main.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import re\n",
    "from collections import Counter\n",
    "import pickle\n",
    "import numpy as np\n",
    "from textblob import TextBlob\n",
    "import nltk\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "from scipy.sparse import csr_matrix, hstack\n",
    "from sklearn.datasets import load_files\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.feature_extraction import DictVectorizer\n",
    "from sklearn.decomposition import LatentDirichletAllocation\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.svm import SVC, LinearSVC\n",
    "from sklearn.ensemble import RandomForestClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def remove_url(text):\n",
    "    pattern = \"((http|ftp|https):\\/\\/)?([\\w_-]+(?:(?:\\.[\\w_-]+)+))([\\w.,@?^=%&:\\/~+#-]*[\\w.,@?^=%&:\\/~+#-])?\"\n",
    "    return re.sub(pattern, \"\", text)\n",
    "\n",
    "def remove_stopwords(text):\n",
    "    stop_words = set(stopwords.words(\"english\"))\n",
    "    tokens = nltk.word_tokenize(text)\n",
    "    filtered_sentence = \" \".join([w for w in tokens if not w in stop_words])\n",
    "    return filtered_sentence\n",
    "\n",
    "def is_too_short(text):\n",
    "    tokens = nltk.word_tokenize(text)\n",
    "    return len(tokens) <= 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def divide_text(text, n):\n",
    "    tokens = nltk.word_tokenize(text)\n",
    "    spilting_length = len(tokens) / n\n",
    "    out = []\n",
    "    x = 0\n",
    "    for i in range(n):\n",
    "        str_list = tokens[x:int(x+spilting_length)]\n",
    "        string = \" \".join(str_list)\n",
    "        out.append(string)\n",
    "        x = int(x+spilting_length)\n",
    "    return out\n",
    "\n",
    "def get_sentiment(arr):\n",
    "    n = len(arr)\n",
    "    polar = []\n",
    "    for i in range(n):\n",
    "        analysis = TextBlob(arr[i])\n",
    "        polar.append(analysis.sentiment.polarity)\n",
    "    return polar\n",
    "\n",
    "def find_sentiment(arr):\n",
    "    n = len(arr)\n",
    "    out = np.empty((len(arr), 6))\n",
    "    for i in range(len(arr)):\n",
    "        analysis = TextBlob(arr[i])\n",
    "        uni_polar = analysis.sentiment.polarity\n",
    "        string_list = divide_text(arr[i], 2)\n",
    "        bi_polar = get_sentiment(string_list)\n",
    "        string_list = divide_text(arr[i], 3)\n",
    "        tri_polar = get_sentiment(string_list)\n",
    "        out[i] = [uni_polar, bi_polar[0], bi_polar[1], tri_polar[0], tri_polar[1], tri_polar[2]]\n",
    "    return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def stem_tokens(tokens, stemmer):\n",
    "    stemmed = []\n",
    "    for item in tokens:\n",
    "        stemmed.append(stemmer.stem(item))\n",
    "    return stemmed\n",
    "\n",
    "def tokenize(text):\n",
    "    tokens = nltk.word_tokenize(text)\n",
    "    stems = stem_tokens(tokens, stemmer)\n",
    "    return stems"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def pos_tag_finder(text):\n",
    "    tokens = nltk.word_tokenize(text)\n",
    "    tags = nltk.pos_tag(tokens)\n",
    "    counts = Counter(tag for word,tag in tags)\n",
    "    total = sum(counts.values())\n",
    "    return dict((word, float(count)/total) for word,count in counts.items())\n",
    "\n",
    "def get_pos_features(arr):\n",
    "    out = np.array([])\n",
    "    for i in range(len(arr)):\n",
    "        pos_tags = pos_tag_finder(arr[i])\n",
    "        out = np.append(out, pos_tags)\n",
    "    return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataset = load_files('container/', encoding=\"utf8\", decode_error=\"replace\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = np.array([])\n",
    "y = np.array([])\n",
    "for i in range(len(dataset.data)):\n",
    "    if not is_too_short(dataset.data[i]):\n",
    "        noisless_text = remove_url(str(dataset.data[i]))\n",
    "        noisless_text = remove_stopwords(noisless_text)\n",
    "        X = np.append(X, noisless_text)\n",
    "        if dataset.target[i] == 0:\n",
    "            y = np.append(y, 'notsarc')\n",
    "        else:\n",
    "            y = np.append(y, 'sarc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sentiment_train = csr_matrix(find_sentiment(X_train))\n",
    "sentiment_test = csr_matrix(find_sentiment(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "stemmer = PorterStemmer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "tfidf = TfidfVectorizer(tokenizer=tokenize, stop_words='english', ngram_range=(1, 5), max_features=1939)\n",
    "tfs_train = tfidf.fit_transform(X_train)\n",
    "tfs_test = tfidf.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "vec = DictVectorizer()\n",
