Day 01 Simple_test_processing.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "colab": {
      "name": "Copy of Copy of practice01_simple_text_processing.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/lagom-QB/M12/blob/master/Simple_test_processing.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1sxIff7iTyPH",
        "colab_type": "text"
      },
      "source": [
        "## Practice 01. Simple text processing."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ijdw6mz9TyPL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "import pandas as pd\n",
        "from IPython import display"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m8eR3l80TyPk",
        "colab_type": "text"
      },
      "source": [
        "### Toxic or not\n",
        "Our main goal in this assignment is to classify, whether the comments are toxic or not. And practice with both classical approaches and PyTorch in the process.\n",
        "\n",
        "*Credits: This homework is inspired by YSDA NLP_course.*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C5kjQylwTyPr",
        "colab_type": "code",
        "outputId": "5bc0f042-9989-4d91-b441-38a029229129",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 145
        }
      },
      "source": [
        "# In colab uncomment this cell\n",
        "! wget https://raw.githubusercontent.com/neychev/harbour_dlia2020/master/day01/utils.py -nc"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-06-15 18:37:15--  https://raw.githubusercontent.com/neychev/harbour_dlia2020/master/day01/utils.py\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 579 [text/plain]\n",
            "Saving to: ‘utils.py’\n",
            "\n",
            "\rutils.py              0%[                    ]       0  --.-KB/s               \rutils.py            100%[===================>]     579  --.-KB/s    in 0s      \n",
            "\n",
            "2020-06-15 18:37:15 (29.7 MB/s) - ‘utils.py’ saved [579/579]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m9duPfoaTyP6",
        "colab_type": "code",
        "outputId": "db4a3325-9803-40f5-cad1-8bb97db801d4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 165
        }
      },
      "source": [
        "try:\n",
        "    data = pd.read_csv('../datasets/comments_small_dataset/comments.tsv', sep='\\t')\n",
        "except FileNotFoundError:\n",
        "    ! wget https://raw.githubusercontent.com/neychev/harbour_dlia2020/master/datasets/comments_small_dataset/comments.tsv -nc\n",
        "    data = pd.read_csv(\"comments.tsv\", sep='\\t')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-06-15 18:37:16--  https://raw.githubusercontent.com/neychev/harbour_dlia2020/master/datasets/comments_small_dataset/comments.tsv\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 353358 (345K) [text/plain]\n",
            "Saving to: ‘comments.tsv’\n",
            "\n",
            "comments.tsv        100%[===================>] 345.08K  --.-KB/s    in 0.06s   \n",
            "\n",
            "2020-06-15 18:37:18 (6.07 MB/s) - ‘comments.tsv’ saved [353358/353358]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q_6l5FGaTyQL",
        "colab_type": "code",
        "outputId": "5b4ee919-acea-4b38-c979-1605de88ecfd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 144
        }
      },
      "source": [
        "texts = data['comment_text'].values\n",
        "target = data['should_ban'].values\n",
        "data[50::200]"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>should_ban</th>\n",
              "      <th>comment_text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>0</td>\n",
              "      <td>\"Those who're in advantageous positions are th...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>250</th>\n",
              "      <td>1</td>\n",
              "      <td>Fartsalot56 says f**k you motherclucker!!</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>450</th>\n",
              "      <td>1</td>\n",
              "      <td>Are you a fool? \\n\\nI am sorry, but you seem t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>650</th>\n",
              "      <td>1</td>\n",
              "      <td>I AM NOT A VANDAL!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>850</th>\n",
              "      <td>0</td>\n",
              "      <td>Citing sources\\n\\nCheck out the Wikipedia:Citi...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     should_ban                                       comment_text\n",
              "50            0  \"Those who're in advantageous positions are th...\n",
              "250           1          Fartsalot56 says f**k you motherclucker!!\n",
              "450           1  Are you a fool? \\n\\nI am sorry, but you seem t...\n",
              "650           1    I AM NOT A VANDAL!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!\n",
              "850           0  Citing sources\\n\\nCheck out the Wikipedia:Citi..."
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j4GeqLCPTyQY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "texts_train, texts_test, y_train, y_test = train_test_split(texts, target, test_size=0.5, random_state=42)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NJ47nBhCTyQu",
        "colab_type": "text"
      },
      "source": [
        "__Note:__ it is generally a good idea to split data into train/test before anything is done to them.\n",
        "\n",
        "It guards you against possible data leakage in the preprocessing stage. For example, should you decide to select words present in obscene tweets as features, you should only count those words over the training set. Otherwise your algoritm can cheat evaluation."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GPexLCWdTyQw",
        "colab_type": "text"
      },
      "source": [
        "### Preprocessing and tokenization\n",
        "\n",
        "Comments contain raw text with punctuation, upper/lowercase letters and even newline symbols.\n",
        "\n",
        "To simplify all further steps, we'll split text into space-separated tokens using one of nltk tokenizers.\n",
        "\n",
        "Generally, library `nltk` [link](https://www.nltk.org) is widely used in NLP. It is not necessary in here, but mentioned to intoduce it to you."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R1o-fqJfTyQx",
        "colab_type": "code",
        "outputId": "9c28c3c8-5763-47bd-ee06-3d2dc944c5ba",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 40
        }
      },
      "source": [
        "from nltk.tokenize import TweetTokenizer\n",
        "tokenizer = TweetTokenizer()\n",
        "preprocess = lambda text: ' '.join(tokenizer.tokenize(text.lower()))\n",
        "\n",
        "text = 'How to be a grown-up at work: replace \"I don\\'t want to do that\" with \"Ok, great!\".'\n",
        "print(\"before:\", text,)\n",
        "print(\"after:\", preprocess(text),)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "before: How to be a grown-up at work: replace \"I don't want to do that\" with \"Ok, great!\".\n",
            "after: how to be a grown-up at work : replace \" i don't want to do that \" with \" ok , great ! \" .\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HaPNuiz7TyRF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# task: preprocess each comment in train and test\n",
        "\n",
        "texts_train = np.array(list(map(preprocess, texts_train))) #<YOUR CODE>\n",
        "texts_test = np.array(list(map(preprocess, texts_test))) #<YOUR CODE> "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KUWrc_ZZV1rX",
        "colab_type": "code",
        "outputId": "c3b8f1dc-674b-4010-9897-21bd1d1da4c5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "texts_train[5]"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'who cares anymore . they attack with impunity .'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RpkQ5xrCTyRZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Small check that everything is done properly\n",
        "assert texts_train[5] ==  'who cares anymore . they attack with impunity .'\n",
        "assert texts_test[89] == 'hey todds ! quick q ? why are you so gay'\n",
        "assert len(texts_test) == len(y_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ynSC8MSTTyRi",
        "colab_type": "text"
      },
      "source": [
        "### Step 1: bag of words\n",
        "\n",
        "One traditional approach to such problem is to use bag of words features:\n",
        "1. build a vocabulary of frequent words (use train data only)\n",
        "2. for each training sample, count the number of times a word occurs in it (for each word in vocabulary).\n",
        "3. consider this count a feature for some classifier\n",
        "\n",
        "__Note:__ in practice, you can compute such features using sklearn. __Please don't do that in the current assignment, though.__\n",
        "* `from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F7EYjAIaTyRn",
        "colab_type": "code",
        "outputId": "dd4d329a-d705-46e7-9ee8-c6ef9402e1d2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 40
        }
      },
      "source": [
        "# task: find up to k most frequent tokens in texts_train,\n",
        "# sort them by number of occurences (highest first)\n",
        "k = min(10000, len(set(' '.join(texts_train).split())))\n",
        "\n",
        "bow_vocabulary = list(set(' '.join(texts_train).split()))                                    #<YOUR CODE>\n",
        "\n",
        "n = 100\n",
        "print(k,'\\nexample features:', len(sorted(bow_vocabulary)[::n]))                             #Words that appear n times"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "5722 \n",
            "example features: 58\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zYy2cQ95ZEWC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "word_to_index = {word:index for index,word in enumerate(bow_vocabulary)}"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ONeLLkauZMyS",
        "colab_type": "code",
        "outputId": "7c6ba0ad-bafd-4bb4-dce9-631eee725c2d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "word_to_index.get('bag')"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2530"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "okjZMi6DZh8W",
        "colab_type": "code",
        "outputId": "6ed120a4-2c84-4d56-ef5c-f18a05a1d10b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "VOCABULARY_SIZE = len(bow_vocabulary)\n",
        "VOCABULARY_SIZE"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5722"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V_9et_UoTyRx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def text_to_bow(text):\n",
        "    \"\"\" convert text string to an array of token counts. Use bow_vocabulary. \"\"\"\n",
        "    #<YOUR CODE>\n",
        "    global VOCABULARY_SIZE, word_to_index\n",
        "    bow_vector = np.zeros(VOCABULARY_SIZE, dtype=float)\n",
        "\n",
        "    if isinstance(text, str) :\n",
        "        text = text.split(' ')\n",
        "\n",
        "    for word in text:\n",
        "        word_index = word_to_index.get(word, -1)\n",
        "        if word_index >= 0:\n",
        "            bow_vector[word_index] += 1.\n",
        "\n",
        "    return bow_vector"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rXovDuPcTyR9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train_bow = np.stack(list(map(text_to_bow, texts_train)))\n",
        "X_test_bow = np.stack(list(map(text_to_bow, texts_test)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XQUV8w2h_btg",
        "colab_type": "code",
        "outputId": "0b1c95f6-0c62-49ad-f967-579839b8a160",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "len(X_train_bow)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "500"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jh_KKjP_TySJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Small check that everything is done properly\n",
        "k_max = len(set(' '.join(texts_train).split()))\n",
        "assert X_train_bow.shape == (len(texts_train), min(k, k_max))\n",
        "assert X_test_bow.shape == (len(texts_test), min(k, k_max))\n",
        "assert np.all(X_train_bow[5:10].sum(-1) == np.array([len(s.split()) for s in  texts_train[5:10]]))\n",
        "assert len(bow_vocabulary) <= min(k, k_max)\n",
        "assert X_train_bow[6, bow_vocabulary.index('.')] == texts_train[6].split().count('.')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cs--TymETySY",
        "colab_type": "text"
      },
      "source": [
        "Now let's do the trick with `sklearn` logistic regression implementation:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XiIB4201TySa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "bow_model = LogisticRegression().fit(X_train_bow, y_train)   #<YOUR CODE>"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6kWGSgr0TySi",
        "colab_type": "code",
        "outputId": "601ecc88-3264-4b9d-96da-51bdf76dd91f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "from sklearn.metrics import roc_auc_score, roc_curve\n",
        "\n",
        "for name, X, y, model in [\n",
        "    ('train', X_train_bow, y_train, bow_model),\n",
        "    ('test ', X_test_bow, y_test, bow_model)\n",
        "]:\n",
        "    proba = model.predict_proba(X)[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hCmtxXD9TySu",
        "colab_type": "text"
      },
      "source": [
        "Seems alright. Now let's create the simple logistic regression using PyTorch. Just like in the classwork."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_MnHPnM6TySv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch\n",
        "from torch import nn\n",
        "from torch.nn import functional as F\n",
        "from torch.optim.lr_scheduler import StepLR, ReduceLROnPlateau\n",
        "\n",
        "from sklearn.metrics import accuracy_score"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jIt8lwQ1TyS3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from utils import plot_train_process"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G5SzTZq5TyTC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = nn.Sequential(nn.Linear(VOCABULARY_SIZE, 2))                                                                   ### YOUR CODE HERE\n",
        "model.add_module('dropout1', nn.Dropout2d(p = .73))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E1gTWyuhTyTM",
        "colab_type": "text"
      },
      "source": [
        "Remember what we discussed about loss functions! `nn.CrossEntropyLoss` combines both log-softmax and `NLLLoss`.\n",
        "\n",
        "__Be careful with it! Criterion `nn.CrossEntropyLoss` with still work with log-softmax output, but it won't allow you to converge to the optimum.__ Next comes small demonstration:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eaI0jHmTTyTO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# loss_function = nn.NLLLoss()\n",
        "loss_function = nn.CrossEntropyLoss()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QXzrj99STyTV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "opt = torch.optim.Adagrad(model.parameters())                                                                           ### YOUR CODE HERE"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mpBuAdvoTyTh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train_bow_torch = torch.tensor(X_train_bow, dtype=torch.float32)                                                      ### YOUR CODE HERE\n",
        "X_test_bow_torch = torch.tensor(X_test_bow, dtype=torch.float32)                                                        ### YOUR CODE HERE\n",
        "\n",
        "y_train_torch = torch.tensor(y_train, dtype=torch.long)                                                                 ### YOUR CODE HERE\n",
        "y_test_torch = torch.tensor(y_test, dtype=torch.long)                                                                   ### YOUR CODE HERE"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SgYipX-tTyTm",
        "colab_type": "text"
      },
      "source": [
        "Let's test that everything is fine"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "83qjBgVpTyTo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# example loss\n",
        "loss = loss_function(model(X_train_bow_torch[:3]), y_train_torch[:3])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AfJLddoATyTy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "assert type(loss.item()) == float"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mlCSrCX4TyT-",
        "colab_type": "text"
      },
      "source": [
        "Here comes small function to train the model. In future we will take in into separate file, but for this homework it's ok to implement it here. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WnBytNbbTyUA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train_model(\n",
        "    model,\n",
        "    opt,\n",
        "    lr_scheduler,\n",
        "    X_train_torch,\n",
        "    y_train_torch,\n",
        "    X_val_torch,\n",
        "    y_val_torch,\n",
        "    n_iterations=500,\n",
        "    batch_size=256,\n",
        "    warm_start=False,\n",
        "    show_plots=True,\n",
        "    eval_every=10\n",
        "):\n",
        "    if not warm_start:\n",
        "        for name, module in model.named_children():\n",
        "            print('resetting ', name)\n",
        "            try:\n",
        "                module.reset_parameters()\n",
        "                print('reset ', name)\n",
        "            except AttributeError as e:\n",
        "                print('Cannot reset {} module parameters: {}'.format(name, e))\n",
        "\n",
        "    train_loss_history = []\n",
        "    train_acc_history = []\n",
        "    val_loss_history = []\n",
        "    val_acc_history = []\n",
        "\n",
        "    local_train_loss_history = []\n",
        "    local_train_acc_history = []\n",
        "    for i in range(n_iterations):\n",
        "\n",
        "        # sample random observations\n",
        "        ix = np.random.randint(0, len(X_train_torch), batch_size)\n",
        "        # ix = SelectKBest(score_func=f_classif, k = batch_size).fit(X_train_torch, y_train_torch).scores_    #### <<<<< ERROR HERE or smthx\n",
        "        \n",
        "        x_batch = X_train_torch[ix]\n",
        "        y_batch = y_train_torch[ix]\n",
        "\n",
        "        # predict log-probabilities or logits\n",
        "        y_predicted = model(x_batch)                                                                                       ### YOUR CODE\n",
        "\n",
        "        # compute loss, just like before\n",
        "        loss = loss_function(y_predicted, y_batch)                                                                          ### YOUR CODE\n",
        "\n",
        "        # compute gradients\n",
        "        loss.backward()                                                                                                     ### YOUR CODE\n",
        "\n",
        "        # Adam step\n",
        "        opt.step()                                                                                                          ### YOUR CODE\n",
        "\n",
        "        # clear gradients\n",
        "        opt.zero_grad()                                                                                                     ### YOUR CODE\n",
        "\n",
        "        local_train_loss_history.append(loss.data.numpy())\n",
        "        local_train_acc_history.append(\n",
        "            accuracy_score(\n",
        "                y_batch.to('cpu').detach().numpy(),\n",
        "                y_predicted.to('cpu').detach().numpy().argmax(axis=1)\n",
        "            )\n",
        "        )\n",
        "\n",
        "        if i % eval_every == 0:\n",
        "            train_loss_history.append(np.mean(local_train_loss_history))\n",
        "            train_acc_history.append(np.mean(local_train_acc_history))\n",
        "            local_train_loss_history, local_train_acc_history = [], []\n",
        "\n",
        "            predictions_val = model(X_val_torch)\n",
        "            val_loss_history.append(loss_function(predictions_val, y_val_torch).to('cpu').detach().item())\n",
        "\n",
        "            acc_score_val = accuracy_score(y_val_torch.cpu().numpy(), predictions_val.to('cpu').detach().numpy().argmax(axis=1))\n",
        "            val_acc_history.append(acc_score_val)\n",
        "            lr_scheduler.step(train_loss_history[-1])\n",
        "\n",
        "            if show_plots:\n",
        "                display.clear_output(wait=True)\n",
        "                plot_train_process(train_loss_history, val_loss_history, train_acc_history, val_acc_history)\n",
        "    return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HafIWRULTyUI",
        "colab_type": "text"
      },
      "source": [
        "Let's run it on the data. Note, that here we use the `test` part of the data for validation. It's not so good idea in general, but in this task our main goal is practice."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uyX3sGsGvy9i",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "lr_scheduler = ReduceLROnPlateau(opt, verbose=True, patience=5)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EWuWdpVY7l8A",
        "colab_type": "code",
        "outputId": "890820fc-9a2f-421b-b70c-6f44439c1891",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "loss"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(0.7432, grad_fn=<NllLossBackward>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yEgN-QOaTyUJ",
        "colab_type": "code",
        "outputId": "a228f1d1-ef39-4329-a8f9-4793e4064e8c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 311
        }
      },
      "source": [
