graphpattern2vec_process-multithread-PREDICTION-Edited2022.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### To make prediction:\n",
    "  - We will use **graphpattern2vec_process-multithread-PREDICTION-Edited2022** notebook\n",
    "    - All the cells should be run until **Link Prediction**.\n",
    "    - Stop using the notebook, and manually run the metapath2vec to generate walk files.\n",
    "    - Run all the cells after **Link Prediction** until *prediction*\n",
    "      - Do not use the figures generated in this step for rvaluation purposes.\n",
    "    - To make prediction run all remaining steps."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "from graphpattern2vec import split_data as sp\n",
    "from graphpattern2vec import biased_walk_multiprocess as BWMP\n",
    "# from graphpattern2vec import check_walks as CW\n",
    "from graphpattern2vec import CheckWalks as CW\n",
    "from graphpattern2vec import link_prediction as LP"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_path = 'data/'\n",
    "main_edge_file = 'df_edges.csv'\n",
    "relation_to_predict = 'hasPathway'\n",
    "train_file = 'df_edges.csv'\n",
    "\n",
    "node_file = 'df_node.csv'\n",
    "relation_file = 'df_relation.csv'\n",
    "type_file = 'df_type.csv'\n",
    "test_file = ''\n",
    "\n",
    "result_path = 'results/'\n",
    "\n",
    "source_types = ['Protein', 'LightKinase', 'DarkKinase']\n",
    "target_types = ['Pathway']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Walks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Types\n",
    "- 0,Protein\n",
    "- 1,Pathway\n",
    "- 2,LightKinase\n",
    "- 3,DarkKinase\n",
    "\n",
    "- 4,PTM\n",
    "- 5,FunctionalDomain\n",
    "- 6,BiologicalProcess\n",
    "- 7,MolecularFunction\n",
    "- 8,CellularComponent\n",
    "- 9,Chemical\n",
    "- 10,Disease"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'graphpattern2vec.biased_walk_multiprocess' from '/hdd/BigIronOnHisHip/r2v/newRegpatterin/RegPattern2Vec_project-main/notebooks/graphpattern2vec/biased_walk_multiprocess.py'>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import imp\n",
    "imp.reload(BWMP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{0: 1, 2: 1, 3: 1},\n",
       " {0: 1, 2: 1, 3: 1, 4: 1, 5: 1, 6: 1, 7: 1, 8: 1, 9: 1, 10: 1},\n",
       " {0: 1, 2: 1, 3: 1},\n",
       " {1: 1}]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph_pattern = [{0:1,2:1,3:1}, {0:1, 2:1, 3:1, 4:1,5:1, 6:1, 7:1, 8:1, 9:1, 10:1}, {0:1,2:1,3:1}, {1:1}]\n",
    "graph_pattern"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loading data ...\n",
      "df_train: (6279373, 5)\n",
      "based on graphpattern df_train_sel: (6279373, 5), set_sel_types: 11\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "    \u001b[0;31m[... skipping hidden 1 frame]\u001b[0m\n",
      "\u001b[0;32m<ipython-input-26-5dee829fc658>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m biasedwalk = BWMP.BiasedWalk(graph_pattern, loop_index, type_loop, data_path = data_path, \n\u001b[0m\u001b[1;32m     11\u001b[0m                             train_file = train_file, postfix='', num_processes = num_processes)\n",
      "\u001b[0;32m/hdd/BigIronOnHisHip/r2v/newRegpatterin/RegPattern2Vec_project-main/notebooks/graphpattern2vec/biased_walk_multiprocess.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, graph_pattern, loop_index, type_loop, data_path, train_file, postfix, degree_biased, num_processes)\u001b[0m\n\u001b[1;32m     37\u001b[0m         \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'loading data ...'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 38\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     39\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/hdd/BigIronOnHisHip/r2v/newRegpatterin/RegPattern2Vec_project-main/notebooks/graphpattern2vec/biased_walk_multiprocess.py\u001b[0m in \u001b[0;36mload_data\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    157\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 158\u001b[0;31m         \u001b[0mh_nodes\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mh_id\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdf_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mh_c_id\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    159\u001b[0m         \u001b[0mt_nodes\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mt_id\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdf_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mt_c_id\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: ",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# loop_index = [1]\n",
    "# type_loop = graph_pattern[1]\n",
    "# num_allow_loop = 3\n",
    "\n",
    "loop_index = []\n",
    "type_loop = {}\n",
    "num_allow_loop = 0\n",
    "num_processes = 0 # zero for number of cores\n",
    "\n",
    "biasedwalk = BWMP.BiasedWalk(graph_pattern, loop_index, type_loop, data_path = data_path, \n",
    "                            train_file = train_file, postfix='', num_processes = num_processes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'biasedwalk' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-6-5e0e6b275ca7>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0minitials\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'v'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'a'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'v'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m3\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'v'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m4\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m6\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m7\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m8\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m9\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m10\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m'f'\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 9\u001b[0;31m walk_file_name = biasedwalk.walk_progress(initials, all_nw = all_nw, all_wl = all_wl, num_source = num_source, \n\u001b[0m\u001b[1;32m     10\u001b[0m                                           \u001b[0mnum_allow_loop\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnum_allow_loop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwalk_file_prefix\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mwalk_file_prefix\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m                                           loop_chance = loop_chance)\n",
      "\u001b[0;31mNameError\u001b[0m: name 'biasedwalk' is not defined"
     ]
    }
   ],
   "source": [
    "all_nw = 40\n",
    "all_wl = 40\n",
    "\n",
    "walk_file_prefix = 'walk_PREDICTION_REL_simplified_sep_probs'\n",
    "loop_chance = 0\n",
    "num_source = 0\n",
    "\n",
    "initials = {0:'v', 1:'a', 2:'v', 3:'v',4:'f',5:'f', 6:'f', 7:'f', 8:'f', 9:'f', 10:'f'}\n",
    "walk_file_name = biasedwalk.walk_progress(initials, all_nw = all_nw, all_wl = all_wl, num_source = num_source, \n",
    "                                          num_allow_loop = num_allow_loop, walk_file_prefix = walk_file_prefix, \n",
    "                                          loop_chance = loop_chance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'walk_file_name' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-7-cbc1fe6aad60>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mwalk_file_name\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'walk_file_name' is not defined"
     ]
    }
   ],
   "source": [
    "walk_file_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Link Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'graphpattern2vec.link_prediction' from '/hdd/BigIronOnHisHip/r2v/newRegpatterin/RegPattern2Vec_project-main/notebooks/graphpattern2vec/link_prediction.py'>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import imp\n",
    "imp.reload(LP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# list_embedding_file = 'emb_walk_final_PREDICTION_10_40_0_0_bwV4_s256_w7_n5_mc5'\n",
