code/classify_lyrics/nb_init_model.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:b52579a7f9fd53079b5188b464a5c0939238a1dcb16b88d84b477139f572b9d9"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%load_ext watermark"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "%watermark -d -v -a 'Sebastian Raschka' -p scikit-learn,nltk,numpy"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Sebastian Raschka 10/12/2014 \n",
        "\n",
        "CPython 3.4.2\n",
        "IPython 2.3.1\n",
        "\n",
        "scikit-learn 0.15.2\n",
        "nltk 3.0.0\n",
        "numpy 1.9.1\n"
       ]
      }
     ],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<font size=\"1.5em\">[More information](https://github.com/rasbt/watermark) about the `watermark` magic command extension.</font>"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Lyrics Mood Classification - Training"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Sections"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "- [Preprocessing](#Preprocessing)\n",
      "    - [Reading the dataset](#Reading-the-dataset)\n",
      "    - [Label Encoder](#Label-Encoder)\n",
      "    - [Feature extraction: Word counts and Vectorizers](#Feature-extraction:-Word-counts-and-Vectorizers)\n",
      "        - [Porter Stemmer](#Porter-Stemmer)\n",
      "        - [Stop word list](#Stop-word-list)\n",
      "        - [Count Vectorizer](#Count-Vectorizer)\n",
      "        - [Tfidf Vectorizer](#Tfidf-Vectorizer)\n",
      "- [Model Selection](#Model-Selection)\n",
      "    - [Grid search and hyperparameter tuning](#Grid-search-and-hyperparameter-tuning)\n",
      "        - [Grid Search](#Grid-Search)\n",
      "        - [Grid Search ROC Curves](#Grid-Search-ROC-Curves)\n",
      "        - [Hyperparameter tuning 1 - max_features](#Hyperparameter-tuning-1---max_features)\n",
      "        - [Hyperparameter tuning 2 - min_df](#Hyperparameter-tuning-2---min_df)\n",
      "        - [Hyperparameter tuning 3 - alpha](#Hyperparameter-tuning-3---alpha)\n",
      "- [Validation](#Validation)\n",
      "    - [Confusion matrix](#Confusion-matrix)\n",
      "    - [ROC AUC, Accuracy, Precision, Recall, and F1-score](#ROC-AUC,-Accuracy,-Precision,-Recall,-and-F1-score)\n",
      "- [Save classifier](#Save-classifier)"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Preprocessing"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Reading the dataset"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pandas as pd\n",
      "\n",
      "df = pd.read_csv('../../dataset/training/train_lyrics_1000.csv')\n",
      "\n",
      "df.head()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>file</th>\n",
        "      <th>artist</th>\n",
        "      <th>title</th>\n",
        "      <th>lyrics</th>\n",
        "      <th>genre</th>\n",
        "      <th>mood</th>\n",
        "      <th>year</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>0</th>\n",
        "      <td> TRAAAAW128F429D538.h5</td>\n",
        "      <td>        Casual</td>\n",
        "      <td>        I Didn't Mean To</td>\n",
        "      <td> Verse One:\\n\\nAlright I might\\nHave had a litt...</td>\n",
        "      <td> Hip Hop/Rap</td>\n",
        "      <td>   sad</td>\n",
        "      <td> 1994</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>1</th>\n",
        "      <td> TRAAAEF128F4273421.h5</td>\n",
        "      <td>      Adam Ant</td>\n",
        "      <td>         Something Girls</td>\n",
        "      <td> Adam Ant/Marco Pirroni\\nEvery girl is a someth...</td>\n",
        "      <td>        Rock</td>\n",
        "      <td> happy</td>\n",
        "      <td> 1982</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>2</th>\n",
        "      <td> TRAAAFD128F92F423A.h5</td>\n",
        "      <td>           Gob</td>\n",
        "      <td>          Face the Ashes</td>\n",
        "      <td> I've just erased it's been a while, I've got a...</td>\n",
        "      <td>        Rock</td>\n",
        "      <td>   sad</td>\n",
        "      <td> 2007</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>3</th>\n",
        "      <td> TRAABJV128F1460C49.h5</td>\n",
        "      <td> Lionel Richie</td>\n",
        "      <td> Tonight Will Be Alright</td>\n",
        "      <td> Little darling \\nWhere you've been so long \\nI...</td>\n",
        "      <td>         R&amp;B</td>\n",
        "      <td> happy</td>\n",
        "      <td> 1986</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>4</th>\n",
        "      <td> TRAABLR128F423B7E3.h5</td>\n",
        "      <td>    Blue Rodeo</td>\n",
        "      <td>                Floating</td>\n",
        "      <td> Lead Vocal by Greg\\n\\nWell, these late night c...</td>\n",
        "      <td>        Rock</td>\n",
        "      <td>   sad</td>\n",
        "      <td> 1987</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 3,
       "text": [
        "                    file         artist                    title  \\\n",
        "0  TRAAAAW128F429D538.h5         Casual         I Didn't Mean To   \n",
        "1  TRAAAEF128F4273421.h5       Adam Ant          Something Girls   \n",
        "2  TRAAAFD128F92F423A.h5            Gob           Face the Ashes   \n",
        "3  TRAABJV128F1460C49.h5  Lionel Richie  Tonight Will Be Alright   \n",
        "4  TRAABLR128F423B7E3.h5     Blue Rodeo                 Floating   \n",
        "\n",
        "                                              lyrics        genre   mood  year  \n",
        "0  Verse One:\\n\\nAlright I might\\nHave had a litt...  Hip Hop/Rap    sad  1994  \n",
        "1  Adam Ant/Marco Pirroni\\nEvery girl is a someth...         Rock  happy  1982  \n",
        "2  I've just erased it's been a while, I've got a...         Rock    sad  2007  \n",
        "3  Little darling \\nWhere you've been so long \\nI...          R&B  happy  1986  \n",
        "4  Lead Vocal by Greg\\n\\nWell, these late night c...         Rock    sad  1987  "
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Label Encoder"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.preprocessing import LabelEncoder\n",
      "import pickle\n",
      "import numpy as np\n",
      "\n",
      "X_train = df['lyrics'].values \n",
      "\n",
      "y_train = df['mood'].values\n",
      "\n",
      "print('before: %s ...' %y_train[:5])\n",
      "\n",
      "le = LabelEncoder()\n",
      "le.fit(y_train)\n",
      "y_train = le.transform(y_train)\n",
      "\n",
      "print('after: %s ...' %y_train[:5])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "before: ['sad' 'happy' 'sad' 'happy' 'sad'] ...\n",
        "after: [1 0 1 0 1] ...\n"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Save object to disk\n",
      "\n",
      "import pickle\n",
      "\n",
      "pickle_out = open('./lyrics_label_encoder_py.pkl', 'wb')\n",
      "pickle.dump(le, pickle_out)\n",
      "pickle_out.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Feature extraction: Word counts and Vectorizers"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Porter Stemmer"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Porter Stemmer\n",
      "\n",
      "import nltk\n",
      "import string\n",
      "import re\n",
      "\n",
      "porter_stemmer = nltk.stem.porter.PorterStemmer()\n",
      "\n",
      "def porter_tokenizer(text, stemmer=porter_stemmer):\n",
      "    \"\"\"\n",
      "    A Porter-Stemmer-Tokenizer hybrid to splits sentences into words (tokens) \n",
      "    and applies the porter stemming algorithm to each of the obtained token. \n",
      "    Tokens that are only consisting of punctuation characters are removed as well.\n",
      "    Only tokens that consist of more than one letter are being kept.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "        \n",
      "    text : `str`. \n",
      "      A sentence that is to split into words.\n",
      "        \n",
      "    Returns\n",
      "    ----------\n",
      "    \n",
      "    no_punct : `str`. \n",
      "      A list of tokens after stemming and removing Sentence punctuation patterns.\n",
      "    \n",
      "    \"\"\"\n",
      "    lower_txt = text.lower()\n",
      "    tokens = nltk.wordpunct_tokenize(lower_txt)\n",
      "    stems = [porter_stemmer.stem(t) for t in tokens]\n",
      "    no_punct = [s for s in stems if re.match('^[a-zA-Z]+$', s) is not None]\n",
      "    return no_punct\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 7
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Test"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "porter_tokenizer(\"Don't !!! --- want swimming. \")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 8,
       "text": [
        "['don', 't', 'want', 'swim']"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Stop word list"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Commented out to prevent overwriting files:\n",
      "#\n",
      "# stp = nltk.corpus.stopwords.words('english')\n",
      "# with open('./stopwords_eng.txt', 'w') as outfile:\n",
      "#    outfile.write('\\n'.join(stp))\n",
      "    \n",
      "    \n",
      "with open('./stopwords_eng.txt', 'r') as infile:\n",
      "    stop_words = infile.read().splitlines()\n",
      "print('stop words %s ...' %stop_words[:5])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "stop words ['i', 'me', 'my', 'myself', 'we'] ...\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Count Vectorizer"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Count Vectorizer\n",
      "\n",
      "from sklearn.feature_extraction.text import CountVectorizer\n",
      "\n",
      "vec = CountVectorizer(\n",
      "            encoding='utf-8',\n",
      "            decode_error='replace',\n",
      "            strip_accents='unicode',\n",
      "            analyzer='word',\n",
      "            binary=False,\n",
      "            stop_words=stop_words,\n",
      "            tokenizer=porter_tokenizer,\n",
      "            ngram_range=(1,1)\n",
      "    )\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Test 1"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "vocab = [\"123 1 The\\n swimmer likes swimming so he swims. Don't didn`t\"]\n",
      "\n",
      "vec = vec.fit(vocab)\n",
      "\n",
      "sentence1 = vec.transform([u'The swimmer likes swimming.'])\n",
      "sentence2 = vec.transform(['The\\nswimmer \\nswims.'])\n",
      "\n",
      "\n",
      "print('TEST:')\n",
      "print('Vocabulary: %s' %vec.get_feature_names())\n",
      "print('Sentence 1: %s' %sentence1.toarray())\n",
      "print('Sentence 2: %s' %sentence2.toarray())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "TEST:\n",
        "Vocabulary: ['didn', 'like', 'swim', 'swimmer']\n",
        "Sentence 1: [[0 1 1 1]]\n",
        "Sentence 2: [[0 0 1 1]]\n"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Fitting the lyrics"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "vec = vec.fit(X_train.ravel())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print('Vocabulary size: %s' %len(vec.get_feature_names()))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Vocabulary size: 8550\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Test 2 (N-grams = 2)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "vec = CountVectorizer(\n",
      "            encoding='utf-8',\n",
      "            decode_error='replace',\n",
      "            strip_accents='unicode',\n",
      "            analyzer='word',\n",
      "            binary=False,\n",
      "            stop_words=stop_words,\n",
      "            tokenizer=porter_tokenizer,\n",
      "            ngram_range=(2,2)\n",
      "    )\n",
      "\n",
      "vocab = [\"123 1 The\\n swimmer likes swimming so he swims. Don't didn`t\"]\n",
      "\n",
      "vec = vec.fit(vocab)\n",
      "\n",
      "sentence1 = vec.transform([u'The swimmer likes swimming.'])\n",
      "sentence2 = vec.transform(['The\\nswimmer \\nswims.'])\n",
      "\n",
      "\n",
      "print('TEST:')\n",
      "print('Vocabulary: %s' %vec.get_feature_names())\n",
      "print('Sentence 1: %s' %sentence1.toarray())\n",
      "print('Sentence 2: %s' %sentence2.toarray())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "TEST:\n",
        "Vocabulary: ['like swim', 'swim didn', 'swim swim', 'swimmer like']\n",
        "Sentence 1: [[1 0 0 1]]\n",
        "Sentence 2: [[0 0 0 0]]\n"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Tfidf Vectorizer"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.feature_extraction.text import TfidfVectorizer\n",
      "\n",
      "tfidf = TfidfVectorizer(\n",
      "            encoding='utf-8',\n",
      "            decode_error='replace',\n",
      "            strip_accents='unicode',\n",
      "            analyzer='word',\n",
      "            binary=False,\n",
      "            stop_words=stop_words,\n",
      "            tokenizer=porter_tokenizer\n",
      "    )"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 15
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Test"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "vocab = [\"123 1 The\\n swimmer likes swimming so he swims. Don't didn`t\"]\n",
      "\n",
      "tfidf = tfidf.fit(vocab)\n",
      "\n",
      "sentence1 = tfidf.transform([u'The swimmer likes swimming.'])\n",
      "sentence2 = tfidf.transform(['The\\nswimmer \\nswims.'])\n",
      "\n",
      "\n",
      "print('TEST:')\n",
      "print('Vocabulary: %s' %tfidf.get_feature_names())\n",
      "print('Sentence 1: %s' %sentence1.toarray())\n",
      "print('Sentence 2: %s' %sentence2.toarray())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "TEST:\n",
        "Vocabulary: ['didn', 'like', 'swim', 'swimmer']\n",
        "Sentence 1: [[ 0.          0.57735027  0.57735027  0.57735027]]\n",
        "Sentence 2: [[ 0.          0.          0.70710678  0.70710678]]\n"
       ]
      }
     ],
     "prompt_number": 16
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Fitting the lyrics"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "tfidf = tfidf.fit(X_train.ravel())\n",
      "\n",
      "print('Vocabulary size: %s' %len(tfidf.get_feature_names()))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Vocabulary size: 8550\n"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Model Selection"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Grid search and hyperparameter tuning"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Models: Multivariate Bernoulli and Multinomial naive Bayes"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.naive_bayes import MultinomialNB\n",
      "from sklearn.naive_bayes import BernoulliNB\n",
      "from sklearn.pipeline import Pipeline"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 18
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Performance metric: F1-score"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Custom scorer methods to account for positive-negative class labels\n",
      "\n",
      "from sklearn import metrics\n",
      "\n",
      "# `pos_label` for positive class, since we have sad=1, happy=0\n",
      "\n",
      "f1_scorer = metrics.make_scorer(metrics.f1_score, greater_is_better=True, pos_label=0)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 19
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Grid Search"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.grid_search import GridSearchCV\n",
      "from pprint import pprint\n",
      "\n",
      "pipeline_1 = Pipeline([\n",
      "    ('vect', CountVectorizer()),\n",
      "    ('clf', BernoulliNB())\n",
      "])\n",
      "\n",
      "parameters_1 = dict(\n",
      "    vect__binary=[True],\n",
      "    vect__stop_words=[stop_words, None],\n",
      "    vect__tokenizer=[porter_tokenizer, None],\n",
      "    vect__ngram_range=[(1,1), (2,2), (3,3)],\n",
      ")\n",
      "\n",
      "grid_search_1 = GridSearchCV(pipeline_1, \n",
      "                           parameters_1, \n",
      "                           n_jobs=1, \n",
      "                           verbose=1,\n",
      "                           scoring=f1_scorer,\n",
      "                           cv=10\n",
      "                )\n",
      "\n",
      "\n",
      "print(\"Performing grid search...\")\n",
      "print(\"pipeline:\", [name for name, _ in pipeline_1.steps])\n",
      "print(\"parameters:\")\n",
      "pprint(parameters_1, depth=2)\n",
      "grid_search_1.fit(X_train, y_train)\n",
      "print(\"Best score: %0.3f\" % grid_search_1.best_score_)\n",
      "print(\"Best parameters set:\")\n",
      "best_parameters_1 = grid_search_1.best_estimator_.get_params()\n",
      "for param_name in sorted(parameters_1.keys()):\n",
      "    print(\"\\t%s: %r\" % (param_name, best_parameters_1[param_name]))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done   1 jobs       | elapsed:    3.9s\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n",
        "[Parallel(n_jobs=1)]: Done  50 jobs       | elapsed:  2.0min\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Performing grid search...\n",
        "pipeline: ['vect', 'clf']\n",
        "parameters:\n",
        "{'vect__binary': [True],\n",
        " 'vect__ngram_range': [(...), (...), (...)],\n",
        " 'vect__stop_words': [[...], None],\n",
        " 'vect__tokenizer': [<function porter_tokenizer at 0x108604d90>, None]}\n",
        "Fitting 10 folds for each of 12 candidates, totalling 120 fits\n",
        "Best score: 0.617"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Best parameters set:\n",
        "\tvect__binary: True\n",
        "\tvect__ngram_range: (1, 1)\n",
        "\tvect__stop_words: ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', 'her', 'hers', 'herself', 'it', 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', 'should', 'now']\n",
        "\tvect__tokenizer: <function porter_tokenizer at 0x108604d90>\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done 120 out of 120 | elapsed:  4.7min finished\n"
       ]
      }
     ],
     "prompt_number": 23
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.grid_search import GridSearchCV\n",
      "\n",
      "pipeline_3 = Pipeline([\n",
      "    ('vect', CountVectorizer()),\n",
      "    ('clf', MultinomialNB())\n",
      "])\n",
      "\n",
      "parameters_3 = dict(\n",
      "    vect__binary=[False],\n",
      "    vect__stop_words=[stop_words, None],\n",
      "    vect__tokenizer=[porter_tokenizer, None],\n",
      "    vect__ngram_range=[(1,1), (2,2), (3,3)],\n",
      ")\n",
      "\n",
      "grid_search_3 = GridSearchCV(pipeline_3, \n",
      "                           parameters_3, \n",
      "                           n_jobs=1, \n",
      "                           verbose=1,\n",
      "                           scoring=f1_scorer,\n",
      "                           cv=10\n",
      "                )\n",
      "\n",
      "\n",
      "print(\"Performing grid search...\")\n",
      "print(\"pipeline:\", [name for name, _ in pipeline_3.steps])\n",
      "print(\"parameters:\")\n",
      "pprint(parameters_3, depth=2)\n",
      "grid_search_3.fit(X_train, y_train)\n",
      "print(\"Best score: %0.3f\" % grid_search_3.best_score_)\n",
      "print(\"Best parameters set:\")\n",
      "best_parameters_3 = grid_search_3.best_estimator_.get_params()\n",
      "for param_name in sorted(parameters_3.keys()):\n",
      "    print(\"\\t%s: %r\" % (param_name, best_parameters_3[param_name]))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done   1 jobs       | elapsed:    3.8s\n",
        "[Parallel(n_jobs=1)]: Done  50 jobs       | elapsed:  2.0min\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Performing grid search...\n",
        "pipeline: ['vect', 'clf']\n",
        "parameters:\n",
        "{'vect__binary': [False],\n",
        " 'vect__ngram_range': [(...), (...), (...)],\n",
        " 'vect__stop_words': [[...], None],\n",
        " 'vect__tokenizer': [<function porter_tokenizer at 0x108604d90>, None]}\n",
        "Fitting 10 folds for each of 12 candidates, totalling 120 fits\n",
        "Best score: 0.620"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Best parameters set:\n",
        "\tvect__binary: False\n",
        "\tvect__ngram_range: (1, 1)\n",
        "\tvect__stop_words: ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', 'her', 'hers', 'herself', 'it', 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', 'should', 'now']\n",
        "\tvect__tokenizer: <function porter_tokenizer at 0x108604d90>\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done 120 out of 120 | elapsed:  4.8min finished\n"
       ]
      }
     ],
     "prompt_number": 24
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.grid_search import GridSearchCV\n",
      "\n",
      "pipeline_4 = Pipeline([\n",
      "    ('vect', TfidfVectorizer()),\n",
      "    ('clf', MultinomialNB())\n",
      "])\n",
      "\n",
      "parameters_4 = dict(\n",
      "    vect__binary=[False],\n",
      "    vect__stop_words=[stop_words, None],\n",
      "    vect__tokenizer=[porter_tokenizer, None],\n",
      "    vect__ngram_range=[(1,1), (2,2), (3,3)],\n",
      ")\n",
      "\n",
      "grid_search_4 = GridSearchCV(pipeline_4, \n",
      "                           parameters_4, \n",
      "                           n_jobs=1, \n",
      "                           verbose=1,\n",
      "                           scoring=f1_scorer,\n",
      "                           cv=10\n",
      "                )\n",
      "\n",
      "\n",
      "print(\"Performing grid search...\")\n",
      "print(\"pipeline:\", [name for name, _ in pipeline_4.steps])\n",
      "print(\"parameters:\")\n",
      "pprint(parameters_4, depth=2)\n",
      "grid_search_4.fit(X_train, y_train)\n",
      "print(\"Best score: %0.3f\" % grid_search_4.best_score_)\n",
      "print(\"Best parameters set:\")\n",
      "best_parameters_4 = grid_search_4.best_estimator_.get_params()\n",
      "for param_name in sorted(parameters_4.keys()):\n",
      "    print(\"\\t%s: %r\" % (param_name, best_parameters_4[param_name]))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done   1 jobs       | elapsed:    4.4s\n",
        "/Users/sebastian/miniconda3/lib/python3.4/site-packages/sklearn/metrics/metrics.py:1771: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 due to no predicted samples.\n",
        "  'precision', 'predicted', average, warn_for)\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stderr",
       "text": [
        "[Parallel(n_jobs=1)]: Done  50 jobs       | elapsed:  2.1min\n",
        "[Parallel(n_jobs=1)]: Done 120 out of 120 | elapsed:  4.8min finished\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Performing grid search...\n",
        "pipeline: ['vect', 'clf']\n",
        "parameters:\n",
        "{'vect__binary': [False],\n",
        " 'vect__ngram_range': [(...), (...), (...)],\n",
        " 'vect__stop_words': [[...], None],\n",
        " 'vect__tokenizer': [<function porter_tokenizer at 0x108604d90>, None]}\n",
        "Fitting 10 folds for each of 12 candidates, totalling 120 fits\n",
        "Best score: 0.335"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Best parameters set:\n",
        "\tvect__binary: False\n",
        "\tvect__ngram_range: (2, 2)\n",
        "\tvect__stop_words: ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', 'her', 'hers', 'herself', 'it', 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', 'should', 'now']\n",
        "\tvect__tokenizer: None\n"
       ]
      }
     ],
     "prompt_number": 25
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Grid Search ROC Curves"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "clf_1 = Pipeline([\n",
      "                  ('vect', CountVectorizer(\n",
      "                                           binary=True,\n",
      "                                           stop_words=stop_words,\n",
      "                                           tokenizer=porter_tokenizer,\n",
      "                                           ngram_range=(1,1),\n",
      "                                           )\n",
      "                 ),\n",
      "                 ('clf', BernoulliNB()),\n",
      "                 ])\n",
      "\n",
      "clf_2 = Pipeline([\n",
      "                  ('vect', CountVectorizer(\n",
      "                                           binary=False,\n",
      "                                           stop_words=stop_words,\n",
      "                                           tokenizer=porter_tokenizer,\n",
      "                                           ngram_range=(1,1),\n",
      "                                           )\n",
      "                 ),\n",
      "                 ('clf', MultinomialNB()),\n",
      "                 ])\n",
      "\n",
      "clf_3 = Pipeline([\n",
      "                  ('vect', TfidfVectorizer(\n",
      "                                           binary=False,\n",
      "                                           stop_words=stop_words,\n",
      "                                           tokenizer=porter_tokenizer,\n",
      "                                           ngram_range=(1,1),\n",
      "                                           )\n",
      "                 ),\n",
      "                 ('clf', MultinomialNB()),\n",
      "                 ])\n",
      "\n",
      "colors = ['#1947D1', '#CC3300', 'k']\n",
      "linestyles = ['-', '--', '-.']\n",
      "classifiers = [clf_1, clf_2, clf_3]\n",
      "labels = ['1: MV Bernoulli NB, stop words, porter stemmer, \\nuni-gram, df', \n",
      "          '2: Multinomial NB, stop words, porter stemmer, \\nuni-gram, tf',\n",
      "          '3: Multinomial NB, stop words, porter stemmer, \\nuni-gram, tf-idf',\n",
      "          ]\n",
      "\n",
      "for clf,col,ls,lab in zip(classifiers, colors, linestyles, labels):\n",
      "    \n",
      "    mean_tpr = 0.0\n",
      "    mean_fpr = np.linspace(0, 1, 100)\n",
      "    all_tpr = []\n",
      "    cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "    for i, (train, test) in enumerate(cv):\n",
      "        probas_ = clf.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "        # Compute ROC curve and area the curve\n",
      "        fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "        mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "        mean_tpr[0] = 0.0\n",
      "        roc_auc = auc(fpr, tpr)\n",
      "\n",
      "    mean_tpr /= len(cv)\n",
      "    mean_tpr[-1] = 1.0\n",
      "    mean_auc = auc(mean_fpr, mean_tpr)\n",
      "    plt.plot(mean_fpr, \n",
      "             mean_tpr, \n",
      "             color=col, \n",
      "             linestyle=ls,\n",
      "             label='%s (ROC AUC = %0.2f)' % (lab, mean_auc), \n",
      "             lw=2\n",
      "    )\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')    \n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.savefig('./images/roc_gridsearch_1.eps', dpi=300)\n",
      "plt.show()\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFgCAYAAACSb/HAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvIUQQggv8/8Bz0Fl6u5W/DwAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x109b6a6d8>"
       ]
      }
     ],
     "prompt_number": 202
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Hyperparameter tuning 1 - max_features"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "\n",
      "colors = ['#1947D1', '#CC3300', 'k', '#339933']\n",
      "linestyles = ['-', '--', '-.', ':']\n",
      "params = [1000,3000,5000,None]\n",
      "labels = ['max features = 1000', \n",
      "          'max features = 3000',\n",
      "          'max features = 5000',\n",
      "          'max features = all (=8550)',\n",
      "          ]\n",
      "\n",
      "for param,col,ls,lab in zip(params, colors, linestyles, labels):\n",
      "\n",
      "    clf = Pipeline([\n",
      "                         ('vect', TfidfVectorizer(\n",
      "                                                  binary=False,\n",
      "                                                  stop_words=stop_words,\n",
      "                                                  tokenizer=porter_tokenizer,\n",
      "                                                  ngram_range=(1,1),\n",
      "                                                  max_features=param,\n",
      "                                                  )\n",
      "                          ),\n",
      "                         ('clf', MultinomialNB()),\n",
      "                        ])\n",
      "    \n",
      "    mean_tpr = 0.0\n",
      "    mean_fpr = np.linspace(0, 1, 100)\n",
      "    all_tpr = []\n",
      "    cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "    for i, (train, test) in enumerate(cv):\n",
      "        probas_ = clf.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "        # Compute ROC curve and area the curve\n",
      "        fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "        mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "        mean_tpr[0] = 0.0\n",
      "        roc_auc = auc(fpr, tpr)\n",
      "\n",
      "    mean_tpr /= len(cv)\n",
      "    mean_tpr[-1] = 1.0\n",
      "    mean_auc = auc(mean_fpr, mean_tpr)\n",
      "    plt.plot(mean_fpr, \n",
      "             mean_tpr, \n",
      "             color=col, \n",
      "             linestyle=ls,\n",
      "             label='%s (ROC AUC = %0.2f)' % (lab, mean_auc), \n",
      "             lw=2\n",
      "    )\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')    \n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.title('Multinomial NB, stop words, porter stemmer, uni-gram, tf-idf')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.savefig('./images/roc_maxfeatures.eps', dpi=300)\n",
      "plt.show()\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFtCAYAAAAu8eIeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd4FFXbwOHfZpNNW1IIKSQEQj30lqCARAgg4CuIFAEF\nfUXFBmLvvor9sysWrCgq9o4NW+jFEECqhxIwlEAggfSyZb4/dgkhBIhCsinPfV1cbKaceWZ2dp6Z\nM2fOmAzDQAghhBB1h5enAxBCCCHEPyPJWwghhKhjJHkLIYQQdYwkbyGEEKKOkeQthBBC1DGSvIUQ\nQog6RpJ3HaKUilNKOZVSCysZ9457XONTlDFDKfWS+/PVSqnr3Z+vVUrdVT2RH7P8h5RSl51imiuU\nUvMqGT7AvY6DKwx/WSn1oPvzu0qp3UqpNUqptUqpzUqpmUop82nE/IBS6sJ/O39NUkptUEr193AM\nbyqlelZT2b2UUrOqo+zaSCkVr5T6zNNxnEzF70QpNUsplaaUeqSSab9XSrWvZPhYpVSy+3Osez9e\no5Q6u3qjr7u8PR2A+MeKgbZKqeZa63QApVQg0A+oykP7Rrnp+gHrAbTWr1dDrMfRWj94mkWUAnOU\nUl211lnuYeXXyQCe01o/B6CU8gWWAeOBD//lMgcCG/99yDWq/LbwlMHAa9VUdiegWTWVXetorVOB\niz0dxylU/E6uAWK11nsrTqi1vqAK5SUBGVrr885QfPWSJO+6xwF8AkwEnnAPGw18DdwGritU4CWt\ndZdK/jYBJqXURcAIYLBSqgiIAMK01jcqpXYC7wCDgObAJ1rru9xlXQPc6I5jPzBNa71VKfUuUAQk\nAFHAp8AB9zKigKu11snu6dZrrZ9VSl2J64duARoD/6e1PtVBfyuwwh3fia6GTeU+WwFfIOMU5aKU\negi4CNcJQhZwBTAGiAeeVkrZgQXAK0A3XEnyR+BerbVDKeUEnsKVvALdw7+qsIzngXyt9f+UUlHA\nXmCQe9tMBC7UWo9XSv0PmADYgS24tvN+pdQCd2ztgVeBZGA24A9o9/rirml4GTjHvT5pwGStdcFJ\n1v8KXPuVE9fBeC9wudY6QynVDJgFtHBv3zla62eUUnHAYmATcORzNPCBUupyd+wvAp0BH+A34A73\n9irBtd92Ay7VWq8uF0s/4FnA7N7OTwB/AA8DQUqpt7XWVymlRgD34dqHCoHbtdYrlFIzgNZAK3c8\nK4Gfgf8CLYE7tdYfV3U6d0z34fqteQE7gRvc2+aY70Rr/coJtu+7uPf9in+f6DdX8bdcobxw9zyt\n3Mvf7y7voQrbdqL7/+N+a+7vfAzg5/7+0nHt39OAdpQ7ET7BOsWW+05mA21w7R8/KaVu0FovqTD9\nTmC01nq1Uuph4FJ37Nvc4wcAjwDBSqnftNaDTrTshk6qzeum94FJ5f6+HHj3nxSgtf4a+BbXj/NV\njr96DdRanwv0BW5USrVQSg0E7gAGaK2747qS/bpcsd2A3rgS+C1Antb6HFwH77vLlW24awuuBs7X\nWvfElaieqmL4NwHtlFJTKxlnAm5xV7mtw3Uw2gssPVmB7oPQTUCC1roXrgP4We4D8SpcSeEbYCZw\nwH0wTXCv8+3lisrXWicA44DZSqkmFRb1JTDM/XkYsA9XsgcYCXymlJrsHpegte4GbODo92sA2Vrr\nTu7Y5gKvu6d7Foh1T9cX6K+17uqOJw04LgFUog9wo9a6k3u9Z7qHzwV+01p3xXVCMEkpNd49LgZ4\nWGuttNZX49reE7XWKcDzwCp3DD2BcOBW93w+wLda6/blE7fbDFz7ZgJwJZCktd4N/A9Y7E7cbYHH\nOLoPXQt8qZQKcJdxjns7dgDOAzporfvjSkwPlVvWKadzn4h0xrVP9MB10vaWe/6K38mJVKwVOeVv\n7iRlgeu7Wa+17ojr6rxPufLKti3wFyf/rfXDdaLaDogExmutBwL/AR49WQBa610c/U6udMcPrmPE\nkkpmMQCUUiNxnQh1c69vIGBorRcAD7jLk8R9EpK86yD3gc6plOrpTjqNtNb/tlrXVOH/I75xL2sv\nkAmE4TrAfXykulprPQeIcV99GcA8rbVDa70fKAB+cpeVhutsv2yZ7ivA4cAI9xn4vbh+wKektS7E\ndcb+mFKqU4XRR6rNe7gTTRMgm6NJ6ER2A38Ca5RSTwNrtdbfVjLdMFxXtGitS3FVD59fbvyRcetx\n3ZI4t8L8S4Fm7qumYbgOjucppXzc0/7oLm+21rrIPc9MYJB7GnBd3aKUCsOVkN9zL3Olex0A1gEO\npdRK9/b9Qmu94hTbAOBXrfUW9+e3gKHuZNgX1xUZWutcXCcT5+Pa3nZg+QnKGw5cq5Rag+tkoBeu\nJHjE4hPM9ynwilLqA1w1H/e5h5ffT88DmgK/u8v/AFeNUBt3XL9orfO01sW4TihOtD9WZbrhuE5M\nV7mXdeTK9FTrUVHF31l5FX9zJ22/gmv7v+GeZx/weYXxi93jTvVbS9Fa79FaG8AOXCeu4Fp/v3In\nQ/9mnU40/WBc+2SB1toBvM2Jj0WiEpK8664jV9+TcB+8yzE49gdgqWR8ZZ/LK6owjancv/JMuM7y\nwVU9W57tBGUb7mrYP3FdKS4G7q+k7BNyn8A8CnyEq8qv0nndCfA9jk+iFacz3Fdb/8VVjfe8UuqF\nSib1qrAsM8fefnJUmNZeYTlOYB6ug+nZwJu4EtDFwDL3gbbidvZyL+PIsHz3/0a58UfY3cvJwXVV\nc5s7pk+UUjdXtu4VVIzfwdF1PtF6l7jXqzJewFj3yVQPXAlwernx+ZXNpLV+A9eJyS/AUGCdUiqo\nkrJ/O1K2u/xzcNVUQNX3x6pM54WrqvnIchI4dp+qdD0qqPi79K0wvrLfXBl3A8w1SqnVSql4XN91\n+e++4neQ757vVL+1kgrz2TkDlFIJ7njXKKUq1qw4OTZ2B+IfkeRdd32Aq2q2soZYB4DmSqlwpZQJ\n133c8o78cO0cm9hPljwNYD4w/khVsLt69yCu+1VVTbxHkkA8kKm1fkxr/Quue+Mopaq8T2qtn8FV\n7TyJo4nsmCTjLm8krnuZJ6SU6qaU2gD8pbX+P+AFoKt7dPntNB+Y6p7HF9d9xF/KFXW5e1xPXPdA\nj3syAPgKuBNYp7W2Ab/juqd75MppPjC53BXPdGCh+0r/yDqitc4GUnFViaKU6g50d38ejuv+8nKt\n9UO4TmCOrM/JJCmlYtyfr8NV9ZqPq53BkfUOBi5zr3dl33vF7XWrUsqklLK41/2GUwWhlFoK9HDX\n7lwLhLj/2Th6spgMDFFKKfc8w4C1nORk7jTMB6YopRq5/54BzCk3virLO4Ar6eP+DfX7JwForbu7\nTx56aldDtu+Bq9zlheH6nVd2Mp7Aaf7WTqH8d1I+3lXlTqzKP31g4KrduFgpFeyO46RPoIjjSfKu\newwoq1rbBGzRWh+uMG4T8DquasrluKoCy99bO/L5R2C6Uupujr8fdxyt9a+47mH+7k50lwHD3dVt\nld3PK/+54vJ/BnYrpbRSajGus/8MjlZ5niiWisMvB/IqjD9yz3s1rm1kBW4GUEpdppR6s5J1+xNX\nVe0qpVQKrnuAt7hHzwOeUa5H3KYDEUqp9biqpjfjuu96xNlKqVRcjcjGu6+AK/odV+OoI0l/Pq4G\ng0cej3sb+BX4Qym1CVdCnniCbXAJMMF9f/8BdzwAP+BqIb/BvT59cCUclFKfuRt6VWY38K57uXG4\nt5t7+YPcy1kJfO5OrBXjAVc7iE+U65G+6biqaNe5/23g6P3Wk+1vdwIPu7/D34EZ2vV0xXKgvVLq\nC/etomuAj5VSa3E1dBrhvq1yqv25st/DyaZ7C/gOWOHe97vhqqU5ZjqlVIh734uqpKyXgKZKqb9w\nnXwnn2S55cs9UXy34NoW63Cd+P2Nq9EeFeaZzz/7rR33Oz7FepV9J5XMXymt9Y+4fiOrcJ0Y5lP1\n70QAJnklqGhIlFJ+wCta66uqoWwnEKW1zjzTZZ9J7pOQg+4DaPnhV+A64Ti/0hlFlSilXgdu1Sdp\n2X+GlnM9sEa7Wtf7AouAB7TW86tpeTWyXqJq5FEx0dB0A56uprLrypmwgevKvrLhdWUdaiWllDeu\nltI1keA2AS8p12OBFuDTakzcNbleogrkylsIIYSoY+SetxBCCFHHSPIWQggh6pg6c887NTVV6veF\nEEI0OPHx8cc9ilhnkjdAfHz8GS0vNTX1jJfZ0Mg2PH2yDU+fbMPTJ9vw9FXHNkxNTa10uFSbCyGE\nEHWMJG8hhBCijpHkLYQQQtQxkryFEEKIOkaStxBCCFHHSPIWQggh6hhJ3kIIIUQdI8lbCCGEqGOq\nPXkrpc5WSh333lql1Ail1B9KqWVKqaurOw4hhBCivqjW5K2UuhN4E/CtMNwHeA44D+gPXKOUiqjO\nWIQQQoj6orqvvLcBo4GK/bJ2ALZprXO01jZgCXBuNccihBBCVIv8/FKysgprbHnV2re51vpLpVRc\nJaOCgJxyf+cBwdUZixBCCFFRTr6DPzYWsCp1H7l7M3CYbBwu3I3VKxKLXyAHgwIo8tpLiL07axe/\nSVbRarxbFJHAeKKDwskrmkeudTOXbs5jjuki/Jp059yejao9bk+9mCQHKL92jYBDp5rpRB20n47q\nKLOhkW14+mQbnj7Zhqevrm3DzZs3s2bbGnYU7mSoGkez5m3Ith1gbf5ykkIu5IsvvsSrSRNyIg8R\nld4PZ2kxB6w+7Ch6l4M/55OTk0Ngr2FEdtnCrB/mYsbJtshAnu/WjJEfa/o39mVix3ewtvuerCVx\nZK9ZiqXxKqK6RtFl3nUMCIGtUVa+OiuGsExNx+5DOHxgCzWxGT2VvP8C2iqlQoECXFXmT59qJnmr\nWO0j2/D0yTY8fbINT19t2IalRUU4AG+LNz5mHwCchpPX3n+NoMU/0zfSyjpnAT8EHmLImsa8ujyF\nP32yaTGpBb8/v5mAZh3wCSkhuMcGGs3bwhb9Cx3beuO8oDGTFzwCwH9bz8HokM7fm1YA4Nu2Fz7h\nOdisUTh9fcHqjV+BgSUkClNkCA9OiufP3AMk9o1l45+3kpG9jULzAdrEfIAtyEqL9gncpc6m47wR\nmPZm1thbxWoqeRsASqlLAKvW+k2l1K3AfFz33d/WWmfUUCxCCCE8oLCwkKy8LP7KTaP5D8k49mzF\ngcET/jvY+ZZm7b5DjL32TtK6r6F34WwMw8BOEUss79LnN02HxrmE+prZdEkXrv57IYN8YIvXQAq2\nW4mI6URopA+Gd1PIGMs5jluJDcsiwGamZFURtsBw8LNy94QurNg1noF3/pdmMY0ZdcHZRMeGEXi9\nHyZvb3oD48vF3AeAdgCMHjgcAJvNhs+Mx45fwb2Z1bwFj6r25K213gn0dX/+qNzw74Dvqnv5Qggh\nas6ePVncMXEcvgWH+E+rNmxvcyl7w9tSlB/IgnnPk1tyiJird3D95wX0LFqDwwRbr+zJqr2uO6dL\nvltBSCc7H87PBMMbk7mYsHND0E2u49mAYKwBXgzWdnZdOIVhYT5MHjaGmGYh+FrKt78+i0PJfpzt\nH4glqgWWqOZ4+fkD0BuYRI9/tW5Op5MNGzawZs0aRo0aRVBQ0GlurX/PU9XmQggh6pGsw3Yee3Qu\nH7xyDZEPtCP085202r6Gr/d3Y915n3N49WQKje4UHfiKom0j+C3GzNqC8RgYRPyxjsFn9SCk6yhi\n2nclppGZqGm+eJXl44+wxHvRvZ0/bWJ9MZkqPsB0vNCkUWd0/fLy8liwYAEZGRn4+/uTn58vyVsI\nIUTtZxgGa7cU8cKid7HnNSLc0Rsfiz97HMvITPmbGzY/xJu2UoyNuXTtdT57myVyYYveODlIu077\nGdFuHM0iJxER6o2X16kTcG1gGAZaa5YvX47NZiMuLo7ExET8/f09GpckbyGEEJVyFBXgLClmz/5i\nfp2/mdTFW3Bm/s2GLiYO5qaQk7qVgNixWNVW/FpkELormDcu7MHI5z8lILxpWTnj6eXBtTg9+fn5\nLF26FC8vLwYMGEDbtm2rdOVf3SR5CyFEPWQYRlmScRpOnl76NHckTMdpK8FRXMAzy5/ler+zKd27\nE2vv8/go/SPi4+PZta+U177cz4rC1xj03QbMu76ltE0j0nuGc92ONAC6+T/JVwUxtB4WRb9BMWSW\nDqZtCx/OufUefMKja0VyO1MaNWrEwIEDCQ8Px2q1ejqcMpK8hRCiHpiVMouJXScS5BtEQWkBE7+c\nyBfjviB1VSpzP5zL0g5LyX/mJy7avQAn8PWUngx783kySiCl61IW9s/g20WHueOF3eQV2mk6aj73\nLXQ9ptQnt5iAYcFk9hxEh5btuPz8c7ih+53llh7mkXWuKS1btvR0CMeR5C2EEHWMYRh8sfkLzo/s\nS8FvX1GyaytfB/5Bx19SaOljpVFCEmaTmR2Hd+B0Opn54kzCB4Tz3Q4fxsQFY/b24bJtEJI0mr02\nX+b9vIaYLlO47vF0AIb0CSbEfzyb/DbRu3cfhg49j5um3YS/j2fv81a3rKwswsLqxomIJG8hhKiF\niu3FFNuLCTIHYJQUsT5/B7uzd9Mvuh/btm1j6cGl5G9eRZenngTggnZhjPh6JfOUk4j1K3j7tV9J\n3+PNmt1BDBw1AwMIbN6EuT1Hli3jNaAgN5PCyAWkLhiMr4+JGddGc/kFjYHbeWrCTVgsFo+sf02y\n2+2sXLmSjRs3MmjQIFq3bu3pkE5JkrcQQtQChbZCcktyaWL48cOcN9hqWs2hrJ2M/nENhq2U7XM+\n4YlZT7D2lbUALNm0BJvZTk6HTewM7E6gyYy313Os73kDh4nk/Yl/k1dkuEt3PzZVCKk/ZldYsjf4\nDiY2wuDdh9rSoeXRq+uGkLgzMzNJTk4mJyeH0NBQjz7+9U9I8hZCiBpSemAvmR/PpCR9C3EPvUdG\n/mGaRTQDYNHfi/g17Vcuf/FT7l6wi4lnBZKdEI5ht+EVGES3yG5Y9loICgoiOjqG774t4Ie1zThw\n6FE44Crfr98oHt9zZGkGLZpa6NMlkC5t/PHxPnEjsgB/LyL90o5J3PWd0+kkNTWVtWvXYhgGXbp0\noVevXnh71420WDeiFEKIOi5n2U/sfOAy7IcP4jDg7cZ38m3BIh7s/ihdEoZRsK8daXu/IbvUj4HN\noymy9KCTeQzfTzwLm28IfAZDzvuSgQPhu8WHmZPsAOx0aePP2EEhxyTnIKuZszsHEhNe9SvnOvZO\nktPmdDpJS0sjMDCQAQMGEB0d7emQ/hFJ3kIIUU3sTjtmp8GeWf9j35wnufeSzjyb1o2PCsJ44v5Z\ndLivA1ffN4fAlq3cc1zPCvP10ML951oAB5B1XNnd2/lz68RIBp3VqF49mlVTvL29GTp0KAEBAXXy\n9oAkbyGEqAaFtkImfTmJRxoNo3jOk5jMZvwahXFo2stkfVaKT+CfZMy7CmurwbSK8SU20ofYSAsx\nERYC/LxOWrZq4Uu/7lZJ2qcpJCTE0yH8a5K8hRDiDDCcTjCZyhJqgE8AozuM5sEf5zGl92Q2N7mQ\n/PWxXPJVMYbhRcQ5nzB2UGNunBBB62a+Ho6+/jIMg+3bt9OiRQt8fHw8Hc4ZI8lbCCFOQ/FOTdb3\nc9j0xTuEP/oKy71yaFY6iuXrC1i+vhfLv3yDXyOHEBDjqhq3+JgYnRTC9AkRxEVL0q5OhYWFLFq0\niPT0dDp37kzfvn09HdIZI8lbCCH+oX3vP8Oema4exuxOg48z4fW9MPDOJ9k9OoycPy2U7O8CQEjr\ny+nepTWDEyPo09VKz/YB+PuevFpcnL60tDQWL15MSUkJ0dHRdO3a1dMhnVGSvIUQogpKS0tJTU2l\nT58+rgGG6xnqnwv8eGF3MQC/pzUhauk9JKjGnDM0jD5drXRv17nCu6ZFdXI6nSxcuJCtW7diNpvp\n27cvnTp1qnftAyR5CyEajPIv6yjPWVzIX1f2pWTvThyF+WXDTSYTPZcWYfL2pri4mCFDz+feZ1cQ\nEDgJbp2IYcDiZftpnjCN/K1jGHnBKB6+LppmEXWv9XJ94eV+CXh4eDhJSUl1ulHayUjyFkLUS38O\njcKRU6E3MZOJ7ovy2KS30LRp07J+rE2+/pTsSWPq2jyW57omfaoVDAx1fbY7DF7/phAixvHcO+vx\nCVLlynQS3acXScN3MXtCXPWvmDilfv36YTabyxJ5fSTJWwhRLx0qKmHeHhtnBUGbch2H2fbvZsaM\nGYwbN45x48YBrivsnKdep+XML1n+4ecA5D10Hy1HTmdvtoNpT/3N1qBHCe44hfFJsYSH+vCzbSp9\nve8jLrQ5E4Y8TTGHPbGaohL1qVX5iUjyFkLUOc7iIgo2rOTAl68T1HsIxBzbGOmll17i3jV28vPh\n5ZkvcvF115WN8/KxcE6/c1hWuoyR9pH4ertafM/c+RnDr7iXPiNfAeDr3Ovh6x188a2VQ3kOmg45\nzDN3+TPhnFgAVn7s5Jrzm9A8uAkAVprUxKoLN4fDwZo1a2jZsmWdeRPYmSTJWwhRJ5Tu28XBr98k\nL3UBBRtWYthKAShYtwwe+PSYaYODg8nPz2fo0KF07xmPl4+FzQc2E2gJpHlwc265+Rau/+56Fqcv\nZnCrwWzdVUzuzp7M+CITe67rfnVgm978tbsEZ7E/SQmNuHn0E3SKaV62jI/GfoTFLPe2PSE7O5vk\n5GSysrI4ePAgw4YN83RINU6StxCiTnCWFpPx1iOuP0wm/Nt1w6drIrp5Ak3N3hwuPoy3lzdWi5VL\nL72UXY130bF1R87pcA4AP2//mWC/YK7ofgUA1yVch5fhx5Pv7uPVzw9gs19IaJCZi0eF4u1tAiZB\nD1At/BgzMAQvr2Mbuvl5+9Xg2gtwtSTfsGEDKSkpOBwOlFJHW/83MJK8hRB1giMsmpgbn8SvhcLa\nIxHv4Mbk5eXRKzaWzz//nFdTXiXMP4xrE67F29ubmOjmzP9zMwc29wMgs6QVu4xCHk7Zy+79NnZn\nBrBjbyk5+ZkAXDqsMfdeGUXjIDks1kaGYfDzzz+Tnp6Ov78/iYmJxMXFeTosj5G9VAhRqzjycwAw\nW4MB1/PVzz//PM+8+gwPvPMANw4YWTbtygMrueeeezh06BCXDbmM11NfB2DH3hLmzunMph2t+bY0\nwz31kbd9HDxmeR1b+vH4tBjO6hRYreslTo/JZCIuLg4vLy8SExPx9284ry+tjCRvIUStUbovna03\nXYBPWCRtXvwBLx8LV199Ne+//z4+IT68v/V9bhx4Y9n0Dy14iHk3zWP7xu3EBsfy6MBH+XrBYe6c\nuZv8QjOxkaEM6xt8zDKaBHsTE+lDswgLzSJ8aNrEp9514FFfKaVQSsn3hSRvIYSHFWxOpXBzKobd\nxr7Zj2E7mAEOO46cLJ76aw5XTL2CP/74g5kzZ5LbLPeYeS/udDHFdlfvZoXFTma8vpcPfnQ92z28\nXzBP39yMYKu5xtdJnL7KOtSRpH2UJG8hhEflLJpHxpsPuT77e2PpnUi3x7/BOygUi9nCasdqNm3a\nVGmHG9PPng7Al5v3ct2Lmj2ZNnx9TMy4NprLL2gsB/s6qKSkhKVLlxIVFUXHjh09HU6tJclbCFHt\nDIeDkt3b8WvR7pjhf+77k/eCdjB91BQA3ir+G0uPdvQLcnVtNrn7ZIrtxSfsKWv3/lLun7WXn1d4\nAzY6tfbj+Vtj6dy6Yd8Prat2797NwoULKSgooLCwkA4dOsgJ2AlI8hZCVJui7RvJ+m4O2T9+gGE4\n6fr9LhxeJtJ3ppOSkkLrxNbkW/1ocZ+roVmnt1/m7SVvU3JBCb6+voT6h1JY7GTrrmL27LexO7OU\n3Zk2du0vZdf+UjZsL6K4xMDf1+CeyTFcMSIMb7Mc7Osau93OypUr2bhxIyaTiYSEBLp37y6J+yQk\neQshzjhHQR7bbxtJ3qrksmG+sW3ITd/KqB+nsfaVtRSmFbL6z9U8PujxsmmmXjmVzgs7U2wz8+js\nPXyzMIeDh+0nXdaIc4MZ0+cgQ5Kkh7O6asGCBaSlpRESEkJSUhLh4eGeDqnWk+QthDjjdj0znbxV\nyZitwYSeN56w4f8lsGsfnE4nub/mEtI3hP/0/A9hIWGEBRzbtWWOqTv9p2wh85Araft4m4gO96FZ\nhA8xERZiIy3ERvoQG2mhRVML0eEWUlMPVhaGqCPi4+OxWq0kJCTg7S1pqSpkKwkhzriICTexfNky\nQi6/GzVuArPXzOY642zMZjNvPfgW2dnZtO3Yj5W6ELTrhR5OJ3z6SzYLUl2v5IzvEMBjN0TTqZU/\nZqkKr9dCQ0Pp3bu3p8OoUyR5CyHOuADVHfuk+3nhvQ8YdukV/Ln/T77Y/AXjOo2je/fu/Lg0h/7X\naopLjOPmDbaaue/KKC4d1vi4LklF3WYYBna7vUG89au6SfIWQpxxuSW59B/en+zsbEwmEw/0fwAv\nkxeGYfD6lwd55K0MDAP6dg0kLPjoYSg63Iep4yJoEiKHpvomLy+PBQsWEBAQwKBBgzwdTp0nvxAh\nxGnJzc3lgQce4MEHHyQ01PWI16q9q3h66dPMnDQTgGZBzbA7DO55eQ/vfe/qROWeK6KYNj5cWhTX\nc4ZhoLVm+fLl2Gw24uLicDgcmM3Sec7pkOQthPjXVq9ezcUXDidtTwYH9u9n7kcfAdAusB/WvQe5\n9uGDeNlcVeP5hU7S95Xi62PihdtjGdk/xJOhixpQWFjIokWLSE9Px8fHhwEDBtC2bVs5YTsDJHkL\nIf4xw+EgZ/F3LH3wVtL2ZNDWH6Yl9WD9/vXs3t6c257bzeH89u6pi8vmaxxs5p0H4+jVUV4C0hCs\nX7+e9PR0oqOjGTBgAFar1dMh1RuSvIUQ/0jWd3PY8/I92A5m0Bf4v24hjL/pbrwvmMCEr65h/7rz\nKMg/j8FnN+K2iZHHtBRvFeNLgF/lvaWJ+ic+Pp7g4GB5mUg1kOQthPhHTL7+2A5m4Bvbhr2jx7Ij\nOIu0ptfz+IP72LnvbvxCMnj4umiuGhkmB+wGztvbm/bt2596QvGPSfIWQlTZ5s2bie09GPX2UgK6\n9Gb9kr85NYFgAAAgAElEQVRY8+ddfPvqTgDimkbx2r1n0bVtgGcDFTXKbreTn59PSIi0Y6gpkryF\nEMcp2PAHu569ibYvzcdsDQJcrYavuvoqvC/2ZvLZt/HpTdv5c6sNeIQmId5cPzac/w4Pk2rxBiYz\nM5Pk5GQMw2DMmDHyDHcNkV+ZEOIYRWmb2Dp9GAXrV3Bw3jtlw00mEz/9+BMq61xemP8rf24tJCLU\nhxnXxLDi3fZcPzZcEncD4nA4SElJ4ZtvviEnJ4cWLVrIbZIaJFfeQogypfvS+fDS/izaeYh1plBm\ntz2HSCB5RzIdwjuwYIWF+avHYXfA0D5BvHxnLIH+8rxuQ5OdnU1ycjJZWVlYrVYGDBhAdHS0p8Nq\nUCR5CyEAsB8+yNZpQ/lky0F+OQRwiIVLltAjIYF1+9cz69dklnwxFoDrxjThviubSp/jDdThw4fJ\nyspCKUWfPn2wWCyeDqnBkeQthABg/4cvULzzL4Z3bE7rfudh7R7Klf+5EsMw2LNmEKlrFmL2giem\nxTDpP2GnLlDUW61atWL06NE0aSKvYfUUSd5CNGCbNm2iQ4cOmEwmoqY8AIbBdeOmUtQokNGfjCaP\nPN7/vJgPvnFg8UnkvYfjSEpo5OmwRS0giduzpHWJEA2Uw+Fg0qRJPPTQQ66W5N9fg23iFCzh0QT7\nBfPKBa+weKWFx97eh8kEM++IlcTdwBQWFrJjxw5PhyEqUW1X3kopL+BVoCtQAlyttd5ebvwo4F7A\nAGZrrV+rrliEEMd7ffXr3PnmnexbvA+AxOaJfP3X19zc+2YAMnbGcOcLrgP3Q9dGc+G58gxvQ5KW\nlsbixYux2WyMHTtWnuGuZaqz2vwiwKK17quUOht41j3siOeAHkABsEkp9ZHWOqca4xGiQSm2F7Mv\nfx9xIXEArM5YTW5JLgPiBgAQERjBhswNzLh5BgtX57H4u0TsNi8u+96VsJevL8DugKnjwrn6Iqki\nbShKSkpYtmwZW7duxWw207t3b4KDgz0dlqigOpP3OcBPAFrrlUqphArjbUAI4ARMuK7AhRBnyM7D\nO3lk4SPMHTMXgFJHKW+lvlWWvBMLgxnc4Qre+z6Le1/Zg9N5fBljB4Vw7+SoGoxaeFJmZia//PIL\nBQUFhIeHk5SUJFfctVR1Ju8gILfc3w6llJfW+sgh4lkgFdeV9xda69yKBQghqs7utPOt/paRaiRm\nLzO+Zl+aBzcvG7/518388cUfbOu2jRYhVjLunEBhqYnn/d/DaW7G9WPD6dPl6Nu+GgWa6dUxQDre\naED8/f2x2+3Ex8fTo0cPvLykWVRtVZ3JOxco37qlLHErpZoD04AWQCHwgVJqrNb682qMR4h6zYSJ\nX7b/woGCA1ybcC0tQ1vyxOAnAEhJSeHKyVcCMO+s9xny13c4crPZ6NOHLJ8Ynru5GROGNPZk+KIW\naNSoEZdccok8t10HVGfyXgqMAD5TSvUG1pUb5wc4gBKttVMplYmrCv2kUlNTz3iQ1VFmQyPb8PT9\n222Yb88nvTCdjkEdARjqM5Sfvv+JVNOx5WmtAZh20VASf3qW4pIC9ntF82bYDGZc7qBt2A5SU+t2\nq2LZD0+fbMPTV1Pb0GQY1XOrWSll4mhrc4DJQDxg1Vq/qZS6BbgUKAa2AVO01vYTlZeammrEx8ef\n0RhTU1M502U2NLINT9/pbMNdObu44usr+OaSb7BarOTk5NCzZ0+efPJJxo4dWzZdVlYW6xf/TqPH\nLwGng2WWQXze8lHeeKw7Ks7vTK2Kx8h++M9kZ2ejtaZ3795lt0VkG56+6tiG7jKPu3dVbVfeWmsD\nuL7C4C3lxj8PPF9dyxeiPjIMg483fMwINQKrxUpscCxT4qeQX5qP1WIlKCiIOXdPJTLy2IqssLAw\nLC3PZ27wLeSWWtjT5b989nBLIhrLG6AaEqfTyYYNG0hJScHhcNCiRQvpk7yOkh7WhKhDTCYTW7O3\n8vqq17mq41UsWbKECcMnAFCydye7npqG/5Lv2R3di5dXtAWTq8GRzW7w/ZIcbObLGDYoiC/ubC5v\nAGtgcnNzWbhwIRkZGfj7+5OYmCiJuw6T5C1ELVdsL2bV3lX0a94PgOlnT+fe5+6lxfktKCoqYsvm\nTXglf8b+tx/By1ZEgcnK3JzB/PjrIQzTsQn62jFNuF9eKNLgHDhwgO+++w6bzUZcXByJiYn4+/t7\nOixxGiR5C1HLpeek82Dyg8wZNYdmQc0I8Qth24Jt5Obm0iPhXOZffiEJJZvxAhZZhvJ+6J2c3a81\nT/dohHe5t3U2j7LQu4vVY+shPKdx48aEh4fTrl072rZtK4//1QOSvIWoxdatW4evry8PJT2Ew+kA\n4K+dxUR3uZp2XME+r86sLppLtK2A5E4ziB93IQvODSEoUN6xLY4ym81ccMEFkrTrEUneQtRC6/ev\n55e0X7Cut/LKK6/w3Vc/sXjueq7ZXMr6bUVAe/CCuKYW1MDpdDn3Xka0kC4shatRY2VJWhJ3/SLJ\nW4haqFVoK1L2pDA2fjLdg6LZMrYlzb282BS2mCCrLyP7hzB2cCgJHaQHNHHU7t27SUlJ4fzzz8fP\nr+4/AihOTJK3ELXAnj17eGvuW/Ts05MRiSM4cNCbDqsvwPT09Ux3pIEXZAR3Y9ZUfwYPa4+fRVqK\ni6PsdjsrV65k48aNmEwmMjIyaNmypafDEtVIkrcQHvbxxx/z3HPPEdwzmDZmxQc/tSZ6ySwmFbwM\nQI41jqhbXmL4yOEejlTURpmZmSQnJ5OTk0NISAhJSUmEh4d7OixRzSR5C+FBdqediObdMJm8se1N\n4MCGtuzNzqaD39nYi98kcOztDJh+P16+UgUqjpeXl8c333yDYRh06dKFXr164e0th/WGQL5lIWqQ\n/XAWeWsWUbB+BQe3bOOVRukkH7qMsMQfMfs2oWOQPxePDmXkgN40cgzDO0Teoy1OrFGjRiQkJBAZ\nGSkdrjQwkryFqAGG3cbmy89izdq1BJuhqa9r+MgQP1KH/MaYziO5+sJI2jYvf4UtiVucWo8ePTwd\ngvAASd5C1ACTtw8mLy9WF5r5aUhLzt3SmwLvTvTo14HHu4UzbGgLT4coarnS0lJ5VacoI8lbiBqQ\nX2Tjs44v8VmmN75+T7NzRDveveI2IsN85DWM4qQMw0BrzYoVKxg2bBhRUVGeDknUApK8hahmLy1+\nn48XbmHHkoux+Jq4ZeDjTBgSRuMAeaOXOLnCwkIWLVpEeno6Pj4+FBUVeTokUUtI8haimjidTv73\n1Dd8sjoOa++PiIu5mNfvaUOXNvJCCHFqaWlpLF68mJKSEmJiYujfvz9Wq/RNL1wkeQtxBixMS8Zq\nsdIzqge24gKuevN8Nn8ZTOqSn2ikbmNI/Es89VIc1gDpc1ycWmlpKUuWLMFut9O3b186deokPemJ\nY0jyFuI0rZ71Mj9sfJLQglJYnwlAUfumZDbzxuRlYeplvXjirtYejlLUJRaLhUGDBhEYGEhISIin\nwxG1kPSxKMS/tC9/H0UlTs57+lcydBCt9hXiwIwDM2O3eOFYFs7s977gifvGezpUUQfFxMRI4hYn\nJFfeQvwLxfZipnw7hQ5MoKjEYGnBKOJvvJF97qrNsGAz2weG4Ocr1eTi5LKysggNDcXLS66lRNVJ\n8hbiH3hnzTuM7zyeAJ8Anj3vBSY+9yUhXR7hoev9mDQy0tPhiTrE4XCwevVq1q5dS69evejevbun\nQxJ1iCRvIU7AcDhY/9MbeH34Jra/NQDvXdyW+F+W0/XON1i7NpQ9a4bSJtaXS0e09XC0oi7Jzs4m\nOTmZrKwsrFYrERERng5J1DGSvIU4gYPfzmbW8qfJaevDdTtKCCx1MGxtBpY2dhwOg5c+OQDAjeMj\n8PKSlsDi1AzDYP369aSkpOBwOFBK0adPH+k5TfxjkryFOIHGwy5l2tdvsm7oufS+9V7MPr50BxwG\nPPDkh2zb1ZnmURYuGiCNikTV/f3331gsFhITE4mLi/N0OKKOkuQtRCVy8nP4YM4HTJ3zB53KDS8q\nKmLqtKm88847+EdfyA03zsbHW666RdWYTCaSkpIwm834+0tnPeLfk+aNQlRwoOAAl3x9CU+8/8Rx\n4woLC3nnnXcAiGo9lPFDQms6PFHHWa1WSdzitMmVtxCAPScbvLzwbhRCeGA49/S5h1l7Z5Fb4ODa\nx/7mj40FADjthVhbXoY5sA13TL8YP4uc/4rK7dixg4iICAIDAz0diqiHJHmLBs/pdPLbLeNxFm7l\nnHs/JKhrXxJbJ3LWzecw8f4dLFtXUG5qfxq1u524phYu+09jj8Usaq+SkhKWLl3Ktm3biIuLY8iQ\nIZ4OSdRDkrxFg1ZUVMS0i0fwwQ+/MeT+9mw9uIDp9MXpNJj+9C6WrSsgItSbL55uTdMmR98C5mcx\nSQtzcZzdu3ezcOFCCgoKCA8P56yzzvJ0SKKekuQtGizDMLh54lhmf/8bPiYYSRJGcDhOp5MHX89g\n3uIcGgV4MfexlrRu5uvpcEUtt2zZMjZs2IDJZCIhIYHu3btLr2mi2kjyFg3WzsevwdZ8I52a+nBf\n/z5MeOgVTCYTr3yWydvfZGHxMTH7wTg6tZLGReLULBYLISEhJCUlER4e7ulwRD0np4WiQSgsLGTy\n5Mk4nc6yYSG9h9HSZiLpseFM+OB3TCYTv/2Ry2Nv78Nkgpl3xHJON3l/sqianj17Mnr0aEncokbI\nlbeoV/Ly8nj//fcZPnw4zZs3LxseEBDAmjVrSF6QTEj7EOKj4wkZOJo7eySS62vCZDaz90Ap05/Z\nBcBd/43iwnOl8xVRdV5eXlJNLmqM7Gmi3sjOziYpKYmpU6eSmpp63PjZs2cT0SqCu369ixW7V2Ay\nmfBtHEF4YDh2h8HUJ3dxKNdB/55Wpo2TqydxPKfTybp169i1a5enQxENnFx5i3rjxRdfJDU1lWbN\nmhETE3PMuNIDe+nZsycAbwa/yb78fceMf27uflZucLUsf+nO5tKSXBwnLy+PBQsWkJGRQVhYGM2a\nNcNkkv1EeIYkb1Fv/O9//6OoqIibbrrpmOR96Pcv+fXFq+h/7Uwi/nMZLUNb0jK0Zdn4xWvyePGj\nTEwmeOXu5jQJkZ+FOMowDLTWLF++HJvNRlxcHImJiZK4hUfJUUrUSTabjXfffZcpU6aUDfP29uap\np57CMAyKtm8kL3UB+asXcmjhN8wb3JzFuz/leeelmL3MZGbbWL6+gBXrC/h24WEMA26dGCEN1MRx\nlixZwubNm/Hx8WHAgAG0bdtWErfwOEneok4qLS3l5ptvPiZ5H+HIyWbT+M68OqQ1I9bsJdZWyoyQ\n89k68AIys51c9XAaa7cUHTPPgHgrt1waWVPhizqkVatW5Obm0r9/f6xWObkTtcMpk7dSqjHwJNAG\nGAc8BdyqtT5UzbEJcRxb1n5yl/9EcXExEwf2oWR3Gr7NWgGwdt9awvzDiA2JJajPUPKb5tPkyivo\n1Hscfi3a0Qq45bldrN1ShL+vibM6BdKnq5U+XQPpqQIwm+VqShwvJiaG6OhoudoWtUpVrrzfBH4G\nzgbygD3AB8AF1RiXEJXacf+l5KX8DsC1QMGmlLLknZ6Tzt2/3s3skbNp+9JPvFxwgGC/YCxmCwBp\ne0r4/NdDmL3g11ntaBktvaaJYxmGUWmSlsQtapuqJO+WWuvXlVLXaa2LgfuVUuuqOzAhKpr1+AzC\nF/1O68aBhA4eB4Bv9NGGZxeqC4kLiSMy0FX9HR547ONez8/dj8MJlwwNlcQtjmG321m5ciU+Pj7S\nH7moE6qSvG1KqeAjfyil2gKO6gtJiOPt2bOHqfc/hH+QN89OHsaEac8R4hdCsb2YxX8vJrFFIgBd\nI7tWOv/W9GK+TD6Mj7eJmy+Re9viqMzMTJKTk8nJyaFx48bEx8djNps9HZYQJ1WV5P0gsABorpT6\nBugDXFmdQQlR0WeffYZhQO9zFckxTpLy9xPiF4Kftx+zVs0ipySH4e2Gn3D+Z+fuxzBcV92xUZYa\njFzUVg6Hg9WrV7N27VoMw6BLly706tVLEreoE06ZvLXWPymlUnHd8zbjutUojdVEjeratStjxoxh\n4sSJdO3flUBLYNm4Ue1H0a95vxPOu3lHEd8uzMHXx8T0CRE1Ea6oA1JSUli3bh1Wq5UBAwYQHR3t\n6ZCEqLKqtDZfrrXuA3zn/tsMrAW6VHNsQpQZOHAgAwcOrHTcxZ0uPum8z7y/H4BJ/2lMdLhcdQuX\nrl274nQ6SUhIwGKR/ULULSdM3kqpZKC/+7Oz3CgH8E01xyXEMQ4XH+aOn+9gTMcxDGszrErz7NpX\nyie/ZPPjslz8fE3cOF6uusVRAQEB9O3b19NhCPGvnDB5a62TAJRSM7XW02suJCGO18jSiEldJ7H5\n4OaTTrc/y0Zyah6f/3qIZesKyoZfPzaciMY+1R2mqIUMw8Bms8nVtahXqtJg7U6l1CjACphw3fdu\nqbV+oFojEwIo3qkp3b+LRmcNon9cf3qE9+PXlbk4nEbZNDn5Dv7YWMiK9fmk7SktG+5nMTGsbzDj\nBofSP156xmqICgsLWbRoESUlJYwYMUJe2Snqjaok7y8Bf6AtsAg4F6k2F9Vk17M3c3jRvLK/396Y\nwbJCG3dOvIyxT8/mhv9L5/eUvBPOH+DnRa+OAQxPDGbEuSEEBUrL4YYqLS2NxYsXU1JSQnR0NKWl\npfj5+Xk6LCHOiKokb4Wra9SZwGzgduD1U86klBfwKtAVKAGu1lpvLze+F/Asrqv5PcDlWuvSysoS\nDUNJSQlFzTpiy5yFYXPtChcHwd7zW/J2+134pa7n9xQDf18T/bofvZL29fGiR/sAencJpEsbf7yl\nm9MGraSkhKVLl7Jt2zbMZjN9+/alU6dO0kuaqFeqkrz3a60NpdRfQFet9RylVFQV5rsIsGit+yql\nzsaVqC8CUEqZgDeAMVrrNKXUFKAloP/daoj64Ouvv+byy2/kthuu5cHbbiob/mFwGBsKdvL8m96A\njckXNuH+q5p6LlBRq23bto1t27YRHh5OUlISISEhng5JiDOuKsl7o1LqJWAWMFcpFQ1UpW/Jc4Cf\nALTWK5VSCeXGtQOygFuVUp2B77XWkrgbuNdee43S0lKatW2Pb7PWzF4zmxC/EEY3G41fVnt+WboF\nXx8T14xq4ulQRS3WsWNHvL29adu2rdzjFvVWVfbs64FPtdabcPW2FgVcWoX5goDccn873FXpAE2A\nvsBLwGBgkFIqqcpRi3qnqKiIIksRzYY1Y9T4UQDkluRSUOpqMf7yJ5kAXDKssbQaFydlMplQSkni\nFvWayTCME45USikgT2u9t9ywSOARrfU1JytYKfUssEJr/Zn7711a61j35/a4Tgi6uv++GfDRWj99\novJSU1NPHKioF/YV7+OzHZ8RYY1gfOx48u35mE1mDh/257rnvTGZ4I1b7USEejpSURs4nU6Ki4sJ\nCAjwdChCVKv4+PjjGmycrJOWGbgap+F+VCzZ/fe9wIoqLG8pMAL4TCnVGyj/JrI0wKqUau1uxJYI\nvFWFFajCYqsuNTX1jJfZ0PzTbViUtomSXdvK/i7Y+AdmazCRl96MyTueC8654LjXMt7x4m6cRjbj\nBody/uDYMxp/bSD74T+XnZ1NcnIy+fn5jB07lr/++ku24WmS/fD0Vcc2TE1NrXT4ye55/xfX42HR\nwCPAXUAkcLHWen4VlvkVcJ5Saqn778lKqUsAq9b6TaXUVcCH7sZrS7XWP1ZtVURdVZq5h80Te5S1\nJC8vQPUg6OzBwLHvTs44aOPTXw5hMsG0ceHHzScaFqfTyYYNG0hJScHhcKCUwsdHbqOIhudkyTtX\na50BZLgf63ofuENrXaXXgWqtDVz3y8vbUm58Mq6XnYgGwhIRQ7tZv5H+5FQsUS3Khgd27cvtmZ/S\nc9U2SndcwII/jib3zGwbNrvBiHODaRMrz+g2ZHl5eSxYsICMjAz8/f1JTEwkLi7O02EJ4REnS97l\n+zM/CNzmTshC/GvW7v3o+NGfZX+vX7+eGU8+yZirxrAmaydvzT4IxrEdq3ibkbeBCYqKiti3bx9x\ncXEkJibi7+/v6ZCE8JiqPCoGUCyJW5wJO3bswNfXt+z1i7GxscTHx6NXarbmjgMjnwlDQrl48NFW\naZFhPrSKqcrTiaI+i4iIYPTo0TRu3Fg6XBEN3smSdyel1A735+hynwEMrXWraoxL1ENbt26la9eu\nzJw5kylTprgG+sGUqVNYvcFg5r07aBTgxf1XN6VxUFXPK0VDEhYW5ukQhKgVTnaEbFdjUYh6q3zL\n8dtuu43i4mI2b9vMxsyNdIroxJL0JTy3/Dmc2yYBCUwbFyGJu4ErKSnh77//pl07OQQJcSIneyXo\nzhqMQ9RT6Y9fR6EP7B08nJ07d9KoUSPGXDmGxxY/xodjPmR4u+EcSm/PPV/vJirMm6sukt7TGrLd\nu3ezcOFCCgoKsFqtZbdXhBDHkkscUS0K9Voy3nqYw8lfkRUSwJPhmhUpK9iyeQuNoxrTcX9HAEpt\nTl6da8dZ1JjbrokkwE96xWqI7HY7K1euZOPGjZhMJhISEoiKqsorFIRomCR5izOqcOs69s76HzmL\nvgXA5OtH92ueYnQc5NvzadG6Mwey7YxrcTtb/i7mx2U5/J1RSptYX8YPaezZ4IVHHD58mPnz55OT\nk0NISAhJSUmEh8sz/UKcTJWSt1KqH9AZeBc4S2u9qDqDEnVX8c6/yFn0LavbRbC3b29uv/glAiOb\nM3J/Kc/NzuSj+Zux2Y9/cOGeyVHyKs8Gyt/fH4fDQZcuXejVqxfe3nJNIcSpnPJX4u53/CJcPa19\nAbyhlHr7ZP2Qi4YrdOAYSqY+TothF/N/61/jw83L2fKxF5/+cgib3cBkgtbNfCn/zoizOwUyrE+Q\n54IWHuXr68vYsWOxWCyeDkWIOqMqp7hX4OoJbYXW+oD71Z4pgCRvcRyT2UzTyfcA0F/fxuj/JOAb\nOYTg9jczOimUmy6JoG1z6SlNHEsStxD/TFVaBzm01iXl/i4G7NUUj6jjPt34KXty9/DT8hymPXOQ\nxme/R8/uHVj4huLlu5pL4m7A8vLyWLRoEQ5HlXpYFkKcRFWuvBe6X+9pVUpdBFwD/F69YYk6o8Ir\nZW0OG5M+v5K0bx7GZvfn+gkdePCagdIjVgNmGAZaa5YvX47NZiMqKkqe4RbiNFXlyvt2YCvwJ3A5\n8ANwW3UGJeoGW9Z+eHIyeWsWlw2zHrqA7d/+D1uJP1PHhfPgNU0lcTdghYWFzJ8/n0WLXG1cBwwY\nQNu2bT0clRB1X1WuvJ8H3tdav1bdwYi6w3Ywgy3XD4Idm/n4k7vZvn0K2/5IZNHyjZi8rdx2RQR3\nXB4pibsBy8vL46uvvqK4uJjo6GgGDBiA1Wr1dFhC1AtVSd5bgReUUmHAXOAD6X2tYSvdv5st1w+k\nJH0rB61teDXtVoyoNzi4vRGH196BnzmfYS/+iskknWw0ZFarlZiYGCIjI+nUqZOcyAlxBp2y2lxr\n/bLWuh8wDFdjtW+UUkuqPTJRK5Vk/I2+pj8l6VvZ7d+BmyxvsT+3Lc33PE1s1lxs+Tto3iyCNm3a\neDpU4WEmk4mBAwfSuXNnSdxCnGFV7aQlGBgMDAHMwPzqDErUXkVb/qQ0Yycp0Z15tUtfQvf688Oj\nilYxvvz++zSmTPmTzz//nICAAE+HKmpQ+RfQlCdJW4jqUZVOWuYBPYEvgf9prVdWe1Si1spTw3g1\n9lUWFbQjvNE8Bl44n1YxZwEwcOBAtNbSQ1YDk5mZyZIlSxg0aBDBwcGeDkeIBqEqR9k3gB+11vJs\ndwNiz8lm3e2TWN5oODuaX1g2PHlVHhkFfeih/PnghifZtmXtMfNJ4m44HA4Hq1evZu3atRiGwZ49\neyR5C1FDTnikVUo9pLV+EBgNjFJKla//MrTWV1Z7dMIjnE6Dr5+aTcs1P+Ljc4iPtsRj2P0BaNTx\nS7rFJvLx/QMoKsjC6XB6OFrhCdnZ2SQnJ5OVlYXVamXAgAHy+k4hatDJLpNWuf9fAFS8cXX8myVE\nvXDwsJ3pT6cz9LePASjtPZZmHW5meux3AOy0d2N57gt4W/ryxXtf8Pbbb/Pjjz/KW6AakNLSUr79\n9ltKS0tRStGnTx/p3lSIGnbC5K21nuf+GKO1frz8OKXUE9UalagRK9bn81Xy4bJO0gzglxW5GAd3\ncZs9FcPHj8vvn8hX3/3A2IFWvL29MZtHUWgbir+PP9dccw1LliyhtLTUo+shapbFYqF37974+fn9\nP3t3HhZV2T5w/DvDvsqugguicFTUVNyNxC01rTSXSqNySy3rZy5l5ZJbmUu9ZflqpZZLuKWZllq5\nv5ik45aoR1wQUZRNlHVgZs7vD/IEgoiyDMjzuS6umrPNPQ849znPec5z4+vra+5wBKFKKqrbfC5Q\nHXhGkqQG/Hv1bQm0A94r+/CEsqIoCuMWxhJzvWDifdt1F8YU8Ah5hvenz2b3V7uxfdWWRYsWMXbs\nWOytckeSW1hYMH78eHx8fMo7fMHMGjZsaO4QBKFKK6rbfBPQGOgK7OPf5G0AZpVxXEIZO39FT8z1\nbFydLZj8yr+TqbhXs8Tv+whWt/LmQsMMbOQ0tUtUqy3ObLrCo0Sv12NjY2PuMARBuEtR3eZ/AX9J\nkrRZluVb5RiTUA52HU4FoGtrJ0J7u+dbZ2r7O2+H/0qUryc+fevw3UffmSFCwdwuXrzIgQMH6Nix\no5h0RxAqmKK6zY/JstwCuClJ0t2rFVmWLco0MqFM7frrNgBdWjsXWKe1ssY9pC/uBdYIVYFeryc8\nPJzz589jYWGBwSCeEhWEiqaoK+8W//xX9JU+YlLTjUScSkerhZCg/IUiTp06hd5eT3Pf5lhoxflZ\nVRMbG8u+fftIT0/H09OTzp074+LiYu6wBEG4S3FmWGsAtAXCgCVAC2C8LMsHitxRqLD2H0vDYIQ2\ngb0OyGIAACAASURBVPY42phQjMbcFRoN23dsZ2XSSmpINdj04iacbJzMG6xQbkwmEwcPHiQjI4NW\nrVrRvHlzMc5BECqo4kyHtQJYBDwDBJBby3sBuQldqITudJmn/dyT51de4P26uct9/m8+kyZO4vE/\nH6dmQE2RuKsYrVZLly5dUBRFPLcvCBVccU6rbWVZXg/0AX6QZXk/xSxoIlQ8JpPC7sOpOMetYtf5\nC1zVg6LRgFarFpFo3749vu6+5g1UMAsPDw+RuAWhEihO8jZIkjSA3OS9TZKkvoCxbMMSysqpC5nc\nSM4h8+JSAJ7qGkKrwyaC/jJiO2AEW+WtpGWnmTlKoazdvHmTnJwcc4chCMJDKk7yHgU8Bbwhy/I1\nYBAwokyjEsrM7sOpaDQaXh75X/6vvcQ7S1azPnI9KVkppGansjd6LzP2zjB3mEIZMZlMnDx5kk2b\nNnHo0CFzhyMIwkO6b/e3LMsnJUn6DAiRJGkcMF+W5ZNlH5pQFnb9lft8d8eBfej56RAAwvaE0a5W\nO+pUq8PCHgsxKaLYyKPo9u3b7Nu3j7i4OOzs7Khdu7a5QxIE4SHd98pbkqRQ4CegHuALbJIkaXgZ\nxyWUgaRbBo7KGVhbaQhu/u8jYgMaD8DZ5t/nvbUaMcL4UXP27Fl+/PFH4uLi8PX1ZcCAAWJeckGo\nxIoz8Gwi0EaW5SQASZJmkztd6rKyDEwofYtX/IbJ5EO7po7Mj5iFk7UTrzZ/lSHNhpg7NKGMxcXF\nARASEoK/v786OFEQhMqpOMlbeydxA8iynChJkhiwVgnsO5qK7kyG+vqLz+aSU60H3dqMolfLkWw+\nu1l8iVcRHTp0oHXr1jg6Ot5/Y0EQKrziJO+TkiT9h9wrbQ0wHDhRplEJJRaXkMnRt16gSfZhddlz\nWg3blcfp0toJH2cPxrYZa8YIhfJkY2MjCowIwiOkODc3RwLZwHJyJ2zJBl4vy6CEkuvboyffH/+V\nhLQE3E25P6+5x7O25iqqOSabOzyhjMTGxpKUlHT/DQVBqNSKvPKWJMkDqAvMkGX5nfIJSSipzMxM\ndH//hdEAtiHDaPrhTHXduZx4BmwezKDAQYxpPcaMUQqlyWAwEBERQWRkJB4eHvTr10/cEhGER1hR\nVcUGknulnQZoJUkaJMvy3vIKTHh4qzaGYzRkY+famGcXfYulxb9f4k3w4acXfuJ62nUzRiiUpvj4\nePbs2cOtW7dwcXEhODhYJG5BeMQVdeU9FWgty/IZSZJ6AB8CIeURlFAyp+MlvDrv4vlOhnyJ+45q\nttWoZlvNDJEJpe348eMcPnwYRVFo2rQprVu3xtJSzF4sCI+6ou55m2RZPgMgy/JOEOWdK4PUdCNb\n96dgYe1Gt372rI9cr67TXdMRcyvGjNEJpc3Ozg4HBwf69OlD+/btReIWhCqiqOSt3PXaUJaBCKVj\ny74UMvUK7Zs6kKaJZtfFXeo6D3sPnlv3nEjgj5CAgAAGDhyIt7e3uUMRBKEcFXWa7ihJ0hP//L8m\nz2sNoPxTXUyoYNb+egOAwT3d6N90MJK7pK6rZluN2V1mU6daHXOFJ5QyjUaDlZWVucMQBKGcFZW8\nrwIzinjduUwiEh7a6YsZ1Nr1As9V09ClxveAK0HeQep6F1sXejboab4AhYeiKAqyLKPRaJAk6f47\nCILwyLtn8pZlOaQc4xBKwZr1f3Ps6il2ubpw7PwShruPpnmN5uYOSyiBjIwM9u/fT0xMDPb29tSv\nX1/c1xYEoVgzrAkV0PkrWYydd4WYuGx1WfukbSwOUMgIaEay1ItbWbfMGKFQUhcvXuTAgQPo9Xq8\nvb0JCQkRiVsQBEAk70op6koWA9+5SOzFvzCkX0Zj6UQDjwb0yNgIQOMBQ3Fv+KyZoxRK4ujRoxw5\ncgQLCws6dOhAYGCgeHZbEARVmSVvSZK0wGKgGaAHRsiyfKGQ7b4GkmRZfq+sYnmURMVkMeDdi8Sc\n2UHKiQkA+Ac0ZMMrr5P0+VmS3J1oHtLPzFEKJVWvXj2uXr1KcHAwLi4u5g5HEIQK5r7JW5IkN+AT\noAEwCJgHjJdl+eZ9du0LWMuy3EGSpLbAwn+W5T32KKAJsPfBQ696zl3OTdyJKQa6dn8Ky0Y6Um4m\nERQUhFtAI3J6DWG2Xzxrd43j66e/xlIrOlYqK1dXV55++mlzhyEIQgVVnG/3b4DfgLZAKrmjzlcD\nve+zX0dgB4AsyxGSJLXKu1KSpA5AG2Ap0PDBwq56Ll7V0/+dCyTdMvJES0dWTPfF1np1vq5U57bd\n2GgycibxjEjclYii3D2lgiAIQtGKU1WsnizLSwGjLMtZsixPAWoXYz9n4Hae18Z/utKRJKkmMA0Y\nS+5z40IRbqcbefXDaDVxf+zyBak/LUbRZ6nbvP7L61xIvoCF1oImXk3MGK1QXEajkcOHDxMdHW3u\nUARBqGSKc3mWI0mSOhG2JEn+gLEY+90GnPK81sqybPrn/wcAHsCvQA3AXpKkM7IsryzqgDqdrhhv\n+2DK4pilyWiC2astkM8n4WZ1ibGtXUn+6AsWPV6X+taptK3dHY1Gw7Xka/wd+Tcp9inlHmNFb8OK\nKDMzk0uXLpGZmYm1tTWHDh0Sk62UkPg7LDnRhiVXXm1YnOQ9ndx70nUkSdoCtAeGFWO/cOBpYIMk\nSe2Ak3dWyLK8CFgEIEnSK0DD+yVugKCgoPtt8kB0Ol2pH7O0zVkexxE5AVvDBRL+nsKFTS1ooij0\nqdGR1dooXmj8Gu727ixvvBwnayesLMo3AVSGNqxITCYTp06d4vjx4xiNRiRJws7OjjZt2pg7tEpN\n/B2WnGjDkiuLNrzXycB9k7csyzskSdKRe3/aAnhNluUbxXjPzUB3SZLC/3k9VJKkFwFHWZa/uWtb\ncdOvEJv23OSr9QlYaGHNosH4Wrbi79A2YKWlw+AP6VyrvnrP283OzczRCsURGRnJoUOHsLOzIzg4\nGF9fX3G1IwjCAyvOaPPp5CbXO/emm0uShCzLM4vaT5ZlBRhz1+JzhWz3fTFjfaQdOZ3OpM9jycr+\n9zzmWkIOADNGe/N4c0diFnzHucYu1KvXDdvaDcwVqlACjRo1IjU1lRYtWmBnZ2fucARBqKSK022e\nd0CZNdATOFQ24VRd81beQL6sL7B82DPuDH3aHUVRuHXuGGd9nDkSaGSloohJOyohS0tLOnToYO4w\nBEGo5IrTbf5h3teSJM0Efi+rgKqis9FZ/O94Gva2Wrb9pwG21rlJOSc7nZoe1mqSbrJ0PwvOHiW1\nto9I3JWAXq/HxsbG3GEIgvAIKs6jYndzoniPignFtHxLIgADu7nS0NcWX28bfL1t+Oo/M5Ekid9+\n+4349Hg0Gg0OjYKo4VjDzBELRdHr9ezevZvNmzeTk5Nj7nAEQXgEFeee96U8LzWAKzC/zCKqYm6m\nGti4+yYoCq80uULaicsAxFy7zueff45WqyXHKYfBPw5mXLtx9AnoY+aIhaLExsayb98+0tPT8fT0\nRK/Xi0fABEEodcW55z0ISPjn/xUgRZZlUa6qlITtSCZLrzDRLYz0iZ8g/7M84Z9iYW+88Qa92/em\nUcNGXE+7brY4haIZDAYiIiKIjIxEo9HQqlUrmjdvjlb7MJ1bgiAIRStO8l4ly7KYvrQMGI0K321N\ngoQ9nDr9H5q7Qc1mrdFYWmJlNBH+7jDavjAcAD9XP/xc/cwcsXAv165dIzIyEhcXFzp37oynp6e5\nQxIE4RFWnOR9XJKkl4EIIPPOQlmWY8osqirit4jbxMbnINWuTprei132Nsz+PkIdjLbx9EY+Dv+Y\nt9u9jYO1g5mjFYpSp04dunTpgq+vr6i5LQhCmSvOt0w7couS3K1eKcdS5dwZqDbk1W6M6BuDKSuT\nvdF7uXjzIsNbDqdng56cSTjD8evH6Vino5mjFe6nQQPx7L0gCOXjnslbkqRXZFn+XpZl33KM55GV\nlmHk4Ml0jMbcSViSbxsIP5GOg52WQd1dAdDa2pGSlUJcWhwAjtaOTO001WwxCwUpikJCQgJeXl7m\nDkUQhCqsqCvvcYCY/ayUTFl8jfV/FCyBPqi7K6nG6yQkZ1HfrT4hviG08RHzXFdEqamp7N27lxs3\nbtCvXz/c3d3NHZIgCFWUuDlXTo6cSQegU0tH7Gy1KCYTN68e5K3nByHfPMSc/XMYGTSSAY0H4Grn\nauZohbwURUGWZf78809ycnLw9fXF3t7e3GEJglCFFZW8G9/1jHdeiizLYuhzMWXqTcReS0cyyiwe\n4EtCciIvvz2ZiFNnUOI3sGXLFtq+2Ba9seD0qIJ5ZWRksH//fmJiYrCysiIkJAR/f38xw50gCGZV\nVPI+DzxF/rnNhYcQFZPFqNuzeVL/ExdfgytZEBGZu+4Jb2cAbCxtsLEUU2lWNEajkbi4OHx8fOjU\nqROOjo7mDkkQBKHI5J0ty/LlcovkESZf1mPSWABgU8efuvbV+LZmKtYujpzp7c6J6yd4rMZjZo5S\nKIyTkxN9+/bFxcVFXG0LglBhFJW8w4tYJzyA7Tt282NyQ/QjDjJzfHsAWgBZhiy2nN3ChtMbRPKu\nwFxdxRgEQRAqlnvO3SjL8tjyDORRFpuokHPrJLcTjuRbbmtpy/NNnmd2l9lmiky4w2AwEBkZiaIo\n999YEATBzMRo83KQZAjEq8sQ2vSvpi776+pfNPZsjKO1uIdqbvHx8ezZs4dbt25hbW2Nv7+/uUMS\nBEEokqiaUMZSUg3EJebgLO3ktnKZbGNuxZEj144w8beJ3Ei7YeYIqy6j0cjhw4fZsmULt27domnT\nptSrJyYOFASh4hNX3mVMvpz7+FfNG+/SomYKltrcJn/K/yna+LTBy0HM1GUO6enp7Ny5k8TERBwd\nHQkJCcHb29vcYQmCIBSLSN5lSB97gcw5o2mv70XtugNpW0tS1/m6+OLr4mu+4Ko4Ozs7ACRJon37\n9lhbW5s5IkEQhOITybsMRW9azjtyBM5e1+lad4i5wxHy0Gq1PP3001hZWZk7FEEQhAcm7nmXEUVR\n2P7Dd5xJV7jRyoIM5wPmDkm4i0jcgiBUViJ5l5FM+Rj7L1wj7VwaMes68kLLPuYOqUrKyMhg7969\nZGVlmTsUQRCEUiO6zctIzML/Y6Q3mFza8rdfb3w8xNSn5e3ixYscOHAAvV5PtWrVaNGihblDEgRB\nKBUieZcig8lAYkYiNRxrENugLl81SCb5dE+aegSIqTXLkV6vJzw8nPPnz2NhYUGHDh0IDAw0d1iC\nIAilRnSbl6LEjESGbxlOTEwM7Xq/RUe717hhXZNG9ezMHVqVkZ2dzcaNGzl//jyenp7079+fJk2a\niJMnQRAeKeLKuxQY026jsbJGq9HibutO8+bNSU5OJv63mmReuYn0jOgyLy/W1tbUr18fKysrWrRo\ngVYrzk8FQXj0iG+2UnD5hwUc6eSI6cfv+aLrF9SuXRuAs5dyB0k19LU1Z3hVTrt27QgKChKJWxCE\nR5b4disFf9w4xLjQpvxkcxUXFxdOnDiByaQgx+Qmb0kk7zIhiogIglBVieRdCgL/jGL2ypM83XSA\nuizmejZZeoUa7pa4Oom7E6UtOTmZn376ievXr5s7FEEQhHInkncJGdNTWXg4mh6HTfwRcU5dfvby\nP1fddcVVd2lSFIWTJ0+yefNmEhISiI2NNXdIgiAI5U5cEpbQzbNHOGdlhUHJoW49P3W5er+7nkje\npSU1NZW9e/cSFxeHnZ0dwcHB+Pr6mjssQRCEcieSdwlFxZ/BemJj6kelsvaAOz/prgBw5HQ6AA3F\nlXepMJlMbNu2jdTUVHx9fQkODlaLiwiCIFQ1InmXUPter3Okx2iaP/8HP19SgJv51jcPsDdPYI8Y\nrVZLhw4d0Ov1+Pv7i+e2BUGo0kTyLgXJt00kptbA0V7LzNH/1oSu5WUtRpqXorp165o7BEEQhApB\nJO8SUBSFv+P/JvmaDwBSHVteeNLNzFFVftnZ2VhaWorntAVBEO5BJO8SSM1OZX74fGJvJgMf0aCO\nmEmtpGJjY9m3bx+BgYE0b97c3OEIgiBUSCJ5l4CTtROjvEax+ZIbMgYC6ogu8odlMBiIiIggMjJS\n3M8WBEG4D5G8S+CG7n+MH/sGkZeScG73KwHiyvuhxMfHs2fPHm7duoWrqyshISF4enqaOyxBEIQK\nSyTvEgj/cQYz3M6xodYcdhg0+Isr74dy+PBhbt26RdOmTWndujWWluLPUhAEoSjiW7IE9llc40yf\nAJJ3+2Bro6GWl5W5Q6qUnnjiCVJTU/H29r7/xoIgCIJI3g8r52YCr+w4g2LryLMODQmsa4NWK+7V\nPgwnJyecnJzMHYYgCEKlIZ7FeUiZ504AkOXVGEWjFYPViuH27dukp6ebOwxBEIRKT1x5P6TfT/zM\nQWcn6lrkdvX6i8Fq96QoCrIs8+eff1K9enV69eolRpQLgiCUgLjyfgC39bfZeHojALqMZLYF1eQ/\nF08CiCvve8jIyGDnzp3s378fgAYNGpg5IkEQhMpPJO8HcFt/m1UnVgHgqQRxbsE50lJbAYgJWgpx\n6dIlNmzYQExMDN7e3gwcOJCAgABx1S0IglBCotu8GHISr5OwcTFK0yD6N+4PQM2aNencpStHE5ph\nZ6nBt6ZI3ndLSkrCYDDQoUMHAgMDRdIWBEEoJSJ5F8NP/1vKr/u/xnWLK59uPwPACy+8QOOgZ+j5\n5nn8fKyxshSJ6W4tW7bE39+fatWqmTsUQRCER0qZJW9JkrTAYqAZoAdGyLJ8Ic/6F4H/AwzA38Dr\nsiwrZRVPSdhdjiHe1ooU2/z1o6Ni9AD41xb3uwuj1WpF4hZKzZUrVxg5ciQtWrTg448/LvZ+cXFx\nnD17ls6dO5dhdMX30Ucf4efnxwsvvADA+vXrWbduHZaWlowZM4aQkBCysrKYNGkSycnJODg4MHfu\nXNzc3Dh+/DgfffQRFhYWdOzYkbFjxxY4fnR0NJs2bWL8+PE0adKEFi1aALlTEJtMJhYuXEitWrXI\nyclh6dKl/Pnnn2i1WjIzM5k2bRrNmjUDcttt7ty5JCcno9frCQwM5P3338fKKv98Fnq9ni5dujBs\n2DCGDx8O5NYomDBhAuvWrVO3CwsLIykpibFjx3Lr1i0++eQTrly5Qk5ODt7e3sycORNHR8cHasvk\n5GQmTpyIXq/Hy8uLjz/+GFvbf7+PDxw4wNdff62+1ul0bNu2DTc3N3r06EFAQAAATz75JKGhocyc\nOZM33ngDd3f3B4rDHMrynndfwFqW5Q7AZGDhnRWSJNkBs4AQWZYfB6oBfcowlhLxjzzPjBOxvND+\n6XzLz8Vk5a6vW7W7zOPj47l27Zq5wxAecTqdjs6dOz9Q4gb4888/OXr0aBlFVXzJycmMGDGCPXv2\nqLeQEhISWLVqFWvXrmXZsmUsXLiQ7OxswsLCkCSJNWvW0LdvX/773/8CMH36dBYuXEhYWBgnT57k\nzJkzBd5n3rx5DBs2DAAXFxdWrVrFqlWrCAsL47nnnmPFihUAfPHFF2RlZbFmzRpWrVrFyJEj+eCD\nD4iNjcVoNPL6668zfPhwVq1axfr167G0tOSLL74o8H47d+6kd+/ebN68GUW59/VX3ttm48ePp0uX\nLupnb9asGdOmTXvgNl28eDHPPPMMa9asoVGjRqxduzbf+uDgYPXzh4SEMHLkSPz8/Dh9+jRPP/20\nui40NBSA0NBQFi5cWNhbVThl2W3eEdgBIMtyhCRJrfKsywLay7KclSeOzDKM5aEpBgMZp4+g0cBz\nI/Of5Z6r4lfeRqORq1evcvToUezt7Xn++efF1KYVTOjUS+w6nFqqx+za2olVs+rdc/2mTZvYs2cP\ner2ehIQEXn75ZXbt2kVUVBTvvPMOXbt2ZfXq1fz+++9kZmbi6urKl19+yfr16zl69CgLFy7k3Xff\n5bHHHmPw4MEAXLt2jaVLl5KVlUWdOnVo2bIlc+bMQVEUXF1dGThwIEajkWnTpnH9+nUSEhLo0qUL\nb731Fl9//TXZ2dm0aNGCFStWMGPGDPz8/NQrwX79+jF69GhcXFzo1KkTwcHB+Y790UcfkZ2dzbhx\n41AUhezsbGbMmEHDhg3Vz7xz505Wr16drx3eeecdmjZtqr7OyMjgrbfeYv/+/WqSO3nyJC1btsTK\nygorKyvq1q2LLMscPXqUkSNHArkJaPHixaSlpZGTk0Pt2rUBePzxxzl48CCNGjVS3+PixYuYTCZc\nXFwK/d1cvXpV7Q3bunUru3fvVtd5eHgwePBgNm/eTLt27ahZs6Z6FQ4wadIkTCZTgWNu3LiRDz74\ngKSkJPbt20dISMg9/zbu/C6TkpLo1q2buuzll18mIyMj33Y6nY7//Oc/+ZYNGzYsXw/K0aNHGTNm\nDJA7U+Nnn33Gq6++WuA9r1+/zpYtW/jxxx8BOHXqFKdOnSI0NBQ3NzemTJmCp6cn9erV4+LFi6Sk\npNyzDSuKsvymdQZu53ltlCRJK8uy6Z/u8QQASZLeBBxkWf6jDGN5aL/9tYadrTxof8uGINf8xTKi\n/rnyrooFSZKTk9mzZw9JSUk4OjoSEhIiEregysjIYNmyZfz666989913rF+/noiICFauXEmXLl1I\nSUnhu+++Q6PRMHz4cE6dOsWQIUM4ePAgkydPxmAwqIkbwNvbm9dee41Lly7x4osvMmjQID7++GPq\n16/Phg0b2LZtG35+fjRv3pyBAwei1+vp1KkT48aNY9SoUVy6dIkuXbqwYsUK9Qow75VgYmIimzdv\nxtLSMt+xN27cyDfffEPLli1xdXVl3rx5nD9/vkCi6dGjBz169CiyTWrVqkWtWrXUxyYB0tPT880u\n6ODgQFpaGmlpaTg4OKjLUlNTSU9Pz9et7ODgwJUrV/K9x+HDh5EkSX2dkpJCaGgo6enppKSk0KNH\nD8aMGUNSUhLVqlVDq83f+Vq7dm1OnjxJfHy8epJwh7W1dYHPFB0dTWZmJpIk0b9/f1asWHHf5B0f\nH0+tWrXyLdNqtQW6zIOCgli1alWRx0pLS1Pb7047FWbFihUMHTpU7fKvX78+TZs2pX379mzdupVZ\ns2apvQr16tXj6NGjdOnSpcj3Nrey/La9DeSd81Iry7J62vbPPfF5QAOgf3EOqNPpSjXA4hzTmGpN\ndb8OZJgc8m2bY4Doa5ZoNXAr/jS6m6UeWoWVmJhITEwMiqLg7u5O7dq1iYuLIy4uztyhVVpl8bcN\nMK5v7k/pSkanS77n2suXL+Pq6opOp+P69etUq1YNnU5HbGwsCQkJHD16lPj4eF599VVsbW2Jjo4m\nMjISk8lEcHAwH374IXPmzCnQJtHR0cTFxaHT6Th37hwTJ04Ecu/l1qxZk0uXLrFnzx62b9+OnZ0d\nWVlZ6HQ6Ll26pO6XlpZGZGQkycnJREdHc/v2bf7++29cXV05cSJ31sSoqKgCxw4ODsbT05MhQ4Zg\naWlJ375983URR0RE8Pvvv+eL98UXX6R+/foF2ufatWtkZGSo7RMdHa1+1qtXrxIbG0tOTg46nY6U\nlBQyMjKwsrIiKiqKxMREddszZ85gMpnytVNkZCS2trbqMgcHB8aNG4fJZGLJkiUkJiZy+vRpDAYD\nCQkJHD58OF8Cv3Nicfv2bU6fPp3v2KmpqURFRdGyZUt1WVhYGMnJyQwaNAhFUYiKiuLXX3/F1taW\nlJSUfPvLsgzkfn9cuHAh37o7JYE7duyYb/sNGzbka7tevXoRFBSkvtZoNISHh+Ps7Mzly5cLtAeA\nyWRi586ddO7cWV1na2uLpaUlOp0Od3d3jh8/rq4zGo0cP378ocfrlNW/5buVZfIOB54GNkiS1A44\nedf6peR2n/cr7kC1vL+00qDT6e57zCtXvNi0YRc+ISH5tj0bnYVJOUc9b2vat3usVOOq6K5evUp8\nfDzBwcEkJSWV+u+lqinO32FlEhMTg8lkIigoiPT0dKKioggKCsLe3h5nZ2ccHR05e/Ys69evJzMz\nk/79+xMQEEDTpk2ZN28es2bNYv369axevTrf4Kg7X85BQUH4+/uzaNEiatSoweHDh9Uk7efnx8SJ\nE7l8+bL6RX/lyhWMRiNBQUG4u7vj6elJUFAQmzdvxtvbm6ZNm+Lk5KT+Dho0aJDv2CkpKRgMBlq2\nbMmUKVM4duwYn332GStXrlRjCwoK4vXXXy9W+/z55594eHgQFBREnTp12Lp1K02bNkWv15OUlMSz\nzz5LWloa8fHxDBo0iF9++YXg4GAef/xxnJ2d8fLyolatWixdupSxY8fm69q+ePEi8fHx6mexsrJS\n/79Zs2Y8++yz9OnTh06dOvHss8+yb98+JkyYgEaj4ddffyU8PJylS5fi4+PDhg0bsLKyolmzZiiK\nwuzZs7Gzs1OPl5OTw6RJk/jpp59wdnYGYMmSJZw8eZLJkyczd+5cXFxcqF+/PkajkcWLFzNq1Cja\ntGnDunXrSElJoWvXrgAsW7aMS5cu8dZbb+Vr07y9L4UJDg4mJSVFTczdunUr8G/p7NmzNGrUiHbt\n2qnLxo8fT/fu3enVqxe7d+8mKChI3W/r1q35Xj+Isvi3fK+TgbJM3puB7pIkhf/zeug/I8wdgSPA\nMGA/sPufbp7PZVn+qQzjeSiHDh1i2bJlxMXF8dJLL6nL1cFqVXBmNR8fH1588UUsLS1JSkoydzhC\nBVRY1/Sd13Xr1sXOzo4hQ4bg6upK48aNuXHjBgsXLqRz584MHDhQfT158uR8+9453ocffsikSZMw\nGo1oNBoGDx6MJElMmDCByMhIvL29adKkCfHx8UiSxJIlSwgMDCQ0NJQZM2ZQs2ZNqlevXiDewo79\n0UcfUa1aNcaPH09YWBhGo7HQUd4P0z6enp68/PLLDB48GJPJxPjx47G2tubFF1/k3XffZfDgwVhb\nW6uDqGbMmMHEiRMxGo08/vjj+RI3QNu2bZkzZ06h72ljY8Ps2bOZPHkybdu2ZeLEiSxatIhBGi8y\n6wAAIABJREFUgwZhZWVFVlYWc+bMUbu0P//8c2bNmkVmZiYZGRm0aNGCcePGqcfbs2cPTZo0URM3\nQL9+/ejXrx9vv/02H3/8Me+//z4ajQaDwUC3bt1o06YNkDuobubMmSxfvpycnBzq1KnDrFmzHrgd\nx4wZw7vvvsv69etxc3NT22n+/Pn06NGDZs2aER0dTZ06dfLtN3HiRN577z3CwsKwt7dn9uzZ6roz\nZ84wadKkB46l3CmKUil+jhw5opS2+x3zeNxx5cnPnlTcO7or06ZNy7du/so4pWaPE8rsZddKPa7K\npCx+L1WNaMOSE234r1GjRimJiYkPvJ9oQ0WJiopSpkyZ8tD7l2GeKpATxfSoRfBz9eP5Zs/z0oCX\nePrp/I+JRV3JHWn+KM9pfvHiRY4dO2buMARBeACTJk1SHwcTHszq1av5v//7P3OHUSxieHARnGyc\nGNZlGHQB3Zl0wnb+O0jnuJw72tS/9qM30lyv1xMeHs758+extLSkYcOG2NnZ3X9HQRDMrn79+uqA\nO+HBfPjhh+YOodhE8i5CzCdjMdyMx3rg+zw33YIcQ/5xdVotNHjEkndsbCz79u0jPT0dT09POnfu\nLBK3IAhCBSOS9z3cSLvBG5b/o1F6Em1OZ5FjsMfPx5rWgQ7qNu2bOuBob2HGKEvX2bNn2b9/PxqN\nhqCgIFq0aFHgOVBBEATB/ETyvgenND3DfzvHLXcXtl/0AtJ47TlPXu5d8ee8fVh169alZs2atGvX\nDk9Pz/vvIAiCIJiFSN73oD99hK+PZKJ1c+ZUYhRoa9KltdP9d6zE7OzsCgzMEwRBECoekbzvIe3U\nIYbVgBMNWnAkzYHGfrbU8io4PWBlpSiKqK8tVCqVvarY+fPnmTp1KgC+vr7Mnj0bCwsLUVXMDFXF\natSowYcffsjVq1cxGAxMnTqVJk2aiKpilVVEbATX065jNBkZarGHzQP8qd5kAFor50fmqttkMnHy\n5El++eWXQosMCEJFVdmrin322WdMmDCBsLAwIHeSE1FVzDxVxZYtW6a278yZM4mKigJEVbFK64e/\nf2Bg4EBqONZg/Qub+Pv4dmavzp36tFubyp+8U1NT2bt3L3FxcdjZ2XH79u0KXzlHKDldq8J7WIKO\nFP5F+6Db301UFct1d1WxRYsWodVqyc7OJiEhAScnJ1FV7B/lXVUsPDycXr16MXz4cBwdHZk+fTqA\nqCpW2SgGA1hY0NqnNV4OXgC4udembqNXiIyTcXbQEtTI4T5HqbgURUGWZf78809ycnLw9fUlODhY\nPAImlBlRVawgrVbLtWvXePXVV3F2dkaSJPbv3y+qilH+VcVu3rxJamoqy5Yt46effuKTTz7hk08+\nAURVsUrl+q/fc23uG3TqP5raE3LnL1cURa2D3CnICSvLynt/ODo6mv3792NlZUVISAj+/v7ifncV\nUtwr5ofd/m4ajUa9GnR0dFQrazk7O6PX69FoNFhZWTF+/Hjs7e25ceMGBoMBgJEjR/LCCy+wefPm\nQo99p1v2woUL6oQaBoMBJycnqlWrxt9//01ERASOjo5kZ2er+xTWnZv3KrJWrVpqSduLFy/mO7av\nry9PPPEE0dHRvP766+q96bx27NjBmjVr8i2bNGlSgbnHvb29+e2339iwYQNz587lySefJD09XV1/\np0Soo6Ojujw9PR1nZ2ccHBzybZuWlpZvXnHITdYeHh7q6zvd5iaTicmTJ2NpaYmdnR0WFhbcunUL\no9GIhcW/j7tGR0fj7e2Nt7c3O3fuzHfsmzdvcvz48XxXvhs2bCAzM5MRI0agKArHjh0jJiYGe3t7\ntf3zfjY7Oztq1qzJ9evX863Lyclhx44d+QbMFnblPXTo0HxJ1dHRkbS0NNzc3NR2upvJZGLv3r1M\nmDAhX7vcOU7nzp355ptv1HVeXl6kpKQUOE5FI5I3sCJ+F1tfasxIU0JutZT9+xk8eDBGa19o8Dld\nK/n9bl9fXx577DECAwMfeECIIDyMok4OZVlm165d+aqK3emO/vjjj5k1axYffvhhgapieROwn58f\n8+fPz1dVbPPmzTg7OzNz5kwuX77M+vXrgdyrujuJ2sbGhvj4eOrVq8fp06epUaOGus0d9erVy3fs\nlJQUIiIi8PT0ZNmyZRw7doxPP/00X1Wxnj170rNnzyLbZPTo0bz33nvUrVsXBwcHtFotzZo147PP\nPiM7Oxu9Xs+FCxcICAigZcuW7N+/n2bNmrF//35atWqFo6MjVlZWXLlyhVq1ahEeHl5gwJqbmxvx\n8fEF3lur1TJr1iyeffZZWrVqRadOnejVq5d6H16j0XDjxg3CwsLUqmKxsbGcPHlSrSr25ZdfYmdn\npybvnJwctm/fXqCq2A8//MDkyZNJT0/nwoULalWxgwcPMmrUKKpXr46rqyu7du1Sq4qtXLmSU6dO\n5UvexbnybtmyJfv27aNfv35qO93t3Llz+Pn55es5aNmyJXv37qVx48YcPnwYf39/dd2tW7cqxYA1\nkbyBl256EPLjCXzeeAGABQsWcPXqVSztNXgGWNGldcGzucpEo9HQtm1bc4chVCGiqlhBo0aNYvLk\nyVhZWamVrDw8PERVMTNUFRs9ejRTpkzhhRdewMrKSu0yB1FVrFJVFTv/zgDlSBBK0vYfFEVRlHnz\n5imdnxyguLb8Sun15rlSf9+ylJmZWa7vJyoRlZxow5ITbfgvUVXs4YmqYpXMrcSrAFh51ARy71W1\n6fMZtp6P07VN5bjqNhgMhIeHs27dunz3xQRBqFpEVbGHJ6qKVTJjmmSQ/Vhzuq7dRS9TSwB+j8gd\nrFYZ7nfHx8ezZ88ebt26hYuLC9nZ2epIVUEQqhZRVezhiapilcySXgfo/+ZufjC58cOpaHW5h4sl\nzfwr7uNURqORo0ePcvz4cRRFoWnTprRu3VodNSsIgiA8msS3PHAmOocsi3p417CioW/u1HoaYFB3\nV7TaivtI1Z1HNxwcHAgJCcHb29vcIQmCIAjloMon7xxjDlfi0wAY1M2Vd16pYeaIis/Dw4Pu3bvj\n7e1d6AQKgiAIwqOpyg9Yi7gawcbMQbi0XoqPl9X9d6hgfH19ReIWBEGoYqr8lXc773bUjlrOodMZ\n+DxbMZOgoijcuHFDnVBCEKqiyl5V7PTp04wePZq6desCMHjwYHr16iWqipmhqpibmxs9evQgICAA\ngCeffJLQ0NBKVVXM7M9vF/enrJ6f+9/e3QqgWLu1Vc5dLt9npIsjPT1d2b59u7J06VLl8uXL5g6n\nAPFsaMmJNiyezZs3K3Pnzi10XVFt+OOPPyoLFiwoq7CKbf369cry5cvzLYuPj1f69OmjZGdnK6mp\nqUqfPn0UvV6vLF++XFm0aJGiKIryyy+/KLNnz1YURVGeeeYZJSYmRlEURRk5cqRy+vTpAu8zZswY\n5ebNm4qiKErHjh3zrVu7dq0yc+ZMRVEUZcGCBcr8+fPVdTt27FD69OmjXLlyRTEYDErfvn2VEydO\nqOtnz55daDtu2bJFmTNnjtK7d2/FZDIpiqIoV65cUQYNGpRvu7CwMPUzDRs2TPn999/VdStWrFDe\nfvvte7bdvcyaNUvZvHmzoiiKsnTpUmXFihX33Pbbb79VPv30U0VRFCU8PFyZNWtWgW0uXryovPfe\new8cxx3l+Zx3lb/yPrb6U9CCtyYRb8+K1W1+8eJFDhw4gF6vx8fHBzc3N3OHJFRCeWcPUwqZ41uj\n0dxzeVH73YuoKpbr7qpikZGRXLp0iV27dlG3bl3ef/99UVXsH+VdVezUqVOcOnWK0NBQ3NzcmDJl\nCp6enqKqWGXyR/XrtPyyJdnrLXCws7j/DuUgOzub8PBwoqKisLCwoEOHDgQGBopiIkKlIaqKFdSs\nWTMGDRpE48aNWbJkCV9++SWNGjUSVcUo/6pi9evXp2nTprRv356tW7cya9YstVa5qCpWSUy74oh+\nx16W1V9k7lBUiqJw7do1PD096dy5c4U/AxQqtvtdNd9r/YNcbeclqorluruqWPfu3dVE0717d2bN\nmkXr1q1FVTHKv6pYu3bt1JLI3bp1UxM3iKpilUZO0nWsAWdvP3OHorKxsaFPnz44OTkVODMWhMpA\nVBUraMSIEXzwwQc0a9aMgwcP0qRJE1FVzExVxaZOnUr37t3p1asXf/75J02aNFHXiapilYCiKKRm\nJVANqObjY+5w8rlzX0oQKiNRVaygGTNmMGPGDCwtLfHy8mLmzJk4ODiIqmJmqCo2ceJE3nvvPcLC\nwtQKb3dUlqpimoftGitvOp1OCQoKKtVj7j+4h6lHxqM1mRjgvZM3BpXvo1hGo5EzZ87QqFGjfF1X\nlYlOp6O0fy9VjWjDkhNt+K/Ro0czZ86cB756FG0I58+f5/vvv3+oEwkomzb855gFurKqdJ+svbUT\ndeKWEbX9a3y8bO+/QylKTk5my5YtHDx4kBMnTpTrewuC8OgSVcUenqgqVklkZWWx7rM+GIwK3p9e\nLZf3NJlMnDp1isOHD2M0GgkICMh3v0UQBKEkRFWxhyeqilUW1tB44G/cuKHBx6vsZ1fLzs5m586d\nxMXFYWdnR3BwML6+vmX+voIgCMKjpUon74jEI2jb/ojL5U5Udys4SrG0WVlZYW1tja+vL8HBweqj\nCoIgCILwIKp08pYsg7m+rQveXhZYWpT9BCgajYauXbtiYWEhJlwRBEEQHlqVHrB281oydqY0anmU\nX0ESS0tLkbgFQRCEEqnSV976rR+wJEXHEePXgH/pHVev59ChQzRv3lw8ry0IpaSyVxUrTN7qW126\ndGHHjh0FpiFdunQpHTt2vOfA1sTERPXZb2dnZxYuXIitrS3fffcdGzduxNXVFYBZs2bh6+tLv379\n1KlIa9euzUcffcTly5f58MMPcXZ2xt/fn+nTpwPw3nvvMWPGDGxsbAq875w5cxgxYgTr169n27Zt\neHl5AbmzvPXu3ZvRo0cDcPbsWRYuXEhWVhY5OTm0bduWsWPHqhPwrFu3jq1bt6rPg7/99tvq8+B5\nffPNN3z//ffs3r1bbaPJkyfTu3dvgoOD1e06duxIeHg4AH/88QcrV65EURT0ej3Dhw+/7zS2hSms\n8lte48ePJyEhgbS0NG7dukWLFi1YuHAhs2fP5ujRozg4OKDRaFi8eDHXrl3jt99+e6i5AvIprFpJ\nRfwpi2otrSY2VJp/0VwZPOKtUjvmlStXlNWrVytLly5VDhw4UGrHrahERaySE21YPJW9qlhh8lbf\n6ty5s6LX6/Otv3btmjJ+/PgijzF//nzlhx9+UBRFUT799FNl1apViqIoysSJE5XIyMh822ZlZSl9\n+/YtcIxRo0ap+02bNk2t+HXgwAG1Elhex44dU38XixYtUtauXauu0+v1Srdu3ZSkpCQlISFBeeqp\np5To6Gh1/ZdffqnMmDFDURRF2bZtm/LWW28pBoNBbY9OnTqpVdHy6tOnj/Lxxx8rmzZtUpdNnjy5\nwPfsnUpqOp1Oeemll5SMjAxFURTl5s2bypNPPqmcP3++wLGLcq/Kb4XZt2+f8uyzzyoJCQmKoijK\niy++WOhnmTRpklod7n5EVbFCpC85z5k0A89MG3f/je/DYDAQERFBZGQkGo2GVq1a0bx581KIUhBK\nptXXrTjy2pFSe30/oqpYrruriv3111989dVXmEwmMjIyWLhwoTqf+r2EhYWp066uWbOGHTt2qOs0\nGg2ffPIJtra26lzcaWlpeHt7A7lVzJYsWUJiYiIhISG89tprnD17lszMTIYPH47BYGD8+PE89thj\nnD59Wq1i9sQTTxAeHk63bt1o3749c+fOLXCVuGrVKrV2N+SfuvbmzZsYjUZsbGxYu3YtAwYMUGuY\nA7zxxht07doVvV7PunXreP/999VJqmrVqsWWLVsK9FhGRETg6+vL888/z6RJk+jXr1+h753Xhg0b\neOWVV9SBwS4uLmzcuDFfIRiAKVOmcPnyZfW1q6trvrnO71X5Le/v9o6NGzcSGhqKh4cHJpOJy5cv\nM3XqVBITExkwYAD9+/cHoFevXqxZsybfDIIPqsomb8VoxEoxYK0B/7umGHxQJpOJLVu2kJSUhKur\nK507d85XHEAQqhpRVayg8+fPM3/+fLy8vFi6dGmBQhyF+euvvxgwYAAAQ4YMYciQIQW2ee6553j+\n+ef55ZdfyMnJ4c033wSgd+/eDBkyBAcHB8aOHcvevXvx9vZm+PDhDBw4kOjoaEaOHMmOHTvyJUB7\ne3u1OpeFhQVubm7IspyvWtnhw4f55JNPgNzkuWLFCn755Rfi4uKoXr06s2fPxsHBgdjYWDp27Fgg\nZk9PTxITEwutMFbYrcYNGzYwYMAA6tWrh7W1tTrnOtx7Hv3CKqPdnbiBfFOjFuZOsZg77lR+u1tS\nUhKRkZF8/vnnAGRmZhIaGsrQoUMxGAy8/PLLNGnSBEmSCAgIYNGikhXDqrLJOz3xKvNaW6NkOFC7\nuXT/HYqg1WoJCAggLS2N1q1b3/dsWhDK091XzSV9fT+iqliuu6uKeXl5qUntxo0btGzZ8r5tefPm\nTXWa09WrVxeo9DVv3jymTJnC3Llz6dixI/v27ePdd99l6dKlvPLKK+q97U6dOnH69Gk6dOigXgX7\n+vri4uJCQkJCvsIsd1fn8vT0LFBly2Qyqe2l0WgYNmwYzz//PJGRkYwfP159j+rVqxMbG5tvX6PR\nSHx8PO7u7vj4+BAXF4e//79jjg4cOEDDhg3x9PQEcguFHDhwgJs3b7Jq1SpSU1NZvXo18+bNw8bG\nBr1eX+D4gHrsvCcdOp0OT0/PfHOdf/DBB8TExKivXVxc8iXWvBXeCmufO3bs2EHHjh3Vkwk7OztC\nQ0OxsbHBxsaGdu3acfbsWSRJKrRNH1SVzTKX488x7+lGOKZrWFm94GCMB1VYF4ogVFWiqlhB06ZN\n448//sDe3p7JkycXq+Squ7s7t2/fxsHBgZdeeomXXnqpwDYZGRlqkvb09OT27dukpaXx9NNP88sv\nv2BnZ8ehQ4cYMGAAmzZtQpZlpk+fzo0bN0hPT8fT05NGjRpx5swZgoKC2L9/P+3bt1ePf/v27QI9\niTY2NiiKov6e73yWwMBARo4cyfjx41m7di39+vVj6NChdO7cmbp166rVyTp16oStrS39+/dn8eLF\nLFiwAAsLCy5dusTUqVPZtGmT+l4///wzAwYMUIuFZGVl0bVrV5KTkwkMDOT3339Xq5MdOXKEBg0a\nALk9EgsXLqRt27bY2dmRlJTE+++/X+CK916FXO4orPJb3pONOw4dOpRvINulS5cYP348mzdvxmg0\notPpeO6559Q2dXNzK/J976fKJm9v3xDO/LYSRzsNzg7Ff2Iu7x+sIAiFE1XFCnrmmWcYMmQIXl5e\n+Pn5kZCQUGgb5dWmTRtOnDhBzZo177nN9OnTmT17NlqtFkVRmD59Oo6OjkyYMIGXX34Za2trOnTo\nwBNPPIHBYOC9995Tu98/+ugjLCwsmDx5MuPGjWPbtm3Ur19fPRExmUzcuHFD7T25o2XLlpw6dUq9\naMn7GQYMGMD27dsJCwtj8ODBzJs3jxkzZpCZmYnBYKBt27a8//77ADz11FPEx8czePBgrKysMBqN\nLFiwIF9i27hxI/Pnz1df29ra8uSTT7Jx40aGDRvGmTNn6Nu3Lw4ODlhbW6tFRZo3b86gQYMYOnSo\nWjVt4sSJBAQE3P+XlYeHh0ehld/Onz/PmjVr1JH5ly5dYtCgQep+9evXp2/fvjz//PNYWlry3HPP\nqe144sQJOnTo8EBxFFDYKLaK+FPaI3L/Pp+h1OxxQgl57Wyx97l165by888/K1FRUaUaS2UmRkqX\nnGjDkntU2/Dq1avKW2+V3tMwRSmsDffu3av897//LbD82LFjyuzZs8sjrEqluH+HEyZMUGJjYx/k\nmAVyYpWdpCUi8hRozlHDw3TfbRVF4ezZs/z444/ExcVx9Wr5FDERBKFq8/b2RpIkTp06Ve7vrSgK\n27Zt49VXXy2wrnnz5hiNRm7cuFHucVV2sixTt25dfHx8SnScKtttvkH3JZ5dThB9KRBYc8/tMjIy\n2L9/PzExMVhZWRESElLo/Q5BEISy8Prrr5vlfTUaTb7u6rtNmzatHKN5dEiSlG8Q3cOqssn78Tpz\nWbwhgXdfqV7kdn/88QfXr1/H29ubkJAQdWCIIAiCIJhLlU3eqRfO4mJS8HEvuuuiXbt2xMfHExgY\nKAaqCYIgCBVClU3ezU4MpUfmVWw5DLjfczsvLy91zl5BEARBqAiqZPI2mUxs7GBPimNjltbPfc7T\nYDCgKEq+50oFQRAEoSIqs+QtSZIWWAw0A/TACFmWL+RZ/zQwFTAAy2VZ/rasYrlbWkICg9acwcnG\nHt/XvYiPj2fPnj14e3vnq04jCMKDi4iIYNy4cTRo0ACNRkNaWhq1a9dmwYIFJTo5njlzJj179iy0\n4tTDOHjwoDrvuYWFBbVq1eKDDz4os3Etmzdvplq1anTp0qVMji9ULWV55d0XsJZluYMkSW2Bhf8s\nQ5IkK+BToBWQAYRLkvSzLMvxZRiP6siBCAafAW97Iz8d13H8+HEURaFOnTpiEhZBKCGNRkOHDh1Y\nuHChumzChAns3r37ocox5j1uaTl79iwLFixgyZIl6m2x7777jm+//ZZx40peqKgweYtpCEJJlWXy\n7gjsAJBlOUKSpFZ51jUCzsuyfAtAkqT/AU8AG8swHtXJU8fwbebLsH6hHDt2DEdHR0JCQtRqPILw\nKPnhhx8KXZ638EdJtr/bnUkk7sjOziYhIYFq1aphMpmYOnVqvspf48aNY/LkyVhbW3P16lUSEhKY\nO3cujRs3JiwsjPXr1+Pm5kZmZiY9e/YkJyeH9957j9jYWEwmE6+++irVq1cnNDSUhg0bEhUVhb29\nPa1ateJ///sft2/fZvny5fnmow4LC+P111/PN54l7/PMeWtrL1iwgPr169OvXz8WLlyITqdT37dn\nz56sWbOGLVu2oNVqadKkCVOmTOG3337j22+/xdLSEi8vLz777DO+/PJLPD098fPz4+uvv8ba2por\nV66ota8vX77M5MmTsbKywsfHh9jYWFatWlWsNheqnrKcpMUZuJ3ntfGfrvQ7627lWZcKFCwlU0ai\nDQk8O/QZqnvXQpIkBgwYIBK3IJSiQ4cOERoaSu/evenfvz9PPvkk7dq1Iy4ujubNm7Ns2TI2bNjA\n2rVrgdyr6lq1arFs2TJCQ0NZt24dycnJfP/996xfv55vvvkGjUaDoiisW7cODw8P1q5dy4oVK/j8\n88/VKliPPfYY3333HdnZ2djZ2bF8+XIaNGjAX3/9lS++q1evqsUprly5QmhoKKGhoYWeoNy54t+3\nbx9Xr17lhx9+4Pvvv2fJkiWkpqayefNmpk2bxtq1a6lfvz5Go5FffvmFESNG8MMPPxASEkJaWlq+\nnoO4uDi1DOq33+beMZw3bx5jxoxh5cqVxSpaIlRtZXnlfRvIW39NK8vynenMbt21zgm4eb8D6nS6\nUgmsR8hQvvpJIbDdbRwdHfn7779L5bhVVWn9XqqysmzDe00Ica/3fNDt73bu3DkCAgJ48803SUtL\n46OPPiIzMxOdTkdGRgZ79uxh+/bt2NnZkZWVhU6nIzExEX9/f3Q6HampqVy7do2dO3fi7u7OyZMn\ngdzZxs6dO8ehQ4do0qSJGo+7uzs3btwgLS0NRVHQ6XQYDAa1GERWVhZnz57F1dVVjdHS0pK9e/eq\nSX/cuHHk5OQwceJEdDod2dnZHDt2DEtLS+Li4tBqtZw4cQKdTqd2f6elpfH7778TGhrK559/TkJC\nAv7+/tSvX5/evXvz888/s3jxYry9vfHw8ODatWtkZGRgMBjw8PDg2LFjQG7ZTZ1Ox+nTp9FoNOh0\nOmxsbEhLSyv3f1vi33LJlVcblmXyDgeeBjZIktQOOJln3VnAX5IkVyCd3C7ze0/l84+goKBSC65X\nt9xGLs1jVkWiDUvuUWtDg8HAsWPH1M9Uq1YtXn75Zfr27cv27dvx8/Nj4sSJXL58mV69ehEUFISH\nhwf+/v4EBQWRnp5OVFQUPXr04LvvviMwMBAbGxu++OILAgIC0Gg0XLlyhaCgINLS0oiPj8fLywtH\nR0eaNGlCvXr1cHNzQ5IkgoKC2LlzJ35+fvna2NbWlunTp9OvXz+19OS+ffuwtbUlKCgIJycnfHx8\n8PHxYfHixXTs2FHt9p85cyYGg4ElS5bQo0cP/vOf//DVV19hbW3N8OHD0Wg0nDlzhpkzZ+Lm5sa0\nadNITExUk7ifnx9Hjx5V47GysiIoKIimTZuiKApBQUFcvnwZR0fHcv27eNT+Ds2hLNrwXicDZZm8\nNwPdJUkK/+f1UEmSXgQcZVn+RpKk8cBOcrvul8myHFeGsQiCUE7yVv+C3OpKoaGhzJ49mzfffLNA\n5a8782PfXYnMzc2NMWPGMHjwYKpVq4alpSUajYZBgwYxdepUBg8eTFZWFmPHji20vnJRAgMDeeed\nd5g8eTI5OTlkZmZSo0YNvvjiCwBGjBjBa6+9ho+PDy4uLmg0Grp06cJff/3FkCFDyMjIoHv37jg4\nOBAQEMDgwYNxcHCgRo0aPPbYY6SlpTFq1CgcHBxwcHAgJCSE1atX37PaGuTW/37//fdZvnw5Tk5O\n4rFVoUiavANLKjKdTqeUxRmNONMsGdGGJSfasOQehTbcunUrjz32GHXq1GHDhg0cP378vrWmS9Oj\n0IbmVlZX3kFBQQXO9qrkJC2CIAgVTc2aNXn77bexs7PDwsKiXBO3UPmI5C0IglABtGrVih9//NHc\nYQiVRJWt5y0IgiAIlZVI3oIgCIJQyYjkLQiCIAiVjEjegiAIglDJiOQtCIIgCJWMSN6CIAiCUMmI\n5C0IgiAIlYxI3oIgCIJQyVSq6VHNHYMgCIIglLfCpketNMlbEARBEIRcottcEARBECqQED2YAAAH\nm0lEQVQZkbwFQRCE/2/v3mPsKMs4jn93KZRKuRQJRCsVNPXRgBWwYHeRS6UahWItlxggQe5ELWmg\nUaGCpRJjNYJCI9JSqiJog1aBekHlsiVtClFsCyX4M0VjTNHUFG0VhdJ2/eN9TzN72J3dWjLnzPb3\nSTbZuZx5n30yZ5/zvjNnXqsZF28zM7OacfE2MzOrGRdvMzOzmhn283lHRCdwOzABeAW4TNLzhe1n\nAjcA24DFkha1JNA2NoQcngfMJOXwGeBTkvw1hoLBcljYbyGwSdJ1FYdYC0M4F48HbgY6gA3AhZK2\ntiLWdjWEHE4HZgO9pP+Jd7Qk0DYXEe8D5kma3LS+kpqyJ/S8PwbsI6kbuJb0xgYgIvYGbgE+CJwC\nXBERh7YkyvZWlsNRwE3AqZLeDxwITG1JlO1twBw2RMSVwNGkf5rWv7JzsQNYCFwk6STgEeDIlkTZ\n3gY7Fxv/E08EZkXEgRXH1/Yi4rPAncDIpvWV1ZQ9oXifCDwEIOlJYGJh27uA9ZI2S3oVWAGcXH2I\nba8shy8DXZJezssjgP9WG14tlOWQiOgGTgAWkHqN1r+yPL4D2ARcExE9wEGSVHmE7a/0XAReBQ4C\nRpHORX+YfK31wFm89r1aWU3ZE4r3AcCWwvL2PGzU2La5sO1fpJ6j9TVgDiX1Svo7QERcBewn6eEW\nxNjuBsxhRLwJ+AIwAxfuwZS9nw8BuoH5wBTgtIiYjDUryyGknvhTwDpgmaTivgZI+jFpWLxZZTVl\nTyjeW4D9C8udknbk3zc3bdsf+EdVgdVIWQ6JiM6I+BpwGnB21cHVRFkOzyEVnp8DnwPOj4gLK46v\nLsryuInU65GkbaTeZXOv0kpyGBHjSB8i3wocARwWEedUHmF9VVZT9oTivRI4HSAiJgFPF7b9Hhgf\nEWMiYh/S8Maq6kNse2U5hDTUOxKYXhg+t74GzKGk+ZIm5htf5gHfl3R3a8Jse2Xn4h+B0RHx9rx8\nEqn3aH2V5XBfYDvwSi7oG0lD6DY0ldWUYf9s83wTS+POSoCLgfcCoyXdGRFTSUOWncBdkr7Vmkjb\nV1kOgd/mn8cLL7lV0v2VBtnmBjsPC/t9AghJs6uPsv0N4f3c+ADUAayUdHVrIm1fQ8jh1cD5pPtZ\n1gOX55EMK4iII0gftLvzN24qrSnDvnibmZkNN3vCsLmZmdmw4uJtZmZWMy7eZmZmNePibWZmVjMu\n3mZmZjXj4m1mZlYzw35WMbN2kL8T+gfg2aZNUyVtGOA1NwK9kubuRrsXkSZK+HNeNQpYTpr5bfsu\nHmsu8BtJP42IxxqzKUXEaknH/r8x5mP0AGOBf+dVB5AeunKBpI0lr7sC2CJpye60b1Y3Lt5m1dmw\ni0Xu9XgIQy9wv6RLYOd0kD3Ap4HbduVAkuYUFk8prN+twp31ApdKehx2PkjkR8A1pJmvBtINPPY6\ntG9WKy7eZi0WEUeTCulo4FDgZknzC9tHAN8Gjsqrbpe0KCIOA+4ADgd2ANdJeqSfJnZOdiJpR0Ss\nAsbnY19MKpC9pMkoZgBbgcX9tPcdUqE8Lr92laSuiNgB7A38BThG0saIOJg0t/s40vSIc/M+fyI9\nsevFsjhzLg4BnshtnZvjHJV/LgP2Ac4EJkfEC6THfC4A3jJIPsxqz9e8zarz5ohYXfiZlddfCtwk\n6QTgA8CXml7XDYyRdBxptqzuvP5WYLGkicA0YEFEjC4LICLeCHwYWBkR7wZmAydLmgC8BMwBugZo\nr5c0jD8TQFJX47h5CP4+4Ny86mzgJ8AY4MvAh/LxfgV8pZ/QOoBFEbEmF+JVed+v59GCK4EzJB2T\nX/+ZXJgfBG6Q9Oucj7t2JR9mdeWet1l1XhhgiHkW8JGIuBZ4D7BfXt+YS3kdEBHxEGnmscYw8pS8\n/ot5eQTwNvpONNEBfDQiVuffO4GlkpZExAzgQUmNWY8Wknr48wZobzDfA74BfBM4j/TBYBKp990T\nEQB7kWb/arZz2DwiuoClwC8az9SOiOn57wjSkH1/z9oeSj7MhgUXb7PW+yGpoC0DlgAfL26U9GJE\nHEUafj4d+F1e7gQmS/onQESMBf7adOxe4IHGNe8mHfQdqu4ERpS0V0rSUxFxcEQcD4yV9ERETANW\nSJqWY9yXvlMmNseDpFURcRtwd0RMAN5Amvzmu6Tr9WtJw/vNhpIPs2HBw+ZmrTcFmCNpGXAq7Lyx\njPz7GcA9kn4GzCTdkX048CjpxjNycV1Luh5c1Fygi3pIvdkxefly4NGS9oq2R8Re/RzzXtJ15x/k\n5SeBrogYn5evB746QDzFG/RuIY1AfJJ0fX47afi9h/SBotH2NtK1dBhaPsyGBRdvs+oMdPf4jcCK\niFgJvBN4Djgy798L/BL4T0Q8SyqGSyWtA64CJkXEWlKxvEDSS/202W+7kp4hFcTlEfEc6etZ15e0\nV/QAsCYiRjYd/17SVJP35Db+BlwC3BcRTwPHkm48KyVpK/B50tSKzwNrcl6Wk4bBx+VdHwZmR8RZ\nQ8yH2bDgKUHNzMxqxj1vMzOzmnHxNjMzqxkXbzMzs5px8TYzM6sZF28zM7OacfE2MzOrGRdvMzOz\nmnHxNjMzq5n/AUyA35nDgQLvAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x109c49da0>"