        "from sklearn.feature_selection import SelectKBest,f_classif\n",
        "\n",
        "train_model(model, \n",
        "            opt, \n",
        "            lr_scheduler, \n",
        "            X_train_bow_torch, \n",
        "            y_train_torch, \n",
        "            X_test_bow_torch, \n",
        "            y_test_torch)"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): Linear(in_features=5722, out_features=2, bias=True)\n",
              "  (dropout1): Dropout2d(p=0.73, inplace=False)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jlwG7X5MTyUW",
        "colab_type": "code",
        "outputId": "eee67d89-ee95-4cfc-a5ab-9d4c54515280",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "from sklearn.metrics import roc_auc_score, roc_curve\n",
        "\n",
        "for name, X, y, model in [\n",
        "    ('train', X_train_bow_torch, y_train, model),\n",
        "    ('test ', X_test_bow_torch, y_test, model)\n",
        "]:\n",
        "    proba = model(X).detach().cpu().numpy()[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zb23q2rE-iBU",
        "colab_type": "code",
        "outputId": "91c83631-c47f-40e2-d315-c24f43f38689",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "X_train_bow_torch.shape"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([500, 5722])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1GqteshFTyUf",
        "colab_type": "text"
      },
      "source": [
        "Try to vary the number of tokens `k` and check how the model performance changes. Show it on a plot."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MybKQpx4xhIs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.feature_selection import SelectKBest,f_classif\n",
        "\n",
        "n = [10,\n",
        "     50,\n",
        "     100,\n",
        "     500]\n",
        "\n",
        "\n",
        "model_ass = nn.Sequential(nn.Linear(VOCABULARY_SIZE, 2), \n",
        "                          nn.Dropout2d(p = .52))   \n",
        "optimizer_ass = torch.optim.AdamW(model_ass.parameters(), lr = 0.001)\n",
        "lr_scheduler_ass = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer_ass, patience=5)\n",
        "\n",
        "def work(token_list):\n",
        "  TOKEN_COUNT =  len(token_list)\n",
        "\n",
        "  print(train_model(model_ass, \n",
        "          optimizer_ass, \n",
        "          lr_scheduler_ass, \n",
        "          X_train_bow_torch, \n",
        "          y_train_torch, \n",
        "          X_test_bow_torch, \n",
        "          y_test_torch,\n",
        "          batch_size = TOKEN_COUNT))\n",
        "  \n",
        "  for name, X, y, model in [\n",
        "    ('train', X_train_bow_torch, y_train, model_ass),\n",
        "    ('test ', X_test_bow_torch, y_test, model_ass)\n",
        "    ]:\n",
        "    proba = model(X).detach().cpu().numpy()[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "    plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "    plt.legend(fontsize='large')\n",
        "    plt.grid()\n",
        "  \n",
        "tokens = []\n",
        "for i in n:\n",
        "    tokens.append(sorted(bow_vocabulary)[:i])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c4d24Hyf7aG2",
        "colab_type": "code",
        "outputId": "97616acd-8350-462a-a535-ed3a44c307ae",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "print([(len(i)) for i in tokens])"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[10, 50, 100, 500]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DAl0yPR5RLKf",
        "colab_type": "code",
        "outputId": "84c8e235-8844-4f51-cdfd-8e535afe3a3b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 559
        }
      },
      "source": [
        "work(tokens[0])"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Sequential(\n",
            "  (0): Linear(in_features=5722, out_features=2, bias=True)\n",
            "  (1): Dropout2d(p=0.52, inplace=False)\n",
            ")\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U-zYpq0mzE8u",
        "colab_type": "code",
        "outputId": "e84c0023-ddaf-4097-9839-dfcc4077edcb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 558
        }
      },
      "source": [
        "work(tokens[1])"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Sequential(\n",
            "  (0): Linear(in_features=5722, out_features=2, bias=True)\n",
            "  (1): Dropout2d(p=0.52, inplace=False)\n",
            ")\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-PYVFdOy3zm8",
        "colab_type": "code",
        "outputId": "ea7eb109-9aaa-43e5-da00-38f294a0141c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 558
        }
      },
      "source": [
        "work(tokens[2])"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Sequential(\n",
            "  (0): Linear(in_features=5722, out_features=2, bias=True)\n",
            "  (1): Dropout2d(p=0.52, inplace=False)\n",
            ")\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a2mTErZujLzA",
        "colab_type": "text"
      },
      "source": [
        ">*There is some witchery going on here*\n",
        ">> But the accuracy should decrease with fewer items selected so ... but still"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "At4UI0Q5TyUm",
        "colab_type": "text"
      },
      "source": [
        "### Step 2: implement TF-IDF features\n",
        "\n",
        "Not all words are equally useful. One can prioritize rare words and downscale words like \"and\"/\"or\" by using __tf-idf features__. This abbreviation stands for __text frequency/inverse document frequence__ and means exactly that:\n",
        "\n",
        "$$ feature_i = { Count(word_i \\in x) \\times { log {N \\over Count(word_i \\in D) + \\alpha} }}, $$\n",
        "\n",
        "\n",
        "where x is a single text, D is your dataset (a collection of texts), N is a total number of documents and $\\alpha$ is a smoothing hyperparameter (typically 1). \n",
        "And $Count(word_i \\in D)$ is the number of documents where $word_i$ appears.\n",
        "\n",
        "It may also be a good idea to normalize each data sample after computing tf-idf features.\n",
        "\n",
        "__Your task:__ implement tf-idf features, train a model and evaluate ROC curve. Compare it with basic BagOfWords model from above.\n",
        "\n",
        "__Please don't use sklearn/nltk builtin tf-idf vectorizers in your solution :)__ You can still use 'em for debugging though."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-59yTqwKTyUn",
        "colab_type": "text"
      },
      "source": [
        "Blog post about implementing the TF-IDF features from scratch: https://triton.ml/blog/tf-idf-from-scratch"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Idppn1gA10wv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Just reinitializing because if not then why not\n",
        "\n",
        "texts_train_, texts_test_, y_train_, y_test_ = train_test_split(texts, target, test_size=0.5, random_state=42)\n",
        "\n",
        "texts_test_ = list(map(preprocess, texts_test_))\n",
        "texts_train_ = list(map(preprocess, texts_train_))\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eGgwGkC4hGaJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "texts_test_data = [sentence.split(' ') for sentence in texts_test_]\n",
        "texts_train_data = [sentence.split(' ') for sentence in texts_train_]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZAczlWcvkALJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# texts_test_data[0], texts_train_data[0]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X3QVjICNTyUo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Compute tf map\n",
        "def computeTFDict(text):\n",
        "    \"\"\"computes the TF score for each word in the corpus, by document\n",
        "    \"\"\"\n",
        "    TFDict = {}\n",
        "    for word in text:\n",
        "        if word in TFDict:\n",
        "            TFDict[word] += 1\n",
        "        else:\n",
        "            TFDict[word] = 1\n",
        "    # Computes tf for each word\n",
        "    for word in TFDict:\n",
        "        TFDict[word] = TFDict[word] / len(text)\n",
        "    return TFDict\n",
        "\n",
        "test_tfDict = [computeTFDict(texts_test_datum) for texts_test_datum in texts_test_data]\n",
        "train_tfDict = [computeTFDict(texts_train_datum) for texts_train_datum in texts_train_data]\n",
        "\n",
        "# Count the occurance of words in the dictionary\n",
        "def computeCount(TFDict):\n",
        "    count = {}\n",
        "    for text in TFDict:\n",
        "        for word in text:\n",
        "            if word in count:\n",
        "                count[word] += 1\n",
        "            else:\n",
        "                count[word] = 1\n",
        "    return count\n",
        "\n",
        "test_wordCount = computeCount(test_tfDict)\n",
        "train_wordCount = computeCount(train_tfDict)\n",
        "\n",
        "# Compute idf\n",
        "def computeIDFDict(text_data, wordCount):\n",
        "    # Computes the IDF score of every word in the corpus.\n",
        "    IDFDict = {}\n",
        "    for word in wordCount:\n",
        "        IDFDict[word] = np.log(len(text_data)/ wordCount[word])\n",
        "    return IDFDict\n",
        "\n",
        "test_idfDict = computeIDFDict(texts_test_data, test_wordCount)\n",
        "train_idfDict = computeIDFDict(texts_train_data, train_wordCount)\n",
        "\n",
        "# TF-IDF(term) = TF(term) * IDF(term)\n",
        "def computeTFIDFDict(TFDict, IDFDict):\n",
        "    TFIDFDict = {}\n",
        "    #For each word , we multiply its tf and its idf.\n",
        "    for word in TFDict:\n",
        "        TFIDFDict[word] = TFDict[word] * IDFDict[word]\n",
        "    return TFIDFDict\n",
        "\n",
        "test_tfidfDict = [computeTFIDFDict(test_tfDict_item, test_idfDict) for test_tfDict_item in test_tfDict]\n",
        "train_tfidfDict = [computeTFIDFDict(train_tfDict_item, train_idfDict) for train_tfDict_item in train_tfDict]\n",
        "\n",
        "# Build a matrix, each representing a tf of each word in the text\n",
        "def computeTFIDFVector(text, text_wordCount):\n",
        "    TFIDFVector = [0.0] * len(text_wordCount)\n",
        "\n",
        "    for idx, word in enumerate(text_wordCount):\n",
        "        if word in text:\n",
        "            TFIDFVector[idx] = text[word]\n",
        "    return TFIDFVector\n",
        "\n",
        "test_words = sorted(test_wordCount.keys())\n",
        "train_words = sorted(train_wordCount.keys())\n",
        "test_tfidfVector = [computeTFIDFVector(test_tfidfDict_item, test_words) for test_tfidfDict_item in test_tfidfDict]\n",
        "train_tfidfVector = [computeTFIDFVector(train_tfidfDict_item, test_words) for train_tfidfDict_item in train_tfidfDict]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sphd6eRgTyUv",
        "colab_type": "text"
      },
      "source": [
        "Same stuff about model and optimizers here (or just omit it, if you are using the same model as before)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YOJsI6ZkTyU4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train_tfidf_torch = torch.tensor(train_tfidfVector, dtype=torch.float32)                  ### YOUR CODE HERE\n",
        "X_test_tfidf_torch = torch.tensor(test_tfidfVector, dtype=torch.float32)                    ### YOUR CODE HERE\n",
        "\n",
        "y_train_torch_ = torch.tensor(y_train_, dtype=torch.long)                                     ### YOUR CODE HERE\n",
        "y_test_torch_ = torch.tensor(y_test_, dtype=torch.long)                                       ### YOUR CODE HERE"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7PRFLzW1fr0s",
        "colab_type": "code",
        "outputId": "64c72700-8f21-4019-c039-71e5df6b8aa8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "X_train_bow_torch.shape, X_train_tfidf_torch.shape, "
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(torch.Size([500, 5722]), torch.Size([500, 4946]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "umdgrV_s1_28",
        "colab_type": "code",
        "outputId": "39f9875d-9b47-412b-979a-bf20d09b74af",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 29
        }
      },
      "source": [
        "ASS_1_SIZE = X_train_tfidf_torch.shape[1] #<< Let's agree to disagree\n",
        "ASS_1_SIZE"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4946"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cKH-D4hUdUCn",
        "colab_type": "code",
        "outputId": "a2c4203f-09b7-4b1c-8ee1-b14611948f80",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 275
        }
      },
      "source": [
        "model_ass_1 = nn.Linear(ASS_1_SIZE, 2)\n",
        "optimizer_ass_1 = torch.optim.SGD(model_ass_1.parameters(), lr = 0.02, momentum=0.9)\n",
        "lr_scheduler_ass_1 = ReduceLROnPlateau(optimizer_ass_1)\n",
        "\n",
        "train_model(model_ass_1, \n",
        "        optimizer_ass_1, \n",
        "        lr_scheduler_ass_1, \n",
        "        X_train_tfidf_torch, \n",
        "        y_train_torch_, \n",
        "        X_test_tfidf_torch, \n",
        "        y_test_torch_,\n",
        "        batch_size = 222)"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Linear(in_features=4946, out_features=2, bias=True)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qwyjkO6r5HHH",
        "colab_type": "text"
      },
      "source": [
        "> When i put a test-size less than 0.5 this doesn't work. Dunno why though. *This happens when i experimented with Kselect maybe it has something to do with that. But it was just experimenting and i probably made an error in invoking so mehhh*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2ZcWQZu17w1A",
        "colab_type": "code",
        "outputId": "44ab14ba-9ca6-41c8-b58d-74200c2bb704",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "for name, X, y, model in [\n",
        "    ('train', X_train_tfidf_torch, y_train_, model_ass_1),\n",
        "    ('test ', X_test_tfidf_torch, y_test_, model_ass_1)\n",
        "]:\n",
        "    proba = model(X).detach().cpu().numpy()[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QboW2J3CbN9j",
        "colab_type": "text"
      },
      "source": [
        ">> I tried almost all optimizers and ended on SGD because it looks like i can improve it. And the graph looks more realistic *IMHO*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_ID_C81pTyVC",
        "colab_type": "text"
      },
      "source": [
        "Fit your model to the data. No not hesitate to vary number of iterations, learning rate and so on.\n",
        "\n",
        "_Note: due to very small dataset, increasing the complexity of the network might not be the best idea._"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tcJ2y3wdTyVD",
        "colab_type": "text"
      },
      "source": [
        "### Step 3: Comparing it with Naive Bayes\n",
        "\n",
        "Naive Bayes classifier is a good choice for such small problems. Try to tune it for both BOW and TF-iDF features. Compare the results with Logistic Regression."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K7HWHxYxTyVE",
        "colab_type": "code",
        "outputId": "f3c8193b-b3c0-49df-ba06-62d6ffb996ff",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "from sklearn.naive_bayes import GaussianNB, MultinomialNB\n",
        "from sklearn.metrics import roc_auc_score, roc_curve\n",
        "\n",
        "naive_baye_model = MultinomialNB()\n",
        "naive_baye_model.fit(X_train_bow, y_train)\n",
        "\n",
        "for name, X, y, model in [\n",
        "    ('train', X_train_bow, y_train, naive_baye_model),\n",
        "    ('test ', X_test_bow, y_test, naive_baye_model)\n",
        "]:\n",
        "    proba = model.predict_proba(X)[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()\n",
        "# prediction = torch.tensor(naive_baye_model.fit(X_train_tfidf_torch, y_train_torch_).predict(X_test_tfidf_torch), dtype=torch.long)\n",
        "# print(\"Misclassified %d out of total %d points\"% ((y_test_torch_ != prediction).sum(),X_test_tfidf_torch.shape[0]))"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0OXyEH6_kSTi",
        "colab_type": "code",
        "outputId": "97438ef3-536f-4d98-e000-9bed7beda7fa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "naive_baye_model_1 = MultinomialNB()\n",
        "naive_baye_model_1.fit(train_tfidfVector, y_train_)\n",
        "\n",
        "for name, X, y, model in [\n",
        "    ('train', train_tfidfVector, y_train_, naive_baye_model_1),\n",
        "    ('test ', test_tfidfVector, y_test_, naive_baye_model_1)\n",
        "]:\n",
        "    proba = model.predict_proba(X)[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()\n",
        "\n",
        "# prediction_1 = torch.tensor(naive_baye_model.fit(X_train_bow_torch, y_train_torch).predict(X_test_bow_torch), dtype=torch.long)\n",
        "# print(\"Misclassified %d out of total %d points\"% ((y_test_torch != prediction_1).sum(),X_test_bow_torch.shape[0]))"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MNTxJmt4p5o9",
        "colab_type": "text"
      },
      "source": [
        "*Ca vas maintenant*\n",
        "> I think the model does well enough on the both models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aE9jknnVTyVM",
        "colab_type": "text"
      },
      "source": [
        "### Step 4: Using the external knowledge.\n",
        "\n",
        "Use the `gensim` word2vec pretrained model to translate words into vectors. Use several models with this new encoding technique. Compare the results, share your thoughts."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xxc2wUNkTyVM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "f05f9437-1922-4861-b3a5-4fb57d583888"
      },
      "source": [
        "from gensim.models import Word2Vec\n",
        "import gensim.downloader as downloader\n",
        "\n",
        "gensim_model  = downloader.load('glove-twitter-25')"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[==================================================] 100.0% 104.8/104.8MB downloaded\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/smart_open/smart_open_lib.py:253: UserWarning: This function is deprecated, use smart_open.open instead. See the migration notes for details: https://github.com/RaRe-Technologies/smart_open/blob/master/README.rst#migrating-to-the-new-open-function\n",
            "  'See the migration notes for details: %s' % _MIGRATION_NOTES_URL\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SMro5lWcWjYn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_phrase_embedding(phrase, model):\n",
        "    \"\"\"\n",
        "    Convert phrase to a vector by aggregating it's word embeddings.\n",
        "    \"\"\"\n",
        "    # 1. lowercase phrase\n",
        "    # 2. tokenize phrase\n",
        "    # 3. average word vectors for all words in tokenized phrase\n",
        "    # skip words that are not in model's vocabulary\n",
        "    # if all words are missing from vocabulary, return zeros\n",
        "    \n",
        "    vector = np.zeros([model.vector_size], dtype='float32')\n",
        "    phrase_tokenized = tokenizer.tokenize(phrase.lower()) \n",
        "    phrase_vectors = [model[x] for x in phrase_tokenized if x in model.vocab.keys()]\n",
        "\n",
        "    if len(phrase_vectors) != 0:\n",
        "        vector = np.mean(phrase_vectors, axis=0)\n",
        "\n",
        "    return vector"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tfUDGrAbB-bb",
        "colab_type": "code",
        "outputId": "79357df2-5de6-4c73-eee9-d251570609d4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        }
      },
      "source": [
        "phrase_vectors = np.asarray([get_phrase_embedding(x, gensim_model ) for x in texts])\n",
        "texts[0], phrase_vectors[0], y_train[0], np.unique(y_train)"
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(\"The picture on the article is not of the actor T.R. Knight who is the subject of the article.  Even a basic Google search turns up pictures of the T.R. Knight about which this article is written.  The photo continually being added to the article is, again, not a photo of this article's subject.  Because this photo is not relevant to the subject of the article, it should and will be removed.  Before my edit or those of others who have made the same edit are reverted, please provide a legitimate reason for this image's inclusion in this article.  If a fair-use image of T.R. Knight, the actor and subject of this article, can be found, then by all means it can be added, but random images should not be used as placeholders.\",\n",
              " array([ 0.2080236 , -0.0312066 , -0.05922682, -0.07017404, -0.05980513,\n",
              "        -0.06995879,  1.143351  ,  0.0306216 , -0.19066091, -0.01105928,\n",
              "         0.03959571,  0.24115683, -4.5852275 ,  0.07428471,  0.07573736,\n",
              "        -0.03880551,  0.1525831 , -0.05122429, -0.32870182, -0.27374995,\n",
              "        -0.10397394, -0.04776035, -0.27536592, -0.06197042, -0.3048497 ],\n",
              "       dtype=float32),\n",
              " 0,\n",
              " array([0, 1]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 73
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-0nxzUZPZ7rR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "ceac788d-ce9c-4554-b8f3-ce89babe9459"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "'''  '''\n",
        "svc_model = SVC(gamma = 'auto', probability=True)\n",
        "svc_model.fit(phrase_vectors, target)\n",
        "\n",
        "for name, X, y, model in [('texts', phrase_vectors, target, svc_model)]:\n",
        "    proba = model.predict_proba(X)[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()\n",
        "# SVC?"
