sim_and_analogies_comp.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*План на 2 апреля:*\n",
    "\n",
    "    *посмотреть на данные аналогий, сравнить пересечения словарей*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import EvaluateEmbeddingsLab as lab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualise_sim_dataset(data, size=(20, 20)):\n",
    "    data_vocab, data_labels = data.get_all_words_and_labels()\n",
    "    \n",
    "    matr = np.zeros((len(data_vocab), len(data_vocab)))\n",
    "\n",
    "    for pair in list(data.data.data_similarity.keys()):\n",
    "        word1, word2 = pair[0], pair[1]\n",
    "        # value = data.data.data_similarity[pair]\n",
    "        word1_idx = data_vocab.index(word1)\n",
    "        word2_idx = data_vocab.index(word2)\n",
    "\n",
    "        matr[word1_idx, word2_idx] = 1\n",
    "        matr[word2_idx, word1_idx] = 1 # value\n",
    "        \n",
    "    plt.figure(figsize=size)\n",
    "    plt.imshow(matr, interpolation='none', cmap='gray')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "simlex = lab.GoldenStandartDataset(\"./SimLex-999-standartized.csv\")\n",
    "men = lab.GoldenStandartDataset(\"./MEN-standartized.csv\")\n",
    "wordsim = lab.GoldenStandartDataset(\"./WordSim-353-standartized.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*В визуализации считаем матрицу симметричной, хотя в датасеты пары входят без симметрии.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Визуально MEN хорош, в нем больше всего концентрация пар.*\n",
    "*Диагональ на визуализации обусловлена принципом построения матрицы - если оба слова в паре (a,b) не встретились до этого, то в список уникальных слов они попадут друг за другом [..., a, b, ...]. Т.е. это диагональ смещенная на 1.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualise_sim_dataset(men, size=(10, 10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualise_sim_dataset(wordsim, size=(10, 10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Тут у simlex хорошо видны три блока - в каждом блоке пары из слов одной части речи.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualise_sim_dataset(simlex, size=(10, 10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Сравним словари.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "men_vocab, men_labels = men.get_all_words_and_labels()\n",
    "wordsim_vocab, wordsim_labels = wordsim.get_all_words_and_labels()\n",
    "simlex_vocab, simlex_labels = simlex.get_all_words_and_labels()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "размеры словарей\n",
      "MEN 751\n",
      "WordSim 437\n",
      "SimLex 1028\n"
     ]
    }
   ],
   "source": [
    "print(\"размеры словарей\")\n",
    "print(\"MEN\", len(men_vocab))\n",
    "print(\"WordSim\", len(wordsim_vocab))\n",
    "print(\"SimLex\", len(simlex_vocab))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "размеры датасетов\n",
      "MEN 3000\n",
      "WordSim 351\n",
      "SimLex 998\n"
     ]
    }
   ],
   "source": [
    "print(\"размеры датасетов\")\n",
    "print(\"MEN\", len(men.data.data_similarity.keys()))\n",
    "print(\"WordSim\", len(wordsim.data.data_similarity.keys()))\n",
    "print(\"SimLex\", len(simlex.data.data_similarity.keys()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "men_vocab_set = set(men_vocab)\n",
    "wordsim_vocab_set = set(wordsim_vocab)\n",
    "simlex_vocab_set = set(simlex_vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "пересечения словарей\n",
      "все три 52\n",
      "men & wordsim 98\n",
      "men & simlex 227\n",
      "simlex & wordsim 124\n"
     ]
    }
   ],
   "source": [
    "print(\"пересечения словарей\")\n",
    "print(\"все три\", len(set.intersection(men_vocab_set, wordsim_vocab_set, simlex_vocab_set)))\n",
    "print(\"men & wordsim\", len(set.intersection(men_vocab_set, wordsim_vocab_set)))\n",
    "print(\"men & simlex\", len(set.intersection(men_vocab_set, simlex_vocab_set)))\n",
    "print(\"simlex & wordsim\", len(set.intersection(simlex_vocab_set, wordsim_vocab_set)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Неплохие пересечения относительно размеров словарей.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Сравним шкалы:*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "минимальные и максимальные значения датасетов\n",
      "MEN 0.0 50.0\n",
      "WordSim 0.23 10.0\n",
      "SimLex 0.23 9.8\n"
     ]
    }
   ],
   "source": [
    "print(\"минимальные и максимальные значения датасетов\")\n",
    "print(\"MEN\", min(list(men.data.data_similarity.values())), max(list(men.data.data_similarity.values())))\n",
    "print(\"WordSim\", min(list(wordsim.data.data_similarity.values())), max(list(wordsim.data.data_similarity.values())))\n",
    "print(\"SimLex\", min(list(simlex.data.data_similarity.values())), max(list(simlex.data.data_similarity.values())))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Посмотрим, на каких парах минимальные значения*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_min_pair(data):\n",
    "    for pair in data.data.data_similarity.keys():\n",
    "        min_val = min(list(data.data.data_similarity.values()))\n",
    "        if data.data.data_similarity[pair] == min_val:\n",
    "            print(pair.to_string())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "king,cabbage\n"
     ]
    }
   ],
   "source": [
    "find_min_pair(wordsim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "new,ancient\n",
      "shrink,grow\n"
     ]
