Text_Analytics_IMDB.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Text Analytics - IMDB"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "####  Functions to support text analytics of imdb reviews\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import requests\n",
    "import re\n",
    "from bs4 import BeautifulSoup\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from collections import defaultdict\n",
    "from itertools import combinations\n",
    "from ggplot import *\n",
    "import unicodedata\n",
    "from textblob import TextBlob\n",
    "import nltk\n",
    "from itertools import izip\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.decomposition import NMF, LatentDirichletAllocation\n",
    "\n",
    "%matplotlib inline \n",
    "matplotlib.style.use('ggplot') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import ner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Modules to support text analytics \n",
    "\n",
    "def requestPage(url): \n",
    "    try:\n",
    "        response = requests.get(url)  \n",
    "        page  = BeautifulSoup(response.text, 'html.parser')\n",
    "        return page\n",
    "        \n",
    "    except(ConnectionError, Exception), e:\n",
    "        print \"Exception is :\", e\n",
    "\n",
    "        \n",
    "def extractRatings(page,i):\n",
    "    data_rating = {}\n",
    "    i = i + 1 \n",
    "    for b in page.select('div[id=\"tn15content\"] img'):\n",
    "        rating = b.get(\"alt\")\n",
    "        if rating is not None and rating != u\"[Next]\" and rating != u\"[Prev]\":\n",
    "            data_rating.update({i:rating.encode('utf-8')})\n",
    "            i = i + 1\n",
    "    return data_rating\n",
    "\n",
    "def extractReviews(page,i):\n",
    "    data_review = {}\n",
    "    i = i + 1 \n",
    "    for a in page.select('div[id=\"tn15content\"] p'):\n",
    "        review = a.get_text().replace(\"\\n\",\" \").replace(\"*** This review may contain spoilers ***\", \"IGNORE\").replace(\"Add another review\",\"IGNORE\")\n",
    "        if review != \"IGNORE\":\n",
    "            data_review.update({i:review.encode('utf-8')})\n",
    "            i = i + 1 \n",
    "    return data_review\n",
    "\n",
    "def tokenize(text):\n",
    "    # first tokenize by sentence, then by word to ensure that punctuation is caught as it's own token\n",
    "    tokens = [word.lower() for sent in nltk.sent_tokenize(text) for word in nltk.word_tokenize(sent)]\n",
    "    filtered_tokens = []\n",
    "    # filter out any tokens not containing letters (e.g., numeric tokens, raw punctuation)\n",
    "    for token in tokens:\n",
    "        if re.search('[a-zA-Z]', token):\n",
    "            filtered_tokens.append(token)\n",
    "    return filtered_tokens\n",
    "\n",
    "def print_top_words(model, feature_names, n_top_words):\n",
    "    for topic_idx, topic in enumerate(model.components_):\n",
    "        print(\"Topic #%d:\" % topic_idx)\n",
    "        print(\" \".join([feature_names[i]\n",
    "                        for i in topic.argsort()[:-n_top_words - 1:-1]]))\n",
    "    print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Extract IMDB movie reviews ( 50 positive , 50 negative ) for movie = Titanic ( 1997 ) \n",
    "   http://www.imdb.com/title/tt0120338/\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Extracted ratings & reviews\n"
     ]
    }
   ],
   "source": [
    "# Movie name = Up\n",
    "\n",
    "p_url = \"http://www.imdb.com/title/tt0120338/reviews?filter=love;filter=love;start=\" # positive reviews link\n",
    "n_url = \"http://www.imdb.com/title/tt0120338/reviews?filter=hate;filter=hate;start=\" # negative reviews link\n",
    "\n",
    "# DataFrame for all extracted reviews\n",
    "imdb = pd.DataFrame()\n",
    "\n",
    "# Pagination for 50 reviews \n",
    "pagination = [0,10,20,30,40] # pagination for imdb pagination to get top 50 reviews in a category \n",
    "\n",
    "for counter in pagination:\n",
    "\n",
    "    positive_url = p_url + str(counter)\n",
    "    negative_url = n_url + str(counter)\n",
    "    \n",
    "    # Request url for reviews \n",
    "    responses_positive = requestPage(positive_url)\n",
    "    responses_negative = requestPage(negative_url)\n",
    "    \n",
    "    # Extract ratings and reviews \n",
    "    ratings_positive  = extractRatings(responses_positive,counter)    \n",
    "    reviews_positive  = extractReviews(responses_positive,counter)\n",
    "    \n",
    "    ratings_negative  = extractRatings(responses_negative,counter+50)\n",
    "    reviews_negative  = extractReviews(responses_negative,counter+50)\n",
    "    \n",
    "    # Copy the extracted data into dataframes\n",
    "    df_p = pd.DataFrame(columns=['ratings','reviews'])\n",
    "    df_p = pd.DataFrame([ratings_positive,reviews_positive]).T\n",
    "    \n",
    "    df_n = pd.DataFrame(columns=['ratings','reviews'])\n",
    "    df_n = pd.DataFrame([ratings_negative,reviews_negative]).T\n",
    "    \n",
    "    # Create master dataframe for text analytics \n",
    "    imdb = imdb.append([df_p,df_n])\n",
    "    \n",
    "print \"\\n Extracted ratings & reviews\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   ratings                                            reviews\n",
      "1    10/10   Every once in a while the conversation will t...\n",
      "2    10/10    To all the miserable people who have done ev...\n",
      "3    10/10    I find it so amazing that even after all the...\n",
      "4    10/10   This movie re-wrote film history in every way...\n",
      "5    10/10   Back in 1997, do I remember that year: Clinto...\n",
      "6    10/10   James Cameron's 'Titanic' is essentially a ro...\n",
      "7    10/10   It's a crying freaking shame that this outsta...\n",
      "8    10/10    When you have a film this big and successful...\n",
      "9    10/10   Thirteen years after Titanic launched and all...\n",
      "10   10/10   It really physically hurts me when i don't se...\n",
      "    ratings                                            reviews\n",
      "91     1/10    Alright... where to begin. This is.. without...\n",
      "92     1/10   This movie is the biggest waste of an opportu...\n",
      "93     1/10    This movie should have never have been nomin...\n",
      "94     1/10    I have said it before, and I'll say it again...\n",
      "95     1/10    I don't see what all the hype is about this ...\n",
      "96     1/10    Your average American loves Titanic. This is...\n",
      "97     1/10    Actually reading a lot of the previous revie...\n",
      "98     1/10    Why does everyone say this is a good movie w...\n",
      "99     1/10    What a dog! A wonderful exercise by Hollywoo...\n",
      "100    1/10    Why is this the biggest grossing movie of al...\n"
     ]
    }
   ],
   "source": [
    "imdb.columns = ['ratings', 'reviews']\n",
    "imdb = imdb.sort_index()\n",
    "print imdb.head(10)\n",
    "print imdb.tail(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Convert user ratings from string to float \n",
    "temp_ratings =  imdb['ratings'].str.split('/', expand = True ).astype(float)\n",
    "temp_ratings['divide'] = temp_ratings[0].div(temp_ratings[1], axis='index')\n",
    "imdb['ratings'] = temp_ratings['divide']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ratings</th>\n",
       "      <th>reviews</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Every once in a while the conversation will t...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>To all the miserable people who have done ev...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>I find it so amazing that even after all the...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>This movie re-wrote film history in every way...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Back in 1997, do I remember that year: Clinto...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.0</td>\n",
       "      <td>James Cameron's 'Titanic' is essentially a ro...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.0</td>\n",
       "      <td>It's a crying freaking shame that this outsta...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.0</td>\n",
       "      <td>When you have a film this big and successful...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.0</td>\n",
       "      <td>Thirteen years after Titanic launched and all...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1.0</td>\n",
       "      <td>It really physically hurts me when i don't se...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    ratings                                            reviews\n",
       "1       1.0   Every once in a while the conversation will t...\n",
       "2       1.0    To all the miserable people who have done ev...\n",
       "3       1.0    I find it so amazing that even after all the...\n",
       "4       1.0   This movie re-wrote film history in every way...\n",
       "5       1.0   Back in 1997, do I remember that year: Clinto...\n",
       "6       1.0   James Cameron's 'Titanic' is essentially a ro...\n",
       "7       1.0   It's a crying freaking shame that this outsta...\n",
       "8       1.0    When you have a film this big and successful...\n",
       "9       1.0   Thirteen years after Titanic launched and all...\n",
       "10      1.0   It really physically hurts me when i don't se..."
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imdb.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### Identify polarity & subjectivity ( sentiment analysis ) of all movie reviews \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ratings</th>\n",
       "      <th>reviews</th>\n",
       "      <th>polarity</th>\n",
       "      <th>subjectivity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>0.1</td>\n",
       "      <td>Alright... where to begin. This is.. without...</td>\n",
       "      <td>0.298106</td>\n",
       "      <td>0.504167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>0.1</td>\n",
       "      <td>This movie is the biggest waste of an opportu...</td>\n",
       "      <td>0.097222</td>\n",
       "      <td>0.701111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>0.1</td>\n",
       "      <td>This movie should have never have been nomin...</td>\n",
       "      <td>0.031250</td>\n",
       "      <td>0.575000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>0.1</td>\n",
       "      <td>I have said it before, and I'll say it again...</td>\n",
       "      <td>0.113091</td>\n",
       "      <td>0.514425</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>0.1</td>\n",
       "      <td>I don't see what all the hype is about this ...</td>\n",
       "      <td>-0.030919</td>\n",