      ],
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "934eb5cc-b962-4056-9012-cd1a49ca9282",
        "id": "9D0oIz6Lcq12",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        }
      },
      "source": [
        "new_lr_cls_1 = LogisticRegression()\n",
        "new_lr_cls_1.fit(phrase_vectors, target)\n",
        "\n",
        "for name, X, y, model in [\n",
        "    ('texts', phrase_vectors, target, new_lr_cls_1)\n",
        "]:\n",
        "    proba = model.predict_proba(X)[:, 1]\n",
        "    auc = roc_auc_score(y, proba)\n",
        "    plt.plot(*roc_curve(y, proba)[:2], label='%s AUC=%.4f' % (name, auc))\n",
        "\n",
        "plt.plot([0, 1], [0, 1], '--', color='black',)\n",
        "plt.legend(fontsize='large')\n",
        "plt.grid()"
      ],
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sDE34patdqQ_",
        "colab_type": "text"
      },
      "source": [
        "Logistic regression is better than SVC in this case."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GmZ4VilzTyVY",
        "colab_type": "text"
      },
      "source": [
        "### Step 5: Visualizing the embeddings\n",
        "\n",
        "Finally, let's visualize the embeddings for every comment. One can use just averaged vector or use some more complex models. We recommend using `umap` (`pip install umap-learn`) to map everything to 2-dimensional space. For inspiration one might refer to the [day 14](https://github.com/neychev/harbour_ml2020/tree/master/day14_Unsupervised_learning) practice of the Machine Learning course."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C5k2pPUPd9RP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "outputId": "323e625e-affb-4c3b-c0fd-392ddfc92e4b"
      },
      "source": [
        "!pip install umap-learn"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: umap-learn in /usr/local/lib/python3.6/dist-packages (0.4.4)\n",
            "Requirement already satisfied: tbb in /usr/local/lib/python3.6/dist-packages (from umap-learn) (2020.0.133)\n",
            "Requirement already satisfied: scikit-learn>=0.20 in /usr/local/lib/python3.6/dist-packages (from umap-learn) (0.22.2.post1)\n",
            "Requirement already satisfied: scipy>=1.3.1 in /usr/local/lib/python3.6/dist-packages (from umap-learn) (1.4.1)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.6/dist-packages (from umap-learn) (1.18.5)\n",
            "Requirement already satisfied: numba!=0.47,>=0.46 in /usr/local/lib/python3.6/dist-packages (from umap-learn) (0.48.0)\n",
            "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.6/dist-packages (from scikit-learn>=0.20->umap-learn) (0.15.1)\n",
            "Requirement already satisfied: llvmlite<0.32.0,>=0.31.0dev0 in /usr/local/lib/python3.6/dist-packages (from numba!=0.47,>=0.46->umap-learn) (0.31.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from numba!=0.47,>=0.46->umap-learn) (47.1.1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hpY5ZJ6pfC-u",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import bokeh.models as bm, bokeh.plotting as pl\n",
        "from bokeh.io import output_notebook\n",
        "output_notebook()\n",
        "\n",
        "def draw_vectors(x, y, radius=10, alpha=0.25, color='red',\n",
        "                 width=600, height=400, show=True, **kwargs):\n",
        "    \"\"\" draws an interactive plot for data points with auxilirary info on hover \"\"\"\n",
        "    if isinstance(color, str): color = [color] * len(x)\n",
        "    data_source = bm.ColumnDataSource({ 'x' : x, 'y' : y, 'color': color, **kwargs })\n",
        "\n",
        "    fig = pl.figure(active_scroll='wheel_zoom', width=width, height=height)\n",
        "    fig.scatter('x', 'y', size=radius, color='color', alpha=alpha, source=data_source)\n",
        "\n",
        "    fig.add_tools(bm.HoverTool(tooltips=[(key, \"@\" + key) for key in kwargs.keys()]))\n",
        "    if show: pl.show(fig)\n",
        "    return fig"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TxrYJJVyFC7A",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "outputId": "694db404-9493-4a76-fe78-66201c061a03"
      },
      "source": [
        "import umap\n",
        "\n",
        "phrase_vectors_2d = umap.UMAP(n_neighbors=3,\n",
        "                              metric = 'cosine').fit_transform(phrase_vectors)\n",
        "\n",
        "draw_vectors(phrase_vectors_2d[:, 0], \n",
        "             phrase_vectors_2d[:, 1],\n",
        "             phrase = [phrase[:50] for phrase in texts],\n",
        "             radius = 20)\n",
        "\n",
        "umap.UMAP?\n",
        "# https://www.youtube.com/watch?v=UMqGP8sewAE&list=RDCDWVeVZa-DA&index=15"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "\n",
              "(function(root) {\n",
              "  function now() {\n",
              "    return new Date();\n",
              "  }\n",
              "\n",
              "  var force = true;\n",
              "\n",
              "  if (typeof root._bokeh_onload_callbacks === \"undefined\" || force === true) {\n",
              "    root._bokeh_onload_callbacks = [];\n",
              "    root._bokeh_is_loading = undefined;\n",
              "  }\n",
              "\n",
              "  var JS_MIME_TYPE = 'application/javascript';\n",
              "  var HTML_MIME_TYPE = 'text/html';\n",
              "  var EXEC_MIME_TYPE = 'application/vnd.bokehjs_exec.v0+json';\n",
              "  var CLASS_NAME = 'output_bokeh rendered_html';\n",
              "\n",
              "  /**\n",
              "   * Render data to the DOM node\n",
              "   */\n",
              "  function render(props, node) {\n",
              "    var script = document.createElement(\"script\");\n",
              "    node.appendChild(script);\n",
              "  }\n",
              "\n",
              "  /**\n",
              "   * Handle when an output is cleared or removed\n",
              "   */\n",
              "  function handleClearOutput(event, handle) {\n",
              "    var cell = handle.cell;\n",
              "\n",
              "    var id = cell.output_area._bokeh_element_id;\n",
              "    var server_id = cell.output_area._bokeh_server_id;\n",
              "    // Clean up Bokeh references\n",
              "    if (id != null && id in Bokeh.index) {\n",
              "      Bokeh.index[id].model.document.clear();\n",
              "      delete Bokeh.index[id];\n",
              "    }\n",
              "\n",
              "    if (server_id !== undefined) {\n",
              "      // Clean up Bokeh references\n",
              "      var cmd = \"from bokeh.io.state import curstate; print(curstate().uuid_to_server['\" + server_id + \"'].get_sessions()[0].document.roots[0]._id)\";\n",
              "      cell.notebook.kernel.execute(cmd, {\n",
              "        iopub: {\n",
              "          output: function(msg) {\n",
              "            var id = msg.content.text.trim();\n",
              "            if (id in Bokeh.index) {\n",
              "              Bokeh.index[id].model.document.clear();\n",
              "              delete Bokeh.index[id];\n",
              "            }\n",
              "          }\n",
              "        }\n",
              "      });\n",
              "      // Destroy server and session\n",
              "      var cmd = \"import bokeh.io.notebook as ion; ion.destroy_server('\" + server_id + \"')\";\n",
              "      cell.notebook.kernel.execute(cmd);\n",
              "    }\n",
              "  }\n",
              "\n",
              "  /**\n",
              "   * Handle when a new output is added\n",
              "   */\n",
              "  function handleAddOutput(event, handle) {\n",
              "    var output_area = handle.output_area;\n",
              "    var output = handle.output;\n",
              "\n",
              "    // limit handleAddOutput to display_data with EXEC_MIME_TYPE content only\n",
              "    if ((output.output_type != \"display_data\") || (!output.data.hasOwnProperty(EXEC_MIME_TYPE))) {\n",
              "      return\n",
              "    }\n",
              "\n",
              "    var toinsert = output_area.element.find(\".\" + CLASS_NAME.split(' ')[0]);\n",
              "\n",
              "    if (output.metadata[EXEC_MIME_TYPE][\"id\"] !== undefined) {\n",
              "      toinsert[toinsert.length - 1].firstChild.textContent = output.data[JS_MIME_TYPE];\n",
              "      // store reference to embed id on output_area\n",
              "      output_area._bokeh_element_id = output.metadata[EXEC_MIME_TYPE][\"id\"];\n",
              "    }\n",
              "    if (output.metadata[EXEC_MIME_TYPE][\"server_id\"] !== undefined) {\n",
              "      var bk_div = document.createElement(\"div\");\n",
              "      bk_div.innerHTML = output.data[HTML_MIME_TYPE];\n",
              "      var script_attrs = bk_div.children[0].attributes;\n",
              "      for (var i = 0; i < script_attrs.length; i++) {\n",
              "        toinsert[toinsert.length - 1].firstChild.setAttribute(script_attrs[i].name, script_attrs[i].value);\n",
              "      }\n",
              "      // store reference to server id on output_area\n",
              "      output_area._bokeh_server_id = output.metadata[EXEC_MIME_TYPE][\"server_id\"];\n",
              "    }\n",
              "  }\n",
              "\n",
              "  function register_renderer(events, OutputArea) {\n",
              "\n",
              "    function append_mime(data, metadata, element) {\n",
              "      // create a DOM node to render to\n",
              "      var toinsert = this.create_output_subarea(\n",
              "        metadata,\n",
              "        CLASS_NAME,\n",
              "        EXEC_MIME_TYPE\n",
              "      );\n",
              "      this.keyboard_manager.register_events(toinsert);\n",
              "      // Render to node\n",
              "      var props = {data: data, metadata: metadata[EXEC_MIME_TYPE]};\n",
              "      render(props, toinsert[toinsert.length - 1]);\n",
              "      element.append(toinsert);\n",
              "      return toinsert\n",
              "    }\n",
              "\n",
              "    /* Handle when an output is cleared or removed */\n",
              "    events.on('clear_output.CodeCell', handleClearOutput);\n",
              "    events.on('delete.Cell', handleClearOutput);\n",
              "\n",
              "    /* Handle when a new output is added */\n",
              "    events.on('output_added.OutputArea', handleAddOutput);\n",
              "\n",
              "    /**\n",
              "     * Register the mime type and append_mime function with output_area\n",
              "     */\n",
              "    OutputArea.prototype.register_mime_type(EXEC_MIME_TYPE, append_mime, {\n",
              "      /* Is output safe? */\n",
              "      safe: true,\n",
              "      /* Index of renderer in `output_area.display_order` */\n",
              "      index: 0\n",
              "    });\n",
              "  }\n",
              "\n",
              "  // register the mime type if in Jupyter Notebook environment and previously unregistered\n",
              "  if (root.Jupyter !== undefined) {\n",
              "    var events = require('base/js/events');\n",
              "    var OutputArea = require('notebook/js/outputarea').OutputArea;\n",
              "\n",
              "    if (OutputArea.prototype.mime_types().indexOf(EXEC_MIME_TYPE) == -1) {\n",
              "      register_renderer(events, OutputArea);\n",
              "    }\n",
              "  }\n",
              "\n",
              "  \n",
              "  if (typeof (root._bokeh_timeout) === \"undefined\" || force === true) {\n",
              "    root._bokeh_timeout = Date.now() + 5000;\n",
              "    root._bokeh_failed_load = false;\n",
              "  }\n",
              "\n",
              "  var NB_LOAD_WARNING = {'data': {'text/html':\n",
              "     \"<div style='background-color: #fdd'>\\n\"+\n",
              "     \"<p>\\n\"+\n",
              "     \"BokehJS does not appear to have successfully loaded. If loading BokehJS from CDN, this \\n\"+\n",
              "     \"may be due to a slow or bad network connection. Possible fixes:\\n\"+\n",
              "     \"</p>\\n\"+\n",
              "     \"<ul>\\n\"+\n",
              "     \"<li>re-rerun `output_notebook()` to attempt to load from CDN again, or</li>\\n\"+\n",
              "     \"<li>use INLINE resources instead, as so:</li>\\n\"+\n",
              "     \"</ul>\\n\"+\n",
              "     \"<code>\\n\"+\n",
              "     \"from bokeh.resources import INLINE\\n\"+\n",
              "     \"output_notebook(resources=INLINE)\\n\"+\n",
              "     \"</code>\\n\"+\n",
              "     \"</div>\"}};\n",
              "\n",
              "  function display_loaded() {\n",
              "    var el = document.getElementById(null);\n",
              "    if (el != null) {\n",
              "      el.textContent = \"BokehJS is loading...\";\n",
              "    }\n",
              "    if (root.Bokeh !== undefined) {\n",
              "      if (el != null) {\n",
              "        el.textContent = \"BokehJS \" + root.Bokeh.version + \" successfully loaded.\";\n",
              "      }\n",
              "    } else if (Date.now() < root._bokeh_timeout) {\n",
              "      setTimeout(display_loaded, 100)\n",
              "    }\n",
              "  }\n",
              "\n",
              "\n",
              "  function run_callbacks() {\n",
              "    try {\n",
              "      root._bokeh_onload_callbacks.forEach(function(callback) {\n",
              "        if (callback != null)\n",
              "          callback();\n",
              "      });\n",
              "    } finally {\n",
              "      delete root._bokeh_onload_callbacks\n",
              "    }\n",
              "    console.debug(\"Bokeh: all callbacks have finished\");\n",
              "  }\n",
              "\n",
              "  function load_libs(css_urls, js_urls, callback) {\n",
              "    if (css_urls == null) css_urls = [];\n",
              "    if (js_urls == null) js_urls = [];\n",
              "\n",
              "    root._bokeh_onload_callbacks.push(callback);\n",
