main.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package punkt to /Users/valerio/nltk_data...\n",
      "[nltk_data]   Package punkt is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.svm import SVC\n",
    "from scipy.stats import f_oneway\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.impute import SimpleImputer\n",
    "from sklearn.model_selection import cross_validate\n",
    "from sklearn.feature_selection import SelectKBest\n",
    "from sklearn.feature_selection import chi2\n",
    "from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer\n",
    "from sklearn.decomposition import PCA\n",
    "from gensim.models import Word2Vec\n",
    "import nltk\n",
    "nltk.download('punkt')\n",
    "from nltk.tokenize import word_tokenize  # You may need to install nltk and download punkt\n",
    "from sklearn.metrics import make_scorer, accuracy_score, precision_score, recall_score, f1_score\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import accuracy_score\n",
    "from scipy.stats import ttest_rel\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import f1_score\n",
    "from sklearn.metrics import precision_recall_fscore_support\n",
    "from utilities import *\n",
    "\n",
    "import torch\n",
    "from transformers import BertModel\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from transformers import BertModel, BertTokenizer\n",
    "from transformers import BertTokenizer, BertForSequenceClassification, AdamW"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Upload the DataFrame\n",
    "df = pd.read_csv('data/modelready_220423.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cleaning dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tot columns = 772, numeric type columns = 765\n"
     ]
    },
    {
     "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>publication_number</th>\n",
       "      <th>company_name</th>\n",
       "      <th>countries_in_family</th>\n",
       "      <th>publn_nr</th>\n",
       "      <th>primary_cpc</th>\n",
       "      <th>abstract</th>\n",
       "      <th>description_text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>US-8623043-B1</td>\n",
       "      <td>Entellus Medical, Inc.</td>\n",
       "      <td>['AU' 'EP' 'CA' 'US']</td>\n",
       "      <td>8623043</td>\n",
       "      <td>A61M29/02</td>\n",
       "      <td>A method of treating a constricted sinus passa...</td>\n",
       "      <td>RELATED APPLICATIONS \\n     This Application i...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  publication_number            company_name    countries_in_family publn_nr  \\\n",
       "0      US-8623043-B1  Entellus Medical, Inc.  ['AU' 'EP' 'CA' 'US']  8623043   \n",
       "\n",
       "  primary_cpc                                           abstract  \\\n",
       "0   A61M29/02  A method of treating a constricted sinus passa...   \n",
       "\n",
       "                                    description_text  \n",
       "0  RELATED APPLICATIONS \\n     This Application i...  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Print non-numerical columns\n",
    "numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']\n",
    "print(f'tot columns = {len(df.columns)}, numeric type columns = {len(df.select_dtypes(include=numerics).columns)}' ) # not too many non-numeric columns\n",
    "df.select_dtypes(include = ['object']).head(1)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Extract unique countries in the df\n",
    "unique_values = set()\n",
    "df['countries_in_family'].apply(lambda x: unique_values.update(x.strip(\"[]\").replace(\"'\", \"\").split())) \n",
    "\n",
    "# Create new columns for each unique value (one-hot encoding)\n",
    "for value in unique_values:\n",
    "    # each country has a column (1 if the patent belong to the country 0 otherwise)\n",
    "    df[value] = df['countries_in_family'].apply(lambda x: 1 if value in x else 0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "missing value in description text 0\n"
     ]
    }
   ],
   "source": [
    "df = df[df.abstract.notna()].copy() # drop all samples without abstract (no way to impute it)\n",
    "print('missing value in description text' , df.description_text.isna().sum()) # description_text doesn't have mssing vales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Encode company names\n",
    "df['company_name_encoded'] = df.company_name.astype('category').cat.codes  # encode companies\n",
    "\n",
    "# Remove non-numeric columns\n",
    "df_columns_dropped = df.drop(['publication_number', 'company_name', 'countries_in_family', 'publn_nr','primary_cpc'], axis = 1)\n",
    "\n",
    "# f0_ has the same value as commercialization, the other two shouldn't be used\n",
    "df_columns_dropped = df_columns_dropped.drop(['f0_', 'centrality', 'similarity'], axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remove text as we can't compute min and max on it\n",
    "text = df_columns_dropped[['abstract', 'description_text']] # putting them aside for later\n",
    "df_columns_dropped.drop(['abstract', 'description_text'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Extracting what we'll try to predict\n",
    "y = df_columns_dropped['commercialized']\n",
    "df_columns_dropped.drop('commercialized', axis= 1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Dropping columns where all the value are the same (only one unique value) they would be zero if I apply min max rescaling\n",
    "nunique = df_columns_dropped.nunique()\n",
    "cols_to_drop = nunique[nunique == 1].index\n",
    "df_clean = df_columns_dropped.drop(cols_to_drop, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "for feature_name  in df_clean.columns: \n",
    "    feature_name.replace('v1', 'ge')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# renaming google embedings for clarity\n",
    "df_clean.columns = [s.replace('v1', 'ge') for s in df_clean.columns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Splitting train and test and trying best preprocessing on training set\n",
    "# All preprocessing will be done on X_train and only in the end tested on X_test\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean, y, test_size=0.20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Select best way to fill missin values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "method:mean\n",
      "Average F1 Score: 0.8757077252216657\n",
      "Average Accuracy: 0.8922643745207782\n",
      "Average Precision: 0.8886133260364527\n",
      "Average Recall: 0.8631878236910417\n",
      "\n",
      "method:median\n",
      "Average F1 Score: 0.8758913138080681\n",
      "Average Accuracy: 0.8924275283730971\n",
      "Average Precision: 0.8888289418314926\n",
      "Average Recall: 0.8633467498937726\n",
      "\n",
      "method:most_frequent\n",
      "Average F1 Score: 0.8759989506682043\n",
      "Average Accuracy: 0.8925673993801008\n",
      "Average Precision: 0.8893386178694819\n",
      "Average Recall: 0.8630815260627441\n",
      "\n",
      "method:zero\n",
      "Average F1 Score: 0.8760295174310129\n",
      "Average Accuracy: 0.8925907175559855\n",
      "Average Precision: 0.8893443637162098\n",
      "Average Recall: 0.863134576453163\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# expected time to run: 3 minutes\n",
    "\n",
    "accuracies = []\n",
    "\n",
    "#rescale \n",
    "scaler = ('scaler', StandardScaler())\n",
    "filling_methods = ['mean', 'median', 'most_frequent', 'zero']\n",
    "\n",
    "for method in filling_methods:\n",
    "    # Make a copy of the original data to avoid modifying it\n",
    "    data_filled = X_train.copy()\n",
    "\n",
    "    # Fill NaN values based on the selected method\n",
    "    if method == 'zero':\n",
    "        # data_filled = data_filled.fillna(0)  # You can choose any constant value\n",
    "        imputer =  ('imputer', SimpleImputer(strategy='constant', fill_value=0))\n",
    "    else:\n",
    "        imputer = ('imputer', SimpleImputer(strategy=method, keep_empty_features=False))\n",
    "    \n",
    "    # Default max_iter was reached, so increased it\n",
    "    model = ('model', LogisticRegression(max_iter=1000))\n",
    "\n",
    "    pipeline = Pipeline([imputer ,scaler, model])\n",
    "\n",
    "    # Perform cross-validation and print the mean score\n",
    "    scores_dict = cross_validate(estimator=pipeline, X=data_filled, y=y_train, n_jobs=5, cv=5,scoring=['f1', 'accuracy', 'precision', 'recall'])\n",
    "\n",
    "    accuracies.append(scores_dict[\"test_accuracy\"])\n",
    "\n",
    "    # Print the results\n",
    "    print(f'method:{method}')\n",
    "    print(f'Average F1 Score: {np.mean(scores_dict[\"test_f1\"])}')\n",
    "    print(f'Average Accuracy: {np.mean(scores_dict[\"test_accuracy\"])}')\n",
    "    print(f'Average Precision: {np.mean(scores_dict[\"test_precision\"])}')\n",
    "    print(f'Average Recall: {np.mean(scores_dict[\"test_recall\"])}')\n",
    "    print() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F-statistic: 0.01288102659911202\n",
      "P-value: 0.9979154702110262\n",
      "There is no significant difference in means between the accuracy vectors.\n"
     ]
    }
   ],
   "source": [
    "# Perform ANOVA test to see if there is significant difference between the accuracies\n",
    "\n",
    "f_statistic, p_value = f_oneway(*accuracies)\n",
    "\n",
    "# Print the results\n",
    "print(f'F-statistic: {f_statistic}')\n",
    "print(f'P-value: {p_value}')\n",
    "\n",
    "# Check if the p-value is below the significance level (e.g., 0.05)\n",
    "if p_value < 0.05:\n",
    "    print('The means of the accuracy vectors are significantly different.')\n",
    "else:\n",
    "    print('There is no significant difference in means between the accuracy vectors.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create new train test split with missing values filled (now that we know best method to fill)\n",
    "df_clean.fillna(0, inplace=True) \n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean, y, test_size=0.20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So, it doesn't matter which filling method we will use. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Vectorization of text (bag of words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Putting text back in the DataFrame df_clean\n",
    "df_clean_text = df_clean.copy()\n",
    "df_clean_text[['abstract', 'description_text']] = text  \n",
    "\n",
    "# Split again\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean_text, y, test_size=0.20, random_state=42)\n",
    "\n",
    "X_no_text = X_train.drop(['abstract', 'description_text'],  axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Expected runtime: 30s \n",
    "\n",
    "# Tokenize the text\n",
    "tokenized_text = [word_tokenize(text) for text in X_train['abstract']]\n",
    "\n",
    "# Train Word2Vec model\n",
    "word2vec_model = Word2Vec(sentences=tokenized_text, vector_size=300, window=5, min_count=1, workers=8)\n",
    "\n",
    "# Encode abstracts using the trained Word2Vec model\n",
    "encoded_abstracts = []\n",
    "for words in tokenized_text:\n",
    "    # Take the mean of word vectors for each abstract\n",
    "    encoded_abstract = np.mean([word2vec_model.wv[word] for word in words if word in word2vec_model.wv] or [np.zeros(100)], axis=0)\n",
    "    encoded_abstracts.append(encoded_abstract)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model:Count Vectorizer\n",
      "Average F1 Score: 0.87398370515502\n",
      "Average Accuracy: 0.8907022288538655\n",
      "Average Precision: 0.8869043934111215\n",
      "Average Recall: 0.8614349300551081\n",
      "\n",
      "model:Word2Vec\n",
      "Average F1 Score: 0.8737803111240663\n",
      "Average Accuracy: 0.8905390282570176\n",
      "Average Precision: 0.8868216611039244\n",
      "Average Recall: 0.8611169987282747\n",
      "\n",
      "model:TF-IDF Vectorizer\n",
      "Average F1 Score: 0.8734059120159636\n",
      "Average Accuracy: 0.8901193695794087\n",
      "Average Precision: 0.885658668031578\n",
      "Average Recall: 0.8614879186095803\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Expected runtime: 3 minutes\n",
    "\n",
    "# Create pipeline, logist regression is used as it is our baseline model\n",
    "pipeline = Pipeline([\n",
    "    ('imputer', SimpleImputer(strategy='constant', fill_value=0)),  # Fill missing values with zero\n",
    "    ('scaler', StandardScaler(with_mean=False)), \n",
    "    ('random_forest', LogisticRegression(random_state=42, max_iter=10000))\n",
    "    ])\n",
    "\n",
    "# Define vectorizers\n",
    "vectorizers = {\n",
    "    'Count Vectorizer': CountVectorizer(max_features = 300),\n",
    "    'Word2Vec' : None,\n",
    "    'TF-IDF Vectorizer': TfidfVectorizer(max_features = 300),\n",
    "}\n",
    "\n",
    "X_train_only_text = X_train[['abstract']]\n",
    "\n",
    "# Tokenize text for Word2Vec\n",
    "tokenized_text = [word_tokenize(text) for text in X_train['abstract']]\n",
    "\n",
    "# Train Word2Vec model\n",
    "word2vec_model = Word2Vec(sentences=tokenized_text, vector_size=100, window=5, min_count=1, workers=4)\n",
    "accuracies = []\n",
    "\n",
    "# Transform text data using different vectorization methods and evaluate models\n",
    "for vectorizer_name, vectorizer in vectorizers.items():\n",
    "    if 'Word2Vec' in vectorizer_name:\n",
    "        # Use Word Embeddings (Word2Vec) for Word Embeddings model\n",
    "        X_train_transformed = np.array([np.mean([word2vec_model.wv[word] for word in words if word in word2vec_model.wv] or [np.zeros(100)], axis=0) for words in tokenized_text])\n",
    "        X_train_transformed = pd.DataFrame(X_train_transformed).reset_index(drop=True)\n",
    "    else:\n",
    "        # Use Count Vectorization or TF-IDF Vectorization for other models\n",
    "        X_train_transformed = vectorizer.fit_transform(X_train_only_text['abstract'].values)\n",
    "        X_train_transformed = pd.DataFrame(X_train_transformed.toarray()).reset_index(drop=True)\n",
    "\n",
    "    \n",
    "    X_train_transformed = pd.concat([X_no_text.reset_index(drop=True), X_train_transformed], axis=1)\n",
    "    X_train_transformed.columns = X_train_transformed.columns.astype(str) \n",
    "\n",
    "    # Evaluate the model using cross-validation\n",
    "    scores_dict = cross_validate(pipeline, X_train_transformed, y_train, cv=2, n_jobs=5, scoring=['f1', 'accuracy', 'precision', 'recall'])\n",
    "\n",
    "    # Print the mean accuracy for each combination\n",
    "    accuracies.append(scores_dict[\"test_accuracy\"])\n",
    "\n",
    "    # Print the results\n",
    "    print(f'model:{vectorizer_name}')\n",
    "    print(f'Average F1 Score: {np.mean(scores_dict[\"test_f1\"])}')\n",
    "    print(f'Average Accuracy: {np.mean(scores_dict[\"test_accuracy\"])}')\n",
    "    print(f'Average Precision: {np.mean(scores_dict[\"test_precision\"])}')\n",
    "    print(f'Average Recall: {np.mean(scores_dict[\"test_recall\"])}')\n",
    "    print() \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section we have considered a dataset made of the features and the words embedings (created with different techniques). CountVectorizer is the one that leads to the best results, but these (even if close) don't improve the baseline model. Thus in the following section we won't consider the word embeddings. <br>\n",
    "Due to computational limitation we could only compute the word embeding on the abstract (as it was shorter) and not on description_text (text description of images). In any case the abstract is probably the one with more relevant information. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "del X_train_transformed, encoded_abstract, word2vec_model, X_train_only_text\n",
    "del X_train, X_test, y_train, y_test, df_clean_text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model selection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "IMPORTANT: while testing on PC with low RAM availability we found out that the python Kernel crushed because it didn't have enough memory. If that's the case for the following cells restart the notebook and only run \"Cleaning and baseline models\" section, then start from here. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create new train test split with missing values filled (now that we know best method to fill)\n",
    "df_clean.fillna(0, inplace=True) \n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean, y, test_size=0.20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model:Random Forest\n",
      "Average F1 Score: 0.9102131674025129\n",
      "Average Accuracy: 0.9215237797425984\n",
      "Average Precision: 0.9158357069219314\n",
      "Average Recall: 0.9046659597030754\n",
      "\n",
      "model:Logistic Regression\n",
      "Average F1 Score: 0.8770147946941019\n",
      "Average Accuracy: 0.8937562485849089\n",
      "Average Precision: 0.8929966775249427\n",
      "Average Recall: 0.861611876988335\n",
      "\n",
      "model:Neural Network\n",
      "Average F1 Score: 0.8934101015357596\n",
      "Average Accuracy: 0.9052969247647209\n",
      "Average Precision: 0.8846392235804516\n",
      "Average Recall: 0.902492046659597\n",
      "\n",
      "model:Gradient Boosted Trees\n",
      "Average F1 Score: 0.8836374543626777\n",
      "Average Accuracy: 0.8999113637542763\n",
      "Average Precision: 0.9038458762535324\n",
      "Average Recall: 0.8643160127253446\n",
      "\n",
      "model:SVM\n",
      "Average F1 Score: 0.8808017062401664\n",
      "Average Accuracy: 0.8979295661180491\n",
      "Average Precision: 0.9050530928948817\n",
      "Average Recall: 0.857847295864263\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Expected execution time: 22 minutes\n",
    "\n",
    "# Model selection\n",
    "accuracies = []\n",
    "\n",
    "pipe_step = [('imputer', SimpleImputer(strategy='constant', fill_value=0)),  # Fill missing values with zero\n",
    "    ('scaler', MinMaxScaler())]\n",
    "\n",
    "models = {\n",
    "    'Random Forest': RandomForestClassifier(random_state=42),\n",
    "    'Logistic Regression': LogisticRegression(random_state=42, max_iter=1000),\n",
