oU//wUc3Y6mFDaXYsKaePBjUz/azonnCe8+7JcWYxfNp5DWYcAnc+79K6l5aK0XKj80zomrppIx4YdWb13tTdIbhDdgH0n9mEzbCgUbuVmcPvB/LLlFwmOhRCiBEmAXM6MGqW/r10LX3+tt7/4QgfOgsD8lOnTYcUKePhhfcwwAkuC1KolM8hVTPqhdM7/8Hz2n9gP6AVwC25b4G0AUp75p3WMXTiWIe2H8N7S97ztsJtUb8Ku47u8QTLgnWWWdthCCFGypIpFOXPWWfDPP4GlfCtcJYtwSkvTaRZm7bvRo3Xy9gUX6DQLfx4PDBp0au2uRYXUuk5rpg6ZitXQ90JaRhrVXqjGdVOuY9PBTWU8usKlZaSROjgVl8dFco9kpq6byj3d7vG2w951fBddm3Rl34l93nbYR7KPUCuylrTDFkKIUiIzyGUgJQWqV4ePP/Y1ogO4+GLdd0PkESyZ++ef9XeA/v2haVOpYlEFvbXkLe7/8f6AfZe1voyfb/65jEZUPP450DaLzTuj/MmKT8AAt0fnX1sMC9nubDo36szmw5t55pJnvJUuzMV7smhPCCGKT1IsyhGHA4YM0YvzPvwQbrxR7+/RQ+chixBMm6Z/iKCrXmzbBg0alO2YRKlTSnHvrHt5b9l7Afs33LeBs+qdVUajOjUFLRjMOJbBr9t+xW6xczj7MDaLDRS80PsFRs8fjYHB9KHTJS9ZCCFOgQTI5YzZYS8x0VfQIT7et2BPFMHt1q2uzZJvV10F3bvr9Isrr4Ru+e51UUkppVizdw2XfX6ZNyf5kQsf4ZXLXynjkZUcR7qDgVMGYjNsHMw6iMXQ2XHVbNUY1mkYSR11F0MpBSeEEMUjAXI51LRpYHOQCtcspKzdcgt8/nn+/Yah6yXfdpuvhJxZiq5tW1+JOknHqFS+/+d7+n3RD4B61eqx+I7FnFnvzDIeVclJnpvMS4teolFMI/Zk7gHAalh5uc/LPDXvKRQKA4PbutzGZ39+RlzDOOIaxQHkK1sndZWFEEKTALmccTj0ROeAAfDVV3qfYcCqVfnXoOVVty40ahT+MZZ78+ZB3756kV5eZqULu13XT1ZKf7/3XvjgAx049+1b+mMWYeP2uGnzZhu2HdkG6NnVkRePZEj7Id4Z11NRzV6NM2qdgWEYJTXUYius25+/+tXqc+DkAQacNYBv//42aGMTq8UqdZWFECKXBMjliNmI7sILdZD89NNw8GDo1xsGvPoqPPRQ+MZYYUydqn+YNpsOlIMFy8GccYbOZ2nYMLzjE6VqxvoZDJk65LTaXQdzQ8cbmHTdpDIJkoN1+7NarDxzyTMk/5xMjlsHye3rt2fd/nU0imnE3sy9XNj8Qn7b8RsR1ghvu2xz2+VxcVPcTfyw8QcJjoUQVZqUeStH0tIgNVV/0p+R4WsiFyql4I03wjK0imfIEGjf3tdMxGYL7bqtW+H668M7NlHqBp49kD/u/MObWlBSvlzzJV//9XWJPmeo/MvCJXVIYuYNM5kxdAYb9m/AbrFjM2xYDSt/7f+Ls+udzZ7MPdgtdn7b8RvVbNXIcefg9DgDtgEmrJrAiG4jJDgWQoggZAa5DH30ESxaBCNG6HTY3buLvmbDBl+PjGPHdLm4Ks2/257LBTk5OkfFndueOG9TEf/Hdjt8+aUuFyf5yJWK2+Nm0upJjF82noMni/HxTB5Hs4+ScSwDgGY1mvFN0je0q9uOGpE1Titt43SZi/YMDEZdMoox88eQ5coix53jbTjStUlXlu1aFpCO4b99frPz2Xxos8wgCyGqNEmxKId++EHPBP/4Y+jXnH22DpIBli6Frl3DM7YKwb/b3qhR+iszU5d9A92aOipKB8wuF2RnBwbPANWqwaxZ0oVPBHXo5CHO/t/Z7M3cG7C/cfXGzL5pNp0adSqTcaUsSvGWgjNrKI+ZP4Zz6p/D6r2r6d26d6E5yE63E4XipT4vSQ6yEKJKkwC5HFq5Em6+WS/MC9WgQTBjht6eOBGGDQvHyCqIYA1EYmN9dfOGDg2sYnHPPfD334HPERGhe3xv2iTVLURQU9dOZei0oSgC/1tZnhqS+NdQNr8v2LqAZjWaAYFVLAD6tu3Lywtf5rwm5wXsz7u9Yf8Gb6ULaUgihKiMJEAuh/bu1emz+/eHfs0TT8DYsXr7ySfh+efDM7ZKx1wZ2bu3L4A2Vaumf5Bjx+rkcJlNFnks3r6YN/94k6lrp+JWvk8gvr/xe7o37x70mhoRNbBb7aU1xJCZi/7cHrd3JlmhAmaZAbLcWcTYYxiTMEYakgghKi0JkMshj0dnABw7BpGRoV3z+ee6/C/Addf5SsSJQpjBcWqqfjxggE7FMBkGREfDzJn6cVqazCaLAt3xzR18vPLjIs+rHlGdrxO/pm/b8ldO0JHu4KpJV9GxYcdC85Sthq45GWWLkoYkQohKSapYlEMWS/Gbg5xzjm97/fqSH1OlZJYNAR0oz5wJV1zhO66UDphffFFXxbDZ9Hnx8WUzXlGuPdXrKd3uuQjHc45z2ze3cSz7WCmMqngSWicwIHYAy3Ytw2pYyXZnewNi/223cuNWbpweJ7H1Yxk0ZRD9v+zPlLVTsFlsJE5LJL6p/P9ECFH5SIBcxpo21aXeQnX22b7tf/7Ra89EEUaO1GkT/oHysmW6PbW/uXN1QWr/PJaUlNIdqyj3WtdpzZtXvkmzGs2oE1Un6JcZQO88tpMn5z1JefukzpHu4JctvzA8bjhu5SbSGkmEJSLottWwkuPOYeSckRzLOcYJ5wmuaHuFLO4TQlRqEiCXsbztpotSowY0b663nU69tkyEyEyXSEyE5GQdMD/xhE6xMCmlq13cfbeeZV64EF5/XQJlEWBE/Ah2PLSDg48dDPr1ybWfeM9964+36P15b9buXVuGI/bxbzzyzYZviLHHYLVYvWXrDAzvts1iI6VvCjbDhlu58SgPtSJrkbouVWooCyEqNclBLiMpKfoT/Cef1PGZudjOTH+dPBnatg2eAtu3r57sBN1IbvDg0ht3hWf+4M0fNOjSIHXqwJYtBV9XvbpOGK9XDw4cgJo14cYb9V8skqcs8lBKceWkK5m9aXbA/o4NOxJbL5anej1F58ady2Rs/hUvzFJxk9dMzlfpwqxiATBp9SSyXdm4ldvbptpiWHix94ts2L8h6LVT1kyhc+PO3ueXnGUhRHkki/TKGXPdWHw8/Pyz7llhtcIzz8Do0Tponj49eEEF/wyA226Dj4teLySCMX8Jycn6B5qcrH+42dmhP8dFF8Htt8PVV0OTJuEbq6hwjucc5+l5T/PWH28FVL4AvYBvVK9RxETEBL22brW69D+rP9UjyrYTUN6GJKMco8h06gWudosdp8cZUPUi77bVYvUG1KMuGSVpGUKIckcC5HLI4YBrrtETk/v367VhVqt+XFBwDPD773DhhXq7fn3dgc9qLb1xVxoFzSZ37Qq//KLLjNjtOpelKG3awOrVuhqGEH7W7F3Dfd/fx/yt84t1XZs6bfjomo+45IxLMPzTgEpR3oYk8U3jWbF7BY/NfYwoaxTHncexGBYMDNzKjc1iQynl3XZ79Ixzh/od2Hhoo3fGGZCZZSFEuSABcjn1wAPw5pt68tHMRX76aRgzpuBrPB5o1szXmnrBAujZM/xjrdTyloIbNEhXtnC5oHZtOHxY74+J0fv9f2Gmvn19uc2SdiH8KKX45+A/pO1M49E5j7LreOgLD+pH16d6RHW6N+vOq5e/SvOazcM40tA8OvtRXl38Kk2qN/G+l4K2DQzvLHKENQKFwsAgx51DhDWCMQljvDPLgATKQohSJQFyOWTGZFddBRMm6KZuOTlQq1bhM8ig15C9/77v8TPP6E7LFll2eWrM2WTwpV2MGQMNG+pyIXa7XsDncunGIllZEBcHf/7pew6bTefGdOwIdesGf50mTfRrSTpGlbX7+G4+XP4hGceCl69xe9xMXjuZo9lH8x2rW60uXRp3KfT5I6wR3NX1LgaePbAkhpuPucjvqnZXMWHVhAJrKPtv2yw2XJ78JXcMDKrZq3FTp5uIrR/L2IVjSe6RjMvjklllIUSpkAC5nPFPfx09GtxunSbhcunvFkvhQfIPP+i0V3/TpsH114d/7JVaSooOdM2c5NGj9V8tbrf+C8bl0jnKFoueyq9eHY4fL95rXH01zJoVnvGLSmHL4S2Mnj+abzd8y8GTB4t9vcWwML7/eC5ofoF3X6Q1knZ1251WuoZ/BYzR80d7UyiCdeHzz0F2e9xku7O9tZUBujbpyrJdywDo26YvczfP5Z5u9zB13VSSeyRLvrIQolRIgFzO+Ke/btoESbpBFbfeCg89BOvWFVzFAnTM1rUrrFnj23f//fDGG2EfeuUX7JczebI+lpQEjzwCy5dD69aQnu6b+i+O337zJZILUQCXx8XOoztJy0jjgR8fKHDWOVSdGnbik2s/oWvTrqd0fXEqYJhVLGLrx/LkvCc5p/45rNi9Apthw2JYcHqcDDx7INPXTwegXrV6HDh5gOFxw/lh4w+SciGEKBUSIFcQt96q84nvuKPoc7OydFrFK6/oxxddBIsWhXV4wj8vZuJE6NMH5syBLl1g82Z46ino3Dn4tW+/Dd9843v8xRcwdKjkxYiQZLmy+GPnH+S4C/5jzO1x8/Dsh1m7r/Cay+fUP4d74+/l3/H/DusCwIJmnM28ZPP7SedJFL5/i57u9TQJrRJInJYos8hCiLCSALmCePFFOHIEXn45tPN37/als1arBkeP6gwBEQZ5y8INGQLvvaf/Mlm2DF54Qe9PTQ2eG/PPP7pXuNuv5Nett8Inn+Q/V4hTtPv4bu6ddS9/H/g7YP/mQ5s56ToZsG9oh6F8fO3HRNvDU32loBlnwDuz3Kd1H37d9isNoxvyz6F/AJ0OEmWLYvrQ6RIcCyHCSgLkCuKrr/SCvRkzQr+mRQvYsUNvr16t14iJMMhbFi4tTf81smmT7txyxx2+/QXlxjz4YP48mAkT4Kabwj58UbVtOriJ//70X37a9FPALHTnxp2ZMXQGZ9Q+o1THk7IoBZvF5l2YN3bhWIa0H8L7y97HrdxUs1Vj1o2zJEAWQoRVuQiQDcNoAXwONAY8wPtKqTcMw6gLTAFaAVuARKXUocKeqzIGyCkp8McfOr7697913BUbC1N0Wh9Dh8KGDTo32b/b3uLFvoD6k0/0pKQoJSkp0KmTLmh9yy0wbJj+xaxcqX9h/r88c7t7d/joIzhxQj+2WPRs9N13+36pRQXaQpyiE84TPDL7Ed5d+q53X9MaTVl+13IaVW9UqmPxn2GOb6qryFw96Wqy3dlE2aLo1LCTN485b37zqW5v2L/BW3vZrMMsOc5CVF3lJUBuAjRRSi03DKMGsAwYCNwKHFRKvWQYxuNAHaXUY4U9V2UMkB0OGDhQp0m88go8+6wuuRulF4aTlaXL8I4ZE9htb9EiXTsZdAbAf/4T+LzVq8O55+rzRQkz0y7sdv2LA10OzuPRvzD/X57/dmRkYMc+i0Uv9rNadXvECRP0L7pPH52WIb88UcI+WPYB//7+3zg9uhFO3zZ9GXXJqBJ7/o4NO1I7qnbI5/vnKz8+93EA7FZ7gRUyTnU7xh7DmIQxjJ4/GgND0jiEqOLKRYCc78UN4xvg7dyvS5VSu3KD6F+UUrGFXVsZA2TQ8Vbv3r6yu06n3gbfttWq4yuzDNyPP+o1Y4VJSoIvvwz/+Kskh0P/Atq1g7Vr8//CCtr2eALzkQvyxBM6v1mIEvb9P9/T74t+YXnuGHsMf97zJ23rtg3pfHM2OaF1Au/88Q73/XCftxufBw92i/7/j9PjPO1tq8VKpDVSgmMhRIEBcpktnzcMoxXQBVgCNFJK7QLI/d6wgGvuMgxjqWEYS/ft21dqYy1NCQl6zVdOjo6jatTQ3/23s7J0STdzHVh8vA6YCzN5sm+CU5SwhATdeW/t2uC/sIK23e6if3EA48ZBJb3fRdm6+syreeziQj+sO2WZzkw+XvFxyOePvHikN1gdET+CTo064fQ48eChRkQNnB4nTo+zRLazXFncG3+vBMdCiAKVyQyyYRjVgfnAC0qprw3DOKyUqu13/JBSqk5hz1GZZ5AHDtT9KFwuHShHRurZZHM7Ozt/t73Jk+GDDwI/tQe9aM8MjH/9FXr0KNW3UzXkbYmY9xdW1LbVGjiTbBg65aJ6dV3SBOD55+HJJ8vm/YlKzel28uS8J/lt+28l8nzHco6xas8qQDcDWXpX8f877Uh3MHDKQFweFy63q9Aufaey7XQ7ibJH8d0N30mQLEQVV9AMcqkXBDMMww58BUxSSn2du3uPYRhN/FIs9pb2uMoDMzg2DHjuOV3jOCdHPzZTUG02/eXx6ElLM0hOSvI1G/F3yy3w+ed6e+VKCZBLXN6WiDExvhxkCPzlFbRtt+v8ZKX0xwMejw6a771Xl40DX4qFBMmihNmtdlL6ppTY8x3JOkK9lHq4lZvlu5azL3MfDWIahHy9GRwbGDyX8ByjHKPI8eRgYHhrNp/Otkd5uKjFRSzZuYRBUwYxfahuVCINSYQQ/ko1QDb0f6E+Av5SSr3md+hb4Bbgpdzv3wS5vNJLS/MFumlpuiBCsCoWs2bptNQ//9TnFdSOGnTPCjNAXrEi7G+h6klL03WP/X95RVWxCLbdubP+ZT/+uA6Qc3L0c9SpA4cOwcmTugbgiBFQt24ZvFEhQlMrqhYXNL+ARdsXoVDM2TyHGzvdGPL1aRlpJHVIIqljEmkZaQzrNKxEq1g40h38tOknAM6qdxb3zrqXjOMZ9GzZk6/WfSWVLoQQQOlXsegB/AqsRpd5A3gCnYecCrQEtgFDlFIHC3uuyppiEYrrr9dx2JAhRZ/rcMBll+nt887T/SxEOWTORKemwoEDvl+uYeiZZVO1ajB8uO7Y16JF2YxViCI8N/85Rv2iK2K0qNmCNnXaeI+1rNWSxA6JtKrdimY1mlGnWqHZdGHx2uLXeHj2w5xZ50z+OfQPUbYoLIbFWzFDKl0IUXWUyyoWp6MqB8j//S80bw4PP1z0uQcPQr16ejsiAo4f9xVTEOWI2YTE/Djgrrt0UjnAxRcH7yFeo4b+qyc1FRoGXdcqRJlYsmMJF3x0QZHn2Sw2hrQfQtcmXWlTpw0DYgdgs5TOB5sDvhjAd/98R4QlwpvCAXhbXhsYRFgj6NOmDw9fqP9ja9ZrllllISoPCZArkf/7P9iyJX9DtoK0agVbt+rtVat0XwtRjpnJ6CdP6pWadjv06gXLl+u/ePJq1w7efVcfl8Yiohxwe9z0+KQHv+/4vVjXnVP/HK4+82pa1GxBYodEmtRoEpbxmTWXz653Ngu3L6R17dakH04HoFPDTqzeuxqAZtWbccx5DLfHjcvj4vYutzNp9STvrDJI7rIQFZ0EyJXIV1/BxIl6gV4oBg6Eb3Kzum++Gd58U1fBEOWQ/0rNUaN0snl2dmCTkbxpF6A/HnjsMd0RBuDMMyEurlSHLoQ/p9vJ0oylZLuzA/Z9/8/3OLY4OHDyADuO7ijweqthZUDsAK4/53qq2ap591sMCxe1uOiUu/75NyQZu3AsXRp3Yc7mOdgtdgyMfFUvmtdszo6jO7AaVtzKTYw9hpk3zAQgcVoiqYNTJf1CiApMAuRKJC0N7rkn9Hzi55/3ddoDXcliwQJpzlYupaToHuNJSTrdYvZsuOIKXfYN9AI+u12XhTMrZRTk+uth2rTwj1mIU7Ro2yJm/TOLrUe2Mv2v6Zx0nQzpunrV6rHgtgW0b9C+2K+ZsigFm8XG2IVjSe6RzOj5o8lx5+D2uLEa1oAc5MfnPo7T46RhdEP2ntDFlayGlUvOuISlu5YyY+gMwJd6IbPJQlQ8EiBXInv2QMeOofeO2L9fl4RbuNC3b8EC6NkzPOMTJeyaa2CmnrGiXj29iA+CzyT7MwwdIF93XfjHKMRp2nZkGzPWz+DQyUPM2zKPBVsXFHp+4+qNuajFRd7HFsPCVe2u4vYutxf5WmbXvrSMNDYd3ERSxyQmr5nMyt0rvVUsAKasnULD6Ib8c+gfIq2ReJTH25o72h7NcwnPeQPtsQvHymyyEBWQBMiVREoKbNwIn36qP4Hfvj14Kbi2bfWar8mT9fajj8Jtt8Fnn+nzbr8dPvqozN6GCFVRnWPyNhmBwMDZZtNJ6Jdd5qsf6H9jSM6yKKfW71/PZys/4++Dfwfsn71pNsdzjhd43deJXzPonEGn9dr+aRij54/Ws8sWK8M6DePdpe8GnBvXMI4NBzbwYu8X2bB/gzfInrJmCp0bd/YG32bZOJllFqJ8kQC5kjDjpePHddD79tuQmRmYohoTA2PG6L4VhuFrJrJkCVyQu7C8enXYvVufK8qpvPnIo0b5ftlWqw6C3W6do2yx+FIurFaoXds30ww6R9lu18W1J03S+957D266qbTflRCn5efNP9P/y/5kubKCHm9aoykfDvgQi6HTkgzDIK5RHI2rNw75NYLNMK/YvYJRjlGMSRjDpys/9S7kA4i0RmKz2FAob6m4KGsUVosVhcJusTPqklEyyyxEOSQBciXicEDfvr60VKfTV7rN3LZa9USjfztqpaB9e1i/Xj+eMEHio3LNPx85LU1vmx8XmI1FnnwSzjlHd4Gx2XQw7XQGn1kO5ttvYcCAsL8VIUpS+qF0lu9a7n3sUR7+88N/2JO5J+j5NSNrsviOxaeUs2wyg2bQi/OuancVE1ZNAPAu4LNb9H+InR5nwPZNnW7ix00/SnAsRDkkAXIl068ffP+93q5dGw4fzr/9wAPw+uuB173wgu4xAXDnnb5Su6KCydvi2u3WQfEzz+hqFi5XaM/TvbuufCHpFqKCm7xmMjd8dUOBx7s3686i2xdhtVhP+TXyVsBI7pHMEz8/EVCpA6B2VG0OZx0GdK1nl8fF072eZkzCGBzpDkmzEKIcKShALtVW06JkOBx6wV10tI6DDh/2paUePuz79P299+Daa/UMssOhJyG7d/c9z6pVZfUOxGkL1uIadG5xZKTe9nj0V926un5ysEV9S5bojxSmT9c3ieQmiwpqaIeh7M3cyw8bf8CjfBVeHOkOnB4nS3Yu4d5Z9/JSn5dOuXtfWkYaqYNTvd8BomxRtG/QnhW7V2AzbFgMC4ezDhNpjfTWT7ZZbKQsSmH/if189udnPJfwHI50h+QmC1GOyQxyBVNYWiroHOSoKH385Eld73jUKBg7VsdT7dtD49xUvJgYOHrUl6ohKri8N8eYMfqGMOsoR0ToihgTJwZed++9MHiw71r/vBwhKjj/ttegc5QX37GYlrVantbzBlvIFywH2eVxke3OxoIFDx6uOesa5m+d703DkNxkIcpWQTPIEhpVMOaE4fTpevZ42DAYN073hIiL09s336y75VksuirYqFE6OAZdxcLsSpyZCenpZfdeRAnLe3NMnw633AJnn61vihkz9I0xbFjgde++q2st+wfHDofOgRaignu8x+Nc0fYK7+OMYxn854f/nPbzmrPILo+LpA5JzLxhJjd1uom4RnGMu3wccY3iGNZpGD8M+4Ez656JBw8GBt/98x1Hs49ywnmCvm37eoNj0HnOQojyQWaQKymHA666Sk8ePvCATrVITNSB8gsvwM8/6/O+/lrXSBZVhDnLnJWly8T5u/xy+OknX35zaqrMJItKQSnFpNWTGD59uHdfmzpt+M/5/2FEtxFE2iLD9trmTPMt597Cm0ve9NZRNjBQKJJ7JNO3TV/pyidEGZEZ5CooMlJXtHjzTR0Em/FOp06+c1avLvh6Ucn4p2CYHyn4mz1bB8YSHItKxjAMboq7iTu73Ondt/nQZv770385/8Pz+WvfX2F5XTM4Th2cSr8z+xFhjQDAbrGjUFgMCymLUhg0ZZAEx0KUMxIgV0LmBOCMGbqggVJ6wtAUF+fblgC5CvFPwahZU5eFy2vqVJ2ofs45MG8evPxy6Y9TiDB5qc9LNK3RNGDfqj2ruOjji1i0bVGJv57/Yr6BUwZis9gY0W0EVouVaHs0NosNt3KT484p4pmEEKVNUiwqoZQU3SwNdKDcpg38+adORx0/HpYu9R0HuOsueOklqHNqC7tFRZSSolsutmkDH34IW7YEP69WLd2uuk+fUh2eEOGy+/huftnyC0szlvK/tP95G44YGHRr2o261erSoUEHnr/searZq5XIa6YsSvE2HEnLSCO+aTwrdq/gyXlPEmWNIseTw02dbmL8gPEl8npCiNBJHeQqxpxFvu46aNcOXn0VjhyBF1+ENWvgk08Cz7/oIl3g4LHHyma8ogzNmqV/+R5P8OM1a8KZZ+r6gUOHBvY1z7vdubOepTbLxcXH65lrKRsnyqEVu1Zw5aQr2Zu5N9+xe7vdy//6/S8sr2umXlx39nWs3L0Sm8XGyj0rSWiVwGWtL2PKGv3/K7NldUHbnRt3BmDB1gX0OqMXQECr62DXbdi/wVtaziwzJ+XlRFUmAXIVk3cW+fhxXdXC5dKfrGdn52+0ds01gXWSC9KsmX7OaiUzuSLKWkqKrv33xx/w3XewdWv+cwxD30BmQxIIvm3WDLRadQ7z/Plwww06rycpyVdCRYhyYuPBjYyYNYJ56fMC6icDjLxoJLWiahX7OW0WG1e1u4pOjToFPe7fle+aL68h05nJ3V3v5qMVH+H0OImy6rqdZrm4grb9y8hF2aKwGJaAMnPBrouxxzAmYQyj54/GwGD60OmS+yyqNAmQqzCHA668Uge26em+RiIej06rOHSo+M/5yCPwyislP1ZRhsyPHTp10tsWS8Gzyqfi/PPh9991sC1EObP/xH5W7FrBHd/ewfaj20/7+WpE1GD9fevz5TznNW/zPPpM6IPVYsWjPHiUB6th1TWVQ9xuW6ctmw5tCvk6m8VGlC1KgmMhkCoWVVpCAtx2mw6OmzfXk30ej55JPnTo1HKPp00r+XGKMuRf2m3ePLjwwpINjkHPUP/wQ8k+pxAlpH50ffq27Uvav9KoH13/tJ/vWM4x/m/x/xV53mVtLuPGTjfi8rjwKA/NazbHrdzF2t50aFOxrst2Z/Of8/8jwbEQhZAZ5CrAjH2uugomTPC1pc7JgUsv1S2nBw6ERo2Kfq433oATJ/T21q3Q8vSaUYnywszJMZuEDByo83CcTh0o+3/sUNi2x6MfWyz6Rhs0SLeyXr5cv45h6AC8fXtJtxDl1ob9G5i8ZjLZ7uxiX7s3cy8frfgIgOoR1dn24LZCW1s70h0MnDIQl8eFy+0ix5NDpDUSpVRI206Pkz5t+jBn85yQrrMaVtzKTZQ1ipvPvZmkjrpNvbS9FlWVpFhUUWZwnJwMo0fr2WOldJySna1nkV94wdeKuqjSt337wty5envCBLjppvC/B1GKQu1lXtS2xaI79n3zjb7hTpzIn/TeooVOev/vf/UqUSEqAY/yEPduHGv3rQXgrHpn0a2p/rfXaljpd2Y/EjskYhiGNzg2MBh1yShGOUaR6cwMaw5ylDUKl3Lh8ri8z6FQ0vZaVFkSIFdR5sRgWhps2uQrMAC6gtf06fCvf4VebOC553TMBPq6998P7/hFKUtJ8d0o5k0TG1t45Qr/7c6d9fmPPw41auigeOhQ/f2jj4K/ps0G770Ht98u+cmiUvhi9RcM+3pYgcfv6XoPb139Fq8tfi2g/Numg5uIrR8b9ioW2a5sPlv1mbebH8CQ9kNwbHFIcCyqHAmQRT5LlsCIEb5Pv0OxYAFcconePussXUpXiHySk3Vx7fvv13k5R47olZ3p6XDsmM7r8e9eA9C6tZ517ts3eBOTVq2gaeELnoQoD5RSjF04lpcXvczR7KNBz+nZsidjEsZ4u+vlVSeqDmfXPxsjTH803v7N7XyyUtf7rBVZiyPZR3i619OMSRgTltcTorySAFnkc+yYru519KivWldRsrL0zHNObuOnHTt0dQwhvBwO6N9frwjdsUOXjjNzm82PKZxOHSTffrv+HgrD0AH300/rmzBYEC1EOXI85zg/b/6Z4znHAZj590ymrJ0S8vV92/Tlw2s+pHZUbe++CGsEUbao0xpXsLznFjVbcNJ1UmaQRZUjAbIIYKZeXHcdXHEF3H23L/WiqD4PCQnwyy96++GHdRMSIYDApPcnn9RBrMej6yCvWwc9e8K+fb6UjPbt9Q22bp3OVQ5V9eowcSJce2343osQJcyjPLz464s888sz+Wouh8rA4MELHmTc5eNOaXY5b97zkz8/iVKKbE825zc9n/UH1tOzZU/2Ze4LSOVI6pgkC/lEpSQBsghgxjH16sH27Xpyzly8pxTY7TrXONjiva+/huuv19uRkfDBBxAR/FPCAnXsCB06lNz7EeWEfzWMceN0WkWbNrB5s17EZ9ZWNhf1Wa26e012tr7pPB5fJQzDCLwx85adi4qC//u/ku+R3r69rgUtRJgs3r6YV357hd3Hdwc97lZu0namefODg3mg+wNc2PzCYr/2txu+Zc/xPVzU4iJv7eSv/voKl8eFW7n1Qj/lwWJYcHqcRFgisFvtspBPVFoSIIt8HA7o10+nWaSn+yb7PB6dCvrTTzo4hsBZZKV08YG0tFN/basVZszQn8SLSuzyy2HOHF1DcM8eX1qE2dLR3O7UCVavzn/cvCE7d9bl4mrWhL352wKXKItFl2i58cbwvo4Qhfjhnx94/OfHST+U7t3n8rg46TpZ6mMxF/MldUxi7ua5pA7W/zDILLKoDCRAFkHdfjt88gnUrw/79wcee/ppPRFo9o/wn0X++Wfo0+f0XrtOHVi5UmopV1rmxxStWsHSpTpv+MgRfcz/hjO3g+3zZ7fDnXfq1IzsbF1+LlxsNl0XsWtXX2WNuDj916QQZSTHnUO/L/oxd/PcMhvDUz2f4rLWl5E4LVFmkUWlIAGyyKewBiKgH0dF6VJwweojT5gA339f/IZrCxbA7txPFi+8EObP17GPqET8c5HHjoUuXfRMst2uA86cHN8N53Tqv7bmzAm8Cf23zVQL0PWTDx3SqRB//qlvIoCDB+Hss09v3EuX6nSQYGw2ePNNXfpFiDJyLPsYzy94ni1HtpTYc+7N3Mvi7YtpXL0xW49sxWJYQIGH/P9xt1lsxNhjpE21qDQkQBYBCmsgYsYkLhdUqwazZhXdQKQ4fvsNevXy9Y3o1QtefBEuuCD0ahqinEtJ0QHl2LG+mywnR//SrdbgOcgF5Sgrpc/xD5QNQ9+c332nX8+8mV2uoot5F2bnTn1DFhQkA7zyis6tFqIScKQ7SJyWSHKPZEbPH43b40ahvA1HgomwRDDxuokM6TCklEcrRMmTAFkEKKyBSGwsPPGEntirUUMXHBg/vuRf/7HHAvd16gQzZ8IZZ5Tsa4kyUtBNtnJlYGMR0B8r9NKNDvIdj43VFTHOOUcfq19fV8IAnSc8e7ZvpjqUdpBFOXZMv/a8eb40j/R02LjRd860ab6VqkJUYCmLUrxVKcymJZPXTPY2HJm4aiJWw8rSXYH/3tosNv648w+6NOlSRiMXomRIgCxCYs4sp6bqIHnwYN3voSTiDn8ej16g98MP+Y9FRcFDD+kW2KKKy5uqMWSI7rp3ySW+WoPnnKMD5pK+Sf0dO6Zv2AULfPtsNl2Xedy48LymEOWEI93BtZOv5VjOsYD9NSNr8sstvxBbP5Zoe3QZjU6I0yMBsgiJ/6Tf5Mk69qhePf+kXlG1kkNx+LCOe1as0F398po7F3r3LoE3JSou/xvSP2VjwwaYNMm3UO/SS3UyvX9r7JK6UU0HD+ryLf4zyaA/9pByLKISM2eZ3/rjLeZunpsvULYZNmLrx3J7l9tDao2dcSyDnmf09M5cS11lUZYkQBbF4nDAgAE6FTQlRddEzswMTAstrFZycf36K9xzj+4XYWrTRpebC1WLFjB8uB6jqITMYBl8s8rmjQlwzTW6vEo4b9T16/Xr/POPb1/LlnDLLYVf17KlvpmrVTv11xaijDnSHQycPJCjOcHbZwNYDb2QxK3c+bYthgUDA6fHyT3d7mHquqkk90iWusqiTEmALIrtp5/gyisDy9GaMYfHo2OUefNK7pNtpSAjQ89aHztW9PnBJCfrBX+iEvNfAJiaqm+am27Sx8xFfv436uDBOh2jJFMw9u7VFTMOHQr9mqefhjFjSub1hSgjjnQHfT7XNT6DVbkIRatardhyZAs3dbqJHzf9KMGxKFMFBciWshiMqBiuuAIGDdKFATwePQnmduvt6tV1vDFiRMnFHIYBzZrpnOdTNXly8ToWiwpo5Eh9U5oB77BhOgiG/DdqjRp6QV1J3qgADRsWP/d4xoySe30hykhC6wRu6HSDNzhuFNOo2M+x5cgWXQlj9URGdBshwbEol2QGWRTI4YCBA3UsYlbZ8i9N27IlnDhR8mujlNLtrNeuDf2alBTfJ+1//w1nnlly4xHlXFE3arNmuoxcOBbxffONrsVcEKX0rLHHo/8CPHxYdwMUooJypDsYOGUgLo8Ll9tFjieHCGsELo8Lj/JgNay6TFyebbOFNUDNiJoczTlKl8Zd2H50u8wgizIlKRaiWMyYwzB0+qZ/DrLb7ev18Mor8PLLJVOC9nQMGOArifv22/Dvf5fNOEQpK+pGtVh0cHz++Tp/+Jln9I0aH+9bvBfum7ZLF73KFXRJur59w/t6QoSJGRwbGIy6ZBSjHKPIdGYSZdULP7LcWQVuR1ojyXZne5/rnPrnsPHgRl7q8xJjF44luUcyLo9LFuqJUicpFqJY0tJ0AYDp03U8MWyY/kQ5Lk6nXlSrptMstmzxrZUy10+VhSuu8G3/+KMvZ1pUckXdqFFREBOjS6Xk5Oh6yjabDqqnTCmdm9bs9AeweHH4X0+IMEnLSCOpQxLTh07H5XExrNMwxl0+jrhGccQ1iit0+9xG53JW3bO8z7V+/3oAMp2ZJPdIZpRjFPFNy/AfESHykBlkcUocDujXT8cX69aFtwRtKP75B846K3CfYcDVV8OHH0LjxmUzLlHGHA5dgq1VK32jmq0ao6J8JVJiY3XZOLMUXEnPLE+cqMurgF71Gqz4txBVwD8H/qHL+C5kOjO9+1rXbs2BkwdI6pAEQGz9WDbs30Dbum2JbxrP5DWTaVu3rcwsi7CRFAtR4rp2heXL4YYb4Isv9D6Ho/Q+ufanlH7N9PT8x1q21OO7+OLSG48oR+68Ez76KHCf3a5Xg5opGXa7Pm/SJP2X1fTp+rzTqZts2rQJ2rXT27Vqwe7dUotQVFlfrP6CYV8H1u80MDiz3plsPLjRm698XpPz2LB/AxbDwvSh0yVHWYSNpFiIEuVw6H4JFgt8+SW89povHbS0Prn2Zxjw6qvQoIHeNgzfsW3boEcPXVe5d2+dgiGqCIdDL6QzZ3AjI3WKhdMJjz6qV5mCzld+91393QyOExNL5kZu00bfmABHjkD79oEd+YSoQm7sdCOPXvRowD6F4u8Df3sX8XmUh6UZSznhPMG0xGkSHIsyITPIotj821EfPaqDYtDBco0aOr4oy3QL048/6nEGq6k8YABEF9EZ9fzz4b//DQy2RQXi36Z69Ggd/FqteqFecrLOSQbo0CGwZMq118KiRfoGh5KZRf7vf+H1132PrVbdcCQiQv8fp1cvPYNts53e6whRAczbPI9rJl/DSefJImsp9zqjF02qN2H/if3Uq1aPLwd/icWQuT1RciTFQpQYs6GZGQQ/+qievQV4+GHfdnmwdase01dfndr1n30GN99csmMSpcS/TfWmTXoxH+gcoEmTdHULpXTg7F8WDuDee3VtZfMvwdP9i8/thvff14sEC2ou0qoVXHCBzlO++Wb5y0xUWmbr6nManMPnKz9nzIIx5LhzcHvcePBgwVJg4HzJGZeQOiSVhjENS3nUorKSAFmEhX8J2qwsHWfMmlU+ZpD9ZWTo+shjxugxh+qCC6TwQKVSUFm4mBgdlL77rj7PMHS94pL+OGTHDh10F3VT3XknvPeeb1GhEJVQQWXjYuwx1IqsRcbxjKDXXdXuKr4f9n0pj1ZUVhIgixLnH2sMHQq//64/qbbbda7vZZfpfGTQxwvb7txZP87IgJ49fRN/5veSWvDn8ehPz3fuLPgcpxPuuEN/B1i9Gjp2LJnXF2UsJcU3m2zOLMfG6pvZ4dB/QY0bp29Emw1uv12fO3myzhvu1Us/z8qVRd/U5nbeChmtW0P37rBnjz5n/XrdNtucvTYNHQoTJuj/QwlRCaUsSmHTwU0kdUwiLSONTQc3eatYbD2yleW7lrPvxD7qRNXh8raXM2XtFO+1/c/qT8+WPaW6hThtEiCLEucfa4BO3Tx2TOfurlihA0xzsX5WVuHbVquvCdo998DUqTpNdOzYsikhN2SI7lAMcMYZ0CiEbqqtW8Nbb/nWY4kKxEzHAP1X38mT+gY+/3wd4GZl6ZSMqCidM+zxFH1Tm9sxMTrwHj3aVyEj7w29c6cOwN97L3ABX7t2ULdu8DFHR+u/HK+6qsR+DEKUF68tfo1HZj+CxbDgVm6ibFG4PW6cHj1zYWAQWz+WmpG+zpQWw8LV7a7mqV5PYUiKkgiRBMgi7BwO6NNHp3Kat5VhFG87MlLHIUlJMHduya6TKo7ZswObj4Sqb1+9ONAia0gqHv/Vp6BvZv9uMzVr6lWpxb2pDUP/BRgTU3TKhscDDzyg20GGIjpa/zWatwi4EBWYI91B4rREknsk8/S8p/EoD1nurNCvv8XBpa0uDd8ARaUiZd5E2CUk6MX6ZlzQuXPxt7OzdTwxebJvYsystuVw6Im+0tCnz6nVTZ4zB1q0gBde0JOHogJJS/N9XJGQAI884jvWubMOjk/lplZKfzxyxx2+4Ligm9ligTffhKefDi3/+MQJXcLOzAcSohJIy0gjdXAqD134EP+98L9kubNoW6dtyNf/suWX8A1OVBkygyxKjP+CPTNdwr84QFHbTqcOTOfM8T1nTAzMnKm3S6qgQKjcbp1/nJ1d9LkTJ+af9HvySXj++fCMTYRZ3pvZ/+Yszk2dk6NziJ1O/f2nn/Tzh3Iz796ty7AUdGzIEF9g/PHHcNttJfojEKKsmTPJV7W7igmrJhBpjSTbnU2MPYaUPil0bdrVe+689Hk8Me8JAK5oewU/3iQF70VoJMVChFVBxQFONQf5gw/0Y4DzztOL/158UaeDmuujwtkduLiys/WM87Jlvn3Nmun4RgoRVDB5b+YxY04/B7lNG/3XltWqr3/pJejSxbcS9VRu4NGj4dln9fYVV0gHHFGp+KdZjJ4/GrfHjdVipVmNZmw9vBW71R7QYW/zoc20fVPPMteKrMXBxw5KvWQREgmQRVgVVBzgVKpYNG2q9zVooLv1ga8BmlK+2CSUtU+l6cgR3bTtllt8+xwOuPTSMhuSOBV5b2YzgD2dKhYAH36ob16r1fdX4TPPnPoNvG2bXkEK+v8ce/dCnTqn/faFKA/MWslmdYukjno1+MsLX6ZmVE3qRNWhbd22/Js8aAAAO0tJREFU3ioWSikaj2vM3sy9AKy7dx3nNDinzMYvKg4JkEWFkHedVL9+uqCAuebJavUFycHWPjkcpb+gL69//xveeUdv166t60JfdFHZjUeUA+as9LFj+sYtqRv4/PP1+SBdbUSll7IoBavFytd/fc2i2xcBeqY5LSONkReP5NrJ1/Lthm8B+Oiaj7i9y+1lOVxRQcgiPVEhmOukQAfKs2bpjrxK6U+m3W4dW7Rq5Vv7FBfniy3MBX1ladgw3/bhwzr14rnnymw4oqyZN+aMGfDgg/oGtlh8N3C7dqd+A19/vW978uSSHrkQ5Up803hGzRtF2s40ctw53jQMm8VGyqIULmx+offcxdulw5M4PTKDLMol/7K0iYm6osWECfqTZIvFtw7K49HrlPr3141KyqJmcl5Kwdln6859/l58ERo31mu1rrwS6tcvm/GJUlbYzWwY+gbu3Rv+/LP4N/DGjXDmmb7Hd9yhP8KIjNTpFzExJftehChjry1+jYdnP0yPFj1Yf2A9yT2SGbtwLKmDU7EYFi797FIAGsY0JP2BdKLt0WU6XlH+yQyyqFDMT5gTE3XDkG++0f/WR0T4agzbbHD55frxd9/50i/NCloOB9x9d+mVhjMZhm4ykreowBNP6MZsw4frpiL+/SBEJVbQzRwd7buBf/5Z/9VkziSHetO2a6fznE0ffaRXtXbooP8a+/TTEn87QpSlhy58iKvPvJqF2xfiUR5Gzx/NdWdfB+CtcAGwN3MviVMTy3KoooKTGWRRbpkTb+aiP7Pjr38Vi4wMnYYBeuY2KkoH0aez9qkk7dun45UdO/Ifi4iAhg0Lv75mTd35+MorwzM+UUqC3cwAr72mb2C7Xd+sL75Y/PaRWVm6jeXs2cGPN2umn7sgVqv+P9RLLxV+nhDlgJlWEWOPYeuRrdgMGy/3fZkx88fg9DhRSnHSdRIAm8VG4+qNQ37ujg07MmHQBOpHy8d7VYks0hOVjpmymZysq3GdPOlrfGYYeoLOzAfOW1GjNEvE/fOP7v1w/Lh+/OOPuoxtqGrW1J++t2oVluGJsuJ/Az/xhC9fqF8/HdTGxoZ+o548qfuc//KLrm6xdm3xx/Pee/ojFyHKKf/Sb2MXjqVVrVYs3RUYB7Su3Zr0w+mn/Bp3dLmDD6/58HSHKioQCZBFpWNOyiUk6MZjzz+vc3/Xr9fHbTZ4+eXgNZntdrjzTpg0yTfLDKVTAWPDBhg8GNasKd51/frp3hA33KBnn0UF538DDx+uu81YrTpH2eHQN+2p3qjHjunrpk71dfUrSrVq+q/Inj19FTaqVz/99ylECUlZlILNYvPmHCe0TuCyzy7DscUBgNWw4lZubBYbLo/rlF/nt9t/o3vz7lJHuYqQAFlUWuZEnLn2yX/xnvmJsVJ626ywZTY3i4qC77/X55iTeS5X+INkpXR6iDnjHcz69fo9ud2B+1u31mmsnTqFd4yilJg3cJcu+Tv1+d+oN9+sZ5XHjvXdqGbaRkE37IEDuh11QVwuGDCg4BnnhAT9f6pmzU7/fQpRAsz6yAmtE3CkOxg4ZSDZrmxcHhdu5fYGyQAR1ggsWLBb7Xww4AMualFwvc07vr2DOZt9bVxb127N+wPep0+bPmF/T6JsSYAsKiX/T6lHj9bBpNl/ITlZxxigg8nVq/V2tWr6E2nT4MH6k+nk5OKnf4bbK68Ej31q1oQLL8y/3190NPznP+XnvYgg/At/JyToyhPbtuljZsc+U8+esHChbjU5dWrJ3bCrVukGKEeOBD/euDGce27wY4ahX/vBB+VjDVGqzODYwGDUJaO8Ochuj5tsd3bAuU/1fIrnLiu81uaKXSs47/3zAvbZLDYmDprI0I5DC7hKVAYSIItKqaC1T5Mn60+ls7P1ZJzbHTgxFxmp/203jzdqpM8pT8Gxae1a2LVLl7FLSdGfnoeqVi29QFA+KS+n/NMszGYi2dl6Zte8af1vVPOvuyuu0H3NS+qG3b4d/u//dIqFmSx/7Fjo6RmXXgrX6UoCVK+ut2vVOv1xCVGAlEUp3g57aRlpxDeNZ8XuFTw570nOqX8OK3avwGbYcCkX1SOq823St9621AV5aeFL/N/v/+ftxgcQbY9mx393UKeadKmsrAoKkFFKldoX8DGwF1jjt68uMAf4J/d7nVCeq2vXrkqIYObNU6pmTaVq1VJq3DilYmKUAqWiovQX6H0jRihlGEpZLHpf//5lPfKirVypVPPmeryhfr3/flmPWhSpoJs22I3aqJH+/vTT4R3TTz/pMRXnZjO/evRQyu0O7/iE8DNv8zxVP6W+GvfbOFVzbE0V80KMqjm2purxUQ8V+VykqjW2lpq3eV5Iz7X9yHbV5o02imdRPIt64/c3wjx6UZaApSpInFmqM8iGYfQCjgOfK6U65u5LAQ4qpV4yDOPx3AD5saKeS2aQRUFSUnyzyebMct4qFg6H/rr5ZvjgA53qefJkYAEB//PzbnfurB8vWKA/nQZf+blg55ol6syCBKezGDA7G377TX8vyOzZekIQdGrrsmVSwatcK+im9b9RP/lE108+cECnX9hsOsViw4bAm684N2ewbf/KGZ9/rnOWBg8OPu7ffiu4TaRUxRClyMxNTstI884sT14zmflb53PSeZLG1RvTuXFn7/62ddt6zx95cf7/GI9fOp57Zt0DwDn1z2HtvWsx5D+ilVK5SbEwDKMV8J1fgLwBuFQptcswjCbAL0qp2KKeRwJkcTpSUnzxRXIyPPus/kTZbtfxQFZWYNWLvNtWq/4UPDtbPzbTRQs6Vyn93KNGlU6e88GDOtDPytKPzU/qC2PGQR9+qH82ohzIe6OOHq1zhHJy9C/U7T69mzPYdkwMjBkTeiHxX3/V5zidurPfjz/q/Yahb7xQtWkDX32lS9EIUQLMPOWj2Ud59tJn+b/F/4fT48RusTPqklEB1TDyOpZ9jKavNeV4jk45irRG5guQ29Rpw9QhU2nfoH2pvB8RHuU5QD6slKrtd/yQUqrIZJ8aNWqorl27BuxLTEzk3nvv5cSJE1x99dX5rrn11lu59dZb2b9/P4ODzIiMGDGCoUOHsn37doYPH57v+MMPP8yAAQPYsGEDdweZGXnqqafo06cPK1eu5MEHH8x3/MUXX+Siiy7it99+44knnsh3/PXXX6dz587MnTuX559/Pt/x8ePHExsby8yZMxk3bly+4xMmTKBFixZMmTKFd999N9/xadOmUb9+fT799FM+DdJh6/vvvyc6Opp33nmH1NTUfMd/+eUXAF599VW+++67gGPVqlXjhx9+AOC5557j559/Djher149vvrqKwCSk5NZvHhxwPHmzZszceJEAB588EFWrlwZcPyss87i/fffB+Cuu+7i7zx9nDt37szrr78OwE033cSOPJ05LrzwQsaOHQvA9ddfz4EDB9i+HWrUgNq1oU2b3kyc+DROJ8BVwEkC9Qceyd2+NO+PBkgE7gVOAPnvPbiVa665lYUL99OixWBq1w48Go57b/162LPHPONF4CLgNyD/vQevA52BudSu/TytW+s4yWrVR+XeK9l7z1/v3r15+umnAbjqqqs4eTLw3usfE8MjjzwCaWlc+sUXeueaNd7yJkXfefprPxBsHngEMBTYDgTceRYLdOrEw889x4Dq1dnw3nvcvXQptGgRcH3AvXf//XoG3PzLjOLcefB89eq6s45fICL3Xhnee/3763sPuPTSS8mrIvybm90km74v9sU9x1cOqGZUTU44T/DOm+/wr37/KvDf3MZJjZmyZwpsQN/AeV0HN/a4kWuc18i9V4HvvZYtW1b8VtOGYdxlGMZSwzCWOnUkI8Qpa9ECb6DaujU85pfYU7Nm8O2LLsq/3/949+7Br6tdG779Vrefzhsch0urVr6JweI4fBhWrIClS/VEpChjCQn6a+RIffPUrg3Nm+tjRd18/sfj4/MfBzjrrOD7zQoaq1bpxYPff6//oiyMxQLnnKM/LjkVx4/D1q2ndq0QQSS0TuCmuJu8j6PsURzNOkqzGs2IbxZfyJVwT7d7OLPumYWe8+vWX0tknKL8KQ8zyJJiIcqcWUDA5dJfZqUL/6oXebedTujTJ3/p2rzbZkGCqCi9wN+cKCiNpiRKBUzmFXre3XfrXhX+br1Vp76KciRY4e9TvTkL2jYLiVss+mOE6Oji9Wz3eApPks/rjTd0ConpmmugZUu93bChbonZpk3ozydELjPNwuVx4XK7yPHkEGOPoZq9WoHpFf6UUmS5Av8j6vK4aDKuCZnOTAC2PriVlrVahu09iPAqKMWiPMwgfwvckrt9C/BNGY5FVEFmcGwYer2ROfllGDo+CLZtt+uyr3Pm+NI+CzsXdOxx7bX6tQYNCpzQCxfD0JXBivqKjoaPPtLlbP2rc336qX6Oiy7SXQlDCbZFGPkX/v7mG18ezKnenMG2bTZ46SX9PGagfPfdxUuat1hCu/HMr0ceCfx45ttv4e239deoUXrRYK1a+qOe227TrTOffloXPF+06PR+pqLS8q+V/FzCc1SzVyPaHs0J5wlGdBtB4rREHOmOQp/DMAyq2asFfNWIrMEFzS/wnrNw28JwvxVRBkq7isWX6ATO+sAe4BlgBpAKtAS2AUOUUgeLei6ZQRYlJZSqF6dbxSI2Fh59VMcedjt06ODbH+6qF6ciKck3Ln99++o1WJby8Kd1VRSs8PfkySVfxQJ0IfGTJ3WQHBGhP06A/Detf9UL86Y9lZv16FG49179usVhtepOPz16FP81RaWWt1bypoObiK0fy3Pzn6N+dH1GxI9gyhp9Lw/tOJQpa6bQuXFnABZsXUCvM3oF1Fk2q1+MvHgkz/7yLKPnjwZgRLcRvNPvnbJ6m+I0lZtFeiVFAmRR0QwbBl98oYPkl1/WE2OZmfmLCbjdOgCNivJVvSitFtim/fvhsst83Qf9xcbqDn233CINSCol/49URo2CJ57wVcSA/FUv7Hbo2lXnKtvtOhUDTj1YXrhQJ8GDDs5//BF++qnwpiXNmhVcUs4wdPAcZKGPqFoc6Q4SpyVyQfML+GnjT0RYI1AoPB4PWe4soqxRWC1WXB4X2e5somxRRFgjeOaSZxg9fzQGBtOHTiehdQJzN8+l74S+3uf+YMAH3HLuLditp5h/L8qMBMhClCHzk/HzztM1is3Sbx5P/m2PRz/u2RP++qvsWmArpbsPnzihZ7/NAgqm66+HadNKbzyilPh/pJKQoAPUq67yVZZQKvCm9W+JPW6cLrxtBtjFyVsuzIkTOhj/+28dQJ88qV//jTd0TcOiWCx6lrlnz9Mfi6jQHOkOrku9DqfbSaYzE6th1UGy8gRst63Tlk2HNmE1rNgsNqJsUd7gGHQZuDov18GtfNUxHuj+AK9f+XoZvTNxqiRAFqKMmMGxGeD26OFLm2zRQnf5zbtdt67+d/+ss/R381qHo/TTLUDPXvfpA/PnB+5fuBAuvrh0xyLKgH/OTUE3rcXi+6pWzRcch/OmnTUL+vcP7dzu3WHxYumYI3hq3lO88OsLtKjZgu1H9f0bbNt/39O9nmZMwpiA57lq0lX8uPFH72ObxcZf//6LdnXbldI7ESVBAmQhyoiZNmrGCgMH6k+szeoWBRUhqFdPpzqcfz4sWZI/0C5tR4/qroOPPOLb17q1nhU31a8P990HHTuW/vhEmIRS4sVi0TO85r8nl1+u0yJK46adN0/PDgf7t8zj0bPaZkWNOnX04r4HH5RAuYoy0yyuancVE1ZNINIaiVKKHE+Od9vpcdKnTR/mbJ5DhDWCHHcOtSJrBcwgA+zN3MvEVRN5ePbD3n1DOwxl8uDJZfHWxCmSAFmIMpY3tbOgHGSzEVpOjv73/OBBPVG3dy+8+GL+zsJw+gsJze1Q1lulp+s85IJKkUdHw/33wyWX6DzmiIjT/9mJMhLKTRsTo1thm40SDEMHq2eeCdu2Bb9pg92cEJ7VqQ8/DK+9Frhv/Hi4667Te15R4ZjBcXKPZEbPH43b4y4yB1kphc1iw+3RqRS92/TmstaXBSzu+2j5R6zbv877Ou3rt+eO8+7wLvoLtb21KBsSIAtRxkKtltGwoe7e27MnNG2qy6+53b7qXWbqZ0m2wza3Q+0y/OSTOu4pSlyczl2uX7/w8+rWPfXeEiKMQrlpHQ6YO1ffcObNYnb9KuimDXZzmoF1SfdkP3BA34gZGb599erpj2VCWWUqN2elkbIoxRugmtUtJq+ZzMrdKwusYhFbP5aRc0ZiMSx0bdKVZbuW4fQ4ibLq/3CagXWOOwcPnoDXMwNuhQqpvbUoGxIgC1FBBEvJOHZMH1PKF284nUVvu1w60F6wILTrrFb9VVQKqcejg/h9+3z7cnJ04LxlS/Hfc926MGOGrKGqkPIu6nM4YMAAvbAOCr5p896cLpe+sc47D9auzT/zfDrl5PbuhZ9/hhtvLP77q1dPdxE8//ziXysqhf/98T/u/+F+Im2RuDwunB6nroCRm47hv+3PwMBmseHyuLgp7iZ+2PiDBMflkATIQlQw/umbX36p83/92Wy+VtBFbRfnXIDBg2Hq1OKnkO7ZA2+9pSfn5s4t3vutUQMeekiPIzISrrtOx0CiAvG/Yb76Cv73v8DjRd2cpshIvd9/5tm/nJzHAy+8oCtmFCcl4+uvdfmV4mrWTPdeb9y4+NeKCs/tcdP0tabszdwL4A16825bsOSbRQYdKCtU0IV+ouxJgCxEBWPOJENgZ2GbTX8iXZLtsP23nU4dfyQl6SD3VD/lXrgQHnsMNm4s/LyjR4N36GvcWFf1qlGj+K8tykhxbtqCbk7QwXDejzf8A+kRI3TujtNZvJQMpfRNOWlS/qA8mMOHfWNq1w7uvFPnC9Wtq4N1sx22qNSCtasOtrgvx5ODxbDgUfmD5Na1W3Ms55jMIJdDEiALUQH5dxYePVrnIoc7B/nSS3X1LNDrrMaPz9+sDU5/2/zEvH59He+YcYi/556Dp54qiZ+kKDWh3LSF5SDn5OQPXqOjfSkbEFh7uX17neJRnBWs/tuFta6cPRuuvDJ4hQybTbfCLqhBiagU/NtVj7pkFKMco8h0ZubLQQ62ndddXe7i6w1fk9wjGZfHJQv2ygkJkIWogELpLAwlW8UiI8MXICsVenBd3G27XU/IffyxjqF699YpGtu363VVoM8ZMgTWrdMVPZo0gV274MILS+gHHETnzvpTeGmnfYpOtR026AWATz4J55yju+nlnXm2WHTwbJZQMW/MiAh9s4T616P/ttn0pKCZ6E8/hQce0B915GWx6LaSMTGn/vOqVw9uu03f4KLcKahddd4W1Xm3h3QYwijHKE66TgI6/cIwDOIaxrFm3xp6tezF4h2LefbSZ3n04kfL5s0JQAJkIUQI/Cf/nnhCpzfs3+8LFoN1/jvVbTO2MdNNITC+KUu9ekHf3C6y0dEwaJCu+SzCqKiZ57zl5Mw0DIvl/9u79/ioqqvh4789yYSEcCcKAlEBkWJBghW0KggCSr0AXrh9eLxULVprq4/6glhFxFfRtHmttn3wVilUKyCKIn2qIgSRtkjAgoKAIGIFgqh445LLZPb7x5rjOUlmkkkyk5lM1ref+eTMmdvOyZau2WedteR5VTv8RbvttK685BJpJLJokby/N5/54EG5CGDHDgnQn38+/CmP+tK2lClpxa4VjPjLiBqfYzCsvX4tg7rqRaCJogGyUqpW3goaTvnYTp1kZRfghBPgk09iuw1yDdTevZUD8WSSmSlpIM4iX7duciGj5kfHUG0rz95ycv36yX5nRbmszE3RgLpPQufb2m23SWe+2q5MPXAAzjxTioLHgs8n5V9yc2PzfippXLHoCl7c+mKNz7m6/9X8eeyfG2dAqhoNkJVSUXMW85xrrKK9wK+u2z6fLBSWlUkq6QehWvvHHy89JkCCU2cB8fDhynFQ585Sau788+VncbGUvP38czlz3bWrFD3o2DG6EnL79knOdUVF7c9t3RruvFNumo7RCKrWZB44UNIwakrJqMuEdCZWdjZMnuwG6JFylE4+WSZBUZE72fburVutwhdegE2bZHvWLOnyp1LK4bLD3L3ybp5890l6dejFps820SGzA0cDR79Pv2jpb0nx7cW0adEmwaNtnjRAVkpFpSEXBtZl2zlj/vjjEteUl0euxNGjh8RGxx4ri3fgxjPhKoJF2/AknPfeg1decbsTL1vmxjDhDB0qBQ4c6elyxnxEzWdWVUPFaqI6KRfl5W5us1Mlo6Qk8lWuteUuR2PBApg0SbY7dZI0j1hLT5dg/9xzY//eqlbe7n2z18xmcr/J/GHdH3h4xMNMe3MaFVa+jV/Q8wJy29R+BuHUTqfyi0G/wGf0W3msaICslIpKQy4MrMt2YaHcrroKnn5aAtn09PDFDsrKpH/Ehg0SjxjjLvqBxCne9AyfT17bsmXdguNwysvl9//wQ/f+K6/Atm2RX5OWJs+56KL6f66qRawmal6ee3HgSSfB5s0ysYyRydSli5xaqDrBnO1Jk6RUXX3qIZaUyPt/9VVDj0bN/H6pyDF0aHw/R1Xj7d43sMtAzso9izaz23DPufew/7v9/HH9H2t/kyr+eOEfuWngTXEYbfOkAbJSKqlEE984bbf79JH9IAuF3mZsZWWVq36ddJJbe/nyy6UBWrgOyQ1pzBYIyHMfeSTyc7Ky5AvAGWfU5+ioRlO1G86YMbB0qTzmTCbvpPJuO+XnTjsNfvvbyhM42gl2771yuiPesrLcHOfOnaWJS9++8f9c9T0nWL5yyZVMO2cak/tNpudjPfm65Os6vU/3dt3Z8csdpPnS4jPQZkYDZKVUkxMpiIaaK4I5BQ5AzpbPny/5y1VTPOqbhuH497+lwZrDWjnT7rTb7thRztS3axf+9V26yAV/KoHC9XYvLXVbX0fqtFNRIc/JyJB9VdMwop1ggQC89pok0MdaMAgzZ8L+/dUf69QJnn0W2lTJe/X55D+qhpSuU2E56Ra9OvTivc/eY9awWTzw9gOM7DGSv+34G1NOm8IPcn4Q9rUWy/QV0zl49CAAj456lDO7nUmrjFb0yemDcU6nqTrTAFkplTKiqQg2erRU4wJJeaiokMAZ3OpgIAtrkyfLdixWmT/8EM46y63lXBNjpMRuQYFe6JdwTnBsjOQUz5pVew6y0+wkN1cKeKeny+udCVY1z6ewsPZ22LG2YYNcxXrwYPSvOf54KXnXpUv8xtVMFX5cyOgFo2nTog3F3xUzud9kXvvotag67P16xa95cM2D1fZP7jeZZy97Nl5DTnkaICulUkZt6RlOYDt/PmzZIq9p1QoOHaq+7ffDQw9JTFR1ldnbzCQYlOd5m7VFCqDfeUfioaNHo/t9fvYzuVhRg+QEClclo6YqFk4HvilTJOUiK8v9g3sn2BlnwNq1lb/VOTk6jeXIEQngQb7BjR9fe7Hx00+Ht96SAF/F1C1/v4XH1j1GukknYAPcM+QeZg2bReHHhTV21iv+rpgTHz2RsorKNbgNhpK7S8hIy2iM4accDZCVUs2KsyAYCLi5ylXPkgeD1XtNeLerVsqA6KtlFBbKY4cPhx/fN9+4F/4BXH01XHtt/X7Xli0lXnPGqBpJbfUQy8tlsgwaBLt2SXBcn2oXsbZypfRxrzo5rYV3361ciPzcc+GOO+SKUz2N32BOmsWAzgNYvms56b502rZoy12D72L2mtm1riS/sOUFfr/u95QESthQvIGglb/VN3d+o2Xi6ilSgIy1tknefvSjH1mllApn5Upr27Sxtm1bawsKrM3OlpA3M1NuIPsKCqz1+51w2No+fcJvG2PtiBGy7fe7r/H5rE1Lk/daudL97ClTrH344ZrHGAhYe+WV7mc09DZ2bNwPq/JaudLanByZRG3ayCRo2bL6BPP55H7r1vK8KVPktd6JsnJl7ROmsdx3X/gJNmGCtSUliR5dk7Zy10qbk59jC/5ZYHPyc+zPl/3cMhPb/XfdrZlpbME/C+r0fjn5OZaZWGZiPzv0WZxGnfqA9TZMnKnrDUqplFNUJGe/nbPlkyeHzy/evt2tiGEtbN3qLgI6286Ff2++KdczffutvId32ykFV1Agi4rBIAwfLiu6kdIw0tJg7lxJ43jmmYb/zi+/DGvWwDnnNPy9VBSKimQl2DvZqub5bN8uzUQqKuC77+SP3rs3XHihe4ripz+F3/xGVpe9uUONmYLh9etfy7ieeUY6DTpnmRculFvfvtJi0+nFrqJWtK+IRVcs+v7nsO7DWP3JarZ8voXBxw8mEAwAssq8YPMCenboGTHdAiAzPfP77ZJALSkzqs40xUIp1SxVvSYrXA5yNM1MnLPo3lJz3soZTh7zc8+5aRjgxkDBoFTcWrJEUkHqas8et+PxRRdJYxOVBLw5x7NnQ/fu8kf3ys2VP+CNN0pXvWRJwXDs2SP/YcydW3l/t24y6TSnp0EKPy5k7MKxHCo7RNAGKTi/gAGdBzB24VgMhiUTltSYbnHy709mx8EdAGy/eTsndzy5sYaeUjQHWSmlPKpekxWuVnJtzUyysyX/eNo0tz21MW4eczBYuYPxq6/Kc7xldxtq2zZp0+38U75pk7TbVgmWny+TxRvwnneeTCiAXr1gxw73gr5LL5Wi38kSHDuslVMj994rF/s5XnwRLrssceNq4pxc5EVXLOJI+REufv5iANJMGq0yWtUaHAOcOudU3j/wPgAbb9hI/8794z7uVKQBslJK1VE01TJAVodLS91g2FkJdmIgkLPrQ4fKe738suxziiU09Iz65ZfDSy/J9pQp8MQT9X8vFUO11Vg+5RQps9K2rVy1ec89jdM0pD6OHIGbboJ589x9q1bB2WfrSnI9OE1DnCD4xmU38sQG+Q936llTeXjkw7W+x6CnBlG0T85KrL1uLWd0065E9aEBslJKxZg3TWPJEmkccvvt8pjTrMRbLcPZf/31lc+oN7Ty15o1MHiwbLdqJZ2RW7du8K+nYiVSjeWyMqk1vHevfIPKzo79t6dY+vRTSRVxTpcAXHGFrHprhYt6c1ItAsEAJeUlZPmzeHXSq7WuIA+ZO4S3//M2AKuuXsW5J57bGMNNOZECZK26qZRS9eRcn+XkFd93n8Q4Z57pPic9XQJiRyAAc+ZA//5ucDx7tsRC9XX22dL8DORs/SOPyNgactuwofZSuSpK3okSCEiQnJkpNZK//FK2KyokYL74Ykm3SE+XXJyGTIxYy82tnlaxeLEU8t6+3c3zUVFzgmODYdmkZUzqN4lARYBLF15K4ceFNb5WL9KLL11BVkqpGIiU0+zkMc+aJYHxzp3ua04+WRqcOWmnDWm09uijcOutMft1AKm4sX595FbZqp68uTtOAHzhhbIKe/QoHHecTIwHH3TzeJz8HqckSqJWlg8ckEoXTz9d/bGbbpIrTlXU8v+Rz0cHP2Ji34kU7Sti1e5VbPtiG8e0PIa8znlM7Dvx+4oWA7sMrNRIZMyCMSzdvhSAlye8zJgfjEnkr9JkaYqFUkolgBMLgdtT4tlnpfHakSPwwx/KRYDO9VxOukVdY6CDB6W4QLTd+6I1c6Zcn6Xi7NZb5VuOw+lOY63bzSYzU1afY5GX01ClpXKqZONGd5/PJw1RTjghMWNq4go/LmTsgrF8W/YtDw9/mAfXPEh5sBy/z8+Mc2dUayQyYfEEFm1ZBMDzlz/PxL4TEzn8JksDZKWUSpCqFb/GjZPScTk58Pnnsii4a1fDq30tWSLpFQ0Nko8edVt0t20Lb7wh8Vrr1pKCqummMVa1I58xcgsG3bIpwaC0vP7gg+QpB7d7N9xyCyxd6u6bNk16sqt6Kfy4kBF/GUGaScNaS8AGGHfKOAp3F1brsnf1y1czf9N8AOaOmcs1edckaNRNmwbISimVIN4z6t6V4vJyqaD1+eduta8BAyQ9I5HXagUCsrLtbYXt6NMHTjvNDZKHDJEcaw2a66nqt6fp0yWFoWoCeOvW0mykb1/Yvz/xwbHX0qUwJnR6v21buOQS2T7lFGlT7fcnbmxN0GULL2PJNrmwITMtk5KKEu4Zcg+zhlWucHLDqzfw5LtPAjDnojncePqNjT7WVKAX6SmlVIJMnSqxzNSpEnwuWgS33QaDBskZ9G7dJDgGqYQRCMjP8eMTc61WerqcyQ9n61Ypa/fss3KbMkXuq3pyOvI5E2PAAFmuHzBAHk9Pl3aPhw5B+/awebOULHGC48JC+QaWSBddBCeeKNvffONOjrvugtGjpfpFfbrgNEOFHxey4uMVtPS3JMOXQUlFCf079WfO+jnVLtrTi/TiSwNkpZRqRE6w7F04LCmp3Lm3pETKxeXmSowxfbrsz8+X191wQ/xjokmT5Gx5//7SeOTUU6VCRzi33QZffx3f8aQs77cnkEkxY4acRsjOluDY75efX30leb5Llki7Z2cSJbrSRVqaTNRwXnsNjj9eLjx08nZUWN6KFvcPu58sfxZ+n59tX2xj+jnTGb94fKUgWQPk+NIUC6WUSgBvozVvbvKTT1YuM+vER4EAXHtt5JbVjeHwYVixQhYJrZWYaO9eeaxFCxg1qvZUixNPhOuuk0wBVUWkzjT79kmXvcGDJXH9z3+WoLRVK5kIyZBqYa0U5N69W+6//TY89VTl55x7rgT1mo8TVtWKFgO7DGTxB4t565O3uKr/VdWqWMxcNZP73roPgHvPvZeZQ2cmcPRNl+YgK6VUkglX7evSS6FHD0mxcDrzgbvdogVcfbWUkGtI1YtYWLxYgvr6GDZMOgBmZMj9Hj1g+PDYjS2leDvyAYwYId9Uhg+HN99M7NgisVZKoBQUyDcrx8UXy+mRoUMTNbIm5akNT/HO3nd4enT1snoPrXmI6Svk9FK03fdUdZqDrJRSSSbS2fVPP5W4IjN0BtXncwPl0lJ4/32JMc46S7oTL1kijdpWr5b0ixtukIvpbrjBTVGNR2rG5ZfDAw/IQmZdFRbCzTdLDvOUKRLzPfZY7MaWUpyJAvJHXLdOTj8UFbmFtp0/cqLzkR3GSOecQ4fgV79y9y9bJr+L9kOPSnZGNofLD4d9TFMs4ksbqCulVBJwrtVyfoKsrv7gB7KanJ4uqRalpfCvf8Exx0jxgLw8WLtWgmnn7HUgIM/bvRvmzZPrve67TyqFDR8u77V9u9tzwuk/UdfVZ2MkzeL222UMBw/W/PzSUll1fvnlymkkjmnTpLRuly7hX9+mjdyarcJC+WP5fJKrsm+frMj6/fLNavx4d/Ikkxkz5KI97wS5/XY4/XTo1Cn8azp2lGLhzVy2P5vDZRogJ4IGyEoplQSc4NR7Ad+MGW776rQ0mDxZ6if7fFIazu+XPg3du8PHH0tAba0EyIMHSxpoeroErwA//znMnw9/+5u856xZsvLszWmujxYtJL00GhMnygr5vHnwySeyb9Uq6TBYUiLdlyPx+eT36t1buvtdeCGcc44cm2ahqMit/zd6tByEffvgpJMq10VuSEvGeOjYUVJBHn9ckuxB0i5Or3ZW25WVJd8AR4xonDEmqahXkCs0QI41TbFQSqkk4638NXEivPqqdLObP1+aiWRkQK9eUke5UycJjtPToaxM9vn9EhxnZLjVtfx+ubarrEzul5fD3Xe7wXFjXueVmyuf/dRTcnvhBRl/bYJBeOstibHy8yWN9Yc/lEYmzYKTajFsmHzb2bdPvjXs3ClXSHq/XSW6skVVAwZIWsU778iYa3P0KNx5p5tb1EzVtIKcle6usOsKcuzpCrJSSiUZ72qyo6hIVnxnz5a839mzZRFx6VL40Y9gwwYJgo2R4HfkSFi+XFZ3rXWDZ5DnOYHy5ZfLey9YICuzCxfK/gkTKm/HIiUjkrw8CZQffFDab4djLRQXV4+Xtm+HCy6Q3+O668KvJh93nKSqpEy/isJCmDsXrrxSUhcyM+WnMfD3v7tpFvn5ybOK7Bg0CObMgd/8JnLLx/37JQdnwwaZnIMGNe4Yk0i0K8hHy2PcY15pFQullGoqwnXkGzdOYiWQmMJZNS4tlbgpI0NSM+bMkef4/e4qsxMwO6kXhw+7FwaWlFTe9vslEH/vPTeX+bzzqgfUeXludTInoI7V2f69e6V4w6FDcqHiX/8K334b3WtbtJB+FgMHSrO3MWMi5zontXCd92bOlC57IEHzT3/q5iMnQwm4urrmGsnBccyfL4W5oznNkGJ2HtzJ+X85n1237Kr22Os7X2fUc6MAGNljJG9c2VxOpcSWlnlTSqkUEqlkrmP1amkD3bu3dC5OS4N+/WRRrrxccpCHD5cVaJCz3sGgxCDWSrDt3XYeh5oDaue5mZmSQ+0tRRfrxcz9++XCvvnz6/7aFi0kn7ldO2md3auXfJno31/6WiRtqd6qtQGHDZOmIXfeKX/YZKuPXB/r1lVPRr/++up1lZuB4u+KyXsij8/u+KzaY2/tfouh84YCMPj4waz+6erGHVyK0ABZKaWaofx8N4B2gunevd2UiZdekrRQgK5d3cYfkbZBYrBgUIJnpwSds+0E0Z07S7rEvfe6149BfK4dW7MG/ud/4Isvqj9WUSG/s3NBYDScVJVIj40YIQu0HTpIjNqxY/3GHRPOivKiRXD//XI/K0uuxGyqATLI1Zf/+EflfS+9JIXCm5FvS7+lS0EXDt11qNpj7+x5hzP/dCYAA7sMZN3P1jX28FKCBshKKaUqKSyUKhaBgNzKyirnLFfd9vnkeU5qRp8+sHVr9e3cXKlU4fPJYqZTeCGRZ/23bZNF1QMH5EK/f/87Nu/booX8Pm3byheO3r3dqh6dO8fmM2rkrCiDJKX37Cn1/SZMaNq1houL4ZlnpNC3E6fk5MjEysys+bUppCJYgf9+PxUzKjBVvrVt2r+JvCfyAOh3bD/e+/l7CRhh06cBslJKqe85wbExkgoxY0bNOcjZ2XDVVdVzmcMF1OXlcPLJskoNkhL76qvJkxJrrZzFLy6W0rz/+pf8/Ppr2X+o+mJdnXXtCh980Eh1m51V5Jtvhk2b4Je/bNo5yF5ffSVXWB44IPeLimouD5eCMv9vJl9N+4osf+W60B9++SG9/9AbgJM6nMSOX+5IxPCavEgBcvPLeFdKKUVRkaRdOKkXkyfXXMXCKZyQne3mMoME2M7CljGSt2wMfPihW0lj7lxZCEyWWM2Yyimu117rbgeD7gp5OLt2SZm5HTtkoXbLlvDP27tXVs6vuioWI66FUxfQWikqPWyY23UmWQ56fbVvD337wsqVcr+2bjQpyKlkUTVA1kYh8aUBslJKNUPePGBvDHXbbeG3AwGp+uAE1KeeGj6gPvZYqcE8cqQ8PytLqnk99phbwjeZ+XyyEh5Jnz7wyCPu/Y0bZaXcaQH+29+6jy1c2EgBsvPHfPddWQaHpnGwo9Whg7vdHAPkUC3knJY5lfZrgBxfGiArpZSqVbQBdX6+u2/8eHjlFUmNvfTS1Dnr75WXJzfHTTdBjx6y/cYbEs9547u4atfODZBTifcqyOYYIEeohawBcnxpJz2llFIx4zR7c876jxwpfR4mTXLP+qey7t3dvhaBgMR2d9zRSA3h2reXnN1UoyvIYbvpaYAcXxogK6WUirmpUyUYLiyEs86Cf/5TAueBA2WVOZVNmFD5fkGBlKKLuzZtpGGIU2svVTT3ADnCCrLf58dnJIwLBAMEgoHGHlpK0wBZKaVUXAwcKGkVrVpJgOwUW3CqkqWqa6+VXGWvF19shA92moR8800jfFgjau4BcoQVZGOMriLHkQbISiml4sIppvDII9LZLxVzkMNp106qW3i7Jb/0UiOmWaRaHnJzD5AjrCCDplnEkwbISiml4mbYMGlNXVoqK6upHhw7jJG86/bt5f6nn8KKFdLZz7nFRSpeqNfcA2R/NkfKj4R9TAPk+NEAWSmlVNwUFsLjj0uM89RTcr+58Pulgodj5EipE+3czj9fSuDFVLt2qXehXjMPkFv6W4ZNsQANkONJA2SllFJx4eQcL1oE/fvD9OlyvzkFyePHR35s+XJpOhJTmmKRcrL9kVMsstLd5iEaIMeWBshKKaXiwin1NmwYdOsGxxzTPEq9ef3kJzBtmrTezsyURiRO50GQxiJlZTH8wFRMsXDyVEAC5EZJ5k4e2RnhL9KDyivIR8tjfTqiedMAWSmlVFw4NZEBunaFPXvkvrfpSKozBh56SLrtHT0quceHD0vHQZBj8txzMfzAVKyFnJUlN5A+4IfDB4upqqYVZE2xiB8NkJVSSsVd166wd2+iR5EcsrLgv//bvf+nP8XwzVNxBRmadZpFtCvIGiDHlgbISiml4q5bNw2Qva65xt3euDGGVS00QE45uoKcGBogK6WUijtdQa6sc2e5gWQMfPRRjN44FVMsoHkHyFoHOSE0QFZKKRV3Tg6ycuXludsbN8boTXUFOeVE6qQHkOXXKhbxogGyUkqpuOvUSRY2Y1qxoYmLW4CsK8gppcYV5DRPFYuAVrGIJQ2QlVJKxV1amgTJxcWJHknyiEuAnIp1kKF5B8jaSS8hjE2SeoLGmFHAo0Aa8LS19qGann/66afb9evXN8rYlFJK1V9+vuTYrl4tZd7GjYMFC+T+kCHynI0bYcIEWLhQ7id62wle4znGefPgvffkfno6nHZaw8Z4/sZ8Wu7cxNCS13m641S6lXxE+x/3puMKeeGXwyckdPurXBlw149Ws7enHNT2n26M6j1a5ubQc9v/AvCfnkP5Zv+RpBtjrLcDFWV8dPYouq/5O+XBAHuGXEL3Nf/LzpzjATi1+APWHtOdsopS+n99kFX9fsh5W7YC8P6gwfRb93adtvM2vonPX8ZHQ4fQfflafP4yvvmv/nR8dYWM65Lh1bZ35LakYvwVpC1azN5jMuk87GL2Fy5j8zU/AWD1J6sZcoIcx437NzKh7wQWbpbf09nO65zHxL4TWbB5AT079GRgl4EU7Sti6tmNUw/SGLPBWnt6tf3JECAbY9KAD4GRwB6gCJhkrf0g0ms0QFZKqaahsBDGjoUjR2DSJFi6FEpKoLTUbZ4RDMq+zNCCWKK309IgEIj/GEs8i37OZxgjt9q2nedbK9tDbSEv2bG04Vtup4BZzCCbw5QgH5pJSUK3K0gjjQAtKKWUTCrwkUYw6vdw+qtYIIgBDD6CMd22+IAgaViC+Agip9pj/TnJO0bwYUPbzieF5meYbWvk74GBef1hzHaYMA5W90gjaINYLD7j+/4PJ58i/3O2fcb3/fPuGXIPv1/3exZdsYhh3YfRGJI9QP4xMNNae0Ho/nQAa+3sSK/RAFkppZqOwkIYNUr6PIAEdTk58MUXEtxZ6wZ6ybQd7zEGgzUft7oaykpWMpwKfKFQSoIokmj7c3I4hi/q/DpT/ddVSeS7DBg9CVZ1r/97dMjqwOJxixstOIbkD5CvAEZZa68P3b8SOMNae3OV500BpoTu9ga2N+pARQ7wRQI+t6nS41U3erzqRo9X3ST4eOV2gWOPk+0jh6BlK/end1+ybSf68+s2xi7sLjuOLzMADpFV0Yqjacm2nejPb5JjNFm0snIhnrMdbl+Dt30ZtlWwzAAcSqeiVYC0aLaLW8G+TA7hR+ZmeWi7PMy+SNuHKeZb9lFdPP/tOsFae0zVnelx+rC6CvfFsFrkbq19Engy/sOJzBizPtw3DRWeHq+60eNVN3q86kaPV/T0WNWNHK8jeryipMerbhLx32OyVLHYA+R67neDsN8glFJKKaWUiqtkCZCLgF7GmO7GmAxgIrA0wWNSSimllFLNUFKkWFhrA8aYm4HXkTJvz1hrtyR4WJEkNMWjCdLjVTd6vOpGj1fd6PGKnh6rutHjVTd6vOqm0Y9XUlykp5RSSimlVLJIlhQLpZRSSimlkoIGyEoppZRSSnlogBwlY8woY8x2Y8xOY8ydiR5PMjLG7DbGvG+M2WiMWR/a18EYs9wYsyP0s32ix5koxphnjDEHjDGbPfsiHh9jzPTQfNtujLkgMaNOnAjHa6YxZm9ojm00xlzoeay5H69cY0yhMWarMWaLMeaW0H6dY2HUcLx0joVhjMk0xqwzxmwKHa/7Qvt1flVRw7HSuVUDY0yaMebfxphlofuJnVvWWr3VckMuHPwI6AFkAJuAUxI9rmS7AbuBnCr78oE7Q9t3Ag8nepwJPD5DgNOAzbUdH+CU0DxrAXQPzb+0RP8OSXC8ZgJ3hHmuHi84DjgttN0a+DB0XHSO1e146RwLf7wM0Cq07QfeAc7U+VWnY6Vzq+bjdhvwV2BZ6H5C55auIEdnELDTWrvLWlsGLADGJHhMTcUYYF5oex4wNnFDSSxr7WrgYJXdkY7PGGCBtbbUWvsxsBOZh81GhOMViR4va4utte+Gtr8DtgJd0TkWVg3HK5LmfrystfZQ6K4/dLPo/KqmhmMVSbM9Vg5jTDfgIuBpz+6Ezi0NkKPTFfjUc38PNf9D2lxZ4A1jzIZQW3CATtbaYpD/QwKOTdjoklOk46NzLrKbjTHvhVIwnFNuerw8jDEnAgOQlSudY7WocrxA51hYoVPgG4EDwHJrrc6vCCIcK9C5FcnvgKlA0LMvoXNLA+ToRNUKW3G2tfY04CfAL4wxQxI9oCZM51x4c4CeQB5QDBSE9uvxCjHGtAJeBG611n5b01PD7Gt2xyzM8dI5FoG1tsJam4d0ux1kjOlbw9Ob9fGKcKx0boVhjLkYOGCt3RDtS8Lsi/nx0gA5OtoKOwrW2n2hnweAJcgpj8+MMccBhH4eSNwIk1Kk46NzLgxr7Weh/+MJAk/hnlbT4wUYY/xIsPectfal0G6dYxGEO146x2pnrf0aWAWMQudXjbzHSudWRGcDo40xu5EU1vOMMc+S4LmlAXJ0tBV2LYwx2caY1s42cD6wGTlOV4eedjXwSmJGmLQiHZ+lwERjTAtjTHegF7AuAeNLKs4/liGXInMM9HhhjDHAn4Ct1tr/53lI51gYkY6XzrHwjDHHGGPahbazgBHANnR+VRPpWOncCs9aO91a281aeyISX6201v4XCZ5bSdFqOtnZptUKO1E6AUvk/3NIB/5qrX3NGFMELDLGXAf8BxiXwDEmlDHmeWAokGOM2QPcCzxEmONjrd1ijFkEfAAEgF9YaysSMvAEiXC8hhpj8pDTabuBG0CPV8jZwJXA+6HcR4C70DkWSaTjNUnnWFjHAfOMMWnI4toia+0yY8y/0PlVVaRj9RedW3WS0H+7tNW0UkoppZRSHppioZRSSimllIcGyEoppZRSSnlogKyUUkoppZSHBshKKaWUUkp5aICslFJKKaWUhwbISimVJIwxfzbGrI/B+1xsjLGhFspKKaXqSOsgK6VU8rgfyEr0IJRSqrnTAFkppZKEtfajRI9BKaWUplgopVTS8KZYGGOuCaVJ9DPGLDfGHDbGbDPGXFblNcYYM9MYc8AY850xZj7QJsx7Zxpj8o0xnxpjSo0xm4wxF3oe/z/GmBJjzCmefYOMMQFjzPVx/LWVUirpaICslFLJ7a/AUuBSYAewwBjTzfP4r4AZwJPAFcBRID/M+ywGrgEeBC4BioCloda3AAXAeqRFbroxJhOYB7xhrX06xr+TUkolNU2xUEqp5PaItfYZAGPMBuAz4GLgcWNMGjANeMJae3fo+a8bY5YDXZ03MMYMBy4Chlpr3wrtfsMYczLwa2CctTZojLkG2AhMB9oDnYDhcf79lFIq6WiArJRSye0NZ8Na+6Ux5gDgrCDnAscBr1R5zUvACM/9EcB+4B/GGO+/+yuQVWXn/XcaY6YBjwBpwFXW2n0x+j2UUqrJ0ABZKaWS29dV7pcBmaHtzqGfB6o8p+r9nNBzy8O8f0WV+y8iAfJB4IW6DFQppVKFBshKKdV07Q/9PLbK/qr3DwJ7gbFRvOfjwH9C7zETuKv+w1NKqaZJA2SllGq6PkWC5DHAa579l1V53grgduCQtXZbpDczxlyF5DefA/QD5hhjXrbWrovpqJVSKslpgKyUUk2UtbbCGJMP/NYY8wXwNnA50KfKU5cDrwPLjTEPA1uQUnB5QKa1dnqoMsajwG+stWuBtaGScvOMMQOstSWN81sppVTiaZk3pZRq2n6HlG67EckfbgVM9T7BWmuRVeVngFuRYPkJ4MfAmtDTngb2APd6Xno9krv8QLwGr5RSycjIv5tKKaWUUkop0BVkpZRSSimlKtEAWSmllFJKKQ8NkJVSSimllPLQAFkppZRSSikPDZCVUkoppZTy0ABZKaWUUkopDw2QlVJKKaWU8tAAWSmllFJKKY//D5n2+Hzmhe9hAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot for % comparison\n",
    "plt.rcParams['figure.figsize'] = [10, 10]\n",
    "\n",
    "plt.plot(list(SHAP_df1.index), SHAP_df1['percent significant'], label= label1, linestyle='-', color='green', linewidth=3)\n",
    "plt.plot(list(SHAP_df2.index), SHAP_df2['percent significant'], label= label2, linestyle='-', marker='x', color='green', linewidth=1)\n",