       "      <td>0.421402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>0.1</td>\n",
       "      <td>Your average American loves Titanic. This is...</td>\n",
       "      <td>0.014734</td>\n",
       "      <td>0.471261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>0.1</td>\n",
       "      <td>Actually reading a lot of the previous revie...</td>\n",
       "      <td>-0.088302</td>\n",
       "      <td>0.528534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>0.1</td>\n",
       "      <td>Why does everyone say this is a good movie w...</td>\n",
       "      <td>-0.027848</td>\n",
       "      <td>0.636391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>0.1</td>\n",
       "      <td>What a dog! A wonderful exercise by Hollywoo...</td>\n",
       "      <td>0.097994</td>\n",
       "      <td>0.432562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.1</td>\n",
       "      <td>Why is this the biggest grossing movie of al...</td>\n",
       "      <td>0.167273</td>\n",
       "      <td>0.290909</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     ratings                                            reviews  polarity  \\\n",
       "91       0.1    Alright... where to begin. This is.. without...  0.298106   \n",
       "92       0.1   This movie is the biggest waste of an opportu...  0.097222   \n",
       "93       0.1    This movie should have never have been nomin...  0.031250   \n",
       "94       0.1    I have said it before, and I'll say it again...  0.113091   \n",
       "95       0.1    I don't see what all the hype is about this ... -0.030919   \n",
       "96       0.1    Your average American loves Titanic. This is...  0.014734   \n",
       "97       0.1    Actually reading a lot of the previous revie... -0.088302   \n",
       "98       0.1    Why does everyone say this is a good movie w... -0.027848   \n",
       "99       0.1    What a dog! A wonderful exercise by Hollywoo...  0.097994   \n",
       "100      0.1    Why is this the biggest grossing movie of al...  0.167273   \n",
       "\n",
       "     subjectivity  \n",
       "91       0.504167  \n",
       "92       0.701111  \n",
       "93       0.575000  \n",
       "94       0.514425  \n",
       "95       0.421402  \n",
       "96       0.471261  \n",
       "97       0.528534  \n",
       "98       0.636391  \n",
       "99       0.432562  \n",
       "100      0.290909  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Find polarity and subjectivity of the reviews ( sentiment analysis)\n",
    "\n",
    "imdb['polarity'] = imdb.reviews.apply(lambda s: TextBlob(unicode(s, errors='ignore')).sentiment.polarity)\n",
    "imdb['subjectivity'] = imdb.reviews.apply(lambda s: TextBlob(unicode(s, errors='ignore')).sentiment.subjectivity)\n",
    "imdb.head(15)\n",
    "imdb.tail(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Co-relation between movie review polarity and star ratings from the users \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "               ratings  polarity  subjectivity\n",
      "ratings       1.000000  0.631849      0.160403\n",
      "polarity      0.631849  1.000000      0.141231\n",
      "subjectivity  0.160403  0.141231      1.000000\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuUAAAF6CAYAAAC+4ib1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsvXl4ZGd5pn9/p6pUJam1tFS9qtXq1b3Z3Xa7bWMbHBwa\nmIRJYJIBBsKMExKYmCXEyYQBEn6YyRgMyZXMkJCQxXHCQBJDAk6AAE7jBK/gtsHubtvd7n1VL6Vd\ntarqfL8/vnNKW1XpVNUpVan13tfVl6TSqXO+Vh2VnvOe531epbXWCIIgCIIgCIJQN6x6L0AQBEEQ\nBEEQFjsiygVBEARBEAShzogoFwRBEARBEIQ6I6JcEARBEARBEOqMiHJBEARBEARBqDMiygVBEARB\nEAShzogoFwRBEARBEIQ6I6JcEARBEARBEOqMiHJBEARBEARBqDMiygVBEARBEAShzogoFwRBEARB\nEIQ6E6z3AmrNhQsX6r0EYR6IRqPEYrF6L0NoYOQcEeZCzhGhFHJ+CHOxevXqqp4vlXJBEARBEARB\nqDMiygVBEARBEAShzogoFwRBEARBEIQ6I6JcEARBEARBEOqMiHJBEARBEARBqDMiygVBEARBEASh\nzogoFwRBEARBEIQ6I6JcEARBEARBEOqMiHJBEARBEARBqDMiygVBEARBEAShzogoFwRBEARBEIQ6\nI6JcEIRFjR4eRMfH670MQRAEYZEjolwQhEWN/SefQj/0l/VehiAIgrDICdZ7AYIgCHVlKIZuaa33\nKgRBEIRFjlTKBUFY3CTikE7VexWCIAjCIkdEuSAIixadnYBMWkS5IAiCUHdElAuCsHhJJsxHEeWC\nIAhCnRFRLgjC4iURNx9FlAuCIAh1RkS5IAiLFxHlgiAIQoMgolwQhMVLclKUa63ruxZBEARhUSOi\nXBCExYsryrWGTKa+axEEQRAWNSLKBUFYtGjXvgKQEQuLIAiCUD9ElAuCsHiZKspTyfqtQxAEQVj0\niCgXBGHxMlWUS7OnIAiCUEdElAuCsHhJjk9+LqJcEARBqCMiyoUFgU4lyX3mf6LPn673UoSrCamU\nC4IgCA2CiHJhYXD5Ahx7GX385XqvRLiK0Ik4WM7boIhyQRAEoY6IKBcWBu449PGx+q5DuLpIxqGz\nCwAtolwQBEGoIyLKhYWBmycdHy+9nSCUQyIOS6PmcxHlgiAIQh0RUS4sCPJ50nGplAs+kpislJOW\nSERBEAShfogoFxYGjn1FS6Vc8JNkHNXZbT5Pp+u7FkEQBGFRI6JcWBi4lfKEVMoFf9DZrLGsLGmD\npiaplAuCIAh1RUS5sDAQT7ngN27zcPMSaIqIp1wQBEGoKyLKhYWBWymX9BXBL9zBQS2tEBZRLgiC\nINQXEeXCgkC7Vc34GFrr+i5GuDpwLvRUSytEmiUSURAEQagrIsqFhYFrX8lOQCZT37UIVwfu3Zfm\nVmgKQ0pEuSAIglA/RJQLC4Op49AlFlHwA/fuS0sLRJohI6JcEARBqB/Bei/A5fnnn+fBBx/Etm1e\n97rX8Za3vGXWNk899RRf/epXUUrR19fHhz70oTqsVKgLyQQEQ6ZSnhiDrmi9VyQscHTC8ZQ3LzGe\n8rHR+i5IEARBWNQ0hCi3bZsHHniA3/md36G7u5uPfvSj7NmzhzVr1uS36e/v5+GHH+Z3f/d3WbJk\nCSMjI3VcsTDvJOMQXQ4Xz0uzp+APriWqpRXVFEFLJKIgCIJQRxrCvnLs2DFWrlzJihUrCAaD3Hbb\nbezfv3/aNt/73vd44xvfyJIlSwDo6Oiox1KFOqC1dkT5SvOAxCIKfpCIg7JMlTwSgYwMDxIEQRDq\nR0NUygcHB+nu7s5/3d3dzdGjR6dtc+HCBQA+/vGPY9s2b33rW7n++uvndZ1CnZjIQDaLWrYCDej4\nGKreaxIWPok4NLegLAsdjkijpyAIglBXGkKUe8G2bfr7+/nEJz7B4OAgn/jEJ/j93/99Wltbp223\nb98+9u3bB8D9999PNCre44VObmiAGNDat4FxoBWb1hmvazAYlNdaKMnMc2TEzjLR1k40GmW8s4t4\nJkV3dzdKySXfYkXeR4RSyPkh1JqGEOVdXV0MDAzkvx4YGKCrq2vWNps3byYYDLJ8+XJWrVpFf38/\nmzZtmrbd3r172bt3b/7rWCxW28ULNUdfPAdAPBiGUBPxy5dIznhdo9GovNZCSWaeI7mhQQhHiMVi\n2LYNWhO7cAEVDtdxlUI9kfcRoRRyfghzsXr16qqe3xCe8o0bN9Lf38/ly5fJZrM89dRT7NmzZ9o2\nN998My+++CIAo6Oj9Pf3s2LFinosV5hv3CEvzS3QugQS4ikXfCARNxnlAOFm81GaPQVBEIQ60RCV\n8kAgwLvf/W7uu+8+bNvmzjvvpLe3l4ceeoiNGzeyZ88edu3axQsvvMA999yDZVm8613voq2trd5L\nF+YDN0+6uRVa29DjEl0n+EAyDstXmc/d6rhM9RQEQRDqREOIcoDdu3eze/fuaY+9/e1vz3+ulOKu\nu+7irrvumu+lCXVGJyaj62htk/QVwR8ScVSLqZSrcDMaRJQLgiAIdaMh7CuCUBI3T9q1r8hET8EP\nEnEzOAikUi4IgiDUHRHlQuMzdciLVMoFH9C5nPGPt8z0lIsoFwRBEOqDiHKh8UkknCEvzU6jp1TK\nhSqZcqEHmAFCIKJcEARBqBsiyoXGJ+kMeVHKeMozGbRMXxSqwe1TaJ4uyrWkrwiCIAh1QkS50Pg4\nohwwlXIQC4tQHU6lXM2qlMvFniAIglAfRJQLDY9OJvI2A9Xabh6UZk+hGhIz7CsRV5RLpVwQBEGo\nDyLKhcYnMT5pM5BKueAHM0V5k5O+khJPuSAIglAfRJQLjU8yMcW+4gyMkkq5UAXanQrrXOwpKwBN\nTZARUS4IgiDUBxHlQuMzZciLWynXIsqFapg6JdYl3CzpK4IgCELdEFEuND7JxBT7ilTKBR9IxkEp\niDRPPtYUFvuKIAiCUDdElAsNjbZtSCWme3+DQfGUC9WRcGI2rSlvgZFmtNhXBEEQhDoholxobFJJ\n0DrvKTdZ5e1SKReqIxGfbl0BE4solXJBEAShTogoFxqbQt7f1iVoqZQLVaCT8cm7Ly7hiEQiCoIg\nCHVDRLnQ2CSN+FYt00W5VMqFqkiMQ8uS6Y+FIzI8SBAEQagbIsqFxiZRqFLeJp5yoToK2FeUVMoF\nQRCEOiKiXGhsnHHoUwWUkkq5UC3J+PS7L+BUysVTLgiCINQHEeVCQ6Pzorxl8kFp9BSqJVHMUy72\nFUEQBKE+iCgXGpuZ49DBeMozafREpj5rEhY0OpczqT6z0leaIZMyMZyCIAiCMM+IKBcam4LpK+4A\nIfGVCxWQcs6pWZXysInflIs9QRAEoQ6IKBcam2QcQk2oUCj/kGp1UjNElAuVUOjuC5hKOYivXBAE\nQagLIsqFxsaZvDiNfKVcfOVCBTiivGCjJ4goFwRBEOqCiHKhsUkmZlc085VyEeVCBSScOyzN03PK\nVV6USyyiIAiCMP+IKBcaGl1oHHpru/meiHKhEpLF7CuuKJcEFkEQBGH+EVEuNDbJQvYV8ZQLlaOL\nesqlUi4IgiDUDxHlQmOTjKNmRddFIBAU+4pQGYnZA6kAqZQLgiAIdUVEudDYFPCUK6VMtVwq5UIl\nJOOgFESapz8eMaJcS6VcEARBqAMiyoXGJlnAUw7Q2oaOj87/eoSFTzIBkRaUNePtr8mplKckfUUQ\nBEGYf0SUCw2Lzk5AJjPbUw4mFlEq5UIlJMZn+8khXyknI6JcEARBmH9ElAuNS7LI5EVw7CviKRfK\np2CiD0BT2HysYaVcZyfQw4M1278gCIKwcBFRLjQuxRryACWVcqFSkvGCF3rKCkBTU02HB+lHv4X9\n8bvRqUTNjiEIgiAsTESUC41LssjkRZBGT6FyEoVFOQDh5traVy5fgFQSjhyq3TEEQRCEBYmIcqFx\nyVfKi3jK00njOxeEckgUiNl0CUdqa18ZGzEfX/xxzY4hCIIgLExElAuNi+spL5i+IgOEhAopYl8B\nIBypbSTi6DAA+qXna3cMQRAEYUEiolxoWHSxcegAre3mozR7CmWg7VzB7Ps84UhthweNDoOy4NJ5\ndOxS7Y4jCIIgLDhElAuNSwn7ipJKuVAJSacKXlKU17hSvm0XAPolsbAIgiAIk4goFxqXZMKZvFjE\nUw5SKRfKI+FcxDUvKfz9cKRm6Ss6k4ZUEnXNDlgaRb8oFhZBEARhEhHlQuOSjEOkefbkRch7yrVU\nyoVyKJXoA6gainKcJk/aO1E7boDDL6BzudocSxAEQVhwiCgXGpdiQ15gSqV8dP7WIyx8EiX6FMBE\nItZKlDtNnqp9KWy/wazl1NHaHEsQBEFYcIgoFxoWnYwXjkMEiDRDICCecqE8SgykAiAcrrkop70D\ntW0nKCXRiIIgCEIeEeVC41IiJUMpBS1LxFMulEXJRB9whgel0bbt/7HzorwTtaQd1m2WZk9BEAQh\nj4hyoXFJlrCvgLGwSKVcKIe57CuRCGgNExn/jz1FlAOo7dfDyVfQCTmHBUEQBBHlQiOTiKOK2VcA\nWpegpVIulEMiXjzRB6ApYj7WIhZxbASaW1ChJgDU9hvAtuHwAf+PJQiCICw4RJTPwP7WV8j98f+u\n9zIEKD3kBaRSLpRPqUQfMJGIUJsBQqPD0NYx+fWGLRBplmhEQRAEARBRPgv9yiF4+Xm01vVeyqJG\naz2nfUW1iqdcKJPEeOlzKlK7SrkeHc5bVwBUMAhbd6Jf/JG83wiCIAgiymcxGINMZjJTWKgP6ZS5\ntV+yUt4uolwoCz3X3RfXvpKqQQLLDFEOjoVl4DJc7vf/eIIgCMKCQkT5FLTWMBQzXwxcru9iFjv5\n6LrSnnJSSXQ2Oz9rEhY+iXhpUe5WyjM1EOVjw6iZonzHDQCSwiIIgiCIKJ9GMj6ZUSyivL4kE+Zj\nsXHoMDlASNIrBK+UGkgFNauU61wOxseme8oBtXwVLFspeeWCIAiCiPJpDMbyn2oR5fUlaYT2XOkr\ngDR7Ct5JxlEeKuXa7wFCrh1uRqUcnGjEwwfR2Ql/jykIgiAsKESUT2VoUpRLpbzOuJXyEgJKuZVy\n8ZULXknEzdCpYoSbzUe/RbmTUT7TvgKgduw2jaUnjvh7TEEQBGFBIaJ8CtqtlC9pRw9cqe9iFjl6\nrnHoMKVSLqJcmBtt25BKlD6nwmHzsUainLbZopwt14FliYVFEARhkSOifCpDMbAsWH+NVMrrTdJL\no6eplMsAIcELOhk30zpLpq+4otzfSERdyr7S0gobtogoFwRBWOQ0jCh//vnn+dCHPsQHP/hBHn74\n4aLb/eAHP+Btb3sbx48f938RgzHo6DLNVwOXJTu4nniwr+QbPcVTLnhAu+dJKUuUFTDC3O/hQW6l\nvIAoByeF5cxx9Niov8cVBEEQFgwNIcpt2+aBBx7gYx/7GH/4h3/Ik08+yblz52Ztl0wm+fa3v83m\nzZtrsg49FIOuKHQtg1RSUj3qSSIOgSA4I8kL0txi7mxIpVzwgO2cJ6qUfQXMVE+/hweNDptzOdJc\n8Ntq+w2gNfplme4pCIKwWGkIUX7s2DFWrlzJihUrCAaD3Hbbbezfv3/Wdg899BBvfvObCYVCtVnI\nYAy1NIqKLjdfi4WlfiRNnrRSqugmSinTtCeVcsEDXirlgCPKfa6Ujw1DW0fx83ndJnMuS165IAjC\noqUhRPng4CDd3d35r7u7uxkcHJy2zYkTJ4jFYuzevbsma8gPDuqKQrcjymMiyutGIl7aT+7S2iaV\ncsETbqXciyjXfnvKC0zznIqyArB1J/rwQV+PKwiCICwcgvVegBds2+aLX/wi73vf++bcdt++fezb\ntw+A+++/n2g06u0YoyNcmcjQuqaP5s3buAK0pBO0eny+4C9DuSx2ewfdc/z8Bzs6UZkUwWDQ82u9\n2NHZLCq4IH71fSV9yAjtrp5eAiXOlcHWJSg7x1Ifz6eBxDhWdHnJfcZ37WH8R0+x1IJAl5zL9UDe\nR4RSyPkh1JqG+Mvc1dXFwMBA/uuBgQG6urryX6dSKc6ePcsnP/lJAIaHh/nsZz/Lhz/8YTZu3Dht\nX3v37mXv3r35r2OxGF7QZ04AkGhqJpHOQDhC/PQJkh6fL/hLbmQImsJzvn65cDMMD5HNZj2/1osZ\nPTqE/ZH3YN39UdR1N9Z7OfNKs5OAMpjKoEqcK7lAEMbHfD2fcoMDqJ51JfepV6wx63v2adTuW307\ntuCdaDQq7yNCUeT8EOZi9erVVT2/IewrGzdupL+/n8uXL5PNZnnqqafYs2dP/vstLS088MADfP7z\nn+fzn/88mzdvLijIq8IdHNQVNb7P7uWSVV5PPNpXlNhXyuNSP0xk0D/893qvZN7JR2c2F262zBNu\n9jWnXNs2jI9AW0fpDdduhGAQLUOEBEEQFiUNUSkPBAK8+93v5r777sO2be688056e3t56KGH2Lhx\n4zSBXivyg4OWOremupfDwKWaH1coQjIxd0oGmAFC0ujpHSeaTx98Dp3LoQKBOi9o/tDxcWhuMf7t\nEqhwGO3n8KDEOORyJT3lACoUgt4N6JMiygVBEBYjDSHKAXbv3j2rifPtb397wW3vvfde/xcwFINA\nANpNNUt1L0cfP+z/cQRvJOOlJy+6tLZBMo7OZWu/pqsA7eZlJ8bh+GG4Zkd9FzSP2I4onxOfK+Vz\nZZRPRW3Ygn78u4vW9y8IgrCYaQj7SkMwFIPO7skqWvcySIyj3SE2wryhczkjiuZKyQBTKQf0uFhY\nPDE6DEpBIIg+8Ey9VzOv6PiYtwu9SKQmolx5EOVs2AKZDJw/7d/xBUEQhAWBiHIHPRibtK4AdK8w\nHyWrfP5Jxs1Hr5GIgD0ukxA9MToMS9rhmh3oF2bPAriasePj3i70mowo17bty3G102BKm4dK+cat\n5jniKxcEQVh0iCh3GYqhpsSQqe5l5hMR5fOPe3fCg4BSriiX8eSecPOy1a6b4eI59OUL9V7SvKET\n42ZAz1xEIubjRMafA5dhX6FrGXQsBRHlgiAIiw4R5UwZHLR0coARzlRPLaJ8/nEq5d4aPY0o11Ip\n98aYI8p33gSwqKrldnzc2zkVdkS5XwOERofBsvJWq1IopWD9FvQJ6WcRBME/DlyM8+f7L9Z7GcIc\niCgHGBuBbBaWLpt8rK0TQk2LvlKuR4aw//bP0Cl/JxyWJOHYV8rwlIt9xSOjw6i2TtSylbB6LfqF\nxeMr1+XYVwBSPvnKR4ehrQNleXu7VRu2wOV+tNz9EQTBJ549P863Xhk2RUihYRFRDvmM8mn2FaWg\ne9mir5Tr738H/W/fQu9/fP4O6tpXyvGUi4DxxpRx72rXTXDsJWPruMrRtu3YVzxYolz7SsYfUa7H\nRjz5yfPH37jFfCLRiIIg+EQqa