task7.1_aspect_sentiment_eval.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Аспектный анализ тональности текстов: используем возможности fasttext"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Если Вы запускаете ноутбук на colab или kaggle,\n",
    "# выполните следующие строчки, чтобы подгрузить библиотеку dlnlputils:\n",
    "\n",
    "# !git clone https://github.com/Samsung-IT-Academy/stepik-dl-nlp.git && pip install -r stepik-dl-nlp/requirements.txt\n",
    "# import sys; sys.path.append('./stepik-dl-nlp')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<torch._C.Generator at 0x7f33a20e4bf0>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "\n",
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "torch.manual_seed(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## BIO-тэггинг для обучения модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from dlnlputils.sentiment_utils import parse_xml_aspect, show_markup,prepare_data, generate_markup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.tokenize import RegexpTokenizer\n",
    "\n",
    "word_tokenizer = RegexpTokenizer('\\w+|\\$[\\d\\.]+|\\S+')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Если Вы запускаете ноутбук на colab или kaggle, добавьте в начало пути ./stepik-dl-nlp\n",
    "xml_train = 'datasets/sentirueval2015/SentiRuEval_car_markup_train.xml'\n",
    "xml_test  = 'datasets/sentirueval2015/SentiRuEval_car_markup_test.xml'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "texts_w_aspect_spans = parse_xml_aspect(xml_train)\n",
    "texts_w_aspect_spans = [(text.lower(),spans) for text, spans in texts_w_aspect_spans]\n",
    "training_data        = prepare_data(texts_w_aspect_spans, word_tokenizer.tokenize)\n",
    "\n",
    "texts_w_aspect_spans = parse_xml_aspect(xml_test)\n",
    "texts_w_aspect_spans = [(text.lower(),spans) for text, spans in texts_w_aspect_spans]\n",
    "test_data            = prepare_data(texts_w_aspect_spans, word_tokenizer.tokenize)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2210, 1922)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(training_data), len(test_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">все так чисто ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">опрятно<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, светло и тепло , что я сразу же очутился в своей тарелке .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('все так чисто , опрятно , светло и тепло , что я сразу же очутился в своей тарелке .',\n",
       "          [Span(0, 15, 'Other'),\n",
       "           Span(16, 23, 'Comfort'),\n",
       "           Span(24, 84, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_recipe, test_tags = training_data[400]\n",
    "\n",
    "text, spans = generate_markup(test_recipe, test_tags) \n",
    "\n",
    "show_markup(text, spans)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Статистика по тэгам"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "from collections import defaultdict\n",
    "\n",
    "def form_vocabulary_and_tagset(words_w_tags):\n",
    "    dictionary = defaultdict(Counter)\n",
    "    for words,tags in words_w_tags: \n",
    "        for word, tag in zip(words, tags):\n",
    "            dictionary[tag].update([word])      \n",
    "    return dictionary, set(dictionary.keys()) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "word_counter, labels = form_vocabulary_and_tagset(training_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "103\tin I-Appearance\n",
      "49\tin I-Safety\n",
      "284\tin I-Comfort\n",
      "169\tin B-Costs\n",
      "200\tin I-Whole\n",
      "97\tin I-Costs\n",
      "464\tin B-Reliability\n",
      "60\tin B-Safety\n",
      "236\tin B-Whole\n",
      "189\tin B-Appearance\n",
      "393\tin I-Driveability\n",
      "300\tin I-Reliability\n",
      "584\tin B-Driveability\n",
      "4933\tin Other\n",
      "478\tin B-Comfort\n"
     ]
    }
   ],
   "source": [
    "for label in labels:\n",
    "    num_words = len(word_counter[label])\n",
    "    print(f'{num_words}\\tin {label}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('в', 29),\n",
       " ('салоне', 17),\n",
       " ('места', 15),\n",
       " ('сиденья', 14),\n",
       " ('багажник', 12),\n",
       " ('сидений', 10),\n",
       " ('салона', 9),\n",
       " ('на', 8),\n",
       " ('для', 6),\n",
       " ('сидения', 5),\n",
       " ('багажника', 5),\n",
       " ('с', 5),\n",
       " ('и', 5),\n",
       " ('комплектации', 5),\n",
       " ('не', 5),\n",
       " ('автомобиля', 4),\n",
       " ('задние', 4),\n",
       " ('управления', 4),\n",
       " ('салон', 4),\n",
       " ('автомобиле', 3),\n",
       " ('тесно', 3),\n",
       " ('отделение', 3),\n",
       " ('передач', 3),\n",
       " ('фар', 3),\n",
       " ('стекло', 3),\n",
       " ('хватает', 3),\n",
       " ('приборов', 3),\n",
       " ('ног', 3),\n",
       " ('двери', 2),\n",
       " ('пластик', 2),\n",
       " ('авто', 2),\n",
       " ('машины', 2),\n",
       " ('колонка', 2),\n",
       " ('все', 2),\n",
       " ('пассажиров', 2),\n",
       " ('панели', 2),\n",
       " ('диване', 2),\n",
       " ('обслуживании', 2),\n",
       " ('заднем', 2),\n",
       " ('водителя', 2),\n",
       " ('от', 2),\n",
       " ('панель', 2),\n",
       " ('эксплуатации', 2),\n",
       " ('универсал', 2),\n",
       " ('-контроль', 2),\n",
