DystSys(1).ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_vMmpCLS5Shi"
      },
      "source": [
        "TODO\n",
        "\n",
        "- Provjeriti dal je na politics apply-an lemmatizer (nije)\n",
        "- Zašto su svi politics tokeni u prvom clusteru (svih 107 usera)\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uw_vHitmsBzs"
      },
      "source": [
        "###### Should always be run"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a5hXskjnIHOv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8ac5591d-5ae3-4349-faa2-83b0fcf43359"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "h35SDM0WIGjV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "837eff9a-8816-4b81-d80a-3ea55f66b1ed"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/MyDrive/Colab Notebooks/p2p tokenizer\n"
          ]
        }
      ],
      "source": [
        "%cd /content/drive/MyDrive/Colab Notebooks/p2p tokenizer/"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mpZc6xR2GxkG",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a8b2ecb0-dd4e-41b2-f6f8-fd1f1625f9ed"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data] Downloading package omw-1.4 to /root/nltk_data...\n"
          ]
        }
      ],
      "source": [
        "import nltk\n",
        "from nltk.stem import WordNetLemmatizer\n",
        "nltk.download(\"wordnet\")\n",
        "nltk.download(\"omw-1.4\")\n",
        "\n",
        "wnl = WordNetLemmatizer()\n",
        "lem = lambda word: wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WN01UidMg_5J"
      },
      "source": [
        "#### Prep\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uf9-WhJ2PCAV"
      },
      "source": [
        "##### General tokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Sjn4gM2ZMHZx"
      },
      "outputs": [],
      "source": [
        "import json\n",
        "from tokenization import FullTokenizer\n",
        "from collections import Counter"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ypS7-QHfcVqJ"
      },
      "outputs": [],
      "source": [
        "tokenizer = FullTokenizer(\"vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HpeuaxbaamVF",
        "outputId": "53bb3c13-7854-4eb9-e8b6-fb3820bf4935"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Counter({'this': 1,\n",
              "         'is': 1,\n",
              "         'my': 1,\n",
              "         'sample': 1,\n",
              "         'sent': 1,\n",
              "         '##ec': 1,\n",
              "         '##nce': 1,\n",
              "         '.': 1})"
            ]
          },
          "execution_count": 6,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "Counter(tokenizer.tokenize(\"This is my sample sentecnce.\"))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "49CsU9JzgORP"
      },
      "outputs": [],
      "source": [
        "with open(\"nba.txt\") as n, open(\"politics.txt\") as p:\n",
        "  nba_comments = \" \".join(json.loads(row)[\"comment\"] for row in n)\n",
        "  politics_comments = \" \".join(json.loads(row)[\"comment\"] for row in p)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lfX1Aq4PgmzE"
      },
      "outputs": [],
      "source": [
        "all_comments = nba_comments + ' ' + politics_comments"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 169
        },
        "id": "rmw0fcdXw2kK",
        "outputId": "e74c4de0-552d-4c02-aa6e-c6c30890bf95"
      },
      "outputs": [
        {
          "ename": "NameError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-9-6bfc24ad5359>\u001b[0m in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset_diff\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m: name 'set_diff' is not defined"
          ]
        }
      ],
      "source": [
        "len(set_diff)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kU8hByjvw8Mm"
      },
      "outputs": [],
      "source": [
        "len(politics_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "E_Hs1xbRxJ0n"
      },
      "outputs": [],
      "source": [
        "politics_comments[:40]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EOD8iEQHPIy-"
      },
      "source": [
        "##### NBA (its actually all, not just NBA)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2xqOJGQw4Ix6"
      },
      "source": [
        "###### Small Tokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tIAXgXTKko7x"
      },
      "outputs": [],
      "source": [
        "tokenized = tokenizer.tokenize(all_comments)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "96a4bYVYkloO"
      },
      "outputs": [],
      "source": [
        "c = Counter(tokenized).most_common(500)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yO84ksINe7UA"
      },
      "outputs": [],
      "source": [
        "for i in c:\n",
        "  if i[0] == \"[UNK]\":\n",
        "    print(i)\n",
        "    break\n",
        "else:\n",
        "  print(\"Not found\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gCvgMthOjecf"
      },
      "outputs": [],
      "source": [
        "small_vocab = \"\\n\".join(key for key, value in c)\n",
        "\n",
        "with open('small_vocab.txt', 'w') as f:\n",
        "  f.write(small_vocab)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LSK4O83whqox"
      },
      "outputs": [],
      "source": [
        "small_tokenizer = FullTokenizer(\"small_vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GHY9k--cPVDS"
      },
      "source": [
        "###### Testing"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "h-HJVVXR1Dm6"
      },
      "outputs": [],
      "source": [
        "small_tokenizer = FullTokenizer(\"small_vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8V0imZLihwGJ"
      },
      "outputs": [],
      "source": [
        "nba_tokenized = small_tokenizer.tokenize(nba_comments)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6DeJ4Hx0QYT_"
      },
      "outputs": [],
      "source": [
        "all_tokens = list(lem(token) for token in Counter(tokenized).keys())\n",
        "nba_tokens = list(lem(token) for token in Counter(nba_tokenized).keys())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kVxXJzLvQxCv"
      },
      "outputs": [],
      "source": [
        "set_diff = set(all_tokens) - set(nba_tokens)\n",
        "list(set_diff)[:10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4tOFGDeNSbmT"
      },
      "outputs": [],
      "source": [
        "clean_token_diff = [i for i in set_diff if \"#\" not in i]\n",
        "clean_token_diff[:10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fZTup246ueA1"
      },
      "outputs": [],
      "source": [
        "word_dict_nba = {len(word): [] for word in nba_tokens}\n",
        "for word in nba_tokens:\n",
        "    word_dict_nba[len(word)].append(word)\n",
        "\n",
        "word_dict_clean = {len(word): [] for word in clean_token_diff}\n",
        "for word in clean_token_diff:\n",
        "    word_dict_clean[len(word)].append(word)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "sIgr3gMwsxMo"
      },
      "outputs": [],
      "source": [
        "unclean_nba_tokens = [token for token in list(nba_tokens) if \"#\" in token]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yq2zlwM72QBN"
      },
      "outputs": [],
      "source": [
        "a = list(nba_tokens)\n",
        "b = list(clean_token_diff)\n",
        "c = list(unclean_nba_tokens)\n",
        "#print_unique(a, b)\n",
        "\n",
        "#nba_tokens"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-gKti7LnwCah"
      },
      "outputs": [],
      "source": [
        "print(a[:10])\n",
        "print(b[:10])\n",
        "print(c[:10])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VZSnGzv0m-vv"
      },
      "outputs": [],
      "source": [
        "import re\n",
        "\n",
        "to_rmv = [\".\", \"^\", \"|\"]\n",
        "\n",
        "for i in to_rmv:\n",
        "  try:\n",
        "    c.remove(i)\n",
        "  except:\n",
        "    pass\n",
        "\n",
        "i = 0\n",
        "not_found_tokens = []\n",
        "count = 0\n",
        "for clean in b:\n",
        "  for reg_token in c:\n",
        "    i += 1\n",
        "    if i % 1000000 == 0:\n",
        "      print(i, len(c) * len(b))\n",
        "\n",
        "    try:\n",
        "      if re.match(reg_token.replace(\"#\", \".\") + r\"\\b\", clean) is not None:\n",
        "\n",
        "        # print(\"1\", reg_token, clean)\n",
        "        break\n",
        "    except:\n",
        "      pass\n",
        "  else:\n",
        "    count += 1\n",
        "    not_found_tokens.append(clean)\n",
        "    # print(clean)\n",
        "\n",
        "\n",
        "print(\"Count:\", count)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vWrKlld4wFX3"
      },
      "outputs": [],
      "source": [
        "print(not_found_tokens[:10])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rDSCy-K_tplR"
      },
      "outputs": [],
      "source": [
        "print(len(a), len(nba_tokens))\n",
        "print(len(b))\n",
        "print(len(c))\n",
        "print(len(not_found_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "88vwyBN8u80y"
      },
      "outputs": [],
      "source": [
        "# First 10 tokens marked as UNK that are smaller then 4 characters in length\n",
        "print(list(i for i in not_found_tokens if 3 < len(i) < 5)[:10])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "I6-Fm0-SvSZ_"
      },
      "outputs": [],
      "source": [
        "atts = [\"hi\", \"#i\", \"h#\"]\n",
        "\n",
        "# 0 means no matches\n",
        "print(sum(att in a for att in atts))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-n8c3ksDxT9a"
      },
      "outputs": [],
      "source": [
        "actual_unks = []\n",
        "for i in not_found_tokens:\n",
        "  if \"[UNK]\" in small_tokenizer.tokenize(i):\n",
        "    actual_unks.append([i, small_tokenizer.tokenize(i)])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1CBDGh5cye1l"
      },
      "outputs": [],
      "source": [
        "len(actual_unks)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RTlJjja69xwO"
      },
      "outputs": [],
      "source": [
        "actual_unks"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ahW1CklJy_jd"
      },
      "outputs": [],
      "source": [
        "small_tokenizer.tokenize(\"Random nintendo zack\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bsbRoeTX0Z81"
      },
      "outputs": [],
      "source": [
        "print((list([i for i in actual_unks if not i[0].isalnum()])))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "S7Wf7BUu3hPb"
      },
      "outputs": [],
      "source": [
        "print((list(i for i in Counter(tokenized).most_common(1000) if \"#\" in i[0])))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OgkYTvuS1dmQ"
      },
      "outputs": [],
      "source": [
        "nba_unks = list([i[0] for i in sorted(actual_unks) if i[0].isalnum()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9wpSVqdz2WTz"
      },
      "outputs": [],
      "source": [
        "nba_word_tokens = [token for token in nba_unks if not any(char.isdigit() for char in token)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RBuCqKoG2zVB"
      },
      "outputs": [],
      "source": [
        "len(nba_unks) - len(nba_word_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gB60mI5R3FHL"
      },
      "outputs": [],
      "source": [
        "[token for token in nba_unks if any(char.isdigit() for char in token)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FCCaA4ff0bJH"
      },
      "outputs": [],
      "source": [
        "# Store unk tokens in a file\n",
        "with open('nba_unk_tokens.txt', 'w') as f:\n",
        "  f.write(\"\\n\".join(nba_word_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BstpX-peA1jf"
      },
      "outputs": [],
      "source": [
        "len(nba_word_tokens)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zbv-Djkk4R4U"
      },
      "source": [
        "##### Politics"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TOOsXZIV4R4V"
      },
      "source": [
        "###### Small Tokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XWu73pziAg0s"
      },
      "outputs": [],
      "source": [
        "tokenized = tokenizer.tokenize(all_comments)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uWB7GGmdAg0u"
      },
      "outputs": [],
      "source": [
        "c = Counter(tokenized).most_common(500)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0mOTFkkMAg0v"
      },
      "outputs": [],
      "source": [
        "for i in c:\n",
        "  if i[0] == \"[UNK]\":\n",
        "    print(i)\n",
        "    break\n",
        "else:\n",
        "  print(\"Not found\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IvS7eaDUAg0w"
      },
      "outputs": [],
      "source": [
        "small_vocab = \"\\n\".join(key for key, value in c)\n",
        "\n",
        "with open('small_vocab.txt', 'w') as f:\n",
        "  f.write(small_vocab)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5x0UQjDOAg0x"
      },
      "outputs": [],
      "source": [
        "small_tokenizer = FullTokenizer(\"small_vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "127sIWT24R4e"
      },
      "source": [
        "###### Testing"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "agpYDX_5Awmo"
      },
      "outputs": [],
      "source": [
        "small_tokenizer = FullTokenizer(\"small_vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "En3_TMGlAwms"
      },
      "outputs": [],
      "source": [
        "politics_tokenized = small_tokenizer.tokenize(politics_comments)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Lkym48Ys4R4g"
      },
      "outputs": [],
      "source": [
        "all_tokens = list(lem(token) for token in Counter(tokenized).keys())\n",
        "politics_tokens = list(lem(token) for token in Counter(politics_tokenized).keys())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RZF4LLW74R4g"
      },
      "outputs": [],
      "source": [
        "set_diff = set(all_tokens) - set(politics_tokens)\n",
        "list(set_diff)[:10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "o8wNlH954R4h"
      },
      "outputs": [],
      "source": [
        "clean_token_diff = [i for i in set_diff if \"#\" not in i]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rPB6T1CB4R4h"
      },
      "outputs": [],
      "source": [
        "word_dict_politics = {len(word): [] for word in politics_tokens}\n",
        "for word in politics_tokens:\n",
        "    word_dict_politics[len(word)].append(word)\n",
        "\n",
        "word_dict_clean = {len(word): [] for word in clean_token_diff}\n",
        "for word in clean_token_diff:\n",
        "    word_dict_clean[len(word)].append(word)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oYxevy5W4R4h"
      },
      "outputs": [],
      "source": [
        "unclean_politics_tokens = [token for token in list(politics_tokens) if \"#\" in token]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rtIXnspS4R4h"
      },
      "outputs": [],
      "source": [
        "a = list(politics_tokens)\n",
        "b = list(clean_token_diff)\n",
        "c = list(unclean_politics_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OwpZh2YfvMAx"
      },
      "outputs": [],
      "source": [
        "print(a[:10])\n",
        "print(b[:10])\n",
        "print(c[:10])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "d6OlFjcU4R4i"
      },
      "outputs": [],
      "source": [
        "import re\n",
        "\n",
        "to_rmv = [\".\", \"^\", \"|\"]\n",
        "\n",
        "for i in to_rmv:\n",
        "  try:\n",
        "    c.remove(i)\n",
        "  except:\n",
        "    pass\n",
        "\n",
        "i = 0\n",
        "not_found_tokens = []\n",
        "count = 0\n",
        "for clean in b:\n",
        "  for reg_token in c:\n",
        "    i += 1\n",
        "    if i % 1000000 == 0:\n",
        "      print(i, len(c) * len(b))\n",
        "\n",
        "    try:\n",
        "      if re.match(reg_token.replace(\"#\", \".\") + r\"\\b\", clean) is not None:\n",
        "\n",
        "        # print(\"1\", reg_token, clean)\n",
        "        break\n",
        "    except:\n",
        "      pass\n",
        "  else:\n",
        "    count += 1\n",
        "    not_found_tokens.append(clean)\n",
        "    # print(clean)\n",
        "\n",
        "\n",
        "print(\"Count:\", count)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "c1L9S90Lv6PP"
      },
      "outputs": [],
      "source": [
        "not_found_tokens[:10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "NjSQ2IFN4R4j"
      },
      "outputs": [],
      "source": [
        "print(len(a), len(politics_tokens))\n",
        "print(len(b))\n",
        "print(len(c))\n",
        "print(len(not_found_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HqZKPlwU4R4j"
      },
      "outputs": [],
      "source": [
        "# First 10 tokens marked as UNK that are smaller then 4 characters in length\n",
        "print(list(i for i in not_found_tokens if len(i) < 4)[:10])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zByTPLzs4R4k"
      },
      "outputs": [],
      "source": [
        "atts = [\"ri\", \"#i\", \"r#\"]\n",
        "\n",
        "# 0 means no matches\n",
        "print(sum(att in a for att in atts))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "G3bKM8y14R4k"
      },
      "outputs": [],
      "source": [
        "actual_unks = []\n",
        "for i in not_found_tokens:\n",
        "  if \"[UNK]\" in small_tokenizer.tokenize(i):\n",
        "    actual_unks.append([i, small_tokenizer.tokenize(i)])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WFfYEsJZ4R4l"
      },
      "outputs": [],
      "source": [
        "len(actual_unks)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9za_v4O54R4m"
      },
      "outputs": [],
      "source": [
        "small_tokenizer.tokenize(\"Random sentance using 495 1969 ncaa naughty national 1769\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "U5ubLCt84R4m"
      },
      "outputs": [],
      "source": [
        "small_tokenizer.tokenize(\"coffee cream in 2014 was op\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "78iUELyJ4R4n"
      },
      "outputs": [],
      "source": [
        "print(\"Tokens marked as UNK:\")\n",
        "for i in actual_unks:\n",
        "  print(i[0])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "UAgrDZWv4R4n"
      },
      "outputs": [],
      "source": [
        "print((list([i for i in actual_unks if not i[0].isalnum()])))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "MeqK2G9-4R4o"
      },
      "outputs": [],
      "source": [
        "print((list(i for i in Counter(tokenized).most_common(1000) if \"#\" in i[0])))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "z4MHJ_224R4o"
      },
      "outputs": [],
      "source": [
        "politics_unks = list([i[0] for i in sorted(actual_unks) if i[0].isalnum()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "W23T7FXR4R4o"
      },
      "outputs": [],
      "source": [
        "politics_word_tokens = [token for token in politics_unks if not any(char.isdigit() for char in token)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0fpZB_3h4R4p"
      },
      "outputs": [],
      "source": [
        "len(politics_unks) - len(politics_word_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VO3DvGoo89KC"
      },
      "outputs": [],
      "source": [
        "set(politics_unks) - set(politics_word_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ImS88xgN4R4p"
      },
      "outputs": [],
      "source": [
        "[token for token in politics_unks if any(char.isdigit() for char in token)]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Bu6mKczs4R4q"
      },
      "outputs": [],
      "source": [
        "# Store unk tokens in a file\n",
        "with open('politics_unk_tokens.txt', 'w') as f:\n",
        "  f.write(\"\\n\".join(politics_word_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-IWkuFQU4R4r"
      },
      "outputs": [],
      "source": [
        "politics_word_tokens[:10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8XlaIv3tZdtZ"
      },
      "source": [
        "#### 200 user Simulation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9X4lTMUx91_x"
      },
      "source": [
        "##### Comparations (checkpoint)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xgStouYe96Si",
        "outputId": "5ab1ac9f-8a5f-4e37-cd60-c7bbc8759bec"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Politics tokens: abbreviate, abbreviation, abdominal, abduct, abduction...\n",
            "Nba tokens:      abbreviate, abbreviation, abdominal, abduct, abduction...\n"
          ]
        }
      ],
      "source": [
        "with open('politics_unk_tokens.txt', 'r') as f:\n",
        "  politics_unk_tokens = f.read().split()\n",
        "\n",
        "with open('nba_unk_tokens.txt', 'r') as f:\n",
        "  nba_unk_tokens = f.read().split()\n",
        "\n",
        "print(f\"Politics tokens: { ', '.join(politics_unk_tokens[:5]) }...\")\n",
        "print(f\"Nba tokens:      { ', '.join(nba_unk_tokens[:5]) }...\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jZlOJ6ToXGk-"
      },
      "source": [
        "###### Strictly same"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IM7od_Ev-yWD",
        "outputId": "bfac8f57-4daa-40c6-c9c1-6d47c137adaa"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "8502\n",
            "8502\n"
          ]
        }
      ],
      "source": [
        "print(len(politics_unk_tokens))\n",
