{ "cells": [ { "cell_type": "markdown", "id": "30487fe4-5659-41fa-b2a7-7ca9d677b169", "metadata": {}, "source": [ "# Prediction:" ] }, { "cell_type": "code", "execution_count": 34, "id": "86ad082a-d600-493f-aac4-310e520cfe84", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "[nltk_data] Downloading package stopwords to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package stopwords is already up-to-date!\n", "[nltk_data] Downloading package punkt to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package punkt is already up-to-date!\n", "[nltk_data] Downloading package wordnet to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package wordnet is already up-to-date!\n", "[nltk_data] Downloading package omw-1.4 to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package omw-1.4 is already up-to-date!\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Accuracy of the loaded model: 0.4909329829172142\n" ] } ], "source": [ "import pandas as pd\n", "import nltk\n", "from nltk.corpus import stopwords\n", "from nltk.tokenize import word_tokenize\n", "from string import punctuation\n", "import re\n", "import joblib\n", "from sklearn.metrics import accuracy_score\n", "\n", "# Load data\n", "df = pd.read_csv('./data.csv', usecols=['title', 'news'])\n", "\n", "# Download necessary NLTK data\n", "nltk.download('stopwords')\n", "nltk.download('punkt')\n", "nltk.download('wordnet')\n", "nltk.download('omw-1.4')\n", "\n", "# Define stop words\n", "stop_words = set(stopwords.words('english'))\n", "\n", "# Text cleaning function\n", "def clean_text(text):\n", " # Tokenize and lowercase\n", " words = word_tokenize(text.lower())\n", " # Remove stop words, punctuation, and digits\n", " words = [\n", " word for word in words\n", " if word not in stop_words and word not in punctuation and not re.search(r'\\d', word)\n", " ]\n", " # Rejoin words\n", " return ' '.join(words)\n", "\n", "# Apply cleaning function to the 'title' column\n", "df['title'] = df['title'].apply(clean_text)\n", "\n", "# Define features (X) and labels (y)\n", "X = df['title'] # Use the cleaned titles as features\n", "y = df['news'].apply(lambda x: 1 if x == 'fox' else 0) # Convert 'news' to binary labels\n", "\n", "# Load the saved vectorizer and model\n", "loaded_vectorizer = joblib.load('vectorizer_bong.pkl')\n", "loaded_model = joblib.load('naive_bayes_model.pkl')\n", "\n", "# Transform the text data using the loaded vectorizer\n", "X_test = loaded_vectorizer.transform(X)\n", "\n", "# Use the loaded model to make predictions\n", "y_pred_loaded = loaded_model.predict(X_test)\n", "\n", "# Evaluate the model\n", "accuracy = accuracy_score(y, y_pred_loaded)\n", "print(\"Accuracy of the loaded model:\", accuracy)" ] }, { "cell_type": "markdown", "id": "eda06eb9-cde6-4370-b189-3d06ffe505c8", "metadata": {}, "source": [ "## Data Processing" ] }, { "cell_type": "code", "execution_count": 35, "id": "23e35ed8-027e-42f8-8902-43591ef7e34e", "metadata": {}, "outputs": [], "source": [ "!pip install geopy > delete.txt\n", "!pip install datasets > delete.txt\n", "!pip install torch torchvision datasets > delete.txt\n", "!pip install huggingface_hub > delete.txt\n", "!pip install pyhocon > delete.txt\n", "!pip install transformers > delete.txt\n", "!rm delete.txt" ] }, { "cell_type": "code", "execution_count": 36, "id": "8d91a73a-cde9-48a6-9d2a-83715ad64a12", "metadata": {}, "outputs": [], "source": [ "%matplotlib inline\n", "import warnings\n", "warnings.filterwarnings(\"ignore\") \n", "from sklearn.model_selection import train_test_split\n", "import numpy as np\n", "import pandas as pd\n", "import re\n", "from datetime import datetime\n", "\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "\n", "from string import punctuation\n", "from nltk.corpus import stopwords\n", "from nltk.tokenize import word_tokenize\n", "import nltk\n", "pd.options.display.max_colwidth = None" ] }, { "cell_type": "code", "execution_count": 37, "id": "d6e65e5d-df2b-4db6-b866-6c473ed85919", "metadata": {}, "outputs": [], "source": [ "df=pd.read_csv('./data.csv',usecols=['title', 'news'])" ] }, { "cell_type": "code", "execution_count": 38, "id": "6c291504-9f0c-46cc-aaee-9ce274ccaafc", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
titlenews
0Jack Carr recalls Gen. Eisenhower's D-Day memo about 'great and noble undertaking'fox
1Bruce Willis, Demi Moore avoided doing one thing while co-parenting, daughter saysfox
2Blinken meets Qatar PM, says Israeli actions are not 'retaliation,' but 'defending the lives of its people'fox
3Emily Blunt says her ‘toes curl’ when people tell her their kids want to act: 'I want to say, don’t do it!'fox
4'The View' co-host, CNN commentator Ana Navarro to host night 2 of Democratic National Conventionfox
\n", "
" ], "text/plain": [ " title \\\n", "0 Jack Carr recalls Gen. Eisenhower's D-Day memo about 'great and noble undertaking' \n", "1 Bruce Willis, Demi Moore avoided doing one thing while co-parenting, daughter says \n", "2 Blinken meets Qatar PM, says Israeli actions are not 'retaliation,' but 'defending the lives of its people' \n", "3 Emily Blunt says her ‘toes curl’ when people tell her their kids want to act: 'I want to say, don’t do it!' \n", "4 'The View' co-host, CNN commentator Ana Navarro to host night 2 of Democratic National Convention \n", "\n", " news \n", "0 fox \n", "1 fox \n", "2 fox \n", "3 fox \n", "4 fox " ] }, "execution_count": 38, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.head()" ] }, { "cell_type": "code", "execution_count": 39, "id": "7106c263-be37-4037-bdfc-8c8308e895fc", "metadata": { "scrolled": true }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "[nltk_data] Downloading package stopwords to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package stopwords is already up-to-date!\n", "[nltk_data] Downloading package punkt_tab to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package punkt_tab is already up-to-date!\n", "[nltk_data] Downloading package wordnet to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package wordnet is already up-to-date!\n", "[nltk_data] Downloading package omw-1.4 to\n", "[nltk_data] C:\\Users\\26656\\AppData\\Roaming\\nltk_data...\n", "[nltk_data] Package omw-1.4 is already up-to-date!\n" ] } ], "source": [ "nltk.download('stopwords')\n", "nltk.download('punkt_tab')\n", "nltk.download('wordnet')\n", "nltk.download('omw-1.4')\n", "\n", "stop_words = set(stopwords.words('english'))\n", "\n", "def clean_text(text):\n", " # Remove punctuation\n", " words = word_tokenize(text.lower())\n", " # Remove stop words and punctuation\n", " words = [\n", " word for word in words\n", " if word.lower() not in stop_words and word not in punctuation and not re.search(r'\\d', word)\n", " ]\n", " \n", " # Rejoin words\n", " return ' '.join(words)\n", "\n", "# Apply the function to the 'title' column\n", "df['title'] = df['title'].apply(clean_text)" ] }, { "cell_type": "code", "execution_count": 40, "id": "91f557c2-1aab-43ee-b464-f3ef7a33556d", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
titlenews
0jack carr recalls gen. eisenhower 's d-day memo 'great noble undertakingfox
1bruce willis demi moore avoided one thing co-parenting daughter saysfox
2blinken meets qatar pm says israeli actions 'retaliation 'defending lives peoplefox
3emily blunt says ‘ toes curl ’ people tell kids want act want say ’fox
4'the view co-host cnn commentator ana navarro host night democratic national conventionfox
.........
3800trump 's lawyers seek post-election day delay court fight immunity decision fallout interference casenbc
3801treat acne scars hyperpigmentation according expertsnbc
3802best vegetarian vegan meal delivery services according expertsnbc
3803trump says presidential civilian award 'better top military honor whose recipients 'dead 'hit bulletsnbc
3804best white elephant secret santa gift ideasnbc
\n", "

3805 rows × 2 columns

\n", "
" ], "text/plain": [ " title \\\n", "0 jack carr recalls gen. eisenhower 's d-day memo 'great noble undertaking \n", "1 bruce willis demi moore avoided one thing co-parenting daughter says \n", "2 blinken meets qatar pm says israeli actions 'retaliation 'defending lives people \n", "3 emily blunt says ‘ toes curl ’ people tell kids want act want say ’ \n", "4 'the view co-host cnn commentator ana navarro host night democratic national convention \n", "... ... \n", "3800 trump 's lawyers seek post-election day delay court fight immunity decision fallout interference case \n", "3801 treat acne scars hyperpigmentation according experts \n", "3802 best vegetarian vegan meal delivery services according experts \n", "3803 trump says presidential civilian award 'better top military honor whose recipients 'dead 'hit bullets \n", "3804 best white elephant secret santa gift ideas \n", "\n", " news \n", "0 fox \n", "1 fox \n", "2 fox \n", "3 fox \n", "4 fox \n", "... ... \n", "3800 nbc \n", "3801 nbc \n", "3802 nbc \n", "3803 nbc \n", "3804 nbc \n", "\n", "[3805 rows x 2 columns]" ] }, "execution_count": 40, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df" ] }, { "cell_type": "code", "execution_count": 41, "id": "e01e6fd5-9314-4000-b467-dd718bc77acb", "metadata": {}, "outputs": [], "source": [ "train_df, test_df = train_test_split(df, test_size=0.2, random_state=41)\n", "X_train = train_df['title']\n", "y_train = train_df['news']\n", "X_test = test_df['title']\n", "y_test = test_df['news']\n", "y_train = y_train.apply(lambda x: 1 if x == 'fox' else 0)\n", "y_test = y_test.apply(lambda x: 1 if x == 'fox' else 0)\n", "accuracy_scores={}" ] }, { "cell_type": "markdown", "id": "0d65ad3b-9185-4462-a6fe-1f33ede24ce2", "metadata": {}, "source": [ "## Word Embedding" ] }, { "cell_type": "markdown", "id": "e267400b-0c90-496d-b392-fe651525150a", "metadata": {}, "source": [ "### TF-IDF/ Bag of worsd/Bag of ngrams\n" ] }, { "cell_type": "code", "execution_count": 42, "id": "12c31464-013a-4221-81d6-7d81ddb6c931", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Feature Names: ['aapi' 'aaron' 'abandon' ... 'zuckerberg' 'zyn' 'zzz']\n", "7471\n", "Feature Names: ['aapi' 'aaron' 'abandon' ... 'zuckerberg' 'zyn' 'zzz']\n", "7471\n", "Feature Names: ['aapi' 'aapi owned' 'aaron' ... 'zyn maker' 'zzz' 'zzz amazon']\n", "29715\n", "Vectorizer saved successfully.\n" ] } ], "source": [ "from sklearn.feature_extraction.text import TfidfVectorizer\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.metrics import accuracy_score\n", "from sklearn.metrics import classification_report\n", "from sklearn.feature_extraction.text import CountVectorizer\n", "\n", "vectorizer = TfidfVectorizer(max_features=10000) # Adjust max_features if needed\n", "X_train_tfidf = vectorizer.fit_transform(X_train)\n", "X_test_tfidf = vectorizer.transform(X_test)\n", "print(\"Feature Names:\", vectorizer.get_feature_names_out())\n", "print(len(vectorizer.vocabulary_))\n", "\n", "vectorizer = CountVectorizer() # Adjust max_features if needed\n", "X_train_bow = vectorizer.fit_transform(X_train)\n", "X_test_bow = vectorizer.transform(X_test)\n", "print(\"Feature Names:\", vectorizer.get_feature_names_out())\n", "print(len(vectorizer.vocabulary_))\n", "\n", "vectorizer = CountVectorizer(ngram_range=(1, 2)) # Adjust max_features&ngram_range if needed\n", "# Fit and transform the training data, then transform the test data\n", "X_train_bong = vectorizer.fit_transform(X_train)\n", "X_test_bong = vectorizer.transform(X_test)\n", "# Display the feature names (words and n-grams)\n", "print(\"Feature Names:\", vectorizer.get_feature_names_out())\n", "print(len(vectorizer.vocabulary_))\n", "joblib.dump(vectorizer, 'vectorizer_bong.pkl')\n", "print(\"Vectorizer saved successfully.\")\n" ] }, { "cell_type": "markdown", "id": "d9c46662-7c07-4a4f-b2dc-803a39153a46", "metadata": {}, "source": [ "### Word2Vec" ] }, { "cell_type": "code", "execution_count": 43, "id": "1564acd6-ea3b-4292-8aad-94cc07e179ac", "metadata": {}, "outputs": [], "source": [ "from gensim.models import KeyedVectors\n", "#install scipy<1.13 to be compatible for gensim\n", "#!pip install \"scipy<1.13\"\n", "word2vec_model = KeyedVectors.load_word2vec_format('./GoogleNews-vectors-negative300.bin.gz', binary=True,limit=500000)\n", "\n", "# Tokenize your text data\n", "X_train_tokenized = [sentence.split() for sentence in X_train]\n", "X_test_tokenized = [sentence.split() for sentence in X_test]\n", "\n", "# Define a function to average word vectors for each sentence\n", "def average_word_vectors(sentence, model, vector_size):\n", " words = [word for word in sentence if word in model]\n", " return np.mean(model[words], axis=0)\n", "# Apply average word vectors on training and test sets\n", "X_train_word2vec = np.array([average_word_vectors(sentence, word2vec_model, 300) for sentence in X_train_tokenized])\n", "X_test_word2vec = np.array([average_word_vectors(sentence, word2vec_model, 300) for sentence in X_test_tokenized])" ] }, { "cell_type": "markdown", "id": "be0674db-2da7-4347-8552-25ce83fd4337", "metadata": {}, "source": [ "### Glove" ] }, { "cell_type": "code", "execution_count": 44, "id": "311c869b-a8cf-4185-9619-9c768ac05e8b", "metadata": {}, "outputs": [], "source": [ "glove_file = \"./glove.6B.100d.txt\"\n", "\n", "# Load GloVe embeddings\n", "def load_glove_embeddings(file_path):\n", " embeddings = {}\n", " with open(file_path, \"r\", encoding=\"utf-8\") as f:\n", " for line in f:\n", " values = line.strip().split(' ')\n", " word = values[0]\n", " try:\n", " vector = np.asarray(values[1:], dtype=\"float32\")\n", " embeddings[word] = vector\n", " except ValueError:\n", " print(f\"Skipping line with invalid vector for word: {word}\")\n", " return embeddings\n", "glove_embeddings = load_glove_embeddings(glove_file)" ] }, { "cell_type": "code", "execution_count": 45, "id": "425dc076-76cc-45fd-93c4-23ab6aea6cd0", "metadata": {}, "outputs": [], "source": [ "\n", "def sentence_to_glove(sentence, embeddings):\n", " words = sentence.split()\n", " vectors = [embeddings[word] for word in words if word in embeddings]\n", " if vectors:\n", " return np.mean(vectors, axis=0)\n", " else:\n", " return np.zeros(300) # Return a zero vector if no words are in embeddings\n", "\n", "# Transform an entire dataset (e.g., train and test sets)\n", "X_train_glove = np.array([sentence_to_glove(sentence, glove_embeddings) for sentence in X_train])\n", "X_test_glove = np.array([sentence_to_glove(sentence, glove_embeddings) for sentence in X_test])\n", "\n", "glove_embeddings = load_glove_embeddings(glove_file)" ] }, { "cell_type": "markdown", "id": "84133d6c-d0d1-4ae4-8912-afed2401675d", "metadata": {}, "source": [ "## Model" ] }, { "cell_type": "markdown", "id": "6f54068e-2318-4174-b2a7-38583c2b141f", "metadata": {}, "source": [ "### LogisticRegression" ] }, { "cell_type": "code", "execution_count": 46, "id": "5cc91140-2e3c-477f-8d7e-b8eed08941db", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Tfidf With Logistic Regression\n", "Accuracy: 0.7963\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.79 0.72 0.75 330\n", " 1 0.80 0.85 0.83 431\n", "\n", " accuracy 0.80 761\n", " macro avg 0.80 0.79 0.79 761\n", "weighted avg 0.80 0.80 0.79 761\n", "\n", "Bag Of Word With Logistic Regression\n", "Accuracy: 0.7911\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.75 0.78 0.76 330\n", " 1 0.83 0.80 0.81 431\n", "\n", " accuracy 0.79 761\n", " macro avg 0.79 0.79 0.79 761\n", "weighted avg 0.79 0.79 0.79 761\n", "\n", "Bag Of N Grams With Logistic Regression\n", "Accuracy: 0.8003\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.76 0.79 0.77 330\n", " 1 0.83 0.81 0.82 431\n", "\n", " accuracy 0.80 761\n", " macro avg 0.80 0.80 0.80 761\n", "weighted avg 0.80 0.80 0.80 761\n", "\n", "Word2Vec With Logistic Regression\n", "Accuracy: 0.7293\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.70 0.66 0.68 330\n", " 1 0.75 0.78 0.77 431\n", "\n", " accuracy 0.73 761\n", " macro avg 0.72 0.72 0.72 761\n", "weighted avg 0.73 0.73 0.73 761\n", "\n", "GloVe With Logistic Regression\n", "Accuracy: 0.7122\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.66 0.68 0.67 330\n", " 1 0.75 0.74 0.74 431\n", "\n", " accuracy 0.71 761\n", " macro avg 0.71 0.71 0.71 761\n", "weighted avg 0.71 0.71 0.71 761\n", "\n" ] } ], "source": [ "# Logistic Regression with TF-IDF\n", "model = LogisticRegression()\n", "model.fit(X_train_tfidf, y_train)\n", "y_pred = model.predict(X_test_tfidf)\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(\"Tfidf With Logistic Regression\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred))\n", "accuracy_scores[\"Tfidf