Prereq_Outcomes.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Prereq_Outcomes.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "ZZ2WJrNwn6tt",
        "a1FDHRriAlEa",
        "Fg5NUDV3oFsg",
        "9l-KgqMmz9c_",
        "8sfWtD6v2euR",
        "p93st8hW8wYK",
        "wuIZ0hroTz0v",
        "aflDbtVHd-ci",
        "QS3o6hJKZLQW",
        "IJObEqvOsUP2",
        "D71GlrsArzHt",
        "cPzg40TEaheE",
        "IdpgLxwoxvZu",
        "_xsS6ZR1vka6"
      ],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "20kY0B5wDbRu"
      },
      "source": [
        "*   Регулярными выражениями взять название книги из списка лит-ры\n",
        "*   Брать названия онлайн курсов для дисциплины, чтобы тоже сделать эмбединг\n",
        "*   Мб если на раздел выделено больше часов, то больший приоритет отдавать схожим с ним уч.сущностям?? Ерунда?\n",
        "*   посмотреть -- если подходящая уч.сущность входит в другую, то ее тоже рекомендовать?.. что-то со связями придумать"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mGEv88lRg6ME",
        "cellView": "form"
      },
      "source": [
        "#@title Загрузка библиотек и данных\n",
        "# for TF-Hub universal encoder\n",
        "%%capture\n",
        "!pip install numpy==1.19.5\n",
        "!pip install PyDrive\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import itertools\n",
        "import seaborn as sns\n",
        "import re # только в черновике\n",
        "\n",
        "from pydrive.auth import GoogleAuth\n",
        "from pydrive.drive import GoogleDrive\n",
        "from google.colab import auth\n",
        "from oauth2client.client import GoogleCredentials\n",
        "\n",
        "# скачиваем данные\n",
        "import glob\n",
        "\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "path = '/content/drive/My Drive/bachelor/!Диплом/d/'\n",
        "all_files = glob.glob(path+'*.csv')\n",
        "\n",
        "dfs = []\n",
        "for f in all_files:\n",
        "  dfs.append(pd.read_csv(f,sep=';'))\n",
        "\n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZZ2WJrNwn6tt"
      },
      "source": [
        "# Присваиваем все таблицы"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a1FDHRriAlEa"
      },
      "source": [
        "## -- Для всего + рек-ии результатов"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PhZP9bKxj6WQ"
      },
      "source": [
        "dp_domains,dp_items,dp_relations,\\\n",
        "wp_aplan,wp_additional,wp_bib_ref,wp_disc_b,wp_disc_bm,\\\n",
        "wp_disc_s,wp_disc_seval,wp_eval,wp_fos,wp_iaplan,\\\n",
        "wp_iaplan_fos,wp_outcomes,wp_outcomes_eval,\\\n",
        "wp_prereq,wp_topic,wp_wp,wp_wpbib_ref,\\\n",
        "wp_wp_disc_bm,wp_wp_isu,wp_wp_fos,wp_cert_eval = dfs"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "0AzhPQeuFGiJ",
        "outputId": "e2a179e4-578c-4d56-924e-09c90ddd588d"
      },
      "source": [
        "# Все РПД\n",
        "wp_wp.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>title</th>\n",
              "      <th>hoursFirstSemester</th>\n",
              "      <th>hoursSecondSemester</th>\n",
              "      <th>qualification</th>\n",
              "      <th>discipline_code</th>\n",
              "      <th>approval_date</th>\n",
              "      <th>authors</th>\n",
              "      <th>description</th>\n",
              "      <th>video</th>\n",
              "      <th>subject_code</th>\n",
              "      <th>credit_units</th>\n",
              "      <th>semester_hour</th>\n",
              "      <th>owner_id</th>\n",
              "      <th>work_status</th>\n",
              "      <th>hours</th>\n",
              "      <th>extra_points</th>\n",
              "      <th>language</th>\n",
              "      <th>structural_unit_id</th>\n",
              "      <th>have_course_project</th>\n",
              "      <th>have_diff_pass</th>\n",
              "      <th>have_exam</th>\n",
              "      <th>have_pass</th>\n",
              "      <th>lab_hours</th>\n",
              "      <th>lecture_hours</th>\n",
              "      <th>old_discipline_code</th>\n",
              "      <th>practice_hours</th>\n",
              "      <th>srs_hours</th>\n",
              "      <th>wp_isu_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>3689</td>\n",
              "      <td>Пищевая химия</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>06.1.5.26.18</td>\n",
              "      <td>2020-08-04 17:55:37.324431+03</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>44</td>\n",
              "      <td>0,0,0,0,4,0,0,0,0,0,0,0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>w</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>11.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0,0.0,0.0,0.0,16.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,32.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>06.1.5.38.18</td>\n",
              "      <td>0.0,0.0,0.0,0.0,16.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,73.6,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>26155.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2648</td>\n",
              "      <td>Современный C++</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>06.0.6.19.18</td>\n",
              "      <td>2020-08-04 17:39:33.064108+03</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>14</td>\n",
              "      <td>0,0,3,0,0,0,0,0,0,0,0,0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>w</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>14.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,16.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>06.0.6.10.18</td>\n",
              "      <td>0.0,0.0,32.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,55.2,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>26505.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2686</td>\n",
              "      <td>Дифференциальные уравнения и ТФКП</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>06.5.6.62.18</td>\n",
              "      <td>2020-08-04 17:39:39.412693+03</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>40</td>\n",
              "      <td>0,0,0,0,4,0,0,0,0,0,0,0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>w</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>14.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0,0.0,0.0,0.0,16.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,32.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>06.5.6.4.18</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,91.19999999999999,0.0,0.0,0.0,...</td>\n",
              "      <td>35677.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3926</td>\n",
              "      <td>Основы конкурентоспособности предприятий</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>06.4.6.27.18</td>\n",
              "      <td>2020-08-04 17:56:38.312651+03</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>48</td>\n",
              "      <td>0,0,0,0,0,0,0,5,0,0,0,0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>w</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>1.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,16.0,0.0,0.0</td>\n",
              "      <td>06.4.6.27.18</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,32.0,0.0,0.0</td>\n",
              "      <td>0.0,0.0,0.0,0.0,0.0,0.0,0.0,127.2,0.0,0.0</td>\n",
              "      <td>26064.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     id  ... wp_isu_id\n",
              "0  3689  ...   26155.0\n",
              "1  2648  ...   26505.0\n",
              "2  2686  ...   35677.0\n",
              "3  3926  ...   26064.0\n",
              "\n",
              "[4 rows x 29 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "QcYVr_VH_Sko",
        "outputId": "c88ba4e2-9bc4-46df-ca1c-38e518b4bade"
      },
      "source": [
        "# разделы с распределением часов\n",
        "wp_disc_s.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>work_program_id</th>\n",