       " ('передних', 2),\n",
       " ('потолок', 2),\n",
       " ('достаточно', 2),\n",
       " ('магнитолу', 2),\n",
       " ('удобно', 2),\n",
       " ('системы', 2),\n",
       " ('сидениях', 2),\n",
       " ('арок', 2),\n",
       " ('chery', 2),\n",
       " ('qq6', 2),\n",
       " ('1', 2),\n",
       " ('.1', 2),\n",
       " ('стучит', 2),\n",
       " ('безопасности', 2),\n",
       " ('\"космо\"', 1),\n",
       " ('руссифицированный', 1),\n",
       " ('удобные', 1),\n",
       " ('колонке', 1),\n",
       " ('всяких', 1),\n",
       " ('опций', 1),\n",
       " ('шумы', 1),\n",
       " ('логане', 1),\n",
       " ('бояться', 1),\n",
       " ('за', 1),\n",
       " ('усталость', 1),\n",
       " ('болела', 1),\n",
       " ('focus', 1),\n",
       " ('2', 1),\n",
       " ('седан', 1),\n",
       " ('автомат', 1),\n",
       " ('уно', 1),\n",
       " ('больше', 1),\n",
       " ('руле', 1),\n",
       " ('подлокотник', 1),\n",
       " (')', 1),\n",
       " ('заднего', 1),\n",
       " ('вида', 1),\n",
       " ('спортивный', 1),\n",
       " ('болид', 1),\n",
       " ('пластика', 1),\n",
       " ('экран', 1),\n",
       " ('бизнес', 1),\n",
       " ('-класс', 1),\n",
       " ('открывать', 1),\n",
       " ('окна', 1),\n",
       " ('полка', 1),\n",
       " ('акустики', 1),\n",
       " ('двигатель', 1),\n",
       " ('зад', 1),\n",
       " ('диваном', 1),\n",
       " ('подлокотником', 1),\n",
       " ('свет', 1),\n",
       " ('топливного', 1),\n",
       " ('бака', 1),\n",
       " ('компьютере', 1),\n",
       " ('-проигрывателе', 1),\n",
       " ('кпп', 1),\n",
       " ('тесноты', 1),\n",
       " ('мультимедийный', 1),\n",
       " ('плеер', 1),\n",
       " ('тихо', 1),\n",
       " ('диванчик', 1),\n",
       " ('только', 1),\n",
       " ('задний', 1),\n",
       " ('дива', 1),\n",
       " ('пакетам', 1),\n",
       " ('поясница', 1),\n",
       " ('-ноги', 1),\n",
       " ('зеркал', 1),\n",
       " ('запаска', 1),\n",
       " ('кнопок', 1),\n",
       " ('легкостью', 1),\n",
       " ('поместиться', 1),\n",
       " ('переднего', 1),\n",
       " ('пассажира', 1),\n",
       " ('положение', 1),\n",
       " ('прогревается', 1),\n",
       " ('температуры', 1),\n",
       " ('нагрева', 1),\n",
       " ('салазках', 1),\n",
       " ('ужасно', 1),\n",
       " ('части', 1),\n",
       " ('хайса', 1),\n",
       " ('люк', 1),\n",
       " ('кондиционера', 1),\n",
       " ('нам', 1),\n",
       " ('мест', 1),\n",
       " ('субары', 1),\n",
       " ('огромной', 1),\n",
       " ('баранкой', 1),\n",
       " ('своих', 1),\n",
       " ('местах', 1),\n",
       " ('неудобств', 1),\n",
       " ('появлялись', 1),\n",
       " ('скрипы', 1),\n",
       " ('часть', 1),\n",
       " ('пространства', 1),\n",
       " ('подогревает', 1),\n",
       " ('зонный', 1),\n",
       " ('климат', 1),\n",
       " ('расположены', 1),\n",
       " ('бардачке', 1),\n",
       " ('оркестр', 1),\n",
       " ('правой', 1),\n",
       " ('панеле', 1),\n",
       " ('подсветкой', 1),\n",
       " ('приносит', 1),\n",
       " ('дискомфорта', 1),\n",
       " ('что', 1),\n",
       " ('вмещается', 1),\n",
       " ('долго', 1),\n",
       " ('отличается', 1),\n",
       " ('просторностью', 1),\n",
       " ('коленки', 1),\n",
       " ('спина', 1),\n",
       " ('лихвой', 1),\n",
       " ('хватит', 1),\n",
       " ('совсем', 1),\n",
       " ('комфортно', 1),\n",
       " ('багаж', 1),\n",
       " ('создает', 1),\n",
       " ('раздражающих', 1),\n",
       " ('звуков', 1),\n",
       " ('бензином', 1),\n",
       " ('включить', 1),\n",
       " ('печку', 1),\n",
       " ('машина', 1),\n",
       " ('дверки', 1),\n",
       " ('кнопки', 1),\n",
       " ('переключения', 1),\n",
       " ('система', 1),\n",
       " ('гарнитурой', 1),\n",
       " ('планка', 1),\n",
       " ('под', 1),\n",
       " ('рукой', 1),\n",
       " ('навязчиво', 1),\n",
       " ('опель', 1),\n",
       " ('автомагнитола', 1),\n",
       " ('водить', 1),\n",
       " ('регулируются', 1),\n",
       " ('комплектацию', 1),\n",
       " ('фары', 1),\n",
       " ('кожа', 1),\n",
       " ('внутри', 1),\n",
       " ('холодно', 1),\n",
       " ('ноги', 1),\n",
       " ('s', 1),\n",
       " ('магнитола', 1),\n",
       " ('сd', 1),\n",
       " ('диски', 1),\n",
       " ('стекла', 1),\n",
       " ('компьютер', 1),\n",
       " ('сидеть', 1),\n",
       " ('греет', 1),\n",
       " ('орет', 1),\n",
       " ('ремонте', 1),\n",
       " ('дверь', 1),\n",
       " ('руля', 1),\n",
       " ('управляется', 1),\n",
       " ('тормоза', 1),\n",
       " ('-подъёмники', 1),\n",
       " ('свободно', 1),\n",
       " ('мягко', 1),\n",
       " ('сиденье', 1),\n",
       " ('магнитолой', 1),\n",
       " ('зеркала', 1),\n",
       " ('маленький', 1),\n",
       " ('контроль', 1),\n",
       " ('жарко', 1),\n",
       " ('холодный', 1),\n",
       " ('воздух', 1),\n",
       " ('замерзают', 1),\n",
       " ('тепловой', 1),\n",
       " ('удар', 1),\n",
       " ('арки', 1),\n",
       " ('pioneer', 1),\n",
       " ('аудио', 1),\n",
       " ('припарковаться', 1),\n",
       " ('кресло', 1),\n",
       " ('управлять', 1),\n",
       " ('шумов', 1),\n",
       " ('сидении', 1),\n",
       " ('кресел', 1),\n",
       " ('задних', 1),\n",
       " ('мало', 1),\n",
       " ('довольно', 1),\n",
       " ('мал', 1),\n",
       " ('как', 1),\n",