    "# list_embedding_file = 'emb_walk_final_PREDICTION_1for1_10_40_0_0_bwV4_s256_w7_n5_mc5'\n",
    "\n",
    "# relation simp\n",
    "list_embedding_file = 'emb_walk_PREDICTION_REL_simplified_sep_probs_40_40_0_0_bwV4_s256_w7_n5_mc5'\n",
    "list_embedding_file = 'GP2VEC_emb_walk_file_40_40_0_0_bwV4_s256_w7_n5_mc5'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['Protein', 'LightKinase', 'DarkKinase'], ['Pathway']]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[source_types, target_types]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__load_files| df_merge: (6279373, 10), df_nodes: (1064097, 4), df_relations: (13, 2), df_types: (11, 2)\n",
      "\n",
      "__read_embeddings| row: 1039768, col:256\n",
      "\n",
      "__read_embeddings| size of emb: 1039767, count line: 1039767, list chk: 1039767\n",
      "\n",
      "Getting examples\n",
      "__get_training_examples| list_yes_tuple: 68856, list_yes_emb: 68845 with 11 errors.\n",
      "__get_training_examples| h_uniq_nodes: 1014602, t_uniq_nodes: 1974\n",
      "__get_training_examples| list_no_emb: 68688 with 2 rejected, and 155 errors.\n",
      "__get_training_examples| release some data [set_yes_edges, list_yes_tuple]\n",
      "__get_training_examples| shuffled.\n",
      "__get_training_examples| release some data[ __list_no_emb, __list_yes_emb]\n",
      "__get_training_examples| Garbage Collector.\n",
      "__get_training_examples| release some data[__list_train_emb] \n",
      "__get_training_examples| Training data generated.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lp = LP.LinkPrediction(data_path, train_file, node_file, relation_file, type_file\n",
    "                , list_embedding_file ,test_file, result_path, \n",
    "                    [source_types, target_types])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "#lp.draw_pca_of_all_nodes()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The validation is for only the link prediction and not for the random walk and does not represent a the overall prefomacne of the whole model\n",
    "The other notebook named \"graphpattern2vec_process-multithread-Edited2022.ipynb\" is for the real validation found in the paper\n",
    "\n",
    "The histagram and Roc cureve in this notebook is not a good represitaion of the the model's prefomance. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_model_cross_validation| train_model| 256 and 1\n",
      "train_model_cross_validation| train_model| X_train: 96273, X_test: 41260, y_train: 96273, y_test:41260\n",
      "train_model_cross_validation| Cross-validation scores: [0.99740341 0.99563772 0.99532613 0.99605277 0.99594889 0.99522177\n",
      " 0.99646827 0.99542952 0.99584502 0.99646827]\n",
      "train_model_cross_validation| Cross-validation Accuracy: 1.00 (+/- 0.00)\n",
      "Cutt off is calculated to be [0.38878835440223275]\n",
      "[[20428    81]\n",
      " [   83 20668]]\n",
      "The test set, precision: 0.996096197407104, recall: 0.9960001927617946, f1 score: 0.9960481927710844\n",
      "__AUC_ROC| roc_auc_lr: 0.9996685740605554\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAikAAAIVCAYAAAAUK2/IAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/NK7nSAAAACXBIWXMAAAsTAAALEwEAmpwYAAByF0lEQVR4nO3dd5xU1f3/8deHuvQmAiICKjZUULFgVEQExRLsXUSNJBq/XzR2/UXxq1GjscQkGlHjghHFhhIbCoolYoGIBQWDir0AovT++f1x7iyzs7OzM7MzO7O77+fjMY/Zuffcc89chp333nvOuebuiIiIiBSbBoVugIiIiEgyCikiIiJSlBRSREREpCgppIiIiEhRUkgRERGRoqSQIiIiIkVJIUVERESKkkKKSB1gZqVm5kkeS81stpndYWbbZ1DfHtE2H5rZz2a20szmm9nDZnasmVkGdW1rZteb2Vtm9r2ZrTGzxWb2HzP7i5ntld27FpG6zjSZm0jtZ2alwGnAWuDH2GJgEzb+MbIGOMXdH0lRTwkwBjg1bvGqaNvWcctmAMe4++cp6moM3AycAzSMFm8AfgZaAo3jir8AHO3uSyt9kyJS7+hMikjd8rq7d44enYASYCgwH2gC3GdmHZNtGIWK5wgBZQNwB7CDuzdz9zZAJ+B8QsjoB0w3sx6V1NUIeAr4H0JAmQDsC5S4e3ugKdALuBj4DhgMdKj2uxeROkUhRaQOc/e17v4ccHK0qAVwdCXFrwMGEALKSe7+W3f/KK6uH9z9NmBvYAHQBXjQzJL9HrkGGAI4MMLdT3D319x9bVSXu/s8d78J2AooreZbFZE6SCFFpH6YDiyLft4hcaWZbQaMil7e6e4TKqvI3T8Efhu93As4MqGuLsDvopd/c/exqRrm7ivc/XTgi6reRJJ2b29mfzezj81shZn9ZGbvm9ntZrZbQtlpUT+dESnqmx+V2T9h+ehoeamZNTCzc6M+Nj9Fy/tGbXAzO7eKNk+Oyt2aZF2TqO5XzexHM1ttZp+b2T8y6VMkUlcopIjUH7HOrg2TrDud0EdkPXBDVRVF/Vo+jl7+OkldTYB1wPXpNs7dN6RbFsDM/gd4P9p/L8IZIAd2JFxmujmT+tLdLfA48Bdg12h/MQ9GzydVurHZpsCg6OX4hHVdgLeiuvcB2gCrgS0Ix/Q/ZnZU9d+CSO2hkCJSP+xNuNQD8GmS9ftHzzPd/as063wyev5F1AclZmBcXd9k1Mo0mdmxwO2EwPUooe9MS3dvR+jbcgowMw+7Pgo4mNAZuHW0v06EYxoLHf0r66sDHBu1+b/u/nZsYdQf6EmgDzCV8O9V4u6tgc2A2wj9i+43s61y/J5EipZCikgdZmaNzewg4J/RorWETqyJYpeA3s2g+vei5+ZA97jlscsSmdSVtugLPXap5EF3Pzah78yP7v6Au1+Qh923BP7X3e909xXR/n5w9yXuPhf4T1TuxEq2jy1/MGH5acDuwKvAUHefHtd/51t3Px+4i3Csz8/d2xEpbgopInXL3mb2XfT4njB8+DmgB+FyyK8rOVPSPnpelMG+Fsb93CHJzz+SH4OAroRLUxflaR+VWQT8I8X62NmUCiHFzLYgnCGJLxdzWvT851g4SeKB6HlwGu0UqRMaVV1ERGqRxoTLD4l+BA5y9xk13J58iE3+9q67f13D+57h7utSrH8IuBHYycx6u/vsuHUnEvq0/Cc66wKUDdfeI3p5l5n9rZK6Y32JumXXdJHaR2dSROqWl93d3N0IfRj6EvpstAfuNbN2lWwXO+uRyVwlmyTZHjaejWlPfsRCWMajgXJgQaqVUWh6JXqZ2IE2dnYl8SxKe0JHYwjHv1Mlj9jxbpZxq0VqKYUUkTrK3Ve7+7vAccBkYGdCv4ZkYn06+mSwi52j5xVA/Myz2dRVW6xPo0yFSz7R8OE+hEtuDyWUj/89vEssZKZ6VOsdiNQiCikidZyHe1/8L+EL9lgzG5Ck2EvR825mtnmaVQ+Lnl9P6EcRX9dmGTe4at9Hz91TlqoodpmmJEWZNpk3p4JHCbcR6Bl3X6JYYHklySWqRWwMP1vkYP8idYZCikg94O4fs3FUzx+SFCkljPxpCFxaVX3REOBtopeJZ2dKCV/SjdKpK67OdM8QvBE972xmXdOtH/gpek4awsxsa6BtBvUl5e6LCZ2VYeMln8ou9RAFvFhfoaHV3b9IXaKQIlJ//Cl6/kXijKrRX/d/iV6ebWbHV1ZJdOki1rnzLWBiQl3fEOb1ADjXzE4jBTNrbmb3kf6ZkanA14RAdVOa20CY+A3gl5WsTztQpSEWRo6LzqZsTQhuj1ZSvjR6HmFmKS+TpehXJFLnKKSI1BPu/g4wJXr5/5IUuQx4jfB7YbyZ/dXMtoutNLOOZjYKeB3oSLjscqK7J+uncQUhTBjhpobjzazcpG9mtrWZXQh8AozI4H2sBWJzoJxoZg8ntLO9mZ1lZrcnbPooYYbYnczsz2bWNiq/aVT2VEL/mlyYRLgNQSc2BrrnorMsydxLOENUArwYtb/srtNm1tnMTjazl9l4+wKROk8hRaR+uTF6HhTXXwIAd19DuCngeMLvht8CH5nZSjP7GfiBcIakLWHSsv7unmz2WqJhuocAdxI6i55ICECrzWyRma0B/ks4E9KZcMfkhcnqqqT+CYSgsoEwi+tHZrbUzBYT+niMYWPH3tg2s9l4hud/gcVR+e8IM8j+mipG72TQvpXAE9HLXaPnCpd64sqvJfTx+TdhtM+YqH2LzGwZ8C1hQr79KD8Vv0idppAiUo+4+wvAO9HL3ydZv9LdTwb6E/qazCX0VWlCGPL7GHAC0M/dP6tiX2vc/RzCvXRuJPS7+BFoTThj8Q5havt+7n64uy+rtLLk9d8C7ALcB8wnzBHjhJlw/0zymVkvIASSdwkT3Tlh5NMB7l6ayf7TEB9KlhHOrlTK3X8g3IX6ZOAZQmBqFa2eA4wjjNSq8t5KInWFhY7/IiIiIsVFZ1JERESkKBVtSDGzVmb2SzO7xsyeNbOFZubRY7uqa6iy/tZmdq2ZfWRmK6Jrv1PN7JhctF9ERESqp2gv95jZESQMbYyzvbvPqUbdmxOmru4ZLVpG6FUfG3lwZ3QtXURERAqkaM+kRH4gdCC7GhiZiwqjCaMeJQSU+cAv3L0VoYPaxYTRAmeb2Vm52J+IiIhkp5jPpDSMn3/BzHoAsdEEWZ9JiTtDswHYzd1nJay/FTiPMCyxezQsU0RERGpYo6qLFEYlE0TlwsnR85TEgBL5E2GypM7AAWyc3rpSm2yyiffo0aPcsvnz55O4TEREpDaZNWsWffv2rbB85syZC929Y773X7QhJY8GRs+Tk61096/NbDZhboe0QkqPHj2YMWNGuWX9+vWrsExERKQ2qey7zMw+T1I854q9T0pOmdmmQIfo5ewURT+MnnfIb4tERESkMvUqpABd4n7+JkW52LouKcqIiIhIHtW3kNIi7ueVKcrFbjLWsrICZjbSzGaY2YwFC3Jyuw8RERGJU99CSs64+xh37+fu/Tp2zHvfIRERkXqnvoWU5XE/N0tRrnn0nNENz0RERCR36ltIie+HslmKcrF13+axLSIiIpJCvQop7r4AWBi97J2iaGxUz4cpyoiIiEge1auQEnkpeh6cbKWZdWVjgJlaIy0SERGRCupjSBkfPQ8xsz5J1v8OMMKlnpeSrBcREZEaUNQhxcw2iT2AdnGr2savM7MGCdt59BidpNongTcJ732ime0VbdPUzC4g3LcH4Crdt0dERKRwin1a/MomIJme8Dp2R+Mqubub2THAK9F2081sGVDCxuPxd3e/O/PmioiISK4U9ZmUfHH3r4C+wHXAHEI4WUq4vHOcu59duNaJiIgIFPmZFHe3fG3n7kuAK6KHiIiIFJl6eSZFREREip9CCmBm/zCzH8zsg0rWm5ndbmbzzOw9M9u1ptsoIiJS3yikBKXAwSnWDwV6RY+RwJ010CYREZF6TSEFcPdXgB9TFBkGjPPgDcIQ6C410zoREZGqzZuX6musdirqjrNFpCvwZdzrr6JlNX5vn7tf+ZTbpnzM8jXra3rXIiJSpJa9v5BFz86n/ZDutOrbEYD5Nxxa4FZVn0JKDo0ZM4YxY8YAsGBBZVO8ZO/uVz7lD898lPN6RUSk9lq3ZA2LJn8ODj9O/pwGJQ1psV37QjcrJ3S5Jz1fA93iXm8eLStn5MiRzJgxgxkzZtCxY8ecNkABRUREkmnUugkdD98SGhiNN21GSffWhW5SzuhMSnomAeea2UPAnsDP7l6jl3pum/JxuddXHLI9Z+23ZU02QUREitjkyfPo128zOnRoXuim5IxCCmBmDwL7A5uY2VfAVUBjAHf/O/AMcAgwD1gBnF7TbYzvg6KAIiJSv61du57GjRuWW3bQQVsXqDX5o5ACuPuJVax34Lc11JwqKaCIiNRfpaWzuOWW6UyZMpxNN21R6ObklfqkiIiI1BKlpbM444wnef/9Hxg4cCw//LC80E3KK4UUERGRWiAWUNzD68aNG9CwYVa3uKs1FFJERESKXGJA6dOnE1OnDq9TnWSTUUgREREpYvU1oIBCioiISNGqzwEFFFJERESKUn0PKKCQIiIiUnSeeea/9T6ggEKKiIhI0RkwoDv7798DqL8BBRRSREREik6LFk34179O5Jxz+tXbgAKacVZERKQotWjRhL/97dBCN6OgdCZFRESkwEpLZ/HnP79R6GYUHZ1JERERKaDEUTyjRu1V2AYVEZ1JERERKZDEgFJa+i6rV68rbKOKiEKKiIhIASSbB2XKlFNp2lQXOWIUUkRERGqYJmpLj0KKiIhIDVJASZ9CioiISA1RQMmMQoqIiEgNGDtWASVTCikiIiI1oEOH5jRqFL52FVDSo5AiIiJSAw47bBsef/x49tijqwJKmjTOSUREpIYcdtg2HHJILxo0