              "      <th>SRO</th>\n",
              "      <th>contact_work</th>\n",
              "      <th>laboratory</th>\n",
              "      <th>lecture_classes</th>\n",
              "      <th>practical_lessons</th>\n",
              "      <th>total_hours</th>\n",
              "      <th>ordinal_number</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>17</td>\n",
              "      <td>Теоретические основы метрологии</td>\n",
              "      <td>5</td>\n",
              "      <td>4.0</td>\n",
              "      <td>2.2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.0</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>18</td>\n",
              "      <td>Измерения, погрешности и неопределенности</td>\n",
              "      <td>5</td>\n",
              "      <td>10.0</td>\n",
              "      <td>13.2</td>\n",
              "      <td>8.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>20</td>\n",
              "      <td>Обработка результатов измерений</td>\n",
              "      <td>5</td>\n",
              "      <td>4.0</td>\n",
              "      <td>4.4</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>12.0</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1208</td>\n",
              "      <td>Особенности охлаждения сверхпроводниковых устр...</td>\n",
              "      <td>12522</td>\n",
              "      <td>25.2</td>\n",
              "      <td>4.4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>29.6</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     id  ... ordinal_number\n",
              "0    17  ...              1\n",
              "1    18  ...              2\n",
              "2    20  ...              4\n",
              "3  1208  ...              5\n",
              "\n",
              "[4 rows x 10 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "8D7I8SPT-O6f",
        "outputId": "a94430b7-fb6c-443a-8550-921e0f203a71"
      },
      "source": [
        "# темы РПД (\"содержание дисциплины\")\n",
        "wp_topic.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>number</th>\n",
              "      <th>description</th>\n",
              "      <th>discipline_section_id</th>\n",
              "      <th>url_online_course_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>42</td>\n",
              "      <td>1</td>\n",
              "      <td>Администрирование БД</td>\n",
              "      <td>10</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>43</td>\n",
              "      <td>2</td>\n",
              "      <td>Привилегии пользователей</td>\n",
              "      <td>10</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>44</td>\n",
              "      <td>3</td>\n",
              "      <td>Виды доступа к БД</td>\n",
              "      <td>10</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>45</td>\n",
              "      <td>4</td>\n",
              "      <td>Технические методы и средства защиты БД</td>\n",
              "      <td>10</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   id  number  ... discipline_section_id  url_online_course_id\n",
              "0  42       1  ...                    10                   NaN\n",
              "1  43       2  ...                    10                   NaN\n",
              "2  44       3  ...                    10                   NaN\n",
              "3  45       4  ...                    10                   NaN\n",
              "\n",
              "[4 rows x 5 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "C9ji4LKWveG5",
        "outputId": "ea9fe66f-66fb-4c6f-935e-4786c5ad9675"
      },
      "source": [
        "# пререквизиты обучения в РПД\n",
        "wp_prereq.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>masterylevel</th>\n",
              "      <th>item_id</th>\n",
              "      <th>workprogram_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>106</td>\n",
              "      <td>1</td>\n",
              "      <td>427</td>\n",
              "      <td>2811</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>107</td>\n",
              "      <td>1</td>\n",
              "      <td>428</td>\n",
              "      <td>2811</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>108</td>\n",
              "      <td>1</td>\n",
              "      <td>409</td>\n",
              "      <td>2811</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>109</td>\n",
              "      <td>2</td>\n",
              "      <td>427</td>\n",
              "      <td>6986</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    id  masterylevel  item_id  workprogram_id\n",
              "0  106             1      427            2811\n",
              "1  107             1      428            2811\n",
              "2  108             1      409            2811\n",
              "3  109             2      427            6986"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        },
        "id": "z78lkHHo35cF",
        "outputId": "c778c402-b1b3-4d4c-f380-b4f0b643fd6d"
      },
      "source": [
        "# результаты обучения в РПД\n",
        "wp_outcomes.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>masterylevel</th>\n",
              "      <th>item_id</th>\n",
              "      <th>workprogram_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>44816</td>\n",
              "      <td>2</td>\n",
              "      <td>9506</td>\n",
              "      <td>12684</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>44817</td>\n",
              "      <td>2</td>\n",
              "      <td>9502</td>\n",
              "      <td>12685</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>303</td>\n",
              "      <td>3</td>\n",
              "      <td>620</td>\n",
              "      <td>2730</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>305</td>\n",
              "      <td>3</td>\n",
              "      <td>1415</td>\n",
              "      <td>2730</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id  masterylevel  item_id  workprogram_id\n",
              "0  44816             2     9506           12684\n",
              "1  44817             2     9502           12685\n",
              "2    303             3      620            2730\n",
              "3    305             3     1415            2730"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "r1stJNif_wof",
        "outputId": "ea75b53d-40f5-46b7-9680-cffcebae1541"
      },
      "source": [
        "# учебные сущности\n",
        "dp_items = dp_items[dp_items['name'].notna()]\n",
        "dp_items"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>150</td>\n",
              "      <td>Эйлеровы графы</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>151</td>\n",
              "      <td>Эйлеров цикл</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>152</td>\n",
              "      <td>Эйлеров путь</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>153</td>\n",
              "      <td>Эйлеровость орграфов</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>155</td>\n",
              "      <td>Графы де Брюина</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18346</th>\n",
              "      <td>18559</td>\n",
              "      <td>Химический состав нутриентов</td>\n",
              "      <td>2</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>59.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18347</th>\n",
              "      <td>18913</td>\n",
              "      <td>Компьютерное моделирование нелинейных финансов...</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>17.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18348</th>\n",
              "      <td>18914</td>\n",
              "      <td>Проектирование тепломассообменного оборудования</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>88.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18349</th>\n",
              "      <td>18916</td>\n",