        "print(len(nba_unk_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FKzYxt9dGYso"
      },
      "outputs": [],
      "source": [
        "shared_tokens = set(politics_unk_tokens).intersection(set(nba_unk_tokens))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t1gHsdCyGjpL",
        "outputId": "a6702ecb-8e50-4559-c1a8-239dc5c1be1c"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "8502"
            ]
          },
          "execution_count": 21,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(shared_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PJ8JA_J6GxR_",
        "outputId": "d8788ee4-9b75-4ecf-81f9-86f9e968b6d2"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "['abbreviate',\n",
              " 'abbreviation',\n",
              " 'abdominal',\n",
              " 'abduct',\n",
              " 'abduction',\n",
              " 'abdul',\n",
              " 'abdullah',\n",
              " 'ability',\n",
              " 'abnormality',\n",
              " 'abolish']"
            ]
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "sorted(list(shared_tokens))[:10]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QfFhN5joG8fe"
      },
      "outputs": [],
      "source": [
        "#politics_tokenizer.tokenize(\"abdul\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k1wgekaDKd4q",
        "outputId": "c408ecd3-06af-4fd6-ff36-0be2ca401eda"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[]\n",
            "0\n"
          ]
        }
      ],
      "source": [
        "nba_specific = set(nba_unk_tokens) - shared_tokens\n",
        "print(sorted(list(nba_specific))[:10])\n",
        "print(len(nba_specific))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8TJNyEuAKzcl",
        "outputId": "89de7ed0-8126-48ab-d3a7-6d811275b6e3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[]\n",
            "0\n"
          ]
        }
      ],
      "source": [
        "politics_specific = set(politics_unk_tokens) - shared_tokens\n",
        "print(sorted(list(politics_specific))[:10])\n",
        "print(len(politics_specific))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9W7i7Bi3UAUb"
      },
      "outputs": [],
      "source": [
        "from difflib import SequenceMatcher\n",
        "\n",
        "def similar(a, b):\n",
        "    return SequenceMatcher(None, a, b).ratio()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V5PYvWy8XNS7"
      },
      "source": [
        "###### Lemmatization"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PY7_zU4CX52n"
      },
      "outputs": [],
      "source": [
        "lemmatized_politics = set()\n",
        "\n",
        "wnl = WordNetLemmatizer()\n",
        "for word in politics_unk_tokens:\n",
        "   lemmatized_politics.add(wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\"))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BTht3q9oYxR9",
        "outputId": "0aa1b973-ff56-4d86-c484-47aaed7b1213"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(8502, 8496)"
            ]
          },
          "execution_count": 28,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(politics_unk_tokens), len(lemmatized_politics)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "igY1zjGWY8x8"
      },
      "outputs": [],
      "source": [
        "lemmatized_nba = set()\n",
        "\n",
        "wnl = WordNetLemmatizer()\n",
        "for word in nba_unk_tokens:\n",
        "   lemmatized_nba.add(wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\"))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NDWNv2HlZEoq",
        "outputId": "9739a4b3-bf2d-4731-fb3b-37f7c39cb84b"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(8502, 8496)"
            ]
          },
          "execution_count": 30,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(nba_unk_tokens), len(lemmatized_nba)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Fm9Q1S9LZMqB"
      },
      "outputs": [],
      "source": [
        "shared_lemm_tokens = lemmatized_politics.intersection(lemmatized_nba)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tuqHPldSZVUH",
        "outputId": "dcfb96a7-676e-47db-ff72-a4caae37ff76"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "8496"
            ]
          },
          "execution_count": 32,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(shared_lemm_tokens)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LBUC3uh1ZfEu"
      },
      "outputs": [],
      "source": [
        "nba_lemm_specific = lemmatized_nba - shared_lemm_tokens\n",
        "politics_lemm_specific = lemmatized_politics - shared_lemm_tokens"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "c7PYK2xxZkyM"
      },
      "outputs": [],
      "source": [
        "# Execution cca 2 min\n",
        "for index, i in enumerate(nba_lemm_specific):\n",
        "  for j in politics_lemm_specific:\n",
        "    if similar(i, j) > 0.8:\n",
        "      print(i, j)\n",
        "\n",
        "  if index % 100 == 0:\n",
        "    print(f\"          {index} / {len(nba_lemm_specific)}\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 215
        },
        "id": "6YEsKVQNad63",
        "outputId": "f318cada-c536-4b99-e43a-0b075a3c0edf"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'fossils'"
            ]
          },
          "execution_count": 35,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "wnl.lemmatize(\"fossils\", pos=\"v\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AOjGoqsMeVIU",
        "outputId": "f6b1de18-cf24-4ab0-9ab2-e32b60826ac6"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(0, 0, 8496)"
            ]
          },
          "execution_count": 36,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# When nba_lemm only contains nba comments and politics_lemm only politic tokens\n",
        "# nba - 1000, politics 3000, shared 5000 (982, 3021, 5282)\n",
        "\n",
        "len(nba_lemm_specific), len(politics_lemm_specific), len(shared_lemm_tokens)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j3e5FEGUaOmf"
      },
      "source": [
        "##### Random partition (based on hash function)\n",
        "x partitions represending \"x networks of users\".\n",
        "Goal is to put all the tokens in one cluster that matches users from same subreddit. (There should not be cluster with 2 users in different subreddits)   "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8chD8tAiXD0Y"
      },
      "outputs": [],
      "source": [
        "from collections import defaultdict"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IsCSLuK3bcU9"
      },
      "outputs": [],
      "source": [
        "# Load unk lemm tokens\n",
        "with open('nba_unk_tokens.txt', 'r') as n:\n",
        "  all_unk_tokens = n.read().split()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AYPiISuFYWuo"
      },
      "outputs": [],
      "source": [
        "def get_hashed_cluster(_word):\n",
        "    return hash(_word) % 10\n",
        "\n",
        "# Main varaible\n",
        "clusters = defaultdict(list)\n",
        "\n",
        "# Partitions will include NBA specific, Politics specific and shared tokens\n",
        "for word in all_unk_tokens:\n",
        "    clusters[\"c\" + str(get_hashed_cluster(word))].append(word)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0tyhjnDodNA5",
        "outputId": "6d2cf535-fbdc-4608-b9ca-05e5f5b13e0c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "c4 - abbreviate, abortion, abundance, abundant, accessibility, accident, accountability, accurately, additive, adjustment...\n",
            "c2 - abbreviation, abrupt, abusive, accustom, acid, acquaintance, active, adaptive, adequate, admiralty...\n",
            "c8 - abdominal, accessible, acoustic, acquit, activate, adrenaline, aesthetic, affection, aggravate, agile...\n",
            "c6 - abduct, ability, abruptly, accidental, accommodation, accountable, accredit, accreditation, accuracy, achieve...\n",
            "c7 - abduction, abdullah, absolute, abu, abuse, accumulation, adaptation, adjoin, administration, adoptive...\n",
            "c1 - abdul, abolitionist, above, absolutely, academic, account, accountant, achievement, adriatic, adultery...\n",
            "c3 - abnormality, abolition, aboriginal, abstraction, accession, acclaim, acquire, activation, activist, additional...\n",
            "c0 - abolish, accomplish, accomplishment, accurate, activism, addict, additionally, adulthood, adventure, affiliate...\n",
            "c9 - absurd, accidentally, accumulate, accusation, accuse, actively, actual, addiction, adequately, adi...\n",
            "c5 - academia, acceleration, acquisition, activity, adjective, adjunct, admiration, advocate, afterlife, aggression...\n"
          ]
        }
      ],
      "source": [
        "for i, val in clusters.items():\n",
        "    print(i, \"-\", \", \".join(val[:10]) + \"...\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oUJq3MXTiFJn"
      },
      "source": [
        "##### 200 users"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "R4VBbHLblR0D"
      },
      "outputs": [],
      "source": [
        "from collections import defaultdict\n",
        "import json"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jKxpNKJRiH1S"
      },
      "outputs": [],
      "source": [
        "with open(\"nba.txt\") as n, open(\"politics.txt\") as p:\n",
        "    nba_comments = defaultdict(list)\n",
        "    politics_comments = defaultdict(list)\n",
        "\n",
        "    for row in n:\n",
        "        nba_comments[json.loads(row)['author']].append(json.loads(row)['comment'])\n",
        "\n",
        "    for row in p:\n",
        "        politics_comments[json.loads(row)['author']].append(json.loads(row)['comment'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SyaA6bKPiUyV"
      },
      "outputs": [],
      "source": [
        "all_comments = nba_comments.copy()\n",
        "all_comments.update(politics_comments)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ZjqA9TvoSod"
      },
      "source": [
        "###### Get UNK tokens only"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hZCGVITBo3gB",
        "outputId": "de7f8f28-fe06-4726-ec97-31b83ef6b39d"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Package wordnet is already up-to-date!\n",
            "[nltk_data] Downloading package omw-1.4 to /root/nltk_data...\n",
            "[nltk_data]   Package omw-1.4 is already up-to-date!\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "execution_count": 30,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "import nltk\n",
        "import json\n",
        "from nltk.stem import WordNetLemmatizer\n",
        "from tokenization import FullTokenizer\n",
        "from collections import Counter\n",
        "\n",
        "nltk.download(\"wordnet\")\n",
        "nltk.download(\"omw-1.4\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1SJ7Ax2vocMa"
      },
      "outputs": [],
      "source": [
        "user_comments = all_comments['AHarshInquisitor']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xiIOuEWHok4L"
      },
      "outputs": [],
      "source": [
        "# nba_tokenizer = FullTokenizer(\"nba_vocab.txt\", True)\n",
        "# politics_tokenizer = FullTokenizer(\"politics_vocab.txt\", True)\n",
        "wnl = WordNetLemmatizer()\n",
        "merged = \" \".join(user_comments)\n",
        "\n",
        "tokenizer = FullTokenizer(\"vocab.txt\", True)\n",
        "tokenized = tokenizer.tokenize(merged)\n",
        "c = Counter(tokenized).most_common(500)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "t0vDscpmzemJ",
        "outputId": "075100df-9220-446b-c935-9fb9f3ead74a"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'user_tokens_nba = set()\\n\\nfor word in merged.split():\\n    if nba_tokenizer.tokenize(word) == [\"[UNK]\"]:\\n        lem = wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\").lower()\\n        if lem in nba_tokens:\\n            user_tokens_nba.add(lem)'"
            ]
          },
          "execution_count": 33,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"\"\"user_tokens_nba = set()\n",
        "\n",
        "for word in merged.split():\n",
        "    if nba_tokenizer.tokenize(word) == [\"[UNK]\"]:\n",
        "        lem = wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\").lower()\n",
        "        if lem in nba_tokens:\n",
        "            user_tokens_nba.add(lem)\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "2sfVgNPi10MY",
        "outputId": "7bf95ab3-c8de-421d-8a67-2ea8377a9ba4"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'user_tokens_politics = set()\\n\\nfor word in merged.split():\\n    if politics_tokenizer.tokenize(word) == [\"[UNK]\"]:\\n        lemmed = wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\").lower()\\n        if lemmed in politics_tokens:\\n            user_tokens_politics.add(lemmed)'"
            ]
          },
          "execution_count": 34,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"\"\"user_tokens_politics = set()\n",
        "\n",
        "for word in merged.split():\n",
        "    if politics_tokenizer.tokenize(word) == [\"[UNK]\"]:\n",
        "        lemmed = wnl.lemmatize(wnl.lemmatize(word, pos=\"n\"), pos=\"v\").lower()\n",
        "        if lemmed in politics_tokens:\n",
        "            user_tokens_politics.add(lemmed)\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2TMnghUuuigr"
      },
      "outputs": [],
      "source": [
        "def get_user_unk_tokens(user):\n",
        "    joined_comments = \" \".join(all_comments[user])\n",
        "    return [n for n in nba_unk_tokens if n in joined_comments]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I6HOBU17uigs",
        "outputId": "ee81af9c-f97c-4459-a34e-27b065624e96"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "1419"
            ]
          },
          "execution_count": 36,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "rels = { user: get_user_unk_tokens(user) for user in [_ for _ in all_comments][:] }\n",
        "len(rels['BizGilwalker'])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZYkHR5j4t_jz"
      },
      "source": [
        "##### The simulation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_1XUll3Srny4"
      },
      "outputs": [],
      "source": [
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1lIS53e8tZob"
      },
      "outputs": [],
      "source": [
        "# Svi UNK Tokeni usera\n",
        "users200 = rels\n",
        "\n",
        "# Proizvoljan broj clustera\n",
        "# U svakom clusteru se nalaze UNK tokeni\n",
        "# Inicijalno stanje je hash random\n",
        "clusters = clusters"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XCRrf5bUumBj"
      },
      "outputs": [],
      "source": [
        "# User set could have: \"I am working\" and UNK token = \"work\". Func will not count it, but should\n",
        "def get_user_cluster_index(username):\n",
        "    user_set = set(users200[username])\n",
        "    cluster_absolute_count = list(map(lambda cluster: len(set(cluster[1]).intersection(user_set)), clusters.items()))\n",
        "    # print(cluster_absolute_count)\n",
        "    return np.argmax(cluster_absolute_count)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EjdaOG4lrrTx"
      },
      "outputs": [],
      "source": [
        "# Da       (+1 vote) - Ako je User u istom clusteru kao i \"iniciator\" i User ima predloženi token (u svom vokabularu)\n",
        "# Svejedno (+0 vote) - Ako je User u istom clusteru kao i \"iniciator\" i User nema predloženi token (u svom vokabularu)\n",
        "# Svejedno (+0 vote) - Ako User nije u istom clusteru kao i \"iniciator\" i User nema predloženi token (u svom vokabularu)\n",
        "# Ne       (-1 vote) - Ako User nije u istom clusteru kao i \"iniciator\" i User ima predloženi token (u svom vokabularu)\n",
        "\n",
        "# x     y\n",
        "# True  True  +1\n",
        "# True  False  0\n",
        "# False False  0\n",
        "# False True  -1\n",
        "\n",
        "# x y\n",
        "# 1 1 +1\n",
        "# 1 0  0\n",
        "# 0 0  0\n",
        "# 0 1 -1\n",
        "\n",
        "# y * 2x - 1\n",
        "def vote_decision(x: bool, y: bool) -> int:\n",
        "    return int(y) * (2 * int(x) - 1)\n",
        "\n",
        "assert vote_decision(True, True) == 1\n",
        "assert vote_decision(True, False) == 0\n",
        "assert vote_decision(False, False) == 0\n",
        "assert vote_decision(False, True) == -1"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mF1DtqUuvffC"
      },
      "outputs": [],
      "source": [
        "def round_of_vote(username, tolerance = 0, batch_size=1):\n",
        "    user = users200[username]  # All UNKs\n",
        "\n",
        "    # Nadi User Cluster\n",
        "    user_cluster_index = get_user_cluster_index(username)\n",
        "\n",
        "    # Odaberi token za koji se vote-a za transport u taj cluster\n",
        "    tokens_not_in_cluster = set(user).difference(clusters[\"c\" + str(user_cluster_index)])\n",
        "\n",
        "    batch_size = min(len(list(tokens_not_in_cluster)), batch_size)\n",
        "    proposed_tokens = np.random.choice(list(tokens_not_in_cluster), batch_size, False)\n",
        "\n",
        "    proposed_tokens_dictkeys = [[index for index, cluster in clusters.items() if proposed_token in cluster][0] for proposed_token in proposed_tokens]\n",
        "\n",
        "    # Svaki user Vote-a batch\n",
        "    votes = [0] * batch_size\n",
        "    for _username in users200:\n",
        "        same_cluster = get_user_cluster_index(_username) == user_cluster_index\n",
        "        proposed_exist = set(users200[_username]).intersection(set(proposed_tokens))\n",
        "        has_proposed_all = [proposed_token in proposed_exist for proposed_token in proposed_tokens]\n",
        "\n",
        "        vote_decisions = [vote_decision(same_cluster, has_proposed) for has_proposed in has_proposed_all]\n",
        "        votes = list(map(lambda x: sum(x), zip(votes, vote_decisions)))\n",
        "\n",
        "    # Prebaci tokene u novi patch\n",
        "    for i, vote in enumerate(votes):\n",
        "        if vote > -tolerance:\n",
        "            clusters[proposed_tokens_dictkeys[i]].remove(proposed_tokens[i])\n",
        "            clusters[\"c\" + str(user_cluster_index)].append(proposed_tokens[i])\n",
        "            # print(\"Success\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9HR_jLjgTnCt",
        "outputId": "2b64befb-2f80-4d89-f311-0aed5184570c"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "['10', '9', '9', '9', '10', '10', '10', '9', '10', '10']"
            ]
          },
          "execution_count": 42,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "total = 8502\n",
        "[str(int(100 * len(cluster) / total)) for cluster in clusters.values()]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZIRR_XMblKy_"
      },
      "outputs": [],
      "source": [
        "from collections import Counter\n",
        "\n",
        "def purities():\n",
        "    users_clusters_keys = [get_user_cluster_index(_username) for _username in users200]\n",
        "    subbreddits = [\"politics\" if _username in list(politics_comments.keys()) else \"nba\" for _username in users200]\n",
        "\n",
        "    #print(users_clusters_keys)\n",
        "    #print(len(users_clusters_keys), len(subbreddits))\n",
        "    #print(list(zip(users_clusters_keys, subbreddits)))\n",
        "    c = Counter(zip(users_clusters_keys, subbreddits))\n",
        "    return c"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HWIoxaJplpQk"
      },
      "outputs": [],
      "source": [
        "cluster_purities = purities()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 544
        },
        "id": "DdnKND68fnFo",
        "outputId": "1917b32a-0b9e-4c1c-97c6-a1d395baab39"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[[], [], [], [], [], (False, 0.49, 98, 104), [], [], [], (False, 0.4, 2, 3)]\n",
            "['9', '8', '9', '9', '11', '12', '9', '10', '8', '10']\n",
            "[(False, 0.48, 95, 104), (False, 0.67, 2, 1), [], [], [], [], [], [], (False, 0.6, 3, 2), []]\n",
            "['12', '10', '8', '10', '10', '10', '8', '8', '10', '8']\n",
            "[(False, 0.25, 2, 6), [], (False, 0.53, 88, 78), [], [], [], [2], [], (False, 0.32, 10, 21), []]\n",
            "['10', '10', '11', '9', '9', '10', '10', '8', '11', '8']\n",
            "[[], [], [], [], (False, 0.47, 91, 103), [], [], [5], [], (False, 0.5, 4, 4)]\n",
            "['9', '8', '10', '8', '12', '9', '10', '10', '9', '10']\n",
            "[[], (False, 0.49, 98, 104), [], [1], [], (False, 0.33, 1, 2), [], [], [], [1]]\n",
            "['8', '11', '9', '10', '10', '11', '9', '9', '8', '9']\n"
          ]
        },
        {