With Logistic Regression\"] = accuracy\n", "\n", "# Logistic Regression with Bag of Words\n", "model = LogisticRegression()\n", "model.fit(X_train_bow, y_train)\n", "y_pred = model.predict(X_test_bow)\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(\"Bag Of Word With Logistic Regression\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred))\n", "accuracy_scores[\"Bag Of Word With Logistic Regression\"] = accuracy\n", "\n", "# Logistic Regression with Bag of N-Grams\n", "model = LogisticRegression()\n", "model.fit(X_train_bong, y_train)\n", "y_pred = model.predict(X_test_bong)\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(\"Bag Of N Grams With Logistic Regression\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred))\n", "accuracy_scores[\"Bag Of N Grams With Logistic Regression\"] = accuracy\n", "\n", "# Logistic Regression with Word2Vec\n", "model = LogisticRegression()\n", "model.fit(X_train_word2vec, y_train)\n", "y_pred = model.predict(X_test_word2vec)\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(\"Word2Vec With Logistic Regression\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred))\n", "accuracy_scores[\"Word2Vec With Logistic Regression\"] = accuracy\n", "\n", "# Logistic Regression with GloVe\n", "model = LogisticRegression()\n", "model.fit(X_train_glove, y_train)\n", "y_pred = model.predict(X_test_glove)\n", "accuracy = accuracy_score(y_test, y_pred)\n", "print(\"GloVe With Logistic Regression\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred))\n", "accuracy_scores[\"GloVe With Logistic Regression\"] = accuracy" ] }, { "cell_type": "markdown", "id": "c67027b3-faaf-4255-8fa6-f3bf8202a39a", "metadata": {}, "source": [ "### Decision Tree" ] }, { "cell_type": "code", "execution_count": 47, "id": "ad692f26-6935-4c5c-adca-2a5f38ae311e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "TFIDF with Decision Tree\n", "Accuracy: 0.7411\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.67 0.79 0.73 330\n", " 1 0.81 0.71 0.76 431\n", "\n", " accuracy 0.74 761\n", " macro avg 0.74 0.75 0.74 761\n", "weighted avg 0.75 0.74 0.74 761\n", "\n", "Bag of Words with Decision Tree\n", "Accuracy: 0.7503\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.71 0.71 0.71 330\n", " 1 0.78 0.78 0.78 431\n", "\n", " accuracy 0.75 761\n", " macro avg 0.75 0.75 0.75 761\n", "weighted avg 0.75 0.75 0.75 761\n", "\n", "Bag of N-grams with Decision Tree\n", "Accuracy: 0.7661\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.75 0.70 0.72 330\n", " 1 0.78 0.82 0.80 431\n", "\n", " accuracy 0.77 761\n", " macro avg 0.76 0.76 0.76 761\n", "weighted avg 0.77 0.77 0.77 761\n", "\n", "Word2Vec with Decision Tree\n", "Accuracy: 0.5782\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.51 0.55 0.53 330\n", " 1 0.63 0.60 0.62 431\n", "\n", " accuracy 0.58 761\n", " macro avg 0.57 0.57 0.57 761\n", "weighted avg 0.58 0.58 0.58 761\n", "\n", "GloVe with Decision Tree\n", "Accuracy: 0.6386\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.58 0.63 0.60 330\n", " 1 0.69 0.65 0.67 431\n", "\n", " accuracy 0.64 761\n", " macro avg 0.64 0.64 0.64 761\n", "weighted avg 0.64 0.64 0.64 761\n", "\n" ] } ], "source": [ "from sklearn.tree import DecisionTreeClassifier\n", "from sklearn.metrics import accuracy_score, classification_report\n", "\n", "# Decision Tree with TF-IDF\n", "dt_model = DecisionTreeClassifier()\n", "dt_model.fit(X_train_tfidf, y_train)\n", "y_pred_dt = dt_model.predict(X_test_tfidf)\n", "accuracy = accuracy_score(y_test, y_pred_dt)\n", "print(\"TFIDF with Decision Tree\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_dt))\n", "accuracy_scores[\"TFIDF with Decision Tree\"] = accuracy\n", "\n", "# Decision Tree with Bag of Words\n", "dt_model = DecisionTreeClassifier()\n", "dt_model.fit(X_train_bow, y_train)\n", "y_pred_dt_bow = dt_model.predict(X_test_bow)\n", "accuracy = accuracy_score(y_test, y_pred_dt_bow)\n", "print(\"Bag of Words with Decision Tree\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_dt_bow))\n", "accuracy_scores[\"Bag of Words with Decision Tree\"] = accuracy\n", "\n", "# Decision Tree with Bag of N-grams\n", "dt_model = DecisionTreeClassifier()\n", "dt_model.fit(X_train_bong, y_train)\n", "y_pred_dt_bong = dt_model.predict(X_test_bong)\n", "accuracy = accuracy_score(y_test, y_pred_dt_bong)\n", "print(\"Bag of N-grams with Decision Tree\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_dt_bong))\n", "accuracy_scores[\"Bag of N-grams with Decision Tree\"] = accuracy\n", "\n", "# Decision Tree with Word2Vec\n", "dt_model = DecisionTreeClassifier()\n", "dt_model.fit(X_train_word2vec, y_train)\n", "y_pred_dt_word2vec = dt_model.predict(X_test_word2vec)\n", "accuracy = accuracy_score(y_test, y_pred_dt_word2vec)\n", "print(\"Word2Vec with Decision Tree\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_dt_word2vec))\n", "accuracy_scores[\"Word2Vec With DecisionTree\"] = accuracy\n", "\n", "# Decision Tree with GloVe\n", "dt_model = DecisionTreeClassifier()\n", "dt_model.fit(X_train_glove, y_train)\n", "y_pred_dt_glove = dt_model.predict(X_test_glove)\n", "accuracy = accuracy_score(y_test, y_pred_dt_glove)\n", "print(\"GloVe with Decision Tree\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_dt_glove))\n", "accuracy_scores[\"GloVe with Decision Tree\"] = accuracy" ] }, { "cell_type": "markdown", "id": "97b8ce80-3405-4605-b7ee-fdd4cc5f7981", "metadata": {}, "source": [ "### Random Forest" ] }, { "cell_type": "code", "execution_count": 48, "id": "23e5725e-cfb3-40f3-b120-6bfac8054b00", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "TFIDF with Random Forest\n", "Accuracy: 0.7792\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.74 0.75 0.75 330\n", " 1 0.81 0.80 0.80 431\n", "\n", " accuracy 0.78 761\n", " macro avg 0.78 0.78 0.78 761\n", "weighted avg 0.78 0.78 0.78 761\n", "\n", "Bag of Words with Random Forest\n", "Accuracy: 0.7963\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.76 0.78 0.77 330\n", " 1 0.83 0.81 0.82 431\n", "\n", " accuracy 0.80 761\n", " macro avg 0.79 0.79 0.79 761\n", "weighted avg 0.80 0.80 0.80 761\n", "\n", "Bag of N-grams with Random Forest\n", "Accuracy: 0.7911\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.74 0.79 0.77 330\n", " 1 0.83 0.79 0.81 431\n", "\n", " accuracy 0.79 761\n", " macro avg 0.79 0.79 0.79 761\n", "weighted avg 0.79 0.79 0.79 761\n", "\n", "Word2Vec with Random Forest\n", "Accuracy: 0.7148\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.67 0.68 0.68 330\n", " 1 0.75 0.74 0.75 431\n", "\n", " accuracy 0.71 761\n", " macro avg 0.71 0.71 0.71 761\n", "weighted avg 0.72 0.71 0.72 761\n", "\n", "GloVe with Random Forest\n", "Accuracy: 0.7293\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.70 0.66 0.68 330\n", " 1 0.75 0.78 0.77 431\n", "\n", " accuracy 0.73 761\n", " macro avg 0.72 0.72 0.72 761\n", "weighted avg 0.73 0.73 0.73 761\n", "\n" ] } ], "source": [ "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.metrics import accuracy_score, classification_report\n", "\n", "# Random Forest with TF-IDF\n", "rf_model = RandomForestClassifier()\n", "rf_model.fit(X_train_tfidf, y_train)\n", "y_pred_rf = rf_model.predict(X_test_tfidf)\n", "accuracy = accuracy_score(y_test, y_pred_rf)\n", "print(\"TFIDF with Random Forest\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_rf))\n", "accuracy_scores[\"TFIDF with Random Forest\"] = accuracy\n", "\n", "# Random Forest with Bag of Words\n", "rf_model = RandomForestClassifier()\n", "rf_model.fit(X_train_bow, y_train)\n", "y_pred_rf_bow = rf_model.predict(X_test_bow)\n", "accuracy = accuracy_score(y_test, y_pred_rf_bow)\n", "print(\"Bag of Words with Random Forest\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_rf_bow))\n", "accuracy_scores[\"Bag of Words with Random Forest\"] = accuracy\n", "\n", "# Random Forest with Bag of N-grams\n", "rf_model = RandomForestClassifier()\n", "rf_model.fit(X_train_bong, y_train)\n", "y_pred_rf_bong = rf_model.predict(X_test_bong)\n", "accuracy = accuracy_score(y_test, y_pred_rf_bong)\n", "print(\"Bag of N-grams with Random Forest\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_rf_bong))\n", "accuracy_scores[\"Bag of N-grams with Random Forest\"] = accuracy\n", "\n", "# Random Forest with Word2Vec\n", "rf_model = RandomForestClassifier()\n", "rf_model.fit(X_train_word2vec, y_train)\n", "y_pred_rf_w2c = rf_model.predict(X_test_word2vec)\n", "accuracy = accuracy_score(y_test, y_pred_rf_w2c)\n", "print(\"Word2Vec with Random Forest\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_rf_w2c))\n", "accuracy_scores[\"Word2Vec With RandomForest\"] = accuracy\n", "\n", "# Random Forest with GloVe\n", "rf_model = RandomForestClassifier()\n", "rf_model.fit(X_train_glove, y_train)\n", "y_pred_rf_glove = rf_model.predict(X_test_glove)\n", "accuracy = accuracy_score(y_test, y_pred_rf_glove)\n", "print(\"GloVe with Random Forest\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_rf_glove))\n", "accuracy_scores[\"GloVe with Random Forest\"] = accuracy" ] }, { "cell_type": "markdown", "id": "aee4f2d2-6d9d-4921-917c-417138843e15", "metadata": {}, "source": [ "### SVM" ] }, { "cell_type": "code", "execution_count": 49, "id": "502902cb-23fa-4a33-b244-ee7d9e24a6d8", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "TFIDF with SVM\n", "Accuracy: 0.8095\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.82 0.72 0.77 330\n", " 1 0.80 0.88 0.84 431\n", "\n", " accuracy 0.81 761\n", " macro avg 0.81 0.80 0.80 761\n", "weighted avg 0.81 0.81 0.81 761\n", "\n", "Bag of Words with SVM\n", "Accuracy: 0.7832\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.70 0.86 0.78 330\n", " 1 0.87 0.72 0.79 431\n", "\n", " accuracy 0.78 761\n", " macro avg 0.79 0.79 0.78 761\n", "weighted avg 0.80 0.78 0.78 761\n", "\n", "Bag of N-grams with SVM\n", "Accuracy: 0.5742\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.50 0.97 0.66 330\n", " 1 0.91 0.27 0.42 431\n", "\n", " accuracy 0.57 761\n", " macro avg 0.71 0.62 0.54 761\n", "weighted avg 0.74 0.57 0.53 761\n", "\n", "Word2Vec with SVM\n", "Accuracy: 0.7792\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.76 0.71 0.74 330\n", " 1 0.79 0.83 0.81 431\n", "\n", " accuracy 0.78 761\n", " macro avg 0.78 0.77 0.77 761\n", "weighted avg 0.78 0.78 0.78 761\n", "\n", "GloVe with SVM\n", "Accuracy: 0.7622\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.75 0.68 0.71 330\n", " 1 0.77 0.83 0.80 431\n", "\n", " accuracy 0.76 761\n", " macro avg 0.76 0.75 0.75 761\n", "weighted avg 0.76 0.76 0.76 761\n", "\n" ] } ], "source": [ "from sklearn.svm import SVC\n", "from sklearn.metrics import accuracy_score, classification_report\n", "\n", "# SVM with TFIDF\n", "svm_model = SVC()\n", "svm_model.fit(X_train_tfidf, y_train)\n", "y_pred_svm = svm_model.predict(X_test_tfidf)\n", "accuracy = accuracy_score(y_test, y_pred_svm)\n", "print(\"TFIDF with SVM\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_svm))\n", "accuracy_scores[\"TFIDF with SVM\"] = accuracy\n", "\n", "# SVM with Bag of Words\n", "svm_model = SVC()\n", "svm_model.fit(X_train_bow, y_train)\n", "y_pred_svm_bow = svm_model.predict(X_test_bow)\n", "accuracy = accuracy_score(y_test, y_pred_svm_bow)\n", "print(\"Bag of Words with SVM\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_svm_bow))\n", "accuracy_scores[\"Bag of Words with SVM\"] = accuracy\n", "\n", "# SVM with Bag of N-grams\n", "svm_model = SVC()\n", "svm_model.fit(X_train_bong, y_train)\n", "y_pred_svm_bong = svm_model.predict(X_test_bong)\n", "accuracy = accuracy_score(y_test, y_pred_svm_bong)\n", "print(\"Bag of N-grams with SVM\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_svm_bong))\n", "accuracy_scores[\"Bag of N-grams with SVM\"] = accuracy\n", "\n", "# SVM with Word2Vec\n", "svm_model = SVC()\n", "svm_model.fit(X_train_word2vec, y_train)\n", "y_pred_svm_w2v = svm_model.predict(X_test_word2vec)\n", "accuracy = accuracy_score(y_test, y_pred_svm_w2v)\n", "print(\"Word2Vec with SVM\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_svm_w2v))\n", "accuracy_scores[\"Word2Vec With SVM\"] = accuracy\n", "\n", "# SVM with GloVeBag of Words with Naive Bayes\n", "svm_model = SVC()\n", "svm_model.fit(X_train_glove, y_train)\n", "y_pred_svm_glove = svm_model.predict(X_test_glove)\n", "accuracy = accuracy_score(y_test, y_pred_svm_glove)\n", "print(\"GloVe with SVM\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_svm_glove))\n", "accuracy_scores[\"GloVe with SVM\"] = accuracy" ] }, { "cell_type": "markdown", "id": "15e26973-193c-45df-8496-18a10b874690", "metadata": {}, "source": [ "### Naive Bayes" ] }, { "cell_type": "code", "execution_count": 50, "id": "11ae28fd-b517-47e0-94c5-f4e7f420e9c9", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "TFIDF with Naive Bayes\n", "Accuracy: 0.8029\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.78 0.75 0.77 330\n", " 1 0.82 0.84 0.83 431\n", "\n", " accuracy 0.80 761\n", " macro avg 0.80 0.80 0.80 761\n", "weighted avg 0.80 0.80 0.80 761\n", "\n", "Bag of Words with Naive Bayes\n", "Accuracy: 0.8055\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.77 0.78 0.78 330\n", " 1 0.83 0.83 0.83 431\n", "\n", " accuracy 0.81 761\n", " macro avg 0.80 0.80 0.80 761\n", "weighted avg 0.81 0.81 0.81 761\n", "\n", "Bag of N-grams with Naive Bayes\n", "Accuracy: 0.8200\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.81 0.77 0.79 330\n", " 1 0.83 0.86 0.84 431\n", "\n", " accuracy 0.82 761\n", " macro avg 0.82 0.81 0.82 761\n", "weighted avg 0.82 0.82 0.82 761\n", "\n", "Word2Vec with Naive Bayes (GaussianNB)\n", "Accuracy: 0.6728\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.63 0.59 0.61 330\n", " 1 0.70 0.74 0.72 431\n", "\n", " accuracy 0.67 761\n", " macro avg 0.67 0.66 0.66 761\n", "weighted avg 0.67 0.67 0.67 761\n", "\n", "GloVe with Naive Bayes (GaussianNB)\n", "Accuracy: 0.6689\n", "Classification Report:\n", " precision recall f1-score support\n", "\n", " 0 0.63 0.56 0.59 330\n", " 1 0.69 0.75 0.72 431\n", "\n", " accuracy 0.67 761\n", " macro avg 0.66 0.66 0.66 761\n", "weighted avg 0.67 0.67 0.67 761\n", "\n" ] } ], "source": [ "from sklearn.naive_bayes import MultinomialNB, GaussianNB\n", "from sklearn.metrics import accuracy_score, classification_report\n", "from sklearn.metrics import precision_score, recall_score\n", "# Naive Bayes with TF-IDF\n", "nb_model = MultinomialNB()\n", "nb_model.fit(X_train_tfidf, y_train)\n", "y_pred_nb_tfidf = nb_model.predict(X_test_tfidf)\n", "accuracy = accuracy_score(y_test, y_pred_nb_tfidf)\n", "print(\"TFIDF with Naive Bayes\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_nb_tfidf))\n", "accuracy_scores[\"TFIDF with Naive Bayes\"] = accuracy\n", "\n", "# Naive Bayes with Bag of Words\n", "nb_model = MultinomialNB()\n", "nb_model.fit(X_train_bow, y_train)\n", "y_pred_nb_bow = nb_model.predict(X_test_bow)\n", "accuracy = accuracy_score(y_test, y_pred_nb_bow)\n", "print(\"Bag of Words with Naive Bayes\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_nb_bow))\n", "accuracy_scores[\"Bag of Words with Naive Bayes\"] = accuracy\n", "\n", "# Naive Bayes with Bag of N-grams\n", "nb_model = MultinomialNB()\n", "nb_model.fit(X_train_bong, y_train)\n", "y_pred_nb_bong = nb_model.predict(X_test_bong)\n", "accuracy = accuracy_score(y_test, y_pred_nb_bong)\n", "print(\"Bag of N-grams with Naive Bayes\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_nb_bong))\n", "accuracy_scores[\"Bag of N-grams with Naive Bayes\"] = accuracy\n", "precision = precision_score(y_test, y_pred_nb_bong, average='weighted')\n", "recall = recall_score(y_test, y_pred_nb_bong, average='weighted')\n", "\n", "\n", "joblib.dump(nb_model, 'naive_bayes_model.pkl')\n", "\n", "\n", "\n", "# Naive Bayes with Word2Vec (using GaussianNB)\n", "nb_model = GaussianNB()\n", "nb_model.fit(X_train_word2vec, y_train)\n", "y_pred_nb_w2v = nb_model.predict(X_test_word2vec)\n", "accuracy = accuracy_score(y_test, y_pred_nb_w2v)\n", "print(\"Word2Vec with Naive Bayes (GaussianNB)\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_nb_w2v))\n", "accuracy_scores[\"Word2Vec with Naive Bayes\"] = accuracy\n", "\n", "# Naive Bayes with GloVe (using GaussianNB)\n", "nb_model = GaussianNB()\n", "nb_model.fit(X_train_glove, y_train)\n", "y_pred_nb_glove = nb_model.predict(X_test_glove)\n", "accuracy = accuracy_score(y_test, y_pred_nb_glove)\n", "print(\"GloVe with Naive Bayes (GaussianNB)\")\n", "print(f\"Accuracy: {accuracy:.4f}\")\n", "print(\"Classification