    "plt.plot(list(SHAP_df3.index), SHAP_df3['percent significant'], label= label3, linestyle='-', color='blue', linewidth=3)\n",
    "plt.plot(list(SHAP_df4.index), SHAP_df4['percent significant'], label= label4, linestyle='-', marker='x', color='blue', linewidth=1)\n",
    "plt.plot(list(SHAP_df5.index), SHAP_df5['percent significant'], label= label5, linestyle='-', color='red', linewidth=3)\n",
    "plt.plot(list(SHAP_df6.index), SHAP_df6['percent significant'], label= label6, linestyle='-', marker='x', color='red', linewidth=1)\n",
    "plt.axhline(y=10, color='black', linestyle='--', label='y = 10')\n",
    "\n",
    "# Add axis labels\n",
    "plt.xlabel('index', fontsize=15)\n",
    "plt.ylabel('percent significant (%sig)', fontsize = 15)\n",
    "\n",
    "# Add a title\n",
    "plt.title('drugs/features', style='italic', fontsize=20)\n",
    "\n",
    "plt.legend(fontsize=20)\n",
    "\n",
    "plt.ylim(0, 90)\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.savefig(f'Targets/%sig/{test_type1}_%significant_comparison_final2')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "5c9878b8-f255-4156-b583-0a91375c61f0",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics1 = pd.read_csv(f'Targets/Deep_NN/model performance/proteomic/drugs_features_proteomic.csv')\n",
    "metrics2 = pd.read_csv(f'Targets/Deep_NN/model performance/full_network/drugs_features_full_network.csv')\n",
    "metrics3 = pd.read_csv(f'Targets/rfr/model performance/proteomic/drugs_features_proteomic.csv')\n",
    "metrics4 = pd.read_csv(f'Targets/rfr/model performance/full_network/drugs_features_full_network.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "94e9d442-89f4-47b9-ad57-085186f60a3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot 2\n",
    "\n",
    "plt.plot(metrics1['rs'].astype(str), metrics1['model_r2'], linestyle='-', marker='x', color='blue', linewidth=1)\n",
    "plt.plot(metrics2['rs'].astype(str), metrics2['model_r2'], linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "plt.plot(metrics3['rs'].astype(str), metrics3['model_r2'], linestyle='-', marker='x', color='green', linewidth=1)\n",
    "plt.plot(metrics4['rs'].astype(str), metrics4['model_r2'], linestyle='-', marker='.', color='green', linewidth=1)\n",
    "\n",
    "plt.title('performance')\n",
    "plt.xlabel('seed')\n",
    "plt.ylabel('r-squared')\n",
    "plt.ylim(0.5, 0.85)\n",
    "\n",
    "plt.savefig('Targets/drugs_features_performance_comparison.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "575e779e-3f60-4f78-913a-d6b45e774281",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## phospho comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "e236aff9-d445-44ed-a364-d1d1f3bbdcfd",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_type1 = 'features'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "9c7cf5a5-fe2c-41a6-acec-1923589c45a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'phospho'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df1 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label1 = f'RF_{analysis}_Phospho'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "83d0a31d-f547-45d1-b23f-6210c3a4eb8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df2 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label2 = f'RF_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "14735bb2-dd42-4bc9-912b-7731565d88e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df3 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label3 = f'RF_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "78037d19-4866-41d1-afa0-bdc681296212",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'phospho'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df4 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label4 = f'DNN_{analysis}_Phospho'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "cc8f27a9-d201-475a-9c1a-ebcc681eb681",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df5 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label5 = f'DNN_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "ed0294ac-819a-44a8-8d03-18778cd709c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df6 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label6 = f'DNN_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "2287750b-7685-4dbd-951b-2441da65c095",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'phospho'\n",
    "analysis = 'IG'\n",
    "SHAP_df7 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label7 = f'DNN_{analysis}_Phospho'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "fe4975a8-22fc-47d7-81b9-1f9a33605579",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'IG'\n",
    "SHAP_df8 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label8 = f'DNN_{analysis}_Proteomic'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "68124b7d-e0cb-4da1-b27e-75a05ac4a853",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'IG'\n",
    "SHAP_df9 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "label9 = f'DNN_{analysis}_Combined_data'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "e2161441-70d4-4f89-ae89-13949f788cd1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plot for % comparison\n",
    "plt.rcParams['figure.figsize'] = [10, 10]\n",
    "\n",
    "plt.plot(list(SHAP_df1.index), SHAP_df1['percent significant'], label= label1, linestyle='-', marker='.', color='green', linewidth=1)\n",
    "#plt.plot(list(SHAP_df2.index), SHAP_df2['percent significant'], label= label2, linestyle='-', marker='x', color='green', linewidth=1)\n",
    "#plt.plot(list(SHAP_df3.index), SHAP_df3['percent significant'], label= label3, linestyle='-', color='green', linewidth=3)\n",
    "plt.plot(list(SHAP_df4.index), SHAP_df4['percent significant'], label= label4, linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "#plt.plot(list(SHAP_df5.index), SHAP_df5['percent significant'], label= label5, linestyle='-', marker='x', color='blue', linewidth=1)\n",
    "#plt.plot(list(SHAP_df6.index), SHAP_df6['percent significant'], label= label6, linestyle='-', color='blue', linewidth=3)\n",
    "plt.plot(list(SHAP_df7.index), SHAP_df7['percent significant'], label= label7, linestyle='-', marker='.', color='red', linewidth=1)\n",
    "#plt.plot(list(SHAP_df8.index), SHAP_df8['percent significant'], label= label8, linestyle='-', marker='x', color='red', linewidth=1)\n",
    "#plt.plot(list(SHAP_df9.index), SHAP_df9['percent significant'], label= label9, linestyle='-', color='red', linewidth=3)\n",
    "plt.axhline(y=10, color='black', linestyle='--', label='y = 10')\n",
    "\n",
    "# Add axis labels\n",
    "plt.xlabel('index', fontsize=15)\n",
    "plt.ylabel('percent significant (%sig)', fontsize=15)\n",
    "\n",
    "# Add a title\n",
    "plt.title('features', style='italic', fontsize=20)\n",
    "\n",
    "plt.legend(fontsize=22)\n",
    "\n",
    "plt.ylim(0, 100)\n",
    "plt.xlim(-20, 200)\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.savefig(f'Targets/%sig/phospho_features_%significant_comparison_final2')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5e8bab1-cc57-48cb-b3ae-9da5959c3319",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# Integrated gradients for targets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b2a1f23e-050c-4a66-a42b-578dc0dedc18",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = 23)\n",
    "print(f'dd_X_main.shape: {dd_X_main.shape}')\n",
    "print(f'X_train.shape: {X_train.shape}, X_test.shape: {X_test.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "53fc34bf-eddb-46a5-8dcf-221bb7f79f4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = 23)\n",