sEI3pVbQmRQREZE8ePTRD/nmm6UVliugpE8hpRa4+5VPC90EERHJQGnpLI477hEGDhybNKhIehRSaoHbpnxc9nOLJg0L2BIREalK/DDjjz9exHnnPVfoJtVaCim1wPI168t+Pu/AbQrYEhERSSVxHpS+fTtz552HFrZRtZhCSi1z1n5bFroJIiKSRLKAMmXKqRrFUw0KKSIiItWkgJIfCikiIiLVoICSPwopIiIiWVJAyS+FFBERkSy4OxMmzFZAySOFFBERkSyYGY8/fhxDhmylgJInmhZfREQkS82aNeaJJ45n5cp1tG/frNDNqXN0JkVERCRNH3zwQ4VlzZo1VkDJE4UUERGRNJSWzmLnne/kxhv/Xeim1BsKKSIiIlWIH8VzySVTKC2dVegm1QsKKSIiIikkG2Z8+OG6RUlNUEgRERGphOZBKSyFFBERkSQUUApPIUVERCSBAkpxUEgRERGJo4BSPBRSREREIt99t4zf/vYZBZQioZAiIiIS6dy5JRMnHk/Tpg0VUIqApsUXERGJM2TIVjz//Kn07t1RAaXAFFJERKReW7VqHSUl5b8O99uve4FaI/F0uUdEROqt0tJZ7LTTnXz55c+FbookoZAiIiL1UmwUz7x5P7L//mMVVIqQQoqIiNQ7icOMW7duSvPmjQvbKKlAIUVEROoVzYNSeyikiIhIvaGAUrsopIiISL2ggFL7KKSIiEidp4BSOymkiIhInfb8858ooNRSCikiIlKnDRjQnUMO6QUooNQ2CikiIlKnNW3aiMceO47f/W4vBZRaRtPii4hInde0aSNuvvmgQjdDMqQzKSIiUqeMHTuLa699pdDNkBwo+jMpZtYZuAw4DOgK/Ay8Bdzm7lOrUe+RwJnAbsAmwCrgE+A54FZ3/76aTRcRkRo2duwsTj89dJJ1d37/+wGFbpJUQ1GfSTGznYEPgP8FtgRWEwLFYcALZnZpFnU2MLN/Ao8DhwKdgZVAM6APcAnwoZntnpM3ISIiNSI+oAA89thHrFy5trCNkmop2pBiZs2ASUAH4B1gR3dvA7QDbgYMuM7MhmRY9VnAydHPtwGd3L01UAIcDHwBtAceNLOiPT4iIrJRYkDp06cTU6cOp1kz3Y+nNivmL+FfA92BZcDh7j4bwN2XuPuFwBOEoHJ9hvWeFD2/6O7nu/sPUb3r3H0ycFq0fitg5+q9BRERybfKAopG8dR+xRxSYmc7xrv710nW3xQ972pm22ZQb6fo+Z1K1s+M+7lFBvWKiEgNU0Cp24oypJhZK0KHVoDJlRR7g9CJFmBQBtXPj553qWR9bL+rgQ8zqFdERGqQAkrdV5QhBdiecCkHYHayAu6+AZgbvdwhg7rvjp4PMLNbzWxTADNrZGYHAWOj9f/n7osza7aIiNSEcePeVUCpB4o1pHSJ+/mbFOVi67qkKFOOuz8GXAGsB84DvjezJYQhyM8BS4HT3f26TBosIiI1p1u31pSUhFk0FFDqrmINKfF9QVamKLciem6ZYf3XAyOA5dHrVkDDuH1vUtXIHjMbaWYzzGzGggULMty9iIhUx8CBPXn66ZPYe+9uCih1WLGGlLyJ+rv8C7gfeAnYE2gN9AB+E/18E/BAqnrcfYy793P3fh07dsxrm0VEpKKBA3vy2munK6DUYcUaUpbH/dwsRbnYJ3NZBnXfQpjEbYq7H+7ub7n7Unf/3N3vAo4CHDjBzIZm1GoREcmLCRM+YP78nyosN7OKhaXOKNaQEt8PZbMU5WLrvk2nUjNrDZwevbwtWRl3fxn4T/RyWDr1iohI/pSWzuLEEx9j//1LkwYVqbuKNaTMIZzNAOidrEDUZyQ2P0q6Q4V7sbHvyWcpyn0aPfdIs14REcmD0tJZnHFGGMXz+ec/M2rUc4VuktSgogwp7r4UmBG9HFxJsT2BNtHP6d5ocEPcz1ukKNc9el6aZr0iIpJj8QEFwiief/zjl4VtlNSoogwpkfHR88lmlmyI8YXR80x3n5tkfTJzCZO0QbiHDwBmdrCZzTWzeWZ2O7BrtOrNaP0WZvaSmb1jZu+Z2SGZvRUREclEsoCiUTz1TzGHlLuAzwnDg58ysx0gjM4xsxsJHVwBLk/c0Mw8eoyOX+7uK9g4WdtRZna3mXUH/kbof3IJYYRPI2AJUBqV/X/Aw+6+C3ACcEeu3qSIiJSngCIxjQrdgMq4+0ozG0a4lLMrMDuadK0lIVw5cLm7P59h1RcQZqjdB/hV9FhPmNk2FtpWA8e6+8JYcwhDkyFcYko1wZyIiGRJAUXiFW1IAXD3d81sR+Ay4DCgK7AIeAu41d3T7YsSX+cyM9sfGE44K7IX4WzNCkJn2q+AHxLCz2jgeTP7H8Jkbwdm+ZZERKQSCiiSqKhDCoC7fweMih7pbpNy4Ly7rwfuA+4zs2OAg939VwBmdiqhU268E4FSd7/ZzPoD95vZjtH9g8qMGTOGMWPGAKBZaEVE0ufuPPXUxwooUk4x90mpKV8D3eJebx4ti3cm8DCAu08HSoBNEisaOXIkM2bMYMaMGWgWWhGR9JkZ48cfzZFHbqeAImWK/kxKDXgb6GVmPQnh5ATgpIQyXwCDgFIz254QUnSqREQkh5o0aciECcewbNka2rVLNdm41Bf1/kyKu68DzgUmAx8RRvHMNrP/M7PYgPwLgLPM7F3gQWCEe+ykpIiIZOM///mWxF+ljRs3VECRMjqTArj7M8AzCcuujPv5Q+AXNd0uEZG6KtZJ9vLL9+WaawbqHjySVL0/kyIiIjUrfhTPH/7wKnfc8XahmyRFSiFFRERqTOIw4759O3PCCTsWtlFStBRSRESkRiQLKFOmnKpRPFIphRQREck7BRTJhkKKiIjklQKKZEshRURE8kYBRapDIUVERPJi4cIVjBr1nAKKZE0hRURE8mKTTZrz9NMn0aJFYwUUyYomcxMRkbzZZ58tePHF09hqq3YKKJKxrEKKhakBjwQGE27O18zdB8WtbwHsBri7v5qLhoqISPFbsWItzZs3Lrdsjz26Fqg1UttlfLnHzHoB7wGPAL8GDgH2Tyi2CrgXmGZm+1SzjSIiUguUls5iu+3+yrx5Pxa6KVJHZBRSzKwdMAXoDbwPXAksSSzn7uuBOwEDjq5+M0VEpJjFRvF8+eUS9t+/lE8+UVCR6sv0TMoFhMs7k4F+7n4tsLKSspOi572zbJuIiNQCicOMO3ZsQdu2JYVtlNQJmYaUYYADF7j7ulQF3X0esAbYOsu2iYhIkdM8KJJPmYaUnsAqd/8wzfJLgVYZ7kNERGoBBRTJt0xDiqe7jZk1AlqTpM+KiIjUbgooUhMyDSmfAU3MbMs0yg4CGgMfZdwqEREpWgooUlMyDSlPE0bsnJ+qUDRPyk2EMy9PZtc0EREpNi+99JkCitSYTEPKzcBi4Bwzu9bMOsSvNLNWZnYsMAPYEfiGMBRZRETqgH337c6xx/YGFFAk/zKacdbdF5rZMOBfwGXAJYQzK5jZj4Q+KBY9fgSOcPflOW2xiIgUTKNGDXjggaPo1as955+/lwKK5FXGM866+2tAH+BBYH1UhwFto5/XAxOA3dx9Zs5aKiIiRaFRowZce+0BCiiSd1ndu8fdvwBOMbOzCPfo6UIIKN8DM9x9We6aKCIihVJaOouPPlrADTccSLhtm0jNqdZdkN19JfBajtoiIiJFJH4Uz4YNzo03DlZQkRqV6b17/mFmt2RQ/kYzuzfzZomISCElDjOeMuUzli9fW9hGSb2TaZ+UEcAJGZQ/NtpGRERqicrmQWnZsklhGyb1TsYdZzNkhLlSRESkFtBEbVJM8hZSzKwBsCmgIcgiIrWAAooUm5QdZ82sNWFocbyGZtaNaH6UZJtF2wwHSoB3q9dEERHJNwUUKUZVje45H7gyYdkmwPwM9nF3Jg0SEZGadf/97yqgSFGq6nKPJTw8ybLEB4Q7H08HznD3f+S+2SIikiu9enWgVaumgAKKFJeUIcXdR7t7g9iDEEK+i1+W5NHQ3du5+z7uXloj70JERLK2116bM3nyKQwY0F0BRYpKppO5jQN+ykM7RESkgPbaa3Neeuk0TdYmRSWj0T3uPsLdz8tTW0REpAY8+OD7zJmzsMJyBRQpNvmeJ0VERIrI2LGzOPnkxxk4cGzSoCJSTLK+d4+Z7Qv8AtgMaEHlQ5Ld3c/Mdj8iIpIbY8fO4vTTwyie775bxvnnT+bZZ08udLNEKpVxSDGzHYHxQO/EVdGzJyxzQCFFRKSA4gMKQJ8+nfjnP48sbKNEqpBRSDGzLsBUoCPwIfACMApYBtwGdAIOALYCFgJ3Aety11wREclUsoAydepwjeKRopfpmZQLCQHlOWCYu681s1HAMncvm/TNzEYCfwV2BQ7LVWNFRCQzCihSm2XacfZgwuWbK9y90nt2u/sY4Iqo/G+zb56IiGRLAUVqu0xDSndgPTArbpkDTZOU/Xu0bnhWLRMRkawpoEhdkGlI2QD87O7xnWOXAa3NrGF8QXdfSpgef5vqNVFERDI1bdrnCihS62XaJ+VrYEsza+DuG6Jl84EdgZ2Bd2IFzawN4W7Iq6rfTBERycQ99xzO+vUbeO+97xVQpNbK9EzKXEKw2T5u2auEocYXJpS9Jnr+MLumiYhItho2bMB99w1j2rQRCihSa2UaUp4nBJL4ETt/AdYCJ5jZ+2b2gJm9S+gw68CdOWmpiIhU6s03v6L8lfgQVNq2LSlQi0SqL9OQMgG4GVgeW+Duc4HTomW9gROBnaLVt7r7vTlop4iIVKK0dBb9+9/LqFHPVQgqIrVZRn1S3H0RcFGS5Q+Z2RRgKLA58DMwxd0/zkkrRUQkqdLSWZxxRhjF85e/vMUWW7Thwgv3LnSzRHIi63v3JHL3hcD9uapPRERSiw8oAH37dub00/sWtE0iuZTXuyCb2eZm9pd87kNEpD5KFlCmTDlVnWSlTslLSDGznmY2BpgHnJOPfYiI1FcKKFJfpHW5x8waAwOBHYCGwKfAs+6+KqHclsBo4ISonAHf5bC9IiL1mgKK1CdVhhQz2xsYD3RLWLXQzE529ylm1gi4FjgPaEwIJx8DfwLG5bTFIiL1lAKK1DcpQ4qZbQ48A7QiBI94HYEnzGwnYAxwQFTmdeAmYJJrLJyISE789NMqLrzweQUUqVeq6pPyv0BrYBFwMiGYbAqcGi1rBkwBBgEfAIPcfR93f1IBRUQkd9q2LWHy5FNo27ZEAUXqjaou9xxImDX2XHd/OG75A2a2nnAZqAfwInCou6/OSytFRITddtuMadNOY/PNWyugSL1QVUjZkhBSnkiybmK0DuAKBRQRkdxaunQ1rVo1LbesT5/OBWqNSM2r6nJPK2CRu69JXBGFkoXRy3dz3TARkfqstHQWvXr9hQ8++KHQTREpmKpCigHrU6zfAJA4FFlERLIXG8Xz/ffLGThwLB9+uKDQTRIpiLzOOCsiIplJHGa8+eat6dSpRWEbJVIg6Uzm1sbM/lHZOoAU6wHc3c/MuGUiIvWM5kERKS+dkFICnFbJutjcKanWO6CQIiKSggKKSEVVhZRX2DiCR0RE8kABRSS5lCHF3fevoXaIiNRLCigilVPHWRGRAnnttS8UUERSKPqQYmadzezPZvaJma0ys+/N7F9mNigHdXc1sxvM7H0zW2Jmy8zsv2Y23syG5aL9IiKV2Xvvbpxxxi6AAopIMul0nC0YM9uZMOV+h2jREmAT4DDgUDO73N1vyLLuY4F7CRPWAawg9L/ZOnpsCjyZfetFRFJr0MAYM+ZwevZsy29+008BRSRB0Z5JMbNmwCRCQHkH2NHd2wDtgJsJI4euM7MhWdR9CPAgIaD8A9jO3Vu4e0tCCDqacPdnEZG8atDAuOKK/RRQRJIo2pAC/BroDiwDDnf32QDuvsTdLyTcT8iA6zOp1MxaA/cADYHr3P1Md58bW+/ui9z9cXe/JTdvQ0QkKC2dxTnnPM2GDRo0KZKOYr7cc3L0PN7dv06y/ibgCGBXM9s2PmhU4XSgC/AVMLq6jRQRSUf8KJ4NG5w77jiUBg2s6g1F6rGiPJNiZq2A3aKXkysp9gbwc/RzJp1oY+HnUXdfm0XzREQykjjM+M03v2bZsgr3bRWRBEUZUoDt2Tib7exkBdx9AxA7e7JDOpWaWQnQN3r5jpltZ2YPmtkP0cihT83sTjPrkX3TRUQ2qmwelNatmxa2YSK1QLGGlC5xP3+TolxsXZcUZeL1ABpHP28D/Ac4AWgNNAF6Ar8B3jWz/eM3NLPjzOxDM5ttZuPT3J+I1GOaqE2keoo1pMTf8nNlinIroueWadbbNu7ny4DFwFDga8Kw4wHAakJoecTM2gOYWa+o/C/cvTdwnpmNNLMZZjZjwQLdRl1EylNAEam+rEKKmfU0s9vN7KNoArR1CevbmtmVZvZ7M2tcWT0F0CDh5+GEfi3z3P1Td38F+DthvpRNgF9FZc8C/ubuiwHc/Qd3H+Pu/dy9X8eOHWvuHYhI0VNAEcmNjEOKmR0JvAf8FtgWaM7G/iMAuPtPwAGE0TO/zKJdy+N+bpaiXOx//LI0640vN9vdpwJdgS/jls8kjPyBjR1ytwG2MbN/m9kbZnZwmvsTkXrmgQfeU0ARyZGMQoqZbQc8QLgcMwbYD1hYSfG7CeHlsCzaFd8PZbMU5WLrvs2i3lRDlhdHz92i50ZAL2B/4ETgbjNrm7jRmDFj6NevH/369UOXgETqpx133JT27cPfVgooItWT6ZmUi4AS4FZ3P9vdXwPWV1J2SvS8RxbtmkO45ALQO1kBM2tAOJMD8GE6lbr7QuD7hMVfszGMAGzOxjMusTZ8BUxy97Xu/hnwMSG0lDNy5EhmzJjBjBkz0CUgkfqpT5/OTJ06nAMP3FIBRaSaMg0pgwhf3DdWVdDdvydctulWVdkk2y4FZkQvB1dSbE+gTfTz1Ayqj4WnWMB5G+gV9bNpQhjtEzM/en6CcBYFM9uEcPnn0wz2KSL1SJ8+nXnhBQUUkerKNKR0BpZGASQdqwlDe7MRG+Z7spklG2J8YfQ8M4PZZgHGRc+9zexAd18HnEuYNO4j4GVgr6hMrK/KZGCRmX0IvARc5O6LMtiniNRR//zne7zzTrpXnEUkE5mGlOVACzNrWFXBaNbYtsCPWbQL4C7gc8JNAJ8ysx1i9ZrZjcBRUbnLk+zbo8foxHXu/jzwQvRyrJkNBp51922AU4CDCMflM+B30Tbu7r9z9x3cfSd3fyjL9yQidUhp6SyGD5/IgQfer6AikgeZhpTZ0Ta7VVUQOD4qOzPTRgG4+0pgGLAI2BWYbWY/Az8R+sY4cFkUOjJ1IvA+oePt88AyM1sCvE64lPMN8Et3X1F5FSJSn8UPM/7xx5VcdNELVW8kIhnJNKQ8TBixc03UcTUpM9sJuIEQJB7ItnHu/i6wI3A7oQ9IU0JoeRoY7O43ZFnvImB34BLgHWADYQTPbOA6YGd3/yDbdotI3ZZsHpQJE44pbKNE6qBM74J8F2GCswOBqWZ2e6yOKJh0J8zgOoIwv8lrwITqNNDdvwNGRY90t6ny1qLuvprQAbjKTsAiIjGaqE2k5mQUUtx9bTSR2STCFPL7xa2eFfezEe5SfJR77L+yiEjtpoAiUrMynnE2OrOxNzCS0IdjLSGUGOGyyVvA2cB+0bwkIiK1ngKKSM3L9HIPANGw3XuAe6KRPu0JgWdRtE5EpM5QQBEpjKxCSjx3Xw9oDngRqbNmzvxGAUWkADIKKWZ2M/CAu/8nT+0RESk6t98+lPXrnenTv1JAEalBmZ5JOR84z8z+S5gR9kF3/2/umyUiUjzMjL/97RCWLVtDq1ZNC90ckXoj046zLxA6x24DXAXMMbO3zey8SqauFxGpdV599XM2bCg/MNHMFFBEalhGIcXdDyLM0nouYWQPhNlnbwa+MLMpZnaGmbWprA4RkWI2duwsBgwo5Ve/mlQhqIhIzcpmCPICd7/D3fcFegCXAe8BDYEDgLuB78xsopkda2b600NEaoWxY2dx+ulhFM99983iD394pdBNEqnXMg4p8dz9S3f/o7vvAuwAXMvG6euHAQ8B6d4xWUSkYOIDCkCfPp0455zdC9sokXquWiElnrvPcfcr3b0X0B/4D2GCt1a52oeISD4kCyhTpw7XKB6RAqv2PCnxzGwP4CTgOKBTLusWEckHBRSR4lXtkGJm2xOCyYlAz9hiYBnwBGGosohI0VFAESluWYUUM9uCEEpOBHaKLQbWAM8Rgskkd1+Vi0aKiOSaAopI8ct0xtlzCGdN+scWEeZNeYUQTB51959y2UARkVxbtmwNl102VQFFpMhleiblr3E//4cQTB5y929y1yQRkfxq2bIJU6YMZ+DAsXTp0lIBRaRIZRpS5hGCyXh3/zgP7RERqRE77NCRl18eQceOzRVQRIpURiHF3bfJV0NERPJp8eKVtGvXrNyy7bbbpECtEZF05GyeFBGRYlVaOouttrqdt9/+utBNEZEMKKSISJ1WWjqLM854ksWLVzF48P288863hW6SiKSp0ss9ZrY++nGOu/dOWJYJd/ecThonIpKOWECJjeLp2bMdW2yh+5+K1BapwoMlPCf+nK5sthERqZbEgNK3b2emTDlVnWRFapFUISU2e+zaJMtERIqWAopI3VBpSHH3z9NZJiJSTBRQROoOdZwVkTpDAUWkbskopJjZBjNLewyfmX1mZusyb5aISGbeeutrBRSROiabMymZdoRVx1kRybvdd9+MUaP2BBRQROqKfA8NbgpkM2xZRCQjZsYttxxEt25tOO20PgooInVA3kKKmXUGNgV+yNc+RKR+c3fMNp6sNTN+97v+KbYQkdokZUgxs/2A/RMWtzSzK1NtBrQFDo5+/nc12iciklRp6SxefPEz7rtvGA0bagyASF1U1ZmUgcBVgMctaxEtSyX2p82PwNXZNU1EJLn4UTzr1zvjxh2hoCJSB1UVUmYBY+NenwasAh5Osc0GYAkwG5jo7ouq00ARkXiJw4w//HABS5euoW3bksI2TERyLmVIcfcngSdjr83sNOBndz893w0TEUlU2TwoCigidVOmHWcHAmvy0RARkVQ0UZtI/ZNRSHH3l/PVEBGRyiigiNRP6mkmIkVNAUWk/qr0TIqZfRr9OM/dhyQsy4S7+1bZNE5E6rcHH3xfAUWkHkt1uadH9LwqybJMeNVFREQq2m23zejcuSXffrtMAUWkHkoVUgZGzyuSLBMRybtttunAtGkjOP/8yYwbd4QCikg9U2lISdZJVh1nRaSmbbNNB55++qRCN0NECkAdZ0WkaNx//7u8/vqXhW6GiBSJnN5g0Mw2AfoR7n78qrv/mMv6RaTuio3iadmyCc89dwp7792t0E0SkQLL6EyKme1lZuPN7JIk604BPgWeBh4HvjAznaMVkSrFDzNeunQNl102FXf1uRep7zK93HMKcDzh3jxlzGxr4B9AS2AdsBpoDpSa2Y45aKeI1FHJ5kF5/PHjMLPUG4pInZdpSNknev5XwvJfEy4dvQx0ANoSbkLYCBhVjfaJSB2midpEJJVMQ0pnYD3wdcLyQwnzoVzl7svcfQ0QuyQ0oHpNFJG6SAFFRKqSaUhpDyz1uIvFZtYe2I5wCejV2HJ3/5wwx8rmOWiniNQhCigiko5MQ8pyoI2ZNYlbFjtTMt0r9nRbQzjzIiICKKCISPoyDSkfAgYcHbdsBOFSz7T4gmbWEmgDfJt980Skrpk7d6ECioikJdN5Uh4G+gNjzGwfoAtwOLAWmJBQdm9CoPlvdRspInXHddcNwh0mT/5EAUVEUsr0TModwCtAC+A3wBHR8v+L+qDEO4FwhuXF6jRQROoWM+P66wfx2munK6CISEoZnUlx97VmNgg4CdiL0Fn2WXd/Jb6cmTUGmgGTqDhcWUTqkRdf/Ix9992Cxo0bli0zM1q0aJJiKxGRLKbFd/f1wP3Ro7Iya4ETq9EuEakDYp1kjz56B8aPP6pcUBERqYpuMCgieRE/iufRRz/kqqumFbpJIlLLZH2DwWj0ziHArkDHaPEC4D/AM+6+rPrNE5HaKNkw4wsu6F/YRolIrZNxSLFwQ43LCDPKtqyk2DIzux74Y5K5U0SkDtM8KCKSK9mcSSkl3GjQgFXATOCraN3mwG5AK+APwPbAadVupYjUCgooIpJLGYUUMzsKOJUwtDh2piTxjsitgUsJZ1pOMbMn3H1ijtorIkVKAUVEci3TjrMjCQHl/7n7FYkBBcDdl7j75cDvCWdbRla/mSJSzBRQRCQfMg0puxHuxfPnNMr+OSrbL9NGiUjtsWLFWkaPnqaAIiI5l2lIaUW4C/KKqgq6+3LCZG+tsmmYiNQOzZs3ZurU4XTr1loBRURyKtOOsz8AXc1sM3f/JlVBM+sKtAVSlhOR2m+rrdrz8ssjaN26qQKKiORMpmdSYtPf3xINRU7lluh5Wob7EJEit3BhxZOpPXu2U0ARkZzKNKT8idBx9lhgmpkdbGZlv5XMrIOZHWNmbwPHABuAm3PWWhEpuLFjZ7Hlln/mlVcS7ykqIpJbGYUUd58FnEMIKvsATwNLzGyRmS0jXA6aQOhg68Bvo21EpA4YO3YWp5/+JEuXrmHo0Ad4442vqt5IRCRLGd+7x93HAPux8TJOA6Ad0Jww5BjgRWDfqKyI1AGxgBIbxdOrV3t69Wpf2EaJSJ2W1b173P11YJCZtQN2ofy9e95x98U5ah9m1pkwDf9hQFfgZ+At4DZ3n5qjfTQE3iScAQK42t1H56JukbogMaD06dOJqVOHqw+KiORV2iEl+iLfFmgN/ATMjcLIi/lpGpjZzlH9HaJFS4BNCIHlUDO73N1vyMGu/sbGgJKqPUcDjwK7u/uMHOxXpOgpoIhIoVR5ucfMSszsFmAx8D7wb2A2sMjMrjGzrO+kXMV+mwGTCAHlHWBHd29DuLR0M+HS0nVmNqSa+9mCMCvu18D30eKOScq1AkYRzriI1AsKKCJSSClDipk1ACYTvpxbEoJB7NEWuBx4OE9t+zXQHVgGHO7us6Fs2v0LgSeidlxfzf3cH9VzDuGGiRDOGCW6BvhjXBmROk0BRUQKraozKSOAfQlf4i8CZwGHAL8h9AsxYJiZHZaHtp0cPY9396+TrL8pet7VzJKFiiqZ2S8JnYC/cPdJcataJ5TbFejm7k9nsx+R2mbWrO8UUESk4KoKKScShhLf7e4Huvu97v5cNGqnP/AvQlA5PpeNii6txPqITK6k2BuETrQAg7LYRwvgr8AaYHqKcg0IE9NdUFWdY8aMoV+/fvTr148FCxZk2iSRotG3b2euvHIAoIAiIoVTVUjZOXq+KnGFuzswOqFcrmzPxuHMs5MVcPcNwNzo5Q5Z7OMaoBtQSujnEi/+7s6tgB0Jk9fNB/YCJplZhRsnjhw5khkzZjBjxgw6dqzQrUWkVhk9en/uvPNQBRQRKZiqOr22A5a4+3eVrP8orlwudYn7OdW9f2LruqQoU4GZ7QL8L/AxcB4w28x6xhWJhR/c/WfCiKLYttOACzW6R+oadyfxbhe/+Y1uYi4ihVPVmZRGwMrKVrr7qrhyudQi7udK9w/EbiDSMt2Ko8s3dwENgXPdfSVwLuGyUteo2AIz+7+oz0pl9Yw0sxlmNkOXdqS2Ky2dxdFHP8yaNesL3RQRkTIZzzhbB/wW2B142N1fAHD3Z9x9G8IwZKJlVyZ0po0t39/dZ7j7GHfv5+79dGlHarPS0lmcccaTTJw4h2OPfURBRUSKRjpnQJqZ2fDqlHH3cZk1i+XxdQNLKykXu1C+LJ1KzWwz4NqovvMzbJNInRMLKLFRPF988TPLl6+hSZNmhW2YiAjphZTWwH0p1nsVZRzINKTE90PZjLg+Igk2i56/TbPe6wlt/T3hxoiJl4liF+SbxNa5e1oBSKS2SQwofft2ZsqUU2nXTgFFRIpDOpd7rJqPbC4pzSGEG4DeSRsV+pbE5kf5MM16u0fP1xDOpiQ+tojWXxa3TKTOqSygaBSPiBSTlAHC3Rvk4pFpo9x9KRAbPTO4kmJ7Am2in3Nyo0GR+kABRURqi2LuODs+ej7ZzJINMb4wep7p7pVdDion6vRqlT2Az6OiV8ctE6kzFFBEpDYp5pByFyE0tAKeMrMdIMxGa2Y3AkdF5S5P3NDMPHqMrqnGihS7CRM+UEARkVqlaENKNH/JMGARsCthwrWfgZ+Aiwh9Vi5z9+cL1kiRWmSvvTane/e2gAKKiNQORRtSANz9XcKU9LcDnwJNCaHlaWCwu99QwOaJ1Crdu7dl2rTTOOKI7RRQRKRWyPVMsTkXTck/Knqku01WfUncvUc224nUFt27t2XixJzeD1REJG+K+kyKiGRv7NhZTJ36aaGbISKSNYUUkTqotHQWp5/+JIcd9qCCiojUWgopInVM/DDjVavWceWV0/DYkB4RkVpEIUWkDkk2D8qkSSdgpil/RKT2UUgRqSM0UZuI1DVZje6x8GfZkYQp67sBzdx9UNz6FsBugLv7q7loqIhUTgFFROqijEOKmfUCHgd2YONdgxMveK8C7gW2NLMB7v5atVopIpVSQBGRuiqjyz1m1g6YQrgz8fvAlcCSxHLuvh64kxBijq5+M0UkGQUUEanLMu2TcgHh8s5koJ+7XwusrKTspOh57yzbJiJV+PbbpQooIlJnZXq5Zxjh0s4F7r4uVUF3n2dma4Cts22ciKR22WX74g6PPPKhAoqI1DmZnknpCaxy9w/TLL+UcBdjEcmTyy/fl+nTz1RAEZE6J9OQ4uluY2aNgNYk6bMiItmZPHkeq1ZVPIlZUlL0t+ESEclYpiHlM6CJmW2ZRtlBQGPgo4xbJSIVlJbOYujQBzjqqAlJg4qISF2TaUh5mjBi5/xUhaJ5Um4inHl5MrumiUhM/CieZ5+dx+WXTy10k0RE8i7TkHIzsBg4x8yuNbMO8SvNrJWZHQvMAHYEviEMRRaRLCUbZnzFFfsWtlEiIjUgo5Di7gsJI3yWAJcB3wEdAczsR0KAeQjYFvgROMLdl+eywSL1ieZBEZH6LON790Szx/YBHgTWR3UY0Db6eT0wAdjN3WfmrKUi9YwCiojUd1kNCXD3L4BTzOwswj16uhACyvfADHdflrsmitQ/CigiIlmGlBh3XwnovjwiOaSAIiISZHy5R0TyZ/XqdVx//WsKKCIiZHgmxcy2yGYn0eUhEalC06aNmDp1OAMHjqVlyyYKKCJSr2V6ueezLPbhWexHpN7afPPWTJt2GiUljRRQRKRey/Ryj2Xx0CUlkRS+/XZphWVdu7ZWQBGRei/TeVIapHoQhiEPBqYS5kkZEi0XkSRKS2ex1Va389xz8wrdFBGRopPTAOHuS9x9qrsPBqYAT5hZ71zuQ6SuiI3iWblyHUcc8RDTps0vdJNERIpKPs9yXAo0B67M4z5EaqXEYcbbbbcJO+20aWEbJSJSZPIWUtx9PvATMCBf+xCpjRIDSp8+nZg6dbj6oIiIJMjbqBszaw60Btbmax8itY0CiohI+vJ5uefcqP5shi2L1DkKKCIimcl0Mrf9qihSAmxOuFPyoYQ5UsZl1zSRukMBRUQkc5le7plGCB5Vsej5ceBPGe5DpE758MMFCigiIlnINKR8QeqQso7QWfZ94GF3fy7LdonUGTvs0JGbbhrMhRe+oIAiIpKBjEKKu/fIUztE6rQLLtibTTZpzmGHbaOAIiKSpkz7pLSOflzu7uvz0B6ROsHdMbNyy047rW9hGiMiUktlOrrnJ8J095vlvikidcPYsbM4+OAHWLFCo+9FRKoj05CyDFji7l/mozEitd3YsbM4/fQnef75Txg27CEFFRGRasg0pHwGNDezvE0CJ1JbxQJKbBTPggXLWbVqXWEbJSJSi2UaUh4GGgNH5L4pIrVXYkCJjeJp375ZYRsmIlKLZRpSbgJmAHeZ2aA8tEek1qksoGgUj4hI9WR62eZS4EVge+B5M3sPmA4sACod7ePu/5d1C0WKmAKKiEj+pAwpZvYisMjdj40WjSZM5hYbW9kH2DlVFVF5hRSpcxRQRETyq6ozKfsD38W9Hkd60+KL1GmPPfahAoqISJ5lOuPsiDy1Q6RW+cUvtmDbbTdhzpyFCigiInmSacdZEQE6d27JSy+dxvHH91ZAERHJE813IpKlzp1b8tBDxxS6GSIidZbOpIikYezYWTz55JxCN0NEpF5J50xKGzP7RzX24e5+ZjW2Fymo0tJZnHHGkzRq1IBHHjmWYcO2K3STRETqhXRCSglwWpb1x4YgK6RIrRQLKO6wdu0GrrvuNQ4/fFsaNLCqNxYRkWpJJ6SsJUzYJlKvxAcUgL59O/PMMycpoIiI1JB0QsqP7j4w7y0RKSLJAsqUKadqFI+ISA1Sx1mRBAooIiLFQSFFJI4CiohI8VBIEYkooIiIFBeFFJHIsmVrFFBERIqIZpwViZx77h4A/OMf7/DCCwooIiKFljKkuLvOtEi9cu65ezBy5G40adKw0E0REan3FEKk3nr66Y9ZvnxNheUKKCIixUEhReql0tJZHH74gxxyyPikQUVERApPIUXqnfhRPK+88jkXXfRCoZskIiJJKKRIvZJsmPE112hCZRGRYqSQIvWG5kEREaldFFIAMzvYzOaa2TwzuzTJ+t+Z2Ydm9p6ZTTWz7oVop2RPAUVEpPap9yHFzBoCfwOGAjsAJ5rZDgnF3gH6ufvOwKPAjTXbSqkOBRQRkdqp3ocUYA9gnrt/6u5rgIeAYfEF3P0ld18RvXwD2LyG2yhZUkAREam9FFKgK/Bl3OuvomWVORN4Nq8tkpxYu3Y9t932hgKKiEgtpWnxM2BmpwD9gAHJ1o8ZM4YxY8YAsGDBghpsmSTTuHFDXnjhVA44YByNGjVQQBERqWWKPqSYWWfgMuAwwhmOn4G3gNvcfWoW9XUEjgIOBHYlXLppaGa/AKYCK4Gvk2x3IHAFMMDdVyere+TIkYwcORKAfv36Zdo0yYOOHVvw4ovDadDAFFBERGqZog4pZrYz8CLQIVq0BNiEEFgONbPL3f2GDKv9hvLvexnQEtguejhwSUI7dgHuAg529x8yfR9Sc7788me6dWtTblnHji0K1BoREamOou2TYmbNgEmEgPIOsKO7twHaATcDBlxnZkMyrLoR8ApwGtDF3VsBhxP6payJq/fvZvbLaJubCEHmETObZWaTqvfuJB9KS2ex9dZ/4dFHPyx0U0REJAeKNqQAvwa6E850HO7uswHcfYm7Xwg8QQgU12dY7wB3H+Du49z9u6jOp9x9C8Klnx8IQaapu0+K1h/o7p3cvW/0+GXl1UshxEbxrFmznhNOeJTnnptX6CaJiEg1FXNIOTl6Hu/uFfqIEM5uAOxqZtumW6m7v5Ji3QLgmejlbunWKYWVOMx4p506sfvumxW2USIiUm1FGVLMrBUbQ8LkSoq9QehECzAoh7tfFD03zGGdkieaB0VEpO4qypACbE+4lAMwO1kBd98AzI1eJs4QWx2x4cUf5LBOyQMFFBGRuq1YQ0qXuJ+/SVEutq5LijJpM7NhhHlQAO7LRZ2SHwooIiJ1X7GGlPgxoytTlItNVd+yujs0s67AmOjlJHd/roryI81shpnN0MRtNUsBRUSkfijWkFKjzKwlYbTQpsDnhKnvU3L3Me7ez937dezYMc8tlJj//ncRZ545SQFFRKQeKNaQsjzu52YpysW+mZZluyMzKwGeJFzmWQAc5O4Ls61P8qtXrw7ceeehgAKKiEhdV6wzzsb3Q9mMjR1kE8XGmX6bzU7MrAnwKHAA8BMwxN0r25cUiZEjd6NduxIOOKCnAoqISB1WrGdS5hCmpwfonayAmTUAYvOjZDzFqJk1Ah4EDiWciTnE3Wdl3FLJu/XrN1RYduyxvRVQRETquKIMKe6+FJgRvRxcSbE9gdhNWjK60WAUcMYSbjS4Evilu0/PoqmSZ6Wls9h//7EsWZL0no4iIlKHFWVIiYyPnk82s2RDjC+MnmdmconGzIwwiuckwr16jnL3l6rVUsmL2Cie1177gqFDH1BQERGpZ4o5pNxFGGnTCnjKzHaAMButmd1IOAsCcHnihmbm0WN0knpvJYzeWQccV9VQYymMxGHGy5evYe3a9YVtlIiI1Khi7TiLu6+MJlebCuwKzDazJYQ5URoQ+qxc7u7Pp1unmW0BjIrtArjLzO5K0YbO2bZfspcYUPr06cTUqcPVB0VEpJ4p2pAC4O7vmtmOwGXAYUBXwr113gJudfeM+qJQ/sxRY6BTThoqOaOAIiIiMUUdUgDc/TvC2Y9RVZWN28YqWT6fjfcEkiKjgCIiIvGKuU+K1CMKKCIikkghRQruiSfmKKCIiEgFCilScPvt152+fUMfZQUUERGJUUiRgmvfvhlTpgxn+PA+CigiIlKm6DvOSv3Qvn0zxo49otDNEBGRIqIzKVLjxo6dxfjx7xe6GSIiUuR0JkVqVGwUT7g7AZx00k4FbpGIiBQrnUmRGhM/zHjDBueWW6azbl3FOxyLiIiAQorUkGTzoEyefAqNGukjKCIiyekbQvJOE7WJiEg2FFIkrxRQREQkWwopkjcKKCIiUh0KKZIXCigiIlJdCimSdwooIiKSDc2TInkxYkRfAP7yl7d4/vlTFFBERCRjCimSNyNG9OWUU3bWMGMREcmKvj0kJ558cg4//bSqwnIFFBERyZa+QaTaSktnceSRExg8+P6kQUVERCQbCilSLfGjeGbM+Ibf/W5yoZskIiJ1hEKKZC1xmHHfvp256abBhW2UiIjUGQopkpVkAWXKlFM1ikdERHJGIUUypoAiIiI1QSFFMqKAIiIiNUUhRdKmgCIiIjVJIUXSsn79Bv7+9xkKKCIiUmM046ykpWHDBjz33CkMGXI/a9duUEAREZG8U0iRtLVtW8Lzz5/K+vUbFFBERCTvdLlHKvXpp4srLGvbtkQBRUREaoRCiiRVWjqLbbb5C6WlswrdFBERqacUUqSC2Cie9eudM854kokTPyp0k0REpB5SSJFyEocZ9+nTmf32617YRomISL2kkCJlNA+KiIgUE4UUARRQRESk+CikiAKKiIgUJYWUek4BRUREipVCSj32+ec/MXLkvxRQRESkKCmk1GPdu7dl7NgjaNDAFFBERKToaFr8eu7EE3eiVaum9O+/uQKKiIgUFYWUembdug00alT+BNphh21ToNaIiIhUTpd76pHS0lnsuec9LFq0otBNERERqZJCSj0RG8Xzn/98y4EH3q+gIiIiRU8hpR5IHGYsIiJSGyik1HGaB0VERGorhZQ6TAFFRERqM43uqaMUUGrOkiVL+OGHH1i7dm2hmyIikrbGjRuz6aab0rp160I3pVIKKXWQAkrNWbJkCd9//z1du3alWbNmmFmhmyQiUiV3Z+XKlXz99dcARRtUdLmnjvnXv+YqoNSgH374ga5du9K8eXMFFBGpNcyM5s2b07VrV3744YdCN6dSCil1zIABPejfvxuggFIT1q5dS7NmzQrdDBGRrDRr1qyoL1UrpNQxrVs35dlnT+ass3ZVQKkhOoMiIrVVsf/+Up+UOqh166aMGXN4oZshIiJSLTqTUsuNHTuLMWNmFroZIiIiOaeQUouNHTuL009/kl//+ikFFcna6NGj2WSTTWpsX2ZW9ujcuTOHHXYY7733Xo3svyo9evQoa1uTJk3o1asXl1xyCcuXL69Q1t0pLS1lzz33pGXLlrRu3ZoBAwYwadKkpHVv2LCBe+65h7333pvWrVtTUlLCjjvuyFVXXcVPP/2U53dWOO5O3759GTt2bKGbUmPuuOMODj30UDp06ICZMW3atLS3ffLJJ9lpp50oKSlhhx12YMKECRXK/Pzzz5x++um0a9eONm3acPLJJ7No0aKy9e7OTjvtxP3335+Lt1NQCim1VCygxEbx/P3vM1i7dn1hGyWShjZt2jB9+nSmT5/Obbfdxscff8zgwYP58ccfC900AE466SSmT5/OlClTGD58OLfeeiujRo2qUO6cc87hV7/6FXvuuScTJ05kwoQJ9OjRg2HDhvHHP/6xXNkNGzZw/PHHc+6559K/f38efvhhnn32Wc444wzGjRvH1VdfXVNvr8Y9/PDD/Pjjj5x00kmFbkqNGTduHD/++CMHHXRQRtu99tprHH300QwcOJBnn32WQw89lBNPPJHnn3++XLnjjjuOadOmcc8991BaWsrbb7/NEUccUbbezLj44ou5+uqrWbduXS7eUuG4ux7VfOy2226eKNmybHW/5Kmyh7t7aek7bjbaITz69LnTFy5cnrP9Sfo+/PDDQjeh2q666irv0KFDwfY1ffp0B/yBBx6okTasWLGi0nXdu3f3Cy64oNyyX//6115SUuLr168vWzZx4kQH/M4776xQx8UXX+wNGjTwmTNnli27/fbbvUGDBv7CCy9UKL9y5UqfMmVKNm+lWtasWePr1q3L+3723ntvv/zyy6tdz7p163z16tU5aFH+xT4r77//vgP+0ksvpbXdkCFDfODAgeWWDR061H/xi1+UvX799dcd8Jdffrls2ZtvvulAuc/XypUrvWXLlj5p0qQq95vq91hl32XADK+B71edSallEs+g9OnTialTh2sUj+TViy++yJ577klJSQmdOnXinHPOYdmyZeXKvPfee+y9996UlJTQu3dvnnnmGfr168eIESNS1t2nTx8Avvzyy3LL77nnHnr37k3Tpk3p3r07N954Y4Vt//rXv9KtWzdatGjBEUccwdSpUyucXjczbrnlFs477zw6duzITjvtlNF779OnD6tWrWLBggVly/785z+z9dZbc9ZZZ1Uof/nll9OqVSv++te/li279dZbOeKIIzjwwAMrlC8pKWHQoEEp2/Dee+9x+OGH07ZtW1q2bMkee+zBCy+8AEBpaSlmVuHfo0ePHlx44YVlr/fff3+OOeYYxowZw1ZbbUVJSQnjx4/HzJg9e3a5bRcvXkyTJk245557ypa9+uqrDBgwgObNm9OhQwfOOussli5dmrLd8+bN4/XXX+eYY44pt3zcuHHss88+tG/fnnbt2jFw4EBmzJhRrsyIESPo168fTzzxBL1796akpIQ333wTCJdE+vXrR0lJCZ07d+biiy8uN4x2zpw5nHDCCXTr1o3mzZvTu3dvbrvtNjZs2JCyvbnSoEHmX62rV6/mpZde4rjjjiu3/IQTTmD69On8/PPPADz77LN06tSJ/fbbr6zMHnvsQc+ePXn22WfLlpWUlHDIIYcwbty4LN9FcdDoHsDMDgb+DDQE7nH3GxLWNwXGAbsBi4Dj3X1+Tbdz2fsLOf3GGQooUqNmz57NwQcfzODBg3nsscf48ssvufTSS/n000957rnnAFixYgUHHXQQnTt35sEHH2TVqlWcf/75LF68mB133DFl/V988QUAPXv2LFt20003cfnll3PxxRez//77M3PmTH7/+9/TvHlzzj33XAAmTpzI//zP/3DOOecwbNgwXnvtNc4888yk+7jpppvYb7/9uP/++zP+ovriiy9o1apVWb+ddevWMX36dM455xwaNmxYoXybNm0YOHAgr7zyChDC12effcZll12W0X5j5syZwy9+8Qu23XZb/v73v9OhQwdmzJhRIdSl49///jeffPIJf/zjH2nevDn77LMPXbp04eGHHy53yWnixIkAHH300WXbHXjggRxxxBE8+uijLFq0iEsvvZTFixfz6KOPVrq/qVOn0qJFi7IgGjN//nyGDx/OVlttxZo1a3jwwQfZd999mT17NltuuWW5chdffDFXXnklnTt3pmfPnjz88MOceOKJ/PrXv+a6667jk08+4bLLLmPDhg386U9/AuDrr79m22235eSTT6ZVq1bMmjWLq666ipUrV6b8d3B31q+v+rJ5o0a5/+r85JNPWLt2Ldttt1255dtvvz0bNmzg448/Zvfdd2fOnDkVysTKzZkzp9yyvffem6uvvhp3L/qhxpWp9yHFzBoCfwMGA18Bb5vZJHf/MK7YmcBid9/azE4A/ggcX5PtXPb+QhY9M7/stQJK8epx6dOFbgLzbzg0Z3Vdc801dO/enUmTJpV9Kbdv357jjz+e6dOn079/f+677z4WLVrEjBkz6Nq1KwBbbbUVe+65Z9I6Y9fJP//8c84991z69u3LsGHDgHCrgauvvpr/9//+H1dddRUAgwcPZsWKFVx77bWcffbZNGzYkOuuu45DDjmEv/3tbwAMGTKEhQsXcuedd1bYX5cuXZJ2QEzG3Vm3bh1r1qzhlVde4e9//ztXXHFF2XtfuHAhq1evpnv37pXW0b1797IAF5t2fIsttkhr/4muvvpq2rRpw6uvvlo2ceDgwYOzquunn35i1qxZdOrUqWzZsccey4QJE8qFlAkTJjBkyBDatWsHwKWXXsree+9d7hh27dqVQYMG8cEHH1QaRGfOnMn2229f4czClVdeWfbzhg0bGDx4MG+99Rb//Oc/y61btGgRU6ZMoW/fvkD4t7nooosYPnw4d9xxR1m5pk2b8tvf/pbLLruMDh06MGjQoLKzU+7OPvvsw4oVK7j77rtThpSXX36ZgQMHVro+5rPPPqNHjx5VlsvE4sWLAWjbtm255bF/g9j6xYsXVygTK/fpp5+WW9anTx8WL17MvHnz6NWrV07bW1N0uQf2AOa5+6fuvgZ4CBiWUGYYEOua/igwyGowliqgSCG99dZbHHnkkeXOGhx99NE0atSI1157DYC3336b3XbbrSygQDgFHf9lGLNo0SIaN25M48aN2XrrrXnnnXd4/PHHadq0KQDTp09n+fLlHHvssaxbt67sccABB/D999/z1VdfsW7dOt555x1++ctflqs78XXMIYcckvb7veWWW2jcuDEtWrRg6NChHHDAAVxyySVpb1+ZbH9lvPjiixx//PE5mdl4t912q/BvcvzxxzN37lzeffddIISw2D4hnCWbPn06xx13XLl/j3322YfGjRszc2blIwu/++67pCPHPvroI4488kg6depEw4YNady4MXPnzuXjjz8uV65r165lAQXg448/5osvvqjQlgMOOIBVq1bxwQcfALBq1Squuuoqtt56a5o2bUrjxo254oor+Oyzz1J2JN1tt914++23q3xsttlmqQ90kYgd+++++67ALclevT+TAnQF4s+bfgUk/vlXVsbd15nZz0AHYGFNNNCaNAQDXAFFat63335b4YutYcOGdOjQoWxEznfffUfHjh0rbJtsWZs2bZgyZQrr16/n3Xff5cILL+Skk07i3//+Nw0aNGDhwvDfqnfv3knb8+WXX9K0aVPWr19fof5k+wOShqXKnHLKKYwaNYrly5czduxY7rvvPu68807OPvtsIPzib9q0KZ9//nmldXz++edlgS32HLuslalFixbRpUuXrLZNlOw49O/fny222IIJEybQp08fHnvsMRo1alQ2WmTx4sWsX7+ec845h3POOafC9qkuO61atYrmzcv/rlq6dClDhgyhU6dO3HLLLXTv3p2SkhJ+9atfsWrVqpTtjX02KgudsbZccskl3HPPPVx11VXsuuuutG3blieffJJrr72WVatW0bJly6Tbt2zZslwoqkw+LvfEzpjE+p7ExM6gxNa3a9euXP+o+HKxMjGx4J94XGsThZQcGjNmDGPGjAFI+iHKVott28GwrVjyxrcKKLVALi+1FIMuXbpUuAHZ+vXrWbRoEe3btwegc+fOzJ07t8K2yf4fNGrUiH79+gGw55570qxZM4YPH84jjzzC8ccfX1bnU089lfRLddttt6VZs2Y0bNiwQv2V/b/L5CxGp06dyto3YMAAPv/8c6688kqGDx9OixYtaNSoEf379+fpp5/mT3/6U4VLGUuWLGHatGkceeSRAHTr1o0tt9ySyZMn86tf/SrtdsR06NCBb7/9ttL1JSUlAKxZs6bc8tiXW7xkx8HMOO6443j44Ye57rrrmDBhAkOHDqVVq1ZAuPxgZowePTppOEh1VqF9+/YV/oqfPn06X331FS+88EK5vhWJX87J2hv7bIwZM4ZddtmlQvlYv6ZHHnmE//mf/+Hiiy8uW/f001Vfhi3k5Z6tttqKxo0bM2fOHAYMGFC2fM6cOTRo0IBtttkGgO22245XX321wvZz5swpNwwZKJt/J3bcaiNd7oGvgW5xrzePliUtY2aNgDaEDrTljBw5khkzZjBjxoxK/6LLVott29H51O0VUKTGxeYBie9Q+Pjjj5ed8gfYfffdmTlzZln/CwiXib7//vsq6z/llFPo3bt32dwi/fv3p1mzZnzzzTf069evwqNVq1Y0atSIXXbZhSeffLJcXZVNpFYd119/PQsXLuTee+8tWzZq1Cg+/vjjcqNfYm644QaWLFlS1sEX4LzzzuPxxx/npZdeqlB+1apVvPjii5Xuf9CgQTz88MOV/jW8+eabA+ESSsybb77JkiVLqn5zkRNOOIFPPvmEp556ipdffpkTTjihbF2LFi3Ya6+9mDt3btJ/j1QhZdttt+Wzzz4rt2zlypXAxr/yAV5//XXmz59fZTu33XZbunbtyvz585O2pUOHDmX7iK9//fr1PPTQQ1XWX8jLPU2bNmXgwIE88sgj5ZZPmDCB/v3706ZNGwCGDh3Kd999V3apFWDGjBl8+umnDB06tNy28+fPp0GDBmy99dY5b2+NqYlxzsX8IJxN+hToCTQB3gV6J5T5LfD36OcTgIfj1+d7nhQpXnVlnpRWrVr5I488UuHxww8/+AcffOCNGzf2ww47zJ9++mm/6667vG3btn7QQQeV1bF8+XLv3Lmz9+3b1ydOnOjjx4/3Xr16eceOHf30008vt69kc7I88MADDpTNF/LHP/7RmzVr5ldccYVPnjzZn332Wf/zn//sRxxxRNk2jz32mAP+29/+1idPnuxXXnmlb7HFFhXmkAD8L3/5S1rHItk8Ke7ugwcP9h49epSbV+Q3v/mNN2rUyEeNGuUvvPCCP/vssz5ixAgH/Prrry+3/fr16/2YY47xkpISv+CCC/y5557zF1980W+99Vbfaqut/Lzzzqu0TXPmzPFWrVr57rvv7g899JC/8MILfuONN/q9997r7u6rV6/2rl27+q677upPP/2033///b7TTjt569aty72XAQMG+NFHH13pfrbeemvv0qWLt2jRwpcvLz/v0quvvupNmjTxU045xZ944gmfOnWq33fffX7MMcf43LlzK61z8uTJDvgPP/xQtuy7777zli1b+qBBg3zy5Ml+7733erdu3bxr167l2nfaaacl/T360EMPeePGjf3cc8/1p59+2l944QW/6667fOjQoWXtPvbYY71Dhw4+btw4f+qpp3zo0KHes2dPB3zp0qWVtjdX3n77bX/kkUf8lltuccBHjx7tjzzyiL/99ttlZcaOHesNGzb0+fPnly179dVXvWHDhj5q1Ch/6aWX/KKLLnIz88mTJ5erf8iQId6zZ09/7LHHfOLEib7NNtv4PvvsU6EdF198se+0005VtreY50kpeEgohgdwCPAx8AlwRbTs/4BfRj+XAI8A84C3gC3jt1dIqb/qSkgBkj5ik1BNmTLF99hjD2/atKl37NjRzz777Aq/7GfNmuX9+/f3Jk2a+DbbbOMTJ070Xr16+ahRo8rtK1lIWbdunffq1cuHDBlStuz+++/3XXfd1UtKSrxt27a+xx57+M0331xuu9tvv927du3qzZo186FDh/rDDz/sgL/zzjtlZXIRUl5++WUHfPz48WXLNmzY4Pfdd5/vscce3rx5c2/ZsqXvt99+/uSTTyate/369X733Xf7nnvu6S1atPCmTZv6jjvu6KNHj/affvopZbveffddHzp0qLds2dJbtmzpe+yxR7kJ4N566y3v16+fN2vWzPv27euvvfZahfdSVUi54oorHPATTjgh6fo33njDDzroIG/VqpU3b97ct99+ez///PNTtn316tXevn17HzduXLnlzz77rPfu3dtLSkp8p5128qeffrpC+yoLKe7uzzzzjO+zzz7evHlzb9Wqlffp08evuOIKX7t2rbuHIHTEEUd4q1atfNNNN/WLLrrIx4wZU2Mh5bTTTkv6/+m0004rK3Pfffc54J999lm5bSdOnOi9e/f2Jk2a+LbbbusPPvhghfoXL17sI0aM8DZt2nirVq38xBNP9AULFlQot/POO/vo0aOrbG8xhxQL+5Lq6NevnydORNSvX78KkxNJ3fPRRx+x/fbbF7oZRemzzz5jm222YcyYMZx++uk1ss9rr72WP/zhD/z44485GQ0j1Tdq1CjmzZuXVp8QyZ25c+fSu3dv5s2bV2X/mVS/xyr7LjOzme7eLxdtTUUdZ0UkJ66//no222wzunfvzhdffMH1119Px44dyyYEy7UFCxZw/fXXM3DgQJo3b86rr77KH//4R84880wFlCJy0UUXsc022/Dxxx+Xdf6U/Lv11ls55ZRTct7Bt6YppIhITpgZV199Nd988w1NmzZl33335U9/+hOtW7fOy/6aNGnCnDlzGDduHD///DNdunRh1KhRXHPNNXnZn2Rn88035x//+AfffvutQkoNcXd69uzJiSeeWOimVFvRX+4xs87AZcBhhPlKfib0C7nN3adWo97WwMXA0UB3YCUwC7jT3Suf5zkJXe6pv3S5R0Rqu2K+3FPUQ5DNbGfgA+B/gS2B1cAmhMDygpldmmW9mxMCyRXAdsB6oDVwAPCImd1R+dYiIiJSE4o2pJhZM2ASYWbXd4Ad3b0N0A64mTAH63VmNiTDeo0wtX1PYD7wC3dvBbQinFnZAJxtZhVvbyqSRLGfjRQRqUyx//4q2pAC/JpwGWYZcLi7zwZw9yXufiHwBCGoXJ9hvcMI095vAI5099ejele5+03A7VG5/zOzJtV+F1KnNW7cuGxyKhGR2mblypU0bty40M2oVDGHlJOj5/HunjgDLMBN0fOuZrZtFvVOcfdZSdb/iTCmvTPh8o9IpTbddFO+/vprVqxYUfR/kYiIxLg7K1as4Ouvv2bTTTctdHMqVZSje8ysFbBb9HJyJcXeIHSibQMMAireOCS52I0Zktbr7l+b2WxgR0JIeS7NeqUeio1c+eabb1i7dm2BWyMikr7GjRvTqVOnvI3Ay4WiDCnA9oRLOQCzkxVw9w1mNhfYA9ghnUrNbFNCH5dK6418SAgpadUr9Vvr1q2L+j+5iEhtVayXe+LvS/5NinKxdenexzxf9YqIiEiOFWtIaRH3c6peiSui55YFrldERERyrFhDStEzs5FmNsPMZixYsKDQzREREalzijWkLI/7OdVNOJpHz8tqul53H+Pu/dy9X8eOHdPcvYiIiKSrWENKfH+RzVKUi637tsD1ioiISI4Va0iZQ5irBKB3sgJm1gCIzY/yYTqVuvsCYGGqeiOxUT1p1SsiIiK5V5RDkN19qZnNAHYHBgOPJym2J2GOFIBMbjT4EnBsVO8tiSvNrCsbA0xa9c6cOXOhmX0et2gToKuZrcmgXVVpBKzLYX21XSbHQ8cuc7FjpmOXvVTHTse1atU9RvX5GOfyvTcxs3eTLM9kEtWsFWVIiYwnhJSTzez/3D3x0suF0fNMd093IrdYvccCQ8ysj7snHvzfEeZo+ZYQaKrk7uU6pZjZDHfvkUGbqhTVmfc7TtYWmRwPHbvMxY6Zjl32Uh07HdeqVfcY1edjXBPvPTqRkHfFerkH4C7gc8KN/54ysx0gzEZrZjcCR0XlLk/c0Mw8eoxOUu+TwJuE9z7RzPaKtmlqZhcA50XlrnL3XJ4JERERkQwU7ZkUd19pZsMIl1x2BWab2RLC3CUNCH1WLnf35zOs183sGOAVwp2Qp5vZMqCEjcfj7+5+d47eioiIiGShmM+kEF2K2ZFwZ+JPgabAIuBpYLC735BlvV8BfYHrCJ10GwFLCZd3jnP3s6vZ9DHV3L6m6qzNMjkeOnaZG5PwLJlLdex0XKtW3WNUn49xTbz3Gjm+pju3ioiISDEq6jMpIiIiUn8ppBQZMzvYzOaa2TwzuzTJ+qZmNiFa/6aZ9ShAM4tSGsfud2b2oZm9Z2ZTzax7IdpZjKo6dnHljo46pdfLURPJpHPszOy46LM328zG13Qbi1Ua/2e3MLOXzOyd6P/tIYVoZzEys3+Y2Q9m9kEl683Mbo+O7XtmtmtNtzEn3F2PJA+gM/Bn4BNgFfA98C9gUDXrbQ1cC3xEuJHhIkLn4GOAhtH+tgSaAO8COyRsfw6hYy/ACcCEQh+rYjjG0bH7DLgCeBRYHT2WR8f6r8DJQPOo/Nm1+djl8vim87mLyrUidDh/A+gXt+0MQkd2B0YX+n3X8DHeDPiJMBJxFbCW0LdtUFyZXsA7QLvo9aZp1t0VuAF4H1hCuE3HfwnTKAwr9HvPwbFL5/fd/cCLUbnVwPoc/R4+EniKMNXEWkKfxFnR8e5U6GOT5nvYjzCo5INK1h8CPEuYUmMv4M2E9a2AXwLXROUWxv0/3i4H7av0uy6jegp9oIvxAeyc8A/2c/Sfw4ENwKVZ1rs5oQNwrN6l0X+Q2OvHgMlx5S8DLkuoYzLQP/q5UdROK/QxK/QxBvpH23ncIxZUYq9XAidG5XcB/l3o41AMxzc6dik/d9Hy24BDgWlsDCnnJRzz0YV+7zV8jH9KOMYbEo8xcCPwqwzrPpYQTGJ1LyeElNjrKYV+/zk4fik/d9HxXRn3nuPff1a/hwlXD/6Z8JldQpj4LPZ6EbB7oY9Pmu+nB5WHlLtiv++i13OBLnGvj0g4DvGPaoUUqv6uuyPdunS5J4GZNQMmAR0If/3s6O5tgHbAzYRUep2ZDcmwXiP8hd8TmA/8wt1bEdLsxYT/dEex8eaGAF8R/pqK1xX4EsDd1xF+MXbIpC2Flqdj3DXa7hXgNOB/gLsJx3Mfwl9JJcA4M9sZOJPw10Odk8XxLftMRSp87qJTxd3c/em4ZZsT/gr7nHCWpt6IO8ZtCF9qsWP8a8JnLf4YbwNsY2b/NrM3zOzgKuo+BHiQ8LvhH4QvjBbu3pIwm/XRwDP5eWc1qtLPXdzxLWHjWcC1wACq8XsYOItwRhVC6O7k7q2j/RwMfAG0Bx6Mbr1Sm1X5/xr4gfBZuhoYmYudpvldd7aZnZVWhYVOgsX2YONfhkuBrknWT4zWz8yw3iOi7dYDfZOsvzVavwJoEi07FfhrQrkPgM3jXn8CbFLo41boY0y4XPZU3Otyxw7oSPhF58DLhEsWTQt9LIrh+EbH7p4Ux64B4exJj+j1NKBfXD2/jH4Z1ZszKXHHeCUwPvHYxR9jwmWFiUBjwi/uL4G2ldTbmnAjVAf+UOj3medjWOnnLu74rgKujpb1J9xPrUE1fg+/HG03tZL1+7Pxr/2+hT5GabyfHlR+JuUpYJ+411OJzoBGrxsmqavaZ1JI/7vuW6LvulSP2p4U8yGWsse7+9dJ1t8UPe9qZpncuyBW7xR3n5Vk/Z8I/3DNgAOiZZsDiW34GugGYGaN2PiXXG2Sj2P8NeFLIKbcsfNwc8nYX597Ar9099XpN7lWyfT4ln2mIomfu1aE+Yqmmdl8wvXtyYRfRk+5+6TcNb3WiB3j5yh/JjN27MqOMSEsTnL3te7+GfAxoZ9KMqcDXQh/9Y7OcZuLTarPXez4rgDuAXD36YQzHpuQ/e/hTtHzO5Wsnxn3c4sM6i1GKf9fu/v6PO033e+6zmz8rquUQkocM2sF7Ba9nFxJsTcIl1gABmVQ/cBU9UZfJrOjl0eaWRNCx9jEL4BJhMsZEP4SedGjeFob5PEYvw30MrOeKY5dw+j5S3f/Ic16a5Usj2/KY+fuP7v7Ju7ew8M9qd4inHpfSbisVq8kHOMHSH7s4o/xIsJf6JjZJoTLP59WUn3sF/yj7r42960vKkk/dwnHdz7R7wAz254QUhaQ/e/h+dHzLpWsj+13NeGsTW02CRgejfLZC/jZK94DLx8y+a5TSMnQ9oRrnbDxIJbj7hsIHZAAdkinUjPblI1/bSWtNxL7T3EyoUf0w+4+28z+z8x+Ga27F+hgZvMIN0OsdLhokcrLMfbQP+dcwn+Myo7dsOh5EzObZWZ18QxAxsc3zWMXryfhL9Lr3H1+LhtfS8Qf4/dJcuwIZ0Hi++ksMrMPCSN/LnL3Cmc/zayEMBM2wDtmtp2ZPRgNM11lZp+a2Z11ZdqByj53hMsBseP7/4CzLNyF90FghAcZ/46IxG53coCZ3Rr9bsbMGpnZQcDYaP3/ufvibN9bTTCzB4HpwLZm9pWZnWlmvzGz30RFniGE4XmE931ODbQp0++6qv/tCn1NrZgehC+x2DW5VinKxa6HPpZmvX3i6t0pRbnYtbqMrrPWpke+jnGG+z240Mehth5fwl+g6whfEE3ils+nnvRJyePvie3i6r2WcKnDCSN7lsat+xnYv9DHobYd37jtLqf8aJ740T0fEIJQwY9DAY57j7hjklWflHx81+lMSnnx1yBXpii3InpuWeB6a6MaPxZm1pWN95mY5O7PVbfOIpa34xuNdriLcNnsXK+/dwnP1zFuG/fzZcBi4CCgpYfREb8g9GdpDTxiZu3TrLe2yffviOuBEYTwB6HPVexScAvCmVZ9N2Yn5/92+oeQOs3MWgJPAJsShsqeWdAG1W6/BXYnnJZ/odCNqYMaJPw83N2f9+hPT3d/ndAPbQOh8+ivar6JtVvU3+VfhEniXiJ0om9NOIvwm+jnmwh9jaQIKKSUtzzu52YpysXmMllW4Hproxo7FtE1/icJw2UXAAe5+8Js66sl8nJ8zWwzwiWIpcD52TWtzsjXZzi+3Gx3n5pYwN3fB6ZELzPpMFqb5PN3xC2ECQmnuPvh7v6Wuy9198/d/S7CXFUOnGBmQzNqtUAe/u0UUsr7Ju7nzVKUi61Lt6d0vuqtjWrkWESjBR4l9B7/CRji7nNTblQ35Ov4Xk/4K/NGYImZtYx/sLGjY5O4ZXVVTfyeSPVZja3rlqJMbZaX42tmrQlDvCFM5FaBu78M/Cd6OSxZGUkp5/92CinlzSGkaIDeyQpE1ypj4/LTGqLmYY6O2F/wSeuNxHo61/ahb6nk5RgnbN+IMBLgUEJSP8STj9evi/J1fLtHz9cQzqYkPraI1l8Wt6yuytfviYVkNnNvrZl6IEP5+gz3YmPfk89SlIsND++RZr0Sycd3nUJKHHdfSrhZGsDgSortSZhADcIMful6KVW9UefO2D9qJvXWKnk+xrFfXmMJp21XEiZtm55FU2ulfB9fyfsxjl3KSTVB2XbR8/wM6q018nh8N8T9vEWlpTYG8roctPMpp991CikVxW6jfrKZdUmy/sLoeWaGlw9i9Q4xsz5J1v+OcMr8Wzb+I9dVeTnG0T0jxgAnAWuAo9y9rh/LZHJ+fN19f3e3yh6ETskQpjCPLavL8vV7Ylz03NvMDkxcaWY7sbEvSl24f09l8nF85xImaYNwD58KontU7Rq9fDPNeqW83H7XFXpsdrE9CJ195rPxvhs7RMtbEa7Hx8aAD0mybWzd6CTrjDBLohNOJ+4VLW8KXMDGO9SeVehjUIuP8W3RurXUgVvZF9vxrWKf87PZrrY+8nmMgeej9V8T/hq1aHl/whdt7HdI80Ifh9p2fAlD6GPr7ybcNBPCTLbDCDcYjM1FU6vuiZblcd4k7rFL3LHZK2FdgwyOcU6/6wp+kIrxQZiQJq3b3Kf7Dxetr+r21XcW+r3X1mNMOH0bW7cG+C7Vo9Dvv7Yd3zT2Nz+b7WrzI4+/JzoA78WVW06YcCz2+mvCXZcLfgxq2/ElzMvxalwZJ/RbWx/3eglJwk9dfCQch1SPHuke42h9zr7rdLknCXd/l3BDtdsJB7op4f4bTwOD3f2GLOv9ijDt9XWEzmGNCP94LwHHufvZ1W58LZGHYxz/WW5MmLY91aNOy9dnWDbK4++JRYT5aC4h3AhvA+F3xWzC746d3f2Dar+BIpeP4+vuywj3UTqDcMZqQVTvSsJss7cSZkp9Pgdvod7K5Xdd7DSiiIiISFHRmRQREREpSgopIiIiUpQUUkRERKQoKaSIiIhIUVJIERERkaKkkCIiIiJFSSFFREREipJCioiIiBQlhRSRGmRmI8zMzWxaodtS20XH0c2sR5bbz4+23z+nDRORnFFIEQHMrDTuSy/V47xCt7UmmNm0JO99g5n9ZGZvmdmVZta+0O1Mxsz6mtloMxtR6LbkSvR+kn0el5vZf81srJntkad9j4j23zcf9Yuk0qjQDRApMmuBH1OsX15TDSkSqwg3dgNoSLgj6u7RY6SZDXL3uQVqW2y/axOW9wWuAl4GSlNs/wnh/a3IdcPyaAPhfjMxHYCto8cpZnaBu9+W432OAAYQbiI5K8d1i6SkkCJS3uvuvn+hG1FEJrj7iNgLM2tNuDnbDUBX4CEz29ULcBMwd9+umtsPylVbatCX7t4j9sLMGgJ7An8jhLObzWxKfbgBodQPutwjImlz9yXRX+p/iBb1BfYqWIPqOXdf7+6vA0cQzig1AE4paKNEckghRSQLZrarmd1gZq+Z2RdmttrMFkV9OX4V/YWbaZ1NzGyUmb0e9f1Ya2bfm9m7ZvY3M+ufYrtzzexVM/sxasvnZvYPM9u++u82qQfjft4toT2dzOxmM5tjZivM7OeoH8sFZta0sgrNbJiZPRO957XRe5lrZg+a2fFJylfoOGtmDtwXvRyQpA/H/nFlK3ScNbO7o2WPpnrzZnZZVO6dStYfbmZPmtl3ZrbGzH4ws3+Z2UGp6s2Wu38OfBy93CFJe5qa2bFmNi76PC00s1XR5+QBM9styTYjouM5IFp0X8KxnJ9km0J8FqUuc3c99Kj3D0LfBQempVl+YVTeCf1UFse9duBpoFGS7UYk2w/h0uu0uO03RHWui1v2UJL6uhD6CcTKrAeWxL1eCRyVxfGItaW0kvXN4vZxedzyPYBFceuWRG2IvZ4FbJqkvj8kHL/E7b5Lsk1sXY+4Zd8R+tA4sCZ6Hf/YO67s/Kjc/nHLBsYdt9Ypjs97UbmLEpY3Bv6Z8F5+Tnj9xyz+PUZH285PUWZ27LOXZN1hCZ+tHxOO71rg1IRtjo+O2Zq49xF/LN+uic+iHvX7oTMpItl5HjgR6OLuLdy9HdASOJXwC/wQ4PwM6juJ8BfriqiO5lGdTYHuwLnAu/EbmFlj4EmgDzAV2BsocffWwGbAbUAJcL+ZbZXd26zUFnE//xS1px3wBNAeeB/YI2pLS+BYQujqAzyQ8D56AJdGL68HOrp7a3dvBmwKHEMIfVVy987AqOjl6+7eOeHxehVVvAx8TThuRyYrYGa9gZ0IX7wPJqy+ETgZmAccB7R09zZAa+AcYClwsZmdmM77SVd0DHtFLz9NUmQZcDuwX9Sm9tHx7U74nDQCxphZ2b+ru0+IjmfsmI1KOJa7x+2/kJ9FqcsKnZL00KMYHmw8k5Lsr+/Y474069o3quuzJOtGkPxMyh3R8jszaPOvom1eARpXUubvUZm/Zng8ppH6TMqf2PgX8m7Rst9HrxcDnZNsMyRumwPilh8XLfsowzZWOJOS6hgn2X4+CWdSouU3R8snV7LdtdH6lxOW9yKcpfgB6FbJtidE236Q4XsdTZIzKYQRV/2Bd+KOx55ZfP7vjba9KsVnYUQhPot61O+HzqSIlNcY6FTJo106Fbj7q4SzCz3MbLM097skeu6SQVtPi57/7O6Jw3BjYmctBmdQb1Jm1tDMepnZ9cB50eLp7j4z+vmY6Pked/8ucXt3fx6YHr08Lm5V7L23MbPm1W1nDoyPngeZ2aZJ1p+QUC5mOGCEEVFfVlL3o8BqoLeZZfJvHdMt6ufynZl9R7iE8jqhAzPAaHd/M4t6/xU9/yKLbaGGP4tSf2gIskh5L3uaQ5DN7FjCqf1dgY6E09mJNgO+SaO6Z4FLgGFmNolwZudld19Uyb4bEfp/ANxlZn+rpN5YB95uabQhmdPM7LRK1s0l+sI2sybAjtHyl1LU9yLhL/9d45a9Segj0QWYHr2XF9z9syzbXC3uPtPM5gLbEsLUX2PrzGxPYCtCH45HEjbdO3o+LfpsVKZx9NwN+DbD5jUgBOZEq4Cj3f2Zyja0MPneb4GhhPfWho2fj5h0Q3V8vTX1WZR6SCFFJEPRL+WHKd9nYTWhM+366HVHwhdKi3TqdPeXzexK4Erg8OiBmc0h9Me4y93/G7dJe6BJ9HOHNHbRLJ12JBE/mdsGQp+K/0ZtGuvusYnQ2rNxtODXKer7KnruGFvg7ovN7FRCh9OdgbsAojMFzwP/cPeXs2x/th4kXGI5kbiQEr2GcCkocdK/2JmRVtGjKtmcNfrco3lSon4gWxL64JxNCAj7uvv8xI3MbAdCQIwPOEvZ2Hm2CeFMYVqf1wQ19VmUekiXe0QydxYhoKwA/pfQ/6DE3Tt61KmQjWdPLN1K3f0aYBvgMmAy4TLIdsAFwIdmNjyuePz/3V3c3ap6ZPleJ/jGjpKbufu27n6Yu98ZF1ASJTujlFJ0BqAnMJIQAL8BOhMuoUwzszFZtj9bsUs5/WNDnM2sARsvUyVe6oGN/ybnp/Pv4e7TqtNAd1/r7nPd/RzgbmBz4MGonYnuIwSU/wAHA608dE7uFH1eY2d+svmc1NRnUeohhRSRzMV+oV/j7n9x96/iV1qYI2WTbCp298/c/QZ3P5jwF+pAQmfERsAdcX0kFrHxrM0WFWuqcT8SzrRA6vZsHj0vSFzh7j+7+93ufry7dwV6E758Ac4ys0Nz1toqRGetZhC+tGN9UPYnnC1ZThjJkuj76LkQ/x6XEM547UUYHVYmGrGzB+Hz8kt3n+zuyxK2T3YJKV3F9lmUOkQhRSRzsS/adypZ/wuyOJuQyMNsotMIc1ysJZyK7xetW0v4EoXQx6Cg3H0NEJuKfWCKogdEz/9Jo84P3X0k8Ea0aECazYmFper+xR47W3JS9By71PNkJWeRYp2CD67mfjPm7osJU+MDjI4uScaUBUN3r+xS3IEpqk95PIvtsyh1i0KKSOZifTR2SlwRfTlcm2mFUcfTyqxh41+q8TO2lkbPI8ysTxX1pzUyqZpis7SOSDZyxcyGEDrNQrikE1ue6r1D6DcB5d97KrHRQm3TLF+Zhwhf0DuZ2S7A0dHyZJd6AMYR+ndsb2a/TlVxnv49/kLoG9WD8lPjxz6vnZKNVjKzndgYxJJJ53iWRs/F8lmUOkIhRSRzL0TPv4+mcm8IYGbbEYZy7kHmd0seZ2b3mdlBZlbW6TLqDzGWcGZmJfBq3Db3Es4ylAAvmtlZFm4AGNu2s5mdbGYvs3GCs3z6K2G0SjPgOTPrF7WjoZkdTfjSB5ji7i/GbXe2mU02s5Piw42ZtTWzywmXWSD000nH7Oh5h2g0Tlbc/VvCHCEA9xA6li4idOZNVv5D4Nbo5R1mdr2Zxc5iYGatzGyImf2TiiODqi0a9n1/9PKyuL4pHxE6LBswwcy2jtrT2MyOInyeEy//xIsdz6PMrE0lZYrtsyh1RaEnatFDj2J4kMG0+IS+IvPYOHnWGjZOfb6OMJnYfJJPFDYi2X4IM7XGT1u+mBB0PK7eU5O0ZVPgtbhy6wlfpMviliWdpKuK9ziNFJO5pdhuD0L/lNh+E6e3f5eEafEJc67Et3UZFW8zcFeSfSWdzC1a93Lc+kXRv8d8YK+4Mkn/jRLqOTOhHSkn2yMMs70jYZufCfPmbIhb9lKGx3U0VUyLH5XbNvoMOHBi3PIj45bH/l1WRz9/TjjzkrR+QuftWNm1hNFb84HXauKzqEf9fuhMikiGPAw93Qu4k41DalcSgsYAdy/NotpLgYuB5wjTmjchfOF9QhiZsau735+4kbv/QOircTLwDKFDauxMzBzCJYjjgBuyaFPG3P0twg3ubiXc8K4xIWDNAC4izIb6Q8Jm4wkjpiYQ/upfS5hK/1tgEqGzZ8rLJ0kcRQgLn0V1dY8emfYVeozwBR3f1kp56Ed0DrAPYUj154TLVCXAF4T3cy4bJ77LKXefG+0D4HIzs2j5REJ/oBcIQ48bR237E7ALGz/HyeqcQ5iA7TlC4OpMOJabJ5Qrqs+i1A3m7oVug4iIiEgFOpMiIiIiRUkhRURERIqSQoqIiIgUJYUUERERKUoKKSIiIlKUFFJERESkKCmkiIiISFFSSBEREZGipJAiIiIiRUkhRURERIqSQoqIiIgUpf8PRdIEk3J0LtQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#from graphpattern2vec import link_prediction as LP\n",
    "ver_name = list_embedding_file[19:] + '.png'\n",
    "# From GridSearch Best: 0.997333 using {'C': 1.0, 'penalty': 'l2', 'solver': 'liblinear'}\n",
    "# from GridSearch Best: 0.991940 using {'C': 1.0, 'penalty': 'l2', 'solver': 'liblinear'}\n",
    "#scoring = ('accuracy', 'f1', 'roc_auc') # accuracy, f1_micro, roc_auc, roc_auc_ovr\n",
    "\n",
    "lp.train_model_cross_validation_with_split('liblinear', 'ovr', C=1, penalty='l2', cv = 10, ver = ver_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  1%|          | 1/151 [00:00<01:21,  1.85it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predict_for_specific_types| source_nodes: 151, target_nodes: 1974\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 151/151 [01:31<00:00,  1.65it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predict_for_specific_types| used cuttoff: [0.38878835440223275]\n",
      "predict_for_specific_types| missing sources: 0, missing targets: {1929, 5273, 3539, 4484}\n",
      "predict_for_specific_types| possitive predictions: 60053 (negative predictions: 237417)\n",
      "predict_for_specific_types| removing the known predictions...\n",
      "predict_for_specific_types| all prediction: 60053, prediction for unkowns: 59561, remained(known): 492\n",
      "predict_for_specific_types| Raw Predictions saved.\n",
      "\n",
      "predict_for_specific_types| all predictions (297470, 5) saved.\n"
     ]
    }
   ],
   "source": [
    "dark_kinases = lp.get_node_type_to_id()['DarkKinase']\n",
    "df_raw_prediction, list_prediction_all ,neg = lp.predict_for_specific_types('DarkKinase', 'Pathway', filter_source=dark_kinases)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>DarkKinase</th>\n",
       "      <th>Pathway</th>\n",
       "      <th>probability</th>\n",
       "      <th>odds</th>\n",
       "      <th>logOdds</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>199418</td>\n",
       "      <td>0.9999999993935897</td>\n",
       "      <td>1649048627.6928833</td>\n",
       "      <td>21.223464369290337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>168176</td>\n",
       "      <td>0.9999999990804023</td>\n",
       "      <td>1087431998.848001</td>\n",
       "      <td>20.807084789177104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Q6P5Z2</td>\n",
       "      <td>5663220</td>\n",
       "      <td>0.9999999976453935</td>\n",
       "      <td>424699410.9191864</td>\n",
       "      <td>19.86689220823094</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>P31152</td>\n",
       "      <td>5673001</td>\n",
       "      <td>0.9999999971097913</td>\n",
       "      <td>345995769.65313166</td>\n",
       "      <td>19.661937106511218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>5683057</td>\n",
       "      <td>0.9999999961227581</td>\n",
       "      <td>257915299.5859731</td>\n",
       "      <td>19.368141792806504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59556</th>\n",
       "      <td>Q8TD19</td>\n",
       "      <td>8949664</td>\n",
       "      <td>0.3888789847328807</td>\n",
       "      <td>0.6363371165740433</td>\n",
       "      <td>-0.45202679856651534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59557</th>\n",
       "      <td>Q9NRP7</td>\n",
       "      <td>8866652</td>\n",
       "      <td>0.3888707948345155</td>\n",
       "      <td>0.6363151876029475</td>\n",
       "      <td>-0.4520612604081764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59558</th>\n",
       "      <td>Q8TD19</td>\n",
       "      <td>113510</td>\n",
       "      <td>0.3888432941456489</td>\n",
       "      <td>0.6362415570685348</td>\n",
       "      <td>-0.45217698103879367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59559</th>\n",
       "      <td>O43781</td>\n",
       "      <td>6811434</td>\n",
       "      <td>0.38883354126627917</td>\n",
       "      <td>0.6362154462335933</td>\n",
       "      <td>-0.4522180210658868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59560</th>\n",
       "      <td>Q9BYP7</td>\n",
       "      <td>9646399</td>\n",
       "      <td>0.3888007970758884</td>\n",
       "      <td>0.6361277881511949</td>\n",
       "      <td>-0.4523558110586626</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>59561 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      DarkKinase  Pathway          probability                odds  \\\n",
       "0         Q9BUB5   199418   0.9999999993935897  1649048627.6928833   \n",
       "1         Q9BUB5   168176   0.9999999990804023   1087431998.848001   \n",
       "2         Q6P5Z2  5663220   0.9999999976453935   424699410.9191864   \n",
       "3         P31152  5673001   0.9999999971097913  345995769.65313166   \n",
       "4         Q9BUB5  5683057   0.9999999961227581   257915299.5859731   \n",
       "...          ...      ...                  ...                 ...   \n",
       "59556     Q8TD19  8949664   0.3888789847328807  0.6363371165740433   \n",
       "59557     Q9NRP7  8866652   0.3888707948345155  0.6363151876029475   \n",
       "59558     Q8TD19   113510   0.3888432941456489  0.6362415570685348   \n",
       "59559     O43781  6811434  0.38883354126627917  0.6362154462335933   \n",
       "59560     Q9BYP7  9646399   0.3888007970758884  0.6361277881511949   \n",
       "\n",
       "                    logOdds  \n",
       "0        21.223464369290337  \n",
       "1        20.807084789177104  \n",
       "2         19.86689220823094  \n",
       "3        19.661937106511218  \n",
       "4        19.368141792806504  \n",
       "...                     ...  \n",
       "59556  -0.45202679856651534  \n",
       "59557   -0.4520612604081764  \n",
       "59558  -0.45217698103879367  \n",
       "59559   -0.4522180210658868  \n",
       "59560   -0.4523558110586626  \n",
       "\n",
       "[59561 rows x 5 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_raw_prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df_node = pd.read_csv(data_path + node_file)\n",
    "# node_id_to_name = dict(zip(df_node.id, df_node.name))\n",