    "    'Neural Network': MLPClassifier(random_state=42, max_iter=1000),\n",
    "    'Gradient Boosted Trees': GradientBoostingClassifier(random_state=42),\n",
    "    'SVM': SVC(random_state=42)\n",
    "}\n",
    "\n",
    "# Perform model selection using cross-validation\n",
    "for model_name, model in models.items():\n",
    "    \n",
    "\n",
    "    pipeline = Pipeline(pipe_step + [(model_name, model)])\n",
    "    # Print the cross-validation scores\n",
    "    scores_dict = cross_validate(estimator=pipeline, X=X_train, y=y_train, cv=5, n_jobs=5, scoring=['f1', 'accuracy', 'precision', 'recall'])\n",
    "\n",
    "    accuracies.append(scores_dict[\"test_accuracy\"])\n",
    "\n",
    "    # Print the results\n",
    "    print(f'model:{model_name}')\n",
    "    print(f'Average F1 Score: {np.mean(scores_dict[\"test_f1\"])}')\n",
    "    print(f'Average Accuracy: {np.mean(scores_dict[\"test_accuracy\"])}')\n",
    "    print(f'Average Precision: {np.mean(scores_dict[\"test_precision\"])}')\n",
    "    print(f'Average Recall: {np.mean(scores_dict[\"test_recall\"])}')\n",
    "    print() \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F-statistic: 48.74186502883015\n",
      "P-value: 4.893122317602524e-10\n",
      "The means of the accuracy vectors are significantly different.\n"
     ]
    }
   ],
   "source": [
    "# Perform ANOVA test to see if there is significant difference between the accuracies\n",
    "f_statistic, p_value = f_oneway(*accuracies)\n",
    "\n",
    "# Print the results\n",
    "print(f'F-statistic: {f_statistic}')\n",
    "print(f'P-value: {p_value}')\n",
    "\n",
    "# Check if the p-value is below the significance level (e.g., 0.05)\n",
    "if p_value < 0.05:\n",
    "    print('The means of the accuracy vectors are significantly different.')\n",
    "else:\n",
    "    print('There is no significant difference in means between the accuracy vectors.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The difference in accuracies between Random Forest and Logistic Regression is statistically significant.\n",
      "The difference in accuracies between Random Forest and Neural Network is statistically significant.\n",
      "The difference in accuracies between Random Forest and Gradient Boosted Trees is statistically significant.\n",
      "The difference in accuracies between Random Forest and SVM is statistically significant.\n",
      "The difference in accuracies between Logistic Regression and Neural Network is statistically significant.\n",
      "The difference in accuracies between Logistic Regression and Gradient Boosted Trees is statistically significant.\n",
      "The difference in accuracies between Logistic Regression and SVM is statistically significant.\n",
      "The difference in accuracies between Neural Network and Gradient Boosted Trees is statistically significant.\n",
      "The difference in accuracies between Neural Network and SVM is statistically significant.\n",
      "There is no statistically significant difference in accuracies between Pair SVM and Gradient Boosted Trees.\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "# Perform paired t-test for each pair of lists\n",
    "num_lists = len(accuracies)\n",
    "model_names = list(models.keys())\n",
    "\n",
    "for i in range(num_lists ):\n",
    "    for j in range(i + 1, num_lists):\n",
    "        acc_list_1 = accuracies[i]\n",
    "        acc_list_2 = accuracies[j]\n",
    "\n",
    "        # Perform paired t-test\n",
    "        t_statistic, p_value = ttest_rel(acc_list_1, acc_list_2)\n",
    "\n",
    "        # Display the results\n",
    "        # print(f\"T-statistic for Pair {i+1}-{j+1}: {t_statistic}\")\n",
    "        # print(f\"P-value for Pair {i+1}-{j+1}: {p_value}\")\n",
    "\n",
    "        # Check for statistical significance (common significance level is 0.05)\n",
    "        if p_value < 0.05:\n",
    "            print(f\"The difference in accuracies between {model_names[i]} and {model_names[j]} is statistically significant.\")\n",
    "        else:\n",
    "            print(f\"There is no statistically significant difference in accuracies between Pair {model_names[i+1]} and {model_names[j-1]}.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We performed cross validation on the different models in order to determine which one was the best fit for our data. We found that the difference in accuracies between random forest and all the other models are statistically significant, thus we can conclude that Random forest is the best model among the considered ones. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Searching for most relevant features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "accuracies_features = []  # store here the accuracies of different subset of features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average F1 Score: 0.9094141204643164\n",
      "Average Accuracy: 0.9208243241513442\n",
      "Average Precision: 0.9163468131766251\n",
      "Average Recall: 0.9026159392487088\n"
     ]
    }
   ],
   "source": [
    "# Baseline with all features\n",
    "accuracies_features.append(train_RF(X_train, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Case 1: No google embeddings\n",
      "Average F1 Score: 0.9045972224326437\n",
      "Average Accuracy: 0.9172339664507978\n",
      "Average Precision: 0.9183869918987828\n",
      "Average Recall: 0.8912307515829531\n",
      "\n",
      "Case 2: No future information\n",
      "Average F1 Score: 0.9098492382364629\n",
      "Average Accuracy: 0.9210807969087165\n",
      "Average Precision: 0.9152514866910568\n",
      "Average Recall: 0.904522522583164\n",
      "\n",
      "Case 3: No vpm_patent_score\n",
      "Average F1 Score: 0.8748255767105299\n",
      "Average Accuracy: 0.8913316583563278\n",
      "Average Precision: 0.8875098202992175\n",
      "Average Recall: 0.8625287898879479\n",
      "\n",
      "Case 4: No future information and vpm\n",
      "Average F1 Score: 0.8746600163281961\n",
      "Average Accuracy: 0.8909819590969301\n",
      "Average Precision: 0.8857166862496431\n",
      "Average Recall: 0.8639059646253191\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Expected execution time: 2 minutes\n",
    "\n",
    "# Case 1 - no google embeddings\n",
    "google_embeddings = ['ge_0', 'ge_1', 'ge_2', 'ge_3', 'ge_4', 'ge_5', 'ge_6', 'ge_7', 'ge_8', 'ge_9', 'ge_10', 'ge_11', 'ge_12', 'ge_13', 'ge_14', 'ge_15', 'ge_16', 'ge_17', 'ge_18', 'ge_19', 'ge_20', 'ge_21', 'ge_22', 'ge_23', 'ge_24', 'ge_25', 'ge_26', 'ge_27', 'ge_28', 'ge_29', 'ge_30', 'ge_31', 'ge_32', 'ge_33', 'ge_34', 'ge_35', 'ge_36', 'ge_37', 'ge_38', 'ge_39', 'ge_40', 'ge_41', 'ge_42', 'ge_43', 'ge_44', 'ge_45', 'ge_46', 'ge_47', 'ge_48', 'ge_49', 'ge_50', 'ge_51', 'ge_52', 'ge_53', 'ge_54', 'ge_55', 'ge_56', 'ge_57', 'ge_58', 'ge_59', 'ge_60', 'ge_61', 'ge_62', 'ge_63']\n",
    "df_no_google = modify_df(X_train, google_embeddings)\n",
    "print(\"Case 1: No google embeddings\")\n",
    "accuracies_features.append(train_RF(df_no_google, y_train))\n",
    "print()\n",
    "\n",
    "# Case 2 - no future info\n",
    "future_info = ['forward_citations','litigated', 'renewed_after_first_term', 'renewed_after_second_term', 'renewed_after_third_term', 'opposed', 'fwd_cites_same_company']\n",
    "df_no_future = modify_df(X_train, future_info)\n",
    "print(\"Case 2: No future information\")\n",
    "accuracies_features.append(train_RF(df_no_future, y_train))\n",
    "print()\n",
    "\n",
    "# Case 3 - no vpm_patent_score\n",
    "vpm_patent_score = ['vpm_patent_score']\n",
    "df_no_vpm = modify_df(X_train, vpm_patent_score)\n",
    "print(\"Case 3: No vpm_patent_score\")                                                                                                                                                                                \n",
    "accuracies_features.append(train_RF(df_no_vpm, y_train))\n",
    "print()\n",
    "\n",
    "# Case 4 - no future info\n",
    "future_info_vpm = ['vpm_patent_score', 'forward_citations','litigated', 'renewed_after_first_term', 'renewed_after_second_term', 'renewed_after_third_term', 'opposed', 'fwd_cites_same_company']\n",
    "df_no_future_vpm = modify_df(X_train, future_info_vpm)\n",
    "print(\"Case 4: No future information and vpm\")\n",
    "accuracies_features.append(train_RF(df_no_future_vpm, y_train))\n",
    "print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average F1 Score: 0.9125050736238183\n",
      "Average Accuracy: 0.9233888995318456\n",
      "Average Precision: 0.9165187441988818\n",
      "Average Recall: 0.9085365853658537\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([0.92609861, 0.92400047, 0.92527396, 0.91862905, 0.92294241])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "one_hot_encoded = pd.get_dummies(X_train['company_name_encoded'], prefix='company')\n",
    "\n",
    "X_train_company_encoded = pd.concat([X_train, one_hot_encoded], axis=1)\n",
    "\n",
    "train_RF(X_train_company_encoded, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Case 5: No company name\n",
      "Average F1 Score: 0.9099735166428109\n",
      "Average Accuracy: 0.9213838842759697\n",
      "Average Precision: 0.9176857095247645\n",
      "Average Recall: 0.9024040748419291\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# No company name for ethics section\n",
    "no_company = ['company_name_encoded']\n",
    "df_no_company_name = modify_df(X_train, no_company)\n",
    "print(\"Case 5: No company name\")\n",
    "train_RF(df_no_company_name, y_train)\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see there is a deacrese in accuracy when we tried to act ethically and remove the company name. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>PCA()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">PCA</label><div class=\"sk-toggleable__content\"><pre>PCA()</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "PCA()"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Case 5 - CPA, taking no future info as baseline as it achieved max accuracy\n",
    "data = df_no_future.copy()\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaled_data = scaler.fit_transform(data)\n",
    "\n",
    "pca = PCA()\n",
    "pca.fit(scaled_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of components selected: 614\n"
     ]
    }
   ],
   "source": [
    "desired_variance = 0.95\n",
    "n_components = np.argmax(np.cumsum(pca.explained_variance_ratio_) >= desired_variance) + 1\n",
    "print(\"Number of components selected:\", n_components)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average F1 Score: 0.8540320714033948\n",
      "Average Accuracy: 0.8753847709645551\n",
      "Average Precision: 0.8817271686750411\n",
      "Average Recall: 0.8280554443038373\n"
     ]
    }
   ],
   "source": [
    "# Expected execution time: 2 minutes\n",
    "\n",
    "pca = PCA(n_components=627)\n",
    "principal_components = pca.fit_transform(scaled_data)\n",
    "\n",
    "# Create a DataFrame with the principal components\n",
    "pc_df = pd.DataFrame(data=principal_components)\n",
    "\n",
    "accuracies_features.append(train_RF(pc_df, y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Explained Variance Ratio:\n",
      "[0.03363899 0.01361137]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkkAAAHHCAYAAACr0swBAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB2iUlEQVR4nO3deVwU9f8H8NdyH8IiIJeKkBeSF+KFR6ahoqamqWne+dUyLU0zs/LO0CxLrdSsX1Zq59cjzTAS7/CWlPBAxRskRUBEzp3fH3x3Y2EXZpZZ9uD1fDx4PGBmduY9yzFvPsf7oxAEQQARERERabExdQBERERE5ohJEhEREZEOTJKIiIiIdGCSRERERKQDkyQiIiIiHZgkEREREenAJImIiIhIByZJRERERDowSSIiIiLSgUkSEZGJPPnkk3jyySdNHQYR6cEkiYiqZMOGDVAoFJoPJycnNGnSBFOnTsWdO3fKHX/nzh28/vrrCAkJgYuLC1xdXREeHo53330XmZmZOq/Rvn17KBQKrFmzRnRcV69e1YrLxsYGnp6e6NOnD+Lj4w29XaO6ffs2FixYgISEBFOHQkQA7EwdABFZh0WLFiE4OBh5eXk4dOgQ1qxZg127diExMREuLi4AgOPHj6Nv377IycnBqFGjEB4eDgA4ceIEli5digMHDuD333/XOm9ycjKOHz+OoKAgbNq0CZMnT5YU14gRI9C3b18UFxfj4sWL+Oyzz9C9e3ccP34cLVq0kOfmZXL79m0sXLgQQUFBaN26tanDIarxmCQRkSz69OmDtm3bAgD+85//wMvLCytWrMD27dsxYsQIZGZmYtCgQbC1tcXp06cREhKi9folS5Zg/fr15c67ceNG+Pj44MMPP8SQIUNw9epVBAUFiY6rTZs2GDVqlObrrl27ok+fPlizZg0+++wzw26WiGoEdrcRkVH06NEDAJCSkgIAWLduHW7duoUVK1aUS5AAwNfXF++880657Zs3b8aQIUPw9NNPQ6lUYvPmzVWKq2vXrgCAy5cva23PzMzE9OnTUb9+fTg6OqJRo0ZYtmwZVCqV1nHff/89wsPD4ebmBnd3d7Ro0QIrV67U7F+wYAEUCkW566q7Ja9evaozrn379qFdu3YAgPHjx2u6CTds2FCFuyWiqmCSRERGoU5CvLy8AAC//PILnJ2dMWTIENHnOHr0KC5duoQRI0bAwcEBgwcPxqZNm6oUlzpJqV27tmZbbm4uunXrho0bN2LMmDFYtWoVOnfujDlz5mDGjBma42JjYzFixAjUrl0by5Ytw9KlS/Hkk0/i8OHDVYoJAJo1a4ZFixYBACZNmoRvv/0W3377LZ544okqn5uIDMPuNiKSRVZWFu7evYu8vDwcPnwYixYtgrOzM55++mkAwLlz59CkSRM4ODiIPufGjRtRv359dO7cGQAwfPhw/N///R8SEhJEj9nJzc3F3bt3UVxcjOTkZE3SUzpZW7FiBS5fvozTp0+jcePGAIAXX3wRAQEBWL58OWbOnIn69evj119/hbu7O3bv3g1bW1vR9yGGr68v+vTpg3nz5iEiIkKri5CITIMtSUQki8jISNSpUwf169fH8OHDUatWLWzduhV169YFAGRnZ8PNzU30+YqKivDDDz/gueee03Rf9ejRAz4+PpJak+bPn486derAz88PXbt2xblz5zTjm9R++ukndO3aFbVr18bdu3c1H5GRkSguLsaBAwcAAB4eHnj48CFiY2NFX5+ILBdbkohIFp9++imaNGkCOzs7+Pr6omnTprCx+ff/MHd3dzx48ED0+X7//Xf8888/aN++PS5duqTZ3r17d3z33XdYtmyZ1vn1mTRpEoYOHYq8vDzExcVh1apVKC4u1jomOTkZZ86cQZ06dXSeIz09HQDw8ssv48cff0SfPn1Qt25d9OrVC8OGDUNUVJTo+yIiy8EkiYhk0b59e83sNl1CQkKQkJCAgoICUV1u6taiYcOG6dy/f/9+dO/evdLzNG7cGJGRkQCAp59+Gra2tnjzzTfRvXt3TbwqlQo9e/bEG2+8ofMcTZo0AQD4+PggISEBu3fvxm+//YbffvsNX331FcaMGYOvv/4aAHQO2gZQLjEjIvPHJImIqkX//v0RHx+P//73vxgxYkSFxz58+BDbt2/Hc889p3Og96uvvopNmzaJSpLKevvtt7F+/Xq88847iImJAQA0bNgQOTk5mmSqIg4ODujfvz/69+8PlUqFl19+GevWrcPcuXPRqFEjzYDwzMxMeHh4aF537dq1Ss+tL8EiItPgmCQiqhYvvfQS/P39MXPmTFy8eLHc/vT0dLz77rsAgK1bt+Lhw4eYMmUKhgwZUu7j6aefxn//+1/k5+dLjsPDwwMvvvgidu/eralsPWzYMMTHx2P37t3ljs/MzERRUREA4N69e1r7bGxs0LJlSwDQxNKwYUMA0IxjAkqSPnVLU0VcXV011yQi02NLEhFVi9q1a2Pr1q3o27cvWrdurVVx+9SpU/juu+8QEREBoKSrzcvLC506ddJ5rgEDBmD9+vX49ddfMXjwYMmxTJs2DR9//DGWLl2K77//HrNmzcIvv/yCp59+GuPGjUN4eDgePnyIs2fP4ueff8bVq1fh7e2N//znP8jIyECPHj1Qr149XLt2DatXr0br1q3RrFkzAECvXr0QGBiICRMmYNasWbC1tcX//d//oU6dOrh+/XqFcTVs2BAeHh5Yu3Yt3Nzc4Orqig4dOiA4OFjyPRKRDAQioir46quvBADC8ePHRR1/+/Zt4bXXXhOaNGkiODk5CS4uLkJ4eLiwZMkSISsrS7hz545gZ2cnjB49Wu85cnNzBRcXF2HQoEF6j0lJSREACMuXL9e5f9y4cYKtra1w6dIlQRAE4cGDB8KcOXOERo0aCQ4ODoK3t7fQqVMn4YMPPhAKCgoEQRCEn3/+WejVq5fg4+MjODg4CIGBgcKLL74opKamap375MmTQocOHTTHrFixQvM+paSkaI7r1q2b0K1bN63Xbt++XQgNDRXs7OwEAMJXX31VwbtJRMakEARBMG2aRkRERGR+OCaJiIiISAcmSUREREQ6MEkiIiIi0oFJEhEREZEOTJKIiIiIdGCSRERERKQDi0mWolKpcPv2bbi5uXF5ACIiIgshCAIePHiAgIAAUQtfi8UkqZTbt2+jfv36pg6DiIiIDHDjxg3Uq1dPtvMxSSrFzc0NQMmb7O7ubuJoiIiISIzs7GzUr19f8xyXC5OkUtRdbO7u7kySiIiILIzcQ2U4cJuIiIhIB7NJkg4cOID+/fsjICAACoUC27Zt03vsSy+9BIVCgY8//lhre0ZGBkaOHAl3d3d4eHhgwoQJyMnJMW7gREREZJXMJkl6+PAhWrVqhU8//bTC47Zu3YojR44gICCg3L6RI0fi77//RmxsLHbu3IkDBw5g0qRJxgqZiIiIrJjZjEnq06cP+vTpU+Ext27dwiuvvILdu3ejX79+WvvOnTuHmJgYHD9+HG3btgUArF69Gn379sUHH3ygM6kiIiIi0sdsWpIqo1KpMHr0aMyaNQuPP/54uf3x8fHw8PDQJEgAEBkZCRsbGxw9elTnOfPz85Gdna31QURERARYUJK0bNky2NnZ4dVXX9W5Py0tDT4+Plrb7Ozs4OnpibS0NJ2viY6OhlKp1HywRhIRERGpWUSSdPLkSaxcuRIbNmyQdXrfnDlzkJWVpfm4ceOGbOcmIiIiy2YRSdLBgweRnp6OwMBA2NnZwc7ODteuXcPMmTMRFBQEAPDz80N6errW64qKipCRkQE/Pz+d53V0dNTURGJtJCIiIirNbAZuV2T06NGIjIzU2ta7d2+MHj0a48ePBwBEREQgMzMTJ0+eRHh4OAAgLi4OKpUKHTp0qPaYiYiIyLKZTZKUk5ODS5cuab5OSUlBQkICPD09ERgYCC8vL63j7e3t4efnh6ZNmwIAmjVrhqioKEycOBFr165FYWEhpk6diuHDh3NmGxERURUVqwQcS8lA+oM8+Lg5oX2wJ2xtrHsxeLNJkk6cOIHu3btrvp4xYwYAYOzYsdiwYYOoc2zatAlTp07FU089BRsbGzz77LNYtWqVMcIlIiKqMWISU7FwRxJSs/I02/yVTpjfPxRRzf1FncMSkyyFIAiCqYMwF9nZ2VAqlcjKyuL4JCIiIpQkSJM3nkLZZEGd3qwZ1abSREmOJKsixnp+W8TAbSIiIqp+xSoBC3cklUuQAGi2LdyRhGKV/vYWdZJVOkECgLSsPEzeeAoxianyBSwzJklERESk07GUjHLJTWkCgNSsPBxLydC5X44ky5SYJBEREZFO6Q/0J0hijqtqkmVqTJKIiIhIJx83pyodV9Uky9TMZnYbERER/cscZoO1D/aEv9IJaVl5OrvMFAD8lCWx6VLVJMvUmCQRERGZGWPPBhPL1kaB+f1DMXnjKSgArURJna7N7x+qN3mrapJlauxuIyIiMiPmNhssqrk/1oxqAz+ldmuPn9Kp0un/6iQL+DepUhOTZJka6ySVwjpJRERkSsUqAV2Wxekd7KxueTk0u0e1JxZV6f6z1DpJ7G4jIiIyE1Jmg0U09NJ7nDHY2igMvmZUc3/0DPUz+RgrqZgkERERmQlLnw1WkaokWabCMUlERERmwtJng1kbJklERERmQj0bTF8nlAIlY3nMdTaYtWGSREREZCYsfTaYtWGSREREZEaqMuWe5MWB20RERGbGUmeDWRsmSURERGbIEmeDWRt2txERERHpwCSJiIiISAcmSUREREQ6MEkiIiIi0oFJEhEREZEOTJKIiIiIdGCSRERERKQDkyQiIiIiHZgkEREREenAJImIiIhIByZJRERERDowSSIiIiLSgUkSERERkQ5MkoiIiIh0YJJEREREpIPZJEkHDhxA//79ERAQAIVCgW3btmn2FRYWYvbs2WjRogVcXV0REBCAMWPG4Pbt21rnyMjIwMiRI+Hu7g4PDw9MmDABOTk51XwnREREZA3MJkl6+PAhWrVqhU8//bTcvtzcXJw6dQpz587FqVOnsGXLFly4cAEDBgzQOm7kyJH4+++/ERsbi507d+LAgQOYNGlSdd0CERERWRGFIAiCqYMoS6FQYOvWrXjmmWf0HnP8+HG0b98e165dQ2BgIM6dO4fQ0FAcP34cbdu2BQDExMSgb9++uHnzJgICAiq9bnZ2NpRKJbKysuDu7i7X7RAREZERGev5bTYtSVJlZWVBoVDAw8MDABAfHw8PDw9NggQAkZGRsLGxwdGjR00UJREREVkqO1MHYIi8vDzMnj0bI0aM0GSMaWlp8PHx0TrOzs4Onp6eSEtL03me/Px85Ofna77Ozs42XtBERERkUSyuJamwsBDDhg2DIAhYs2ZNlc4VHR0NpVKp+ahfv75MURIREZGls6gkSZ0gXbt2DbGxsVr9jn5+fkhPT9c6vqioCBkZGfDz89N5vjlz5iArK0vzcePGDaPGT0RERJbDYrrb1AlScnIy9u7dCy8vL639ERERyMzMxMmTJxEeHg4AiIuLg0qlQocOHXSe09HREY6OjkaPnYiIiCyP2SRJOTk5uHTpkubrlJQUJCQkwNPTE/7+/hgyZAhOnTqFnTt3ori4WDPOyNPTEw4ODmjWrBmioqIwceJErF27FoWFhZg6dSqGDx8uamYbERERUWlmUwJg37596N69e7ntY8eOxYIFCxAcHKzzdXv37sWTTz4JoKSY5NSpU7Fjxw7Y2Njg2WefxapVq1CrVi1RMbAEABERkeUx1vPbbJIkc8AkiYiIyPKwThIRERFRNWKSRERERKQDkyQiIiIiHZgkEREREenAJImIiIhIByZJRERERDowSSIiIiLSgUkSERERkQ5MkoiIiIh0YJJEREREpAOTJCIiIiIdmCQRERER6cAkiYiIiEgHJklEREREOjBJIiIiItKBSRIRERGRDkySiIiIiHRgkkRERESkg52pAyAiIstXrBJwLCUD6Q/y4OPmhPbBnrC1UZg6LKIqYZJERERVEpOYioU7kpCalafZ5q90wvz+oYhq7m/CyIiqht1tRERksJjEVEzeeEorQQKAtKw8TN54CjGJqSaKjKjqmCQREZFBilUCFu5IgqBjn3rbwh1JKFbpOoLI/DFJIiIigxxLySjXglSaACA1Kw/HUjKqLygiGTFJIiIig6Q/0J8gGXIckblhkkRERAbxcXOS9Tgic8MkiYiIDNI+2BP+Sifom+ivQMkst/bBntUZFpFsmCQREZFBbG0UmN8/FADKJUrqr+f3D2W9JLJYTJKIiMhgUc39sWZUG/gptbvU/JROWDOqDeskkUVjMUkiIqqSqOb+6Bnqx4rbZHWYJBERUZXZ2igQ0dDL1GEQyUpykpSfn4+jR4/i2rVryM3NRZ06dRAWFobg4GBjxEdERERkEqKTpMOHD2PlypXYsWMHCgsLoVQq4ezsjIyMDOTn5+Oxxx7DpEmT8NJLL8HNzc2YMRMREREZnaiB2wMGDMBzzz2HoKAg/P7773jw4AHu3buHmzdvIjc3F8nJyXjnnXewZ88eNGnSBLGxsZIDOXDgAPr374+AgAAoFAps27ZNa78gCJg3bx78/f3h7OyMyMhIJCcnax2TkZGBkSNHwt3dHR4eHpgwYQJycnIkx0JEREQkKknq168fUlJS8P7776Nr165wdnbW2v/YY49h7NixiImJwZ49e2BjI33S3MOHD9GqVSt8+umnOve///77WLVqFdauXYujR4/C1dUVvXv3Rl7ev5VcR44cib///huxsbHYuXMnDhw4gEmTJkmOhYiIiEghCILZrTyoUCiwdetWPPPMMwBKWpECAgIwc+ZMvP766wCArKws+Pr6YsOGDRg+fDjOnTuH0NBQHD9+HG3btgUAxMTEoG/fvrh58yYCAgIqvW52djaUSiWysrLg7u5utPsjIiIi+Rjr+W0RdZJSUlKQlpaGyMhIzTalUokOHTogPj4eABAfHw8PDw9NggQAkZGRsLGxwdGjR3WeNz8/H9nZ2VofRERERIDEJOmzzz5DZGQkhg0bhj179mjtu3v3Lh577DFZg1NLS0sDAPj6+mpt9/X11exLS0uDj4+P1n47Ozt4enpqjikrOjoaSqVS81G/fn0jRE9ERESWSHSStGrVKsyaNQshISFwdHRE3759ER0drdlfXFyMa9euGSVIY5kzZw6ysrI0Hzdu3DB1SERERGQmRJcAWLduHdavX4/nn38eADB58mQ888wzePToERYtWmS0AAHAz88PAHDnzh34+/9b4v7OnTto3bq15pj09HSt1xUVFSEjI0Pz+rIcHR3h6OhonKCJiIjIooluSUpJSUGnTp00X3fq1AlxcXH4/PPPMWfOHKMEpxYcHAw/Pz+tLr7s7GwcPXoUERERAICIiAhkZmbi5MmTmmPi4uKgUqnQoUMHo8ZHRERE1kd0S5K3tzdu3LiBoKAgzbbmzZsjLi4OPXr0wO3bt6sUSE5ODi5duqT5OiUlBQkJCfD09ERgYCCmT5+Od999F40bN0ZwcDDmzp2LgIAAzQy4Zs2aISoqChMnTsTatWtRWFiIqVOnYvjw4aJmthERERGVJjpJ6tKlC7Zs2YKuXbtqbQ8NDcWePXvQvXv3KgVy4sQJrXPMmDEDADB27Fhs2LABb7zxBh4+fIhJkyYhMzMTXbp0QUxMDJyc/l15etOmTZg6dSqeeuop2NjY4Nlnn8WqVauqFBcRERHVTKLrJJ05cwYnT57E+PHjde5PTEzEf//7X8yfP1/WAKsT6yQRERFZHmM9v82ymKSpMEkiIiKyPCYvJpmcnIwRI0boLLiYlZWF559/HleuXJEtMCIiIiJTEp0kLV++HPXr19eZoakLMS5fvlzW4IiIiIhMRXSStH//fgwdOlTv/mHDhiEuLk6WoIiIiIhMTXSSdP369XLLfpSmLhFAREREZA1EJ0lKpRKXL1/Wu//SpUsc7ExERERWQ3SS9MQTT2D16tV6969atapcDSUiIiIiSyU6SZozZw5+++03DBkyBMeOHdMsCnv06FE8++yz2L17t9GXJyEiIiKqLqIrboeFheHnn3/GCy+8gK1bt2rt8/Lywo8//og2bdrIHiARERGRKYhOkgDg6aefxrVr1xATE4NLly5BEAQ0adIEvXr1gouLi7FiJCIiIqp2kpIkQRBw8+ZNNGnSBP3794ednaSXExEREVkM0WOSUlJS0LJlS4SEhKBly5Zo2LAhTpw4YczYiIiIiExGdJI0a9YsFBUVYePGjfj5559Rr149TJo0yZixERERmY1ilYD4y/ewPeEW4i/fQ7GKS59aO9H9ZYcOHcLPP/+MLl26AAA6duyIevXq4eHDh3B1dTVagERERKYWk5iKhTuSkJqVp9nmr3TC/P6hiGrub8LIyJhEtySlp6ejcePGmq/9/f3h7OyM9PR0owRGRERkDmISUzF54ymtBAkA0rLyMHnjKcQkppooMjI20S1JCoUCOTk5cHZ21myzsbHBgwcPkJ2drdnGqttERKZXrBJwLCUD6Q/y4OPmhPbBnrC1UZg6LItTrBKwcEcSdHWsCQAUABbuSELPUD++v1ZIdJKknu5fdltYWJjmc4VCgeLiYnkjJCIiSdg1JJ9jKRnlWpBKEwCkZuXhWEoGIhp6VV9gVC1EJ0l79+41ZhxERCQDdddQ2ZYPddfQmlFtmChJkP5Af4JkyHFkWUQnSd26dTNmHEREVEXsGpKfj5uTrMeRZRE1cPvhw4eSTir1eCIiqjopXUMkTvtgT/grnaAvpVSgpCuzfbBndYZF1URUktSoUSMsXboUqan6R/ALgoDY2Fj06dMHq1atki1AIiISh11D8rO1UWB+/1AAKJcoqb+e3z+ULXNWSlR32759+/DWW29hwYIFaNWqFdq2bYuAgAA4OTnh/v37SEpKQnx8POzs7DBnzhy8+OKLxo6biIjKYNeQcUQ198eaUW3KDYb342B4q6cQBEF0ydDr16/jp59+wsGDB3Ht2jU8evQI3t7eCAsLQ+/evdGnTx/Y2toaM16jys7OhlKpRFZWFksZEJHFKVYJ6LIsDmlZeTrHJSlQ8mA/NLsHWz4MwLIK5stYz29JSZK1Y5JERJZOPbsNgFaipH6Uc3YbWSNjPb9FV9wmIiJ5GWMtMHXXkJ9Su0vNT+nEBIlIItElAIiISD7GLPgY1dwfPUJ88W38VVzLyEUDTxeMjgiCgx3/LyaSgt1tpbC7jYiqg76Cj3J1iRkrAeOYHDJXxnp+syWJiKgaGbvgo7EqbnOpE6qJ2PZKRFSNxBZ83HA4RfIYpcoSMKAkAZN6XnXiVTZudeIVk6i/hh6RJZOcJKlUKr3br1+/XuWAiIismdhCjot/PYcuy+IkJSDGqLhtrMSLyBKITpKys7MxbNgwuLq6wtfXF/PmzUNxcbFm/z///IPg4GCjBElEZC2kFHKU2lJjjIrbXOqEajLRSdLcuXPx119/4dtvv8WSJUvwzTffYODAgSgoKNAcwzHgREQVU68FJobUlhpjVNzmUidUk4lOkrZt24Z169ZhyJAh+M9//oMTJ07gn3/+Qf/+/ZGfnw8AUCiMN8uhuLgYc+fORXBwMJydndGwYUMsXrxYKzETBAHz5s2Dv78/nJ2dERkZieTkZKPFREQkla2NAgNaiR/oLKWlxhiLsXKpE6rJRCdJ//zzDxo0aKD52tvbG3/88QcePHiAvn37Ijc31ygBqi1btgxr1qzBJ598gnPnzmHZsmV4//33sXr1as0x77//PlatWoW1a9fi6NGjcHV1Re/evZGXx/9wiMg8FKsE/PKX9IHOYlpqjLEYqzESLyJLITpJCgwMxLlz57S2ubm54ffff8ejR48waNAg2YMr7c8//8TAgQPRr18/BAUFYciQIejVqxeOHTsGoKQV6eOPP8Y777yDgQMHomXLlvjmm29w+/ZtbNu2zaixERGJVdkYH33EttTIXXHbGIkXkaUQnST16tULX331VbnttWrVwu7du+HkZNym1k6dOmHPnj24ePEiAOCvv/7CoUOH0KdPHwBASkoK0tLSEBkZqXmNUqlEhw4dEB8fr/Oc+fn5yM7O1vogIjImqWN3DGmpiWruj0Oze+C7iR2xcnhrfDexIw7N7mFwPSMudUI1lehikgsXLsTt27d17nNzc0NsbCxOnTolW2Blvfnmm8jOzkZISAhsbW1RXFyMJUuWYOTIkQCAtLQ0AICvr6/W63x9fTX7yoqOjsbChQuNFjMRWRc5Kk5LGbtTlZYaWxsFIhp6SXpNRaKa+6NnqF+V759Vu8mSiE6Sateujdq1a+vd7+bmhm7duskSlC4//vgjNm3ahM2bN+Pxxx9HQkICpk+fjoCAAIwdO9agc86ZMwczZszQfJ2dnY369evLFTIRWRG5Kk6rx/ikZeXprD1Ump+ZVbSuauLFqt1kaSym4vasWbPw5ptvYvjw4WjRogVGjx6N1157DdHR0QAAPz8/AMCdO3e0Xnfnzh3NvrIcHR3h7u6u9UFEVJacFacrGuOjNqFzUJW7yMwNq3aTJbKYJCk3Nxc2Ntrh2traaiqABwcHw8/PD3v27NHsz87OxtGjRxEREVGtsRKR9TBGxWl9Y3z8lU5YO6oN5vZ/HBENvaymG4pVu8lSWcwCt/3798eSJUsQGBiIxx9/HKdPn8aKFSvwwgsvACip0TR9+nS8++67aNy4MYKDgzF37lwEBATgmWeeMW3wRGSxpFScltIVJdcYH0tgrPeQyNgsJklavXo15s6di5dffhnp6ekICAjAiy++iHnz5mmOeeONN/Dw4UNMmjQJmZmZ6NKlC2JiYow+846IrJcxK07LPbhabnINsmbVbrJUkpMkW1tbpKamwsfHR2v7vXv34OPjo7Wem5zc3Nzw8ccf4+OPP9Z7jEKhwKJFi7Bo0SKjxEBENU9NrTgt5yDrmvoekuWTPCZJ3/ps+fn5cHBwqHJARETmpCZWnJZ7kHVNfA/JOohuSVq1ahWAktaaL774ArVq1dLsKy4uxoEDBxASEiJ/hEREJqSejTZ54ykoAK3Bx9ZYcbqyQdYKlAyy7hnqJ/qea9p7SNZDdJL00UcfAShpSVq7di1sbW01+xwcHBAUFIS1a9fKHyERkYmpZ6OV7X4ytzpGcjDmQPWa8h6S9RCdJKWkpAAAunfvji1btlRYWJKIyNrUlNloxhxkXVPeQ7Iekgdu79271xhxEBGZPXOfjSYHYw+yrgnvIVkPyUlScXExNmzYgD179iA9PV1TzFEtLi5OtuCIiKh6VbZsigIlXWQcZE01geQkadq0adiwYQP69euH5s2bQ6FgMykRkbXgIGuifykEfXP69fD29sY333yDvn37Gismk8nOzoZSqURWVhbXcSOiGo2L0ZIlMdbzW3JLkoODAxo1aiRbAEREZH44yJrIgCRp5syZWLlyJT755BN2tRERGYFcy4FUFQdZU00nOUk6dOgQ9u7di99++w2PP/447O3ttfZv2bJFtuCIiGoaU3ZzmUtyRmQuJCdJHh4eGDRokDFiISKq0dTLgZQdKKpeDmTNqDZGS5Q4BomoPMkDt60ZB24TkakUqwR0WRant9q1eur9odk9ZG/d0Zecqa9izOSMSA7Gen5LXuAWAIqKivDHH39g3bp1ePDgAQDg9u3byMnJkS0wIqKaRMpyIHKqbK02oGSttmIV/5+mmkdyd9u1a9cQFRWF69evIz8/Hz179oSbmxuWLVuG/Px8rt9GRGQAYy4HUhFjrdVGZA0ktyRNmzYNbdu2xf379+Hs7KzZPmjQIOzZs0fW4IiIagpjLweij6mSMyJLILkl6eDBg/jzzz/h4OCgtT0oKAi3bt2SLTAioprEVMuBmCo5I7IEkluSVCoViouLy22/efMm3NzcZAmKiKimUS8HAvw7YFrNmMuBqJMzfWdVoGSWG9dqo5pIcpLUq1cvfPzxx5qvFQoFcnJyMH/+fKtcqoSIqLpENffHmlFt4KfUbrXxUzoZbYaZqZIzIksguQTAzZs30bt3bwiCgOTkZLRt2xbJycnw9vbGgQMH4OPjY6xYjY4lAIjIHJiiqCPrJJElM9bz26A6SUVFRfj+++9x5swZ5OTkoE2bNhg5cqTWQG5LxCSJiKyBoUkWK26TpTKrJMlaMUkiIkvHFiGqiYz1/JY8uw0AkpOTsXfvXqSnp0OlUmntmzdvniyBERGRNFVZ1oStSETlSU6S1q9fj8mTJ8Pb2xt+fn5QKP79JVIoFEySiKhGMnWSUVnlbAVKKmf3DPUrFxdbn4h0k5wkvfvuu1iyZAlmz55tjHiIiCyOOSQZhlbONuWiukTmTnIJgPv372Po0KHGiIWIyOKok4yyCYo6yYhJTK2WOAypnM1124gqJjlJGjp0KH7//XdjxEJEZFHMKckwpHK2qRbVJbIUkrvbGjVqhLlz5+LIkSNo0aIF7O3ttfa/+uqrsgVHRGTOzGlxWEOWNeG6bUQVk5wkff7556hVqxb279+P/fv3a+1TKBRMkoioxjCnJENdOXvyxlNQAFqJkr7K2Vy3jahikpOklJQUY8RBRGRxzC3JUC9rUnYQuZ+eQeSmWlSXyFIYVCdJTV2HsnQZACKimsIck4yo5v7oGeonqhyBIa1PRDWJ5IHbAPDNN9+gRYsWcHZ2hrOzM1q2bIlvv/1W7tiIiMyauS4Oa2ujQERDLwxsXRcRDb0qvL4pFtUlshSSk6QVK1Zg8uTJ6Nu3L3788Uf8+OOPiIqKwksvvYSPPvrIGDFq3Lp1C6NGjYKXlxecnZ3RokULnDhxQrNfEATMmzcP/v7+cHZ2RmRkJJKTk40aExHVbNaQZEQ198eh2T3w3cSOWDm8Nb6b2BGHZvfQir1YJSD+8j1sT7iF+Mv3WBaAagTJa7cFBwdj4cKFGDNmjNb2r7/+GgsWLDDamKX79+8jLCwM3bt3x+TJk1GnTh0kJyejYcOGaNiwIQBg2bJliI6Oxtdff43g4GDMnTsXZ8+eRVJSEpycKh8TwLXbiMhQpq64bUzmUCyTqCJms8Ctk5MTEhMT0ahRI63tycnJaNGiBfLyjDOL480338Thw4dx8OBBnfsFQUBAQABmzpyJ119/HQCQlZUFX19fbNiwAcOHD6/0GkySiMgcmFPCpa8itzoaS2ktI+tmrOe35O62Ro0a4ccffyy3/YcffkDjxo1lCUqXX375BW3btsXQoUPh4+ODsLAwrF+/XrM/JSUFaWlpiIyM1GxTKpXo0KED4uPjdZ4zPz8f2dnZWh9ERKYUk5iKLsviMGL9EUz7PgEj1h9Bl2Vx1Va5uzRzKpZJZAqSZ7ctXLgQzz33HA4cOIDOnTsDAA4fPow9e/boTJ7kcuXKFaxZswYzZszAW2+9hePHj+PVV1+Fg4MDxo4di7S0NACAr6+v1ut8fX01+8qKjo7GwoULjRYzEZEU5raOmjkVyyQyBcktSc8++yyOHj0Kb29vbNu2Ddu2bYO3tzeOHTuGQYMGGSNGAIBKpUKbNm3w3nvvISwsDJMmTcLEiROxdu1ag885Z84cZGVlaT5u3LghY8REROKZY6uNORXLJDIFg+okhYeHY+PGjXLHUiF/f3+EhoZqbWvWrBn++9//AgD8/PwAAHfu3IG//7//ad25cwetW7fWeU5HR0c4OjoaJ2AiIgnMsdXG3IplElU3g5Kk4uJibN26FefOnQMAhIaGYuDAgbCzq1Jtygp17twZFy5c0Np28eJFNGjQAEDJrDs/Pz/s2bNHkxRlZ2fj6NGjmDx5stHiIiKSgzm22phjsUyi6iS5u+3vv/9GkyZNMHbsWGzduhVbt27F2LFj0bhxYyQmJhojRgDAa6+9hiNHjuC9997DpUuXsHnzZnz++eeYMmUKgJKq39OnT8e7776LX375BWfPnsWYMWMQEBCAZ555xmhxERHJwRxbbcy1WCZRdZFcAiAiIgJ16tTB119/jdq1awMoqWE0btw4/PPPP/jzzz+NEigA7Ny5E3PmzEFycjKCg4MxY8YMTJw4UbNfEATMnz8fn3/+OTIzM9GlSxd89tlnaNKkiajzswQAEelj7Gn5xSoBXZbFVdpqc2h2j2pPSlgnicyd2dRJcnZ2xokTJ/D4449rbU9MTES7du3w6NEj2YKrbkySiEiX6koS1LPbAN3rqJmyJpE51W4iKsts6iQ1adIEd+7cKbc9PT29XIFJIiJLp05cyg6qVk/Ll7N+kTkvcSJlPTgiayF5pHV0dDReffVVLFiwAB07dgQAHDlyBIsWLcKyZcu0CjKyNYaILFll0/IVKJmW3zPUT7akIaq5P3qG+rHVhsgMSO5us7H5t/FJoSj5pVWfovTXCoUCxcXFcsVZLdjdRkSlxV++hxHrj1R63HcTO7KYIpEJGev5Lbklae/evbJdnIjInJnjtHwiqj6Sk6Ru3boZIw4isiA1ZRCvOU7LJ6LqY1D1x7y8PJw5cwbp6elQqVRa+wYMGCBLYERknmrSdHAWUySq2SQnSTExMRgzZgzu3r1bbp8ljkMiIvHMbQFWY1MXU5y88RQU0D0tn8UUiayX5BIAr7zyCoYOHYrU1FSoVCqtDyZIRNbLHBdgrQ7mPC2fiIxLckvSnTt3MGPGDPj6+hojHiIyU+a4AGt14bR8oppJcpI0ZMgQ7Nu3Dw0bNjRGPERkpmr6TC91MUUiqjkkJ0mffPIJhg4dioMHD6JFixawt7fX2v/qq6/KFhwRmQ/O9CKimkZykvTdd9/h999/h5OTE/bt26cpIAmUDNxmkkRknTjTi4hqGskDt99++20sXLgQWVlZuHr1KlJSUjQfV65cMUaMRGQG1DO9gH9ndqlxphcRWSPJSVJBQQGee+45reVJiKhm4EwvIqpJJK/d9tprr6FOnTp46623jBWTyXDtNiJxakrFbSKyDGazdltxcTHef/997N69Gy1btiw3cHvFihWyBUdE5okzvYioJpCcJJ09exZhYWEAgMTERK19pQdxExEREVkyyUnS3r17jREHERERkVmp0ujrmzdv4ubNm3LFQkRERGQ2JCdJKpUKixYtglKpRIMGDdCgQQN4eHhg8eLFUKlUxoiRiIiIqNpJ7m57++238eWXX2Lp0qXo3LkzAODQoUNYsGAB8vLysGTJEtmDJLJUnAVGRGS5JJcACAgIwNq1azFgwACt7du3b8fLL7+MW7duyRpgdWIJAJJTTGIqFu5I0loU1l/phPn9Q1lPiIhIRsZ6fkvubsvIyEBISEi57SEhIcjIyJAlKCJLF5OYiskbT2klSACQlpWHyRtPISYx1USRERGRWJKTpFatWuGTTz4pt/2TTz5Bq1atZAmKyJIVqwQs3JGkc30z9baFO5JQrJLUiCuLYpWA+Mv3sD3hFuIv3zNJDERElkLymKT3338f/fr1wx9//IGIiAgAQHx8PG7cuIFdu3bJHiCRpTmWklGuBak0AUBqVh6OpWRUa0FGdv8REUkjuSWpW7duuHjxIgYNGoTMzExkZmZi8ODBuHDhArp27WqMGIksSvoD/QmSIcfJgd1/RETSSW5JAkoGb3MWG5FuPm5OlR8k4biqqqz7T4GS7r+eoX6ceUdEVIrolqTk5GSMGDEC2dnZ5fZlZWXh+eefx5UrV2QNjsgStQ/2hL/SCfrSDQVKurnaB3tWSzxSuv+IiOhfopOk5cuXo379+jqn1imVStSvXx/Lly+XNTgiS2Rro8D8/qEAUC5RUn89v39otbXamGP3HxGRJRCdJO3fvx9Dhw7Vu3/YsGGIi4uTJSgiSxfV3B9rRrWBn1K7S81P6YQ1o9pU60Bpc+v+IyKyFKLHJF2/fh0+Pj5693t7e+PGjRuyBEVkSfRV1Y5q7o+eoX4mr7it7v5Ly8rTOS5JgZLkrbq6/4iILIXoliSlUonLly/r3X/p0qVqrVK9dOlSKBQKTJ8+XbMtLy8PU6ZMgZeXF2rVqoVnn30Wd+7cqbaYqOaJSUxFl2VxGLH+CKZ9n4AR64+gy7I4zWwxWxsFIhp6YWDruoho6GWSgdHm1v1HRGQpRCdJTzzxBFavXq13/6pVq6qtBMDx48exbt06tGzZUmv7a6+9hh07duCnn37C/v37cfv2bQwePLhaYqKax5Km1ZtT9x8RkaUQvXbb6dOnERERgaeffhpvvPEGmjZtCgA4f/483n//ffz666/4888/0aZNG6MGnJOTgzZt2uCzzz7Du+++i9atW+Pjjz9GVlYW6tSpg82bN2PIkCGa2Jo1a4b4+Hh07Nix0nNz7TYSq1gloMuyOL2zxtRdWIdm9zCrFhouuEtE1sjka7eFhYXh559/xoEDBxAREQFPT094enqiU6dOOHjwIH788UejJ0gAMGXKFPTr1w+RkZFa20+ePInCwkKt7SEhIQgMDER8fLzR46KaxVKn1ZtD9x8RkaWQVEzy6aefxrVr1xATE4NLly5BEAQ0adIEvXr1gouLi7Fi1Pj+++9x6tQpHD9+vNy+tLQ0ODg4wMPDQ2u7r68v0tLSdJ4vPz8f+fn5mq911YAi0oXT6omIrJ/kitvOzs4YNGiQMWKp0I0bNzBt2jTExsbCyUmeqcrR0dFYuHChLOeimoXT6omIrJ/ktdtM5eTJk0hPT0ebNm1gZ2cHOzs77N+/H6tWrYKdnR18fX1RUFCAzMxMrdfduXMHfn5+Os85Z84cZGVlaT5YwoDEMreq2kREJD+D1m4zhaeeegpnz57V2jZ+/HiEhIRg9uzZqF+/Puzt7bFnzx48++yzAIALFy7g+vXriIiI0HlOR0dHODo6Gj12sj7qafWTN56CAtCqPyRlWj0HUhMRmS+LSZLc3NzQvHlzrW2urq7w8vLSbJ8wYQJmzJgBT09PuLu745VXXkFERISomW1EUqmn1S/ckaQ1iNtP6YT5/UMrnVYfk5ha7rX+Il9rLEzaiIj+ZTFJkhgfffQRbGxs8OyzzyI/Px+9e/fGZ599ZuqwyIoZWlVbXWOpbP0NdY0lU9QuMsekzZiYEBJRZUTVSZIy68uS6wuxThJVB6k1lqrjYa4vaVNfxdoKTta0hJDI2hnr+S2qJcnDwwMKRcV/lAVBgEKhQHFxsSyBEVkrKTWWsh4VGP1hXqwSsHBHks513QSUJEoLdyShZ6ifVbS0mGMrHhGZJ1FJ0t69e40dB1GNIbZ2UmxSGr46fNXoD3MpSVtEQ68qX8+UalpCSERVIypJ6tatm7HjIKoxxNZO2pZwu1oe5jWpMKYcCeH1u7mIWrkfjwpVcLa3Qcy0bgj0Nn4xXSKqfgYP3M7NzcX169dRUFCgtb3sorNEpE1dYyktK09nEqQAUNvVHhkPC3TsLSFn605NKoxZ1YSw0Vu/okj179e5hSo88cFe2NkAl97rJ0eIRGRGJBeT/Oeff/D000/Dzc0Njz/+OMLCwrQ+iKhi6hpLAMoVo1R/Pah1XVHnkqN1pyYVxqxKQlg2QSqtSFWyn4isi+Qkafr06cjMzMTRo0fh7OyMmJgYfP3112jcuDF++eUXY8RIZHXUNZb8lNoPYz+lE9aMaoPIUN1V4suSo3VHTNImpjCmJTA0Ibx+N1dvgqRWpCo5joish+Tutri4OGzfvh1t27aFjY0NGjRogJ49e8Ld3R3R0dHo149NzmS+zKk2TkU1lopVQqVdcn4ytu5UtTCmpTC0UnrUyv2izh+1cj+SFveRJVYiMj3JSdLDhw/h4+MDAKhduzb++ecfNGnSBC1atMCpU6dkD5BILuZYG8fWRqFzTJFcy55IYWhhTEtjSEL4qLCSZiSJxxGRZZCcJDVt2hQXLlxAUFAQWrVqhXXr1iEoKAhr166Fv791/LdJ1scSa+OYonVHX9JmbaQmhM72NsgVkQA521vMmuFEJIKoitulbdy4EUVFRRg3bhxOnjyJqKgoZGRkwMHBARs2bMBzzz1nrFiNjhW3rZPUCtfmprq6CM2pK9LcXL+biyc+qLxe3IHXu7McAJEJmLTidmmjRo3SfB4eHo5r167h/PnzCAwMhLe3t2yBEcnF0oslVkfrjjl2RZqTQG8X2NmgwsHbdjZggkRkZarUNiwIApydndGmTRsmSGS2qlobp1glIP7yPWxPuIX4y/dQrJLU+Gr21F2RZRNJdVdkTGKqiSITp7q+P5fe6wc7PX8xWSeJyDoZVEzyyy+/xEcffYTk5GQAQOPGjTF9+nT85z//kTU4IjlUpTaOtbewWPoyHdX9/bn0Xj9W3CaqQSQnSfPmzcOKFSvwyiuvICIiAgAQHx+P1157DdevX8eiRYtkD5KoKsRUuNY1nd4SB3tLZcldkab6/gR6u3CaP1ENITlJWrNmDdavX48RI0Zotg0YMAAtW7bEK6+8wiSJzI4h0+ktvYVFLEtdt82Svj8cEE9kuSQnSYWFhWjbtm257eHh4SgqKpIlKCK5SZ1Ob8ktLFJY6rptlvL9sfbuWiJrJzlJGj16NNasWYMVK1Zobf/8888xcuRI2QIjkpuU2jiW2sIilaFdkaZmCd+fmtBdS2TtDB64/fvvv6Njx44AgKNHj+L69esYM2YMZsyYoTmubCJFZGpip9NbaguLVKao7C0Hc//+WFJ3IBHpJzlJSkxMRJs2bQAAly9fBgB4e3vD29sbiYmJmuMUCv7ik+WqrIUFAGwUwP2H+dUalzH0DPXD9Mgm+OpwCjIfFWq2m/O6bebeAmYp3YFEVDHJSdLevZVXnSWydKVbWPRRCcCUzaexxkZhlomEGLrGzHg422N85yBM7dHYbFs5zL0FzBK6A4moclxoiEiPqOb++PT5NqjsObtwR5JFFpjUV0Qy61EhPv4jGbFJaSaKTBz1YHw/pXaXmp/SyeTjfcy9O5CIxBHVkjR48GBs2LAB7u7uGDx4cIXHbtmyRZbAiKpDZdOza7s6oKL8R45uE1NMEbeWMTNSF6qtLubeHUhE4ohKkpRKpWaMkVKpNGpAVLOYsoaMmOnZxu42MdUUcWsaM1Mda9tJZcruQEN+px4VFOO9XUm4ei8XQV4ueKtvKJwdbGWPjcjSKARBsLx+AiMx1irCpJspa8jom56tfpSou2viL9/DiPVHKj3fdxM7Sn5Qi41BbsUqAR/FXsQney9VeuzK4a0xsHVd2WOoKar7Z9yQ60385jhik9LLbe8Z6oP1Y9rJHiORMRjr+S05SUpJSUFRUREaN26stT05ORn29vYICgqSLbjqxiSp+pgqQQBKkoQuy+L0tqSou0IOze4BAOiyLK7SbpNDs3tIahWQEoOcrQ26HqIVMST5I23V1VpqyO+UvgRJjYkSWQpjPb8lD9weN24c/vzzz3Lbjx49inHjxskRE1m5ysbDABUPhi4oUuHLg1cwb3sivjx4BQVFKknXl9LVpO42Af592KhVpdtESgxy0TdQWxcFSlogOGam6tTdgQNb10VEQy+jdbFJ/Z16VFBcYYIEALFJ6XhUUCxfoEQWRnKSdPr0aXTu3Lnc9o4dOyIhIUGOmMjKVSVBiN6VhJC5v2Hxr+fwTfw1LP71HELm/oboXUmiry91nJExZlFV9xTxih6iZZnDFHqSxpDfqfdE/s6IPY7IGkmuk6RQKPDgwYNy27OyslBczP84qHKGJgjRu5Kw7kBKueNUAjTb5/QNrfS8hkzPlnsWVXVPEa/sIVqaOReRJN0M+Z26ei9X1GvEHkdkjSS3JD3xxBOIjo7WSoiKi4sRHR2NLl26yBocWSdDEoSCIhXWHyyfIJW2/mCKqK439fRsfemNvq4mObtNDI3BUGIfolO7N8Sh2T2YIFkYQ36ngrxcRL1G7HFE1khykrRs2TLExcWhadOmGD9+PMaPH4+mTZviwIEDWL58uTFiJCtjSILwbfzVCusVASUtSt/GX630+sYaZyRFdccg9iHauVEddrFZIEN+p94S0eoq5TgiayQ5SQoNDcWZM2cwbNgwpKen48GDBxgzZgzOnz+P5s2bGyNGsjKGJAjXMsQ1+R9M/kfUceZQrbk6Y6juliuqXob8Tjk72KJnqE+F5+0Z6sN6SVSjsU5SKSwBUL2k1HT58uAVLP71nKjzrpWQYEidnm2M6dzVPUUc0F3c0NRLeVDVsU4S1VRmUycJADIzM3Hs2DGkp6dDpdIeAzJmzBjZgistOjoaW7Zswfnz5+Hs7IxOnTph2bJlaNq0qeaYvLw8zJw5E99//z3y8/PRu3dvfPbZZ/D19RV1DSZJ1U9sglBQpELI3N8q7XIDSh4KuuoLVTUZMWXxS7lYwz1QxT/LrLhNNZHZJEk7duzAyJEjkZOTA3d3d81yJUDJzLeMDPnqupQWFRWF4cOHo127digqKsJbb72FxMREJCUlwdXVFQAwefJk/Prrr9iwYQOUSiWmTp0KGxsbHD58WNQ1mCSZN32z23QpWwSxqsmBKYtfys2US8GYmjXcOxNdovLMJklq0qQJ+vbti/feew8uLqab9fDPP//Ax8cH+/fvxxNPPIGsrCzUqVMHmzdvxpAhQwAA58+fR7NmzRAfH4+OHTtWek4mSebvhQ3HEHe+8nFHpZfTqCjBEQD0ae6LhnXcENHQCx0fKz9rzVTVsUle1pBcWFOyTiQns6m4fevWLbz66qsmTZCAkrpMAODpWTLQ9OTJkygsLERkZKTmmJCQEAQGBiI+Pt4kMZL8JnZtKOo49WwuMZWIf0u8g0/2XsLIL44i/N1YxCSmao4pVgnYcDil2qtjk7z0VRtPy8rD5I2ntL7n5qqqleqJSDrJSVLv3r1x4sQJY8QimkqlwvTp09G5c2fNjLq0tDQ4ODjAw8ND61hfX1+kpaXpPE9+fj6ys7O1Psi8SZmlJSbBKSsztxAv/e+hGZOYii7L4kQPGJerOjbJy1qSC1MsZUNU00muuN2vXz/MmjULSUlJaNGiBezt7bX2DxgwQLbg9JkyZQoSExNx6NChKp0nOjoaCxculCkqqg7qqc6TN57SdJeplZ7qHJuUJmkh17LmbDmL+7mFkl4jV3VskpeU5MKcF/Ot7qVsiMiAJGnixIkAgEWLFpXbp1AojL40ydSpU7Fz504cOHAA9erV02z38/NDQUEBMjMztVqT7ty5Az8/P53nmjNnDmbMmKH5Ojs7G/Xr1zda7CQPdX2hskmQejkNADrHbUghJUFSj0lijSHzZC3JRXUvZUNEBiRJZaf8VxdBEPDKK69g69at2LdvH4KDg7X2h4eHw97eHnv27MGzzz4LALhw4QKuX7+OiIgIned0dHSEo6Oj0WMn+elaSy28QW0cv5qBKZuqliBJoW69Gt4uEDvP3LbYGVPWzFqSC3VXc1pWns6fbybrRPKTnCSZypQpU7B582Zs374dbm5umnFGSqUSzs7OUCqVmDBhAmbMmAFPT0+4u7vjlVdeQUREhKiZbWR51GupASUDc7st32tw95qhPFzsIQD46I+Lmm1yzZiyhunq5sBckouqfj/FdjXzZ4RIPqJKAKxatQqTJk2Ck5MTVq1aVeGxr776qmzBlVa6HlNpX331FcaNGwfg32KS3333nVYxSX3dbWWxBIBl0jct2tiGtKmLn0/dKrddjunY1jBd3ZyYutq4nN9P/mwQlWfSOknBwcE4ceIEvLy8ynVzaZ1MocCVK1dkC6661fQkyRJbLiqrYWQoDxd7ZOUWVtjyIAgC0rLzdb6+KrWTWAvHOEyVXBjj+2mJv6tExmSs57eo7raUlBSdn5P1MJf/TqX+8a9s5pJUtV3sET24BQBU2K0xvF2gVhdbWYbOmKpsuroCJdPVe4b68aEoka5xbMZOLoz1/Szd1UxExiNpTFJhYSFCQkKwc+dONGvWzFgxUTXT95+uutBedbVcGJKoyTUjaVDrAAxpW1+r4nZFM+jyi8RNYJAan7VMVzdX1Z1c8PtJZNkkJUn29vbIyzPvabIkjbm0XBiaqFV1RpK6W+yDYa3L3V9FLQ/xl++JOr/U+KxlurrcLLV7id9PIssmeXbblClTsGzZMnzxxRews7OYyXGkhzn8p1uVRK2ymUuVEVDxjCB9LQ/GmjFlLdPV5WQuXcGG4PeTyLJJXpbk+PHj2LJlCwIDA9G7d28MHjxY64Msizn8p1uV5RbU06INNe2pRgY9aEtft2x6VXY6drFKQPzle9iecAvxl+9VuPxF6/oeoq4v9jhLZ+lrrklZRoeIzI/kpiAPDw9NsUayfObwn64ciZrSxR6ZEpcRAVCl9boqq/wd1dxfcivI5qPXRF1789FrmND1MYNjtwTFKgFvbjlr8q7gqmBtIyLLJjlJ+uqrr4wRB5mIORTaq0qiVvUaSdIfTmXHx+yf1R0nr90vN17GkHFW1zJyRcVwNCUD3m6OFjU+Rx99440+iUuuMPG1lEHPYpJpIjJPopMklUqF5cuX45dffkFBQQGeeuopzJ8/H87OzsaMj4zMHP7TNTRRq2gsk1hSH64VtQwNbF1XVGwVtYI08HQRFcfvSXfwe9Idretb4sNW3/s5t18ovjp8VdQ5pHQFm2oAuCnKDxBR1Ykek7RkyRK89dZbqFWrFurWrYuVK1diypQpxoyNqon6P10/pXZLjZ/SqVqm/0sZ31NaVWsk1XaxR8fHxCdJUsbHGDrOanREEKQ+Ny1lfE5ZFb2fL28+hcxH4rpPxbZExiSmosuyOIxYfwTTvk/AiPVH0GVZXLW9b+pJAANb10VEQy8mSEQWQHRL0jfffIPPPvsML774IgDgjz/+QL9+/fDFF1/Axkby+G8yM3L/pyv1P3ZDuiSqOph8yTMtRN+f1JYhQ8dZOdjZYGLXYKw7IL5oq6WMzymtsvdTLA8Xe1FdweZSC4yILIvoJOn69evo27ev5uvIyEgoFArcvn0b9erVM0pwVL3kKrRn6JRtqYlaVQeTL/41CTY2EPVwlFoqoSrjrOb0LWlVW38wBWLHlVvK+Bw1uSqlj+8UXGlSKCYhW/DL3xaTYBJR9RGdJBUVFcHJSfsPur29PQoLpc8oIutRtsXo/sMCTNls+H/sUhI19VgmQx+2pWOqLDmT2jJU1QHxc/qGYmavEHwbfxXXMnLx4FEhtibcFn19cydHnB4u9pjao1Glx4lJyNKy8/FJ3CVMi2xc5biIyHqITpIEQcC4cePg6Oio2ZaXl4eXXnoJrq6umm1btmyRN0IyW7pajGwUurtLjNElpB7L9NL/VneXSh3Tm/89gzlbzuJ+qZlUZVu+pLYMyTEg3sHORjPNP/7yPVFJkqeLg6g4TU1KK2DZ909t6WBx3aViE7KP/riIpn612O1GRBqiBxONHTsWPj4+UCqVmo9Ro0YhICBAaxsZR0GRCl8evIJ52xPx5cErKNCzdpiUwoVVoW/QbUWXq6gopKGimvtjQucgg18vAMh8VKSVIAElcZYeDG1IUUA5B8RXdn21sV8dQ/SuJNHnlULsz6AYYt/Pz54PK/f++SudsFbC+yclIVu4I8lovzNEZHkUgiDwL8L/ZGdnQ6lUIisrC+7u7qYORyN6V1K58SkKAP1a+mPl8DDNf9MVTaeu7eog29TjYpWALsviDO7mWjm8tdZ0+ao6fOkuRn5xVLbzlebn7ojDbz6lVfcI0N0ypC/xkWvaub7r6/LiE8GasU1y0PUzaKMAJnY1/Dpi38+qvn9Sf16/m9jRIsZ1EdG/jPX8ZpJUijkmSdG7kiqc6eTiYIsVw1oBgOiiilWtqxN/+R5GrD9i0GsB+R9Ch5PvYuSXxkmSAOC1yCaasSqmXkcsJjEV87f/jTsP8is8zkYBnF/cBw52VZ95WtnPYFUSsup6P2MSU0V3y8qdxBOR8Rnr+c0Vas1YQZEK6w9WPBU8t6AYL208BQ8Xe9FTp6s67dnQQbfGqt5992HFCUNVlR6rYuqigFHN/XEj4xGW7DpX4XEqAfg2/mqVly4R8zO4/mAKZvYKMSghq673M6q5P16LbIyP/kiu9FguNktEaixwZMa+jb8qegq4lHXL1Kc0ZPxFsUrA3UpaMXQxZvXu6niolX6vTF0U8MZ9cUuXiF3ipCJifgbVCZlYZcfNAaiW93Nqj8bwc9f/s8LFZomoLLYkmTE5HnL6GFJXR1fXiD42Cu1B3MZcp6qqpQDEMKcaRGKXLhF7XEXE/gyKPc6U3ZW2NgosGBBa4TgoLjZLRKUxSTJjcjzkKlNZ15l60GxsUhr+T8RaWurHyycjwlDb1dEoXSi6BvKqp9sbc4CdudQgGh0RhCW7zlXYwmOjKDmuqsT+DNavXfkajuZQ9ZqLzRKRFEySzJiYh2FVVdRVVTJIOBF3HhSIPl9lD5tilYA/L93Ff0/dRG5BEdoFeWFspyDR41kqaonQ9fCTk7mMVRGzdMnErsGyDNoW+zP45aGrqO/povm+l01kwxvUNmjBX2Mw9bgyIrIcnN1WijnObpu6+RR2npF/AU71IOpDs3vofDhImQ2kNrdfM4zrXLJMhK7WntikNMz48S/kFhRrx6IAJomYSq6vJaL0lHH1wy8t6xHu5hRgdVwysvOKJN2HLv4VvFemYoxp+WqPCorx3q4kXL2Xi3s5eUhKzanw+NLfAwDlklVPVwdkPKw82eb0eyIyBGe31VA9Q31lT5IqG39RrBIw48e/JJ/X281RU0+o7EPSw8Ve7+ByQYCmVUTfw72gSIW3tiaKaolQP2TjL9+TJUECgD7NS5Ivc2pxKLt0SQNPF4yOEN8qp8/Eb44jNild0ms01cu3nEVWbmG575OYBAkwny5NNV0tYiev3WcLFFENwSTJTBUUqfBt/FUcTblXpfM429vA0d5WK0GprEvs4z8ulGvtEePq3Yd6W3vEzL7TN5U8JjEVb209i4yH+s+hayC6nA/c/zt8Ff93+Gq11kQSo/TSJXIwJEFSEyBtlqUu5tKlCehfdqd0y525/TwQkbyYJJkhXd0ohnpUqMKjQhVa1VPiMW9XDG5TD50aeev97zcmMRWr4y4bdK2P/kiWVK+pLF21ffQlXfqUToyM8cCtzkHG1e1RQbHBCVJVGauGlqH0/dyV/Z205p8HImKSVC2C3vxV5/arS/uV21ZZdWND/XUzC3/dzML2v27rHbNSrBKwcEfV1v2qaktC6ank6nikJF1pWXkoVgmwtVFoSgOkZeVJTtzKthioVTTIWK7lR0zlPSOt+VYZc5t+L+XnrroHnRNR9WIxSSPTlyDp2iemunFVqf43/kfXIqjHUjKMWmtIjNJTzg2JJ/q38widF4NdZ27D1kaB+f2lDWAeE9EAc/s1k7xQb0xiKrosi8OI9Ucw7fsEjFh/BF2WxWkWyLUEV+8Zry5XaZ6u9lpfG7LgrzFJ/bkzxsLNRGQe2JJkRBUlSKWPWf5sC9TzdEXirSyjTvcvTdf4H1MPmi1b28fQePKLVHh582lExF9DYz83PNumLuLO/4OM3MoHD/dp7i/6uurjzKH+jxyCvFxwsPJVOwCUtJ6ULcYooGSAvq6B2+pj/JRO2D+ru1kPfjb0587Uvz9EJD8mSUYiJkFSm/XfswBKBllXF13jf0w9aPbJpnVw8tp9zUOzqvHEp2QgvtR/9052NsgrUuk9vraLPdoHe4puEbh6N7fCrpnSy79YQlfMW31D8e2R65Uet2pYa0TvPq+zGCNQstCyriQKKOlSc7CzMetp/ob+3Jn694eI5MckyYw8KtT/ADeGsktJVGUMjxzizv+DuPP/aGYMnbx2X9bzV5QgAcD93ELEJqWhZ6gfXBxskFtQ8fFfx6egoLi40q4Zc1rSpCLODrboGepT4eDtnqE+GNCmLvq1DtA7/srSK1pL/T0wt0HnRCQfJkk1WG6+dg0h9RgeXS0B1SktK09yIUs5qAfgdmvig0eVJEgAkPGwEJ/uFTcTMC3bMrpi1o9pp7cMQM9QH6wf0w7Av4v86mLpFa2l/B6Y26BzIpIXk6Qa7I9z6TicfBcd//ewO5aSgfwiFaZHNsZ3x64jLTtfc6yjLeDqaA8HOxtkPypErhFbvUyVnKkH4L63S9qMOjEycvIrP8hMrB/TTqvidpCXC97qGwpnB1vR56goibIE+tZ4q86Fm4nI9KwySfr000+xfPlypKWloVWrVli9ejXat29v6rDMTuajQoz88ig8XEpmG2kVnHR3wtMt/XEo+S4yHxUivxjIr+L0fkthjFlenq4Osp/TmJwdbLH4mRamDsOkdLWIseI2Uc1idUnSDz/8gBkzZmDt2rXo0KEDPv74Y/Tu3RsXLlyAj49PtcXhCuBhtV2tanTVNkrLzjPKmnGWQMosL7H8lM7ynrAMU9RosvS6UGLoahGz5BYyIpLG6ha47dChA9q1a4dPPvkEAKBSqVC/fn288sorePPNNyt8rdwL5EmZ4Uampx6A+3rPppj5s/S16/Qx9uK4upbPMPZyGTGJqVjwy99aXbK1Xeyw5JkW6NsywCjXJCLShwvcilBQUICTJ09izpw5mm02NjaIjIxEfHx8uePz8/ORn//vH/ns7OxqiZPM1+MBbrImSApoD+qVu/WlshpNnz7fBrVdHWRt7YlJTNU5sP5+bhFe3nwaL97M1LtQMRGRJbGqJOnu3bsoLi6Gr6+v1nZfX1+cP3++3PHR0dFYuHCh0eK5/F5fNHxrl9HOT4CdjQJFMlTgrOVoi4iGXrKuXeZib4sVz7XStObI3eIjpkbT1O9Oyboga7FKwJtbzlZ4zLoDKWhVrzb6tuRgZiKybDV6WZI5c+YgKytL83Hjxg1Zz29t4zPMkRwJEgAoIGDPOXkXd32tZxOtBGnyxlPlaiqpW3wMWb5EzPIZ+hZkNXS5lCOX74lan++dbWdRXF3l44mIjMSqkiRvb2/Y2trizp07Wtvv3LkDPz+/csc7OjrC3d1d60NuuhaxJfPzIF8l+5IwXq4OiL98DwVFKlFVuaUmFYYsg1GV6wFA/JW7oo7LyC2UdS2zgiIVVu25gFYLYhAydxciP9yLjJzKl5khIqoKq+puc3BwQHh4OPbs2YNnnnkGQMnA7T179mDq1Kkmi2vvjCfRfcU+k12fTGPGTyVjmzxdHZDxUP8DvfQCqWJnThUUqXDiqmFJiPp6Gw6nwNvNUeJYJfGto3KtZRa9KwnrDmgv/Hzpn1y0eTcWzvY2+GJsO3R8zIstt0QkO6tKkgBgxowZGDt2LNq2bYv27dvj448/xsOHDzF+/HiTxRTs42qya5PpVZQglSY2qYjelYT1B1Oq3PK1+Ndzms/FjlWKaOiFT/ZeEnV+OdYy05UglfaoUIWRX5TU+lo6uAWLOhKRrKyquw0AnnvuOXzwwQeYN28eWrdujYSEBMTExJQbzF3d2O1GlRGTVKiTBrm7BsWOVer4mJem+GhF/Nwdq7yWWUGRqsIEqbTM3EK8VIWxVkREulhdkgQAU6dOxbVr15Cfn4+jR4+iQ4cOpg4JQMlsN6KyFChpyaksqSgoUmH9QXFJg9SeJ7FjlWxtFFg6uPJK3AsGPF7l7q9v469Kfo2hY62IiHSxyiTJXNnaKPBK94amDoMM4OtuvGVFBIhbIPXb+KuiWpBGdwzEJyPCoICUEUTaY6MqEtXcH2tHtdHZouThYo+1o9rI0u11LUP68jBi4iciEsvqxiSZu+k9m2LN/iuyTV0n4/N1c8Cfb0aioEiF/3x9HIcv35P1/K9FNhaVVIhNGhQKBfq2DMAaG0W5ukxiiBkbpV7X7MiVe4i/fA+AgIjHvNGxoXwDqBt4uhj0OrkGjBMRMUmqZrY2CnzyfJjOisVknlrW94CtjQL7L6bjT5kTJAAI8hY3sF9s0qA+ruwCrXcf5GsN1tZH7IBrWxsFOjfyRudG3qKOl2p0RJCoeMuSY8A4ERHA7jaTiGruj8+eDzN1GCTSnnPpeFRQrLfWUVWJfaiPjgiqdKyRjaLkODX1Aq0DW9fFuM7B8Fc66e2CEzs2qro42NngxSeCJb3GnOInIsvHJMlE+rYMwEfDWpk6DBJBJQDv7ZLebVUZqUmJg50NJnatOGmY2DUYDna6f61tbRSY3z9Uc+2ysQDixkYBJYPIvzx4BfO2J+LLg1dQUKSq9DWGmNM3VFKiJDZ+IiIxFIIgcHDM/xhrFeGKDFh9EGducWFdc9fUzw0X0h5Ieo0CwCs9GmF1XEldIaHMPgBYY8AgZ111kmwUJQmSmIVlq7qGXFWvb4iCIhVe/+EUfjl7R+f+2i72iGadJKIay1jPbyZJpZgiSQKYKFmrF58oSRrkXtgWKEkavo2/imsZuWjg6YLREUF6W5B0KVYJmrFKUipuV1bcUX3PxlKsEnDkyj0cvnQXtzMfoa6HMzo18mbFbaIajklSNTBVkgQA20/fwrQfEqr1mmQ8CgAX3u2jSVzEJiWGJi/VoaBIhZC5v1VYhsBGAZxf3EdSwkZEVFXGen5zdpuZGBhWF56uDhj9f8dMHQrJQEBJXaMJXR8D8O8A6ooYo8VJF0MTMTF1mlSC9n0bMx4pzDn5JCLzxSTJjHRq5A0PF3tk5haaOhSSQem6RpV1j8UkpmLyxlPlZs+plwsxZOySLlVJxMTWaZJSBLI6EsPqSj6JyPqwTdyMiF3ygSyDul5R9K4khMz9DYt/PYdv4q9h8a/nEDL3N0TvSgJQ0sqhr7yA2OVCxFAnYmVn6VW2bltWbiGe/ewwtp6+Keo6Yus5GRqPFNVxDSKyXkySzIx6yQc/dxbEs2TqekX6FqRVCcC6AymI3pWEYykZFZYXELtcSEUMTcS6LY9Dq0W/4+T1TDzIK670OmXrNMkdjxTVlXwSkfVikmSGopr74/CbPfDdxI5YObw1eoXWMXVIJJG6nlFlC9KuP5iC1MxHos5ZleU2DEnEui2Pw7V74mJTq+/pLGrQdnUkhtVxDSKybhyTZKZKD/Qd2Loudvx1G698d9rEUVFlStcL+vLgFVEDnU/fuC/q3FVZbkNsgqU+Liu3UHKCBADX7j3Co4JiODvYyhqPIarjGkRk3diSZCH6twrA2lFt4FLJw4dMp2czH/y9MEpTJ+jqvYeiXqcShEqXC/Fzd0TirSyDK1yLTbDUx72wwfBZlu/9b6yVnPEYojquQUTWjUmSBYlq7o+zC3pj+lON4OrIZMncxJ5LR48P90keDGyjqHi5EAHAnQf5WLJL98BvtYqWCrn/sKDCdd/KLpFyuwpLsFy9V/nstvbBnkZfR646rkFE1o1JkoWxtVFges+mODO/Nzb9pwOUTuwxNSelZ021rl9b1Gta16+NqOb+WDOqDfyU2q0a6pbDsiVfSw/8BoAlv/6Npu/onkEXk5iKlzefqrTrr/S6ZwFKw1tXgrwqn90m5zpyprwGEVk3VtwuxZQVtw2lr74OmY4CgJ/SCR8MaYWRXx6t9PjvJnbUjD8rXfTQ08UBY786VmmF6+5N62DP+X/0HuNoZ4P8CrrnFArg0xFh6NsyQLMtK7cQrRb9XmnsupxbFFXpmCQ11kkiIjmw4jbppG6BKPsQINNRz5qCouRhXNH3pWx3T+kB+2IHfleUIAGoMEECSlqplC4OWtuULvZo4OUsefB295A6ohMkoOTnt2eon1GrYVfHNYjIOrG7zQpENffHodk9MLdfM1OHQqXczcnH/P6hFY6Jqai7R0rl6qqKv3yv3Lb9s3og0FN8t1udWg74alx7yddWJ4YDW9dFREPjLFRbHdcgIuvDJMlK2NooMK5zcIUDVal6+bg5Iaq5PyY9EQyFjm9KZS0uYitXy6N8k1VMYioKy9SP1Pez1aOpN46/01P+sIiITIjdbVZEPVB18sZTmllRVP3UY5LaB3siJjEV6w7oLiiZW1CMlzaegou9LerVdsL3kzrBs9a/3V6jI4KwZNe5Srvc5BDxmLfW1xWNdVMAeKKJN1RCySDtt/qGSupiIyKyFGxJsjL6Zkk52tnAyZ7fbmMrPWsKABb8UnnNoNzCYlxMf4g278ai3buxmu0Odjaayt3G5OFij47/GwcFiFvO4+KdHGwY3x6Ln2nBBImIrBZbkqyQvoGqQMlSDbsTU7Eh/pqJo7R8tV3sIQDIzC3UbPMrNWsq/vI9pGVLG0z/T04B2r0bq+m6mtM3FFfuPkRsUnq5Yx1sFXBxtNO6vi6Rzergj3P6B3cvHdxCa4yOlOU8IkolV0RE1oZJkpUqPUuqtIiGXoho6IX2wZ6Y+t3paunKMVe+bg74J6dA8nvQp7kfRnVsgI6Plby/+mZNGbrcxT85BcjIKYBnLQfEJKbqTJAAoKBYQEElCVJdDyfkFwl4orE3zqdmIz2nQLPPz90RCwY8Xm4aPJfzICIqwSSphurbMgCfQIGXN58ydSgm4+HqiIOzn8JbW87i51M3Rb/ut8Q0JNzI1LQY6WtNuXpX3LIkugz//E/8Nr2bqO46D2c7ONnbabVa2dkoUKQScCszD7cy/93etoEHRkcEVTgNnst5EBGV4CCVGqxvS3+sHdUG/lWormwsT7c0fpG/wNrOsLVR4PDlu5JfW7qyti4xian46I9kg2NLf1CAYykZorrrMh8V4cOhrfDdxI5YObw1wht4oEhP89iJa5nYeeZ2hdPguZwHEVEJJkk1nCE1loxdYiYi2BMrh4cZvZzBR8+FVTr+Rh91CrJwRxKKyyQk6oHPVeHj5iCpO+vuw3xENPRCr1A/nLyWWeGxsUnpeFRQrHc/l/MgIirBJIlE11jycLbHpv90wPnFfTStFl+Pb4fn29dHi7ruaOZbC0qnqv9IfT2hg+ZBbawhUy3ruaOWk12VxtWUHsBcmqGJV2nfT+okqTtLfex7u8QlZ5Udp2+WpJ/SCWtGteFyHkRUI3BMEgGouMaSOnFa+mwLdG5UUk+n9Dicbk19NJ+XXnvs7oN8LP71nKQ4JnYNgoNdSaIV1dwfLQPcceZ2tiG3pFfLeu74ZWpXAPKMqymbaFV1QHOdWg7wrOWA9i6e8HN3qrTLzc/dUdP1dfWeuCrdYo7jch5EVNOxJYk05Gg9KL38g9QK4D1DffB2v8e1tt1/VPHsLSkaeDojcUFvTYIEVD7+RoyyiVZVEq86tRw00/9tbRRYMCC00tcsGPC4JnEJ8hJXpVvscVzOg4hqMoUgCDV4Erg2Y60ibGlKtwZVtfVAXbkZ0F8B3M3JDu8NaoH+rQLK7Xt61UEkVrElyUYBTOwajDl9dScc+mKsrGq5urL2odk9tN6fYpWALsvikJaVp7ditZ/SCb9M6YKRX8Qj/UEBfNwcylXcLh3fm1vOlquH5OFij6WDW2glr48KitFsXkwFUZc4tyiKRSCJyGoY6/ltEUnS1atXsXjxYsTFxSEtLQ0BAQEYNWoU3n77bTg4/PtQOXPmDKZMmYLjx4+jTp06eOWVV/DGG2+Ivg6TJOOISUzFwh1JWuN0vFwdMLB1AHqG+lWYhO1NuoPx35yQfM0ApSMiQ/3QwNMFoyP+7cKTEqO/0gkDWvnj8/8tK6KrC1JfC1tFiVdFr9OnWCXgyJV7/1uIVkDEY97oqKdlZ+I3x/XWVgJKWuzWj2kn+tpERObOWM9vixiTdP78eahUKqxbtw6NGjVCYmIiJk6ciIcPH+KDDz4AUPIG9erVC5GRkVi7di3Onj2LF154AR4eHpg0aZKJ76Bmq8rYlidCfDQ1f8SyUQAHZz9VrnWnoutXFGNYYO1yCVTpytr67nnNqDaSX6ePrY0CnRt5a8aEVWT9mHZ6EyUmSERE4llES5Iuy5cvx5o1a3DlyhUAwJo1a/D2228jLS1N07r05ptvYtu2bTh//ryoc7IlyTzFJKbipY3ii16uLdNKE5OYigW//I207HzNttou9ljyTAv0FVmPydAuSDm7LqV6VFCM93Yl4eq9XC5ES0RWrUa3JOmSlZUFT89/i9nFx8fjiSee0Op+6927N5YtW4b79++jdu3a5c6Rn5+P/Px/H5zZ2fLOoiJ5RDUvKXo5b1ui1rIaZelaZkNfgnU/txAvbz6FF2/qH6tUmr5lXoz1Ojk4O9hi8TMtTHJtIiJrYJFJ0qVLl7B69WpNVxsApKWlIThYe8V0X19fzT5dSVJ0dDQWLlxo3GBJFrq6w8Ib1MbJa/f1ttIUqwS8ueVshedddyAFrep5oG/L8oPGiYioZjNpCYA333wTCoWiwo+yXWW3bt1CVFQUhg4diokTJ1bp+nPmzEFWVpbm48aNG1U6HxlX2enoDnY2FU5PP3L5XrkZYbq8ve1suarZREREJm1JmjlzJsaNG1fhMY899pjm89u3b6N79+7o1KkTPv/8c63j/Pz8cOfOHa1t6q/9/Px0ntvR0RGOjo4GRE6WIP6KuDXZ7ucW4VhKhsm6xYiIyDyZNEmqU6cO6tSpI+rYW7duoXv37ggPD8dXX30FGxvtRrCIiAi8/fbbKCwshL29PQAgNjYWTZs21dnVRjWB+AHSVa2STURE1sciKm7funULTz75JAIDA/HBBx/gn3/+QVpaGtLS0jTHPP/883BwcMCECRPw999/44cffsDKlSsxY8YME0ZOpiSlZUiO5UmIiMi6WMTA7djYWFy6dAmXLl1CvXr1tPapKxgolUr8/vvvmDJlCsLDw+Ht7Y158+axRlIN1vExL3i42Fc6Lqn02mdERERqFlsnyRhYJ8n6iKmxVLauEhERWRZjPb8toruNyFDqGkseLvbl9nm42DNBIiIivSyiu42oKtQ1lsSufUZERAQwSaIaQsraZ0RERAC724iIiIh0YpJEREREpAOTJCIiIiIdmCQRERER6cAkiYiIiEgHJklEREREOjBJIiIiItKBSRIRERGRDkySiIiIiHRgxe1S1Gv9ZmdnmzgSIiIiEkv93FY/x+XCJKmUBw8eAADq169v4kiIiIhIqgcPHkCpVMp2PoUgd9plwVQqFW7fvg03NzcoFNW38Gl2djbq16+PGzduwN3dvdquW914n9aF92ldasp9AjXnXmvafSYlJaFp06awsZFvJBFbkkqxsbFBvXr1THZ9d3d3q/5BVuN9Whfep3WpKfcJ1Jx7rSn3WbduXVkTJIADt4mIiIh0YpJEREREpAOTJDPg6OiI+fPnw9HR0dShGBXv07rwPq1LTblPoObcK++z6jhwm4iIiEgHtiQRERER6cAkiYiIiEgHJklEREREOjBJIiIiItKBSZKJffrppwgKCoKTkxM6dOiAY8eOmTqkKomOjka7du3g5uYGHx8fPPPMM7hw4YLWMXl5eZgyZQq8vLxQq1YtPPvss7hz546JIpbH0qVLoVAoMH36dM02a7nPW7duYdSoUfDy8oKzszNatGiBEydOaPYLgoB58+bB398fzs7OiIyMRHJysgkjNkxxcTHmzp2L4OBgODs7o2HDhli8eLHWWlCWeK8HDhxA//79ERAQAIVCgW3btmntF3NPGRkZGDlyJNzd3eHh4YEJEyYgJyenGu+ichXdZ2FhIWbPno0WLVrA1dUVAQEBGDNmDG7fvq11Dku/z7JeeuklKBQKfPzxx1rbreU+z507hwEDBkCpVMLV1RXt2rXD9evXNfvl+BvMJMmEfvjhB8yYMQPz58/HqVOn0KpVK/Tu3Rvp6emmDs1g+/fvx5QpU3DkyBHExsaisLAQvXr1wsOHDzXHvPbaa9ixYwd++ukn7N+/H7dv38bgwYNNGHXVHD9+HOvWrUPLli21tlvDfd6/fx+dO3eGvb09fvvtNyQlJeHDDz9E7dq1Nce8//77WLVqFdauXYujR4/C1dUVvXv3Rl5engkjl27ZsmVYs2YNPvnkE5w7dw7Lli3D+++/j9WrV2uOscR7ffjwIVq1aoVPP/1U534x9zRy5Ej8/fffiI2Nxc6dO3HgwAFMmjSpum5BlIruMzc3F6dOncLcuXNx6tQpbNmyBRcuXMCAAQO0jrP0+yxt69atOHLkCAICAsrts4b7vHz5Mrp06YKQkBDs27cPZ86cwdy5c+Hk5KQ5Rpa/wQKZTPv27YUpU6Zovi4uLhYCAgKE6OhoE0Ylr/T0dAGAsH//fkEQBCEzM1Owt7cXfvrpJ80x586dEwAI8fHxpgrTYA8ePBAaN24sxMbGCt26dROmTZsmCIL13Ofs2bOFLl266N2vUqkEPz8/Yfny5ZptmZmZgqOjo/Ddd99VR4iy6devn/DCCy9obRs8eLAwcuRIQRCs414BCFu3btV8LeaekpKSBADC8ePHNcf89ttvgkKhEG7dulVtsUtR9j51OXbsmABAuHbtmiAI1nWfN2/eFOrWrSskJiYKDRo0ED766CPNPmu5z+eee04YNWqU3tfI9TeYLUkmUlBQgJMnTyIyMlKzzcbGBpGRkYiPjzdhZPLKysoCAHh6egIATp48icLCQq37DgkJQWBgoEXe95QpU9CvXz+t+wGs5z5/+eUXtG3bFkOHDoWPjw/CwsKwfv16zf6UlBSkpaVp3adSqUSHDh0s6j4BoFOnTtizZw8uXrwIAPjrr79w6NAh9OnTB4B13auamHuKj4+Hh4cH2rZtqzkmMjISNjY2OHr0aLXHLJesrCwoFAp4eHgAsJ77VKlUGD16NGbNmoXHH3+83H5ruE+VSoVff/0VTZo0Qe/eveHj44MOHTpodcnJ9TeYSZKJ3L17F8XFxfD19dXa7uvri7S0NBNFJS+VSoXp06ejc+fOaN68OQAgLS0NDg4Omj9MapZ4399//z1OnTqF6Ojocvus5T6vXLmCNWvWoHHjxti9ezcmT56MV199FV9//TUAaO7FGn6O33zzTQwfPhwhISGwt7dHWFgYpk+fjpEjRwKwrntVE3NPaWlp8PHx0dpvZ2cHT09Pi73vvLw8zJ49GyNGjNAs/Got97ls2TLY2dnh1Vdf1bnfGu4zPT0dOTk5WLp0KaKiovD7779j0KBBGDx4MPbv3w9Avr/BdnIGTlTalClTkJiYiEOHDpk6FNnduHED06ZNQ2xsrFYfuLVRqVRo27Yt3nvvPQBAWFgYEhMTsXbtWowdO9bE0cnrxx9/xKZNm7B582Y8/vjjSEhIwPTp0xEQEGB191qTFRYWYtiwYRAEAWvWrDF1OLI6efIkVq5ciVOnTkGhUJg6HKNRqVQAgIEDB+K1114DALRu3Rp//vkn1q5di27dusl2LbYkmYi3tzdsbW3LjbS/c+cO/Pz8TBSVfKZOnYqdO3di7969qFevnma7n58fCgoKkJmZqXW8pd33yZMnkZ6ejjZt2sDOzg52dnbYv38/Vq1aBTs7O/j6+lrFffr7+yM0NFRrW7NmzTQzSNT3Yg0/x7NmzdK0JrVo0QKjR4/Ga6+9pmkptKZ7VRNzT35+fuUmkxQVFSEjI8Pi7ludIF27dg2xsbGaViTAOu7z4MGDSE9PR2BgoObv0rVr1zBz5kwEBQUBsI779Pb2hp2dXaV/m+T4G8wkyUQcHBwQHh6OPXv2aLapVCrs2bMHERERJoysagRBwNSpU7F161bExcUhODhYa394eDjs7e217vvChQu4fv26Rd33U089hbNnzyIhIUHz0bZtW4wcOVLzuTXcZ+fOncuVcLh48SIaNGgAAAgODoafn5/WfWZnZ+Po0aMWdZ9AyQwoGxvtP4m2traa/1qt6V7VxNxTREQEMjMzcfLkSc0xcXFxUKlU6NChQ7XHbCh1gpScnIw//vgDXl5eWvut4T5Hjx6NM2fOaP1dCggIwKxZs7B7924A1nGfDg4OaNeuXYV/m2R71kgcZE4y+v777wVHR0dhw4YNQlJSkjBp0iTBw8NDSEtLM3VoBps8ebKgVCqFffv2CampqZqP3NxczTEvvfSSEBgYKMTFxQknTpwQIiIihIiICBNGLY/Ss9sEwTru89ixY4KdnZ2wZMkSITk5Wdi0aZPg4uIibNy4UXPM0qVLBQ8PD2H79u3CmTNnhIEDBwrBwcHCo0ePTBi5dGPHjhXq1q0r7Ny5U0hJSRG2bNkieHt7C2+88YbmGEu81wcPHginT58WTp8+LQAQVqxYIZw+fVozq0vMPUVFRQlhYWHC0aNHhUOHDgmNGzcWRowYYapb0qmi+ywoKBAGDBgg1KtXT0hISND625Sfn685h6Xfpy5lZ7cJgnXc55YtWwR7e3vh888/F5KTk4XVq1cLtra2wsGDBzXnkONvMJMkE1u9erUQGBgoODg4CO3btxeOHDli6pCqBIDOj6+++kpzzKNHj4SXX35ZqF27tuDi4iIMGjRISE1NNV3QMimbJFnLfe7YsUNo3ry54OjoKISEhAiff/651n6VSiXMnTtX8PX1FRwdHYWnnnpKuHDhgomiNVx2drYwbdo0ITAwUHBychIee+wx4e2339Z6iFrive7du1fn7+TYsWMFQRB3T/fu3RNGjBgh1KpVS3B3dxfGjx8vPHjwwAR3o19F95mSkqL3b9PevXs157D0+9RFV5JkLff55ZdfCo0aNRKcnJyEVq1aCdu2bdM6hxx/gxWCUKqcLBEREREB4JgkIiIiIp2YJBERERHpwCSJiIiISAcmSUREREQ6MEkiIiIi0oFJEhEREZEOTJKIiIiIdGCSRGSmgoKC8PHHH8t2vnHjxuGZZ56R7XwAsG/fPigUinLrI5F5unfvHnx8fHD16tVqv3ZSUhLq1auHhw8fVvu1iQzFJInIyMaNGweFQgGFQgEHBwc0atQIixYtQlFRUYWvO378OCZNmiRbHCtXrsSGDRtkO58Up0+fxtChQ+Hr6wsnJyc0btwYEydOxMWLF00Sj7kSmxh//vnnePLJJ+Hu7i4pSV2yZAkGDhyoWez06tWrmp9NhUIBLy8v9OrVC6dPn9Z63aVLlzB+/HjUq1cPjo6OCA4OxogRI3DixAmtc3fq1AkuLi7w8PAod+3Q0FB07NgRK1asEBUrkTlgkkRUDaKiopCamork5GTMnDkTCxYswPLly3UeW1BQAACoU6cOXFxcZItBqVTqfHgZ286dO9GxY0fk5+dj06ZNOHfuHDZu3AilUom5c+dWezzWIDc3F1FRUXjrrbckvebLL7/EhAkTyu37448/kJqait27dyMnJwd9+vTRJF4nTpxAeHg4Ll68iHXr1iEpKQlbt25FSEgIZs6cqTlHQUEBhg4dismTJ+uNYfz48VizZk2l/yAQmY2qr7BCRBUZO3asMHDgQK1tPXv2FDp27Ki1/9133xX8/f2FoKAgQRDKr7kEQFi/fr3wzDPPCM7OzkKjRo2E7du3a503MTFR6Nevn+Dm5ibUqlVL6NKli3Dp0iWdcXTr1k2YMmWKMGXKFMHd3V3w8vIS3nnnHUGlUmmO+eabb4Tw8HChVq1agq+vrzBixAjhzp07mv3q9ZXu37+v894fPnwoeHt7C88884zO/aVft2/fPqFdu3aCg4OD4OfnJ8yePVsoLCzUinfq1KnCtGnTBA8PD8HHx0f4/PPPhZycHGHcuHFCrVq1hIYNGwq7du0qF9/OnTuFFi1aCI6OjkKHDh2Es2fPasXx888/C6GhoYKDg4PQoEED4YMPPtDa36BBA2HJkiXC+PHjhVq1agn169cX1q1bp3XM9evXhaFDhwpKpVKoXbu2MGDAACElJUWzX/3+L1++XPDz8xM8PT2Fl19+WSgoKNDcH8qsU1WZyt7/0n766SehTp06WtvUa5qdPn1as+3w4cMCACEmJkZQqVTC448/LoSHhwvFxcXlzqnrul999ZWgVCp1xpCfny84OjoKf/zxR6XxEpkDtiQRmYCzs7OmxQgA9uzZgwsXLiA2NhY7d+7U+7qFCxdi2LBhOHPmDPr27YuRI0ciIyMDAHDr1i088cQTcHR0RFxcHE6ePIkXXnihwv/av/76a9jZ2eHYsWNYuXIlVqxYgS+++EKzv7CwEIsXL8Zff/2Fbdu24erVqxg3bpzo+9y9ezfu3r2LN954Q+d+dcvWrVu30LdvX7Rr1w5//fUX1qxZgy+//BLvvvtuuXi9vb1x7NgxvPLKK5g8eTKGDh2KTp064dSpU+jVqxdGjx6N3NxcrdfNmjULH374IY4fP446deqgf//+KCwsBACcPHkSw4YNw/Dhw3H27FksWLAAc+fOLdc1+eGHH6Jt27Y4ffo0Xn75ZUyePBkXLlzQvE+9e/eGm5sbDh48iMOHD6NWrVqIiorS+j7v3bsXly9fxt69e/H1119jw4YNmuts2bIF9erVw6JFi5CamorU1FTR77MYBw8eRHh4eKXHOTs7AyhpGUpISMDff/+NmTNnwsam/ONCasukg4MDWrdujYMHD0p6HZHJmDpLI7J2pVtwVCqVEBsbKzg6Ogqvv/66Zr+vr6/WavOCoLsl6Z133tF8nZOTIwAQfvvtN0EQBGHOnDlCcHCwpmWiojgEoaTlolmzZlotR7NnzxaaNWum916OHz8uANCsGF5ZS8ayZcsEAEJGRobecwqCILz11ltC06ZNtWL59NNPhVq1amlaMLp16yZ06dJFs7+oqEhwdXUVRo8erdmWmpoqABDi4+O14vv+++81x9y7d09wdnYWfvjhB0EQBOH5558XevbsqRXPrFmzhNDQUM3XDRo0EEaNGqX5WqVSCT4+PsKaNWsEQRCEb7/9tlz8+fn5grOzs7B7925BEEre/wYNGghFRUWaY4YOHSo899xzWtcpu2J7RaS0JA0cOFB44YUXtLaVbUm6f/++MGjQIKFWrVpCWlqa8MMPPwgAhFOnTomOqaKWJEEQhEGDBgnjxo0TfT4iU2JLElE12LlzJ2rVqgUnJyf06dMHzz33HBYsWKDZ36JFCzg4OFR6npYtW2o+d3V1hbu7O9LT0wEACQkJ6Nq1K+zt7UXH1bFjRygUCs3XERERSE5ORnFxMYCSVpb+/fsjMDAQbm5u6NatGwDg+vXros4vCIKo486dO4eIiAitWDp37oycnBzcvHlTs630/dva2sLLywstWrTQbPP19QUAzXtS+r7UPD090bRpU5w7d05z7c6dO2sd37lzZ633oey1FQoF/Pz8NNf566+/cOnSJbi5uaFWrVqoVasWPD09kZeXh8uXL2te9/jjj8PW1lbztb+/f7lYjeXRo0dwcnLSua9Tp06oVasWateujb/++gs//PADfH19RX//pHB2di7X0kdkruxMHQBRTdC9e3esWbMGDg4OCAgIgJ2d9q+eq6urqPOUTYAUCgVUKhWAf7tJ5PLw4UP07t0bvXv3xqZNm1CnTh1cv34dvXv31upCqkiTJk0AAOfPn9dKVAyl6/5Lb1MnWer3RE4Vvfc5OTkIDw/Hpk2byr2uTp06os5hbN7e3rh//77OfT/88ANCQ0Ph5eWl1YVW+vsXFhYmSxwZGRlo2LChLOciMja2JBFVA1dXVzRq1AiBgYHlEiS5tGzZEgcPHtSMtRHj6NGjWl8fOXIEjRs3hq2tLc6fP4979+5h6dKl6Nq1K0JCQiS3evTq1Qve3t54//33de5Xz6Bq1qwZ4uPjtVouDh8+DDc3N9SrV0/SNXU5cuSI5vP79+/j4sWLaNasmebahw8f1jr+8OHDaNKkiVarT0XatGmD5ORk+Pj4oFGjRlofSqVSdJwODg5arVdyCgsLQ1JSks599evXR8OGDcuNMWrdujVCQ0Px4Ycf6kzmDKmPlZiYKFvCRWRsTJKIrMTUqVORnZ2N4cOH48SJE0hOTsa3336rGVysy/Xr1zFjxgxcuHAB3333HVavXo1p06YBAAIDA+Hg4IDVq1fjypUr+OWXX7B48WJJMbm6uuKLL77Ar7/+igEDBuCPP/7A1atXceLECbzxxht46aWXAAAvv/wybty4gVdeeQXnz5/H9u3bMX/+fMyYMUPngGGpFi1ahD179iAxMRHjxo2Dt7e3prDmzJkzsWfPHixevBgXL17E119/jU8++QSvv/666POPHDkS3t7eGDhwIA4ePIiUlBTs27cPr776qlZ3YWWCgoJw4MAB3Lp1C3fv3tV7XFpaGhISEnDp0iUAwNmzZ5GQkKAZxK9L79698ffff+ttTdJFoVDgq6++wsWLF9G1a1fs2rULV65cwZkzZzQ1l9SuX7+OhIQEXL9+HcXFxUhISEBCQgJycnI0x1y9ehW3bt1CZGSk6BiITIlJEpGV8PLyQlxcHHJyctCtWzeEh4dj/fr1FY5RGjNmDB49eoT27dtjypQpmDZtmqaAZZ06dbBhwwb89NNPCA0NxdKlS/HBBx9IjmvgwIH4888/YW9vj+effx4hISEYMWIEsrKyNLPX6tati127duHYsWNo1aoVXnrpJUyYMAHvvPOOYW9GGUuXLsW0adMQHh6OtLQ07NixQzMGrE2bNvjxxx/x/fffo3nz5pg3bx4WLVokaRafi4sLDhw4gMDAQAwePBjNmjXDhAkTkJeXB3d3d9HnWbRoEa5evYqGDRtqddOVtXbtWoSFhWHixIkAgCeeeAJhYWH45Zdf9L6mRYsWmnuVon379jhx4gQaNWqEiRMnolmzZhgwYAD+/vtvrcKX8+bNQ1hYGObPn4+cnByEhYUhLCxMq+Dkd999h169eqFBgwaSYiAyFYVgjJF5RGT2nnzySbRu3VrWpU/Mzb59+9C9e3fcv3/fJIU0zc2vv/6KWbNmITExUZYWOikKCgrQuHFjbN68udxAeSJzxYHbREQ1RL9+/ZCcnIxbt26hfv361Xrt69ev46233mKCRBaFSRIRUQ0yffp0k1xXPZCdyJKwu42IiIhIBw7cJiIiItKBSRIRERGRDkySiIiIiHRgkkRERESkA5MkIiIiIh2YJBERERHpwCSJiIiISAcmSUREREQ6MEkiIiIi0uH/AY4tACpDJo/MAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Case 5 - PCA, taking no future info as baseline as it achieved max accuracy\n",
    "data = df_no_future.copy()\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaled_data = scaler.fit_transform(data)\n",
    "\n",
    "pca = PCA(n_components=2)\n",
    "principal_components = pca.fit_transform(scaled_data)\n",
    "\n",
    "pc_df = pd.DataFrame(data=principal_components, columns=['PC1', 'PC2'])\n",
    "\n",
    "# Display the explained variance ratio\n",
    "print(\"Explained Variance Ratio:\")\n",
    "print(pca.explained_variance_ratio_)\n",
    "\n",
    "# Plot the results\n",
    "plt.scatter(pc_df['PC1'], pc_df['PC2'])\n",
    "plt.title('PCA Result')\n",
    "plt.xlabel('Principal Component 1 (PC1)')\n",
    "plt.ylabel('Principal Component 2 (PC2)')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average F1 Score: 0.9104697680452611\n",
      "Average Accuracy: 0.9217335971240482\n",
      "Average Precision: 0.9172309768475302\n",
      "Average Recall: 0.9038339492378042\n"
     ]
    }
   ],
   "source": [
    "# Case 6 - using chi-square for feature selection\n",
    "data = df_no_future.copy()\n",
    "scaler = MinMaxScaler()\n",
    "scaled_data = scaler.fit_transform(X_train)\n",
    "\n",
    "k_best = SelectKBest(score_func=chi2, k='all')\n",
    "k_best.fit(scaled_data, y_train)\n",
    "\n",
    "selected_features_indices = k_best.get_support(indices=True)\n",
    "\n",
    "x_train_selected = k_best.transform(scaled_data)\n",
    "\n",
    "accuracies_features.append(train_RF(x_train_selected, y_train))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F-statistic: 270.3659549991617\n",
      "P-value: 1.9056022828871137e-23\n",
      "The means of the accuracy vectors are significantly different.\n"
     ]
    }
   ],
   "source": [
    "# Perform ANOVA test to see if there is significant difference between the accuracies\n",
    "f_statistic, p_value = f_oneway(*accuracies_features)\n",
    "\n",
    "# Print the results\n",
    "print(f'F-statistic: {f_statistic}')\n",
    "print(f'P-value: {p_value}')\n",
    "\n",
    "# Check if the p-value is below the significance level (e.g., 0.05)\n",
    "if p_value < 0.05:\n",
    "    print('The means of the accuracy vectors are significantly different.')\n",
    "else:\n",
    "    print('There is no significant difference in means between the accuracy vectors.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Statistical test for difference between baseline accuracies and reduced features accuracies\n",
      "\n",
      "P-value for baseline and no google embeddings: 0.008015696383196595\n",
      "The difference in accuracies between baseline and no google embeddings is statistically significant.\n",
      "\n",
      "P-value for baseline and no future information: 0.7282614966698906\n",
      "There is no statistically significant difference in accuracies between Pair baseline and no future information.\n",
      "\n",
      "P-value for baseline and No vpm_patent_score: 4.076270341827097e-05\n",
      "The difference in accuracies between baseline and No vpm_patent_score is statistically significant.\n",
      "\n",
      "P-value for baseline and no future info and vppm: 3.6248810150422344e-06\n",
      "The difference in accuracies between baseline and no future info and vppm is statistically significant.\n",