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-55-62f80ef60c53>\u001b[0m in \u001b[0;36m<cell line: 4>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0m_\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvote_simulation_count\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mrandom_username\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mchoice\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0musers200\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m     \u001b[0mround_of_vote\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrandom_username\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m100\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m     \u001b[0;31m#if _ % 200 == 0:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-41-9bfee8a95e95>\u001b[0m in \u001b[0;36mround_of_vote\u001b[0;34m(username, tolerance, batch_size)\u001b[0m\n\u001b[1;32m     17\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0m_username\u001b[0m \u001b[0;32min\u001b[0m \u001b[0musers200\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     18\u001b[0m         \u001b[0msame_cluster\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_user_cluster_index\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_username\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0muser_cluster_index\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 19\u001b[0;31m         \u001b[0mproposed_exist\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0musers200\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0m_username\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintersection\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mproposed_tokens\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     20\u001b[0m         \u001b[0mhas_proposed_all\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mproposed_token\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mproposed_exist\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mproposed_token\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mproposed_tokens\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     21\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ],
      "source": [
        "purity_req = 0.8\n",
        "vote_simulation_count = 200\n",
        "\n",
        "for _ in range(vote_simulation_count):\n",
        "    random_username = np.random.choice(list(users200.keys()))\n",
        "    round_of_vote(random_username, 10, 100)\n",
        "\n",
        "    #if _ % 200 == 0:\n",
        "        #print(_)\n",
        "\n",
        "    if _ % 10 == 0:\n",
        "        cluster_purities = purities()\n",
        "        cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\n",
        "        print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\n",
        "        print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Y-8qBGe9hACs",
        "outputId": "d65dd68e-6d9a-4fab-8282-3208771b11a5"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "Counter({(5, 'nba'): 96,\n",
              "         (9, 'nba'): 4,\n",
              "         (5, 'politics'): 100,\n",
              "         (9, 'politics'): 7})"
            ]
          },
          "execution_count": 46,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "purities()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ydywj63rhL2J",
        "outputId": "5c26a24e-6122-4a7c-cdbd-b48477852798"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[[], [], [], [], [], (False, 0.49, 96, 100), [], [], [], (False, 0.36, 4, 7)]\n",
            "['9', '8', '9', '8', '11', '12', '9', '10', '9', '10']\n"
          ]
        }
      ],
      "source": [
        "cluster_purities = purities()\n",
        "cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\n",
        "print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\n",
        "print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 107
        },
        "id": "T7CInM1ZgBtF",
        "outputId": "e9a236e6-82d7-4c7f-e509-d428357d4c59"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "\"\\n\\npurity_req = 0.8\\nvote_simulation_count = 5000\\n\\nfor _ in range(vote_simulation_count):\\n    random_username = np.random.choice(list(users200.keys()))\\n    round_of_vote(random_username, 10, 100)\\n    \\n    #if _ % 200 == 0:\\n        #print(_)\\n\\n    if _ % 100 == 0:\\n        cluster_purities = purities()\\n        cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\\n        print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\\n        print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])\\n\\n\\n[[1], [1], (False, 0.4, 2, 3), (True, 0.82, 9, 2), (False, 0.21, 9, 33), (False, 0.64, 44, 25), (False, 0.58, 15, 11), (False, 0.29, 2, 5), (False, 0.42, 17, 23), (False, 0.2, 1, 4)]\\n['9', '10', '10', '9', '9', '9', '10', '9', '10', '9']\\n[[1], (False, 0.55, 36, 30), [1], [4], (True, 0.1, 3, 27), (False, 0.4, 22, 33), [4], [], (False, 0.75, 3, 1), (False, 0.62, 26, 16)]\\n['8', '12', '9', '9', '10', '9', '9', '9', '10', '11']\\n[[], (False, 0.47, 73, 82), [], [], (False, 0.44, 4, 5), (False, 0.23, 6, 20), [], [], [], [17]]\\n['5', '17', '6', '5', '15', '15', '6', '5', '6', '16']\\n[[], [1], [], [], (True, 0.92, 72, 6), [2], [], [], [], (False, 0.2, 25, 101)]\\n['4', '18', '4', '4', '18', '16', '4', '4', '4', '20']\\n[[], (False, 0.51, 35, 33), [], [], (True, 0.03, 1, 32), (False, 0.27, 12, 32), [], [], [], (True, 0.84, 52, 10)]\\n['3', '19', '4', '3', '19', '18', '4', '3', '4', '20']\\n[[], (False, 0.44, 31, 39), [], [], (False, 0.44, 23, 29), (False, 0.32, 14, 30), [], [], [], (False, 0.78, 32, 9)]\\n['3', '19', '3', '3', '20', '19', '3', '3', '3', '20']\\n[[], [], [], [], (True, 0.14, 2, 12), (False, 0.51, 98, 95), [], [], [], []]\\n['2', '18', '3', '3', '21', '20', '3', '3', '3', '20']\\n[[], (False, 0.72, 53, 21), [], [], (True, 0.1, 3, 26), (False, 0.7, 43, 18), [], [], [], (True, 0.02, 1, 42)]\\n['2', '20', '3', '2', '21', '20', '3', '2', '3', '20']\\n[[], (False, 0.48, 29, 32), [], [], (False, 0.48, 29, 32), (False, 0.38, 18, 29), [], [], [], (False, 0.63, 24, 14)]\\n['2', '21', '2', '2', '21', '20', '2', '2', '2', '20']\\n[[], (False, 0.51, 21, 20), [], [], (False, 0.51, 18, 17), (False, 0.55, 33, 27), [], [], [], (False, 0.39, 28, 43)]\\n['2', '21', '2', '2', '21', '20', '2', '2', '2', '21']\\n[[], [], [], [], (False, 0.48, 100, 107), [], [], [], [], []]\\n['2', '21', '2', '2', '23', '18', '2', '2', '2', '21']\\n[[], (False, 0.54, 22, 19), [], [], [1], [50], [], [], [], (False, 0.23, 27, 88)]\\n['2', '22', '2', '2', '21', '19', '2', '2', '2', '22']\\n[[], [], [], [], (False, 0.42, 39, 54), (False, 0.4, 34, 51), [], [], [], (True, 0.93, 27, 2)]\\n['1', '21', '2', '1', '23', '21', '2', '2', '2', '20']\\n[[], [5], [], [], (False, 0.22, 22, 77), (True, 0.89, 67, 8), [], [], [], (False, 0.21, 6, 22)]\\n['1', '21', '2', '1', '23', '20', '2', '2', '2', '21']\\n[[], [6], [], [], (False, 0.63, 70, 41), (True, 0.13, 9, 60), [], [], [], [21]]\\n['1', '22', '2', '1', '22', '22', '2', '1', '2', '20']\\n[[], (False, 0.47, 88, 99), [], [], (False, 0.5, 1, 1), (False, 0.61, 11, 7), [], [], [], []]\\n['1', '23', '2', '1', '20', '23', '2', '1', '2', '20']\\n[[], [], [], [], (False, 0.51, 99, 94), [], [], [], [], (True, 0.07, 1, 13)]\\n['1', '21', '1', '1', '23', '21', '2', '1', '2', '22']\\n[[], (True, 0.1, 4, 36), [], [], (False, 0.49, 26, 27), (True, 0.12, 5, 36), [], [], [], (True, 0.89, 65, 8)]\\n['1', '22', '1', '1', '22', '22', '2', '1', '2', '21']\\n[[], (False, 0.67, 2, 1), [], [], [], (False, 0.48, 98, 106), [], [], [], []]\\n['1', '22', '1', '1', '21', '23', '1', '1', '2', '21']\\n[[], (True, 0.88, 73, 10), [], [], (True, 0.15, 9, 53), (False, 0.33, 1, 2), [], [], [], (False, 0.29, 17, 42)]\\n['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\\n[[], (True, 0.17, 6, 29), [], [], (True, 0.91, 52, 5), (True, 0.92, 23, 2), [], [], [], (False, 0.21, 19, 71)]\\n['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\\n[[], (False, 0.73, 40, 15), [], [], (True, 0.12, 3, 21), (False, 0.45, 57, 71), [], [], [], []]\\n['1', '22', '1', '1', '22', '22', '1', '1', '1', '21']\\n[[], (False, 0.5, 1, 1), [], [], (False, 0.24, 12, 39), [1], [], [], [], (False, 0.57, 87, 66)]\\n['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\\n[[], (False, 0.22, 11, 38), [], [], (False, 0.65, 41, 22), (False, 0.24, 9, 29), [], [], [], (False, 0.68, 39, 18)]\\n['1', '22', '1', '1', '23', '22', '1', '1', '1', '21']\\n[[], [], [], [], (False, 0.49, 99, 104), [1], [], [], [], [3]]\\n['1', '21', '1', '1', '23', '22', '1', '1', '1', '22']\\n[[], (False, 0.33, 1, 2), [], [], [], [], [], [], [], (False, 0.49, 99, 105)]\\n['1', '22', '1', '1', '21', '22', '1', '1', '1', '23']\\n[[], (False, 0.77, 51, 15), [], [], (True, 0.12, 4, 29), (False, 0.51, 40, 39), [], [], [], (True, 0.17, 5, 24)]\\n['1', '22', '1', '1', '22', '22', '1', '1', '1', '23']\\n[[], (False, 0.25, 1, 3), [], [], [27], (False, 0.41, 71, 104), [], [], [], [1]]\\n['1', '22', '1', '1', '23', '22', '1', '1', '1', '21']\\n[[], (False, 0.58, 41, 30), [], [], [1], (False, 0.43, 58, 76), [], [], [], [1]]\\n['1', '22', '1', '1', '22', '23', '1', '1', '1', '22']\\n[[], [4], [], [], (False, 0.78, 18, 5), (False, 0.44, 77, 100), [], [], [], (False, 0.33, 1, 2)]\\n['1', '21', '1', '1', '23', '24', '1', '1', '1', '21']\\n[[], (True, 0.05, 1, 18), [], [], (False, 0.42, 17, 23), (False, 0.66, 80, 42), [], [], [], (True, 0.08, 2, 24)]\\n['1', '22', '1', '1', '23', '23', '1', '1', '1', '21']\\n[[], (True, 0.9, 28, 3), [], [], (False, 0.4, 2, 3), [], [], [], [], (False, 0.41, 70, 101)]\\n['1', '22', '1', '1', '23', '22', '1', '1', '1', '23']\\n[[], (False, 0.47, 27, 31), [], [], (False, 0.64, 39, 22), (True, 0.04, 1, 24), [], [], [], (False, 0.52, 33, 30)]\\n['1', '21', '1', '1', '23', '22', '1', '1', '1', '23']\\n[[], (False, 0.56, 36, 28), [], [], (False, 0.46, 17, 20), (False, 0.56, 31, 24), [], [], [], (False, 0.31, 16, 35)]\\n['1', '22', '1', '1', '23', '23', '1', '1', '1', '22']\\n[[], (False, 0.24, 31, 96), [], [], [], [], [], [], [], (True, 0.86, 69, 11)]\\n['1', '23', '1', '1', '23', '21', '1', '1', '1', '22']\\n[[], (False, 0.49, 100, 106), [], [], [], [1], [], [], [], []]\\n['1', '23', '1', '1', '23', '22', '1', '1', '1', '22']\\n[[], (False, 0.47, 91, 102), [], [], [], (False, 0.64, 9, 5), [], [], [], []]\\n['1', '24', '1', '1', '22', '23', '1', '1', '1', '20']\\n[[], (False, 0.37, 22, 37), [], [], (False, 0.7, 30, 13), (False, 0.33, 13, 26), [], [], [], (False, 0.53, 35, 31)]\\n['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\\n[[], [], [], [], (False, 0.48, 98, 107), [2], [], [], [], []]\\n['1', '21', '1', '1', '24', '22', '1', '1', '1', '23']\\n[[], (True, 0.12, 6, 46), [], [], (False, 0.56, 50, 40), [1], [], [], [], (False, 0.69, 44, 20)]\\n['1', '23', '1', '1', '23', '21', '1', '1', '1', '22']\\n[[], (False, 0.49, 99, 105), [], [], [2], [1], [], [], [], []]\\n['1', '24', '1', '1', '23', '22', '1', '1', '1', '21']\\n[[], [], [], [], (False, 0.49, 99, 103), [], [], [], [], (False, 0.2, 1, 4)]\\n['1', '21', '1', '1', '24', '23', '1', '1', '1', '22']\\n[[], [], [], [], (False, 0.24, 12, 38), (False, 0.57, 87, 66), [], [], [], (False, 0.25, 1, 3)]\\n['1', '21', '1', '1', '24', '24', '1', '1', '1', '22']\\n[[], (True, 0.88, 78, 11), [], [], (True, 0.02, 1, 57), (True, 0.88, 7, 1), [], [], [], (False, 0.27, 14, 38)]\\n['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\\n[[], [1], [], [], [], [], [], [], [], (False, 0.49, 100, 106)]\\n['1', '22', '1', '1', '23', '23', '1', '1', '1', '23']\\n[[], [], [], [], (False, 0.5, 1, 1), (False, 0.48, 99, 106), [], [], [], []]\\n['1', '21', '1', '1', '23', '24', '1', '1', '1', '22']\\n[[], (False, 0.74, 28, 10), [], [], (False, 0.41, 41, 58), (False, 0.67, 2, 1), [], [], [], (False, 0.43, 29, 38)]\\n['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\\n[[], (False, 0.26, 12, 34), [], [], (False, 0.53, 28, 25), (False, 0.8, 56, 14), [], [], [], (True, 0.11, 4, 34)]\\n['1', '22', '1', '1', '24', '23', '1', '1', '1', '22']\\n[[], (False, 0.67, 41, 20), [], [], (False, 0.39, 13, 20), (False, 0.48, 31, 33), [], [], [], (False, 0.31, 15, 34)]\\n['1', '22', '1', '1', '23', '22', '1', '1', '1', '23']\\n[[], (False, 0.75, 21, 7), [], [], [], (False, 0.69, 25, 11), [], [], [], (False, 0.38, 54, 89)]\\n['1', '23', '1', '1', '23', '22', '1', '1', '1', '23']\\n\""
            ]
          },
          "execution_count": 48,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"\"\"\n",
        "\n",
        "purity_req = 0.8\n",
        "vote_simulation_count = 5000\n",
        "\n",
        "for _ in range(vote_simulation_count):\n",
        "    random_username = np.random.choice(list(users200.keys()))\n",
        "    round_of_vote(random_username, 10, 100)\n",
        "\n",
        "    #if _ % 200 == 0:\n",
        "        #print(_)\n",
        "\n",
        "    if _ % 100 == 0:\n",
        "        cluster_purities = purities()\n",
        "        cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\n",
        "        print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\n",
        "        print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])\n",
        "\n",
        "\n",
        "[[1], [1], (False, 0.4, 2, 3), (True, 0.82, 9, 2), (False, 0.21, 9, 33), (False, 0.64, 44, 25), (False, 0.58, 15, 11), (False, 0.29, 2, 5), (False, 0.42, 17, 23), (False, 0.2, 1, 4)]\n",
        "['9', '10', '10', '9', '9', '9', '10', '9', '10', '9']\n",
        "[[1], (False, 0.55, 36, 30), [1], [4], (True, 0.1, 3, 27), (False, 0.4, 22, 33), [4], [], (False, 0.75, 3, 1), (False, 0.62, 26, 16)]\n",
        "['8', '12', '9', '9', '10', '9', '9', '9', '10', '11']\n",
        "[[], (False, 0.47, 73, 82), [], [], (False, 0.44, 4, 5), (False, 0.23, 6, 20), [], [], [], [17]]\n",
        "['5', '17', '6', '5', '15', '15', '6', '5', '6', '16']\n",
        "[[], [1], [], [], (True, 0.92, 72, 6), [2], [], [], [], (False, 0.2, 25, 101)]\n",
        "['4', '18', '4', '4', '18', '16', '4', '4', '4', '20']\n",
        "[[], (False, 0.51, 35, 33), [], [], (True, 0.03, 1, 32), (False, 0.27, 12, 32), [], [], [], (True, 0.84, 52, 10)]\n",
        "['3', '19', '4', '3', '19', '18', '4', '3', '4', '20']\n",
        "[[], (False, 0.44, 31, 39), [], [], (False, 0.44, 23, 29), (False, 0.32, 14, 30), [], [], [], (False, 0.78, 32, 9)]\n",
        "['3', '19', '3', '3', '20', '19', '3', '3', '3', '20']\n",
        "[[], [], [], [], (True, 0.14, 2, 12), (False, 0.51, 98, 95), [], [], [], []]\n",
        "['2', '18', '3', '3', '21', '20', '3', '3', '3', '20']\n",
        "[[], (False, 0.72, 53, 21), [], [], (True, 0.1, 3, 26), (False, 0.7, 43, 18), [], [], [], (True, 0.02, 1, 42)]\n",
        "['2', '20', '3', '2', '21', '20', '3', '2', '3', '20']\n",
        "[[], (False, 0.48, 29, 32), [], [], (False, 0.48, 29, 32), (False, 0.38, 18, 29), [], [], [], (False, 0.63, 24, 14)]\n",
        "['2', '21', '2', '2', '21', '20', '2', '2', '2', '20']\n",
        "[[], (False, 0.51, 21, 20), [], [], (False, 0.51, 18, 17), (False, 0.55, 33, 27), [], [], [], (False, 0.39, 28, 43)]\n",
        "['2', '21', '2', '2', '21', '20', '2', '2', '2', '21']\n",
        "[[], [], [], [], (False, 0.48, 100, 107), [], [], [], [], []]\n",
        "['2', '21', '2', '2', '23', '18', '2', '2', '2', '21']\n",
        "[[], (False, 0.54, 22, 19), [], [], [1], [50], [], [], [], (False, 0.23, 27, 88)]\n",
        "['2', '22', '2', '2', '21', '19', '2', '2', '2', '22']\n",
        "[[], [], [], [], (False, 0.42, 39, 54), (False, 0.4, 34, 51), [], [], [], (True, 0.93, 27, 2)]\n",
        "['1', '21', '2', '1', '23', '21', '2', '2', '2', '20']\n",
        "[[], [5], [], [], (False, 0.22, 22, 77), (True, 0.89, 67, 8), [], [], [], (False, 0.21, 6, 22)]\n",
        "['1', '21', '2', '1', '23', '20', '2', '2', '2', '21']\n",
        "[[], [6], [], [], (False, 0.63, 70, 41), (True, 0.13, 9, 60), [], [], [], [21]]\n",
        "['1', '22', '2', '1', '22', '22', '2', '1', '2', '20']\n",
        "[[], (False, 0.47, 88, 99), [], [], (False, 0.5, 1, 1), (False, 0.61, 11, 7), [], [], [], []]\n",
        "['1', '23', '2', '1', '20', '23', '2', '1', '2', '20']\n",
        "[[], [], [], [], (False, 0.51, 99, 94), [], [], [], [], (True, 0.07, 1, 13)]\n",
        "['1', '21', '1', '1', '23', '21', '2', '1', '2', '22']\n",
        "[[], (True, 0.1, 4, 36), [], [], (False, 0.49, 26, 27), (True, 0.12, 5, 36), [], [], [], (True, 0.89, 65, 8)]\n",
        "['1', '22', '1', '1', '22', '22', '2', '1', '2', '21']\n",
        "[[], (False, 0.67, 2, 1), [], [], [], (False, 0.48, 98, 106), [], [], [], []]\n",
        "['1', '22', '1', '1', '21', '23', '1', '1', '2', '21']\n",
        "[[], (True, 0.88, 73, 10), [], [], (True, 0.15, 9, 53), (False, 0.33, 1, 2), [], [], [], (False, 0.29, 17, 42)]\n",
        "['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\n",
        "[[], (True, 0.17, 6, 29), [], [], (True, 0.91, 52, 5), (True, 0.92, 23, 2), [], [], [], (False, 0.21, 19, 71)]\n",
        "['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\n",
        "[[], (False, 0.73, 40, 15), [], [], (True, 0.12, 3, 21), (False, 0.45, 57, 71), [], [], [], []]\n",
        "['1', '22', '1', '1', '22', '22', '1', '1', '1', '21']\n",
        "[[], (False, 0.5, 1, 1), [], [], (False, 0.24, 12, 39), [1], [], [], [], (False, 0.57, 87, 66)]\n",
        "['1', '22', '1', '1', '23', '21', '1', '1', '1', '22']\n",
        "[[], (False, 0.22, 11, 38), [], [], (False, 0.65, 41, 22), (False, 0.24, 9, 29), [], [], [], (False, 0.68, 39, 18)]\n",
        "['1', '22', '1', '1', '23', '22', '1', '1', '1', '21']\n",
        "[[], [], [], [], (False, 0.49, 99, 104), [1], [], [], [], [3]]\n",
        "['1', '21', '1', '1', '23', '22', '1', '1', '1', '22']\n",
        "[[], (False, 0.33, 1, 2), [], [], [], [], [], [], [], (False, 0.49, 99, 105)]\n",
        "['1', '22', '1', '1', '21', '22', '1', '1', '1', '23']\n",
        "[[], (False, 0.77, 51, 15), [], [], (True, 0.12, 4, 29), (False, 0.51, 40, 39), [], [], [], (True, 0.17, 5, 24)]\n",
        "['1', '22', '1', '1', '22', '22', '1', '1', '1', '23']\n",
        "[[], (False, 0.25, 1, 3), [], [], [27], (False, 0.41, 71, 104), [], [], [], [1]]\n",
        "['1', '22', '1', '1', '23', '22', '1', '1', '1', '21']\n",
        "[[], (False, 0.58, 41, 30), [], [], [1], (False, 0.43, 58, 76), [], [], [], [1]]\n",
        "['1', '22', '1', '1', '22', '23', '1', '1', '1', '22']\n",
        "[[], [4], [], [], (False, 0.78, 18, 5), (False, 0.44, 77, 100), [], [], [], (False, 0.33, 1, 2)]\n",
        "['1', '21', '1', '1', '23', '24', '1', '1', '1', '21']\n",
        "[[], (True, 0.05, 1, 18), [], [], (False, 0.42, 17, 23), (False, 0.66, 80, 42), [], [], [], (True, 0.08, 2, 24)]\n",
        "['1', '22', '1', '1', '23', '23', '1', '1', '1', '21']\n",
        "[[], (True, 0.9, 28, 3), [], [], (False, 0.4, 2, 3), [], [], [], [], (False, 0.41, 70, 101)]\n",
        "['1', '22', '1', '1', '23', '22', '1', '1', '1', '23']\n",
        "[[], (False, 0.47, 27, 31), [], [], (False, 0.64, 39, 22), (True, 0.04, 1, 24), [], [], [], (False, 0.52, 33, 30)]\n",
        "['1', '21', '1', '1', '23', '22', '1', '1', '1', '23']\n",
        "[[], (False, 0.56, 36, 28), [], [], (False, 0.46, 17, 20), (False, 0.56, 31, 24), [], [], [], (False, 0.31, 16, 35)]\n",
        "['1', '22', '1', '1', '23', '23', '1', '1', '1', '22']\n",
        "[[], (False, 0.24, 31, 96), [], [], [], [], [], [], [], (True, 0.86, 69, 11)]\n",
        "['1', '23', '1', '1', '23', '21', '1', '1', '1', '22']\n",
        "[[], (False, 0.49, 100, 106), [], [], [], [1], [], [], [], []]\n",
        "['1', '23', '1', '1', '23', '22', '1', '1', '1', '22']\n",
        "[[], (False, 0.47, 91, 102), [], [], [], (False, 0.64, 9, 5), [], [], [], []]\n",
        "['1', '24', '1', '1', '22', '23', '1', '1', '1', '20']\n",
        "[[], (False, 0.37, 22, 37), [], [], (False, 0.7, 30, 13), (False, 0.33, 13, 26), [], [], [], (False, 0.53, 35, 31)]\n",
        "['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\n",
        "[[], [], [], [], (False, 0.48, 98, 107), [2], [], [], [], []]\n",
        "['1', '21', '1', '1', '24', '22', '1', '1', '1', '23']\n",
        "[[], (True, 0.12, 6, 46), [], [], (False, 0.56, 50, 40), [1], [], [], [], (False, 0.69, 44, 20)]\n",
        "['1', '23', '1', '1', '23', '21', '1', '1', '1', '22']\n",
        "[[], (False, 0.49, 99, 105), [], [], [2], [1], [], [], [], []]\n",
        "['1', '24', '1', '1', '23', '22', '1', '1', '1', '21']\n",
        "[[], [], [], [], (False, 0.49, 99, 103), [], [], [], [], (False, 0.2, 1, 4)]\n",
        "['1', '21', '1', '1', '24', '23', '1', '1', '1', '22']\n",
        "[[], [], [], [], (False, 0.24, 12, 38), (False, 0.57, 87, 66), [], [], [], (False, 0.25, 1, 3)]\n",
        "['1', '21', '1', '1', '24', '24', '1', '1', '1', '22']\n",
        "[[], (True, 0.88, 78, 11), [], [], (True, 0.02, 1, 57), (True, 0.88, 7, 1), [], [], [], (False, 0.27, 14, 38)]\n",
        "['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\n",
        "[[], [1], [], [], [], [], [], [], [], (False, 0.49, 100, 106)]\n",
        "['1', '22', '1', '1', '23', '23', '1', '1', '1', '23']\n",
        "[[], [], [], [], (False, 0.5, 1, 1), (False, 0.48, 99, 106), [], [], [], []]\n",
        "['1', '21', '1', '1', '23', '24', '1', '1', '1', '22']\n",
        "[[], (False, 0.74, 28, 10), [], [], (False, 0.41, 41, 58), (False, 0.67, 2, 1), [], [], [], (False, 0.43, 29, 38)]\n",
        "['1', '22', '1', '1', '24', '22', '1', '1', '1', '22']\n",
        "[[], (False, 0.26, 12, 34), [], [], (False, 0.53, 28, 25), (False, 0.8, 56, 14), [], [], [], (True, 0.11, 4, 34)]\n",
        "['1', '22', '1', '1', '24', '23', '1', '1', '1', '22']\n",
        "[[], (False, 0.67, 41, 20), [], [], (False, 0.39, 13, 20), (False, 0.48, 31, 33), [], [], [], (False, 0.31, 15, 34)]\n",
        "['1', '22', '1', '1', '23', '22', '1', '1', '1', '23']\n",
        "[[], (False, 0.75, 21, 7), [], [], [], (False, 0.69, 25, 11), [], [], [], (False, 0.38, 54, 89)]\n",
        "['1', '23', '1', '1', '23', '22', '1', '1', '1', '23']\n",
        "\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "id": "9njFvXx95EDA",
        "outputId": "1e914867-6e59-4d72-f606-e985c692f060"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'\\npurity_req = 0.85\\nvote_simulation_count = 20000\\n\\nfor _ in range(vote_simulation_count):\\n    random_username = np.random.choice(list(users200.keys()))\\n    round_of_vote(random_username, 0, 100)\\n    \\n    if _ % 200 == 0:\\n        print(_)\\n\\n    if _ % 1000 == 0:\\n        cluster_purities = purities()\\n        cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\\n        print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\\n        print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])\\n'"