8T4hC2ivJERUQ4wODk4aBpdyyG2yEX5s0+Yjz/8/vwd062Ue7Ea\nNLeAsqSq6AGdSppUEVeU77wZcjn0oR/VeWXzQCoJWns7p8LOcCE/K+Ve/OQufZvAstDHRZQLguAP\n6ZxpXM9kRZQ3MiLKAT00Y3CQg4ouh8HFO9VTnz8D/Wdh2Up45dDkgKVak/RuX1GWBa2tYl/xwsyG\nww3XmCSWxeArL+OcIhw2H/2KRRwd9haH6KDCEVizToYICYLgG+msI8qlUt7QiCgHUykPBo1AmUrX\nMhgb8Xe63wJCP/sEKAvr3b8OWqP3PzY/B3Yr5REP9hWAljbfGz3tf/pb7Me+4+s+686MvGxlBVDX\n3Yg+ZKZ7XtU455TyJMqdSrkPv/daa6e5tqOs56kNW+HkK2j7Kn9dBEGYF1z7SibrT9SrUBtElIOT\nvBKd3YgVdbLKF2G1XGuNfu5JuGYHatN2WH/N/FlYkgkIN3sfAd+6xNdKuc5m0Y98Hf30v/u3z8EY\n9hf/uL4XeAWi+dSum810z2Mv12lR80Q5liinUu7La5VOmmFA5dhXwAwRSiWh/1z1axAEYdGTr5Tn\npFLeyIgop8DgIId8Vvli9JVfMNYVted2ANQtPwFnT6IvnKn9sZNxb02eLq1t/jZ6nj0BmbSvyTt6\n3z+hH38Ejhz0bZ9lr6FQXvb2GxbHdM+k08zqJac8Xyn3IXHI/ZmX0egJTgILoI9LNKIgCNXjesrd\nj0JjIqIcZg0OyuNM9VyMCSyudUXtvhUAddOrTUPlPFTLdTLuzfvroFqXTE5N9OP4R18ynwwPoLMT\n1e8vO4H+wb+bz08fr3p/FZMXiJNWCtXcAluuRR+4un3lupyYzXxOebr6A8+wDHlm+SpY0iZDhARB\n8IW8fUUq5Q3Nohfl2rZheGD64CCXjqUQCC66rHKttRHlW65FtS8FMB+37UL/8Pu1zzlNJsqrlC9p\nxx4f8+3weVGuNQwNVL/DA/tNFn4ggD59rPr9VcrYMLS2oYLBaQ+rnTfDxfPoS1fxdM9yRHmT2+jp\nR6XcuVgs11OeHyIkolwQhOpx7SsTIsobmkUvyhkdhlxu+uAgB2VZJiZxkYlyzp+Gi+dRN94+7WF1\ny0+Yn8XxGvuPE3FvNgOXliXo+JgvTXFaazj2EnSbia7ELlW9T/uJfdDZhdp9G9SxUq6LRPOpXe50\nz6vYwlJG87CyLCPM/Wj0LGQZ8ojasAX6zy6KHHlhYfHYqVEOXUrUexlCGbiiXOwrjY2I8gKDg6bR\nvXzR2VdmWldc1O5XQVNT7S0sybixVXjFGSCUF17VcPE8jI+ibnktUL11SQ8PwKEfoW79SVh/jbHE\njAxVv85KKCbKoyugp29BWFh0dgL9wjPl361xzinPzcPhiD+RiK4oX1JepRwmfeWcPFr9OgTBR/7f\n81f4yqF5isgVfGEyfUUq5Y2MiPLBwhnlLqp7OQwsnvQVY115ErZeN8sHqyItqF23oJ99Ap3N1m4R\niXIbPZ2qug8WFn3MWFfUza8BZVV9l0Q//W+gbdSr96L6NpkH62VhKZGXrXbeBEdfNKPoGxj96Dex\n//h/w7lT5T0xEUe5F29e8EuUF7EMeWL9NaCUWFiEhiOeyXFmJFPvZQgeydk6P8kzI5XyhmbRi/Ji\ng4PydC+HkUH0xCJ5Azp3Ci6dz6euzETd8hNG/L7045ocXmttPOXlNHpGHavJxbPVL+Doi6YRcvVa\n02dQhSjXWqOf2GdiJZevhrUbjMiql4WlxGRJtetmsG30oefmeVHe0VqbBBso21akx0aw2trn3tAl\nHEH7MNFTj45UZF0Bpwl39dqSolzbOexvfQV9/nSlSxSEsrC1JjFhM5TMMp6WHP2FwFTLijR6NjaL\nXpQzFINQk0k6KITrLZ6vaZZ1Rj/7JFgW6oZbC2+w4wZobaudhWUiA7mstzxpl77NEGqabNCsAn3s\nZdi0zTTadS+rzr5y9CW4fAF1+14AVKQZVvTUpdlTZ9Im97qtiI1i/WbzvUa2sBx72diLoPzpsoNX\nsNy5A16INEPGJ/tKhaIcHAvLiSOmIb0A+uEvox/+Evrfv13xMQShHBITNq6sOzPiQ0KRUHOmWlZE\nlDc2IsqdjHKlVMFvK1eUD1Tf8NfoTKauXIcqIt5UMITaczv6+R+iUz6kU8yknCEv7ppCIUKbt1Ut\nyvXwAFy5aIYlAap7RVUZ9frJfRBpntYwq/o21se+MkfDoZnuucdM93zpx+jzp9Hjo7VP2ikD/fgj\nRiwHgzBUpqVsKEbA/V32gk+NnqUsQ55Yf40Z7nR5djKO/cPvo7/9D+buy/lTlR9DEMognpmsjoso\nXxikslMr5f7ZV54+M8ZwsoZW1kXIohfleihmElaK4QwQ0ovBV372pKns7nl1yc3UzT8BmTT6+R8U\n/L6emMD+1lewv/v18teQdDr6y/GUA03bdsGZ41VdKOijJlVGbd5hHoguh6GBivzzOpVAP/sE6qbX\noNzca4B1m2B4cP6bPd287I6lRTdRe26HRBz7Dz+Bfe8Hse95F/b7fp7cR36F3Kd/C/vv/rxuIl0n\nxtHPPWHOvaXRsu5c6VQCEnECy8qolIebzZ2FahkbKX53wgNq41aAWRYWffoY+m/+CDZtN03E5083\n1AWUcPUSz0yKOvGVLwymi3J/3icSEznuf/w8/3p82Jf9CYZFL8oZiqEKZZS7LI2CZS2KqZ762SdK\nW1dcNm2DrmUFLSz6lUPY/+vXzC31b/x9+ULBiX9T5UQiAqHt15toy2qa4o69ZCqkvevN193LQdv5\nhJ5y0PufgEw6b11xUWvr1OzpIZpPXbcH69N/gfVbn0K997dQb/9l1Ot+1lykZCfQj34TLp2fpwVP\nRz/zGGQyqNe8HpZGJ3tBvOAI+HLsKyoSMVNdq0BPZMx02moq5SvXmAvUKee1HhnC/vynoK0D6+6P\nQN9Gc4fJj0x9QZiD+ISplFsKzg5LpXwhkM75b18ZTprzYOpFmlA9FUQCXD1oOwfDgwUzyl1UIACd\n3TB4dYtyrTX6uSdh6y7UHA1xyrJQt9yB/u7X0c7teR0fR//jXxuLQXQF6tY7TfLI4JVJX74XKqyU\nh7ZeZyaOHn0Jtf36sp7roo++CBu35pMyVPdy450cvALLVpa3ryf3wapecGPtXKY0e6qdN1W0zkrw\nmpetoivM6zfz+ZcvYP/2r6IPH0CtXFObRZZAP/6IuVjq24TqiqJfedH7kx1RHijHU94Uqb5Snh8c\nVIWn3LJg/TXo40aU64kJ7D/9NMTHsP7nZ4w1pmedOU/Pny59108QfMAVYes6w2JfWSCka2BfGUmZ\nO8iJCRHlfrK4K+XDQ2Dbc/8hiy5HX+2V8rMn4HJ/0dSVmahbXmvSOvY/gf3MY9j/3/vQT+5DvfHn\nsO79Y9Rr3mg2LDMVQifLmLw4BaulFXrXGWFdAToRh3OnUZu2TT7oXEyU+9rr/rNw/LCJQZzRq6Ai\nzbByzfw3e7qivK1Cgbhslbk78vIB/9bkEX36OJw5gXrNG8zPc2nU5L17HBalB431LBAt4+LQh0p5\n3jJUTaUcp9nz/Gl0Kon+8p/C8cNYv/Qh1NoNZoM1fQDiKxfmhXHHU75tWTPDqRyjKfEUNzq1sK8M\np8x5kBSa7FKSAAAgAElEQVRR7iuLW5TPNTjIQXUtv/or5Xnryqs8ba96+mDNOvQ/PIj+i9+HpVGs\n3/4DrP/8i6hw2EQKAvrCmfIWkiy/0TO/ps074OQRdHai7Ody4rDJE3f95GAu1pQqu8lXP7kPAgHU\nq15beJ31aPYcHYaWVlQoVNHTlVKorTvhyMGiSSC1Qj/+XQg1mThOMK+LbcOIRy/jUAyUwuoqfkds\nFk5OeVX/1zH3QqhyTzmA2rAVtI3+68+ZC9//+PZpfR+qtQ06u8q+ABaESnAr5duWmbuZ4itvfNJT\n01d8Gh40nK+USyymn4goB5NHXYrochgarO3AnDqSHxi0bRdqifcsZ/Xan4ZgEPX2X8H62O9NVu4A\n1brEEQplivJ8+kp59hUAtXk7ZDIVjbLXR182vQPrr5ncXzBkrEtlVMp1NmtsO9fdhGov0lTZt3H+\nmz2rjOYDYNtOiI/BuZP+rMkDOp1CP/MY6sbb830GyrWbDXpsvh6MQUdXeQN83ObcKqrlXi1Dc7J+\ns9nfc0/C9a9C/cw7Zm/T04c+J6JcqD3xiRwK2BJtBiSBZSHg5pS3hKxpmeXVMOJUysW+4i+LWpTn\ns45LeMoB6FpWccPfgqD/rIkC3H1bWU9Td7wR6//+Hdben0VZBcaXr+6roFKeMOJ4amKJVzabKEN3\nKmc56GMvwtqNxl4yle7l5WWVH3oORoexXr236Cb1aPbUY9WLcrV1p9nXPFpY9LNPQjKBes0bJh90\n72x5/H3Ug1fK91rnRXkVsYjVWoYc1JJ26OmDnj6sX/514zOfuU3POrh49qotHAiNQzxj09Jksaw1\nSEvI4ow0ezY8rn2lPRzw0b5i3muS2fJE+VAyy9/8+DI5W9KiCrGoRTlDMfPHdw7/snIbxKocud6o\n6GNOFOCW68p6nlKqoEDIf3/1Wug/69n7Cxj7SnNr0dz4kutpX2qG85SZV64nJuDk0Xw++bR9di8r\n63W3n9wHHUvh2huLb+Q2e56aRwvL6DCqWnHY2W388IfnUZQ/8Qis7MlfcAF5ge15gNBgDFWOdQVM\nJCJANVM9R4ch3GzsXFVi/cbvYn3kM6hIkTtIPX2QzRbMMxe8kc7a2BIrOSfxTI7WUAClFL0dTZyV\nSnnD49pX2sIBJnyqlFfqKd9/fpyvvTTIabmYK8iiFuV6jsFBea72rPITh2FJOyxf5e9+e9aaCZ3l\njERPxMtu8pyK2rwdjr5Unhf49DGYyJjnzqR7BQzF0Lm5Lyz02Cgc2I+69SdNak+xNbrNnmfKt9lU\njB/2FZxq+dEXK/Ptl4nuPwvHXka9+g3Tf0dblpjoSg+Vcq212a7MSnleSFczQGhsBNqr85Pn19Pe\nWVyQAyrf7CkWlkrI2Zr3/NNxvnHo6h8SVy3xCZvWJiMdejvC4ilfAKSyNgpoDVnT4hGrodL0FVfE\nxxK1/xuyEFnUopyh2Nx+cjD2FqWu3kr58SOwYUtF1elSKKfZsxxfuU4mKvKT59m83WSdl2Gbydtd\npiavuESXm6bCYQ8Z0CePgG2jrtsz56bz2eypJybMxY4fonzbTiNUTx31YWWl0Y8/AoEg6tY7p69B\nKeiKequUj4+aC8Ol5dpXnEp5FaJc+3Qh5IlVvcb2Jb7yiohncoykclwY8WGK61VOPJOjtckUHdZ2\nhBlN5/JWBqExSWdtwkFFOGj5nr5Srih3c+5jCTlnCrHoRflcyStgxrjTsXRBiHL7n/8O+9+/7Xl7\nHR+Di+dM7JrfrO41xyinepeIV5S84uKmp5RjYdFHX4KVPQWj65Sbse6h2VOfdRog3eFDpXCbPYcH\nPa+zYsZ8ajgE2HKdsd7U2FeuJyZMw+z1NxeOFFwa9eYpH3QTlsq1rzie8moq5fMoylWoyVi3JBax\nIsacRJFktv5JEpfHJxo6ZnA8Y9MaMtJhbae5oyS+8sYmndOEgxZNAeV7TnnW1mVZYlwRH4tLpbwQ\ni1aU62wWRobmbvJ0Kbfhrw7o+Bj6X76KfqSM8fYnXgEmx3n7iYq0mKzvcpo9k9WJcqIrTOqLx7xy\nbdvGIlHAT27252SVe3jt9dkTsGwlykOlX/WZRI1KkmLKxqe8bHDi93o3oA+/UPW+SqGf/yGMj2K9\n+g0Fv6+6onnBXRI3oaXSRs90FQOEfPDxl4Pq6ZNYxAoZSzv+2Ez9RfmnHzvHgz9u3L810yvlTQCc\nFQtLQ5PK2oQDFk0BfyrlEzmb+IRNZ8ScB+X4yhMZ177SuBee9WTRinJGBkFrz3+sVffyhq+U6x89\nDbksXLmI9ujj1icOg7Jg3ebaLGr12vISWJJxT6K2GEop1OYd6KMvGT/xXPSfNXaXQn5ymLxo8/La\nnzlhmji90Lvemew5DxYWv6L5HNS2nXD8CDpdu+qYfuIRc0FXbDrr0mUwOjSntz1vcamwUl7p/1Hn\nciY+cr7sK2CaPWOX0KnE/B3zKsEdiJMqM0miFgynclwab9wqYjwz6Snvag7S2mRJLGKDk85qIkHl\nVMqrF+WudWVVm7koK8fCkhBPeUkWryh3b2t78ZSDEQhDsfKSROYZvf/xfJOk14QMffwwrOmbHQXo\nE6qnDy6e9x7VlkxU1egJGIE9POCpwdS1uRSrlKtQyFTe5xggpJMJEyvZ602Uz2ezp2952Q5q605z\n8VdB9KQX9MAVeOl51O17i6f7dEXNRfVc9p/BKxAMlT/Ap9pK+fioWd88inK32bPs2QACo26lvAEG\noaSydl70NBo5W5PM2ixxKuVKKdZ2hMW+0uAYT7ljX/HhwtPtIVjVZobRlSfKxVNeikUryvWQx4xy\nl+7lkMvNLQLqhB4ZgsMHUXe+yfjfPXh+tZ2Dk6/UxLqSZ/VaI+A8RLXp0SEjyr1eKBXBTVHx5Cs/\n+hJ0dMGylcW36V6OnstT7vjJlddKOfPY7OmzKGfzDggEahaNqA88A4C66TVFt1Fu4+ZcFhanmbvs\nJuZqPeU+WoY807MOkASWSshXyussyrXWjihvTMHiii/XUw6m2fPsSNrbnUmhLqTyotzYV6p9rUZm\nVMrLySp3z6GBRFYiSAuwaEV5vkmsHPsKQIPGIurnnjRj4m++A7VlJ/rIgbl/8S6chVQSNtROlKse\nJ4HFg4VFHzlknjN11H0lrO4z1XYPlVx97CXUpm0lRZvqXjGnfaWsJk+Xvk1VNXvqRBztZeLk6DBE\nmlFN1edlA6hwBNZvQb9cG1+5PrAfVvSgVvYU38jNKp+j2dMMDirTugImchEqF+Vuc225Ffpq6F5u\nLiZElJdN3lNe5+mEWVtja2MR8StP2k/izsWL6ykH6O1oYixjM9Sg1X3BNHpGAoqmoEJjzrNqcC8a\nV7uivAL7StbWjMo5MwvPotwuJ/d5ITAYg+YW7/7lbrfhrzFzbPUzj0FPn4kh3HqdaWK9cLb0c04c\nBkBtrEHyisvKNaAstJdb6kcOQqTZiNUqUJYFm7aj52j21ANXjL1hrouA7mVzW5fOHjcCrKPL+zrd\n/2cFzZ56fBT7k7+G/uIfz71xDVJA1LadcOYEOj7u6351KgmHD6B2zhEr2eWxUj7oLWFpJsqyjDCv\nUJT7bRnygrIs08MhorxsXFGeqnP6SjI7KZYa0cISdyvlTVMq5ZLA0vC49pVwwLxu1frKZ1bKy7Kv\nZHJ0OA2iV8RXPgvPovy9730vDz74IMePz+PAkxqiB2PQWYZNottt+Gu8SrkeuAzHD6NuvgMAtW2X\neXwue8HxI2Zo0DKfhwZNQTWFYdlKT82e+shB2Lyj5OAdz8fdvN142UeHih/PqaSrzQXyyacSndu6\npM+ehN4N5dkkKmz21Fpj//XnYPCK6QmYa/taiPKtu0Db8MohX/fL4Rcgm0XtvKn08SMtJqVnqPjv\no3Zfs0oq5WCqzpVO9KyDKAdQa9bB+VNiJSiTsUxjVMrTU2wAjWhhcW0+S0KT79FrO4wol8mejYtr\nXwkFzN+nakX5cCpLJKjy6SuJMmxfiQk7f86Ir3w2nkX5xz72MSzL4jOf+Qz33HMPX/va14jFPI65\nbkTKnPKnmsKmEtqACSz62SeASQ+uiq4wQniO2Dp94jBs3Or70KBZrF4LF0pX7/TwAFw8j9pynS+H\nzFtgjr5c+Hi2bX5ukWZYs670vrpXmE+K+Mp1dgIunEGVY12h8mZP/b1vwAvPmIExsUumybQUtcjL\n3nANNIV9t7DoF/ab4VHFIiqnMtcAoeFBc+FQQaUcMOdGpgpRHgxVNwirEnr6YHzM3CkTPJOvlNfZ\nUz41/WWkESvlefvKpHTojARokwSWhiadm0xfAarOKh9O5eiIBGkJuaLc2/6ytiad0/m7K5JVPhvP\nonzDhg3cddddfOELX+Cuu+7i3Llz/OZv/iaf/OQnefTRR0lVWlGqF0Ox8geKLFuJvjR3w+J8o595\nHNZfg5rSrKi27oQjh4qOhzdDg87XZmjQDFTPWrjcj54onmWb95Nv9UeU07cRmpoKWli0baO/9Cfw\n/A9Rb/w5lDVHZb57jqzy/nOQzXqPQ5yC6tsEp7xXyvXpY+h/+GvYdTPWz/+iefDcqdJPGhv2veFQ\nBUOwebuvzZ7attEHn0Xt2I0KBud+wlwDhJwquvLazD2TpjC64kr5CLR31P6Cdwaqx01gEQtLOYw3\nSKU81eCV8njGta9MvmcqpejtCHNmWLLKG5V01ibiNHqCEenVMJLK0hkJEAkqFN5/b9ztVi4J0RRQ\nUikvQNmNnpZl0dPTQ09PD+3t7QwODvLEE09w991389hjj9Vijb6jJyZMJavM0duqpw/ONdatYX3x\nPJw5PjupYutOM4inWBX2xBGgNkODZtHTZ0bVXzxffJsjB40docxqczFUMGSaEWcksLiCXD/+COqn\n34Z609vm3lneulS4n0CfPWGO6TEOcRp9G2HEW7OnTiaw/+yz0N6J9Yu/lv9Z6XMniz8nmzWV0xoM\nsVFbd0L/Wf+mkp45biq8c1hX8sfvisLQQNHvV5xR7hJprjgSUY8N1+RnPif5BJZT83/sBYxbKc/k\nbHJVNsFVw3RR3oCV8onZlXKAvk5JYGlUcrYmk9OEAxZht1Kerda+YirlSimaQ5ZnUe7aXFpCFt0t\nQckqL4CHcpRhfHycp59+mscee4zz589z66238oEPfIAtW0yl9dixY9x3333ccccdNVusbww7f8jL\njd7rXQ+PP2KEQKW3xH1G738clELtefW0x9XWnWiMr1ytv2b2844fBquGQ4OmrmX1WrOWEhYPffgA\nXLNj7qp1OcfdvB39ra+ikwlUc8tsQf6WX/BUyVShJhMzWayf4MwJ0xS4onxvvurbhAbT7NlZvElU\na22q+wOXsf7Hp1BL2s0fwNa2fBxjQcZHzMcaeJvVtl3mdT1yEHXLT1S9P31gPygLde2N3p6wNApj\nI+iJjHmNZlLpNE+XcKS6SMQymn79QrW1m3P1nFTKy2E0baMADaRzNi0+vg+VQ2piaqNn41UR4xkb\nS0FzcLoo7+0IE5+wGUxm6W4J1Wl1QiFc/3g4qGhyXrdqk31GUlm2RE1sbHPI8mxfmYzUDBBtCTEg\nlfJZeK6U33333Tz33HP81E/9FH/2Z3/Ge97znrwgB9i0aRN79syRmNAoOLe8y01lUGscQVmiMjmf\naK1N6srmHbOGIKn2TujpK+r51SeOwJp1Jt6u1qxYDYFA0VvqevCKGbzjk5/cRW3ebjzFxw8bQf7l\nP3UE+Vs9C/I83cuL2lf02ZPmZ1nJH/Le9SadZo5mT/3Ev6KfeQz1s+/M57ArpWDNusk4xkLUMi+7\nd72JnvTJV64PPAsbtxhh6QX397eYhWUwBs2tlU+IbYqYyNBKGB1Gtc9jHOJUevokgaUMJnImG7yr\n2dSoUlVWEathmqc82YCV8kyO1pA1671zbae5KD4zIhaWRsNtHg4HLZos87pVY1/J2ZrRdI7OiPl9\naSlHlDv2p+aQRbQlKJ7yAngW5Z/73Of4yEc+wm233UYoNHklPDw8nP/8/e9/v7+rqxH529rlek0d\nv2ZJETSfnD8FF88VHbKitu6EYy/P8nJrOwcnXkHVMJ982jqCIVi+umgCS95P7rMoZ8NWsCz00ReN\nIH/su44gf1fZXl/VvbzghFCtNZw9WdbQoGn7jTTDyp6SolyfP4P++z+HbbtQP/Xz05/fuwEunC4e\n11jDFBBlBWDLdb74yvXwAJw+NmfqyrTjzzFAyGSUV35HS0Uqq5RrrWFsZN6TV1xUT5+xFTXw9OFG\nwvWTR1vN37W0DxMPKyXtVDC7m4MNWymf6id3cdM0JBax8XAv9CJBi6Zg9Y2e45kctiYfa9gSsjwP\nD0pMidTsbgkxkMzW1S7WiHgW5b/+679e8PF77rnHt8XMG1cumo/lVspbWk3TX4NUofQzj4NloW68\nreD31bZdMJGBmbF5F84Yr2wt88lnrqWnr/gAoSMHjA1jjhSUso8ZaYa1G9GPfL0qQQ6YWMTBGHpm\nXn/skvHuV+GFV32b4NjL2N98CPvxR9AHn0WfOY4eGUKnkth//lkIN2P98m/Mrsb3roNMBi71F9x3\nrfOy1dadMHAZ7f5OVYg+8KzZXxmi3PWKF01gGYqV3TcyjUrtK4lxE6FZJ1FOzzrze3+58DkhTMeN\nQ1zW6lbK6yfK841wbaGGFOXjmdwsPzlARyRIRzggCSwNSHqqfcWHnHK318GtlDeHAiQ9phbF857y\nANGWILZuTJtWPfHsKS/UwJFIJLCshTcUVL9yCNasr8y60bu+ISrlWmvjJ9+2C1VsauA115pK8eED\nRjy5zz3uNHnOU6UcMLGIzz6BTqdm/dz14YOOn9z/c0ldcy361NHqBDlA13LIZU0j4lSrkHMuVNTk\n6a7xxlvRh55F/9OXAeNrnYn1659EdSyd/dw1642v+9xJ1Ko1s59Ya1Hu+spffmFa+k+56AP7zQXv\n6rXen+S+DiXsK2rd7H4Kz1Qqyt2feT0aPQG