    "xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = 23)\n",
    "y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = 23)\n",
    "\n",
    "print(f'dd_x_all.shape: {dd_x_all.shape}')\n",
    "print(f'xo_train.shape: {xo_train.shape}, X_test.shape: {xo_test.shape}')\n",
    "\n",
    "print(f'dd_x_drug.shape: {dd_x_drug.shape}')\n",
    "print(f'xd_train.shape: {xd_train.shape}, X_test.shape: {xd_test.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ac6966f4-727f-4cb6-9658-ce066a1e8c10",
   "metadata": {},
   "source": [
    "## build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54f036dd-607a-49ea-ab8b-0b6ace198eab",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "## fit the model\n",
    "\n",
    "model = build_Deep_NN(xo_train, xd_train, learning_rate, momentum, initializer_seed, complexity = 4)\n",
    "history = model.fit([xo_train, xd_train], y_train,\n",
    "                    validation_data=([xo_test, xd_test], y_test),\n",
    "                    epochs=num_epochs, \n",
    "                    batch_size=None, \n",
    "                    verbose=0, \n",
    "                    callbacks=callbacks_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d6116a41-f023-4a43-aca6-a8ac107e5efa",
   "metadata": {},
   "outputs": [],
   "source": [
    "#initialise IG explainer\n",
    "ig  = IntegratedGradients(model,\n",
    "                          layer=None,\n",
    "                          target_fn=None,\n",
    "                          method=\"gausslegendre\",\n",
    "                          n_steps=50,\n",
    "                          internal_batch_size=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b39ce207-fb87-4837-8dec-8321713d2f99",
   "metadata": {},
   "source": [
    "## test model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c4eeea26-5607-4a02-ba41-e68e2e03e905",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "indexes, targets, percent_sig, sig_targs, insig_targs = IG_targets_NN(dd_X_main, ig, [xo_test, xd_test], y_test, dtype = 'mix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f4a27514-fb43-4ede-87a7-047e347b8d8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_targs = [len(x) for x in targets]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3aa0492e-4d50-4046-9c8c-b0a44df59b35",
   "metadata": {},
   "outputs": [],
   "source": [
    "IG_df = table_make('IG',\n",
    "           [indexes, targets, number_of_targs, percent_sig, sig_targs, insig_targs],\n",
    "           ['cl:drug', 'drug targets', 'number of targets', 'percent significant', 'significant targets', 'insignificant targets'], file = False)\n",
    "IG_df = IG_df.sort_values('percent significant', ascending=False)\n",
    "IG_df.to_csv(f'Targets/{model_type}/explainability/IG/{dtype}/features_{dtype}.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a386d8e-5823-489b-92f3-8ed1192b2746",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "# EMDRs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b62df1a4-2172-4f99-bc9b-a5562a8cb03a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "X_train, X_test, y_train, y_test = cell_line_split(dd_X_main, dd_y_main, test_size=0.2, random_state = 23)\n",
    "print(f'dd_X_main.shape: {dd_X_main.shape}')\n",
    "print(f'X_train.shape: {X_train.shape}, X_test.shape: {X_test.shape}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cc6b4221-fe2d-484b-a472-deee77385266",
   "metadata": {},
   "outputs": [],
   "source": [
    "#train-test-split\n",
    "xo_train, xo_test = cell_line_split(dd_x_all, y=None, test_size=0.2, random_state = 23)\n",
    "xd_train, xd_test = cell_line_split(dd_x_drug, y=None, test_size=0.2, random_state = 23)\n",
    "y_train, y_test = cell_line_split(dd_y_main, y=None, test_size=0.2, random_state = 23)\n",
    "\n",
    "print(f'dd_x_all.shape: {dd_x_all.shape}')\n",
    "print(f'xo_train.shape: {xo_train.shape}, X_test.shape: {xo_test.shape}')\n",
    "\n",
    "print(f'dd_x_drug.shape: {dd_x_drug.shape}')\n",
    "print(f'xd_train.shape: {xd_train.shape}, X_test.shape: {xd_test.shape}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50040606-48ed-436c-953e-81746e0821b6",
   "metadata": {
    "tags": []
   },
   "source": [
    "## build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "80b421ed-b285-49ce-9f37-c3deb99e7722",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%time\n",
    "\n",
    "if model_type == 'Deep_NN':\n",
    "    ## fit the model\n",
    "\n",
    "    model = build_Deep_NN(xo_train, xd_train, learning_rate, momentum, initializer_seed, complexity = 4)\n",
    "    history = model.fit([xo_train, xd_train], y_train,\n",
    "                        validation_data=([xo_test, xd_test], y_test),\n",
    "                        epochs=num_epochs, \n",
    "                        batch_size=None, \n",
    "                        verbose=0, \n",
    "                        callbacks=callbacks_list)\n",
    "    shap.explainers._deep.deep_tf.op_handlers[\"AddV2\"] = shap.explainers._deep.deep_tf.passthrough\n",
    "\n",
    "    # DeepExplainer to explain predictions of the model\n",
    "    explainer = shap.DeepExplainer(model, [np.array(xo_train), np.array(xd_train)])\n",
    "\n",
    "elif model_type == 'rfr':\n",
    "\n",
    "    #run rfr\n",
    "    classify = RandomForestRegressor(n_jobs=-1, max_depth=300, n_estimators=200)\n",
    "\n",
    "    #fit model\n",
    "    classify.fit(X_train, y_train)\n",
    "\n",
    "    #creates explainer instance\n",
    "    explainer = shap.explainers.Tree(classify, X_train)\n",
    "\n",
    "elif model_type == 'xgb':\n",
    "    #run xgb\n",
    "    xgb = XGBRegressor(max_depth = 75, \n",
    "                       n_estimators = 300, \n",
    "                       seed = 42, \n",
    "                       min_child_weight = 3, \n",
    "                       gamma = 0, \n",
    "                       colsample_bytree = 0.3, \n",
    "                       reg_alpha = 0.1,\n",
    "                       n_jobs=-1)\n",
    "\n",
    "    #fit model\n",
    "    xgb.fit(X_train, y_train)\n",
    "\n",
    "    #creates explainer instance\n",
    "    explainer = shap.explainers.Tree(xgb, X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f3699ee6-f982-4e20-a73b-a64df1ac58fc",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## significant EMDR finder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b006563-4a16-4585-9b52-4d927c41a39d",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "indexes, features, targets = SHAP_EMDR_NN(dd_X_main, explainer, [xo_test, xd_test], y_test, dtype = 'proteomic', strict=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54ecec1b-b5b4-44c7-ba2e-5576d95bdf98",
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_targs = [len(x) for x in targets[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dbf60f3b-64b4-4dfb-80a3-2139c10ceae9",
   "metadata": {},
   "outputs": [],
   "source": [
    "SHAP_df = table_make('SHAPs',\n",
    "           [indexes, features[0], features[1], features[2], features[3], targets[0], number_of_targs, targets[1], targets[2], targets[3]],\n",
    "           ['cl:drug',' total number of features', 'EMDR percentage of total', 'total number of significant features', 'EMDR percentage of significant', 'EMDRs', 'number of EMDRs', 'percent significant', 'significant EMDRs', 'insignificant EMDRs'], file = False)\n",
    "SHAP_df = SHAP_df.sort_values('percent significant', ascending=False)\n",