              "      <td>Программируемые логические контроллеры</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>52.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18350</th>\n",
              "      <td>18915</td>\n",
              "      <td>Программируемые логические контроллеры</td>\n",
              "      <td>2</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>52.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>18350 rows × 6 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "          id  ... domain_id\n",
              "0        150  ...      15.0\n",
              "1        151  ...      15.0\n",
              "2        152  ...      15.0\n",
              "3        153  ...      15.0\n",
              "4        155  ...      15.0\n",
              "...      ...  ...       ...\n",
              "18346  18559  ...      59.0\n",
              "18347  18913  ...      17.0\n",
              "18348  18914  ...      88.0\n",
              "18349  18916  ...      52.0\n",
              "18350  18915  ...      52.0\n",
              "\n",
              "[18350 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CGAydWB2M-UT",
        "outputId": "7551149a-da93-46f6-c055-048fe6c0127e"
      },
      "source": [
        "len(('oejo wo fjow ojo').split())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "4"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "i0DWqBhKMuOO",
        "outputId": "c207dc4b-75b0-41f9-9b1a-78246e362bcc"
      },
      "source": [
        "dp_items1 = dp_items.copy()\n",
        "dp_items1['len'] = dp_items1.name.apply(lambda x: len(x.split()))\n",
        "dp_items1[dp_items1.len > 2]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>len</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>155</td>\n",
              "      <td>Графы де Брюина</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>156</td>\n",
              "      <td>Деревья Эйлерова обхода</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>179</td>\n",
              "      <td>Кратчайшие пути в графах</td>\n",
              "      <td>23</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>207</td>\n",
              "      <td>Алгоритм вырезания соцветий</td>\n",
              "      <td>13</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>119</td>\n",
              "      <td>Метод ближайших соседей</td>\n",
              "      <td>81</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>15.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18346</th>\n",
              "      <td>18559</td>\n",
              "      <td>Химический состав нутриентов</td>\n",
              "      <td>2</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>59.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18347</th>\n",
              "      <td>18913</td>\n",
              "      <td>Компьютерное моделирование нелинейных финансов...</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>17.0</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18348</th>\n",
              "      <td>18914</td>\n",
              "      <td>Проектирование тепломассообменного оборудования</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>88.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18349</th>\n",
              "      <td>18916</td>\n",
              "      <td>Программируемые логические контроллеры</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>52.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18350</th>\n",
              "      <td>18915</td>\n",
              "      <td>Программируемые логические контроллеры</td>\n",
              "      <td>2</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>52.0</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10899 rows × 7 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "          id                                               name  ...  domain_id len\n",
              "4        155                                    Графы де Брюина  ...       15.0   3\n",
              "5        156                            Деревья Эйлерова обхода  ...       15.0   3\n",
              "11       179                           Кратчайшие пути в графах  ...       15.0   4\n",
              "20       207                        Алгоритм вырезания соцветий  ...       15.0   3\n",
              "21       119                            Метод ближайших соседей  ...       15.0   3\n",
              "...      ...                                                ...  ...        ...  ..\n",
              "18346  18559                       Химический состав нутриентов  ...       59.0   3\n",
              "18347  18913  Компьютерное моделирование нелинейных финансов...  ...       17.0   5\n",
              "18348  18914    Проектирование тепломассообменного оборудования  ...       88.0   3\n",
              "18349  18916             Программируемые логические контроллеры  ...       52.0   3\n",
              "18350  18915             Программируемые логические контроллеры  ...       52.0   3\n",
              "\n",
              "[10899 rows x 7 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "f-5sWYcgOROd",
        "outputId": "9e29716d-5f59-4211-904c-5f0c0e0e66e7"
      },
      "source": [
        "dp_items1.len.value_counts(normalize=True)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2     0.327357\n",
              "3     0.268229\n",
              "4     0.169373\n",
              "5     0.093134\n",
              "1     0.078692\n",
              "6     0.036022\n",
              "7     0.017057\n",
              "8     0.005886\n",
              "9     0.001853\n",
              "10    0.000981\n",
              "11    0.000599\n",
              "14    0.000436\n",
              "12    0.000163\n",
              "13    0.000109\n",
              "20    0.000054\n",
              "17    0.000054\n",
              "Name: len, dtype: float64"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fg5NUDV3oFsg"
      },
      "source": [
        "---\n",
        "## -- Для рек-ии пререквизитов:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Uzqj-2UWnqtQ",
        "outputId": "fd39c478-b0e8-4766-8e2b-8210b62dbcea"
      },
      "source": [
        "# учебные планы\n",
        "# academic_plan_id -- id для сайта; id -- id для таблиц\n",
        "wp_iaplan.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>year</th>\n",
              "      <th>academic_plan_id</th>\n",
              "      <th>period_of_study</th>\n",
              "      <th>op_isu_id</th>\n",
              "      <th>ns_id</th>\n",
              "      <th>ap_isu_id</th>\n",
              "      <th>educational_profile</th>\n",
              "      <th>language</th>\n",
              "      <th>qualification</th>\n",
              "      <th>title</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>531</td>\n",
              "      <td>2019</td>\n",
              "      <td>473</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6960</td>\n",
              "      <td>3893</td>\n",
              "      <td>11687</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Нейротехнологии и программирование</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>493</td>\n",
              "      <td>2018</td>\n",
              "      <td>435</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6215</td>\n",
              "      <td>3108</td>\n",
              "      <td>10555</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Информатика и программирование</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>528</td>\n",
              "      <td>2020</td>\n",
              "      <td>470</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6959</td>\n",
              "      <td>3892</td>\n",
              "      <td>13280</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Компьютерные технологии в дизайне</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    id  year  ...  qualification                               title\n",
              "0  531  2019  ...       bachelor  Нейротехнологии и программирование\n",