              "    if (root._bokeh_is_loading > 0) {\n",
              "      console.debug(\"Bokeh: BokehJS is being loaded, scheduling callback at\", now());\n",
              "      return null;\n",
              "    }\n",
              "    if (js_urls == null || js_urls.length === 0) {\n",
              "      run_callbacks();\n",
              "      return null;\n",
              "    }\n",
              "    console.debug(\"Bokeh: BokehJS not loaded, scheduling load and callback at\", now());\n",
              "    root._bokeh_is_loading = css_urls.length + js_urls.length;\n",
              "\n",
              "    function on_load() {\n",
              "      root._bokeh_is_loading--;\n",
              "      if (root._bokeh_is_loading === 0) {\n",
              "        console.debug(\"Bokeh: all BokehJS libraries/stylesheets loaded\");\n",
              "        run_callbacks()\n",
              "      }\n",
              "    }\n",
              "\n",
              "    function on_error() {\n",
              "      console.error(\"failed to load \" + url);\n",
              "    }\n",
              "\n",
              "    for (var i = 0; i < css_urls.length; i++) {\n",
              "      var url = css_urls[i];\n",
              "      const element = document.createElement(\"link\");\n",
              "      element.onload = on_load;\n",
              "      element.onerror = on_error;\n",
              "      element.rel = \"stylesheet\";\n",
              "      element.type = \"text/css\";\n",
              "      element.href = url;\n",
              "      console.debug(\"Bokeh: injecting link tag for BokehJS stylesheet: \", url);\n",
              "      document.body.appendChild(element);\n",
              "    }\n",
              "\n",
              "    for (var i = 0; i < js_urls.length; i++) {\n",
              "      var url = js_urls[i];\n",
              "      var element = document.createElement('script');\n",
              "      element.onload = on_load;\n",
              "      element.onerror = on_error;\n",
              "      element.async = false;\n",
              "      element.src = url;\n",
              "      console.debug(\"Bokeh: injecting script tag for BokehJS library: \", url);\n",
              "      document.head.appendChild(element);\n",
              "    }\n",
              "  };\n",
              "\n",
              "  function inject_raw_css(css) {\n",
              "    const element = document.createElement(\"style\");\n",
              "    element.appendChild(document.createTextNode(css));\n",
              "    document.body.appendChild(element);\n",
              "  }\n",
              "\n",
              "  \n",
              "  var js_urls = [\"https://cdn.pydata.org/bokeh/release/bokeh-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-widgets-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-tables-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-gl-1.4.0.min.js\"];\n",
              "  var css_urls = [];\n",
              "  \n",
              "\n",
              "  var inline_js = [\n",
              "    function(Bokeh) {\n",
              "      Bokeh.set_log_level(\"info\");\n",
              "    },\n",
              "    function(Bokeh) {\n",
              "    \n",
              "    \n",
              "    }\n",
              "  ];\n",
              "\n",
              "  function run_inline_js() {\n",
              "    \n",
              "    if (root.Bokeh !== undefined || force === true) {\n",
              "      \n",
              "    for (var i = 0; i < inline_js.length; i++) {\n",
              "      inline_js[i].call(root, root.Bokeh);\n",
              "    }\n",
              "    } else if (Date.now() < root._bokeh_timeout) {\n",
              "      setTimeout(run_inline_js, 100);\n",
              "    } else if (!root._bokeh_failed_load) {\n",
              "      console.log(\"Bokeh: BokehJS failed to load within specified timeout.\");\n",
              "      root._bokeh_failed_load = true;\n",
              "    } else if (force !== true) {\n",
              "      var cell = $(document.getElementById(null)).parents('.cell').data().cell;\n",
              "      cell.output_area.append_execute_result(NB_LOAD_WARNING)\n",
              "    }\n",
              "\n",
              "  }\n",
              "\n",
              "  if (root._bokeh_is_loading === 0) {\n",
              "    console.debug(\"Bokeh: BokehJS loaded, going straight to plotting\");\n",
              "    run_inline_js();\n",
              "  } else {\n",
              "    load_libs(css_urls, js_urls, function() {\n",
              "      console.debug(\"Bokeh: BokehJS plotting callback run at\", now());\n",
              "      run_inline_js();\n",
              "    });\n",
              "  }\n",
              "}(window));"
            ],
            "application/vnd.bokehjs_load.v0+json": "\n(function(root) {\n  function now() {\n    return new Date();\n  }\n\n  var force = true;\n\n  if (typeof root._bokeh_onload_callbacks === \"undefined\" || force === true) {\n    root._bokeh_onload_callbacks = [];\n    root._bokeh_is_loading = undefined;\n  }\n\n  \n\n  \n  if (typeof (root._bokeh_timeout) === \"undefined\" || force === true) {\n    root._bokeh_timeout = Date.now() + 5000;\n    root._bokeh_failed_load = false;\n  }\n\n  var NB_LOAD_WARNING = {'data': {'text/html':\n     \"<div style='background-color: #fdd'>\\n\"+\n     \"<p>\\n\"+\n     \"BokehJS does not appear to have successfully loaded. If loading BokehJS from CDN, this \\n\"+\n     \"may be due to a slow or bad network connection. Possible fixes:\\n\"+\n     \"</p>\\n\"+\n     \"<ul>\\n\"+\n     \"<li>re-rerun `output_notebook()` to attempt to load from CDN again, or</li>\\n\"+\n     \"<li>use INLINE resources instead, as so:</li>\\n\"+\n     \"</ul>\\n\"+\n     \"<code>\\n\"+\n     \"from bokeh.resources import INLINE\\n\"+\n     \"output_notebook(resources=INLINE)\\n\"+\n     \"</code>\\n\"+\n     \"</div>\"}};\n\n  function display_loaded() {\n    var el = document.getElementById(null);\n    if (el != null) {\n      el.textContent = \"BokehJS is loading...\";\n    }\n    if (root.Bokeh !== undefined) {\n      if (el != null) {\n        el.textContent = \"BokehJS \" + root.Bokeh.version + \" successfully loaded.\";\n      }\n    } else if (Date.now() < root._bokeh_timeout) {\n      setTimeout(display_loaded, 100)\n    }\n  }\n\n\n  function run_callbacks() {\n    try {\n      root._bokeh_onload_callbacks.forEach(function(callback) {\n        if (callback != null)\n          callback();\n      });\n    } finally {\n      delete root._bokeh_onload_callbacks\n    }\n    console.debug(\"Bokeh: all callbacks have finished\");\n  }\n\n  function load_libs(css_urls, js_urls, callback) {\n    if (css_urls == null) css_urls = [];\n    if (js_urls == null) js_urls = [];\n\n    root._bokeh_onload_callbacks.push(callback);\n    if (root._bokeh_is_loading > 0) {\n      console.debug(\"Bokeh: BokehJS is being loaded, scheduling callback at\", now());\n      return null;\n    }\n    if (js_urls == null || js_urls.length === 0) {\n      run_callbacks();\n      return null;\n    }\n    console.debug(\"Bokeh: BokehJS not loaded, scheduling load and callback at\", now());\n    root._bokeh_is_loading = css_urls.length + js_urls.length;\n\n    function on_load() {\n      root._bokeh_is_loading--;\n      if (root._bokeh_is_loading === 0) {\n        console.debug(\"Bokeh: all BokehJS libraries/stylesheets loaded\");\n        run_callbacks()\n      }\n    }\n\n    function on_error() {\n      console.error(\"failed to load \" + url);\n    }\n\n    for (var i = 0; i < css_urls.length; i++) {\n      var url = css_urls[i];\n      const element = document.createElement(\"link\");\n      element.onload = on_load;\n      element.onerror = on_error;\n      element.rel = \"stylesheet\";\n      element.type = \"text/css\";\n      element.href = url;\n      console.debug(\"Bokeh: injecting link tag for BokehJS stylesheet: \", url);\n      document.body.appendChild(element);\n    }\n\n    for (var i = 0; i < js_urls.length; i++) {\n      var url = js_urls[i];\n      var element = document.createElement('script');\n      element.onload = on_load;\n      element.onerror = on_error;\n      element.async = false;\n      element.src = url;\n      console.debug(\"Bokeh: injecting script tag for BokehJS library: \", url);\n      document.head.appendChild(element);\n    }\n  };\n\n  function inject_raw_css(css) {\n    const element = document.createElement(\"style\");\n    element.appendChild(document.createTextNode(css));\n    document.body.appendChild(element);\n  }\n\n  \n  var js_urls = [\"https://cdn.pydata.org/bokeh/release/bokeh-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-widgets-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-tables-1.4.0.min.js\", \"https://cdn.pydata.org/bokeh/release/bokeh-gl-1.4.0.min.js\"];\n  var css_urls = [];\n  \n\n  var inline_js = [\n    function(Bokeh) {\n      Bokeh.set_log_level(\"info\");\n    },\n    function(Bokeh) {\n    \n    \n    }\n  ];\n\n  function run_inline_js() {\n    \n    if (root.Bokeh !== undefined || force === true) {\n      \n    for (var i = 0; i < inline_js.length; i++) {\n      inline_js[i].call(root, root.Bokeh);\n    }\n    } else if (Date.now() < root._bokeh_timeout) {\n      setTimeout(run_inline_js, 100);\n    } else if (!root._bokeh_failed_load) {\n      console.log(\"Bokeh: BokehJS failed to load within specified timeout.\");\n      root._bokeh_failed_load = true;\n    } else if (force !== true) {\n      var cell = $(document.getElementById(null)).parents('.cell').data().cell;\n      cell.output_area.append_execute_result(NB_LOAD_WARNING)\n    }\n\n  }\n\n  if (root._bokeh_is_loading === 0) {\n    console.debug(\"Bokeh: BokehJS loaded, going straight to plotting\");\n    run_inline_js();\n  } else {\n    load_libs(css_urls, js_urls, function() {\n      console.debug(\"Bokeh: BokehJS plotting callback run at\", now());\n      run_inline_js();\n    });\n  }\n}(window));"
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "\n",
              "\n",
              "\n",
              "\n",
              "\n",
              "  <div class=\"bk-root\" id=\"d1ad7c93-6c1e-4ceb-90fb-759c1b30cdd8\" data-root-id=\"1217\"></div>\n"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "(function(root) {\n",
              "  function embed_document(root) {\n",
              "    \n",
              "  var docs_json = {\"9f29e7bb-8f83-43e3-97d5-f9702adb4848\":{\"roots\":{\"references\":[{\"attributes\":{\"below\":[{\"id\":\"1226\",\"type\":\"LinearAxis\"}],\"center\":[{\"id\":\"1230\",\"type\":\"Grid\"},{\"id\":\"1235\",\"type\":\"Grid\"}],\"left\":[{\"id\":\"1231\",\"type\":\"LinearAxis\"}],\"plot_height\":400,\"renderers\":[{\"id\":\"1252\",\"type\":\"GlyphRenderer\"}],\"title\":{\"id\":\"1275\",\"type\":\"Title\"},\"toolbar\":{\"id\":\"1242\",\"type\":\"Toolbar\"},\"x_range\":{\"id\":\"1218\",\"type\":\"DataRange1d\"},\"x_scale\":{\"id\":\"1222\",\"type\":\"LinearScale\"},\"y_range\":{\"id\":\"1220\",\"type\":\"DataRange1d\"},\"y_scale\":{\"id\":\"1224\",\"type\":\"LinearScale\"}},\"id\":\"1217\",\"subtype\":\"Figure\",\"type\":\"Plot\"},{\"attributes\":{},\"id\":\"1236\",\"type\":\"PanTool\"},{\"attributes\":{},\"id\":\"1237\",\"type\":\"WheelZoomTool\"},{\"attributes\":{\"formatter\":{\"id\":\"1277\",\"type\":\"BasicTickFormatter\"},\"ticker\":{\"id\":\"1227\",\"type\":\"BasicTicker\"}},\"id\":\"1226\",\"type\":\"LinearAxis\"},{\"attributes\":{\"overlay\":{\"id\":\"1280\",\"type\":\"BoxAnnotation\"}},\"id\":\"1238\",\"type\":\"BoxZoomTool\"},{\"attributes\":{},\"id\":\"1239\",\"type\":\"SaveTool\"},{\"attributes\":{},\"id\":\"1222\",\"type\":\"LinearScale\"},{\"attributes\":{},\"id\":\"1281\",\"type\":\"Selection\"},{\"attributes\":{\"ticker\":{\"id\":\"1227\",\"type\":\"BasicTicker\"}},\"id\":\"1230\",\"type\":\"Grid\"},{\"attributes\":{},\"id\":\"1282\",\"type\":\"UnionRenderers\"},{\"attributes\":{\"dimension\":1,\"ticker\":{\"id\":\"1232\",\"type\":\"BasicTicker\"}},\"id\":\"1235\",\"type\":\"Grid\"},{\"attributes\":{\"fill_alpha\":{\"value\":0.1},\"fill_color\":{\"value\":\"#1f77b4\"},\"line_alpha\":{\"value\":0.1},\"line_color\":{\"value\":\"#1f77b4\"},\"size\":{\"units\":\"screen\",\"value\":20},\"x\":{\"field\":\"x\"},\"y\":{\"field\":\"y\"}},\"id\":\"1251\",\"type\":\"Scatter\"},{\"attributes\":{\"bottom_units\":\"screen\",\"fill_alpha\":{\"value\":0.5},\"fill_color\":{\"value\":\"lightgrey\"},\"left_units\":\"screen\",\"level\":\"overlay\",\"line_alpha\":{\"value\":1.0},\"line_color\":{\"value\":\"black\"},\"line_dash\":[4,4],\"line_width\":{\"value\":2},\"render_mode\":\"css\",\"right_units\":\"screen\",\"top_units\":\"screen\"},\"id\":\"1280\",\"type\":\"BoxAnnotation\"},{\"attributes\":{\"fill_alpha\":{\"value\":0.25},\"fill_color\":{\"field\":\"color\"},\"line_alpha\":{\"value\":0.25},\"line_color\":{\"field\":\"color\"},\"size\":{\"units\":\"screen\",\"value\":20},\"x\":{\"field\":\"x\"},\"y\":{\"field\":\"y\"}},\"id\":\"1250\",\"type\":\"Scatter\"},{\"attributes\":{\"text\":\"\"},\"id\":\"1275\",\"type\":\"Title\"},{\"attributes\":{},\"id\":\"1232\",\"type\":\"BasicTicker\"},{\"attributes\":{},\"id\":\"1279\",\"type\":\"BasicTickFormatter\"},{\"attributes\":{\"formatter\":{\"id\":\"1279\",\"type\":\"BasicTickFormatter\"},\"ticker\":{\"id\":\"1232\",\"type\":\"BasicTicker\"}},\"id\":\"1231\",\"type\":\"LinearAxis\"},{\"attributes\":{\"callback\":null,\"tooltips\":[[\"phrase\",\"@phrase\"]]},\"id\":\"1254\",\"type\":\"HoverTool\"},{\"attributes\":{},\"id\":\"1240\",\"type\":\"ResetTool\"},{\"attributes\":{\"callback\":null},\"id\":\"1218\",\"type\":\"DataRange1d\"},{\"attributes\":{},\"id\":\"1277\",\"type\":\"BasicTickFormatter\"},{\"attributes\":{},\"id\":\"1227\",\"type\":\"BasicTicker\"},{\"attributes\":{\"source\":{\"id\":\"1216\",\"type\":\"ColumnDataSource\"}},\"id\":\"1253\",\"type\":\"CDSView\"},{\"attributes\":{\"data_source\":{\"id\":\"1216\",\"type\":\"ColumnDataSource\"},\"glyph\":{\"id\":\"1250\",\"type\":\"Scatter\"},\"hover_glyph\":null,\"muted_glyph\":null,\"nonselection_glyph\":{\"id\":\"1251\",\"type\":\"Scatter\"},\"selection_glyph\":null,\"view\":{\"id\":\"1253\",\"type\":\"CDSView\"}},\"id\":\"1252\",\"type\":\"GlyphRenderer\"},{\"attributes\":{\"callback\":null,\"data\":{\"color\":[\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\",\"red\"],\"phrase\":[\"The picture on the article is not of the actor T.R\",\"Its madness. Shes of Chinese heritage, but JAPANES\",\"Fuck You. Why don't you suck a turd out of my assh\",\"God is dead\\nI don't mean to startle anyone but God\",\"THIS USER IS A PLANT FROM BRUCE PERENS AND GROKLAW\",\"\\\"\\n\\n JSTOR \\n\\nHi there. You're one of the first 100 \",\"It was a lie, a fucking lie.