            ]
          },
          "execution_count": 49,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"\"\"\n",
        "purity_req = 0.85\n",
        "vote_simulation_count = 20000\n",
        "\n",
        "for _ in range(vote_simulation_count):\n",
        "    random_username = np.random.choice(list(users200.keys()))\n",
        "    round_of_vote(random_username, 0, 100)\n",
        "\n",
        "    if _ % 200 == 0:\n",
        "        print(_)\n",
        "\n",
        "    if _ % 1000 == 0:\n",
        "        cluster_purities = purities()\n",
        "        cluster_purities_tpl = [[(i[0][1], i[1]) for i in cluster_purities.items() if i[0][0] == c1] for c1 in range(len(clusters.keys()))]\n",
        "        print(list(map(lambda y: ((z:=round(y[0] / sum(y), 2)) > purity_req or z < (1-purity_req), z, y[0], y[1]) if len(y) > 1 else y, [list(map(lambda c: c[1], i)) for i in cluster_purities_tpl])))\n",
        "        print([str(int(100 * len(cluster) / total)) for cluster in clusters.values()])\n",
        "\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "qnEYvIecrp1H",
        "outputId": "aab7fbcb-6607-41ab-b578-707d1b0da0b6"
      },
      "outputs": [
        {
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'[str(int(100 * len(cluster) / total)) for cluster in clusters.values()]'"
            ]
          },
          "execution_count": 50,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "\"[str(int(100 * len(cluster) / total)) for cluster in clusters.values()]\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8UoRvvUadTap"
      },
      "outputs": [],
      "source": [
        "#with open(\"clusters.txt\", \"w\") as c:\n",
        "    #c.write(json.dumps(clusters))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "edZGCr40EHqh"
      },
      "outputs": [],
      "source": [
        "#for key, value in clusters2.items():\n",
        "    #print(key, value)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Eow0uWwEEXyS"
      },
      "outputs": [],
      "source": [
        "# import copy\n",
        "# clusters2 = copy.deepcopy(clusters)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "89zA2eecAIcM"
      },
      "outputs": [],
      "source": [
        "# [len(users200[user]) for user in users200.keys()]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l9fMkZ884x3T"
      },
      "source": [
        "#### K-means alternative (fail)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2f3e-0CC9vuM"
      },
      "outputs": [],
      "source": [
        "import json\n",
        "import numpy as np\n",
        "from tokenization import FullTokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "baMGr5ohYxs8"
      },
      "outputs": [],
      "source": [
        "# Za funkciju treba dobiti 2d array.\n",
        "# Prva dimenzija su svih 200 usera.\n",
        "# Druga dimenzija je array relativnih frekvencija UNK tokena (postotak pojavljivanja i-tog UNK tokena nasprem svih UNK tokena (od svih 200 usera))\n",
        "#    [0.3%, 0%, 0.6%, 0.02%, ...] gdje je svaki element count(UNKs[i]) / sum(UNKs))\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1YYSu8_V1klT"
      },
      "source": [
        "##### Random user's all comments"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fOd4S_fB0Dvf"
      },
      "outputs": [],
      "source": [
        "from collections import defaultdict"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ulbYhXGx0Dvh"
      },
      "outputs": [],
      "source": [
        "with open(\"nba.txt\") as n, open(\"politics.txt\") as p:\n",
        "    nba_comments = defaultdict(list)\n",
        "    politics_comments = defaultdict(list)\n",
        "\n",
        "    for row in n:\n",
        "        nba_comments[json.loads(row)['author']].append(json.loads(row)['comment'])\n",
        "\n",
        "    for row in p:\n",
        "        politics_comments[json.loads(row)['author']].append(json.loads(row)['comment'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Nw8erwzhYzma"
      },
      "outputs": [],
      "source": [
        "with open('politics_unk_tokens.txt', 'r') as p, open('nba_unk_tokens.txt', 'r') as n:\n",
        "  politics_unk_tokens = p.read().split()\n",
        "  nba_unk_tokens = n.read().split()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "A3D0pbOW0Dvi"
      },
      "outputs": [],
      "source": [
        "all_comments = nba_comments.copy()\n",
        "all_comments.update(politics_comments)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(all_comments.keys())\n",
        "print(politics_comments.keys())"
      ],
      "metadata": {
        "id": "zq-8SloPkMRU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ff4a11cf-f19b-4867-8ab7-d89043006da3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "dict_keys(['BizGilwalker', 'Laaakerrrsss', 'f5_szn', 'HokageEzio', 'allstar96', 'themariokarters', 'DeathBySuplex', 'wackysama', 'vrythngsgngtblrght', 'OrangeKookie', 'ZandrickEllison', 'clevelandfansblow', 'klawhileonard', 'Falt_ssb', 'ShaolinCheesecake', 'Jahsay', 'snacktastic2', 'bball84958294', 'WheresMySaucePlease', 'Whydidideletemyaccou', 'Hyperactivity786', 'sop1232', 'LuckysBestMan', 'DonEYeet', 'JewRepublican69', 'tapk69', 'jcw4455', 'Roger--Smith', 'lakerswiz', 'ward0630', 'BenjaminJamesGrimm', 'victor396', 'sunglao', 'The_Natural_Log', 'LanaRhoades-', 'BIizard', 'nakedsamurai', 'friskydongo', 'WordsAreSomething', 'TOP_5_DOA', 'Sayter', 'TuneHD', 'Mysteriagant', 'NE_ED', 'gotdemmadsquirtsyo', 'MacDerfus', 'PENIS__FINGERS', 'PormanNowell', 'IncaseAce', 'Bladex10', 'ireddit270', 'toftr', 'IndigoRivers', 'BuzzedWoody', 'TheOriginalElTigre', 'magikarptoothbrush', 'jack12345618', 'ashwinr136', 'oakled', 'jbranton30', 'ChrowanWestbrook', 'chad12341296', 'Anti_Thon', 'Byrie_Swirving', 'augustineiver', 'jaylen_browns_beard', 'gookieboy', 'Duche07', 'DeerOnTheRocks', 'WildYams', 'jhall0310', 'mrsuns10', 'DeanBlandino', 'keefstrong', 'DeArmani_DeBooker', 'Dctr_K', 'WiltAbdulOlajuwon', 'TheHeartfeltKid', 'IdEgoLeBron', 'livefreeordont', 'SwarleymonLives', 'RedSoxEatPoop', 'ITsLoverBoy', 'skepticdoubt', 'BigPoppaPuff', 'TheLastSecondShot', 'Iwasbornathalfcourt', 'anser1233', 'DummedDownInThot', 'ReplyGod', 'burnerfret', 'brisketis', 'Saprano44', 'definitelywrongbud', 'LanceWillMakeEmDance', 'alarmed__fan', 'oftenevil', 'back2god', 'supremehat13', 'sourcreamonionchipz', 'muffler48', 'AisleOfRussia', 'TrumpSucksMyBalls', 'spolio', 'Lamont-Cranston', 'kevie3drinks', '9001_Dalmatians', 'BenedictD0nald', 'TheDemonrat', 'DBDude', 'ryanbuck_', 'NatashaStyles', 'StatisticalAstronaut', 'AlternativeSuccotash', 'TAINT-TEAM_dorito', 'Whose_asking', 'michaelcharlie8', 'Acidporisu', 'Enjolras55', 'CarmineFields', 'ScholarOfTwilight', 'goforthandconquer', 'TowelCarryingTourist', 'DJTHatesPuertoRicans', 'geodynamics', 'Smallmammal', 'BringOn25A', 'Jump_Yossarian', 'nyonpa108', 'blue_crab86', 'mr_fn_wonderful', 'objectivedesigning', '__NamasteMF__', 'hup-hup-hup', 'OneReportersOpinion', 'whitenoise2323', 'derGropenfuhrer', 'Memetic1', 'Nomandate', 'out_of_ideas123', 'autotldr', 'Spacedman-Spliff', 'IUseThisForThings8', 'lofi76', 'masstrip', 'captaincanada84', 'SprayTanCaligula', 'dismayedcitizen', 'Geralt_of_Rivia1', 'thelizardkin', 'JijiLV29', 'DesperateDem', 'Gingold', '38SSM', 'JeremyMo88', 'TrumpUberAlles', 'BigBoyWithTheButter', 'accountabilitycounts', 'Lochmon', 'slapnflop', 'ConanTheProletarian', 'MBAMBA0', 'A_Dogfish', 'txyesboy', 'makoivis', 'El_Hamaultagu', 'charmed_im-sure', '2018Blunami', 'Jeffersons_Mammoth', 'Ruebarbara', 'drvondoctor', 'SwingJay1', 'CoreWrect', 'gpl2017', 'Kimball_Kinnison', 'Hoxha-Posadist', 'Pint_and_Grub', 'PhillyIndy', 'zzzigzzzagzzziggy', 'SlimLovin', 'tiredofwinning12345', 'VROF', 'AHarshInquisitor', 'slakmehl', 'ubix', 'imnotanevilwitch', 'Nostromo1905', 'FDRs_ghost', 'CovfefeForAll', 'Thymdahl', 'PoliticalPleionosis', 'JollyMurderousGhoul', 'WatchingDonFail', 'EnlightenedMind_420', 'FUCK_BALLS_SHIT_ASS', 'AndIAmEric', 'Biptoslipdi', 'jimmydean885', 'wyvernwy', 'Corvus_Clanculum', 'F3715H', 'RussianTrollPatrol', 'TrumpIsATraitor420', 'ISuspectFuckery', 'granola_robot', 'footsurgery', 'PunkAssMutha'])\n",
            "dict_keys(['muffler48', 'AisleOfRussia', 'TrumpSucksMyBalls', 'spolio', 'Lamont-Cranston', 'kevie3drinks', '9001_Dalmatians', 'BenedictD0nald', 'TheDemonrat', 'DBDude', 'ryanbuck_', 'NatashaStyles', 'StatisticalAstronaut', 'AlternativeSuccotash', 'TAINT-TEAM_dorito', 'Whose_asking', 'michaelcharlie8', 'Acidporisu', 'Enjolras55', 'CarmineFields', 'ScholarOfTwilight', 'goforthandconquer', 'TowelCarryingTourist', 'DJTHatesPuertoRicans', 'geodynamics', 'Smallmammal', 'BringOn25A', 'Jump_Yossarian', 'nyonpa108', 'blue_crab86', 'mr_fn_wonderful', 'objectivedesigning', '__NamasteMF__', 'hup-hup-hup', 'OneReportersOpinion', 'whitenoise2323', 'derGropenfuhrer', 'Memetic1', 'Nomandate', 'out_of_ideas123', 'autotldr', 'Spacedman-Spliff', 'IUseThisForThings8', 'lofi76', 'masstrip', 'captaincanada84', 'SprayTanCaligula', 'dismayedcitizen', 'Geralt_of_Rivia1', 'thelizardkin', 'JijiLV29', 'DesperateDem', 'Gingold', '38SSM', 'JeremyMo88', 'TrumpUberAlles', 'BigBoyWithTheButter', 'accountabilitycounts', 'Lochmon', 'slapnflop', 'ConanTheProletarian', 'MBAMBA0', 'A_Dogfish', 'txyesboy', 'makoivis', 'El_Hamaultagu', 'charmed_im-sure', '2018Blunami', 'Jeffersons_Mammoth', 'Ruebarbara', 'drvondoctor', 'SwingJay1', 'CoreWrect', 'gpl2017', 'Kimball_Kinnison', 'Hoxha-Posadist', 'Pint_and_Grub', 'PhillyIndy', 'zzzigzzzagzzziggy', 'SlimLovin', 'tiredofwinning12345', 'VROF', 'AHarshInquisitor', 'slakmehl', 'ubix', 'imnotanevilwitch', 'Nostromo1905', 'FDRs_ghost', 'CovfefeForAll', 'Thymdahl', 'PoliticalPleionosis', 'JollyMurderousGhoul', 'WatchingDonFail', 'EnlightenedMind_420', 'FUCK_BALLS_SHIT_ASS', 'AndIAmEric', 'Biptoslipdi', 'jimmydean885', 'wyvernwy', 'Corvus_Clanculum', 'F3715H', 'RussianTrollPatrol', 'TrumpIsATraitor420', 'ISuspectFuckery', 'granola_robot', 'footsurgery', 'PunkAssMutha'])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "W5MEze1F0IJX"
      },
      "outputs": [],
      "source": [
        "user_comments = nba_comments['Anti_Thon']\n",
        "user_comments_str = \". \".join(user_comments)\n",
        "small_tokenizer = FullTokenizer(\"small_vocab.txt\", True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JCgLwdy-PHLb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a1eefcc2-b80c-469c-ca60-1cef3cd7939e"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[['[UNK]', ',', 'is', 'this', 'really', 'w'],\n",
              " ['##e', 'than', 'the', '[UNK]', 'that', '[UNK]'],\n",
              " ['[UNK]', 'that', '[UNK]', '[UNK]', 'w', '##r'],\n",
              " ['[UNK]', 'w', '##r', '##o', '##t', '##e'],\n",
              " ['##o', '##t', '##e', '?', '[UNK]', ','],\n",
              " ['?', '[UNK]', ',', 'i', 'was', 'to'],\n",
              " ['i', 'was', 'to', '##l', '##d', '[UNK]'],\n",
              " ['##l', '##d', '[UNK]', 'was', 'a', 'la'],\n",
              " ['##k', '##er', 'in', '[UNK]', '.', '[UNK]'],\n",
              " ['[UNK]', '.', '[UNK]', 'l', '##m', '##f'],\n",
              " ['with', 'a', 'top', '3', '[UNK]', '[UNK]'],\n",
              " ['3', '[UNK]', '[UNK]', 'and', 'the', 'd'],\n",
              " ['a', '6', '##0', 'win', '[UNK]', 'when'],\n",
              " ['win', '[UNK]', 'when', 'he', 'played', 'he'],\n",
              " ['.', '.', '.', '[UNK]', 'didn', '’'],\n",
              " ['[UNK]', 'didn', '’', 't', 're', '-'],\n",
              " ['##c', '##e', ',', '[UNK]', 'and', 'west'],\n",
              " ['[UNK]', 'and', 'west', 'left', '.', 'from'],\n",
              " ['left', '.', 'from', 'that', '[UNK]', 'team'],\n",
              " ['that', '[UNK]', 'team', ',', 'it', 'was'],\n",
              " ['just', 'him', 'and', '[UNK]', '.', 'did'],\n",
              " ['[UNK]', '.', 'did', 'anyone', 'even', 'watch'],\n",
              " ['part', '1', 'of', 'his', '[UNK]', '?'],\n",
              " ['his', '[UNK]', '?', 'they', 'made', 'it'],\n",
              " ['there', '.', 'he', '[UNK]', 'like', 'they'],\n",
              " ['[UNK]', 'like', 'they', 'didn', '’', 't'],\n",
              " ['.', 'they', 'could', '[UNK]', 're', '##g'],\n",
              " ['[UNK]', 're', '##g', '##r', '##e', '##t'],\n",
              " ['if', '(', 'go', '##d', '[UNK]', ')'],\n",
              " ['##d', '[UNK]', ')', '[UNK]', 'gets', '[UNK]'],\n",
              " ['[UNK]', 'gets', '[UNK]', 'again', 'we', 'have'],\n",
              " ['again', 'we', 'have', 'no', '[UNK]', 'but'],\n",
              " ['no', '[UNK]', 'but', 'i', 'think', '[UNK]'],\n",
              " ['i', 'think', '[UNK]', ',', 'his', 'best'],\n",
              " ['2', '+', '1', '[UNK]', 'being', '[UNK]'],\n",
              " ['[UNK]', 'being', '[UNK]', 'for', 'this', '.'],\n",
              " ['.', '.', 'actually', '[UNK]', ',', 'the'],\n",
              " ['[UNK]', ',', 'the', 'lakers', 'fa', '##n'],\n",
              " ['should', 'see', 'the', 'hate', '[UNK]', 'i'],\n",
              " ['hate', '[UNK]', 'i', 'get', 'in', 'my'],\n",
              " ['##x', 'from', 'them', '[UNK]', 'fuck', '.'],\n",
              " ['[UNK]', 'fuck', '.', 'this', 'is', '[UNK]'],\n",
              " ['this', 'is', '[UNK]', 'for', 'us', '.'],\n",
              " ['t', '##a', '##x', '[UNK]', 'is', '[UNK]'],\n",
              " ['[UNK]', 'is', '[UNK]', 'but', '[UNK]', 'probably'],\n",
              " ['but', '[UNK]', 'probably', 'would', 'have', 'got'],\n",
              " ['it', '’', 's', 'a', '[UNK]', 'con'],\n",
              " ['a', '[UNK]', 'con', '##t', '##r', '##a'],\n",
              " ['im', '##o', 'and', 'we', 'have', '[UNK]'],\n",
              " ['we', 'have', '[UNK]', 'run', '##n', '##ing'],\n",
              " ['run', '##n', '##ing', '[UNK]', '[UNK]', '.'],\n",
              " ['[UNK]', '[UNK]', '.', '.', '[UNK]', 'fans'],\n",
              " ['.', '[UNK]', 'fans', 'love', '[UNK]', 'i'],\n",
              " ['love', '[UNK]', 'i', 'agree', 'but', 'it'],\n",
              " ['’', 's', 'such', 'a', '[UNK]', 'con'],\n",
              " ['a', '[UNK]', 'con', '##t', '##r', '##a'],\n",
              " ['##c', '##t', '.', '.', '.', '[UNK]'],\n",
              " ['.', '.', '[UNK]', 'm', '##d', '##a'],\n",
              " ['does', 'a', 'good', '[UNK]', 'of', 't'],\n",
              " ['[UNK]', 'of', 't', '##a', '##k', '##ing'],\n",
              " ['##a', '##k', '##ing', 'care', 'of', '[UNK]'],\n",
              " ['care', 'of', '[UNK]', 'w', '##t', '##f'],\n",
              " ['deal', '?', 'this', '[UNK]', 'love', '##s'],\n",
              " ['[UNK]', 'love', '##s', 'his', '[UNK]', '[UNK]'],\n",
              " ['his', '[UNK]', '[UNK]', ':', 'fuck', 'it'],\n",
              " ['it', 'they', 'run', 'the', '[UNK]', 'so'],\n",
              " ['the', '[UNK]', 'so', 'well', 't', '##h'],\n",
              " ['’', 't', 'want', '[UNK]', 'di', '##p'],\n",
              " ['[UNK]', 'di', '##p', '##o', 'gonna', 'be'],\n",
              " ['##o', 'gonna', 'be', 'an', '[UNK]', '[UNK]'],\n",
              " ['an', '[UNK]', '[UNK]', 'in', 'the', '[UNK]'],\n",
              " ['in', 'the', '[UNK]', 'when', 'you', 'feel'],\n",
              " ['m', '##a', '##d', 'about', '[UNK]', 'what'],\n",
              " ['about', '[UNK]', 'what', '’', 's', 'w'],\n",
              " ['##e', '?', 'the', '[UNK]', 'that', '[UNK]'],\n",
              " ['[UNK]', 'that', '[UNK]', '[UNK]', 'w', '##r'],\n",
              " ['[UNK]', 'w', '##r', '##o', '##t', '##e'],\n",
              " ['that', 'the', 'win', '##k', '[UNK]', 'him'],\n",
              " ['##k', '[UNK]', 'him', '5', '##0', '##0'],\n",
              " ['you', 'guys', 'and', 'your', '[UNK]', '[UNK]'],\n",
              " ['your', '[UNK]', '[UNK]', 'lo', '##l', '.'],\n",
              " ['i', 'remember', 'him', 'getting', 'into', '[UNK]'],\n",
              " ['getting', 'into', '[UNK]', '[UNK]', 'in', 'game'],\n",
              " ['[UNK]', 'in', 'game', '1', 'and', 'k'],\n",
              " ['1', 'and', 'k', '##d', '[UNK]', '.'],\n",
              " ['##d', '[UNK]', '.', 'very', 'likely', 'now'],\n",
              " ['##k', '##ed', 'up', '[UNK]', '.', 'it'],\n",
              " ['[UNK]', '.', 'it', '’', 's', 'the'],\n",
              " ['’', 's', 'the', 'right', '[UNK]', 'to'],\n",
              " ['right', '[UNK]', 'to', 'make', '.', '[UNK]'],\n",
              " ['make', '.', '[UNK]', ',', 'i', '[UNK]'],\n",
              " [',', 'i', '[UNK]', '[UNK]', 'does', 'mel'],\n",
              " ['[UNK]', 'does', 'mel', '##o', 'some', '[UNK]'],\n",
              " ['##o', 'some', '[UNK]', '[UNK]', 'and', 'as'],\n",
              " ['[UNK]', 'and', 'as', '##k', '##s', 'him'],\n",
              " ['s', 'will', '##ing', 'to', '[UNK]', 'his'],\n",
              " ['to', '[UNK]', 'his', '[UNK]', 'for', '[UNK]'],\n",
              " ['[UNK]', 'for', '[UNK]', '[UNK]', '.', 'lakers'],\n",
              " ['[UNK]', '.', 'lakers', 'also', '[UNK]', 'a'],\n",
              " ['also', '[UNK]', 'a', 'lot', '##t', '##er'],\n",
              " ['they', 'might', 'get', '[UNK]', 'also', 'mel'],\n",
              " ['[UNK]', 'also', 'mel', '##o', '[UNK]', 'in'],\n",
              " ['##o', '[UNK]', 'in', 'an', 'off', '##s'],\n",
              " ['we', 'sign', '##ed', '[UNK]', 'to', 'a'],\n",
              " ['[UNK]', 'to', 'a', '3', 'year', 'deal'],\n",
              " ['.', 'we', 'have', 'all', 'the', '[UNK]'],\n",
              " ['all', 'the', '[UNK]', 'in', 'this', '[UNK]'],\n",
              " ['in', 'this', '[UNK]', 'lo', '##l', '.'],\n",
              " ['##e', '##n', '##g', 'or', '[UNK]', 'l'],\n",
              " ['or', '[UNK]', 'l', '##m', '##f', '##a'],\n",
              " ['’', 're', 'gonna', '[UNK]', 'them', 'i'],\n",
              " ['[UNK]', 'them', 'i', '[UNK]', 'it', 'was'],\n",
              " ['[UNK]', 'it', 'was', 'at', 'this', 'point'],\n",
              " ['but', 'i', 'am', 'also', '[UNK]', 'he'],\n",
              " ['also', '[UNK]', 'he', '’', 's', 'about'],\n",
              " ['.', 'i', 'mean', 'just', 'go', '[UNK]'],\n",
              " ['just', 'go', '[UNK]', 'their', 'sub', 'l'],\n",
              " ['’', 's', 'just', '[UNK]', 't', '##h'],\n",
              " ['[UNK]', 't', '##h', '##o', 'lo', '##l'],\n",
              " ['##n', '.', '.', '.', '[UNK]', 'lo'],\n",
              " ['.', '[UNK]', 'lo', '##y', '##a', '##l'],\n",
              " ['yeah', '.', '.', 'it', 'was', '[UNK]'],\n",
              " ['it', 'was', '[UNK]', 'and', 'not', 'the'],\n",
              " ['##o', '##m', 'of', '[UNK]', '(', 'r'],\n",
              " ['[UNK]', '(', 'r', '##o', '##d', '##n'],\n",
              " ['##o', '##d', '##n', '##e', '##y', '[UNK]'],\n",
              " ['##e', '##y', '[UNK]', ')', 'that', 'he'],\n",
              " ['”', 'might', 'be', '[UNK]', 'but', 'i'],\n",
              " ['[UNK]', 'but', 'i', '’', 'd', 'r'],\n",
              " ['##e', '##a', '##d', '[UNK]', 'in', '[UNK]'],\n",
              " ['[UNK]', 'in', '[UNK]', '/', 'd', '##e'],\n",
              " ['s', 'ok', ',', 'you', 'guys', '[UNK]'],\n",
              " ['you', 'guys', '[UNK]', 'it', 'r', '/'],\n",
              " ['nba', '’', 's', '[UNK]', 'for', 'that'],\n",
              " ['[UNK]', 'for', 'that', 'trade', ':', '-'],\n",
              " ['trade', ':', '-', '[UNK]', '/', '[UNK]'],\n",
              " ['[UNK]', '/', '[UNK]', '[UNK]', ':', '[UNK]'],\n",
              " ['[UNK]', ':', '[UNK]', 'got', '[UNK]', '-'],\n",
              " ['got', '[UNK]', '-', '[UNK]', '2018', ':'],\n",
              " ['[UNK]', '2018', ':', 'w', '##o', '##w'],\n",
              " ['w', '##o', '##w', ',', '[UNK]', 'won'],\n",
              " [',', '[UNK]', 'won', 'the', 'trade', 'and'],\n",
              " ['the', 'trade', 'and', '[UNK]', 'is', 'a'],\n",
              " ['[UNK]', 'is', 'a', 'la', '##k', '##er'],\n",
              " ['la', '##k', '##er', '-', '[UNK]', '2018'],\n",
              " ['-', '[UNK]', '2018', ':', 'trade', 'is'],\n",
              " ['a', 'la', '##k', '##er', 'in', '[UNK]'],\n",
              " ['##er', 'in', '[UNK]', '?', 's', '##t'],\n",
              " ['i', 'believe', 'that', '[UNK]', 'will', 'do'],\n",
              " ['[UNK]', 'will', 'do', 'mel', '##o', 'a'],\n",
              " ['mel', '##o', 'a', '[UNK]', 'and', 'as'],\n",
              " ['[UNK]', 'and', 'as', '##k', 'him', 'if'],\n",
              " ['go', 'to', 'a', '[UNK]', 'team', 'and'],\n",
              " ['[UNK]', 'team', 'and', 'as', '##k', 'him'],\n",
              " ['s', 'will', '##ing', 'to', '[UNK]', 'his'],\n",
              " ['to', '[UNK]', 'his', '[UNK]', '.', 'he'],\n",
              " ['[UNK]', '.', 'he', '’', 's', '[UNK]'],\n",
              " ['’', 's', '[UNK]', 'so', 'i', 'di'],\n",
              " ['##r', '##e', '##e', '[UNK]', ',', 'i'],\n",
              " ['[UNK]', ',', 'i', 'get', 'way', 'too'],\n",
              " ['get', 'way', 'too', 'much', 'hate', '[UNK]'],\n",
              " ['much', 'hate', '[UNK]', 'in', 'my', 'in'],\n",
              " ['##b', '##o', '##x', 'from', '[UNK]', 'lakers'],\n",
              " ['from', '[UNK]', 'lakers', 'fa', '##n', 'to'],\n",
              " ['fa', '##n', 'to', 'have', '[UNK]', 'about'],\n",
              " ['have', '[UNK]', 'about', 'this', 'yeah', '.'],\n",
              " ['im', '##o', '.', '[UNK]', 'its', 'am'],\n",
              " ['[UNK]', 'its', 'am', 'when', 'it', '[UNK]'],\n",
              " ['when', 'it', '[UNK]', ':', '[UNK]', 'got'],\n",
              " [':', '[UNK]', 'got', '[UNK]', 'all', 'star'],\n",
              " ['[UNK]', 'all', 'star', '[UNK]', ':', '[UNK]'],\n",
              " ['[UNK]', ':', '[UNK]', 'won', 'the', 'trade'],\n",