       " ('болтается', 1),\n",
       " ('бензин', 1),\n",
       " ('баке', 1),\n",
       " ('колесных', 1),\n",
       " ('отсеке', 1),\n",
       " ('тахометра', 1),\n",
       " ('упираюсь', 1),\n",
       " ('так', 1),\n",
       " ('много', 1),\n",
       " ('мульимедиа', 1),\n",
       " ('климатом', 1),\n",
       " ('мелочей', 1),\n",
       " ('плпстиком', 1),\n",
       " ('поддержка', 1),\n",
       " ('со', 1),\n",
       " ('двора', 1),\n",
       " ('шума', 1),\n",
       " ('вибраций', 1),\n",
       " ('ваши', 1),\n",
       " ('позвоночники', 1),\n",
       " ('размер', 1),\n",
       " ('дисплей', 1),\n",
       " ('обогревом', 1),\n",
       " ('\"калины\"', 1),\n",
       " ('при', 1),\n",
       " ('вождении', 1),\n",
       " ('чистится', 1),\n",
       " ('всех', 1),\n",
       " ('вещей', 1),\n",
       " ('есть', 1),\n",
       " ('место', 1),\n",
       " ('автомагнитолы', 1),\n",
       " ('небольшой', 1),\n",
       " ('-контроля', 1),\n",
       " ('гремит', 1),\n",
       " ('корпуса', 1),\n",
       " ('автомобилю', 1),\n",
       " ('дверей', 1),\n",
       " ('полная', 1),\n",
       " ('тишина', 1),\n",
       " ('kangoo', 1),\n",
       " ('удобны', 1),\n",
       " ('ремни', 1),\n",
       " ('присутствует', 1),\n",
       " ('помещается', 1),\n",
       " ('поддержки', 1),\n",
       " ('спины', 1),\n",
       " ('круиз', 1),\n",
       " ('контороли', 1),\n",
       " ('багажного', 1),\n",
       " ('отдела', 1)]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_counter['I-Comfort'].most_common()#[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TagConverter():\n",
    "    def __init__(self, tags):\n",
    "        self.idx_to_tag  = sorted(tags)\n",
    "        self.tag_to_idx  = {tag:idx for idx,tag in enumerate(self.idx_to_tag)}\n",
    "        \n",
    "    def tags_to_index(self, words):\n",
    "        return torch.tensor([self.tag_to_idx[w] for w in words], dtype=torch.long)\n",
    "    \n",
    "    def indices_to_tags(self, indices):\n",
    "        return [self.idx_to_tag[i] for i in indices]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "tag_converter = TagConverter(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">автомобиль<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">принадлежит дяде .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('автомобиль принадлежит дяде .',\n",
       "          [Span(0, 10, 'Whole'), Span(11, 29, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([ 6, 14, 14, 14])\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">автомобиль<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">принадлежит дяде .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('автомобиль принадлежит дяде .',\n",
       "          [Span(0, 10, 'Whole'), Span(11, 29, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "test_recipe, test_tags = training_data[321]\n",
    "\n",
    "# оригинал\n",
    "text, spans = generate_markup(test_recipe, test_tags) \n",
    "show_markup(text, spans)\n",
    "\n",
    "# кодируем только тэги:\n",
    "encoded_tags = tag_converter.tags_to_index(test_tags)\n",
    "print(encoded_tags)\n",
    "print()\n",
    "\n",
    "# декодируем тэги:\n",
    "decoded_tags = tag_converter.indices_to_tags(encoded_tags)\n",
    "\n",
    "text, spans = generate_markup(test_recipe, decoded_tags) \n",
    "show_markup(text, spans)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## fastText: забираем функции embedding слоя"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import zipfile\n",
    "import gensim\n",
    "import wget"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_url = 'http://vectors.nlpl.eu/repository/11/187.zip'\n",
    "wget.download(model_url)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### модель fasttext: редкие и  выдуманные слова"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "with zipfile.ZipFile('187.zip', 'r') as zip_ref:\n",
    "    zip_ref.extractall('datasets/187/')\n",
    "    # Colab:\n",
    "    # zip_ref.extractall('./stepik-dl-nlp/datasets/187/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "w2v_model = gensim.models.KeyedVectors.load('datasets/187/model.model')\n",
    "\n",
    "# Colab:\n",
    "# w2v_model = gensim.models.KeyedVectors.load('./stepik-dl-nlp/datasets/187/model.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "тачаночка\n",
      "тачанка 0.7318679094314575\n",
      "тачанк 0.6545170545578003\n",
      "бричка 0.627240777015686\n",
      "двуколка 0.610139012336731\n",
      "тачка 0.5886926054954529\n",
      "бараночка 0.5879709720611572\n",
      "паночка 0.573016881942749\n",
      "эмка 0.5701597929000854\n",
      "саничка 0.565357506275177\n",
      "каретка 0.5643386244773865\n",
      "\n",
      "\n",
      "двиганчик\n",
      "движка 0.4994381070137024\n",
      "танчик 0.4963422417640686\n",
      "двигай 0.4941864609718323\n",
      "двигатели 0.48835229873657227\n",
      "движа 0.4803929924964905\n",
      "двигун 0.46693849563598633\n",
      "движок 0.4630592465400696\n",
      "моторчик 0.45433372259140015\n",
      "шарк 0.4522787034511566\n",
      "движ 0.452262818813324\n",
      "\n",
      "\n",
      "ковервертолет\n",
      "вертолет 0.743363618850708\n",
      "ковер-самолет 0.6644095778465271\n",
      "вертолетная 0.6616967916488647\n",
      "вертолетчика 0.6364192962646484\n",