              "1  493  2018  ...       bachelor      Информатика и программирование\n",
              "2  528  2020  ...       bachelor   Компьютерные технологии в дизайне\n",
              "\n",
              "[3 rows x 11 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-EXGJrW76avC",
        "outputId": "b1010ea1-a4b9-410f-c109-98307447f8e6"
      },
      "source": [
        "# предметная область\n",
        "wp_fos.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>number</th>\n",
              "      <th>qualification</th>\n",
              "      <th>education_form</th>\n",
              "      <th>educational_profile</th>\n",
              "      <th>title</th>\n",
              "      <th>faculty</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>15734</td>\n",
              "      <td>01.03.02</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Прикладная математика и информатика</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>15735</td>\n",
              "      <td>01.04.02</td>\n",
              "      <td>master</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Прикладная математика и информатика</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>15736</td>\n",
              "      <td>02.04.03</td>\n",
              "      <td>master</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Математическое обеспечение и администрирование...</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>15737</td>\n",
              "      <td>07.04.04</td>\n",
              "      <td>master</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>Градостроительство</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id    number  ...                                              title  faculty\n",
              "0  15734  01.03.02  ...                Прикладная математика и информатика      NaN\n",
              "1  15735  01.04.02  ...                Прикладная математика и информатика      NaN\n",
              "2  15736  02.04.03  ...  Математическое обеспечение и администрирование...      NaN\n",
              "3  15737  07.04.04  ...                                 Градостроительство      NaN\n",
              "\n",
              "[4 rows x 7 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fUlvzgRVnqkw",
        "outputId": "8a8c2ded-b05a-4906-e451-bbdc4d881b3a"
      },
      "source": [
        "# связь уч.плана и предметной области\n",
        "wp_iaplan_fos.head(5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>implementationacademicplan_id</th>\n",
              "      <th>fieldofstudy_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>34</td>\n",
              "      <td>493</td>\n",
              "      <td>15734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>35</td>\n",
              "      <td>494</td>\n",
              "      <td>15734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>36</td>\n",
              "      <td>495</td>\n",
              "      <td>15734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>37</td>\n",
              "      <td>496</td>\n",
              "      <td>15734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>38</td>\n",
              "      <td>497</td>\n",
              "      <td>15735</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   id  implementationacademicplan_id  fieldofstudy_id\n",
              "0  34                            493            15734\n",
              "1  35                            494            15734\n",
              "2  36                            495            15734\n",
              "3  37                            496            15734\n",
              "4  38                            497            15735"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VL7joB6u-02D"
      },
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VVoV1cl9nqyG",
        "outputId": "f7020509-b223-4127-edfd-195d23015fd6"
      },
      "source": [
        "# модули в блоках в уч.плане\n",
        "wp_disc_bm.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>descipline_block_id</th>\n",
              "      <th>order</th>\n",
              "      <th>description</th>\n",
              "      <th>type</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>8461</td>\n",
              "      <td>Универсальный модуль</td>\n",
              "      <td>1695</td>\n",
              "      <td>2</td>\n",
              "      <td>NaN</td>\n",
              "      <td>faculty_module</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>8462</td>\n",
              "      <td>Модуль «Философия+Мышление»</td>\n",
              "      <td>1695</td>\n",
              "      <td>5</td>\n",
              "      <td>NaN</td>\n",
              "      <td>faculty_module</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>8463</td>\n",
              "      <td>Модуль «Цифровая культура»</td>\n",
              "      <td>1695</td>\n",
              "      <td>7</td>\n",
              "      <td>NaN</td>\n",
              "      <td>faculty_module</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>8464</td>\n",
              "      <td>Модуль «Предпринимательская культура»</td>\n",
              "      <td>1695</td>\n",
              "      <td>8</td>\n",
              "      <td>NaN</td>\n",
              "      <td>faculty_module</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     id                                   name  ...  description            type\n",
              "0  8461                   Универсальный модуль  ...          NaN  faculty_module\n",
              "1  8462            Модуль «Философия+Мышление»  ...          NaN  faculty_module\n",
              "2  8463             Модуль «Цифровая культура»  ...          NaN  faculty_module\n",
              "3  8464  Модуль «Предпринимательская культура»  ...          NaN  faculty_module\n",
              "\n",
              "[4 rows x 6 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-2Jmg593oFFC",
        "outputId": "a8e2b55b-ee84-414d-8873-61f11d31a004"
      },
      "source": [
        "# соответствие уч.плана блокам\n",
        "wp_disc_b.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>academic_plan_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1695</td>\n",
              "      <td>Блок 1. Модули (дисциплины)</td>\n",
              "      <td>435</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1696</td>\n",
              "      <td>Блок 2. Практика</td>\n",
              "      <td>435</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1697</td>\n",
              "      <td>Блок 3. ГИА</td>\n",
              "      <td>435</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1698</td>\n",
              "      <td>Блок 4. Факультативные модули (дисциплины)</td>\n",
              "      <td>435</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "     id                                        name  academic_plan_id\n",
              "0  1695                 Блок 1. Модули (дисциплины)               435\n",
              "1  1696                            Блок 2. Практика               435\n",
              "2  1697                                 Блок 3. ГИА               435\n",
              "3  1698  Блок 4. Факультативные модули (дисциплины)               435"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AviEjamPLg6p"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lSr8p-5tovi3",
        "outputId": "0b3f7a45-5607-4dda-83da-dd9d84aaa9c5"
      },
      "source": [
        "# распр-ие РПД по модулям в уч.плане\n",
        "wp_wp_fos.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>work_program_id</th>\n",
              "      <th>work_program_change_in_discipline_block_module_id</th>\n",
              "      <th>id_str_up</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>41411</td>\n",
              "      <td>2623</td>\n",
              "      <td>32043</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>41412</td>\n",
              "      <td>2624</td>\n",