\",\"HA! \\n\\nHA! I still continue to edit yet! Serves ya \",\"\\\"I personally think it was the luckiest goal in NH\",\"I'm no expert, but I think use like that is allowe\",\"\\\"==E.E.S.==\\n\\nPlease do not remove the template at \",\"\\\"\\nMy userspace gets alot of vandalism, so I just h\",\"harassment?? \\n\\nyou know what you don't understand \",\"Your submission at Articles for creation \\n Thank y\",\"2006 (UTC)\\n\\nThe linked article is incorrect - the \",\"a sincere question\\nWhy do you do what you do?\\n\\nsla\",\"If Kenny, JFD, and Freedom skies can have their sa\",\"Hey, you are preaching to the choir here, dude )  \",\"You better block me off for a year or forever Fasl\",\"I will fucking cut your heart out. I will set you \",\"Image\\n\\nI would like to use the second paragraph as\",\"\\\"\\nI am slowly working on many of the mountains of \",\"EFFIGY STUDIOS \\n\\nSorry to report this, but accordi\",\"SUP NIGGA \\n\\nI WUZ JUS CHILLIN WIT MY OG'S ON DA WE\",\"Request \\n\\nDo you have any versions of File:Heteroc\",\"LOADING THAT FUCKING GARBAGE HERE\",\"\\\":Lsi john, I've already invited you to bring this\",\"You are a nazi. \\n\\nYour defense of the Latin Americ\",\"This guy is cool, you are all a bunch of faggy fat\",\", although if you want to cite a reference you cou\",\"Your blatant POV pushing \\n\\nNeither of you guys has\",\"please \\n\\nplease don't be a cunt\",\"WIKIPEDIA AT IT'S FINEST:\\n\\n99.141.246.39 I want yo\",\"Nay, you ARE homophobic and indeed very much so. I\",\"formating problem \\n\\nThe value is supposed to be 17\",\"HAHAHA THAT'LL TEACH YOU TO WP:AGF 94.8.193.32  \\nL\",\"Nah, I'm just gonna go to the beach or pier. It's \",\"FUCKING SUCK YOU LIBERAL BASTARD WHO THINKS MICHAE\",\"Twit \\n\\nYour just another petty fascist.\\nYou clowns\",\"You must be logged in to upload files. \\n\\nSo In oth\",\"\\\"\\n\\n Errors in Listing \\n\\nThere are a number of erro\",\"Whether or not it's scientific, it bears at least \",\"fuck off\\nseriously you tosser have you nothing bet\",\"\\\"\\nKeep honorary members. Collect is laboring under\",\"\\\"\\nWith pleasure, cunt, and I'll leave you to finis\",\"| decline=Nobody on Wikipedia wants your moronic e\",\"suxk my balls \\n\\nsssss\",\"\\\"\\n\\nRfD nomination of User:-Zeus-/Contributions\\nI h\",\"BUT HEY I AM TAY AND I DO HAVE AIDS WTF YOU GUYS A\",\"\\\"\\n\\n new thread at WP:AN that relates to you. \\n\\nI w\",\"\\\"Those who're in advantageous positions are the bi\",\"\\\"\\n Cut the bullshit You are doing personal favor h\",\"\\\"\\n\\nWhilst I don't agree with your reasoning, I am \",\"Also, please not that you are an idiot. Thank you \",\"IM GOING TO KILL YOU \\n\\nI JUST WANT YOU TO KNOW THA\",\"Teahouse Talkback reminder\",\"\\\"\\n\\nFascism\\nYou obviously are quite well read and s\",\"Land Area \\n\\nThe land area mentioned in this articl\",\"Thanks for blocking me. I called Tao an idiot and \",\"What I understand so far, is that I should mind my\",\"And what words do you think you have revelation fr\",\"No, this is not needed \\n\\nIn fact, there is another\",\"\\\"\\nThat is absolutely false.\\nYou posted a \\\"\\\"civilit\",\"See also section\\nI am trying to clean that up per \",\"As of today (January 25, 2009), the Training sente\",\"bollocks to this you ignorant cunt.\",\"Hey F**K T*RD, STOP deleting my page bi*ch. You mu\",\"Are You Kidding Me?? \\nThere is no reliable source \",\"EXCUSE ME\\n\\nWhy do u act like u have dick up ass?\",\"Antaeus Feldspar reverts \\nPlease explain why are y\",\"They seem to be reliable enough for this topic any\",\"support - mostly.  I think there is room for a gen\",\"See it wasn't 5 mins until all my suggestions were\",\"What amuses me is if I'd acted like a gormless new\",\"Its Not inapproprite, its constructive editing, if\",\"Butts \\n\\nScrew my butthole mr krabs\",\"You're doing good work! Have a great weekend. Best\",\"\\\" Go fuck yourselves and hope you never try to \\\"\\\"e\",\"of the Islamophobia page\",\"\\\"\\n\\nMORE ON LAWSUITS\\n\\nI do think that it is time th\",\"<-WHAT A FAGGOT, I'M GLAD HE'S DEAD!!!\",\"JdWolff (the mentally imbalanced loser) wryfully s\",\"BEWARE OF FLYING DICKS, FUDGING NIGGER.\",\"hi \\n\\ni want to assfack you\",\"It is my religious view that she will burn in hell\",\"fuck god youn shit asshole\",\"Fuck you Jasper you know he's a cunt and if you do\",\"\\\"\\n{| style=\\\"\\\"border-spacing:8px;margin:0px -8px\\\"\\\" \",\"Hilleri groves\\nCategory:WelcomeBotResearch\\nHello, \",\", and to suggest that is flabbergastingly arrogant\",\"\\\"\\n\\n Concern: bad faith by POV editor on this page.\",\"PS. Since when is it your responsibility to just d\",\"Maybe they are similar, but you can't call any typ\",\"I fantasize about seeing my girlfriend with anothe\",\"\\\"::Or the disruption is borderline or the sock is \",\"I also choose to reject blatant vandalism and prop\",\"\\\"\\\"\\\"changes\\\"\\\" to the article have related to format\",\"TO BE EVEN CLEARER \\n\\nU IZ A MASTURBATE\",\"Five minutes later?!? Thanks. \\u00a0(my\\u00a0talk)\",\"Sneering, transphobic sarcasm is not my idea of hu\",\"\\\"\\n\\n Season or Series? \\n\\nThe first paragraph states\",\"Single-issue nut-jobs \\n\\nFor some reason, this page\",\"Wikipedia doesn't like it when stupidity of modera\",\"kids incorporated \\n\\nSHUT THE HELL UP.\",\"Apology \\n\\nHey, I apologize for that little inciden\",\"ddduu \\n\\nNeilN lives in Canada (eastern time zone) \",\"\\\"\\nThe Graceful Slick....\\nIs non other than an ungr\",\"\\\"::::::: Malleus, you dirty devil; how can you spe\",\"pile of poop, and not particularly constructive po\",\"No, it doesn't. It needs to be organized.\",\"there is a mistake on the front page, fix it plz. \",\"Croke Park is a GAA arena but since it falls into \",\"Do you enjoy being a terrorist? \\n\\n Why? You hate A\",\"\\\"\\n\\n You currently appear to be engaged in an edit \",\"Cuz this admin is a maggot and slave to is POV.68.\",\"Your  A little insignificant Mother's  Lover\\n\\nyou \",\"\\\"\\n\\nyou, sir, are a moron.  it does not, will not, \",\"People died so this starlet's philanthropy would e\",\"I don't think it will be that quick for all the ra\",\"Please answer me. \\n\\nWhy are you reverting my edits\",\"Why not admit it? The only reason you disagree wit\",\"The hell I care. 202.124.135.206\",\"\\\"\\n\\n Swedish Wikipedia \\n\\nPerhaps you might want to \",\"HEY BITCH STOP TAKING DOWN OUR FUCKING EVENT PAGES\",\". All while real people are spilling real blood an\",\"Weird, I listened to the conversation with Richard\",\"your right bro, your so right.....not! i hate you.\",\"Not really sure what the issue is here \\n\\nWhat is T\",\"Insult to Injury \\n\\nlegobatman3609, I hope you've m\",\"\\\"\\nNo, it's a serious breach of policy, he hasn't a\",\"YOU'RE BLOCKED FOR 700 DAYS. \\n SHUSHING COW\",\"No. Controversy shouldn't be a criterion for censo\",\"\\\"\\n\\nPresumptive arsehole \\\"\\\"Fred\\\"\\\", aren't you. The \",\"BIG YAWN. Stop making ridiculous changes to the ar\",\"a region of the celestial sphere close to the ecli\",\"You guys suck stop faking it.\",\"Why did Vladimir Putin kill them all? \\n\\nIn film Di\",\"\\\"\\n\\nI've got information from fellow wikipedians th\",\"\\\"\\n\\nreligious discrimination\\nwe are struggling to j\",\"Please don't tell me about emotions when you are c\",\"Who do you think you are!!!!!!!!!!!!!!!!!!!! \\n\\nI w\",\"Cut the crap. Only warning.\",\"I'm afraid I still am...\",\"\\\"\\n\\n The future of the article \\n\\nI stripped the art\",\"I'm back\\nCheers, mate\",\"\\\"\\n\\n @\\\"\\\"Dougweller\\\"\\\" Please, don't provoke me to go\",\"Run the reflinks tool \\n\\nCan you run the Reflinks t\",\"\\\"\\n\\n Is that the best you can say? \\n\\nWOW! Making a \",\"\\\"\\nNope. \\\"\\\"Killing by hunger\\\"\\\" is not good English,\",\"Sorry, didn't see that at first.  I just saw that \",\"I haz a question...why are you an idiot XD\",\"\\\"\\n\\nThe TM topic\\nHi IRWolfie and welcome to the TM \",\"\\\"\\n\\n Stephan Schulz is a moron \\nI attest that Steph\",\"\\\"\\n\\n Episode List Layout. \\n\\nI think the episode lis\",\"\\\"\\nA little something for my bestest buddy: Admin T\",\"\\\"\\nGee, 5 minutes of reverting redirects cost you p\",\"I DONT CARE WHAT THE FUCK U DO!! \\n\\nFUCK you u Raci\",\"Article suggestion\\n\\nAs the 'notability query' has \",\"\\\"\\n\\nBy the way, nobody told you to \\\"\\\"edit the page \",\"Well, I really give a damn about a political syste\",\"Moreover, this person has abused his/her discretio\",\"Well done Jvstvs you just convinced an admin that \",\"\\\":In response to the interwiki and internal incomi\",\"For sanity's sake I hope santa gets you a new hobb\",\"August 2011 \\n Please do not add unsourced content,\",\"I haven't.  And I'm right between books right now.\",\"So now you are fabricating sources as well? I sugg\",\"This is David Speaking.\",\"Piss off \\n\\nHomo, no one wants to hear from you, al\",\"\\\"Ha, ha, ha. Really now. Please do not change the \",\"i am willing to accept under your conditions, but \",\"Dont delete talk page entries fucking asshole. You\",\"Or, the Arbcom could always simultaneously accept \",\"\\\"\\n\\nSo, you point out the contradiction between the\",\"Go to hell, you power-spinning, pompous arsewipe.\",\"\\\"\\n\\n How about an RFC then? \\n\\nCan we try an RFC?  I\",\"\\\"\\nElizabeth I was referred to as Elizabeth I ever \",\"\\\"\\n\\nComment\\n\\\"\\\"But it's easier to be lazy and accuse\",\"Screw you.149.254.186.158\",\"\\\"\\nNote the phrase \\\"\\\"near the common\\\"\\\"  TALK  \\\"\",\"Way too much speculation here. For the purposes of\",\"Dear Piotrus,\\n\\nhe's a stupid illiterate from Bucha\",\"\\\"\\n\\nExcuse me for coming on hard, but the thought o\",\"Note: I've placed merge tags on the articles, and \",\"\\\"\\nAh, Jeanne la spirituelle! (spirituelle = \\\"\\\"witt\",\"Good a.m. - I did leave a note with the tea pot fe\",\"Fight Club! F**k Yeeaaaaahh!!!\",\"do so sexually with me\",\"SO why did you change my shit\",\"\\\"\\nPS: Virtually the entire body of the Hope articl\",\"Headline text OMFG NO NO NO NO NO NO OMFG NO NO!!!\",\"\\\", 4 September 2014 (UTC)\\n My apologies for not re\",\"User Syrthiss is a mot*herfuc*ker assh*ole ha ha h\",\"Would looking at the source code not be original r\",\"It was signed Moron - what part of that didn't you\",\", and welcome to Wikipedia! Thank you for your con\",\"FundRasing my Ass. \\n\\nSetting aside the protypical \",\"I HAVE NO LIFE, AND IF I DON'T GET TO EVER HAVE SE\",\"Yes, complain to your rabbi and then go shoot some\",\"You watch me, soon your I.P Address will be banned\",\"He was a racial Jew, you fool. If you want to argu\",\"will not be intentionally inconstructive}}\",\"Good news\\ni would replace the shit that is the rea\",\"Does it matter honestly?\",\"Semi-protected edit request on 12 October 2014 Why\",\"Why do you love hate? \\n\\nAll I want to do is inform\",\"Maybe you should quit being a douche bag, and then\",\"based on your own request\",\"I applued the fact that you don't work with any on\",\"pretending to be anti-fascist\",\"He is too, as well as being very sexy and attracti\",\"\\\"\\n\\nIt seems to me, to use a Christian metaphor, th\",\"Proposed approach \\n\\nThis section is intended to be\",\"|politician-work-group = yes \\n|needs-photo = yes\\n|\",\"Small minded and ignorant Wiki contributors. \\n\\nIt \",\"\\\"\\n\\n My order for Mother's Day \\n\\nEarlier today I (I\",\"|decline=Absolutely. Adding links in conflict of i\",\"Surname extinction \\nHi. I do not agree with you on\",\"Bloody Idiot \\n\\nYou Bloody Freak, Do not Bicker and\",\"it should be biased against those soulless fuckers\",\"Nice edit\\n\\nHi Elizabeth, nice reverts on my edits \",\"\\\"\\n\\nNo Rodhullandemu, as I've proven in my posts, i\",\"\\\"\\n\\n Livery Names \\n\\nTo user/IP address 78.146.90.16\",\"\\\"\\\"\\\", except for you because you know nothing about\",\"Hi, Please sign your posts\\nAs a courtesy for other\",\"\\\"*E-dog, on another page, you wrote: \\n\\nThanks for \",\"I removed it for you. Is this what you were lookin\",\"Oh, shit! I've been blocked! Whatever will I do? *\",\"Andy Swan (US) Deleted \\n\\nWhy did you delete Andy S\",\"You better fuck off \\n\\nRuddy shit fuck!\",\"Samantha555, in case you haven't noticed, this pow\",\"I see the Bad Idea Inc. thing too. Surely that's n\",\"January 2010 \\n Please stop your disruptive editing\",\"Rape. \\n\\nNaconkantari is DEAD, or soon will be... \\n\",\"\\\"\\n\\n Stolen funds \\nA big claim like \\\"\\\"Most of the m\",\"many unsecured wi-fi hot spots, so many idiot wiki\",\"See also at Wikipedia_talk:Manual_of_Style/Dates_a\",\"Batteries \\n\\nHow many D batteries can you put in yo\",\"I have been ill from 2012 onwards, but intend to g\",\"Well, that's a disappointment, re-blocked user.\",\"Hey Fat and Beardy \\n\\nI check the Sock Puppet page \",\"That may be appalling but what's really appalling \",\"Hooray! You created your Teahouse profile!\",\"Niggers \\n\\nniggers everywhere69.233.92.38\",\"I noticed you post a lot about taking dumps in the\",\"my page\\n\\ncan you remove the message from my user p\",\"\\\"\\n\\n \\\"\\\"SCIENTIFIC\\\"\\\" SKEPTIC ???! \\n\\nHe never went to\",\"For your information: you\\u00b4re an idiot too!! and a \",\"RACIST \\n\\nnow you know who you are and where you be\",\"etc CM Punk DID adopt the GO TO SLEEP from KENTA f\",\"Fartsalot56 says f**k you motherclucker!!\",\"Pig Fenian \\n\\nNice wrist action boys, keep it up.\",\"\\\"\\n\\n Marketing crap \\n\\\"\\\"2006 - when calls should be \",\"Thank you so much sir. It's a very strong feeling \",\"England Society (Keele University) \\n\\nRegan, You ha\",\"\\\"\\n\\nRe: Martin muljana\\nThis is not German.  Why wou\",\"I am going to eat your toes.\",\"Warning \\nDo not tell me what to do or you will be \",\", you are making a fool of yourself!! This is not \",\"\\\"\\n  The Gay Barnstar F\\u00e6, I am pleased to offer you\",\"i wanna fuck you \\n\\nfuck u fuck u fuck u\",\"religious conversion\\n\\nhow come there is no refernc\",\"I had not seen that actually.  