      "вертолёт 0.6328296065330505\n",
      "самолет 0.6245843768119812\n",
      "ковре-самолет 0.6201825141906738\n",
      "вертолетчик 0.6189783811569214\n",
      "гидросамолет 0.6109312176704407\n",
      "вертолетный 0.6071040630340576\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "words = ['тачаночка', 'двиганчик', 'ковервертолет']\n",
    "\n",
    "for word in words:\n",
    "    print(word)\n",
    "\n",
    "    for i in w2v_model.most_similar(positive=[word], topn=10):\n",
    "        nearest_word      = i[0]\n",
    "        cosine_similarity = i[1]\n",
    "        print(nearest_word, cosine_similarity)\n",
    "    print('\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Создадим конвертер слов в векторы fasttext:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "class WordConverter:\n",
    "    def __init__(self, w2v_model):\n",
    "        self.w2v_model = w2v_model\n",
    "\n",
    "    def convert(self, words):\n",
    "        embeddings = [self.w2v_model.get_vector(word) for word in words]\n",
    "        embeddings = torch.FloatTensor(embeddings)\n",
    "        return embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([27, 300])"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_converter = WordConverter(w2v_model)\n",
    "\n",
    "words,_ = training_data[333]\n",
    "vectors = word_converter.convert(words)\n",
    "\n",
    "vectors.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##  LSTM: теперь без embedding слоя"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "EMBEDDING_DIM = 300\n",
    "HIDDEN_DIM    = 32\n",
    "TAGSET_SIZE   = len(tag_converter.tag_to_idx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LSTMTagger(nn.Module):\n",
    "    def __init__(self, embedding_dim, hidden_dim, output_dim, word_converter):\n",
    "        super(LSTMTagger, self).__init__()\n",
    "        tagset_size = len(tag_converter.tag_to_idx)\n",
    "        \n",
    "        self.word_converter = word_converter\n",
    "        self.lstm           = nn.LSTM(embedding_dim, hidden_dim, bidirectional=True)\n",
    "        self.hidden2tag     = nn.Linear(2*hidden_dim, output_dim)\n",
    "\n",
    "    def forward(self, words):\n",
    "        embeds      = self.word_converter.convert(words)\n",
    "        lstm_out, _ = self.lstm(embeds.view(len(words), 1, -1))\n",
    "        tag_space   = self.hidden2tag(lstm_out.view(len(words), -1))\n",
    "        tag_scores  = F.log_softmax(tag_space, dim=1)\n",
    "        return tag_scores\n",
    "    \n",
    "    def predict_tags(self, words):\n",
    "        with torch.no_grad():\n",
    "            tags_pred = model(words).numpy()\n",
    "            tags_pred = np.argmax(tags_pred, axis=1)  \n",
    "        return tags_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "103\tin I-Appearance\n",
      "49\tin I-Safety\n",
      "284\tin I-Comfort\n",
      "169\tin B-Costs\n",
      "200\tin I-Whole\n",
      "97\tin I-Costs\n",
      "464\tin B-Reliability\n",
      "60\tin B-Safety\n",
      "236\tin B-Whole\n",
      "189\tin B-Appearance\n",
      "393\tin I-Driveability\n",
      "300\tin I-Reliability\n",
      "584\tin B-Driveability\n",
      "4933\tin Other\n",
      "478\tin B-Comfort\n"
     ]
    }
   ],
   "source": [
    "for label in labels:\n",
    "    num_words = len(word_counter[label])\n",
    "    print(f'{num_words}\\tin {label}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('Other', 4933),\n",
      " ('B-Driveability', 584),\n",
      " ('B-Comfort', 478),\n",
      " ('B-Reliability', 464),\n",
      " ('I-Driveability', 393),\n",
      " ('I-Reliability', 300),\n",
      " ('I-Comfort', 284),\n",
      " ('B-Whole', 236),\n",
      " ('I-Whole', 200),\n",
      " ('B-Appearance', 189),\n",
      " ('B-Costs', 169),\n",
      " ('I-Appearance', 103),\n",
      " ('I-Costs', 97),\n",
      " ('B-Safety', 60),\n",
      " ('I-Safety', 49)]\n",
      "tensor([ 26.7865,  10.5913,  29.9565,   8.6689,  10.9109,  84.3775,  21.4519,\n",
      "         49.1519,  17.8262,  52.1923,  12.8821,  16.8755, 103.3194,  25.3133,\n",
      "          1.0263])\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter\n",
    "from pprint import pprint\n",
    "\n",
    "tag_counter = Counter()\n",
    "for label in labels:\n",
    "    tag_counter[label] = len(word_counter[label])\n",
    "        \n",
    "pprint(tag_counter.most_common())\n",
    "\n",
    "\n",
    "class_weights = torch.ones(15)\n",
    "class_divs    = torch.ones(15)\n",
    "\n",
    "for tag, inv_weight in tag_counter.most_common():\n",
    "    tag_idx             = tag_converter.tags_to_index([tag])\n",
    "    class_divs[tag_idx] = inv_weight\n",
    "    \n",
    "norm       = torch.norm(class_divs, p=2, dim=0).detach()\n",
    "class_divs = class_divs.div(norm.expand_as(class_divs))\n",
    "\n",
    "class_weights /= class_divs\n",
    "\n",
    "print(class_weights.detach())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "model         = LSTMTagger(EMBEDDING_DIM, HIDDEN_DIM, TAGSET_SIZE, word_converter)\n",
    "loss_function = nn.NLLLoss(class_weights) \n",
    "optimizer     = optim.SGD(model.parameters(), lr=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "negative log likelihood loss:\n",
      "training   (min:    0.002, max:    3.925, cur:    0.425)\n"
     ]
    }
   ],
   "source": [
    "from livelossplot import PlotLosses\n",