              " ['won', 'the', 'trade', 'and', '[UNK]', 'is'],\n",
              " ['and', '[UNK]', 'is', 'a', 'la', '##k'],\n",
              " ['a', 'la', '##k', '##er', '[UNK]', '2018'],\n",
              " ['##er', '[UNK]', '2018', '/', 'free', 'a'],\n",
              " ['trade', '-', 'r', '/', 'nba', '[UNK]'],\n",
              " ['/', 'nba', '[UNK]', '.', '[UNK]', 'or'],\n",
              " ['.', '[UNK]', 'or', '[UNK]', 'le', '##bron'],\n",
              " ['[UNK]', 'le', '##bron', 'go', '##es', '.'],\n",
              " ['us', 'to', 'give', 'you', 'an', '[UNK]'],\n",
              " ['you', 'an', '[UNK]', 'con', '##t', '##r'],\n",
              " ['##r', '##o', '##w', 'in', '2', '[UNK]'],\n",
              " ['in', '2', '[UNK]', 'first', '[UNK]', '?'],\n",
              " ['first', '[UNK]', '?', '[UNK]', ',', 'take'],\n",
              " ['[UNK]', ',', 'take', 'your', '[UNK]', 'ass'],\n",
              " ['your', '[UNK]', 'ass', 'to', 'r', '/'],\n",
              " ['to', 'r', '/', 'lakers', 'and', '[UNK]'],\n",
              " ['lakers', 'and', '[UNK]', 'them', 'in', 'a'],\n",
              " ['dr', '##a', '##f', '##t', 'where', '[UNK]'],\n",
              " ['##t', 'where', '[UNK]', 'high', 's', '##c'],\n",
              " ['##t', ')', 'this', 'is', '[UNK]', 'up'],\n",
              " ['is', '[UNK]', 'up', 'la', '##k', '##er'],\n",
              " ['##a', '##s', '##e', '[UNK]', 'at', 'its'],\n",
              " ['[UNK]', 'at', 'its', '[UNK]', 'we', 'could'],\n",
              " ['[UNK]', 'we', 'could', 'also', 'just', 's'],\n",
              " ['##t', '##c', '##h', 'him', 'and', '[UNK]'],\n",
              " ['him', 'and', '[UNK]', 'him', 'on', 'our'],\n",
              " ['##t', '##c', '##h', 'and', '[UNK]', 'him'],\n",
              " ['and', '[UNK]', 'him', '[UNK]', 'which', 'would'],\n",
              " ['[UNK]', 'which', 'would', 'all', '##o', '##w'],\n",
              " ['us', 'to', 'keep', 'our', 'first', '[UNK]'],\n",
              " ['our', 'first', '[UNK]', 'pick', '##s', '.'],\n",
              " ['away', 'from', 'this', '[UNK]', 'fa', '##n'],\n",
              " ['[UNK]', 'fa', '##n', '##b', '##a', '##s'],\n",
              " ['is', 'a', 'fa', 'in', '[UNK]', '.'],\n",
              " ['in', '[UNK]', '.', 'w', '##o', '##w'],\n",
              " ['##t', '##h', '##o', '##m', 'the', '[UNK]'],\n",
              " ['##m', 'the', '[UNK]', 'in', 'this', 'com'],\n",
              " ['this', 'is', 'a', '[UNK]', 'post', 'and'],\n",
              " ['[UNK]', 'post', 'and', 'a', '[UNK]', 'trade'],\n",
              " ['a', '[UNK]', 'trade', '.', 'there', '’'],\n",
              " ['.', 'there', '’', 's', 'no', '[UNK]'],\n",
              " ['s', 'no', '[UNK]', 'to', 'our', 'team'],\n",
              " ['if', 'the', 't', '##a', '##x', '[UNK]'],\n",
              " ['##a', '##x', '[UNK]', 'is', 'really', 'that'],\n",
              " ['problem', ',', 'we', 'can', 'just', '[UNK]'],\n",
              " ['can', 'just', '[UNK]', '&', 'amp', ';'],\n",
              " ['nba', 'is', 'getting', '[UNK]', 'of', 'one'],\n",
              " ['[UNK]', 'of', 'one', 'and', 'done', 'around'],\n",
              " ['reason', 'why', 'the', '[UNK]', 'and', 'many'],\n",
              " ['[UNK]', 'and', 'many', 'other', 'teams', 'will'],\n",
              " ['next', 'season', '.', '[UNK]', 'him', '##s'],\n",
              " ['[UNK]', 'him', '##s', '##e', '##l', '##f'],\n",
              " ['##e', '##l', '##f', '[UNK]', '[UNK]', 'to'],\n",
              " ['[UNK]', '[UNK]', 'to', 't', '##a', '##n'],\n",
              " ['right', 'there', '.', '[UNK]', 'but', 'it'],\n",
              " ['[UNK]', 'but', 'it', 'less', '##e', '##n'],\n",
              " ['less', '##e', '##n', '##s', 'the', '[UNK]'],\n",
              " ['##s', 'the', '[UNK]', '.', 'the', 'team'],\n",
              " ['##l', '##l', '##ed', 'out', 'for', '[UNK]'],\n",
              " ['out', 'for', '[UNK]', 'and', '[UNK]', '.'],\n",
              " ['and', '[UNK]', '.', 'i', 'think', 'they'],\n",
              " ['of', 'a', 'team', 'that', 'just', '[UNK]'],\n",
              " ['that', 'just', '[UNK]', 'out', 'on', 'a'],\n",
              " ['out', 'on', 'a', '[UNK]', 'free', 'a'],\n",
              " ['[UNK]', 'free', 'a', '##g', '##e', '##n'],\n",
              " ['##g', '##e', '##n', '##t', '[UNK]', 'never'],\n",
              " ['##t', '[UNK]', 'never', 'won', 'that', 'trade'],\n",
              " ['won', 'that', 'trade', '(', 'in', '[UNK]'],\n",
              " ['(', 'in', '[UNK]', ')', '.', 'we'],\n",
              " ['k', '##n', '##e', '##w', 'the', '[UNK]'],\n",
              " ['##w', 'the', '[UNK]', 'of', 't', '##r'],\n",
              " ['##a', '##d', '##ing', 'for', '[UNK]', '.'],\n",
              " ['for', '[UNK]', '.', 'teams', 'would', 'make'],\n",
              " ['times', '(', 'i', 'know', 'the', '[UNK]'],\n",
              " ['know', 'the', '[UNK]', 'would', ',', 's'],\n",
              " ['##n', '##d', '##er', 'too', '.', '[UNK]'],\n",
              " ['too', '.', '[UNK]', 'is', 'in', '[UNK]'],\n",
              " ['is', 'in', '[UNK]', 'now', '.', 'just'],\n",
              " [',', 'he', 'was', 'in', '[UNK]', '[UNK]'],\n",
              " ['in', '[UNK]', '[UNK]', 'in', 'game', '1'],\n",
              " ['end', '##ed', 'up', 'with', 'like', '[UNK]'],\n",
              " ['with', 'like', '[UNK]', '+', 'point', '##s'],\n",
              " ['+', 'point', '##s', ',', '[UNK]', '.'],\n",
              " [',', '[UNK]', '.', '[UNK]', '.', 'a'],\n",
              " ['[UNK]', '.', 'a', 'lot', 'of', 'great'],\n",
              " ['of', 'the', 'warriors', 'and', 'the', '[UNK]'],\n",
              " ['and', 'the', '[UNK]', '[UNK]', '/', '7'],\n",
              " ['[UNK]', '/', '7', '##6', '##ers', 'teams'],\n",
              " ['##6', '##ers', 'teams', '.', 'it', '[UNK]'],\n",
              " ['.', 'it', '[UNK]', '.', 'ok', '##c'],\n",
              " ['team', 'at', 'all', '.', 'it', '[UNK]'],\n",
              " ['.', 'it', '[UNK]', 'the', 'team', 'more'],\n",
              " ['.', 'it', '’', 's', '[UNK]', 'that'],\n",
              " ['s', '[UNK]', 'that', 'someone', 'would', 'think'],\n",
              " ['s', 'even', 'more', '[UNK]', 'to', 'believe'],\n",
              " ['[UNK]', 'to', 'believe', 'that', 'this', 'trade'],\n",
              " ['that', 'this', 'trade', 'can', '[UNK]', 'if'],\n",
              " ['can', '[UNK]', 'if', 'some', 'other', 'even'],\n",
              " ['##t', 'doesn', '’', 't', '[UNK]', '.'],\n",
              " ['t', '[UNK]', '.', '[UNK]', 'is', '[UNK]'],\n",
              " ['[UNK]', 'is', '[UNK]', 'different', '##ly', ','],\n",
              " ['should', 'take', 'a', '[UNK]', 'on', 'ka'],\n",
              " ['[UNK]', 'on', 'ka', '##w', '##hi', 'because'],\n",
              " ['(', 'the', 'same', 'way', '[UNK]', 'was'],\n",
              " ['way', '[UNK]', 'was', 'actually', ')', '.'],\n",
              " ['just', 'here', 'to', 'see', 'people', '[UNK]'],\n",
              " ['see', 'people', '[UNK]', 'his', '##to', '##r'],\n",
              " ['a', 'guy', 'on', '[UNK]', '[UNK]', 'con'],\n",
              " ['[UNK]', '[UNK]', 'con', '##t', '##r', '##a'],\n",
              " ['that', '.', 'everyone', '[UNK]', 'at', 'the'],\n",
              " ['[UNK]', 'at', 'the', '[UNK]', 'we', '[UNK]'],\n",
              " ['[UNK]', 'we', '[UNK]', 'to', 'ka', '##n'],\n",
              " ['##m', '##s', 'and', '[UNK]', 'read', 'through'],\n",
              " ['[UNK]', 'read', 'through', 'this', 'your', '##s'],\n",
              " ['/', 'comments', '/', '[UNK]', '/', '[UNK]'],\n",
              " ['[UNK]', '/', '[UNK]', '_', 'ok', '##l'],\n",
              " ['##a', '##h', '##o', '##ma', '_', '[UNK]'],\n",
              " ['##ma', '_', '[UNK]', '_', '[UNK]', '_'],\n",
              " ['_', '[UNK]', '_', '[UNK]', '_', '[UNK]'],\n",
              " ['[UNK]', '_', '[UNK]', '_', 'has', '/'],\n",
              " ['##7', '##m', '/', '[UNK]', '_', '[UNK]'],\n",
              " ['[UNK]', '_', '[UNK]', '_', 'ok', '##c'],\n",
              " ['_', 'with', '_', '[UNK]', '_', 'a'],\n",
              " ['[UNK]', '_', 'a', '##d', '##a', '##m'],\n",
              " ['##d', '##a', '##m', '##s', '/', '[UNK]'],\n",
              " ['##s', '/', '[UNK]', ',', 'it', 'seems'],\n",
              " ['.', '.', '.', 'so', 'true', '[UNK]'],\n",
              " ['so', 'true', '[UNK]', '.', '.', '.'],\n",
              " ['see', 'the', 'lakers', '[UNK]', '?', '&'],\n",
              " ['[UNK]', '?', '&', 'gt', ';', '[UNK]'],\n",
              " ['gt', ';', '[UNK]', 'in', 'every', 'game'],\n",
              " ['you', 'played', 'go', '[UNK]', 'the', 'm'],\n",
              " ['[UNK]', 'the', 'm', '##a', '##vs', '[UNK]'],\n",
              " ['##a', '##vs', '[UNK]', 't', '##a', '##n'],\n",
              " ['like', '6', '##0', 'games', '[UNK]', '5'],\n",
              " ['games', '[UNK]', '5', 'point', '##s', 'and'],\n",
              " ['point', '##s', 'and', '[UNK]', '/', 'm'],\n",
              " ['[UNK]', '/', 'm', '##a', '##vs', 'will'],\n",
              " ['##a', '##n', '##k', '##ed', 'a', '[UNK]'],\n",
              " ['##ed', 'a', '[UNK]', 'of', 'those', 'with'],\n",
              " ['of', 'those', 'with', 'bad', '[UNK]', 'in'],\n",
              " ['bad', '[UNK]', 'in', 'the', 'last', '[UNK]'],\n",
              " ['the', 'last', '[UNK]', 'of', '[UNK]', 'of'],\n",
              " ['of', '[UNK]', 'of', 'the', 'game', '.'],\n",
              " ['##d', '##d', '##ed', '[UNK]', 'and', 'likely'],\n",
              " ['[UNK]', 'and', 'likely', 'd', '##e', '##a'],\n",
              " ['##n', '##dr', '##e', '+', '[UNK]', '+'],\n",
              " ['+', '[UNK]', '+', 'd', '##s', '##j'],\n",
              " ['d', '##s', '##j', '##r', '+', '[UNK]'],\n",
              " ['##r', '+', '[UNK]', '+', '[UNK]', 'are'],\n",
              " ['+', '[UNK]', 'are', 'a', 'much', 'better'],\n",
              " ['team', '.', 'i', '[UNK]', 'in', 'new'],\n",
              " ['[UNK]', 'in', 'new', '[UNK]', '[UNK]', 't'],\n",
              " ['[UNK]', '[UNK]', 't', '##h', '##o', 'you'],\n",
              " ['are', 'one', 'of', 'the', '[UNK]', ','],\n",
              " ['the', '[UNK]', ',', 'im', '##o', '.'],\n",
              " ['m', 'not', 'trying', 'to', '[UNK]', 'an'],\n",
              " ['to', '[UNK]', 'an', '[UNK]', 'fa', '##n'],\n",
              " ['[UNK]', 'fa', '##n', '##b', '##a', '##s'],\n",
              " ['##b', '##a', '##s', '##e', 'but', '[UNK]'],\n",
              " ['##e', 'but', '[UNK]', 'fuck', 'some', 'of'],\n",
              " ['you', 'guys', 'are', '[UNK]', '.', 'we'],\n",
              " ['[UNK]', '.', 'we', 'wouldn', '’', 't'],\n",
              " ['wouldn', '’', 't', 'have', 'been', '[UNK]'],\n",
              " ['have', 'been', '[UNK]', '.', 'we', 'k'],\n",
              " ['##s', 'fans', 'in', '[UNK]', '.', 'warriors'],\n",
              " ['[UNK]', '.', 'warriors', 'fans', 'and', 'r'],\n",
              " ['fans', 'in', '2018', '.', 'now', '[UNK]'],\n",
              " ['.', 'now', '[UNK]', 'will', 'be', 'a'],\n",
              " ['will', 'be', 'a', '[UNK]', 'of', 'the'],\n",
              " ['[UNK]', 'of', 'the', 'warriors', ',', 'r'],\n",
              " ['and', 'lakers', 'fans', '.', '[UNK]', '.'],\n",
              " ['.', '[UNK]', '.', '.', '.', 'you'],\n",
              " ['.', '.', 'you', 'know', 'they', '[UNK]'],\n",
              " ['know', 'they', '[UNK]', 'him', 'right', '?'],\n",
              " ['s', 'not', 'just', '[UNK]', '.', 'no'],\n",
              " ['[UNK]', '.', 'no', '[UNK]', '[UNK]', '!'],\n",
              " ['[UNK]', '[UNK]', '!', 'he', 'gets', '“'],\n",
              " ['he', 'gets', '“', '[UNK]', '”', 'because'],\n",
              " ['[UNK]', '”', 'because', 'he', '[UNK]', 'the'],\n",
              " ['he', '[UNK]', 'the', 'warriors', 'and', 'not'],\n",
              " ['i', '’', 'm', 'gonna', '[UNK]', 'in'],\n",
              " ['gonna', '[UNK]', 'in', 'this', 'com', '##m'],\n",
              " ['##s', '##t', 'was', '[UNK]', 'that', '[UNK]'],\n",
              " ['[UNK]', 'that', '[UNK]', 'too', '##k', 'a'],\n",
              " ['3', '+', '1', '[UNK]', 'a', '1'],\n",
              " ['[UNK]', 'a', '1', '+', '1', '.'],\n",
              " ['+', '1', '.', '[UNK]', 'left', 'a'],\n",
              " ['[UNK]', 'left', 'a', 'lot', 'of', 'money'],\n",
              " ['it', '’', 's', '[UNK]', 'that', 'he'],\n",
              " ['[UNK]', 'that', 'he', 'too', '##k', 'this'],\n",
              " ['.', 'it', '’', 's', 'also', '[UNK]'],\n",
              " ['s', 'also', '[UNK]', 'that', 'he', 'think'],\n",
              " ['##k', 'a', '4', 'year', 'deal', '[UNK]'],\n",
              " ['year', 'deal', '[UNK]', 'a', '1', '+'],\n",
              " ['##c', '##k', '##ed', 'or', '[UNK]', 'at'],\n",
              " ['or', '[UNK]', 'at', '[UNK]', '[UNK]', 'to'],\n",
              " ['[UNK]', '[UNK]', 'to', 'ok', '##c', '.'],\n",
              " ['deal', '.', 'as', 'a', '[UNK]', 'fa'],\n",
              " ['a', '[UNK]', 'fa', '##n', ',', 'i'],\n",
              " ['##d', '##k', 'why', 't', '##f', '[UNK]'],\n",
              " ['t', '##f', '[UNK]', 'too', '##k', 'a'],\n",
              " ['s', '##e', '##n', '##s', '##e', '[UNK]'],\n",
              " ['##s', '##e', '[UNK]', 'but', 'i', 'guess'],\n",
              " ['long', 't', '##er', '##m', '[UNK]', 'after'],\n",
              " ['##m', '[UNK]', 'after', 'that', '[UNK]', '[UNK]'],\n",
              " ['that', '[UNK]', '[UNK]', 'a', '[UNK]', 'of'],\n",
              " ['a', '[UNK]', 'of', 'years', 'a', '##g'],\n",
              " ['+', '1', '.', 'go', '[UNK]', 'our'],\n",
              " ['go', '[UNK]', 'our', 'sub', '##r', '##ed'],\n",
              " ['##dit', 'in', 'the', '[UNK]', 'we', '##e'],\n",
              " ['[UNK]', 'we', '##e', '##k', '.', '[UNK]'],\n",
              " ['##k', '.', '[UNK]', 'of', 'yes', '##t'],\n",
              " ['##er', '##d', '##a', '##y', 'and', '[UNK]'],\n",
              " ['##y', 'and', '[UNK]', '(', 'his', '[UNK]'],\n",
              " ['(', 'his', '[UNK]', ')', '[UNK]', 'news'],\n",
              " [')', '[UNK]', 'news', ',', 'we', 'thought'],\n",
              " [',', 'we', 'thought', '[UNK]', 'was', 'go'],\n",
              " ['[UNK]', 'was', 'go', '##n', '##e', '.'],\n",
              " ['##n', '##e', '.', 'we', 'already', '[UNK]'],\n",
              " ['we', 'already', '[UNK]', 'to', 't', '##er'],\n",
              " ['it', '.', 'i', 'love', '[UNK]', 'and'],\n",
              " ['love', '[UNK]', 'and', '[UNK]', '[UNK]', 'should'],\n",
              " ['[UNK]', '[UNK]', 'should', 'be', 'an', '[UNK]'],\n",
              " ['be', 'an', '[UNK]', 'to', 'this', '[UNK]'],\n",
              " ['to', 'this', '[UNK]', 'thing', '.', 'i'],\n",
              " ['mean', 'he', '’', 's', '[UNK]', '('],\n",
              " ['s', '[UNK]', '(', '[UNK]', 'for', 'the'],\n",
              " ['[UNK]', 'for', 'the', 'd', '##e', '##n'],\n",
              " ['real', ',', 'i', 'hate', '[UNK]', '_'],\n",
              " ['hate', '[UNK]', '_', 'ball', '##s', '&'],\n",
              " ['gt', ';', 'most', '[UNK]', 'fans', 'hate'],\n",
              " ['[UNK]', 'fans', 'hate', '[UNK]', 'more', 'than'],\n",
              " ['[UNK]', 'more', 'than', '[UNK]', 'is', 'this'],\n",
              " ['[UNK]', 'is', 'this', '[UNK]', 'or', 'no'],\n",
              " ['[UNK]', 'or', 'no', 'yeah', 'man', ','],\n",
              " ['.', 'i', 'hate', '[UNK]', '_', 'ball'],\n",
              " ['[UNK]', '_', 'ball', 'because', 'the', 'con'],\n",
              " ['##c', '##t', 'isn', '’', 't', '[UNK]'],\n",
              " ['’', 't', '[UNK]', 'to', 'be', '[UNK]'],\n",
              " ['to', 'be', '[UNK]', 'in', 'its', 'first'],\n",
              " ['can', '’', 't', '[UNK]', 'the', 'last'],\n",
              " ['[UNK]', 'the', 'last', '4', 'years', 'of'],\n",
              " ['i', 'am', 'not', 'going', 'to', '[UNK]'],\n",
              " ['going', 'to', '[UNK]', 'no', '##r', 'd'],\n",
              " ['this', 'because', 'that', 'would', 'be', '[UNK]'],\n",
              " ['would', 'be', '[UNK]', '[UNK]', 'the', 'm'],\n",
              " ['[UNK]', 'the', 'm', '##o', '##d', '##s'],\n",
              " ['fuck', 'until', 'you', '[UNK]', 'something', 'to'],\n",
              " ['[UNK]', 'something', 'to', 'their', '[UNK]', 'which'],\n",
              " ['their', '[UNK]', 'which', 'is', '[UNK]', 'because'],\n",
              " ['is', '[UNK]', 'because', 'this', 'sub', 'has'],\n",
              " ['it', 'has', 'probably', '[UNK]', '-', '[UNK]'],\n",
              " ['[UNK]', '-', '[UNK]', 'people', 'on', 'at'],\n",
              " ['a', 'time', 'because', 'of', '[UNK]', 'season'],\n",
              " ['of', '[UNK]', 'season', '.', 'they', 'aren'],\n",
              " ['are', '.', 'i', '[UNK]', 'any', 'of'],\n",
              " ['[UNK]', 'any', 'of', 'them', 'know', '##ing'],\n",
              " ['an', '##s', '##w', '##er', 'that', '[UNK]'],\n",
              " ['##er', 'that', '[UNK]', '?', 'm', '##a'],\n",
              " ['##vs', 'also', 'i', 'am', 'just', '[UNK]'],\n",
              " ['am', 'just', '[UNK]', 'as', 'you', 'are'],\n",
              " ['but', 'i', 'think', 'him', 'and', '[UNK]'],\n",
              " ['him', 'and', '[UNK]', 'would', 'like', 'to'],\n",
              " ['still', 'a', 'good', '[UNK]', ',', 'im'],\n",
              " ['[UNK]', ',', 'im', '##o', '.', 'he'],\n",
              " ['##o', '.', 'he', 'has', 'his', '[UNK]'],\n",
              " ['has', 'his', '[UNK]', 'like', 'every', '[UNK]'],\n",
              " ['like', 'every', '[UNK]', 'in', 'the', 'nba'],\n",
              " ['##m', '##es', 'had', 'us', '[UNK]', 'the'],\n",
              " ['us', '[UNK]', 'the', 'top', '[UNK]', 'before'],\n",
              " ['top', '[UNK]', 'before', 'r', '##o', '##b'],\n",
              " ['it', '’', 's', 'a', '[UNK]', 'if'],\n",
              " ['a', '[UNK]', 'if', 'anything', ',', 'it'],\n",
              " ['anything', ',', 'it', 'was', '[UNK]', 'by'],\n",
              " ['was', '[UNK]', 'by', '[UNK]', '[UNK]', 'com'],\n",
              " ['[UNK]', '[UNK]', 'com', '##m', '##e', '##n'],\n",
              " ['##r', '##n', '##a', '##m', '##e', '[UNK]'],\n",
              " ['##m', '##e', '[UNK]', 'out', 'he', 'said'],\n",
              " ['t', 'like', 'the', '[UNK]', 'or', '[UNK]'],\n",
              " ['[UNK]', 'or', '[UNK]', '.', 'it', 'was'],\n",
              " ['just', 'him', 'and', '[UNK]', '.', 'he'],\n",
              " ['[UNK]', '.', 'he', 'could', '##n', '’'],\n",
              " ['t', 'win', 'with', 'those', '[UNK]', '”'],\n",
              " ['those', '[UNK]', '”', 'yeah', 'and', 'he'],\n",
              " ['was', 'just', 'him', 'and', '[UNK]', 'on'],\n",
              " ['and', '[UNK]', 'on', 'that', 'team', '.'],\n",
              " ['he', 'made', 'his', '[UNK]', 'yet', 'lo'],\n",
              " ['[UNK]', 'yet', 'lo', '##l', 'yeah', 'but'],\n",
              " ['all', 'good', 'and', 'it', '[UNK]', 'that'],\n",
              " ['it', '[UNK]', 'that', 'you', 'got', '[UNK]'],\n",
              " ['you', 'got', '[UNK]', 'for', 'a', '[UNK]'],\n",
              " ['for', 'a', '[UNK]', 'question', 'le', '##bron'],\n",
              " ['question', 'le', '##bron', 'going', 'to', '[UNK]'],\n",
              " ['going', 'to', '[UNK]', 'would', 'help', 'the'],\n",
              " ['ka', '##w', '##hi', '[UNK]', '.', '[UNK]'],\n",
              " ['[UNK]', '.', '[UNK]', 'and', '[UNK]', 'would'],\n",
              " ['and', '[UNK]', 'would', 'have', 'a', '[UNK]'],\n",
              " ['have', 'a', '[UNK]', 'war', '.', 'lo'],\n",
              " ['the', 'w', '##c', '##f', '[UNK]', 'lo'],\n",
              " ['##f', '[UNK]', 'lo', '##s', '##ing', 'in'],\n",
              " ['ll', 'be', 'a', '[UNK]', 'agree', '##m'],\n",
              " ['[UNK]', 'agree', '##m', '##e', '##n', '##t'],\n",
              " ['##e', '##n', '##t', 'that', 'he', '[UNK]'],\n",
              " ['that', 'he', '[UNK]', 'on', 'defense', 'and'],\n",
              " ['on', 'defense', 'and', 'just', '[UNK]', 'lo'],\n",
              " ['just', '[UNK]', 'lo', '##b', '##s', '.'],\n",
              " ['##s', 'a', 'lot', 'of', '[UNK]', 'so'],\n",
              " ['of', '[UNK]', 'so', 'we', 'may', 'even'],\n",
              " ['we', 'may', 'even', 'see', '[UNK]', 'get'],\n",
              " ['see', '[UNK]', 'get', 'some', 'post', '-'],\n",
              " ['some', 'post', '-', 'up', '[UNK]', 'where'],\n",
              " ['up', '[UNK]', 'where', 'he', 'can', '[UNK]'],\n",
              " ['he', 'can', '[UNK]', 'out', 'of', 'the'],\n",
              " ['##bron', ',', 'be', '##n', ',', '[UNK]'],\n",
              " ['##n', ',', '[UNK]', 'not', 'until', 'the'],\n",
              " ['not', 'until', 'the', '[UNK]', '[UNK]', 'we'],\n",
              " ['[UNK]', '[UNK]', 'we', 'all', 'are', '[UNK]'],\n",
              " ['all', 'are', '[UNK]', 'on', 'this', '[UNK]'],\n",
              " ['on', 'this', '[UNK]', 'day', 'or', '[UNK]'],\n",
              " ['day', 'or', '[UNK]', '’', 's', '[UNK]'],\n",
              " ['’', 's', '[UNK]', 'they', '’', 're'],\n",
              " ['they', 'have', 'a', '[UNK]', '[UNK]', '%'],\n",
              " ['[UNK]', '[UNK]', '%', 'too', '.', 'i'],\n",
              " ['##d', '##er', 'if', '[UNK]', '[UNK]', 'says'],\n",
              " ['[UNK]', '[UNK]', 'says', 'to', 'the', '7'],\n",
              " ['”', 'there', 'was', 'some', '[UNK]', 'that'],\n",
              " ['some', '[UNK]', 'that', '[UNK]', 'to', '##l'],\n",
              " ['[UNK]', 'to', '##l', '##d', 'ok', '##c'],\n",
              " ['##ing', 'in', 'ok', '##c', 'in', '[UNK]'],\n",
              " ['##c', 'in', '[UNK]', 'they', '’', 've'],\n",
              " ['been', 'playing', 'like', 'this', '(', '[UNK]'],\n",
              " ['this', '(', '[UNK]', '[UNK]', '%', ')'],\n",