Report:\\n\", classification_report(y_test, y_pred_nb_glove))\n", "accuracy_scores[\"GloVe with Naive Bayes\"] = accuracy" ] }, { "cell_type": "code", "execution_count": null, "id": "3bf8cec4-17ee-4342-a43b-163fd1b36d68", "metadata": {}, "outputs": [], "source": [ "\n" ] }, { "cell_type": "markdown", "id": "2a7e35c1-1444-4098-8866-450af4216066", "metadata": {}, "source": [ "## LSTM" ] }, { "cell_type": "code", "execution_count": 51, "id": "5852a896-7b5d-42dc-845d-3e58aa7007c4", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/7\n", "96/96 [==============================] - 2s 18ms/step - loss: 0.6869 - accuracy: 0.5345 - val_loss: 0.6719 - val_accuracy: 0.5848\n", "Epoch 2/7\n", "96/96 [==============================] - 1s 7ms/step - loss: 0.6759 - accuracy: 0.5700 - val_loss: 0.6655 - val_accuracy: 0.5926\n", "Epoch 3/7\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.6712 - accuracy: 0.5700 - val_loss: 0.6686 - val_accuracy: 0.5940\n", "Epoch 4/7\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.6711 - accuracy: 0.5834 - val_loss: 0.6692 - val_accuracy: 0.5598\n", "Epoch 5/7\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.6671 - accuracy: 0.5949 - val_loss: 0.6579 - val_accuracy: 0.6018\n", "Epoch 6/7\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.6683 - accuracy: 0.5880 - val_loss: 0.6681 - val_accuracy: 0.6150\n", "Epoch 7/7\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.6654 - accuracy: 0.5920 - val_loss: 0.6621 - val_accuracy: 0.6084\n", "24/24 [==============================] - 0s 1ms/step - loss: 0.6621 - accuracy: 0.6084\n", "Test Loss: 0.6620948314666748\n", "Test Accuracy: 0.6084100008010864\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABKUAAAHqCAYAAADVi/1VAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdd1xV9f/A8ddlbwRkyXaCigsU90jTcKSWintXalquSjPLzH6W5chKLcVtjtL8mpscufcW3AoqIKICKhvO74+bN684UJHDeD8fj/Pw3s/9nHPe58LBz33fz9AoiqIghBBCCCGEEEIIIUQ+MlA7ACGEEEIIIYQQQghR/EhSSgghhBBCCCGEEELkO0lKCSGEEEIIIYQQQoh8J0kpIYQQQgghhBBCCJHvJCklhBBCCCGEEEIIIfKdJKWEEEIIIYQQQgghRL6TpJQQQgghhBBCCCGEyHeSlBJCCCGEEEIIIYQQ+U6SUkIIIYQQQgghhBAi30lSSgjxVPPnz0ej0XDo0CG1QykSrly5gkajeeI2btw4tUPE29ub1q1bqx2GEEIIkS+mT5+ORqOhcuXKaociXpHGjRs/se3l7e2tdniMGzcOjUZDfHy82qEIke+M1A5ACCGKoyFDhtC1a9cc5e7u7ipEI4QQQhRfc+fOBeD06dPs37+foKAglSMSr0Lp0qVZsmRJjnJTU1MVohFCPCBJKSGEyGMpKSmYmZmh0WieWMfT05PatWvnY1RCCCGEeNShQ4c4fvw4rVq1Yt26dYSGhhbYpFRycjIWFhZqh1EgKYpCamoq5ubmT6xjbm4ubS8hCiAZvieEyBO7du2iadOmWFtbY2FhQd26dVm3bp1eneTkZEaOHImPjw9mZmbY29sTGBjI0qVLdXUuXbpE586dKVWqFKampjg7O9O0aVOOHTv2zBjWrFlDnTp1sLCwwNramtdff529e/fqXl+9ejUajYYtW7bk2HfmzJloNBpOnDihKzt06BBvvvkm9vb2mJmZUb16dVasWKG334PhjZs3b6Zv3744OjpiYWFBWlpabt+6J2rcuDGVK1dm586d1K5dG3Nzc9zc3Bg7dixZWVl6dW/fvs2gQYNwc3PDxMSE0qVLM2bMmBxxZGdn8+OPP1KtWjXMzc0pUaIEtWvXZs2aNTnOv3HjRmrUqIG5uTm+vr66b5IfyM3PUwghhCjIQkNDAfjmm2+oW7cuy5YtIzk5OUe969ev8+677+Lh4YGJiQmlSpWiQ4cO3LhxQ1cnISGBESNGULp0aUxNTXFycqJly5acOXMGgO3bt6PRaNi+fbvesR8M7Z8/f76urHfv3lhZWXHy5EmaN2+OtbU1TZs2BSAsLIy2bdvi7u6OmZkZZcuW5b333nvs0K8zZ87QpUsXnJ2dMTU1xdPTk549e5KWlsaVK1cwMjJi4sSJOfbbsWMHGo2G33///anvX1RUFN27d8fJyQlTU1P8/PyYPHky2dnZAGRkZODk5ESPHj1y7JuQkIC5uTnDhw/XlSUlJenaFiYmJri5uTF06FDu37+vt69Go2Hw4MHMmjULPz8/TE1NWbBgwVNjzY0H7bqwsDD69OmDvb09lpaWtGnThkuXLuWoP3fuXKpWraprB7Vv356IiIgc9fbv30+bNm1wcHDAzMyMMmXKMHTo0Bz1bty4QZcuXbC1tcXZ2Zm+ffuSmJioV+f3338nKCgIW1tbLCwsKF26NH379n3paxdCLdJTSgjx0v755x9ef/11qlSpQmhoKKampsyYMYM2bdqwdOlSQkJCABg+fDiLFi1iwoQJVK9enfv373Pq1Clu3bqlO1bLli3Jyspi0qRJeHp6Eh8fz549e0hISHhqDL/99hvdunWjefPmLF26lLS0NCZNmkTjxo3ZsmUL9evXp3Xr1jg5OTFv3jxdw+6B+fPnU6NGDapUqQLAtm3beOONNwgKCmLWrFnY2tqybNkyQkJCSE5Opnfv3nr79+3bl1atWrFo0SLu37+PsbHxU+PNzs4mMzMzR7mRkf6f5djYWDp37syoUaMYP34869atY8KECdy5c4effvoJgNTUVJo0acLFixf58ssvqVKlCjt37mTixIkcO3ZMLznYu3dvFi9eTL9+/Rg/fjwmJiYcOXKEK1eu6J33+PHjjBgxglGjRuHs7MycOXPo168fZcuWpWHDhkDufp5CCCFEQZWSksLSpUupWbMmlStXpm/fvvTv35/ff/+dXr166epdv36dmjVrkpGRwaeffkqVKlW4desWmzZt4s6dOzg7O3P37l3q16/PlStX+OSTTwgKCuLevXvs2LGDmJgYfH19nzu+9PR03nzzTd577z1GjRqlazdcvHiROnXq0L9/f2xtbbly5QpTpkyhfv36nDx5UtcGOX78OPXr16dkyZKMHz+ecuXKERMTw5o1a0hPT8fb25s333yTWbNm8fHHH2NoaKg7908//USpUqVo3779E+O7efMmdevWJT09na+++gpvb2/Wrl3LyJEjuXjxIjNmzMDY2Jju3bsza9Ysfv75Z2xsbHT7L126lNTUVPr06QNov+xq1KgR165d073Pp0+f5vPPP+fkyZP8/fffer3QV69ezc6dO/n8889xcXHBycnpme/p49peBgYGGBjo99Xo168fr7/+Or/99htXr17ls88+o3Hjxpw4cYISJUoAMHHiRD799FO6dOnCxIkTuXXrFuPGjaNOnTocPHiQcuXKAbBp0ybatGmDn58fU6ZMwdPTkytXrrB58+Ycsbz99tuEhITQr18/Tp48yejRo4H/hpju3buXkJAQQkJCGDduHGZmZkRGRrJ169ZnXrsQBZYihBBPMW/ePAVQDh48+MQ6tWvXVpycnJS7d+/qyjIzM5XKlSsr7u7uSnZ2tqIoilK5cmWlXbt2TzxOfHy8AijTpk17rhizsrKUUqVKKf7+/kpWVpau/O7du4qTk5NSt25dXdnw4cMVc3NzJSEhQVcWHh6uAMqPP/6oK/P19VWqV6+uZGRk6J2rdevWiqurq+48D96fnj175irWy5cvK8ATt507d+rqNmrUSAGU//3vf3rHeOeddxQDAwMlMjJSURRFmTVrlgIoK1as0Kv37bffKoCyefNmRVEUZceOHQqgjBkz5qkxenl5KWZmZrrjK4qipKSkKPb29sp7772nK3vWz1MIIYQoyBYuXKgAyqxZsxRF0bYbrKyslAYNGujV69u3r2JsbKyEh4c/8Vjjx49XACUsLOyJdbZt26YAyrZt2/TKH7QN5s2bpyvr1auXAihz58596jVkZ2crGRkZSmRkZI42w2uvvaaUKFFCiYuLe2ZMf/75p67s+vXripGRkfLll18+9dyjRo1SAGX//v165QMHDlQ0Go1y9uxZRVEU5cSJEwqg/Prrr3r1atWqpQQEBOieT5w4UTEwMMjR5vzjjz8UQFm/fr2uDFBsbW2V27dvPzXGBx60qR639evXT1fvQbuuffv2evvv3r1bAZQJEyYoiqIod+7cUczNzZWWLVvq1YuKilJMTU2Vrl276srKlCmjlClTRklJSXlifF988YUCKJMmTdIrHzRokGJmZqZrS3///fcKoNeOFaKwk+F7QoiXcv/+ffbv30+HDh2wsrLSlRsaGtKjRw+uXbvG2bNnAahVqxYbNmxg1KhRbN++nZSUFL1j2dvbU6ZMGb777jumTJnC0aNHdd2/n+bs2bNER0fTo0cPvW+6rKysePvtt9m3b5+uK37fvn1JSUlh+fLlunrz5s3D1NRUN/H4hQsXOHPmDN26dQO036o92Fq2bElMTIzumh54++23n+dt48MPP+TgwYM5tmrVqunVs7a25s0339Qr69q1K9nZ2ezYsQOArVu3YmlpSYcOHfTqPejN9WC44oYNGwB4//33nxlftWrV8PT01D03MzOjfPnyREZG6sqe9fMUQgghCrLQ0FDMzc3p3LkzoG03dOzYkZ07d3L+/HldvQ0bNtCkSRP8/PyeeKwNGzZQvnx5mjVrlqcxPq59ERcXx4ABA/Dw8MDIyAhjY2O8vLwAdEPHkpOT+eeff+jUqROOjo5PPH7jxo2pWrUqP//8s65s1qxZaDQa3n333afGtnXrVipWrEitWrX0ynv37o2iKLreO/7+/gQEBDBv3jxdnYiICA4cOKA37Gzt2rVUrlyZatWq6bW9WrRo8dhhj6+99hp2dnZPjfFhZcqUeWzba+zYsTnqPmgDPlC3bl28vLzYtm0boO2xlJKSkqPnvIeHB6+99pqu7XXu3DkuXrxIv379MDMze2aMj7b5qlSpQmpqKnFxcQDUrFkTgE6dOrFixQquX7+eu4sXogCTpJQQ4qXcuXMHRVFwdXXN8VqpUqUAdMO5pk+fzieffMLq1atp0qQJ9vb2tGvXTtfwezDfU4sWLZg0aRI1atTA0dGRDz74gLt37z4xhgfHf1IM2dnZ3LlzB4BKlSpRs2ZNXcMoKyuLxYsX07ZtW+zt7QF080OMHDkSY2NjvW3QoEEAOeZteNy5n8bd3Z3AwMAc28OJPQBnZ+cc+7q4uOhd961bt3BxcckxsbqTkxNGRka6ejdv3sTQ0FC3/9M4ODjkKDM1NdVLPD3r5ymEEEIUVBcuXGDHjh20atUKRVFISEggISFB9wXPw/Mo3rx585mr4+amzvOysLDQG+4G2uH/zZs3Z9WqVXz88cds2bKFAwcOsG/fPgDd/9N37twhKysrVzF98MEHbNmyhbNnz5KRkcHs2bPp0KHDM9sLt27dylX7D7RfCu7du1c3v9aDLwS7dOmiq3Pjxg1OnDiRo+1lbW2Noigv3fYyMzN7bNvrQULvYY+7dhcXF72215NiKFWqlF7bC3K/uvKj7a8HKwM++Lk2bNiQ1atXk5mZSc+ePXF3d6dy5coyn6co1CQpJYR4KXZ2dhgYGBATE5PjtejoaABKliwJgKWlJV9++SVnzpwhNjaWmTNnsm/fPtq0aaPbx8vLi9DQUGJjYzl79izDhg1jxowZfPTRR0+M4cF/4E+KwcDAQO+btD59+rBv3z4iIiLYuHEjMTExuvkMHo539OjRj/1G7XE9mp620t7LeHgC1QdiY2OB/67bwcGBGzduoCiKXr24uDgyMzN11+Po6EhWVpZu/5eVm5+nEEIIURDNnTsXRVH4448/sLOz022tWrUCYMGCBbpFRRwdHbl27dpTj5ebOg96yjy6CMnjJiiHx7ctTp06xfHjx/nuu+8YMmQIjRs3pmbNmjmSGfb29hgaGj4zJtD2wHZwcODnn3/m999/JzY2Nle9qh0cHHLV/gPo0qULpqamzJ8/n6ysLBYtWkS7du302mclS5bE39//iW2vR3s0vaq2F/DYtlJsbKxe2wue3PZ8uO0F5OrnkFtt27Zly5YtJCYmsn37dtzd3enatave4j5CFCaSlBJCvBRLS0uCgoJYtWqVXi+a7OxsFi9ejLu7O+XLl8+xn7OzM71796ZLly6cPXv2sSvdlC9fns8++wx/f3+OHDnyxBgqVKiAm5sbv/32m15i5v79+6xcuVK3It8DXbp0wczMjPnz5zN//nzc3Nxo3ry53vHKlSvH8ePHH/uNWmBgINbW1s/9Xr2Iu3fv5lgZ77fffsPAwEA34XjTpk25d+8eq1ev1qu3cOFC3esAwcHBgHalwbyWm5+nEEIIURBkZWWxYMECypQpw7Zt23JsI0aMICYmRjfsPTg4mG3btuUYuv+w4OBgzp0799QJp729vQH0VvoFHrsC7pM8SMQ86EHzwC+//KL33NzcnEaNGvH7778/Men1gJmZGe+++y4LFixgypQpVKtWjXr16j0zlqZNmxIeHp6jjbZw4UI0Gg1NmjTRldnZ2dGuXTsWLlzI2rVriY2NzbFiXOvWrbl48SIODg6PbXs9eP/yw5IlS/Se79mzh8jISBo3bgxAnTp1MDc3Z/HixXr1rl27xtatW3Vtr/Lly1OmTBnmzp2bJyszP8zU1JRGjRrx7bffAnD06NE8Pb4Q+UVW3xNC5MrWrVtzrNAG2tXyJk6cyOuvv06TJk0YOXIkJiYmzJgxg1OnTrF06VJdAyooKIjWrVtTpUoV7OzsiIiIYNGiRbqk0YkTJxg8eDAdO3akXLlymJiYsHXrVk6cOMGoUaOeGJuBgQGTJk2iW7dutG7dmvfee4+0tDS+++47EhIS+Oabb/TqlyhRgvbt2zN//nwSEhIYOXJkjlVXfvnlF4KDg2nRogW9e/fGzc2N27dvExERwZEjR565RPKzREVF6braP8zR0ZEyZcronjs4ODBw4ECioqIoX74869evZ/bs2QwcOFA351PPnj35+eef6dWrF1euXMHf359du3bxf//3f7Rs2VI3v0WDBg3o0aMHEyZM4MaNG7Ru3RpTU1OOHj2KhYUFQ4YMea5reNbPUwghhCiINmzYQHR0NN9++60uyfCwypUr89NPPxEaGkrr1q0ZP348GzZsoGHDhnz66af4+/uTkJDAxo0bGT58OL6+vgwdOpTly5fTtm1bRo0aRa1atUhJSeGff/6hdevWNGnSBBcXF5o1a8bEiROxs7PDy8uLLVu2sGrVqlzH7uvrS5kyZRg1ahSKomBvb89ff/1FWFhYjroPVuQLCgpi1KhRlC1blhs3brBmzRp++eUXvS/YBg0axKRJkzh8+DBz5szJVSzDhg1j4cKFtGrVivHjx+Pl5cW6deuYMWMGAwcOzPGlZN++fVm+fDmDBw/G3d09x/xbQ4cOZeXKlTRs2JBhw4ZRpUoVsrOziYqKYvPmzYwYMYKgoKBcv1ePSklJeWzbC6B27dp6zw8dOkT//v3p2LEjV69eZcyYMbi5uemmcShRogRjx47l008/pWfPnnTp0oVbt27x5ZdfYmZmxhdffKE71s8//0ybNm2oXbs2w4YNw9PTk6ioKDZt2pQj+fUsn3/+OdeuXaNp06a4u7uTkJDADz/8gLGxMY0aNXrOd0SIAkK9OdaFEIXBg1VInrRdvnxZURRF2blzp/Laa68plpaWirm5uVK7dm3lr7/+0jvWqFGjlMDAQMXOzk4xNTVVSpcurQwbNkyJj49XFEVRbty4ofTu3Vvx9fVVLC0tFSsrK6VKlSrK1KlTlczMzGfGunr1aiUoKEgxMzNTLC0tlaZNmyq7d+9+bN3NmzfrruHcuXOPrXP8+HGlU6dOipOTk2JsbKy4uLgor732mm6Vnoffn6etTviwZ62+161bN13dRo0aKZUqVVK2b9+uBAYGKqampoqrq6vy6aef5lgV8NatW8qAAQMUV1dXxcjISPHy8lJGjx6tpKam6tXLyspSpk6dqlSuXFkxMTFRbG1tlTp16uj9rLy8vJRWrVrliL1Ro0ZKo0aNdM+f9fMUQgghCqJ27dopJiYmT12VrnPnzoqRkZESGxurKIqiXL16Venbt6/i4uKiGBsbK6VKlVI6deqk3LhxQ7fPnTt3lA8//FDx9PRUjI2NFScnJ6VVq1bKmTNndHViYmKUDh06KPb29oqtra3SvXt35dChQ49dfc/S0vKxsYWHhyuvv/66Ym1trdjZ2SkdO3ZUoqKiFED54osvctTt2LGj4uDgoJiYmCienp5K7969c7QPFEVRGjdurNjb2yvJycm5eRsVRVGUyMhIpWvXroqDg4NibGysVKhQQfnuu+/0VkN+ICsrS/Hw8HjqSsD37t1TPvvsM6VChQq6doq/v78ybNgw3c9CUbSr773//vu5jvNpq+8BunbVg3bd5s2blR49eiglSpTQrbJ3/vz5HMedM2eOUqVKFV2sbdu2VU6fPp2j3t69e5Xg4GDF1tZWMTU1VcqUKaMMGzZM9/qD1fdu3rypt9+DeB60t9euXasEBwcrbm5uiomJieLk5KS0bNlSb/VmIQobjaI8MgmJEEKIAqFx48bEx8dz6tQptUMRQgghRBEWFxeHl5cXQ4YMYdKkSWqHo5r58+fTp08fDh48SGBgoNrhCFEsyPA9IYQQQgghhCiGrl27xqVLl/juu+8wMDDgww8/VDskIUQxIxOdCyGEEEIIIUQxNGfOHBo3bszp06dZsmQJbm5uaockhChmZPieEEIIIYQQQgghhMh30lNKCCGEEEIIIYQQQuQ7SUoJIYQQQgghhBBCiHwnSSkhhBBCCCGEEEIIke9k9b0XlJ2dTXR0NNbW1mg0GrXDEUIIIUQ+URSFu3fvUqpUKQwM5Pu9p5H2khBCCFE85ba9JEmpFxQdHY2Hh4faYQghhBBCJVevXsXd3V3tMAo0aS8JIYQQxduz2kuSlHpB1tbWgPYNtrGxUTkaIYQQQuSXpKQkPDw8dG0B8WTSXhJCCCGKp9y2lyQp9YIedEG3sbGRRpYQQghRDMlwtGeT9pIQQghRvD2rvSQTIQghhBBCCCGEEEKIfCdJKSGEEEIIIYQQQgiR7yQpJYQQQgghhBBCCCHyncwpJYQQotDKysoiIyND7TBEEWNsbIyhoaHaYRQrci+Lokr+ngghxNNJUkoIIUShoygKsbGxJCQkqB2KKKJKlCiBi4uLTGb+ism9LIoD+XsihBBPJkkpIYQQhc6DD7FOTk5YWFhIQ1/kGUVRSE5OJi4uDgBXV1eVIyra5F4WRZn8PRFCiGeTpJQQQohCJSsrS/ch1sHBQe1wRBFkbm4OQFxcHE5OTjL05hWRe1kUB/L3RAghnk4mOhdCCFGoPJh3xsLCQuVIRFH24PdL5jl6deReFsWF/D0RQognk6SUEEKIQkmG+YhXSX6/8o+816Kok99xIYR4MklKCSGEEEIIIYQQQoh8J0kpIYQQohBr3LgxQ4cOzXX9K1euoNFoOHbs2CuLSQjx/OReFkIIURxJUkoIIYTIBxqN5qlb7969X+i4q1at4quvvsp1fQ8PD2JiYqhcufILnS+35AOzKKqK2738sObNm2NoaMi+ffvy7ZxCCCGKNll9TwghhMgHMTExusfLly/n888/5+zZs7qyBys0PZCRkYGxsfEzj2tvb/9ccRgaGuLi4vJc+wgh/lNc7+WoqCj27t3L4MGDCQ0NpXbt2vl27sfJ7fsqhBCiYJOeUkIIIUQ+cHFx0W22trZoNBrd89TUVEqUKMGKFSto3LgxZmZmLF68mFu3btGlSxfc3d2xsLDA39+fpUuX6h330SE/3t7e/N///R99+/bF2toaT09Pfv31V93rj/Zg2r59OxqNhi1bthAYGIiFhQV169bV+5ANMGHCBJycnLC2tqZ///6MGjWKatWqvfD7kZaWxgcffICTkxNmZmbUr1+fgwcP6l6/c+cO3bp1w9HREXNzc8qVK8e8efMASE9PZ/Dgwbi6umJmZoa3tzcTJ0584ViEeB7F9V6eN28erVu3ZuDAgSxfvpz79+/rvZ6QkMC7776Ls7MzZmZmVK5cmbVr1+pe3717N40aNcLCwgI7OztatGjBnTt3dNc6bdo0veNVq1aNcePG6Z5rNBpmzZpF27ZtsbS0ZMKECWRlZdGvXz98fHwwNzenQoUK/PDDDzlinzt3LpUqVcLU1BRXV1cGDx4MQN++fWndurVe3czMTFxcXJg7d+4z3xMhhBAvT5JSBVBGVjbLD0aRkp6ldihCCFEoKIpCcnqmKpuiKHl2HZ988gkffPABERERtGjRgtTUVAICAli7di2nTp3i3XffpUePHuzfv/+px5k8eTKBgYEcPXqUQYMGMXDgQM6cOfPUfcaMGcPkyZM5dOgQRkZG9O3bV/fakiVL+Prrr/n22285fPgwnp6ezJw586Wu9eOPP2blypUsWLCAI0eOULZsWVq0aMHt27cBGDt2LOHh4WzYsIGIiAhmzpxJyZIlAZg+fTpr1qxhxYoVnD17lsWLF+Pt7f1S8YiCQa17OS/vYyh697KiKMybN4/u3bvj6+tL+fLlWbFihe717OxsgoOD2bNnD4sXLyY8PJxvvvkGQ0NDAI4dO0bTpk2pVKkSe/fuZdeuXbRp04asrOdr637xxRe0bduWkydP0rdvX7Kzs3F3d2fFihWEh4fz+eef8+mnn+rFNnPmTN5//33effddTp48yZo1ayhbtiwA/fv3Z+PGjXq939avX8+9e/fo1KnTc8UmhFDZ3RuQFK12FOIFyPC9Aqj/gkP8c+4mN++mMfi1cmqHI4QQBV5KRhYVP9+kyrnDx7fAwiRv/jsdOnQob731ll7ZyJEjdY+HDBnCxo0b+f333wkKCnricVq2bMmgQYMA7YfjqVOnsn37dnx9fZ+4z9dff02jRo0AGDVqFK1atSI1NRUzMzN+/PFH+vXrR58+fQD4/PPP2bx5M/fu3Xuh67x//z4zZ85k/vz5BAcHAzB79mzCwsIIDQ3lo48+IioqiurVqxMYGAigl3SKioqiXLly1K9fH41Gg5eX1wvFIQoete7lvLyPoejdy3///TfJycm0aNECgO7duxMaGqo7zt9//82BAweIiIigfPnyAJQuXVq3/6RJkwgMDGTGjBm6skqVKj31nI/TtWtXvSQbwJdffql77OPjw549e1ixYoUuqTRhwgRGjBjBhx9+qKtXs2ZNAOrWrUuFChVYtGgRH3/8MaDtEdaxY0esrKyeOz4hhEpOrIC/PoSMFCjdCGr0BN/WYGSqdmQiF6SnVAH0Vg03AGZuv8jNu2kqRyOEECK/PEjAPJCVlcXXX39NlSpVcHBwwMrKis2bNxMVFfXU41SpUkX3+MHQori4uFzv4+rqCqDb5+zZs9SqVUuv/qPPn8fFixfJyMigXr16ujJjY2Nq1apFREQEAAMHDmTZsmVUq1aNjz/+mD179ujq9u7dm2PHjlGhQgU++OADNm/e/MKxCPEqFLV7OTQ0lJCQEIyMtIm7Ll26sH//ft3QwGPHjuHu7q5LSD3qQU+pl/Xo+wowa9YsAgMDcXR0xMrKitmzZ+ve17i4OKKjo5967v79++uGBsfFxbFu3bociS8hRAGVmQ7rP4JV70BGMqDApe3wR1+YXAE2fAKxp9SOUjyD9JQqgNpUKUXorsucuJbID1vOMaGdv9ohCSFEgWZubEj4+BaqnTuvWFpa6j2fPHkyU6dOZdq0afj7+2NpacnQoUNJT09/6nEenfxXo9GQnZ2d6300Gg2A3j4Pyh54meFOD/Z93DEflAUHBxMZGcm6dev4+++/adq0Ke+//z7ff/89NWrU4PLly2zYsIG///6bTp060axZM/74448XjkkUDGrdy3l5H0PRupdv377N6tWrycjI0Bvql5WVxdy5c/n2229zTO7+qGe9bmBgkCOOjIyMHPUefV9XrFjBsGHDmDx5MnXq1MHa2prvvvtONyzyWecF6NmzJ6NGjWLv3r3s3bsXb29vGjRo8Mz9hBAqS7wOv/eCa//OR9noE6jaGY4vg6OLIek67J+l3UrVgBo9oHIHMLNRN26Rg/SUKoAMDDR82tIPgKUHrnIh7sWGRwghRHGh0WiwMDFSZXv0A15e2rlzJ23btqV79+5UrVqV0qVLc/78+Vd2viepUKECBw4c0Cs7dOjQCx+vbNmymJiYsGvXLl1ZRkYGhw4dws/PT1fm6OhI7969Wbx4MdOmTdOb5NnGxoaQkBBmz57N8uXLWblypW4+KlF4qXUvv8r7GAr3vbxkyRLc3d05fvw4x44d023Tpk1jwYIFZGZmUqVKFa5du8a5c+cee4wqVaqwZcuWJ57D0dFRb16npKQkLl++/Mzr2blzJ3Xr1mXQoEFUr16dsmXLcvHiRd3r1tbWeHt7P/XcDg4OtGvXjnnz5jFv3jzdkEQhRAF2aTv80kCbkDIrAV1/hyafgn1p7b9DT0K3lVCxLRgYQ/QRWDtM23vqz4EQuQfyeC5B8eKkp1QBVbu0A838nPg7Io5vN55hds+c3ZWFEEIUbWXLlmXlypXs2bMHOzs7pkyZQmxsrF7iJj8MGTKEd955h8DAQOrWrcvy5cs5ceKE3pwxT/Loyl8AFStWZODAgXz00UfY29vj6enJpEmTSE5Opl+/foB2rpuAgAAqVapEWloaa9eu1V331KlTcXV1pVq1ahgYGPD777/j4uJCiRIl8vS6hcgrhfleDg0NpUOHDlSuXFmv3MvLi08++YR169bRtm1bGjZsyNtvv82UKVMoW7YsZ86cQaPR8MYbbzB69Gj8/f0ZNGgQAwYMwMTEhG3bttGxY0dKlizJa6+9xvz582nTpg12dnaMHTtWN0n605QtW5aFCxeyadMmfHx8WLRoEQcPHsTHx0dXZ9y4cQwYMAAnJyeCg4O5e/cuu3fvZsiQIbo6/fv3p3Xr1mRlZdGrV68XeGeFEPkiOxt2T4WtE0DJBpcqELII7Lz16xkYQrlm2u1+vLb31JGFEH8Wjv+m3RzKQvUeUK0rWDmpcjlCS5JSBdioYF+2nb1JWPgNDly+TS0fe7VDEkIIkY/Gjh3L5cuXadGiBRYWFrz77ru0a9eOxMTEfI2jW7duXLp0iZEjR5KamkqnTp3o3bt3jh4Xj9O5c+ccZZcvX+abb74hOzubHj16cPfuXQIDA9m0aRN2dnYAmJiYMHr0aK5cuYK5uTkNGjRg2bJlAFhZWfHtt99y/vx5DA0NqVmzJuvXr8fAQDqAi4KpsN7Lhw8f5vjx48yePTvHa9bW1jRv3pzQ0FDatm3LypUrGTlyJF26dOH+/fuULVuWb775BoDy5cuzefNmPv30U2rVqoW5uTlBQUF06dIFgNGjR3Pp0iVat26Nra0tX331Va56Sg0YMIBjx44REhKCRqOhS5cuDBo0iA0bNujq9OrVi9TUVKZOncrIkSMpWbIkHTp00DtOs2bNcHV1pVKlSpQqVSrX76cQIh+lJMDqgXB2vfZ59R7Q8nswNnv6fpYloe5gqPO+tmfVkYVwahXcugB/fwFbv4Lyb2iPV7YZGEqKJL9plLxeA7eYSEpKwtbWlsTERGxsXt241E//PMlv+6Oo6lGC1YPqvvLu5UIIUdClpqZy+fJlfHx8MDN7RkNEvDKvv/46Li4uLFq0SO1QXomn/Z7lVxugKHjaeyX3csFQ1O/l3EhOTqZUqVLMnTs3x6qJeUF+14V4SbGnYHl3uHMZDE2h1ffaFfZeVNpdOP0nHFkE1x5Kylu7antOVe+uHQooXkpu20uSBizghjYrx+qj1zl+NYG1J2JoU1W+vRFCCJG/kpOTmTVrFi1atMDQ0JClS5fy999/ExYWpnZoQojnIPeyvuzsbGJjY5k8eTK2tra8+eabaockhHjUsaXa+aAyU6CEJ3RaCKWqv9wxTa21Sa0aPSHuDBxdBMeXwt0Y2DlZu3k30L7u1waMn71ognhxkpQq4JyszXivYRmm/n2OSZvO0LySM6ZGebtCjBBCCPE0Go2G9evXM2HCBNLS0qhQoQIrV66kWbNmaocmhHgOci/ri4qKwsfHB3d3d+bPn4+RkXw0EqLAyEyDjaPg0Fzt87Kvw1u/gkUeT2nj5AstvoamX2iHBh5dBBe2wJWd2s3MFvw7aVfvc62at+cWgCSlCoV3GvqwZH8kV2+nsGhvJP0bSFdCIYQQ+cfc3Jy///5b7TCEEC9J7mV93t7eyEwmQhRACVdhRU/tqnlooPFoaPgRvMq5I41MoFI77ZZwFY79BkcXQ2IUHJyt3Vyrauee8u8I5iVeXSzFjMwIWghYmBgxonl5AH7ceoHE5AyVIxJCCCGEEEIIIfLYhS3wS0NtQsrcDrr9AY0/ebUJqUeV8NCe88Pj0ONPqPQWGJpAzHFYPxImV4BV78LlnSCJ7ZcmSalCokOABxWcrUlMyeDn7RfUDkcIIYQQQgghhMgb2dnwz3ew+G1Iua2dN+q9HVBOxeHFBgZQ5jXoOA9GnIU3vgGnipCZCieWw4LW8GMN7RxUSTHqxVnISVKqkDA00DCqpS8A83df4ertZJUjEkIIIYQQQgghXlLKHVjaGbZNABQI6A19NmonNi8oLOyh9kAYuAf6b9XGaGINty/BlvEwtRL81hnOrIMsGdn0PCQpVYg0Lu9IvbIOpGdl8/3ms2qHI4QQQgghhBBCvLiY4/BLIzi/CYzMoO0MaPMDGJupHdnjaTTgHqCNceRZbbyedUDJgnMbYFlXbYIq7AuIlxFOuSFJqUJEo9EwOtgPjQb+dyyaE9cS1A5JCCGEEEIIIYR4fkcXQ2hzSIgEO2/oFwbVu6kdVe6ZWGrj7bsR3j8IdT8AS0e4dwN2T4OfAmBeSzi2FNJlpNOTSFKqkKnsZkv7am4AfL0uQlYMEUIIIYQQQghReGSkwpoP4H/va+dnKv8GvLsdXKuoHdmLcywPzb+C4REQshjKtQCNAUTuhtUDtJOjrx0G14/I5OiPkKRUITSiRQVMjAzYf/k2WyLi1A5HCCFEPmrcuDFDhw7VPff29mbatGlP3Uej0bB69eqXPndeHUcIIfeyEKKYuhMJc1vAkQWABl77DDov1a60VxQYGoNfG+i2Aoad1l6fnTekJcGhuTC7CcyqD/t/geTbakdbIEhSqhByK2FO33o+AEzcEEFmVrbKEQkhhHiWNm3a0KzZ41eQ2bt3LxqNhiNHjjz3cQ8ePMi77777suHpGTduHNWqVctRHhMTQ3BwcJ6e61Hz58+nRIkSr/QcQrwMuZefT0pKCnZ2dtjb25OSkpIv5xRCFFDn/4ZfG0HMMTC3hx6roOFH2lXuiiKbUtrrG3IUeq4B/45gaAo3TsGGj2GyL/zRDy5t164+WEwV0Z9+0TeoSRnsLIy5ePM+yw9dVTscIYQQz9CvXz+2bt1KZGRkjtfmzp1LtWrVqFGjxnMf19HREQsLi7wI8ZlcXFwwNTXNl3MJUVDJvfx8Vq5cSeXKlalYsSKrVq3Kl3M+iaIoZGZmqhqDEMVSdjZs/waWdNCutOcWAO/tgDKvqR1Z/jAwgNKN4O052snRg78DZ3/ISoNTf8DCtjC9GvzzHSReVzvafCdJqULKxsyYD5qWA2Bq2Hnupcl/sEIIUZC1bt0aJycn5s+fr1eenJzM8uXL