    }
   ],
   "source": [
    "find_min_pair(simlex)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "bakery,zebra\n"
     ]
    }
   ],
   "source": [
    "find_min_pair(men)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Приведем к одной шкале и попробуем посмотреть на объединения датасетов (wordsim и men объединять законно - там relatedness, у simlex - similarity, но тут спорно, создатели men считают similarity частным случаем relatedness, так что можем оставить такое допущение)*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "for pair in men.data.data_similarity.keys():\n",
    "    men.data.data_similarity[pair] /= 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Важно: выбрать один или несколько способов преобразования сходств в различия.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "common_vocab_set = set.union(men_vocab_set, wordsim_vocab_set, simlex_vocab_set)\n",
    "common_vocab = list(common_vocab_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1819"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(common_vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualise_sim_data_list(data_list, data_vocab, size=(20, 20)):\n",
    "    matr = np.zeros((len(data_vocab), len(data_vocab)))\n",
    "    \n",
    "    for data in data_list:\n",
    "    \n",
    "        #data_vocab, data_labels = data.get_all_words_and_labels() \n",
    "\n",
    "        for pair in list(data.data.data_similarity.keys()):\n",
    "            word1, word2 = pair[0], pair[1]\n",
    "            # value = data.data.data_similarity[pair]\n",
    "            word1_idx = data_vocab.index(word1)\n",
    "            word2_idx = data_vocab.index(word2)\n",
    "\n",
    "            matr[word1_idx, word2_idx] = 1\n",
    "            matr[word2_idx, word1_idx] = 1 # value\n",
    "        \n",
    "    plt.figure(figsize=size)\n",
    "    plt.imshow(matr, interpolation='none', cmap='gray')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1800 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualise_sim_data_list([simlex, men, wordsim], common_vocab, size=(25, 25))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Теперь посмотрим на аналогии.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "prefix = \"./BATS_3.0/\"\n",
    "directory = \"3_Encyclopedic_semantics\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['1_Inflectional_morphology',\n",
       " '2_Derivational_morphology',\n",
       " '3_Encyclopedic_semantics',\n",
       " '4_Lexicographic_semantics']"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "directories = sorted(os.listdir(prefix))\n",
    "directories[1:-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['E01 [country - capital].txt',\n",
       " 'E02 [country - language].txt',\n",
       " 'E03 [UK_city - county].txt',\n",
       " 'E04 [name - nationality].txt',\n",
       " 'E05 [name - occupation].txt',\n",
       " 'E06 [animal - young].txt',\n",
       " 'E07 [animal - sound].txt',\n",
       " 'E08 [animal - shelter].txt',\n",
       " 'E09 [things - color].txt',\n",
       " 'E10 [male - female].txt']"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "files = sorted(os.listdir(prefix + directory))\n",
    "files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_analogy(filepath):\n",
    "    vocab = []\n",
    "    output = []\n",
    "    with open(filepath, newline='\\n') as file:\n",
    "        for string in file:\n",
    "            pairlike = string[:-1]\n",
    "            word, shifted_words = pairlike.split()\n",
    "            shifted_words = shifted_words.split('/')\n",
    "            output.append([word, shifted_words])\n",
    "            vocab.append(word)\n",
    "            for word in shifted_words:\n",
    "                vocab.append(word)\n",
    "    return output, vocab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compare_vocabs(vocab):\n",
    "    print(\"men\", len(set.intersection(men_vocab_set, vocab)))\n",
    "    print(\"simlex\", len(set.intersection(vocab, simlex_vocab_set)))\n",
    "    print(\"wordsim\", len(set.intersection(vocab, wordsim_vocab_set)))\n",
    "    print(\"common\", len(set.intersection(vocab, common_vocab_set)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3_Encyclopedic_semantics\n",
      "\n",
      "country - capital\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 0\n",
      "men 0\n",
      "simlex 1\n",
      "wordsim 0\n",
      "common 1\n",
      "\n",
      "\n",
      "country - language\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 0\n",
      "men 0\n",
      "simlex 0\n",
      "wordsim 0\n",
      "common 0\n",
      "\n",
      "\n",
      "UK_city - county\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 0\n",
      "men 2\n",
      "simlex 1\n",
      "wordsim 0\n",
      "common 2\n",
      "\n",
      "\n",
      "name - nationality\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 0\n",
      "men 0\n",
      "simlex 0\n",
      "wordsim 0\n",
      "common 0\n",
      "\n",
      "\n",
      "name - occupation\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 0\n",
      "men 1\n",
      "simlex 9\n",
      "wordsim 5\n",
      "common 13\n",
      "\n",
      "\n",
      "animal - young\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 15\n",
      "men 33\n",
      "simlex 18\n",
      "wordsim 4\n",
      "common 39\n",
      "\n",
      "\n",
      "animal - sound\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 3\n",
      "men 27\n",
      "simlex 14\n",
      "wordsim 2\n",
      "common 33\n",
      "\n",
      "\n",
      "animal - shelter\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 22\n",
      "men 35\n",
      "simlex 22\n",
      "wordsim 3\n",
      "common 44\n",
      "\n",
      "\n",
      "things - color\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 28\n",
      "men 37\n",
      "simlex 18\n",
      "wordsim 8\n",
      "common 44\n",
      "\n",
      "\n",
      "male - female\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 15\n",
      "men 13\n",
      "simlex 37\n",
      "wordsim 11\n",
      "common 41\n",
      "\n",
      "4_Lexicographic_semantics\n",
      "\n",
      "hypernyms - animals\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 23\n",
      "men 30\n",
      "simlex 18\n",
      "wordsim 10\n",
      "common 39\n",
      "\n",
      "\n",
      "hypernyms - misc\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 19\n",
      "men 44\n",
      "simlex 38\n",
      "wordsim 24\n",
      "common 72\n",
      "\n",
      "\n",
      "hyponyms - misc\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 27\n",
      "men 120\n",
      "simlex 101\n",
      "wordsim 35\n",
      "common 189\n",
      "\n",
      "\n",
      "meronyms - substance\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 34\n",
      "men 49\n",
      "simlex 49\n",
      "wordsim 18\n",
      "common 69\n",
      "\n",
      "\n",
      "meronyms - member\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 23\n",
      "men 39\n",
      "simlex 44\n",
      "wordsim 26\n",
      "common 69\n",
      "\n",
      "\n",
      "meronyms - part\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 27\n",
      "men 83\n",
      "simlex 79\n",
      "wordsim 24\n",
      "common 132\n",
      "\n",
      "\n",
      "synonyms - intensity\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 10\n",
      "men 21\n",
      "simlex 32\n",
      "wordsim 7\n",
      "common 44\n",
      "\n",
      "\n",
      "synonyms - exact\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 21\n",
      "men 38\n",
      "simlex 35\n",
      "wordsim 22\n",
      "common 63\n",
      "\n",
      "\n",
      "antonyms - gradable\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 14\n",
      "men 20\n",
      "simlex 36\n",
      "wordsim 1\n",
      "common 51\n",
      "\n",
      "\n",
      "antonyms - binary\n",
      "Пар, из которых можно составлять аналогии, целиком из слов общего словаря 4\n",
      "men 8\n",
      "simlex 8\n",
      "wordsim 2\n",
      "common 17\n",
      "\n",
      "Всего пар (из которых потом составляются четверки), которые состоят из слов общего словаря 285\n"
     ]
    }
   ],
   "source": [
    "global_vocab = []\n",
    "global_both_counter = 0\n",
    "\n",
    "for directory in directories[3:-1]:\n",
    "    print(directory)\n",
    "    for file in sorted(os.listdir(prefix + directory)):\n",
    "        print('')\n",
    "        print(file[5:-5])\n",
    "        analogy, vocab = create_analogy(prefix + directory + \"/\" + file)\n",
    "        global_vocab += vocab\n",
    "        \n",
    "        words_counter = 0\n",
    "        both_counter = 0\n",
    "        for pairlike in analogy:\n",
    "            word = pairlike[0]\n",
    "            words = pairlike[1]\n",
    "            left_flag = False\n",
    "            right_flag = False\n",
    "            if word in common_vocab:\n",
    "                words_counter += 1\n",
    "                left_flag = True\n",
    "            for word in words:\n",
    "                if word in common_vocab:\n",
    "                    words_counter += 1\n",
    "                    right_flag = True\n",
    "            if left_flag and right_flag:\n",
    "                both_counter += 1\n",
    "        print(\"Пар, из которых можно составлять аналогии, целиком из слов общего словаря\", both_counter)\n",
    "        global_both_counter += both_counter\n",
    "        compare_vocabs(set(vocab))\n",
    "        print('')\n",
    "        \n",
    "print(\"Всего пар (из которых потом составляются четверки), которые состоят из слов общего словаря\", global_both_counter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "men 369\n",
      "simlex 391\n",
      "wordsim 141\n",
      "common 657\n"
     ]
    }
   ],
   "source": [
    "compare_vocabs(set(global_vocab))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1819"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(common_vocab)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*С объединенным датасетом больше трети слов пересекается со словарем датасета аналогий.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Вопрос в том, что нам нужно, чтобы хотя бы два слова из разных пар в аналогии были в близостях. Тогда мы достраиваем дополнительные слова. Если есть все четыре слова, то корректируем.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*Как выяснилось, ~285 пар состоят только из слов общего словаря, то есть потенциально корректируют уже имеющиеся представления без добавления новых слов. Можно работать*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*План на 3 апреля:*\n",
    "    \n",
    "    *Строгая постановка задачи и введение ВКР*\n",
    "    \n",
    "    *Сравнить полносвязный слой с основными методами MDS*"
   ]
  }
 ],
 "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": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}