    "\n",
    "liveplot = PlotLosses()\n",
    "\n",
    "for epoch in range(5): \n",
    "    for i, (recipe, tags) in enumerate(training_data):\n",
    "        \n",
    "        model.zero_grad()\n",
    "        \n",
    "        tag_predicted = model(recipe) # <- теперь передаем предложение, а не список индексов слов\n",
    "        tag_ground    = tag_converter.tags_to_index(tags) \n",
    "        \n",
    "        \n",
    "        loss = loss_function(tag_predicted, tag_ground)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if i % 100 == 0:\n",
    "            liveplot.update({'negative log likelihood loss': loss})\n",
    "            liveplot.draw()   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_tags(model, recipe):\n",
    "    encoded_tags   = model.predict_tags(recipe)      # предсказанные тэги (номера)\n",
    "    decoded_tags   = tag_converter.indices_to_tags(encoded_tags) # номер тэга -> тэг\n",
    "    return decoded_tags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">но это не значит , что<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">не подходит мужской половине .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('но это не значит , что машина не подходит мужской половине .',\n",
       "          [Span(0, 22, 'Other'), Span(23, 29, 'Whole'), Span(30, 60, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">но это не значит , что<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">не подходит<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">мужской<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">половине .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('но это не значит , что машина не подходит мужской половине .',\n",
       "          [Span(0, 22, 'Other'),\n",
       "           Span(23, 29, 'Whole'),\n",
       "           Span(30, 41, 'Other'),\n",
       "           Span(42, 49, 'Appearance'),\n",
       "           Span(50, 60, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">заплатил 1250<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">гривен<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">за<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">ремонт<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('заплатил 1250 гривен за ремонт',\n",
       "          [Span(0, 13, 'Other'),\n",
       "           Span(14, 20, 'Costs'),\n",
       "           Span(21, 23, 'Other'),\n",
       "           Span(24, 30, 'Costs')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">заплатил 1250 гривен<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">за<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">ремонт<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('заплатил 1250 гривен за ремонт',\n",
       "          [Span(0, 20, 'Costs'),\n",
       "           Span(21, 23, 'Other'),\n",
       "           Span(24, 30, 'Reliability')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">вообще ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">под мой характер -<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">педаль газа<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">как сенсорная , только прикоснулся , она<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">рвет с места<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, а<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">тормоз<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">надо немного больше нажать .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('вообще , машина под мой характер - педаль газа как сенсорная , только прикоснулся , она рвет с места , а тормоз надо немного больше нажать .',\n",
       "          [Span(0, 8, 'Other'),\n",
       "           Span(9, 15, 'Driveability'),\n",
       "           Span(16, 34, 'Other'),\n",
       "           Span(35, 46, 'Driveability'),\n",
       "           Span(47, 87, 'Other'),\n",
       "           Span(88, 100, 'Driveability'),\n",
       "           Span(101, 104, 'Other'),\n",
       "           Span(105, 111, 'Driveability'),\n",
       "           Span(112, 140, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">вообще ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">под мой характер -<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">педаль<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">газа<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">как<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">сенсорная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, только прикоснулся , она<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">рвет<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">с<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">места<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, а<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">тормоз<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">надо немного больше<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">нажать<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">.<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('вообще , машина под мой характер - педаль газа как сенсорная , только прикоснулся , она рвет с места , а тормоз надо немного больше нажать .',\n",
       "          [Span(0, 8, 'Other'),\n",
       "           Span(9, 15, 'Whole'),\n",
       "           Span(16, 34, 'Other'),\n",
       "           Span(35, 41, 'Reliability'),\n",