6devH7du3aJLly64u7tjYWGBv78/S5cufepxHx3yc/78eRo2bIiZmRkVK1YkLCwsxz6ffPIJ5cuXx8LCgtKlSzN27FgyMrTLF8+fP58vv/yS48ePo9Fo0Gg0upgfHfJz8uRJXnvtNczNzXFwcODdd9/l3r17utd79+5Nu3bt+P7773F1dcXBwYH3339fd64XERUVRdu2bbGyssLGxoZOnTpx48YN3evHjx+nSZMmWFtbY2NjQ0BAAIcOHQIgMjKSNm3aYGdnh6WlJZUqVWL9+vUvHEtBM2PGDHx8fDAzMyMgIICdO3c+tX5aWhpjxozBy8sLU1NTypQpw9y5c3Wvnz59mrfffhtvb280Gs0zh5YVF3IvP9+9HBoaSvfu3enevTuhoaE5Xj99+jStWrXCxsYGa2trGjRowMWLF3Wvz507l0qVKmFqaoqrqyuDBw8G4MqVK2g0Go4dO6arm5CQgEajYfv27QBs374djUbDpk2bCAwMxNTUlJ07d3Lx4kXatm2Ls7MzVlZW1KxZk7///lsvrrS0ND7++GM8PDwwNTWlXLlyhIaGoigKZcuW5fvvv9erf+rUKQwMDPRiF0KgHaL2W0fYPhFQILAf9NkAJTzUjkwd5nYQ9C4M2KmdRyuwH5jaaid73zYBplWGxR0gfA1kpqsdbb4wUjsA8eK6BXmxYM8VrtxK5tcdlxj+enm1QxJCCHUoCmSotKqJsYV2eeBnMDIyomfPnsyfP5/PP/8czb/7/P7776Snp9OtWzeSk5MJCAjgk08+wcbGhnXr1tGjRw9Kly5NUFDQM8+RnZ3NW2+9RcmSJdm3bx9JSUl6c9Y8YG1tzfz58ylVqhQnT57knXfewdramo8//piQkBBOnTrFxo0bdR/SbG1tcxwjOTmZN954g9q1a3Pw4EHi4uLo378/gwcP1vuwvm3bNlxdXdm2bRsXLlwgJCSEatWq8c477zzzeh6lKArt2rXD0tKSf/75h8zMTAYNGkRISIjuQ2i3bt2oXr06M2fOxNDQkGPHjmFsbAzA+++/T3p6Ojt27MDS0pLw8HCsrKyeO46CaPny5QwdOpQZM2ZQr149fvnlF4KDgwkPD8fT0/Ox+zxI6IWGhlK2bFni4uL0epEkJydTunRpOnbsyLBhw/LnQtS6l3N5H4Pcy89zL1+8eJG9e/eyatUqFEVh6NChXLp0idKlSwNw/fp1GjZsSOPGjdm6dSs2Njbs3r1b93s4c+ZMhg8fzjfffENwcDCJiYns3r37me/foz7++GO+//57SpcuTYkSJbh27RotW7ZkwoQJmJmZsWDBAtq0acPZs2d190vPnj3Zu3cv06dPp2rVqly+fJn4+Hg0Gg19+/Zl3rx5jBw5UneOuXPn0qBBA8qUKfPc8QlRZEUfheU9ITEKjMyhzTSo2lntqAoGjQZKVdduzSdAxBo4sggid8GFMO1mURKqdYHqPbUTqRdVisp+/vlnxdvbWzE1NVVq1Kih7Nix46n1U1NTlU8//VTx9PRUTExMlNKlSyuhoaF6daZOnaqUL19eMTMzU9zd3ZWhQ4cqKSkpute/+OILBdDbnJ2dnyvuxMREBVASExOfa7+8tu5EtOL1yVrF97MNSmxiyrN3EEKIQi4lJUUJDw/X+7uupN1TlC9s1NnS7uU69oiICAVQtm7dqitr2LCh0qVLlyfu07JlS2XEiBG6540aNVI+/PBD3XMvLy9l6tSpiqIoyqZNmxRDQ0Pl6tWrutc3bNigAMqff/75xHNMmjRJCQgI0D3/4osvlKpVq+ao9/Bxfv31V8XOzk65d++/61+3bp1iYGCgxMbGKoqiKL169VK8vLyUzMxMXZ2OHTsqISEhT4xl3rx5iq2t7WNf27x5s2JoaKhERUXpyk6fPq0AyoEDBxRFURRra2tl/vz5j93f399fGTdu3BPP/bDH/p79q6C0AR5Wq1YtZcCAAXplvr6+yqhRox5bf8OGDYqtra1y69atXB3/4d+z5/G096pA3cvPcR8ritzLivLse1lRFOXTTz9V2rVrp3vetm1bZcyYMbrno0ePVnx8fJT09PTH7l+qVCm9+g+7fPmyAihHjx7Vld25c0cBlG3btimKoijbtm1TAGX16tVPjVNRFKVixYrKjz/+qCiKopw9e1YBlLCwsMfWjY6OVgwNDZX9+/criqIo6enpiqOj4xP/9jzt74kQRdah+YoyvqT2b+wP1RQl5qTaERUO8RcUJewLRfmunP7/U3NeV5QjixQl9a7aEeZabttLqg7fe/Ct3pgxYzh69CgNGjQgODiYqKioJ+7TqVMntmzZQmhoKGfPnmXp0qX4+vrqXl+yZAmjRo3iiy++ICIigtDQUJYvX87o0aP1jlOpUiViYmJ028mTJ1/Zdb5KwZVdqOFZgpSMLKaGnVM7HCGEEE/h6+tL3bp1dUOkLl68yM6dO+nbty8AWVlZfP3111SpUgUHBwesrKzYvHnzU/9ffFhERASenp64u7vryurUqZOj3h9//EH9+vVxcXHBysqKsWPH5vocD5+ratWqWFpa6srq1atHdnY2Z8+e1ZVVqlQJQ0ND3XNXV1fi4l5s5diIiAg8PDzw8Pivy3/FihUpUaIEERERAAwfPpz+/fvTrFkzvvnmG72hNB988AETJkygXr16fPHFF5w4ceKF4iho0tPTOXz4MM2bN9crb968OXv27HnsPmvWrCEwMJBJkybh5uZG+fLlGTlypExEnUtyLz/7Xs7KymLBggV0795dV9a9e3cWLFhAVlYWAMeOHaNBgwa63owPi4uLIzo6mqZNmz7X9TxOYGCg3vP79+/z8ccf6/5+WFlZcebMGd17d+zYMQwNDWnUqNFjj+fq6kqrVq10P/+1a9eSmppKx44dXzpWIQq9jBT43/vw1weQlQ4VWsE728ClstqRFQ4OZaDZOBgWDl2WQYWWoDGEq/u17+vkCrBmCFw7pO1dXASoOnxvypQp9OvXj/79+wMwbdo0Nm3axMyZM5k4cWKO+hs3buSff/7h0qVL2NvbA9rx9w/bu3cv9erVo2vXrrrXu3TpwoEDB/TqGRkZ4eLi8gquKn9pNBrGtPLj7Zl7WXHoKn3r+1De2VrtsIQQIn8ZW8Cn0eqd+zn069ePwYMH8/PPPzNv3jy8vLx0H7omT57M1KlTmTZtGv7+/lhaWjJ06FDS03M3p4DymMaJ5pEhSfv27aNz5858+eWXtGjRAltbW5YtW8bkyZOf6zoURclx7Med89EPmxqNhuwXXGHmSed8uHzcuHF07dqVdevWsWHDBr744guWLVtG+/bt6d+/Py1atGDdunVs3ryZiRMnMnnyZIYMGfJC8RQU8fHxZGVl4ezsrFfu7OxMbGzsY/e5dOkSu3btwszMjD///JP4+HgGDRrE7du39eaVel5paWmkpaXpniclJT3fAdS6l5/zPga5l591L2/atInr168TEhKiV56VlcXmzZsJDg7G3Nz8ifs/7TUAg39X63r4vXrSHFcPJ9wAPvroIzZt2sT3339P2bJlMTc3p0OHDrqfz7PODdC/f3969OjB1KlTmTdvHiEhIfk2Ub0QBdbty7CiJ8SeAI0BvDYW6g0tuqvrvUqGRlAhWLvdjYVjv8HRRXD7EhxZqN0c/aBGD6jSGSwd1I74han22/GqvtWrX78+hw8f1iWhLl26xPr162nVqpXesc6fP0+pUqXw8fGhc+fOXLp06anxpqWlkZSUpLcVFAFe9rxRyYVsBSauj1A7HCGEyH8aDZhYqrPlch6aBzp16oShoSG//fYbCxYsoE+fProPfjt37qRt27Z0796dqlWrUrp0ac6fP5/rY1esWJGoqCiio//7UL937169Ort378bLy4sxY8YQGBhIuXLlcqwiZmJiouvJ8LRzHTt2jPv37+sd28DAgPLlX828Bw+u7+rV/1adDQ8PJzExET8/P11Z+fLlGTZsGJs3b+att95i3rx5utc8PDwYMGAAq1atYsSIEcyePfuVxKqGRxMLT0s2ZGdno9FoWLJkCbVq1aJly5ZMmTKF+fPnv1RvqYkTJ2Jra6vbHu7Vlitq3cvPeR+D3MvPEhoaSufOnTl27Jje1q1bN92E51WqVGHnzp2PTSZZW1vj7e3Nli1bHnt8R0dHAGJiYnRlD096/jQ7d+6kd+/etG/fHn9/f1xcXLhy5YrudX9/f7Kzs/nnn3+eeIyWLVtiaWnJzJkz2bBhg66XnBDF1tmN8GsjbULKoiT0+BMaDJeEVF6wdtG+l0OOQO/12iSUkTncjIBNn2p7T63oBRe2aFc6LGRU+w15mW/1Tp06xZ9//sm0adP4448/eP/993V1OnfuzFdffUX9+vUxNjamTJkyNGnShFGjRunqBAUFsXDhQjZt2sTs2bOJjY2lbt263Lp164nxvnQj6xX7JNgXIwMN287eZM+FeLXDEUII8QRWVlaEhITw6aefEh0dTe/evXWvlS1blrCwMPbs2UNERATvvffeE/9PfJxmzZpRoUIFevbsyfHjx9m5cydjxozRq1O2bFmioqJYtmwZFy9eZPr06fz55596dby9vbl8+TLHjh0jPj5er+fLA926dcPMzIxevXpx6tQptm3bxpAhQ+jRo0eO/9ufV1ZWVo4PsuHh4TRr1owqVarQrVs3jhw5woEDB+jZsyeNGjUiMDCQlJQUBg8ezPbt24mMjGT37t0cPHhQl7AaOnQomzZt4vLlyxw5coStW7fqJbMKq5IlS2JoaJjjdyUuLu6JPwtXV1fc3Nz0Jr728/NDURSuXbv2wrGMHj2axMRE3fZwArGokXv5yW7evMlff/1Fr169qFy5st7Wq1cv1qxZw82bNxk8eDBJSUl07tyZQ4cOcf78eRYtWqQbNjhu3DgmT57M9OnTOX/+PEeOHOHHH38EtL2ZateuzTfffEN4eDg7duzgs88+y1V8ZcuWZdWqVRw7dozjx4/TtWtXvV5f3t7e9OrVi759+7J69WouX77M9u3bWbFiha6OoaEhvXv3ZvTo0ZQtW/axwyuFKBays2DrBFgaAqmJ4F4T3tsBpRurHVnRo9GAdz146xcYcQZaTQbXapCdAeGrYfFb8EMV2DYREp5vKLeaVE9b5vW3etu3b+frr79mxowZHDlyhFWrVrF27Vq++uor3XGCg4N5++238ff3p1mzZqxbtw6ABQsWPDHOgt7I8ilpSffaXgB8vT6C7OyiMb5UCCGKon79+nHnzh2aNWumtzLa2LFjqVGjBi1atKBx48a4uLjQrl27XB/XwMCAP//8k7S0NGrVqkX//v35+uuv9eq0bduWYcOGMXjwYKpVq8aePXsYO3asXp23336bN954gyZNmuDo6PjYpewtLCzYtGkTt2/fpmbNmnTo0IGmTZvy008/Pd+b8Rj37t2jevXqelvLli11y9jb2dnRsGFDmjVrRunSpVm+fDmg/ZB469YtevbsSfny5enUqRPBwcF8+eWXgDbZ9f777+Pn58cbb7xBhQoVmDFjxkvHqzYTExMCAgIICwvTKw8LC6Nu3bqP3adevXpER0dz7949Xdm5c+cwMDDQm8foeZmammJjY6O3FWVyLz/ewoULsbS0fOx8UE2aNMHa2ppFixbh4ODA1q1buXfvHo0aNSIgIIDZs2frhgr26tWLadOmMWPGDCpVqkTr1q31epzNnTuXjIwMAgMD+fDDD5kwYUKu4ps6dSp2dnbUrVuXNm3a0KJFC2rUqKFXZ+bMmXTo0IFBgwbh6+vLO++8o9ebDLQ///T0dOklJYqv+7dgSQfY8Z32ea33tD15bN3Ujas4MC8BNfvDe//Aezuh1rtgZguJV+Gfb2BaFVjUHk7/CZk5v5AoSDTK4wat54P09HQsLCz4/fffad++va78ww8/5NixY4/tLturVy92797NhQsXdGURERFUrFiRc+fOUa5cORo0aEDt2rX57rvvdHUWL17Mu+++y71793Tjzx/1+uuvU7ZsWWbOnJmr+JOSkrC1tSUxMbHANLhu30+n0aRt3E3LZGpIVdpXf/FGpRBCFFSpqalcvnwZHx8fzMzM1A5HFFFP+z0riG2A5cuX06NHD2bNmkWdOnX49ddfmT17NqdPn8bLy4vRo0dz/fp1Fi5cCGgTf35+ftSuXZsvv/yS+Ph4+vfvT6NGjXRDGtPT0wkPDwe0Q5W6detGt27dsLKyomzZsrmK62nvldzLorDbvXs3jRs35tq1a0/tVSa/66JIunZYO39U0jXtvHxtpkMVmexfVRmpcGYtHFkAl3f8V25uD1U7Q/Ue4Fwx38LJbXtJtZ5Sr+pbveTk5ByJJ0NDQxRFeeykkaCdLyoiIgJXV9eXuSTV2VuaMLBJGQC+33SO1IynzyEghBBCiKIhJCSEadOmMX78eKpVq8aOHTtYv349Xl7aXtQxMTF6q7JZWVkRFhZGQkICgYGBdOvWjTZt2jB9+nRdnejoaF1PtZiYGL7//nuqV6+uW6BGiOIqLS2NCxcuMHbsWDp16vTSQ5aFKFQUBQ6Gwrw3tAkph7LQf4skpAoCYzPw7wC9/oIPjkHDj8C6FKTchn0zYGYdmN0UDs+HtLtqR6ujWk8peDXf6o0bN44pU6bw66+/EhQUxIULFxg4cCABAQG67v0jR46kTZs2eHp6EhcXx4QJE/jnn384efKkrvH2LAXxW1KA1IwsXvt+O9GJqXzyhi8DG5dROyQhhMhT8o2zyA+FradUQSU9pURRNH/+fPr160e1atVYs2YNbm5PH6okv+uiyEhPhnXD4fi/Q4H92kDbGWAm/xcWWNlZ2gnQjyyAcxshO1NbbmwJldprV+/zCHqhBT+eJbftJaM8P/NzCAkJ4datW4wfP56YmBgqV66cq2/1hgwZQmBgIA4ODnTq1Elv/Phnn32GRqPhs88+4/r16zg6OtKmTRu9cfjXrl2jS5cuxMfH4+joSO3atdm3b1+uE1IFmZmxISOaV2DE78eZse0CITU9sLc0UTssIYQQQgghioTevXvrTWwvRLFw66J2uN6NU6AxhGbjoO6QV5LMEHnIwBDKN9du9+K0CcUji+DWeTi2WLu5BWh7u6n0s1S1p1RhVpC/Jc3OVmj94y7CY5LoXdebcW9WUjskIYTIM/KNs8gP0lMqb0hPKSHkd10UAWfWw58DIC0RLB2h43zwrq92VOJFKQpc3Q9HFmonQvfvCG9Of/Z+z6nAzyklXh0DAw2fttQucb14XyRX4u8/Yw8hhBBCCCGEEOIhWZnw95ewrIs2IeVRW7vSmySkCjeNBjxrQ7sZMOIsNBmjajiSlCqi6pcrSaPyjmRmK0zadEbtcIQQIs9lZ2erHYIowuT3K//Iey2KOvkdF4XSvZuwuD3smqJ9XnsQ9F4LNoV7cTDxCDMbsFZ3sQZV55QSr9bolr7sPH+T9SdjORx5hwAvO7VDEkKIl2ZiYoKBgQHR0dE4OjpiYmKCRuYzEHlEURTS09O5efMmBgYGmJjIvIyvitzLoqiTvyei0Lp6UDt/1N1o7YTYbX+Eym+rHZUooiQpVYT5utjQIcCdFYeu8X/rI/hjQB1p7AkhCj0DAwN8fHyIiYkhOjpa7XBEEWVhYYGnpycGBtKp/FWRe1kUF/L3RBQaigIH58DG0ZCdASXLQ6dF4OSrdmSiCJOkVBE3/PUKrDkezeHIO2w6HcsblaW7pRCi8DMxMcHT05PMzEyysrLUDkcUMYaGhhgZGckXOflA7mVR1MnfE1FopN+Hv4bCyRXa5xXbQtufwdRa1bBE0SdJqSLOxdaMdxqU5setF/h241ma+jljbCjf0gghCj+NRoOxsTHGxsZqhyKEeAlyLwshhMriL8CKHhAXDhpDaP6Vdg4pSaaKfCDZiWLgvUZlKGllwuX4+/y2P0rtcIQQQgghhBBCFAQRf8GvjbUJKStn7WTmdd6XhJTIN5KUKgasTI34sFl5AH7Ycp6k1AyVIxJCCCGEEEIIoZqsTAj7HJZ3h/S74FkX3tsBXnXVjkwUM5KUKiY61/SgjKMlt++nM2v7RbXDEUIIIYQQQgihhntxsKgd7P5B+7zOYOi1BqxdVA1LFE+SlComjA0NGBXsB0DorstEJ6SoHJEQQgghhBBCiHwVtQ9mNYArO8HECjougBZfg6HM6yfUIUmpYqSZnxO1fOxJy8xm8uZzaocjhBBCCCGEECI/KArsmwnzW8G9WChZAd7ZBpXaqR2ZKOYkKVWMaDQaxrTU9pZadfQap6MTVY5ICCGEEEIIIcQrlXYPVvaDjaMgOxMqvQXvbAXH8mpHJoQkpYqbqh4laFO1FIoCE9efQVEUtUMSQgghhBBCCPEq3DwHc5rCqZVgYARvfAsd5oKpldqRCQFIUqpY+rhFBUwMDdh1IZ5/zt1UOxwhhBBCCCGEEHnt9GqY3QRungFrV+i9DmoPAI1G7ciE0JGkVDHkYW9BzzpegLa3VFa29JYSQgghhBBCiCIhKwM2jYHfe0H6PfBuAO/tAM/aakcmRA6SlCqmBr9WFhszI87euMvKw9fUDkcIIYQQQgghxMu6GwsL3oS9P2mf1/sQeqwGKydVwxLiSSQpVUyVsDBhyGvlAJgcdpbk9EyVIxJCCCGEEEII8cIi98AvDSFqD5hYQ8hieH08GBqpHZkQTyRJqWKsZ10v3O3MuZGURujOy2qHI4QQQgghhBDieSkK7PkJ5reGezfAqSK8ux382qgdmRDPJEmpYszUyJCPWlQAYNY/F7l5N03liIQQQgghhBBC5FraXe3cUZvHgJIF/h2h/99QsqzakQmRK5KUKubaVClFFXdb7qdn8cOWc2qHI4QQQgghhBAiN+LOwOzXIPx/YGAMLb+Ht2aDiaXakQmRa5KUKuYMDDR82tIPgKUHrnIh7p7KEQkhhBBCCCGEeKpTK7UJqfhzYOMGfTZArXdAo1E7MiGeiySlBLVLO9DMz4msbIVvN55ROxwhhBBCCCGEEI+TmQ4bRsEffSHjPvg0hPd2gEdNtSMT4oVIUkoAMCrYF0MDDWHhNzhw+bba4QghhBBCCCGEeFhSNCxoDftnap/XHw49VoNlSVXDEuJlSFJKAFDWyZqQmh4AfL0+AkVRVI5ICCGEEEIIIQQAl3fCLw3h6n4wtYXOv0GzL8DAUO3IhHgpkpQSOkOblcPSxJDjVxNYeyJG7XCEEEIIIYQQonhTFNj9AyxsC/dvgnNleHcb+LZSOzIh8oQkpYSOk7UZ7zUqA8C3G8+QlpmlckRCCCGEEEIIUUylJsLy7hD2OShZULUL9AsDhzJqRyZEnpGklNDTv4EPTtamXLuTwqK9kWqHI4QQQgghhBDFz41w+LUJnFkLhibQagq0mwkmFmpHJkSekqSU0GNhYsSI5uUB+HHrBRKTM1SOSAghhBBCCCGKkRMrYE5TuH0RbNyhz0ao2Q80GrUjEyLPSVJK5NAhwIMKztYkpmTw07bzaocjhBBCCCGEEMXD7umw6h3ISIbSTeC9HeAeoHZUQrwykpQSORgaaBjV0heABXsiuXo7WeWIhBBCCCGEEKKIS0+GHd9rH9cfBt1XgqWDujEJ8YpJUko8VuPyjtQr60B6VjbfbTqrdjhCCCGEEEIIUbSdXgVpiWDnDa99DgaGakckxCsnSSnxWBqNhtHBfmg0sOZ4NCeuJagdkhBCCCGEEEIUXYfmaf+t0QsM5KO6KB7kN108UWU3W9pXcwPg63URKIqickRCCCGEEEIIUQTFnoTrh8DACKp3VzsaIfKNJKXEU41oUQETIwP2X77Nlog4tcMRQgghxBPMmDEDHx8fzMzMCAgIYOfOnU+tn5aWxpgxY/Dy8sLU1JQyZcowd+5cvTorV66kYsWKmJqaUrFiRf78889XeQlCCFF8Pegl5dsarJzUjUWIfCRJKfFUbiXM6VvPB4CJGyLIzMpWOSIhhBBCPGr58uUMHTqUMWPGcPToURo0aEBwcDBRUVFP3KdTp05s2bKF0NBQzp49y9KlS/H19dW9vnfvXkJCQujRowfHjx+nR48edOrUif379+fHJQkhRPGRdg9OrNA+DuyjbixC5DONImOyXkhSUhK2trYkJiZiY2OjdjivVFJqBo0mbeNOcgZft69MtyAvtUMSQgghVFMQ2wBBQUHUqFGDmTNn6sr8/Pxo164dEydOzFF/48aNdO7cmUuXLmFvb//YY4aEhJCUlMSGDRt0ZW+88QZ2dnYsXbo0V3EVxPdKCCEKnMML4K8PwL40DD4s80mJIiG3bQD5bRfPZGNmzAdNywEwNew899IyVY5ICCGEEA+kp6dz+PBhmjdvrlfevHlz9uzZ89h91qxZQ2BgIJMmTcLNzY3y5cszcuRIUlJSdHX27t2b45gtWrR44jGFEEK8oMP/Dt0L6C0JKVHsGKkdgCgcugV5sWDPFa7cSubXHZcY/np5tUMSQgghBBAfH09WVhbOzs565c7OzsTGxj52n0uXLrFr1y7MzMz4888/iY+PZ9CgQdy+fVs3r1RsbOxzHRO081SlpaXpniclJb3oZQkhRPEQfQyij4KhCVTrpnY0QuQ7ScOKXDExMuDjN7TzTMzecYkbSakqRySEEEKIh2k0Gr3niqLkKHsgOzsbjUbDkiVLqFWrFi1btmTKlCnMnz9fr7fU8xwTYOLEidja2uo2Dw+Pl7giIYQoBh70kvJrA5Yl1Y1FCBVIUkrkWnBlF2p4liAlI4spm8+pHY4QQgghgJIlS2JoaJijB1NcXFyOnk4PuLq64ubmhq2tra7Mz88PRVG4du0aAC4uLs91TIDRo0eTmJio265evfqilyWEEEVf2l04+Yf2cYBMcC6KJ0lKiVzTaDSMaeUHwO+Hr3I29q7KEQkhhBDCxMSEgIAAwsLC9MrDwsKoW7fuY/epV68e0dHR3Lt3T1d27tw5DAwMcHd3B6BOnTo5jrl58+YnHhPA1NQUGxsbvU0IIcQTnPwd0u+BQznwrq92NEKoQpJS4rkEeNkTXNmFbAUmbohQOxwhhBBCAMOHD2fOnDnMnTuXiIgIhg0bRlRUFAMGDAC0PZh69uypq9+1a1ccHBzo06cP4eHh7Nixg48++oi+fftibm4OwIcffsjmzZv59ttvOXPmDN9++y1///03Q4cOVeMShRCiaFEUOPTQBOdPGRotRFEmSSnx3D5+wxcjAw3bz95k94V4tcMRQgghir2QkBCmTZvG+PHjqVatGjt27GD9+vV4eXkBEBMTQ1RUlK6+lZUVYWFhJCQkEBgYSLdu3WjTpg3Tp0/X1albty7Lli1j3rx5VKlShfnz57N8+XKCgoLy/fqEEKLIuX4EYk+AoSlU66p2NEKoRqMoiqJ2EIVRUlIStra2JCYmFsuu6ePWnGb+nitUKmXDX4PrY2AgmX0hhBDFQ3FvAzwPea+EEOIJ/vc+HF0MVULgrV/VjkaIPJfbNoD0lBIv5IOm5bA2NeJ0dBKrj11XOxwhhBBCCCGEKBxSE+HUKu1jmeBcFHOSlBIvxN7ShIFNygDw/aazpGZkqRyREEIIIYQQQhQCJ1ZARjI4+oJnbbWjEUJVkpQSL6xvPR9K2ZoRnZjKvN1X1A5HCCGEEEIIIQo2vQnO+8gE56LYk6SUeGFmxoaMaF4BgBnbLnD7frrKEQkhhBBCCCFEAXbtIMSdBiMzqBqidjRCqE6SUuKltK/uRkVXG+6mZTJ9y3m1wxFCCCGEEEKIgutBL6lKb4G5nbqxCFEASFJKvBQDAw2ftvQDYPG+SK7E31c5IiGEEEIIIYQogFLuwOl/JzgPlAnOhQBJSok8UL9cSRqVdyQzW2HSpjNqhyOEEEIIIYQQBc/x5ZCZCk6VwL2m2tEIUSBIUkrkidEtfTHQwPqTsRyOvKN2OEIIIYQQQghRcCgKHP536F6gTHAuxAOSlBJ5wtfFhg4B7gB8vS4cRVFUjkgIIYQQQgghCoiofXDzDBhbQJVOakcjRIGhelJqxowZ+Pj4YGZmRkBAADt37nxq/bS0NMaMGYOXlxempqaUKVOGuXPn6tWZNm0aFSpUwNzcHA8PD4YNG0ZqaupLnVc82/DXK2BmbMCRqAQ2nopVOxwhhBBCCCGEKBge9JKq/BaY2aobixAFiKpJqeXLlzN06FDGjBnD0aNHadCgAcHBwURFRT1xn06dOrFlyxZCQ0M5e/YsS5cuxdfXV/f6kiVLGDVqFF988QURERGEhoayfPlyRo8e/VLnFc/mYmvGOw1KA/DtxjOkZ2arHJEQQgghhBBCqCz5NpxerX0c0FfVUIQoaDSKiuOsgoKCqFGjBjNnztSV+fn50a5dOyZOnJij/saNG+ncuTOXLl3C3t7+scccPHgwERERbNmyRVc2YsQIDhw4oOsN9bznfZykpCRsbW1JTEzExsYmV/sUB/fSMmn83Tbi76Uzrk1FetfzUTskIYQQIk9JGyD35L0SQghg78+w6VNw8Yf3dsp8UqJYyG0bQLWeUunp6Rw+fJjmzZvrlTdv3pw9e/Y8dp81a9YQGBjIpEmTcHNzo3z58owcOZKUlBRdnfr163P48GEOHDgAwKVLl1i/fj2tWrV64fOCdthgUlKS3iZysjI1Ymiz8gD8sOU8SakZKkckhBBCCCGEECpRFDj079C9AJngXIhHGal14vj4eLKysnB2dtYrd3Z2Jjb28fMRXbp0iV27dmFmZsaff/5JfHw8gwYN4vbt27p5pTp37szNmzepX78+iqKQmZnJwIEDGTVq1AufF2DixIl8+eWXL3PJxUbnmh7M232ZizfvM3P7RT55w/fZOwkhhBBCCCFEURO5G26dB2NL8O+odjRCFDiqT3SueSRTrChKjrIHsrOz0Wg0LFmyhFq1atGyZUumTJnC/Pnzdb2ltm/fztdff82MGTM4cuQIq1atYu3atXz11VcvfF6A0aNHk5iYqNuuXr36IpdbLBgZGjAq2A+AubsuE52Q8ow9hBBCCCGEEKIIetBLyr8DmMkwZiEepVpSqmTJkhgaGubonRQXF5ejF9MDrq6uuLm5YWv732oFfn5+KIrCtWvXABg7diw9evSgf//++Pv70759e/7v//6PiRMnkp2d/ULnBTA1NcXGxkZvE0/WzM+JWj72pGVm8/3ms2qHI4QQQgghhBD56/4tiFijfRzYJ1e7pKRnER6dhIpTPwuRr1RLSpmYmBAQEEBYWJheeVhYGHXr1n3sPvXq1SM6Opp79+7pys6dO4eBgQHu7u4AJCcnY2Cgf1mGhoYoioKiKC90XvH8NBoNY1pqe0v9efQ6p6MTVY5ICCGEEEIIIfLRsSWQlQ6u1aBU9WdW33MxnmZT/qHl9J00nfwPc3ZeIiE5/dXHKYSKVB2+N3z4cObMmcPcuXOJiIhg2LBhREVFMWDAAEA7ZK5nz566+l27dsXBwYE+ffoQHh7Ojh07+Oijj+jbty/m5uYAtGnThpkzZ7Js2TIuX75MWFgYY8eO5c0338TQ0DBX5xV5o6pHCdpULYWiwMT1ZyTbL4QQQgghhCgesrPh8Hzt42f0kkpJz2LcmtN0nb2f6/9OfXIp/j4T1kUQ9H9bGLHiOEei7sjnKVEkqTbROUBISAi3bt1i/PjxxMTEULlyZdavX4+XlxcAMTExREVF6epbWVkRFhbGkCFDCAwMxMHBgU6dOjFhwgRdnc8++wyNRsNnn33G9evXcXR0pE2bNnz99de5Pq/IOx+3qMCmU7HsuhDPP+du0riCk9ohCSGEEEIIIcSrdWUH3L4IJtZQucMTqx2OvMPI349zOf4+AN2CPPmwaTnCIm6weF8UETFJrDxyjZVHruHnakP32p60reaGlamqH+WFyDMaRdKtLyQpKQlbW1sSExNlfqlnmLA2nDm7LlPB2Zr1HzbA0ECWQRVCCFF4SRsg9+S9EkIUWyt6QfhqCOwLrafmeDktM4upYef5dcdFshVwsTHj2w5VaFTeUVdHURSOXk1gyb4o1p6IJi0zGwBLE0PaVXeje20v/Fzlb6somHLbBpD0qnjlBr9WlhWHrnL2xl1WHr5Gp5oeaockhBBCCCGEVnYWpNyB5FsPbbf1H6fcBp9GUGeQ2tGKwuBeHJxZq30c2DfHy6euJzJixXHO3rgLwFs13PiiTSVszY316mk0Gmp42lHD046xrf344/A1ftsfxaX4+yzZH8WS/VHU8CxBtyAvWlVxxczY8JVfmhB5TZJS4pUrYWHCkNfK8fX6CCaHnaV1VVcsTORXTwghhBBC5LHsbEhNeCSpdCtngunhspQEIBeDR85thNKNwLnSK74IUegdXQzZmeAWCC7+uuKMrGxmbLvIj1vPk5mtUNLKhK/b+9OiksszD1nCwoT+DUrTr74Pey/dYsm+KDadjuVIVAJHohL4al04HWq40zXIk9KOVq/y6oTIU5IZEPmiZ10vFuy9wrU7KczZeZkPmpZTOyQhhBBCCFGQKQqkJur3XHo0ofRo8inlDijZL3Y+M1uwcHhks9f+e2ELXNkJ/0yCTgvy9jpF0ZKdDUf+/R15aILz8zfuMnzFcU5e165KHlzZhQntKuNgZfpch9doNNQtU5K6ZUoSdzeV3w9pe09dT0hhzq7LzNl1mXplHegW5MXrFZ0xNlR1bTMhnknmlHpBMkfC8/vfset8uOwYliaGbP+oCY7Wz/cHWAghhCgIpA2Qe/JeCR1FgfR7T04m6coeKk+5re1t8iJMbf5LKFk4gLm9fpLp0c3cDgyf8n39jdMwsy6ggUF7wcnvxeISRd+FLbD4LTC1hRFnyDIyJ3TXJb7ffI70zGxszY0Z37YSb1YthUaTN3PtZmUr/HMujiX7oth6No4Hn/AdrU3pXNODzrU8cSthnifnEiK3ZE4pUeC0qVKK0F2XOXEtkWl/n+Pr9v7P3kkIIYQQQhQsigIZyY8klB6dk+lWzh5OWekvdj5jy6cklB6TbDK3ByOTvL1m50rg9yZErNH2luo4L2+PL4qOw//+blQN4UqSwsjf93Io8g4Ar/k6MfEtf5xtzPL0lIYGGl7zdeY1X2eu3Ulm2YGrLDt4lZt30/hx6wV+3naB13yd6BbkRcPyjrLwlChQpKfUC5Jv/l7Mvku36PzrPgwNNGwa2pCyTjLeWQghROEibYDck/eqkMhIeUzPpUd6LD1anpn6YucyMgOLko8kmJ70+N8Ek3HefoB/YbEnYVZ9QAPv7wfHCmpHJAqau7EwpSIoWayp+wef7MwiJSMLK1MjPm9dkY6B7nnWO+pZ0jOzCQu/wZL9key5eEtX7m5nTpdannQK9JCRK+KVkp5SokCqXdqBZn7O/B1xg282nGFOr0C1QxJCCCGEKB5SErS9OC7v0B8ul3H/xY5naPLkZNLD5eYPvWZikaeXlK9c/MG3tXZVtR3fwdtz1I5IFDRHF4GSxTmTinywVdszsE5pB77rWAV3u/z93TcxMqBVFVdaVXHl4s17/LY/ij8OX+PanRS+23SWqWHnaFHZhW5BntQp7ZBvyTIhHiU9pV6QfPP34i7E3aPFtB1kZSssf7c2QaUd1A5JCCGEyDVpA+SevFcFRFIM7PsZDs2H9LuPr2Ng9PhhceaPJpseTjBZQnH7IBtzHH5pCBoDeP8AlJTFe4SWkpVJ8vf+WKZEMzx9AOsNGzM62I8etb0wKCDD5VIzslh3IobF+yM5GpWgKy/taEm3IC861HDH1sJYvQBFkZLbNoAkpV6QNLJezpg/T7JkfxRV3W35c1C9AvOHWgghhHgWaQPknrxXKos/D7t/gBPL/5vPydEPavaDEl7/Jpj+TTKZ2hS/BNOLWtoFzq6HKiHw1q9qRyMKgLikVJYsDmVY3KckKJYMcFrMxJAgfEpaqh3aE52OTuS3/VGsPnqd++lZAJgaGdCmaim6BXlSzaOE9J4SL0WSUq+YNLJezs27aTT+bhv307OY3qU6b1YtpXZIQgghRK5IGyD35L1SybVDsGsqnFkH/NvU96wD9YZCueZgIEvEv5Too/Br4397Sx2EkmXVjkio6K/j0Yz93ykmZXxDc8PDnHDvQqW+MwvNZOL30jJZffQ6i/dFcib2v56UlUrZ0C3Ii7bVSmFpKrP+FCWZWdmExyRx6ModDkXextfFhg+a5n2vT0lKvWLSyHp507ecZ0rYOdztzNkyohGmRoZqhySEEEI8k7QBck/eq3ykKNql6HdPgys7/yuv0FKbjPIMUiuyoum3EDi3Eap2gfaz1I5GqOD2/XTG/u8U607E4Mxt9ph9gCHZ2mGdhXASfEVROBKVwJL9kaw9EUN6ZjYAVqZGtK/uRrfanvi6yN/xwuheWiZHo+5w8ModDkfe5mhUAsn/9o4DqOJuy5rB9fP8vDLRuSjw+jfwYfG+SK7dSWHR3kj6NyitdkhCCCGEEIVLViaEr4Zd0+DGSW2ZgRH4d4J6H4CTn5rRFV2NPtYmpU6sgIYfgUMZtSMS+Sgs/AajV50k/l4aRgYappU9gWFUNnjWLZQJKQCNRkOAlx0BXnaMbVWRlUeusWR/FJfj77NoXySL9kUS4GVH99qeBFd2xcxYOhQUVLGJqRy8cpvDkXc4eOU2ETFJZD/SFcnazIhALzsCve2p5WOvTqD/kp5SL0i++csbyw9G8cnKk9