This user is really\",\"You have uploaded the damaged file, I had to rever\",\"\\\"\\n\\nHello \\\"\\\"Alf\\\"\\\"\\nIt's very easy to call someone a \",\"Whoever used 'she' instead of neutral they is a li\",\"ur  msg\\ni wasnt tryin to get\\n rid of it at all.. i\",\"\\\"\\nWelcome!\\n\\nHello, , and welcome to Wikipedia! Tha\",\"any source? what source? cumans(kipchaks) and late\",\"There's a better answer to the question: the peopl\",\"\\\"\\n\\ni find it funny that Mexia, if prounounced corr\",\"Ignorant & Unhelpful \\n\\ni've made numerous valid st\",\"\\\"\\n No name-calling on my part, though I have been \",\"Can I also mention that if you are going to tag WP\",\"\\\"\\n\\nWelcome!\\n\\nHello, and welcome to Wikipedia! Than\",\"\\\"\\n\\nfirst. come on. I should be ashamed because I l\",\"I apologise for that. I was reverting quite abit o\",\"\\\"\\nI'm sorry, I don't understand.  What is the \\\"\\\"it\",\"Don \\n\\nNews to me.  I'll file it for future use, bu\",\"\\\"\\n\\n Mother Jones \\n\\nI am puzzled why you removed:\\n\\n\",\"\\\"\\n\\nSUSPICIOUS MASS EDITS-IS ANY ADMINISTRATOR FOLL\",\"Vaginas \\n\\nDo you have a problem with vaginas? 68.0\",\"REDIRECT Talk:Mi'kmaq hieroglyphic writing\",\"Take yourwarning and shake salt on it and shove it\",\"o wait i r just a big ideot wid a psp slim i r so \",\"\\\"\\n\\n Please do not vandalize pages, as you did with\",\"YO B1TCH! \\n\\nyou absolute fuk1ng little b1tch you w\",\"Sidenote: Xizer is not a racist. Calling this guy \",\"you're not an Editor\\n\\nYou are a stalker with with \",\"Hey, moron, why don't you read the entire edit req\",\"dead \\nnice one dude!!\",\"I'm very concerned about the attempt to edit-war c\",\"ocean. In the end, Michael tragically hangs himsel\",\"Last warning \\n\\nPlease stop. If you continue to van\",\"Speedy deletion of Template:Omani political partie\",\"\\\"\\n\\n2010 Pretoria runaway\\nHappy to be of assistance\",\"\\\"\\n\\n\\\"\\\"\\\"\\\"the generally accepted cause of the death o\",\"Timeslot hopping \\n\\nSomething should be said in the\",\"\\\"\\n\\nAnd something else, \\\"\\\"I, too, don't like the Or\",\"PEOPLE ON THIS SITE SUCK  \\n\\nFUCKING BAN ME. GO ON \",\"HOW DARE YOU ABUSE AUTOGRAPH'S PAGE, YOU FAG!!!! D\",\"Attention Biaatch=\\nI know you want to be me, but u\",\"Paoli Dam \\nHere the whole article has been rewritt\",\"And you might want to check the spelling of douche\",\"\\\"\\n\\n On the goodness of description and evil of pre\",\"I love penis \\n I love penis \\n I love penis \\n I lov\",\"and you're gay too, faggot\",\"Hebrew Name of Lydia \\nAppologies to Til Eulenspieg\",\"I'm in favour of your proposal too Javits. Now I u\",\"this user needs to get a life lol \\n\\nlol get a real\",\"This individual is either mentally ill, or is a cr\",\"\\\"\\nYOu can't use the apple ARTICLE to reference tha\",\"You first piss off a user and then pretend to advi\",\"LEAVE 4 A FUKIN SNAK AND I RETRN WITH A FUKIN BLCK\",\"Go fuck yourself you faggot -\",\"\\\"\\n\\n Point taken. Page 35 of the PDF guide to Surat\",\"Hey Gwernol, go suck a dick you fuckin piece of sh\",\"\\\"\\n\\nAWARD OF THE PURPLE HEAD TO TOMSTAR81\\n\\nWe of Wi\",\"Discussion\\nAny additional comments:\\nI noticed that\",\"So it's OK with someone with NO first-hand knowled\",\"My user page \\n\\nHey jews bitch!! I was expecting th\",\"The internet, as a whole, is less biased than any \",\"Go to the devil, cultist! You don't own this place\",\"\\\"\\n\\nRpo.castro ... you're just another Kante4's ali\",\"ya'll stupid muthafuckas need to mind ya'll own bu\",\"Although not officially released, they deserve an \",\"Diannaa wants us to be blocked and she's far from \",\"heY TOdds! Quick Q? \\n\\nWhy Are You So Gay\",\"\\\"\\nMore weasel or pov wording. The paragraph beginn\",\"Please stop vandalising my writings in the dog art\",\"What is wrong with a neutrally coloured map showin\",\"\\\"\\n\\nMinnesota Meetup\\n\\n2009Proposed date: Saturday, \",\"http://en.wikipedia.org/wiki/Wikipedia:LIVING#Priv\",\"\\\"== Bhagavad Gita Talk page ==\\n\\nHello! Welcome to \",\"\\\"::You ought to get cool, it's not like I held up \",\"FINE \\n\\nCUNT FUCK FAGGOT JIZZ BITCH CUNT WHORE SHIT\",\"SO WHAT IF I LOVE MICHAEL JACKSON@@@@@@@@@@@@@@@@@\",\"Talk:Chelsea Manning \\n\\nThis talk page is for discu\",\"Actually, there's no reason why you have to be the\",\"International Property Rights Index \\n\\nDoes anyone \",\"\\\"\\\"\\\"John Henry Haines was born on April 20, 1420, i\",\"Unblock Request To Wikieditor 006 and Ale_jrb \\n\\nHo\",\"What RealDealBillMcNeal really is \\n\\nRealDealBillMc\",\"A misery=\\nYou have made my life a misery on Wikipe\",\"Why are you such a jerk?\",\"\\\"\\n\\nEWWWWWWW. It's not Tim, dummy.\\n\\n~ \\\"\",\"Like i would listen to you, you are a loony dickhe\",\"\\\"::::::::: Right, I've been working butt off to tr\",\"\\\"\\n\\nI did cite the CIA resource, \\\"\\\"https://www.cia.\",\"\\\".  \\\"\\\"For the gay market\\\"\\\" makes no sense and seem\",\"Pontic Greeks? \\n\\nI wonder if there is any connecti\",\"yo man your pissing me off\",\"You're a motherfucking prick! \\n\\nSo, you don't agre\",\"To me a nemonic is supposed to provide an easy way\",\"Your opinion is unfair and biased.\",\"\\\"\\nThat is a false \\\"\\\"balance\\\"\\\", similar to cutting \",\"12th Doctor \\n\\nLOOK! I HAVE HAD ENOUGH YOU SUPREMAC\",\"This article seems to be just a dictionary definit\",\"\\\"\\n\\n Why \\\"\\\"Citation Needed\\\"\\\" everywhere? \\n\\nIs there\",\"There should be no mystery. This article is List o\",\"professional pedophile and rapist of cats.\",\"Napoleon Factor \\n\\nThis movie is like Napoleon Dyna\",\"Hey dumb dick ==\\n\\nI left Michael a message, and yo\",\"oh wow thanks for defending that guy who deleted e\",\"burn in hell motherfucker\",\"YOUR MOM's Va jay jay\",\"Stay out of private business between me and Fridge\",\"I'm sorry, but I have only been reacting to other \",\"going to have to agree with the first guy this is \",\"FURTHERMORE.... I HAVE JUST VISITED RAGIB'S PAGE A\",\"pussy bitch. delete things\",\"Welcome to the Nazi website.  Where's Hitler?\",\"Febuary 2009 \\n\\nI didn't vandalize the fucking page\",\"\\\"\\n Your submission at Articles for creation \\n Herb\",\"Sexuality \\n\\nMabye it should be added that he is ac\",\"Very well then, it shall be. Cheers!\",\"\\\"Oh, and don't you dare ever come to my talk page \",\"spam \\n\\nim spamming you, cause you're a cunt and de\",\"Silliness\\n\\nYou did something silly. Please don't b\",\"\\\": The proverb is commonly cited as a truism, i.e.\",\"Niggus luv dem wyt biches \\nfa reel\",\"JUST MIND YOUR OWN FUCKIN BUSINESS ASSHOLE\\n\\ntion\",\"Does this mean that Wagner will Try to get raises \",\"\\\"\\nIf you do, they will be deleted. \\u0394T The only con\",\"Bodhidarama new section \\n\\nHe has to be added to th\",\"REDIRECT Talk:The Suppression of Religious Houses \",\"Harry the Dirty Dog  - CONFLICT OF INTEREST \\n\\nHarr\",\"Thanks for your kind comments - I'm under attack b\",\"editor responsible for the change\",\"So you admit to forgetting about Melbourne existin\",\"Dear DoorsAjar - Please tell BongWarrior that the \",\"oh why dear johnny boy part three \\n\\nshit im gonna \",\"Why? \\n\\nWhy did you do that?\",\"\\\"\\n\\nAllow me to reiterate: whether the intro is \\\"\\\"c\",\"I haven't called you names, but if you want to sta\",\"henry lee lucas was my dad im pretty sure i know t\",\"Page Redirect \\nI would like this page to exist aga\",\"Stick it up your ass \\n\\na video for ...............\",\"Image size \\n\\nKindly go to Nathan Altman, Mary Cass\",\"why am I not aloud to put whatever i want on my us\",\"REDIRECT Talk:Athanasiou Diakou Street\",\"You changed it back idiot, I did leave my reasons \",\"\\\"\\n\\n Real or fiction? \\n\\n\\\"\\\"The Mascots are the world\",\"OK, not for now then. This is headed to some form \",\"What the hell? \\nWoah.. what was this edit about? -\",\"Check this page in wikipedia for false traits and \",\"Conversation with me \\n\\nSeems like you forgot conve\",\"bring it \\n\\nhello tommy remember me? im back asshol\",\"Of course it's also possible that you are, in fact\",\"keep this shit up one day something bad will happe\",\"\\\"\\n\\n Sir etc in sequence \\n\\nSame goes for Lord/Lady \",\"GET SOME GLASSES YOU MENTAL FAGOTs, IM Calling you\",\"Good Morning America\",\"It is not an improvement. There is longstanding co\",\"u r sooooooooooooo gay\",\"Monster attacking Troy \\n\\nIt says here that the mon\",\"Yarmouk camp ==\\n\\nNews from Yarmouk camp reported t\",\"\\\"\\n{| style=\\\"\\\"background-color:#F5FFFA; padding:0;\\\"\",\"\\\"\\n Your submission at Articles for creation \\n Nata\",\"Formal Definition \\n\\nThe Dunning-Kruger effect is t\",\"\\\"\\\"\\\"Stars and Bars\\\"\\\" only applies to the Flag of th\",\"\\\" May 2009 (UTC)\\nI did it because although \\\"\\\"pregn\",\"Yes you stupid asshole, it does count. Just becaus\",\"\\\"\\n\\n Fansites \\nPer WP: External Links, these should\",\"\\\"\\n\\nAwww thats cute Magog but lets get real DS just\",\"Cirkey!  It would seem she is a bit more than her \",\"r u Bigees cock sock?\",\"Words to Watch \\n\\nSlim and I are ready to take it l\",\"Seriously, stop. There was no Isaias last year. \\u266b \",\"Your mom plays gay music.. linkin park is the shit\",\"you fucking little retard!!!\",\"Dude did you see what he wront on my page in the h\",\"HIS INFERNAL MAJESTY rock alternative????? \\n\\nalter\",\"check the links on the page. ( )\",\"So basically you treat me like shit and ban me fro\",\"\\\"\\n\\n\\\"\\\"Bitch, suck my dick before I slap you with it\",\"Personally, I'd say, yes it is. In the absence of \",\"I suspect Calabe is a troll.\",\"do I talk so much shit. The second question is wha\",\"\\\"\\nsources disagree hirshhorn says: \\\"\\\"Three-Piece N\",\"Piss off you sanctimonious weed. You forget that 1\",\"Stop vandalising people's hard work. You don't kno\",\"Considering the fact that I already had a warning \",\"Oh wait, you're an American asshole... Let me expl\",\"WHY THE FUCK ARE THESE TWO THE SAME? THEY ARE INDI\",\"Notability \\n\\nSee the references for notability. Do\",\"Not sure, but they seemed to be when I was working\",\", 21 April 2008 (UTC)\\n\\nWhat the fuck are you talki\",\"Just another bitter Zionist pig. Leave Nableezy al\",\"get lost tommy u gay fucker\",\"Are you a fool? \\n\\nI am sorry, but you seem to be a\",\"Why should we users have to create a user name or \",\"Goal Count\\n\\nI don't understand people that don't u\",\"\\\"\\n\\n A knife for you! \\n\\n  A knife for you! One of t\",\"\\\"\\n\\n That wasn't nice \\n\\nI am reposting this to piss\",\"Hi! \\n\\nHow come you're gay?\",\"m\\n\\nWhatever. You're an stupid redneck hillbilly wh\",\"Even first year students would not produce such ba\",\"\\\"\\n\\n Mangojuice, I am not against articles on coats\",\"AfD voting pattern \\n\\nHi Unrepentant, \\n\\nI've notice\",\"Hi \\n\\nHi,\\nI know Moukity is going around claiming t\",\"P.S don't talk to NeilN he called you and me stupi\",\"YOU TOO JEFF G, FUCKIN HOMOS.\",\"Regarding Your Message \\n\\nCan you please be more sp\",\"I want to know how long i have to wait before i ca\",\"test for time \\n\\nTime is\",\"\\\"\\n\\nunder Wikipediapoliciy harrasment is \\\"\\\"Harassme\",\"GTFO U LITTLE NOOBLET\",\"\\\"\\nComment It's official: this stays as \\\"\\\"Hole in o\",\"don't care. Come kill me c'mon!!!!!\",\"File:GoogleChromeWindows5037538.png\\nI have tagged \",\"\\\"\\n I don't think you counted right, because per WP\",\"help \\n\\ni cant find anything on this mold i need he\",\"I'm not talking about wikipedia, I'm talking about\",\"Lamellae \\n\\nLamellae (cf lamina) are structures lik\",\"ISIS massacred many civilians, including children.\",\"Stop Fucking men up the arse James! \\nStop doing it\",\"I'm sick of wikipedia. It's full of illiterates an\",\"this is a stupid idea. all it is doing is adding j\",\"1:55, 20 March 2006 (UTC)\",\"Lol fred, now why on earth would I want to be unbl\",\"It's better than nothing.\",\"African Black penis size compared to white penises\",\"Beagel you have not justified the removal of sourc\",\"forst of all get off your fat ass and get a job! s\",\"Which bit of 'terrorist' do you not understand? Ha\",\"hgfhfghdghgfgdh \\n\\nI hope you will suck my piss, fu\",\"Hanibal911You're a bastard Pro-Assad. The rebels A\",\"guess what ==\\n\\nyour retarded\\n\\n==\",\"\\\", 24 November 2011 (UTC)\\n\\nMy apology if you find \",\"npov\\nYou yell at me about NPOV, but your the fucki\",\"Orangemarlin, you're an asshole and that's all the\",\"\\\"\\n\\nCentury formats\\nRe: Alcohol laws of New Jersey.\",\"Would you do me a favor and STOP FUCKING WITH MY I\",\"well, another proof that wikipedia is (over)run by\",\". Crzrussian, slit your wrists\",\"\\\"\\n\\n Image caption - Homosexual? \\n\\nThe second image\",\"Hello \\n\\nHi, Ursasapien. I was looking through WP:B\",\"I assume that you have never met foreigners. What \",\"\\\"\\nUnsurprisingly, trimming it down to only gaffes \",\"VfD\\nOn May 4, 2005, this article was nominated for\",\"See page 48 of the Freeh report for Freeh's explan\",\"\\\"\\n\\n This is of no consequence, I have listed my so\",\"nigger \\n\\n{{unblock|reason=Liar 98.162.163.72  }\",\"This is what i mena SOME TOTAL BULLYING LITTLE HIT\",\"uhhhh YOU DON'T OWN THIS SITE. SO DON'T TOUCH MY A\",\"Idiocy\\n\\nI made a number of changes, clearly descri\",\"blow me, you facist douche\",\"dont show me attitude i give u standing ovation of\",\"\\\"\\n\\nSpeedy deletion of Talk:The distance from top t\",\"Well... \\nYou could for instance beef up 'this with\",\"And there you go again, tell me the most important\",\"\\\"::::I note  and friends are now organising aperso\",\"\\\"\\n Mark identifies Jesus as \\\"\\\"the son of the carpe\",\"hey \\n\\nyou r a douche\",\"block me i don't give a shit\",\"You don't want the truth?? She IS and WAS a tramp.\",\"Blind as bats \\n\\nNot one of you has seen what I hav\",\"Rklaton is faggot and loves to suck cock and get h\",\"Oh Sh*t! I'm shaking in my boots, big bad admin! G\",\"Piotrus' and propaganda\",\"You need a life \\n\\nREALLY BAD......................