              " ['[UNK]', '%', ')', 'for', 'a', 'while'],\n",
              " ['##r', '##o', '##w', '[UNK]', 'in', 'for'],\n",
              " ['[UNK]', 'in', 'for', 'the', 'spurs', 'to'],\n",
              " ['le', '##bron', 'would', 'want', '[UNK]', 'go'],\n",
              " ['want', '[UNK]', 'go', '##n', '##e', 'any'],\n",
              " ['##w', '##a', '##y', '##s', '[UNK]', '1'],\n",
              " ['##s', '[UNK]', '1', 'pick', 'and', '[UNK]'],\n",
              " ['pick', 'and', '[UNK]', 'en', '##g', '##e'],\n",
              " ['##n', '##d', 'can', '[UNK]', 'his', 'shot'],\n",
              " ['[UNK]', 'his', 'shot', '.', 'he', 'might'],\n",
              " ['not', 'be', 'read', '##y', 'to', '[UNK]'],\n",
              " ['##y', 'to', '[UNK]', 'to', 'a', '[UNK]'],\n",
              " ['to', 'a', '[UNK]', '[UNK]', 'team', 'just'],\n",
              " ['[UNK]', 'team', 'just', 'yet', '.', 'the'],\n",
              " ['yet', '.', 'the', 'same', 'way', '[UNK]'],\n",
              " ['same', 'way', '[UNK]', 'wasn', '’', 't'],\n",
              " ['it', 'can', 'you', '[UNK]', 'on', 'this'],\n",
              " ['[UNK]', 'on', 'this', '?', 'don', '’'],\n",
              " ['a', 'star', '.', '[UNK]', 'didn', \"'\"],\n",
              " ['[UNK]', 'didn', \"'\", 't', 'know', 'lo'],\n",
              " ['the', '2', '##n', '##d', '[UNK]', '/'],\n",
              " ['##d', '[UNK]', '/', 'w', '##c', '##f'],\n",
              " ['w', '##c', '##f', '[UNK]', 'lo', '##s'],\n",
              " ['[UNK]', 'lo', '##s', '##e', 'in', 'the'],\n",
              " ['finals', 'or', 'i', '[UNK]', 'the', '\"'],\n",
              " ['[UNK]', 'the', '\"', 't', '\"', 'ok'],\n",
              " ['##a', '##o', 'before', '[UNK]', '[UNK]', '5'],\n",
              " ['[UNK]', '[UNK]', '5', ':', '[UNK]', '[UNK]'],\n",
              " [':', '[UNK]', '[UNK]', '.', '.', '.'],\n",
              " ['.', '.', '.', 'they', 'got', '[UNK]'],\n",
              " ['they', 'got', '[UNK]', 'of', 'him', '?'],\n",
              " ['##ing', 'lakers', 'for', '[UNK]', 'like', 'every'],\n",
              " ['[UNK]', 'like', 'every', 'other', 'star', 'fa'],\n",
              " ['other', 'star', 'fa', '.', '[UNK]', '[UNK]'],\n",
              " ['.', '[UNK]', '[UNK]', 'is', 'probably', 'in'],\n",
              " ['##t', '##ing', '##s', 'tell', '##ing', '[UNK]'],\n",
              " ['tell', '##ing', '[UNK]', 'to', 'get', 'ka'],\n",
              " ['##s', '##e', 'does', 'a', 'good', '[UNK]'],\n",
              " ['a', 'good', '[UNK]', 'of', 'making', '[UNK]'],\n",
              " ['of', 'making', '[UNK]', 'look', 'bad', 'already'],\n",
              " ['##b', '##l', '##e', '[UNK]', '[UNK]', 'to'],\n",
              " ['[UNK]', '[UNK]', 'to', 'the', '[UNK]', '[UNK]'],\n",
              " ['the', '[UNK]', '[UNK]', ',', 'why', 'us'],\n",
              " ['lot', 'of', 'his', '[UNK]', '.', 'he'],\n",
              " ['[UNK]', '.', 'he', 'was', 'much', 'more'],\n",
              " ['was', 'much', 'more', '[UNK]', '[UNK]', '-'],\n",
              " ['[UNK]', '[UNK]', '-', '[UNK]', '.', 'he'],\n",
              " ['[UNK]', '.', 'he', '’', 's', 'not'],\n",
              " ['is', '.', 'this', 'was', 'a', '[UNK]'],\n",
              " ['was', 'a', '[UNK]', 'com', '##m', '##e'],\n",
              " ['&', 'gt', ';', '[UNK]', 'at', 'the'],\n",
              " ['[UNK]', 'at', 'the', 'time', 'probably', 'because'],\n",
              " ['don', '’', 't', 'believe', 'these', '[UNK]'],\n",
              " ['believe', 'these', '[UNK]', '.', 'do', 'you'],\n",
              " ['this', 'sub', 'and', 'the', '[UNK]', '[UNK]'],\n",
              " ['the', '[UNK]', '[UNK]', '.', 'con', '##g'],\n",
              " ['##s', '.', '.', '.', '?', '[UNK]'],\n",
              " ['.', '?', '[UNK]', '.', 'is', 'the'],\n",
              " ['.', 'is', 'the', '[UNK]', 'in', '?'],\n",
              " ['[UNK]', 'in', '?', 'they', '[UNK]', 'don'],\n",
              " ['they', '[UNK]', 'don', '’', 't', '.'],\n",
              " ['dr', '##a', '##w', '[UNK]', 'or', 'warriors'],\n",
              " ['[UNK]', 'or', 'warriors', 'in', 'the', '2'],\n",
              " ['in', 'the', '2', '##n', '##d', '[UNK]'],\n",
              " ['##n', '##d', '[UNK]', '.', '.', '.'],\n",
              " ['##d', 'to', 'you', 'guys', '[UNK]', 'at'],\n",
              " ['guys', '[UNK]', 'at', 'other', 'teams', '[UNK]'],\n",
              " ['other', 'teams', '[UNK]', 'players', 'that', 'aren'],\n",
              " ['on', 'the', 'trade', '[UNK]', 'it', '’'],\n",
              " ['[UNK]', 'it', '’', 'll', 'be', 'lakers'],\n",
              " ['probably', '7', '##6', '##ers', '-', '[UNK]'],\n",
              " ['##ers', '-', '[UNK]', '[UNK]', '-', 'r'],\n",
              " ['[UNK]', '-', 'r', '##o', '##c', '##k'],\n",
              " ['##e', '##t', '##s', '[UNK]', '-', 'someone'],\n",
              " ['[UNK]', '-', 'someone', 'else', '[UNK]', ','],\n",
              " ['else', '[UNK]', ',', 'no', 'you', 'have'],\n",
              " ['##t', 'la', '##k', '##er', '.', '[UNK]'],\n",
              " ['##er', '.', '[UNK]', 'or', '[UNK]', '[UNK]'],\n",
              " ['or', '[UNK]', '[UNK]', '[UNK]', '?', 'lo'],\n",
              " ['[UNK]', '?', 'lo', '##n', '##z', '##o'],\n",
              " ['as', 'far', 'as', '[UNK]', '.', '[UNK]'],\n",
              " ['[UNK]', '.', '[UNK]', 'is', 'a', '[UNK]'],\n",
              " ['is', 'a', '[UNK]', '4', 'and', 'i'],\n",
              " ['is', 'trying', 'to', '[UNK]', 'to', 'the'],\n",
              " ['[UNK]', 'to', 'the', 's', '##ma', '##l'],\n",
              " ['##n', '##o', '##c', '##k', 'on', '[UNK]'],\n",
              " ['##k', 'on', '[UNK]', '.', 'it', '’'],\n",
              " ['s', 'just', 'about', '[UNK]', '.', 'maybe'],\n",
              " ['[UNK]', '.', 'maybe', 'a', '3', 'team'],\n",
              " ['##e', '##n', '##d', '##s', '[UNK]', 'else'],\n",
              " ['##s', '[UNK]', 'else', '##w', '##h', '##er'],\n",
              " ['could', 'he', 'be', 'a', '[UNK]', '?'],\n",
              " ['a', '[UNK]', '?', 'or', 'is', 'it'],\n",
              " ['##e', '##a', '##t', '[UNK]', 'under', 'any'],\n",
              " ['[UNK]', 'under', 'any', '[UNK]', '?', 'i'],\n",
              " ['[UNK]', '?', 'i', '[UNK]', 'k', '##d'],\n",
              " ['[UNK]', 'k', '##d', '[UNK]', '[UNK]', 'this'],\n",
              " ['[UNK]', '[UNK]', 'this', 'year', 'if', 'that'],\n",
              " ['##o', '##w', 'on', 'his', '[UNK]', 'and'],\n",
              " ['his', '[UNK]', 'and', 'he', 'might', 'not'],\n",
              " ['he', 'might', 'not', '[UNK]', 'games', 'over'],\n",
              " ['[UNK]', 'games', 'over', '[UNK]', '.', 'we'],\n",
              " ['[UNK]', '.', 'we', '’', 'd', 'like'],\n",
              " ['##r', '##o', '##w', 'out', 'a', '[UNK]'],\n",
              " ['out', 'a', '[UNK]', '-', 'r', '##o'],\n",
              " ['##b', '##es', '-', '[UNK]', '-', '[UNK]'],\n",
              " ['[UNK]', '-', '[UNK]', '-', 'a', '##d'],\n",
              " ['##a', '##m', '##s', '[UNK]', 'to', '[UNK]'],\n",
              " ['[UNK]', 'to', '[UNK]', 'the', 'game', 'out'],\n",
              " ['the', 'reason', '##s', 'you', '[UNK]', '.'],\n",
              " ['you', '[UNK]', '.', '.', '.', 'i'],\n",
              " ['##ed', 'with', 'the', '[UNK]', 'but', 'it'],\n",
              " ['[UNK]', 'but', 'it', 'doesn', '’', 't'],\n",
              " ['want', '##ed', 'to', 'play', 'with', '[UNK]'],\n",
              " ['play', 'with', '[UNK]', 'i', '[UNK]', 'with'],\n",
              " ['i', '[UNK]', 'with', 't', '##h', '##o'],\n",
              " ['it', '’', 's', '[UNK]', 'dr', '##a'],\n",
              " ['[UNK]', 'dr', '##a', '##y', 'i', 'hate'],\n",
              " ['about', 'a', 'fa', '##k', '##e', '[UNK]'],\n",
              " ['##k', '##e', '[UNK]', 'hard', '##a', '##w'],\n",
              " ['hard', '##a', '##w', '##a', '##y', '[UNK]'],\n",
              " ['##a', '##y', '[UNK]', 'not', 'about', 'le'],\n",
              " ['.', 'it', '’', 's', '[UNK]', 'what'],\n",
              " ['s', '[UNK]', 'what', 'people', 'would', 'do'],\n",
              " ['##r', '##ma', 'on', 'this', 'sub', '[UNK]'],\n",
              " ['this', 'sub', '[UNK]', ',', '[UNK]', ','],\n",
              " [',', '[UNK]', ',', 'r', '##o', '##b'],\n",
              " ['##d', '##a', '##m', '##s', ',', '[UNK]'],\n",
              " ['##s', ',', '[UNK]', 'and', '[UNK]', 'not'],\n",
              " ['and', '[UNK]', 'not', 'sure', 'if', 's'],\n",
              " ['but', 'le', '##bron', 'does', 'need', '[UNK]'],\n",
              " ['does', 'need', '[UNK]', 's', '##p', '##a'],\n",
              " ['will', 'be', 'the', 'first', 'time', '[UNK]'],\n",
              " ['first', 'time', '[UNK]', 'has', 'the', 'same'],\n",
              " ['has', 'the', 'same', '[UNK]', '2', '[UNK]'],\n",
              " ['[UNK]', '2', '[UNK]', 'years', '.', '[UNK]'],\n",
              " ['years', '.', '[UNK]', 'to', 'see', 'what'],\n",
              " ['.', '&', 'gt', ';', '[UNK]', 'red'],\n",
              " [';', '[UNK]', 'red', '##dit', '[UNK]', '[UNK]'],\n",
              " ['##dit', '[UNK]', '[UNK]', 'are', 'you', 'even'],\n",
              " ['##s', '##e', 'or', '[UNK]', 'you', 'know'],\n",
              " ['[UNK]', 'you', 'know', 'how', 'to', 'use'],\n",
              " ['how', 'to', 'use', '[UNK]', 'but', 'not'],\n",
              " ['[UNK]', 'but', 'not', 'how', 'to', 'do'],\n",
              " ['##h', '?', 'w', '##t', '##f', '[UNK]'],\n",
              " ['##t', '##f', '[UNK]', 'are', 'my', 's'],\n",
              " ['.', 'i', 'can', 'see', 'the', '[UNK]'],\n",
              " ['see', 'the', '[UNK]', 'getting', 'there', 'as'],\n",
              " ['as', 'well', '.', 'he', '[UNK]', 'up'],\n",
              " ['he', '[UNK]', 'up', 'every', '[UNK]', 'we'],\n",
              " ['every', '[UNK]', 'we', 'had', 'that', '[UNK]'],\n",
              " ['had', 'that', '[UNK]', 'and', 'when', 'your'],\n",
              " ['mel', '##o', 'shot', '6', '-', '[UNK]'],\n",
              " ['6', '-', '[UNK]', 'on', '[UNK]', '3'],\n",
              " ['on', '[UNK]', '3', '##s', ')', 'then'],\n",
              " ['is', 'this', 'a', '[UNK]', 'and', 'on'],\n",
              " ['[UNK]', 'and', 'on', 'the', 'other', '[UNK]'],\n",
              " ['the', 'other', '[UNK]', ',', '[UNK]', 'would'],\n",
              " [',', '[UNK]', 'would', 'probably', 'make', 'it'],\n",
              " ['probably', 'make', 'it', 'every', 'year', '[UNK]'],\n",
              " ['every', 'year', '[UNK]', 'for', 'a', '##d'],\n",
              " ['for', 'a', '##d', '’', 's', '[UNK]'],\n",
              " ['’', 's', '[UNK]', 'year', 'maybe', 'on'],\n",
              " ['##c', '##e', 'if', '[UNK]', 'gets', '[UNK]'],\n",
              " ['[UNK]', 'gets', '[UNK]', '.', 'it', '’'],\n",
              " ['.', 'it', '’', 's', '[UNK]', 'to'],\n",
              " ['s', '[UNK]', 'to', 'tell', '.', 'they'],\n",
              " ['tell', '.', 'they', 'might', 'play', '[UNK]'],\n",
              " ['might', 'play', '[UNK]', 'and', '[UNK]', 'in'],\n",
              " ['and', '[UNK]', 'in', 'the', 'same', 'playoffs'],\n",
              " ['off', 'a', 'long', '[UNK]', '[UNK]', '[UNK]'],\n",
              " ['[UNK]', '[UNK]', '[UNK]', 'and', 'win', '[UNK]'],\n",
              " ['and', 'win', '[UNK]', '[UNK]', '.', 'it'],\n",
              " ['[UNK]', '.', 'it', '’', 's', 'not'],\n",
              " ['’', 's', 'not', '[UNK]', '[UNK]', '.'],\n",
              " ['[UNK]', '[UNK]', '.', 'it', '’', 's'],\n",
              " ['it', '’', 's', '[UNK]', 'the', '7'],\n",
              " ['[UNK]', 'the', '7', '##6', '##ers', '/'],\n",
              " ['##6', '##ers', '/', '[UNK]', '.', 'they'],\n",
              " ['[UNK]', '.', 'they', '’', 're', 'talking'],\n",
              " ['you', 'can', 'tell', 'how', '[UNK]', 'the'],\n",
              " ['how', '[UNK]', 'the', '[UNK]', 'to', '[UNK]'],\n",
              " ['[UNK]', 'to', '[UNK]', '[UNK]', '[UNK]', 'is'],\n",
              " ['[UNK]', '[UNK]', 'is', 'yeah', ',', 'it'],\n",
              " ['t', 'make', 'a', '[UNK]', 'post', 'about'],\n",
              " ['[UNK]', 'post', 'about', 'it', '3', 'teams'],\n",
              " ['lakers', 'also', 'go', 'on', 'the', '[UNK]'],\n",
              " ['on', 'the', '[UNK]', '[UNK]', '(', 'or'],\n",
              " ['[UNK]', '(', 'or', 'is', 'it', 'the'],\n",
              " ['is', 'it', 'the', '[UNK]', '?', ')'],\n",
              " ['[UNK]', '?', ')', 'r', '##o', '##a'],\n",
              " ['r', '##o', '##a', '##d', '[UNK]', 're'],\n",
              " ['##d', '[UNK]', 're', '##p', '##o', '##r'],\n",
              " ['##p', '##h', '##e', '##n', 'a', '[UNK]'],\n",
              " ['##n', 'a', '[UNK]', 'so', '.', '.'],\n",
              " ['.', 't', '##b', '##h', '[UNK]', 'will'],\n",
              " ['##h', '[UNK]', 'will', 'likely', 'have', 'the'],\n",
              " ['want', 'that', 'trade', ',', '[UNK]', '[UNK]'],\n",
              " [',', '[UNK]', '[UNK]', '.', 'after', 'that'],\n",
              " ['##ers', '##hi', '##p', '[UNK]', 'in', 'more'],\n",
              " ['[UNK]', 'in', 'more', 'guys', 'that', 'were'],\n",
              " ['been', 'there', ')', '.', '[UNK]', 'guys'],\n",
              " ['.', '[UNK]', 'guys', 'like', '[UNK]', 'and'],\n",
              " ['like', '[UNK]', 'and', '[UNK]', 'won', '’'],\n",
              " ['[UNK]', 'won', '’', 't', 'let', 'him'],\n",
              " ['&', 'gt', ';', '[UNK]', 'from', 'for'],\n",
              " ['[UNK]', 'from', 'for', '##m', '##er', 'team'],\n",
              " ['&', 'gt', ';', 'a', '##d', '[UNK]'],\n",
              " ['a', '##d', '[UNK]', 'at', '##t', '##a'],\n",
              " ['/', 'nba', '&', 'gt', ';', '[UNK]'],\n",
              " ['gt', ';', '[UNK]', 'from', 'for', '##m'],\n",
              " ['gt', ';', 'a', '##d', '[UNK]', 'at'],\n",
              " ['##d', '[UNK]', 'at', '##t', '##a', '##c'],\n",
              " ['##k', 'of', 'a', '[UNK]', 'that', 'was'],\n",
              " ['[UNK]', 'that', 'was', 'la', '##t', '##er'],\n",
              " ['la', '##t', '##er', '[UNK]', 'between', 'the'],\n",
              " ['[UNK]', 'between', 'the', 'two', '&', 'gt'],\n",
              " ['##w', '##e', '##e', '##t', '[UNK]', '.'],\n",
              " ['##t', '[UNK]', '.', 'i', 'know', 'this'],\n",
              " ['i', 'know', 'this', 'is', 'gonna', '[UNK]'],\n",
              " ['is', 'gonna', '[UNK]', 'so', '##o', '##n'],\n",
              " ['.', 'he', 'has', '[UNK]', 'but', 'his'],\n",
              " ['[UNK]', 'but', 'his', 'con', '##t', '##r'],\n",
              " ['##a', '##c', '##t', 'is', 'really', '[UNK]'],\n",
              " ['is', 'really', '[UNK]', 'and', 'i', 'would'],\n",
              " ['not', 'trade', 'him', '[UNK]', '.', 'how'],\n",
              " ['[UNK]', '.', 'how', 'do', 'you', 'know'],\n",
              " ['he', '’', 's', 'not', '[UNK]', 'on'],\n",
              " ['not', '[UNK]', 'on', 'the', 'court', '?'],\n",
              " ['what', 'if', 'he', 'just', '[UNK]', '?'],\n",
              " ['just', '[UNK]', '?', 'he', 'could', 'have'],\n",
              " ['start', '##ed', 'in', 'was', '[UNK]', 'got'],\n",
              " ['was', '[UNK]', 'got', 'up', '##s', '##e'],\n",
              " ['##t', 'and', 'now', '[UNK]', 'is', 'on'],\n",
              " ['[UNK]', 'is', 'on', 'the', 've', '##r'],\n",
              " ['you', 'can', 'if', 'you', 'have', '[UNK]'],\n",
              " ['you', 'have', '[UNK]', 's', '##c', '##o'],\n",
              " ['s', '##c', '##o', '##r', '##ers', '[UNK]'],\n",
              " ['##r', '##ers', '[UNK]', 'go', '##b', '##er'],\n",
              " ['##t', 'is', 'a', '[UNK]', '[UNK]', 'for'],\n",
              " ['[UNK]', '[UNK]', 'for', 'us', '.', '[UNK]'],\n",
              " ['us', '.', '[UNK]', '/', '[UNK]', ','],\n",
              " ['/', '[UNK]', ',', 'on', 'the', 'other'],\n",
              " ['on', 'the', 'other', '[UNK]', ',', 'would'],\n",
              " ['[UNK]', ',', 'would', 'love', 'to', 'see'],\n",
              " ['go', '##b', '##er', '##t', '[UNK]', 'them'],\n",
              " ['##t', '[UNK]', 'them', '.', 'did', '[UNK]'],\n",
              " ['.', 'did', '[UNK]', 'even', 'do', 'his'],\n",
              " ['even', 'do', 'his', '[UNK]', '[UNK]', 'and'],\n",
              " ['[UNK]', '[UNK]', 'and', 'try', 'to', 'get'],\n",
              " ['no', '##e', '##l', '?', '[UNK]', 'w'],\n",
              " ['?', '[UNK]', 'w', '##a', '##l', '##l'],\n",
              " ['is', 'probably', 'the', 'best', '[UNK]', 'point'],\n",
              " ['best', '[UNK]', 'point', '[UNK]', 'that', '[UNK]'],\n",
              " ['[UNK]', 'that', '[UNK]', 'will', 'ever', 'play'],\n",
              " ['stop', 'watch', '##ing', 'this', '[UNK]', '-'],\n",
              " ['this', '[UNK]', '-', '[UNK]', 'game', 'i'],\n",
              " ['[UNK]', 'game', 'i', 'don', '’', 't'],\n",
              " ['.', 'i', '’', 'm', '[UNK]', 'we'],\n",
              " ['m', '[UNK]', 'we', 'got', 'no', '##e'],\n",
              " ['##l', 'lo', '##l', '.', '[UNK]', '’'],\n",
              " ['.', '[UNK]', '’', 's', 'a', 'go'],\n",
              " ['guess', 'this', 'is', 'where', '[UNK]', 'play'],\n",
              " ['where', '[UNK]', 'play', '##s', 'a', 'r'],\n",
              " ['.', 'another', 'fa', 'that', '[UNK]', 'ok'],\n",
              " ['that', '[UNK]', 'ok', '##c', 'over', 'la'],\n",
              " ['##e', '##t', 'by', '##e', '[UNK]', 'by'],\n",
              " ['##e', '[UNK]', 'by', '##e', '[UNK]', 'they'],\n",
              " ['##e', '[UNK]', 'they', '[UNK]', 'in', 'other'],\n",
              " ['[UNK]', 'in', 'other', 'guys', 'into', 'the'],\n",
              " ['own', '##ers', '##hi', '##p', '[UNK]', 'since'],\n",
              " ['##p', '[UNK]', 'since', 'then', 'http', ':'],\n",
              " ['/', '/', 'www', '.', '[UNK]', '.'],\n",
              " ['.', '[UNK]', '.', 'com', '/', '[UNK]'],\n",
              " ['com', '/', '[UNK]', '-', 'a', '##d'],\n",
              " ['d', '##e', '##e', '##p', '-', '[UNK]'],\n",
              " ['##p', '-', '[UNK]', '/', '?', '?'],\n",
              " ['##a', '##l', '##t', 'for', 'these', '[UNK]'],\n",
              " ['for', 'these', '[UNK]', 'also', 'i', '’'],\n",
              " ['from', 'ok', '##c', '[UNK]', '.', '.'],\n",
              " ['[UNK]', '.', '.', '.', '?', '?'],\n",
              " ['?', 'w', '##t', '##f', '?', '[UNK]'],\n",
              " ['##f', '?', '[UNK]', 'we', '’', 're'],\n",
              " ['not', 'doing', 'that', 'his', '[UNK]', '[UNK]'],\n",
              " ['his', '[UNK]', '[UNK]', 'out', '.', 'i'],\n",
              " ['##t', '##ing', 'on', '[UNK]', ',', '[UNK]'],\n",
              " ['[UNK]', ',', '[UNK]', 'or', 'the', '[UNK]'],\n",
              " ['or', 'the', '[UNK]', 'ok', '##c', 'team'],\n",
              " [',', 'it', '’', 's', '[UNK]', 'down'],\n",
              " ['s', '[UNK]', 'down', 'to', 'r', '##o'],\n",
              " ['a', '##l', '##e', '##x', '[UNK]', '?'],\n",
              " ['##x', '[UNK]', '?', 'this', 'com', '##m'],\n",
              " ['##d', 'about', 'ok', '##c', 'than', '[UNK]'],\n",
              " ['##c', 'than', '[UNK]', 'about', 'his', '[UNK]'],\n",
              " ['about', 'his', '[UNK]', 'd', '##e', '##l'],\n",
              " ['##h', '##er', 'in', 'the', '[UNK]', 'court'],\n",
              " ['the', '[UNK]', 'court', 'but', 'if', 'lo'],\n",
              " ['doesn', '’', 't', '[UNK]', 'his', 'shot'],\n",
              " ['[UNK]', 'his', 'shot', 'then', 'they', '’'],\n",
              " ['then', 'they', '’', 'll', 'have', '[UNK]'],\n",
              " ['ll', 'have', '[UNK]', 'problem', '##s', 'in'],\n",
              " ['problem', '##s', 'in', 'the', '[UNK]', 'court'],\n",
              " ['the', '[UNK]', 'court', 'during', 'the', 'playoffs'],\n",
              " ['so', 'well', '.', 'the', '[UNK]', 'is'],\n",
              " ['the', '[UNK]', 'is', 'one', 'of', 'the'],\n",
              " ['’', 'll', 'see', 'the', '[UNK]', 'of'],\n",
              " ['the', '[UNK]', 'of', 'the', '“', 's'],\n",
              " ['##a', '##m', '##p', '##l', '##e', '[UNK]'],\n",
              " ['##l', '##e', '[UNK]', '”', 'me', '##m'],\n",
              " ['they', '’', 're', '[UNK]', 'of', 'an'],\n",
              " ['[UNK]', 'of', 'an', 'an', '##o', '##ma'],\n",
              " ['their', 'team', 't', '##h', '##o', '[UNK]'],\n",
              " ['##h', '##o', '[UNK]', '.', 'i', 'still'],\n",
              " ['##k', '##ing', ',', '[UNK]', 'and', 'the'],\n",
              " ['[UNK]', 'and', 'the', '[UNK]', 'would', 'still'],\n",
              " ['[UNK]', 'would', 'still', 'put', 'a', '[UNK]'],\n",
              " ['put', 'a', '[UNK]', 'on', 'the', 'court'],\n",
              " ['##r', '##t', '##h', '[UNK]', 'a', '[UNK]'],\n",
              " ['[UNK]', 'a', '[UNK]', 'for', '.', '[UNK]'],\n",
              " ['for', '.', '[UNK]', 'that', '’', 's'],\n",
              " ['that', '’', 's', 'not', '[UNK]', '.'],\n",
              " ['not', '[UNK]', '.', 'i', '[UNK]', 'the'],\n",
              " ['i', '[UNK]', 'the', 'team', 'will', '[UNK]'],\n",
              " ['team', 'will', '[UNK]', 'after', 'the', 'west'],\n",
              " ['##b', '##r', '##o', '##o', '##k', '[UNK]'],\n",
              " ['##o', '##k', '[UNK]', '.', 'they', '[UNK]'],\n",
              " ['.', 'they', '[UNK]', 'la', '##n', '##c'],\n",
              " ['t', 'even', 'a', 'good', '[UNK]', 'next'],\n",
              " ['good', '[UNK]', 'next', 'to', 'le', '##bron'],\n",
              " ['of', 'that', ',', '[UNK]', 'is', 'a'],\n",
              " ['[UNK]', 'is', 'a', 'be', '##n', '##c'],\n",
              " ['be', '##n', '##c', '##h', '[UNK]', '.'],\n",
              " ['##h', '[UNK]', '.', 'they', 'still', 'have'],\n",
              " ['they', 'still', 'have', 'a', '[UNK]', 'in'],\n",