    "SHAP_df.to_csv(f'EMDR/{model_type}/explainability/SHAP/{dtype}/{test_type}_{dtype}.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "08671e34-4c59-4571-a41e-fdd489e86d20",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## significance plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be79f890-d7b6-45f6-be84-82eea015afad",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'phospho'\n",
    "test_type1 = 'drugs'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df1 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}_notStrict.csv')\n",
    "label1 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "18a75c7e-a9b8-44fc-a9ee-f345bef4f364",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "test_type1 = 'drugs_features'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df2 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}.csv')\n",
    "label2 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5682f059-f1ef-42ef-ab25-464ab15394ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "test_type1 = 'drugs'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df3 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}.csv')\n",
    "label3 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "021faf57-9cad-4a00-9599-0b94e829a057",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "test_type1 = 'drugs_features'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df4 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}.csv')\n",
    "label4 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a171e062-296b-4945-99ae-879b8dd53803",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "test_type1 = 'drugs'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df5 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}.csv')\n",
    "label5 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25f67a8e-8046-454e-bbe5-6e05d3a65b0a",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'EMDR'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "test_type1 = 'drugs_features'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df6 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}.csv')\n",
    "label6 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "25e37a0d-3503-466e-9909-53c4aa2a5e1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot for % comparison\n",
    "plt.rcParams['figure.figsize'] = [15, 15]\n",
    "\n",
    "plt.plot(list(SHAP_df1.index), SHAP_df1['percent significant'], label= label1, linestyle='-', marker='.', color='green', linewidth=1)\n",
    "#plt.plot(list(SHAP_df2.index), SHAP_df2['percent significant'], label= label2, linestyle='-', marker='x', color='green', linewidth=1)\n",
    "plt.plot(list(SHAP_df3.index), SHAP_df3['percent significant'], label= label3, linestyle='-', marker='.', color='blue', linewidth=1)\n",
    "#plt.plot(list(SHAP_df4.index), SHAP_df4['percent significant'], label= label4, linestyle='-', marker='x', color='blue', linewidth=1)\n",
    "plt.plot(list(SHAP_df5.index), SHAP_df5['percent significant'], label= label5, linestyle='-', marker='.', color='red', linewidth=1)\n",
    "#plt.plot(list(SHAP_df6.index), SHAP_df6['percent significant'], label= label6, linestyle='-', marker='x', color='red', linewidth=1)\n",
    "plt.axhline(y=10, color='black', linestyle='--', label='y = 10')\n",
    "\n",
    "# Add axis labels\n",
    "plt.xlabel('index')\n",
    "plt.ylabel('percent significant')\n",
    "\n",
    "# Add a title\n",
    "plt.title('% significant features determine by post-hoc analysis')\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.ylim(0, 80)\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.savefig('EMDR/EMDR_drugs_%significant_comparison')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "427c7a08-0baa-4cb9-8fab-1a49fe030521",
   "metadata": {
    "tags": []
   },
   "source": [
    "# significant drug plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9a69207-9960-40ce-9995-69700299ca00",
   "metadata": {},
   "outputs": [],
   "source": [
    "#set target\n",
    "targ = 'Fulvestrant'\n",
    "test_type1 = 'features'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "024fa8a5-f7a1-46ad-87be-cab20a9912f4",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df1 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes1, sigs1 = drug_or_cl_sig(targ, SHAP_df1, isDrug=True)\n",
    "\n",
    "label1 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "28f0900c-89d8-4d1d-84e0-627865360e1e",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'rfr'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df2 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes2, sigs2 = drug_or_cl_sig(targ, SHAP_df2, isDrug=True)\n",
    "\n",
    "label2 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45901b3b-95d8-4196-8c53-bbf0379f9810",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df3 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes3, sigs3 = drug_or_cl_sig(targ, SHAP_df3, isDrug=True)\n",
    "\n",
    "label3 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4f63bbc9-109d-4455-9fe9-f287bcab09ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'SHAP'\n",
    "SHAP_df4 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes4, sigs4 = drug_or_cl_sig(targ, SHAP_df4, isDrug=True)\n",
    "\n",
    "label4 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "494806c1-e118-451c-87e7-5596eae257c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'full_network'\n",
    "analysis = 'IG'\n",
    "SHAP_df5 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes5, sigs5 = drug_or_cl_sig(targ, SHAP_df5, isDrug=True)\n",
    "\n",
    "label5 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a1842bef-f728-42fb-889e-e3fa47c303ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "method = 'Targets'\n",
    "model_type1 = 'Deep_NN'\n",
    "dtype1 = 'proteomic'\n",
    "analysis = 'IG'\n",
    "SHAP_df6 = pd.read_csv(f'{method}/{model_type1}/explainability/{analysis}/{dtype1}/{test_type1}_{dtype1}2.csv')\n",
    "\n",
    "indexes6, sigs6 = drug_or_cl_sig(targ, SHAP_df6, isDrug=True)\n",
    "\n",
    "label6 = f'{model_type1}_{analysis}_{dtype1}_{test_type1}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae47cec1-5a10-43ee-b021-00d0b6c5a60c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot for % comparison\n",
    "fig, axes = plt.subplots(1, 2, figsize=(15, 10))\n",
    "\n",
    "plt.scatter(indexes1, sigs1, label= label1, linestyle='-', marker='.', color='green', linewidth=3)\n",
    "plt.scatter(indexes2, sigs2, label= label2, linestyle='-', marker='x', color='green', linewidth=3)\n",
    "plt.scatter(indexes3, sigs3, label= label3, linestyle='-', marker='.', color='blue', linewidth=3)\n",
    "plt.scatter(indexes4, sigs4, label= label4, linestyle='-', marker='x', color='blue', linewidth=3)\n",
    "plt.scatter(indexes5, sigs5, label= label5, linestyle='-', marker='.', color='red', linewidth=3)\n",
    "plt.scatter(indexes6, sigs6, label= label6, linestyle='-', marker='x', color='red', linewidth=3)\n",
    "plt.axhline(y=10, color='black', linestyle='--', label='y = 10')\n",
    "\n",
    "# Add axis labels\n",
    "plt.xlabel('significance rank')\n",
    "plt.ylabel('percent significant')\n",
    "\n",
    "# Add a title\n",
    "plt.title(f'Distribution of significance for {targ}')\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.ylim(0, 70)\n",
    "plt.xlim(-20, 400)\n",
    "\n",
    "# Adjust spacing between subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "#plt.savefig(f'Targets/example_targs/{test_type1}_distribution_significance_{targ}')\n",
    "\n",
    "# Display the plots\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}