      "\n",
      "P-value for baseline and PCA: 618 features: 4.445888390291907e-05\n",
      "The difference in accuracies between baseline and PCA: 618 features is statistically significant.\n",
      "\n",
      "P-value for baseline and chi-square feature selection: 0.036299385348087417\n",
      "The difference in accuracies between baseline and chi-square feature selection is statistically significant.\n",
      "\n",
      "P-value for no google embeddings and no future information: 0.005871261908203966\n",
      "The difference in accuracies between no google embeddings and no future information is statistically significant.\n",
      "\n",
      "P-value for no google embeddings and No vpm_patent_score: 5.7910345338820684e-05\n",
      "The difference in accuracies between no google embeddings and No vpm_patent_score is statistically significant.\n",
      "\n",
      "P-value for no google embeddings and no future info and vppm: 1.3965043324922337e-05\n",
      "The difference in accuracies between no google embeddings and no future info and vppm is statistically significant.\n",
      "\n",
      "P-value for no google embeddings and PCA: 618 features: 2.4884445037234855e-05\n",
      "The difference in accuracies between no google embeddings and PCA: 618 features is statistically significant.\n",
      "\n",
      "P-value for no google embeddings and chi-square feature selection: 0.0011127901956611284\n",
      "The difference in accuracies between no google embeddings and chi-square feature selection is statistically significant.\n",
      "\n",
      "P-value for no future information and No vpm_patent_score: 8.04113280432112e-05\n",
      "The difference in accuracies between no future information and No vpm_patent_score is statistically significant.\n",
      "\n",
      "P-value for no future information and no future info and vppm: 2.772264181432662e-05\n",
      "The difference in accuracies between no future information and no future info and vppm is statistically significant.\n",
      "\n",
      "P-value for no future information and PCA: 618 features: 4.988667013654389e-05\n",
      "The difference in accuracies between no future information and PCA: 618 features is statistically significant.\n",
      "\n",
      "P-value for no future information and chi-square feature selection: 0.26889852032184736\n",
      "There is no statistically significant difference in accuracies between Pair no future information and chi-square feature selection.\n",
      "\n",
      "P-value for No vpm_patent_score and no future info and vppm: 0.7487317990360418\n",
      "There is no statistically significant difference in accuracies between Pair No vpm_patent_score and no future info and vppm.\n",
      "\n",
      "P-value for No vpm_patent_score and PCA: 618 features: 0.00034440400032392814\n",
      "The difference in accuracies between No vpm_patent_score and PCA: 618 features is statistically significant.\n",
      "\n",
      "P-value for No vpm_patent_score and chi-square feature selection: 4.7096838265974095e-05\n",
      "The difference in accuracies between No vpm_patent_score and chi-square feature selection is statistically significant.\n",
      "\n",
      "P-value for no future info and vppm and PCA: 618 features: 0.0012001392112429108\n",
      "The difference in accuracies between no future info and vppm and PCA: 618 features is statistically significant.\n",
      "\n",
      "P-value for no future info and vppm and chi-square feature selection: 6.832463167426587e-06\n",
      "The difference in accuracies between no future info and vppm and chi-square feature selection is statistically significant.\n",
      "\n",
      "P-value for PCA: 618 features and chi-square feature selection: 3.895057026566266e-05\n",
      "The difference in accuracies between PCA: 618 features and chi-square feature selection is statistically significant.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# pairwise statistical difference of accuracies between different set of features\n",
    "\n",
    "feature_selection_type = ['baseline', 'no google embeddings', 'no future information', 'No vpm_patent_score', 'no future info and vppm', 'PCA: 618 features', 'chi-square feature selection']\n",
    "\n",
    "\n",
    "print('Statistical test for difference between baseline accuracies and reduced features accuracies')\n",
    "print()\n",
    "\n",
    "# Perform paired t-test for each pair of lists\n",
    "num_lists = len(accuracies_features)\n",
    "\n",
    "for i in range(num_lists - 1):\n",
    "    for j in range(i + 1, num_lists):\n",
    "        acc_list_1 = accuracies_features[i]\n",
    "        acc_list_2 = accuracies_features[j]\n",
    "\n",
    "        # Perform paired t-test\n",
    "        t_statistic, p_value = ttest_rel(acc_list_1, acc_list_2)\n",
    "\n",
    "        # Display the results\n",
    "        # print(f\"T-statistic for Pair {i+1}-{j+1}: {t_statistic}\")\n",
    "        print(f\"P-value for {feature_selection_type[i]} and {feature_selection_type[j]}: {p_value}\")\n",
    "\n",
    "        # Check for statistical significance (common significance level is 0.05)\n",
    "        if p_value < 0.05:\n",
    "            print(f\"The difference in accuracies between {feature_selection_type[i]} and {feature_selection_type[j]} is statistically significant.\")\n",
    "        else:\n",
    "            print(f\"There is no statistically significant difference in accuracies between Pair {feature_selection_type[i]} and {feature_selection_type[j]}.\")\n",
    "\n",
    "        print()\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "baseline 0.9208243241513442\n",
      "no google embeddings 0.9172339664507978\n",
      "no future information 0.9210807969087165\n",
      "No vpm_patent_score 0.8913316583563278\n",
      "no future info and vppm 0.8909819590969301\n",
      "PCA: 618 features 0.8753847709645551\n",
      "chi-square feature selection 0.9217335971240482\n"
     ]
    }
   ],
   "source": [
    "for name, accuracy in zip(feature_selection_type, accuracies_features): \n",
    "    print(name, np.mean(accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the following analysis we won't consider \"chi-square feature selection\" as that didn't drop any feature. <br>\n",
    "Some differences between accuraccies are statistically, but not all of them. The interesting thing to notice is that the regularization that comes from dropping a certain subset of features doesn't lead to an improvement in performances.<br>\n",
    "Removing google embedings and future information didn't lead to any statistical significant reduction in the accuracy which means that these information aren't of central importance for predicting commercialization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9234427452443118\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAJbCAYAAAAMiytuAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAACRXElEQVR4nOzdd1xT1/8/8FdANjIcgCgKFBwICIoDFw6sVutuHXXi6FJEse5VJ+rHgVWrdVetq2rd4sCtOAEnTlSoiqBWqKKicH5/+CNfUxIgJCGEvJ6PRx6ae+/73vclIbxz7rnnSIQQAkRERER6xEDbCRAREREVNhZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd4poe0EiqKsrCw8fvwYJUuWhEQi0XY6RERElA9CCPz7779wdHSEgUHubTwsgOR4/PgxnJyctJ0GERERFUBiYiIqVKiQ6zYsgOQoWbIkgI8/QCsrKy1nQ0RERPmRlpYGJycn6d/x3LAAkiP7speVlRULICIiIh2Tn+4r7ARNREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd0poOwF95Dx6r1LbP5jZRkOZEBER6Se2ABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R2tF0CLFy+Gs7MzTE1NUbduXZw/f17httevX0fnzp3h7OwMiUSC8PBwuds9evQIPXv2ROnSpWFmZgYvLy9cvHhRQ2dAREREukarBdDmzZsRGhqKSZMmITo6GjVq1EDLli2RnJwsd/v09HS4urpi5syZcHBwkLvNP//8gwYNGsDIyAj79+/HjRs3MHfuXNja2mryVIiIiEiHaHUk6Hnz5mHgwIEICgoCACxduhR79+7FqlWrMHr06Bzb165dG7Vr1wYAuesBYNasWXBycsLq1auly1xcXDSQPREREekqrbUAZWRk4NKlSwgMDPy/ZAwMEBgYiKioqALvd9euXfDz88PXX38NOzs7+Pr6Yvny5bnGvHv3DmlpaTIPIiIiKr60VgA9e/YMmZmZsLe3l1lub2+PpKSkAu83Pj4eS5Ysgbu7Ow4cOIAffvgBQ4YMwe+//64wJiwsDNbW1tKHk5NTgY9PRERERZ/WO0GrW1ZWFmrWrIkZM2bA19cX3377LQYOHIilS5cqjBkzZgxSU1Olj8TExELMmIiIiAqb1gqgMmXKwNDQEE+fPpVZ/vTpU4UdnPOjXLly8PDwkFlWrVo1JCQkKIwxMTGBlZWVzIOIiIiKL60VQMbGxqhVqxYiIyOly7KyshAZGQl/f/8C77dBgwa4deuWzLLbt2+jUqVKBd4nERERFS9avQssNDQUffr0gZ+fH+rUqYPw8HC8fv1aeldY7969Ub58eYSFhQH42HH6xo0b0v8/evQIsbGxsLS0hJubGwBg2LBhqF+/PmbMmIEuXbrg/PnzWLZsGZYtW6adkyQiIqIiR6sFUNeuXZGSkoKJEyciKSkJPj4+iIiIkHaMTkhIgIHB/zVSPX78GL6+vtLnc+bMwZw5cxAQEIBjx44B+Hir/F9//YUxY8ZgypQpcHFxQXh4OHr06FGo50ZERERFl0QIIbSdRFGTlpYGa2trpKamaqQ/kPPovUpt/2BmG7XnQEREVNwo8/e72N0FRkRERJQXFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3ikSBdDixYvh7OwMU1NT1K1bF+fPn1e47fXr19G5c2c4OztDIpEgPDw8133PnDkTEokEQ4cOVW/SREREpLO0XgBt3rwZoaGhmDRpEqKjo1GjRg20bNkSycnJcrdPT0+Hq6srZs6cCQcHh1z3feHCBfz222/w9vbWROpERESko7ReAM2bNw8DBw5EUFAQPDw8sHTpUpibm2PVqlVyt69duzb+97//oVu3bjAxMVG431evXqFHjx5Yvnw5bG1tNZU+ERER6SCtFkAZGRm4dOkSAgMDpcsMDAwQGBiIqKgolfY9aNAgtGnTRmbfirx79w5paWkyDyIiIiq+tFoAPXv2DJmZmbC3t5dZbm9vj6SkpALvd9OmTYiOjkZYWFi+tg8LC4O1tbX04eTkVOBjExERUdGn9Utg6paYmIiQkBD88ccfMDU1zVfMmDFjkJqaKn0kJiZqOEsiIiLSphLaPHiZMmVgaGiIp0+fyix/+vRpnh2cFbl06RKSk5NRs2ZN6bLMzEycOHECixYtwrt372BoaCgTY2Jikmt/IiIiIipetNoCZGxsjFq1aiEyMlK6LCsrC5GRkfD39y/QPps3b46rV68iNjZW+vDz80OPHj0QGxubo/ghIiIi/aPVFiAACA0NRZ8+feDn54c6deogPDwcr1+/RlBQEACgd+/eKF++vLQ/T0ZGBm7cuCH9/6NHjxAbGwtLS0u4ubmhZMmS8PT0lDmGhYUFSpcunWM5ERER6SetF0Bdu3ZFSkoKJk6ciKSkJPj4+CAiIkLaMTohIQEGBv/XUPX48WP4+vpKn8+ZMwdz5sxBQEAAjh07VtjpExERkQ6SCCGEtpMoatLS0mBtbY3U1FRYWVmpff/Oo/cqtf2DmW3UngMREVFxo8zf72J3FxgRERFRXlgAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpnSJRAC1evBjOzs4wNTVF3bp1cf78eYXbXr9+HZ07d4azszMkEgnCw8NzbBMWFobatWujZMmSsLOzQ4cOHXDr1i0NngERERHpEq0XQJs3b0ZoaCgmTZqE6Oho1KhRAy1btkRycrLc7dPT0+Hq6oqZM2fCwcFB7jbHjx/HoEGDcPbsWRw6dAjv37/H559/jtevX2vyVIiIiEhHSIQQQpsJ1K1bF7Vr18aiRYsAAFlZWXByckJwcDBGjx6da6yzszOGDh2KoUOH5rpdSkoK7OzscPz4cTRu3DjPnNLS0mBtbY3U1FRYWVnl+1zyy3n0XqW2fzCzjdpzICIiKm6U+ftdoBagkydPomfPnvD398ejR48AAOvWrcOpU6eU2k9GRgYuXbqEwMDA/0vIwACBgYGIiooqSGpypaamAgBKlSold/27d++QlpYm8yAiIqLiS+kCaNu2bWjZsiXMzMwQExODd+/eAfhYZMyYMUOpfT179gyZmZmwt7eXWW5vb4+kpCRlU5MrKysLQ4cORYMGDeDp6Sl3m7CwMFhbW0sfTk5Oajk2ERERFU1KF0DTpk3D0qVLsXz5chgZGUmXN2jQANHR0WpNTh0GDRqEa9euYdOmTQq3GTNmDFJTU6WPxMTEQsyQiIiIClsJZQNu3boltx+NtbU1Xr58qdS+ypQpA0NDQzx9+lRm+dOnTxV2cFbG4MGDsWfPHpw4cQIVKlRQuJ2JiQlMTExUPh4RERHpBqVbgBwcHHD37t0cy0+dOgVXV1el9mVsbIxatWohMjJSuiwrKwuRkZHw9/dXNjUpIQQGDx6Mv/76C0eOHIGLi0uB90VERETFj9ItQAMHDkRISAhWrVoFiUSCx48fIyoqCj/99BMmTJigdAKhoaHo06cP/Pz8UKdOHYSHh+P169cICgoCAPTu3Rvly5dHWFgYgI8dp2/cuCH9/6NHjxAbGwtLS0u4ubkB+HjZa8OGDdi5cydKliwp7U9kbW0NMzMzpXMkIiKi4kXpAmj06NHIyspC8+bNkZ6ejsaNG8PExAQ//fQTgoODlU6ga9euSElJwcSJE5GUlAQfHx9ERERIO0YnJCTAwOD/GqoeP34MX19f6fM5c+Zgzpw5CAgIwLFjxwAAS5YsAQA0adJE5lirV69G3759lc6RiIiIipcCjwOUkZGBu3fv4tWrV/Dw8IClpaW6c9MajgNERESke5T5+610C1BqaioyMzNRqlQpeHh4SJe/ePECJUqU0EjBQERERKROSneC7tatm9xbyrds2YJu3bqpJSkiIiIiTVK6ADp37hyaNm2aY3mTJk1w7tw5tSRFREREpElKF0Dv3r3Dhw8fcix///493rx5o5akiIiIiDRJ6QKoTp06WLZsWY7lS5cuRa1atdSSFBEREZEmKd0Jetq0aQgMDMTly5fRvHlzAEBkZCQuXLiAgwcPqj1BIiIiInVTugWoQYMGiIqKgpOTE7Zs2YLdu3fDzc0NV65cQaNGjTSRIxEREZFaKd0CBAA+Pj74448/1J0LERERUaEoUAGUlZWFu3fvIjk5GVlZWTLr5E2USkRERFSUKF0AnT17Ft988w0ePnyI/w4iLZFIkJmZqbbkiIiIiDRB6QLo+++/h5+fH/bu3Yty5cpBIpFoIi8iIiIijVG6ALpz5w62bt0qnXmdiIiISNcofRdY3bp1cffuXU3kQkRERFQolG4BCg4OxvDhw5GUlAQvLy8YGRnJrPf29lZbckRERESaoHQB1LlzZwBAv379pMskEgmEEOwETURERDpB6QLo/v37msiDiIiIqNAoXQBVqlRJE3kQERERFZoCDYQIADdu3EBCQgIyMjJklrdr107lpIiIiIg0SekCKD4+Hh07dsTVq1elfX8ASMcDYh8gIiIiKuqUvg0+JCQELi4uSE5Ohrm5Oa5fv44TJ07Az88Px44d00CKREREROqldAtQVFQUjhw5gjJlysDAwAAGBgZo2LAhwsLCMGTIEMTExGgiTyIiIiK1UboFKDMzEyVLlgQAlClTBo8fPwbwsXP0rVu31JsdERERkQYo3QLk6emJy5cvw8XFBXXr1sXs2bNhbGyMZcuWwdXVVRM5EhEREamV0gXQ+PHj8fr1awDAlClT8OWXX6JRo0YoXbo0Nm3apPYEiYiIiNRN6QKoZcuW0v+7ubnh5s2bePHiBWxtbTkzPBEREekEpfsA9evXD//++6/MslKlSiE9PV1megwiIiKiokrpAuj333/Hmzdvcix/8+YN1q5dq5akiIiIiDQp35fA0tLSIISAEAL//vsvTE1NpesyMzOxb98+2NnZaSRJIiIiInXKdwFkY2MDiUQCiUSCypUr51gvkUgwefJktSZHREREpAn5LoCOHj0KIQSaNWuGbdu2oVSpUtJ1xsbGqFSpEhwdHTWSJBEREZE65bsACggIwIcPH9CnTx/4+fnByclJk3kRERERaYxSnaBLlCiBrVu3qn3C08WLF8PZ2RmmpqaoW7cuzp8/r3Db69evo3PnznB2doZEIkF4eLjK+yQiIiL9ovRdYM2aNcPx48fVlsDmzZsRGhqKSZMmITo6GjVq1EDLli2RnJwsd/v09HS4urpi5syZcHBwUMs+iYiISL9IhBBCmYClS5di8uTJ6NGjB2rVqgULCwuZ9e3atVMqgbp166J27dpYtGgRACArKwtOTk4IDg7G6NGjc411dnbG0KFDMXToULXtE/h4x5u1tTVSU1NhZWWl1Pnkh/PovUpt/2BmG7XnQEREVNwo8/db6ZGgf/zxRwDAvHnzcqyTSCRKXR7LyMjApUuXMGbMGOkyAwMDBAYGIioqStnUCrzPd+/e4d27d9LnaWlpBTo2ERER6QalL4FlZWUpfCjbN+jZs2fIzMyEvb29zHJ7e3skJSUpm1qB9xkWFgZra2vpgx28iYiIijelC6DiaMyYMUhNTZU+EhMTtZ0SERERaVCBCqDjx4+jbdu2cHNzg5ubG9q1a4eTJ08qvZ8yZcrA0NAQT58+lVn+9OlThR2cNbFPExMTWFlZyTyIiIio+FK6AFq/fj0CAwNhbm6OIUOGYMiQITAzM0Pz5s2xYcMGpfZlbGyMWrVqITIyUrosKysLkZGR8Pf3VzY1je2TiIiIihelO0FPnz4ds2fPxrBhw6TLhgwZgnnz5mHq1Kn45ptvlNpfaGiodHDFOnXqIDw8HK9fv0ZQUBAAoHfv3ihfvjzCwsIAfOzkfOPGDen/Hz16hNjYWFhaWsLNzS1f+yQiIiL9pnQBFB8fj7Zt2+ZY3q5dO4wdO1bpBLp27YqUlBRMnDgRSUlJ8PHxQUREhLQTc0JCAgwM/q+h6vHjx/D19ZU+nzNnDubMmYOAgAAcO3YsX/skIiIi/aZ0AeTk5ITIyEhpa0u2w4cPF/juqcGDB2Pw4MFy12UXNdmcnZ2Rn6GLctsnERER6TelC6Dhw4djyJAhiI2NRf369QEAp0+fxpo1a7BgwQK1J0hERESkbkoXQD/88AMcHBwwd+5cbNmyBQBQrVo1bN68Ge3bt1d7gkRERETqpnQBBAAdO3ZEx44d1Z0LERERUaEoUAEEABcvXkRcXBwAwMPDA7Vq1VJbUkRERESapHQB9Pfff6N79+44ffo0bGxsAAAvX75E/fr1sWnTJlSoUEHdORIRERGpldIDIQ4YMADv379HXFwcXrx4gRcvXiAuLg5ZWVkYMGCAJnIkIiIiUiulW4COHz+OM2fOoEqVKtJlVapUwcKFC9GoUSO1JkdERESkCUq3ADk5OeH9+/c5lmdmZsLR0VEtSRERERFpktIF0P/+9z8EBwfj4sWL0mUXL15ESEgI5syZo9bkiIiIiDRBIvIzrPInbG1tkZ6ejg8fPqBEiY9X0LL/b2FhIbPtixcv1JdpIUpLS4O1tTVSU1M1MjO88+i9Sm3/YGYbtedARERU3Cjz91vpPkDh4eEFzYuIiIioSFC6AOrTp48m8iAiIiIqNAUeCDE5ORnJycnIysqSWe7t7a1yUkRERESapHQBdOnSJfTp0wdxcXE5ZmWXSCTIzMxUW3JEREREmqB0AdSvXz9UrlwZK1euhL29PSQSiSbyIiIiItIYpQug+Ph4bNu2DW5ubprIh4iIiEjjlB4HqHnz5rh8+bImciEiIiIqFEq3AK1YsQJ9+vTBtWvX4OnpCSMjI5n17dq1U1tyRERERJqgdAEUFRWF06dPY//+/TnWsRM0ERER6QKlL4EFBwejZ8+eePLkCbKysmQeLH6IiIhIFyhdAD1//hzDhg2Dvb29JvIhIiIi0jilC6BOnTrh6NGjmsiFiIiIqFAo3QeocuXKGDNmDE6dOgUvL68cnaCHDBmituSIiIiINEHp2eBdXFwU70wiQXx8vMpJaRtngyciItI9Gp0N/v79+wVOjIiIiKgoULoPEBEREZGuy1cLUGhoKKZOnQoLCwuEhobmuu28efPUkhgRERGRpuSrAIqJicH79++l/1eEE6MSERGRLshXAfTpbe+8BZ6IiIh0HfsAERERkd5hAURERER6hwUQERER6Z0iUQAtXrwYzs7OMDU1Rd26dXH+/Plct//zzz9RtWpVmJqawsvLC/v27ZNZ/+rVKwwePBgVKlSAmZkZPDw8sHTpUk2eAhEREekQrRdAmzdvRmhoKCZNmoTo6GjUqFEDLVu2RHJystztz5w5g+7du6N///6IiYlBhw4d0KFDB1y7dk26TWhoKCIiIrB+/XrExcVh6NChGDx4MHbt2lVYp0VERERFmNIF0O+//469e/9vKoeRI0fCxsYG9evXx8OHD5VOYN68eRg4cCCCgoKkLTXm5uZYtWqV3O0XLFiAVq1aYcSIEahWrRqmTp2KmjVrYtGiRdJtzpw5gz59+qBJkyZwdnbGt99+ixo1auTZskRERET6QekCaMaMGTAzMwMAREVFYfHixZg9ezbKlCmDYcOGKbWvjIwMXLp0CYGBgf+XkIEBAgMDERUVJTcmKipKZnsAaNmypcz29evXx65du/Do0SMIIXD06FHcvn0bn3/+udx9vnv3DmlpaTIPIiIiKr6UngssMTERbm5uAIAdO3agc+fO+Pbbb9GgQQM0adJEqX09e/YMmZmZsLe3l1lub2+Pmzdvyo1JSkqSu31SUpL0+cKFC/Htt9+iQoUKKFGiBAwMDLB8+XI0btxY7j7DwsIwefJkpXInIiIi3aV0C5ClpSWeP38OADh48CBatGgBADA1NcWbN2/Um10BLVy4EGfPnsWuXbtw6dIlzJ07F4MGDcLhw4flbj9mzBikpqZKH4mJiYWcMRERERUmpVuAWrRogQEDBsDX1xe3b99G69atAQDXr1+Hs7OzUvsqU6YMDA0N8fTpU5nlT58+hYODg9wYBweHXLd/8+YNxo4di7/++gtt2rQBAHh7eyM2NhZz5szJcfkMAExMTGBiYqJU7kRERKS7lG4BWrx4Mfz9/ZGSkoJt27ahdOnSAIBLly6he/fuSu3L2NgYtWrVQmRkpHRZVlYWIiMj4e/vLzfG399fZnsAOHTokHT79+/f4/379zAwkD01Q0NDZGVlKZUfERERFU9KtwDZ2NjI3HGVraB9aEJDQ9GnTx/4+fmhTp06CA8Px+vXrxEUFAQA6N27N8qXL4+wsDAAQEhICAICAjB37ly0adMGmzZtwsWLF7Fs2TIAgJWVFQICAjBixAiYmZmhUqVKOH78ONauXcuZ6omIiAiAEgVQQkKCzPOKFSuqJYGuXbsiJSUFEydORFJSEnx8fBARESHt6JyQkCDTmlO/fn1s2LAB48ePx9ixY+Hu7o4dO3bA09NTus2mTZswZswY9OjRAy9evEClSpUwffp0fP/992rJmYiIiHSbRAgh8rOhgYEBJBIJhBCQSCTIzMzUdG5ak5aWBmtra6SmpsLKykrt+3cevTfvjT7xYGYbtedARERU3Cjz9zvfLUDsP0NERETFhdanwiAiIiIqbAUqgNatW4cGDRrA0dFROv1FeHg4du7cqdbkiIiIiDRB6QJoyZIlCA0NRevWrfHy5UtpXyAbGxuEh4erOz8iIiIitVO6AFq4cCGWL1+OcePGwdDQULrcz88PV69eVWtyRERERJqgdAF0//59+Pr65lhuYmKC169fqyUpIiIiIk1SugBycXFBbGxsjuURERGoVq2aOnIiIiIi0iilR4IODQ3FoEGD8PbtWwghcP78eWzcuBFhYWFYsWKFJnIkIiIiUiulC6ABAwbAzMwM48ePR3p6Or755hs4OjpiwYIF6NatmyZyJCIiIlIrpQqgDx8+YMOGDWjZsiV69OiB9PR0vHr1CnZ2dprKj4iIiEjtlOoDVKJECXz//fd4+/YtAMDc3JzFDxEREekcpTtB16lTBzExMZrIhYiIiKhQKN0H6Mcff8Tw4cPx999/o1atWrCwsJBZ7+3trbbkiIiIiDRB6QIou6PzkCFDpMv0ZZZ4IiIiKh6ULoDu37+viTyIiIiICo3SBVClSpU0kQcRERFRoVG6AFq7dm2u63v37l3gZIiIiIgKg9IFUEhIiMzz9+/fIz09HcbGxjA3N2cBREREREWe0rfB//PPPzKPV69e4datW2jYsCE2btyoiRyJiIiI1ErpAkged3d3zJw5M0frEBEREVFRpJYCCPg4SvTjx4/VtTsiIiIijVG6D9CuXbtkngsh8OTJEyxatAgNGjRQW2JEREREmqJ0AdShQweZ5xKJBGXLlkWzZs0wd+5cdeVFREREpDFKF0BZWVmayIOIiIio0CjdB2jKlClIT0/PsfzNmzeYMmWKWpIiIiIi0iSlC6DJkyfj1atXOZanp6dj8uTJakmKiIiISJOULoCyJz39r8uXL6NUqVJqSYqIiIhIk/LdB8jW1hYSiQQSiQSVK1eWKYIyMzPx6tUrfP/99xpJkoiIiEid8l0AhYeHQwiBfv36YfLkybC2tpauMzY2hrOzM/z9/TWSJBEREZE65bsA6tOnDwDAxcUF9evXh5GRkcaSIiIiItIkpW+DDwgIkP7/7du3yMjIkFlvZWWlelZEREREGqR0J+j09HQMHjwYdnZ2sLCwgK2trcyjIBYvXgxnZ2eYmpqibt26OH/+fK7b//nnn6hatSpMTU3h5eWFffv25dgmLi4O7dq1g7W1NSwsLFC7dm0kJCQUKD8iIiIqXpQugEaMGIEjR45gyZIlMDExwYoVKzB58mQ4Ojpi7dq1SiewefNmhIaGYtKkSYiOjkaNGjXQsmVLJCcny93+zJkz6N69O/r374+YmBh06NABHTp0wLVr16Tb3Lt3Dw0bNkTVqlVx7NgxXLlyBRMmTICpqanS+REREVHxIxFCCGUCKlasiLVr16JJkyawsrJCdHQ03NzcsG7dOmzcuFFua0xu6tati9q1a2PRokUAPo407eTkhODgYIwePTrH9l27dsXr16+xZ88e6bJ69erBx8cHS5cuBQB069YNRkZGWLdunVK5ZEtLS4O1tTVSU1M1cknPefRepbZ/MLON2nMgIiIqbpT5+610C9CLFy/g6uoK4GN/nxcvXgAAGjZsiBMnTii1r4yMDFy6dAmBgYH/l5CBAQIDAxEVFSU3JioqSmZ7AGjZsqV0+6ysLOzduxeVK1dGy5YtYWdnh7p162LHjh0K83j37h3S0tJkHkRERFR8KV0Aubq64v79+wCAqlWrYsuWLQCA3bt3w8bGRql9PXv2DJmZmbC3t5dZbm9vj6SkJLkxSUlJuW6fnJyMV69eYebMmWjVqhUOHjyIjh07olOnTjh+/LjcfYaFhcHa2lr6cHJyUuo8iIiISLcoXQAFBQXh8uXLAIDRo0dj8eLFMDU1xbBhwzBixAi1J6is7Mla27dvj2HDhsHHxwejR4/Gl19+Kb1E9l9jxoxBamqq9JGYmFiYKRMREVEhU/o2+GHDhkn/HxgYiJs3b+LSpUtwc3ODt7e3UvsqU6YMDA0N8fTpU5nlT58+hYODg9wYBweHXLcvU6YMSpQoAQ8PD5ltqlWrhlOnTsndp4mJCUxMTJTKnYiIiHSX0i1An3r79i0qVaqETp06KV38AB9HkK5VqxYiIyOly7KyshAZGalwVGl/f3+Z7QHg0KFD0u2NjY1Ru3Zt3Lp1S2ab27dvo1KlSkrnSERERMWP0gVQZmYmpk6divLly8PS0hLx8fEAgAkTJmDlypVKJxAaGorly5fj999/R1xcHH744Qe8fv0aQUFBAIDevXtjzJgx0u1DQkIQERGBuXPn4ubNm/j5559x8eJFDB48WLrNiBEjsHnzZixfvhx3797FokWLsHv3bvz4449K50dERETFj9IF0PTp07FmzRrMnj0bxsbG0uWenp5YsWKF0gl07doVc+bMwcSJE+Hj44PY2FhERERIOzonJCTgyZMn0u3r16+PDRs2YNmyZahRowa2bt2KHTt2wNPTU7pNx44dsXTpUsyePRteXl5YsWIFtm3bhoYNGyqdHxERERU/So8D5Obmht9++w3NmzdHyZIlcfnyZbi6uuLmzZvw9/fHP//8o6lcCw3HASIiItI9Gh0H6NGjR3Bzc8uxPCsrC+/fv1d2d0RERESFTukCyMPDAydPnsyxfOvWrfD19VVLUkRERESapPRt8BMnTkSfPn3w6NEjZGVlYfv27bh16xbWrl0rMz0FERERUVGldAtQ+/btsXv3bhw+fBgWFhaYOHEi4uLisHv3brRo0UITORIRERGpVb5bgOLj4+Hi4gKJRIJGjRrh0KFDmsyLiIiISGPy3QLk7u6OlJQU6fOuXbvmGJGZiIiISBfkuwD6793y+/btw+vXr9WeEBEREZGmqTQVBhEREZEuyncBJJFIIJFIciwjIiIi0jX57gQthEDfvn2ls