              " ['a', '[UNK]', 'in', 'their', 'start', '##ing'],\n",
              " ['their', 'start', '##ing', '[UNK]', 'at', '[UNK]'],\n",
              " ['[UNK]', 'at', '[UNK]', '.', '[UNK]', ','],\n",
              " ['.', '[UNK]', ',', 's', '##t', '##a'],\n",
              " ['##n', '##d', 'over', 'this', '[UNK]', 'so'],\n",
              " ['this', '[UNK]', 'so', 'i', 'can', 'get'],\n",
              " ['##l', '##t', 'on', 'my', '[UNK]', 'more'],\n",
              " ['my', '[UNK]', 'more', 'players', 'to', 'play'],\n",
              " ['team', ',', 'more', '[UNK]', ',', 'con'],\n",
              " ['[UNK]', ',', 'con', '##t', '##r', '##a'],\n",
              " ['are', '10', '##0', '%', '[UNK]', 'we'],\n",
              " ['%', '[UNK]', 'we', 'a', '##d', '##d'],\n",
              " ['##ed', 'another', 'own', '##er', 'in', '[UNK]'],\n",
              " ['##er', 'in', '[UNK]', 'that', 'was', 'will'],\n",
              " ['’', 't', 'have', 'that', 'in', '[UNK]'],\n",
              " ['that', 'in', '[UNK]', '.', 'it', 'was'],\n",
              " ['.', 'it', 'was', 'just', '[UNK]', ','],\n",
              " ['just', '[UNK]', ',', '[UNK]', '.', 'i'],\n",
              " ['[UNK]', '.', 'i', '’', 'm', 'from'],\n",
              " ['’', 'm', 'from', '[UNK]', '.', 'i'],\n",
              " ['[UNK]', '.', 'i', '[UNK]', 'my', '[UNK]'],\n",
              " ['[UNK]', 'my', '[UNK]', 'fa', '##n', '##d'],\n",
              " ['##o', '##m', 'when', 'i', '[UNK]', 'the'],\n",
              " ['i', '[UNK]', 'the', 'team', 'would', 'go'],\n",
              " ['i', '’', 've', 'been', '[UNK]', 'my'],\n",
              " ['been', '[UNK]', 'my', '[UNK]', 'player', '('],\n",
              " ['[UNK]', 'player', '(', 'of', 'all', 'time'],\n",
              " ['.', 'can', '’', 't', '[UNK]', 'for'],\n",
              " ['t', '[UNK]', 'for', 'the', 'r', '##o'],\n",
              " ['##c', '##g', '##e', '##e', '[UNK]', 'it'],\n",
              " ['##e', '[UNK]', 'it', 'could', 'mean', 'a'],\n",
              " ['could', 'mean', 'a', '[UNK]', 'things', '.'],\n",
              " ['[UNK]', 'things', '.', '.', '.', '[UNK]'],\n",
              " ['.', '.', '[UNK]', 'on', 'in', 'the'],\n",
              " ['then', 'defense', 'is', '[UNK]', 'with', 'it'],\n",
              " ['[UNK]', 'with', 'it', 'it', 'could', 'also'],\n",
              " ['mean', 'that', 'his', '[UNK]', 'shot', 'is'],\n",
              " ['[UNK]', 'shot', 'is', 'pretty', '[UNK]', 'as'],\n",
              " ['pretty', '[UNK]', 'as', 'far', 'as', '[UNK]'],\n",
              " ['far', 'as', '[UNK]', 'and', 're', '##l'],\n",
              " ['’', 's', 'much', '[UNK]', 'to', 'con'],\n",
              " ['[UNK]', 'to', 'con', '##t', '##es', '##t'],\n",
              " ['yeah', ',', 'they', 'have', 'a', '[UNK]'],\n",
              " ['have', 'a', '[UNK]', 'of', '[UNK]', '[UNK]'],\n",
              " ['of', '[UNK]', '[UNK]', '[UNK]', 'le', '##bron'],\n",
              " ['[UNK]', 'le', '##bron', ',', 'la', '##n'],\n",
              " ['##a', '##c', '##e', 'the', '[UNK]', 'probably'],\n",
              " ['the', '[UNK]', 'probably', 'but', 'i', '[UNK]'],\n",
              " ['but', 'i', '[UNK]', 'it', '[UNK]', 'on'],\n",
              " ['it', '[UNK]', 'on', '##to', 'the', 'team'],\n",
              " ['a', '##d', '##d', '[UNK]', 'around', 'him'],\n",
              " ['[UNK]', 'around', 'him', 'that', 'just', 'don'],\n",
              " ['that', 'just', 'don', '’', 't', '[UNK]'],\n",
              " ['’', 't', '[UNK]', '?', 'he', '’'],\n",
              " ['play', 'with', 'le', '##bron', '.', '[UNK]'],\n",
              " ['##bron', '.', '[UNK]', 'also', 'sign', '##ed'],\n",
              " ['##o', '##l', '##e', '.', '[UNK]', '.'],\n",
              " ['.', '[UNK]', '.', 'he', 'might', 'start'],\n",
              " ['##n', '##g', 'to', 'a', '[UNK]', '3'],\n",
              " ['a', '[UNK]', '3', '##r', '##d', 'team'],\n",
              " ['##a', '##m', 'or', '[UNK]', '+', 'ass'],\n",
              " ['[UNK]', '+', 'ass', '##e', '##t', '##s'],\n",
              " ['##w', '##hi', 'to', 'la', 'and', '[UNK]'],\n",
              " ['la', 'and', '[UNK]', 'up', 'a', 's'],\n",
              " ['##l', '##o', '##t', 'for', '[UNK]', '?'],\n",
              " ['for', '[UNK]', '?', 'what', 'would', 'these'],\n",
              " ['what', 'would', 'these', '[UNK]', 'teams', '[UNK]'],\n",
              " ['[UNK]', 'teams', '[UNK]', 'back', 'to', 'the'],\n",
              " ['at', 'the', 'trade', '[UNK]', '.', '.'],\n",
              " ['[UNK]', '.', '.', '.', '?', 'he'],\n",
              " ['.', '?', 'he', 'left', '[UNK]', 'part'],\n",
              " ['left', '[UNK]', 'part', '##ly', 'for', 'this'],\n",
              " ['##n', '##z', '##o', 'can', '[UNK]', 'to'],\n",
              " ['can', '[UNK]', 'to', '[UNK]', 'on', 'a'],\n",
              " ['[UNK]', 'on', 'a', 't', '##a', '##n'],\n",
              " ['##er', 'team', 'much', 'as', '[UNK]', '.'],\n",
              " ['as', '[UNK]', '.', 'ok', '##c', 'i'],\n",
              " ['he', '##a', '##r', '##d', '[UNK]', '[UNK]'],\n",
              " ['##d', '[UNK]', '[UNK]', 'so', 'i', '##d'],\n",
              " ['##e', 'from', 'the', '[UNK]', '/', '[UNK]'],\n",
              " ['[UNK]', '/', '[UNK]', 'i', 'have', '##n'],\n",
              " ['’', 't', 'either', '.', '[UNK]', ','],\n",
              " ['.', '[UNK]', ',', 'it', '’', 's'],\n",
              " ['it', '’', 's', '[UNK]', 'up', 'by'],\n",
              " ['[UNK]', 'up', 'by', 'w', '##o', '##j'],\n",
              " ['mel', '##o', 'be', 'our', '[UNK]', 'in'],\n",
              " ['our', '[UNK]', 'in', 'the', '2', '##n'],\n",
              " ['the', '2', '##n', '##d', '[UNK]', '[UNK]'],\n",
              " ['##d', '[UNK]', '[UNK]', 'u', 'right', 'about'],\n",
              " ['got', '##t', '##a', '[UNK]', 'that', '[UNK]'],\n",
              " ['[UNK]', 'that', '[UNK]', '/', 'r', '##o'],\n",
              " ['##r', '##d', 'not', '[UNK]', 'by', 'any'],\n",
              " ['[UNK]', 'by', 'any', 'big', '[UNK]', 're'],\n",
              " ['big', '[UNK]', 're', '##p', '##o', '##r'],\n",
              " ['##t', '##er', 't', '##h', '##o', '[UNK]'],\n",
              " ['##h', '##o', '[UNK]', 'that', 'this', '[UNK]'],\n",
              " ['that', 'this', '[UNK]', 'is', 'used', 'to'],\n",
              " ['is', 'used', 'to', '[UNK]', 'another', '[UNK]'],\n",
              " ['[UNK]', 'another', '[UNK]', 'player', 'but', 'any'],\n",
              " ['##bron', 'is', 'pretty', 'ball', '-', '[UNK]'],\n",
              " ['ball', '-', '[UNK]', '(', 'nothing', 'wrong'],\n",
              " ['##z', '##o', 'need', '##s', 'to', '[UNK]'],\n",
              " ['##s', 'to', '[UNK]', 'a', '[UNK]', '[UNK]'],\n",
              " ['a', '[UNK]', '[UNK]', 'in', 'or', '##d'],\n",
              " ['in', 'or', '##d', '##er', 'to', '[UNK]'],\n",
              " ['##er', 'to', '[UNK]', 'le', '##bron', 'well'],\n",
              " ['le', '##bron', 'well', '.', '“', '[UNK]'],\n",
              " ['.', '“', '[UNK]', '[UNK]', '”', 's'],\n",
              " ['[UNK]', '”', 's', '##a', '##ma', '##n'],\n",
              " ['##t', '##h', '##a', ',', '[UNK]', 'doing'],\n",
              " [',', '[UNK]', 'doing', '[UNK]', 't', '##h'],\n",
              " ['[UNK]', 't', '##h', '##o', 'this', '[UNK]'],\n",
              " ['##o', 'this', '[UNK]', 'give', '##s', 'them'],\n",
              " ['##e', '##a', '##t', '/', '[UNK]', 'run'],\n",
              " ['/', '[UNK]', 'run', '##n', '##er', 'but'],\n",
              " ['i', 'think', 'the', '[UNK]', 'the', 'lakers'],\n",
              " ['[UNK]', 'the', 'lakers', 'are', 'going', 'with'],\n",
              " ['are', 'going', 'with', 'is', '[UNK]', 'play'],\n",
              " ['is', '[UNK]', 'play', '##ma', '##k', '##ers'],\n",
              " ['and', 'let', '##t', '##ing', 'everyone', '[UNK]'],\n",
              " ['##ing', 'everyone', '[UNK]', 'for', '[UNK]', 'other'],\n",
              " ['for', '[UNK]', 'other', '.', 'good', '[UNK]'],\n",
              " ['.', 'good', '[UNK]', '.', 'but', 'still'],\n",
              " ['##ers', '.', 'no', '[UNK]', 't', '##b'],\n",
              " ['[UNK]', 't', '##b', '##h', '.', 'i'],\n",
              " ['##w', 'you', 'were', 'a', '[UNK]', '[UNK]'],\n",
              " ['a', '[UNK]', '[UNK]', 'fa', '##n', 'no'],\n",
              " [',', 'it', '’', 's', 'one', '[UNK]'],\n",
              " ['s', 'one', '[UNK]', 'of', 'a', '[UNK]'],\n",
              " ['of', 'a', '[UNK]', 'league', 'game', 'because'],\n",
              " ['##r', '##a', '##c', '##t', 'is', '[UNK]'],\n",
              " ['##t', 'is', '[UNK]', '[UNK]', 'now', 'that'],\n",
              " ['[UNK]', 'now', 'that', 'le', '##bron', 'isn'],\n",
              " ['’', 't', 'there', 'yes', ',', '[UNK]'],\n",
              " ['yes', ',', '[UNK]', 'they', '’', 're'],\n",
              " ['they', '’', 're', 'good', '[UNK]', 'game'],\n",
              " ['good', '[UNK]', 'game', ',', 'im', '##o'],\n",
              " ['he', '’', 'll', '[UNK]', '[UNK]', 'or'],\n",
              " ['[UNK]', '[UNK]', 'or', 'take', 'a', 'better'],\n",
              " ['##y', 'if', 'he', 'gets', 'the', '[UNK]'],\n",
              " ['gets', 'the', '[UNK]', '.', '[UNK]', ','],\n",
              " ['.', '[UNK]', ',', 'i', 'need', '##a'],\n",
              " ['##o', '##w', '##ing', '[UNK]', 'yeah', '1'],\n",
              " ['[UNK]', 'yeah', '1', '[UNK]', '=', '1'],\n",
              " ['[UNK]', '=', '1', '[UNK]', 'i', 'think'],\n",
              " ['[UNK]', 'i', 'think', 'the', 'only', 'reason'],\n",
              " ['he', 'probably', 'got', 'from', 'the', '[UNK]'],\n",
              " ['from', 'the', '[UNK]', 'of', 'the', 'league'],\n",
              " ['something', 'that', 'isn', '’', 't', '[UNK]'],\n",
              " ['’', 't', '[UNK]', 'to', 'what', 'he'],\n",
              " ['’', 's', 'something', '[UNK]', 'to', 'the'],\n",
              " ['[UNK]', 'to', 'the', '5', '.', '[UNK]'],\n",
              " ['5', '.', '[UNK]', 'he', 'got', 'with'],\n",
              " ['this', 'little', 'but', '[UNK]', 'a', 'free'],\n",
              " ['[UNK]', 'a', 'free', 'r', '##ing', '”'],\n",
              " ['m', 'not', 'sure', '[UNK]', '.', '[UNK]'],\n",
              " ['[UNK]', '.', '[UNK]', 'maybe', 's', '##a'],\n",
              " ['maybe', 's', '##a', '##w', 'a', '[UNK]'],\n",
              " ['##w', 'a', '[UNK]', 'with', 'r', '##a'],\n",
              " ['him', 'up', '.', '[UNK]', '[UNK]', 'back'],\n",
              " ['[UNK]', '[UNK]', 'back', 'to', '[UNK]', 'and'],\n",
              " ...]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ],
      "source": [
        "user_tokenized = small_tokenizer.tokenize(\" \".join(user_comments))\n",
        "[user_tokenized[i * 3: i * 3 + 6] for i in range(int(len(user_tokenized) / 3)) if \"[UNK]\" in user_tokenized[i * 3: i * 3 + 6]]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zpPtxV5l0Pur",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f6182b8b-43d2-48c0-ac44-482865b6c034"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['Hmm,',\n",
              " 'be',\n",
              " 'this',\n",
              " 'really',\n",
              " 'worse',\n",
              " 'than',\n",
              " 'the',\n",
              " 'haiku',\n",
              " 'that',\n",
              " 'Magic']"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "user_lemmed_comments = [lem(word) for word in user_comments_str.split()]\n",
        "user_lemmed_comments[:10]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qtYjHvW31xIb"
      },
      "source": [
        "##### Random user's UNKs"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WKjopRSS1we7"
      },
      "outputs": [],
      "source": [
        "from collections import Counter\n",
        "\n",
        "\n",
        "all_unk_tokens = nba_unk_tokens #+ politics_unk_tokens\n",
        "\n",
        "\n",
        "\"\"\"\n",
        "all_tokens_count = len(all_unk_tokens)\n",
        "all_tokens_user_count = [0] * all_tokens_count\n",
        "for user in _users:\n",
        "    user_bin = \"0\" * (all_tokens_count - len(bin(user)[2:])) + bin(user)[2:]\n",
        "    for i in range(len(user_bin)):\n",
        "        all_tokens_user_count[i] += int(user_bin[i])\n",
        "all_unk_tokens = list(map(lambda x: x[1], filter(lambda token_key_value: all_tokens_user_count[token_key_value[0]] <= 10, enumerate(all_unk_tokens))))\n",
        "\"\"\"\n",
        "\n",
        "def get_user_unk_tokens(user):\n",
        "    joined_comments = \" \".join(all_comments[user])\n",
        "    return [joined_comments.count(n) for n in all_unk_tokens]\n",
        "\n",
        "def get_user_unk_tokens_optimized(user):\n",
        "    joined_comments = \" \".join(all_comments[user])\n",
        "\n",
        "    word_counts = Counter(joined_comments.split())\n",
        "    return [word_counts[unk] for unk in all_unk_tokens]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oS3MlF_bZROT"
      },
      "outputs": [],
      "source": [
        "# rels = { user: get_user_unk_tokens(user) for user in all_comments }"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "a = [10, 20, 15]\n",
        "list(map(lambda y: y[1], list(filter(lambda x: x[1] > 10, enumerate(a)))))"
      ],
      "metadata": {
        "id": "RMhelgFt_uYf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "995f9cd0-8072-46ee-feee-6791bea60850"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[20, 15]"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "rels = { user: get_user_unk_tokens_optimized(user) for user in [i for i in all_comments] }"
      ],
      "metadata": {
        "id": "X8u6F93DfLcO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QEu8OYTyed2b",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "58658985-3188-4b92-c94d-1c33e226e251"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "398572"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "my_sum = 0\n",
        "for r in rels.values():\n",
        "    my_sum += sum(r)\n",
        "my_sum"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KrXKgkQiftbC"
      },
      "outputs": [],
      "source": [
        "#rels2 = [list(map(lambda x: round(x / sum(val) if sum(val) > 0 else 0, 5), val)) for val in rels.values()]\n",
        "#len(rels2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "E3wLyRKlncxb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "76286d71-e5a5-40d8-8662-46dde27271ec"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'\\nimport json\\nwith open(\\'rels.txt\\', \\'w\\') as r:\\n    for i in rels2:\\n        r.write(json.dumps(i) + \"\\n\")\\n'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 20
        }
      ],
      "source": [
        "\"\"\"\n",
        "import json\n",
        "with open('rels.txt', 'w') as r:\n",
        "    for i in rels2:\n",
        "        r.write(json.dumps(i) + \"\\n\")\n",
        "\"\"\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mteel81OxMBB"
      },
      "outputs": [],
      "source": [
        "#_users = list(set(i) for i in rels.values())\n",
        "\n",
        "bool_arr_to_int = lambda arr: int(\"\".join(map(lambda x: str(int(x)), arr)), 2)\n",
        "\n",
        "_users = list(bool_arr_to_int(list(map(bool, i))) for i in rels.values())"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##### Optimized token search with binary digits"
      ],
      "metadata": {
        "id": "HTIx3Dv2A3S1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# def matching(source_user, target_user):\n",
        "#    return len(source_user.intersection(target_user)) / (len(source_user) + len(target_user))\n",
        "\n",
        "def matching(source_user: int, target_user: int):\n",
        "    source_user_shift = 2 ** (len(bin(source_user)) - 2) + source_user\n",
        "    return bin(source_user_shift & target_user).count(\"1\") / bin(source_user).count(\"1\")\n"
      ],
      "metadata": {
        "id": "1GewuRjJfK5d"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(all_unk_tokens)"
      ],
      "metadata": {
        "id": "INBkxyzO-g_h",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "77178d52-09ce-4e2c-9102-03e0f88b2d34"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8502"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "imDmfP4RhnO6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "67aa0345-7b58-471c-ed28-4cb441639424"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'\\n     u1, u2, u3, ..., un\\nu1    1 0.2 xxx\\nu2  0.2   1 yyy      ...\\nu3  xxx yyy   1\\n...\\nun      ...            1\\n\\n'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 27
        }
      ],
      "source": [
        "# focus = _users[80:120]\n",
        "focus = _users\n",
        "matrix = list(list(matching(user_j, user_i) for user_i in focus) for user_j in focus)\n",
        "\"\"\"\n",
        "_user1 = {token1, token2, token3, token4}\n",
        "_user2 = {token4, token5, token6, token7, token12}\n",
        "...\n",
        "_usern = {token3, token6, token7}\n",
        "\n",
        "_users = [user1, user2, ..., usern]\n",
        "\n",
        "matrix = [\n",
        "    [matching(user1, user1), matching(user1, user2), ..., matching(user1, usern)],\n",
        "    [matching(user2, user1), matching(user2, user2), ..., matching(user2, usern)],\n",
        "    ...\n",
        "    [matching(usern, user1), matching(usern, user2), ..., matching(usern, usern)],\n",
        "]\n",
        "\"\"\"\n",
        "\n",
        "# User x with all others\n",
        "\n",
        "# User x with user y\n",
        "\n",
        "\"\"\"\n",
        "     u1, u2, u3, ..., un\n",
        "u1    1 0.2 xxx\n",
        "u2  0.2   1 yyy      ...\n",
        "u3  xxx yyy   1\n",
        "...\n",
        "un      ...            1\n",
        "\n",
        "\"\"\"\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "W06_xByCyR05",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "86926752-72b8-4253-f41d-12b89367197e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "ax = sns.heatmap(matrix)\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EKzy_ySHcw0B",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "outputId": "830448a6-71e6-4856-ba94-e73253d57c77"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "531.91\n",
            "892.83\n",
            "515\n",
            "830\n"
          ]
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import statistics\n",
        "\n",
        "nba_users_unk_counts = [bin(user).count(\"1\") for user in _users[1:100]]\n",
        "politics_users_unk_counts = [bin(user).count(\"1\") for user in _users[101:200]]\n",
        "\n",
        "indices = range(len(nba_users_unk_counts))\n",
        "fig, ax = plt.subplots()\n",
        "ax.plot(indices, nba_users_unk_counts)\n",
        "\n",
        "ax.set_xlabel('X-axis')\n",
        "ax.set_ylabel('Y-axis')\n",
        "ax.set_title('NBA Plot')\n",
        "\n",
        "plt.show()\n",
        "\n",
        "indices = range(len(politics_users_unk_counts))\n",
        "fig, ax = plt.subplots()\n",
        "ax.plot(indices, politics_users_unk_counts)\n",
        "\n",
        "ax.set_xlabel('X-axis')\n",
        "ax.set_ylabel('Y-axis')\n",
        "ax.set_title('Politics Plot')\n",
        "\n",
        "plt.show()\n",
        "\n",
        "print(sum(nba_users_unk_counts) / 100)\n",
        "print(sum(politics_users_unk_counts) / 100)\n",
        "\n",
        "# Calculate middle values\n",
        "print(statistics.median(nba_users_unk_counts))\n",
        "print(statistics.median(politics_users_unk_counts))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##### Simulation setup"
      ],
      "metadata": {
        "id": "WWYcGGMwK5hQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "len(bin(_users[0])[2:])"
      ],
      "metadata": {
        "id": "fuIbbD6O-tKd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "db47d8ed-2b69-445b-99a3-bd61c8f02806"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8495"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from numpy.random import randint\n",
        "\n",
        "def get_random_cluster():\n",
        "    token_cluster_indices = [randint(0, 10) for unk_token in all_unk_tokens]\n",
        "\n",
        "    clusters_count = 10\n",
        "    cluster_str_len = len(all_unk_tokens)\n",
        "    clusters_arr = [list('0' * cluster_str_len) for i in range(clusters_count)]\n",
        "\n",
        "    for index, token_cluster_index in enumerate(token_cluster_indices):\n",
        "        clusters_arr[token_cluster_index][index] = \"1\"\n",
        "\n",
        "    return list(map(lambda cluster_arr: int(''.join(cluster_arr), 2), clusters_arr))\n",
        "\n",
        "clusters = get_random_cluster()"
      ],
      "metadata": {
        "id": "EptdoTAAbh8P"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "# Matching score of user x in regards to a cluster\n",
        "def get_cluster_index_with_max_matching_score(_user):\n",
        "  matching_scores = [matching(_user, cluser) for cluser in clusters]\n",
        "  return np.argmax(matching_scores)\n",
        "\n",
        "# Matching score of user x in regards to a cluster\n",
        "def get_max_matching_score(_user, _clusters):\n",
        "  matching_scores = [matching(_user, cluser) for cluser in _clusters]\n",
        "  return max(matching_scores)\n",
        "\n",
        "get_cluster_index_with_max_matching_score(_users[0])"
      ],
      "metadata": {
        "id": "f4U2zne5ckvQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "20bcef06-3ca8-4e6d-d49a-3e817d3b3bf2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def get_better_distributed_clusters(min_count=0, max_count=9999):\n",
        "    global clusters\n",
        "    while True:\n",