1ps+cQ+dPm+FdQkncJs+o44WupyhPO9aZFUuaODvi\n72AuP4hP2LSGCtv0ep1mT6GxcM/ncMCaEolYjSg3ItqtlDcHLQY8Nmy6F50tIYtlzp2pWEL6EKYy\npyi/++67AchkMvnPXcbHx7n99ttrs7IaoScycOxl1Gt/uqLnqzXr0S/sR2fSvo0tr4iTr0DsEupn\n/kvRTVRzC6zbPNtecPywyVyvQkSVi+oxzZ70n53WXKpjl2DgMur1b67Ncd/0NnNBcu3uquLpVHS5\nWf/ApWmiXJ89AcrKW5sq2vf1ryJw/aucVKAhk689MoQeGYKRQdTaDagdNxR+8qpe49c/exIK2ZhG\nh6EpbO4a1IKVa0xD4+EDcMcbK9qFzqTh5RdQt7+urNdINYXN8KsClXKdTsP4aHUN2eEIZNJo2y7v\ngnHUNNfWxMfvhVW9pk/h3GnUjQvr/bkejKenV8rTdfaUWwqWtwYZS+fI2ZqANb+xmqWIZ3L0tBdo\nqgbWdjTxbydG0VrPexSoUJz0VPuKDznlIzMq5eV4yuNTRHm0xTw/Fp9gS7RGf58WIHOK8g9+8INo\nrfn0pz/NBz/4wWnf6+zsZPXq1TVbXE04fhiyE9Mqx+Wg1qxDaxvOn4H1tU8uKYbe/zgEg6jrX1Vy\nO7V1J/o7/5hPIAGnyXPDlvl943QqoPr8GdRUUX7koFmn335yB9XSCtd5TPMohTNASMcuo6YMttFn\nT8LKHl8u0FQoZKrFTi66l1dHhUJmAFGxrPJaDA6aenylTLX84LPobNZbvvhMjhyCTBq18+byn9sV\nRReqlA9VGYcIEHb+UGTSJh7RI/rUUfNJNdaZKlBNYVi+SmIRPeLaVxqhUp7K2oQDFp2RIBoYSefy\nDai1YCKniU9MNu3NRTFPORhfeTJrE0tk81VQof64F5n+2VdMpXyqKPcciZjJEVDQFFD53zfJKp/O\nnOWf7du3s2PHDh544AG2b98+7d+CE+Q40XuWBdfsqGwHvevMfuqYwKLtHHr/E3DtHiM6S6C27jQZ\n4a+YITp6fBQunZ+ffPKpLFtlJhzO9JUfPmiqneXYFupBPqt8RgLLmRMVN3n6herdAE5W+kx0jUU5\ngNp9K8TH4JWDFT1fH9hvqtJbri3/yUujk9GHU3EeKzdhaRph50KrDAuLTiXQ3/lHYytb2VP5saul\np69hel8andF8pdxfUX5mOM23jpQ3WTWVtZ3x5UbwDCdrK1i+eWSQD3zzJLbHfPH4RI4lJUQ5SLNn\no5Gamr7iQ075SMoIa7e3wM0p95JRn5iwaWkKoJSitckiElSSVT6DkpfHX/va1/i5n/s5AB5++OGi\n27397W/3d1U1RB8+AOs2Vx6TFl1pBMRcUxRrydGXjK3h5sKpK9PYtA1CTcZXvuumyaFB8+knB1Qg\nAKvWoC9MCgWtNfqVg7Dl2pr4yf1ENYWN5WeKKNdjo6YiW4Wf3Bd618EP/g09Njo7TnB0uPY2pR03\nQLgZ/dxTqO1FbDZF0FqjDzwD264vnDU+B6orWnhqq1spr6rR06mOp5PAbD9/IfS//jOMj2L9p/9a\n+XF9QPX0oX/8dMEeDmE64+kcQcuMiwd/7Ctaa/7v0/0cG0zxHzZ3eragpLOaSMjKr6XWTXD9YxOM\npXOMp3O0z1Etz9qaVFbTGir8Xt3rjE4/M5Lmxp4lvq9VqAw30ScyxVOersK+MpzK0hEJYjl32ptD\nFhoTJdocKn2eJyZsWpzzRylTLZdK+XRKKqGBgYFpnxf7t1DQqQScfKVi6wpgxGNPX10q5TqTxn5y\nH/aXvwDhiKekChVqgk3b0IdNlrQ+fsQZGrSp1sudvZbVa6dXyq9cNM14Wyp/PeaV6IrpWeXuJM96\nV8pL5eePDtfc26yawqide9A/ehqdKzOG7/wpcw7srHDGwdJlkIijZ+aJD1YvyvNiNu2t8qfHRtGP\nfB1ueFXBgV3ziVrTB1rDhbN1XcdCYCxjqr9hZ7CKH5XyH5wd59igucPiNS7O3TYStOh0LCu1nurp\n3iXwcpxExp3mWbhS3h4OsDQSkKzyBsO9yIwEFUopmgKKiWrsK8lsvskTyIvshIcElqmiHDBTPSWr\nfBolL43f8573AGDbNnfccQdbtmyZllG+4Dj6Eth2VaIcnGbPZx+ft4YWffkC+vvfQT+xz8StrV6L\n9e57PFfA1Jbr0A9/CT06jD5xuPLkmWrp6YMffh+diKNaWqf4ySuwLdQB1b0cfWbSJpJP4VlTeRyi\nLzhxjPrsSROD6aBzOdPsOA8Nh+rG202fwyuHYMoa5kK/sN88/7oKRfnUAUKreicfH7wC7Z3Gc18p\neVHubYCQ/s4/QDqF9ZZ3VX5Mv+hZB4A+fwpVx96XhcBY2qYtHCDikyjP2ZovvTBpqUpO2EUtHzNJ\nO6K8Y54q5WNps/+hVJa1lO6LGc9MZkwXQxJYGo+p9hUwfu6qGj3T03sQWpw0Hi++8sREbpooj7aE\n+HF/vOK1XI148gxYlsVnP/vZhS3IMbFtBENQrZ+6dx0k4kWHlviBtm30C8+Q+z+fwP7tX0V/7xuo\n7ddj/Y9PYd37R8bH6xFXqOmXX4CTR1HzmE8+bR2ub7zfqd4dPmgE41Qx1ch0L4fBy5NZ5WdOwNKo\n9wmUNUK1dZgElJmV8vFRUy2djxSQa2+EpjD6uSfLepo++Cz0bUJ1VjaSvtgAIT0Yq67JE6aI8rlF\nhh6MoR/9FupVr508z+vJshXQ1FR8im4uh37hmXleVGMylsnR1hQgaClCAVW1KP/+qVHOjWa4bW0b\nUJ7IT2VtwkGLZsf/OzJflXIP3nU3Y7pYJCIYX/nZkbRnj7pQe9I5GwV560ooYFU10bNYpdzLHSFT\nKZ98brQ1yFAyS1YGCOXx3Na9bds2XnnlFa65pja3ZZ9//nkefPBBbNvmda97HW95y1umff+b3/wm\n3/ve9wgEArS3t3P33XezbFl5f3T14QOwcWvVSRlqzToTj3fu5GQDoM/ob/w9+pt/D51dZrT6a16P\n6uye+4mF6NsEzS3o733DVP3m2U+eJ5/Acho2bEEfOYjact3Cic+KLods1sQWdnabOMQ6W1fyFMrP\nd/Ky5yOaT4XDqOscC8s7//vsIUcF0GMjcOII6mfeUfmBnUq5HrwyPa1m8AqsrvJir4xKuf7WQ6A1\n6mffWd0xfUJZAVi11vyuTUHbOfQPHzPvLZf7sT7y2flv+m4wxtI5Vi4xBadIMFDVRM+JnM3fHbjC\nxq4Id65v56kzY56TKcCI8q6WIEopOiOBmlfKR8qwr8Q9VMrXdoRJZTVX4hOsWFJ+j4jgP+msJuxY\nVwDCAVVx+orWelalvDlvX/EmyltnVMo1MJjIsnzJwi76+oVnUb5s2TI+/elPs2fPHrq7u6cJqWob\nPW3b5oEHHuB3fud36O7u5qMf/Sh79uxhzZrJwRfr1q3j/vvvJxwO88gjj/ClL32prGmienzUjEN/\n8y9UtVYgP3lSnzuF2lVBjNsc6MEY+rtfQ914O+pXfrOymLkpqEDADBJyKmN1+yPcvRyawsZXfuk8\njAxCjaIQa4HqdrLKY5fRLUvg4vmi01TnG9W7Dv3yC+jsBCrovLnN9xCbG2+H556Eoy97SlLRB581\nQracKZ4z6ewGpaYNENJaw1CseLa7VxxRrlOpkvGU+tIF9BP/ivqJn0JFV1R3TB9Ra/ryk1K1baOf\nexL9z38HF8/BmvVY7/9t2FCfu2aNxHg6R1u3ea2bQxapKho9Hzk2wuV4lvfdsoqQ09xZjqc8ldV5\nG01HJFhTT7mtdX5wkhfxH5/DUw4mqxzg7EhGRHmD4MZsuhhPeWUXnsmsTSan843IYIYHgXdR3jxN\nlBttM5CYEFHu4DnyIpPJcNNNN6GUYnBw0NdGz2PHjrFy5UpWrFhBMBjktttuY//+/dO2ufbaawk7\nEWWbN29mcHCwvIO4/uUy/K7FUJEWk2hRo8me+p+/DNpG/edfrFqQu+R99G0dUCfhoCwLVq9FXziD\nPglIVsQAACAASURBVHLIPLaARLmbH64HLhtbgLarmuTpK2vWm4mj/efyD+kaT/OcibruRmhqQj/3\nhKft9YH90NlV1d0GFQxC+9Lp9pVE3MQYVmtfiTiV8kzpSET9T1+GYAj1H99W3fH8pmcdjI1gP/4I\n9v/6EPrPfw8sC+tXP4L18T9EXX/LwrlLVUPcRk+ASChQsX0llbX5yqEY165o4fqVLXnxUW6l3BXl\nnZEgIzWslCcyNq5rwJMon/DmKQeJRWwk0o4lyqUpYFVcKXftVB3TPOXeznOtNYlMbtpFXdSJIb0i\nCSx5PCu+973vfTVbxODgIN3dk9aM7u5ujh49WnT7Rx99lOuvv76sY+jDB0zEWZ9PqSM964oPbKkC\nfe4U+qlHUa9/s69VN3cc+rwPDZq5jp616EM/MtnknV2wYgFl3TuinIHLaNfS0FvnJk8H1bvejLs/\nexLlrmm+RXmkGa690VhY/st7S8Zc6vFROPRj1M2vqf58nDlAyI+McoAm175SXJTrMyfQ+x9H/fRb\nUe3eYhPnC9XTZ86JL/4xrOgxd91uerUna9FiIe1U/trCzsjwkFWxfeWbh4cYTuX46B3LUErlK4iV\ni/IAxwa8NRlXgusnBxhOzl2RH3cq5aWaVpc0BVjaHOTcqCSwNAopx77i0hRQFXvK3d6DzgrSVzI5\nTU5TsFIuWeWTlF2GTSaTjI2NTQuKX7Fi/iqvjz32GCdOnODee+8t+P19+/axb98+AO6//36iUfOH\nOXb0RQLXXs/Slf5kNo9v2U78hR/S3bbE1ySToT/9NBMtS4i+61exfGwg1N3dDN94G813/jSRaH0m\nDQLEN29j/MnvwYs/IrLndjrK7AsoRjAYzL/WteRyeyeR+Cgkxkm1LCG6dUdDVBv10k4uNzXRPHCR\nNufnMDaRJhFqItq7dt7WmHrtf2DkR0/TEbtA0/biF86jX32A5ESGrp//rwSrfN2GV64me+Zk/vVP\nnzzCMNC5YTOhKfsu9xzRts3lYAgeeZiIguY3vIVAdPm0bYa+8GkmWtuIvvNXsFrbqvp/+I2+5dWM\n/uSbaLp2N5E7Xo8K1G4y5ELl8pip6K7q6iAajdIc6ienKPu9ZDSV5eHDR7l9fRev3mZ6GVRzBjhB\nINziaX85W5PJabraW4lGo6zuijN6YoSu7u58JrSfXJwYBSAUUIxl5/4/6+A4AUvRs2JZyfeT9d39\nXEra8/J+PN/M198ZP9HWJZZEdH7drc0XiWdyFf0/csOm+LFuZZRo1GTRt2dt4Bgq1FxynwNxc6G2\nYmn7tO1am04Qtxfez7VWeH6XPnfuHJ/73Oc4fXp2N/9DDz1U1SK6urpmZaJ3dc1OYzhw4ABf//rX\nuffee4smwezdu5e9e/fmv47FYuihAezzZ7Bv20ss5k9iiu5aAVoTO/Aj3zKJ9csvYP/oadRbf4nB\ndAbSPqe7/OpHGAfGffoZVILucBrzEnHS667x7fWIRqO+7asUemmU5LkzkErAmnWNldO/uo/E0ZdJ\nOz8H+1I/tHXM6xr1+i0QDDH86Lexlq8pvM3JV7D/9Z9Rr38zw60dUOXrZre2o2OXuHLlCkop7FPH\nARi2gqgp+67kHLHu+ST2d79O/B/+hvg/fhF1w62on3wTbN4Bx17Gfu5p1M/9NwaTaUg24C37d/x3\nJoD40HC9V9KQnBkyd0HURIJYLEYkaDEwniz7PPl/z19hPJ3jbdva8891bTCx4VFisbn9sm6lMZdJ\nEYvFaNJpchpOnr80zS7gF2cujQHQ09ZEbDw95//5yvA4rSFrzveTFc2Kx0/H87+PVxPz9XfGT8aS\nKQJK5det7CyJ1ERF/4+zl82EWp0aIxabvIMYtBSxkbGS+zw3at4fc+nEtO26mwOcHSj93IVEtZPu\nPXvK//Iv/5IdO3bwV3/1V7S0tPDggw/y+te/nve///1VLQBg48aN9Pf3c/nyZbLZLE899RR79kzP\nLT558iR/8Rd/wYc//GE6OjrK2r8+fACg6nzyabjNnj75yrVtY//Dg9C9HHXnm3zZZ0MyJS5uQfnJ\nXaLLIXYRzp2q+9Cgmaje9XD2RP4ulh4dnjfrSn4NkRa4drdpKrRn37bXds4Mv+pYivqZ/+LPQZdG\njcUk4eTdDl2BQMCX/7u65loCH/w41n1/htr7ZnPh/Hsfw/7kr2F/6U/M/+Mnf6bq4wj1wbVwuPaV\niDMyvByGklm+cXiQ16xrZ93Sybum4YBC4b3Rc3LIy6SnHGo3QMj9v6/tDDOSys4ZYxjP2CX95C49\n7U2MZ+x8sotQX1JZTTgw3b7ip6ccjIVlrt8b9/szIzVlqud0PIvy06dP8wu/8Au0traitaalpYV3\nvetdVVfJAQKBAO9+97u57777uOeee7j11lvp7e3loYce4tlnTXrAl770JVKpFH/wB3/Ab/3Wb/GZ\nz3zG+wEOH4DWtryQ9oXoCuNR98lXrp95DM6cQL3lXRWNG18wLO2G5lYTZVfr8e81QHWvgMv9kEk3\njJ88z5p1MD4Gw04TdB1EOZhBQgwPwokjs76nH/9XOH0M9dZ3GwHvx/HyA4ScgS0DMejs9tU7rZat\nxHrrL2F99kHUf/uAmYp74QzqZ96BClcXsSrUjzHHJ93W5HrKy49E/OqhGFlb886d02+/K6WIBL2L\nfLeyPluU10aw5EV5RxO2Jp/EUoz4RK5kRrnLmg7z+3BeJns2BOkpfQrgNnpW1jcxnMrS1mQRtKbf\nAWkJWXOmr7iRmlOHB4HJKh8QT3kez/fEQqEQuVyOYDBIW1sbsViM1tZWxsfHfVnI7t272b1797TH\npkYtfvzjH69ov1prUynfcl3JxrNyUZYFa/rQhUabl4meyKC//v9g7UbUzXf4sLrGRSmFuvVOM21x\nId7anOIpbpjkFQe1Zv1kfv7SbhgbRq3zqbG5nHXsvAkdDKKfewq1aVv+cT02iv7aF83v4k2v8e+A\nUwcIrVmPHroyOenTZ1Q4jHrNG9Cvfj0MXJ5s/hUWJONpIxQmGz0DpMqoIl4en+C7x4bZu7GTVW2z\niynNIctzpXxSlJv3RbeZzstgn0oYS+cIWYqVTnThcCpX0iYz7rFSvqbd7O/caIYdK/y58BYqZ3b6\nSuWV8mLniJfz3L04bZlxDkVbQgynckzkbEIB/zTaQsXzT2Dr1q08/fTTALzqVa/iU5/6FPfeey87\nduyo2eJ84Uo/DF5BbfPRuuKg1qyDc6enNb1Wgn70WzB4Bes//6KvFw6NivWO92K9qcHi4zyiXBEW\nDMKqwp7pujHFUqVtG8ZG6lMpb2mFHbvRP5puYdFf+xtIJ7He+d/9vSBb6g4QcjyJgzFUtXGIc6CU\nQkVXLMwLSyGPWx2ejEQsL33lxcsJsja8aUvh5J3mMuwwxSvltbOvtIdNWoo5TmnxH58RZ1eM7pYg\n4YDKe4gXEomJHH974Mqcdw0WEqlcIftKZZXykVR2WvKKS3PQQ6Xc6ZmYWSnvzmeVi4UFyqiU/8Zv\n/Eb+83e84x309vaSSqW4447GruzWxE/usmY9fP87pmJWYXyhjo+h/+UrcO2NvmSoCzXGFeWr104O\n6WkQVEurWd+5UxAfA9uuiygHULtvM2PcTx0101uPHzYDdt74n/wfQ9+51NhJhmJoOwfDAzWrlAuV\ncfBSnL7OCO3hxopjHMvkaAqofCWxORgw0W22JmDNfcHlCpFCQsXsz7son+kpb22yCFq1ta+0RwL5\ntQ/NUZGPT9gs8VApt5Sip72J8wssFjFna37/iQs8dyHOitYQr9tYn/dOvylsX9ForcsuKgyncqxf\nOtuu1xKyGJrjPM1Xygt4ygFiiSwrC9xtWmxUVJa1LIs77riDN7zhDUQi/sUB1oTDB5087B7fd53P\ng67CV67/5auQTGD9/F3+LEqoLd2mAtto1pU8vRvQZ0/Me0b5TNT1N0Mg6DR85rD/9gvG5/0fq5v+\nW/BYVsD8jg/GzP87l4Olta2UC965MJrh4/vO8n+eulD1XUW/GUvn8tYVMJVygLTHSuKk0Cj8p9RM\nCPW4rxmVcqUUHeHaTfUcTZn/e6dTKR+Z4zjxjDdPOcCa9vCCyyr/mx9f5rkLpll8oV1QFMPWJmZz\nqigPBxS2hkocLMUq5S2hgIdKeXFPOUhWuUvJSvkf/dEfebqS+sAHPuDbgvxGHz6A2rG7NreZe/rM\nMc6dRF1/S+HjZyfQ3/26EQx2zlQv7Rzkcqaq98IzqNt+0lhhhIZHRVpQP/tO1LU31nspBVG960yF\nOnbJfF0vUd6yBLZfj37uKXMX6cwJ1Hs/7Ftz5yyWmgFCasAdHCSivFH45pFBNPDchTjPnBvnlt7G\nyXMfy+TyTZ5gPOVgqtYtHm6EJSZyBC1V1AsbCVqexUZ6higH6GwO1Gyq52g6y4bWCK0h07hXqiI/\nkTNDlrx4ygF6Opp4/PToLD9zo/LIsWH+6fAQb9qylBf64wvugqIY7t2XpqnDg5zPMzmbYBnN8BM5\nm/iEnbdVTcWLTSuRyREOqFl3oJa5lfK42FdgDlG+0qdBO3VlbARqYV3BmWC4bCX67Kmi2+hv/yP6\nn//WjLcPBMAKmFvtVsB8vWk76i3vqsn6hNpg+RXlVwPUmvVobedtW/WqlAOoG29DH3wW/ZW/gm27\nUHtur92xupahTx0Fd7Kn2FcagvFMju+dGOGOde2cGkrxl89d4vpVrQ0j1MaLVMq9VrcTE3bRKjlU\n5imfOn2xMxKsrX0lHEApRWckUNK+4iZnePGUg2n21ED/WGZaTGQjcvBSnC88c5EbVrXyy7uX85nH\nz181lXL3js/UC72Q07eW8Xjh6eLesXHvrEzFS/pKYsKmpcD5Ew5atDV5v3i92ikpyt/61rf+/+zd\ne3gchXkv/u9cdvam3dVKK8m62JZsbAMGExscCDE2EKdNTTiH5Jy0JQmpk5aQQEhC25zShCZ5Si8k\nwHHalASSQJLy4zlNTtP6pEmbNITaQMzFYJuLARtfJNuyLVn3ve/O5ffH7Ix2tTOzM7sjjWS9n+fJ\nE2yvd0fS7Pqdd97LXB3HrGIunMV52Ev7TMtXlNMnoPz8x2A2XgP2k1+YvWMgRFMqqVIO7ld/7WVQ\n/o4roXAcoChgb3a5uXOmeALY/zyUsdJYRMqUzwu/PjqJnKjgAxe1IFOU8aUnT+CfD47iI5fNj5/P\nVF7C0th0jWyA1zLlLgXlDYxEBNR50P0T7jdMSrKCVEHWa/zjQesymVSpSS9s8bWWK5/AMp+D8jPJ\nAr729CA6IwK+sKlL3VgaFfDy6ZTtvoL5zOjui3bRZ7dES6PPKDfoCwn61Dp1UVaqxiVqrN4riTDN\nKtc4WhN28OBB7N69G+Pj44jH49i8eTMuueSS2To2d7QtmZ6YMQuYnj4o+5+Hks+B8U9/+CiyBPkf\n/wEIBMH8/q2z9vqEVGhtBwJB4MxJgOOBUJNnh8KEI2De9z+ASDOY2Z5U05IAxCIwcAzwB4BQeHZf\nj9QkyQp+dmgMl7QHsaJF/Wzc3BvFv7wxhutXxAxHCM61VEGqaD7Vyldyor2C20xRRrBWptz2SMTK\nRk9AbSCdzEl1NeVZ0eazR/28/jpWQZHTTHlnRAADzOsykFRBwr27TgEMg3uu7dG/tu6oAFEGhtPF\neXGONkI7pyqnr6jnV9FhUbl2x8YsUw6oPRYRk2Zuq6C8NchTprzE9j3EX//619ixYweam5vxzne+\nE/F4HH/3d3+HJ598cjaPr2GzPdGE6ekFFKUqW678138AR98C83t/5FldL1l81Pn5pQbkSMzzkX3s\nTR8F+573z/rrMNpYxKNvAvGE5183AV48lcJwWsT7L2zRf2/7+jb4WAbffWnI86ZPRVGQzEsVE0WC\nDstXskWpZvmKNs2lllxRho+trLltDvAQZUUPit0yc5NpLMBbzkNPl4J4uzXlfp5Fe5Nv3i4QkmQF\n9z8ziKFUAX9+TXdF8N1dyvKfDyUseb0kqnJOOQDHs8q1oNwsUw6oPRZmMkXJ9E4LZcqn2Q7Kf/rT\nn+Kee+7Bhz/8Ybz3ve/FzTffjHvuuQc//elPZ/P4GjdL9eQ6bTb0YL/+W8roMJR//Udg7XowV107\nu69PyAzM0l71PxbTxaBWQz46TKUr88RP3xpDe9iHd3ZP361pDflw87qE3vTppawoQ1JQkdmbzpS7\nV76ivVYtOVHWFwdp9AVCLteVJ/V16aXylQCPybwE2eRCyWmmHFBLWObrrPJHXx7CgbMZfGrjkqoF\nR93R0kbS8yAoN+pT0ILyestXamXKzah3lYzPn0SIRzIvOd6mez6yHZQnk0n09FTegu7q6nJto+ds\nYdbMcnlNokMtFyg1eyqKAvn/+xYAgL3lDsrYkbmnZcoXY1AOgKEmT88dGc3hjXNZvH9NvKou94Y1\ncSyLCfjey0Oe/iOcnJEtBoBAXUG5eaAatBGsaPJS5TxpYDoAcnssopYp10p3moMcZAWmS3O0xS9N\nDoJybVa5WaDvlUxRws8PT+C3LojhvRdUf0ZG/Rwifu68CMpnzr4HpstX6smUB3im6hwFpi9mLYNy\ni42w5bPKFzvbQfmaNWvwwx/+EPm8euWby+Xw+OOPY/Xq1bN2cG5gosab1lx7foYBenqhnDoOAFBe\n2AW8vg/MB26Z1Vp2Qsxo8/MXVdlUU0zdsgpQpnwe+LdDYwjwLLaujFX9Gc8yuG3jEgynRfzzwVEP\njk6VzKsBROVIxNKccps15dkamfKAg0x5tqhUTaWZ3urpbrBSFZSXXsdsAktKy5TbbPQE1FnleUmZ\nd5sah1Nq7fJlS8z7TnqiAgbnaZbfCS0bbli+YvMc10zmJMQMxiEC05lyqwksVv0XNKt8mu132Cc/\n+UmcOHEC27dvx6233oqPf/zjGBgYwCc/+cnZPL4FgelRJ7AoUxNQfvQ9YMUaMNdt8/qwyGLVtVwN\nUBdRxphhWXUCC7Covu75aCwr4tmBKbxnZcy03OGSjhA290bxr2+M4UzSm4xkquBOptyq0dPObX3N\nzM2LwOyVr0zl1efTvvZ4wDojny6o89gFzv6d3/IJLPPJUCko72gynwfYHRXm3XHXQy9f4arLVwqy\ns7tUEyaLg4DymnLj55RkBVlRNq8pL2XK59sFnBdqTl/J5/P4yU9+gpMnT2Lt2rX47Gc/q09faW1t\nnYtjnP+W9gK7spAf+RqQzYL92J3qlkFCPMD4/WD/19eA9k6vD2VuxRPAubN60yfxxn8cHockA+9f\nbX2Xcvv6Nuw9lcJ3XxrCX1zbM+elflq2uKm8ppzXMuW1A5aiJKMoK/Zqym0E5TlR1uekayJ+DiwD\nTGTdL18J8KxeyhALWgf/6VLpgZOfUXesFJRP5rG+c/5MQxpOq0F5e9giKI8IeDInqVtMHZTszDdG\n5Sta1ryeTHm7yYWMfvFp8r7Rft+s1Ks1VMqUpylTXjMof/TRR3H06FGsX78eL7zwAlKpFD7xiU/M\nxbEtGEx3LxQAOHwQzI03g+le5vUhkUWO6Vvl9SHMOaYlob4PqXzFMwVJxi/fnsAV3U3oilqPk9Oa\nPh/bN4zf/dFhhHwsgj4WQV79/5CPRUvQhz+8vH1Wlg1pmfJoWdDFcyx41l6mPFO0DjQAZ9NccqKM\nlhlNdCzDIOrnZqV8pXwUpFa+MmmWKS9KCFt8nUZifg5NAjvvMs5DqSICPFPx9c9UPoFldSI4V4fm\nOqM55Y1MX1mdMJ45X+uOUFZ/rxi/jwWORcxvPZZzsagZlB84cABf+9rXEI/H8b73vQ9f+cpXKCif\nqacXYBigcymYbf/T66MhZHFKdKjbcilT7pmn+6cwmZfw3y6018vz/jVx+DgGQ6kiskVZ/Z8oIVOU\nMZwSsXcwrZe6uC1pkCkH1ExizkbAogXlteaUA3Yz5dU15YC21dPdTHlyRlAe9rHwsYxpTXnaoknP\nDMMw6I76511QPpwuoj3ss8z6uxmUT+VEFGUFrU7WZ7okV6op95WVr/j0Rk/75SuSrGAqLyHmNw4Z\ntaDfbCSiNlLT6q5SIkyzygGb5SvxuPoBm0gkkMlkZv2gFhrGHwDzh38MpncVGH7u33iEEIB5z41g\nVl8Cxu+v/WDiOkVR8G9vjWN5sx+XzhgzZ4ZjGWwzKXORFQV/tPMonu6fnLWgPMizVRsIAxxrq3yl\nVvYPcNbomReN69ObA+5nyidzlUE5wzCWr1NvGUdPVMC+0/NrQttQqmhZTw4AHU0CWMadsYgP7x3C\ncLqIB97X2/BzOZUXFfg5BmzZBYhfH4loP1OeKkiQFXVKjxGWUaeymNWU6+8Vi3MoEfLhbJKC8ppB\nuSRJeP311/Vfy7Jc8WsA83+r5xxgr9zi9SEQsqgxTVFglpeFEXOvDWXQP5HHZ65c4kp9OMsw2Lw8\nip++NVZVbuGGZEEy3D4Y8LG2MtsZG0G5s0y5bJopP510Nxk2lZf0Rkz9dYLmGfl0UUabRQ22mZ6o\ngF8fk5AqSI7GKc4WRVEwlCpWzSafyccxWNLkw6ALTcinkwWcSRZd38pqh1HzsJY1LzrIlGvnRbPJ\n9BVAfR+YBeVpG++VRIjH60OU9K0ZlMdiMXz729/Wf93U1FTxa4Zh8A//8A+zc3SEEEIWhJ8dGkfU\nz7ma1d7cG8W/vjmGPSem8L5V7o63TeYlRPzVQUKAZ2xlyrVb9VY15WqW0n5QHjQKykvBsptB3VRe\nQmTGJI3mgHn5QL1BtdbsOThVwJp5UJudKsjIijI6bFxgaHPWGzWSEZETZSQLsusXlrWoF3qV5wzL\nMPCxjKOa8kltm6fJ9BVADbjNznM7F7CJkA/pooxMUbJ8T53vagblDz300FwcByGEkAVKVhS8cjaD\n6/qirjZl9sX96IkK2H3c/aA8VZAqZpRr/JyzmnKrQIMp3davVb5SlBRICqoCKEANhAqSOlLOjWCl\nIMnIidUBYnOAw5HRbNXjFUWpq6YcUGeVA/MnKNfGIZpNESnXHfXjlbMZSLJStQDLrrwo670LQ6kC\nov65/R7kJeM+BYFjHJWv2MmUBy2D8to15doEltGMiFBs8Qbl7re0E0IIWVROTxWQE2Vc0Go8naFe\nDMNgS28Ub5zL4pzL49KSeamqyRNQ68DdqikH1LGItaav5AymZGhqTUZxSgsSZzbtNQd4TOYlSHJl\nsFaQFIiyUldNeUeTDzyrjkWcD4bSaubbbqa8ICkNNR+Wz90e9mDcn1H5CqAG5QUH23Qn9Ey5dVBu\nVr5iZ1JRIkxbPQEKygkhhDToyFgOAHBBi7tBOQBcUyqHeaZ/ytXnTRZk40y5jSAasDd9RfvzWuUr\n1kF5aYa4yWQUp2Zu89TEgzxkRa21L6fVAzvZ5qnhWQZLmubPIp5hR5ny6dKbepUH9FqWfi7lRLli\ncZBG4FkUHZWvSOAYoMniboll+UpBBsuopWFmEg3MKv/BvmHsOj7p+O/NRxSUE0IIaciRsRwEjsHS\nmPuTbzojAla3BvD0gHtBuawoSJs1etquKZfBMai55TLANxqUW2/bdMosKDcL/rVxdvUu0emJuVOb\n7YahVBFhgbVVH+9OUF6WKXcpKE8VJPzVrlO27hzlTcZsOi9fEREL8BVTXGYK+VhkReNzNFOUEPRZ\nL59qDfnAAI7vTKTyEv7fW2P49TEKygkhhBAcHc2hL+6vu/a2li19URwfz+PEhDtlEOmCDFmBSVDO\nImdj26HakFZ7y2XQV7umXF+HbpBJbA5qQblLmfJScG/U6Km+zoxMeUE9NqssqZWeqB9nkgWIsrNl\nNbNBm1FuR8zPISywDQXlo6UAsycquFa+cnAog72DKRwcrj2pJGdavsI6mlM+mRMtmzwBtUzLqnyl\n1p0WnmXQHOQdl6+8MpSGrACDk/Pjwq9RFJQTQgipmyQrODaem5XSFc2mZVGwjLqcyA3aNs9Gy1eC\nNhov7ZSvaOvQjaavxPwcGLgYlJtlyk2C/0Yz5d1RAZICnE15HzTZmVGuYRgG3ZHGsvwjGRERP4el\nMb9r5SsDpQtT7edoJW8wfQUo1ZQ7bPS0avIEgKCPQ6YoQ1Gqn9fueyUR4h2Xr+w/nQYAjGZF0+VF\nCwkF5YQQQup2OllATlSwchaD8uYgj3VLwnh6YMrwH32ntIDGrHylICmQa7xOtijXbPIE1EDbfqa8\n+vk4lkHEz7lWvqI1es68INHLV2YE5Sk9KK83U14qA/E4k6koiqNMOdD4WMTRTBGJEI+OJh+G00VX\nzt3+UlBup/E3LynwcyaNni5nykM+FrICw2DfTqYcKAXlDjLliqJg/5m0/j70ukzqyGiu4eegoJwQ\nQkjdjpaaPGczKAeALb1RDKWKOOzCP3wpy6Bc/WcxX6OEJWM3KPexyDVQUw6oYxHdy5SLaBLYqlKj\nkI+Fj2UwkTVu9GyqcxyjVpvtdbPnZE5CQVJsZ8oB9dhHs6KtOfNGRjIiEiEe7WEfCpLiyoWVlilP\n2siUmy2kUstX7F0gKIpiK1OuvReMSli0Uq9aEmEfzqWLtkudTk0VMJIRsXVlTP21xxd+T7xyruHn\noKCcEEJI3Y6Mzl6TZ7mrljbBxzLY7cKUhaRF+YoWGNcqYbEdlDvIlJtNp2gO8FXBcr3MtqMyDIN4\nkMO4SflKqM5MeVjgEA/yngflQ6WyiI6wUOOR07Q566fr3Ow5khHRGvLpFwKN1pUXJFk/lsm89UWa\nrCgoSIrhOeWkfCUryihISu2acovttep7pfZF3aXtIeQlBa+eTds6tgNn1Mf9zqo4OMb7C79+F3pe\nKCgnhBBSt6NjOfTFA7PW5KkJ+Ths7GnCswPJqlnaTmlZRqM55X6bQXnW5ubBgE/NSlodc61MebOL\nmfLJvISI3zjrGQvwho2eAsdAMCiDsKsnKmBwyttZ5U4WB2n0LH8dc9a1xUGtIV5/zUbryk9NFqCd\nRlM1su5a0G2UKffz9ueUT9pYHARMB+WGmfKCbOuibkNXGGEfi2cGkraObf+ZNLoiArqiApZEa8JN\niAAAIABJREFUBJzy8BybyokYc2FsKQXlhBBC6qI3ebq8NMjMlt4oJvMSXrGZSTOTLEhgYDx7W8ss\n1hqLqDav2cuUA8YZRI027cU8KK8OluuVNMmUa68zOTNTXpTqmlFerieqzip3o6a6XvqMcgc15Z0R\nH1gGGKwjU64tDkqEfPprNjoWUcvELo/5azZ66n0KBhdTPtZ++Yo2IrPZRk05AMNmS7t3lXwciyuX\nRvD8ySSKNWreC5KM14YyWN8VBlA6xzwsX3EjSw5QUE4IIaROg6Umz9mcvFLu8lImrdEpLMm8hLBB\nXTVQXr7iXk05AMsSlrwog4H5zPPmAI+cKNuan17LVM48KI8HOYxXzSmX6568oumOCkgXZNe2ktZj\nKF1A1M/ZupDS+DgW7WFfXQ2E2rztRIhHgGcRC3D6RtF6DUzkIXAMVicCNYPyvEVJlMDbn1M+kbeX\nKdfuGs28+CxKMoqyYuu9AgDXLI8gU5Sx77T1hfcbw1kUJAUbOtWgvDsq4GzKu9GbAxSUE0II8dLR\n0blp8tT4OBbvWhbBcydTDQWoqbxs2OQJ2CtfKUpqva7dmnLAOlOeLTXkmc08bw4aT0ZxSlEUTOUl\n0/rg5gCPqbxUUWqTLkh1T17R9JT6Dcxqfn9+aByf/umxqgsCNw07GIdYrt4JLCNlmXJAzdC7kSlf\nGhMQD/JIzvg5zZS3uPvi59iamWiNlim3M30FqL741MpZ7JR6AcC6JWFE/RyeqbEsbP+ZNHgWWNse\nAgAsjfkhyt5sTgXUn03M5DPFCQrKCSGE1OXIWA5+jtHH3s2FLb1R5EQZLw2m6n6OqYJkutVxevqK\ndRANwF75is1MedBiBblbWz1zooKirJhekDQHeMjKdCMsAKQKMsJ1Tl7R9OgTWKqzif/vzTF856Uh\nnE4WsLeBn2ktQw7HIWq6ogJOTxVqjsicSVsc1FpaH9/R5NObTes1MJ7D8mY/on4OCqabcI1YjdkU\nOAaSAlu9GdrdjVitmnLeuKZ8Oii3F27yLIOrl0Xw4qmU5YXxgTNpXNQW0t9fjdT/u6F/PI/eeOPN\n7hSUE0IIqctcNXmWW9seQjzIY3cDJSwpi7pqbdmKVUCQLdXNOipfqVFTbhQ8aWImM8SdmipN7DCt\nKdcy8mUZ63TR/ALGrtYQDz/HVGXK//ngKB7bN4x3L4ugNcRj3+nZCcolWcG5dJ2Z8oiAvKToNeJ2\naYuDtJ9re2ncn9PgXjOZEzGek9DbHNB/fpMWJSx5yXxLrFYmlbeRLZ/IiYgILPga73GzRk9tI6zd\noBwANi2PIC8p2HvK+HwYy4ron8hjfal0BfB29KYkKzgxmUdvMwXlhBBCPCDJCo6NzV2Tp4ZjGWxa\nFsHLp9MoOthKWC5ZkAzHIQJlmXKL53ZyS14vX7EI8s3WoWu0THmjNdlm2zxnvk55Rl6tKW8sVGAZ\nRi0DKTXiKYqCf3ptBI8fOIfNvVH8ybu7cHlXGK+czcxKTfB4ToQoO2vy1PTESsuPHAZ72uIgTUeT\nD6KMuid0aDXLy5v9etbaqq7cqnxFm6Rjp9lzIifVzJKrz8mAY6ovPrXGTycjNS9uUy+8nz1hfOGt\njUIsD8qbBA7xAOdJUH4mWUBBUtAbb/yzkIJyQgghjg1OFZCXZneTp5kVLQGIslJ341wyLxmOQwTs\nzSnXgnJH5SuWmfJaQXl1BrseST0oNw6ypoNy9XUURSnVlDdeK9sT9esTWJ54ZQT/59URXL8iis+/\nqxMcy2BDVxMyRRmHzmUbfq2ZtDrj+mrK1eyn06B8JCOiNTj9fW50Akt5UK5dVFmNRbSavqJlyu1c\n1E7mxJqTVwB1zn3Ix+p3kTTae8VJCZR+4T2YNizR2X86jeYAV1Uu0hPzezJ6U5u8QplyQgghnjhS\n2uQ5V5NXymm3qk/XkRUTZQWZonmjpx6UWzVmaoGGjeyfrZGIRdl0cRCgNriGBdaF8hXrTHm8VL6i\nNVzmJQWSYjw60qnumIBz6SK+9/Iw/u/BUbx3ZQx3XtWplz5dtiQEjgH2nWls3KWR4TpmlGviAQ5B\nnnUc7I1kRCTKMvMdTeo5W28jotZI2BzgEC0FydaZcnfKVybz9jLlABD0caY15U6m3gDANb1RFGUF\nL8woYZEVBQfOpvGOJWGwMxqjvRq92T+eB8tM31VpBAXlhBBCHNOaPLvnsMlT0x2pr6QAAFIW2zwB\ntdGMZ2chU27Z6GldUw4AMX/js8prBeVBnoXAMfrraFlKdzLlAhQAPzs0jt9Z1Yzbr1xSEVSFfBwu\nagvi5VmoK9caLOspX2G00hsH51r54iBNW1j973q3eg5M5LG82Q+GYcpqys0v0qxm3wul3yvUGPsJ\nqP0XZhewMwV9rEFQXjqHHAblq1sDaA/zeHbGFJZjY3lM5SV9Pnk5bfSmWzP97eqfyKMnKjS0YEtD\nQTkhhBDHjo7msKJlbps8NU1+DjE/V9f681QpMLUKNPw8i5zFrX199byNW/ICx4BlrDPvOUnWM+pm\n3NjqOZWTwDLm9b0Mw1S8jtak12hNOQCsKNXb3nhhHLdt7KjKcgLA+q4mHB/P2667fns0q9cXWxlK\nFREP8nUHTU6D8tEZ4xABtY67JcjXlSmXFQUnJvJYXirXEDgWAZ61mSk3CMpL79laNeWKoiBVsL88\nSi1fMcuUO7uwYxgGm5ZHceBMuuLr3H9GvWh7R2d1UD49enNuS1gGJnLobXbnjiEF5YQQQhyRZAXH\nx3OelK5ouuqcH63VVTdZBJoBjrUeiehgzBvDMAjyrHWjZ1GumSlvDrqTKY/4OcOAWH+dAK/Xrqdc\nzJR3RQU8/j8uwB9d3mE6j/3yUvZzv41suawoePA3p/G/f3O65kST4XQRHXVkyTXdUQHnMiJyBtsq\njZQvDirXHvbVlSkfShWRl5SKmuWon0PS4nzQSlOMFlIJvBaUW5evaOVLdqfvhIwy5QUZAsfAZ7IY\ny8o1y6OQFOD5k0n99/afSWNF3G+4zEgfvTmHmz3TBQnDaVG/YGoUBeWEEEIcOeVhk6emKyLUVVM+\nVbCZKa9RvsIygN9moBHgqzOI5dRGT+vnciVTnhdNS1f01ykL/rVMudUFjBPRGrXJvc1+xIM8Xq6x\nzREAXjmbwZlkEZN5CcfHrTOjw6lCXfXkGq1E6+SEvSbUmYuDNB1Nvroy5f3j002emliAsxyJmBOV\n0l0ao5pye9NXnJYvBX3VF5+Zouy4nlzTF/ejKyLgmdL400xRwlvnshVTV8q1hngEeKaui/V6DbjY\n5AlQUE4IIcSho6Umz5VzPA6xXHdUwHhO0mtW7UrVqKsG1NXkVpnyjKgGGmYZ35mMghWNrCjISwoC\nNQKX5gCPdEG2vYnRyJTFfPbp1ykrX9HrgRvPlNvBMAw2dIZx4Gy65mKb/zg8rpfV7LcoYZFkBSMZ\nseFMOQCcGLcXlM9cHKRpD/swkinaWtpTbmAiDwbAslhlprxW+YrZRB/tYrJWpjzl8KIsyBvXlNfb\nKMwwDK7pjeC1oQzGsiJeO5uBpBiXrgDTozdPzmFQrk9eoUw5IYQQLxwZzSLAM3rDpRe6ovU1e2rb\nKq1uyQd4Vm+UM5ItSgjVKDcpFzSotdXo86Rr1Du7sdXTXlDOY6q0wt3NmnK7Lu8KI12QcXjUPAAe\nyRSxdzCF376gGX1xv2VQPpIpQlbqG4eo6YoIYOAkKFcX7swsSepo8kFWpstb7OqfyKMz4qt4vqif\nw5TFnZO8ZH73ZbYy5WYjEe30XpjZtDwKBcCeE1PYfyaNAM/goraQ6eO7o34MzuFWz/7xPJoEtmL8\nZSMoKCeEEOLIkbE8VszxJs+ZtAsCpyUsybxaemJVD26nfMVJoBG0eD4tI187U974Vk81KLcOHpoD\nPGRFrb130tDqlsuWhMEywD6LEpb/PDIBRQHet6oZ6zvDeOtcxvSiZ6iBcYgaP8+iLczbDspHMsWK\ncYga7Ric1pUPTOSwfEYjYa1MuVq+YnxO+Th7jZ7TPQV2Gz055ESl4k6A+l6pP9RcFvOjt9mPZ/qT\n2H8mjUs7Qpb16T1a/b/F+9dN/RPqJk+7d81qoaCcEEKIbVqTp5elKwDQGfGBARxPYEnm1W2eVv+I\nBnjrRs9MUXa0odAqU64FD1bLgwC11huof6unrChI2smUB6eD/3RRhr/OJr16Nfk5rEkETYNyUVbw\nn0cmsaErjI4mAe/oDEOUgdeGjB+vBcCNlK8AQFfUjwHbQblomDnVjsFJXXlelHEmWayqWY4GeOQl\nxfQ8tVO+YnWOA+U9BfZryoHKcaKZgrP3ipFNyyN4aySLs6ki1nc2WT5WmxVeT7+JU7KiqKMqXdjk\nqaGgnBBCiG0nJ/MoSIqnk1cAdaFOe5PPcflKqlB77nKAZywzbVmH2T+r6Ss5iyUv5RrNlGcKMmQF\n+uIZM/FSmcx4TlLH4bkwecWpDV1hHBnLGX6tL5xKYjwr4ndWxQEAF7cF4ecY09GIQ6kiWAaGmWsn\nuqMCToxnbS2mmbk4SJMI+8AyzjLlJybzUFDZ5AlM90SYZcvzFs3D2pzyWhs99Uy5zXNdC8rL68oz\nRamhTDmglrBozJo8NT1RbSzi7Aflw6kicqLsWpMnQEE5IYQQB/QmT4+DcqA0gaWeTHnNoNy6ptzp\nRImAZaZcfZ3ac8pLwbLNGd4z1VocNPN1JrIi0gVpTuvJNRtK2dD9BtnyXxyeQHuYx4bS+EQfx+KS\njpBpXflQqojWIA++wVKr7oiAbFGqOUPdaHGQhmcZtDqcVT5g0kgY0xYImdw5yVkspBJslq+k9c21\n9mvKgcrttY3WlANAZ0TAqtYAljT50BmxvrjqjKgXPifnoK683+XJKwAF5YQQQhw4MpZDgGfR5WGT\np0Zd6lJ0tFY7WZBq3o6vWVNecJb9C1qMRMxZLHmZeUyJEI8Tdc5gth2Ul5evFOQ5m7xSbkWLH7EA\nh30zAu1Tk3m8OpTBb13QXNHPsL4zjNPJIoZS1d+b4XSxoSZP/ZhKQfGR0kWpGS1onzkOUdPe5MOw\ng6C8fyIPP8dUfQ3TmXLji4S8ZD77nmUY8CyjzzI3kypICPCs7d6R0IxMuaIoDdeUa76wqQv3XNtT\ns3Zb4Fi0h53fQatHvzYVh4JyQgghXjg6lsOKuN/TJk9NV0RATpRtb4AE7GbKGRQkxXQpjeNGTx+L\noqxANBiFZ7emHAD64gEcH7cOCs1owVutrz3IsxA4BhM5Cemi5NqMcidYhsH6zjD2n6kcjfiLtyfA\ns8B7VzZXPF4raTCqQx9KFRtq8tSsaAmAZxkcHrH+/p9LGy8O0nQ0+TDkoHxlYDyPZc3+qnnj2sz3\nespXADVbXnv6iuzo5z9dvqIeU1aUocDekq1aOpoELI3ZC36XxoQ5KV/pH1en4th579pFQTkhhBBb\n1CbPPC7wuMlTo82PdlLCojZ61s5KA9PjCstJsjpX3En5it4AZ5AtdxaU+zE4VajZoGfEbqacYRh9\nq2e6IHtSUw4Al3c1IZmX9HKpvCjjqeOTeNfSiN70qumOCmgL8ThwtjIoL0rqBVtHuPG7On6exaq2\nMN4asW72HC0tDmo1y5SHfRjLiDXruTUDE/mqenKgdk15TlTgtxizKXBMzWNIO+wp0C5Utf6J6c23\nc3sOdUf9OD1VcDwP3imjqTiNoqCcEEKILVqT53yoJwemg3K7t6oLkoy8pNiqKQeMp1NMBxrOylcA\nGDZ7Tgflte88rIgHICtq859T00F57XnK2gIhr2rKAeAdS0JgMJ39fmZgCumCrDd4lmMYBuu7wnj1\nbKbibsRwWg2Q3ciUA8DaJREcGc1aBnvaDHLzTLkABfZmlU9kRUzmJcOa5bDAgmXMa8rzonn5CqCW\nedQqX0kXnN0p0c/z0nskXcd7xQ09UQFFWanZUNvI2MScNhXHpaVBGgrKCSGE2KLV03o9eUXTGuIh\ncIzt8WfJfO3FQcB0UG70j3amnqDcoAFOo2Xja9WUA2qmHEDNtfJGpnISfCxjK/iPB3mMZyWki97U\nlANqecaq1gBePp0CAPz74Qksiwm4uD1o+Pj1nWFkijIOl2WytRrzRschatYuiSAnKpYXRWaLgzTt\nDsYiao2ERplylmEQ8XP6OV1O3xLbYPlKyuGdkpk15RkPlk8BalAOWF+sv3Uuiw//+DAODmXqeo0T\nE+pUHDebPIF5FJQfOHAAn/vc53DnnXdi586dVX9eLBaxY8cO3HnnnfjiF