       ]
      }
     ],
     "prompt_number": 23
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Hyperparameter tuning 2 - min_df"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "\n",
      "colors = ['#1947D1', '#CC3300', 'k', ]\n",
      "linestyles = ['-', '--', '-.']\n",
      "params = [1, 0.1, 0.01]\n",
      "labels = ['no cutoff',\n",
      "          'min. df = 0.1', \n",
      "          'min. df = 0.01',\n",
      "          ]\n",
      "\n",
      "for param,col,ls,lab in zip(params, colors, linestyles, labels):\n",
      "\n",
      "    clf = Pipeline([\n",
      "                         ('vect', TfidfVectorizer(\n",
      "                                                  binary=False,\n",
      "                                                  stop_words=stop_words,\n",
      "                                                  tokenizer=porter_tokenizer,\n",
      "                                                  ngram_range=(1,1),\n",
      "                                                  min_df=param,\n",
      "                                                  )\n",
      "                          ),\n",
      "                         ('clf', MultinomialNB()),\n",
      "                        ])\n",
      "    \n",
      "    mean_tpr = 0.0\n",
      "    mean_fpr = np.linspace(0, 1, 100)\n",
      "    all_tpr = []\n",
      "    cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "    for i, (train, test) in enumerate(cv):\n",
      "        probas_ = clf.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "        # Compute ROC curve and area the curve\n",
      "        fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "        mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "        mean_tpr[0] = 0.0\n",
      "        roc_auc = auc(fpr, tpr)\n",
      "\n",
      "    mean_tpr /= len(cv)\n",
      "    mean_tpr[-1] = 1.0\n",
      "    mean_auc = auc(mean_fpr, mean_tpr)\n",
      "    plt.plot(mean_fpr, \n",
      "             mean_tpr, \n",
      "             color=col, \n",
      "             linestyle=ls,\n",
      "             label='%s (ROC AUC = %0.2f)' % (lab, mean_auc), \n",
      "             lw=2\n",
      "    )\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')    \n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.title('Multinomial NB, stop words, porter stemmer, uni-gram, tf-idf')\n",
      "plt.savefig('./images/roc_mindf.eps', dpi=300)\n",
      "plt.show()\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFtCAYAAAAu8eIeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYyI+AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x106c1bcf8>"
       ]
      }
     ],
     "prompt_number": 29
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "Hyperparameter tuning 3 - alpha"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "\n",
      "colors = ['#1947D1', '#CC3300', 'k', '#339933']\n",
      "linestyles = ['-', '--', '-.', ':']\n",
      "params = [0.05, 0.1, 1.0, 2.0]\n",
      "labels = ['alpha = 0.05', \n",
      "          'alpha = 0.1',\n",
      "          'alpha = 1.0',\n",
      "          'alpha = 2.0',\n",
      "          ]\n",
      "\n",
      "for param,col,ls,lab in zip(params, colors, linestyles, labels):\n",
      "\n",
      "    clf = Pipeline([\n",
      "                         ('vect', TfidfVectorizer(\n",
      "                                                  binary=False,\n",
      "                                                  stop_words=stop_words,\n",
      "                                                  tokenizer=porter_tokenizer,\n",
      "                                                  ngram_range=(1,1),\n",
      "                                                  \n",
      "                                                  )\n",
      "                          ),\n",
      "                         ('clf', MultinomialNB(alpha=param)),\n",
      "                        ])\n",
      "    \n",
      "    mean_tpr = 0.0\n",
      "    mean_fpr = np.linspace(0, 1, 100)\n",
      "    all_tpr = []\n",
      "    cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "    for i, (train, test) in enumerate(cv):\n",
      "        probas_ = clf.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "        # Compute ROC curve and area the curve\n",
      "        fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "        mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "        mean_tpr[0] = 0.0\n",
      "        roc_auc = auc(fpr, tpr)\n",
      "\n",
      "    mean_tpr /= len(cv)\n",
      "    mean_tpr[-1] = 1.0\n",
      "    mean_auc = auc(mean_fpr, mean_tpr)\n",
      "    plt.plot(mean_fpr, \n",
      "             mean_tpr, \n",
      "             color=col, \n",
      "             linestyle=ls,\n",
      "             label='%s (ROC AUC = %0.2f)' % (lab, mean_auc), \n",
      "             lw=2\n",
      "    )\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')    \n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.title('Multinomial NB, stop words, porter stemmer, uni-gram, tf-idf')\n",
      "plt.savefig('./images/roc_alpha.eps', dpi=300)\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFtCAYAAAAu8eIeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nf5eKNGPVUuvTAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x105f149b0>"
       ]
      }
     ],
     "prompt_number": 210
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 4,
     "metadata": {},
     "source": [
      "n-gram comparison"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "\n",
      "colors = ['#1947D1', '#CC3300', 'k', ]\n",
      "linestyles = ['-', '--', '-.',]\n",
      "params = [(1,1), (2,2), (3,3),]\n",
      "labels = ['1-gram', \n",
      "          '2-gram',\n",
      "          '3-gram',\n",
      "          ]\n",
      "\n",
      "for param,col,ls,lab in zip(params, colors, linestyles, labels):\n",
      "\n",
      "    clf = Pipeline([\n",
      "                         ('vect', TfidfVectorizer(\n",
      "                                                  binary=False,\n",
      "                                                  stop_words=stop_words,\n",
      "                                                  tokenizer=porter_tokenizer,\n",
      "                                                  ngram_range=param,\n",
      "                                                  \n",
      "                                                  )\n",
      "                          ),\n",
      "                         ('clf', MultinomialNB(alpha=1.0)),\n",
      "                        ])\n",
      "    \n",
      "    mean_tpr = 0.0\n",
      "    mean_fpr = np.linspace(0, 1, 100)\n",
      "    all_tpr = []\n",
      "    cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "    for i, (train, test) in enumerate(cv):\n",
      "        probas_ = clf.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "        # Compute ROC curve and area the curve\n",
      "        fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "        mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "        mean_tpr[0] = 0.0\n",
      "        roc_auc = auc(fpr, tpr)\n",
      "\n",
      "    mean_tpr /= len(cv)\n",
      "    mean_tpr[-1] = 1.0\n",
      "    mean_auc = auc(mean_fpr, mean_tpr)\n",
      "    plt.plot(mean_fpr, \n",
      "             mean_tpr, \n",
      "             color=col, \n",
      "             linestyle=ls,\n",
      "             label='%s (ROC AUC = %0.2f)' % (lab, mean_auc), \n",
      "             lw=2\n",
      "    )\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')    \n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.title('Multinomial NB, stop words, porter stemmer, uni-gram, tf-idf')\n",
      "plt.savefig('./images/roc_ngrams.eps', dpi=300)\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFtCAYAAAAu8eIeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8hZCCCEczP8DOTR91mnwD4oAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x105c19048>"
       ]
      }
     ],
     "prompt_number": 212
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.metrics import roc_curve, auc\n",
      "import seaborn as sns\n",
      "import matplotlib.pyplot as plt\n",
      "%matplotlib inline\n",
      "from sklearn.cross_validation import StratifiedKFold\n",
      "from scipy import interp\n",
      "\n",
      "sns.set()\n",
      "sns.set_style(\"whitegrid\")\n",
      "\n",
      "classifier = Pipeline([\n",
      "               ('vect', TfidfVectorizer(\n",
      "                                           binary=False,\n",
      "                                           stop_words=stop_words,\n",
      "                                           tokenizer=porter_tokenizer,\n",
      "                                           ngram_range=(1,1),\n",
      "                                           )\n",
      "                 ),\n",
      "                 ('clf', MultinomialNB()),\n",
      "              ])\n",
      "\n",
      "\n",
      "cv = StratifiedKFold(y_train, n_folds=10, random_state=123)\n",
      "\n",
      "mean_tpr = 0.0\n",
      "mean_fpr = np.linspace(0, 1, 100)\n",
      "all_tpr = []\n",
      "\n",
      "for i, (train, test) in enumerate(cv):\n",
      "    probas_ = classifier.fit(X_train[train], y_train[train]).predict_proba(X_train[test])\n",
      "    # Compute ROC curve and area the curve\n",
      "    fpr, tpr, thresholds = roc_curve(y_train[test], probas_[:, 1])\n",
      "    mean_tpr += interp(mean_fpr, fpr, tpr)\n",
      "    mean_tpr[0] = 0.0\n",
      "    roc_auc = auc(fpr, tpr)\n",
      "    plt.plot(fpr, tpr, lw=1, label='ROC fold %d (area = %0.2f)' % (i+1, roc_auc))\n",
      "\n",
      "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Random Guessing')\n",
      "\n",
      "mean_tpr /= len(cv)\n",
      "mean_tpr[-1] = 1.0\n",
      "mean_auc = auc(mean_fpr, mean_tpr)\n",
      "plt.plot(mean_fpr, mean_tpr, 'k--',\n",
      "         label='Mean ROC (area = %0.2f)' % mean_auc, lw=2)\n",
      "\n",
      "plt.xlim([-0.05, 1.05])\n",
      "plt.ylim([-0.05, 1.05])\n",
      "plt.xlabel('False Positive Rate')\n",
      "plt.ylabel('True Positive Rate')\n",
      "plt.title('Receiver Operator Characteristic of the Lyrics Classifier')\n",
      "plt.legend(loc=\"lower right\")\n",
      "plt.savefig('./images/roc_final.eps', dpi=300)\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAFtCAYAAAAu8eIeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XlcVFUbwPHfMICgqIiipqm4cSqXVLTUIrfqrcwyNU1L\nW7TFNkuyPTNfbbOy/W21tDLNrTKtNMMlTENyScuTprliqLggOzP3/eMOxDLAADMMA8/38/Ejc++d\ne565MPPMPffc51gMw0AIIYQQvsPP2wEIIYQQomwkeQshhBA+RpK3EEII4WMkeQshhBA+RpK3EEII\n4WMkeQshhBA+xt/bAQjPUUrZge2ADTCA2sBpYLzWOsED7W0G+mitT7t734793wXcBQRgvp5fgSe0\n1gc80Z6T9scBAVrr/1VwP6HAVKAPYMd8LW9qrWc51v8NDNFa/1qhgMse12Rgi9b66zI+bxkQo7Xe\nWcI2K4AbtNbJrmxfhrb/B/wH+Exr/VS+5a2BGVrrYUqpCOA3rXXdCrTTF3hDa92pnM+PAh7VWl9f\n3hjy7csKTABGYn6GBwJLgcla6yyl1MeYr/fliraVr81BwKVa6wlKqS7AIuAEMBtop7We4K62hGsk\neVd/fbXWybkPlFIxwBtAb3c3pLXu6u595lJKvQR0AgZqrQ8ppSzATcDPSqkLtdaHPNV2PhcDv1Vk\nB0qpIGAN8AnQVWttV0q1BFYppQyt9UeYydxS4WjLrj+wo6xP0loPdGGzS3G8Jhe3d9UdQAut9eFC\ny1sByo3tVIjjy3KFE7fD/4D6QH+tdYpSqjbwGfABMAbz78etBTy01ksxvyAAXAP8qLW+3Z1tiLKR\n5F395SUBpZQ/5ofa8XzLngCGYF5C+Ru4W2udqJRqCryD+QFoB97RWr+hlKoPvAZ0xDwDXgVM0lrb\nHGf64Zhv8pe11oscbTwPoLV+VCk1FhjvaO84cK/WWjvOFsKANsBSrfVj+WI8G7gTOFtrfcqxLwP4\nxHFG8xhwr+OMdQlmkg11xPCOYx+DgCcwz1LSgIe01huUUlOAXkBTYCvwEPAe0NixbB8w3LHPQcCl\nSqk04H1gJmbCswEbgQe11mcccWwAOgOPaa2/yvf7GAGc1lq/lLtAa71fKTXccTxz3amU6uqI4xOt\n9ZNKKT9HmxcCdR2/23Fa6/WFjx/wEfAWUAdoBmwBRmitM5VSFwKvY/bEZDle87lAFDBDKZUDLAde\nBC4BrMBm4H5Hssj/+h4HXsX8G/rT0W47zL+ZBMfvbZbjNf2olBoI/ISjZ0EpdRsw0XEMjwE3a60P\n5jsOKKU6AG86Xp/h+L1+opRa5zgG3yml7tZa/+TY3oqZyJoppb7F7K2xOs7SL8D825iktV7s2N7p\newAXKKXeB5K01k84Ht8IDMV8j7wOnHH8Dh52xN1JKRXCv1+gc4AvtdZPKKUuBl52HG8DeC43xnzt\ntQZGAU211mcAtNZpjl6pXk7iuw3zC06g4/g9r7V+x/H+ngM0dGy6TGs9uYTltzhe1+eY71+rUioY\nWAkM01oPKuWzIRP4EjgfGFXZvUrVkVzzrv5ilVJblFKHAI35oXorgFJqDOYb7QLHWfO3mB96AG8D\nO7XW52J+KNyhlGqLmTw2aa27A90wk/XEfO0ZmMnvFkcbVuBG4H2lVB/MM4NorXU3YAaQ/8MpSGvd\nMX/idrgQ+CM3cReyCjOx5rYdorXugdklPVUp1VEp1R6YDlzpaPdOYLHjjAWgBeZZ8BjgBiBOa91b\na90GM9GP1lovAb4GXnF0mz+Fmdw7Y34g+TleT24cv2mtzyuUuAG6A3GFX4TWerPW+pd8i9Idr+MC\nIEYp1dxxHJpqrXtqrTtgfsg+WszxGwd8pLXujZlMWwNXKaUCMD9Epzi6gG/HTL7/AzZhfqn5CvML\nUbbWOkpr3QVIBJ538vq+5N+zvOscx78r0MOxrLXW+lbHz/0cidkAUEqd79jnf7TW5zuO7xP5j4vj\nC+fXwGuOba4EnnX0tkQ7Nuubm7gdx9IGjAX+0lpfiZngg4AVWusoIAbzi0lp7wFXvAnc4vhiBebf\n1v8cbXbAvFTQBfNLUq6pmMn0HKALcJHjvTEF8++rO3Ab0M9Je92AHbmJO99r/sfxu8ijlKqD+XeQ\n+3d/Q+7rxvy9/+U4HtFAO6VUvRKW57YzF/NL/Tyt9U0U7CEq6bMhAPhaa32OJG73kDPv6q+v4zpj\nF8wPpp+11scc667G/JDdpJQC8xt/sGPdAMwzMhzXsDsBKKWuBno4zqBxbG8v1OYC4CWlVBPMs7ld\nWuu/lFJ3YiaS9Y72ABoopRpgfqD/RPECilleq1D7bzliPqyU+g64HMgAzsI888vdzuaIxQA2aK3t\njue9rpSKVkpNBNpjfrBvcNLuFcDjjkSBUuoNzKSYa10x8dowj3Np5jri+Ucp9Q/QWGv9s1LquFJq\nPOYZdl/MMQxQ9Pg9AlyulJqE2XvSDAjB/D3maK2/dez/V8wvH+Q7NmD+bdRXSl3meBwI/FPK61sH\nTFdKxWKekb2qtd5TzOuzYP6NfZd7yUNr/ZqT7SKBWrmJydErtAgziW8sZt+5+88vy/EFDMwelsaO\nn0t6D5RKa71VKbUXuFoptQs4S2u90nGN/IB2Ph5jAGYvjQFkY/4eUWYAbzl6iX6g0BcZBxuunXRZ\ntNapjvfrIKVUO8wvCnUc678Fljsu2fyA2UN02tFT4Wx5gX3j/LJOaZ8Nxb0nRDnImXcNobXeAjwI\nfKCUauVY7IfZjdbVcdbRHbObFMzuvDxKqdaOb+B+mN1kuc/pCdxfqK1UzAQ+CvMM/P187X2S77nd\ngJ5a6xOO9anFhL8BaO/4MlBYP2B9vse2fD9b+ffDblVuu462L8IczFegXaXUC8AzmInqXWAFzj+o\n/Aott1LwC8YZnNuAecwKUEpdo5R6Md+i7Hw/G4DF0eW8DPMD8UvMM6D87+H8x28e5lnU38ArmIP7\nLI79FrgeqpQ6z9FDkp8fZjd57vG6EPPyQbGvT2v9N+YXoueAesAPSqmhhbcr5jWilKqllIp0Ekdh\nVsp+4lHkeObbf3HvAVe9hXmmfCvm30yu4v4GCr+3miulwrTW72F+uVqJOQhvW/6zXod44FxH13vh\nfXzjGFMBYDguN23F7FlaBzzJv+MONmH2xrwHRAC/KKV6Fbe8lNef+/dU2mdDccdDlIMk7xpEaz0P\n+BmzmxTge+B2pVTuKNwpmKNHwfzWndu9Xh+ze7qd4zkTlVIWpVQg5jXmu500977j+b0wR6aCmQhH\nOq6rgZlcVjh+LnaAluPM7HXgc6VUs9zlSqlbMa9VvpBv8zGOdS2ByzCv3f6IeRaqHOuuwLwGHOSk\n3csxzxg/A4469pGb2HIwz0BxHIe7lFL+ji7Te/K9lpIswjyjnZTb1eq4HPEK8HsJz7NgDvpaqrV+\nF/N68nX5YnP2OqZqrRc4Hl/o2FZjfrBf6mi7GxCL+VlQ+PXdp5QKdMT5Dualh2Ljc/QIfKS1XqG1\nftSxjw6O9bZ8+wbzAz8WcwxB7t/DeP699JBLA1lKqesc8TbD/J2vLCEWHK+luN6a/Ep6D7hqIdDV\nEdesUrYF8711s+M9VAvzb6KPUioO8/LNbMzu91DHvzyO98JnwKzcmB0J/m3gmNY6w7GpBbPXK0lr\nPV1rvRJzzAZKKT9ljkN5ynGJ5AHMgYqRSqnnnCxvT8kD4HL/9lz9bBBuIMm7enP2hrsXuNLRHfoB\n8A2wQSm1HbP79OZ8252rlNqK2R37rKOL9X7Mrrdtjn/b+fc6Wl57jm2zgUVa6yzHshWYiXalY783\nYSag3OcW+wGhtX4c+BT4Sin1m1LqT8zBYj0LdU22VEolAN8BE7TWu7TWv2MO2pmnlNoC/BcYpLVO\nc9LuVMwu/w2Y1y4XYn5pAbOr8X6l1CPANOAI5peA3zETY6m3y2itszGTcAfgN8dxWICZaD8u4akG\nZgLto8xb8pZjJrAIZY68L/w6HgeWKKXWA5MxE0Q7x+9iCPC0Yz//A65zxLXU8dpHO47R35gD1XZg\nflbElBLfbMyBTL8rpeIxB9XldoUvBtYpc/BZ7rHYDkzCHHC2BfMLx51OjtdgYILjWK0EntFar8nX\nrjPbAZvj9+hsu9zHJb0HCm9/rlIqJd+/00qpuo4YF2Jekkou9BxnbT6DeQ18K2aPyDJHl/7DmOM0\nfsX8wjlFa73fSSx3Y/7NrXf8Djc4Xu+4Qm2tAA4qpbQyB/dlYo5dyB270kUp9Rvm2fwezEs1rzpZ\nPq9Q/EYxP7v02SDcwyJTgorqwnHtcUShgV9CeJQyB4atwayfEO/teETNIGfeQghRTkqp/wD7Me97\nlsQtKo2ceQshhBA+Rs68hRBCCB8jyVsIIYTwMT5TpCUhIUH694UQQtQ4UVFRRW6l9ZnkDRAVFeXW\n/SUkJLh9nzWNHMOKk2NYcXIMK06OYcV54hgmJDifAFK6zYUQQggfI8lbCCGE8DGSvIUQQggfI8lb\nCCGE8DGSvIUQQggfI8lbCCGE8DGSvIUQQggfI8lbCCGE8DEeT95KqQuVUrFOlg9SSv2ilFqvlBrn\n7LlCCCGEKMqjyVsp9TDwPlCr0PIA4BXgMqAPcIdSqrEnYxFCCCGqC0+fee8GhgCF67KeC+zWWp/S\nWmcDPwGXeDgWIYQQwiPOnDlDenp6pbXn0eSttV4M5DhZVQ84le9xClDfk7EIIYQQpUlPT2fXrl3s\n2rWLhIQEFi9ezMyZM5k7d67T7ZctW0ajRo2oW7cuX3/9daXF6a2JSU4BdfM9rgucKO1JxRVorwhP\n7LOmkWNYcXIMi/dREmQaRSZVKsQCy3+tlHiqr4ofw2ZrDmPN8fwEkIZhcOLUEU6c+oeTp5M4dfoo\nflYrfXveUGTb5JNHiP25aOJtUK8J/S+6scjyfYd+Z9b8R4ssjzi7E7sT6hbd/uAOjh8/jr81AL1l\nX+W9lw3D8Oi/yMjIiMjIyJ8LLQuIjIz8MzIyskFkZGRgZGTkpsjIyLNK2s+mTZsMd/PEPmsaOYYV\nJ8ewZOOWJZS6jRzDinPHMXxm4teutfX9QyWuT0tLMwa+eaPTdRs3bjQaN25sAAX+RUVFOd0+Pj6+\nyLaFt7964pcFtm/Xrp3Rrl07o3PnzsagQYOMe++913j0/G5O95+enm4cPnzYsNlsxk/XDCntpZeZ\n4/dSJCdW1pm3AaCUGgmEaK3fV0pNBL7H7Lr/UGudWEmxCCGEqCJycnIYN24cqamp7Nu3j3379pGU\nlIR/UADc82mR7du0aUNSUhJNmjTBFubHld0vJSIigi5dujjdf+vWrZkzZ06R5WFhYU637969O7t2\n7SqyPO7aoXk/Z2dnExAQAEBQUBBnnXWWS6/VnTyevLXWfwO9HT9/nm/5N8A3nm5fCCFE5UpLS+P4\n8eN5jw3DYMuWLVxzzTVFtjUMg9mzZxdYFhAQQFBYHaf7btSoEXv37qVVq1aM+OJu5oz4X4mxNGzY\nkNGjR5fjVRRlt9vZvn07mzdv5rrrrqNevXpu2W95eOuatxBCiGrIbrdz7rnnsn///iLrNm/eXOQM\n2Wq18tFHH1GrVi1atWpFq1ataNq0KSMX3ltsGxEREe4Ou1RZwUEsW7aMxMREgoODOXPmjCRvIYTw\ndS8++R0Z6dneDqNCls09XOo2J079Q8Jv35OWfpr2EVGc275X3jp/WyY/X3c9F1gDSK0VhNXy70DD\nJsHBrL93AqkNGwEQ99a/3dDtHf8bwN+OfxOAuLn/buOMK9u44lEg7tqiXeu5MZ1s0Zwjl1yEPTGR\nukf+odlvO9i78Ev2Ft44KKjCsbhKkrcQQrhBRno2k18eVOp2w+eP54tSunpzDYr5iqUvX1vR0Ii7\ndigXfbWoxG0SEhKIiooqdv1ff/3FZbdfzYF1u8nJMe8A/s+g7kye/u9rjrt2KKva3cL8JQPx8yv+\nTuSEFZOIunxGGV9FUe46PiVJSUnhiy++wOrnxyUXXUT79u2xWJzf/VCZd41I8hZCiGrIbrfz999/\n06ZNG6fr7rvvvgLLjh49SuPGjXnzzTeLbL9y5UquvPJKbDYbfn5+3HTTTURHR9O1a1enbZeUuH1N\n3bp16d+/P+Hh4YSEhHg7nDySvIUQohp46qmnOHnyJAD79+9n3bp1ZGVlceKE8xIab7/9dpFlFovF\nafKOjo6mWbNmWNvWZuX739CuXTv3Bl/FtW7d2tshFCHJWwghqoGPP/6YgwcPFljWsmVLDhw4UGRb\nZ0l6//79tGzZ0um+g4KC0Fpz89cTq3XiPn78OA0bNvR2GC6R5C2EEG4yfP74Urex5gQwNWZpidtc\nsmcutqw0LjuRzLhzHuNM9r8D4TqFNaRvs+ZFnjO6UWPSQ83EUzcggPMbNqJp7dokPjgJKHifMkDh\nu6K7AOzcTdyKH4uNq7QBYul+gcRjZ1DMVyW+PrgYvi9tm9KFBAdUeB9g3mu+ceNGduzYwYABA2jb\ntq1b9utJkryFEMJNcgeiTY1ZWvLgtaJVOQuIu/Zjfr64J48//niRdQ2vuJyL3nor37bmYLSLnOyn\nLAO6ShuwBq4NNLvUpdaqjqSkJGJjYzl16hQNGjTw6u1fZSHJWwghvGTr1q3s3buXwYMHF1nXt29f\nQls34uZrbizQnX3++edXZojVlt1uJyEhgS1btmAYBp06daJHjx74+/tGWvSNKIUQopo4cuQIc+fO\nZfbs2Wzbtg2llNPk3atXLy57bhivjnjVC1FWf3a7nT179lCnTh369u1Ls2bNvB1SmUjyFkKISmAY\nBhEREQUqj4WFhXHppZdis9mwWq1ejK7m8ff35z//+Q+1a9cmMDDQ2+GUmSRvIYRHTFi5lbRsW5mf\ndzplDgaZBZYNn+/CE3d/UOa23KkjV+UNWMv/c37JWafwC7DStEsL7vBvSp+GTQk4kMiGIcOLbBt3\n7dASB4i92noEGdZa0G4MFDNAzF0Duqqr0NBQb4dQbpK8hRAekZZt4/2rupX5ecPnv+9yBbJ/n+N6\n1bKSVKRi19RflvLFiP+RlpbG5FVfOI3n+KXTaNCgAX5+fiVWPXOlItrzlVBdrDowDIO//vqLVq1a\n5c0EVh1UnzI4QgjhZTt37qRHjx58/vV0srKyiqxv2LBhtao+VtWlpaXx/fff8+OPPxIfH+/tcNxK\nzryFEMINTp5OYsCAuzh8+DCNws7myJEjxRY9EZ63Z88e1q1bR2ZmJs2aNaNz587eDsmtJHkLIUQF\nHT9+nE8WT+FY8mEuueQSos+/VxK3l9jtdtasWcOuXbuwWq307t2bDh06FDuZiK+S5C1EDTLyyeWc\nqaRpK5sOaFFqpa2gbquw+BeNx5VKZfkZOQEuVPUqnSsDvApP/ZmVncmchU9xLPkgberW47E69ai7\nbx5x184rdV+Fq57l8q9CE2D4mtzLEuHh4fTr18+nB6WVRJK3EDXImfTsShvkdPvyX0tta/j874oM\n7CrP4LOEhASibiy5Opi7FJ76Mzs7m0Mpy1iy6Bhr/9hC8+bNXRpwJjzn4osvxmq1VuvxBZK8hRDC\niVmzZnHs2DFOnjzJ33//zb59+9i3bx83XvNCge0CAgKYNWsWjWp9QvPmRWuOi8pXnUaVF0eStxCi\nxvnjjz+YPXs2n3/+OT/88APt27cvss3MmTPZvn17keWnTicVWWaxWKhbp4FHYhXO2Ww2Nm/eTOvW\nrX1mJjB3kuQthKgREhISWL16NV988QW//PJL3vJ9+/Y5Td633norR44cISQkhFatWuX9+/j1rZUZ\ntnAiOTmZ2NhYjh8/zrFjx7jiiiu8HVKlk+QthJc9v/AQGXMPlr6hG5S34tatS2JIzUor8/NcqYxW\neHBancDaZW5nxcJEFnz4Gdv1TzQNj6Bl8/OKbPP9mo9Yn7AEgFqBwXSIvJgu5/XnjeUpvPats8Fu\nbc1/pyB+G7DtFLAN/CxOB8f1wO/f5SVUPXOX3N/llh8nY8tJd8s+E1aUPMjO6h/slnbKy263s337\nduLj47HZbCil6NWrl1dj8hZJ3kJ4WUaWUeUrZaVmpZV5ENnty38tV4W1sjpy5AjLfpjFlj9WkJqa\nyqRJk5j8YtHpOLsuhW++CadPnz4MHjyY2rXNLwkVqaqW39SYpXn7qcwBa7ac9FKn6XSFK1OCepNh\nGKxYsYL9+/cTHBxMdHQ0ERER3g7LayR5CyF81vjx4/n444/JyMgA4LLLLuOii5zNbA2DBg1i0KAS\n5tgWVZrFYiEiIgI/Pz+io6MJDvZuL4C3SfIWQlRphw8fLna6xnfeeQeAc9peyJzP36BHjx6VGZqo\nZEoplFLVruBKeVTfm+CEED7LMAw2b97MTTfdRMuWLdm8ebPT7R5++GG2bt3KyGufkMRdzRiGUWSZ\nxWKRxO0gZ96ixqnMKmOuCAqsvA+jwtXBXHaBeU0316FLzsIeUPJ3f79se4HnuCIl9QTb/ljN1t9/\n5J9j+8z9+FmZPOl9epx/ZZHt63AxX87eh789s9hqZaV5FIi7dk65nltAu1vyYnClQpq7Bpp5exCZ\nu2VmZhIXF0fTpk0577yiAw+FSZK3qHEqs8qYKxISEiqtrcLVwVw1fP7yAs9zeTBaGQ/zK6+8wkvv\nfgSYM3CNGjWKiRMnljowqbgBYq4MHHPXgLVVMUvLNEjNXQPNqpODBw+yZs0aUlNTSUtL49xzz5Uz\n7WJI8hZCVKoDBw7w3nvv8cwzzxQpX3njjTeybt06xowZw8CBAwkMDPRSlKIy5eTksHHjRnbs2IHF\nYqF79+506dJFEncJJHkLISpNQkICgwYNIjExkbCwMB588MEC65s0acKSJUu8FJ3wltWrV7Nnzx5C\nQ0Pp168f4eHh3g6pypPkLYSoFIsWLWL06NGkp6fTt29fbr75Zm+HJKqIqKgoQkJC6N69O/7+kpZc\nIUdJiCqmvNXMXHKBef26rMpT9SxXamoqY8eOZf58s9za2R36U/v88dw8fZ3T7SfsmUewPavM7Tgb\nsJbuF1jqVKHd8CvzwDpngspZvU5AgwYN6Nmzp7fD8CmSvIWoYspTzQzM0eDlGYzmabVr1yYxMZHg\n4GBaXzCS7bEflHgtM+7aOWWuTlbSwLRLS3luVT1u1ZFhGOTk5NSIWb88TZK3EMKjLBYLs2bNwmq1\nct8bW2UQUg2VkpLC6tWrqV27NgMGDPB2OD5PkrcQwuPatm3r+Elm5KppDMNAa83PP/9MdnY2ERER\n2Gw2rFart0PzaZK8hRBuY7PZyMrKqvF1p4UpLS2NtWvXsn//fgICAujbty/t27eX3hc3kOQtqhVX\nqqeVd1rMiipxINruD/J+rMjgMHdwFmd6Qn+wFbznetCqAwUedzHg2x/+x9Hj+xl57ZPUDq5bZN89\nXBkc1u4WVpV1AFl5nuPgiYFmrlRPq26V0Zz57bff2L9/P82aNaNv376EuFB5TrhGkreoVqpa9bT8\nihuIVtWmYnQW56BfClYhy19hLT09nX379jFu9BMk/PY9QUFBDB7Txuno4cIDy9w1dWZVO4ZSPc0U\nFRVF/fr1ZTIRD5DkLYQot9jYWPr375/32M/Pj3nz5sltPwIAf39/zjnnHG+HUS3JrGJCiHI7++yz\nCQgIoG3btrRpeT4LFizg2murZs+H8JycnBxOnjzp7TBqFDnzFkKUW7t27UhPT8dqtTI1ZilDhsj9\n0jVNUlISsbGxGIbB0KFD5R7uSiLJWwgf4ErVtY5cxfD54yvcVnED5pzd3mOxWOSWnxrKZrPx66+/\nsmXLFgzDoFOnTnJduxJJ8hbCB/gFDqV+raC8x83WHMaaYxTZruMvV7mlvak//Tty2263EfTnTzRt\nfC/DBz5Mk/AIWgBTVx0q8JyqXh7UXfNnu6K6jyRPTk4mNjaW48ePExISQt++fWnWrJm3w6pRJHkL\n4QP8LEEF5s+euupQkZKe7i7zabPZmDdvHtOmTWPnzp0ApBDPWy/f57Y2KpOMAHefkydPcvz4cZRS\n9OrVS6Zu9QJJ3kKIIrZs2cKIESP4888/AQiu15jXXprGLbfc4t3ARJXQpk0bhgwZQqNGjbwdSo0l\nyVsIUUTLli1JTEykTZs2PPHEEyzaVp/bby86a5eouSRxe5ckbyE86MUnvyMjt+LbBRRbXWzZ3MMl\n72hA8yLPLfy4PNecz5w5A1Ck8lVYWBhr1qyhY8eOBAQEsKSUaTVF9ZSWlsY///xD69atvR2KKMRj\nyVsp5Qe8DXQGMoFxWuu/8q2/DngcMIBZWut3PBWLEGU1YeVW0rJtLm9/OmUOBplFV3T690cLtTgw\noHm54qkdYC1wPduV69vFlYpNP32Ug3+s5p+/4jn1z2469r+DVudfUcxe9gPeKykLrg80S1gxr8T1\n1X0Qmbvt2bOHdevWkZ2dzbBhwwgNDfV2SCIfT555DwYCtda9lVIXAi87luV6BegKpAK/K6U+11qf\n8mA8QrgsLdtWYIBYaYbPf99p6VNXkuzw+ePLNX93aQqXil27di1TpkwhNjY2b5nVamXQhWFMn151\nC6u4MtCsqpVH9WWZmZmsX7+eXbt2YbVa6dmzJ/Xr1/d2WKIQTybvi4DvALTWG5VS3QutzwZCATtg\nwTwDF0JU0KlTp5x+2GZkZBAbG0utWrUYPHgwo0aNol+/ftStW3QCEVEzJSUlsXLlSlJTUwkPD6df\nv35yxl1FeTJ51wNO53tsU0r5aa3tjscvAwmYZ96LtNanC+9ACOG6nJwcnn76aWbNmsXmzZuLrB8w\nYAAfffQRgwcPlg9k4VRwcDA5OTlERUXRtWtX/PykgnZV5cnkfRrI/5U+L3ErpVoC9wKtgDTgU6XU\nMK31Qg/GI3xcVZvu05XBaJVVuOTAgQOMHDmSuLg4/Pz8+PHHH4E6BbaxWq1yq5coUd26dRk5cqTc\nt+0DPJm844BBwAKlVE9gW751QYANyNRa25VSSZhd6CVKSEhwe5Ce2GdNU1nH8Ex6NlNGnV3qdu6J\nx5K3n9f2fEKGveTBaEF+tRg4ynmFKVfiKU/MCQkJnDx5ktmzZ7Nw4ULS09MJDw9n+vTpKKUg4WCV\n/PvOH1OJiMk5AAAgAElEQVStsa1IWDGp5CdYAj12DEVBcgwrrrKOoSeT9xLgMqVUnOPxrUqpkUCI\n1vp9pdRsYL1SKgPYDXxc2g7dPSBFBrlUXKUew7kHK6+t5b/mtZWx+4NyD0Zzye4Pyvy6ls09TFRU\nFLt37+azzz7Dbrdz3XXX8d577/17/21lHi8XxVHwfZxwfJ5bqp7Je7lskpOT0VrTs2fPvHrkcgwr\nzhPHsLgvAx5L3lprAyg8S8Kf+dbPBGZ6qn0haoJ27doxc+ZMevXqRY8ePbwdjqji7HY727dvJz4+\nHpvNRqtWraQmuY+SIi1C+Ahns3oB3H///V6IRvia06dPs2bNGhITEwkODiY6OloStw+T5C1EBbhj\nCs4gv1olrk9LS2PChAmkpqYyd+7cCrcnap6jR4/yzTffkJ2dTUREBNHR0QQHS9EaXybJW4gKcEdx\nlcLXtPKPYk85k8zcr6Zx+J/dWK3+NKndjwb1m7ptFPvGG28mx1EitVLU8iswQM3IcL2KnSi/sLAw\nwsPDiYyMpH379jLvdjUgyVuIKiYjPZvJLw9i27ZtXH31PRz+5wCtW7dm6dKldOjQwa1t5Zw5w0Vf\nLXLrPnMlrJhU6mC0uGuHwjUeaV7kY7VaGThwoCTtakSStxBV0IYNG7jssss4c+YMvXv35ssvvyQ8\nPNzbYQkfYBiG0yQtibt6kfI5QlRBHTp0oE2bNtxwww2sWrVKErdwycGDB/nyyy/JyMjwdijCw+TM\nW1QJVa16GuSrmFZM9bQca5bH2q5bty6rV68mNDTUp8+Y4q4teQ5w/0JTkYryycnJYePGjezYsQOL\nxUJiYqJM41nNSfIWVULhGbCqgtwCLMPnL3dajMUcaV5ycqqIBg0aeGzflcVT19PFv5KSkoiNjeXU\nqVOEhobSr18/6ampASR5C+FFOTk5fPLJJ0RGRsrsXqLMUlJS+OqrrzAMg06dOtGjRw/8/eVjvSaQ\n37IQXrJv3z5GjRrF+vXrOXHiBLNnz/Z2SMLH1K1bl+7du9OkSRMpuFLDSPIWwgsWLVrEuHHjOHny\nJI0bN+a2227zdkjCR3Xt2tXbIQgvkOQtiuXKIDIA5h6scFuVORjt6UcWY80ppb0BzQtUT3NWSa1O\nYO0yt20YBtdffz2LFpnXggcNGsSECRPo06dPic8rVzGVdmNKHTAGpQ8qK6+ge9p4ZL81VVZWlkzV\nKfJI8hbFcmUQmS/ORHSkd0vsASXfJVk7wMr7juppw+ePd0sltdzKaft2pRLgX4tLo28mqt1A4pan\nEbf839HszqqnFS6m4soXqwf2zis1Jv+QEC78zDPd9aVO9SlcYhgGWms2bNjAFVdcQdOmTb0dkqgC\nJHmLGsce4Mf7V3Wr9HZzK6fd+c8FWCwWGjduDJTvC5ArX6zirp0jo719XFpaGmvXrmX//v0EBASQ\nnp7u7ZBEFSHJWwgPyM7OZuXKlVx11VVF1jVp0sQLEQlfs2fPHtatW0dmZibNmzenT58+hMh98cJB\nkrcQbrB06VLWrVuX93j58uXs2LGDtWvXEh0d7cXIhC/Kysrip59+Iicnh969e9OhQwefLtYj3E+S\nt/AZ+WfbqpABzSu+j0JiY2OZOXNmgWVt2rTBbre7vS1R/QUGBjJgwADq1KlDaGiot8MRVZAkb1El\nTFi5lbTsUqaH7N3YLW3ZjX/rPt+6JIbUrLQStz/318uY+kvR8qj5JR8I47Lom/MeBwfVpfM5fVjz\n9WnWfG0+113TeIqaoXlz93/JFNWHJG9RJaRl20odRDY1ZqnTMqVlZd721RuA1Ky0UkeST/3FlXYr\nHpeomY4fP06DBg3w85N5ooTrJHkLUYq/9m3hySc35j3u2rUrQ4YMkWuQokJsNhu//vorW7ZsoUeP\nHnTp0sXbIQkfIslbiBKsWLGCOYsmF1gWFhZGu3btOP/8870UlfB1ycnJxMbGcvz4cUJCQvJuGxTC\nVZK8hSjBRx99BMCNN97IOeeck7f82LFj3gpJ+DDDMPjtt9+Ij4/HZrOhlKJXr15SOU2UmSRv4XGu\nDEarHWCtpGjK5tNPPyXzZFNmz34Jq7VyYlw1ZCRBtqJzhecvY/ooZhGWknhyruwtP07GllNywRCr\nf7DH2vdl+/btIzAwkOjoaCIiIrwdjvBRkryFx7kyGK2qslqtdDmvf6UlboAgW1aRymhx1w6tUtXS\nbDnpRF0+w9th+ByLxUK/fv2wWq0EB8uXG1F+kryFEKISSZU04Q5yb4IQ+RiG4e0QRDWxd+9eUlNT\nvR2GqKYkeQvhsG7dOqKiokhMTPR2KMKHZWZm8uOPP7Jy5Uri4uK8HY6opqTbXPgcV6qilSZ3fu7c\nObnnzp3LrbfeSlZWFjNnzuTFF18s0/5cnvu8OPnmRH+0/HsRXnbw4EHWrFlDamoq4eHhXHDBBd4O\nSVRTkryFz3GlKlpJ8s/Pbbfb+e9//8vkyea93Pfeey/PPvtsmffpyhSdxSk8JWhpo8hF1bR+/Xq2\nb9+OxWKhe/fudOnSRaqmCY+R5C1qLLvdzpAhQ/jqq6+wWCy8+uqr3H///d4OS/iowMBAQkND6dev\nH+Hh4d4OR1RzkrxFjeXn50dUVBRr165lzpw5XH311d4OSfiwbt260aVLF/z95WNVeJ78ldVQrlyj\nDakBs2A9/vjjjB07lmbNmnk7FOHj/Pz8pJtcVBpJ3jVURa7RVidWq7XSEvfGG28m58wZp+vyj0n2\nZGU0UTF2u53t27fToEEDWrRo4e1wRA0myVvUGCkpKdStW9dr7eecOeO0SlrhAWuiakpJSWH16tUk\nJibSsGFDzj77bJlZTniN9PGIGuHjjz9GKcXWrVu9HYrwMYZhsHPnThYuXEhiYiIRERFcddVVkriF\nV8mZt6j2li5dyrhx47DZbGbRjIbejkj4kp9++ok//viDgIAA+vbtS/v27SVxC6+T5C0qZMXCRJbN\nXVryRgOaMzWmlG1cEOTCALr8BVziZnzHyX3HSD+RhmGzc+513Vjd8Le8wixVRcaLrxCXkVHiNnId\n3HvatGnD6dOn6dOnj9QlF1VGqclbKRUGvAC0A4YDLwITtdYnPByb8AHZWQaTXx5UctWzFDjlpkJT\nw+d/xZ0LjxE3d6jT9ePy/bz9r1McPmkOEBvaug0P5LTCMjcJgLiPnT/fqXa3FJiO05kKTdGZkVGl\nZgyT6T4Lat68Oc2aNZOzbVGluHLm/T6wArgQSAEOAZ8CAz0Yl/AxJVU9u335r26dEjRubvHTY+av\nnvbD4cNkZWURFBRE06ZNnW4/KOarUkfdr4pZWqWSq6fV5Ok+DcNwmqQlcYuqxpUBa6211u8CNq11\nhtb6SUDukRBVxv79+0lLK3rW36xZMyIiIopN3ELkysnJIS4ujvj4eG+HIoRLXDnzzlZK1c99oJRq\nD9g8F5KoDF2xuOU6dEBg5Z6RHDlyhG/2/83v778PmGdKL7zwAk2aNOGbb76p1FhE9ZCUlERsbCyn\nTp0iLCyMqKgorFart8MSokSuJO+ngdVAS6XUV0Av4DZPBiU8zx8Lk18eVOH9JCQkuCEa123bto0F\ne/7irzvuKLA8LCyMgIDqXxFOuI/NZuPXX39ly5YtGIZBp06d6NGjhyRu4RNKTd5a6++UUgmY17yt\nwJ2ADFar4kqdNvMCGD5/uXsa2/1BpY3gvvzyy3m6Ww/WtovIW9aoUSNiYmK8WoBF+J74+Hi2bdtG\nSEgIffv2lRK5wqe4Mtr8Z611L+Abx2MrsAXo5OHYRAWUNm3m1JilbjvzruzqYG3q1WO0o9tciPLq\n3Lkzdrud7t27ExgY6O1whCiTYpO3UioW6OP42Z5vlQ34ysNxCSGER9WuXZvevXt7OwwhyqXY5K21\n7geglHpday2THAtRDiVNRpInKKhygqmhDMMgOztbzq5FteLKgLWHlVLXASGABfO6d2ut9WSPRiaE\nQ05ODjk5OQT5YJIrbjKS/Cp70F9NkpaWxtq1a8nMzGTQoEEyZaeoNlxJ3ouBYKA9sBa4BOk296pS\nB6NBlSsBWhGLFy/mnnvu4ZlnnuHuu+/2djjCR+zZs4d169aRmZlJs2bN8gr2CFEduJK8FWZp1NeB\nWcBDwLulPkkpP+BtoDOQCYzTWv+Vb30P4GXMs/lDwBitdVZZX0BNVNpgtOrm9ddf59ixY94OQ/iI\nzMxM4uLi2L17N1arld69e9OhQwepkiaqFVf6kP7RWhvATqCz1vow4ErJqsFAoNa6N2bp55dzVyil\nLMB7wC1a62hgFdC6rMGL6i8hIYG4uDjq16/PmDFjvB2O8AG7d+9m9+7dhIeHM3ToUDp27CiJW1Q7\nrpx571BKvQH8D/hMKdUMqOXC8y4CvgPQWm9USnXPty4SOA5MVEp1BJZprXXZQq+ZRj65HDqZNbkr\noocPTOX+7LPP8tJLLwEwduxYmdFJuOS8887D39+f9u3byzVuUW25krzHA7201r8rpZ4GBgCjXHhe\nPeB0vsc2pZSf1toONAJ6A/cAfwHfKKU2aa1jyxZ+zXMmPZtgKHUyjdK4ozSqpyUnJ3PixAlCQkK4\n9957vR2O8BEWiwWllLfDEMKjLIZhFLtSme+AFEdXee6yJsB/tdZ3FPtEc7uXgQ1a6wWOxwe01i0c\nP58DfKG17ux4/AAQoLUudiqjhISE4gOtQabMPUjwBd/xSLtxpW9cgmVzDzNwVMUrSn2UBJlGyV2S\ntSwGtzZ2bX/PLzxERpb5q85KP01OVjoBQXUJqPXvALxHd8/h+Xbu6ULvigV/So4/INDC5cPOKtf+\nM6Y+S9Dkx8v1XI9IXgxGKUNLLIEQNqRy4qkAu91ORkYGtWtXn8GZQjgTFRVV5EOqpCItUzAHp+G4\nVSzW8fhxYIML7cUBg4AFSqmewLZ86/YAIUqpto5BbNHABy68ABeadZ03qoNV2NyDQMWPxbK5h93y\n2t9x83SfGXMPltqrEHftnGK3yT8lqCvcVWmuOHGU/ruqzL/DhBXzqsV0n8nJycTGxnLmzBmGDRvG\nzp07fe+9XMX45OdhFeOJY1jcraQldZvfjHl7WDPgv8AjQBPgeq319y60uQS4TCkV53h8q1JqJBCi\ntX5fKTUWmOsYvBantf7WtZciqqP33nuPhIQEzmSf7+1QRBVmt9vZvn078fHx2Gw2lFIyIY2okUpK\n3qe11olAouO2rk+ASVprl6YDdYxQH19o8Z/51sdiTnYiaji73c6MGTPYvXs3UYMe9XY4oopKSUlh\n9erVJCYmEhwcTHR0NBEREd4OSwivKCl5569nfgyIcSRkIdzq+++/Z/fu3bRs2ZImbXt4OxxRRaWn\np3PkyBEiIiKIjo4mODjY2yEJ4TWujDYHyJDELTzljTfeAODuu+/mpySZS1k417hxY4YMGUJYWJjc\nty1qvJKSdwel1F7Hz83y/QxgaK3beDAuUUPs2rWLb7/9lqCgIMaNG4ftv7Gl38bW7hZWFbNNR65i\n6i+u3wYXFFx9rpdu+XEytpz0Erex+vv22WrDhg29HYIQVUJJyTuy0qIQNdbPP/+Mv78/o0aNomHD\nhuRYa1Vo9HdZR5tXJ7ac9GoxkjwzM5N9+/YRGSkfQUIUp6QpQf+uxDhENXb06FHmzp1LcnIyzzzz\nTIF1Y8aM4dJLL8VmK30cZGVOyOLSVJ4u8JeqcGVy8OBB1qxZQ2pqKiEhITRrVvFaBEJUR65e8xai\nzFJTUxk3bhwLFy7Mm9Jz4sSJ1K9fv8B2rn5AV+aELK5M5SncJycnh40bN7Jjxw4sFgvdu3enaVNX\nplAQomaSwr/CY+6//37mzZuHYRgMHDiQOXPmyAhhUcTJkydZtGgRO3bsIDQ0lMGDB9OtWzepS+7E\nxo0b6dWrF6NHj2bMmDEMGTKECRMmkJ2dDZjFax555BFGjx7NjTfeSExMTIEZ+TZt2sRtt93G6NGj\nGTZsGHPnzi3SxoEDB7jiiit47LHHnMZw8OBBRowYUWT5hx9+yJIlS5w+Z+XKlcTExDhdZxgGjz32\nGGlpJfeqVYY333yT66+/nhtuuIFt27YVWZ+QkMDw4cMZMWIEr776at7yxYsXM3z4cJ588knee+89\nANauXcvChQs9FqtLZ95KqYuBjsDHwAVa67Uei0hUCwsWLGDWrFkEBQWxYcMGzj9fiq8UK3kxCSvm\nVXg3vjoYLTg4GJvNRqdOnejRowf+/tIhWByLxULv3r15+eW8SRqJiYnhxx9/5PLLL+fee+9l3Lhx\n9O/fHzDHlNx5550sWLCAQ4cOMX36dD788EPCwsLIzMxkzJgxtGzZkosvvjhvfwkJCfTr149HHnmk\nzLE5M23aNOLi4jjvvPOcrv/222/p2LGj18vc7tixg/j4eBYsWEBiYiL33XdfkeT7wgsv8Nxzz9G2\nbVtGjRrFn3/+SVBQEPPmzePTTz9l69at/PTTT9hsNi655BJuv/12rrjiCo9MqlTqu8RRd3wwZqW1\nRcB7SqkPS6pDLsSAAQMYPHgwV155pSTu0hhZ1WKgWXnVqlWLYcOGERgY6O1QqjzDMMg/H0VWVhZH\njx6lfv36bN++nXr16uUlboBevXrRsmVL4uPjiY+PZ/DgwYSFhQHmcf/www8LJM3Dhw/z7rvvkpGR\nQcuWLTn//POZNm0aVquVwMBApk2bViCeH374gbfffpvQ0FAsFgtXX311kZi7devGZZddxvz5852+\npk8//ZS33noLgF9++YW33noLu91OWloaL7/8Mv7+/owfP57Q0FD69OlDdHQ006dPxzAMGjRowLPP\nPktwcDCTJ0/myJEjHD16lP79+/PAAw8UaOeuu+4iNTU173H79u2ZPHly3uOEhIS8LzFnnXUWNpuN\nEydO0KBBg7xtatWqxcmTJ8nKyiIzMxOr1cr69evp2LEjDz/8MHv37uWhhx7CajVvee3Tpw9Llixh\n9OjRTl97RbjyFfcWzEpoG7TWRx1Te8YDNffTxkNGPrmcM+nZJW4TEhxAaUO7XnzyOzJK2Y+nb5EK\nCwtj8eLFHm1DVB++mrjvmfEj+4+kuG1/LZvW5a1J/UvcZsOGDYwePZrk5GT8/PwYMWIEPXv2ZPny\n5bRo0aLI9i1atODw4cMkJSVx7rnnFlhX+IywWbNm3HHHHezdu5eRI0cyZMgQnn32Wc455xxWrVrF\nc889l3dGbrPZeO6551i8eDH169cvtlv8qquuYuPGjU7XZWRkkJiYmJcgd+/ezYwZM2jcuDHvvvsu\n3333HYMGDeLYsWMsWbIEf39/hg8fnnf2u2DBAt5//32GDx9Oly5duP7668nMzKRPnz5Fkvc777xT\n4nFNTU0lNDQ073GdOnU4c+ZMgeQ9duxY7rrrLkJDQznnnHNo06YNK1asID4+nvnz57Nx40amTZvG\nwoULqVu3Lkop5syZ47XkbdNaZ+abYi8DyHF7JIIz6dkuTfU5fH7J9zFnpGd7dLINV0khDZFfSkoK\nmzdv5qKLLso7M/F1pSVaT+jZsyevvPIKJ0+e5LbbbqN58+YANG3alEOHDhXZ/u+//6Z3794kJSVx\n5MiRAut27tyJYRhFknru2f3Ro0c555xzAOjevXuB7vrjx49Tt27dvAGoPXqUvTriqVOnCiTHxo0b\nM23aNOrUqcM///xDt27mpEdnn3123uWUPXv2MGXKFMAc6BgREUH9+vX57bff2LhxIyEhIWRlFZ05\n78477yxwXb1du3Y8/fTTeY9DQkIKnJmnpqZSt27dvMcZGRlMmzaN5cuXEx4ezowZM5g1axYNGjTg\nwgsvpHbt2tSrV4+2bdvy999/06lTJxo1asTJkyfLfFxc4UryXuOY3jNEKTUYuAP40SPRCCGqHcMw\n0Frz888/k52dTdOmTeUebjcIDQ1lxowZjBkzhi+//JJu3bpx7NgxYmNj6devH2AOmjpw4AAXXngh\nLVq04J577uHKK68kLCyM1NRUnn76ae65554CyTt/t3zjxo3RWqOUIj4+ntatW+eta9iwISkpKRw/\nfpyGDRuydetWLrjggjK9hgYNGhRImJMnT+aHH36gdu3aPProo3mx5B+82Lp1a2bMmEHTpk2Jj4/n\n5MmTLFmyhHr16jF16lT27dvHF198UaStd999t8RYunXrxowZMxg7diyJiYnY7fYCZ+J2u52cnJy8\nQbfh4eGcOHGC6OhoPvvsM7KyssjIyOCvv/6iVatWAJw+fTrvMoW7uZK8H8JM2FuBMcByoOT+B1Hj\nfPnll/Tv35969ep5O5QqxZWqZ1h8s8vYFWlpaaxdu5b9+/cTEBBA3759ad++vbfD8lkWi6VAj1bb\ntm0ZPXo006ZN47XXXuOdd95h+vTpeYnqrLPO4t1338VisdC8eXMmTZrEfffdh5+fH6mpqQwfPpxL\nLrmk2DamTZvGf//7XwzDwN/fP+9as8ViwWq1MmXKFG6//Xbq1atH7dq1i+1tKxx3rsDAQBo1akRy\ncjJhYWFcc8013HjjjTRu3Jg2bdpw9OjRvOfnmjJlCpMmTcJms2GxWHj22Wdp3bo1MTEx7Nixg2bN\nmtGxY0eSkpJo3Lixy8e2Q4cOdO/enREjRmC32/POyjds2EBCQgL33HMPDz30ELfccgvBwcHUq1eP\n559/nrp16zJs2DBuuOEGUlNTufvuu/M+B7du3Urv3r1djqFMcgdAFPcvMjLytcjIyO6lbefpf5s2\nbTLczRP7rIirJ37p0nbXz7urxPXPTPzaHeG4ZNyyBOOdd94xAOOiiy4ysrKyKrS/kmIv7XW700/X\nDHHLfjZ9/1Dp21Sxv0N3OX36tDF79mzj3XffNZYuXWqkpKR4rK3qegwrk7eO4TfffGN89NFHXmnb\n3Qofw7Fjxxpnzpxxxz6L5ERXzrx3Aa8qpRoCnwGf1qTqa64MInOXEB+ss/37N1/wwf9eAGDIkCHl\nnls5t3paR65i+PzCM8maXKmeJpXRqo6QkBCaN29OkyZN6NChg4yBEE4NHDiQhx9+mLS0NK/fLuZO\na9as4T//+Q916tTxyP5LTd5a6zeBN5VSrYDrga+UUila64tLeWq14Oogspro008/Zb0jcc+cObPI\n6M6yyK2eNvWXpRWqoiaV0aoOi8VC//79JWmLUr344oveDsHt+vTp49H9u1TCSClVH7gUuBywAt97\nMihR9W3dupVx48YB8Morr1QocQvfZxjOZwyWxC2EZ7hSpGUp0A1YDDyltXZ+w56oUZRSjBw5kg1J\nKTz44IPeDscrasIUnK5ISkrip59+YsCAAUXq1gshPMOVa97vAd9qreXebpEnKCiIWbNmMe6bTd4O\nxWuqyxSc5WWz2fj111/ZsmULhmFw6NAhSd5CVJJik7dS6hmt9dPAEOA6pVT+/i9Da32bx6MTXpeW\nlkZOjvm9rfBtYBaLBT+rtdgBZmXhrsFoMtCsciQnJxMbG8vx48cJCQmhb9++Mn2nEJWopDPv3FOq\n1UDhC1fOL3CJauWtt97ivvvuwzAMAgMDyczMdLqdTNNZs2RlZfH111+TlZWFUopevXr5bHlTX7Nx\n40YeeOAB2rVrh8Vi4cyZM7Ro0YKXXnqJgIAAkpOTeeGFFzh8+DB2u52mTZvy2GOP0ahRI8CcVezt\nt98mOzub9PR0hgwZwqhRowq0ceDAAW6//Xa6du3Kc889VySGgwcPEhMTU6RWee6EJ9ddd13espSU\nFCZNmkRqairZ2dk8+uijdOnSpcDzDMPg8ccf56mnnvL6aPM333yTNWvWYLVaefzxx+ncuXOB9fnL\nnO7Zs4ehQ4cyceJExo8fz8mTJ8nIyCA8PJz33nuPtWvXkpSUxLBhwzwSa7HJW2udW4Ozudb62fzr\nlFJFf6OiWjEMg5kzZ2IYBnXq1CEoKMjbIYkqIjAwkJ49exIUFERERIS3w6lRfG1WsY8//pjevXsz\nZswY9u7dS0xMTJE5D3xpVrFPPvkEML/gPPjgg4wfb/Y67t+/n2XLlpGQkEBUVBSA92YVU0o9DzQB\nrlFKtePfs29/oCfgfLJXUS388ssv/PXXXzRt2pSDBw+WWovalclQXOFvyyRhxSSn64LuaVPsOm+o\nCYPRipNb71pUrtwCHbmq+qxit9xyS16vTE5ODrVq1SrymnxpVrFc06dP56GHHiI4OJhjx45x+vRp\n7rrrLg4fPszEiRPp27cv4L1ZxRYD5wEDgDX8m7xzgP+6PRJRpSxZsgSAG264waVJJEqaDCVhxSSX\nB3bFXTuUqMudd43HXTtUus0rWWZmptMPXGGK+XYqB04num1/LeqdxctXTi5xG1+aVSx3Yo+jR4/y\n8MMP88QTTxRY72uzioE5mUtqaio9e/YEzC8lY8eOZcyYMaxZs4bnnnuOzp07ExYW5p1ZxbTWvwC/\nKKWWaK1Pub1lUS4vPvkddIKpMcXPLOaO6T6nT5/OZZddlldgX9Q8e/bsYd26dVx00UW0a9fO2+FU\nSaUlWk/wtVnFtNbExMTwyCOP0L179wLrfGlWsVxff/01I0aMyHvcqFEjRowYgZ+fH/Xq1ePcc89l\n7969hIWFeWdWMaXUZq11V+BEvulAcxla6+oxp5+Pye2a9vSUn1arlQEDBni0DVE1ZWZmEhcXx+7d\nu7FarXl3G4iqxRdmFdu9ezcTJkzgtddew0ke8alZxXJt2LCBO++8M+/x+vXr+fTTT3nvvffIyMhg\n165dtG3bFvDSrGKOxI3W2qUqbEII33fw4EHWrFlDamoq4eHh9OvXz+kHmPAOX5tV7JVXXiE7Ozvv\nWnm9evXyrm+D780qBmaPQ/56Bpdccgk///wzI0aMIC0tjZiYmLz3jCdnFbMUV9Ywl2Ow2oXA55hT\ngXYFJmqt13kkomIkJCQYuaP43LhPStvnoJivqlRt86kxS9l+wfJKuT1rwsqtpGXbStzGbmTw4cDe\nTI1ZWuo1b1fu0073C+TSJZ87XVddr3m78ndYGex2OwsXLuTUqVNERUXRpUuXAmc8VVlVOYa+zFvH\ncOe1kugAACAASURBVNmyZRw9epRbbrml0tt2t8LHcNy4cbz22msVmpzEsc8iQ/ldqbD2EfAGcA0Q\nCcQAL2EmdFGNpWXbeP+qbiVuYxZoce2bpSv3aQ+K+YpLXQ1QuJWfnx/9+/fHMAzCw8O9HY6oIWRW\nsfJxJXkHaa2/UEp9AMzVWq9VSrnyPOGDvvzySy688ELOOussgFKrpwX5yUjk6iS3mIcQlUlmFSs7\nV5JwjlJqGHA1MFkpNRgouS9V+KTjx49z/fXXY7FYSEpKAkqvnpaQkFAZoQk3O3HiBCEhIeWef10I\n4V2uJO87gQeAe7TWh5VSw4Fxng1LVKa33nqL5ORkfv/9d3Jycrj88stlkFI1Zbfb2b59O/Hx8URG\nRhIdHe3tkIQQ5VBq8tZab1NKzQT6KqUeAGZorbd5PjRRWV599VV2796d9/imm27K+9mVymnL5h7G\nhnm92rmL4fuveLTEbUwhbrhHXTh3+vRp1qxZQ2JiIsHBwU4LegghfIMr83mPBqYAXwF+wASl1DSt\n9Ycejk1Ukrvvvpvk5GQAwsPDC0xUUFLlNPh3dGVJo/JzR5vHXTunSo3cr0l27tzJzz//THZ2NhER\nEURHRxMcXHPLuwrh61zpNn8IuEBrfRxAKTUNs1yqJG83c+Us1x3V0wp78MEH3b5PUbUkJpolPPv2\n7Uv79u2dTiIhqj5fm1Us977nlJQUAgICeP7552nSpEmB5/nSrGL79u1jypQpZGdnU6tWLV555RXq\n16/P888/z6ZNm8jMzOTxxx+nV69e3ptVLB+/3MQNoLU+ppSSAWseUNpZbq7h80vuehaisN69e9Oj\nRw+PzG4kKo+vzSq2YMECOnXqxN13382SJUv44IMPitQ396VZxZ566ikeeughOnfuzIoVK9izZw+n\nTp1i9+7dLFy4kNWrVzNlyhSWL1/uvVnF8tmmlHoV80zbAowFtro9ElFpUlJSnNbsFdVXrVq1ZIKR\nasDXZhW7+eabsdvtABw6dKhAZbJcVXlWsdzKb2BOonLixAlWrVrFjBkz6NixI5MmTWLWrFl5z8ut\n9b5r1y7OOeccr80qlut2zGveszCvef8I3O32SAQAty6JITUrrcRt6gSW/g21uC74HFs2r35wO40b\nteL6gZMIDir+G2ELPNNNLzzn4MGDBAcH07BhQ2+HUiNsvu8B0vYfcNv+ardsQdc3Xi1xG1+aVQzM\n4j8333wzu3btYtasWQXWVfVZxVJTU/OS96lTp9i1axdPPfUUDz74IE888QRLlizh3HPPZdasWdx0\n0038888/7Nq1i/T0dADvzCrmaLgR0Ap4Rmv9sNtbF0WkZqW5pfRpcV3wX3zxBSmvJdMy4iyeffOG\nEq993r7811IrrImqIScnh40bN7Jjxw4aNWrEddddJ9e1K0FpidYTfG1WMYDZs2ezZ88e7rzzTlau\nXJm33JdmFatfvz516tTJm3ylX79+xMXFMXToUH777TdGjx5NaGgoHTp0yHtN3ppV7HrM0qhnAD+l\n1HCt9WqPRFEDeGswWmG53z7vuusu+XCvJpKSkoiNjeXUqVOEhoYSHR0tv9sawBdmFXv33Xdp0qQJ\ngwcPpnbt2litBSej9KVZxYKCgoiIiGDTpk10796dX375hfbt27N3716aNm3K559/zooVK3j99deJ\niIgAvDSrGPAU0ENr/YdS6j+YXed9PRJFDVAVBqPt3LmT2NhYateu7ZFuHFH5tmzZQnx8PIZh0KlT\nJ3r06JF3hiKqH1+bVWzYsGE88sgjLFq0CLvdXmT0uq/NKjZ9+nSmTp2KzWbj7LPP5uGHH8ZutzNz\n5kw+//xz0tLSePbZZ/P26clZxfIGQBT+FxkZuaXQ49+K27Yy/m3atMlwN1f2efXEL93S1jMTv3Zp\nu+vn3eWx9p5++mkDMG6//XaX9jFuWUKp2+Qew5KO06bvHzIMwzB+umaIS+0Wp6LPr6oq8re9c+dO\n47PPPjMOHTrkxoh8jyc+H2oabx3Db775xvjoo4+80ra7FT6GY8eONc6cOeOOfRbJiSV9RS88V2iO\nZ74+iMoyatQonnnmGcaPH+/ydJ+umrBnHnHXzil2fdxbQ/GX25TcLjIykjZt2kiNcuGzZFax8ikp\neYcopXL7Uyz5HlsAQ2u91iMRCY9p2LAhc+fOpWvXrrztwmC0skz3GWzPKna6z9wKa8L9LBaLJG7h\n82RWsbIrKXkfAp4p4XE/j0Tkg6rKYLRcmf9v787jY7r3P46/skpkQ+xLxZajtZRELbGF0