              "      <td>32044</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>41413</td>\n",
              "      <td>2625</td>\n",
              "      <td>32045</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id  ...  id_str_up\n",
              "0  41411  ...        NaN\n",
              "1  41412  ...        NaN\n",
              "2  41413  ...        NaN\n",
              "\n",
              "[3 rows x 4 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ToYwLdLqXJ3k",
        "outputId": "463829c6-a264-4527-99d3-23f1eff3afac"
      },
      "source": [
        "# распр-ие РПД по модулям в уч.плане (2)\n",
        "wp_wp_disc_bm.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>discipline_block_module_id</th>\n",
              "      <th>change_type</th>\n",
              "      <th>semester_hour</th>\n",
              "      <th>credit_units</th>\n",
              "      <th>code</th>\n",
              "      <th>subject_code</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>32043</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 3.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>32044</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 2.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>32045</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.0, 1.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>32046</td>\n",
              "      <td>8462</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 0, 0, 0, 3.0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      id  discipline_block_module_id  ...  code  subject_code\n",
              "0  32043                        8461  ...   NaN             1\n",
              "1  32044                        8461  ...   NaN             2\n",
              "2  32045                        8461  ...   NaN             3\n",
              "3  32046                        8462  ...   NaN             4\n",
              "\n",
              "[4 rows x 7 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9l-KgqMmz9c_"
      },
      "source": [
        "# Предобработка текст.данных"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "cellView": "form",
        "id": "VrUju5ts0x-a"
      },
      "source": [
        "#@title Загрузка библиотек\n",
        "%%capture\n",
        "!pip install pymorphy2\n",
        "!pip install -U pymorphy2-dicts-ru\n",
        "\n",
        "from nltk.corpus import stopwords\n",
        "from gensim.utils import simple_preprocess\n",
        "import nltk\n",
        "import gensim\n",
        "import pymorphy2\n",
        "nltk.download('stopwords')\n",
        "nltk.download('punkt')\n",
        "\n",
        "import spacy\n",
        "from spacy.lang.en.stop_words import STOP_WORDS\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ufQqtHle0s1g"
      },
      "source": [
        "def preprocess(df,clmn_name):\n",
        "  '''\n",
        "    Нижний регистр, удаление стоп-слов и пунктуации, лемматизация. \n",
        "    Аргументы: \n",
        "        df -- таблица; \n",
        "        clmn_name -- название колонки с текст.данными для предобработки. \n",
        "    Возвращает: копию таблицы с предобработанными текст.данными в колонке 'prepro'.\n",
        "  '''\n",
        "\n",
        "  df_new = df.copy()\n",
        "\n",
        "  ru_stopw = stopwords.words('russian')\n",
        "  morph = pymorphy2.MorphAnalyzer()\n",
        "  nlp = spacy.load('en')\n",
        "\n",
        "  # токенизация + убираем пунктуацию\n",
        "  #df_new['prepro'] = df_new[clmn_name].apply(lambda x: (simple_preprocess(str(x), deacc=True)))\n",
        "\n",
        "  # англ: токенизация, лемматизация, убираем стоп-слова/пунктуцацию\n",
        "  df_new['prepro'] = df_new[clmn_name].apply(lambda x:[i.lemma_ for i in nlp(x) \n",
        "                                                                  if not i.is_stop and not i.is_punct])\n",
        "\n",
        "  # рус: лемматизация, убираем стоп-слова, объединяем в строку\n",
        "  df_new['prepro'] =  df_new['prepro'].apply(lambda x:' '.join([morph.parse(word)[0].normal_form for word in x \n",
        "                                                                  if not morph.parse(word)[0].normal_form in ru_stopw]))\n",
        "\n",
        "  return df_new"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BV8iCaVF5zHj"
      },
      "source": [
        "'''\n",
        "#30мин где-то выполняется\n",
        "wp_topic_prepro = preprocess(wp_topic,'description')\n",
        "wp_disc_s_prepro = preprocess(wp_disc_s,'name')\n",
        "dp_items_prepro = preprocess(dp_items,'name')'''"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KBX8MyOJzBet"
      },
      "source": [
        "'''from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "path = '/content/drive/My Drive/!Диплом/d/prepro/'\n",
        "\n",
        "with open(path+'wp_topic_prepro.csv', 'w') as f:\n",
        "  wp_topic_prepro.to_csv(f, sep=';')\n",
        "  f.close()\n",
        "\n",
        "with open(path+'wp_disc_s_prepro.csv', 'w') as f:\n",
        "  wp_disc_s_prepro.to_csv(f, sep=';')\n",
        "  f.close()\n",
        "\n",
        "with open(path+'dp_items_prepro.csv', 'w') as f:\n",
        "  dp_items_prepro.to_csv(f, sep=';')\n",
        "  f.close()\n",
        "\n",
        "drive.flush_and_unmount()'''"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fWzX45A5hqMk"
      },
      "source": [
        "# сохраняем предобработанные таблицы"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qN6lSdWag-0R",
        "outputId": "f0fac0d3-1699-4428-8235-0bda57662b7c"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "path = '/content/drive/My Drive/!Диплом/d/prepro/'\n",
        "\n",
        "with open(path+'wp_topic_prepro.csv') as f:\n",
        "  wp_topic_prepro = pd.read_csv(f, sep=';')\n",
        "  f.close()\n",
        "\n",
        "with open(path+'wp_disc_s_prepro.csv') as f:\n",
        "  wp_disc_s_prepro = pd.read_csv(f, sep=';')\n",
        "  f.close()\n",
        "\n",
        "with open(path+'dp_items_prepro.csv') as f:\n",
        "  dp_items_prepro = pd.read_csv(f, sep=';')\n",
        "  f.close()\n",
        "  \n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Rydq2duZ_03C",
        "outputId": "a770e982-f242-47e1-f172-f054bdf75bb1"
      },
      "source": [
        "wp_disc_s_prepro.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>work_program_id</th>\n",
              "      <th>SRO</th>\n",
              "      <th>contact_work</th>\n",
              "      <th>laboratory</th>\n",
              "      <th>lecture_classes</th>\n",
              "      <th>practical_lessons</th>\n",
              "      <th>total_hours</th>\n",
              "      <th>ordinal_number</th>\n",
              "      <th>prepro</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>17</td>\n",
              "      <td>Теоретические основы метрологии</td>\n",
              "      <td>5</td>\n",
              "      <td>4.0</td>\n",
              "      <td>2.2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>8.0</td>\n",
              "      <td>1</td>\n",
              "      <td>теоретический основа метрология</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>18</td>\n",
              "      <td>Измерения, погрешности и неопределенности</td>\n",
              "      <td>5</td>\n",
              "      <td>10.0</td>\n",
              "      <td>13.2</td>\n",
              "      <td>8.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>2</td>\n",
              "      <td>измерение погрешность неопределённость</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>20</td>\n",
              "      <td>Обработка результатов измерений</td>\n",
              "      <td>5</td>\n",
              "      <td>4.0</td>\n",
              "      <td>4.4</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>12.0</td>\n",
              "      <td>4</td>\n",
              "      <td>обработка результат измерение</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>1208</td>\n",