7+I4eFhD46SEEIWryOj\npSbPqPdNnoAamDiZwJK0WcKhjSc0DsrV53A0fcVWprz287U3+RDysThWo9nQiLbN086Sk+YAj7Op\nAmRl7gOqcpd3NeHt0Rz2nU7h6FgO71sVNz3+daWlQ+VTWLRMqWuZ8k51NN8hixIWs8VBmg4HC4QG\nakz3iPk5TBo0emrBtlX5ip9nUKg5p9xhpnxmUF7He8UN3TFtLKL5xdP/efUcJAVVJU92zcbkFWCe\nBOWyLOPRRx/FF7/4RezYsQO/+c1vcOrUqYrHPPXUUwiHw/jmN7+JG264AU888YRHR0sIIYtTWODw\nzu6mqqYzL3VFBdvlK6lS5s5OsyMA5A0yiVq220kG2SpTnhNl8Cxja2QfyzDoi/vry5TnpZozyjWx\nAKd/7V7VlAPA+q4wFADffP4sAjyD61ZETR/bJHBY1RqsCMqHUkXwLNDi0gr0rqgfMT+HQxbNnmaL\ngzQtQR4cYz9THg9welPnTFE/hymD8pW8jYk+Ppa1NRLRyXnOsQz8HKOf55k63ituiPo5xPwcTplM\nKnrrXBYHzqoZcqsLLCv9E3kEeNa1Cz7NvAjKjxw5giVLlqCjowM8z+Pqq6/G3r17Kx7z0ksv4dpr\nrwUAXHXVVXj99dcdjcEihBDSmFve0YY/2dTl9WFU6IoIOJsqGk42mclu+YrfRvmKo0ZP3joot1NS\noumLB9A/kTOdDGNmysY2T028LKj0MlN+QUsAET+HsayILb2xmg2DGzrDODKa0+vnh9NFJEI+1yYF\nMQyDNW1By0BuNCOaNnkCauDaFrY3FnFgIme5LTIa4A0bPe0spBJ46/IVSVbHGdrd5qkp315bz3vF\nLd0WF+v/9NoIon4O1/ZFcXgkV1dD6MB4DssNpuI0al4E5WNjY2htbdV/3drairGxMdPHcByHUCiE\nZDI5p8dJCCFkfumOCpAV4KzBjOqZkqUNhXYbPY2mpdTTvKZnyg3LVxRHI9X64n7kRAVnkvbH6gFA\nMi/aDsqbyzLqToMyN3Esg/VL1HGHv7O6ucajpzPrr5Sy5UMpd2aUl1vTGsTgVMGwljsvypjKS0iE\nrTPz7eHaYxElWcHJyYJleUTUzxkG5VqfgtV55ecYFCwaPfUst8OLspCP1ctWtP/34sKuJybgpEFQ\nfmgki/1n0vjvF7XgsiVhZEXZ8fhERVHQP5F3vXQFANy5pzOPPPnkk3jyyScBAPfddx8SiYTHR0Tm\nAs/z9LMmlugcOT9dXBSA584ghQASiVbLx8r9atlBb1c7ggZZV+0cSXMZAP3wBcNV5wx3Rq3h7elI\nINFk7x9lLlQAcAycP1j1fAp7DmG/z/a5uUEOAM+fxYjow2UOzudk4Qjam5tsvc7yvA/AaQBAd1sr\nEokm26/jtk9eE8LGvklsXNVZ87HxFgWRXafw5riID1yRwLnMUVyzIuba+57neWxc2YHHXzmHoaKA\nvu7KSTCnJtQMem973PI1lyXG8dzxMcvHDIxnUJAUrO1pNX3ckngaybcnEG9prbgbcE5UE5btLc1I\nJFoM/24kNILhTNr0uXOlr2VJq7PvXyR4CiJT+qzl02AA9Cxpn/OStzWdefznkUnw4Riag9MXZvf9\n5iCiAR63vGslRtNF/N1zZzCY43C5g69xOJlHqiDjEoufTb3mRVDe0tKC0dFR/dejo6NoaWkxfExr\nayskSUImk0EkEql6rq1bt2Lr1q36r0dGRmbvwMm8kUgk6GdNLNE5cn4Ky2o27q3BEVwYtb4NfXZs\nCj6WQWpiDGmDIEE7R7KlcXUjE1MYGanM8g2PTwIAsskJjOTsZZG1Gt9zE8mqc3Aqk4OPUWyfmxFF\nBscAr54YwWUt9gIdSVaQzIsQ5KKt12EL05nDYmYKIzUW5symCIDN3T7b359LO0J4/vgYTp0dxkS2\niBgvufa+TyQSaOPzYBlg77GzuKCpMkt9uNQ0GJBzlq8Z4ySMZooYPDtsWvd9YGBKfU3e/GfGS+oE\nkIHTQxV152dH1ePIpZMYGTHOhitiAdmC+XOfGlV/5nI+4+j7JzAyJtPq1z8ymULQx2KsLL6bK3FO\nfQ+/evwMLm4PAQDeHs3iuf5xfPSyBLJTEwgqCiICi5f7z+HqJfbD4X2D6kSghK/6+9fV1Vh537wo\nX1m5ciXOnDmD4eFhiKKIPXv24Iorrqh4zOWXX45du3YBAJ5//nmsXbvWVhc5IYSQ81fEzyHq52w1\neyYL6jbPWv92BDjrmnIG9qalaASOAcu4U1Pu41gsjfkdbfa0W7ajiZWVr3jZ6FmP9Z1hjGZFvFwK\nnNpdGoeoCfk4LIv58ZbBhUqtxUEarTnwnEUJS/+EGvz3xMwnHcVKgfjkjBIWO+Ur6pxy84vYVMFe\n/8VMIR+rl2llirIn9eTA9PetvDTlR6+NoElgccMa9Q4HwzBYnbDuETDSP24+qrJR8yIo5zgOn/jE\nJ/DXf/3XuOuuu/Cud70LS5cuxY9+9CO89NJLAIDrr78eqVQKd955J372s5/hIx/5iMdHTQghZD7o\nigi2ZpWr2zxrBxlWjZ7ZUqDh5HY8wzAI8qzpSEQ7M8rL9cX9OOZgAovdbZ6aIM9C4NSvL+xRUFWv\n9Z1qDfov3p4AoC7rcduaRBBvj2Srmm1rLQ7SaHPTrcYiDkzk0RURIFiMNTTb6pm3MWaz1kbPtFYP\n7vDnH+TZipGIXp0/iZAPAsfgVGmm/JHRHPYOpvHfL2ypaBhekwji5GRBvwixo38ih/YwPysXrPOi\nfAUANmzYgA0bNlT83u/93u/p/y0IAv74j/94rg+LEELIPNcVFSpG4ZlJFSRE/LWDBB/HgGOmM47l\n6s3+BcqmUpRz2ugJACtaAviv41OYyIpVK+eNJEtj8+yORGQYBvEgj8mc5NrkkrnSFvahJyrg1dJS\nGLcbPQFgTSKAXx6ZwOBUAUtj09nSWouDNFqm3Gos4sBEvubmXD0onzEW0db0lRqNnml98Y/z6SuZ\nso2etSbmzBaOVXcYaJnyH70+gnBZllyzJqEuozoymsM7Shd0tfRP5LG8eXYWqC2sS2BCCCFkhu6o\ngPGsqE97MDOWFW2XcAR41rR8pZ614UGT51PLV5xnygHg+IS9bLnTTDmgTmDxchxiI7RsucAxFZNk\n3KIFcjPLHkZqjEPUxIM8fCxjmil/9WwaZ1PF2kF5wCRTLtXeEivwLEQZpuMAGypfKb0Ps3W+V9zS\nE1PHIh4by+HFUyn8twtbqi4yVrUGwMD+vPKCJGNwynoqTiMW5juOEEIIKemOqCUKp6fMM4/Hx3M4\nkyzi0g572TCzoDxblOrK/gVNM+XOasoBoK+UpTtuc7NnPUF5Z5OANhsB5nykBeXtYd+s9J51RQWE\nBRaHZ9SVj2SKNUtXAHUJVFvYZ5gpH8kU8cCzp9ETFbBtddzgb0/Tfp4zt3ra2RIrlO6AmM0qTxdk\nsAwcn5shHwdRBoqSjHRBRsjDC7ueqIChVBGPHziHsI/F+9dUfz/DAoelMcF2UH5qUt102xufnaB8\n3pSvEEIIIfXoipaC8mQBF7QaZxefOjYJngWu6TXfClnOz7N6bW65TFGuq5bULCjP15Epb/JzaA/z\nOGaz2VML2uzeJQCAWzd22FrINB9d0hGCj2Vcb/LUsAyD1a1BvDUjkBvNiFjdGrT1HB1NvqpMeVFS\n8PVnBpGXFPzN5u6aZVICxyLAs6Y15VpfgOHfLQXbRcm4HCtdkNAk1G6Knkl7rkxRLl3AeheUd0f9\nUADsO5PG71/aapr1X50I4vmTSSiKUvPr7S/dnaJMOSGEEGKgM+IDA2BwyricQ5QV7O6fwsbuJtvZ\n4gDPuF6+MrPRsygpEGVnk1w0ffEAjtts9pzKSwjwrGXT4ExNAodmk/Xu852fZ7F9Q1tV/bCb1iQC\nODGR10um7C4O0rQbZMof2zeEQyM5fPZdS9ATsxf0xQJcVU15XlRKE3+sasrVc8FsAkuqINVVvlQe\nlHtZUw6omXJALam5cY3xvHZALUdKFWSctrGQ681zGQR4Bp0R9xuIAQrKCSGELHACx6It7DMtX9l3\nOoXJnITrV8RsP6davuJeo2eQr86U5/WGvHqCcj8GpwqGFw4zJXNSxZjDxeD9a1pwRffsLT1akwhC\ngdogCKj9CoA69cOOjiYfpvKSfk7sOj6Jfz88gZsuasG7l9m7mwMYb/W006egZdGtylfq2eaqXbCm\nChIKkuJxplxA2MfiAxe1oMniYvxCkx6BmfKijN8MJHHV0sisNUBTUE4IIWTB644KGEwaZ46fOjaF\nmJ/Dhi77QZrf7UZPX3WmPFeaflFPkN8XD6iLY2w0e07lJUf15KQ2rUxFC+TsjkPUtJeNRewfz+Gh\nF85ibXsQH3tHm6PjUIPyypryvCTDb1G6AgD+UqbcbAJLus5xhtp7Y6Q0s93LoNzPs/juTSvxoUus\nN/32xASEfCwO1wjKXziVQroo4z0OLu6doqCcEELIgtcVFTA4VYQyY3Z0Mi9h72AKm/ui4B1ktwI8\nU1VTLskKcmL9QXluRqZc+3WtAMqIPoHFRl05BeXua/Jz6IlONwiOpO0tDtJooxqPj+dw3zODCAsc\nvrCp23EGNuo3Ll+pdffFVyNTnirU3zsBAKOlixQvg3JAbeSsVSfOMgxWtQaqegRm+q9jk2gL8bik\nI+TmIVYey6w9MyGEEDJHuiMCcqKM8RkByjMDUxBlBdf3Octu+Q3KV7TMeV3TV3gWRVlHkr26AAAg\nAElEQVSpaJ7M2di8aKY97ENYYG3VlU/lJUdNnsQedRtkDoqi6Ns8bWfKS0H5I3uHMJwq4n9t6kLc\nxsz5mWIBvmqjp7PyFZNMeanR0yktKNcuUkILZCPsmkQQAxN503Kw0UwRB86mcd2KmKPFYU5RUE4I\nIWTB6y41dc1s9nzq2CR6m/1YUWPm80wBg+kr2lKUumrKS3+nvK5c37xYx/MxDFNq9qRMuVfWJAKY\nyks4mypiJFO0tThIE/Nz8HMMMkUZ2ze04+L2+rKvET+HgqRUnKt5UbZcHARM9zEUDPomFEUpZcrr\nKV9RzzOtnGehbIRdkwhCVqZ7BGbadXwKsgJHfSn1WBjfLUIIIcRCl8Gs8lOTebw9mqvrH1KjOeVa\nQF1PoKFlLsuD8qyNedJW+uJ+9I/nTRfAAGomNCfKFJTPgvIlQnYXB2kYhsElHSG8Z0UMNzYwJSam\nzSovu0OUl2pvibVq9CxI6h2despXQnr5ipopr+cC1gurLZo9FUXBU8cmcXFbcNamrmgW5rwjQggh\npEwizEPgGJxOFvTfe+rYJFgG2GJzNnm5AM8gLymQFUW/Xe1Kplw0yJTXGZSviAeQlxScSRXQEzUe\noZfUFwfRP/duWxbzI8AzODyStb04qNxfXNvT8HIj7WJrKi/pJTE5Ua45BWZ6JGJ1ucb0Nk/n56Wf\nY8Ay042eYQ9HIjoR9XPoivgMg/LDozmcmirgM1cumfXjWBiXMIQQQogFllFnB2vlK5KsYNfxKWzo\nDKO5jlpd/fZ+WSZRm0ldb005gIrs+/TmxfoCM63Z89iYeV15Pds8iT0cy2BVaxBvjeQw6jBTDsCV\nbaPRgBaUT09gydvYEmuVKU/rd4ScnzMMwyDIsxjLzo9GTyfUHoFsVbP4U8cmIXAM3r08MuvHsHC+\nW4QQQoiFrog6gQUAXh3KYDQr4vqV9dWABgyCaC1TXu/0FaCyfEVr9KxnTjkA9ET94FnrCSwUlM+u\nNYkgjo/nHC0OcpN2B6R8VnletF++UjQKykvPZTXb20rQx0J724TqyLZ7ZU0iiImcVLFptSDJeGZg\nClcvjczJIqSF890ihBBCLHRHBQylChBltQa0SWDxzjoXyOhBebE6KHer0VML+IN1BuU+jsHSmN9y\nAsurZzMAUNdkD1Lb6kQAWkm/3cVBbjKqKc+Jcs0LPavylXQDvRPA9EUrzwK+WVqyMxumlwhNX+S+\neCqFdEGu++LeKQrKCSGEnBe6owIkRc0cP38yiWuWR+FzsFq+nDa9ojxTnm0kU85X15TnRBkMprOW\n9VhhMYHlxVNJ/OTgKLb0RtEZmfuAcTFYU1oiBACtDmvK3RASWLDMdKZcVhTkJaXm9BWr8hWtprye\nRk8ACJYyyiFf7Rnh88nyZj8EjqmoK3/q2CQSIR6XzuJs8nIUlBNCCDkvaBNY/u/royhICq5rYHxZ\nQM8kVteU19OYGTDJlPt5pqHApS/ux0ROwni2cqvjqak8duw5gxUtftxx5ZIFFRwtJM1BXl8E5EWm\nnGWYiq2eWjlKoMbFKMcy4FmgYDCXO11Qf6+eRk9g+q7QQqonB7QegYAelI9mith/Jo3rZ3k2ebmF\n9R0jhBBCTHSVZpW/cCqF7qiA1a3OZpOXM6spD/Cs462LQFmmvGJOee3a31pWxNWvsTxbni5I+Jvd\ng/CxDP58c0/dNevEHm00otPpK25Rg3L1glE7X+38zH0si4LBOM1GM+WhBRqUA9M9AgVJxu7SbPLr\nHC4ea8TC+44RQgghBqJ+Tt9ceX1frKHssN8kKK830BBKo+LKy1eyNjYv1tKrTWAp1ZXLioIde87g\nbLKAP7umG21hKluZbdtWN+N/rm317OInGuAxVaopz+vNw7XPfYFnDJcHpQsSAjwDvs56cD0oXyDb\nPMutSQQhysDRsRx+fWwSF7UF9Yv9uUBBOSGEkPNGV0QAA+DaFc5nk5fTRsrly4KWbANBOcMwCPrY\nqjnljQZyYYFDR5NPz5T/02sj2DuYwh9e3oG1c1QHu9hd1BbCLe9o8+z1KzLlkv3Z936OQcFwTrnc\n0HzxhVq+AkwvEfr5oXGcmirM+gbPmagdmxBCyHnjur4oLmoLNlzfq9WAu5UpB9QSlpk15Y1mygG1\nrvzYWB7PnUziR6+NYuvKGLatbm74ecnCEPNzmMxrmfJS+YqNBmeBY03mlEtoaiDLvZDLV1qCPNrD\nPJ4ZSKqzyZfN/mzychSUE0IIOW/8zur6V5aX0xs9K4JyqbGg3DdbQXkAL5xM4Rt7zmB1awC3beyg\nxs5FJOLnkMpLkGTFUfmKzyRTni7ICDcwX1zrn1iIQTmgZsuH00m8a2mk7rr6ei3M7xghhBAyi8xq\nyoMN3NYP8OyMkYiNN3oCaqZcARDkGdy9uVufQU0Wh1iAgwK1QXN6S2z9mfJUQWooGA2VjURciLR5\n5XNdugJQppwQQgip4uMYcMz01k3AhfKVGZnyvEuZ8rVtIaxbEsJH1rU5XvVOFr7yrZ55B9NX1Jpy\no0ZPGcubGzvPgYWbKd+6shkRP4fLlsx9TwYF5YQQQoiBAM9WLQ9qtKZ8Mje9wtuN6SuAug793vcs\na/h5yMIULU0cmspJ+lz9gJ3pKxyDdFGq+v10YfHWlAPqRcW1czgGsdzC/I4RQgghs8xfFpTLioJs\nUdazgPUwzpRT7TdpjB6U56fLV2w1evJsxXQhQD3PM8UGa8pL75G5rsc+H1BQTgghhBgI8IxeDpAT\nZShoLPsXLKsplxW1KY8W+5BGxQJq8DuZFx0tDxLY6vKVTEE9zxvJlHdHBCRCPHqb/XU/x2JF5SuE\nEEKIATVTrgYtWoa7kea18kx5QVKgwF5DHiFWyjPlculGjN3lQcUZ01e0cpZwAxefzUEej37ggrr/\n/mJGQTkhhBBiIMCzeqY8owfljWXKRVlBUVIcTckgxIqPYxHkWUzlJPAsU9oea6emvHr6SqqgnpdU\neuIN+jQghBBCDJQ3eroSlJf+blaU9WCfasqJG6IBTq8pt1sSJXCM3hiqSRfUTHkj5SukfhSUE0II\nIQbUmnI1aHEzKM8VZb2MhTLlxA3R0lbPvKQgwNm70PNz6p0bSZ4OzFOloLyRRk9SP/quE0IIIQb8\nPIucpGXK1WCloekrfFmmXB9dR/8Mk8ZF/RySeRF5B5lyXyl4F8uC8nSpfIUy5d6gTwNCCCHEQHn5\niluNntpzUaacuCkW4DCZkxwF5UIpKC8vYaFMubfou04IIYQYCPAsckV3Gz2ByppyGolI3BD182pN\nuaTY7lPQzr1C2QSWdEEGy0yfq2Ru0XedEEIIMeDn1UY4baEK0Fj5SkDPlE8veQn4qNGTNC7q51CQ\nFEzlRFuLg4DpTHmhbIFQuigh7GPB2JjeQtxHQTkhhBBiIMBpmUR1m6efY8Cx9Qcreqa8KOvzz6l8\nhbhBm1U+nC46Ll8pz5SnCjKNQ/QQzSknhBBCDGjBTU6UkSlKCDUYrOjTV0RFb66joJy4IVra6pkT\n7ZevCGUXnZp0QaKg3EMUlBNCCCEGtOAmL8pIF+SG6smBykZPWaGgnLhHy5QD9vsUpjPl5Y2eMpqo\nydMz9J0nhBBCDAT46cx2tth4UO5jGXCM2uiZFWXwLMA3UA5DiCbmn86x2r3Qm86Ulzd6UqbcS5Qp\nJ4QQQgwEKspX5IaaPAGAYRgEfSyypZnnlCUnbinPlAs2lwcZjURMFyTKlHuIvvOEEEKIAS1ozouy\nK5ly7TmzotroSeMQiVvCAgvtpovtTDmvTV8py5QXZYQbmMVPGkOfCIQQQoiBqkZPF4JyNVMuIyfK\nlCknrmEYRs+W2w7KWfVxxVLTcUGSUZAU2ubpIfpEIIQQQgxojZ45UUFGlBF0IYMY5FlkRYWCcuI6\nra7cb3f6it7IrAbl6YKaMadtnt6h7zwhhBBioDxTni3KCLkQRFdmyqnJk7gnUhqL6Hh5UKnRM1VQ\nex2o0dM7FJQTQgghBrRM9mROhKzAtfKVXGl5EGXKiZtiDstX/DPmlGuZcmr09A5NXyGEEEIMaJns\n8awIAAi5EKyo5SsSBIVFgPc1/HyEaLSacrvlKxzLgGWmg3LKlHuPgnJCCCHEAF8KWsayarAScqGm\nPMCr5SuSYn/JCyF2aFs9nZxXAsfq5StpPSin89IrFJQTQgghBhiGQYBnMaZlyt2aviIqUEA15cRd\nTqevAICfY6bLV4ql8hUaiegZCsoJIYQQEwGe1ctXGl0epD2HKCvIFKmmnLhrVWsQbSEerSH7oZ3A\nMQaNnnReeoWCckIIIcREgGdwLu1iprwUiMsKbfQk7lqTCOJ7H7jA0d8ReLZiJKLAMfDZnN5C3Eff\neUIIIcSEn2f15Spula9oKCgnXhM4Rj+/UwWJFgd5jD4RCCGEEBPlgbMry4MoKCfziMAxKIhao6dM\npSseo+8+IYQQYqJ8koWb5Svqc1OjJ/GWOn1FK1+hTLnXPK8pT6VS2LFjB86dO4e2tjbcddddaGpq\nqnhMf38/vvvd7yKbzYJlWXzwgx/E1Vdf7dERE0IIWSy0CSkCx4BnGw+iy4PyIGXKiccEjsFEaepK\nuighHvA8LFzUPP/u79y5E5deeiluuukm7Ny5Ezt37sRHP/rRiscIgoDPfOYz6OzsxNjYGO6++25c\ndtllCIfDHh01IYSQxSBQanpzI0sOVJav0Jxy4jU1U14EAKQKMpZGKVPuJc8/Efbu3YstW7YAALZs\n2YK9e/dWPaarqwudnZ0AgJaWFsRiMUxNTc3pcRJCCFl8tMB5NoJyqiknXhM4BsWy8hWqKfeW55ny\nyclJxONxAEBzczMmJyctH3/kyBGIooiOjg7DP3/yySfx5JNPAgDuu+8+JBIJdw+YzEs8z9PPmlii\nc4TUYnSOxCNTACYQCfpdOX/4cBHAMQBAZ1sLEgm647tQnI+fIdHwOIrncmhpbUW6IKOtOXLefY0L\nyZwE5ffeey8mJiaqfv/3f//3K37NMAwYxrxmb3x8HN/85jdxxx13gGWNr+a2bt2KrVu36r8eGRmp\n86jJQpJIJOhnTSzROUJqMTpHFDEPABAY2ZXzp1ha1AIA2eQkRpBt+DnJ3DgfP0NksYBcUcTJM8NQ\nALBi/rz7GudSV1dXQ39/ToLyv/iLvzD9s1gshvHxccTjcYyPjyMajRo+LpPJ4L777sPNN9+M1atX\nz9ahEkIIITq3a8p5lgHHABItDyLzgJ9jUBAV2uY5T3j+3b/iiiuwe/duAMDu3buxcePGqseIoogH\nHngAmzdvxlVXXTXXh0gIIWSR0mrKgy4F5QzD6M8VcOk5CamXtjwoXVDv4IRpJKKnPK8pv+mmm7Bj\nxw489dRT+khEADh69Ch+9atf4VOf+hT27NmDN998E8lkErt27QIA3HHHHejt7fXuwAkhhJz3tJGI\nbmXKAXUUYqq00pwQL/lKd4LGsyIAIEwXip7yPCiPRCL48pe/XPX7K1euxMqVKwEAmzdvxubNm+f6\n0AghhCxyAX36insZxKCPhZ9jwFr0UBEyF/ylC8OxUlBOy4O8RZdEhBBCiAm3RyICalBOpStkPhBm\nZsqpptxT9N0nhBBCTARmIyjnWWryJPOCQJnyecXz8hVCCCFkvvKXasrdavQEgI4mAaV9LYR4SgvK\nx3MiGLh7nhPnKCgnhBBCTCyN+fHOniZc3BZy7Tn/6Ip2yBSUk3mgvHwlLLDU5+AxCsoJIYQQEwGe\nxZe29Lj6nFogRIjXhNKdoLGMSOMQ5wH6ZCCEEEIIWYSmy1ckNFGTp+foJ0AIIYQQsgj5S3dtRFlB\n2MWxn6Q+FJQTQgghhCxCvrIFVlS+4j0KygkhhBBCFiGhIiinkNBr9BMghBBCCFmE/GVNxzSj3HsU\nlBNCCCGELELa9BWAMuXzAf0ECCGEEEIWofLxnNTo6T0KygkhhBBCFiGOAdhSspxGInqPfgKEEEII\nIYsQwzB6sydNX/EeBeWEEEIIIYuUVsJCjZ7eo6CcEEIIIWSRms6UU0joNfoJEEIIIYQsUpQpnz8o\nKCeEEEIIWaQoUz5/0E+AEEIIIWSREjim9D8KCb1GPwFCCCGEkEVK4FmEfRQOzgf0UyCEEEIIWaT8\nHEPjEOcJ3usDIIQQQggh3njvBc1I5iWvD4OAgnJCCCGEkEXrXUsjXh8CKaHyFUIIIYQQQjxGQTkh\nhBBCCCEeo6CcEEIIIYQQj1FQTgghhBBCiMcoKCeEEEIIIcRjFJQTQgghhBDiMQrKCSGEEEII8RgF\n5YQQQgghhHiMgnJCCCGEEEI8RkE5IYQQQgghHqOgnBBCCCGEEI9RUE4IIYQQQojHKCgnhBBCCCHE\nY4yiKIrXB0EIIYQQQshiRplycl64++67vT4EMs/ROUJqoXOEWKHzg9TS6DlCQTkhhBBCCCEeo6Cc\nEEIIIYQQj1FQTs4LW7du9foQyDxH5wiphc4RYoXOD1JLo+cINXoSQgghhBDiMcqUE0IIIYQQ4jHe\n6wMgxKmRkRE89NBDmJiYAMMw2Lp1K7Zt24ZUKoUdO3bg3LlzaGtrw1133YWmpiavD5d4RJZl3H33\n3WhpacHdd9+N4eFhfOMb30AymcSKFStw5513gufpI3CxSqfTePjhh3Hy5EkwDINPf/rT6Orqos8Q\novvZz36Gp556CgzDYOnSpbj99tsxMTFBnyOL2Le+9S3s27cPsVgMDz74IACYxh6KouD73/8+9u/f\nD7/fj9tvvx0rVqywfH7uq1/96lfn4OsgxDX5fB6rV6/GzTffjM2bN+ORRx7BpZdeil/84hdYunQp\n7rrrLoyPj+PVV1/FunXrvD5c4pGf//znEEURoihi06ZNeOSRR3Ddddfhtttuw2uvvYbx8XGsXLnS\n68MkHvnOd76DSy+9FLfffju2bt2KUCiEnTt30mcIAQCMjY3hO9/5Dh544AFs27YNe/bsgSiK+OUv\nf0mfI4tYOBzGddddh7179+K3f/u3AQA//vGPDT839u/fjwMHDuBv/uZv0NfXh8ceewzvec97LJ+f\nylfIghOPx/WrzWAwiO7uboyNjWHv3r3YsmULAGDLli3Yu3evl4dJPDQ6Oop9+/bpH4CKouDgwYO4\n6qqrAADXXnstnR+LWCaTwZtvvonrr78eAMDzPMLhMH2GkAqyLKNQKECSJBQKBTQ3N9PnyCJ38cUX\nV909M/vceOmll7B582YwDIPVq1cjnU5jfHzc8vnpngtZ0IaHh3H8+HFccMEFmJycRDweBwA0Nzdj\ncnLS46MjXvnBD36Aj370o8hmswCAZDKJUCgEjuMAAC0tLRgbG/PyEImHhoeHEY1G8a1vfQsDAwNY\nsWIFtm/fTp8hRNfS0oIbb7wRn/70pyEIAi677DKsWLGCPkdIFbPPjbGxMSQSCf1xra2tGBsb0x9r\nhDLlZMHK5XJ48MEHsX37doRCoYo/YxgGDMN4dGTESy+//DJisVjN2j2yeEmShOPHj+O3fuu38PWv\nfx1+vx87d+6seAx9hixuqVQKe/fuxUMPPYRHHnkEuVwOBw4c8PqwyDzX6OcGZcrJgiSKIh588EFc\nc801uPLKKwEAsVgM4+PjiMfjGB8fRzQa9fgoiRcOHTqEl156Cfv370ehUEA2m8UPfvADZDIZSJIE\njuMwNjaGlpYWrw+VeKS1tRWtra1YtWoVAOCqq67Czp076TOE6F577TW0t7fr58CVV16JQ4cO0ecI\nqWL2udHS0oKRkRH9caOjozXPF8qUkwVHURQ8/PDD6O7uxvvf/37996+44grs3r0bALB7925s3LjR\nq0MkHvrwhz+Mhx9+GA899BA+//nP45JLLsFnP/tZrF27Fs8//zwAYNeuXbjiiis8PlLilebmZrS2\ntuL06dMA1ACsp6eHPkOILpFI4O2330Y+n4eiKPo5Qp8jZCazz40rrrgCTz/9NBRFweHDhxEKhSxL\nVwBaHkQWoLfeegtf/vKXsWzZMv020c0334xVq1Zhx44dGBkZoXFmBABw8OBB/Nu//RvuvvtuDA0N\n4Rvf+AZSqRT6+vpw5513wufzeX2IxCP9/f14+OGHIYoi2tvbcfvtt0NRFPoMIbof//jH2LNnDziO\nQ29vLz71qU9hbGyMPkcWsW984xt44403kEwmEYvF8Lu/+7vYuHGj4eeGoih49NFH8corr0AQBNx+\n++01J/VQUE4IIYQQQojHqHyFEEIIIYQQj1FQTgghhBBCiMcoKCeEEEIIIcRjFJQTQgghhBDiMQrK\nCSGEEEII8RgF5YQQQuaFf/mXf8HDDz/s9WEQQognaCQiIYTU6Y477sBtt92GdevWYdeuXfjWt76F\nbdu2Yfv27fpj9u7di/vvvx9btmzBHXfcgeHhYXzmM5+B3+8HAAQCAaxcuRLbtm3DunXrKp57YmIC\nLMuC53msXr0at956KxKJhOGx7NmzB//+7/+O/v5+XHDBBfjqV79a8ef9/f349re/jcHBQXR3d+PT\nn/40ent7AagLuZ544gk89dRTAIDrr78eH/nIRwzXRR88eBB/+Zd/CUEQwDAM4vE4brrpJlx33XUN\nfCcJIYRQppwQQlzS0dGB5557DpIk6b+3e/dudHZ2Vj32Bz/4AR5//HHcf//9WLduHe6//37s2rWr\n4jF/9md/hscffxyPPPIIYrEYHnvsMdPXbmpqwrZt23DTTTdV/Zkoivj617+Oa665Bt///vexZcsW\nfP3rX4coigCAJ598Ur94eOCBB/Dyyy/jV7/6lelrxeNxPP744/jhD3+IP/iDP8Ajjzyib8ckhBBS\nHwrKCSHEJc3NzVi2bBleeeUVAEAqlcKhQ4csV3E3Nzdj27Zt+NCHPoQnnngCsixXPUYQBFx11VUY\nHBw0fZ5169bh6quvNlzjfPDgQUiShBtuuAE+nw/btm2Doih4/fXXAagXDjfeeCNaW1vR0tKCG2+8\nUV8bbYVhGGzYsAFNTU0YGBjQf39wcBD33nsvPv7xj+Nzn/sc9uzZAwB4++23ceutt1Z8jS+++CL+\n9E//FIC6QfHv//7v9T87fPgw7rnnHmzfvh1f+MIXcPDgQQDA66+/jj/5kz/RH3fvvffiz//8z/Vf\nf/nLX8aLL74IANi5cyduu+02fOxjH8PnPvc5vPbaazW/LkII8QIF5YQQ4qItW7boAe1vfvMbbNy4\n0dYa7iuvvBKTk5OGGed8Po89e/Zg1apVdR3TyZMnsXz58opylOXLl+PkyZMVf270Z1ZkWcZLL72E\nZDKJJUuWAAByuRz+6q/+Cps2bcL3vvc9fP7zn8ejjz6KU6dOYdWqVQgEAvrFAAA8++yz2LRpU9Vz\nj42N4b777sMHP/hBPPbYY7jlllvw4IMPYmpqCqtXr8bZs2cxNTUFURRx4sQJjI2NIZvNolAo4OjR\no7joootw+vRp/PKXv8Tf/u3f4h//8R/xpS99CW1tbXV9DwkhZLbxXh8AIYScT975znfihz/8ITKZ\nDJ5++mnccsstOHDgQM2/p2W4U6mU/nv3338/OI5DPp9HNBrFl770pbqOKZfLIRQKVfxeKBRCNps1\n/PNQKIRcLgdFUQzrysfHx7F9+3YUCgVIkoSPfexj6OvrAwDs27cPbW1teo15X18frrzySjz33HP4\n0Ic+hHe/+9149tlnsW7dOmSzWezfvx+33HJL1Ws8/fTTWL9+PTZs2ABAvROwcuVK7Nu3D9deey1W\nrlyJN998E/F4HMuXL0coFMKhQ4fA8zw6OzsRiUSQTqdRLBZx6tQpRKNRtLe31/X9I4SQuUBBOSGE\nuEgQBKxfvx4/+clPkEwmceGFF9oKysfGxgCoteGaL3zhC1i3bh1kWcbevXvxla98BTt27EBzc7Oj\nYwoEAnoArslkMggGg4Z/ns1mEQgEDANyQL2AePjhh1EsFvHEE0/g9ddfxw033AAAOHfuHN5+++2K\nZldJkrB582YAwKZNm3DPPffg1ltvxQsvvIC+vj7D7PXIyAief/55vPzyyxXPs3btWgDAxRdfjIMH\nD6K1tRUXX3wxwuEw3njjDfA8j4svvhjA/9/e/YWy/sdxHH/62rhQVkyJLcma/MuKUMaSFi6UCzdq\n4YaUpBRX3Lhz6UbsbrJc4I6265WbIS6Qi11pppTlhgtb9rvQ+dbpOD9OnexcvB536/P+fNp7F9t7\nn969v1BZWcnk5CR7e3skk0laW1sZHx+nrKzsjz4/EZHvoKJcROQv8/l8rK6uMjo6+uU98Xgcm81G\nVVXVL2uGYdDZ2UkwGOTm5oaurq4/ej9Op5PDw8Ofbr5vb28ZHBw0139MbYH3SS1Op/PTc61WK4FA\ngPn5eeLxOB0dHWaRvLKy8uEeh8NBRUUF5+fnHB8ff9i6AlBeXk5PTw8zMzMfrjc2NrK9vY3dbmdk\nZISSkhK2trawWq0MDAyYcV6vF6/Xy8vLC8FgkHA4zNzc3Ke5iYh8N/WUi4j8ZY2NjSwvLzM0NPRp\n7NPTE9FolP39fcbGxjCMX7+Wc7kcJycnPD8/U11d/eE5b29vZjtJLpfj9fXVnK7S1NSEYRhEIhEy\nmQzRaBSA5uZmAHp7ezk6OiKdTpNOpzk8PMTn830pV4vFwvDwMAcHBwC0tbVxf39PLBYjm82SzWZJ\nJBIkk0lzT3d3N5FIhOvr69/+wejp6eHs7IyLiwszt6urKx4fHwFwu92kUikSiQQulwun02ne0jc0\nNACQSqW4vLwkk8lQVFRkjnEUEfkX6aZcROQvKygooKWl5X9jfrR3FBcXU1dXx8LCAh6P56eYtbU1\nDMOgoKCAiooKZmdnf3uDHYvF2NjYMF8HAgFzNrrFYmFxcZHNzU3C4TAOh4PFxUUslvefAL/fz8PD\ngznRpL+/H7/f/+V8+/r62Nvb4/T0lPb2dpaXlwmFQoRCIXK5HDU1NUxMTJjxXq+X3d1dPB4PpaWl\nH55pt9tZWlpiZ2eH9fV1DMPA5XIxNTUFvLfc1NbWUlRUZObhdrtJJpPYbDYAs22LVU4AAABxSURB\nVL3m7u6OwsJC6uvrmZ6e/nJeIiLfSQ8PEhERERHJM7WviIiIiIjkmYpyEREREZE8U1EuIiIiIpJn\nKspFRERERPJMRbmIiIiISJ6pKBcRERERyTMV5SIiIiIieaaiXEREREQkz1SUi4iIiIjk2X9R/ZYB\nSLPbEAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe92aa9cdd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print imdb[['ratings','polarity','subjectivity']].corr()\n",
    "\n",
    "ax = imdb['polarity'][1:50].plot(figsize=(12, 6))\n",
    "ax = imdb['polarity'][51:100].plot(figsize=(12, 6))\n",
    "ax.set_xlabel(\"IMDB 100 Reviews\", fontsize=12)\n",
    "ax.set_ylabel(\"Polarity\", fontsize=12)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2     0.152778\n",
       "3     0.240171\n",
       "4     0.441667\n",
       "5     0.257593\n",
       "6     0.133479\n",
       "7     0.168599\n",
       "8     0.168932\n",
       "9     0.210227\n",
       "10    0.285317\n",
       "11    0.219643\n",
       "12    0.198482\n",
       "13    0.300154\n",
       "14    0.367708\n",
       "15    0.259101\n",
       "16    0.354945\n",
       "17    0.303926\n",
       "18    0.318220\n",
       "19    0.185188\n",
       "20    0.375143\n",
       "21    0.395614\n",
       "22    0.320290\n",
       "23    0.274561\n",
       "24    0.650000\n",
       "25    0.331250\n",
       "26    0.236441\n",
       "27    0.172656\n",
       "28    0.292273\n",
       "29    0.351263\n",
       "30    0.238229\n",
       "31    0.236032\n",
       "32    0.266685\n",
       "33    0.143882\n",
       "34    0.330898\n",
       "35    0.694333\n",
       "36    0.197098\n",
       "37    0.187546\n",
       "38    0.230517\n",
       "39    0.485459\n",
       "40    0.403030\n",
       "41    0.300079\n",
       "42    0.327754\n",
       "43    0.199018\n",
       "44    0.210559\n",
       "45    0.166440\n",
       "46    0.320625\n",
       "47    0.363582\n",
       "48    0.399097\n",
       "49    0.364315\n",
       "50    0.384444\n",
       "Name: polarity, dtype: float64"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    " imdb['polarity'][1:50]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text = \" this is a awesome review, but i would consider this as pathetic\"\n",
    "TextBlob(unicode(text, errors='ignore')).sentiment.polarity"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Stopwords = [ NLTK stopwords + Extended words from multiple iterations of text analysis ) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'i', u'me', u'my', u'myself', u'we', u'our', u'ours', u'ourselves', u'you', u'your', u'yours', u'yourself', u'yourselves', u'he', u'him', u'his', u'himself', u'she', u'her', u'hers', u'herself', u'it', u'its', u'itself', u'they', u'them', u'their', u'theirs', u'themselves', u'what', u'which', u'who', u'whom', u'this', u'that', u'these', u'those', u'am', u'is', u'are', u'was', u'were', u'be', u'been', u'being', u'have', u'has', u'had', u'having', u'do', u'does', u'did', u'doing', u'a', u'an', u'the', u'and', u'but', u'if', u'or', u'because', u'as', u'until', u'while', u'of', u'at', u'by', u'for', u'with', u'about', u'against', u'between', u'into', u'through', u'during', u'before', u'after', u'above', u'below', u'to', u'from', u'up', u'down', u'in', u'out', u'on', u'off', u'over', u'under', u'again', u'further', u'then', u'once', u'here', u'there', u'when', u'where', u'why', u'how', u'all', u'any', u'both', u'each', u'few', u'more', u'most', u'other', u'some', u'such', u'no', u'nor', u'not', u'only', u'own', u'same', u'so', u'than', u'too', u'very', u's', u't', u'can', u'will', u'just', u'don', u'should', u'now', u'd', u'll', u'm', u'o', u're', u've', u'y', u'ain', u'aren', u'couldn', u'didn', u'doesn', u'hadn', u'hasn', u'haven', u'isn', u'ma', u'mightn', u'mustn', u'needn', u'shan', u'shouldn', u'wasn', u'weren', u'won', u'wouldn', \"'ll\", \"'d\", \"'m\", \"'re\", \"'s\", \"'ve\", \"ca n't\", 'r', \"n't\", 'ca', 'see', 'get', 'movies', 'movie', 'go', 'say', 'come', 'many', 'another', 'could', 'would', 'made', 'really', 'want', 'even', 'odd', 'films', 'plot', 'ever', 'actually', 'also', 'movie', 'film']\n"
     ]
    }
   ],
   "source": [
    "stop_words = nltk.corpus.stopwords.words('english')\n",
    "extended_stopwords = ['\\'ll','\\'d','\\'m','\\'re','\\'s','\\'ve','ca n\\'t','r','n\\'t','ca','see','get','movies','movie','go','say','come','many','another','could','would','made','really','want','even','odd','films','plot','ever','actually','also','movie','film']\n",
    "stopwords = stop_words + extended_stopwords\n",
    "\n",
    "print stopwords"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "###  Document Term Matrix using TF-IDF weighing scheme ( unigrams & bigrams ) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 344 ms, sys: 0 ns, total: 344 ms\n",
      "Wall time: 386 ms\n",
      "(100, 23)\n"
     ]
    }
   ],
   "source": [
    "term_idf_vectorizer       = TfidfVectorizer(max_df=0.9, max_features=2000,min_df=0.3, stop_words=stopwords, use_idf=True, tokenizer=tokenize, ngram_range=(1,2))\n",
    "%time term_idf_matrix     = term_idf_vectorizer.fit_transform(imdb.reviews) \n",
    "term_idf_feature_names    = term_idf_vectorizer.get_feature_names()\n",
    "print (term_idf_matrix).shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TF-IDF Features on model params  ( min_df = 0.3 , max_df = 0.9 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'best', u'cameron', u'dicaprio', u'effects', u'good', u'jack', u'james', u'james cameron', u'kate', u'like', u'love', u'love story', u'much', u'one', u'people', u'rose', u'seen', u'ship', u'special', u'story', u'time', u'titanic', u'winslet']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(u'titanic', 23.816388065259229),\n",