\",\"Who the HELL do you think you are? \\n\\nHow dare you \",\"HiDrNick  \\nPresent for you fatty. Relax. don't get\",\"SkyTeam/KLM\\nNot a problem. I figured I'd just pree\",\"now that he's blocked me so that I can't respond t\",\"I dont care I just want unblocked Damn it. After m\",\"And on another note, what right does he have to se\",\"I am not attacking you man, i'm just sayinmg who y\",\"Thanks. The last group wasn't converted yet. For t\",\"suck my dick irish guy you faggot\",\"Ah, so it should be assessed as start.\",\"You sound like you are sexually frustrated. When d\",\"Oh, I'm scared.  Who made you the Wikipedia police\",\"\\\"\\nAbove user has no en talk page, but does have on\",\"do not tell me to delete stuff you stinker!!!!!!!!\",\"Chavez sure isnt a hero. He's classic scum. And yo\",\"Adolf... \\n\\nOne of these days gonna come pay you a \",\"\\\"\\n Forgive me if I'm wrong, BlazikenMaster, but yo\",\"\\\"\\n\\n Russians in Georgia \\n\\nFirst, here's some stati\",\"I agree \\nAt least I can't get blocked for typing t\",\"Will do Maralia. I edited as I saw it, but I suppo\",\"Miz \\n\\nCut the shit, son. Stop reverting the change\",\", 11 March 2012 (UTC)\\n They can be fixed ;-) Jane \",\"\\\"==GA Review==\\n:This review is transcluded from Ta\",\"Anodea Judith - using  and\",\"\\\"\\n\\nYes! Thank you Salix Alba! This is what I mean.\",\"other diseases and\\n               misfortunes, who\",\"Mermaid: The Body Found rewrite \\n\\nHi. Thanks for m\",\"Now who's all up in arms about cultural minutiae. \",\"Are you a Jew \\n\\nAre you a jew? Don't pollute Obama\",\"are you retarded? \\n\\nhow did I attack an editor? Gr\",\"hey dickwads. stop hijacking this wiki. these are \",\"\\\"\\n\\n There should at least be a mention of this alt\",\"Megadeth album article\\nThis clearly isn't the firs\",\"Fuck shit kwa kwa kwa!\",\"suicides\\n\\njust looking at this article make me wan\",\"!\\n\\nIt is now inside your bones, your head. Haunt y\",\"Free childcare\\n\\nThis should mention that the main \",\"Ed Cole\\n\\nFuck you and ed cole...Harry Sweoger, New\",\"Nazis murdered six million of my people. Why shoul\",\"], as well as [http://en.wikipedia.org/w/index.php\",\", 18 January 2008 (UTC)\\n\\nTry Eppy's page. That's a\",\"I removed something written by obnoxious teenagers\",\"\\\"\\nP.S.: regarding the word \\\"\\\"orthodox\\\"\\\", a word's \",\"\\\"\\n\\n The 24 hour block that you placed on my accoun\",\"A turd for you! \\n\\nFor all your hard work. Found th\",\"Contact us to any inquiries on my contact No.93230\",\"\\\"\\n\\n I agree, the map is very inacurate, it clearly\",\"== GO DIE YOU WORTHLESS COCKSUCKING FAGGOT ==\",\"Get a reality grip! \\n\\nYou suck, try to be responsi\",\"Fuck off!!!!- () Wolfkeeper (Talk)\",\"\\\"== Please clarify ==\\n\\n\\\"\\\"According to the 2006 cen\",\"I did tell her politely where to place her head in\",\"You're mad?. Let me complete this article to this \",\"6D vs 7D \\n\\nregarding my 6D vs 7D 'undo', see my re\",\"\\\"\\n\\nLiz Jones (born 5 September 1958Anne, I was bor\",\"and Divinity. They are seeking the Light! Anyway, \",\"Why is\\n\\nnot in the template:IPA naviagtion?\",\"THE PERSON WHOSE EDITS YOU REVERTED \\n\\nHY. \\n\\nI JUST\",\"HELLO MR POV PUSHER\\nWhy do you hide like a cowerin\",\"Nobody likes you JEff. You look like a pedaphile.\",\"{{unblock|Hello, I am writing this to request that\",\"\\\"\\n\\n As for \\\"\\\"some\\\"\\\", consensus is achieved through\",\"I'ma smack ya upside da head wit a shovel \\n\\nI'm ta\",\"acting\\nfrom now on nobody change my edit of honor \",\"Fuck you Guoguo12! You have no life and no friends\",\"We need to talk... \\n\\nDo you suck dick private? BUL\",\"Get a real job \\n\\nBitch\",\"\\\"Mena appears on lots of goddess-ey webpages, but \",\"I attempted some changes in punctuation and gramma\",\"SarekOfVulcan RFA \\n\\nThank you for !voting on my Rf\",\"\\\"\\nI dont know how to message u without disrupting \",\"Actually, as I considered it, it seemed an issue t\",\"Premiership winning captain \\nI am following Wikipe\",\"Mongo, I demand a clear and clean apology for this\",\", 19 June 2008 (UTC)\\nlast warning? i didn't got th\",\"tocBOOTHY IS A FUCKER\",\"WOE TO YOU CPTNONON(you did not passed, so DON'T P\",\"Both of you are useless cunts and vandals who  who\",\"\\\"\\n\\n Shameful advert \\n\\nHi there. I notice you're in\",\"October 2009 (UTC)\\n\\nYou really are a stupid person\",\"\\\"\\nDan, we are on equal footing here. What I was re\",\"How dare you do that?! How would you like it if I \",\"\\\"\\n\\n Offensive words!!!!! \\n\\nThe second paragraph st\",\"I cannot resist another man's cock! I love it when\",\"ECW new, Gangrel/Jannetty, valet, Bischoff \\n\\nHey s\",\"Internal classification of the Romance languages \\n\",\"\\\"AIG Loan & IQ==\\nShould facts should be added abou\",\"Hey can you please stop it bugs me when some one d\",\"Edit summary\\nPlease leave an edit summary to let o\",\"Merge? Why should this be merged with Mr. T? He's \",\"Stop fucking up Wikipedia, faggot.\",\"Fuck off \\n\\nyou can fuck right off mate!\",\"Hi I have found some more links for you http://www\",\"American Jews \\n\\nThe next time you restore your cha\",\"\\\"\\n}}\\n|style=\\\"\\\"font-size: 85%\\\"\\\"|This is a Wikipedia\",\"Well done, u pussbag! Have a cookie!\\nI bet you peo\",\"\\\"I do not agree that what I am doing is \\\"\\\"besmirch\",\"Mean ferret \\n\\nmeanie ferret!\",\"ASSHAT FUCKING STUPID ASSHOLE PAIR OF CLOWNSHOES\",\"One puzzle being why (if Robert Jenkin is innocent\",\"REDIRECT Talk:IPA vowel chart with audio\",\"I almost forgot. A few comments on your remarks ab\",\"For instance: The Government of BiH declared indep\",\"I am sure wikipedia can see through this obvious a\",\"probably that you are a german... fucking germans,\",\"That can equally be said of \\u2018\\u00f1\\u2019 and \\u2018\\u00e7\\u2019; however, \",\"OK, I have just put another edit on the article pa\",\"yes hello, i am a college student fluent in englis\",\"When non-free media is used on Wikipedia, a justif\",\"A question for EJF\\n\\nWhy is it you persist in delet\",\"Hey donkey \\n\\nHey donkey, your Dad is coming.\",\"Who The F*ck Put You In Charge\\n\\nI create an accoun\",\"Illegal zionist entity \\n\\nhi, 58 country not recogn\",\"\\\"\\n\\n Logo for Simple: \\n\\nCan you put together a logo\",\"Fuck you you dumb sack of shit\",\"Support. I don't see how the reasoning would be ir\",\"dude, why did you erase my thing?\\n\\nwhat's up with \",\"|decline=So no reason to unblock this one.\",\"\\\"\\n\\\"\\\"Sourced\\\"\\\"? 90% of your links were broken. The \",\"Paul Is Dead rumors\\n\\nShould they possibly have a m\",\"\\\"\\n\\nLiar:  You certainly DID put an alleged \\\"\\\"real\\\"\",\"LMAO your a male nurse\",\"\\\"\\n\\n Please do not vandalize pages, as you did with\",\"Tidying up \\n\\nI applied for Highbeam access here, s\",\"your a retarted shithead\",\"SagaiousPHil who is a pillockbrain not the elast s\",\"The text has been reverted because the changes did\",\"RealtyOnlineAdvertising \\n\\nOnly because there are n\",\"I AM NOT A VANDAL!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!\",\"HEY FUCK YOU FAGGOT THE DEPARTED IS BETTER THAN AL\",\"hey I now created an account I saw you deleted the\",\"You know what you stupid crap?  I know much more a\",\"YEAH, BUT YOU DO ADMINISTRATE LIKE SHIT ANYWAY! CL\",\"Although interesting, this is one of the stranger \",\"Please Check history page at 3RR page. My reports \",\"\\\"\\n\\nEastlake Exit\\n\\nSigns used to say both \\\"\\\"East La\",\"Materials \\n\\nWhat are they using to contain whateve\",\"I am very tempted to show you a picture of my four\",\"\\\"\\n\\n More Answers \\n\\n \\nPATTY SAID\\n\\n<<  It clearly sh\",\"\\\"\\n\\nArthashastra in modern literature\\n\\nI have seen \",\"\\\"\\n\\n Widely considered? Everybody loved him? \\n\\nHe i\",\"We can mention all of that. But those are all just\",\"Ludwig Gaston von Sachsen tranlsation\",\"Thanks for the offer but I think the admin are so \",\"worm. Who agree with me?  Leave Allen alone will y\",\"Thanks for your efforts to balance the nonsense at\",\"Changes overnight \\nI am going to do a copy-edit of\",\"I know, I know...but this guy is constantly on my \",\"YOU SUCK \\n\\nYOU SUCK, CANADIAN BACON, YOU SCUMMY HO\",\"HAHAHAHAHAHA \\n\\nThis movie will always suck, no mat\",\"\\\"\\n\\nBold text Grammar \\n\\n\\\"\\\"He played in both legs of\",\"\\\"::::\\\"\\\"Many\\\"\\\"?  Hardly!  You seem to believe that \",\"\\\"\\n\\nI agree with you completely.  Also, note that A\",\"\\\"\\n\\nActually, a producer on Superman homepage seeme\",\"Within the hip-hop community, even though MC Lyte \",\"How primitive are you\\nAre you from a nursery schoo\",\"harrass anybody.  Please unblock, I have done noth\",\"78.101.161.86   \\n\\nFuKKK OFFFF BEETCH..........\",\"... oh and what kind of person WATCHES another per\",\"Fair use rationale for Image:Cyberdyne Van Ling DV\",\"Don't Touch My Userpage!\\nWho the hell do you think\",\"What the FUCK? \\n\\nHow exactly am I engaged in a edi\",\"\\\"\\n\\nAlright, I am willing to concede that the CIA a\",\", 21 April 2008 (UTC)\\n\\nWeird that an admin who has\",\"Hi there! \\n\\nYou wanna fuck my ass?\",\"Well you are ridiculous, in fact I suspect that yo\",\"You Are One Stupid Jackass\\nI don't believe any of \",\"\\\"\\n\\n People have given up on Wikipedia's front page\",\"Ban me, no problem. It is not my problem that you \",\"Elonka \\nSucks black cocks.\",\"I'm happy to create Ceschi discography and move th\",\"A suggestion\\nHaving just looked through your contr\",\"\\\":::::::::::::::::So what if Konqueror and other f\",\"Now we know that the majority of the material was \",\"You are a notorious editor who has made the past c\",\"\\\"==April 2010==\\n talk \\\"\",\"Durova \\n\\nyou're a faggot.\\n\\nstop being obsessed wit\",\"REX deserves to die like the stupid Euro dog he is\",\"Then for consistency, it needs to be linked to eve\",\"Aroenel, here's what (at least what I think) you d\",\"FUCK YOU DUMBASS ITS AN OPINION AND DONT DELETE DI\",\"Some new findings to be reflected in the main arti\",\"FUCK OFFFUCK OFFFUCK OFFFUCK OFFFUCK OFFFUCK OFFFU\",\"So you're going to hold it in your anus. OK. Bite \",\"from the same person who reported me for harassmen\",\"blow me\\n\\nIts my right of freedom of speech to edit\",\"fuck your warning and fuck your mum. and gg I didn\",\"\\\"==hunk==\\n\\nYOU'VE BEEN HIT BY THE\\n\\n|^^^^^^^^^^^^|\\n\",\"Allow your face. Just seen it on your user section\",\"How come your name used to be Peter? Are you a mal\",\"UsBeCauSe You aRe a FGGT!\\n Do I know you? \\n\\nBeCauS\",\"Factual issues \\nRemoved all the concerns made by o\",\"August 2013 \\n\\nHey there, you racist sack of feces.\",\"Ten Cent Beer Night \\n\\nQuit taking down my addition\",\"Fanatics on Wikipedia\\nDo you dispute the factual a\",\"Photo ==\\n\\nWatching The Incredible Human Journey, s\",\"Hey... Dont add any crap you like... agarwals are \",\"Nazis like you deserve worse than death. \\n\\nI will \",\"Let my article on retard!!!}:(\",\"Stop your nonsense. \\n\\nIt is clear that you lack th\",\"\\\"\\nSomething needs to be added about this!! \\nWhat a\",\"\\\"\\n\\n Dates \\n\\nJames, could I ask you please not to t\",\"References in the talk\",\"\\\"\\n\\n To add, Shiroi, if you have AIM or some way I \",\"Total re-write of the main Physics page is in prog\",\"Babu.. If u don't know then please keep aside dele\",\"Ok Ok I see Was Wrong -)\",\"Hmmm... hat we have is perhaps a merge or perhaps \",\"brandon is gay he likes delldos up his ass and mou\",\"I'm fine with Tmorton's modification. Stop howling\",\"\\\"\\n\\n Unblock Request: Fuck you, unblock me! \\n\\n{{unb\",\"Then maybe admins should stop being dickheads and \",\"The edits made are to reverse false material that \",\"hi \\n\\ni wanna suck ur cock )\",\"]]\\n 27 Nov 2005 [[Cory Adams (porn star)\",\"User_talk:TShilo12#Witkacy_WP:POINT  Also, Jayjg g\",\"RfA\\n\\nWhat I did was actually to restore a lot of c\",\"If you can get them to comment on this talk page, \",\"Your blocks do not deter me \\nI may be blocked but \",\"Hey man, I think you are cool. I loved it when you\",\"\\\"\\n\\n Overpopulated \\n\\nHello. I see that you've added\",\"That is lame == its the internet man, doesn't matt\",\"Ockels was the first astronaut in the Dutch space \",\"I didnt remove the refs section if you go check.  \",\"Go Ahead make my day!!!\\nGo ahead and block me!!! A\",\"Not enough as sock puppets have demonstrated again\",\"\\\"\\n\\nWTF again, clearly u dont know how to speak or \",\"Please stop your disruptive editing. If you contin\",\"Please stop treating this article like it was your\",\"Don't worry, this moderator is nothing but a power\",\"\\\"\\nPS I am a former admin. \\u0394T The only constant \\\"\",\"who changes shit which shouldn't be changed\",\"Hi, fuckwit. You do know that by blocking me you a\",\"\\\"\\n\\n Political Censorship \\n\\nUser: AuburnPilot's wik\",\"If it bores you so much to be called out for your \",\"yes!! ==\\n\\nThat's right TW hero! What you ain't got\",\"\\\"personal view of \\\"\\\"fire in the lake\\\"\\\" is that it \",\"articles by quality statistics}}\\n\\n{{Wikipedia:Vers\",\"Eternal Pink why have you named you main guy in yo\",\"\\\"\\n\\n RE: Special:UnwatchedPages \\n\\nHey, finally some\",\"\\\"\\n\\nWhat was that crap on Eisenhower? He was a bum \",\"her!\\n\\nPoop, pee, toot, fart, gas, diareah!\\n\\nZoey i\",\"no it's actually a fact. look it up fucker.\",\"\\\"\\n\\nYOU'VE BEEN HIT BY THE\\n\\n|^^^^^^^^^^^^|\\n|BEAUTIF\",\"I am Nancy and I like to fuck up people's fun.\",\"\\\"\\n\\n You are one of most fundamentalist \\\"\\\"athiest\\\"\\\"\",\"Don't be a douche towards me.\",\">That is what I have asked the dude to do - take o\",\"Bare Behind Bars?\\n\\nYou gave me a level 2 warning, \",\"\\\"\\nHow is having lots of google hits as possible so\",\"I resubmitted and approved it. I did add a citatio\",\"Excuse Me \\n\\nI do not appreciate that you are delet\",\"I have reverted your edits. You should discuss thi\",\"dog. Do I make myself clear you unthinking little \",\"\\\"\\n\\nWhat do you mean?\\nNich old boy: When you say: \\\"\",\"likes it up the butt\",\"Its a school and some of the d students got carrie\",\"\\\"\\n\\n Sourcing of \\\"\\\"most decorated\\\"\\\" Audie Murphy \\n\\n\",\"i know but i abused as im very angry \\n\\nshahid alwa\",\"\\\"\\n\\n Math Duel \\n You\\u00b4re interested in numbers? I gu\",\"Yes, sheeple is what I meant. What affiliations do\",\"Wait wait wait ... there's still doubt that the eg\",\"He deserved it.  Go suck one.