        "        cluster_user = [list() for i in range(10)]\n",
        "        clusters = get_random_cluster()\n",
        "\n",
        "        for index, user in enumerate(_users):\n",
        "            cluster = get_cluster_index_with_max_matching_score(user)\n",
        "            cluster_user[cluster].append(index)\n",
        "\n",
        "        if min(list(map(len, cluster_user))) >= min_count and max(list(map(len, cluster_user))) <= max_count:\n",
        "            return clusters, cluster_user\n",
        "\n",
        "clusters, cluster_user = get_better_distributed_clusters(5, 60)\n",
        "list(map(len, cluster_user))"
      ],
      "metadata": {
        "id": "F5PPk8HSFwNI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a5559a16-489a-4437-a209-755bbe57f3fb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[46, 14, 19, 13, 44, 8, 14, 9, 30, 10]"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "all_tokens_count = len(all_unk_tokens)\n",
        "\n",
        "all_tokens_user_count = [0] * all_tokens_count\n",
        "for user in _users:\n",
        "    user_bin = \"0\" * (all_tokens_count - len(bin(user)[2:])) + bin(user)[2:]\n",
        "    for i in range(len(user_bin)):\n",
        "        all_tokens_user_count[i] += int(user_bin[i])\n",
        "\n",
        "def get_x_tokens_not_in_cluster(_user, _cluster, x=100):\n",
        "    tokens_int = _user & ~_cluster # Exists on user and not in cluster\n",
        "    tokens_cut = bin(tokens_int)[2:]\n",
        "    missing_leading_zeroes = all_tokens_count - len(tokens_cut)\n",
        "    tokens = \"0\" * missing_leading_zeroes + tokens_cut\n",
        "\n",
        "    one_indices = [i for i, val in enumerate(tokens) if val == \"1\" and all_tokens_user_count[i] <= 10]\n",
        "\n",
        "    return random.sample(one_indices, min(x, len(one_indices)))"
      ],
      "metadata": {
        "id": "6UNv9ekL3-5O"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import bisect\n",
        "\n",
        "def binary_search(arr, target):\n",
        "    index = bisect.bisect_left(arr, target)\n",
        "    if index < len(arr) and arr[index] == target:\n",
        "        return index\n",
        "    else:\n",
        "        return -1  # Target not found"
      ],
      "metadata": {
        "id": "RapiVgbI-51H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# User voing:\n",
        "#       Ako naraste max matching score od matching scora njegovog clustera +1 (barem 5% veci)\n",
        "#       Ako padne max matching score od matching scora njegovog clustera -1 (barem 5% manji)\n",
        "# Random user predlaze 100 tokena koji nisu u njegovom clusteru (on je u clusteru get_cluster_index_with_max_matching_score(user) ) (test sa 10 tokena prvo)\n",
        "\n",
        "import math\n",
        "\n",
        "# Map to fetch cluster by token\n",
        "def get_unk_tokens_cluster_map():\n",
        "    all_tokens_count = len(all_unk_tokens)\n",
        "    unk_tokens_cluster_map = [-1] * len(all_unk_tokens)\n",
        "    for i, c in enumerate(clusters):\n",
        "        one_indices = [j + all_tokens_count - len(bin(c)[2:]) for j, val in enumerate(bin(c)[2:]) if val == \"1\"]\n",
        "        for index in one_indices:\n",
        "            unk_tokens_cluster_map[index] = i\n",
        "\n",
        "    return unk_tokens_cluster_map\n",
        "\n",
        "def get_new_clusters_after_transfering_token(cluster_from_index, cluster_to_index, token_to_transfer):\n",
        "    # token transition from cluster to cluster\n",
        "    temp_clusters = [i for i in clusters]\n",
        "    temp_clusters[cluster_from_index] = temp_clusters[cluster_from_index] - (2 ** (all_tokens_count - token_to_transfer - 1))\n",
        "    temp_clusters[cluster_to_index] = temp_clusters[cluster_to_index] + (2 ** (all_tokens_count - token_to_transfer - 1))\n",
        "    return temp_clusters\n",
        "\n",
        "\n",
        "def user_propose_tokens(user, batch_size):\n",
        "    cluster_index = get_cluster_index_with_max_matching_score(user)\n",
        "    cluster = clusters[cluster_index] # User cluster\n",
        "\n",
        "    return get_x_tokens_not_in_cluster(user, cluster, batch_size) # Tokens to transfer\n",
        "\n",
        "\n",
        "def user_vote(user, token_index, unk_tokens_cluster_map, min_increase=0):\n",
        "    matching_score = get_max_matching_score(user, clusters)\n",
        "    cluster_index = get_cluster_index_with_max_matching_score(user)\n",
        "\n",
        "    # TODO treba vidjet dal se glasa token po token ili jedan glas za cijeli batch\n",
        "\n",
        "    new_clusters = get_new_clusters_after_transfering_token(unk_tokens_cluster_map[token_index], cluster_index, token_index)\n",
        "    new_matching_score = get_max_matching_score(user, new_clusters)\n",
        "\n",
        "    if new_matching_score > matching_score + min_increase:\n",
        "        #print(new_matching_score - matching_score)\n",
        "        return 1\n",
        "\n",
        "    if new_matching_score < matching_score - min_increase:\n",
        "        return -1\n",
        "\n",
        "    return 0\n",
        "\n",
        "\n",
        "def do_vote(user, unk_tokens_cluster_map, batch_size, min_inc):\n",
        "    proposed_tokens = user_propose_tokens(user, batch_size)\n",
        "\n",
        "    votes = [0] * len(proposed_tokens)\n",
        "    for _user in _users:\n",
        "        for i, proposed_token in enumerate(proposed_tokens):\n",
        "            votes[i] += user_vote(_user, proposed_token, unk_tokens_cluster_map, min_inc)\n",
        "\n",
        "    return proposed_tokens, votes"
      ],
      "metadata": {
        "id": "zSEqzMpXAp27"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### The simulation"
      ],
      "metadata": {
        "id": "DQn2LtfJJWs2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "len(all_unk_tokens)"
      ],
      "metadata": {
        "id": "hTZMd8b4BX7G",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "360d27b8-9aaf-4986-8ac9-b827e053a862"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "8502"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def save_clusters():\n",
        "    global clusters_copy\n",
        "    clusters_copy = [i for i in clusters]\n",
        "save_clusters()\n",
        "#unk_tokens_cluster_map = get_unk_tokens_cluster_map()\n",
        "#unk_tokens_cluster_map"
      ],
      "metadata": {
        "id": "GNRM53uhJpC8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def load_clusters():\n",
        "    global clusters\n",
        "    clusters = [i for i in clusters_copy]"
      ],
      "metadata": {
        "id": "a69uiGgnJsJw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def cluster_user_status():\n",
        "    cluster_user = [list() for i in range(10)]\n",
        "\n",
        "    for index, user in enumerate(_users):\n",
        "        cluster = get_cluster_index_with_max_matching_score(user)\n",
        "        cluster_user[cluster].append(index)\n",
        "\n",
        "    return list(map(len, cluster_user))"
      ],
      "metadata": {
        "id": "hLRsE8Lcs9Tf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_purities():\n",
        "    purity_cluster_user = [list() for i in range(10)]\n",
        "\n",
        "    for index, user in enumerate(_users):\n",
        "        c = get_cluster_index_with_max_matching_score(user)\n",
        "        purity_cluster_user[c].append(index)\n",
        "\n",
        "    return list(map(lambda cluster: round(2 * abs(50 - sum(100 * list(map(lambda user: user <= (len(_users) / 2), cluster))) / len(cluster)), 1), purity_cluster_user))\n",
        "\n",
        "get_purities()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fpI8M9fcPPG5",
        "outputId": "9c091454-34f5-4c1a-c65f-42d20eb65af5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[4.3, 57.1, 5.3, 38.5, 63.6, 0.0, 14.3, 33.3, 20.0, 80.0]"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "def simulation_loop(vote_batch_size, min_inc):\n",
        "    global clusters\n",
        "    user = random.choice(_users)\n",
        "    unk_tokens_cluster_map = get_unk_tokens_cluster_map()\n",
        "    tokens, votes = do_vote(user, unk_tokens_cluster_map, vote_batch_size, min_inc)\n",
        "\n",
        "    user_cluster_index = get_cluster_index_with_max_matching_score(user)\n",
        "    for i in range(len(tokens)):\n",
        "        if votes[i] > 0:\n",
        "            unk_tokens_cluster_map = get_unk_tokens_cluster_map()\n",
        "            clusters = get_new_clusters_after_transfering_token(unk_tokens_cluster_map[tokens[i]], user_cluster_index, tokens[i])"
      ],
      "metadata": {
        "id": "hWtvURa5oq-c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "load_clusters()\n",
        "print(\"Initial array\", cluster_user_status())\n",
        "for i in range(1000):\n",
        "    simulation_loop(100, 0)\n",
        "    print(\"Loop\", i)\n",
        "    cus = cluster_user_status()\n",
        "    gp = get_purities()\n",
        "\n",
        "    print(list(map(lambda x: \" \" * (5 - len(str(x))) + str(x), cus)), \"Clusters over 30 users:\", len(list(filter(lambda x: x > 30, cus))))\n",
        "    print(list(map(lambda x: \" \" * (5 - len(str(x[1]))) + (str(x[1]) if cus[x[0]] > 30 else (\" \" * len(str(x[1])))), enumerate(gp))))\n",
        "    print()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "vbukIxTztLjC",
        "outputId": "0883dc35-e0cc-4c8c-8576-05e28904d9db"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Initial array [46, 14, 19, 13, 44, 8, 14, 9, 30, 10]\n",
            "Loop 0\n",
            "['   46', '   14', '   20', '   15', '   43', '    8', '   13', '    6', '   27', '   15'] Clusters over 30 users: 2\n",
            "['  4.3', '     ', '     ', '     ', ' 62.8', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 1\n",
            "['   44', '   13', '   32', '   14', '   39', '    8', '   13', '    6', '   25', '   13'] Clusters over 30 users: 3\n",
            "['  4.5', '     ', ' 18.8', '     ', ' 74.4', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 2\n",
            "['   45', '   13', '   31', '   15', '   39', '    8', '   12', '    6', '   25', '   13'] Clusters over 30 users: 3\n",
            "['  6.7', '     ', ' 22.6', '     ', ' 74.4', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 3\n",
            "['   44', '   13', '   39', '   14', '   38', '    7', '    9', '    7', '   25', '   11'] Clusters over 30 users: 3\n",
            "['  9.1', '     ', ' 23.1', '     ', ' 73.7', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 4\n",
            "['   44', '   13', '   38', '   14', '   38', '    8', '    9', '    7', '   25', '   11'] Clusters over 30 users: 3\n",
            "['  9.1', '     ', ' 26.3', '     ', ' 73.7', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 5\n",
            "['   43', '   24', '   37', '   14', '   35', '    8', '    7', '    7', '   21', '   11'] Clusters over 30 users: 3\n",
            "[' 11.6', '     ', ' 24.3', '     ', ' 77.1', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 6\n",
            "['   41', '   35', '   31', '   12', '   36', '    7', '    7', '    6', '   21', '   11'] Clusters over 30 users: 4\n",
            "['  2.4', ' 42.9', ' 16.1', '     ', ' 77.8', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 7\n",
            "['   35', '   54', '   28', '   12', '   34', '    7', '    7', '    3', '   17', '   10'] Clusters over 30 users: 3\n",
            "['  2.9', ' 48.1', '     ', '     ', ' 76.5', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 8\n",
            "['   36', '   54', '   27', '   12', '   33', '    7', '    7', '    3', '   20', '    8'] Clusters over 30 users: 3\n",
            "['  0.0', ' 48.1', '     ', '     ', ' 75.8', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 9\n",
            "['   29', '   72', '   23', '    9', '   32', '    7', '    6', '    3', '   20', '    6'] Clusters over 30 users: 2\n",
            "['     ', ' 50.0', '     ', '     ', ' 81.2', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 10\n",
            "['   29', '   72', '   23', '    9', '   32', '    7', '    6', '    3', '   20', '    6'] Clusters over 30 users: 2\n",
            "['     ', ' 50.0', '     ', '     ', ' 81.2', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 11\n",
            "['   22', '   88', '   21', '    8', '   28', '    7', '    5', '    3', '   19', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 45.5', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 12\n",
            "['   22', '   91', '   20', '    8', '   28', '    7', '    3', '    3', '   19', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 45.1', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 13\n",
            "['   22', '   91', '   20', '    8', '   29', '    7', '    3', '    3', '   18', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 45.1', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 14\n",
            "['   26', '   90', '   17', '    8', '   29', '    7', '    3', '    3', '   18', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 44.4', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 15\n",
            "['   30', '   84', '   15', '    8', '   29', '    5', '    3', '    3', '   24', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 45.2', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 16\n",
            "['   29', '   85', '   15', '    8', '   29', '    6', '    3', '    3', '   23', '    6'] Clusters over 30 users: 1\n",
            "['     ', ' 43.5', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 17\n",
            "['   28', '   90', '   14', '    8', '   28', '    6', '    3', '    3', '   22', '    5'] Clusters over 30 users: 1\n",
            "['     ', ' 37.8', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 18\n",
            "['   27', '   88', '   14', '    8', '   29', '    6', '    3', '    3', '   24', '    5'] Clusters over 30 users: 1\n",
            "['     ', ' 40.9', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 19\n",
            "['   30', '   87', '   14', '    7', '   27', '    6', '    3', '    3', '   25', '    5'] Clusters over 30 users: 1\n",
            "['     ', ' 40.2', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 20\n",
            "['   35', '   81', '   14', '    7', '   28', '    6', '    3', '    2', '   26', '    5'] Clusters over 30 users: 2\n",
            "[' 31.4', ' 45.7', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 21\n",
            "['   35', '   81', '   14', '    7', '   28', '    6', '    3', '    2', '   26', '    5'] Clusters over 30 users: 2\n",
            "[' 31.4', ' 45.7', '     ', '     ', '     ', '     ', '     ', '     ', '     ', '     ']\n",
            "\n",
            "Loop 22\n",
            "['   36', '   69', '   14', '    6', '   29', '    6', '    3', '    1', '   37', '    6'] Clusters over 30 users: 3\n",
            "[' 27.8', ' 44.9', '     ', '     ', '     ', '     ', '     ', '     ', ' 18.9', '     ']\n",
            "\n",
            "Loop 23\n",
            "['   33', '   81', '   12', '    6', '   29', '    6', '    3', '    1', '   31', '    5'] Clusters over 30 users: 3\n",
            "[' 39.4', ' 45.7', '     ', '     ', '     ', '     ', '     ', '     ', ' 16.1', '     ']\n",
            "\n",
            "Loop 24\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "ZeroDivisionError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mZeroDivisionError\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-44-5e136242cb2b>\u001b[0m in \u001b[0;36m<cell line: 3>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Loop\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[0mcus\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcluster_user_status\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m     \u001b[0mgp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_purities\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;34m\" \"\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m5\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcus\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Clusters over 30 users:\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mx\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m30\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcus\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-41-06a00abfa7f8>\u001b[0m in \u001b[0;36mget_purities\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m         \u001b[0mpurity_cluster_user\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mcluster\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mabs\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m50\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m100\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0muser\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0muser\u001b[0m \u001b[0;34m<=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_users\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcluster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcluster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpurity_cluster_user\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mget_purities\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-41-06a00abfa7f8>\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(cluster)\u001b[0m\n\u001b[1;32m      6\u001b[0m         \u001b[0mpurity_cluster_user\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mcluster\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mabs\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m50\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m100\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0muser\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0muser\u001b[0m \u001b[0;34m<=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_users\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcluster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcluster\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpurity_cluster_user\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mget_purities\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mZeroDivisionError\u001b[0m: division by zero"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "a = [1, 3, 2]\n",
        "b = [0.1, 0.2, 0.3]\n",
        "sorted(zip(a, b), key=lambda x: x[0], reverse=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hRb2X54AcpRi",
        "outputId": "cccdc77d-9b39-4581-9688-a8d13c9670c4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[(3, 0.2), (2, 0.3), (1, 0.1)]"
            ]
          },
          "metadata": {},
          "execution_count": 134
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# TODO Simulation Loop (barem 10*200 = 2000 ciklusa)\n",
        "#\n",
        "#\n",
        "#\n",
        "#   Ako je suma glasanja >0, prebaci taj token (optimizirati za sve tokene odjednom - batch voting)\n",
        "#\n",
        "#   Prati mean(max_matching_score) - MMMS i reportaj purity jer imaš labele. MMMS utječe na simulated annealing parameter"
      ],
      "metadata": {
        "id": "bmwHM95J7fk3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "all_tokens_user_count = [0] * len(all_unk_tokens)\n",
        "for user in _users:\n",
        "    user_bin = \"0\" * (len(all_unk_tokens) - len(bin(user)[2:])) + bin(user)[2:]\n",
        "    for i in range(len(user_bin)):\n",
        "        all_tokens_user_count[i] += int(user_bin[i])\n",
        "\n",
        "print(all_tokens_user_count[:100])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ohHaNEIN3DKO",
        "outputId": "26128bfd-8750-4de0-da97-e7f4eefb73ef"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[1, 4, 0, 2, 1, 1, 0, 148, 0, 33, 7, 1, 0, 76, 162, 3, 17, 109, 171, 2, 64, 0, 6, 2, 88, 31, 5, 24, 5, 5, 7, 1, 33, 13, 45, 0, 2, 43, 28, 141, 22, 51, 8, 0, 1, 9, 7, 13, 87, 34, 19, 46, 0, 59, 12, 10, 1, 0, 1, 32, 14, 0, 111, 6, 3, 102, 99, 12, 37, 44, 178, 5, 1, 1, 7, 12, 72, 59, 3, 0, 23, 11, 0, 4, 0, 3, 41, 2, 7, 9, 89, 20, 1, 5, 2, 0, 4, 19, 22, 134]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(\">10 =\", 100 * len([i for i in all_tokens_user_count if i > 10]) / len(all_tokens_user_count))\n",
        "print(\">50 =\", 100 * len([i for i in all_tokens_user_count if i > 50]) / len(all_tokens_user_count))\n",
        "print(\">100 =\", 100 * len([i for i in all_tokens_user_count if i > 100]) / len(all_tokens_user_count))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kat2zQQR3SI3",
        "outputId": "c8188058-4959-4c2a-8d1e-8ec1d97e87f6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            ">10 =  34.227240649258995\n",
            ">50 =  10.597506469066102\n",
            ">100 =  4.222535873912021\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Isklljucio sam tokene koji se ponavljaju kod vise od 10 usera\n",
        "# Promjenio sam da user vote-a 10 tokena a ne 100.\n",
        "\n",
        "# I dalje se desava da useri odu u 1 cluster\n",
        "# ŠProbati pratiti puritz clustera na svaku iteraciju"
      ],
      "metadata": {
        "id": "VYW3iNKRHu0P"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [
        "Uw_vHitmsBzs",
        "WN01UidMg_5J",
        "8XlaIv3tZdtZ",
        "9X4lTMUx91_x",
        "j3e5FEGUaOmf",
        "oUJq3MXTiFJn",
        "ZYkHR5j4t_jz",
        "HTIx3Dv2A3S1",
        "WWYcGGMwK5hQ"
      ],
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}