              "      <td>Особенности охлаждения сверхпроводниковых устр...</td>\n",
              "      <td>12522</td>\n",
              "      <td>25.2</td>\n",
              "      <td>4.4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>29.6</td>\n",
              "      <td>5</td>\n",
              "      <td>особенность охлаждение сверхпроводниковый устр...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>28</td>\n",
              "      <td>Системная и программная архитектура</td>\n",
              "      <td>2811</td>\n",
              "      <td>29.5</td>\n",
              "      <td>27.5</td>\n",
              "      <td>16.0</td>\n",
              "      <td>9.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>56.0</td>\n",
              "      <td>2</td>\n",
              "      <td>системный программный архитектура</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Unnamed: 0  ...                                             prepro\n",
              "0           0  ...                    теоретический основа метрология\n",
              "1           1  ...             измерение погрешность неопределённость\n",
              "2           2  ...                      обработка результат измерение\n",
              "3           3  ...  особенность охлаждение сверхпроводниковый устр...\n",
              "4           4  ...                  системный программный архитектура\n",
              "\n",
              "[5 rows x 12 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sJs8tqUnz757"
      },
      "source": [
        "\n",
        "\n",
        "---\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8sfWtD6v2euR"
      },
      "source": [
        "# 1 вариант: Поиск уч.сущностей в названиях тем/разделов"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xxx-1_b2B4np"
      },
      "source": [
        "убрали нан из таблицы сущностей, т.к. [item for item in items_list if item in x] -- ошибка"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 49
        },
        "id": "on0O5VK36M8q",
        "outputId": "febacda9-c9b0-41d9-e701-445607afd0b8"
      },
      "source": [
        "dp_items_prepro[dp_items_prepro.prepro.isna()]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>prepro</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "Empty DataFrame\n",
              "Columns: [id, name, value, source, author_id, domain_id, prepro]\n",
              "Index: []"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DeeVwhHH6Upc"
      },
      "source": [
        "dp_items_prepro.iloc[3038,-1]='can'\n",
        "dp_items_prepro.iloc[5540,-1]='система can'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G2YM1t91FCVt"
      },
      "source": [
        "#dp_items_prepro.drop('Unnamed: 0',axis=1, inplace=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 81
        },
        "id": "iu2nnkOW6izF",
        "outputId": "09655705-be45-49d2-f8b1-22ce9b584261"
      },
      "source": [
        "dp_items_prepro[dp_items_prepro.id==6152]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>prepro</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>5540</th>\n",
              "      <td>6152</td>\n",
              "      <td>Система CAN</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>система can</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "        id         name  value    source  author_id  domain_id       prepro\n",
              "5540  6152  Система CAN      6  uploaded        1.0       17.0  система can"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 95
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AU0B9W3krZZR"
      },
      "source": [
        "def itemDiscMatches(discipline_id = 12223, to_return=10):\n",
        "  '''\n",
        "    Поиск уч.сущностей, встречаемых в названиях тем/разделов дисциплины.\n",
        "    Аргументы:\n",
        "        discipline_id -- id РПД;\n",
        "        to_return -- кол-во рекомендаций;\n",
        "    Возвращает: dataframe со встреченными уч.сущностями.\n",
        "  '''\n",
        "\n",
        "  # разделы дисциплины\n",
        "  needed_sections = wp_disc_s_prepro[wp_disc_s_prepro['work_program_id']==discipline_id].copy()\n",
        "  section_ids = needed_sections.id.values\n",
        "  # названия тем дисциплины\n",
        "  needed_topics = wp_topic_prepro[wp_topic_prepro['discipline_section_id'].isin(section_ids)].copy()\n",
        "\n",
        "  # учебные сущности в виде списка\n",
        "  items_list = dp_items_prepro['prepro'].unique().tolist()\n",
        "\n",
        "  # выписываем упоминаемые в темах/разделах уч.сущности в отд.колонки\n",
        "  needed_topics['items'] = needed_topics['prepro'].apply(\n",
        "      lambda x: [item for item in items_list if item in x])\n",
        "  needed_sections['items'] = needed_sections['prepro'].apply(\n",
        "      lambda x: [item for item in items_list if item in x])\n",
        "  # удаляем темы/разделы, где не упоминались уч.сущности\n",
        "  needed_topics = needed_topics[needed_topics['items'].astype(bool)]\n",
        "  needed_sections = needed_sections[needed_sections['items'].astype(bool)]\n",
        "\n",
        "  # только уникальные встреченные уч.сущности\n",
        "  topics_sections = needed_sections['items'].append(needed_topics['items'])\n",
        "\n",
        "  # считаем частоту\n",
        "  item_names, count = np.unique(np.array([item for sublist in topics_sections \n",
        "                                          for item in sublist]), return_counts=True)\n",
        "\n",
        "  # добавляем значения частоты уч.сущностей в таблицу\n",
        "  fin_df = dp_items_prepro[dp_items_prepro['prepro'].isin(item_names)].copy()\n",
        "  len = fin_df.shape[0]\n",
        "  fin_df['val'] = 0\n",
        "  fin_df['perc'] = 0\n",
        "  for item,value in zip(item_names,count):\n",
        "    fin_df.loc[fin_df['prepro']==item,'val'] = value\n",
        "    fin_df.loc[fin_df['prepro']==item,'perc'] = value/len\n",
        "  # т.к. есть уч.сущности с пробелами на конце\n",
        "  fin_df = fin_df[-fin_df['name'].str.strip().duplicated()]\n",
        "  return fin_df.sort_values('val',ascending=False).nlargest(to_return,'val')\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 426
        },
        "id": "VOuwlW1bFgyL",
        "outputId": "0b4a6aed-030b-4566-a34f-7667843dd247"
      },
      "source": [
        "matched = itemDiscMatches(12223,15)\n",
        "matched"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>prepro</th>\n",
              "      <th>val</th>\n",
              "      <th>perc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>13167</th>\n",
              "      <td>13932</td>\n",
              "      <td>Дизайн</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>дизайн</td>\n",
              "      <td>6</td>\n",
              "      <td>0.428571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17158</th>\n",
              "      <td>17923</td>\n",
              "      <td>Игра</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>игра</td>\n",
              "      <td>6</td>\n",
              "      <td>0.428571</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>490</th>\n",
              "      <td>633</td>\n",
              "      <td>История</td>\n",
              "      <td>70</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>16.0</td>\n",
              "      <td>история</td>\n",
              "      <td>4</td>\n",
              "      <td>0.285714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2932</th>\n",
              "      <td>3249</td>\n",
              "      <td>Конфликт</td>\n",
              "      <td>14</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>29.0</td>\n",
              "      <td>конфликт</td>\n",
              "      <td>2</td>\n",
              "      <td>0.142857</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6679</th>\n",
              "      <td>7146</td>\n",