6a/ffsW33//PSwsLGS22759u3ozJCIiIlKzfBdAffr0kXnes2dPtSdDREREVBjyXQCtXr1ak3kQERERFRp2giYiIiK9wwKIiIiI9A4LICIiItI7LICIiIhI77AAIiIiIr3DAoiIiIj0DgsgIiIi0jssgIiIiEjvsAAiIiIivcMCiIiIiPQOCyAiIiLSOyyAiIiISO+wACIiIiK9wwKIiIiI9A4LICIiItI7LICIiIhI77AAIiIiIr3DAoiIiIj0DgsgIiIi0jssgIiIiEjvFIkCaPHixXB2doapqSnq1q2L8+fP57r9n3/+iapVq8LU1BReXl7Yt2+fwm2///57SCQShIeHqzlrIiIi0lVaL4A2b96M0NBQTJo0CdHR0ahRowZatmyJ5ORkudufOXMG3bt3R//+/RETE4MOHTqgQ4cOuHbtWo5t//rrL5w9exaOjo6aPg0iIiLSIVovgObNm4eBAwciKCgIHh4eWLp0KczNzbFq1Sq52y9YsACtWrXCiBEjUK1aNUydOhU1a9bEokWLZLZ79OgRgoOD8ccff8DIyKgwToWIiIh0hFYLoIyMDFy6dAmBgYHSZQYGBggMDERUVJTcmKioKJntAaBly5Yy22dlZaFXr14YMWIEqlevnmce7969Q1pamsyDiIiIii+tFkDPnj1DZmYm7O3tZZbb29sjKSlJbkxSUlKe28+aNQslSpTAkCFD8pVHWFgYrK2tpQ8nJyclz4SIiIh0idYvganbpUuXsGDBAqxZswYSiSRfMWPGjEFqaqr0kZiYqOEsiYiISJu0WgCVKVMGhoaGePr0qczyp0+fwsHBQW6Mg4NDrtufPHkSycnJqFixIkqUKIESJUrg4cOHGD58OJydneXu08TEBFZWVjIPIiIiKr60WgAZGxujVq1aiIyMlC7LyspCZGQk/P395cb4+/vLbA8Ahw4dkm7fq1cvXLlyBbGxsdKHo6MjRowYgQMHDmjuZIiIiEhnlNB2AqGhoejTpw/8/PxQp04dhIeH4/Xr1wgKCgIA9O7dG+XLl0dYWBgAICQkBAEBAZg7dy7atGmDTZs24eLFi1i2bBkAoHTp0ihdurTMMYyMjODg4IAqVaoU7skRERFRkaT1Aqhr165ISUnBxIkTkZSUBB8fH0REREg7OickJMDA4P8aqurXr48NGzZg/PjxGDt2LNzd3bFjxw54enpq6xSIiIhIx0iEEELbSRQ1aWlpsLa2Rmpqqkb6AzmP3qvU9g9mtlF7DkRERMWNMn+/i91dYERERER50folMFIOW4+IiIhUxxYgIiIi0jssgIiIiEjvsAAiIiIivcMCiIiIiPQOO0HrEXagJiIi+ogtQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHdYABEREZHeYQFEREREeocFEBEREekdFkBERESkd1gAERERkd5hAURERER6hwUQERER6R0WQERERKR3ikQBtHjxYjg7O8PU1BR169bF+fPnc93+zz//RNWqVWFqagovLy/s27dPuu79+/cYNWoUvLy8YGFhAUdHR/Tu3RuPHz/W9GkQERGRjtB6AbR582aEhoZi0qRJiI6ORo0aNdCyZUskJyfL3f7MmTPo3r07+vfvj5iYGHTo0AEdOnTAtWvXAADp6emIjo7GhAkTEB0dje3bt+PWrVto165dYZ4WERERFWESIYTQZgJ169ZF7dq1sWjRIgBAVlYWnJycEBwcjNGjR+fYvmvXrnj9+jX27NkjXVavXj34+Phg6dKlco9x4cIF1KlTBw8fPkTFihXzzCktLQ3W1tZITU2FlZVVAc9MMefRe5Xa/sHMNlqPJSIiKuqU+fut1RagjIwMXLp0CYGBgdJlBgYGCAwMRFRUlNyYqKgome0BoGXLlgq3B4DU1FRIJBLY2NjIXf/u3TukpaXJPIiIiKj40moB9OzZM2RmZsLe3l5mub29PZKSkuTGJCUlKbX927dvMWrUKHTv3l1hNRgWFgZra2vpw8nJqQBnQ0RERLpC632ANOn9+/fo0qULhBBYsmSJwu3GjBmD1NRU6SMxMbEQsyQiIqLCVkKbBy9TpgwMDQ3x9OlTmeVPnz6Fg4OD3BgHB4d8bZ9d/Dx8+BBHjhzJ9VqgiYkJTExMCngWREREpGu02gJkbGyMWrVqITIyUrosKysLkZGR8Pf3lxvj7+8vsz0AHDp0SGb77OLnzp07OHz4MEqXLq2ZEyAiIiKdpNUWIAAIDQ1Fnz594Ofnhzp16iA8PByvX79GUFAQAKB3794oX748wsLCAAAhISEICAjA3Llz0aZNG2zatAkXL17EsmXLAHwsfr766itER0djz549yMzMlPYPKlWqFIyNjbVzokRERFRkaL0A6tq1K1JSUjBx4kQkJSXBx8cHERER0o7OCQkJMDD4v4aq+vXrY8OGDRg/fjzGjh0Ld3d37NixA56engCAR48eYdeuXQAAHx8fmWMdPXoUTZo0KZTzIiIioqJL6wUQAAwePBiDBw+Wu+7YsWM5ln399df4+uuv5W7v7OwMLQ9tREREREVckSiAqOjjIIpERFScsAAijWPxRERERU2xHgeIiIiISB4WQERERKR3WAARERGR3mEBRERERHqHBRARERHpHRZAREREpHd4GzwVabyFnoiINIEFEBVbyhZPAAsoIiJ9wQKISAFVWp/YckVEVLSxACIqYlg8ERFpHgsgomKExRMRUf7wLjAiIiLSO2wBIiIA2uvzpGpndbZ6EVFBsAAiIr2lK0UfizYi9WMBRESkQ1h4EakHCyAiIsoTL3NSccMCiIiIii1dvMypi7HqiC9svAuMiIiI9A4LICIiItI7LICIiIhI77AAIiIiIr3DAoiIiIj0DgsgIiIi0jssgIiIiEjvsAAiIiIivcMCiIiIiPQOCyAiIiLSOyyAiIiISO+wACIiIiK9UyQKoMWLF8PZ2RmmpqaoW7cuzp8/n+v2f/75J6pWrQpTU1N4eXlh3759MuuFEJg4cSLKlSsHMzMzBAYG4s6dO5o8BSIiItIhWi+ANm/ejNDQUEyaNAnR0dGoUaMGWrZsieTkZLnbnzlzBt27d0f//v0RExODDh06oEOHDrh27Zp0m9mzZ+OXX37B0qVLce7cOVhYWKBly5Z4+/ZtYZ0WERERFWFaL4DmzZuHgQMHIigoCB4eHli6dCnMzc2xatUqudsvWLAArVq1wogRI1CtWjVMnToVNWvWxKJFiwB8bP0JDw/H+PHj0b59e3h7e2Pt2rV4/PgxduzYUYhnRkREREVVCW0ePCMjA5cuXcKYMWOkywwMDBAYGIioqCi5MVFRUQgNDZVZ1rJlS2lxc//+fSQlJSEwMFC63traGnXr1kVUVBS6deuWY5/v3r3Du3fvpM9TU1MBAGlpaQU+t9xkvUtXavtP82Cs5mK1eWzGKkcX82ascnQxb32OVUe8OmTvUwiR98ZCix49eiQAiDNnzsgsHzFihKhTp47cGCMjI7FhwwaZZYsXLxZ2dnZCCCFOnz4tAIjHjx/LbPP111+LLl26yN3npEmTBAA++OCDDz744KMYPBITE/OsQbTaAlRUjBkzRqZVKSsrCy9evEDp0qUhkUgKJYe0tDQ4OTkhMTERVlZWjNVQrDaPzdjCidXmsRmrHF3Mm7HKUTVeWUII/Pvvv3B0dMxzW60WQGXKlIGhoSGePn0qs/zp06dwcHCQG+Pg4JDr9tn/Pn36FOXKlZPZxsfHR+4+TUxMYGJiIrPMxsZGmVNRGysrqwK/SRirG8dmbOHEavPYjNWNYzO2cGLVEa8Ma2vrfG2n1U7QxsbGqFWrFiIjI6XLsrKyEBkZCX9/f7kx/v7+MtsDwKFDh6Tbu7i4wMHBQWabtLQ0nDt3TuE+iYiISL9o/RJYaGgo+vTpAz8/P9SpUwfh4eF4/fo1goKCAAC9e/dG+fLlERYWBgAICQlBQEAA5s6dizZt2mDTpk24ePEili1bBgCQSCQYOnQopk2bBnd3d7i4uGDChAlwdHREhw4dtHWaREREVIRovQDq2rUrUlJSMHHiRCQlJcHHxwcRERGwt7cHACQkJMDA4P8aqurXr48NGzZg/PjxGDt2LNzd3bFjxw54enpKtxk5ciRev36Nb7/9Fi9fvkTDhg0REREBU1PTQj+//DIxMcGkSZNyXIpjrHpjtXlsxhZOrDaPzVjdODZjCydWHfGaJBEiP/eKERERERUfWh8IkYiIiKiwsQAiIiIivcMCiIiIiPQOCyAiIiLSOyyAiIiISO+wACKtKuhNiPfv38eHDx/UnE3Rlt/zff/+Pfr164f79+8X6Bhr167NMdp6Qeni6/T06VMkJSVp5di69rPSN/+dOLs4y8zMxNOnT5GSkqLyvorqzeYsgLRo3bp1aNCgARwdHfHw4UMAQHh4OHbu3Cl3e19fX9SsWTNfj9y8fv26QPkeOnQIkyZNwpEjRwAAJ06cwBdffIFmzZph9erVCuPevXuHn376CY0bN8asWbMAANOmTYOlpSVKliyJb775RulZgatUqYI7d+7kud2TJ0+wfv167Nu3DxkZGTLrXr9+jSlTpuTreI8fP8akSZPQo0cP/PTTT7h582aeMTdu3MCPP/4IX19flCtXDuXKlYOvry9+/PFH3LhxQ2FcREQErl69CuDjyOhTp05F+fLlYWJiggoVKmDmzJm5fqAYGRlh27Zt+Tqv/ypRogS+//57vH37tkDx/5Xf1+lTBflZF8SLFy/w1VdfoWLFivjhhx+QmZmJAQMGoFy5cihfvjzq16+PJ0+e5LqPy5cvo3fv3nB1dYWZmRksLCzg5eWFCRMm5PqeVvU1Luh7KzvnadOm4ddff8WzZ89k1qWlpaFfv365xisSFxcHV1dXhesL+vmhbYcOHULr1q1ha2sLc3NzmJubw9bWFq1bt8bhw4cLvN979+6hWbNmasz0o+TkZJnnsbGx6NOnDxo0aICvvvoKx44dyzV+7969aNy4MSwsLODo6AgHBwfY2NigV69eSEhIUBinic95jctzulTSiF9//VWUKVNGTJs2TZiZmYl79+4JIYRYvXq1aNKkidyYn3/+WfoYPXq0sLKyEvXq1RPDhg0Tw4YNE/7+/sLKykqMHj0612NbWFiIoKAgcfLkyXznu27dOlGiRAlRs2ZNYWlpKVavXi1sbGzEgAEDRL9+/YSxsbH4888/5cYOGzZMODo6iuHDh4tq1aqJH3/8UVSsWFGsX79ebNiwQbi5uYng4GC5sR07dpT7MDAwEIGBgdLn8pw/f17Y2NgIKysrYWZmJtzc3MS1a9ek65OSkoSBgYHcWDMzM5GcnCyEEOL69evC2tpauLm5ia+//lpUrVpVmJubi8uXLyv8ee3bt08YGxuLevXqiUmTJolff/1V/Prrr2LSpEmifv36wsTERERERMiNrVKlijhx4oQQQogZM2aI0qVLi3nz5on9+/eL8PBwYW9vL2bOnKnw2EII0bt3bzFv3rxct1EkICBA7NixQ6kYVV4nVX/Wn/rw4YP4888/xZQpU8SUKVPEn3/+Kd6/fy932379+glPT0+xcOFCERAQINq3by+8vb3FqVOnxJkzZ0Tt2rVF7969FR4rIiJCmJmZic6dO4uePXsKc3NzMXjwYDFq1Cjh5uYmPvvsM/HkyRO5saq8xqq8tw4cOCCMjY1F9erVRcWKFUXp0qXFkSNHpOtz+53IS2xsrMJYVT4/PrV//36Zz61FixaJGjVqiO7du4sXL17kGrtmzRqxZ88e6fMRI0YIa2tr4e/vLx48eKAwpkSJEqJbt25i9erVYt++fWLfvn1i9erVonv37sLIyEisXbs2z7zlye3npUqsgYGBePr0qRBCiNOnTwsjIyMREBAgRowYIVq0aCFKlCghjh8/Ljd27dq1omTJkmL48OFi3LhxwsHBQYwePVosWbJEBAQEiDJlyojbt2/LjVXlc15bWABpSbVq1cRff/0lhBDC0tJSWgBdvXpVlC5dOs/4/v37i/Hjx+dYPnHiRBEUFJRr7F9//SXat28vjIyMhLu7uwgLCxOPHj3KNcbHx0csWLBACCHE4cOHhZmZmcwf2Dlz5ogGDRrIjXVychKHDh0SQghx7949YWBgIPMH9uDBg6JSpUpyYyUSiQgICBB9+/aVeRgYGIgOHTpIn8sTGBgogoKCRGZmpkhLSxM//PCDKF26tIiOjhZC5P5hL5FIpB8i7du3F23btpX+Ic3MzBTdunUTX375paIfl/D29hYTJkxQuH7SpEnCy8tL7joTExPx8OFDIYQQnp6eYsuWLTLr9+zZI9zc3BTuWwghpk6dKmxsbETnzp3FjBkzxIIFC2Qeudm8ebNwdXUVCxcuFGfOnBGXL1+Wecijyuuk6s8627Vr14Srq6swNzcXvr6+wtfXV1hYWAhnZ2dx9erVHNuXK1dOnD59Wgjx8b0gkUjEwYMHpetPnTolypcvr/B4Pj4+YsmSJdLnBw8eFFWrVhVCCJGRkSGaN2+u8JxVeY1VeW/5+/uLsWPHCiGEyMrKErNmzRKWlpZi//79Qojcfyeyv2gpevTs2VNhrCqfH5/y9PQUe/fuFUIIceXKFWFiYiLGjBkj6tWrp/Bnna1y5coiMjJSCCHEmTNnhLm5ufjtt99E27ZtFRbn7u7uYtGiRQr3uXjxYoWv039/5/77GDlypEoFkEQikbvu09+nFi1aiH79+smsDwkJEc2aNZMbW7VqVbFp0ybp8wsXLogKFSqIrKwsIYQQXbt2VfizUuVzXltYAGmJqamp9FvHpwXQ7du3hampaZ7xVlZWcivx27dvCysrq3zlkJycLObOnSu8vLxEiRIlRJs2bcS2bdvkfmO2sLAQ8fHx0udGRkYyfwzj4uIUFm5mZmbSD/vs2E9bYu7fvy/Mzc3lxm7cuFFUqFBBrFq1SmZ5iRIlxPXr13M9P1tbW3Hr1i2ZZWFhYcLW1lacP38+3wWQk5OT9Nt6tujoaFGuXDmFxzY1NRU3b95UuP7mzZsKX+dy5cqJqKgoIYQQ9vb20oIt2+3bt4WZmZnCfQshhLOzs8KHi4tLrrESiSTHw8DAQPqvPKq8Tqr+rLPVq1dPtG3bVqYl4MWLF6Jdu3bC398/x/bm5uYy3/yNjIxkCqX4+HhhYWGh8Himpqbi/v370udZWVnCyMhIPH78WAghxIkTJ0TZsmXlxqryGqvy3rKyshJ3796VWfbHH38ICwsLsXv37lx/JwwMDETNmjVFkyZN5D78/PwUxqry+fHf/WT/zCdNmiQ6d+4shBDi0qVLwt7ePtfYTz+HRo4cKXr16iWE+Fg4lylTRm6MiYlJgX/WEolEODo6Kvw9dHR0VPjzUtSimv1o1qxZvj67Pn2fZcvtfM3MzGTe00J8/B3O/oJ87tw5YWNjozC2oJ/z2sI+QFri4uKC2NjYHMsjIiJQrVq1POPNzMxw+vTpHMtPnz6d7znPypYti9DQUFy5cgXz5s3D4cOH8dVXX8HR0RETJ05Eenq6dFsjIyOZPjQmJiawtLSUef7mzRu5x6lYsSKioqIAABcuXIBEIsH58+el68+dO4fy5cvLje3WrRtOnjyJlStXonPnzvjnn3/ydW7Z/tuXZfTo0Rg7diw+//xznDlzRmGcRCKBRCIBABgYGMDa2lpmvY2NTa65ODs7Y+/evQrX7927F5UqVZK7rmPHjpg+fToyMzPRvn17/PrrrzL9QRYuXAgfHx+F+wY+dj5W9IiPj1c6Nj4+PtdYVV4nVX/W2WJjYxEWFgZbW1vpMltbW0yfPh0xMTE5tnd3d8eePXsAAPv374epqSkOHjwoXX/gwAG4uLgoPF758uVx69Yt6fN79+4hKysLpUuXBgBUqFABr169khurymusynvLxMQEL1++lFn2zTffYMWKFejatSv++usvhft1c3PDsGHDcPToUbmP5cuXK4xV5fPjU8bGxtLPpcOHD+Pzzz8HAJQqVSrP/iWWlpZ4/vw5AODgwYNo0aIFAMDU1FThsatXr46VK1cq3OeqVavg4eEhd12lSpUwf/58hb+Hub2Gu3fvxtu3b2FtbS338enPTp5///0XaWlpMDU1zTEHl6mpqcxn+6ecnZ1x8eJF6fPo6GgYGBhI5+YsVaoU3r9/LzdWlc95rdF2Baavli9fLsqXLy82bdokLCwsxMaNG8W0adOk/89LWFiYMDU1FcHBwWLdunVi3bp1YvDgwcLc3FyEhYXlK4ekpCQxa9YsUa1aNWFubi569Oghjhw5ItauXSuqV68uWrRoId3Wz89PpjkzNTVV2iwqhBCHDh0SlStXlnuc+fPnC1NTUxEYGChsbW3FL7/8IhwcHMTIkSPF6NGjhbW1tZgyZUquuWZmZoqJEycKJycnERERIYyMjPJsWWjUqJHMJYpPzZo1S5iYmOT6LcrGxkbY2toKIyMjsW7dOpn1Bw8eFM7OzgqPvWXLFlGiRAnRtm1bsWDBArFp0yaxadMmsWDBAtGuXTthbGwstm7dKjf25cuXws/PT7i5uYlevXoJU1NTUalSJdGiRQvh4uIirK2txdmzZ3M9d20pyOuk6s86m7e3t/QSx6ciIyOFp6dnjuXr168XhoaGws3NTZiYmIg///xTODo6ii5duohu3boJY2PjXC9/TJ48WVSoUEEsWbJErFq1Snh6espcHti+fbvw8PCQG6vKa6zKe6tFixbif//7n9x1GzZsEEZGRgp/J7755hsxdOhQhT+P3C7LqPL58am2bduKli1biilTpggjIyPx999/CyE+9m1yd3fPNfabb74RNWvWFP379xfm5ubi2bNnQgghdu7cKapXry435ujRo8LCwkJ4eXmJYcOGiZkzZ4qZM2eKYcOGCW9vb2FpaamwP03nzp3FyJEjFeaT28/Ly8tLrFixQmFsTExMrp9dBgYG0lbbZcuWyazfuXOnwst2ixYtEtbW1mLkyJFi4sSJwtHRUfTv31+6fv369cLX11durDo+5wsbCyAtWr9+vXBzc5NeZihfvnyub/r/2rx5s6hfv76wtbUVtra2on79+mLz5s15xm3btk18+eWXwsjISNSoUUMsXLhQ/PPPPzLb3L17VxgZGUmfb9++XeEvuhAfCzJ5fZKy/fHHH2Lw4MFiw4YNQoiPHyyNGjUStWrVEj///LPIzMzMM28hhDh58qRwcXEREokkzz+sy5cvFz179lS4fubMmQr/sK5Zs0bm8d9m5ClTpohhw4blevzTp0+Lrl27iooVKwpjY2NhbGwsKlasKLp27SrOnDmTa2xGRoZYsmSJaN26tahataqoXLmyCAgIEGPHjhWJiYm5xmZLTEwUixcvFqNGjcrRXyM/rl+/Lvbv3y927twp88iP7NfJwMAgz9dJHT9rIYTYu3evqF69uvjzzz9FYmKiSExMFH/++afw8vISe/fuFampqdJHtlOnTok5c+ZI+wJdv35d9OrVS3Tu3FmsWbMm1+O9f/9ejBw5Ujg6OorSpUuLb775RqSkpEjXnzt3LtffGVVe44K+t7Zv355rEfPHH38ovAnjyZMnCjsL50XVz49sDx8+FG3atBHe3t4yn5VDhw7Ns4PtP//8IwYNGiTatWsn7fMkxMd+k9OmTVMYd//+fTFy5EjRuHFjUblyZVG5cmXRuHFjMWrUqByXiz51/fp1ceHCBYXrMzIyFP48+/btK3788UeFsTdu3FD42XXs2DGZx3+7AYSHh4vZs2cr3Pevv/4q6tevL2rVqiXGjh0r3rx5I113+/ZtERcXpzBWXZ/zhYWzwWvBhw8fsGHDBrRs2RL29vZIT0/Hq1evYGdnVyjHt7a2Rrdu3TBgwADUrl1b7jZv3rzB7NmzMWnSpELJSRmvXr3CvXv3UK1aNRgbG2s7nSIpMjIS7dq1g6urK27evAlPT088ePAAQgjUrFlTeiuyPPHx8ejYsSOuXr0KiUQivTSTfZkqMzMzXzkU9utkYPB/V/Szc/1v7kIISCSSfJ8DkTa8e/cOmZmZMDc313YqxRoLIC0xNzdHXFycwmv1+fHy5Uts3boV8fHx+Omnn1CqVClER0fD3t4+12ut6enpOv+LtWbNGnTs2DFHf5HiTJlzrlOnDr744gtMnjwZJUuWxOXLl2FnZ4cePXqgVatW+OGHHxTGtm3bFoaGhlixYgVcXFxw/vx5PH/+HMOHD8ecOXPQqFEjdZ6W2hw/fjzf2wYEBGgsDyEEsrKyYGhoqBOxd+7cQUJCAipVqgQ3N7dCiwWU/z02NDTEkydPcnxZfP78Oezs7PIsbE+ePInffvsN8fHx+PPPP1G+fHmsW7cOLi4uaNiwYb5yUPWcNWHmzJn4/vvvYWNjo+1UcijSn9VaannSewEBAdLb4Avi8uXLomzZssLNzU2UKFFCehfZuHHjpHc35CYzM1PcunVLnDx5Uhw/flzmocjevXtF//79xYgRI3I0g7548UI0bdpUbty5c+fEhw8fpM93794tGjduLBwdHUWtWrXE77//np9TlmFkZCRu3LiR6zaqHFfVnLV1ztksLS2ld/vY2NhI78aIjY3N81bU0qVLS+/QsbKykt4FExkZKXx8fBTGffr++G+eub0//hurzHtLk/Iap+X9+/di3LhxonHjxmLixIlCCCFmz54tzM3NhbGxsejdu7d49+5dkYqdMWOGOHz4sBDi48+1efPmMnf6tWrVKsflcHXEKqLMe1oI2TucPvXo0aM8757dunWrMDMzEwMGDBAmJibSz8yFCxeKL774Qm7Mf8+5WbNm+T5nTXwGKFKyZEnp+eRFU+MPKaLsa1yYWABpSUHGWvlU8+bNxYgRI4QQsrfRnz59Os8/cFFRUdL+GfJud5bnjz/+EIaGhqJNmzaiYcOGwtTUVKxfv166Pq/bZ7M/tHbt2iUMDAxE7969xeLFi8WAAQNEiRIlxPbt2+XGZvdv+u9DIpEIa2tr6XN1H1eVWG2eczZ7e3vph061atWkfXdiY2NzvbVbiI8FU/Yty66urtKB8u7evavw1mxV3h+qxP7XixcvxP/+9z/Rr18/0a9fPzFnzhzx/PnzfMX+V26dVIUQYvz48cLe3l6EhoYKDw8P8f333wsnJyexfv168fvvv4vy5cuLWbNmFanYChUqSG+5HzBggPD19RXR0dHizZs3IjY2VtSrV0+m06u6YlV9T2ePnWNgYCCmT58uM57OvHnzRIcOHXItzoX4OBZRdtHx6WdmdHS0wlvoVTlnVT9DlPHp+eQlr/d1QWPV8blV2FgAaUlBxlr51KfjeXz65n/w4IEwMTHJNbZGjRri66+/Fjdu3BD//POPePnypcxDnk8HMhPiYwFnYWEh7YiY3zF1GjZsmGOk6unTp4t69erJjbW0tBRt2rSR6SS7evVqYWhoKKZPny5dpu7jqhKrzXPO1r59e+ndH8OHDxdubm5i2rRpombNmqJ58+a5xjZs2FDaOtm9e3fRqlUrcerUKdG7d2+Fd8uo8v5QJfZTx48fF1ZWVsLJyUk6XkrFihWFlZWV3JZNVcZaEeJjcbh7924hhBB37twRBgYGMoPIbd68We7dZ9qMNTExkXa8dXZ2zvFzuXjxosIxl1SJVfU9nT12jkQiEU5OTjLj6VSuXFl8/vnned4Z+ekYN59+Zt67d0/hZ6Yq56zqZ4gyPj0fVd7XqsSq43OrsLEA0pIHDx7k+shL2bJlpd9MPn3zHzx4UFSoUCHXWHNzc3Hnzh2l8v3vQGZCCHHkyBFhaWkplixZku8CyM7OTly8eFFm/c2bNxUOrnXnzh3plAT//vuvdLmyA+wpe1xVYlWNV+Wcs927d0/akvjq1Svx3XffCS8vL9GpU6c8318RERFi27Zt0lyqVKkiJBKJKFOmjNzbzIVQ7f2hSuynPD09xcCBA2UuO3z48EF8++23cguCEiVKiC+++CLH6NXZj3bt2uV6XFNTU5GQkCDz/NPLd/Hx8aJkyZJFKrZy5crS6SBcXFykd79li4mJUTiQqiqx6nhPCyFEkyZN8pzyQhEXFxfpSMWffmb+/vvvolq1anJjVDlnVT9DlPHp+ajyvlYlVl2vcWFiAaSj+vfvLzp06CAyMjKEpaWliI+PFw8fPhS+vr4iJCQk19imTZvK3AaaH/JGFBXi4y2XlpaWYty4cbkWQEePHhWXL18WlSpVEufPn5dZf/PmTWFpaanw2Nm3G3/22Wfi1KlTQoj8F0AFPa6qOWvrnDXl+fPnMuO2/Jcq7w9VYj+laIRkRaP1qjLWihAfLzNeuXJF+rx+/frScWmE+Di6saI/jtqK/d///ieqVasm7ty5I+bOnSv8/f2lLcnx8fGiSZMm4quvvlJ7rBDqfU+/e/dO3Lx5U+E8b/LMmDFDeHh4iLNnz4qSJUuKkydPivXr14uyZcuKX375RW6MKues6meAMj4tgFR5X6v6O1HUPrfywgJIi+7evSsGDx4smjdvLpo3by6Cg4NzDFOvyMuXL0VgYKCwsbERhoaGwsnJSRgZGYnGjRuLV69e5RqbPUDb6tWrxcWLF/PV/6h9+/bSDpf/lT1YWF4Dc2Vf6ps/f77M+o0bNyocMO5TkZGRomLFimLMmDH5HmCvoMdVNWdtnfOn/vnnH7F8+XIxevRoaT+YS5cuyfyxzM2dO3dERESESE9PF0KIXAsgVd4fqsR+qn79+nJvLPjrr79E3bp1cyxXZawVIT5+kcitSX/Lli2iVq1aRSpWCCGCg4OFkZGRqFq1qjA1NRUGBgbC2NhYGBgYCD8/P4UTuKoam02V93R6erro16+fMDQ0FIaGhtI/+oMHD85zANisrCzpYLPZv5empqZ5jj9U0HNW12dAfnxaAKnyvlb1dyKbKq9xYeJt8Fpy4MABtGvXDj4+PmjQoAGAj9NYXL58Gbt375YO056XU6dO4cqVK3j16hVq1qyJwMDAPGM+HS8lW/Z4L4rGSDl+/DjOnDmDMWPGyN3n0aNHsXbtWqxevTrHuocPH8o8t7S0lE4XAABr164FAPTu3TvP3J8/f46BAwfi6NGjOHv2LKpUqaJwW1WOq2rO2jrnbFeuXEFgYCCsra3x4MED3Lp1C66urhg/fjwSEhKkx1d0vC5duuDo0aOQSCS4c+cOXF1d0a9fP9ja2mLu3Lk5YlR5f6gS+6nNmzdj5MiRCA4ORr169QAAZ8+exeLFizFz5kyZKWa8vb1VHmvl9u3bMDIyUjhdxoYNG1CiRAl06dKlyMRmi4uLw549exAfH4+srCyUK1cODRo0QGBgoHTMJE3EZivIexoAQkJCcPr0aYSHh6NVq1a4cuUKXF1dsXPnTvz8889ypzwBPo5ddfr0aXh7e8Pc3Bx3797Fq1ev4OHhkee0EgU9Z3V+BuSldevWWLlyJcqVK6fS+1qd4w8V9DUuVFouwPSWj4+PGDVqVI7lo0aNUjjUuLqo2v8oP8LCwpS+JVZfYzVxbFXuEuzVq5do2bKlSExMlImNiIhQ2zdWTfys5d1YUNCbDNSds6rxjP2oYsWK0suln74379y5o7DfUzYTE5Mcfc3UTRPnfPfuXTFu3DjRrVs3aZ+iffv2yUw0qoqi9hoXJrYAaYmpqSmuXr0Kd3d3meW3b9+Gt7d3jkk8AeCXX37J9/6HDBmico6qsLKyQmxsLFxdXRmrhWNbW1sjOjoan332mXQgRFdXVzx8+BBVqlSR+/7K5uDggAMHDqBGjRoysfHx8fD29lY4waeqOasa+99v3LkpyACkRe011sdYc3NzXLt2Da6urjLvzcuXL6Nx48ZITU1VuE8/Pz/MmjULzZs3VzofVfMuaOzx48fxxRdfoEGDBjhx4gTi4uLg6uqKmTNn4uLFi9i6dWuRy7kwYtWlhNaOrOfKli2L2NjYHAVQbGyswikx5s+fL/M8JSUF6enp0tE/X758CXNzc9jZ2eWrALpx4wYSEhJkZmkGgHbt2ilxJvKpUlfrW6wmjm1iYiJ3duzbt2+jbNmyue7v9evXcpvAX7x4kWNm6YLSxM9alVHVVTluYcQz9iM/Pz/s3bsXwcHBAP5vipMVK1bA398/131OmzYNP/30E6ZOnYpatWrBwsJCZr2VlVWB880r74LGjh49GtOmTUNoaChKliwpXd6sWTMsWrSowMfK67hFPVZdWABpycCBA/Htt98iPj4e9evXB/CxD9CsWbMQGhoqN+b+/fvS/2/YsAG//vorVq5cKb22euvWLQwcOBDfffddrsdW11xPVHS1a9cOU6ZMwZYtWwB8fG0TEhIwatQodO7cOdfYRo0aYe3atZg6dao0NisrC7Nnz0bTpk01nruqNFnYk3bNmDEDX3zxBW7cuIEPHz5gwYIFuHHjBs6cOZPnVCitW7cG8PF98Gm/HVGE54e7evUqNmzYkGO5nZ0dnj17poWMihcWQFoyYcIElCxZEnPnzpV2/nR0dMTPP/+cr9abCRMmYOvWrTIdy6pUqYL58+fjq6++Qo8ePRTGhoSEwMXFBZGRkXLneiLdN3fuXHz11Vews7PDmzdvEBAQgKSkJPj7+2P69Om5xs6ePRvNmzfHxYsXkZGRgZEjR+L69et48eIFTp8+XUhnoDwW9sVfw4YNERsbi5kzZ8LLywsHDx5EzZo1ERUVBS8vr1xjjx49WkhZqo+NjQ2ePHmSo8N7TExMrvM9Uv6wANISiUSCYcOGYdiwYfj3338BQKaJMy9PnjzBhw8fcizPzMzE06dPc42NiorCkSNHUKZMGRgYGMDAwAANGzZEWFgYhgwZovBOCtId1tbWOHToUIHuEvT09MTt27exaNEilCxZEq9evUKnTp0waNAglCtXrhCyLxgW9vrhs88+w/Lly5WO0+QEuJrSrVs3jBo1Cn/++ae0Jfb06dP46aef1HL3mL5jAaQl9+/fx4cPH+Du7i5T+Ny5cwdGRkZwdnbONb558+b47rvvsGLFCtSsWRMAcOnSJfzwww95/pHLzMyUHrNMmTJ4/PgxqlSpgkqVKuHWrVuqnRgVCfHx8XB1dUXDhg3zPcv1p6ytrTFu3DgNZKY5LOz1w71797B69WrEx8cjPDwcdnZ22L9/PypWrIjq1asrjDtx4kSu+23cuLG6U1XZjBkzMGjQIDg5OSEzMxMeHh748OEDevTogfHjx2s7PZ2Xc0AYKhR9+/bFmTNnciw/d+4c+vbtm2f8qlWr4ODgAD8/P5iYmMDExAR16tSBvb09VqxYkWusp6cnLl++DACoW7cuZs+ejdOnT2PKlClq65HfqFEjmJmZMVZLx3Zzc0PTpk2xfv36XO/4ksfZ2RlTpkxBYmJigfLJD038rOUV9gDUVtgXtddYH2OPHz8OLy8vnDt3Dtu2bZPekXj58mVMmjQp1302adIkx6Np06bShzqo+5yNjY2xfPlyxMfHY8+ePVi/fj1u3bqFdevWwdDQUB0pF7nXuFAV1v32JKtkyZJy5+O6c+eOsLa2zvd+bt26JXbs2CF27twpbt26la+Ygsz19ClVxqXQt1htHTsmJkYMGTJElC1bVlhbW4tvv/02z8kis82fP1/UqFFDGBoaisDAQLFx40bx9u3bfMWqkrOqsQWZxFUdx1U1nrH5j61Xr56YO3euEEJ2HKBz586J8uXL5xr730mfU1JSxMGDB0XdunXF4cOHNZp3QWOHDRsm9xEaGirGjh0rVq1aJR3lvajkrGpsYWIBpCVWVlbSyUw/dfHiRaXnh8nKysp1moL8yGuup2zHjh0TZmZmIjAwUBgbG0s/gMLCwkTnzp0ZW4SOLcTHuXm2bdsm2rZtK4yMjET16tXF3LlzRXJycp6xly5dEsHBwaJMmTLC1tZWDBo0SFy6dKnInm9BC3tdfY31LVYI2YlzPy2A7t+/r3BG97wcO3ZM1KxZU2N5qxLbpEkTYWVlJSwsLETNmjVFzZo1haWlpbC2thZ169YVNjY2wtbWVu5UE7r6GhcmFkBa8uWXX4qvv/46x8zVnTt3Fq1atcrXPn7//Xfh6ekpTExMhImJifDy8hJr165VKo+EhASZmaXzoso3MH2L1faxP/X27Vsxb948YWJiIiQSiTAxMRG9evUSjx8/zjM2IyNDhIeHCxMTE2FgYCBq1KghVq5cKbdgLirnm01eYZ+YmCgyMzPVdlxdfG/qYqwQQpQvX146I/un8du3bxeurq55xssTFxcnLCwsNJa3KrHz588XnTp1EqmpqdJlL1++FF999ZUIDw8Xr1+/Fu3btxeff/55kclZE7/HmsICSEuuX78uSpcuLT777DPRt29f0bdvX/HZZ5+JsmXLiqtXr+YZP3fuXGFubi5Gjhwpdu7cKXbu3ClGjBghzM3Nxbx583KNff/+vRg/frywsrISBgYGwsDAQFhZWYlx48aJjIyMXGNV+Qamb7HaPrYQQly4cEH88MMPwtbWVlSoUEGMGzdOxMfHixMnTojmzZuL2rVrK4zNyMgQmzdvFq1atRKGhoaiQYMGYtWqVWLKlCnC3t5edO/evcidb36ULFlSum91HFcX35u6GCuEEMOHDxcNGzYUT548kXYjOHXqlHB1dRU///xzrrH/nfQ5NjZW7N+/XwQEBIgGDRoUyXN2dHSU27pz7do14ejoKIT42FJbunTpIpNzYf0eqwM7QWuJh4cHrly5gi5duiA5ORn//vsvevfujZs3b8LT0zPP+IULF2LJkiWYNWsW2rVrh3bt2mH27Nn49ddf85wyIzg4GMuWLcPs2bMRExODmJgYzJ49GytXrsxzDKLscSn+Kz/jUuhbrDaPPW/ePHh5eaF+/fp4/Pgx1q5di4cPH2LatGlwcXFBo0aNsGbNGkRHR+eIjY6ORnBwMMqVK4fBgwejevXquHbtGk6dOoWgoCBMmDABhw8fxl9//VVkzlcZ4j8j0Orqa6xvscDHu6KqVq0KJycn6WSmjRs3Rv369fO8K8rHxwe+vr7w8fGR/r9169bIyMjI88YRbZ1zamoqkpOTcyxPSUmRjvRuY2OTY9BPbeZcWL/HaqHtCowKxsTERG4n6tu3b+dZZVtZWYl9+/blWL53715hZWWVa6wq38D0LVabx3ZzcxMzZszI9RLXu3fvxJo1a3IsNzAwEC1bthRbtmxR2CL46tUr0bdv3yJzvsr49FupOo6ri+9NXYz91MOHD8XevXvF5s2bxe3bt/MV899JnxMSEsSbN2/yFautc/7mm2+Ei4uL2L59u0hMTBSJiYnSy309e/YUQgixceNGUatWrSKTc2H9HqsDCyAt2b9/vzh58qT0+aJFi0SNGjVE9+7dxYsXL/KMr169upg+fXqO5VOnThWenp65xpYtW1bcuHEjx/IbN26IMmXK5Br77t07MWDAAFGiRAkhkUiEkZGRkEgkomfPnjL9mRir/WPnxw8//CBSUlJklj148CBfsRs2bBCvXr1SW86Fcb5C5CyAdPU11rdYbdLWOf/7779iwIABwtjYWNpdwdjYWAwcOFD6uxcTEyNiYmKKTM669BpzNngt8fLywqxZs9C6dWtcvXoVfn5+GD58OI4ePYqqVati9erVucZv27YNXbt2RWBgIBo0aADg41xikZGR2LJlCzp27KgwdsqUKbh58yZWr14tndzy3bt36N+/P9zd3fMcTwMAEhMTcfXqVbx69Qq+vr45JnVlbNE5dm40NZtzUT1fADKziKvzuLr43tS12MzMTKxZswaRkZFITk5GVlaWzPojR44ojFXUNUAikcDU1BRubm5o3LhxruPraOvn9erVK8THxwMAXF1dYWlpWSjHLcq/x+rAAkhLLC0tce3aNTg7O+Pnn3/GtWvXsHXrVkRHR6N169ZISkrKcx+XLl3C/PnzERcXBwCoVq0ahg8fDl9f31zjOnbsiMjISJiYmKBGjRoAPg4klpGRgebNm8tsu337dpnniiZq/fRDpH379ihVqlSObfQtVtvHzg9FxUBBY4v6+QI5CzddfY31LRYABg8ejDVr1qBNmzYoV66czKSmADB//ny5cQDg4uKClJQUpKenw9bWFgDwzz//wNzcHJaWlkhOToarqyuOHj0KJyenInPOBaWrr3Gh0m4DlP76dOyGBg0aiN9++00I8bGnvJmZmUaPnX3XWX4e/6XKuBT6FqvtY+fHfy8HqRpb1M9XXt66+hrrW6wQQpQuXVrs3btX7rq8bNiwQTRp0kTcvXtXuuzOnTuiWbNmYtOmTSIxMVE0aNBA7lg1uvC+Lio5a+t8C4IFkJa0bdtWtGzZUkyZMkUYGRmJv//+WwghxIEDB4S7u3ue8Xv37hURERE5lkdERMjt4KwuqoxLoW+x2j52fqi7ACpK55uamir++uuvHP3dEhISZPoi6OprrG+xQghRrly5fI94/1+urq5y+8pER0cLFxcXIYQQp0+fFg4ODmrNuzB+j+XR1de4MLEA0pKHDx+KNm3aCG9vb7FixQrp8qFDh4rg4OA84728vOR+E9q/f7/w9vZWa66fUmVcCn2L1fax80PdBZA2z/frr78WCxcuFEIIkZ6eLtzd3YWRkZEoUaKE2Lp1q8Lz0NXXWN9ihRBizpw54scffyzQyPdmZmbiwoULOZafP39e2up+//59uYMiFvXfY3l09TUuTBwHSEsqVqyIPXv24PLly+jfv790+fz582U6682cORMvX77MEX/nzh14eHjkWF61alXcvXs312M/f/4cgwYNgoeHB8qUKYNSpUrJPHKjyrgU+har7WNrgzbP98SJE2jUqBEA4K+//oIQAi9fvsQvv/yCadOmaSRnVeMZm3dsp06dpI/Tp0/jjz/+wGeffYa2bdvKrOvUqVOO/X6qadOm+O677xATEyNdFhMTgx9++AHNmjUDAFy9ehUuLi5aP2d10KXXWGu0XYFR7v47am02e3t7ufMbHTp0SJQtWzbXfX7xxRfC3d1dzJw5U6xevVqsWbNG5pEbVcal0LdYbR774cOHcr8lZ2VliYcPH0qff//99zlug8+v6tWr55hGRZs/a1NTU2k+vXr1EqNGjRJCfPxZ5DbVga6+xvoSq0qfxU89efJEBAYGColEIoyNjaW3lrdo0UIkJSUJIYQ4cuSIOHDggNbPWR106TXWFhZARZyiSxTffvut8PLyytGhz9vbW/Tv3z/PfcbGxhYoH1XGpdC3WG0e28DAQDoL86eePXsmDAwM5OaqDtr8Wbu7u4vNmzeLV69eibJly0q/IMTGxuba3K6rr7G+xapLXFycdPqgmzdv5itGF89Zn1/j/OJt8EWcotuUU1NT0apVK1y8eBEVKlQAAPz9999o1KgRtm/fDhsbG4X7rF27NhYuXIh69eoVOC9VxqXQt1htHNvAwABPnz5F2bJlZZY/fPgQHh4eeP36tcJYW1vbHLcXA7K3sfbt2xdBQUFqzVnV2F9//RUhISGwtLREpUqVEB0dDQMDAyxcuBDbt2/H0aNHNXJcdcQzNn+xb968gRAC5ubmAD6+n//66y94eHjg888/z/fxC0qbnyEFpWuvcWFiAVTE5TZOixAChw4dwuXLl2FmZgZvb280btw4z31euHABo0ePxsSJE+Hp6QkjIyOZ9VZWVmrLnwpX9hgcCxYswMCBA6V/KICPg8idO3cOhoaGOH36tMJ9zJ8/H9OnT8cXX3yBOnXqAADOnz+PiIgIDBs2DPfv38e6deuwcOFCDBw4ULMnpKSLFy8iMTERLVq0kH7g7t27FzY2NtIBQ0l3ff755+jUqRO+//57vHz5ElWqVIGxsTGePXuGefPm4YcffpDZPjQ0FFOnToWFhYXC8WmyzZs3T5OpUxFUQtsJUMFJJBJ8/vnnSn/zsbGxQVpamrTjXzYhBCQSCTIzM9WZJhWi7A6eQghcvXoVxsbG0nXGxsaoUaMGfvrpp1z3cerUKUybNg3ff/+9zPLffvsNBw8exLZt2+Dt7Y1ffvmlyBVAfn5+8PPzk1nWpk0bLWVD6hYdHS0d7HDr1q1wcHBATEwMtm3bhokTJ+YogGJiYvD+/XtprLyWTQAKl1PxxhagIi63FqDIyEiFQ8KvWrVK4T7r1KmDEiVKICQkBPb29jl++QMCAtSTPGlNUFAQFixYUKDWPEtLS8TGxsLNzU1m+d27d+Hj44NXr17h3r178Pb2zvVSWmFTZZoE0g3m5ua4efMmKlasiC5duqB69eqYNGkSEhMTUaVKFaSnp8tsf+XKFXh6esLAgDc8U05sASriGjVqBDMzsxzLJ0+ejClTpsDPz0/ukPC5uXbtGmJiYlClShV1pkpFSF5zyeWmVKlS2L17N4YNGyazfPfu3dJhEl6/fo2SJUuqlKO6hYSESKdJ8PT05Lf6YsjNzQ07duxAx44dceDAAel7NDk5WW6x7+vriydPnsDOzg6urq64cOECSpcuXdhpUxHFAkhLDA0Npb+Yn3r+/Dns7Oykl6H27dsnN37p0qVYs2YNevXqpfSx/fz8pN+YqHh6/fo1Zs6cqbA1JLtzojwTJkzADz/8gKNHj0r7AF24cAH79u3D0qVLAQCHDh0qci2FmzZtwpYtW9C6dWttp0IaMnHiRHzzzTcYNmwYmjdvDn9/fwDAwYMH5c6BaGNjg/v378POzg4PHjzI8XtA+o0FkJYouvL47t07mX4bimRkZKB+/foFOnZwcDBCQkIwYsQIeHl55egE7e3tXaD9UtExYMAAHD9+HL169VK6hXDgwIHw8PDAokWLpJPhVqlSBcePH5e+54YPH66RvFVhbGyc47IdFS9fffUVGjZsiCdPnkgncgaA5s2bo2PHjtLnf//9NxwdHdG5c2cEBARIfwf8/PwUzvae25cCKp7YB6iQZY/yPGzYMEydOlXm1sDMzEycOHECDx48kBmtVJ5Ro0bB0tISEyZMUDoHedfDJRIJO0EXIzY2Nti7d69e3fk0d+5cxMfHY9GiRbz8peesrKwQGxsLV1dXRERE4O7duxgyZAimTJmi8NJtSEhIIWdJ2sYWoEKWfQeDEAJLly6V+TZibGwMZ2dn6WWG3Lx9+xbLli3D4cOH4e3tnaMVJ7dbOu/fv1/A7ElX2Nra5jmtSW6ysrJw9+5duZfP8jPUgjacOnUKR48exf79+1G9evUcvxPZrVlU/H36vb5Vq1YAgEuXLiEkJCTPvmvZrUfsOF38sQVIS5o2bYrt27fD1ta2wPGKSCQS3vGi59avX4+dO3fi999/lxkLKD/Onj2Lb775Bg8fPsxxqbYotxDmNjAjoFrHcNItud09m5dPW4+oeGMBpKfu3buH8PBwxMXFAQA8PDwQEhKCzz77TMuZkTr4+vri3r17EELA2dk5R2tIdHS0wlgfHx9UrlwZkydPltt/yNraWiM5E6mLKgWQKrGkW3gJTEvUNWbJ3bt3ce/ePTRu3BhmZmbSfjy5OXDgANq1awcfHx9pH5HTp0+jevXq2L17N1q0aFGwk6Iio0OHDgWOvXPnDrZu3aqzHYpTUlJw69YtAB87b/93OhAiIoAFkNaoOmbJ8+fP0aVLFxw9ehQSiQR37tyBq6sr+vfvD1tbW8ydO1dh7OjRozFs2DDMnDkzx/JRo0axACoGJk2aVODYunXr4u7duzpXAL1+/RrBwcFYu3at9AuFoaEhevfujYULFyp9KZB0FzvBU36wANISVccsGTZsGIyMjJCQkIBq1apJl3ft2hWhoaG5FkBxcXHYsmVLjuX9+vVDeHh4gfKh4iM4OBjDhw9HUlKSTg2TEBoaiuPHj2P37t3Sls1Tp05hyJAhGD58OJYsWaLlDKmwsGcH5QcLIC1RdcySgwcP4sCBA9KZ4LO5u7vj4cOHucaWLVsWsbGxcHd3l1keGxubY2BG0k2ZmZmYP38+tmzZgoSEBGRkZMisf/HihcLYzp07A/hYEGfThWEStm3bhq1bt6JJkybSZa1bt4aZmRm6dOnCAkiP3LhxA46OjgWKZeuR/mABpCXDhw/HggULCjxmyevXr+U26b948QImJia5xg4cOBDffvst4uPjpQPbnT59GrNmzcpzxmTSDZMnT8aKFSswfPhwjB8/HuPGjcODBw+wY8cOTJw4MddYXR0mIT09Hfb29jmW29nZ5ZgjinTT27dvsXDhQhw9elRu38nszv1OTk4FPgZbj/QH7wLTko4dO+Lo0aMoVapUgcYsad26NWrVqoWpU6eiZMmSuHLlCipVqoRu3bohKysLW7duVRgrhEB4eDjmzp2Lx48fAwAcHR0xYsQIDBkyhN+AioHPPvsMv/zyC9q0aYOSJUsiNjZWuuzs2bPYsGGDtlNUu+bNm6N06dJYu3YtTE1NAQBv3rxBnz598OLFCxw+fFjLGZKqevTogYMHD+Krr76SO5GzKn3fsiUmJsLR0VHhiNFUfLAA0hJVxyy5du0amjdvjpo1a+LIkSNo164drl+/jhcvXuD06dP5vp3933//BYAiN7ElqcbCwgJxcXGoWLEiypUrh71796JmzZqIj4+Hr68vUlNTZbbftWsXvvjiCxgZGWHXrl257rtdu3aaTL3Arl27hpYtW+Ldu3fSaRIuX74MU1NTHDhwANWrV9dyhqQqa2tr7Nu3r0AjnOe39Yj0By+BaYmqg7J5enri9u3bWLRoEUqWLIlXr16hU6dOGDRoEMqVK5dr7P379/Hhwwe4u7vLFD537tyBkZERnJ2dVcqNtK9ChQp48uQJKlasiM8++wwHDx5EzZo1ceHCBbmXSDt06ICkpCTY2dnlegt9Ue4D5OnpiTt37uCPP/7AzZs3AQDdu3dHjx49YGZmpuXsSB3Kly9f4C9r/fv3l7Ye1alThy3dBAjSmvfv34tDhw6JpUuXirS0NCGEEI8ePRL//vuv2o7xww8/iJSUFJlljRs3FmvWrMmx7bp160RAQIDajk3aM2rUKDF9+nQhhBCbNm0SJUqUEG5ubsLY2FiMGjVKy9kRFcy+fftEq1atxIMHD5SOtbKyEqdOndJAVqSreAlMSx4+fIhWrVohISEB7969w+3bt+Hq6oqQkBC8e/cuX/OB5Ye8Yd2trKwQHR2d4y60u3fvws/PDy9fvlTLsanoOHv2LM6cOQN3d3e0bdtW2+moTXG4dEf5l5KSgi5duuDEiRMwNzfP0Xcyt7sbPTw8sGnTpiI7jAMVPl4C05KQkBD4+fnh8uXLKF26tHR5x44dMXDgQLUdR159K5FIpH1/PpWamlpkL29Q/r1//x7fffcdJkyYABcXFwBAvXr1UK9ePYUxv/zyS773P2TIEJVzVJficOmO8q979+549OgRZsyYIbcTdG7mzp2LUaNGYenSpahUqZIGsyRdwRYgLSldujTOnDmDKlWqyMw98+DBA3h4eKjttl1589q0bdsWZmZm2Lhxo/ROh8zMTHTt2hWvX7/G/v371XJs0h5ra2vExsZKC6C85Hc7iUSC+Ph4VVIjKjBzc3NERUVJO7krQ5XWIyqe2AKkJVlZWXK/kf79998avyNr1qxZaNy4MapUqYJGjRoBAE6ePIm0tDTOIl9MdOjQATt27MCwYcPytb2ujv3zqbVr16Jr1645OnlnZGRg06ZN6N27t5YyI3WpWrUq3rx5U6BYVVqPqHhiC5CWdO3aFdbW1li2bJl0HJ+yZcuiffv2qFixosp3iWVTNLPx48ePsWjRIly+fBlmZmbw9vbG4MGDUapUKbUcl7Rr2rRpmDt3Lpo3b45atWrBwsJCZn1RuoylLoaGhnjy5EmO0cyfP38OOzs7XgIrBg4ePIjJkydj+vTpcqdpsbKyUhirSusRFU8sgLTk77//RsuWLSGEwJ07d+Dn54c7d+6gTJkyOHHihNqmpFBUAOXHjz/+iClTpqBMmTJqyYUKT26XtORdxgoNDcXUqVNhYWGR52jg8+bNU0uO6mZgYICnT5/mmP398uXLaNq0KS9xFAMGBgYAck5XIfIxTUvNmjXx66+/5toXjvQLCyAt+vDhAzZv3ozLly/j1atXqFmzptrHLPnhhx8wderUAhUx8u4go+KpVKlSuH37NsqUKYOmTZsq3E4ikRS5y6S+vr6QSCS4fPkyqlevjhIl/u/KfmZmJu7fv49WrVrJnQCYdMvx48dzXR8QEKBwnSqtR1Q8sQDSkhMnTqB+/foyH9bAx6LozJkzaNy4cZ77OHnyJH777Tfcu3cPW7duRfny5bFu3Tq4uLigYcOGKueoSusRaZeiVhyJRAJTU1O4ubmhffv20kueBgYG0rupXF1dceHCBZm7E4uyyZMnS/8dPnw4LC0tpeuMjY3h7OyMzp07w9jYWFspUhGgSusRFU/sBK0lTZs2ldtfITU1FU2bNs3zl3Hbtm3o1asXevTogZiYGLx7904aP2PGDOzbt09juVPRFxMTg+joaGRmZqJKlSoAgNu3b8PQ0BBVq1bFr7/+iuHDh+PUqVPw8PCAra0t7t+/Dzs7Ozx48CDHNAFFWfb8T87Ozujatat0HjAqfk6cOJHr+ty+OB49elTd6ZCOYwuQlijqr3D79m34+fkhLS0t13hfX18MGzYMvXv3lmmpiYmJwRdffIGkpCSVc2QLkO4KDw/HyZMnsXr1amnTfmpqKgYMGICGDRti4MCB+Oabb/DmzRscOHAA3377LdauXYty5cohISEBFSpUUDgZJG+DJ23JbsX51KctOmzFIWWwBaiQderUCcDHX9q+ffvK3LKbmZmJK1euoH79+nnu59atW3K/7VhbW3MkZ8L//vc/HDp0SKZfg7W1NX7++Wd8/vnnCAkJwcSJE/H5558DAJYtW4ZOnTrh7t27GDJkCAYOHKhzE+RmZmZi/vz52LJlCxISEpCRkSGznp2gdd8///wj8/z9+/eIiYnBhAkTMH369FxjVWk9ouKJBVAhs7a2BvDxunPJkiVlOjwbGxujXr16+RoJ2sHBAXfv3s0xcempU6fYYkNITU1FcnIyPDw8ZJanpKRIWxdtbGxkioRWrVoBAC5duoSQkBCdK4AmT56MFStWYPjw4Rg/fjzGjRuHBw8eYMeOHZg4caK20yM1yP78/FSLFi1gbGyM0NBQXLp0SWFskyZNcixj65GeK+S5x+j/+/nnn8WrV68KHD9jxgzh4eEhzp49K0qWLClOnjwp1q9fL8qWLSt++eUXteT4/fff55hIlXTDN998I1xcXMT27dtFYmKiSExMFNu3bxeurq6iZ8+eQgghNm7cKGrVqqXlTNXH1dVV7NmzRwghhKWlpbh7964QQogFCxaI7t27azM10rC4uDhhYWGR6zYvX76UeaSkpIiDBw+KunXrisOHDxdSplSUsA+QjhJCYMaMGQgLC5NOm2FiYoKffvoJU6dOzTNe03eQkXa9evUKw4YNw9q1a/HhwwcAQIkSJdCnTx/Mnz8fFhYWiI2NBQD4+PhoL1E1srCwQFxcHCpWrIhy5cph7969qFmzJuLj4+Hr64vU1FRtp0gqunLlisxzIQSePHmCmTNn4sOHDzh16pTS+zx+/HierUdUPPESmBZt3bpVYX+F6OjoXGMlEgnGjRuHESNG4O7du3j16hU8PDxkbgFWhHeQFX+WlpZYvnw55s+fL+207OrqKvP+KC6FT7YKFSrgyZMnqFixIj777DMcPHgQNWvWxIULF3JMj0G6ycfHBxKJJMckz/Xq1cOqVasKtE97e3vcunVLHemRjmEBpCW//PILxo0bh759+2Lnzp0ICgrCvXv3cOHCBQwaNCjf+zE2Ns7RzyMv06ZNw9KlS9G7d29s2rRJurxBgwaYNm2aUvuios3S0hLe3t7aTqNQdOzYEZGRkahbty6Cg4PRs2dPrFy5EgkJCfmeE42Ktv/OWWdgYICyZcvma+iD3FqPituXAcofXgLTkqpVq2LSpEno3r27zO3mEydOxIsXL7Bo0aJc49++fYuFCxfi6NGjSE5OzjFuS24tSObm5rhx4wacnZ1ljh0fHw8PDw+8fftWLedIpE1nz57FmTNn4O7ujrZt22o7HVKTyMhIREZGyv3cy60VyMDAINfWo6pVq2okXyq62AKkJQkJCdLb3c3MzPDvv/8CAHr16oV69erlWQD1798fBw8exFdffYU6deooNbMx7yCj4ui/o6vXq1cP9erVw4cPH3DixAne5lwMTJ48GVOmTIGfnx/KlSun1OeeKq1HVDyxANISBwcHvHjxApUqVULFihVx9uxZ1KhRA/fv38/xDUWePXv2YN++fWjQoIHSxx44cCBCQkKwatUqSCQSPH78GFFRUfjpp58wYcKEgpwOkdapOro6FX1Lly7FmjVr0KtXL6VjK1WqVODWIyqeWABpSbNmzbBr1y74+voiKCgIw4YNw9atW3Hx4kXpYIm5KV++fIHHaRk9ejSysrLQvHlzpKeno3HjxtI7yIKDgwu0TyJtE/9/Tqf/ev78OSwsLLSQEalbRkZGvgaKlUeV1iMqntgHSEuysrKQlZUlba7ftGmTtL/Cd999l+fEjfv378cvv/yCpUuXolKlSgXKISMjQ+k7yIiKmuwvDDt37kSrVq3kjq5epUoVREREaCtFUpNRo0bB0tKyQC3V5cqVw+zZswvUekTFE1uAtMTAwEBmXptu3bqhW7du+Y738/PD27dv4erqCnNzcxgZGcmsz8+w/wW5g4yoqFHX6OpU9L19+xbLli3D4cOH4e3tneNzb968eQpjVWk9ouKJLUBa9M8//2DlypWIi4sDAHh4eCAoKAilSpXKMzYwMBAJCQno378/7O3tczTn9unTR2GsKneQERVVkydPxk8//cTLXcVY06ZNFa6TSCQ4cuSIwvWqtB5R8cQCSEtOnDiBdu3awcrKCn5+fgA+zsH08uVL7N69O887VszNzREVFYUaNWoofewePXpI7yCTVzxNmjRJ6X0SERVlISEhWLt2Lby9vZVuPaLiiQWQlnh5ecHf3x9LliyBoaEhgI/9FX788UecOXMGV69ezTW+Zs2a+PXXX1GvXj2lj21tbV3gO8iIipKaNWsiMjIStra28PX1zbVjK1s29ZsqrUdUPLEPkJbcvXsXW7dulRY/AGBoaIjQ0FCsXbs2z/iZM2di+PDhmD59Ory8vHJ8m7GyslIYq8odZERFSfv27aWdnjt06KDdZKhIO3r0qLZToCKGLUBa0qBBA4wYMSLHh/aOHTswc+ZMnD17Ntf47A7U//3Gm30rcG5jnqjjDjIiIiJdxhYgLRkyZAhCQkJw9+5d6WWss2fPYvHixZg5c6bMvDXy5nJS5duMOu4gIypqLly4gKysLNStW1dm+blz52BoaCjta0dEBLAFSGs+vQVenuw5a/JqzSkIVe4gIyqq6tSpg5EjR+Krr76SWb59+3bMmjUL586d01JmRFQUsQDSkocPH+Z729wuU6WnpyMhIQEZGRkyy3ObAVyVO8iIiipLS0tcuXIlx3x29+/fh7e3t3S+PSIigJfAtCa/fW/atGmDFStWoFy5cjLLU1JSEBQUhP3798uNy63VqGrVqnjz5k3+kyXSASYmJnj69GmOAujJkyfSEdeJiLLlfh2GtO7EiRNyi5WhQ4fi5cuXOHfuHMzMzBAREYHff/8d7u7u2LVrV677zL6D7NixY3j+/DnS0tJkHkS66PPPP8eYMWOQmpoqXfby5UuMHTsWLVq00GJmRFQU8RJYEVeyZElcvnw5x7facuXKYefOnahTpw6srKxw8eJFVK5cGbt27cLs2bNx6tQphftU5Q4yoqLq0aNHaNy4MZ4/fw5fX18AQGxsLOzt7XHo0CE4OTlpOUMiKkrYLqyjXr9+DTs7OwCAra0tUlJSULlyZXh5eeU54BvHw6DiqHz58rhy5Qr++OMPXL58GWZmZggKCkL37t1z3OlIRMQCSEdVqVIFt27dgrOzM2rUqIHffvsNzs7OWLp0aY7+Qv8VEBBQSFkSFS4LCwt8++23uW6jqF8dEekXFkA6KiQkBE+ePAHwce6uVq1a4Y8//oCxsTHWrFmTr30U5A4yIl2nqF8dEekXFkA6qmfPntL/16pVCw8fPsTNmzdRsWJFlClTJtdYVe4gIyIiKg54F1gRN3bsWJQqVSrP7czNzVGzZs08ix9AtTvIiIiIigPeBaZFjx8/xqlTp5CcnIysrCyZdUOGDMk1NjMzE2vWrEFkZKTc+NxmNlblDjIiXafozkoi0i+8BKYla9aswXfffQdjY2OULl1a5pZ0iUSSZwEUEhKCNWvWoE2bNvD09MxxS3tuVLmDjIiIqDhgAaQlEyZMwMSJEzFmzJg85wWTZ9OmTdiyZQtat26tdKwqd5AREREVByyAtCQ9PR3dunUrUPEDAMbGxnBzcytQrDruICPSVfntV0dExRv7AGnJyJEjUapUKYwePbpA8XPnzkV8fDwWLVqk1OUvedLT0/N9BxlRUaJMp/127dppMBMi0jUsgLQkMzMTX375Jd68eQMvL68cI9XOmzcv1/iOHTvi6NGjKFWqFKpXr54jfvv27UrlcvXqVVSqVAm2trb5PwkiLftvC6pEIsGnH2mffjng8A5E9CleAtOSsLAwHDhwAFWqVAGAHJ2g82JjY4OOHTsW6NhDhw6Fl5cX+vfvj8zMTDRu3BhRUVEwNzfHnj170KRJkwLtl6iwfXr34+HDhzFq1CjMmDED/v7+AICoqCiMHz8eM2bM0FaKRFREsQVIS2xtbTF//nz07dtXo8c5ffo0/Pz8YGJiIl1WoUIF7NixA35+ftixYwcGDRqEo0ePYt26dThy5AhOnz6t0ZyINMHT0xNLly5Fw4YNZZafPHkS3377LeLi4rSUGREVRRwIUUtMTEzQoEEDjR/niy++wKNHj2SWPXv2DA4ODgCAffv24euvv0blypXRr18/XL16VeM5EWnCvXv3YGNjk2O5tbU1Hjx4UOj5EFHRxgJIS0JCQrBw4UKNH0deA5+9vT1u3LiBzMxMREREoEWLFgA+doY2NDTUeE5EmlC7dm2Ehobi6dOn0mVPnz7FiBEjUKdOHS1mRkRFEfsAacn58+dx5MgR7NmzR+VOzMoKCgpCly5dUK5cOUgkEgQGBgIAzp07h6pVq2rsuESatHLlSnTq1AkVK1aEk5MTACAxMRHu7u7YsWOHdpMjoiKHBZCW2NjYoFOnTlo59s8//wxPT08kJibi66+/lvYPMjQ0LPBt+UTa5u7ujitXruDQoUO4efMmAKBatWoIDAxUeagIIip+2Am6mFNl3iMvLy/s27dP+m2aqKh6//49zMzMEBsbC09PT22nQ0Q6gC1AWpacnIxbt24B+DhFRfYcXeqiyjffBw8e4P3792rMhkgzjIyMULFiRY71Q0T5xk7QWpKWloZevXqhfPnyCAgIQEBAAMqXL4+ePXsiNTVVbcdhAx/pi3HjxmHs2LF48eKFtlMhIh3AAkhLBg4ciHPnzmHPnj14+fIlXr58iT179uDixYv47rvv8oyfNGkSHj58mOd2//77b4EufxHpmkWLFuHEiRNwdHRElSpVULNmTZkHEdGn2AdISywsLHDgwAG5g7a1atUKr1+/zjXex8cH165dQ0BAAPr374/OnTvLDHaoDqr0HyIqbJMnT851/aRJkwopEyLSBewDpCWlS5eGtbV1juXW1tb5mo8rNjYWMTExWL16NUJCQjBo0CB069YN/fr1Q+3atTWRMlGRxgKHiJTBS2BaMn78eISGhiIpKUm6LCkpCSNGjMCECRPytQ9fX1/88ssvePz4MVauXIm///4bDRo0gLe3NxYsWKDWvkRERETFCS+BaYmvry/u3r2Ld+/eoWLFigCAhIQEmJiYwN3dXWbb6OjoXPeVkZGBv/76C6tWrcKRI0dQv359PH78GE+fPsXy5cvRtWvXPPN5+fJljmkENmzYgPbt28PCwkK5kyPSgszMTMyfPx9btmxBQkICMjIyZNazczQRfYqXwLSkQ4cOKu/j0qVLWL16NTZu3AgTExP07t0bixcvhpubGwBg4cKFGDJkSI4CaNasWXB2dpYu79KlC7Zt2wYHBwfs27cPNWrUAAB88803KudIVFgmT56MFStWYPjw4Rg/fjzGjRuHBw8eYMeOHZg4caK20yOiIoYtQFoyYMAA9OjRA02bNi1QvJeXF27evInPP/8cAwcORNu2bXPM4/Xs2TPY2dkhKytLZrmLiwv++OMP1K9fH4cOHUKXLl2wefNm6TfngwcPFvi8iLTls88+wy+//II2bdqgZMmSiI2NlS47e/YsNmzYoO0UiagIYQuQlqSkpOCLL75A2bJl0b17d/To0UPa8pIfXbp0Qb9+/VC+fHmF25QpUyZH8QN87GuUPbrznj170KVLF3z++edwdnZG3bp1lT8ZoiIgKSkJXl5eAABLS0tpH7gvv/wy3/3qiEh/sBO0luzcuRNPnjzBhAkTcP78edSsWRPVq1fHjBkz8ODBgzzjJ0yYkGvxkxtbW1skJiYCACIiIqSToQohOJIu6awKFSrgyZMnAD62BmW3ZF64cEHtQ0QQke7jJbAi4u+//8bGjRuxatUq3LlzBx8+fMh1+8zMTKxZswaRkZFITk7O0dJz5MgRhbGDBw/Gnj174O7ujpiYGDx48ACWlpbYtGkTZs+enWena6KiaPTo0bCyssLYsWOxefNm9OzZE87OzkhISMCwYcMwc+ZMbadIREUIL4EVAe/fv8fFixdx7tw5PHjwAPb29nnGhISEYM2aNWjTpg08PT2VmvNr/vz5cHZ2RmJiImbPng1LS0sAwJMnT/Djjz8W+DyItOnTAqdr166oVKkSzpw5A3d3d7Rt21aLmRFRUcQWIC06evQoNmzYgG3btiErKwudOnVCjx490KxZszwLmjJlymDt2rVo3bp1IWVLRERUfLAPkJaUL18erVu3xrNnz7Bs2TI8ffoUq1atQvPmzfPVmmNsbCy93b0g1q1bh4YNG8LR0VE6p1h4eDh27txZ4H0SaVPFihXRu3dvrFy5Evfu3dN2OkRUxLEA0pKff/4ZT548wV9//YWvvvpK6U6aw4cPx4IFCwo02/uSJUsQGhqKL774Ai9fvpR2fLaxsUF4eLjS+yMqCmbMmAFTU1PMmjUL7u7ucHJyQs+ePbF8+XLcuXNH2+kRURHDS2A6qmPHjjh69ChKlSqF6tWrw8jISGb99u3bFcZ6eHhgxowZ6NChg8yEp9euXUOTJk3w7NkzTadPpFFPnjzB8ePHsWfPHmzevBlZWVm8w5GIZLATtI6ysbFBx44dCxR7//59+Pr65lhuYmKS5yz0REVZeno6Tp06hWPHjuHo0aOIiYmBp6cnmjRpou3UiKiIYQGko1avXl3gWBcXF8TGxqJSpUoyyyMiIlCtWjVVUyPSivr16yMmJgbVqlVDkyZNMHr0aDRu3Bi2trbaTo2IiiAWQHooNDQUgwYNwtu3byGEwPnz57Fx40aEhYVhxYoV2k6PqEBu3rwJCwsLVK1aFVWrVkW1atVY/BCRQuwDpMO2bt2qcObrvAYz/OOPP/Dzzz9L75ZxdHTE5MmT0b9/f43lS6RJQghcvXoVx44dw/Hjx3HixAkYGxsjICAATZs2xcCBA7WdIhEVISyAdNQvv/yCcePGoW/fvli2bBmCgoJw7949XLhwAYMGDcL06dPztZ/09HS8evUKdnZ2Gs6YqPAIIXDp0iUsWrQIf/zxBztBE1EOvA1eR/36669YtmwZFi5cCGNjY4wcORKHDh3CkCFDpJNAKtKsWTO8fPkSAGBubi4tftLS0tCsWTNNp06kEdHR0Zg3bx7atWuH0qVLw9/fH1euXEFwcHCud0USkX5iC5COMjc3R1xcHCpVqgQ7OzscOnQINWrUwJ07d1CvXj08f/5cYayBgQGSkpJytPokJyejfPnyeP/+vabTJ1K7EiVKwNfXFwEBAQgICEDjxo1hbW2t7bSIqIhiJ2gd5eDggBcvXqBSpUqoWLEizp49ixo1auD+/fsKB0e8cuWK9P83btxAUlKS9HlmZiYiIiIKPMM8kba9ePECVlZWctcJIZSaL4+Iij8WQDqqWbNm2LVrF3x9fREUFIRhw4Zh69atuHjxIjp16iQ3xsfHBxKJBBKJRO6lLjMzMyxcuFDTqRNpxG+//YYRI0bkWJ6ZmYmePXti48aNWsiKiIoqXgLTUVlZWcjKykKJEh9r2E2bNklnvv7uu+9gbGycI+bhw4cQQsDV1RXnz59H2bJlpeuMjY1hZ2cHQ0PDQjsHInWys7NDWFiYzJ2MmZmZ6NatG65du4a4uDgtZkdERQ0LICIqFi5cuIDPP/8cy5cvx1dffYUPHz6gS5cuuHnzJo4cOQIHBwdtp0hERQgvgemwf/75BytXrpR+s/Xw8EBQUBBKlSqVY9tdu3bhiy++gJGREXbt2pXrftu1a6eRfIk0qXbt2ti2bRs6dOgAY2NjrFy5Enfv3sXRo0dhb2+v7fSIqIhhC5COOnHiBNq1awcrKyv4+fkBAC5duoSXL19i9+7daNy4scz2n975ZWCgePQDiUTC8VJIp+3YsQNff/01qlWrhiNHjqBMmTLaTomIiiAWQDrKy8sL/v7+WLJkibTfTmZmJn788UecOXMGV69e1XKGRJqnqMP/2bNn4ebmJlP8cCwgIvoUCyAdZWZmhtjYWFSpUkVm+a1bt+Dj44M3b95oKTOiwhMUFJTvbVWZQJiIih/2AdJRNWvWRFxcXI4CKC4uDjVq1Mg1dsiQIXBzc8OQIUNkli9atAh3795FeHi4utMl0ggWNURUUGwB0lGbN2/GyJEjERwcjHr16gH42Oy/ePFizJw5E9WqVZNu6+3tLRNbvnx57Nq1C7Vq1ZJZHh0djXbt2uHvv//W/AkQqdn9+/fx4cMHuLu7yyy/c+cOjIyM4OzsrJ3EiKhIYgGko3LryAx87MycPfrtfzs1m5qa4tq1a3Bzc5NZfvfuXXh6euLt27dqz5dI0wICAtCvXz/06dNHZvn69euxYsUKHDt2TDuJEVGRxEtgOur+/fsFjnVzc0NERAQGDx4ss3z//v1wdXVVNTUirYiJiUGDBg1yLK9Xr16O9zoREQsgHVWpUqUCx4aGhmLw4MFISUmRTokRGRmJuXPnsv8P6SyJRIJ///03x/LU1FQO7UBEOfASmA57/PgxTp06heTkZGRlZcms+28H5/9asmQJpk+fjsePHwMAnJ2d8fPPP6N3794ay5dIk9q2bQszMzNs3LhRZmiIrl274vXr19i/f7+WMySiooQFkI5as2aNdM6v0qVLy8x0LZFIEB8fn6/9pKSkwMzMDJaWljnWnT59Gn5+fjAxMVFb3kSacuPGDTRu3Bg2NjZo1KgRAODkyZNIS0vDkSNH4OnpqeUMiagoYQGko5ycnPD9999jzJgxeXaILigrKyvExsayXxDpjMePH2PRokW4fPkyzMzM4O3tjcGDB8udHoaI9BsLIB1VunRpnD9/Hp999pnGjlGyZElcvnyZBRARERU77ASto/r3748///wTo0eP1nYqREVKeno6EhISkJGRIbP8v+NhEZF+YwuQjsrMzMSXX36JN2/ewMvLC0ZGRjLr582bp/Ix2AJEuiQlJQVBQUEKOzvzTjAi+hRbgHRUWFgYDhw4IJ0K47+doIn0zdChQ/Hy5UucO3cOTZo0wV9//YWnT59i2rRpmDt3rrbTI6IihgWQjpo7dy5WrVqFvn37auwYLKRIlxw5cgQ7d+6En58fDAwMUKlSJbRo0QJWVlYICwtDmzZttJ0iERUhmrl9iDTOxMRE7qi36sSro6RLXr9+DTs7OwCAra0tUlJSAABeXl6Ijo7WZmpEVASxANJRISEhWLhwoUaP8e+//7L/D+mMKlWq4NatWwCAGjVq4LfffsOjR4+wdOlSlCtXTsvZEVFRw07QOqpjx444cuQISpcujerVq+foBL19+3aZ576+vvm+pMVvy6SL1q9fjw8fPqBv3764dOkSWrVqhefPn8PY2Bi///47unbtqu0UiagIYR8gHWVjY4NOnTrle/sOHTpI///27Vv8+uuv8PDwgL+/PwDg7NmzuH79On788Ud1p0pUKHr27Cn9f82aNfHw4UPcvHkTFStWRJkyZbSYGREVRWwB0kMDBgxAuXLlMHXqVJnlkyZNQmJiIlatWqWlzIhUs3LlSsyfPx937twBALi7u2Po0KEYMGCAljMjoqKGBZCOS0lJkfZ7qFKlCsqWLZtnjLW1NS5evAh3d3eZ5Xfu3IGfnx9SU1M1kiuRJk2cOBHz5s1DcHCwtGUzKioKixYtwrBhwzBlyhQtZ0hERQkvgemo169fIzg4GGvXrpXOBG9oaIjevXtj4cKFMDc3VxhrZmaG06dP5yiATp8+DVNTU43mTaQpS5YswfLly9G9e3fpsnbt2sHb2xvBwcEsgIhIBgsgHRUaGorjx49j9+7d0tvhT506hSFDhmD48OFYsmSJwtihQ4fihx9+QHR0NOrUqQMAOHfuHFatWoUJEyYUSv5E6vb+/Xv4+fnlWF6rVi18+PBBCxkRUVHGS2A6qkyZMti6dSuaNGkis/zo0aPo0qWLdAwURbZs2YIFCxYgLi4OAFCtWjWEhISgS5cumkqZSKOCg4NhZGSUYxqYn376CW/evMHixYu1lBkRFUVsAdJR6enpsLe3z7Hczs4O6enpecZ36dKFxQ7pvNDQUOn/JRIJVqxYgYMHD6JevXoAPrZsJiQkoHfv3tpKkYiKKLYA6ajmzZujdOnSWLt2rbTfzps3b9CnTx+8ePEChw8fVhg7ceJENG3aFP7+/uzzQzqtadOm+dpOIpHgyJEjGs6GiHQJCyAddfXqVbRq1Qrv3r1DjRo1AACXL1+GiYkJDh48iOrVqyuMbdGiBaKiovDhwwfUrl0bAQEBaNKkCRo0aAAzM7PCOgUiIiKtYQGkw9LT0/HHH3/g5s2bAD724+nRo0e+ipgPHz7g3LlzOHHiBI4fP44zZ87g3bt3qF27Nk6dOqXp1ImIiLSKfYB0VFhYGOzt7TFw4ECZ5atWrUJKSgpGjRqVa3yJEiXQoEEDlC1bFqVKlULJkiWxY8cOaTFFRERUnHEyVB3122+/oWrVqjmWV69eHUuXLs01dtmyZfjmm29Qvnx51K9fHxEREWjYsCEuXryY591jRERExQEvgekoU1NTxMXFwcXFRWZ5fHw8PDw88PbtW4WxBgYGKFu2LIYPH44ff/wRlpaWmk6XiIioSGELkI5ycnLC6dOncyw/ffo0HB0dc43dvn07evTogU2bNqFs2bKoX78+xo4di4MHD+brFnoiIiJdxz5AOmrgwIEYOnQo3r9/j2bNmgEAIiMjMXLkSAwfPjzX2A4dOkhnh09NTcXJkyfx559/4ssvv4SBgUGurUdERETFAQsgHTVixAg8f/4cP/74IzIyMgB8vCw2atQojBkzJs/458+f4/jx4zh27BiOHTuG69evw9bWFo0aNdJ06kRERFrHPkA67tWrV4iLi4OZmRnc3d1hYmKSZ4yXlxfi4uJga2uLxo0bo0mTJggICIC3t3chZExERKR9LID00OLFixEQEABPT09kv/wSiUTLWRERERUedoLWQ4MGDcK5c+fg6ekJU1NTmJqawtPTEytWrNB2akRERIWCfYD00MSJEzFv3jwEBwfD398fABAVFYVhw4YhISEBU6ZM0XKGREREmsVLYHqobNmy+OWXX9C9e3eZ5Rs3bkRwcDCePXumpcyIiIgKBy+B6aH379/Dz88vx/JatWrhw4cPWsiIiIiocLEA0kO9evXCkiVLcixftmwZevTooYWMiIiIChf7AOmJ0NBQ6f8lEglWrFiBgwcPol69egCAc+fOISEhAb1799ZWikRERIWGfYD0RNOmTfO1nUQiwZEjRzScDRERkXaxACIiIiK9wz5AREREpHdYABH9v/buL6SpPo7j+PtMsrFVE2E0A/9MjAoy2VAqSaToz4J1E9KgoBVmmERIhHk1GymSsBVLVpHgbrzIuqiQgjAIyrtCgoiktJEX3SXFknKu57mIxrPs6enJ/oz2ecGBnd/O+e3Lufrsd36/c0REJOcoAImIiEjOUQASERGRnKMAJCIiIjlHAUhEfrt9+/ZhGMac7dmzZ/PuOxaLUVBQMP8iReSPogchikhW8Hg89Pf3Z7TZ7fbfVM2XJZNJFixY8LvLEJEfQCNAIpIVFi5ciMPhyNjy8vK4du0abrcbs9lMeXk5wWAw45114XCYyspKrFYrxcXFtLS0kEgkALhz5w779+/n9evX6VGlEydOAB8f+nn16tWMGgoKCojFYgDE43EMw+DSpUvU19djNpsZGBgAoK+vj1WrVmE2m1m5ciXRaDTdx8zMDIcPH6aoqAiz2UxpaSnd3d0/78KJyHfRCJCIZK27d++yd+9eIpEIdXV1jI+Pc/DgQQA6OjoAMJlMRCIRnE4nExMTtLS00NbWRjQapba2ljNnzhAIBBgbGwNg0aJF/6uG9vZ2QqEQLpcrHYICgQC9vb24XC5GR0dpamrCarXi9/uJRCJcv36dwcFBSkpKmJycZHJy8sdeGBGZNwUgEckKQ0NDGeFk+/btTE1N0d7ejt/vB6C8vJyTJ0/S1taWDkCtra3pc8rKyujs7KS5uZloNEp+fj42mw3DMHA4HN9VV2trKzt37kzvd3R0EAqF0m1Op5PHjx9z4cIF/H4/L168YPny5WzYsAHDMCgtLf2u3xWRn0sBSESywsaNGzl37lx632q1smbNGkZGRujq6kq3p1Ip3r17x/T0NBaLheHhYbq7u3ny5Alv3rxhdnY24/v5qq6uTn9++/Yt4+PjNDY20tTUlG6fnZ3FZrMBHyd0b9myhRUrVuDxePB6vWzdunXedYjIj6UAJCJZwWq1UlFRkdGWSCQIBoMZIzCfmM1m4vE4Xq+XQ4cO0dXVRWFhIffu3aOxsZGZmZmvBiDDMPj8VYjJZPKLdf2zHoCLFy+ydu3ajOPy8vIAcLvdPH/+nJs3bzI8PMyuXbvYvHkzV65c+Y8rICK/kgKQiGQtt9vN2NjYnGD0yYMHD/jw4QOhUAiT6eOajsHBwYxj8vPzSaVSc8612+28fPkyvf/06VOmp6e/Ws/SpUtZtmwZExMT7Nmz51+PW7JkCT6fD5/PR0NDAx6Ph1evXlFYWPjV/kXk11EAEpGsFQgE8Hq9lJSU0NDQgMlk4uHDhzx69IjOzk4qKipIJpOcPXuWHTt2MDIywvnz5zP6KCsrI5FIcPv2baqqqrBYLFgsFjZt2kRvby/r168nlUpx/Pjxb1riHgwGOXLkCDabDY/Hw/v377l//z5TU1McPXqUcDhMUVERLpcLk8nE5cuXcTgcehaRSJbRMngRyVrbtm1jaGiIW7duUVNTw7p16zh9+nR6YnFVVRXhcJhTp06xevVqBgYG5iw5r62tpbm5GZ/Ph91up6enB4BQKERxcTF1dXXs3r2bY8eOfdOcoQMHDtDX10d/fz+VlZXU19cTi8VwOp0ALF68mJ6eHqqrq6mpqSEej3Pjxo30CJWIZAfjr89vgouIiIj84fSXRERERHKOApCIiIjkHAUgERERyTkKQCIiIpJzFIBEREQk5ygAiYiISM5RABIREZGcowAkIiIiOUcBSERERHKOApCIiIjkHAUgERERyTl/AxO1pLf6AymIAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot 30 most important features when all features are taken into consideration\n",