       "           Span(42, 46, 'Whole'),\n",
       "           Span(47, 50, 'Other'),\n",
       "           Span(51, 60, 'Comfort'),\n",
       "           Span(61, 87, 'Other'),\n",
       "           Span(88, 92, 'Driveability'),\n",
       "           Span(93, 94, 'Other'),\n",
       "           Span(95, 100, 'Reliability'),\n",
       "           Span(101, 104, 'Other'),\n",
       "           Span(105, 111, 'Driveability'),\n",
       "           Span(112, 131, 'Other'),\n",
       "           Span(132, 138, 'Reliability'),\n",
       "           Span(139, 140, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">через два месяца<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">полетела подвеска<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">ремонт<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">обошелся в 25000<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">руб<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('через два месяца полетела подвеска , ремонт обошелся в 25000 руб',\n",
       "          [Span(0, 16, 'Other'),\n",
       "           Span(17, 34, 'Reliability'),\n",
       "           Span(35, 36, 'Other'),\n",
       "           Span(37, 43, 'Costs'),\n",
       "           Span(44, 60, 'Other'),\n",
       "           Span(61, 64, 'Costs')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">через<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">два месяца полетела подвеска<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">ремонт<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">обошелся в 25000<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">руб<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('через два месяца полетела подвеска , ремонт обошелся в 25000 руб',\n",
       "          [Span(0, 5, 'Appearance'),\n",
       "           Span(6, 34, 'Reliability'),\n",
       "           Span(35, 36, 'Other'),\n",
       "           Span(37, 43, 'Costs'),\n",
       "           Span(44, 60, 'Other'),\n",
       "           Span(61, 64, 'Costs')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">уже на следующий день я стала счастливым обладателем<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">гранты<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('уже на следующий день я стала счастливым обладателем гранты',\n",
       "          [Span(0, 52, 'Other'), Span(53, 59, 'Whole')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">уже на следующий день я стала счастливым обладателем гранты<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('уже на следующий день я стала счастливым обладателем гранты',\n",
       "          [Span(0, 59, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">а так в другом , все отлично ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">красивая<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, так сказать ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">элегантная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">надежная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('а так в другом , все отлично , красивая , так сказать , элегантная , надежная',\n",
       "          [Span(0, 30, 'Other'),\n",
       "           Span(31, 39, 'Appearance'),\n",
       "           Span(40, 55, 'Other'),\n",
       "           Span(56, 66, 'Appearance'),\n",
       "           Span(67, 68, 'Other'),\n",
       "           Span(69, 77, 'Reliability')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">а так в другом , все отлично ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">красивая<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, так сказать ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">элегантная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">надежная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('а так в другом , все отлично , красивая , так сказать , элегантная , надежная',\n",
       "          [Span(0, 30, 'Other'),\n",
       "           Span(31, 39, 'Appearance'),\n",
       "           Span(40, 55, 'Other'),\n",
       "           Span(56, 66, 'Appearance'),\n",
       "           Span(67, 68, 'Other'),\n",
       "           Span(69, 77, 'Reliability')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">по<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">двигателю<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">такого не скажу .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('по двигателю такого не скажу .',\n",
       "          [Span(0, 2, 'Other'),\n",
       "           Span(3, 12, 'Reliability'),\n",
       "           Span(13, 30, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">по<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">двигателю<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">такого не скажу .<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('по двигателю такого не скажу .',\n",
       "          [Span(0, 2, 'Other'),\n",
       "           Span(3, 12, 'Driveability'),\n",
       "           Span(13, 30, 'Other')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">точнее его отсутствии , он настолько маленький что его ну ни как не назвать<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">багажником<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('точнее его отсутствии , он настолько маленький что его ну ни как не назвать багажником',\n",