              "      <td>Механика</td>\n",
              "      <td>13</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>20.0</td>\n",
              "      <td>механика</td>\n",
              "      <td>2</td>\n",
              "      <td>0.142857</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>559</th>\n",
              "      <td>729</td>\n",
              "      <td>Данные</td>\n",
              "      <td>98</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>дать</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2498</th>\n",
              "      <td>2801</td>\n",
              "      <td>Структуры Ли</td>\n",
              "      <td>15</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>11.0</td>\n",
              "      <td>структура</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3326</th>\n",
              "      <td>3650</td>\n",
              "      <td>Интерактивные элементы</td>\n",
              "      <td>12</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>интерактивный элемент</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6460</th>\n",
              "      <td>6930</td>\n",
              "      <td>Структура ПО</td>\n",
              "      <td>7</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>структура</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10398</th>\n",
              "      <td>11140</td>\n",
              "      <td>Ресурсы</td>\n",
              "      <td>11</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>25.0</td>\n",
              "      <td>ресурс</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11954</th>\n",
              "      <td>12712</td>\n",
              "      <td>РИД</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>рид</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15551</th>\n",
              "      <td>16303</td>\n",
              "      <td>Геймдизайн</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>геймдизайн</td>\n",
              "      <td>1</td>\n",
              "      <td>0.071429</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id                    name  ...  val      perc\n",
              "13167  13932                  Дизайн  ...    6  0.428571\n",
              "17158  17923                    Игра  ...    6  0.428571\n",
              "490      633                 История  ...    4  0.285714\n",
              "2932    3249                Конфликт  ...    2  0.142857\n",
              "6679    7146                Механика  ...    2  0.142857\n",
              "559      729                  Данные  ...    1  0.071429\n",
              "2498    2801            Структуры Ли  ...    1  0.071429\n",
              "3326    3650  Интерактивные элементы  ...    1  0.071429\n",
              "6460    6930            Структура ПО  ...    1  0.071429\n",
              "10398  11140                 Ресурсы  ...    1  0.071429\n",
              "11954  12712                     РИД  ...    1  0.071429\n",
              "15551  16303              Геймдизайн  ...    1  0.071429\n",
              "\n",
              "[12 rows x 9 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 520
        },
        "id": "OxY3jekOr-vg",
        "outputId": "e86af3cf-8cd5-42ec-ef5e-c322297ec85f"
      },
      "source": [
        "matched = itemDiscMatches(6986,15)\n",
        "matched"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>prepro</th>\n",
              "      <th>val</th>\n",
              "      <th>perc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>549</th>\n",
              "      <td>717</td>\n",
              "      <td>CSS</td>\n",
              "      <td>57</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>28.0</td>\n",
              "      <td>css</td>\n",
              "      <td>6</td>\n",
              "      <td>0.400000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>548</th>\n",
              "      <td>716</td>\n",
              "      <td>HTML</td>\n",
              "      <td>46</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>28.0</td>\n",
              "      <td>html</td>\n",
              "      <td>5</td>\n",
              "      <td>0.333333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17537</th>\n",
              "      <td>18305</td>\n",
              "      <td>UX</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ux</td>\n",
              "      <td>3</td>\n",
              "      <td>0.200000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12467</th>\n",
              "      <td>13227</td>\n",
              "      <td>Проект</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>проект</td>\n",
              "      <td>2</td>\n",
              "      <td>0.133333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13167</th>\n",
              "      <td>13932</td>\n",
              "      <td>Дизайн</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>дизайн</td>\n",
              "      <td>2</td>\n",
              "      <td>0.133333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16306</th>\n",
              "      <td>17064</td>\n",
              "      <td>Проектирование ПО</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>проектирование</td>\n",
              "      <td>2</td>\n",
              "      <td>0.133333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17359</th>\n",
              "      <td>18126</td>\n",
              "      <td>Проектирование</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>проектирование</td>\n",
              "      <td>2</td>\n",
              "      <td>0.133333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17538</th>\n",
              "      <td>18306</td>\n",
              "      <td>UI</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ui</td>\n",
              "      <td>2</td>\n",
              "      <td>0.133333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>279</th>\n",
              "      <td>461</td>\n",
              "      <td>Классы</td>\n",
              "      <td>26</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>класс</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1559</th>\n",
              "      <td>1813</td>\n",
              "      <td>Bootstrap</td>\n",
              "      <td>23</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>bootstrap</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3573</th>\n",
              "      <td>3925</td>\n",
              "      <td>Анимация</td>\n",
              "      <td>55</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>17.0</td>\n",
              "      <td>анимация</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9952</th>\n",
              "      <td>10694</td>\n",
              "      <td>Библиотеки</td>\n",
              "      <td>0</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>библиотека</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12478</th>\n",
              "      <td>13238</td>\n",
              "      <td>Разработка ПО</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>разработка</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12876</th>\n",
              "      <td>13638</td>\n",
              "      <td>ST</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>st</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16411</th>\n",
              "      <td>17170</td>\n",
              "      <td>ИМА</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>има</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066667</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id               name  value  ...          prepro  val      perc\n",
              "549      717                CSS     57  ...             css    6  0.400000\n",
              "548      716               HTML     46  ...            html    5  0.333333\n",
              "17537  18305                 UX      0  ...              ux    3  0.200000\n",
              "12467  13227             Проект      0  ...          проект    2  0.133333\n",