    "\n",
    "pos_train = vec.fit_transform(get_pos_features(X_train))\n",
    "pos_test = vec.transform(get_pos_features(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lda = LatentDirichletAllocation(n_topics=10, learning_method='online')\n",
    "\n",
    "topic_train = lda.fit_transform(tfs_train)\n",
    "topic_test = lda.transform(tfs_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "final_train = hstack([sentiment_train, tfs_train, pos_train, topic_train])\n",
    "final_test = hstack([sentiment_test, tfs_test, pos_test, topic_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.581218274112\n"
     ]
    }
   ],
   "source": [
    "logistic_clf = LogisticRegression(C=0.2)\n",
    "logistic_clf = logistic_clf.fit(final_train, y_train)\n",
    "predict = logistic_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.558375634518\n"
     ]
    }
   ],
   "source": [
    "svm_clf = SVC(C=4, gamma=1.3)\n",
    "svm_clf = svm_clf.fit(final_train, y_train)\n",
    "predict = svm_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.591370558376\n"
     ]
    }
   ],
   "source": [
    "linear_svm_clf = LinearSVC(C=0.1)\n",
    "linear_svm_clf = linear_svm_clf.fit(final_train, y_train)\n",
    "predict = linear_svm_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.571065989848\n"
     ]
    }
   ],
   "source": [
    "rf_clf = RandomForestClassifier(n_estimators=100)\n",
    "rf_clf = rf_clf.fit(final_train, y_train)\n",
    "predict = rf_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.596446700508\n"
     ]
    }
   ],
   "source": [
    "logistic_clf = LogisticRegression()\n",
    "logistic_clf = logistic_clf.fit(final_train, y_train)\n",
    "predict = logistic_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.497461928934\n"
     ]
    }
   ],
   "source": [
    "svm_clf = SVC()\n",
    "svm_clf = svm_clf.fit(final_train, y_train)\n",
    "predict = svm_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.565989847716\n"
     ]
    }
   ],
   "source": [
    "linear_svm_clf = LinearSVC()\n",
    "linear_svm_clf = linear_svm_clf.fit(final_train, y_train)\n",
    "predict = linear_svm_clf.predict(final_test)\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "796 779\n"
     ]
    }
   ],
   "source": [
    "count = 0\n",
    "for i in range(len(y_train)):\n",
    "    if y_train[i] == 'sarc':\n",
    "        count += 1\n",
    "print(count, len(y_train) - count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "c_range = np.linspace(0.001, 0.1, num=10)\n",
    "c_range = np.append(c_range, np.linspace(0.1, 2, num=20))\n",
    "c_range = np.append(c_range, np.linspace(2, 10, num=30))\n",
    "acc = np.array([])\n",
    "for i in range(len(c_range)):\n",
    "    logistic_clf = LogisticRegression(C=c_range[i])\n",
    "    logistic_clf = logistic_clf.fit(final_train, y_train)\n",
    "    predict = logistic_clf.predict(final_test)\n",
    "    acc = np.append(acc, [accuracy_score(y_test, predict), c_range[i]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.598984771574\n",
      "0.6\n"
     ]
    }
   ],
   "source": [
    "maxv = 0\n",
    "c = 0\n",
    "for i in range(0, len(acc), 2):\n",
    "    if maxv < acc[i]:\n",
    "        maxv = acc[i]\n",
    "        c = acc[i+1]\n",
    "print(maxv)\n",
    "print(c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.581218274112\n"
     ]
    }
   ],
   "source": [
    "print(acc[22])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.621827411168\n"
     ]
    }
   ],
   "source": [
    "from sklearn.naive_bayes import GaussianNB, BernoulliNB\n",
    "gs_clf = BernoulliNB()\n",
    "gs_clf = gs_clf.fit(final_train.toarray(), y_train)\n",
    "predict = gs_clf.predict(final_test.toarray())\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.532994923858\n"
     ]
    }
   ],
   "source": [
    "gs_clf = GaussianNB()\n",
    "gs_clf = gs_clf.fit(final_train.toarray(), y_train)\n",
    "predict = gs_clf.predict(final_test.toarray())\n",
    "print(accuracy_score(y_test, predict))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "c_range = np.array([1.2, 2, 3, 4, 5, 6])\n",
    "gamma_range = np.array([0.9, 1.3])\n",
    "acc = np.array([])\n",
    "for i in range(len(c_range)):\n",
    "    for j in range(len(gamma_range)):\n",
    "        svm_clf = SVC(C=c_range[i], gamma=gamma_range[j])\n",
    "        svm_clf = svm_clf.fit(final_train, y_train)\n",
    "        predict = svm_clf.predict(final_test)\n",
    "        acc = np.append(acc, [accuracy_score(y_test, predict), c_range[i], gamma_range[j]])\n",
    "        if accuracy_score(y_test, predict) >= 0.60:\n",
    "            print([accuracy_score(y_test, predict), c_range[i], gamma_range[j]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.560913705584\n",