       "          [Span(0, 75, 'Other'), Span(76, 86, 'Comfort')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">точнее его отсутствии , он настолько маленький что его ну ни как<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">не<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">назвать<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">багажником<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('точнее его отсутствии , он настолько маленький что его ну ни как не назвать багажником',\n",
       "          [Span(0, 64, 'Other'),\n",
       "           Span(65, 67, 'Reliability'),\n",
       "           Span(68, 75, 'Other'),\n",
       "           Span(76, 86, 'Comfort')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">ужасная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">шумоизоляция<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, всё время<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">скрипят панели приборов<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, слабая<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">ходовая часть<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">печка в салоне<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">проработала по пробегу 40 000 км , из -за некачественного топлива на заправках<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">свечи накрываются<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">через каждые 10 000 км ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">машина дорогая<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">в<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">техобслуживании<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Costs</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('ужасная шумоизоляция , всё время скрипят панели приборов , слабая ходовая часть , печка в салоне проработала по пробегу 40 000 км , из -за некачественного топлива на заправках свечи накрываются через каждые 10 000 км , машина дорогая в техобслуживании',\n",
       "          [Span(0, 7, 'Other'),\n",
       "           Span(8, 20, 'Comfort'),\n",
       "           Span(21, 32, 'Other'),\n",
       "           Span(33, 56, 'Comfort'),\n",
       "           Span(57, 65, 'Other'),\n",
       "           Span(66, 79, 'Driveability'),\n",
       "           Span(80, 81, 'Other'),\n",
       "           Span(82, 96, 'Reliability'),\n",
       "           Span(97, 175, 'Other'),\n",
       "           Span(176, 193, 'Reliability'),\n",
       "           Span(194, 218, 'Other'),\n",
       "           Span(219, 233, 'Costs'),\n",
       "           Span(234, 235, 'Other'),\n",
       "           Span(236, 251, 'Costs')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">ужасная<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">шумоизоляция<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">, всё время<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">скрипят<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">панели<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">приборов<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">, слабая ходовая часть ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">печка<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">в салоне проработала по пробегу 40 000 км , из -за<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">некачественного<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">топлива<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">на<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">заправках свечи<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">накрываются<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">через каждые<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">10 000 км ,<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">дорогая<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Appearance</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">в<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">техобслуживании<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('ужасная шумоизоляция , всё время скрипят панели приборов , слабая ходовая часть , печка в салоне проработала по пробегу 40 000 км , из -за некачественного топлива на заправках свечи накрываются через каждые 10 000 км , машина дорогая в техобслуживании',\n",
       "          [Span(0, 7, 'Other'),\n",
       "           Span(8, 20, 'Comfort'),\n",
       "           Span(21, 32, 'Other'),\n",
       "           Span(33, 40, 'Comfort'),\n",
       "           Span(41, 47, 'Reliability'),\n",
       "           Span(48, 56, 'Comfort'),\n",
       "           Span(57, 81, 'Reliability'),\n",
       "           Span(82, 87, 'Comfort'),\n",
       "           Span(88, 138, 'Other'),\n",
       "           Span(139, 154, 'Reliability'),\n",
       "           Span(155, 162, 'Comfort'),\n",
       "           Span(163, 165, 'Other'),\n",
       "           Span(166, 181, 'Comfort'),\n",
       "           Span(182, 193, 'Reliability'),\n",
       "           Span(194, 206, 'Driveability'),\n",
       "           Span(207, 218, 'Other'),\n",
       "           Span(219, 225, 'Driveability'),\n",
       "           Span(226, 233, 'Appearance'),\n",
       "           Span(234, 235, 'Other'),\n",