       " (u'one', 17.816273201816202),\n",
       " (u'people', 16.28637155526151),\n",
       " (u'love', 15.994226712869134),\n",
       " (u'ship', 15.680991540890769),\n",
       " (u'time', 15.181084930271583),\n",
       " (u'story', 14.158611453695693),\n",
       " (u'like', 12.692362389482325),\n",
       " (u'cameron', 12.570285579352678),\n",
       " (u'rose', 12.314379553851616),\n",
       " (u'jack', 10.536474022225097),\n",
       " (u'best', 10.176133248342458),\n",
       " (u'good', 9.9813549064206892),\n",
       " (u'james', 9.9129111421178884),\n",
       " (u'love story', 9.7936953537828355),\n",
       " (u'effects', 9.7270116987082638),\n",
       " (u'much', 9.1663660037724402),\n",
       " (u'seen', 9.1566845823009686),\n",
       " (u'winslet', 8.511426163671123),\n",
       " (u'special', 8.2773013661909314),\n",
       " (u'kate', 8.2138779927123799),\n",
       " (u'dicaprio', 7.8400281414858632),\n",
       " (u'james cameron', 7.3140359337772081)]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print term_idf_feature_names\n",
    "\n",
    "tfidf = zip(term_idf_vectorizer.get_feature_names(),np.asarray(term_idf_matrix.sum(axis=0)).ravel())\n",
    "term_idf_frequency = sorted(tfidf,key=lambda x: x[1],reverse=True)\n",
    "term_idf_frequency"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Document Term Matrix using Term Frequency (TF)  weighing scheme ( unigrams & bigrams ) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 332 ms, sys: 20 ms, total: 352 ms\n",
      "Wall time: 344 ms\n",
      "(100, 23)\n"
     ]
    }
   ],
   "source": [
    "count_vectorizer    = CountVectorizer(max_df=0.9, max_features=2000,min_df=0.3, stop_words=stopwords, tokenizer=tokenize, ngram_range=(1,2))\n",
    "%time term_frequency_matrix = count_vectorizer.fit_transform(imdb.reviews) \n",
    "term_frequency_feature_names    = count_vectorizer.get_feature_names()\n",
    "print (term_frequency_matrix).shape "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Term Frequency  Features on model params  ( min_df = 0.3 , max_df = 0.9 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[u'best', u'cameron', u'dicaprio', u'effects', u'good', u'jack', u'james', u'james cameron', u'kate', u'like', u'love', u'love story', u'much', u'one', u'people', u'rose', u'seen', u'ship', u'special', u'story', u'time', u'titanic', u'winslet']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(u'titanic', 202),\n",
       " (u'one', 123),\n",
       " (u'ship', 120),\n",
       " (u'love', 111),\n",
       " (u'story', 102),\n",
       " (u'time', 102),\n",
       " (u'rose', 89),\n",
       " (u'cameron', 85),\n",
       " (u'people', 85),\n",
       " (u'like', 72),\n",
       " (u'jack', 71),\n",
       " (u'james', 61),\n",
       " (u'effects', 56),\n",
       " (u'much', 51),\n",
       " (u'best', 49),\n",
       " (u'good', 49),\n",
       " (u'love story', 47),\n",
       " (u'special', 46),\n",
       " (u'winslet', 46),\n",
       " (u'seen', 43),\n",
       " (u'dicaprio', 42),\n",
       " (u'kate', 42),\n",
       " (u'james cameron', 41)]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print term_frequency_feature_names\n",
    "\n",
    "tf = zip(count_vectorizer.get_feature_names(),np.asarray(term_frequency_matrix.sum(axis=0)).ravel())\n",
    "term_frequency = sorted(tf,key=lambda x: x[1],reverse=True)\n",
    "term_frequency"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Topic extraction with Non-negative Matrix Factorization using TF-IDF Features "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 52 ms, sys: 228 ms, total: 280 ms\n",
      "Wall time: 118 ms\n",
      "\n",
      "Fitting the Non-negative Matrix Factorization model with tf-idf features: \n",
      "\n",
      "Topic #0:\n",
      "one time like much ship seen story best cameron effects people jack james dicaprio\n",
      "Topic #1:\n",
      "love love story story much time good titanic cameron like jack james people best seen\n",
      "Topic #2:\n",
      "people seen much dicaprio story titanic jack winslet like james cameron cameron effects good james\n",
      "Topic #3:\n",
      "rose jack story kate ship love like winslet titanic time one much cameron dicaprio\n",
      "Topic #4:\n",
      "cameron james james cameron like dicaprio seen much story titanic one effects rose kate winslet\n",
      "Topic #5:\n",
      "best kate winslet good dicaprio like seen rose love james one james cameron cameron effects\n",
      "Topic #6:\n",
      "effects special good seen titanic cameron one rose people winslet james cameron dicaprio jack james\n",
      "Topic #7:\n",
      "titanic ship time love winslet dicaprio story jack good kate people james effects james cameron\n",
      "()\n"
     ]
    }
   ],
   "source": [
    "# Fit the NMF model\n",
    "%time nmf = NMF(n_components=8, random_state=1,alpha=.1, l1_ratio=.5).fit(term_idf_matrix)\n",
    "\n",
    "print(\"\\nFitting the Non-negative Matrix Factorization model with tf-idf features: \\n\")\n",
    "print_top_words(nmf, term_idf_feature_names, 14)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Topic extraction with Latent Dirichlet Allocation using TF Features "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 148 ms, sys: 0 ns, total: 148 ms\n",
      "Wall time: 147 ms\n",
      "\n",
      "Topics using Latent Dirichlet Allocation model with Term frequencies: \n",
      "\n",
      "Topic #0:\n",
      "titanic ship one time love story people rose jack like cameron best james much\n",
      "Topic #1:\n",
      "rose kate winslet time people ship james story love dicaprio seen james cameron jack good\n",
      "Topic #2:\n",
      "people rose ship time jack story one dicaprio much love story special like titanic seen\n",
      "Topic #3:\n",
      "best one much like cameron james time james cameron story people love dicaprio kate love story\n",
      "Topic #4:\n",
      "love story rose titanic winslet one special jack effects ship dicaprio seen kate time\n",
      "Topic #5:\n",
      "like cameron much one time titanic good kate james james cameron winslet jack people effects\n",
      "Topic #6:\n",
      "titanic cameron effects story special love james love story one james cameron seen rose good ship\n",
      "Topic #7:\n",
      "jack rose love kate titanic story dicaprio winslet time love story much special people one\n",
      "()\n"
     ]
    }
   ],
   "source": [
    "lda = LatentDirichletAllocation(n_topics=8, max_iter=5,learning_method='online',learning_offset=50.,random_state=0)\n",
    "%time lda.fit(term_frequency_matrix)\n",
    "print(\"\\nTopics using Latent Dirichlet Allocation model with Term frequencies: \\n\")\n",
    "print_top_words(lda, term_frequency_feature_names, 14)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  LDA Topic modeling using only Bigram features  based on frequency ( avoid unigram noise ) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 376 ms, sys: 12 ms, total: 388 ms\n",
      "Wall time: 375 ms\n",
      "(100, 22)\n",
      "\n",
      " Bigram Features based on Term Frequency \n",
      " \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(u'love story', 47),\n",
       " (u'james cameron', 41),\n",
       " (u'special effects', 41),\n",
       " (u'kate winslet', 29),\n",
       " (u'jack rose', 25),\n",
       " (u'leonardo dicaprio', 20),\n",
       " (u'billy zane', 16),\n",
       " (u'first time', 13),\n",
       " (u'james horner', 12),\n",
       " (u'celine dion', 11),\n",
       " (u'jack dawson', 11),\n",
       " (u'kathy bates', 11),\n",
       " (u'sinking ship', 10),\n",
       " (u'every time', 8),\n",
       " (u'historically accurate', 8),\n",
       " (u'ship sinking', 8),\n",
       " (u'bill paxton', 7),\n",
       " (u'gloria stuart', 7),\n",
       " (u'dicaprio kate', 6),\n",
       " (u'dicaprio kate winslet', 6),\n",
       " (u'dicaprio winslet', 6),\n",
       " (u'much money', 6)]"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigram_vectorizer    = CountVectorizer(max_df=0.9, max_features=200,min_df=0.06, stop_words=stopwords, tokenizer=tokenize, ngram_range=(2,3))\n",
    "%time dtm_bigram_matrix = bigram_vectorizer.fit_transform(imdb.reviews) \n",
    "bigram_frequency_feature_names    = bigram_vectorizer.get_feature_names()\n",
    "print (dtm_bigram_matrix).shape\n",
    "print (\"\\n Bigram Features based on Term Frequency \\n \")\n",
    "tf = zip(bigram_vectorizer.get_feature_names(),np.asarray(dtm_bigram_matrix.sum(axis=0)).ravel())\n",
    "bigram_term_frequency = sorted(tf,key=lambda x: x[1],reverse=True)\n",
    "bigram_term_frequency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#from IPython.display import Image\n",
    "#Image(filename='up_imdb_review.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Topics = 5 , Displaying top 10 words based on probability "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 104 ms, sys: 0 ns, total: 104 ms\n",
      "Wall time: 102 ms\n",
      "\n",
      "Topics using Latent Dirichlet Allocation model with Term frequencies: \n",
      "\n",
      "Topic #0:\n",
      "special effects love story kate winslet leonardo dicaprio billy zane james cameron kathy bates james horner jack dawson historically accurate bill paxton gloria stuart\n",
      "Topic #1:\n",
      "james cameron jack rose love story celine dion first time ship sinking much money sinking ship every time dicaprio winslet special effects james horner\n",
      "()\n"
     ]
    }
   ],
   "source": [
    "lda = LatentDirichletAllocation(n_topics=2, max_iter=4,learning_method='batch',learning_offset=50.,random_state=0)\n",
    "%time lda.fit(dtm_bigram_matrix)\n",
    "print(\"\\nTopics using Latent Dirichlet Allocation model with Term frequencies: \\n\")\n",
    "print_top_words(lda, bigram_frequency_feature_names,12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  Recommendation based on movie attributes [ Titanic( 1997)  ] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "    [ Floating house, Paradise falls, Thousands Balloons, Old man & Boy , Retirement Home flying to South America, Boy scout, bird = Kevin ] \n",
    "     \n",
    "     Very innovative animation movie ( may be better than toy story ) which will attract young and old because of its characters. The director needs to retain the theme where he effectively utilized talking animals, explorer , adventures, bonding ( young kid with old man ) which got him success. \n",
    "     \n",
    "     Pixar , Disney should make more animation movies with actors Edward Asner , Christopher Plummer & Jordan Nagai\n",
    "    "
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}