qIoXSh+\n1FaqqoqothdFq1VB7622t1zVRYui1mpRvdYgiOVq1EFDYomEkMhkX87vj8lMM5JMElkmk3yej4dH\nzZztO19pPud8z5nvOyWFNZvn4fX4PUaMGGF8SMTT05Nhw4YV67GSFy8jKFl/hR707PO5ruM0uXGe\nywqjol65a5qGqqrY2NigKIqlmyOEKAPyLN6qqvqXYjusWkEfRistGzdu5PLVUyxdurTkH0xLTqbz\ntp8IevZ5s1feeS0T5iUmJnLw4EEiIiKoXLkyTZo0kQfShBAFmqRFWJlPP/0UgClTpshXhqxYWFgY\nhw4dIiUlhbp16+Lv7y+FWwgBSPEud4KDgzlx4gTOTm45AgcK6w0PF4KGvAgpmWbXC3r2eZJsHQnZ\nMzPX5Xb2kl5VWKdOneLkyZPY2dnh5+dHixYt5ERMCGFUYsVbURRb4AugNZACjFdV9a9c1lsJxKiq\n+k5JtaUiMVx1+7TsXeQnN51tbUhOyTQ75G0YLh8wYxs7npK4z+LSqFEjbty4QdeuXU0mihAVm7Wl\nihn89ddfDB06lCNHjuDo6GiyTLOiVLGQkBA+/vhjbGxs6NSpk3EKVsN2ycnJLFy4kNatW1s+VUxR\nlGrAx0BTYAiwGJiuquq9fDYdBDiqquqnKEoHYGnWe9n3PRFoCewvfNPFg9LS0ggJCcHW1pb2bfoV\n237NPmzm5FRsxxF/q1q1KgMGlJ3nKETZYG2pYgA6nY6PP/44z2Aca0oV+/jjj43zqg8fPpyLFy+S\nlpZm3G7Pnj0sWLCATZs2lYlUsf8Ae4AOQDz6p86/B57JZ7vOwC4AVVWDFUVpl32hoih+QHtgBdC8\ncM0WuXFwcCA0NJSTJ0+ye0N0se3X3JV3SEhIsR2notK0B6dUECJ3hgk6DMp6qpimacydO5fp06fz\n2mu551mV5VSxe/fumcy9XqlSJWJjY0lNTSUlJQU7OzuOHTtm3M7T09NkO0unijVSVXWFoiivqqqa\nDMxWFOV/BdjOHbif7XWGoii2qqpmKopSB5gLDAaGFr7Z4vbt2wwcOJBVq1bRokUL4/t2dnZ06NCB\n3Rt2WLB1oiAyMjI4deoU4eHhtGvXLv8NRJnz78D93L4VX2z7q1HbjUkz/c2uY02pYp999hndu3c3\nhps8qKyniul0OpPiPW7cOF599VWqVKlC8+bNady4MXv27MlzO4ulimVJUxTFGMKqKEozwPzUXHr3\ngeyh0baqqhqefHoBqA78AtQGKiuK8qeqqnlP0UXJXOUVZJ/Fsc6/wtaQnJmS736cbCvluy9N05g1\naxbHjh3j0KFDJCc/MBPa3c1AxzwfINMbls9y0JL1/8wDZuQz3/ra6zg52shVeCEkJSVx5coVkpKS\ncHR05NixYzLZShFZ4uevfU83TH/NFZ25z3Hx4kW8vb2ZMmUKOp2ODz/8kKSkJEJCQoiLiyM0NDTH\n9qdPnzbe8z5+/LhJoQkPD0fTNGOQRkhICFevXiUyMpKQkBBu3rxJQkICISEh2Nvb88cff3Du3Dl0\nOh379u3Dzs6Oy5cvA1CrVi2uXr1qcvyNGzfi6enJN998Q3R0NEOGDGHOnDnG5ffu3cPR0dG4TXx8\nPDNmzMDJyYl79+6hKArnzp2jatWqnD17FoBLly4REBAA6FPF6tSpw5UrV9i3bx+//vorzs7OJCcn\n5+iHwMBAUlL+/h1cr149xowZY9KWqKgo43a3b98mLCyM6Gj9KGZqaipz5sxh0aJFVKlShXXr1rFg\nwQIcHBzy3C4qKopr166VzM9mbnOmZv/j7e3dx9vb+7S3t3eMt7f3Nm9v72hvb+/+BdjuOW9v76+z\n/t7R29t7Zx7rvezt7f1Rfvsry3ObF2Te8uKas1zTNG3dunUaoDk7O2sXL17Msfzk7oB821SQecsP\nD3wu3/nEZU7pwsnIyNDOnj2rrVq1SluxYoW2f/9+LTg42NLNsnoV5efw2LFj2rRp04yvL1++rPn5\n+WnR0dGapmnaiy++qP33v/81Lj9w4IA2aNAgLTMzU7t+/br27LPPajExMZqmaZpOp9OGDBmiHThw\nQNO0v/vwp59+0pYsWaJpmqY999xz2oULFzRN07TffvtNe+2117Tr169rQ4YM0dLT07WePXtqd+7c\n0TRN095++21t8+bNeba9R48eWkpKisl7KSkp2lNPPWV83alTJy0hIUHTNE2bNWuW9umnnxqPZ/DC\nCy9okZGRmqZp2vHjx7U9e/Zo3333nRYYGKhpmqZdvXpVe/TRRwvWodn88ccf2ssvv6xlZmZqN27c\n0AYOHGiyPCEhQevevbsWHx+vaZqmff3119qyZctMttu1a5fJdmfOnNHeeOONQrclu4eZ2xwAVVV3\nKYoSgv7+tB0wQVXVqAKcF2wBeiuKEpT1eoyiKMMAV1VV//PgOUQhzjeKzaJNN0hee93sOq6lODNa\nQdy6dYvJkycD8M9//pNmzZo99L7ym/Us2dEGp1S5H1ucQkNDOXbsGM7OznTt2hUvLy8ZsRAFZm2p\nYg/u90HWlioWEBDA6NGjcXZ2xt3dnUWLFuHm5mbcLj4+noULFxr3aZFUMe3vK+P3vL2952b91/D3\nufltV9x/SuLMujQTw4rjyjszM1MbOHCgBmhPPfWUlpmZmet6xXXl/eL6V+XKu5ilpaVpQUFBWmJi\novE96cOikz4sOkkVK7rSTBWzzbe664NIDKc9jsCzQK2SOZUQ5qSkpFC5cmXc3d1ZtWqVTNphhezt\n7fHz88PZWSauEQL0qWLnz58nMTHR0k0pVpZMFQNAVdV52V8rirIA+K1EWiPMcnJyYt26dbwz5Tu+\n/ucZ4Ewea3Yr1SAUkbuUlJQ8v9sqhPibpIoV3sPMsOYG5Pw+gig1zo5VzQahSASnZaWkpBAUFER0\ndDTPP/+8PEUuhCh2BZlh7Uq2lzZAVUAqQzmgz+vOm4ujZWc8skbXr1/nwIEDJCQkUKNGDVJSUqR4\nCyGKXUGuvIcAt7P+rgGxqqrGlVyTRGnZMPTf+a4T9E3Rc7grgvT0dIKDgwkNDcXGxoZ27drRpk0b\nbG0L8liJEEIUTkGK9xpVVWX6UgvIzMzkjTfe4LXXXuOxxx6zdHOEGTdv3iQ0NJQqVarQo0cPatSo\nYekmCSHKsYIU7zOKoowCgoEkw5uqqkaUWKuszJ8+vzHkx1/MrvMwQ9D/+te/+Pzzz9m5cyeXLl0q\ncJbz1+nP8eUvp8yu45icxLDZv6BLSjO73tsFbm3F9sgjj9CzZ0+8vLwkc1uUGGtLFdM0jW7duhln\ncGvbti3Tp0832U6zolSx8PBw5s2bR1paGpUqVWLZsmV4eHiwefNm1q9fT3x8PIMHD2bChAmWTxUD\nOqIPJXlQo2Jui9XKsE8r0BB0QWzZsoUPP/wQgP/9Tz+F/KefflqogpBCJf7Tz8fsOkHPPs+ipqPY\nsdR8jGfQs2ZnrBXZNG3a1NJNEOWctaWKRURE0KJFC7PziltTqticOXMICAigdevW7Nmzh7CwMDw9\nPVm/fj3ff/89Z8+e5fDhw2RkZFguVUxRlJdVVf1WVVWvYj+qyNOdO3c4efKk8fX48eMfKhoyv4fR\nphZ6jwL0Vwm3b98u1MxNQhQXwwQdBmU9VSw0NJTo6GhGjRqFk5MT77zzDo0amV73leVUMcPMb6AP\nUbl37x6///47gYGBtGzZkpkzZ7JhwwZatmzJW2+9xZUrVwgICMDOzg7AYqlibwLfFvsRhVmDBg2i\nTZs2gD7is1WrVg+1n/xGAoLWyoNohRUfH8/+/fuJiopi8ODBeHp6WrpJwsJCg5aQnFCQ2aILxsml\nFi06B5hdx5pSxWrWrMnEiRN5+umnCQkJYebMmSZXs2U9VSwhIcFYvOPi4rh06RJz5sxh2rRp/OMf\n/2DLli3ExsZy4sQJfvzxR4KDg/nggw/YtGkTbm5uFk8VE6WoRo0a8rBTGaNpGqqqcvToUdLS0vDy\n8rL4EJ8oG/IrtCWhY8eOLFu2jNjYWMaOHUu9evUAqF27Njdu3Mix/tWrV/Hz8yM6Oppbt26ZLLtw\n4QKapuUo6oar+9u3bxvjPNu1a2cyXB8TE4ObmxseHvrQySeeeCLHsVu2bGm8CvX19TUmdBnExcWZ\nRG7WrFmTDz74ABcXF6KiovDx0d/+q1+/vvHWYVhYGPPmzQP03/Lw8vLCw8ODc+fOERwcjKurK6mp\nqTnaMnHiRJNZ3Jo2bWqcvxz0JzLZr8wTEhJwc/s7Mc7DwwMXFxfat28PQI8ePQgKCkJRFDp06GCc\n/bJJkyZcvXqVVq1aUb16dWJjY3O0pTiYK96PPfAd7+w0VVUbl0SDyqvFs3eRnM/DYQWVTv4xnQN+\nv2Z+J01HlbnQlbIoMTGRgwcPEhERgYODA/7+/jRr1kymphUWV6VKFQIDAxk1ahRbt27Fx8eHO3fu\nsG/fPnr06AHAwYMHuXbtGh06dKBBgwZMnjyZvn37Uq1aNRISEnjvvfeYPHmySfHOPixfs2ZNVFVF\nURROnDhhMuTt6elJfHw8MTExeHp6cvbsWWNhM/j888/x8PBg/PjxXLhwgbp165osr1q1qknBnDt3\nLnv37qVy5cq8/fbbxrZk/8plo0aNCAwMpHbt2pw4cYLY2Fi2bNmCu7s7CxYsIDw8nA0bNuToL0NY\nS158fHwIDAxk3LhxREZGkpmZaXIl7uTkhJeXFydPnqRdu3YcP36cZs2a4ePjww8//EBqairJycn8\n9ddfNGzYEID79+8br9yLm7nifRnox9/zmosiSE5Ky3VWtFWrVlG1alUGDRpkPEM155VfTnHr92tm\nHzQb8uOk/IfNn32eztt+yr/hFVxGRgaRkZHUq1eP7t27l8iDJ0IUlLWlik2YMIGZM2dy4MAB7O3t\nczy9bm2pYgsXLmTBggVkZGRQv3593nrrLezt7XnhhRd46aWXSEhI4LXXXsPd3R0o2VQxm+xnWdkp\ninJaVdW2JXLUhxASEqL5+voW6z4HzNiW79PWBVGQYrlgxo4cxdvwNY/Y2FiCg4NznLXmxlC8ndvv\nynMdF8fKfD14aZ7LoeDFO7/1QkJCKO5/l7Lm3r17xgdySkJF6MOSJn1YdJbqw507d3L79m1Gjx5d\n6scubg/24fjx4/nXv/5VpHCSrH3m+OVj7so7yMwyUQxWr15NbGwsfn5+BSrc2RXXV9NE/rLfkxNC\nFK9nnnmGt956i8TExHL1LInFUsVUVX29RI4oAP2DFv/85z8BCAgo/YdehKn09HRUVeWxxx6T+9lC\nlDJJFSs8edrcQmbNmsWVK1do0qQJAwcOtHRzKrTo6Gj27dtHXFwcjo6ONGvWzNJNEkIIs6R452PM\nlhkkpJoPibdLL/xT287OzgC89dZbxgfVpv52lit7I9DSM81uW5CnxINHvEy6Tmd2naBn8/+ut305\nfkArIyODU6dOcebMGTRNo1WrVjkmkBBCiLJIinc+ElITC/QwGiMKt9//+7//o1evXiZDK4lpGWjp\nmWYfoitoVne6Tmf2QbOK/rR5QkICu3fv5s6dO7i6uuLv75/jayxCCFFWSfEuYRkZGXzzzTekpnmY\nvN+8eXPj5Aei9BlGPhRFoVOnTjg6Olq4RUIIUXBSvEvY5s2bGT9+PPVqe/PBpyWTLiMKz9bWlgED\nBuDgIBPVCOthbalihilUQ0NDSUtL44033sjxvXJrShXLPs1pWFgYzz//PNOnT+eTTz7h6NGjJCQk\n8P777+Pj41MmUsXEQ9I0zZgQ1qZFz3zWFqVNCrewNtaWKrZt2zYyMjJYt24dUVFR/PrrrzmKtzWl\niq1ZswbQn+BMmzaNSZMm8ddff3Hs2DE2bNjAzp07+eCDD9i8ebPlUsUqAief3xnyY96TncDD5XAb\n/Prrr5w5c4batWvTtkWvh96PeHiJiYkcP36cjh074uTkZOnmCFEk1pYqFhQURLNmzZg4cSKapjFn\nzpwcn6ksp4rdu3cv13keFi5cSEBAAM7Ozjg6OpKUlERqaiqJiYkmt+AslSpW7tkUYw73gzRNY+HC\nhYD+zDgx0jHfmE4Pt1cA83Ges6rmfwYXPOJlwPzT5OX5KXKDsLAwDh06REpKCh4eHrRtW2YmDBTl\nxHsHz3NTl1xs+6vr6sT8bo+ZXceaUsXu3btHREQEK1as4MSJE7zzzjt8//33xuVlPVVMp9PlKN4X\nLlwgISGBjh07GvvX29ubPn36EBsby6JFi4zrSqqYFTpx4gRHjhyhatWqvPrqqyx7b1++Jwqv/HIK\nMD97Wsiemfke2/AVsYr6NHlKSgpBQUFcvnwZOzs7/Pz8aNGihaWbJcqh/AptSbCmVLEqVarg7+9v\nXH716lWT5daUKmawfft2hg4dany9Y8cObGxs+P333zl8+DCLFy/m8ccfp1atWhZLFRNF8MQTT7B3\n714iIyMlzKIUpaamsmnTJhISEqhRowY9evQwOZsWorywhlQxX19fDhw4wFNPPWX1qWIGx44dY+LE\nicbXSUlJxlAWJycnHBwcSEpKAiyXKiYKKCMzg5SUFCpVqmR8z8bGhieffNKCraqYHB0dadKkCQ4O\nDrRt29bkf3ohrJ21pYq9+OKLzJs3z3ilOn/+fJPl1pYqBvoRB8NoA8DgwYM5deoUL730EvHx8bz4\n4ot4eXkBFkoVK2tKIlWsIGlgBTFqyELW/jSPls278lyfabmuk47GafLvayf7NN5+MjjP5Xb2zrTp\nucDsPgz3uktj2FzSnIpO+rDopA+LTlLFiq6spIqJArp16xIZmel09n8018xu0J8o7CjAPe//9PMB\n5PvgBWG4AhBCWC9JFXs4UryLwY1bl4DcH9gQJePu3bscOHCATp06Ubt2bUs3RwhRBJIqVnhSvIuB\nFO/So2ka586d48SJE2RkZHD9+nUp3kKICkeKdxFFRUURF38bV1dXmau8hMXHx7N//34iIyNxdnam\na9euxgdDhBCiIpHiXUTh4eFUdnbH17etMdpTFL/MzEx+/vln4uPj8fLyomvXrsZwESGEqGikeOdj\n8exdJCelmV3nrVfX8OZ7JXt/w6AgOd32rq75rmNtbG1t8fPzIyUlhWbNmsmDakKICk2Kdz6Sk9Ly\nfILcYMGMHSbf+ytJ+eV0G5ibGtVaNWzY0NJNEMKirC1VbOXKlRw6dAjQT1gSExPD4cOHTbYrD6li\noB+FnTVrFnv37gWQVDFR8aSmpmJvby8TrAjxAGtLFZswYQITJkwA9MEgue3T2lPFALZu3cqaNWuI\nj483riupYqJCuX79OgcOHKBFixa0adPG0s0RokyxtlQxgz179uDh4ZHrbGPWnioG+qlqv//+e+M8\n7gaSKlYOTNx0m6C15oeyRwNB5qffBQo2JG5tqWHp6ekEBwcTGhoq97OF1Zgc+F8ibsXnv2IBPVLb\njc9n9jS7jjWlihmsXLmSTz75JMf75SFVDMhRtA0kVayMOnnyJC4uLmRqmfmu65Sq5Xuv+pVfTjF6\nxUKz69i7utLhh28L1c6yLjo6mn379hkThvz9/alRo4almyVEvvIrtCXBmlLFQF+Q3d3dcz2xKA+p\nYuZIqlgZNWPGDA4ePMiIQXOBZ4tlnxUxxvPEiRPExcXRqlUrnnjiCeP/pEKIvFlDqhjAkSNHcoSf\nGJSHVDFzJFWsDMrIyCAkJASAerWbWbg11q1bt27Ex8fniAsUQpiytlQx0F/5d+7cOdfPUx5Sxcwp\nyVQx4wMQZf3PyZMnteL24vpX811n/vTtub5/7tw5DdAaNmyY5zrZHR74XL7rjN8Zku86ZU1J/LtU\nNNKHRSd9WHSW6sOff/5Z+/rrry1y7OL2YB+OGzdO0+l0xbHPHDVRrrzz0S1sLUHPfpPj/Z0RVwFo\nnJYO5P8QWUolJ1755ZTZdSqR8lBttBb379/Hzs6uxFJ2hBDWR1LFHo4U73w4ZKbmeh/6+0mT4Mwp\n+r45laRb8K/h5odnPNxeyYr7zFvInplAp6I0t0zSNA1VVTl69Ci1atWib9++8kS5EMJIUsUKT4r3\nQ2revDldunTBz8+P3zffZUMBsrorosTERA4ePEhERAQODg40bdrU0k0SQgirJ8X7IU2dOpWpU6cC\n8PvmHRZuTdl05coVDh48SEpKCnXr1sXf379EZhoSQoiKRop3ARw5coS9e/fSr18/2rVrZ+nmWI2Y\nmBjS09Px8/OjRYsWMlQuhBDFRIp3AWzbto3FixeTmpqaZ/HW3682Z1i+62jJGQ/ZwrLJx8eHZs2a\nlVpoixBCVBQlVrwVRbEFvgBaAynAeFVV/8q2fBgwFUgHzgGvqaqq5bYvSwsODgYwmQ7vQb5PBZrf\nyS+nSP48zOwq9q6uMLDQzSuzbG1tpXALUYysLVUsKSmJ6dOnEx8fj4ODA4GBgca2GGhWlCoWEhLC\nxx9/jI2NDZ06dTKZgrU8pYoNAhxVVfVTFKUDsDTrPRRFcQbeB1qqqpqsKMpaoD9Q5m4eZ+i/Yw5A\nhw4dirSv8jp7WnR0NOnp6TLJihAlzNpSxbZu3UqTJk0ICAhg48aNfPXVVzn2a02pYh9//LFxXvXh\nw4dz8eJFvL29y12qWGdgF4CqqsGKomQfb04GOqmqmpytHUkl2JaHduX+fRISEmjUqJHMt/2AjIwM\nbty4walTp6hcuTJDhw6VqU2FKEGGCToMynqqmJOTk3Fub8PV94OsKVWsUqVKxMbGkpqaSkpK3ntH\nFwAAIABJREFUCnZ2dkD5SxVzB+5ne52hKIqtqqqZWcPjtwEURZkCuKiqurcE2/LQzsfeBYp+1V3e\n3L17l3379hETE4Orqyv+/v5SuEWFM+PXBVy7H1ls+2vgXoelfeeaXceaUsV69+7NypUreeaZZ4iL\ni+OHH34wWW5tqWLjxo3j1VdfpUqVKjRv3pzGjRsD5S9V7D6QPZLFVlVVY/xW1j3xxUBTIP+MSzDO\nJV6c8ttn2uNj6FPVB/uU+iyYkfuofjpaAdpmUyLtt4Q7d+4QERGBpml4enrSoEEDIiMjiYwsvl9i\nFU15+dmwJEv04fCaz0DBp88uEHOfwzBMO2XKFHQ6HR9++CFJSUmEhIQQFxdHaGhoju1Pnz5tvM98\n/PhxkwIVHh6Opml4eXkZj3316lUiIyMJCQnh5s2bJCQkEBISgr29PX/88Qfnzp1Dp9Oxb98+7Ozs\nuHz5MgC1atXi6tWrJsf/z3/+Q8+ePenZsycRERGMHz+eRYsWGZffu3cPR0dH4zbx8fHMmDEDJycn\n7t27h6IonDt3jqpVq3L27FkALl26REBAAKBPFatTpw5Xrlxh3759/Prrrzg7O5OcnJyjHwIDA0lJ\n+XsWy3r16jFmzBiTtkRFRRm3u337NmFhYURHRwP6UY45c+awaNEiqlSpwrp161iwYEGO0Ybsx42K\niuLatWsl8rNZksU7CBgAbFQUpSPwvweWr0A/fD64oA+q+fr6Fm8LL69i59qbZlepUt0Lt37nSSGK\nP/gl13WSjvfh6QUfmj/WxH8Uf/st5MaNG0RHR9O1a1diYmLKzeeylJCQEOnDIqoofZiens7p06eN\nn7V+/fqMGjWKQYMG4evry5YtW7h//75JqphOp2PkyJHcvHmTyZMnM3HiRGOq2OLFi5k8eTK+vr7G\nPgwPDyczMxNfX1/q1auHq6sriqKwd+9eHn30UVq1aoWrqys9evTgo48+wsvLC09PTzZt2kT79u1N\n/h3Wrl1LixYt8PX15ZFHHmH58uUmy1NTUwkMDDS+N2XKFJNUsdq1a9OqVSvc3NyM6zRt2pTly5eb\npIpduXKFxo0bExAQQHh4OH379s3x87B+/Xqzfevk5ERgYCA+Pj5ERkbi5ORkMktaYmIidnZ2dOjQ\nAVdXV86dO8e9e/dyHCf767Nnz9KwYcMi/WzmVfhLsnhvAXorihKU9XpM1hPmrsBJYCxwEPivoigA\n/1JVdWsJtidXc5cOMLt8wYwd+c6eNuD4tnwfRvumHM2wVq9ePYYNG4a9vT0xMTGWbo4QFYa1pYpN\nnz6dOXPm8MMPP5CRkcHChQtNlltbqlhAQACjR4/G2dkZd3d3k1GE3EiqWBlMFcuu//St+a5jjYlh\nBSFpTkUnfVh00odFJ6liRVeaqWK2+Zd3UVGFhYVx+vRpSzdDCFGOPfPMM5w/f57ExERLN6VYSapY\nCZq46TZBa/N5Vq7p6FJpS1mSkpJCUFAQly9fxt7enubNm+Ps7GzpZgkhyilJFSu8Cl28nVK1PO9V\na5pGly5diFeX8OZ73XF3dy/l1lnG9evXOXDgAAkJCdSoUYMePXpI4RZCiDKmQhdvc65fv86RI0dw\nquRSYZKwLly4wMGDB7GxscHX15e2bdtiayt3VoQQoqyR4p0Hw3zm9Wp7V5gC1rBhQ+rUqUPHjh1l\nNjkhhCjDpHjnwVC869f2tnBLSo+zszMDBpj/6pwQQgjLq/DFO6+Yzt936b/QX7+OdwHiPrvks7zs\n0bK+qymEsB7WlioWFxfHrFmzuH//Ps7Ozrz//vs5Aow0K0oVyz7NaVhYGM8//zzTp08HyleqmFXI\nLcozIyODi1f0E7PUq+2N71P/Z34nu7eVRNNKRGZmJn/88QcRERH069evwtwSEKI8sLZUsS+//BIf\nHx8mTJjA0aNH+eCDD/jiiy9M1rGmVLE1a9YA+hOcadOmMWnSJIBylypmtezs7IiIiODs2bMc+lln\n6eYUm/j4ePbv309kZCTOzs7cv3/fZJ5jIUTZZpigw6Csp4r99ddfTJs2DYC2bdvyxhtv5PhM1pQq\nZrBw4UICAgKM38Qpb6liViMtLQ17e3uTM8dq1arRo0cPDv1c5iLGC03TNFRV5ejRo6SlpeHl5UXX\nrl3lK2BCFNHpKW+SGHGt2PZX+ZEGtF3+T7PrWFOq2KOPPsrvv//Oo48+yn//+1+SkkyTn60tVQz0\n38pJSEigY8eOxvfKW6qY1Vi6dCnvvfcec+bMYfbs2ZZuTrG7evUqBw8exMHBAX9/f5o1ayb3u4Uo\nBvkV2pLQsWNHli1bRmxsLGPHjqVevXoA1K5dmxs3buRY/+rVq/j5+REdHc2tW7dMll24cAFN03IU\ndcPV/e3bt2nevDkA7dq1Mxmuj4mJwc3NDQ8PDwCeeOKJHMeeMGECH3zwAf/3f/9H9+7dqVOnjsny\nuLg4k+JYs2ZNPvjgA1xcXIiKisLHxwfQB7AYIofDwsKYN28eoA9q8fLywsPDg3PnzhEcHIyrqyup\nqak52jJx4kSTWdyaNm1qnL8c9Ccy2a/MExIScHNz40Hbt29n6NChOd7PTfXq1Y155sVNijdw69Yt\nUlNTi3TP5ZV8gkcqO9g99L6LysvLi8cff5wWLVpUmO+sC1HeValShcDAQEaNGsXWrVvx8fHhzp07\n7Nu3zyRV7Nq1a3To0IEGDRowefJk+vbta0wVe++995g8ebJJ8c4+LF+zZk1UVUVRFE6cOEGjRo2M\nyzw9PYmPjycmJgZPT0/Onj1L+/btTdp44sQJhgwZQtu2bdm9ezft2rUzWV61alWTgjl37lyTVDFD\nW7I/m9OoUSMCAwNNUsW2bNmCu7s7CxYsIDw8nA0bNuToL0NYS158fHwIDAxk3LhxREZGkpmZmett\nxWPHjjFx4kSz+zK4f/++8TZFcavwxTvo2ec5d1L/tbC4H9YTdCDIdIWmowl6Nr8pVEfxn34+JdTC\norOxsaFDhw6WboYQooisLVWscePGzJo1C03TqFKlCh9+aBqdbG2pYqAfcTCMNuRHUsVKKFXs8MDn\nNE3TtG7dummA9vvvv+dYp7hSxUpLUlJSqR5P0pyKTvqw6KQPi05SxYpOUsVKmeE+UO3atS3ckoeX\nnp5OUFAQP/74o8kwlBBClGWSKvZwKvywOegfmgByPExhLaKjo9m3bx9xcXFUqVKF1NTUEvuBEUKI\n4iapYoUnxRuIjIwkPj4+1ycLy7KMjAxOnTrFmTNn0DSNVq1a8cQTTxifyhRCCFE+yW959A9DWGPk\n57179zhz5gwuLi74+/vnmHZQCCFE+STF24pVr16d3r17U7duXRwdHS3dHCGEEKVEireV8/LysnQT\nhBBClDJ52twKaJqWY2YkIUTFExwcTKdOnRg5ciSjRo3iueeeY+rUqaSlpQFw9+5dZs2axciRIxkx\nYgQzZszgzp07xu1PnjzJ2LFjGTlyJC+88AJr167NcYxr167Rp08f3nnnnVzbcP369VxnGPvqq6/Y\nsmVLrtv89ttvJtOnnjlzhiFDhjBs2DA+++yzXLe5evUqy5Yty7szSsndu3cZO3YsI0aMYNq0aSQn\nJ5ssP3ToECNHjmTkyJG8//77PPbYY4SFhRmXHz9+3GT61AULFhATE1PkdlX4K2+dTkelSpVwcHDI\nc50BM8ynhtnYl9w5UGJiIgcPHiQiIoI+ffrwyCOPlNixhBBlm7WlioF+opegoCAee+wx43vz5s1j\n+fLlNGjQgAkTJvDnn3/mmKJ18eLFOSZ1sYQvvviCgQMHMmjQIFauXMn69esZPXq0cXnXrl3p2rUr\nAPPnz6d79+40btwY0D8M/fXXX5Oenm5cf+TIkSxdurTIn63CF+9p06axatUqVq9ezZgxY3JdZ8fS\nZ83uI7+pUR9WWFgYhw4dIiUlhXr16pXYNHtCCOtgmKDDoKynioF+2tHevXsb8791Oh2pqanGEJUu\nXbpw5MgRk+IdFhZmMj3p0qVLCQ0NJTY2FkVR+Oijj1i+fDmnT58mMTGRhQsXcuTIEXbu3Anovzs+\ncuRILl68yMcff2xMCJs3bx5t27Y1HickJIR//tN0fvqxY8cap5cFOHXqlDH6s1u3bnzyyScmxdvg\n1q1bHD58mF9++QWAlJQU5s2bx4IFC3j++b9n6WzUqBFhYWHExsYWKdWxwhfvyMhIgDJVGFNTUwkK\nCuLSpUvY2dnh5+dHixYtJExEiDLm34H7uX0rPv8VC6hGbTcmzfQ3u441pYoB9OvXj+DgYONrnU5n\nclwXFxeuXTNNZjtx4gSKohjX9/DwYPXq1WRmZtK/f3+ioqKwsbGhadOmvPvuu1y+fJlff/2VdevW\nkZmZydixY+nSpQuXL19m1qxZeHt78/PPP7N582aT4u3r62vM6M6LTqczfo3YxcXFJLM7u6+//pp+\n/foZR3EXLFjAuHHjqFWrVo51GzVqxKlTp0xOtAqrwhfvsji7mqZp3Lx5kxo1atCjRw/J3BaijMqv\n0JYEa0oVy82D6V06nS7HV3VjY2OpXr06oB8hiImJYcaMGVSuXJnExETjMLThgd2LFy9y8+ZNRo0a\nBUB8fDzh4eHUrFmTL774AicnJxISEnKcrOR25T1mzBiTourq6opOpzOGueT2teLMzEz279/P/Pnz\nAYiKiiIkJISIiAjj55kxY4ax/2rWrFnktDEp3mWweFeqVIn+/fvj5uZmkqYjhBAG1pAqlhtXV1cc\nHBy4du0a9evXJygoiNdff91knWrVqhEdHW38DLdu3eKTTz7h7t27/PbbbznSxho3bkzTpk1ZtWoV\noL8KVhSFyZMnExgYSJMmTVi+fHmOk5uCXHn7+Phw4MABBg8ezMGDB3Mko4H+5KFx48bGq+5atWqx\na9cu4/IuXbqYnPjExcXh6emZb1+ZU6GLd6amERUVBZSt4g0UOLVGCFFxWFuqWF7tnj9/PgEBAWRk\nZNClSxdat25tsn6HDh1YuHAhAI8//jj//ve/GTVqFDVq1ODxxx83FnbDPps3b06nTp0YNmwYKSkp\ntGnThlq1ajFw4EDefPNNateuTcuWLY0pZYUxadIkZs2axYYNG6hWrZqxCAcGBvL000/TunVrrl69\nWqiHif/8809mzpxZ6LaYyC2tpCz+KYnEm1/69NeqV6+uVatWLc91CpIqNn5nyEMdPz09XTt37pyW\nnp7+UNuXBZLmVHTSh0UnfVh0Za0PJ06cqN25c8fSzSiUgvThpUuXtNmzZxd2n5Iqlp27oyO3b982\nXn2Xprt377Jt2zaOHDnC2bNnS/34QghRls2cOZOvv/7a0s0odt9//z1Tp04t8n4q9LC5QWkGeWRm\nZvLHH39w4sQJMjIy8Pb2pmXLlqV2fCGEsAZNmjQhICDA0s0odvPmzSuW/UjxLkWpqans3r2byMhI\nnJ2d6dq1q0xvKoQQotCkeJciBwcHHB0d8fLyomvXrjg7O1u6SUIIIayQFO9SZGNjw5NPPomdnZ1M\nuCKEEOKhVegH1mKSk7l//77J9xpLmr29vRRuIYQQRVKhr7znhRzntIcHv/32G7169Sq2/aakpHDs\n2DHatGkj39cWQhSb4OBg3nzzTZo2bYqNjQ06nY4GDRqwZMkSs+FK+VmwYAGNGzfG19e3WNp55MgR\nVq5cSWpqKnZ2dtSvX59//OMfOWY4Ky5btmzBw8OjSNONWpsKXbzvpqQAxTtBy/Xr1zlw4AAJCQnY\n2dmZpPUIIURRmEsVe/rpp4u03+Jy4cIFlixZwpdffknNmjUB+Oabb1i1ahVvvvlmsR0nu8GDB5fI\nfsuyCl28Y1L0uazFUbzT09MJDg4mNDQUGxsb2rVrR5s2bYq8XyFE2ZVbHjbA8OHDi2X9Bxkm6DDI\nniqWmZnJnDlzuHXrFrdv36Znz568+eabvP322zg6OnLjxg1u377NokWLeOyxx1i3bp1x1rCkpCQa\nNWpEWloa77zzDtevXyczM5PRo0fTr18/Ro4cSfPmzbl06RKVK1emXbt2HD58mPv377N69WqT+b7X\nrVvHa6+9ZizcgEkKV8+ePdm1axeOjo4sWbKEJk2aMHjwYJYuXUpISIjxuH369OGHH35g27Zt2Nra\n0rJlS2bPns2ePXtYtWoV9vb21KxZk08++YTPPvuMGjVq0LhxY1auXImjoyPXrl3jmWee4dVXXyU8\nPJy3334bBwcH6tWrx/Xr1/OdFrWsq7DFOykpCV1aGna29nz2YRC2Nrnf/k9Hyzfys7K9Ldu2bSMm\nJoaqVavSo0cP46T6QghRnPJKFbtx4wZt2rThxRdfJCUlhe7du/Pmm29iY2ND/fr1WbBgARs3buTH\nH39k6tSpfPvtt+zYsQM7OztGjhwJwI8//kj16tVZsmQJCQkJPPfcc3Tq1AnQT1P6j3/8g/Hjx+Ps\n7Mzq1at5++23OX78uMltxxs3bhinCr127RrvvvsuoE8he/DkxXDFf+DAAW7cuMHatWtJSUlh6NCh\ndO7cmS1btjBv3jxatmzJunXryMjIYOfOnYwfP56nnnqKrVu3otPpTEYOIiMj2bFjBykpKXTt2pVX\nX32VxYsXM2nSJLp168bGjRu5fv16yf0DlZIKW7wNs6rVqVubecvyzuseMGMbO/r55Lu/c+fs0Ol0\nPPHEE6U66YsQwnIKesX8sOvnJq9UMQ8PD86dO0dwcDCurq6kpqYatzEEj9SuXZtTp04RHh5OkyZN\njPfJfXz0v+PCwsLw8/MD9PGXTZo0McZ1PvbYYwC4u7vTtGlT49+zHwf086lfu3YNb29vGjRowJo1\na0hNTaVPnz45PothFOHSpUuEhoYaTyIyMjK4ceMGH330EatXr+b69eu0adMGTdN45513WLFiBWvW\nrKFx48Y5nlfy9vbG1tYWZ2dnnJycjJ/L8Bl9fHzYvn174Tu+jKmwT5vrdDo83GrQsGHDYtlfq1at\n6NSpkxRuIUSpMKSKzZ49m9u3b7N582bc3d1ZsmQJY8aMITk5Occ2hmLZsGFDLl++THJyMpqm8b//\n/Q/Qz2p28uRJQP878uLFi9SvXx8o+H3xl156iX//+98mISBHjx41JoBVqlSJ6OhoNE3jzz//BPSp\nYB06dGDNmjV8/fXXPP300zRo0IANGzYwf/581qxZw/nz5zl9+jQ//vgjU6ZMYc2aNWiaxt69e02O\nn1s7mzVrxqlT+hHU8jIddYWtNC1btmT6K18xd+mAQm2nZSXqCCFEaTOXKjZlyhRmzJhBaGgodevW\npWXLlsYRRsM2hv9Wq1aNSZMmMXz4cDw8PIwXHUOGDGHOnDkMHz6c5ORkXn/9dapVq1aoNrZo0YK3\n3nqLt99+m7S0NJKSkqhduzaffvopAOPHj2fChAnUq1ePKlWqYGNjQ8+ePTl+/DgjRowgMTGR3r17\n4+Ligre3N8OHD8fFxYXatWvz+OOPo9PpmDhxIi4uLri4uODv78/333+f4zNmN3PmTN59911Wr16N\nm5tbkZ7MLzNySyspi39KIvGmIIlh/advNf49Li5O2759u3bp0qVib4u1KmtJRNZI+rDopA+Lrjz3\n4fbt27Xw8HBN0zRtw4YN2rvvvlsixymJPswrVazCXnkXhqZpqKrK0aNHSUtLM7nnI4QQomyrU6cO\n06ZNw9nZGTs7O2NWuDWT4p2PSnYZ7N69m4iICBwcHPD396dZs2aWbpYQQogCateuHT/99JOlm1Gs\npHjnw7duHBERd6hbty7+/v4lNkOQEEIIUVAVtnirqkq87i6ZmZnGpyBzE3rbjbeHNadFixbyoJoQ\nQogyocJ+Vax3794sWTma8PBws+vFJTvQsmVLKdxCCCHKjApZvDVN49atW8DfU6Omp6eTlpZmyWYJ\nIYQQBVJiw+aKotgCXwCtgRRgvKqqf2VbPgCYA6QDq1VVXVVSbXnQ3bt3SUtLw9XVFWdnZ6Kjo9m3\nbx9169ala9eupdUMIYQolODgYF5++WWWLVtGv379jO8PGDCAli1b8tFHHxXr8TZv3synn35KgwYN\nAIiPj8fHx4e5c+cC+ulPFy9eTGxsLOnp6TRv3pyAgABcXFwA2Lt3L9999x2appGSksK4ceNyDVDZ\nvn07zs7O9O7du1jbX1hnzpzhww8/xM7Ojs6dO/P666+bLF+5ciWHDh0C4P79+8TExHD48GF+++03\nFi9ejKurK66urkydOpVWrVrx3nvvsWjRohJpa0ne8x4EOKqq6qcoSgdgadZ7KIriACwD2gGJQJCi\nKNtVVY0uwfYYGa66PT09OXHiBGfOnEHTNB555BGZhEUIUaY1btyYnTt3Gou3qqq5zqZWHGxsbBg4\ncCDTp08H9KOWw4cPJzQ0lCZNmvDaa6+xcOFCWrduDcDWrVuZMWMGX375JadOneLbb79l5cqVODs7\nExsby9ChQ2natClNmjQxHiMxMZFt27bx1VdflchnKIx58+axfPlyGjRowIQJE/jzzz+NU8sCTJgw\ngQkTJgDw6quvMmvWLAD++OMPZs6ciaenp0msatu2bdm6dSuDBg0q9raWZPHuDOwCUFU1WFGUdtmW\nPQpcVlU1DkBRlMNAN2BTCbbHyFC8nZ2dOX36NK6urvj7+1O3bt3SOLwQopzI60Rfy5b8VZT1c9u+\nefPmXL16FZ1Oh6urK9u3b2fAgAFERkYC8Ouvv/Ltt99ia2uLr68vM2bM4NatW8ybN8+YQjZ16lR6\n9erFgAED6NChA6qqotPpWLNmTY5v1GRvm06nIz4+Hjc3N/bv30+HDh2MhRtg0KBBrFu3juvXr7Nx\n40ZefvllnJ2dAf10rps2bcLNzc1k/zt27DCOeOp0OmbPnk18fDzR0dEMHz6cYcOGMXLkSDw9Pbl/\n/z5ffvkl8+bNIyIigszMTN58803at2/Prl27WLt2Lenp6djY2PDZZ59RtWpV43F++OEHdu3aZdKX\nH3/8MXXq1DEeOzU11TjK0KVLF44cOWJSvA327NmDh4eHcR748+fPc+HCBW7duoWfnx8BAQHY2dnR\nt29fxo8fb3XF2x24n+11hqIotqqqZmYti8u2LB7wKMG2mMjMzKRhw4ZUr14dRVHo1KkTjo6OpXV4\nIYQokqeeeoo9e/bw3HPPce7cOV555RUiIyOJi4vjs88+Y/PmzVSqVIm33nqLI0eOYGNjw9ixY2nf\nvj2nT59m+fLl9OrVi4SEBPr378/s2bMZO3YsBw8eNBmO1zSNn3/+mTNnzhAdHY2rqyuTJk3ikUce\nYc+ePcZ5z7OrX78+N2/eJDo62lgIDR4s3ADHjx/nhRdeACAiIoJnnnmG3r17ExUVxahRoxg2bBgA\n/fv3p1evXqxdu5Zq1arx4Ycfcu/ePUaOHMnPP/9MeHg4K1euxMnJiblz53L48GEGDPh7+usRI0Yw\nYsSIPPvUcDJk4OLiYgxledDKlSv55JNPjK87d+5Mr169iIqKYtu2baxfv54RI0bg7u7OvXv3cuy7\nOJRk8b4PZP+XMhRu0Bfu7MvcgHv57TAkJKRYGlatWjU2bNjA+n3XWLI9Frb/mue6To42xXbc8kr6\np+ikD4vOEn1oCPF4UF5tKez6D1JVlbt379KwYUNWr15NYmIi9erV49KlS9y5c4fdu3cTHR3NSy+9\nBOijjw0XKVu3bmXFihXY2Nhw7949QkJCSE1NJTk5mZCQEDw9PVFVlVq1ahmPFx4eTrt27XjppZeM\nWeCJiYmEhISQmJjIX3/9RatWrUzaeP78eWJiYnB0dOTQoUPodDqT9lepUiXHMaKioggJCeHu3bv8\n+OOPrF+/HmdnZxISEggJCUGn0xmPGxQUhKqqBAUFAZCQkMD+/fuJj49nwoQJODk5cfPmTapWrWrS\nr7t37+bEiRMmbZ00aRKenp6Afvj+zp07xm3+/PNPMjMzc/zbGOJEo6OjiY7W3+lt0qSJcR75hg0b\ncvDgQZo3bw6Ak5MThw8fpkaNGgX6Ny6okizeQcAAYKOiKB2B/2VbdgFopihKVSAB/ZB5YH47zH4v\noTjY2dkV+z4rmpCQEOnDIpI+LLqK0ofp6emcPn2avn37smbNGo4fP86MGTMIDw/n0qVL9OrVi/Xr\n17Nx40bs7OzYtGkTLVu25PPPP2fMmDF069aNn376ia1bt+Lr64ujoyM+Pj44Ojqyfv16vLy8TPrR\nMDRteM/Dw4PZs2ezc+dOmjdvzpAhQ3BwcDAOnW/cuJEGDRrQt29f6tSpw9KlSxkxYgTOzs7ExMTw\n7rvvsnz5cry9vY3HaNSoEfXr18fX15dFixbx5JNPMmzYMI4dO8b58+fx9fXF1dWVVq1a0ahRI86f\nP0/r1q2ZOHEiOp2O1atX065dO+bOncuBAwfIzMxk7NixNGzY0OSzFOTnw93dnZo1a1K/fn1WrFjB\n66+/bnJbACA0NJT+/fsb96dpGk8++aTxdsHt27fx9/c3Ls/MzKR79+7G2weFldeJXUkW7y1Ab0VR\ngrJej1EUZRjgqqrqfxRFmQ7sRv91ta9UVY0swbYIIYTVy54q1q9fP7Zv307Dhg2JiIgA9KOKo0eP\nZsSIEWRmZlK/fn369+9Pnz59WLx4Md999x1t2rQhNjY2z/2be69Tp074+fmxfPlyZs6cyZdffsmH\nH35IbGwsGRkZNG/enGXLlgHQpk0bhgwZwpgxY3BwcCA5OZmAgACTwg3Qvn17zp49S7t27ejRowcf\nfPABe/fupWnTpri4uOTICx86dChz5sxh5MiR6HQ6hg8fjqurKz4+PgwdOpRq1arRqFEjk0jSgpo/\nfz4BAQFkZGTQpUsXY+EeN24cK1aswN7enqtXr9K5c2eT/lm4cCFTpkwhNTUVHx8fhgwZAuifSHdz\nc3vowm2OTUEflLC0kJAQrbjPrCvK2XpJkj4sOunDopM+LDpL9WFCQgKTJ0/mm2++KfVjF7cH+/CH\nH37A3d3d5N77Q+4zx1lVhZykRQghRNng4uLCoEGD2LNnj6WbUqySk5M5ffp0kQq3ORUIaRrSAAAI\nOUlEQVR2bnMhhBBlQ0l8lcrSnJycWLJkSYntX668hRBCCCsjxVsIIYSwMlK8hRBCCCsjxVsIIYSw\nMlK8hRBCCCsjxVsIIYSwMlK8hRBCCCsjxVsIIYSwMlY1Paql2yCEEEKUttymR7Wa4i2EEEIIPRk2\nF0IIIayMFG8hhBDCykjxFkIIIayMFG8hhBDCykjxFkIIIaxMuc/zVhTFFvgCaA2kAONVVf0r2/IB\nwBwgHVitquoqizS0DCtAHw4DpqLvw3PAa6qqytcYssmvD7OttxKIUVX1nVJuolUowM/iE8BSwAa4\nAYxSVTXVEm0tqwrQh4OBdwEN/e/ELy3S0DJOUZQOwCJVVXs88H6p1JSKcOU9CHBUVdUPeBv9/9gA\nKIriACwDegPdgQmKotS0SCvLNnN96Ay8D/irqtoF8AD6W6SVZVuefWigKMpEoCX6X5oid+Z+Fm2A\nlcBoVVW7Ar8DjSzSyrItv59Fw+/EzsAMRVE8Srl9ZZ6iKG8B/wEqPfB+qdWUilC8OwO7AFRVDQba\nZVv2KHBZVdU4VVXTgMNAt9JvYplnrg+TgU6qqiZnvbYHkkq3eVbBXB+iKIof0B5Ygf6qUeTOXD96\nAzHAdEVR9gNVVFVVS72FZZ/Zn0UgDagCOKP/WZSTyZwuA8+R8//VUqspFaF4uwP3s73OyBo2MiyL\ny7YsHv2VozCVZx+qqqqpqnobQFGUKYCLqqp7LdDGsi7PPlQUpQ4wF3gdKdz5Mff/c3XAD1gO9AKe\nVBSlB+JB5voQ9FfiIcAfwA5VVbOvKwBVVTejHxZ/UKnVlIpQvO8Dbtle26qqmpn197gHlrkB90qr\nYVbEXB+iKIqtoihLgCeB50u7cVbCXB++gL7w/ALMAoYrijKqlNtnLcz1Ywz6qx5VVdV09FeXD15V\nCjN9qCjKI+hPIhsCXkAtRVFeKPUWWq9SqykVoXgHAf0AFEXpCPwv27ILQDNFUaoqiuKIfnjjaOk3\nscwz14egH+qtBAzONnwuTOXZh6qqLldVtV3Wgy+LgLWqqn5nmWaWeeZ+FsMAV0VRmmS97or+6lGY\nMteHTkAGkJJV0KPRD6GLgim1mlLu5zbPeojF8GQlwBjAF3BVVfU/iqL0Rz9kaQt8parqvy3T0rLL\nXB8CJ7P+HMy2yb9UVd1aqo0s4/L7Ocy23suAoqrqu6XfyrKvAP8/G06AbIAgVVWnWaalZVcB+nAa\nMBz98yyXgVeyRjJENoqieKE/0fbL+sZNqdaUcl+8hRBCiPKmIgybCyGEEOWKFG8hhBDCykjxFkII\nIayMFG8hhBDCykjxFkIIIayMFG8hhBDCypT7VDEhyoKs74ReBEIfWNRfVdUbeWwzD9BUVZ1fhOOO\nRh+UEJ71ljNwAH3yW0Yh9zUfOKGq6s+KouwzpCkpinJaVdW2D9vGrH3sB+oBuqy33NFPujJCVdVo\nM9tNAO6rqrq+KMcXwtpI8Rai9NwoZJErjkkYNGCrqqpjwRgHuR+YDHxamB2pqvpetpfds71fpMKd\nRQPGqap6EIwTiWwCpqNPvsqLH7CvGI4vhFWR4i2EhSmK0hJ9IXUFagJLVVVdnm25PfA10CLrrS9U\nVV2lKEot4EugAZAJvKOq6u+5HMIYdqKqaqaiKEeBZln7HoO+QGrowyheB1KB1bkc7xv0hdIna9uj\nqqp2UhQlE3AArgFtVFWNVhSlGvps90fQxyPOz1rnCvoZu+6aa2dWX1QHjmUd68Wsdjpn/RkPOAID\ngB6KotxEP83nCqB+Pv0hhNWTe95ClJ66iqKczvZnRtb744D3VVVtD/QEFj6wnR9QVVVVH/RpWX5Z\n7/8LWK2qajvgWWCFoiiu5hqgKIon0AcIUhSlFfAu0E1V1dZAAvAe0CmP42noh/GnAqiq2smw36wh\n+A3Ai1lvPQ9sAaoCHwFPZe1vD/BxLk2zAVYpinImqxAfzVr3k6zRgonAM6qqtsnafmZWYd4OzFFV\n9bes/viqMP0hhLWSK28hSs/NPIaYZwB9FUV5G3gccMl635Cl/AegKIqyC33ymGEYuVfW+wuyXtsD\njTENmrABBiqKcjrr77bAT6qqrlcU5XVgu6qqhtSjleiv8Bflcbz8rAH+CXwODEN/YtAR/dX3fkVR\nAOzQp389yDhsrihKJ+An4FfDnNqKogzO+hwK+iH73ObaLkh/CFEuSPEWwvI2oi9oO4D1wNDsC1VV\nvasoSgv0w8/9gFNZr22BHqqqxgIoilIPiHxg3xqwzXDP+wE2mA5V2wL2Zo5nlqqqIYqiVFMU5Qmg\nnqqqxxRFeRY4rKrqs1ltdMI0MvHB9qCq6lFFUT4FvlMUpTVQGX34zbfo79efRT+8/6CC9IcQ5YIM\nmwtheb2A91RV3QH4g/HBMrL+/gzwvaqqO4Gp6J/IbgD8F/2DZ2QV17Po7wdn92CBzm4/+qvZqlmv\nXwH+a+Z42WUoimKXyz5/QH/feV3W62Cgk6IozbJezwYW59Ge7A/oLUM/AjEJ/f35DPTD7/vRn1AY\njp2O/l46FKw/hCgXpHgLUXryenp8HnBYUZQgoDnwJ9Aoa30N2A0kKooSir4Y/qSq6h/AFKCjoihn\n0RfLEaqqJuRyzFyPq6rqOfQF8YCiKH+i/3rWbDPHy24bcEZRlEoP7P8H9FGT32cd4xYwFtigKMr/\ngLboHzwzS1XVVOAf6KMV/wLOZPXLAfTD4I9krboXeFdRlOcK2B9ClAsSCSqEEEJYGbnyFkIIIayM\nFG8hhBDCykjxFkIIIayMFG8hhBDCykjxFkIIIayMFG8hhBDCykjxFkIIIayMFG8hhBDCyvw/LEpm\n4h71508AAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10a020be0>"
       ]
      }
     ],
     "prompt_number": 22
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "Validation"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "final_clf = Pipeline([\n",
      "                ('vect', TfidfVectorizer(\n",
      "                                         binary=False,\n",
      "                                         stop_words=stop_words,\n",
      "                                         tokenizer=porter_tokenizer,\n",
      "                                         ngram_range=(1,1),\n",
      "                                         )\n",
      "                ),\n",
      "                ('clf', MultinomialNB(alpha=1.0)),\n",
      "               ])\n",
      "final_clf.fit(X_train, y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 19,
       "text": [
        "Pipeline(steps=[('vect', TfidfVectorizer(analyzer=u'word', binary=False, charset=None,\n",
        "        charset_error=None, decode_error=u'strict',\n",
        "        dtype=<type 'numpy.int64'>, encoding=u'utf-8', input=u'content',\n",
        "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
        "        ngram_range=(1, 1), norm...rue,\n",
        "        vocabulary=None)), ('clf', MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True))])"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Confusion matrix"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import matplotlib as mpl\n",
      "import numpy as np\n",
      "\n",
      "cm = metrics.confusion_matrix(y_train, final_clf.predict(X_train))\n",
      "\n",
      "np.set_printoptions(suppress=True)\n",
      "mpl.rc(\"figure\", figsize=(4, 2))\n",
      "\n",
      "hm = sns.heatmap(cm, \n",
      "            cbar=False,\n",
      "            annot=True, \n",
      "            square=True,\n",
      "            fmt='d',\n",
      "            yticklabels=['happy','sad'],\n",
      "            xticklabels=['happy','sad'],\n",
      "            cmap='Blues'\n",
      "            )\n",
      "plt.title('Confusion matrix - Training dataset')\n",
      "plt.ylabel('actual class')\n",
      "plt.xlabel('predicted class')\n",
      "plt.tight_layout()\n",
      "plt.savefig('./images/confmat_training.eps', dpi=300)\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAMcAAACJCAYAAACYVyzaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nngD+H1qUZfitmQ/bAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x108cb2860>"
       ]
      }
     ],
     "prompt_number": 30
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df = pd.read_csv('../../dataset/validation/valid_lyrics_200.csv')\n",
      "\n",
      "X_valid = df['lyrics'].values \n",
      "y_valid = df['mood'].values\n",
      "\n",
      "y_valid = le.transform(y_valid)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 31
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cm = metrics.confusion_matrix(y_valid, final_clf.predict(X_valid))\n",
      "\n",
      "np.set_printoptions(suppress=True)\n",
      "mpl.rc(\"figure\", figsize=(4, 2))\n",
      "\n",
      "hm = sns.heatmap(cm, \n",
      "            cbar=False,\n",
      "            annot=True, \n",
      "            square=True,\n",
      "            fmt='d',\n",
      "            yticklabels=['happy','sad'],\n",
      "            xticklabels=['happy','sad'],\n",
      "            cmap='Blues'\n",
      "            )\n",
      "plt.title('Confusion matrix - Validation dataset')\n",
      "plt.ylabel('actual class')\n",
      "plt.xlabel('predicted class')\n",
      "plt.tight_layout()\n",
      "plt.savefig('./images/confmat_valid.eps', dpi=300)\n",
      "plt.show()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAANEAAACJCAYAAACy5911AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJE+iiCKRPIkiikTy5P8Bhtsryl9b13AAAAAASUVORK5CYII=\n",
       "text": [
        "<matplotlib.figure.Figure at 0x10a69b860>"
       ]
      }
     ],
     "prompt_number": 32
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "ROC AUC, Accuracy, Precision, Recall, and F1-score"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Custom scorer methods to account for positive-negative class labels\n",
      "\n",
      "from sklearn import metrics\n",
      "\n",
      "# `pos_label` for positive class, since we have sad=1, happy=0\n",
      "\n",
      "acc_scorer = metrics.make_scorer(metrics.accuracy_score, greater_is_better=True)\n",
      "pre_scorer = metrics.make_scorer(metrics.precision_score, greater_is_better=True, pos_label=0)\n",
      "rec_scorer = metrics.make_scorer(metrics.recall_score, greater_is_better=True, pos_label=0)\n",
      "f1_scorer = metrics.make_scorer(metrics.f1_score, greater_is_better=True, pos_label=0)\n",
      "auc_scorer = metrics.make_scorer(metrics.roc_auc_score, greater_is_better=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 263
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "d = {'Data':['Training', 'Validation'],\n",
      "     'ACC (%)':[],\n",
      "     'PRE (%)':[],\n",
      "     'REC (%)':[],\n",
      "     'F1 (%)':[],\n",
      "     'ROC AUC (%)':[],\n",
      "}"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 307
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "d['ACC (%)'].append(acc_scorer(estimator=final_clf, X=X_train, y_true=y_train))\n",
      "d['PRE (%)'].append(pre_scorer(estimator=final_clf, X=X_train, y_true=y_train))\n",
      "d['REC (%)'].append(rec_scorer(estimator=final_clf, X=X_train, y_true=y_train))\n",
      "d['F1 (%)'].append(f1_scorer(estimator=final_clf, X=X_train, y_true=y_train))\n",
      "d['ROC AUC (%)'].append(auc_scorer(estimator=final_clf, X=X_train, y_true=y_train))\n",
      "\n",
      "d['ACC (%)'].append(acc_scorer(estimator=final_clf, X=X_valid, y_true=y_valid))\n",
      "d['PRE (%)'].append(pre_scorer(estimator=final_clf, X=X_valid, y_true=y_valid))\n",
      "d['REC (%)'].append(rec_scorer(estimator=final_clf, X=X_valid, y_true=y_valid))\n",
      "d['F1 (%)'].append(f1_scorer(estimator=final_clf, X=X_valid, y_true=y_valid))\n",
      "d['ROC AUC (%)'].append(auc_scorer(estimator=final_clf, X=X_valid, y_true=y_valid))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 308
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df_perform = pd.DataFrame(d)\n",
      "df_perform = df_perform[['ACC (%)', 'PRE (%)', 'REC (%)', 'F1 (%)', 'ROC AUC (%)']]\n",
      "df_perform.index=(['Training', 'Validation'])\n",
      "df_perform = df_perform*100\n",
      "df_perform = np.round(df_perform, decimals=2)\n",
      "df_perform"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "html": [
        "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
        "<table border=\"1\" class=\"dataframe\">\n",
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
        "      <th>ACC (%)</th>\n",
        "      <th>PRE (%)</th>\n",
        "      <th>REC (%)</th>\n",
        "      <th>F1 (%)</th>\n",
        "      <th>ROC AUC (%)</th>\n",
        "    </tr>\n",
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>Training</th>\n",
        "      <td> 80.0</td>\n",
        "      <td> 99.60</td>\n",
        "      <td> 55.38</td>\n",
        "      <td> 71.18</td>\n",
        "      <td> 77.60</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>Validation</th>\n",
        "      <td> 54.5</td>\n",
        "      <td> 88.89</td>\n",
        "      <td> 15.24</td>\n",
        "      <td> 26.02</td>\n",
        "      <td> 56.57</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 309,
       "text": [
        "            ACC (%)  PRE (%)  REC (%)  F1 (%)  ROC AUC (%)\n",
        "Training       80.0    99.60    55.38   71.18        77.60\n",
        "Validation     54.5    88.89    15.24   26.02        56.57"
       ]
      }
     ],
     "prompt_number": 309
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "df_perform.to_csv('./clf_performance.csv', index_label=False)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 310
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "<br>\n",
      "<br>"
     ]
    },
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "Save classifier"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "[[back to top](#Sections)]"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "lyrics_clf_1000 = final_clf\n",
      "\n",
      "pickle_out = open('./lyrics_clf_1000_py27.pkl', 'wb')\n",
      "pickle.dump(lyrics_clf_1000, pickle_out)\n",
      "pickle_out.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 20
    },
    {
     "cell_type": "heading",
     "level": 1,
     "metadata": {},
     "source": [
      "New pickle objects for webapp"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pickle\n",
      "\n",
      "pickle_out = open('./lyrics_label_encoder.pkl', 'rb')\n",
      "le = pickle.load(pickle_out)\n",
      "pickle_out.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 13
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "\n",
      "from sklearn.naive_bayes import MultinomialNB\n",
      "\n",
      "with open('./stopwords_eng.txt', 'r') as infile:\n",
      "    stop_words = infile.read().splitlines()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": []
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Porter Stemmer\n",
      "\n",
      "import nltk\n",
      "import string\n",
      "import re\n",
      "\n",
      "\n",
      "porter_stemmer = nltk.stem.porter.PorterStemmer()\n",
      "\n",
      "def porter_tokenizer(text, stemmer=porter_stemmer):\n",
      "    \"\"\"\n",
      "    A Porter-Stemmer-Tokenizer hybrid to splits sentences into words (tokens) \n",
      "    and applies the porter stemming algorithm to each of the obtained token. \n",
      "    Tokens that are only consisting of punctuation characters are removed as well.\n",
      "    Only tokens that consist of more than one letter are being kept.\n",
      "    \n",
      "    Parameters\n",
      "    ----------\n",
      "        \n",
      "    text : `str`. \n",
      "      A sentence that is to split into words.\n",
      "        \n",
      "    Returns\n",
      "    ----------\n",
      "    \n",
      "    no_punct : `str`. \n",
      "      A list of tokens after stemming and removing Sentence punctuation patterns.\n",
      "    \n",
      "    \"\"\"\n",
      "    lower_txt = text.lower()\n",
      "    tokens = nltk.wordpunct_tokenize(lower_txt)\n",
      "    stems = [porter_stemmer.stem(t) for t in tokens]\n",
      "    no_punct = [s for s in stems if re.match('^[a-zA-Z]+$', s) is not None]\n",
      "    return no_punct\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pandas as pd\n",
      "df = pd.read_csv('../../dataset/training/train_lyrics_1000.csv')\n",
      "X_train = df['lyrics'].values \n",
      "y_train = df['mood'].values\n",
      "\n"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.preprocessing import LabelEncoder\n",
      "from sklearn.pipeline import Pipeline\n",
      "\n",
      "le = LabelEncoder()\n",
      "le.fit(y_train)\n",
      "y_train = le.transform(y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "ename": "NameError",
       "evalue": "name 'y_train' is not defined",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-4-f4949c0aa6db>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mle\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mLabelEncoder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0mle\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m \u001b[0my_train\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mle\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mNameError\u001b[0m: name 'y_train' is not defined"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.feature_extraction.text import TfidfVectorizer\n",
      "\n",
      "\n",
      "\n",
      "final_clf = Pipeline([\n",
      "                ('vect', TfidfVectorizer(\n",
      "                                         binary=False,\n",
      "                                         stop_words=stop_words,\n",
      "                                         ngram_range=(1,1),\n",
      "                                         )\n",
      "                ),\n",
      "                ('clf', MultinomialNB(alpha=1.0)),\n",
      "               ])\n",
      "final_clf.fit(X_train, y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 26,
       "text": [
        "Pipeline(steps=[('vect', TfidfVectorizer(analyzer=u'word', binary=False, charset=None,\n",
        "        charset_error=None, decode_error=u'strict',\n",
        "        dtype=<type 'numpy.int64'>, encoding=u'utf-8', input=u'content',\n",
        "        lowercase=True, max_df=1.0, max_features=None, min_df=1,\n",
        "        ngram_range=(1, 1), norm...rue,\n",
        "        vocabulary=None)), ('clf', MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True))])"
       ]
      }
     ],
     "prompt_number": 26
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.externals import joblib\n",
      "from sklearn.pipeline import Pipeline\n",
      "\n",
      "import pickle\n",
      "\n",
      "pickle_out = open('./lyrics_label_encoder_np.pkl', 'wb')\n",
      "pickle.dump(le, pickle_out)\n",
      "pickle_out.close()\n",
      "\n",
      "joblib.dump(final_clf, 'lyrics_clf_1000_np.pkl') "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 27,
       "text": [
        "['lyrics_clf_1000_np.pkl',\n",
        " 'lyrics_clf_1000_np.pkl_01.npy',\n",
        " 'lyrics_clf_1000_np.pkl_02.npy',\n",
        " 'lyrics_clf_1000_np.pkl_03.npy',\n",
        " 'lyrics_clf_1000_np.pkl_04.npy',\n",
        " 'lyrics_clf_1000_np.pkl_05.npy',\n",
        " 'lyrics_clf_1000_np.pkl_06.npy',\n",
        " 'lyrics_clf_1000_np.pkl_07.npy']"
       ]
      }
     ],
     "prompt_number": 27
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.externals import joblib\n",
      "\n",
      "final_clf = joblib.load('lyrics_clf_1000_jb.pkl') "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "final_clf.predict(X_train)[:3]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 6,
       "text": [
        "array([1, 0, 1])"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "joblib.dump(le, 'lyrics_label_encoder_jb.pkl') "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 13,
       "text": [
        "['lyrics_label_encoder_jb.pkl', 'lyrics_label_encoder_jb.pkl_01.npy']"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.externals import joblib\n",
      "lyrics_label_encoder = joblib.load('lyrics_label_encoder_jb.pkl') "
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}