    "# len(node_id_to_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# [ node_id_to_name[path] for path in {6854, 8272, 3664, 4791, 3768, 4217, 3547, 8830, 2271}]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Uniprot to Gene Name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "df_uniprot_mapping: (44616058, 3)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(44614225, 'CAMKK1')"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "df_uniprot_mapping= pd.read_csv(data_path + 'original/GeneNameMapping.txt', sep='\\t', names=['uniprot', 'mapping', 'GeneName'])\n",
    "print('df_uniprot_mapping: {}'.format(df_uniprot_mapping.shape))\n",
    "\n",
    "uniprot_to_gene = dict(zip(df_uniprot_mapping.uniprot, df_uniprot_mapping.GeneName))\n",
    "len(uniprot_to_gene), uniprot_to_gene['Q8N5S9']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_reactome = pd.read_csv(data_path + 'original/' + 'UniProt2Reactome_All_Levels.txt'\n",
    "                          , sep='\\t', names=['uniprot', 'pathway_id', 'url', 'pathway_name', 'abr', 'd'])\n",
    "# df_reac_human = df_reactome[]\n",
    "df_reac_human = df_reactome[df_reactome['pathway_id'].str.contains('R-HSA')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(20969, 2649)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# list_pathway_ids = list(df_reactome['pathway_id'].unique())\n",
    "\n",
    "pathway_id_to_name = dict()\n",
    "\n",
    "pathway_code_to_id = dict()\n",
    "\n",
    "for i in df_reactome.itertuples():\n",
    "    pa_id = i[2]\n",
    "    pa_name = i[4]\n",
    "    if pa_id in pathway_id_to_name and pa_name != pathway_id_to_name[pa_id]:\n",
    "        print('err')\n",
    "    pathway_id_to_name[pa_id] = pa_name\n",
    "    \n",
    "    pa_code = pa_id[pa_id.rindex('-') + 1:]\n",
    "    if pa_code not in pathway_code_to_id:\n",
    "        pathway_code_to_id[pa_code] = list()\n",
    "    pathway_code_to_id[pa_code].append(pa_id)\n",
    "    \n",
    "len(pathway_id_to_name), len(pathway_code_to_id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "node_map_file = ''\n",
    "final_prediction_names = ['uniprot', 'ProteinName', 'Pathway_id', 'Pathway_Name', 'Score']\n",
    "prediction_file_path = data_path + result_path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uniprot</th>\n",
       "      <th>ProteinName</th>\n",
       "      <th>Pathway_id</th>\n",
       "      <th>Pathway_Name</th>\n",
       "      <th>Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>MKNK1</td>\n",
       "      <td>199418</td>\n",
       "      <td>Negative regulation of the PI3K/AKT network</td>\n",
       "      <td>0.9999999993935897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>MKNK1</td>\n",
       "      <td>168176</td>\n",
       "      <td>Toll Like Receptor 5 (TLR5) Cascade</td>\n",
       "      <td>0.9999999990804023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Q6P5Z2</td>\n",
       "      <td>PKN3</td>\n",
       "      <td>5663220</td>\n",
       "      <td>RHO GTPases Activate Formins</td>\n",
       "      <td>0.9999999976453935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>P31152</td>\n",
       "      <td>MAPK4</td>\n",
       "      <td>5673001</td>\n",
       "      <td>RAF/MAP kinase cascade</td>\n",
       "      <td>0.9999999971097913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Q9BUB5</td>\n",
       "      <td>MKNK1</td>\n",
       "      <td>5683057</td>\n",
       "      <td>MAPK family signaling cascades</td>\n",
       "      <td>0.9999999961227581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>P78368</td>\n",
       "      <td>CSNK1G2</td>\n",
       "      <td>176974</td>\n",
       "      <td>Unwinding of DNA</td>\n",
       "      <td>0.9999999936017137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>O75747</td>\n",
       "      <td>PIK3C2G</td>\n",
       "      <td>1483249</td>\n",
       "      <td>Inositol phosphate metabolism</td>\n",
       "      <td>0.9999999928222838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Q99640</td>\n",
       "      <td>PKMYT1</td>\n",
       "      <td>69231</td>\n",
       "      <td>Cyclin D associated events in G1</td>\n",
       "      <td>0.9999999922670001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Q13188</td>\n",
       "      <td>STK3</td>\n",
       "      <td>390471</td>\n",
       "      <td>Association of TriC/CCT with target proteins d...</td>\n",
       "      <td>0.9999999922031578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Q9HC98</td>\n",
       "      <td>NEK6</td>\n",
       "      <td>2555396</td>\n",
       "      <td>Mitotic Metaphase and Anaphase</td>\n",
       "      <td>0.9999999896012494</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  uniprot ProteinName Pathway_id  \\\n",
       "0  Q9BUB5       MKNK1     199418   \n",
       "1  Q9BUB5       MKNK1     168176   \n",
       "2  Q6P5Z2        PKN3    5663220   \n",
       "3  P31152       MAPK4    5673001   \n",
       "4  Q9BUB5       MKNK1    5683057   \n",
       "5  P78368     CSNK1G2     176974   \n",
       "6  O75747     PIK3C2G    1483249   \n",
       "7  Q99640      PKMYT1      69231   \n",
       "8  Q13188        STK3     390471   \n",
       "9  Q9HC98        NEK6    2555396   \n",
       "\n",
       "                                        Pathway_Name               Score  \n",
       "0        Negative regulation of the PI3K/AKT network  0.9999999993935897  \n",
       "1                Toll Like Receptor 5 (TLR5) Cascade  0.9999999990804023  \n",
       "2                       RHO GTPases Activate Formins  0.9999999976453935  \n",
       "3                             RAF/MAP kinase cascade  0.9999999971097913  \n",
       "4                     MAPK family signaling cascades  0.9999999961227581  \n",
       "5                                   Unwinding of DNA  0.9999999936017137  \n",
       "6                      Inositol phosphate metabolism  0.9999999928222838  \n",
       "7                   Cyclin D associated events in G1  0.9999999922670001  \n",
       "8  Association of TriC/CCT with target proteins d...  0.9999999922031578  \n",
       "9                     Mitotic Metaphase and Anaphase  0.9999999896012494  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list_prediction_name = list()\n",
    "errs = []\n",
    "for i in df_raw_prediction.itertuples():\n",
    "\n",
    "    s_name = uniprot_to_gene[i[1]]\n",
    "    path_code = i[2]\n",
    "    try:\n",
    "        list_pa_ids = pathway_code_to_id[path_code]\n",
    "    except KeyError as e:\n",
    "        errs.append(path_code)\n",
    "        continue\n",
    "        sd\n",
    "        \n",
    "    if len(list_pa_ids) == 1:\n",
    "        t_name = pathway_id_to_name[list_pa_ids[0]]\n",
    "    else:\n",
    "        hum = [i for i in list_pa_ids if 'R-HSA' in i]\n",
    "        if len(hum) > 0 :\n",
    "            t_name = pathway_id_to_name[hum[0]]\n",
    "        else:\n",
    "            t_name = pathway_id_to_name[list_pa_ids[0]]\n",
    "\n",
    "    list_prediction_name.append([i[1], s_name, i[2], t_name, i[3]])\n",
    "\n",
    "    \n",
    "df_prediction_name = pd.DataFrame(list_prediction_name, columns=final_prediction_names)\n",
    "df_prediction_name.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1441, 5)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_prediction_name[df_prediction_name.ProteinName == 'STK40'].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0, 0)"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(errs), len(set(errs)) # pathway_code_to_id errors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'data/results/df_prediction_name_GP2VEC_emb_walk_file_40_40_0_0_bwV4_s256_w7_n5_mc5.csv'"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction_file_path + 'df_prediction_name_{}.csv'.format(list_embedding_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_prediction_name.to_csv(prediction_file_path + 'df_prediction_name_{}.csv'.format(list_embedding_file), index = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}