      "1.2\n",
      "0.9\n"
     ]
    }
   ],
   "source": [
    "maxv = 0\n",
    "c = 0\n",
    "gamma = 0\n",
    "for i in range(0, len(acc), 3):\n",
    "    if maxv < acc[i]:\n",
    "        maxv = acc[i]\n",
    "        c = acc[i+1]\n",
    "        gamma = acc[i+2]\n",
    "print(maxv)\n",
    "print(c)\n",
    "print(gamma)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "c_range = np.linspace(0.001, 0.1, num=10)\n",
    "c_range = np.append(c_range, np.linspace(0.1, 2, num=20))\n",
    "c_range = np.append(c_range, np.linspace(2, 10, num=30))\n",
    "acc = np.array([])\n",
    "for i in range(len(c_range)):\n",
    "    lsvm_clf = LinearSVC(C=c_range[i])\n",
    "    lsvm_clf = lsvm_clf.fit(final_train, y_train)\n",
    "    predict = lsvm_clf.predict(final_test)\n",
    "    acc = np.append(acc, [accuracy_score(y_test, predict), c_range[i]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.598984771574\n",
      "0.067\n"
     ]
    }
   ],
   "source": [
    "maxv = 0\n",
    "c = 0\n",
    "for i in range(0, len(acc), 2):\n",
    "    if maxv < acc[i]:\n",
    "        maxv = acc[i]\n",
    "        c = acc[i+1]\n",
    "print(maxv)\n",
    "print(c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 0, ..., 1, 0, 0])"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_plt = np.array([])\n",
    "for i in range(len(y_train)):\n",
    "    if y_train[i] == 'sarc':\n",
    "        y_plt = np.append(y_plt, 0)\n",
    "    else:\n",
    "        y_plt = np.append(y_plt, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1575,)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_plt.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1575, 2000)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb580a94a20>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.scatter(final_train.toarray().T[0], final_train.toarray().T[1], y_plt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngcFxHMep4n8yAwVqJMjMjgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb580a98400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb580980a90>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.scatter(final_train.toarray().T[2], final_train.toarray().T[3], y_plt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd4VFX6wPHvSQ9JSAIhlITQey+CIEhHxIK69nXVtaC7\n6s+6Ytm169rWtuoq9oq9gCgdlE5ChwRISEIK6T2ZTD+/P2YImRRSJhXez/PkYebec+89Mwnzzj3t\nVVprhBBCiBM8WrsCQggh2hYJDEIIIVxIYBBCCOFCAoMQQggXEhiEEEK4kMAghBDChQQGIYQQLiQw\nCCGEcCGBQQghhAuv1q5AY4SFhenevXu3djWEEKJd2blzZ67Wuktd5dplYOjduzcxMTGtXQ0hhGhX\nlFLH6lNOmpKEEEK4kMAghBDChQQGIYQQLiQwCCGEcCGBQQghhIsmCQxKqQ+VUtlKqQO17FdKqTeU\nUglKqX1KqbGV9t2glIp3/tzQFPURQgjReE11x/AxMO8U+88HBjh/FgL/A1BKdQIeByYCE4DHlVKh\nTVQnIYQQjdAkgUFr/QeQf4oiC4BPtcM2IEQp1R04D1ittc7XWhcAqzl1gBGCPamFvPfHUWz2tpWW\ndsmOFG75JBqz1d7aVRHCLS3VxxABpFZ6nubcVtv2apRSC5VSMUqpmJycnGarqGj7nvkllmd/PURc\nRnFrV8XF07/EsiYumzWxWa1dFSHc0m5mPmutFwOLAcaPH9+2viqKFvXspSPYm1rI0O4dW7sqLl6+\nfBTrj2Qzb3jX1q6KEG5pqcCQDvSs9DzSuS0dmF5l+4YWqpNopwZ1C2JQt6DWrkY180d2Z/7I7q1d\nDSHc1lJNSUuB652jk84GirTWGcBKYK5SKtTZ6TzXuU0IIUQraZI7BqXUEhzf/MOUUmk4Rhp5A2it\n3wF+BeYDCYAB+KtzX75S6mkg2nmqp7TWp+rEFkII0cyaJDBora+pY78G7qhl34fAh01RDyGEEO6T\nmc9CCCFcSGAQQgjhQgKDEEIIFxIYhBBCuJDAIIQQwoUEBiGEEC4kMAghhHAhgUEIIYQLCQxCCCFc\nSGAQQgjhQgKDEEIIFxIYhBBCuJDAIIQQwoUEBiHagI3xOcx/fSMJ2SWtXRUhJDAI0RbsOlZAbEYx\nCdmlrV0VIZosUc884HXAE3hfa/18lf2vAjOcTzsA4VrrEOc+G7DfuS9Fa31xU9RJiPbkzpkDuHh0\nBH3CAlq7KkK4HxiUUp7AW8AcIA2IVkot1VrHniijtb63Uvm7gDGVTlGutR7tbj2EaM88PZQEBdFm\nNEVT0gQgQWudqLU2A18BC05R/hpgSRNcVwghRDNoisAQAaRWep7m3FaNUqoX0AdYV2mzn1IqRim1\nTSl1SRPURwghhBuapI+hAa4GvtNa2ypt66W1TldK9QXWKaX2a62PVj1QKbUQWAgQFRXVMrUVQogz\nUFPcMaQDPSs9j3Ruq8nVVGlG0lqnO/9NBDbg2v9QudxirfV4rfX4Ll26uFtnIYQQtWiKwBANDFBK\n9VFK+eD48F9atZBSajAQCmyttC1UKeXrfBwGnAPEVj1WCCFEy3G7KUlrbVVK3QmsxDFc9UOt9UGl\n1FNAjNb6RJC4GvhKa60rHT4EeFcpZccRpJ6vPJpJCCFEy1Oun9Ptw/jx43VMTExrV0M0wL9/jeOH\n3eksv2sK4R39Wrs6zeLpX2JZtvc4v949lbBA39auToUnlh7ktwMZ/Pp/U+nchuolWp5SaqfWenxd\n5WTms2gRuaVm8kpNmG321q5Ks8ktNZFXZsbSxl5jXqmJvFIzVnv7+xIoWofcMYgWobXGYtP4eJ2+\n30Xa6mvUWmO1a7w921a9RMur7x1DSw9XFWcopRQ+Xqq1q9Gs2uprVErh7dn26iXaLvkKIc5YWmva\n4x2zEM1NAoM4I9ntmnNeWMcFb2yq9zH5ZWaik/ObsVZCtA0SGMQZy8/bE98G9Afc980ernhnK/vT\nipqxVkK0PuljEGckDw/FuvunN+iY6yZG0dHPm75dZBVUcXqTwCBEPc0e2o3ZQ7u1djWEaHbSlCRE\nE7HbNQaztdmvU2Zq/mu0B/I+NB8JDEI0kbuW7GLYYytJLyxvtmusic1i2OMr+WLbsWa7Rnvwvw0J\nDHt8JVuO5rZ2VU5LEhiEaCJBvt6EBnjj14wT3EIDvAkL9DltlxWpr+7B/nQJ8iXY37u1q3Jakj4G\nIZrI4ewS8sssZJeYmm1NonG9OhHzzznNcu725JIxEVwypsZ8YKIJSGAQook8Mn8I2xPzGBAe2NpV\nEcItEhiEaCJn9e7EWb07tXY1hHCb9DEIIYRwIYFBCCGEiyYJDEqpeUqpw0qpBKXUQzXsv1EplaOU\n2uP8uaXSvhuUUvHOnxuaoj5CCCEaz+0+BqWUJ/AWMAdIA6KVUktrSNH5tdb6zirHdgIeB8YDGtjp\nPLbA3XoJIYRonKa4Y5gAJGitE7XWZuArYEE9jz0PWK21zncGg9XAvCaok2jH8svMbDySg8lqa+2q\n1Mlu18QeL8Zu1yz+/Qhz/rOh1rLH8sooNlowWqysOJDRqExvWmviMopJzTeQV2qq2HYosxizteHn\nyy8z1zohr8hgITXfAECpyUpSblmD63oos7jW11litJDcwHOKltEUgSECSK30PM25rao/KaX2KaW+\nU0r1bOCx4gwy77U/+MuHO3hi6cHWrkqdvopOZf4bG/ls2zGe+y2e+JyyGoNDZpGRaS9t4PoPdrDw\n053c/vkuHvhmb4Ov99OedM5/fSNTX1zPxW9uBmD94WzmvbaRF1ccavD5Fry5iSnPr6txKY+rFm9l\n6ovrySs1sfDTGGa8vKFBH+Rr4hz1ennl4Rr33/xJDNNf3lARfETb0VLDVZcBS7TWJqXUbcAnwMyG\nnEAptRBYCBAVFdX0NRRtxrzh3Vi25zizBoW3dlXqNKpnMGOiQhgTFVKx7YrxPaqVC+ngzfSBXTin\nfxgBvp7sPFbA7KFdK/ZnFhmx2Oz07NThlNcb3iOYMT1D6ODjyfCIYAAGhAcxLiqUcwd2aXD9LxzZ\ng2P5Zfh5edawrzsRIf4E+Xlz/vBueChFl6D6T9wb1DWIsVEhTB1Qc73mD++Gn5cHnQN9Glxv0bzc\nzvmslJoEPKG1Ps/5/GEArfW/aynvCeRrrYOVUtcA07XWtzn3vQts0FovOdU1JeezON2MfXo1ReUW\nDj89Dy/JzSyaSX1zPjfFX2A0MEAp1Ucp5QNcDSytUpnulZ5eDMQ5H68E5iqlQpVSocBc5zYhzihX\njIvkyvGReHpIbmbR+txuStJaW5VSd+L4QPcEPtRaH1RKPQXEaK2XAv+nlLoYsAL5wI3OY/OVUk/j\nCC4AT2mtJXeiaJfySk2kFZQzqmdI3YWreHj+kGaokRCN43ZTUmuQpiTRFl357lZ2JOWz5r5z6R8e\n1NrVEaKalmxKEkIAf54YxUUjuxMZWr0D2Wqz8/jSA/yy93iDzvnq6iN8vCWpqaooRL3IInrijKe1\nRin32/YXjI5gweiaR1tnFhv5ZMsxdiTlc+Go6qOWamK22nl9bTwhHby5cXIft+vX3JrqfRStT+4Y\nxBnt4R/2MeSxFWQVG90+1+aEXF5ccajGCV2RoR34auHZLP5LnXfx2O2a11Yf4Y8j2Sy98xy+vW2S\n23Vrbr8fyaHfI7+ydE96a1dFNAEJDOKM5qEUXh4enPiem19mZnVsFna7o+9ty9FcUvLqNwHr3q/3\n8PaGoxzNKa22b0dSPuFBvnXOUwDH3cVra+N5fsUhRkaGMKBr6/dX7Eop4EhWScXz3FITayq9Tx4K\nPD1UnXcMx/LKmiQdZ1G5hZUHM7E2Yva4qJsEBnFGe/bSERx48ryKVJlP/xLLrZ/G8Ht8DmkFBq59\nbzu3fla/gQ7ZJY4lKl6pMtM3r9TEle9u5caPoms6rJoeIf58dONZvHPduAa8kuZjtNi47O0tXLN4\nW8W2x5ce5JZPY9hyNA+AqQO6EP/sfC6qo5nslk9juPa97Rx3My/2a6uPcNtnO/n1QKZb5xE1kz4G\n0aJS8gz4eXu02ZzFf5nUCx8vD8ZGhRLo68XNU/owLiq0Xsd27+hDRrGZR6oMPQ3t4MPt0/oyqAHf\n/GcMbjuzvv28Pbl39gCXWc83Tu5NkK8Xo3oGN+hc98wawO7UQrq6+fu/fHwkZWYrk/t1dus8omYy\nXFW0GKPFxpB/raBbsB9bH57VLNcoM1m55K3NTO7XmScXDG/w8Td9HE2pycorV45iT2oh84d3x6Oe\nk87Ofm4NmcUmfvr7ZEbXM5gI0ZLqO1xV7hhEi/H18uCysZF0D26+uwWT1U5iblmD1vSp7EhWCSVG\nK4/+uJ/fj+Ty7e1+9U7XWWCwAJBRZGR0Pa9XVG7B18sDP+/qaxUJ0VrkjkG0iMScUqKT87FruGBk\ndzr6ebt9Tq01S/ceZ3hEMP26BFZsLzNZ8fXyaNSaQyarDa3h4PEiVh7M4r45A+v9oV1utpFRVE7f\nSnVJzTcQnZzPJaMjqt15lJqsjH5yFf3DA1lxz7k1vrYREcEu56u6v+prb0/WxGay5Wge/7xgCB4e\n0t3ZEuSOQbQpD32/nx3JjtVOisot3D6tn9vnjMso4e6v9hDawZu7Zg7gpimOsf4Bvo3/s/Z1rjI6\nrlcnxvWq353CCf4+ntU+xB9fepB1h7KJCPFnYl/X9nAfTw8iQv05lFnC0r3HubhSx+3B48Xc/dUe\nzuodyre3T652rdgMx/7xvUL57m+O/Uv3HOeDTYksvn682234LeHOJbsxWuxMGdCFmW2oT0XIqCTR\nQh6cN4hbp/bh1ql9WDC6fhO86jKwayC3T+tLgcHCV9EpTXLOpnb3rAH838z+Na6f5OPlwTOXDCc8\nyJfsIiMXv7mpYqjroG5B3Dt7AA/OG1zjeQd1deyfMTiczCLHHIyNCTnsTSsiraB95De4f+5Azunf\nmSn9w1q7KqIKaUoS7V58VgkhHXwa3a9Q2efbjlFusXLrVPfvaBri1dVHeH1tPG//eSzzR3Sv+wAg\nJjmfy9/ZyoxBXfjorxMwW+0cLyynd1hAM9dWtFeyVpI47b2/MZH/W7KbPmEBNQYFo8XGws9i+LoB\ndxPP/RrHs8sPVUzcail3zezP6nvP5fzh3Vy23/f1Hs579Y8a05wO6xHMDZN6VTTL+Xh54OvtwZYE\n9yeQiTObBAbRZlhtdj7enERcRnG9yn8VncrSvccpKrfUuD+zyMiqg1l8vzOt3nX4euEkvv/bZJeO\n4s0Jufy0u+6lHn7clc6dX+7C2ojcy16eHgzoGlRt5nBSXhnJeWVYbdUDlb+PJ08uGO7Sd3Hnl7u5\n9v3txGeVsOFwNssauGifECCdz6IN2ZdexBPLYpncrzPv/mUcgb5ep1xiYcmtZ1NUbqZz4Mm7BavN\nzr9+PsC4XqFcPq4nK+6ZSrcGdMSOiKw+Yeuer/aQU2pi1pBwgk4xmurRn/ZhMNuZN6xbvRfKq8u3\nt03Catf4elX/Dmex2bHZtcuoqdun9WXD4RyiOnfg8ne2UlRuYe6wrhWd6pWVmqwE+HhyOKuE8CA/\nOgVUT7GptabUZD3l6xanHwkMos0YFRnCUwuGEejrxYgnVnHXzP7cP3dQreW7BPlWa0LKKTWxZEcq\nu1IKuXxcTwZ36+h2vV6/ejQ5pSb86xi26u3hAdgJ7+h+X8cJXp4eWGw2xj29mmERwXx288SKfRe/\nuYnkXAO7H5tTERzmDO3GnKGO5qj/XjOGYqMFH08PbHbtkh1u57F8/vS/rVw/KYpPt6YwKjKYn++c\nUu36i/9I5N+/HeKzmyfUmrtZnH6apClJKTVPKXVYKZWglHqohv33KaVilVL7lFJrlVK9Ku2zKaX2\nOH+WVj1WnDk8PRTXT+rNyMhgunb0JTLUv8Hn6B7sz49/n8zHfz2ryerVI8Sf6OR8+j/6G3tTCwHH\nCqiHM0uoPHijf9dAvDyUyx1MTbTWDVr8TSlHgPCqMg+ib1gg/cMDq20/4dyBXbhwZA/u+XoPg//1\nm8sKssH+3nTt6MugrkHMH9Gdq87qWeM5IkL8CQ/yJbRD9buJyv710wGmv7SeEmPNzXqifXF7VJJS\nyhM4AswB0nCk6bxGax1bqcwMYLvW2qCU+hswXWt9lXNfqda6QTN0ZFSSaEln/3stmUVGAn29+Pb2\nSQzp3pGPNifx5LJYXvzTSK6s5UO1Nvd8tZvl+zPYtGhmi8w3+NdPB/jtQAYr7zm3zqDVWH/9aAcb\n43PZ/sgst67x0spDhHTw4dapfZuwduKElhyVNAFI0Fonaq3NwFfAgsoFtNbrtdYnBldvAyKb4LpC\ntIg7pvfj1ql92P/EXIZ0dzRNje4ZwqjIYIZFnGyqWvhpDCMeX0FhmeOb+dfRKdz08Q5ikvOx2zVv\nrovnjyM5dPT3Jtjf26Vp51T+OJLDm+viK0ZK2e2a/66NZ2N8DgA2u+ah7/exZHvNo6+evmQ4Mf+c\n0yxB4duYVL6OTuX5P43k+km9MJirj56qzZaEXF5fE4/N+bosNjtvrT/KW+sSmryep4ON8a5/B82p\nKfoYIoDUSs/TgIm1lAW4Gfit0nM/pVQMYAWe11r/VNNBSqmFwEKAqKgotyosREP8ZVLvatvGRIVW\na5NfFZsFwJvrjvLPi4bxr58OYLZpDGYbL10+ipdXHWFg10BW3TuNpxqwwN/Ty2OJzyplwegIenbq\nQGqBgf+sPsLgbkGsuKcLeWUmvopOZeexAib27cTDP+7nwfMGM66Xewv5FRst/HdtPH8aF1lrX80j\nP+7HZtc8cdFQPtycTEc/b+6ZM7Be53/utzgOpBdzwcju9A8PxNvTg+X/N6XGjnIBzyyP43BmCReP\niiCqc915PdzRop3PSqnrgPHAtEqbe2mt05VSfYF1Sqn9WuujVY/VWi8GFoOjKalFKizEKZSZrHh7\nelBusRHk68V1E6NYE5fJonmOD8bPb5nIH0dyuHxcT3p26sA7142lT1jD1zV685qxJOWWVST56dU5\ngHeuG1ux/EZ4kB/L7pxCWJAP2xPz2Z6Yz6b4nFoDw1WLt2Kx2vn+b5NPOeprS0Iu721MoqjcwouX\nj6qxzOc3T8SuYVTPYPy8vThvWLcay9XktatGcySrlH5dTk7IG9ajYct4n0neuHoMSbmlzR4UoGkC\nQzpQuZE10rnNhVJqNvAoME1rbTqxXWud7vw3USm1ARgDVAsMQrQlpSYrY55aRZ+wAI5klfKXs6NY\nti+DonILmxMLmD4onAl9OjOhz8k5BvOG129Gc1WDugUxqFsQB48Xcc9Xe3j8omEu53pxRRyLf0/k\nhzvOYcHoHvTtElDR5GW22vGpMtQ1vaAcUz3mWswa0pVXrxrFlP61j0aqPIeioX0t/cOD6B/e+tnp\n2osTfwctoSn6GKKBAUqpPkopH+BqwGV0kVJqDPAucLHWOrvS9lCllK/zcRhwDhCLEG2ct6diYNcg\n+ocHEhnqT78ugcwZ0pVOHbwZHtE833qTcw3EZ5cSm1Hksv2TLcewavhi2zGUUoyMDMHb04Pnfo1j\n4D9/43BmiUv59Q9MZ8tDM+tMw+nt6cGlYyKbZKkR0b64HRi01lbgTmAlEAd8o7U+qJR6Sil1sbPY\nS0Ag8G2VYalDgBil1F5gPY4+BgkMotkczizhgW/3Viw8V9kba+L52xc7eWt93Z2fqw5mcXbfzvz3\nmrFsWjSTG8/pg9Fixa419elSttjs/LQ7ndxSx83zM8sO0v+R5SzbW/MM6/c3JpKSX8amRTOqjdi5\naUpvPBSEVepcLjFa2JaYR5CvZ7XJcd6eHpSZrDz43V72OIffVqa15vnf4vh+Z2q1fW2d1pqXVx5q\ns4sqthdNMo9Ba/2r1nqg1rqf1vpZ57bHtNZLnY9na627aq1HO38udm7forUeobUe5fz3g6aojzi9\nHcsr47f9GS5zCOrr5z3pfLczjT+cI3pOMFvtvLLmCL/tz+TllYfrHPnxxtp4PtiURHbJyQDz64FM\nCsut9Vo+Y9XBLO75eg+vrD4CwBc7UrDa4bXV8RVlLDY7Mcn52Oya/6w6wksrDxMR4l/tm/6Q7sHY\nNRRVmkOwN7WIfWlFnDsovMZF9aKTC/gmJo0vtx+rtq/AYOGd3xN5bW37Gx1UZrbx5vqjvLrqSGtX\npV2T1VVFu3PFO1uITi5g+f9NaXBnZanJyuaEXGYODse7SiKfPamFZBSWExHqz8jI6stkV3Ysr4yM\nIiNnV2pjn/7SOpLzyllxz9QaR/HsTS3k/Y2J/OvCofh6efLamsMM7t6RC0b24GhWCU8si+W9G8YR\nFuiY2/D+xkSeWR7H05cMZ1xUKDa7rnHJDptdszYui4l9OxPs71i6QmvNukPZjO4ZUuMwVbtds/ZQ\nNmf1DiWkhslr0cn5dA7wqTFJUFu381gBwf5e0n9RA0nUI9yWWWTkSFYJUweE1dke3ZLumzOIjfE5\nDOzasP/45WYbd36xixmVgkJGUTkJ2aVMHdCF0T1DGF1D3oQTbHbHh+2E3p3o1TmAXp0d38RLjBa2\nHs3DYjs5z6AmP+85zrJ9GZw3vBsXjuzBWX068/cvdrE7pZDn/zSSH+84x6X85H5hTO0fxsTeofQP\nD3JZ2G/L0Vx6hnagZ6cOeHoo5lYZDaSUYtaQrrW+Fg8PxZyhte+vms60xGjB2/NkCtKjOaUUGszk\nl1mwa92g0UjNzd1hukICgziFe7/ew9bEPH65a0qzdag2xqR+nZnUr3PdBavILTWx4UgOBrOVGyb3\nBuD/luwmOrmg1m/5la2OzeL2z3dy7YQonrtsRMX2h77fx/L9mRXtsvHZpQyt4U7mvrkDOXdgWMWa\nQ0m5jqQ8tcXcoT068tktE7lm8TYOHi9i+yOz8ffxJDXfwLXvbWdo9478evfUhr0JOO4mjhcZ6RHs\nV6+Ab7LaGP/MGiJC/Vl65xTyS8385YPtHC804uWhsGvN0efmt6kvD8I9EhhErW47ty99uwTQP7z9\nNScAHC8s558/HeCOGf0Y16sTPTt1YN390wirNMrmb9P6MahbDn2qtMPHZ5Xw3K9xPDhvcMXQzwl9\nOnHFuEiuHO86cT88yNH04+UBZjt0CfAhOjmf/204ynOXjqBbsGN/oK8X0wedTGF58agIEnPKuKWW\n5R+eXR6Lr5cnAb6eBPp5VQSQ7sF+XDiyO+f0q1/msx92pbHuUDYvXzEKP29Pft5znHu+3sPTC4bV\nOHmvKi8PD0ZHhRAR7M8tn0SzLTGfh84fTJHBwtheIYA6Y4KC0WLjH9/uZcbgcC4be/ou4CCBQdRq\n7aFsvt+Vxh0z+tMjpOEL2rW23SmFrDuUTb8uARX5m6u2mc8c0pWZlZpctNa8t9GxoqjWjruTE4Eh\n2N+bSf06c90H25k6oAv/u24cAA/MHURemYllezMASCkwcCSrlHWHstl1rID5I2uev9CzUwf+c+Xo\nGvdprflgUxL+3p4cfGqey77MYiO/7MsgOa+MaybWvQrAkh0pRCcXcN+cgfTtEkifsAD6dglgQD2b\n4jw9FF8vnMSKA5lsTsh1LnbYiw4+dX985JWa6BTgc9oEjowiI8v2ZZBRZJTAIM4sRouNozmleCqF\np1K1NnW0decP78Y3t01iZA0dtrW55+s9/LznON6eiscvGsaV409O2lodm8l93+zFx1Ph5XnyTTmc\nXcLSvRl4KbBq6NUpgAWjI1kTl8UTyw4wb3g3l/6B2pisNuKzShnWoyNKKe6fO4iXVh7mp93pXDIm\noqJceJAfl42NYHI9m9PeuW4cxwuN9Ajx56Hv93HesK6su396vd+TE3W7/fOdBPt7s/fxufU6ZmN8\nDn/5YAf3zh7A3bPrt0xGW9cnLIBld06hR0jzL37YmiQwiGqeXBbLkh0pfH7zRJ5YMKy1q9NoHh6K\nCX061V2wkm4d/ege7Meyu6a4zAsAmNinM1ef1ZM/T+zlMjpoTM8Q3r9hHLuPFXLgeBFje4Xi6+1J\nlyA/igw1L0P90+40nloWy6tXj2HaQEefw/O/HeKjzcl8cMN4Zg3pypioECJD/Imosvy4j5cHr9Ry\np1GTAoOFlHwDoPkqOpXUAgMzBlfveE7JM7DuUBbXTIyqtl6Rr5cnr101mo7+9f/I6NrRj56h/u22\nKbI2NY0MO93IcFVRzfpDWSz+I4nXrh7dIstCnw5eXHGItzc4VnJZftcUhlXqrF++7zivrD7Ce9eP\nr2jKGvrYCgxmG49fNIS/nuPoY9gUn8t/18fzypWjiail6W75vgxCA7yZXM/+BYB5r/3BocwSVtw9\nhbwyCwPCAwl3/l63J+bx3sZEbpnShw82JbE6Lpv3rx/P7FOMWKpJsdHCN9GpXDImolpAFW1HSy67\nLU4zMwZ3ZcnCs1kXl83Vi7dS3IjkKzuS8rnkrc3EZ5XUXbgB3lqXwF8/2oG5AXmVbXbNbZ/F8NLK\nQw26VpHBzJinVvHXj3bUWXZ4pVFI8dmO1/zoj/u5/5s9HEgv4mhOGRf9dxOvr43HaLFhMNvw9fLg\nxsl9AMcQ1z5dAvh64aRag4LBbOWOL3dx15e7Xban5hu49O3NvL8xkf+ui6fQYCY138Dl/9vC8n3H\nsVQkBVKc0z+sIigALNt3nDVx2Vz93nZijhVw96wBvPP7UX7Y5ciTvTkhl0ve2kxSblmNdXr4h308\n+N1eftqdzjPL4/hsa/UJc6L9kaYkUavVsVlsS8wnq8hIxwbm/I1OzmNPaiGxGcX17uSsj18PZHDw\neDHFRku9v5mWW2ysPJjFkaxS/nHe4FOWvfx/Wyg1Wfnt7qmk5BsoMFjYnpRf5zXmj+wOXzoeH0gv\n4pIxkfy0Ox2T1c6hp+YRnZxPzLFCVuzP4O5ZA3jtqtGEdPCu6JR9e0MCL686wuK/jKs2J+GEDj5e\nvHLlqGoT1uKzS9idUkhGoZHMYiNvrU8g0M+L3BIzvTp34ImLh7ElIa/GJp1H5g/hwhHd+e1AJuEd\n/Zg2sAuvr40nvKMfl42NZMvRXPakFnI4s6TayC2AX/ZlYLNrHj5/MCVGK5ePO307ZM8kEhhErd78\n8xiyik2rijldAAAgAElEQVQ1fiDUpXenAIL8vOgR3LRNUV/eejbF5fUPCuAYJrr5oZkE+NS9zn9h\nuYXicgtaw9bEPAAeu3BIva5z1fhIVsdlcdesgUQn52Ox2bDaHX0dn948kW9j0iqGmK4/nI2Xh6oY\nvjo8Ipj+4ScnzZ2gtcZm13g5J+TVNBJm5uCurLr3XLTWHDxezLcxqQR38OGBuYOIDPXHz9uz1nzN\nHXy8OLtfGGdXappae/+0iruWe2cP5OJREQzsWnM/wdr7HSvohwb4cseM/vV6n0TbJ4FB1KqDjxd9\nwhr3J1JotFBitFJqqn9Gr/oIdmY/q68fdqURGdqh3p3Qq+45F43jw3xAeBBRnTowPCKErUfzyCo2\nuowOqupAehH5ZRY2J+Tw9y92M7R7EIvOH4KHh6LAYOHxpQcZ1qMjv9w1hTWxWXh5KB77+QCXjIlg\n+qBwpg8KZ0dSPou+28dlYyNIyCllT0ohP+xO5/d/TCcytPZ1+E/MAh/UrWOtwyhtdo3WJ4NMZQaz\ntWL4ab9KQ3q9PD1OudTziTkc7UmZycqSHSnMH9G9XQ7DbgkSGESz+PPEXvxpbGTFEgqtIa/UxH3f\n7CUyxJ9ND82s1zGVh5XOGBzOjMGOb/RnP7eWzGIjMwaFE9zBNTBprbnlkxiS8x3Za19YcYgREcHc\nOaN/xYij8CBfrjqrJxN6h6KUYuOimWxPzONvX+xi/aEsVt83HT9vTx78bi/JeQa2JuWRkmfgT+Mi\nCPH3xtvTA5tdc+NHO4jq1IFnLx1BQ81/YyNZxUZiHp3tEhw+23qMf/18oGI01Olu5cFMnlkeR1pB\nOU9c3H5H3TUnCQyi2bRGUMgvNXPOC+sY3C2IH+84h5cuH3nKb9pV3fnlLspMNj68cbzLpKxXrhxF\nRpGxWlAAxzfxzUdzMVocnbzH8so5f3gw5w0/2Vfg7enBC38aWfG8wGDm3785OsNTC4y8+/tRftpz\nnKxix2qtI7oHceuUPvx5Yi8Sc0sJ9PXCYrOz9WgeaQXlLtdPyC7l1k9j+MfcQbVOpgPoGuSLh6La\nZLPwjr50DvCpcTG909HcYd14+HwTF47q0dpVabMkMIjTSm6ZiXKLrWJk0BXjG5ZVLDo5n1KjFa1d\n1zCa3P9kG3xCdik3fxLNA3MHcdGoHnh5erDt4VmMfmo1AFeOi+CZS0dWPbWLzCIjKfkGxvcKYXhE\nCB5KkZRbxtMLhnEks4Sbp/alR4g/S3ak8OhPBwj09WLv43OJfnR2tYxsmUVGknLLiM0oPmVg+PTm\niRzLK8NsteNfqb/lvGHdKhbBs9k1Sbml9OsSeNrMVq4q0NeL26b1a+1qtGlNMlxVKTVPKXVYKZWg\nlHqohv2+Sqmvnfu3K6V6V9r3sHP7YaXUeU1RH3HmGtg1iLX3T2Pzovo1HVW19v7pbH1k1ilnKmcX\nGzmWZ3DJjBbSwYeFU/rQPdiPRy8YVu3Du6pz+oex45FZfHHL2aw4kMEf8Tlsf2QWQX7efLY9hbiM\nYj7fdoxHfzqAh3IsF34gvYjQAB8CfF2/z00ZEMb2R2Zx35xTzy4+klXCtJc2cNeSXbWWeWt9ArNf\n+YNf92ee8lxVZRYZWXEgs1E5MkTb43ZgUEp5Am8B5wNDgWuUUkOrFLsZKNBa9wdeBV5wHjsURyrQ\nYcA84G3n+YRotH5dAgluZLNIoK9XnUNzJ/cPI/rR2dw/1/WD+JELh7L14Vk1NjdVdmKGcXhHPzw8\nHAvQKRwzhU90rq84mMlHm5OYPSScGyf3xttD4e3pQV6piRd+O0Sqsz/jhK7Oc51K1yA/zuodyqwa\nZj2fML53KCMighnUrWGzlR/+YR+3f76TmGMFDTpOtFFaa7d+gEnAykrPHwYerlJmJTDJ+dgLyAVU\n1bKVy53qZ9y4cVqIhsovNemPNiXqQoO5Vetx8X836l6LftHxWcX6ky1J+sp3tujictc63fpptO61\n6BedVmBw2f75tmTda9Ev+uWVh5qsPvFZxbqgzFSvsgfSC/VD3+/VeaWu5bcn5uknfj6gDSZrk9VL\nND0gRtfjc70p+hgigMrJYdOAibWV0VpblVJFQGfn9m1Vjq19PKCbpr6wjtSCcj66cXyNa8U0l98P\nZ3Pjx9HMH96Nt/48rl7HpOQZuPydLfz1nN58uzONcVGhvHTFqAZfu89Dy9HAtwvP5vYvdvHERcO4\naHTdnW6xx4uZ/8ZGAJ67dBjXTuzdoOuuPJDBIz8eYPH142tMnDLmqVWYrHZunNyLtzckMn1gGA+c\nN5iL39yEpwKLHfy9FHHPzAcgrcAxk/fGyX14Y108BrONxdePY+7Qkx28//huL0t3p2OyaQJ8Pejo\n58PVZ/VkwageTP/P7xXlXlx5mM9vnsBTv8SRWVROdomJB88bxPL9mXQJ8uXDG89iV0oBt34Sw7OX\nDmdUzxAueXMzZWYrVrvmyYuHsSe1kB93pWO02hkZ0ZHCcqtzPSIHb09FpwAfIkL8eWT+EG76aDvF\nppOztRVw39d7MVttHMoqZcQTqwAI9vPEw8OD4nILXYO8mfrCOuwaQjt4Y7HaKTU7hv8mZJcw4NFf\n6dbRj6yicsz2k+dVgB34zxUj+XBzMoczS1hy60R6dQ7gkrc2c8X4SGx2zZvrj1bUx0PBkO4duX1a\nX/71434KjTY6+Hhy+bgIPtuawitXjubSsRF8tSOVJTtSOad/GBeO7EHvh5YDkPz8BRVDgm/9NIbj\nheX8+7IR3PRxNI/OH8KlNQyhPXHsBcO78tZ147ny3a0AfHPbJIoMFi58cyNzhnTlsYuqjxx69Id9\nfBmdymtXjWbB6IZ9ZBxIL+KGD3fw2IVDWXCK4cc1KSgzc9Gbmzh/eDcevaBqw0h1j/98kA1Hsll2\n15QGTxKtau6rv5OYU8bGB2fQvZmH2babzmel1EJgIUBUVN1LDdekwGAGILvY1GT1qo/cUhNaQ3ph\n9QT0tSk1WckuMZFWUE5qvoHwoMatP3OixTe31ERemZmskvrVoaTSMhiZRfWv9wnZJWbyyswUlZtr\nOb8Vm12T4XxPMouNFBrM2DWcSIBmsp1srzaYbWQWmzheWI7J4vhwzK5Sr9R8Q8UxRrOdMpORtMJy\n8spO1mHGoDDWH86lwGAhOa8Mo9mGXUNaYTnH8sowOs9dZLCQV2Ymp8SEwWwjq+Tk30xWkYnU/HKM\nzmU5sktM5JS6vk6LTVNosDj+LTe7BAVw/F6OF5Vz05Q+HFpxuGJ7mdmGxo5NQ1bJyd/BiffrhLwy\nMxabJqvYVBEUTpz3RKmcEjOZRUasdk2BwUJ4RxvHi4ykFpRTtStAa0jOLSOr2Eih0fEeGC020guM\naKgYLXX/3IFM7te5WvY3s9mMj4+j+e5IZgnH8g2sic0it9RMZvGp/34SnQmLjuWdXHbDZLWRXlBO\napURWCekFzpeQ3Y9/54rKyq3NOj/QmUmq53jheWk5tdcr6pS8w2kFZRjstjBzSkfuaVmrHZNiclK\n7UMMmkh9bitO9UM7akqyWq06Ld9Qd8FmkFNSrm02W4OOKTNZtN1u1+Vmq7ba7I2+7nFnc0SZydKg\nYwtLDfpAWmGjrlvX9cpMFl1S7tifkldasT2vxKRNZoveeCRLm0yuzRUGk9XxfpRbdEJWSbVzWm12\nXVxu1il5pdpgslaU11rrg+kFOqOwxKVeJotNG81WnVti1Ha7XRstVm22nvwdVa6/43wWXeJs8rHa\nHL+XlLxSbXM+7rXoF91r0S96f2qBLigz6XKzVZsstopzJeWU6C3x2XrFvnSdX2qsuFaZyaJfXhmn\nF/z3D22z2XWZyaKLnMcXlJl0RmGZNpgsusxo0ccLDDqzyPH7LDKYtNFi1TabXb+7/ojWWuvcEqMu\nNZh1Tkm51lprs9Xm0kxU+T05nFGki8rKHXU1WrTRYq2oT1aRoaJZ6MS5apKUXaKTsl1/F6sPZupe\ni37RX2w7dsq/gWM5hfr9349UPDdZbBXvl9Zal5sdr602p6pXXRr6f6GyuupV2Ym/jaZgsdjq3eRX\nG+rZlOT26qpKKS/gCDALSAeigWu11gcrlbkDGKG1vl0pdTVwmdb6SqXUMBwrzEwAegBrgQFa61NO\nl5XVVU9vdrvmj/gcxkSFNmiWs7tWHMggOrmAh88fXOPs4LpMe2k9x/IMLL3zHEZGnswdXVRu4ZEf\n9nPF+EiXDG4ux764nmP5BvY9MdftJocT4jKKyS8zc07/+q/E2lSMFlurTm4UNWux1VW11lbgThzf\n9uOAb7TWB5VSTymlLnYW+wDorJRKAO4DHnIeexD4BogFVgB31BUUxOlv3aFsbvwomud+jWvR676x\nNoEPNiVxvAFNfpX5OXMYeHm4/rc6klXC8v0ZPPT9fj7fllzjsd/ePok1901rsqAA8NePovnz+9sp\nKm/46rjukqDQvkk+BlGrJ5Ye5Ludaay+71y6B7fcmjJ5pSae+iWW6yf1qkjJ2RJS8gyk5BuYMsDx\nDdtosaEU1ZLW1Ca9sJyE7NKKZTBO0Frz6/4M7vhyNyMigll215QG1+3CNzZSbrGx5r5pp5x4llNi\nYtZ/NnDByO5M7htGSoGBv0/vV+sxBWVmZry8gZmDw3nlqvon/xHtU33vGNpN57NoeTa7xmKzV+uo\nbG6dA315/eoxLXtRIKpzB6I6O5bPsNs1Zz+3lkA/LzbVc7JcRIh/jbkUlFJcMLIH/cODCAts3PwK\ni01jsdX9i9DaUS6j0Mje9EIemDvolIFEAxabHau9/X1BFM1H7hjEKWmtW3VphOwSI99Ep3LNhKhq\neQhqk1Vs5JuYVK6b2IvQgMZ9EGutuWrxNgJ8vPjor2c16hxN6cT/0/r8LrTW/OWDHWxKyOWHv09m\nbFT14cJVy5+uy18IV5LBTTSJ1v7A+G5nGi+vOsKPu9MrtuWXmdmTWljrMV/tSOE/q46wdO/xRl9X\nKcU3t01yCQovrjhUYxa4vamFzHh5A1sScht9vfrUp76/i4wiI6N7hvDqlaMY0zOkzvJN/TvOLjaS\nkmeou6BosyQwtIDsYiNXvbuVNbENW39GwDVnRfHYhUO5YtzJxfDu+GIXl7y1mSO1pA297uxePHbh\nUC4d27RzJd/fmMQHm5KqbU/OKyMpt4zDmcUNSjlalc2usTrTcJqstgatO2Synhyz8f7GRN5cn4BN\nOz70tdZu1auhLn5zM9NeWl8xJ6Q2less2hYJDC0gIaeU7Un5rInLbu2qtDuhAT7cNKWPy/pD106M\n4oKR3elZy3LanQN9uWlKnyYd4QOw4p6p/Hb3udW2LxgdwZaHZrI7tZAhj60go6h+k5+qmv/GRsY/\nu4b9aYUM+ucKXlhRvxzVMcn5DPrnCv63IQGAW6b25eHzBzPPuez3XUt2M+SxFY2aqNgYV4yP5LKx\nkfieYiHBn3anM+ifK1hxQL4stUXS+dwCJvcL47e7pzYqRaao7qJRPbioFdbS79ul9oXleoT4Ex7k\nR1igD96NmAMBJ/Ml+Pt40SnAx6VPxWixsXxfBrOGhFfLm+Dv40lIB286OftTeoT4uywr3SXQly6B\nvnWu+NpU7p87qM4yHf29CPb3JshPPoLaIul8FqKJbEnIZVNCLvfOGdjo4FCbr3ak8NAP+/n79H48\nOG8w4Gh6+m5nKhP6dG62Lx2ObGexvHf9eAZ369gs1xAtRzqfhWhhL686zNsbjnI0x7H2z79+PsA/\nvt3bJOeePbQrN0/pw5WVEg/tPFbAou/38+TSg6c40j3x2SWk5peT0ULNUKJtkPs40e5YbXZKTdY2\nl4qyd+cOxGeV0NnZpLNsz3EMZhsv/GlknbkS6hIW6Mu/LnRdzXN0zxAemDuw1mU2msId0/tz9VlR\nhNUyVNhgtgLQwad9fZTklZroFODT6qPu2iq5YxDtzp1f7mb0U6vb3JDIrUfzKDHZKpLorLrvXP54\ncAYeHorsYiNvb0ig0FDzarON4ePlwZ0zBzA8IrjJzlmVUqrWoABw7ovrmfz8unaVuW3D4WzGPbOG\nN9cntHZV2iwJDKLdGR7RkQHhgc3acVlqsvLBpiSy61gyurK+XQLx9fIgspNj9nN4kB/dgh1rLX+x\nPYUXVxx2a25FbcrNNj7clER6YeNGQ9VkbVwW6w/VPYpudM8QxvQMbVffvLsF+9Gzkz8DwxuWpe5M\nIp3PQtRgyY4UHv5hP3+b1o9F5w+u1zF9H16OXcMj5w9mYZVk8zklJn7ek84V43s2+Yqxy/Ye564l\nu7l2QhTPXTaiQcdabHaeWHqQs/t2rhjppbWm3yO/4umhiH92fpPWVbQuWStJtGt2u2bp3uOMjQqt\nWL+oJc0f0Z0Cg5lLG5Dh684Z/Vkdl8WNk3pX29clyJdbpvat8xxaa276OJpOAT7858r6LWo3c3A4\ni+YN5oIRDU/fklFo5IvtKexNLawIDEop3rp2bLPfBbyxNp41cVl8cctEgpp4zolwjwQG0SbtSing\nnq/3MG1gFz65aUKLXz/Y35u/T+/foGMiO3Wge7A/VjSN6RaPyyjmiaUH2Z1acMp2/aoCfL342/R+\ndResQVTnDnx7+6Rqi/+d34Ag894fiYQF+TYoiAJEJ+ezL62IEqNVAkMbI4FBtEkjIoO5a2Z/Zg6u\ne8TNH0dyWPhZDK9fPYbzhnWrs3xz+XFXOlsT88goMtLvFJPhahOTnM/2pHwWzRvMXyb1aoYa1uys\n3o1f2txosfHsr3F0CvBpcGB47/rxlBitdGlk2lrRfCQwiDbJ18uzXjNoAcxWO0aLvUXXA6rJO38Z\nR2YjgwLAtRN7MSIyhBERwXg2YHhrSp6Bu5bs4p7ZA5gxuGvdBzQhP29PvmxkU5Cft6ck9Gmj3AoM\nSqlOwNdAbyAZuFJrXVClzGjgf0BHwAY8q7X+2rnvY2AaUOQsfqPWeo87dRJnntlDu5L43Hy35wq4\nK9jf262OZU8Pxeg6VkM1WW2sjs3ieGE5e1OLePWq0RzNKWVvWhFbjua1eGAAmNwKqUNF83L3juEh\nYK3W+nml1EPO54uqlDEA12ut45VSPYCdSqmVWusT6yb/Q2v9nZv1EGe41gwK8VkldAnybZEJd9/v\nTOeRH/fTJdCHnFIz/7xwCDMGh7PmvnPp1VnW4hJNw93AsACY7nz8CbCBKoFBa32k0uPjSqlsoAtQ\n+4L6QrRBCdmldPTzIryjX8W2tAIDc179g7FRjlnIq2OzePzi4RX7S4wWknMNjIh0bxJaWoEBux1m\nDQnn2vQo/jQugiA/74qUq/3Dg2o9ttBgJqPIyJDubWuto4IyM5nFba9ewv0Jbl211hnOx5nAKe9j\nlVITAB/gaKXNzyql9imlXlVKSS+UaJOKyi3MfuV3rnh3q8v2sEBf5gztymVjI7j2/R18tOUYr62u\n+C7EP77dy0VvbmJ3SkHVU2JrQDrN81/byKxXNhAe5Mtzl41gXK9ODOxaezCobOGnMZz/+kaScsvq\nfb2WcIuzXsfy2la9RD3uGJRSa4Cahno8WvmJ1lorpWr9S1dKdQc+A27QWp/oJXwYR0DxARbjuNt4\nqpbjFwILAaKiouqqthBNKtDXi0vHRDCwq2vHsp+3J+9d75gv9MHGJJLzDMwfcfK/yyVjIjBZ7fQN\ncz1uV0oBl729hYfnD+a2c+seanrNxCjMVnuj5hZcPr4nIQE+dKt0p9MWXDEuki6BPnRtY/USbs58\nVkodBqZrrTOcH/wbtNbVhpIopTriaGZ6rrb+BKXUdOABrfWFdV1XZj6L9u5AehFXL97GQ+cP5rqz\nW25oqjiztdSy20uBG5yPbwB+rqEiPsCPwKdVg4IzmKAcX4MuAQ64WR8h2oXhEcEcePK8iqCQVWys\nSOt5OioxWhqd2U60PHcDw/PAHKVUPDDb+Ryl1Hil1PvOMlcC5wI3KqX2OH9OzPX/Qim1H9gPhAHP\nuFkf0YaZrDbWHcqqMxfwmcBu16w/nE2x0cL+tCImPreWh3/Y36hzZRcbeXN9PMXlTbdya2MYLTZu\n+SSaz7clV9t31bvbmPTvdRSUnaxjRlE52xPzWrCGor7cCgxa6zyt9Syt9QCt9Wytdb5ze4zW+hbn\n48+11t5a69GVfvY4983UWo/QWg/XWl+ntS51/yWdeYrKLa1dhXr5YlsKN30cw2dbjzXrdUpN1gZ9\n+7bZNfvTihrUGVzbeUqM9ftdrIrN5K8fRfPiikN0CfJlULcgxvUKbdR1b/kkhpdXHuEf3+1r0HFF\n5ZY6l8vWWtf77yunxMSauGx+3FV9BdlZQ8I5d0AYAb4nuzVv/2wnVy3e1uY6xYUsu93ufbY1mVFP\nrmLlwbafVH3m4HAuGtmdWUOaL7FMfpmZkU+s5PoPd9T7mM+3HeOiNzfx5Y4Ut659xxe7GPHEKtIK\n6s4TMb53JxaM6sGfxkbSLdiPlfecy9UTGjeoYuG0vvQM9efGyb3rfUxMcj6jnlzFSysPn7LcO78n\nMurJVfxxJKfOc/bs1IHV957LBzdWb8K+f+4gPr15okve6dvO7cc1E3pWW6dJtD5ZEqOdiwj1p2tH\n33ax3kzvsAD+e+3YZr2Gr5cH/cMD6z2UE2Bcr1DO6h3K2KhTzzquS//wAHpndiCgHtnMwgJ9ef2a\nMW5d74QLR/bgwpE9GnRMSAcfegT71ZkrOqqTP906+tE5sH6T9wY04H2fP7I780c2fEVY0fwkH4MQ\nNcgtNfHO70e5bmIvetfx4XnCvV/tZtm+DDYumlEx8UyItqSlRiUJcVpaE5vF+xuT+Domtd7HhAb4\n0CnAB2/Plv1vZbHZ+Wl3Onmlpha9rjh9yR2DEDUwWmz8si+D2UPCW2QNJHf8su84d365mz9PjOLZ\nSxuWwU2cWSSDmxBu8PP25PJxka1djXqZ2r8LN53Tm6vOkhUBRNOQwCBEOxfcwZvHLhrW2tUQpxHp\nYxBCCOFCAoNoVWUmKx9tTiK72NjaVamT0WLj481JpBc6lnbYm1rI9zvTKiaJxR4v5uvoFOyNmCiX\nVWzko81JlJmsTVpnIRpDAoNoVcv3ZfDkslje25jUJOcrN9u4+L+b+PevcU1yvsrWHcrmiWWxvLE2\nHoAHvt3L/d/uJa3AESge/Wk/i77fz+Gskgaf+72NiTy5LJbl+zPqLixEM5M+BtGqzhvejZxSE5c0\nMJF8bcrMVvanF+Hn3fTfeaYP6sKD5w1i/gjHpKxnLx3B4cxiIkMdcxYev2gYu47lN2hy3Qk3Tu5N\niL8384bXtMK9EC1LhquK006hwYy/jye+XpJoXojKZLiqOGO19XkHQrR10scghBDChQQGIYQQLiQw\nCCGEcOFWYFBKdVJKrVZKxTv/rTHTiFLKVil729JK2/sopbYrpRKUUl8704AKIYRoRe7eMTwErNVa\nDwDWOp/XpLxS9raLK21/AXhVa90fKABudrM+Qggh3ORuYFgAfOJ8/AlwSX0PVEopYCbwXWOOF81v\nc0IuTy07iMnasjmaiwwW/vXTAQ6kF7l9rp3HCnj85wMtMqP4rfUJXPjGRkqNDb/WmthM5rzyO4cy\nik9Zzmix8eSyg2w92jS5kq02O8//Fsea2PplAFx/KIvnfo3D4kyd+smWZD7a3LjJiesPZfHvX+Ma\nlIZVtAx3A0NXrfWJqZqZQNdayvkppWKUUtuUUic+/DsDhVrrE/+L0oBaZzkppRY6zxGTk1N3mkHh\nvjfWxfPh5mTis1o2Ffe2pDw+23aMT7cmu32u9zcm8snWY+xOKXT7XHV5e30CB44Xszkht8HHvrEu\ngfjsUj7anES5ufZAHJtRzEebk3l7Q4I7Va2QXljOO78n8srq+HqVf3VNPIv/SORYniN96bPL43jm\nl7g6c0fX5JXV8bz7RyLH8utOhSpaVp0T3JRSa4CapmM+CnyitQ6pVLZAa12tn0EpFaG1TldK9QXW\nAbOAImCbsxkJpVRP4Det9fC6Ki0T3FpGWoGBuIwSZg8Jx3GD1zKsNjsrD2Zxdt9OdA50L2VpdrGR\nXSkFzB3aDQ+P5n0N2xPz2JyQyz2zB+Dh0bDvXKkFBj7YmIRdaz7deozl/zeFYT2Cq5XTWrM6Novh\nEcH0aKJcyRsOZ9O7c0C9MtWl5Bk4mlPKjMGOvN17Uwuxa82YqBq7F0/pWF4ZiTllFecSza++E9zc\nmvmslDoMTNdaZyilugMbtNaD6jjmY+AX4HsgB+imtbYqpSYBT2itz6vruhIYxOnqg02JvPN7It/d\nPoleneuXUlSI+mqp1J5LgRucj28Afq6hIqFKKV/n4zDgHCBWOyLSeuDyUx0vxJnk5il9iX50tgQF\n0arcDQzPA3OUUvHAbOdzlFLjlVLvO8sMAWKUUntxBILntdaxzn2LgPuUUgk4+hw+cLM+4jSQW2ri\n71/sJDo5v7WrIsQZya21krTWeTj6C6pujwFucT7eAtSYiFZrnQhMcKcO4vSzN7WQX/dn0tHPm7N6\nd2rt6ghxxpFF9ESbM3NwOF/cMpGRkdU7X4UQzU8Cg2hzlFKc0z+stashxBlL1koSQgjhQgKDaHVm\nq53bPovh/Y2JrV2VNuupZQd59Mf9rV0NcYaQwCBaXaHBzMqDWfy4O721q9JmfbczjW9j0ho1w1iI\nhpI+BtHqwjv6se7+aYRK5rVarbp3GjatW3QGujhzSWAQbULfLoGtXYU2rVuwX2tXQZxBpClJCCGE\nCwkMQgghXEhgEEII4UICgxBCCBcSGIQQQriQwCCEEMKFBAZxRll5MJOzn1vLvrSmT/X58qrDzHh5\nAwVl5iY/txAtSQKDOKOkFxjILDaSW2pq8nMnZJWSlFtGmdlad2Eh2jB3U3t2Ar4GegPJwJVa64Iq\nZWYAr1baNBi4Wmv9kzPN5zQc+Z8BbtRa76nrupLaUzSW1ppio5Vgf+8mP7fNrjGYrQT5Nf25hWgK\nLZXa8yFgrdZ6ALDW+dyF1nq91nq01no0MBMwAKsqFfnHif31CQpCuEMp1SxBAcDTQ0lQEKcFdwPD\nAp1OyKwAAApjSURBVOAT5+NPgEvqKH858JvW2uDmdYUQQjQTdwNDV611hvNxJtC1jvJXA0uqbHtW\nKbVPKfWqUsrXzfoI0Wqe+Pkg455eTU5J0/dfCNGS6lxETym1BuhWw65HKz/RWmulVK0dFkqp7jhy\nP6+stPlhHAHFB1gMLAKequX4hcBCgKioqLqqLUSLKzZZKDZasNllaWzRvrnb+XwYmK61znB+8G/Q\nWg+qpezdwDCt9cJa9k8HHtBaX1jXdaXzWbRFWmvs2tHXIERb1FKdz0uBG5yPbwB+PkXZa6jSjOQM\nJijHIvOXAAfcrI8QrUYpJUFBnBbcDQzPA3OUUvHAbOdzlFLjlVLvnyiklOoN9AR+r3L8F0qp/cB+\nIAx4xs36iDbObtfYT9OmFrvdjtlqb6ZzV3/ftD5930vRutxK1KO1zgNm1bA9Bril0vNkIKKGcjPd\nub5of2a/+jsmi51Ni2acdtnIpr+8gZT8cpbeeQ4jI0Oa9NxzXv2dcouNzYtmVrxvl/1vC0m5ZWx/\nZBa+Xp5Nej1xZpMMbqJFdfTzxuTVPN+qW5u/tycK8PFs+gUFgvy88a5y3iBfL4L9vPE4zQKsaH1u\ndT63Ful8FkKIhmupzmchhBCnGQkMQgghXEhgEEII4UICgxBCCBcSGIQQQriQwCCEEMKFBAYhGsls\ntZNRVN4s584rNVFqkkxwonVIYBCike7/dg+T/r2OuIziJj2vwWxlwnNrWfDmpiY9rxD1JTOfhWik\niX06k5RTRpegpk0j4uvlyeR+nekbFtCk5xWivmTmsxBCnCFk5rMQQohGkcAghBDChQQGIYQQLiQw\nCCGEcOFWYFBKXaGUOqiUsiulau3QUErNU0odVkolKKUeqrS9j1Jqu3P710opH3fqI4QQwn3u3jEc\nAC4D/qitgFLKE3gLOB8YClyjlBrq3P0C8KrWuj9QANzsZn2EEEK4ya3AoLWO01ofrqPYBCBBa52o\ntTYDXwELlCM/4UzgO2e5T4BL3KmPEKeSWWRkb2phs53/UEYxX25PabbzC9FSWmKCWwSQWul5GjAR\n6AwUaq3/v72zDZGrOuP471/DJhSt7prWJM3bBlJtpBB1EW1LrRrU+mET6ZpuQbrxBatt/SIFE/LB\nUhDbfhFKC2mwal/Al26RblEJiUnwS1bNBzXGmGQTK00aExtfqJSub48fzplyZjN3Z2bn3jubnecH\nw9z7nJf7n+ecuc89556Z+3FiP+W50I6TF+sefoHX3/oPoxuuZt7Zc3Kvf2DTLj4Y/5gvndXFqhXz\ncq/fccqibmCQtA2o1cs3mtnf8peUqeN24HaAxYsXl3VYZwZxyzd72f2Pd5l7ZjG3sm68ZCFb9x2n\nb2lPIfU7TlnUDQxmtqrFYxwFFiX7C6PtJHCOpFlx1FCxZ+nYDGyG8MvnFjU5HcjavkWs7VtUP+MU\nubf/Qu7tv7Cw+h2nLMpYrvoisDyuQOoCBoERC//FsQMYiPmGgNJGII7jOE5tWl2ueoOkI8DlwFOS\ntkT7AklPA8TRwE+ALcA+4Akz2xuruAe4W9IY4Z7D71vR4ziO47SO/4me4zhOh+B/ouc4juNMCQ8M\njuM4ThUeGBzHcZwqPDA4juM4VXhgcBzHcao4LVclSXobeHOKxecC/85RTl64ruZwXc3huppjpupa\nYmZfrJfptAwMrSBpdyPLtcrGdTWH62oO19Ucna7Lp5Icx3GcKjwwOI7jOFV0YmDY3G4BGbiu5nBd\nzeG6mqOjdXXcPQbHcRxncjpxxOA4juNMwowMDJJulLRX0qeSMu/gS7pO0n5JY5LWJ/ZeSc9H++Px\n78Lz0NUjaaukg/G9u0aeKyW9lLz+J2lNTHtE0htJ2sqydMV8nyTHHkns7fTXSkm7Ynu/Iul7SVqu\n/srqL0n67Pj5x6I/liZpG6J9v6RrW9ExBV13S3ot+udZSUuStJptWpKudZLeTo5/W5I2FNv9oKSh\nknU9kGg6IOm9JK0Qf0l6SNIJSa9mpEvSr6PmVyRdnKTl7yszm3Ev4KvA+cBOoC8jzxnAIWAZ0AW8\nDKyIaU8Ag3F7E3BnTrp+BayP2+uBX9bJ3wO8A3w+7j8CDBTgr4Z0AR9k2NvmL+ArwPK4vQA4BpyT\nt78m6y9Jnh8Bm+L2IPB43F4R888GemM9Z5So68qkD91Z0TVZm5akax3wmxple4DD8b07bneXpWtC\n/ruAh0rw17eAi4FXM9KvB54BBFwGPF+kr2bkiMHM9pnZ/jrZLgXGzOywmX0IPAasliTgKmA45vsD\nsCYnaatjfY3WOwA8Y2b/zen4WTSr6/+0219mdsDMDsbtfwEngLo/4JkCNfvLJHqHgaujf1YDj5nZ\nuJm9AYzF+krRZWY7kj40SnhaYtE04q8srgW2mtk7ZvYusBW4rk26vg88mtOxMzGz5wgXgVmsBv5o\ngVHC0y/nU5CvZmRgaJAvA/9M9o9E27nAexYeMJTa8+A8MzsWt98CzquTf5BTO+V9cSj5gKTZJeua\nI2m3pNHK9BbTyF+SLiVcBR5KzHn5K6u/1MwT/fE+wT+NlC1SV8qthCvPCrXatExd343tMyyp8tzV\naeGvOOXWC2xPzEX5qx5ZugvxVd1nPk9XJG0D5tVI2mhmbXtE6GS60h0zM0mZS8Li1cDXCE++q7CB\ncILsIixbuwf4eYm6lpjZUUnLgO2S9hBOflMmZ3/9CRgys0+jecr+molIugnoA65IzKe0qZkdql1D\n7vwdeNTMxiX9kDDauqqkYzfCIDBsZp8ktnb6qzRO28BgZqtarOIokD4ZfmG0nSQM02bFq76KvWVd\nko5Lmm9mx+KJ7MQkVa0FnjSzj5K6K1fP45IeBn5api4zOxrfD0vaCVwE/JU2+0vSF4CnCBcFo0nd\nU/ZXDbL6S608RyTNAs4m9KdGyhapC0mrCMH2CjMbr9gz2jSPE11dXWZ2Mtl9kHBPqVL22xPK7sxB\nU0O6EgaBH6eGAv1Vjyzdhfiqk6eSXgSWK6yo6SJ0ghELd3R2EOb3AYaAvEYgI7G+Ruo9ZW4znhwr\n8/prgJorGIrQJam7MhUjaS7wDeC1dvsrtt2ThPnX4QlpefqrZn+ZRO8AsD36ZwQYVFi11AssB15o\nQUtTuiRdBPwO6DezE4m9ZpuWqGt+sttPeCY8hFHyNVFfN3AN1SPnQnVFbRcQbubuSmxF+qseI8AP\n4uqky4D344VPMb7K8876dHkBNxDm2saB48CWaF8APJ3kux44QIj4GxP7MsIXdwz4CzA7J13nAs8C\nB4FtQE+09wEPJvmWEq4EPjeh/HZgD+EE92fgzLJ0AV+Px345vt86HfwF3AR8BLyUvFYW4a9a/YUw\nNdUft+fEzz8W/bEsKbsxltsPfCfn/l5P17b4Paj4Z6Rem5ak635gbzz+DuCCpOwt0Y9jwM1l6or7\nPwN+MaFcYf4iXAQei335COFe0B3AHTFdwG+j5j0kqy2L8JX/8tlxHMepopOnkhzHcZwaeGBwHMdx\nqvDA4DiO41ThgcFxHMepwgOD4ziOU4UHBsdxHKcKDwyO4zhOFR4YHMdxnCo+A6oTaEOVRruqAAAA\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb580ac1358>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7fb5808a5198>"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.scatter(final_train.toarray().T[4], final_train.toarray().T[5], y_plt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGIQQQpRQYBBCCFFCgUEIIUSJX0u/8mOlQ7skAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb5809e2400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}