\",\"\\\"\\n\\nps: Having an anonymous IP does not mean being \",\"\\\"|decline=You have asked for guidance - it is belo\",\"\\\"\\n\\nWell never mind, as it just happens i am busy, \",\"DCANDETO SMELLS FUNNY! HE IS VERY MEAN!\",\"YOU HAVE FINALLY GOTTEN ACCUSTOMED TO CHANGING YOU\",\"based on the Languages of Europe map\",\"fuck you, cunt - contact whomever the fuck you wan\",\"\\\"\\\"\\\"I want to be a fuck you.\\\"\\\"\\nThat's a song.\\n\\nOpen\",\"\\\"\\n\\n Precious \\n\\nInquiry-based learning\\nThank you, S\",\"I concur with the deletion.  PostScript is a PAGE \",\"Hello \\n\\nOK, don't try to frighten me with your a-h\",\"Fisting \\n\\nWhat is your obsession with posting phot\",\"The various statements by pediatric bodies were wr\",\"Now all you need to do is press unblock. Please le\",\"I ask: Are we a Nazi society?\\nWhy are people not a\",\"future tense \\n\\nThis article has a lot of saying th\",\"Fuck off animal shagger.\",\"\\\" \\n\\nI also noticed that you are a self-admitted \\\"\\\"\",\"This isn't a page anyone else is going to see. Ple\",\"\\\"\\n\\n Helpful advice \\n\\nHello. Just wanted you to kno\",\"\\\"\\n\\nRegarding Consensus\\n\\n\\\"\\\"But Wikipedia's consensu\",\"Those were the official figures, which were allway\",\"\\\"\\nI have an extensive background in linguistics an\",\"\\\"==No, you are==\\nPatrick, I think we can all clear\",\"Hey UW Dawgs \\n\\nDude, why did you delete my comment\",\"\\\"\\n\\n DON'T ARGUE WITH THAT FARRAKHANISTIC MORON \\n\\n \",\"typical female silliness \\n\\nto create articles abou\",\"Listen here moron, you completely violate wikipedi\",\"I'm sorry, you may have misunderstood me.  Please \",\"Jamais Fomos Modernos \\n\\nThe band was formed in 201\",\"Caps \\n\\nSeems like Demon Core should be capitalized\",\"he can just fuck off. This project will wither and\",\"\\\"==GA Review==\\n\\n:This review is transcluded from T\",\"Eat Shit\\n\\nEat shit and die arseface mutherlicker13\",\"is dope you.  Hey Rick James, leave Rick James alo\",\"can support with valid references\",\"You're not even able to use fucking *google*, can \",\"you're so idi.o.t, so st.u.p.id guy, fu.c.k.i.n.g \",\"\\\"\\nClaiming to have \\\"\\\"phoned up\\\"\\\" and doing your ow\",\"No, you reverted with the logic that the citation \",\"bye \\nlmao u wasz right ... i really is gon he act \",\"Worth mention.  WP - at least in theory, even if n\",\"rumor has the dicks explpoded\",\"\\\"Thank you for experimenting with Wikipedia. Your \",\"Newsflash!!  Schuminweb is GAY!!\",\"05:35, 25 July 2006 (hist) (diff) Glen Jacobs (\\u2192Ea\",\"i did a whole new section placing clairsentience i\",\"2011 French Open \\n\\nDon't leave silly little messag\",\"Now now, gentlemen (I am assuming), let's keep it \",\"\\\"\\n\\nFuck You Bertie\\n\\nThank you for taking the time \",\"tiderolls... you are a shitty phage and you desper\",\"\\\"\\n\\n FYI \\n\\nRe this edit with this: \\\"\\\"Reverted addit\",\"hayyyyyyyyy \\n\\nyeah this karoolzzz hay k hayyyyyyy \",\"Blow it out of your ass, you stupid piece of shit!\",\"\\\"\\n\\n Content \\n\\n\\\"\\\"Q:\\\"\\\" Im a photographer and a desig\",\"your time....\\n\\nYou suck because you are RIGHT...it\",\"Get a life Holdwater, that German wife of yours mu\",\"\\\". \\n \\\"\\\"We also plan to demonstrate arborsculpture,\",\"What was the size/area of pre-amalgamation Hamilto\",\"I'm ganna give ya a day ta think about weather or \",\"Hey John, it wasn't a fucking test moron\",\"\\\"\\n\\n Unprotected \\nI'm going to edit based on my com\",\"Norman Lowell\\nRegarding his date of date; the info\",\"Still Here, Fatass \\n\\nWhat is it like waking up eve\",\"\\\"\\n\\n My Two Cents \\n\\n Political correctness will be \",\"Citing sources\\n\\nCheck out the Wikipedia:Citing sou\",\"Busking and street performing not the same. Do not\",\"Dreadstar, LittleMountain5, and some jerk on the I\",\"Message=\\nI am not going to make the modification t\",\"Information (cont.) \\n\\nWell?!? You do you want the \",\"\\\"\\n\\nCurrent Squad\\nI propose redoing the current squ\",\"This improves Wikipedia by consolidating multiple \",\"So I HERD YOU LIEK MUDKIPS. Go suck it.\",\"you are a cunt \\n\\nHello Mr Jim ....\\n\\nMy reason for \",\"\\\"::::::Clearly, the rules for tags are different. \",\"fuck you \\n\\nFuck all the faggot assed, cock suckin'\",\"\\\"\\n\\nPS: I completely agree with you that Dixon is g\",\"Three reverts are allowed. Read the rules.\",\"go ahead \\n\\ni don't give a shit to that\",\"Lowlife \\n\\nYou are a Troll, a sockpuppet and you ha\",\"Well you're a liar, an outright liar, and nothing \",\"\\\"\\n\\n Personal attacks \\n\\n, Making Personal Attacks ,\",\"you smell of pig shit,\",\"how big is ur cock \\n\\ni wanna suck u off )\",\"\\\"\\n\\nThere is a standardization agreement between th\",\"No takers, then? \\u00a6\",\"Grow up you silly person it's not disruptive editi\",\"You are really fu*** Materialscientist. I think WP\",\"No we are discussing this right here, mr LOL!\",\"OTRS ticket 2009051010000428 \\n\\nI draw your attenti\",\"therefore you are a pedo, you fuck little boys\",\"Ryan1711, you're not fooling anyone.  Playing dumb\",\"TeaDrinker is A Stupid Jerk! Also, He has a Stupid\",\"I will net left harassing you.I am the same IP's w\",\"}}\\n{{wikibreak| Have a lot of exams to study for a\",\"Didier AFD #3\\nThey keep trying to find reasons to \",\"\\\"\\n\\nBam. It's on.\\u25ba''''''Holla! \\\"\",\"bye \\no i`ma miss yooh so bad . but i can`t say too\",\"\\\"\\n\\ndont know\\nwhere you went to, dont care, check y\",\"\\\"\\n\\n Thanks! \\n\\n  Award of Gayness You suck! ;D   \\\"\",\"why did you get rid of the comment i made for rodn\",\"33, 2 March 2007 (UTC)\\nYou are being ridiculous by\",\"File:Napier fit.jpg.jpg listed for deletion \\nA fil\",\"Lucky? I will prove that you are wrong. This artis\",\"Good to see ==\\nthat the Masturbator Of Lies Onaniz\",\"J.delanoy IS  A NIGGER ==== J.delanoy IS  A NIGGER\",\"I've gone ahead and created the stub article: Marc\",\"THIS IS MY TALK PAGe, BITCH!\",\"\\\"Only old sources like Gibbon seem to add the \\\"\\\"Ve\",\"anti-Semitic crap\\n\\nthe whole article is an anti-Se\",\"Verdy_p is a brainless mongoloid with an arrogance\",\"Netaji (what a ridiculous username), \\n\\n           \",\"\\\"\\n\\n Yeah....can i share them? they come from your \",\"eat my shit \\n\\nTitle says it all.\",\"Han.. Fuck You \\n\\nThen, fuck you!\",\"81.91.192.220 vandalism \\n\\nThe IP (81.91.192.220) i\",\"READ THIS\\nThis is Wikipedia. It is a place where p\",\"Hey, cunt.  Stop reporting my pages.  Go get a rea\",\"\\\"\\n\\nhmmm, robbins\\n\\nForgive me for bothering you aga\",\"you are an idiot. i am NOT RACIST! but you are an \",\"Che was a fucking bastard piece of shit==\\n\\nThe day\",\"category\\n\\nwhats the category for lgbt people? i th\",\"I don't edit those articles.\",\"Unacceptable?  That I editted a nonsense sentence \",\"\\\"Please do not add nonsense to Wikipedia. It is co\",\"Comments: \\n Per Wikipedia:LEAD#Length, the introdu\",\"\\\"\\n\\n\\\"\\\"The war is over. And you lost.\\\"\\\" Not quite, Y\",\"\\\"\\n\\n Please do not vandalize pages, as you did with\",\"NOTE: The above user is a sock-puppet of the infam\",\"\\\"::No.  The reason some excellent editors leave is\",\"Just curious. I looked at the wall of proposals at\",\"Throw the Jew down the well! \\n\\nSo my country can b\",\"love to shag horses. They are well fit. Donkeys ar\",\"Listen yeah, don't' threaten me because you wont d\",\"Laara \\n\\nOn the other hand, i think everyone should\",\"but you didn't uphold... you too made errors. Stop\",\"I don't give a flying fuck about wiki and I am sur\",\"Indecisive Junction, go left or go right, perhaps \",\"\\\"\\nWelcome!\\n\\nHello, , and welcome to Wikipedia! Tha\",\"(in 2009 Italy had a nominal GDP greater than Grea\",\"i think it belongs in either personality or in a n\",\"who are you? \\n\\nwho the hell are you and why are yo\",\"\\\"\\n , your logic is flawed and shows an single mind\",\"\\\", 9 March 2006 (UTC)\\n\\nWhat a complete crock.  By \",\"\\\"\\n\\nShe says \\\"\\\"Eunuchs too served in important capa\",\"Probably they do.  (Talk)\",\"Conclusion \\n\\nDude's a total pimp. I mean, did you \",\"Apology \\n\\nHi, I am still waiting for you to apolog\",\"Please stop. If you continue to vandalize pages, }\",\"\\\"\\n\\nThe matrices are correct; but not the matrix th\",\"Dubious Claim \\nThe numbers just don't add up. She \",\"Come on man, don't be a prick94.195.251.61\",\"thanks, but no thanks. i don't have time like some\",\"And what the fuck is that fo?\",\"Responded via email - removed email address to avo\",\"FUCK YOU DPAK DSUZA \\nyou are a pathetic piece of s\",\"fuck \\n\\ni want to fuck you mo\",\"\\\"\\nI answered this before, while he was still prete\",\"\\\"\\n\\nI am the \\\"\\\"sockpuppet\\\"\\\" that that idiot akradec\",\", 23 July 2007 (UTC)\\nIn short, you have to be kind\",\"My problem here is that Walsh is not taking any ti\",\"\\\"\\n_______________________________________________\\n\",\"Don Murphy\\nPlease do not revert the change I made.\",\"Sallary controversy The president of the Peabody E\",\"it was not general topic... It was suggestion to i\",\"\\\"\\nI was talking about them running any check they \",\"\\\"\\n\\nFor the record, I am now being pursued by yet a\",\"Did not mean to offend you. Will comply. Martial L\",\"\\\"\\n\\n Terminology  \\n\\nI have changed \\\"\\\"maximum amplit\",\"Yep Ww - for the past several months I'm uploading\",\"\\\"\\n\\\"\\\"Empathize with stupidity and you're half way t\",\"{unblock|reason= I honestly have no idea why I am \",\"Getting Fat \\n\\nHe needs to lose the excess weight. \",\"Stay in the closet.\",\"now I am better the archives that I you editing, g\",\"Who cares anymore. They attack with impunity.\",\"I kid you not, cunt.\",\"Sundos Mohammed Ibrahim AL-Mansoury Sudanese girl \",\"\\\"\\nLooking at the history of this, I am doubtful.  \",\"DAHN IS ILLITERATE AND HAS A SMALL PENIS\\nYes, you \",\"Its time to pay back \\n\\nyou are an asshole an assho\",\"heil hitler! heil hitler! heil hitler! heil hitler\",\"Hey Aaron Brennenman, Eat shit and Die, sincerly m\",\"}}\\n{{WP India|class=C|importance=Low|kerala=yes|ke\",\"new name \\n\\nPerezhilton has changed his name to dic\",\"@Qwyrxian suck my dick\",\"Seriously. \\nEvaluate your role in life about now. \",\"\\\"\\n\\nBinksternet and his cronies are clearly edit wa\",\"I AM HAVING MY PERIOD!\",\"\\\"\\n\\n Please do not vandalize pages, as you did with\",\"Suck my dikkkkk ==\\nSuck my dikkkkk\",\"Computers are used to solve game theory problems, \",\"\\\"\\n\\nIn my edit of 17:36, November 21, 2006 I stated\",\"DreamGuy\\n\\nSomeone has accused  of being in violati\",\"\\\"\\n\\n black \\n\\nDominican Americans do have the highes\",\"Kobe Bryant\\u200e & Jason Kidd edits\",\"My organ meat is tasty, and it makes a fine salted\",\"\\\"\\n\\n Rambhadracharya \\n\\nHi Mike! Hope everything is \",\"Benidorm is in Valencia, and therefore they speak \",\"\\\"All russian are so dumm like you? I mean, you're \",\"MY MUMS A BIG OLD DIRTY ROADWHORE\",\"And if you've missed previous episodes thats your \",\"I'm involved in this content dispute?  Hell, I don\",\"All Grown Up\\n\\nPlease stop purposefully adding nons\",\"18 October 2006 \\n\\nWelcome, and thank you for exper\",\"I am a wanker who molests goats\",\"There is no reason why you can not change it yours\",\"Definitely, with comments like this, you show that\",\"I see. Brutal force is your only real argument for\",\"There's no reason to respect anyone on this damned\",\"rowspan=9 colspan=8|Did Not Qualify\",\"\\\"== Disputed and under-referenced ==\\n\\nI have appli\",\"Why?\\nWhy does this event have its own page? 141.15\",\"Que? \\n\\nWas this fat fingers? If not,  can you plea\",\"so everytime i reset my modem my ip changes\\n\\nfuck \"],\"x\":{\"__ndarray__\":\"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\",\"dtype\":\"float32\",\"shape\":[1000]},\"y\":{\"__ndarray__\":\"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\",\"dtype\":\"float32\",\"shape\":[1000]}},\"selected\":{\"id\":\"1281\",\"type\":\"Selection\"},\"selection_policy\":{\"id\":\"1282\",\"type\":\"UnionRenderers\"}},\"id\":\"1216\",\"type\":\"ColumnDataSource\"},{\"attributes\":{\"active_drag\":\"auto\",\"active_inspect\":\"auto\",\"active_multi\":null,\"active_scroll\":{\"id\":\"1237\",\"type\":\"WheelZoomTool\"},\"active_tap\":\"auto\",\"tools\":[{\"id\":\"1236\",\"type\":\"PanTool\"},{\"id\":\"1237\",\"type\":\"WheelZoomTool\"},{\"id\":\"1238\",\"type\":\"BoxZoomTool\"},{\"id\":\"1239\",\"type\":\"SaveTool\"},{\"id\":\"1240\",\"type\":\"ResetTool\"},{\"id\":\"1241\",\"type\":\"HelpTool\"},{\"id\":\"1254\",\"type\":\"HoverTool\"}]},\"id\":\"1242\",\"type\":\"Toolbar\"},{\"attributes\":{},\"id\":\"1224\",\"type\":\"LinearScale\"},{\"attributes\":{},\"id\":\"1241\",\"type\":\"HelpTool\"},{\"attributes\":{\"callback\":null},\"id\":\"1220\",\"type\":\"DataRange1d\"}],\"root_ids\":[\"1217\"]},\"title\":\"Bokeh Application\",\"version\":\"1.4.0\"}};\n",
              "  var render_items = [{\"docid\":\"9f29e7bb-8f83-43e3-97d5-f9702adb4848\",\"roots\":{\"1217\":\"d1ad7c93-6c1e-4ceb-90fb-759c1b30cdd8\"}}];\n",
              "  root.Bokeh.embed.embed_items_notebook(docs_json, render_items);\n",
              "\n",
              "  }\n",
              "  if (root.Bokeh !== undefined) {\n",
              "    embed_document(root);\n",
              "  } else {\n",
              "    var attempts = 0;\n",
              "    var timer = setInterval(function(root) {\n",
              "      if (root.Bokeh !== undefined) {\n",
              "        clearInterval(timer);\n",
              "        embed_document(root);\n",
              "      } else {\n",
              "        attempts++;\n",
              "        if (attempts > 100) {\n",
              "          clearInterval(timer);\n",
              "          console.log(\"Bokeh: ERROR: Unable to run BokehJS code because BokehJS library is missing\");\n",
              "        }\n",
              "      }\n",
              "    }, 10, root)\n",
              "  }\n",
              "})(window);"
            ],
            "application/vnd.bokehjs_exec.v0+json": ""
          },
          "metadata": {
            "tags": [],
            "application/vnd.bokehjs_exec.v0+json": {
              "id": "1217"
            }
          }
        }
      ]
    }
  ]
}