    "\n",
    "plot_features_importance(X_train, X_test, y_train, y_test, [])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8978925773964939\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot 30 most important features when without VPM score\n",
    "\n",
    "plot_features_importance(df_no_vpm, X_test, y_train, y_test, vpm_patent_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9188735546437896\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot 30 most important features when without google embedings\n",
    "plot_features_importance(df_no_google, X_test, y_train, y_test, google_embeddings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8956546064901156\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# withouth future information and VPM scores\n",
    "\n",
    "plot_features_importance(df_no_future_vpm, X_test, y_train, y_test, future_info_vpm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## BERT to encode abstracts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define your dataset class\n",
    "class CustomDataset(Dataset):\n",
    "    def __init__(self, features, texts, labels, tokenizer, max_length):\n",
    "        self.features = features\n",
    "        self.texts = texts\n",
    "        self.labels = labels\n",
    "        self.tokenizer = tokenizer\n",
    "        self.max_length = max_length\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.features)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        text = list(self.texts.iloc[idx])\n",
    "        label = int(self.labels[idx])\n",
    "\n",
    "        # tokenize text of interest\n",
    "        encoding = self.tokenizer(text, truncation=True, padding='max_length', max_length=self.max_length, return_tensors='pt')\n",
    "        \n",
    "        return {\n",
    "            'features': torch.tensor(self.features.iloc[idx]),\n",
    "            'input_ids': encoding['input_ids'].flatten(),\n",
    "            'attention_mask': encoding['attention_mask'].flatten(),\n",
    "            'label': torch.tensor(label, dtype=torch.long)\n",
    "        }\n",
    "\n",
    "\n",
    "class CustomBERTClassifier(torch.nn.Module):\n",
    "    def __init__(self, num_classes, feature_dim):\n",
    "        super(CustomBERTClassifier, self).__init__()\n",
    "        self.bert = BertModel.from_pretrained('bert-base-uncased', num_labels=num_classes)\n",
    "        \n",
    "        # Freeze all BERT layers except the last three\n",
    "        for param in self.bert.parameters():\n",
    "            param.requires_grad = False\n",
    "\n",
    "        # unfreeze last encoder layer\n",
    "        for param in self.bert.encoder.layer[-1:].parameters():\n",
    "            param.requires_grad = True\n",
    "            \n",
    "        self.fc = torch.nn.Linear(feature_dim + self.bert.config.hidden_size, 128)  \n",
    "        self.final_layer = torch.nn.Linear(128, num_classes)\n",
    "\n",
    "    def forward(self, input_ids, attention_mask, features):\n",
    "        outputs = self.bert(input_ids, attention_mask=attention_mask)\n",
    "        pooled_output = outputs.pooler_output\n",
    "\n",
    "        # Concatenate BERT output with additional features\n",
    "        combined_features = torch.cat([pooled_output, features], dim=1)\n",
    "\n",
    "        # Fully connected layer\n",
    "        combined_features = torch.relu(self.fc(combined_features))\n",
    "\n",
    "        # Final classification layer\n",
    "        logits = self.final_layer(combined_features)\n",
    "\n",
    "        return logits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_clean[['abstract', 'description_text']] = text  # incorporate text\n",
    "df_clean.fillna(0, inplace=True)\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_clean, y, test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5fea8cb201de4f7e82eb9a9bbfcae120",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "tokenizer_config.json:   0%|          | 0.00/28.0 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cbf2de65d8114e75b99592fb3c0a9e93",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "vocab.txt:   0%|          | 0.00/232k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0a5c0868a6444773b67c4c5e1393e78d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "tokenizer.json:   0%|          | 0.00/466k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f7988ea961c14c84bf57ec762872dddb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "config.json:   0%|          | 0.00/570 [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cf308a3a04694fefbe57107f96494e10",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "model.safetensors:   0%|          | 0.00/440M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loaded\n"
     ]
    }
   ],
   "source": [
    "# load model\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "# device  = 'cpu'\n",
    "\n",
    "# Set up your data and model\n",
    "tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\n",
    "max_length = 128  # Adjust as needed\n",
    "num_classes = 2  # Binary classification\n",
    "\n",
    "# reset indexes, otherwise it doesn't work\n",
    "X_train.reset_index(drop=True, inplace=True)\n",
    "y_train.reset_index(drop=True, inplace=True)\n",
    "\n",
    "features = X_train.drop(['abstract', 'description_text'], axis = 1)\n",
    "\n",
    "# only using abstract as text description doesn't seem too useful and would slow down training\n",
    "texts = X_train[['abstract']]   \n",
    "\n",
    "scaler = StandardScaler()\n",
    "features = pd.DataFrame(scaler.fit_transform(features))\n",
    "\n",
    "# instanciate dataset and dataloader\n",
    "dataset = CustomDataset(features, texts, y_train, tokenizer, max_length)\n",
    "dataloader = DataLoader(dataset, batch_size=32, shuffle=True)\n",
    " \n",
    "model = CustomBERTClassifier(num_classes, len(features.columns))\n",
    "model.to(device)  # Move the model to GPU if available\n",
    " \n",
    "print('loaded')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of trainable parameters in the model: 7294594 out of 109688962 total parameter\n"
     ]
    }
   ],
   "source": [
    "num_trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
    "tot_params = sum(p.numel() for p in model.parameters())\n",
    "\n",
    "print(\"Number of trainable parameters in the model:\", num_trainable_params, f'out of {tot_params} total parameter')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "if torch.cuda.is_available() == False: \n",
    "    model = torch.load('models/bert_trained.pth', map_location=torch.device('cpu'))\n",
    "\n",
    "elif os.path.isfile('models/bert_trained.pth'): \n",
    "    model = torch.load('models/bert_trained.pth')\n",
    "    model.to(device) \n",
    "\n",
    "else: \n",
    "\n",
    "    losses = []\n",
    "    optimizer = AdamW(model.parameters(), lr=3e-5)\n",
    "\n",
    "    num_epochs = 12 # Adjust as needed\n",
    "\n",
    "    for epoch in range(num_epochs):\n",
    "        # print('epoch:', epoch)\n",
    "        model.train()\n",
    "\n",
    "        # Iterate through batches in the data loader\n",
    "        for num, batch in enumerate(dataloader):\n",
    "            input_ids = batch['input_ids'].to(device)\n",
    "            attention_mask = batch['attention_mask'].to(device)\n",
    "            features = batch['features'].to(device)\n",
    "            labels = batch['label'].to(device)\n",
    "\n",
    "            optimizer.zero_grad()\n",
    "\n",
    "            features = features.to(torch.float32)       \n",
    "\n",
    "            # Forward pass through the model\n",
    "            outputs = model(input_ids, attention_mask=attention_mask, features=features)\n",
    "\n",
    "            # Compute the cross-entropy loss\n",
    "            loss = torch.nn.functional.cross_entropy(outputs, labels)\n",
    "\n",
    "            # Append the loss to the list for later analysis\n",
    "            losses.append(loss.item())\n",
    "\n",
    "            # Backward pass and optimization step\n",
    "            loss.backward()\n",
    "            optimizer.step()    \n",
    "\n",
    "            # Explicitly release memory to prevent RAM overflow\n",
    "            del input_ids, attention_mask, features, labels, outputs, loss\n",
    "\n",
    "    plt.plot(pd.Series(losses).rolling(20, min_periods=1).sum())\n",
    "    plt.xlabel('iterations')\n",
    "    plt.ylabel('cross entropy loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Expected execution time: 1 minutes\n",
    "# Evaluation\n",
    "\n",
    "# Reset index to ensure correct alignment\n",
    "X_test.reset_index(drop=True, inplace=True)\n",
    "y_test.reset_index(drop=True, inplace=True)\n",
    "\n",
    "# Extract features and texts for testing\n",
    "features_test = X_test.drop(['abstract', 'description_text'], axis = 1)\n",
    "texts_test = X_test[['abstract']]\n",
    "\n",
    "# Standardize numerical features\n",
    "scaler = StandardScaler()\n",
    "features_test = pd.DataFrame(scaler.fit_transform(features_test))\n",
    "\n",
    "# Create a dataset and dataloader for testing\n",
    "test_dataset = CustomDataset(features_test, texts_test, y_test, tokenizer, max_length)\n",
    "dataloader_test = DataLoader(test_dataset, batch_size=32) # withour shuflling in order to have prediction in the right order\n",
    "\n",
    "# Predictions\n",
    "y_pred = []\n",
    "\n",
    "with torch.no_grad():\n",
    "    for num, batch in enumerate(dataloader_test): \n",
    "        input_ids = batch['input_ids'].to(device)\n",
    "        attention_mask = batch['attention_mask'].to(device)\n",
    "        features = batch['features'].to(device)\n",
    "        labels = batch['label'].to(device)\n",
    "\n",
    "        features = features.to(torch.float32)   \n",
    "        y_pred.append(model(input_ids, attention_mask =attention_mask, features=features))\n",
    "\n",
    "        # Explicitly release memory to prevent RAM overflow\n",
    "        del input_ids, attention_mask, features, labels\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_pred = []\n",
    "\n",
    "for pred in y_pred: \n",
    "    final_pred.extend(pred.argmax(dim= 1).to('cpu').tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.9038604998135025\n",
      "precision 0.9026772030656401\n",
      "recall 0.9026263646899115\n",
      "f1 0.9026517107451546\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "# print('f1-score', f1_score(y_test, final_pred))\n",
    "precision, recall, f_score, support = precision_recall_fscore_support(y_test, final_pred, average='macro')\n",
    "print('accuracy', (final_pred == y_test).sum() / len(y_test))\n",
    "print('precision', precision)\n",
    "print('recall', recall)\n",
    "print('f1', f_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model if it didn't exists\n",
    "if not os.path.isfile('models/bert_trained.pth'): \n",
    "    torch.save(model, 'models/bert_trained.pth')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# BERT without google embedings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loaded\n"
     ]
    }
   ],
   "source": [
    "# load model\n",
    "df_clean[['abstract', 'description_text']] = text  # incorporate text\n",
    "df_clean.fillna(0, inplace=True)\n",
    "df_no_google = modify_df(df_clean, google_embeddings)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(df_no_google, y, test_size=0.20, random_state=42)\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "# device  = 'cpu'\n",
    "\n",
    "# Set up your data and model\n",
    "tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\n",
    "max_length = 128  # Adjust as needed\n",
    "num_classes = 2  # Binary classification\n",
    "\n",
    "# reset indexes, otherwise it doesn't work\n",
    "X_train.reset_index(drop=True, inplace=True)\n",
    "y_train.reset_index(drop=True, inplace=True)\n",
    "\n",
    "features = X_train.drop(['abstract', 'description_text'], axis = 1)\n",
    "\n",
    "# only using abstract as text description doesn't seem too useful and would slow down training\n",
    "texts = X_train[['abstract']]   \n",
    "\n",
    "scaler = StandardScaler()\n",
    "features = pd.DataFrame(scaler.fit_transform(features))\n",
    "\n",
    "# instanciate dataset and dataloader\n",
    "dataset = CustomDataset(features, texts, y_train, tokenizer, max_length)\n",
    "dataloader = DataLoader(dataset, batch_size=32, shuffle=True)\n",
    "\n",
    "print('loaded')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# training bert without google embedings\n",
    "\n",
    "if torch.cuda.is_available() == False: \n",
    "    model = torch.load('models/bert_trained_no_ge.pth', map_location=torch.device('cpu'))\n",
    "\n",
    "elif os.path.isfile('models/bert_trained_no_ge.pth'): \n",
    "    print('using saved parameters')\n",
    "    model = torch.load('models/bert_trained_no_ge.pth')\n",
    "    model.to(device) \n",
    "\n",
    "else: \n",
    "    print('training from scratch')\n",
    "    losses = []\n",
    "    optimizer = AdamW(model.parameters(), lr=3e-5)\n",
    "\n",
    "    num_epochs = 12 # Adjust as needed\n",
    "\n",
    "    for epoch in range(num_epochs):\n",
    "        print('epoch:', epoch)\n",
    "        model.train()\n",
    "\n",
    "        # Iterate through batches in the data loader\n",
    "        for num, batch in enumerate(dataloader):\n",
    "            input_ids = batch['input_ids'].to(device)\n",
    "            attention_mask = batch['attention_mask'].to(device)\n",
    "            features = batch['features'].to(device)\n",
    "            labels = batch['label'].to(device)\n",
    "\n",
    "            optimizer.zero_grad()\n",
    "\n",
    "            features = features.to(torch.float32)       \n",
    "\n",
    "            # Forward pass through the model\n",
    "            outputs = model(input_ids, attention_mask=attention_mask, features=features)\n",
    "\n",
    "            # Compute the cross-entropy loss\n",
    "            loss = torch.nn.functional.cross_entropy(outputs, labels)\n",
    "\n",
    "            # Append the loss to the list for later analysis\n",
    "            losses.append(loss.item())\n",
    "\n",
    "            # Backward pass and optimization step\n",
    "            loss.backward()\n",
    "            optimizer.step()    \n",
    "\n",
    "            # Explicitly release memory to prevent RAM overflow\n",
    "            del input_ids, attention_mask, features, labels, outputs, loss\n",
    "\n",
    "    plt.plot(pd.Series(losses).rolling(20, min_periods=1).sum())\n",
    "    plt.xlabel('iterations')\n",
    "    plt.ylabel('cross entropy loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Expected execution time: 2 minutes\n",
    "# Evaluation\n",
    "\n",
    "# Reset index to ensure correct alignment\n",
    "X_test.reset_index(drop=True, inplace=True)\n",
    "y_test.reset_index(drop=True, inplace=True)\n",
    "\n",
    "\n",
    "# Extract features and texts for testing\n",
    "features_test = X_test.drop(['abstract', 'description_text'], axis = 1)\n",
    "texts_test = X_test[['abstract']]\n",
    "\n",
    "# Standardize numerical features\n",
    "scaler = StandardScaler()\n",
    "features_test = pd.DataFrame(scaler.fit_transform(features_test))\n",
    "\n",
    "# Create a dataset and dataloader for testing\n",
    "test_dataset = CustomDataset(features_test, texts_test, y_test, tokenizer, max_length)\n",
    "dataloader_test = DataLoader(test_dataset, batch_size=32) # withour shuflling in order to have prediction in the right order\n",
    "\n",
    "# Predictions\n",
    "y_pred = []\n",
    "\n",
    "with torch.no_grad():\n",
    "    for num, batch in enumerate(dataloader_test): \n",
    "        input_ids = batch['input_ids'].to(device)\n",
    "        attention_mask = batch['attention_mask'].to(device)\n",
    "        features = batch['features'].to(device)\n",
    "        labels = batch['label'].to(device)\n",
    "\n",
    "        features = features.to(torch.float32)   \n",
    "        y_pred.append(model(input_ids, attention_mask =attention_mask, features=features))\n",
    "\n",
    "        # Explicitly release memory to prevent RAM overflow\n",
    "        del input_ids, attention_mask, features, labels\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy 0.9056322267810518\n",
      "precision 0.9037562112759829\n",
      "recall 0.9062987703686612\n",
      "f1 0.9047775276576897\n"
     ]
    }
   ],
   "source": [
    "final_pred = []\n",
    "\n",
    "for pred in y_pred: \n",
    "    final_pred.extend(pred.argmax(dim= 1).to('cpu').tolist())\n",
    "\n",
    "# print('f1-score', f1_score(y_test, final_pred))\n",
    "precision, recall, f_score, support = precision_recall_fscore_support(y_test, final_pred, average='macro')\n",
    "print('accuracy', (final_pred == y_test).sum() / len(y_test))\n",
    "print('precision', precision)\n",
    "print('recall', recall)\n",
    "print('f1', f_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save model if it didn't exists\n",
    "if not os.path.isfile('models/bert_trained_no_ge.pth'): \n",
    "    torch.save(model, 'models/bert_trained_no_ge.pth')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.11.2 64-bit",
   "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": "3.11.2"
  },
  "vscode": {
   "interpreter": {
    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}