iaG7PjoyYysZ4QQogCT9oAuSfv1SuUngzHlsCeHyEhUltmbAkBvaHOILB1VzW8YmFxB7gQBtW6Q7uf1Y5G5IOk1AzG/xXOH4evAVDe2YrJb/vj/0cDSLoGb82GKp1UjjLvKIrCnou3WLI/ks2nb5D5b2ajhIUxHQPc6RrkVaDnzSoOsrMVzsXd1Q7Fu3Kbg1fucD0hJUc9dztzXRKqprc95ZysXvm8zjJ87xWTRlbeyMpWaPnDTs7euMs7DXwY06qi2iEJIYQQTyVtgNyT9+oVSL4NB0Nh/0xIvqUts3CAoAFQs7924nKRP64dgjlNQWMIQw6DvY/aEYlXaOf5m3z8xwliElPRaODdhqUZ1qw8ZpfCYGmIdqXK4RFgbKZ2qK9EXFIqyw9eZemBKKITU3Xl9cuWpFuQJ80qOmNsKPPVvWqpGVkcv5rAoUhtEupw5B2SUjP16hhowM/Vhpre9gR42RHobYerrXm+xyrD90ShYGigYVRLX/rMO8iCPZH0rOONh72F2mEJIYQQQhQsiddg7ww4PB8y7mvLSnhC3Q+gWjcwkfZTvnMPhDJN4eIW2DkZ2v6kdkTiFbiflsnEDREs3hcFgLeDBd93rEqg978J4MPztP9W61pkE1IATjZmDGlajkFNyrL9bByL90Wy/dxNdl2IZ9eFeJysTelc04POtTwpVSL/EyBF1a17aRyKvKMbinfqeiIZWfr9isyNDanhVYIAL3tqettR3dMOq0I0Ob30lHpB8s1f3lEUhe6h+9l94RZvVi3F9C7V1Q5JCCGEeKKC2gaYMWMG3333HTExMVSqVIlp06bRoEGDx9bdvn07TZo0yVEeERGBr68vABkZGUycOJEFCxZw/fp1KlSowLfffssbb7yR65gK6ntVqNw8C7t/gBPLIfvfb8OdK2snL6/UHgwLzwePIunqAQh9XTuP15AjYOeldkQiDx24fJuRvx8n6nYyAL3qePFJsC8WJv/ed4nXYJo/KNkw+BCUzPsVzAqyq7eTWXYwiuUHrxJ/Lx3Q9tJ5zdeZbrU9aVTO8ZUPEStKFEXhyq1kDl65zaErtzkUeYdLN+/nqOdobUpNbzsCvewJ9LbDz9WmQPZSk55SotDQaDSMDvajzU+7WHM8mn71fajqUULtsIQQQohCY/ny5QwdOpQZM2ZQr149fvnlF4KDgwkPD8fT0/OJ+509e1avoejo6Kh7/Nlnn7F48WJmz56Nr68vmzZton379uzZs4fq1eULpFcuar92Jb2z6/8r826gTUaVbQoa+aBXIHjUgtJN4NI2bW+pN6erHZHIA6kZWUzefJY5uy6jKOBWwpxJHapQr2xJ/YpHFmoTUt4Nil1CCsDD3oKPWvjyYdPybA6PZcm+KPZeusXfETf4O+IG7nbmdA3ypFOgByWtTNUOt8BJz8zmdHSidj6oyNscunKHW/fTc9Qr52RFoLc9gV521PS2x8PeHE0R+j9Aekq9IPnmL+8NX36MVUevU8vHnuXv1i5SN5oQQoiioyC2AYKCgqhRowYzZ87Ulfn5+dGuXTsmTpyYo/6DnlJ37tyhRIkSjz1mqVKlGDNmDO+//76urF27dlhZWbF48eJcxVUQ36sCTVHg/GbtSnpRe/4t1IBvK6g/TDtcTBQ8kXth3htgYAwfHNEOqxSF1vGrCQxfcYyL//ZQCQn04LPWflibPbIgU1YmTKsMd2Pg7VDw76BCtAXPhbh7/LY/ij8OX9XNdWRsqOGNyq50C/IkyMe+2H7OS0rN4EjkHQ5d0Q7FO34tgdSMbL06JoYGVPWw1Q3FC/Cyo4SFiUoRvxzpKSUKnREtKrD2ZAwHLt/m74g4Xq/orHZIQgghRIGXnp7O4cOHGTVqlF558+bN2bNnzxP20qpevTqpqalUrFiRzz77TG9IX1paGmZm+vOjmJubs2vXriceLy0tjbS0NN3zpKSk57mU4isrA06t1A7TiwvXlhkYQ9XOUO/DYtkDo1DxqgM+DeHyDtg1FVpPVTsi8QLSM7P5cet5Zmy/SFa2gqO1Kd++7c9rvk/4THJuozYhZeEAfm3yN9gCrKyTFZ+3qchHLSqw9kQ0S/ZHcexqAn8dj+av49GUdbKiW5Anb9Vwx9a8aK+8fj0h5d8V8bS9oM7euMujXYJKWBjrVsUL9LKjspstZsaG6gSsEklKiQLDrYQ5fev5MOufi3yzIYImFRwxKoBjY4UQQoiCJD4+nqysLJyd9T84OTs7Exsb+9h9XF1d+fXXXwkICCAtLY1FixbRtGlTtm/fTsOGDQFo0aIFU6ZMoWHDhpQpU4YtW7bwv//9j6ysrCfGMnHiRL788su8u7iiLv0+HFkEe3+CxKvaMhMrCOwDtQeBTSl14xO512iUNil1ZBE0GAG27mpHJJ5DREwSw1ccJyJGm0h/s2opvnyzEnaWT+mhopvgvBsYydC0R5mbGNIx0IOOgR6cup7Ikv1R/O/YdS7E3ePLv8L5duMZ2lQpRffaXlRxty30vaeyshXOxCb9OyG5dmW8mIdWKXzAy8FCNxdUTW87Spe0KvbzbsnwvRck3dFfjaTUDBpN2sad5AwmtKtM99oyWaQQQoiCpaC1AaKjo3Fzc2PPnj3UqVNHV/7111+zaNEizpw5k6vjtGnTBo1Gw5o1awC4efMm77zzDn/99RcajYYyZcrQrFkz5s2bR3Jy8mOP8bieUh4eHgXmvSow7t+CA79qt5Tb2jJLR6g9EAL7gXkJVcMTL2h+a7iyE2r2h1aT1Y5G5EJmVja/7LjEtL/PkZGlYGdhzIR2/rSq4vr0He9Ewg9VAUU7wb1DmXyJt7C7m5rB6mPRLNkXyZnYu7ryym42dAvyom21Uv9NIl/AJadncuxqwr/zQd3hSOQd7qVl6tUxNNBQqZQNgQ+G4nnb4WRddFdofJQM3xOFko2ZMR82Lce4v8KZ9vc52lV3K1TLWQohhBD5rWTJkhgaGuboFRUXF5ej99TT1K5dW2+uKEdHR1avXk1qaiq3bt2iVKlSjBo1Ch8fnycew9TUFFNT6THwRAlRsPdn7eTIGf8m9ux8oO6Qf5eTl2XUC7VGn2iTUkcWQv3hYOumdkTiKS7evMeIFcc5djUBgGZ+zkx8yx9H61z8DTuyAFCgdGNJSD0HazNjetT2onuQJ0ei7rBkXxRrT8Zw6noSo1ed5P/WRdC+hhvdgryo4GKtdrh64u6mcvjfBNShK7c5HZ1EZrZ+/x4rUyOqe5ag5r9D8ap5lig0STY1yTskCpyuQV7M33OFK7eS+fWfiwxvXkHtkIQQQogCy8TEhICAAMLCwmjfvr2uPCwsjLZt2+b6OEePHsXVNWfvADMzM9zc3MjIyGDlypV06tQpT+IuVm6c1s4XdfIPUP4d/uhaVbuSXsW2YFC85g8psnwagFc9iNyt/Xm3nKR2ROIxsrMV5u+5wrcbz5CWmY21mRHj2lTirRpuuRtClpUBR/9N4Af0ebXBFlEajYYAL3sCvOz5rHVFVh6+xpL9kVy5lczCvZEs3BtJoJcd3Wt78UZll3yfY0lRFC7evPfvhOR3OBx5myu3cvYQdrEx+3cYnnY4nq+LDYbFfCjei5CklChwTIwM+OQNXwYuOcLsnZfpVtsLZ5vi081RCCGEeF7Dhw+nR48eBAYGUqdOHX799VeioqIYMGAAAKNHj+b69essXLgQgGnTpuHt7U2lSpVIT09n8eLFrFy5kpUrV+qOuX//fq5fv061atW4fv0648aNIzs7m48//liVayx0FAWi9mpX0ju/6b9yn0ZQfyiUbgKFfA4V8RiNPoaFbeHwfO2KiTbPGAYm8tXV28mM/P04+y9rh802KFeSb9+uQqkSz9FL8ex6uHcDLJ20K2OKl2JvacI7DUvTr74Pey7eYsn+SDaH39D2SIq8g91fxnQK9KBLLU+8S1q+khjSMrM4dT3x37mgtEmoO8kZenU0GqjgbE2gt51uTii3EuaFfi6sgkCSUqJAeqOyCwFedhyOvMOUzef4tkMVtUMSQgghCqyQkBBu3brF+PHjiYmJoXLlyqxfvx4vL+3cjDExMURFRenqp6enM3LkSK5fv465uTmVKlVi3bp1tGzZUlcnNTWVzz77jEuXLmFlZUXLli1ZtGgRJUqUyO/LK1yys+HcBm0y6tqBfws12h5R9T4EtxpqRideNZ9G4FEbru7T9pYK/kbtiATani9LD1zl63Xh3E/PwsLEkDGt/Ohay/P5kwqH/p3gvHp3MCzaq8flJwMDDfXLlaR+uZLcSEpl+cGrLD0QRUxiKr/suMQvOy7RoFxJugV50czP6aUWxEpITudw5H9D8Y5fSyQ9M1uvjqmRAdU8SmiTUN721PC0K/KrBapFJjp/QQVtktOi6HDkbd6euRcDDWz4sGGBG1cshBCieJI2QO4Vq/cqMx1O/q5NRMSf1ZYZmmjniqr7gcw7U5xc3AqL2oORGXx4HKxd1I6oWItNTOXjlSfYce4mALW87fm+Y1U8HSye/2C3L8H06oAGPjwGdt55Gap4RGZWNtvO3mTJ/kj+OXeTB5kLZxtTQmp60qWWB662T+/lpigKV2+ncCjytm4o3rkb93LUc7A0IcDrv6F4lUrZYmIkK8G/DJnoXBR6AV72BFd2YcOpWCZuiGB+n1pqhySEEEIIoS/tLhxeAPtmQNJ1bZmpDdTsB0EDwTr3k82LIqJ0E3Cvpe0pt3s6vPF/akdULCmKwupj1/nif6dJSs3ExMiAj1tUoG89HwxedN6fwwu0/5Z5TRJS+cDI0IDXKzrzekVnrt5O5rcDUaw4eJUbSWlM33Ken7aep6mfM92CPGlYzhEDAw2ZWdlExNzl4JXbHI68w8Ert4m7m5bj2KVLWuoNxfMpaSlD8VQiPaVeULH65k9Fl+Pv8/qUf8jMVljSP4h6ZUuqHZIQQohiTtoAuVek36v78bB/FhyYDakJ2jIrF6g9EAL7gJmtquEJlV34Gxa/DUbmMPQEWDmpHVGxEn8vjTF/nmTT6RsAVHW3ZXKnqpR1eomRF5npMLUi3L8JIYvBr00eRSueR3pmNptOx7J4X6RubjAAD3tzPO0tOBqVQHJ6lt4+xoYaKrvZEuilHYoX4GVHSStZKfZVk55SokjwKWlJ99ra1fj+b30Efw2u/+LfbAghhBBCvKw7V2DPT3B0EWSmasscymqH6FXtDEbyQUcAZZqCWyBcPwR7pkPzCWpHVGxsOBnDmNWnuH0/HWNDDR82LceARmVeag4iAM6s1SakrFyg/Bt5E6x4biZGBrSpWoo2VUtxIe4uS/ZHsfLwNa7eTuHq7RQArM2MdEPxArzsqOpeAnMTWeW0oJKklCjwPmhajpWHr3E6OonVx67zVg13tUMSQgghRHETe1I7efnpP0H591v4UjW0K+n5tgYD+cAjHqLRQKNP4LeOcDAU6n4IVo5qR1WkJSSn88Wa0/zvWDQAvi7WTO5UlUql8qjX4uF/Jziv0UMmOC8gyjpZ80WbSnzcwpdNp2O5n55JgJcd5Z2spSNDISJJKVHg2VuaMLBJGSZtPMv3m87S0t8VM2Np+AkhhBDiFVMUuLILdk2Fi1v+Ky/TVJuM8m6gTT4I8TjlXodS1SH6KOz9EV4fr3ZERda2M3F8svIEcXfTMNDAwMZl+KBpOUyN8ugzw62LcHkHoIEaPfPmmCLPmJsY0q66m9phiBck08mLQqFvPR9K2ZoRnZjK3N2X1Q5HCCGEEEVZdjaEr4E5TWFBa21CSmMAld+G93ZAj1Xg01ASUuLpNBpoNEr7+MAcuH9L3XiKoLupGYxaeYI+8w8SdzeN0o6WrBxYl49a+OZdQgr+6yVV7nUo4Zl3xxVCSFJKFA5mxoaMaF4BgJnbLnLrXs4VFIQQQgghXkpmGhxZCD/XhBU94PphMDKDmv1hyBHoMBdcq6odpShMyrcA12qQcV/bW0rkmT0X43lj2k6WHbyKRgP96vuw/oMGVPe0y9sTZabBsd+0jwP65O2xhRCSlBKFR/vqblR0teFuWiY/br2gdjhCCCGEKCpSk2D3D/BDVVgzBG5d0K6e12AkDD0FrSaDvY/aUYrC6MHcUqBdqTH59tPri2dKSc9i3JrTdJ29n+sJKXjYm7P0ndqMbV3x1UzxEfEXJN8C61JQrnneH1+IYk7mlBKFhoGBhk9b+tE9dD+L90XSq643PiUt1Q5LCCGEEIXV3RuwfyYcnAtpidoy61JQ530I6AWmL7F8vBAPVAgGF3/tZPl7f4amY9WOqNA6HHmHkb8f53L8fQC6BnnyaUs/rExf4cfaQw8mOO8JhvLxWYi8JneVKFTqlytJo/KO/HPuJpM2nmFm9wC1QxJCCCFEYXPrIuz5UTskJ+vfKQFKlod6H4J/JzAyUTc+UbQ86C21vDvs/0Wb9LSwVzuqQiUtM4upYef5dcdFshVwsTHj2w5VaFT+Fa9oePMcRO7SziknE5wL8UpIUkoUOqNb+rLz/E02nIrlcORtArzkP3UhhBBC5EL0Mdg9DcL/B0q2tsy9JtQfBuWDwUBmthCvSIVW4FwZbpyCfTPhtTFqR1RonLqeyIgVxzl74y4Ab1V344s2lbC1MH71Jz88X/tvuRZgK6u7CfEqyP+8otDxdbGhY4AHAF+vi0BRFJUjEkIIIUSBpShwaTssbAu/NoLTf2oTUuWaQ+/10C8MfFtJQkq8WgYG0Ohj7eP9syDljrrxFAIZWdn88Pd52v28m7M37uJgacKs7gFMCamWPwmpjBQ4/u8E54EywbkQr4r0lBKF0vDm5VlzPJojUQlsPBVLsL+r2iEJIYQQoiDJzoKINbBrGsQc05ZpDMG/g3aYnnMlNaMTxZFvG3CqCHHhsG8WNBmtdkQF1rkbdxmx4jgnr2vneguu7MKEdpVxsDLNvyDC/6dNHtp6QNlm+XdeIYoZ+UpIFErONma800C7Cs63G8+QnpmtckRCCCGEKBAyUuHQXPgpEH7vrU1IGZlDrffgg6Pw1q+SkBLqeLi31L6ZkJqobjwFUFa2wq87LtL6x12cvJ6IrbkxP3SuxoxuNfI3IQX6E5wbvIJV/YQQgPSUEoXYu43K8NuBKK7cSua3/ZH0ridLNQshhBDFVmoiHAzVfti/H6ctM7fTJqNqvQuWDurGJwSAX1tw9IWbZ7STnj9IUgmuxN9n5O/HORSpHdrYpIIj37xdBWcbs/wPJi4Cru7T9q6s3iP/zy9EMSI9pUShZWVqxNBm5QH4Yct5klIzVI5ICCGEEPkuKQY2j4UplWDLl9qElI07vPENDDutHSIlCSlRUBgYQMOPtI/3/gypSerGUwBkZyss3HuF4B92cijyDlamRnz7tj9ze9dUJyEF//WSqhAMNjJNiBCvkiSlRKHWuaYHZRwtuZOcwcztF9UORwghhBD5Jf4CrBkCP1SBPdMh/S44+kH7X+DDY1B7IJhYqh2lEDlVag8ly0NqAhz4Re1oVHU9IYUec/fz+f9Ok5KRRZ3SDmz4sAEhNT3RaDTqBJWeDMeXaR/LBOdCvHKSlBKFmpGhAaOC/QCYu+sy1xNSVI5ICCGEEPnin2/gyELISgfPOtBlOQzcA1U7g2E+rMwlxIsyMISG/w7b2/szpN1VNx4VKIrCikNXeWPqDnZfuIWZsQHj2lRkSf8gPOwt1A3u9J+QlgglvKD0a+rGIkQxIEkpUeg183Oilo89aZnZTN50Vu1whBBCCJEf6g2F8sHQdzP03QgV3tAOjRKiMKj8FjiU067uduBXtaPJV3FJqfRfcIiP/zjB3bRMqnuWYP0HDehdzwcDA5V6Rz3s8L9D9wJ6yd8UIfKB3GWi0NNoNIxpqe0tterodaZvOU92tqJyVEIIIYR4pVwqQ9dl4BmkdiRCPD8Dw//mltrzE6TdUzeefPLX8WiaT9vBljNxmBgaMCrYlz8G1KW0o5XaoWnFnoJrB8HACKp1VzsaIYoFSUqJIqGqRwneaaBdfW9K2Dn6LjjInfvpKkclhBBCCCHEE1R+G+xLQ8ptODhH7Wheqdv303l/yRGGLD1KQnIGlUrZ8NeQ+gxoVAbDgtA76oEHvaR8W4G1s7qxCFFMSFJKFBljWlVk0ttVMDUyYPvZm7T+cRfHryaoHZYQQgghhBA5GRo91FvqR0i/r248r0hY+A2aT/2HdSdjMDTQ8GHTcqx+vx4VXKzVDk1f+n04sUL7OEAmOBciv0hSShQpnWp6sGpQXbwcLLiekELHWXtZtC8SRZHhfEIIIYQQooDx7wR2PpAcDwdD1Y4mT91Ly2TEiuO8s/AQ8ffSKedkxepB9Rj2enmMDQvgx9BTKyEtSfvz8GmkdjRCFBsF8K+BEC+nUilb/hpSn+YVnUnPymbs6lMMW36M+2mZaocmhBBCCCHEfwyNoOFI7eM90yE9Wd148sjFm/do9/NuVh65hkYD7zUszV9D6uPvbqt2aE926MEE571lgnMh8pHcbaJIsjEz5pceAYxp6YehgYbVx6Jp9/NuLsQVvyV3hRBCCCFEAVYlBEp4wf2bcGiu2tG8tLDwG7T7aTcX4u7hYmPGivfqMLqlH2bGhmqH9mQxxyH6CBgYQ7VuakcjRLEiSSlRZGk0Gt5pWJql79TGydqU83H3ePOn3fx1PFrt0IQQQgghhNAyNP6vt9TuHyAjRd14XlB2tsKUsHO8s/AQd9Myqeltx5oh9ajpba92aM/2oJeUXxuwclQ3FiGKGdWTUjNmzMDHxwczMzMCAgLYuXPnU+unpaUxZswYvLy8MDU1pUyZMsyd+983Co0bN0aj0eTYWrVqpaszbty4HK+7uLi8smsU6qrlY8/aD+pTu7Q9yelZDFl6lHFrTpOema12aEIIIYQQQkCVzmDrCffj4PB8taN5bokpGbyz8BDTt5wHoFcdL5b0r42TtZnKkeVC2l04+bv2caBMcC5EfjNS8+TLly9n6NChzJgxg3r16vHLL78QHBxMeHg4np6ej92nU6dO3Lhxg9DQUMqWLUtcXByZmf/NFbRq1SrS09N1z2/dukXVqlXp2LGj3nEqVarE33//rXtuaFiAu5OKl+ZkbcbifkFMCTvHjO0Xmb/nCseuJvBztxq4lTBXOzwhhBBCCFGcGZlAg+GwdijsmqZd/c24ECR0gPM37vLuosNcjr+PiZEB/9fenw4B7mqHlXsn/4D0e+BQFrwbqB2NEMWOqkmpKVOm0K9fP/r37w/AtGnT2LRpEzNnzmTixIk56m/cuJF//vmHS5cuYW+v7Qbq7e2tV+dB+QPLli3DwsIiR1LKyMhIekcVM0aGBnz8hi81PO0YvuIYx64m0Hr6TqZ1rk6j8tJNVwghhBBCqKhaN9g5GRKvwpEFEPSe2hE904aTMYz8/Tj307MoZWvGLz0CC/Zk5o9z+KEJzjUaVUMRojhSbfheeno6hw8fpnnz5nrlzZs3Z8+ePY/dZ82aNQQGBjJp0iTc3NwoX748I0eOJCXlyeOuQ0ND6dy5M5aWlnrl58+fp1SpUvj4+NC5c2cuXbr08hclCoVmFZ1Z90EDKrvZcCc5g97zDjA17BxZ2YraoQkhhBBCiOLKyATqD9M+3jUVMlLVjecpsrIVJm08w8AlR7ifnkXt0vYFf3W9x7l+RDvJuaEJVO2qdjRCFEuqJaXi4+PJysrC2dlZr9zZ2ZnY2NjH7nPp0iV27drFqVOn+PPPP5k2bRp//PEH77///mPrHzhwgFOnTul6Yj0QFBTEwoUL2bRpE7NnzyY2Npa6dety69atJ8ablpZGUlKS3iYKLw97C/4YUJcutTxRFPhhy3l6zzvA7fvpz95ZCCGEEEKIV6F6d7Bxg7sxcHSR2tE8VkJyOn3mH2TG9osA9K/vw+J+QThYmaoc2Qt40EuqYluwdFA3FiGKKdUnOtc80kVSUZQcZQ9kZ2ej0WhYsmQJtWrVomXLlkyZMoX58+c/trdUaGgolStXplatWnrlwcHBvP322/j7+9OsWTPWrVsHwIIFC54Y58SJE7G1tdVtHh4ez3upooAxMzZk4lv+TO5YFTNjA3aej6fV9J0cibqjdmhCCCGEEKI4MjLV7y2VmaZuPI+IiEnizZ92s+PcTcyMDfihczU+a10RI0PVP1Y+v9QkOLlS+zhAJjgXQi2q/fUoWbIkhoaGOXpFxcXF5eg99YCrqytubm7Y2v7XLdTPzw9FUbh27Zpe3eTkZJYtW5ajl9TjWFpa4u/vz/nz559YZ/To0SQmJuq2q1evPvO4onB4O8Cd1e/Xw6ekJTGJqYT8spf5uy+jKDKcTwghhBBC5LPqPcDaFZKuw9HFakejs+Z4NG/N2EPU7WTc7cxZObAubau5qR3Wizu5AjLuQ8kK4FVX7WiEKLZUS0qZmJgQEBBAWFiYXnlYWBh16z7+j0K9evWIjo7m3r17urJz585hYGCAu7v+Cg8rVqwgLS2N7t27PzOWtLQ0IiIicHV1fWIdU1NTbGxs9DZRdPi62LBmcD1a+ruQkaUw7q9whiw9yr20zGfvLIQQQgghRF4xNnukt5S600tkZmXzf+sj+GDpUVIysmhQriR/Da5PpVKFbP6ohykKHJIJzoUoCFTtZzl8+HDmzJnD3LlziYiIYNiwYURFRTFgwABA2zupZ8+euvpdu3bFwcGBPn36EB4ezo4dO/joo4/o27cv5ubmescODQ2lXbt2ODjkHBs8cuRI/vnnHy5fvsz+/fvp0KEDSUlJ9OrV69VesCjQrM2M+blrDca2roiRgYa1J2Jo+9Muzt24q3ZoQgghiiBvb2/Gjx9PVFSU2qEIIQqaGr3AykW7Et+xJaqFcft+Or3mHeDXHdpFoQY0KsP8PrWwszRRLaY8ce0Q3DgFhqZQtbPa0QhRrKmalAoJCWHatGmMHz+eatWqsWPHDtavX4+XlxcAMTExeg01KysrwsLCSEhIIDAwkG7dutGmTRumT5+ud9xz586xa9cu+vXr99jzXrt2jS5dulChQgXeeustTExM2Ldvn+68ovjSaDT0q+/Dsndr42xjysWb92n7027+d+y62qEJIYQoYkaMGMH//vc/Spcuzeuvv86yZctIS3vx+WNmzJiBj48PZmZmBAQEsHPnzifW3b59OxqNJsd25swZvXrTpk2jQoUKmJub4+HhwbBhw0hNLbgrgglRZBibQf2h2sc7p6jSW+rU9UTa/LiL3RduYWFiyM9dazAq2BdDgyLQq+jBBOeV3wILe3VjEaKY0ygycc4LSUpKwtbWlsTERBnKV0TF30vjw2VH2X1Buypj99qejG1dEVMjQ5UjE0IIoaa8bgMcP36cuXPnsnTpUjIzM+natSt9+/alRo0auT7G8uXL6dGjBzNmzKBevXr88ssvzJkzh/DwcDw9PXPU3759O02aNOHs2bN61+Do6Iihofb/uSVLltCvXz/mzp1L3bp1OXfuHL179yYkJISpU6fmKi5pLwnxEjJS4IeqcO8GvPkj1Oj57H3yyKoj1xi96iRpmdl4OVjwa49AKrhY59v5X6mUBJjsC5kp0HczeAapHZEQRVJu2wCFcJkEIfJHSStTFvYNYshrZQFYvC+KTrP2cu1OssqRCSGEKEqqVq3KDz/8wPXr1/niiy+YM2cONWvWpGrVqsydOzdXC29MmTKFfv360b9/f/z8/Jg2bRoeHh7MnDnzqfs5OTnh4uKi2x4kpAD27t1LvXr16Nq1K97e3jRv3pwuXbpw6NChl75mIUQuGJtDvQ+1j3d8D1kZr/yUGVnZjFtzmuErjpOWmU2TCo6seb9+0UlIAZxYrk1IOVUEj1rPri+EeKUkKSXEUxgaaBjRvALz+tSkhIUxx68l0mr6LradiVM7NCGEEEVERkYGK1as4M0332TEiBEEBgYyZ84cOnXqxJgxY+jWrdtT909PT+fw4cM0b95cr7x58+bs2bPnqftWr14dV1dXmjZtyrZt2/Req1+/PocPH+bAgQMAXLp0ifXr19OqVasnHi8tLY2kpCS9TQjxEgL6gKUjJERqkymvUPy9NLrN2c/8PVcA+OC1soT2qomthfErPW++0pvgvI9McC5EAWCkdgBCFAZNKjixdkh9Bi05wolrifSZf5Ahr5VlaLPyRWNcvRBCiHx35MgR5s2bx9KlSzE0NKRHjx5MnToVX19fXZ3mzZvTsGHDpx4nPj6erKwsnJ2d9cqdnZ2JjY197D6urq78+uuvBAQEkJaWxqJFi2jatCnbt2/Xna9z587cvHmT+vXroygKmZmZDBw4kFGjRj0xlokTJ/Lll1/m9i0QQjyLiQXU/QDCxmp7S1XpDIZ5/xHu+NUEBiw+TExiKlamRkzuVJUWlVzy/Dyqu7ofbkaAkTlU6aR2NEIIpKeUELnmbmfB7wPq0KO2dkL8H7deoOfc/cTfe/FJaYUQQhRfNWvW5Pz588ycOZNr167x/fff6yWkACpWrEjnzrlbGUrzyDf+iqLkKHugQoUKvPPOO9SoUYM6deowY8YMWrVqxffff6+rs337dr7++mtmzJjBkSNHWLVqFWvXruWrr756YgyjR48mMTFRt129ejVXsQshnqJmP7AoCXcuw8kVeX74FQev0vGXvcQk/n979x1XZfn/cfx12EPAgSCi4p44IffMUVqamWXONBuOypEN2/nzm00zKy1NM0dpzixNw9yamXvvASqKk6Eyz/n9cStJLkDgPsD7+Xich/e5uc99vzmlXHzOdX+ueMoW9WTBwIZ5syAF/86SCn4M3AuaGkVEDJopJZIBrk6O/F/HYEJLF+L1uTtZd+g8D49dy9fdaxMSpJU7REQk/Y4cOXLXlX89PT35/vvv73iMr68vjo6ON82KioqKumn21J3Ur1+f6dOnpz5/++236dmzJ8888wwA1atX5/Llyzz33HO8+eabODjc/Nmmq6srrq6u6b6miKSDiyc0fBGWvQurP4HqT2TJbKnEZCsjftvN9A3Gauetq/oz+omaeLnlodv1bnTlAuyeb2yH9jE3i4ik0kwpkUx4pFYgv7zQiLJFPTkdE0+Xbzcwae3RdDWjFRERAaNo9Pfff9+0/++//85QM3EXFxdCQkIICwtLsz8sLIyGDRum+zxbt24lICAg9fmVK1duKjw5Ojpis9n0804kp933DLgXhgtHYNfcez5dVEw8XSduYPqGcCwWGNq6It/2CMm7BSmA7TMhJQH8q0NgiNlpROQaFaVEMqmivxcLX2jMwzUCSLba+L/f9jDwxy3Exmf/yigiIpL7DRw48Ja3t508eZKBAwdm6FxDhw7lu+++Y/Lkyezdu5chQ4YQHh5Ov379AOO2ul69/l1OfsyYMSxYsICDBw+ye/duhg8fzty5c3nhhRdSj2nfvj3jx49n5syZHD16lLCwMN5++206dOiQZpU+EckBrgWg4bW/n6s/AWtKpk+1+fhFHv5yLZuPX8TLzYlJT4XyUssKOOTlPqk2G2y+Nus0tLcanIvYEd2+J3IPCrg68WXX2txXujAjF+1h8c7T7I2MZXyPOlQu5m12PBERsWN79uyhTp06N+2vXbs2e/bsydC5unTpwvnz5xkxYgSRkZEEBwezePHi1NsDIyMjCQ8PTz0+MTGRYcOGcfLkSdzd3alWrRqLFi2iXbt2qce89dZbWCwW3nrrLU6ePEnRokVp3749//vf/zL5HYvIPan7HKz/Es4fhF3zoMbjGXq5zWbjx43hvLdwN0kpNir4FWBCr1DK+HpmU2A7cnw9nDsAzp7G7Y8iYjcsNs2/zpSYmBh8fHyIjo7G21vFB4Et4Rd5YcYWTkXH4+bswP86VuexkBJmxxIRkSyWVWOAIkWK8Ntvv9GgQYM0+9evX89DDz3ExYsX7zWq6TReEsliqz+B5SPBtyIM2AAO6Zu1mJCcwru/7GbmP8bszLbBxfjk8ZoUcM0ncxTmPgM7Z0OdXtDhS7PTiOQL6R0D6PY9kSxSp1QhfnupCU0q+BKfZOXl2dsZPm8n8UmZn14tIiJ5V+vWrVNXq7vu0qVLvPHGG7Ru3drEZCJit+o+D24FjVk/15t230Vk9FW6fLuBmf9EYLHAqw9WYlz3OvmnIHX5POz5xdgOUYNzEXujopRIFirs6cKUPnUZ1LICFgv8tDGczt+sJ+LCFbOjiYiInfnss8+IiIggKCiIFi1a0KJFC8qUKcPp06f57LPPzI4nIvbIzRsaXOs5t/oTsFrvePjfR87T/su1bIu4hI+7M1P61GVA8/JY8lNPpe0/QkoiBNSEwJtvmRYRc6koJZLFHB0sDGldkSl96lLIw5ldJ2N4aOwalu05Y3Y0ERGxI4GBgezYsYOPP/6YqlWrEhISwhdffMHOnTspWbKk2fFExF7Vex5cfeDsPtj7yy0Psdls/LD+GN2/+5tzcYlULubFry80plnFojkc1mQ2G2yeYmxrlpSIXVJPqUxSjwRJj1OXrjJgxha2RVwCoH/zcrzcuiJOjqoHi4jkVhoDpJ/eK5FssmIUrPoQ/KpCv3Xg8O/YMj4phTfm72TelpMAdKhZnA8fq46HSz65Xe9GR1fDD+3BpQC8vA9cvcxOJJJvpHcMkA//ZRLJOcULuvPz8w34YPFepqw/xviVh9kWfomxXWtT1MvV7HgiImIH9uzZQ3h4OImJiWn2d+jQwaREImL36veDDeMgag/s+xWqPgLAiYtX6Dd9M7tOxuBggTfaVaFv4zL563a9G2363viz+uMqSInYKRWlRLKZi5MD73WoRp2gQrw+dwd/HTnPQ2PX8FW3OtQtU9jseCIiYpIjR47w6KOPsnPnTiwWC9cnr1//5TElRQtliMhtuBeCev1g9cew6mOo3J71Ry/wwo9buXA5kcKeLnzVtTYNy/uandQ8cWdh76/Gdqhu3ROxV5m6hygiIoITJ06kPt+4cSODBw9mwoQJWRZMJK/pULM4C19oRAW/AkTFJtB14gYmrD6M7qAVEcmfBg0aRJkyZThz5gweHh7s3r2b1atXExoaysqVK82OJyL2rn5/cPGCM7sImz+ZnpM2cuFyIsGB3ix8oVH+LkgBbJsB1iQoXsdoci4idilTRalu3bqxYsUKAE6fPk3r1q3ZuHEjb7zxBiNGjMjSgCJ5SXk/LxYMbMQjtYqTYrXxweJ9PD9tMzHxSWZHExGRHPbXX38xYsQIihYtioODAw4ODjRu3JhRo0bx0ksvmR1PROydR2GSQp8DoPj2saRYrXSqE8icfg0pUcjD5HAms1ph87Vb9zRLSsSuZaootWvXLurWrQvAzz//THBwMOvXr+fHH39kypQpWZlPJM/xdHViTJda/F/HYFwcHfhjzxnaf7mWPadizI4mIiI5KCUlhQIFCgDg6+vLqVOnAAgKCmL//v1mRhORXCDiwhW67w4hzuZGNYfjTKp3ls8er4mbs6PZ0cx3dCVcPAau3hD8mNlpROQOMlWUSkpKwtXVaNK8bNmy1EaclStXJjIyMuvSieRRFouFnvWDmN2vAYEF3Tl+/gqPjlvHz5sizI4mIiI5JDg4mB07dgBQr149Pv74Y9atW8eIESMoW7asyelExJ6tPnCW9l+tZeMZmO3QFoCWUd+TT9uZ3+x6g/MaT4CLp7lZROSOMlWUqlatGt988w1r1qwhLCyMBx98EIBTp05RpEiRLA0okpfVLFmQ315sTPNKRUlItvLqnB28Omc78Ulqbisikte99dZbWK1WAEaOHMnx48dp0qQJixcvZuzYsSanExF7ZLPZGL/yML2/38ilK0nULFmQts+PBGdPiNwOB5aaHdF8sWdg/2JjO0S37onYu0wVpT766CO+/fZbmjdvTteuXalZ02gct3DhwtTb+kQkfQp5ujD5qft4uXVFLBb4edMJOo1bz/Hzl82OJiIi2eiBBx6gU6dOAJQtW5Y9e/Zw7tw5oqKiuP/++01OJyL25nJCMgN/3MJHS/ZhtUGX0JLMeq4+xYqVgLrPGAet+hDy+yI6W6eBNRlK1IViwWanEZG7sNgyufRXSkoKMTExFCpUKHXfsWPH8PDwwM/PL8sC2quYmBh8fHyIjo7G29vb7DiSR6w9eI5BM7dy/nIiXm5OfPp4TR6oVszsWCIicoOsGAMkJyfj5ubGtm3bCA7Ou780abwkkjWOnrvM89M2ceBMHM6OFt7rUI1udUthsVy7Ye/yORhTHZKuQLfZULGNuYHNYrXC2JpwKRw6joda3cxOJJJvpXcMkKmZUlevXiUhISG1IHX8+HHGjBnD/v3780VBSiS7NK7gy28vNSYkqBCx8ck8P20zoxbvJTnFanY0ERHJQk5OTgQFBZGSotu1ReTOlu87Q4ev1nLgTBxFvVyZ+Vx9utcL+rcgBeDpC/f1NbZXfZR/Z0sdXm4UpNx8oNqjZqcRkXTIVFHqkUceYerUqQBcunSJevXq8dlnn9GxY0fGjx+fpQFF8psAH3dmPlefvo3LAPDt6iN0m/g3UTHxJicTEZGs9NZbbzF8+HAuXLhgdhQRsUNWq42xfx6k7w+biI1PJiSoEItebExIUOFbv6DhS+DkDic3weE/czasvdh8rcF5za7g7G5uFhFJl0wVpbZs2UKTJk0AmDNnDv7+/hw/fpypU6eqMadIFnB2dODth6syrnsdCrg6sfHYBdqNXctfh8+bHU1ERLLI2LFjWbNmDcWLF6dSpUrUqVMnzUNE8q/Y+CSen76Z0WEHsNmgR/1S/PRsffy83W7/ogJ+EPq0sb0yH86WiomE/b8b22pwLpJrOGXmRVeuXMHLywuAP/74g06dOuHg4ED9+vU5fvx4lgYUyc/aVQ+gcjEv+k/fwv4zsXT/bgOvPFCZ55uWxcFBi/6KiORmHTt2NDuCiNihQ1FxPD9tE4fPXsbF0YGRHYN54r6S6Xtxo0GwaRKc2AhHVkC5fLRowtZpYEuBUg3Ar7LZaUQknTJVlCpfvjwLFizg0UcfZenSpQwZMgSAqKgoNbEUyWJlixZgwcBGvLlgJ/O2nOSjJfvYfPwinz1eEx8PZ7PjiYhIJr377rtmRxARO/PH7tMM/Xk7cQnJFPN245ueIdQqWTD9J/DyN2YJ/T3emC1VtgVY8sEHmdYU2PyDsa1ZUiK5SqZu33vnnXcYNmwYpUuXpm7dujRo0AAwZk3Vrl07SwOKCLi7OPLZ4zUZ1ak6Lk4OLNt7hoe/WsOuk9FmRxMRERGRe2S12hj9x36em7aZuIRk6pYpzK8vNs5YQeq6RoPA0RUiNsDR1Vme1S4dWgYxJ8C9EFR9xOw0IpIBmSpKde7cmfDwcDZt2sTSpUtT97ds2ZLPP/88y8KJyL8sFgtd65Zibr+GlCjkTsSFq3Qav56fNoZjy289A0RE8gAHBwccHR1v+xCR/CH6ahLPTN3E2OWHAOjTqDQznqlHUS/XzJ3QOwBCehvbqz7KmpD2btP1BufdwPkOfbdExO5k6vY9gGLFilGsWDFOnDiBxWIhMDCQunXrZmU2EbmF6iV8WPRiE4b+vI0/90UxfN5ONh27yMiOwbi76JcYEZHcYv78+WmeJyUlsXXrVn744Qfef/99k1KJSE46cCaW56Zu4tj5K7g6OTCqU3U61Slx7yduPNhYie74Oji6Bso0ufdz2qvoE3Dw2kSJ68U4Eck1MlWUslqtjBw5ks8++4y4uDgAvLy8ePnll3nzzTdxcMjUBCwRSScfD2cm9grlm9WH+XTpfuZuOcHuU9GM616HskULmB1PRETS4ZFHbr7FpHPnzlSrVo1Zs2bRt29fE1KJSE5ZvDOSYbO3cyUxhcCC7nzbM4TgQJ+sObl3cajTC/75zpgtlZeLUlumgc0KQY2haEWz04hIBmWqevTmm2/y1Vdf8eGHH7J161a2bNnCBx98wJdffsnbb7+d1Rnzn5Rk2DIVrFazk4gdc3CwMKB5eaY/Uw/fAi7sOx1Lh6/W8fvOSLOjiYjIPahXrx7Lli0zO4aIZJMUq42PluxjwIwtXElMoVH5Ivz6YuOsK0hd13gIODjDsTVwbF3WntteXP+9CSBUDc5FcqNMFaV++OEHvvvuO/r370+NGjWoWbMmAwYMYOLEiUyZMiWLI+YzNhvMexYWvghLXjeei9xBw3K+LHqpCXVLFyYuIZn+M7bwf7/tISlFRU0Rkdzm6tWrfPnll5QokQW374iI3bl0JZHe329k/MrDADzXtCw/9KlLYU+XrL+YTwmo09PYXvVh1p/fHhz8A2JPgUcRqNLe7DQikgmZun3vwoULVK5c+ab9lStX5sKFC/ccKl+zWKDig7B7Hmz8Fly9oKVmn8md+Xu7MePZenyydD8TVh9h0tqjbI+4xFfd6lDMR80eRUTsUaFChbDcsFS7zWYjNjYWDw8Ppk+fbmIyEckOe07F8Pz0TURcuIqbswMfd65Jh5rFs/eijYcat7cdXQ3H/4KgBtl7vZy2+VqD81rdwCmTjeFFxFSZKkrVrFmTr776irFjx6bZ/9VXX1GjRo0sCZav1ewCCTGweBis+RTcvI2lXUXuwNnRgTfaVaFOqUK8Mns7m45f5OEv1/DFk7VpVN7X7HgiIvIfn3/+eZqilIODA0WLFqVevXoUKlTIxGQiktV+2XaS1+buID7JSqnCHnzbM4QqAd7Zf+GCJaF2d9g8xegt1WtB9l8zp1wKh4NhxnaIbt0Tya0yVZT6+OOPeeihh1i2bBkNGjTAYrGwfv16IiIiWLx4cVZnzJ/qPguJcbDsPQh7x5gxFfq02akkF3gwuBiVi3nRf8YW9kbG0HPS3wxtXZEBzcvj4GC5+wlERCRH9O7d2+wIIpLNklOsfPj7Pr5bexSAphWLMvbJWhT0yIbb9W6n8VDYOh2OrICIjVAyj6yYvmUqYIMyTaFIObPTiEgmZaqnVLNmzThw4ACPPvooly5d4sKFC3Tq1Indu3fz/fffZ3XG/KvxEOOHCMBvQ2HHbHPzSK5R2teT+QMa8nhICaw2+PSPA/T94R8uXUk0O5qIiFzz/fffM3v2zT/bZ8+ezQ8//GBCIhHJSufjEug1eWNqQWpA83J83/u+nC1IARQKgppdje2VeaS3VErSvw3ONUtKJFez2GxZ10l7+/bt1KlTh5SUlKw6pd2KiYnBx8eH6OhovL2zceqtzQaLX4F/JoLFEbpMh8rtsu96kuf8/E8Eb/+yi4RkK4EF3Rnfow41ShQ0O5aISK6VVWOASpUq8c0339CiRYs0+1etWsVzzz3H/v377zWq6XJsvCRiZ3adjOb5aZs5eekqHi6OfPZ4TdpWDzAv0IWj8GUI2FLgmT+hRKh5WbLCnoXwc0/wLApD9oBTDhf6ROSu0jsGyNRMKclBFgu0/RhqPGn8EJndG46sNDuV5CJP3FeSeQMaElTEg5OXrtJ5/F9M33CcLKxHi4hIJhw/fpwyZcrctD8oKIjw8HATEolIVpi7+QSPjV/PyUtXKePryYKBjcwtSAEULpO3ZkulNjjvroKUSC6nolRu4OAAj3wNlR+GlAT4qZtxP7hIOlUr7sPCFxrTpqo/iSlW3lqwiyGztnElMdnsaCIi+Zafnx87duy4af/27dspUqSICYlE5F4kpVh5b+FuXp69nYRkKy0r+7FgYCMq+nuZHc3Q9GXjzotDYXBys9lpMu/CUTi83NgOecrcLCJyz1SUyi0cnaDzZCjbApIuw4zOcHqn2akkF/Fxd+bbniEMb1sZRwcLC7adouPX6zh8Ns7saCIi+dKTTz7JSy+9xIoVK0hJSSElJYXly5czaNAgnnzySbPjiUgGnI1NoPvEv5my/hgAL7WswMReofi4O5sb7EaFy0KNLsb2qo/NzXIvtlzruVfufuN7EpFcLUOr73Xq1OmOX7906dK9ZJG7cXKFJ2fAtE4QsQGmPQp9loBvebOTSS5hsVh4vlk5apUsyAs/beXAmTg6fLmWx0NL4urkgIODBScHCw4WC44O/z6u73Ny/M/Xbtjn5GD59/XXv/affTeex9Fy633Xz53ma9f23bh0uohIbjdy5EiOHz9Oy5YtcXIyhmRWq5VevXrxwQcfmJxORNJrW8Ql+k3bzOmYeAq4OvF5l1q0rupvdqxbazoMdsyEA0vg1FYoXtvsRBmTnGisJAhqcC6SR2So0XmfPun7i58fVuAztXHn1Uvww8PGTCnvEvD0EihYMmczSK4XFRvPSz9tZcORC2ZHSTcHC2kKYtcLVzfuc0xHwSu18GX5z+sdLDg6OOBowfjTIe31/t3ngIeLIx1rBVKqiIfZb4uI5LCsHgMcPHiQbdu24e7uTvXq1QkKCsqClPZBjc4lr5v1TzhvL9hNYoqVckU9mdArlHJFC5gd687mPQc7ZkGldtD1J7PTZMzu+UaP3QL+MGQ3ONrRTDQRSSO9Y4AsXX0vPzF9kHX5HHzfFs4dMKat9lkCXnb6iYzYreQUK7M3n+D4+SukWK2kWMFqs5F8bTvlxj9t15/bbrPvhofNRnKKDast7b6UlGt/XtuXbLVhtabdZ81F/yK5Ozvy6oOVeKpBaRwcNItLJL8wfQyQi+i9krwqMdnK+7/uZsbfxqIEbar689kTNfFyywVFknMH4eu6YLPC86shoKbZidLvhw5wdBU0GQYt3zY7jYjcQXrHABm6fU/siKcv9FwAkx+EC0eMW/l6/wYehc1OJrmIk6MDXeuWMjtGGjbbDQUr27+FqxsLWNcLXjd+7cbi143HJ9+mIJamWPbffbfbf8O+HScu8c+xi7z/6x4W74zk4841KePrafbbJyK5SOfOnQkNDeX1119Ps/+TTz5h48aNzJ49O0PnGzduHJ988gmRkZFUq1aNMWPG0KRJk1seu3LlSlq0aHHT/r1791K5cmUAmjdvzqpVq246pl27dixatChD2UTykjMx8QyYsYXNxy9iscDLrSsyoHn53PMBlW8FCH4Mds42eks9OcPsROlz/rBRkMKiBucieYiKUrmZTyD0WmDMmIraDTMeN5672skKHyKZYLl2u52To9lJ7sxqtTFjYzgfLt7LP8cu8uCY1QxrU4mnG5fBMbcMSkXEVKtWreLdd9+9af+DDz7Ip59+mqFzzZo1i8GDBzNu3DgaNWrEt99+S9u2bdmzZw+lSt3+w4f9+/en+fSyaNGiqdvz5s0jMTEx9fn58+epWbMmjz/+eIayieQlm45doP+MLZyNTcDbzYkvutamRSU/s2NlXNNXYOcc2PcbnN4FxYLNTnR3m6cYf5ZvBQXt60NVEck8rb6X2xUpZ8yYci8EJzfBT10hKd7sVCJ5noODhZ71g1g6pCmNy/uSkGzlf4v30vmb9RyK0oqGInJ3cXFxuLi43LTf2dmZmJiYDJ1r9OjR9O3bl2eeeYYqVaowZswYSpYsyfjx4+/4Oj8/P4oVK5b6cHT89xOBwoULp/laWFgYHh4eKkpJvmSz2Zi+4ThdJ27gbGwClfy9WPhC49xZkAIoWgmCry1iteojc7OkR3ICbLs2oytUDc5F8hLNlMoL/KtCj7nwwyNwbI3R/K/LNDX+E8kBJQp5MK1vXWb9E8HIRXvZGn6JdmPXMKRVRZ5tUgYnR9X+ReTWgoODmTVrFu+8806a/TNnzqRq1arpPk9iYiKbN2++6TbANm3asH79+ju+tnbt2sTHx1O1alXeeuutW97Sd92kSZN48skn8fS8/a3KCQkJJCQkpD7PaHFNst+KfVEs2HYSdZXNmAuXE1l76BwAD1UP4OPONfB0zeW/SjV9BXbNg70L4cxu8K9mdqLb2/srXDkPXsWhwgNmpxGRLJTL/yWVVIEh0G0mTH8MDvwO8/tBpwngYOf3QInkARaLhSfrlqJpxaIMn7eTVQfO8tGSfSzZZfSaqlRMt9SKyM3efvttHnvsMQ4fPsz9998PwJ9//smPP/7InDlz0n2ec+fOkZKSgr9/2gVP/P39OX369C1fExAQwIQJEwgJCSEhIYFp06bRsmVLVq5cSdOmTW86fuPGjezatYtJkybdMcuoUaN4//33051dclZsfBIvzdxKbHyy2VFyJQcLvPZgZZ5rWhaLJQ/cqu9XBao+AnsWGL2lnvjB7ES3d/3WvTo9wVG/workJfobnZeUbgxPTIOZXWHXHHDxhPZfQF74oSmSCxQv6M6UPvcxZ/MJRvy2h+0nonn4yzUMalmB55uVw1mzpkTkBh06dGDBggV88MEHzJkzB3d3d2rWrMny5csztVLdf39Jttlst/3FuVKlSlSqVCn1eYMGDYiIiODTTz+9ZVFq0qRJBAcHU7du3TtmGD58OEOHDk19HhMTQ8mSJTPybUg2mr4hnNj4ZEoX8aBXg9Jmx8l17itdmOolfMyOkbWavWoUpfb8AlF7jUKVvTl30LgbxOIAdXqZnUZEspiKUnlNxTbQaSLM7QtbfgA3b2j9fypMieQQi8XC46ElaVqxKG/M28mf+6L49I8D/L7rNJ90rknV4loSXUT+9dBDD/HQQw8BcOnSJWbMmMHgwYPZvn07KSkp6TqHr68vjo6ON82KioqKumn21J3Ur1+f6dOn37T/ypUrzJw5kxEjRtz1HK6urri6uqb7mpJz4pNSmLT2CAAvtaxApzolTE4kdsG/GlTpYNzCt/oT6DzZ7EQ3uz5LqkIb8NH/tyJ5jT62z4uCO0H7scb2+i9hdcZW8BGRe+fv7cZ3T4XyeZea+Lg7s/tUDB2+WsvnYQdITLaaHU9E7Mjy5cvp0aMHxYsX56uvvqJdu3Zs2rQp3a93cXEhJCSEsLCwNPvDwsJo2LBhus+zdetWAgICbtr/888/k5CQQI8ePdJ9LrE/s/6J4FxcIiUKudO+ZnGz44g9afaa8eeueXB2v7lZ/isp/t8G5yFqcC6SF2mmVF5VpyckxMLS4bBiJLgWgPr9zU4lkq9YLBYerV2CRuV9eXvBLpbuPsMXfx5k6e7TfPp4TYID89gtACKSbidOnGDKlClMnjyZy5cv88QTT5CUlMTcuXMz1OT8uqFDh9KzZ09CQ0Np0KABEyZMIDw8nH79+gHGbXUnT55k6tSpAIwZM4bSpUtTrVo1EhMTmT59OnPnzmXu3Lk3nXvSpEl07NiRIkWK3Ns3LaZJTLby7arDALqdXG5WLBgqPwz7fjNmSz32ndmJ/rV3IVy9CN4loEJrs9OISDYw/SfSuHHjKFOmDG5uboSEhLBmzZo7Hp+QkMCbb75JUFAQrq6ulCtXjsmT/51m2rx5cywWy02P61PjM3vdXKnBAGg+3Nhe8jpsvXlKvohkPz8vN77pEcKXXWtT2NOFfadjeeTrdXy6dD8Jyem7PUdE8o527dpRtWpV9uzZw5dffsmpU6f48ssv7+mcXbp0YcyYMYwYMYJatWqxevVqFi9eTFBQEACRkZGEh4enHp+YmMiwYcOoUaMGTZo0Ye3atSxatIhOnTqlOe+BAwdYu3Ytffv2vad8Yq5ftp3kVHQ8vgVceTxEtz/JLTR71fhz11yjh5O92PS98WedXlrASSSPsths5i0IO2vWLHr27Mm4ceNo1KgR3377Ld999x179uyhVKlSt3zNI488wpkzZxg5ciTly5cnKiqK5OTk1OnpFy5cIDExMfX48+fPU7NmTb777jt69+6d6ev+V0xMDD4+PkRHR2eqGWmOsdngj7fgr6+M5oCdv4dqHc1OJZJvnYtL4N1fdrNoZyQAFf0L8EnnmtQsWdDcYCKSbvc6BnBycuKll16if//+VKhQIXW/s7Mz27dvz9RMKXuVa8ZLeViK1Ubrz1dx5OxlhretzPPNypkdSezVT11h/2Ko0cVYxdtsUftgXD2wOMKQXeCt205FcpP0jgFMnSk1evRo+vbtyzPPPEOVKlUYM2YMJUuWZPz48bc8fsmSJaxatYrFixfTqlUrSpcuTd26ddP0SyhcuDDFihVLfYSFheHh4cHjjz+e6evmahYLtBlpfLpgs8LcZ+DgMrNTieRbvgVc+bp7HcZ3r4NvARcOnInj0XHr+PD3fcQnadaUSH6wZs0aYmNjCQ0NpV69enz11VecPXvW7FiSRy3dfZojZy/j7eZE9/pBZscRe3Z9ttTO2XDukLlZ4N8G5xUfVEFKJA8zrSiVmJjI5s2badOmTZr9bdq0Yf369bd8zcKFCwkNDeXjjz8mMDCQihUrMmzYMK5evXrb60yaNIknn3wST0/PTF8317NY4OExUK0TWJNgVg84ts7sVCL5WtvqAfwxpBmP1CqO1QbfrDrMQ2PXsPn4RbOjiUg2a9CgARMnTiQyMpLnn3+emTNnEhgYiNVqJSwsjNjYWLMjSh5hs9n4eoVRXOjdqAwFXNVOVu6geG2jAGSzwprPzM2SdBW2/2hsh6rBuUheZlpR6ty5c6SkpNy0VLG/v/9NSxpfd+TIEdauXcuuXbuYP38+Y8aMYc6cOQwcOPCWx2/cuJFdu3bxzDPP3NN1wehlFRMTk+aRqzg4GtNwKzwAyVfhxy5wcovZqUTytcKeLnzxZG0m9AyhqJcrh89epvM36xn52x6uJmrWlEhe5+HhwdNPP83atWvZuXMnL7/8Mh9++CF+fn506NDB7HiSB6w6cJbdp2LwcHGkT8PSZseR3OD6Snw7ZsGFI+bl2D0f4qPBpxSUu9+8HCKS7UxvdG6xWNI8t9lsN+27zmq1YrFYmDFjBnXr1qVdu3aMHj2aKVOm3HK21KRJkwgODqZu3br3dF2AUaNG4ePjk/ooWbJker49++LoDE/8AKWbQGIsTH/MuFdbREzVploxwoY0pVOdQGw2+G7tUdp+sZqNRy+YHU1EckilSpX4+OOPOXHiBD/99JPZcSSPuD5Lqnu9UhTydDE5jeQKgXWgQhuwpcBqE2dLXW9wHqIG5yJ5nWlFKV9fXxwdHW+anRQVFXXTLKbrAgICCAwMxMfn32XUq1Spgs1m48SJE2mOvXLlCjNnzkwzSyqz1wVjKeXo6OjUR0RERLq+T7vj7A5df4LAELh6AaY+AheOmp1KJN8r6OHC6CdqMbl3KMW83Th2/gpdJvzFewt3cyUx2ex4IpJDHB0d6dixIwsXLjQ7iuRyG49e4J9jF3FxdOCZJmXNjiO5yfXZUtt/Muf3hDO74cRGcHCC2j1z/voikqNMK0q5uLgQEhJCWFhYmv1hYWFpGpffqFGjRpw6dYq4uLjUfQcOHMDBwYESJdIub/vzzz+TkJBAjx497vm6AK6urnh7e6d55FquXtB9DvhVhbjTMLUDxJwyO5WIAPdX9mfpkKY8EVoCmw2mrD/Gg2PW8Nfh82ZHExGRXOT6LKnOoSXw93YzOY3kKiVCoVxLY7aUGb2lrs+SqtQOvIrl/PVFJEeZevve0KFD+e6775g8eTJ79+5lyJAhhIeH069fP8CYndSrV6/U47t160aRIkXo06cPe/bsYfXq1bzyyis8/fTTuLu7pzn3pEmT6NixI0WKFMnwdfMFj8LQcz4UKgOXwmFqR7h8zuxUIgL4uDvzceea/PB0XYr7uBF+4QpdJ27grQU7iUvQrCkREbmzXSejWXXgLA4W6Ne0nNlxJDdq/rrx5/af4OLxnLtu4mWjnxWowblIPmFqUapLly6MGTOGESNGUKtWLVavXs3ixYsJCjKWq42MjCQ8PDz1+AIFChAWFsalS5cIDQ2le/futG/fnrFjx6Y574EDB1i7di19+/bN1HXzDa9i0OsX8A6Ec/theiejoaCI2IVmFYuydEhTutUrBcD0DeE88Plq1h5UAVlERG5v3EpjllSHmsUpVcTD5DSSK5WsC2VbgDUZ1o7OuevumgcJMVCoNJRpnnPXFRHTWGw2m83sELlRTEwMPj4+REdH5+5b+QDOHYTJD8KVc1CqAfSYBy4awIjYk3WHzvHa3B2cuGgs6tC1bkmGt6uCt5uzyclE8p88NQbIZnqvct6hqDhaf74Kmw2WDm5KpWJeZkeS3Cp8A0x+AByc4aWtUDAHFnqaeD+c3Ayt3oPGQ7L/eiKSbdI7BjB99T2xA74VjFv5XH0g/C+Y1R2SE8xOJSI3aFTel6WDm9KrgTGj86eNETzw+WpW7o8yOZmIiNiTb1YdxmaD1lX9VZCSe1OqPpRpBtaknJktFbnDKEg5OEOtHnc/XkTyBBWlxBBQA7rPBmcPOLwc5vaFFPWuEbEnnq5OjHgkmJnP1SeoiAeR0fH0/v4fhs3eTvSVJLPjiYiIyU5cvMKCrScBGNiivMlpJE+4vhLflmkQfeLOx96rzdcanFd5GAoUzd5riYjdUFFK/lWqHjz5Izi6wN5fYeGLYLWanUrE/lhTjEacJqlftgi/D2rC043KYLHAnM0naP35KpbtOWNaJhERMd+E1UdIttpoXN6XWiULmh1H8oLSjaB0k2uzpT7PvuskxMGO2cZ2iBqci+QnTmYHEDtTrgV0/h5+7gXbfwRXL2j7EVgsZicTyV5WK1y9AHFREHcGLp/9z/YZiDsLl6OM5zYrPPgR1Ddn1U4PFyfeaV+VdtWL8eqcHRw5d5lnpm7i0dqBvNu+KgU9XEzJJSIi5oiKjWfmPxEADGihFfckCzV7DY6tgS1TocnL4F0866+xaw4kxkLhclCmadafX0TslopScrMqD0PH8TD/Odj4rVGYavm22alEMs5qhasXjULSjUWlW26fBVtKxs6/fCRU7wyevtmTPx1CSxdm8aAmjA47wHdrjjB/60nWHDzHyI7BPBhczLRcIiKSsyavPUZispXapQrSoGwRs+NIXlKmCQQ1guPrYO0YaPdx1l9j07Vb90J668NwkXxGRSm5tZpdjE8rFr0Maz4FN29oNMjsVCJgsxmFprioa0WlqNvPbrp81ljKOCPcC0MBf6OXgaff7bd/6gKR22H1p9D2w+z5XtPJzdmRN9pVoW1wMV6Zs4NDUXH0m76Zh2sE8H6HahQp4GpqPhERyV7RV5KYvuE4AAObl8eiX+olqzV7DaZ2gM1TjFXxvAOy7tyntkLkNqOFSK3uWXdeEckVVJSS27vvGUiIhWXvQdg7xoyp0KfNTiV50fVC0023zEXdUHy6YUaTNYNNvd0LGQUlz6JQwO/2255FwdE5feds9T5M6wj/fAf1nofCZTL8bWe12qUK8duLjfniz4NMWH2E33ZE8tfh84x4JJiHamTh4FFEROzK1L+OEZeQTOViXtxf2c/sOJIXlWkKJetDxAZY90XWfiB3fZZUlQ7gqVl+IvmNilJyZ42HQHyMsQzsb0PBxQtqPG52KskNbDaIv3SXW+auF53OQkpixs7vVvA/RaXbzGjyLApO2dBfqVwLKHe/sVrl8pHQeVLWXyMT3Jwdee3Bysasqdk72H8mloE/buG3HcUY8UgwRb00a0pEJC+5kpjM5HVHAejfvBwODpolJdnAYoHmr8G0R41V8hoPBq8saBMQHwM75xjboWpwLpIfqSgld9fyHWPG1D8TYf7z4OIBlR8yO5WYwWaDhJg73zJ34+ymDBeafG5xy5zfrWc0OdlBcaXV+3B4hdGcs+ELULy22YlS1ShRkIUvNuLr5YcYt/Iwv+86zYYj53mvQzU61CyuWztERPKInzZGcPFKEkFFPHioumbFSjYq2wJK1IUTG2HdWHjwg3s/587ZkHQZfCsafatEJN9RUUruzmKBth9DYhxs/wlm94bus6Fsc7OTSVaw2Yyi4397NN1qRlNcFKQkZOz8rj5GgSnNLXN+t57R5OyWPd9jdgmoATWegB2zIOxd6PWLXTXndHVyZGibSrSpZvSa2hsZw6CZ2/htRyT/6xiMn3cue79FRCSNhOQUJq4+AkD/ZuVwcnQwOZHkaddnS01/DDZNNmZLFbiH20VtNmPWFajBuUg+pqKUpI+DA3T4yihe7PsNfuoGvRZAybpmJ5PMSk4wmlWuHQOxpzL2WlfvO98yd+N2bis0ZVSLN2H3fDi6yriVr3xLsxPdJDjQh18GNmL8ysN8teIgYXvOsPHoBd55uCqd6gRq1pSISC41b8tJTsfEU8zbjUfrBJodR/KDci0hMBROboL1Y6HNyMyf6+QWOL0THF2hZtesyygiuYqKUpJ+jk7QeTL82AWOrIAZnaH3IihW3exkkhEpSbBtBqz6BGJO/LvfxesWM5pus+3sbl5+e1MoCO57FjZ8DcveNaa2O9jfJ9UuTg4MalWBB4L9GTZ7O7tOxvDy7O38tuMUozrVoJhPHi8eiojkMckpVr5ZdRiAZ5uWxdXJ0eREki9YLND8deP3gH8mQcNBxvgxMzZPNv6s1hE8CmdZRBHJXVSUkoxxcoUnZ8C0TsbqG9MehT5LwLe82cnkbqwpRiPJlaPgotEQFa/i0OwVqP4EuBYwN19u1nQYbJ1ufNq3czbU7GJ2otuqXMyb+QMaMWH1Eb5YdpAV+8/S+vNVvP1QVR4PLaFZUyIiucSinZEcP3+FQh7OdK1b0uw4kp+Ub2X00Ty1Ff76ElqPyPg5rl6CnXON7RA1OBfJz+zv43yxfy6e0G0WFKthNLae+ghcijA7ldyO1Qp7foHxDWH+c0ZBysMXHhgFL22F0KdVkLpXHoWNvgpgrMSXFG9qnLtxdnRgYIvy/PZSY2qWLEhsfDKvzt3BU9//w8lLV82OJyIid2G12hi3wpgl9XSjMni46HNmyUEWCzR73dje+B1cPp/xc+z4GZKvQtHKUKp+1uYTkVxFRSnJHPeC0HO+sVJGzAmY2gFiz5idSm5ks8GBP2BCM/i5F5zdZ6xu1/IdGLQdGgzI+/2eclL9/sbMs+hw+Oc7s9OkS0V/L+b2a8DwtpVxcXJg9YGzPPD5amb8fRybzWZ2PMG4PedQVBwXLmdwJUsRydOW74ti/5lYCrg60atBabPjSH5U8QEIqGWsnPfXVxl7bZoG533U4Fwkn7PY9JtHpsTExODj40N0dDTe3t5mxzFP9EmY/KDxi7hfNej9m+4JtwdHVxszdiL+Np67FID6A6DBQKOgKNljy1RY+CK4F4KXtuWq9/pQVByvztnOlvBLADQqX4QPO9WgZGEPc4PlIwnJKRw4HceuU9HsPhXNrpMx7DsdQ3ySFRdHBx4LCeT5puUo7etpdtR8T2OA9NN7lfVsNhuPjlvPtohL9GtWjtfbVjY7kuRX+xbDzK7GOHPwzvT/DhD+N0xuA05u8PI+Y9wkInlOescAKkplkgZZN7hwBCa3hbjTxmocvRaAq5fZqfKniI2w/P+MohQYP+zrPguNhoBnEXOz5QcpyfBNI2NWWuMh0Oo9sxNlSIrVxvfrjvLpH/uJT7Li4eLI620r06NeEA4O+hQzK11OSGZvZAy7T8Ww62Q0u07FcPBMLMnWm38kezpZuZxsASw4WKBt9QAGNC9HteI+OR9cAI0BMkLvVdZbf/gc3Sb+jauTA2tfu5+iXq5mR5L8ymaDb5vC6R3QZBi0fDt9r5vfD7b/BDW7waPjszejiJhGRalspkHWf0Tthe/bwtWLULoJdJ+jW8NyUuR2WP4/OLjUeO7gDCG9jQbcXsVMjZbv7P8dfnrSKAi+uAV8ct8S3cfOXebVOTvYeOwCAPXKFObjzjUIKqIZOpkRfSXJmPl0bfbT7lPRHDl3mVv99C3o4UxwcR+qBXpTrbgPNQomErSgI1dtLrxT4B3mHPq3ONisYlEGNC9H3TKF1aA+h2kMkH56r7Jej+/+Zu2hczzVIIj3Hwk2O47kd3t/g1ndjVWcB++4+2ypqxfhs8qQHA99l0HJ+3Imp4jkOBWlspkGWbdwcjP88AgkxkLFttBlGjg6m50qbzu7H1b8z2hkDmBxhFrdoNmrULCUudnyK5sNvm8H4euhdg945GuzE2WK1Wpj2objfPj7Pq4mpeDm7MCrD1Smd8PSmjV1B1Gx8ew+acx+2n0qhl2nojlx8dbN4/29XY0CVHFvqgX6EBzoQ3Eft7QFpgUDYdt0Y7tgKQ61m8nYzYn8tuMU1ydVhQQVYkDzctxf2U/FqRyiMUD66b3KWtsiLtHx63U4OVhY+UpzShTSLdZiMqsVvm0CZ3ZB01fh/jfvfPyG8bDkdfAPhn5r1U9KJA9TUSqbaZB1G8fWwvTHjE8/gjtDpwng4Gh2qrznwhFY+RHs/BlsVsAC1TsbK6H4ljc7nUT8A5NagcUB+q8HvypmJ8q08PNXeG3uDv46YqysExpUiI8716Bs0fy9YqPNZuPExavsPhWd5ha8s7EJtzy+VGEPqhX3JjjwWhGquM/db7m53nMDwDsQYk6CT0l46leO2/z4dvUR5mw6QWKKFYBK/l70b16Oh2sE4OSodUyyk8YA6af3Kms9O3UTYXvO0DmkBJ8+XtPsOCKGPb8Yi+q4ehuzpW7XI8pmg6/rwbn90O5To8WEiORZKkplMw2y7uDAH0bTQ2sy1HkK2n+hT0GySvQJWP0JbJ1uvL8AlR+GFm+AfzVzs0las3rA3l+h4oPQbZbZae6J1Wrjx43hjFq8l8uJKbg6OfBym4r0bVwWx3wwayrFauPouctpClC7T8UQfTXppmMdLFCuaIEbClA+VC3ujY97BmeNpiTDxOZweifU7mn8HZ/yMFw4DN4l4KmFUKQcUTHxTFp3lBkbwolLMP5NKFnYneealuPxkBK4OetDgeygMUD66b3KOvtPx/LAmNVYLBA2pBnl/fL3hwNiR6xWo6dm1B5oPhyav37r446vN9p9OHsYDc7d1BtRJC9TUSqbaZB1F7vnw5ynjVk8DV+E1v+nwtS9iIuCNaNh0yRIubY0fPlW0OJNCKxjbja5tXMHjU8DbSnQezGUbmR2ont24uIVhs/byZqD5wCoVbIgnz5eg/J+eWdhg8RkKwejYtl9Kobd12Y/7Y2M4Upiyk3HOjtaqOjvRXBxH4IDvala3IcqAV54uDjde5C/J8Dvr4BbQXhxM3j6Quxp+KE9nDsAXsXhqV9TZ0ZGX01i2l/HmLzuGBcuG/9G+BZwpW/jMvSoXwovN91KnZU0Bkg/vVdZZ/DMrSzYdop21YsxrnuI2XFE0to9H2b3NgpNg3feuuA091ljln/tnvDIVzkeUURylopS2UyDrHTYMg0WvmBst3gLmr1ibp7c6MoFWPcFbJwASVeMfUGN4f63IKiBudnk7n4bApsmQ4n7oG9YnijM2mw2Zv0Twf8W7SU2IRkXJwcGt6rAc03K5rpbxq4mprDvdAy7rhWgdp+KYf/p2NTb4W7k5uxA1YC0t99V9PfCxSkbvue4KPgyFBKi4aHRcF/ff78WewamdjBWeCxQzChMFa2Y5nua9U84E9cc5eQlo5eVl5sTPesH8XTjMvgW0CpdWUFjgPTTe5U1ws9fofmnK7Da4LcXGxMcqBkmYmesVhjfEM7uNT40bfZq2q9fuWA0OE9JgGeXQ6AKqyJ5nYpS2UyDrHT6axwsHW5sP/gh1O9vbp7cIj4GNoyDv76GhBhjX2AI3P82lG2eJ4ob+ULsGRhbG5IuwxNToeojZifKMpHRVxk+bycr958FoEYJHz7pXJNKxexz1lRMfBJ7rt16t+daA/JDUXGpzcJv5OXmlDr7qdq1P8v4Fsi5WxXn94ftP0JALWPg/t++fHFnYeojELUbPP2MwpRf5TSHJKVYWbjtFONXHeZQVBwArk4OdLmvJM82KUvJwmqOfC80Bkg/vVdZ4435O/nx73CaVSzKD0/XNTuOyK3tmmvcKeFW8NpsqRv+zq//Cv54E4rVgOdXaywrkg+oKJXNNMjKgJUfwcoPjO1HvjZWJJNbS7xizIpaN8ZYMheM1Unuf8voTaQf4LnPig9g1UdQuBwM/DtPrUhps9mYu+UkI37dTUx8Ms6OFl68vwL9m5fD2cRZU+fjEozZT6eijZXwTkVz/PyVWx7rW8AldfaTUYjyoUQhd/NWsQvfAJMfACzwzJ9Q4jafJF8+bxSmzuwED1+jMOVf9abDrFYbYXvPMG7lYbZHXALA0cHCIzWL0695OSr622cR0d5pDJB+eq/u3ZmYeJp8tILEFCuz+zXgvtKFzY4kcmvWFBhX37jN/P63oOm1uyRsNvjqPjh/EB7+HEKfNjeniOQIFaWymQZZGWCzwR9vwV9fGauRdZ4M1R41O5V9SU6AzVNg9adwOcrYV6SC0dy4akdwyF23RckNEmLhi1pw5Rw89Bnc94zZibLcmZh43py/k2V7jf93qxX35pPONalaPHv/bbTZbERGx9/QfNy4BS8yOv6WxwcWdE9tQH59FpSfl6t5Baj/SkmGCc2NQlOdXtDhyzsff+UCTOsIkdvBvbDR/LxY9VsearPZ+OvwecatPMzaQ+dS97eu6k//5uWoU+o2KyXJLWkMkH56r+7dyN/28N3ao9QtXZif++nWfbFzO2bDvGeMFfgG7wRXLzi6Bn54GFwKGA3OXfWBiEh+oKJUNtMgK4NsNvj1JdgyFRycoetMqNDK7FTmS0mCbT/Cqo8h5oSxr2ApY+WS6k+AYxY0TBbzXW9a7ekHL20F17y3YpLNZuOXbad479fdXLqShJODhQEtyvNCi/JZ0nfJarVx/MIVdp+KZtfJmNQC1PWm3jeyWKBMEU+qBfoQfK3/U7Xi3hTydLnnHNnq72/h91evNTffAp5F7v6aqxdh2qNwaqvxC0DPBVC81h1fsuPEJcavPMyS3ae5PgKoX7YwA5qXp0kFX/sp0tkxex0DjBs3jk8++YTIyEiqVavGmDFjaNKkyS2PXblyJS1atLhp/969e6lc+d/bQS9dusSbb77JvHnzuHjxImXKlOGzzz6jXbt26cpkr+9VbnHxciINP1zO1aQUpvS5j+aV/MyOJHJn1hRjoZfzB6Hlu9BkqHFL3665ENLbWJVbRPKF9I4B9Buv5AyLBR4eAwlxsHsezOoBPebmiRXJMsWaYvxwXjkKLhwx9nkFGNOca/cEJzv/5VkyJqS30SPs4lGjT1jz18xOlOUsFgsdawfSsHwR3lmwmyW7TzP2z4P8sfs0n3SuSfUS6W/Km5xi5fDZy+w6Gc2ua8WnPadiiEtIvulYJwcL5f0KGLOfintTLdCHKgHeFHDNZT/e4qJg+Uhju9W76StIwb+FqOmPwclNRhP0ngvuuCpnjRIFGd8jhENRcXy76jDzt55kw5ELbDiykeqBPvRvXo4HqhXLuR5akiVmzZrF4MGDGTduHI0aNeLbb7+lbdu27Nmzh1KlSt32dfv3708zUCxatGjqdmJiIq1bt8bPz485c+ZQokQJIiIi8PLSLIec8v36Y1xNSqFacW+aVSx69xeImM3B0RjPzn8O1n8JwZ1gz0LjayF9zM0mInZJM6UySZ/8ZVJKEszsDgeXgouXcbvJHX55ynNsNtj7q9Fn6OxeY5+HLzR5GUL7gLO7ufkk++yaB3P6GFPXX9oGBfLuLxc2m41FOyN555fdXLiciKODhX7NyvJSywq4OqVt2h2flMKBM7Gps592nYphX2QMCck3r4Dn6uRA5QDv1NlPwYHeVPT3ws3Z8aZjc535/WD7T1C8ttFL6r/Nze8mPgZmdIaIv8HVB3rOgxKh6XrpqUtXmbjmCDM3RnA1KQWAsr6e9GtWjo61A7NnhcFczh7HAPXq1aNOnTqMHz8+dV+VKlXo2LEjo0aNuun46zOlLl68SMGCBW95zm+++YZPPvmEffv24eycuX549vhe5RZxCck0HPUnMfHJjOteh3bVA8yOJJI+KcnwdV24cNhoR3H+oPHz7bmVZicTkRyk2/eymQZZ9yDpKsx4HI6tMfqg9Pn9ppWj8hybDQ6GwYqRRv8XADcfaDQI6j6fJ2/nkv+w2WDi/XBqC9z3LDz0qdmJst35uATeWbibRTsiAajgV4ChrStyJiaeXdf6QB2KiiP5FkvgFXB1ompx7zQNyMsV9cTJxAbq2eb4X/D9g4AFnv0z88tkJ8Qa/7aG/2UU/XvMhVL10v3yC5cTmbL+GD+sP0b01SQAinm78UyTMnStWwrP3Db7LBvZ2xggMTERDw8PZs+ezaOP/tuzcdCgQWzbto1Vq1bd9JrrRanSpUsTHx9P1apVeeutt9Lc0teuXTsKFy6Mh4cHv/zyC0WLFqVbt2689tprODreunCakJBAQkJC6vOYmBhKlixpN+9VbvLtqsOM+n0fZYt6EjakmWYvSu6y7SdY0O/f5+3HQshT5uURkRyn2/fEfjm7Q9efjJWjTm42/nx6CRQuY3ay7HF0tXFbTsTfxnOXAlB/ADQYCO4FTY0mOchigdYjjEafm7+H+v2hSDmzU2WrIgVc+bpbHdrXiOStBbs4GBVH/xlbbjqukIfztRXw/m1AHlTYA4f88AtYSjIsHmZshzyV+YIUGI1ju8+BH7vA8bUwvZPxPCh9jZELe7owtHVFnmtalp/+Due7tUc4HRPPyEV7+WrFIXo3LE3vhqUp6KHbi+3NuXPnSElJwd/fP81+f39/Tp8+fcvXBAQEMGHCBEJCQkhISGDatGm0bNmSlStX0rRpUwCOHDnC8uXL6d69O4sXL+bgwYMMHDiQ5ORk3nnnnVued9SoUbz//vtZ+w3mQ/FJKUxccxSA/s3KqSAluU/1x43Vhy8eNT4oCX7M7EQiYqc0UyqT7O1T0lzpygWY8hBE7TGaez+9FLyLm50q60T8A8v/D45e+4TayQ3qPguNhqS/X4zkPdM7w6EwYwXKx6eYnSbHXLycyAeL9/LPsQuU9yuQ2nw8ONCHAB+3/Ntce8M3sOQ1ozfUi1vAIwuWek+8DD89aRTEnT2h+89QunGGT5OQnMK8LSf5dtVhjp2/AoCHiyPd6pbimSZlKebjdu9Zcyl7GwOcOnWKwMBA1q9fT4MG/xYh//e//zFt2jT27duXrvO0b98ei8XCwoVG/5eKFSsSHx/P0aNHU2dGjR49OrWZ+q1oplTWmLbhOG8v2EVgQXdWvtIc57w4S1TyvuutCxoPNfoliki+oplSYv88CkPP+fB9W6PZ99SO0GcxePqanezeRO6AFf+DA0uM5w7ORqPrpsPAq5ip0cQOtHoPDi2D3fOh4Yv3NjMmFynk6cInj9c0O4Z9iT1j/FsBxgpFWVGQAnDxhK6zYGY3OLLCKIR2mwVlm2XoNK5OjnStW4onQkvy+65Ixq04zJ7IGL5be5Qf/jpGp9oleL5ZWcoW1e3HZvP19cXR0fGmWVFRUVE3zZ66k/r16zN9+vTU5wEBATg7O6e5Va9KlSqcPn2axMREXFxunjXn6uqKq6trJr4LuS4pxcq3qw4D8HyzsipISe4V3Mn4UMQjl4/tRSRb6aecmMurGPT6BbwD4dx+Y2nz+GizU2XO2f3w81PwbROjIGVxhNo94MXNRv8gFaQEoFgw1OxqbIe9a/Sakvwp7B1IiIHidaBOr6w9t4sHdJ0J5VtB8lX48Qk4vDxTp3J0sPBwjeIseqkxU/rcR90yhUlKsTFrUwQtR69i4Iwt7DqZS//dziNcXFwICQkhLCwszf6wsDAaNmyY7vNs3bqVgIB/m2k3atSIQ4cOYbX+u/DAgQMHCAgIuGVBSrLGwm2nOHHxKr4FXHgitKTZcUTuTQE/cNCvnCJye/oXQsxXsJRRmPLwhdM7jH4oiVfMTpV+F47CvOdhXH3YswCwQHBnGLgRHvkaCgWZnVDsTYs3wNHVaPZ/MOzux0vec3w97JgJWIyidUZX20sPZzd48keo8AAkx8OPT8LBZZk+ncVioXklP35+vgFz+zegZWU/bDZYtDOSh79cS6/JG/nr8HnUFcAcQ4cO5bvvvmPy5Mns3buXIUOGEB4eTr9+RqPh4cOH06vXv8XPMWPGsGDBAg4ePMju3bsZPnw4c+fO5YUXXkg9pn///pw/f55BgwZx4MABFi1axAcffMDAgQNz/PvLL6xWG+NWHgKgb+OyeWN1URERkTvQ7XtiH3wrGLfyTXnYWDlqVnfjU34nO74FIPokrP4Ytk4Ha7Kxr/LDRsHBv5q52cS+FSwJ9Z6H9WNh2btQvmX2FCXEPqUkw6Lrzc17Z+8tnE6u0GUazO4D+xfBzK7QZQZUbHNPpw0JKsyk3oXZdzqGb1Ye5tcdkaw+cJbVB85Su1RBBjQvT8vKfvmjWb2d6NKlC+fPn2fEiBFERkYSHBzM4sWLCQoyPhiJjIwkPDw89fjExESGDRvGyZMncXd3p1q1aixatIh27dqlHlOyZEn++OMPhgwZQo0aNQgMDGTQoEG89tprOf795Rd/7DnN4bOX8XJzokf9UmbHERERyXZqdJ5J9tbkNM8I/xumdYSkK1ClPXSeAo52VjuNi4I1o2HTZEi51sy1fCto8SYE1jE3m+QeVy7A2FrG7aodx0OtbmYnkpyyYTwseT1rm5vfTXIizH0a9v5q9Ll7YipUbnf316VT+PkrTFhzmJ83nSAx2bjVq6J/Afo3L0f7GsVxymM9cTQGSD+9V+lns9no8NU6dp6M5sX7y/Nym0pmRxIREcm09I4B8tYoUXK/UvWM200cXYxfnha+CDf0sjDVlQuw7D34oib8Pd4oSAU1gj6/Q4+5KkhJxngUhiYvG9vL/wdJ8ebmkZwRexpWfGBst3ovZwpSAE4u0Pl7qNoRrEnwc0/j39gsUqqIByM7Vmftay3o16wcBVydOHAmjiGzttP805VM++sY8UkpWXY9kbxozcFz7DwZjbuzI30alTE7joiISI5QUUrsT7kW8PgUo1H49h+NGQVmTuiLj4GVHxnFqLWfG7O4AkOg5wLovQiC0t9EViSNus+DdwmIOQEbvzU7jeSE683NA0OgdhY3N78bR2d4bBIEP2bccvzzU8YqkFnIz8uN19tWZt3r9/PKA5Uo4unCiYtXefuX3TT+aDlfrzhETHxSll5TJK/4eoXRS6pr3VIU9lQjeRERyR9UlBL7VPkh45YmMH5ZXz4y5zMkXoF1XxjFqJUfGL9I+gcbva6e+dMonlnUL0XugbMb3P+msb3mM2M2nuRdx9bBjlmABdp9as5qRI5O8OgEqNEFbCkwpy/snJPll/Fxd2Zgi/Kse/1+RjxSjcCC7pyLS+STpftpNGo5Hy3Zx9nYhCy/rkhutenYBf4+egFnRwvPNtUsKRERyT9UlBL7VbMLPPSZsb3mU1g7Jmeum5wAf08w+v2EvQNXL0CRCsatL8+vgUptVYySrFOjC/hVM3pLrR1tdhrJLilJsPhac/PQPube7uvoZBT9a3YzClPznoUdP2fLpdycHenVoDQrX2nO6CdqUsGvALEJyYxfeZjGHy3n7QW7iLiQi1ZbFckm41YeBqBzSAkCfNxNTiMiIpJzVJQS+3bfM0bfFTBWKftnUvZdKyUJtkyFL0Pg91cg7gwULGX88jZgAwR3Mmdmg+RtDo7//j/+9wS4FGFqHMkmGydC1B5wLwz3v212GuP/u0e+hto9wWaFec/Bth+z7XLOjg50qlOCpYObMrFXKLVLFSQh2cq0Dcdp/ulKhszaxv7Tsdl2fRF7tvtUNMv3ReFggeebljM7joiISI6ys2XNRG6h8RCjr9Pa0bDoZXD1hhqPZ935rSmway6sHAUXjhj7vAKg6SvGL2xO6usg2axCayjdBI6tMZpgPzre7ESSlcxqbn43Dg7QfqxRoNo8BRYMMP49rNMzGy9poXVVf1pV8WPDkQuMW3mINQfPMX/rSeZvPUmrKn70b16ekKBC2ZZBxN5cnyX1cI3ilPb1NDmNiIhIzlJRSnKHlu9AQiz8MxHmPw8uHkbfqXthsxmrT634AM7uNfZ5+EKToRD6NDhr+rzkEIsFWr8PE++H7T9Bg4FQLNjsVJJV/ngbEmOvNTfPvoJPpjg4wEOfg4MT/PMdLHzBaIIe2idbL2uxWGhQrggNyhVh54loxq86xO+7TrNsbxTL9kZRr0xhBrQoT9MKvlh0u7TkYUfOxrF4ZyQA/ZtrlpSIiOQ/uhdJcgeLBdp+DDW7Gj1QZveGIyszdy6bDQ6GwYRmxrLoZ/eCm49xS82g7UZBQAUpyWmBIVDtUcAGy94zO41klWNrYefPgMXokWePtwA7OBiN1+v1N57/NtgoUOWQ6iV8GNc9hGVDm9EltCTOjhb+PnqBpyZv5OEv1/LbjlOkWE1cgVUkG32z6jA2G7Sq4keVAG+z44iIiOQ4Oxwdi9yGgwN0+AoqPwwpifBTN4jYmLFzHF0Dkx+EGZ0hcju4FDBu0xu0A5oOA9cC2ZNdJD3uf9uYsXIoDI6uNjuN3KuUJFh0vbn501C8trl57sRigQdHQYMXjOeLXoa/v83RCOWKFuCjzjVY/WoL+jYug7uzI7tPxfDCj1tpNXoVMzeGk5CckqOZRLLTyUtXmbflJAADWpQ3OY2IiIg5VJSS3MXRCTpPhnL3Q9Jlo7h0eufdXxfxD0x9BH54GCI2gJMbNHzRmBl1/1vgXjDbo4vcVZFyEHLttqmwd8BqNTeP3JuNE4yZmB5FjH9n7J3FAm1GQqNBxvPfX4W/vs7xGAE+7rz9cFXWv34/g1pWwMfdmaPnLvP6vJ00/XgF3605wuWE5BzPJZLVJq4+QrLVRoOyRahTSn3UREQkf1JRSnIfJ1foMh1K1of4aJj2KJw7dOtjI3fAj11gUivjdj8HZ7jvWXhpm/HLl6dvTiYXubtmrxkz+E5thT0LzE4jmRUTCStGGdv21Nz8biwWaPU+NHnZeL70DVj3hSlRCnm6MKR1Rda/fj9vPVSFYt5unIlJYOSivTT8cDmjww5w8XKiKdlE7tW5uARm/hMOwEDNkhIRkXxMRSnJnVw8odssKFYDLp81ZkFdivj362f3w89PwbdN4MASsDhC7R7w4mZ46FPwDjAvu8idFCgKDV8ytv8cAcn6pTtXCrve3DwUavUwO03GWCzGraTNXjOeh70Daz4zLY6nqxPPNCnLqleb89Fj1Snj60n01STG/nmQhh8uZ8Sve4iMvmpaPpHM+H7dUeKTrNQsWZBG5YuYHUdERMQ0KkpJ7uVeEHrOB9+KEHMCpnYwekzN7wfj6l+bZWKB4M4wcCM88jUUCjI5tEg6NBgInn5w8ShsnmJ2Gsmoo2tg52zsurn53Vgs0OINaP6G8fzPEbDqY1MjuTo50uW+Uiwb2oyvu9WhWnFvrialMHndUZp+vIJX52zn8Nk4UzOKpEf01SSmrj8OwMDm5bTCpIiI5Gu5cKQscgNPX+i5AHxKwYUjMKk1bP8JbFajIXr/ddB5EvhqarzkIq4FoPnrxvaqjyA+xtw8kn4pSbD4WnPz+/pC8VqmxrlnzV8zZk0BrPgfrPjAWMHURI4OFh6qEcBvLzbmh6frUq9MYZJSbPy86QStRq9iwIzN7DwRbWpGkTuZvuE4sQnJVPQvQKsq/mbHERERMZWKUpL7+QTCU79AgWLG83It4dnl8OQM8K9mbjaRzKrTC4qUhyvnYP2XZqeR9Pr7Gzi7L/c0N0+PpsOMPlNgFEmXjzS9MAVgsVhoVrEos55vwNz+DWlVxR+bDRbvPE37r9bSc9LfrD98DpsdZBW57mpiCpPWHgVgQPPyODholpSIiORvKkpJ3lC4LAz4CwZsgJ7zIDDE7EQi98bRGVq+Y2z/9RXEnjE3j9xdTCSs/NDYbvU+uOeh1bQaD4YHPjC213wKy96zi8LUdSFBhfjuqVCWDm7Ko7UDcXSwsObgOT5dut/saCJpzPwnnAuXEylZ2J2Ha6i/pYiIiOlFqXHjxlGmTBnc3NwICQlhzZo1dzw+ISGBN998k6CgIFxdXSlXrhyTJ09Oc8ylS5cYOHAgAQEBuLm5UaVKFRYvXpz69ffeew+LxZLmUaxYsWz5/iQHeRQGvypmpxDJOlU6QIn7IOkKrPrQ7DRyN3+8BYlxxn+zWt3NTpP1GgyEttf6Sq0bY3y/dlSYAqhUzIvPu9Ri5bDm9GoQxIstK6hfj9iNxGQrE1YfAaBfs3I4OZo+DBcRETGdk5kXnzVrFoMHD2bcuHE0atSIb7/9lrZt27Jnzx5KlSp1y9c88cQTnDlzhkmTJlG+fHmioqJITk5O/XpiYiKtW7fGz8+POXPmUKJECSIiIvDy8kpznmrVqrFs2bLU546OjtnzTYqIZJbFAq1HwPdtYfMPUH8A+FYwO5XcytHVsGsOWByg3ae5s7l5etR73vgeFw8zZvBZU+DBUcb/q3akZGEPRjwSbHYMkTQWbD1JZHQ8fl6uPFanhNlxRERE7IKpRanRo0fTt29fnnnmGQDGjBnD0qVLGT9+PKNGjbrp+CVLlrBq1SqOHDlC4cKFAShdunSaYyZPnsyFCxdYv349zs7OAAQF3bzimpOTk2ZHiYj9C2oIFdvCgd/hz/ehy3SzE8l/pSTBomvNzUPzQHPzu6n7LDg4wW+D4e/xYE2Gdp/YXWFKxJ6kWG2MX3UYgGeblMXNWR+GioiIgIm37yUmJrJ582batGmTZn+bNm1Yv379LV+zcOFCQkND+fjjjwkMDKRixYoMGzaMq1evpjmmQYMGDBw4EH9/f4KDg/nggw9ISUlJc66DBw9SvHhxypQpw5NPPsmRI0fumDchIYGYmJg0DxGRHNHqXWN2yt5fIWKj2WnkvzaMh3P7wcMX7n/T7DQ5I7QPdPgKsMA/E2HRULBazU4lYrd+3xXJ0XOXKejhTLd6t74bQEREJD8yrSh17tw5UlJS8PdPuxSuv78/p0+fvuVrjhw5wtq1a9m1axfz589nzJgxzJkzh4EDB6Y5Zs6cOaSkpLB48WLeeustPvvsM/73v/+lHlOvXj2mTp3K0qVLmThxIqdPn6Zhw4acP3/+tnlHjRqFj49P6qNkyZL3+A6IiKSTXxWo1c3YDnvX7vr45Gsxp4wV6QBa57Hm5ndTpyd0HAdYYNNk+G2QClMit2Cz2fh6hTFLqk/DMni6mnqjgoiIiF0xvenFfxuQ2my22zYltVqtWCwWZsyYQd26dWnXrh2jR49mypQpqbOlrFYrfn5+TJgwgZCQEJ588knefPNNxo8fn3qetm3b8thjj1G9enVatWrFokWLAPjhhx9um3P48OFER0enPiIiIu71WxcRSb/mb4CTG4SvhwNLzE4j16U2N68LNbuZnSbn1eoGj35rzOTbMhUWvmD0mRKRVCv2R7E3MgZPF0eeanhzSwkREZH8zLSilK+vL46OjjfNioqKirpp9tR1AQEBBAYG4uPjk7qvSpUq2Gw2Tpw4kXpMxYoV0zQur1KlCqdPnyYxMfGW5/X09KR69eocPHjwtnldXV3x9vZO8xARyTE+gVC/v7G97D1ISb7j4ZIDjqyCXXONgsxDebi5+d3U7AKdJoLFEbbNgAX9VZgSucZms/HV8kMA9KgfREEPF5MTiYiI2BfTRtAuLi6EhIQQFhaWZn9YWBgNGza85WsaNWrEqVOniIuLS9134MABHBwcKFGiROoxhw4dwnrDLQQHDhwgICAAF5dbDwQSEhLYu3cvAQEB9/ptiYhkn0aDjdvDzu6D7T+anSZ/S06Exa8Y2/c9AwE1zc1jtuqdofMkozC1YxbMe06FUxHg76MX2BJ+CRcnB/o2LmN2HBEREbtj6se6Q4cO5bvvvmPy5Mns3buXIUOGEB4eTr9+/QDjlrlevXqlHt+tWzeKFClCnz592LNnD6tXr+aVV17h6aefxt3dHYD+/ftz/vx5Bg0axIEDB1i0aBEffPBBmr5Tw4YNY9WqVRw9epS///6bzp07ExMTw1NPPZWzb4CISEa4F4Sm1wohKz6AxCumxsnX/r6huXmLfNLc/G6qPQqPTzFW5ts1B+b2NVYmFMnHvl5hzJJ6IrQEft5uJqcRERGxP6YWpbp06cKYMWMYMWIEtWrVYvXq1SxevJigION++8jISMLDw1OPL1CgAGFhYVy6dInQ0FC6d+9O+/btGTt2bOoxJUuW5I8//uCff/6hRo0avPTSSwwaNIjXX3899ZgTJ07QtWtXKlWqRKdOnXBxcWHDhg2p1xURsVv3PQM+pSA20iiMSM6LPgkrrzU3b/N/RrFQDFU7wBNTwcEZ9iyAOX2MWWUi+dCOE5dYc/Acjg4Wnm9azuw4IiIidslis2kZp8yIiYnBx8eH6Oho9ZcSkZy1fRbMfw5cvWHQdvAobHai/GV2H9g9D0rWgz5L8m8vqTs5sBRm9YCURKj0kDGDyinv9NLRGCD98vN71W/aZpbsPk2n2oGM7lLL7DgiIiI5Kr1jAI2kRURym+qPQ7HqkBADqz81O03+cmSlUZCyOEC7fNzc/G4qPgBP/gSOrrB/EfzcE5ITzE4lkmMOnollye7TWCwwoIVmSYmIiNyORtMiIrmNgwO0et/Y/mciXDxubp78Ik1z82choIa5eexdhVbQbSY4ucGBJTCzOyTFm51KJEeMX3kYgAeqFqO8n5fJaUREROyXilIiIrlR+ZZQtrlxe9TykWanyR82jINzB8CzKLR4w+w0uUO5+6Hbz+DkDofCYGZXSLpqdiqRbBVx4Qq/bD8FaJaUiIjI3agoJSKSW7V6z/hz588Qud3UKHle9ElY9bGx3VrNzTOkbDPoMQecPeHwcvixi1aOlDzt29WHSbHaaFLBlxolCpodR0RExK6pKCUiklsVrw3BnY3tZe+ZGiXP++NNSLoMJetDzSfNTpP7lG5sFKZcCsDRVfDjE5AQZ3YqkSwXFRPPz5tOADCwRXmT04iIiNg/FaVERHKzlm+Dg7MxA+XwCrPT5E2HV8Du+UZz84c+BYvF7ES5U1BD6DEPXLzg2BqY8TgkxJqdSiRLTVp7lMRkKyFBhahXRiujioiI3I2KUiIiuVmh0nDfM8Z22DtgtZoaJ8+5sbl53eeMVQ8l80rVg14LwNUbwtfD9McgPsbsVCJZ4tKVRKZvMBaeGNiiHBYVsEVERO5KRSkRkdyu6SvG7JPTO2DXXLPT5C0bvobzB8HTD5oPNztN3lAi1ChMuflAxN8w7VGIjzY7lcg9+2H9cS4nplAlwJsWlfzMjiMiIpIrqCglIpLbeRaBxoOM7eUjIDnB3Dx5RfSJf5ubt1Fz8ywVGAK9FoJ7ITi5CaZ2hKsXzU6V640bN44yZcrg5uZGSEgIa9asue2xK1euxGKx3PTYt29f6jFTpky55THx8fE58e3kKpcTkvl+/VFAs6REREQyQkUpEZG8oP4AKFAMLoXDpslmp8kblr4BSVegVAOo0cXsNHlP8Vrw1K/gXhhObYGpj8CVC2anyrVmzZrF4MGDefPNN9m6dStNmjShbdu2hIeH3/F1+/fvJzIyMvVRoUKFNF/39vZO8/XIyEjc3Nyy81vJlX78O5xLV5Io4+tJ2+AAs+OIiIjkGipKiYjkBS6e0OLa7WWrPtbtUPfq8HLY8wtYHKGdmptnm2LVofdv4OELkdthage4fN7sVLnS6NGj6du3L8888wxVqlRhzJgxlCxZkvHjx9/xdX5+fhQrViz14ejomObrFoslzdeLFSuWnd9GrhSflMLENUcA6N+sHI4O+vdCREQkvVSUEhHJK2r1AN+KcPUCrPvC7DS5V3LCf5qbB5ubJ6/zrwa9Fxl9u07vhB/aw+VzZqfKVRITE9m8eTNt2rRJs79NmzasX7/+jq+tXbs2AQEBtGzZkhUrbl7BMy4ujqCgIEqUKMHDDz/M1q1b73i+hIQEYmJi0jzyurlbThAVm0CAjxsdaweaHUdERCRXUVFKRCSvcHSCVu8Z23+Ng5hTpsbJtf76Gs4fMookLdTcPEf4VTYKUwWKQdRumPIwxEWZnSrXOHfuHCkpKfj7+6fZ7+/vz+nTp2/5moCAACZMmMDcuXOZN28elSpVomXLlqxevTr1mMqVKzNlyhQWLlzITz/9hJubG40aNeLgwYO3zTJq1Ch8fHxSHyVLlsyab9JOJadY+WbVYQCea1oWFycNrUVERDJCPzlFRPKSSu2gZD1IvgorR5mdJve5FAGrPzG224w0VoiTnFG0olGY8gqAs3thykMQe+uCitzaf5tr22y22zbcrlSpEs8++yx16tShQYMGjBs3joceeohPP/009Zj69evTo0cPatasSZMmTfj555+pWLEiX3755W0zDB8+nOjo6NRHRERE1nxzduq3HZFEXLhKYU8XnryvlNlxREREch0VpURE8hKLBVr/n7G9dTqc3W9untwmtbl5Q6jxhNlp8h/f8kZhyjsQzh0wClOa8XdXvr6+ODo63jQrKioq6qbZU3dSv379O86CcnBw4L777rvjMa6urnh7e6d55FVWq41xKw8B0LdxGdxdHO/yChEREfkvFaVERPKaUvWg8sNgs8Ky981Ok3sc+hP2LjSamz+k5uamKVLOKEz5lDRuo5zyEESfMDuVXXNxcSEkJISwsLA0+8PCwmjYsGG6z7N161YCAm6/cpzNZmPbtm13PCY/Wbb3DAfOxOHl6kSP+kFmxxEREcmVnMwOICIi2aDlu7D/d9i/CMI3QKn6Zieyb8kJ8Purxna9543m22KewmWMwtQPD8OFI/B9O2OVvoK6Pep2hg4dSs+ePQkNDaVBgwZMmDCB8PBw+vXrBxi31Z08eZKpU6cCMGbMGEqXLk21atVITExk+vTpzJ07l7lz56ae8/3336d+/fpUqFCBmJgYxo4dy7Zt2/j6669N+R7tic1m4+uVRi+pXg2D8HF3NjmRiIhI7qSilIhIXlS0ItTpCZunwB9vQ98/NPPnTv76ypiVU8Afmr9udhoBKBQEvRcbq/FdPArfPwS9f4VCpc1OZpe6dOnC+fPnGTFiBJGRkQQHB7N48WKCgowZPJGRkYSHh6cen5iYyLBhwzh58iTu7u5Uq1aNRYsW0a5du9RjLl26xHPPPcfp06fx8fGhdu3arF69mrp16+b492dv1h06z/aIS7g5O9CnURmz44iIiORaFpvNZjM7RG4UExODj48P0dHRebpfgojkYjGRMLa20fS8y3So0t7sRPbpUgR8dZ/xPnWaqF5S9ib6pFGYunDYuKXvqYVQuKypkTQGSL+8+l51nbCBv46cp3fD0rzXQTMrRURE/iu9YwD1lBIRyau8A6DBQGN72fuQkmxuHnu1dLhRkApqBNUfNzuN/JdPoHErX5EKEB0BUx6G84fNTiX52ObjF/nryHmcHCw819TcAqmIiEhup6KUiEhe1mgQeBSB8wdh6zSz09ifQ8tg769Gc/N2am5ut7wDjMKUbyWIOWk0Pz93+xXgRLLT+Gsr7nWqE0jxgu4mpxEREcndVJQSEcnL3Lyh6bUG3itHQeJlc/PYk+QEWHy9uXk/8K9qbh65My9/ozDlVxViI43C1Nn9ZqeSfGZvZAzL9kZhsUC/ZuXMjiMiIpLrqSglIpLXhT5tNIeOOwN/jTM7jf1Y/6XRp0jNzXOPAkXhqV/BP9j4/3nKQxC11+xUko+Mv7biXrvqAZQtWsDkNCIiIrmfilIiInmdkwvc/7axve4LuHzO3Dz24FI4rP7U2G7zP2NGmeQOnr5GYapYdbh81ihMnd5ldirJB46du8xvO04BMLB5eZPTiIiI5A0qSomI5AfVOkFALUiMhVUfm53GfEuuNzdvDNU7m51GMsqjMPRaaPw/feW8sTpf5A6zU0ke982qw1htcH9lP6oWVyFbREQkK6goJSKSHzg4QOv3je1Nk+HCEXPzmOlgGOz77Vpz80/U3Dy38igMvX6BwBC4esEoTJ3aZnYqyaMio68yd8sJAAa2UC8pERGRrKKilIhIflG2OZRrCdYkWD7S7DTmSE6A3681N6/fX83Nczv3gtBzPpS4D+IvwdQOcHKz2akkD5q4+ihJKTbqlSlMSFBhs+OIiIjkGSpKiYjkJ63fByyway6c3GJ2mpy3fqwxS6xAMWj2mtlpJCu4+UCPeVCyPsRHw9SOEPGP2akkDzkfl8BPG8MBGNhCvaRERESykopSIiL5SbHqUKOLsb3sXbDZzM2Tky4eh9WfGdsPqLl5nuLmDT3mQlAjSIiBaY9C+AazU0keMWX9Ma4mpVA90IcmFXzNjiMiIpKnqCglIpLftHgDHF3g6Go49KfZaXLO0jeM5ualm0DwY2ankazmWgC6zzb++ybGwrROcGyd2akkl4uNT2LK+mOA0UvKoh50IiIiWUpFKRGR/KZQENR9zthe9i5YrebmyQkH/jCamzs4qbl5XubiCd1+NvqnJV2GRUPBmmJ2KsnFpm8IJzY+mfJ+BWhTtZjZcURERPIcFaVERPKjJi+Dqw+c2QU7fzY7TfZKiv+3uXm9fuBXxdw8kr1cPKDrTKjVw/jTwdHsRJJLxSelMGmtsVLpgOblcHBQMVtERCSrqSglIpIfeRSGJkOM7eUjjcJNXrX+S7h4FLwCoPnrZqeRnODsDh2/hsJlzE4iudisfyI4F5dIiULutK9Z3Ow4IiIieZKKUiIi+VW9fuBVHKIj4J+JZqfJHhePw5pPje02I8HVy9w8IpIrJCZb+XbVYQCeb1YOZ0cNmUVERLKDfsKKiORXzu5G03OA1Z/C1Yvm5skOS4ZDcryam4tIhvyy7SSnouPxLeDK4yElzI4jIiKSZ6koJSKSn9XqBkWrQPwlWDvG7DRZ68BS2L/oWnPzT9XcXETSJcVqY/y1WVLPNimDm7P6komIiGQXFaVERPIzB0do9Z6x/fc3EH3C1DhZ5sbm5vX7g19lc/OISK6xdPdpjpy9jLebE93rB5kdR0REJE9TUUpEJL+r+AAENTJuc1sxyuw0WWPdF3DxmNHcvNlrZqcRkVzCZrPx9YpDAPRuVIYCrk4mJxIREcnbVJQSEcnvLBZoPcLY3v4jnNljbp57dfEYrB1tbD/wPzU3F5F0W3XgLLtPxeDh4kifhqXNjiMiIpLnqSglIiJQIhSqdACbFZa9Z3aae3O9uXmZplCtk9lpRCQXGbfC6CXVvV4pCnm6mJxGREQk71NRSkREDC3fBYsjHFwKx9aanSZz9i+B/YvV3FxEMmzj0QtsPHYBF0cHnmlS1uw4IiIi+YKKUiIiYvAtDyG9je2wd8BmMzVOhiVdvaG5+QAoWsncPCKSq1zvJdU5tAT+3m4mpxEREckfVJQSEZF/NX8dnD3h5GbY84vZaTJm3Rdw6Th4FYdmr5qdRkRykV0no1l14CwOFujXtJzZcURERPINFaVERORfBfyg4YvG9p8jICXJ3DzpdeEorP3c2FZzcxHJoHErjVlSHWoWp1QRD5PTiIiI5B8qSomISFoNXwDPonDhMGyeYnaa9Eltbt4Mqj1qdhoRyUUORcXx+67TAPRvXt7kNCIiIvmLilIiIpKWqxc0e83YXvURJMSam+du9v8OB34HB2c1NxeRDPtm1WFsNmhd1Z9KxTTLUkREJCepKCUiIjcL6Q2Fy8Lls/DX12anub2kq/D7tQJagwFQtKK5eUQkVzlx8QoLtp4EYGALzZISERHJaSpKiYjIzRydoeU7xva6sRAXZW6e21k75t/m5k3V3FxEMmbi6iMkW200Lu9LrZIFzY4jIiKS76goJSIit1a1IwSGQNJl4zY+e3PhyL/NzR/8AFwLmJtHRHKVs7EJzPwnAoABLbTinoiIiBlUlBIRkVuzWKD1CGN78xQ4f9jUODdZMhxSEqBsc6OAJiKSAZPWHiUh2UrtUgVpULaI2XFERETyJRWlRETk9ko3hgptwJoMf75vdpp/7f8dDiwxmpu3/UTNzUUkQ6KvJDF9w3EABjYvj0X/hoiIiJhCRSkREbmzVu8BFtjzC5zYZHaaa83Nr/WPaviCmpuLSIZN/esYcQnJVC7mxf2V/cyOIyIikm+pKCUiInfmXw1qdTO2w94Fm83cPGs/h0vh4B0ITV8xN4uI5DpXEpOZvO4oAP2bl8PBQbOkREREzGJ6UWrcuHGUKVMGNzc3QkJCWLNmzR2PT0hI4M033yQoKAhXV1fKlSvH5MmT0xxz6dIlBg4cSEBAAG5ublSpUoXFixff03VFRPK1Fm+AkxscXwsH/zAvx/nDxop7AA98AC6e5mURsTMZGdusXLkSi8Vy02Pfvn23PH7mzJlYLBY6duyYTelzzk8bI7h4JYmgIh48VD3A7DgiIiL5mpOZF581axaDBw9m3LhxNGrUiG+//Za2bduyZ88eSpUqdcvXPPHEE5w5c4ZJkyZRvnx5oqKiSE5OTv16YmIirVu3xs/Pjzlz5lCiRAkiIiLw8vK6p+uKiORrPiWg3vOw7gtY9h6UbwUOjjmbwWaDJa9fa27eAqo+krPXF7FjmR3b7N+/H29v79TnRYsWvemY48ePM2zYMJo0aZIt2XNSQnIKE1cfAaBfs3I4OZr++ayIiEi+ZrHZzLsPo169etSpU4fx48en7qtSpQodO3Zk1KhRNx2/ZMkSnnzySY4cOULhwoVvec5vvvmGTz75hH379uHs7Jwl172VmJgYfHx8iI6OTjOYExHJs65ehC9qQfwleORrqN0jZ6+/bzHM7Go0Nx/wF/hWyNnri1xjj2OAjI5tVq5cSYsWLbh48SIFCxa87XlTUlJo1qwZffr0Yc2aNVy6dIkFCxakO5e9vVczN4bz+rydFPN2Y9WrzXF1yuHiuoiISD6R3jGAaR8PJSYmsnnzZtq0aZNmf5s2bVi/fv0tX7Nw4UJCQ0P5+OOPCQwMpGLFigwbNoyrV6+mOaZBgwYMHDgQf39/goOD+eCDD0hJScn0dcG4bTAmJibNQ0QkX3EvBE1eNrZXfGA0HM8piVfg99eM7YYvqiAlcoPMjm0AateuTUBAAC1btmTFihU3fX3EiBEULVqUvn37ZmlmMySnWBm/6jAAzzYtq4KUiIiIHTDt9r1z586RkpKCv79/mv3+/v6cPn36lq85cuQIa9euxc3Njfnz53Pu3DkGDBjAhQsXUvtKHTlyhOXLl9O9e3cWL17MwYMHGThwIMnJybzzzjuZui7AqFGjeP99O1oOXUTEDHWfg40TIDoC/v4WGg/Omeuu/Ryiw8G7BDQdljPXFMklMjO2CQgIYMKECYSEhJCQkMC0adNo2bIlK1eupGnTpgCsW7eOSZMmsW3btnRnSUhIICEhIfW5PX2It2hnJMfPX6GQhzNd65Y0O46IiIhgck8pAIsl7YonNpvtpn3XWa1WLBYLM2bMwMfHB4DRo0fTuXNnvv76a9zd3bFarfj5+TFhwgQcHR0JCQnh1KlTfPLJJ7zzzjuZui7A8OHDGTp0aOrzmJgYSpbUgEZE8hlnN2jxJizoB2tHQ51e4HHr26mzzPnDsG6Msf2gmpuL3E5GxjaVKlWiUqVKqc8bNGhAREQEn376KU2bNiU2NpYePXowceJEfH19053BXj/Es1ptjFthzJJ6ulEZPFxMHwKLiIgIJt6+5+vri6Oj402f4EVFRd30Sd91AQEBBAYGphakwOiXYLPZOHHiROoxFStWxNHRMc0xp0+fJjExMVPXBXB1dcXb2zvNQ0QkX6rxBPgHQ3w0rPkse69lsxm37aUkQrn7oUqH7L2eSC6U2bHNf9WvX5+DBw8CcPjwYY4dO0b79u1xcnLCycmJqVOnsnDhQpycnDh8+PAtzzF8+HCio6NTHxEREZn/xrLQ8n1R7D8TSwFXJ3o1KG12HBEREbnGtKKUi4sLISEhhIWFpdkfFhZGw4YNb/maRo0acerUKeLi4lL3HThwAAcHB0qUKJF6zKFDh7BarWmOCQgIwMXFJVPXFRGRGzg4Qqv3jO2NE+BSePZda/9iOBRmNDdv+wncYUarSH6VVWObrVu3EhAQAEDlypXZuXMn27ZtS3106NCBFi1asG3bttvOFrfHD/FsNhtfrTgEQI/6Qfh43HohHBEREcl5ps5dHjp0KD179iQ0NJQGDRowYcIEwsPD6devH2B82nby5EmmTp0KQLdu3fi///s/+vTpw/vvv8+5c+d45ZVXePrpp3F3dwegf//+fPnllwwaNIgXX3yRgwcP8sEHH/DSSy+l+7oiInIX5VtB6SZwbA0s/x90+jbrr5F4BX5/3dhu9BL4ls/6a4jkERkdU40ZM4bSpUtTrVo1EhMTmT59OnPnzmXu3LkAuLm5ERwcnOYa11fp++9+e/fXkfNsi7iEq5MDfRuXMTuOiIiI3MDUolSXLl04f/48I0aMIDIykuDgYBYvXkxQUBAAkZGRhIf/+wl8gQIFCAsL48UXXyQ0NJQiRYrwxBNPMHLkyNRjSpYsyR9//MGQIUOoUaMGgYGBDBo0iNdeey3d1xURkbuwWKD1CJjYAnbMgoYvQLHqWXuNtaP/bW5+fdU/EbmljI6pEhMTGTZsGCdPnsTd3Z1q1aqxaNEi2rVrZ9a3kG2u95Lqcl9Jinq5mpxGREREbmSx2Ww2s0PkRjExMfj4+BAdHW0XU9NFREwxuw/snmfMnOoxN+vOe/4wjKtv9JJ6YhpUVS8psR8aA6Sf2e/VtohLdPx6HU4OFla+0pwShTxyPIOIiEh+lN4xgGk9pUREJA9o+TY4OMGhZXBkVdac02aDxa9ca27eEqq0z5rziki+8/W1XlIdaweqICUiImKHVJQSEZHMK1wWQp82tsPegRsWmci0fb/B4T/B0QXaqbm5iGTO/tOxhO05g8UC/ZqVMzuOiIiI3IKKUiIicm+avgouBSBym3Er371IvAJLhhvbDV+CIvpFUkQyZ/xKY5ZU2+BilPcrYHIaERERuRUVpURE5N4UKAqNBhnby/8PkhMzf641n0F0BPiUVHNzEcm08PNXWLj9FAADmmvlThEREXulopSIiNy7BgOhgD9cPAabv8/cOc4dgvVjje0HR4GL+r+ISOZ8s/owVhs0q1iU4EAfs+OIiIjIbagoJSIi987FE5q/bmyv+gjiYzL2epsNfr/W3Lx8K6j8cNZnFJF84UxMPHM2nQBgYAvNkhIREbFnKkqJiEjWqN0TipSHK+f/nfGUXnt/hcPLjebmbT9Wc3MRybTv1hwhMcXKfaULUbdMYbPjiIiIyB2oKCUiIlnD0Rlavmts//U1xJ5O3+sSL8PSN4ztRoPU3FxEMu3i5URm/B0OaJaUiIhIbqCilIiIZJ0q7aFEXUi6Ais/TN9rUpubl4LGQ7M3n4jkad+vP8aVxBSqFfemWcWiZscRERGRu1BRSkREso7FAq1HGNtbpsK5g3c+/twhWKfm5iJy7+ISkpmy7ihgzJKy6DZgERERu6eilIiIZK2gBlCpHdhSYNl7tz/uenNzaxKUbw2VH8qxiCKS98zYcJyY+GTKFvXkgWrFzI4jIiIi6aCilIiIZL2W74LFAfb9BuF/3/qYvQtvaG7+kZqbi0imxSelMHGNMUuqf7NyODro3xMREZHcQEUpERHJen6VoVZ3YzvsHWNW1I0SL8OS683NB6u5uYjck9mbT3AuLoHAgu50rB1odhwRERFJJxWlREQke7R4A5zcIWID7P897ddWfwoxJ6BgKWg8xJx8IpInJKVY+XbVYQCea1oWZ0cNb0VERHIL/dQWEZHs4V0c6vc3tpe9BynJxva5g7D+S2P7wQ/V3FxE7smv209x4uJVfAu40OW+kmbHERERkQxQUUpERLJP48HgXhjO7YdtM4zb+BZfa25eoY3REF1EJJOsVhvjVhqzpPo2Loubs6PJiURERCQjVJQSEZHs4+YDTV8xtleOgu0z4cgKcHRVc3MRuWd/7DnNoag4vNyc6FG/lNlxREREJINUlBIRkex1X1+jd1RsJPwywNjXeDAULmtqLBHJ3Ww2G1+vMGZJ9W5YGi83Z5MTiYiISEapKCUiItnLyRXuf9vYtlnV3FxEssSag+fYeTIad2dH+jQqY3YcERERyQQVpUREJPsFd4bAEGO77Sfg7G5uHhHJ9Wb9EwFA17qlKOzpYnIaERERyQwnswOIiEg+4OAAPedDTCT4VTY7jYjkAaO71KRxBV+aVypqdhQRERHJJBWlREQkZ7j5GA8RkSzg6uRI17pqbi4iIpKb6fY9ERERERERERHJcSpKiYiIiIiIiIhIjlNRSkREREREREREcpyKUiIiIiIiIiIikuNUlBIRERERERERkRynopSIiIiIiIiIiOQ4FaVERERERERERCTHqSglIiIiIiIiIiI5TkUpERERERERERHJcSpKiYiIiIiIiIhIjlNRSkREREREREREcpyKUiIiIiIiIiIikuNUlBIRERERERERkRynopSIiIiIiIiIiOQ4FaVERERERERERCTHOZkdILey2WwAxMTEmJxEREREctL1n/3XxwJyexoviYiI5E/pHS+pKJVJsbGxAJQsWdLkJCIiImKG2NhYfHx8zI5h1zReEhERyd/uNl6y2PQxX6ZYrVZOnTqFl5cXFoslS88dExNDyZIliYiIwNvbO0vPnZfpfcscvW+Zo/ct4/SeZY7et8zJzvfNZrMRGxtL8eLFcXBQJ4Q7yc7xEujvR2boPcscvW+Zo/ct4/SeZY7et8yxh/GSZkplkoODAyVKlMjWa3h7e+svVCbofcscvW+Zo/ct4/SeZY7et8zJrvdNM6TSJyfGS6C/H5mh9yxz9L5ljt63jNN7ljl63zLHzPGSPt4TEREREREREZEcp6KUiIiIiIiIiIjkOBWl7JCrqyvvvvsurq6uZkfJVfS+ZY7et8zR+5Zxes8yR+9b5uh9yx/03znj9J5ljt63zNH7lnF6zzJH71vm2MP7pkbnIiIiIiIiIiKS4zRTSkREREREREREcpyKUiIiIiIiIiIikuNUlBIRERERERERkRynopQdGjduHGXKlMHNzY2QkBDWrFljdiS7tnr1atq3b0/x4sWxWCwsWLDA7Eh2b9SoUdx33314eXnh5+dHx44d2b9/v9mx7N748eOpUaMG3t7eeHt706BBA37//XezY+Uqo0aNwmKxMHjwYLOj2L333nsPi8WS5lGsWDGzY9m9kydP0qNHD4oUKYKHhwe1atVi8+bNZseSbKDxUsZovJRxGi9ljsZLWUNjpvTReClz7Gm8pKKUnZk1axaDBw/mzTffZOvWrTRp0oS2bdsSHh5udjS7dfnyZWrWrMlXX31ldpRcY9WqVQwcOJANGzYQFhZGcnIybdq04fLly2ZHs2slSpTgww8/ZNOmTWzatIn777+fRx55hN27d5sdLVf4559/mDBhAjVq1DA7Sq5RrVo1IiMjUx87d+40O5Jdu3jxIo0aNcLZ2Znff/+dPXv28Nlnn1GwYEGzo0kW03gp4zReyjiNlzJH46V7pzFTxmi8lDH2Nl7S6nt2pl69etSpU4fx48en7qtSpQodO3Zk1KhRJibLHSwWC/Pnz6djx45mR8lVzp49i5+fH6tWraJp06Zmx8lVChcuzCeffELfvn3NjmLX4uLiqFOnDuPGjWPkyJHUqlWLMWPGmB3Lrr333nssWLCAbdu2mR0l13j99ddZt26dZszkAxov3RuNlzJH46XM03gp/TRmyhiNlzLO3sZLmillRxITE9m8eTNt2rRJs79NmzasX7/epFSSH0RHRwPGgEHSJyUlhZkzZ3L58mUaNGhgdhy7N3DgQB566CFatWpldpRc5eDBgxQvXpwyZcrw5JNPcuTIEbMj2bWFCxcSGhrK448/jp+fH7Vr12bixIlmx5IspvGSmEXjpYzTeCnjNGbKOI2XMsbexksqStmRc+fOkZKSgr+/f5r9/v7+nD592qRUktfZbDaGDh1K48aNCQ4ONjuO3du5cycFChTA1dWVfv36MX/+fKpWrWp2LLs2c+ZMtmzZotkLGVSvXj2mTp3K0qVLmThxIqdPn6Zhw4acP3/e7Gh268iRI4wfP54KFSqwdOlS+vXrx0svvcTUqVPNjiZZSOMlMYPGSxmj8VLmaMyUcRovZZy9jZecTLmq3JHFYknz3Gaz3bRPJKu88MIL7Nixg7Vr15odJVeoVKkS27Zt49KlS8ydO5ennnqKVatWaaB1GxEREQwaNIg//vgDNzc3s+PkKm3btk3drl69Og0aNKBcuXL88MMPDB061MRk9stqtRIaGsoHH3wAQO3atdm9ezfjx4+nV69eJqeTrKbxkuQkjZcyRuOljNOYKXM0Xso4exsvaaaUHfH19cXR0fGmT/mioqJu+jRQJCu8+OKLLFy4kBUrVlCiRAmz4+QKLi4ulC9fntDQUEaNGkXNmjX54osvzI5ltzZv3kxUVBQhISE4OTnh5OTEqlWrGDt2LE5OTqSkpJgdMdfw9PSkevXqHDx40OwodisgIOCmX3iqVKmi5td5jMZLktM0Xso4jZcyTmOmrKHx0t3Z23hJRSk74uLiQkhICGFhYWn2h4WF0bBhQ5NSSV5ks9l44YUXmDdvHsuXL6dMmTJmR8q1bDYbCQkJZsewWy1btmTnzp1s27Yt9REaGkr37t3Ztm0bjo6OZkfMNRISEti7dy8BAQFmR7FbjRo1umm59gMHDhAUFGRSIskOGi9JTtF4KetovHR3GjNlDY2X7s7exku6fc/ODB06lJ49exIaGkqDBg2YMGEC4eHh9OvXz+xodisuLo5Dhw6lPj969Cjbtm2jcOHClCpVysRk9mvgwIH8+OOP/PLLL3h5eaV+2uzj44O7u7vJ6ezXG2+8Qdu2bSlZsiSxsbHMnDmTlStXsmTJErOj2S0vL6+bem94enpSpEgR9eS4i2HDhtG+fXtKlSpFVFQUI0eOJCYmhqeeesrsaHZryJAhNGzYkA8++IAnnniCjRs3MmHCBCZMmGB2NMliGi9lnMZLGafxUuZovJQ5GjNljsZLGWd34yWb2J2vv/7aFhQUZHNxcbHVqVPHtmrVKrMj2bUVK1bYgJseTz31lNnR7Nat3i/A9v3335sdza49/fTTqX83ixYtamvZsqXtjz/+MDtWrtOsWTPboEGDzI5h97p06WILCAiwOTs724oXL27r1KmTbffu3WbHsnu//vqrLTg42Obq6mqrXLmybcKECWZHkmyi8VLGaLyUcRovZY7GS1lHY6a703gpc+xpvGSx2Wy2nCyCiYiIiIiIiIiIqKeUiIiIiIiIiIjkOBWlREREREREREQkx6koJSIiIiIiIiIiOU5FKRERERERERERyXEqSomIiIiIiIiISI5TUUpERERERERERHKcilIiIiIiIiIiIpLjVJQSEREREREREZEcp6KUiEg2slgsLFiwwOwYIiIiInZL4yWR/EtFKRHJs3r37o3FYrnp8eCDD5odTURERMQuaLwkImZyMjuAiEh2evDBB/n+++/T7HN1dTUpjYiIiIj90XhJRMyimVIikqe5urpSrFixNI9ChQoBxlTx8ePH07ZtW9zd3SlTpgyzZ89O8/qdO3dy//334+7uTpEiRXjuueeIi4tLc8zkyZOpVq0arq6uBAQE8MILL6T5+rlz53j00Ufx8PCgQoUKLFy4MPVrFy9epHv37hQtWhR3d3cqVKhw06BQREREJDtpvCQiZlFRSkTytbfffpvHHnuM7du306NHD7p27crevXsBuHLlCg8++CCFChXin3/+Yfbs2SxbtizNIGr8+PEMHDiQ5557jp07d7Jw4ULKly+f5hrvv/8+TzzxBDt27KBdu3Z0796dCxcupF5/z549/P777+zdu5fx48fj6+ubc2+AiIiIyF1ovCQi2cYmIpJHPfXUUzZHR0ebp6dnmseIESNsNpvNBtj69euX5jX16tWz9e/f32az2WwTJkywFSpUyBYXF5f69UWLFtkcHBxsp0+fttlsNlvx4sVtb7755m0zALa33nor9XlcXJzNYrHYfv/9d5vNZrO1b9/e1qdPn6z5hkVEREQySOMlETGTekqJSJ7WokULxo8fn2Zf4cKFU7cbNGiQ5msNGjRg27ZtAOzdu5eaNWvi6emZ+vVGjRphtVrZv38/FouFU6dO0bJlyztmqFGjRuq2p6cnXl5eREVFAdC/f38ee+wxtmzZQps2bejYsSMNGzbM1PcqIiIikhkaL4mIWVSUEpE8zdPT86bp4XdjsVgAsNlsqdu3Osbd3T1d53N2dr7ptVarFYC2bdty/PhxFi1axLJly2jZsiUDBw7k008/zVBmERERkczSeElEzKKeUiKSr23YsOGm55UrVwagatWqbNu2jcuXL6d+fd26dTg4OFCxYkW8vLwoXbo0f/755z1lKFq0KL1792b69OmMGTOGCRMm3NP5RERERLKSxksikl00U0pE8rSEhAROnz6dZp+Tk1Nqc8zZs2cTGhpK48aNmTFjBhs3bmTSpEkAdO/enXfffZennnqK9957j7Nnz/Liiy/Ss2dP/P39AXjvvffo168ffn5+tG3bltjYWNatW8eLL76YrnzvvPMOISEhVKtWjYSEBH777TeqVKmShe+AiIiIyJ1pvCQiZlFRSkTyA7ZBwAAAATpJREFUtCVLlhAQEJBmX6VKldi3bx9grPQyc+ZMBgwYQLFixZgxYwZVq1YFwMPDg6VLlzJo0CDuu+8+PDw8eOyxxxg9enTquZ566ini4+P5/PPPGTZsGL6+vnTu3Dnd+VxcXBg+fDjHjh3D3d2dJk2aMHPmzCz4zkVERETSR+MlETGLxWaz2cwOISJiBovFwvz58+nYsaPZUURERETsksZLIpKd1FNKRERERERERERynIpSIiIiIiIiIiKS43T7noiIiIiIiIiI5DjNlBIRERERERERkRynopSIiIiIiIiIiOQ4FaVERERERERERCTHqSglIiIiIiIiIiI5TkUpERERERERERHJcSpKiYiIiIiIiIhIjlNRSkREREREREREcpyKUiIiIiIiIiIikuNUlBIRERERERERkRz3/5A4ZFD+yjzEAAAAAElFTkSuQmCC", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import LabelEncoder\n", "from tensorflow.keras.preprocessing.text import Tokenizer\n", "from keras_preprocessing.sequence import pad_sequences\n", "from keras.models import Sequential\n", "from keras.layers import Embedding, LSTM, Dense, Dropout\n", "from gensim.models import Word2Vec\n", "from keras.layers import Embedding, LSTM, Dense, Dropout, Bidirectional, GlobalMaxPooling1D\n", "\n", "\n", "# Encode labels\n", "le = LabelEncoder()\n", "df['news'] = le.fit_transform(df['news'])\n", "\n", "# Tokenize the titles\n", "tokenizer = Tokenizer()\n", "tokenizer.fit_on_texts(df['title'])\n", "sequences = tokenizer.texts_to_sequences(df['title'])\n", "word_index = tokenizer.word_index\n", "\n", "# Padding sequences\n", "max_len = max(len(x) for x in sequences)\n", "X = pad_sequences(sequences, maxlen=max_len, padding='post')\n", "y = df['news'].values\n", "\n", "# Split the data\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Train Word2Vec model on the training set only\n", "sentences_train = [df['title'][i].split() for i in range(len(X_train))]\n", "word2vec_model = Word2Vec(sentences_train, vector_size=100, window=5, min_count=1, workers=4)\n", "\n", "# Create embedding matrix\n", "embedding_matrix = np.zeros((len(word_index) + 1, 100))\n", "for word, i in word_index.items():\n", " if word in word2vec_model.wv:\n", " embedding_matrix[i] = word2vec_model.wv[word]\n", "#-------------------------------\n", "# Updated LSTM Model\n", "# Simple LSTM Model\n", "model = Sequential()\n", "\n", "# Embedding layer with pre-trained Word2Vec weights\n", "model.add(Embedding(input_dim=len(word_index) + 1, \n", " output_dim=100, \n", " weights=[embedding_matrix], \n", " input_length=max_len, \n", " trainable=False))\n", "\n", "# Single LSTM layer\n", "model.add(LSTM(64))\n", "\n", "# Dropout for regularization\n", "model.add(Dropout(0.3))\n", "\n", "# Output layer\n", "model.add(Dense(1, activation='sigmoid')) # Binary classification\n", "\n", "# Compile the model\n", "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n", "#---------------------\n", "\n", "history=model.fit(X_train, y_train, epochs=7, batch_size=32, validation_data=(X_test, y_test))\n", "\n", "# Evaluate the model\n", "loss, accuracy = model.evaluate(X_test, y_test)\n", "print(f\"Test Loss: {loss}\")\n", "print(f\"Test Accuracy: {accuracy}\")\n", "accuracy_scores['Word2Vec With LSTM']=accuracy\n", "def plot_metrics(history):\n", " # Plot loss\n", " plt.figure(figsize=(12, 5))\n", " plt.subplot(1, 2, 1)\n", " plt.plot(history.history['loss'], label='Training Loss')\n", " plt.plot(history.history['val_loss'], label='Validation Loss')\n", " plt.title('Loss over Epochs')\n", " plt.xlabel('Epochs')\n", " plt.ylabel('Loss')\n", " plt.legend()\n", "\n", " # Plot accuracy\n", " plt.subplot(1, 2, 2)\n", " plt.plot(history.history['accuracy'], label='Training Accuracy')\n", " plt.plot(history.history['val_accuracy'], label='Validation Accuracy')\n", " plt.title('Accuracy over Epochs')\n", " plt.xlabel('Epochs')\n", " plt.ylabel('Accuracy')\n", " plt.legend()\n", "\n", " plt.tight_layout()\n", " plt.show()\n", "\n", "# Call the plot_metrics function\n", "plot_metrics(history)" ] }, { "cell_type": "code", "execution_count": 52, "id": "01cbb05d-cb1e-4002-8894-a34725cfd66a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/4\n", "96/96 [==============================] - 2s 7ms/step - loss: 0.6274 - accuracy: 0.6393 - val_loss: 0.5472 - val_accuracy: 0.7214\n", "Epoch 2/4\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.5293 - accuracy: 0.7418 - val_loss: 0.5232 - val_accuracy: 0.7490\n", "Epoch 3/4\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.4808 - accuracy: 0.7727 - val_loss: 0.4907 - val_accuracy: 0.7582\n", "Epoch 4/4\n", "96/96 [==============================] - 0s 3ms/step - loss: 0.4403 - accuracy: 0.8003 - val_loss: 0.4851 - val_accuracy: 0.7582\n", "24/24 [==============================] - 0s 1ms/step - loss: 0.4851 - accuracy: 0.7582\n", "Test Loss: 0.48506996035575867\n", "Test Accuracy: 0.7582128643989563\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import LabelEncoder\n", "from tensorflow.keras.preprocessing.text import Tokenizer\n", "from keras_preprocessing.sequence import pad_sequences\n", "from keras.models import Sequential\n", "from keras.layers import Embedding, LSTM, Dense, Dropout\n", "\n", "# Load your dataset (assuming df is your DataFrame with 'title' and 'news' columns)\n", "# df = pd.read_csv('your_dataset.csv')\n", "\n", "# Encode labels\n", "le = LabelEncoder()\n", "df['news'] = le.fit_transform(df['news'])\n", "\n", "# Tokenize the titles\n", "tokenizer = Tokenizer()\n", "tokenizer.fit_on_texts(df['title'])\n", "sequences = tokenizer.texts_to_sequences(df['title'])\n", "word_index = tokenizer.word_index\n", "\n", "# Padding sequences\n", "max_len = max(len(x) for x in sequences)\n", "X = pad_sequences(sequences, maxlen=max_len, padding='post')\n", "y = df['news'].values\n", "\n", "# Split the data\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n", "\n", "# Load GloVe embeddings\n", "embedding_dim = 100\n", "glove_path = \"./glove.6B.100d.txt\" # Update this path with the correct location of your GloVe file\n", "\n", "# Create an embedding index\n", "embedding_index = {}\n", "with open(glove_path, \"r\", encoding=\"utf-8\") as f:\n", " for line in f:\n", " values = line.split()\n", " word = values[0]\n", " coefs = np.asarray(values[1:], dtype=\"float32\")\n", " embedding_index[word] = coefs\n", "\n", "# Create the embedding matrix\n", "embedding_matrix = np.zeros((len(word_index) + 1, embedding_dim))\n", "for word, i in word_index.items():\n", " embedding_vector = embedding_index.get(word)\n", " if embedding_vector is not None:\n", " embedding_matrix[i] = embedding_vector\n", "\n", "# Build the simple LSTM model\n", "model = Sequential()\n", "\n", "# Embedding layer with pre-trained GloVe weights\n", "model.add(Embedding(input_dim=len(word_index) + 1,\n", " output_dim=embedding_dim,\n", " weights=[embedding_matrix],\n", " input_length=max_len,\n", " trainable=False))\n", "\n", "# Single LSTM layer\n", "model.add(LSTM(64))\n", "\n", "# Dropout for regularization\n", "model.add(Dropout(0.3))\n", "\n", "# Output layer\n", "model.add(Dense(1, activation='sigmoid')) # Binary classification\n", "\n", "# Compile the model\n", "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n", "\n", "# Train the model\n", "history=model.fit(X_train, y_train, epochs=4, batch_size=32, validation_data=(X_test, y_test))\n", "\n", "# Evaluate the model\n", "loss, accuracy = model.evaluate(X_test, y_test)\n", "print(f\"Test Loss: {loss}\")\n", "print(f\"Test Accuracy: {accuracy}\")\n", "accuracy_scores['Glove With LSTM']=accuracy\n", "def plot_metrics(history):\n", " # Plot loss\n", " plt.figure(figsize=(12, 5))\n", " plt.subplot(1, 2, 1)\n", " plt.plot(history.history['loss'], label='Training Loss')\n", " plt.plot(history.history['val_loss'], label='Validation Loss')\n", " plt.title('Loss over Epochs')\n", " plt.xlabel('Epochs')\n", " plt.ylabel('Loss')\n", " plt.legend()\n", "\n", " # Plot accuracy\n", " plt.subplot(1, 2, 2)\n", " plt.plot(history.history['accuracy'], label='Training Accuracy')\n", " plt.plot(history.history['val_accuracy'], label='Validation Accuracy')\n", " plt.title('Accuracy over Epochs')\n", " plt.xlabel('Epochs')\n", " plt.ylabel('Accuracy')\n", " plt.legend()\n", "\n", " plt.tight_layout()\n", " plt.show()\n", "\n", "# Call the plot_metrics function\n", "plot_metrics(history)" ] }, { "cell_type": "code", "execution_count": 53, "id": "4eba3a75-8e33-4453-abaa-526413c02f4c", "metadata": {}, "outputs": [], "source": [ "ranked_models = sorted(accuracy_scores.items(), key=lambda x: x[1], reverse=True)" ] }, { "cell_type": "code", "execution_count": 54, "id": "a883f144-71ef-48c2-af43-00590f811235", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[('Bag of N-grams with Naive Bayes', 0.8199737187910644),\n", " ('TFIDF with SVM', 0.80946123521682),\n", " ('Bag of Words with Naive Bayes', 0.8055190538764783),\n", " ('TFIDF with Naive Bayes', 0.8028909329829172),\n", " ('Bag Of N Grams With Logistic Regression', 0.8002628120893561),\n", " ('Tfidf With Logistic Regression', 0.7963206307490145),\n", " ('Bag of Words with Random Forest', 0.7963206307490145),\n", " ('Bag Of Word With Logistic Regression', 0.7910643889618922),\n", " ('Bag of N-grams with Random Forest', 0.7910643889618922),\n", " ('Bag of Words with SVM', 0.783180026281209),\n", " ('TFIDF with Random Forest', 0.7792378449408672),\n", " ('Word2Vec With SVM', 0.7792378449408672),\n", " ('Bag of N-grams with Decision Tree', 0.7660972404730617),\n", " ('GloVe with SVM', 0.7621550591327201),\n", " ('Glove With LSTM', 0.7582128643989563),\n", " ('Bag of Words with Decision Tree', 0.7503285151116952),\n", " ('TFIDF with Decision Tree', 0.7411300919842313),\n", " ('Word2Vec With Logistic Regression', 0.7293035479632063),\n", " ('GloVe with Random Forest', 0.7293035479632063),\n", " ('Word2Vec With RandomForest', 0.7148488830486203),\n", " ('GloVe With Logistic Regression', 0.7122207621550591),\n", " ('Word2Vec with Naive Bayes', 0.6727989487516426),\n", " ('GloVe with Naive Bayes', 0.668856767411301),\n", " ('GloVe with Decision Tree', 0.6386333771353482),\n", " ('Word2Vec With LSTM', 0.6084100008010864),\n", " ('Word2Vec With DecisionTree', 0.5781865965834428),\n", " ('Bag of N-grams with SVM', 0.5742444152431012)]" ] }, "execution_count": 54, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ranked_models" ] }, { "cell_type": "code", "execution_count": null, "id": "6e9e9a35-8788-451a-a1f8-d1b33506699b", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.15" } }, "nbformat": 4, "nbformat_minor": 5 }