       "           Span(236, 251, 'Comfort')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "истинные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">никаких повреждений<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">(даже<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">пыльников<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">) не было выявлено !<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">уверенно<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">держит дорогу<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">и обладает всеми необходимыми качествами !<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">салон<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">достаточно<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">просторный<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('никаких повреждений (даже пыльников ) не было выявлено ! машина уверенно держит дорогу и обладает всеми необходимыми качествами ! салон достаточно просторный',\n",
       "          [Span(0, 19, 'Reliability'),\n",
       "           Span(20, 25, 'Other'),\n",
       "           Span(26, 35, 'Reliability'),\n",
       "           Span(36, 56, 'Other'),\n",
       "           Span(57, 63, 'Driveability'),\n",
       "           Span(64, 72, 'Other'),\n",
       "           Span(73, 86, 'Driveability'),\n",
       "           Span(87, 129, 'Other'),\n",
       "           Span(130, 135, 'Comfort'),\n",
       "           Span(136, 146, 'Other'),\n",
       "           Span(147, 157, 'Comfort')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "предсказанные тэги:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div class=\"tex2jax_ignore\" style=\"white-space: pre-wrap\"><span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">никаких повреждений<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">(даже<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">пыльников<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">) не было выявлено !<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffcdd2; background: #ffebee\">машина<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #e57373;\">Whole</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">уверенно<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #bbdefb; background: #e3f2fd\">держит дорогу<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #64b5f6;\">Driveability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">и обладает всеми необходимыми<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #c8e6c9; background: #e8f5e9\">качествами<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #66bb6a;\">Reliability</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">!<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">салон<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #d7ccc8; background: #efebe9\">достаточно<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #a1887f;\">Other</span></span> <span style=\"padding: 2px; border-radius: 4px; border: 1px solid #ffe0b2; background: #fff3e0\">просторный<span style=\"vertical-align: middle; margin-left: 2px; font-size: 0.7em; color: #ffb74d;\">Comfort</span></span></div>"
      ],
      "text/plain": [
       "BoxMarkup('никаких повреждений (даже пыльников ) не было выявлено ! машина уверенно держит дорогу и обладает всеми необходимыми качествами ! салон достаточно просторный',\n",
       "          [Span(0, 19, 'Reliability'),\n",
       "           Span(20, 25, 'Other'),\n",
       "           Span(26, 35, 'Reliability'),\n",
       "           Span(36, 56, 'Other'),\n",
       "           Span(57, 63, 'Whole'),\n",
       "           Span(64, 72, 'Other'),\n",
       "           Span(73, 86, 'Driveability'),\n",
       "           Span(87, 116, 'Other'),\n",
       "           Span(117, 127, 'Reliability'),\n",
       "           Span(128, 129, 'Other'),\n",
       "           Span(130, 135, 'Comfort'),\n",
       "           Span(136, 146, 'Other'),\n",
       "           Span(147, 157, 'Comfort')])"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for i in range(0,10):\n",
    "\n",
    "    recipe, tags = test_data[np.random.randint(0,1000)]\n",
    "    \n",
    "    tags_pred    = predict_tags(model, recipe)\n",
    "\n",
    "    print('истинные тэги:')\n",
    "    text, spans = generate_markup(recipe, tags) \n",
    "    show_markup(text, spans)\n",
    "\n",
    "    print('предсказанные тэги:')\n",
    "    text, spans = generate_markup(recipe, tags_pred) \n",
    "\n",
    "    show_markup(text, spans)\n",
    "    print()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from dlnlputils.sentiment_utils import plot_confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = []\n",
    "y_true = []\n",
    "\n",
    "for sentence, tags in test_data:\n",
    "    y_pred += predict_tags(model, sentence)\n",
    "    y_true += tags"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "ordered_tags = [\n",
    " 'Other',\n",
    " 'B-Whole',\n",
    " 'I-Whole',\n",
    " 'B-Driveability',\n",
    " 'I-Driveability',\n",
    " 'B-Appearance',\n",
    " 'I-Appearance',\n",
    " 'B-Reliability',\n",
    " 'I-Reliability',\n",
    " 'B-Comfort',\n",
    " 'I-Comfort',\n",
    " 'B-Costs',\n",
    " 'I-Costs',\n",
    " 'B-Safety',\n",
    " 'I-Safety']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(y_true, y_pred, classes=ordered_tags, normalize=True, \n",
    "                      title='Нормализованная матрица ошибок')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}