              "13167  13932             Дизайн      0  ...          дизайн    2  0.133333\n",
              "16306  17064  Проектирование ПО      0  ...  проектирование    2  0.133333\n",
              "17359  18126     Проектирование      0  ...  проектирование    2  0.133333\n",
              "17538  18306                 UI      0  ...              ui    2  0.133333\n",
              "279      461             Классы     26  ...           класс    1  0.066667\n",
              "1559    1813          Bootstrap     23  ...       bootstrap    1  0.066667\n",
              "3573    3925           Анимация     55  ...        анимация    1  0.066667\n",
              "9952   10694         Библиотеки      0  ...      библиотека    1  0.066667\n",
              "12478  13238      Разработка ПО      0  ...      разработка    1  0.066667\n",
              "12876  13638                 ST      0  ...              st    1  0.066667\n",
              "16411  17170                ИМА      0  ...             има    1  0.066667\n",
              "\n",
              "[15 rows x 9 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 100
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a_fRGHm_0dva",
        "outputId": "bf0ba4a0-eccd-40a1-98b7-5f6a7f323de4"
      },
      "source": [
        "%%timeit\n",
        "itemDiscMatches()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10 loops, best of 5: 77.8 ms per loop\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p93st8hW8wYK"
      },
      "source": [
        "# 2 вариант: Сходство уч.сущностей и тем"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rTdIGFvk9tHj"
      },
      "source": [
        "Векторные представления уч.сущностей + тем.\n",
        "- можно попробовать взять не предобученные модели, а обучить на учебниках\\текстах онлайн-курсов?\n",
        "- (X: не нашла multilingual word emb-s, которые нормально работают) попробовать word2vec,GloVe,fasttext -- либо среднее векторов слов для предложения, либо doc2vec, либо...\n",
        "- можно выводить самые частые домены?чтобы преподаватель знал, где искать вручную уч.сущности. Или это и так им очевидно должно быть"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wuIZ0hroTz0v"
      },
      "source": [
        "## -- Ф-ии"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Humy1zPXT_ig"
      },
      "source": [
        "def plot_similarity(topics, items, all_vectors):\n",
        "  '''\n",
        "    Матрица корелляции + ее визуализация. \n",
        "    Аргументы: \n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей. \n",
        "    Возвращает: матрицу корелляции и график.\n",
        "  '''\n",
        "  len = topics.shape[0]\n",
        "\n",
        "  #оставляем только часть сверху справа.\n",
        "  corr = np.inner(all_vectors[:len], all_vectors[len:])\n",
        "  sns.set(font_scale=0.6)\n",
        "  #PuBuGn\n",
        "  g = sns.heatmap(corr,vmin=0,vmax=1,cmap=\"YlGnBu\")\n",
        "\n",
        "  return corr,g"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nBR7AcO9TZiS"
      },
      "source": [
        "def corrItems(labels,topics,items,all_vectors,threshold=0.8):\n",
        "  '''\n",
        "    Поиск уч.сущностей, которые похожи на названия тем \n",
        "    (выбирает только одну сущность с макс.значением для каждой темы).\n",
        "    Аргументы: \n",
        "        labels -- список названий, которые будут выведены в итоге;\n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей;\n",
        "        threshold -- порог значений. \n",
        "    Возвращает: таблицу с темами и соответствующими уч.сущностями \n",
        "                + визуализацию матрицы корелляции.\n",
        "  '''\n",
        "  corr_m, plot = plot_similarity(topics,items,all_vectors)\n",
        "  corr_m_new = corr_m.copy()\n",
        "\n",
        "  new_df = pd.DataFrame(labels[0],columns=['name'])\n",
        "  new_df['value']=corr_m_new.max(axis=1)\n",
        "  new_df['item'] = corr_m_new.argmax(axis=1)\n",
        "  new_df['item_name'] = new_df['item'].apply(lambda x: labels[1][x])\n",
        "\n",
        "  return new_df[new_df.value>threshold], plot"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9X2zWK3iTZiT"
      },
      "source": [
        "def corrItemsList(topics,items,all_vectors,threshold=0.8):\n",
        "  '''\n",
        "    Поиск уч.сущностей, которые похожи на названия тем. \n",
        "    Выбирает все сущности, значение которых выше порога.\n",
        "    Аргументы: \n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей;\n",
        "        threshold -- порог значений. \n",
        "    Возвращает: таблицу с уч.сущностями + визуализацию матрицы корелляции.\n",
        "  '''\n",
        "  corr_m, plot = plot_similarity(topics,items,all_vectors)\n",
        "  corr_m_new = corr_m.copy()\n",
        "\n",
        "  corr_items = np.where(np.any(corr_m_new > threshold, axis=0))\n",
        "\n",
        "  return np.unique(corr_items), plot"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QSoLsyD2TZiU"
      },
      "source": [
        "def itemDiscSim(discipline_id, vec_func, items_vec=False, threshold=0.65, to_return=10):\n",
        "  '''\n",
        "    Поиск уч.сущностей, которые похожи на названия тем/разделов дисциплины. \n",
        "    Выбирает все сущности, значения которых выше порога.\n",
        "    Аргументы: \n",
        "        discipline_id -- id дисциплины;\n",
        "        vec_func -- функция векторизации;\n",
        "        items_vec -- готовые векторы для уч.сущностей;\n",
        "        threshold -- порог значений;\n",
        "        to_return -- кол-во рекомендаций. \n",
        "    Возвращает: таблицу с уч.сущностями\n",
        "  '''\n",
        "  # разделы дисциплины\n",
        "  needed_sections = wp_disc_s[wp_disc_s['work_program_id']==discipline_id].copy()\n",
        "  section_ids = needed_sections.id.values\n",
        "  # названия тем дисциплины\n",
        "  needed_topics = wp_topic[wp_topic['discipline_section_id'].isin(section_ids)].copy()\n",
        "  # будем искать соответствия И в темах, И в разделах\n",
        "  all_headers = np.append(needed_sections['name'].values, needed_topics['description'].values)\n",
        "\n",
        "  # если были поданы на вход векторы уч.сущностей\n",
        "  if items_vec is not False:\n",
        "    headers_vecs = vec_func(all_headers)\n",
        "  else:\n",
        "    headers_vecs, items_vec = vec_func(all_headers,dp_items['name'].values)\n",
        "\n",
        "  # строим матрицу сходства\n",
        "  len = all_headers.shape[0]\n",
        "  sim_m = np.inner(headers_vecs, items_vec)\n",
        "\n",
        "  # ищем id и значения подходящих элементов\n",
        "  ids = np.where(np.any(sim_m > threshold, axis=0))\n",
        "  vals = np.max(sim_m[:,ids],axis=0)[0]\n",
        "  \n",
        "  # добавляем данные в новую таблицу\n",
        "  df = dp_items.iloc[ids].copy()\n",
        "  df['similarity'] = vals\n",
        "\n",
        "  return df.sort_values('similarity',ascending=False).nlargest(to_return, 'similarity')\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UFsFTNlwoidH"
      },
      "source": [
        "## -- Elmo"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aOxHrwuopqAD"
      },
      "source": [
        "Рез-ты плохие + тк длина векторов зависит от длины самого большого предложения, то нельзя заранее представить все уч.сущности в виде векторов?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tLSX_Y2OqQTY"
      },
      "source": [
        "%%capture\n",
        "!pip install --upgrade simple_elmo"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BvyWx4UDqdDd"
      },
      "source": [
        "from google.colab import drive\n",
        "from simple_elmo import ElmoModel\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "model1 = ElmoModel()\n",
        "model1.load('/content/drive/My Drive/!Диплом/212.zip')\n",
        "\n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8NgcSkR_su6R"
      },
      "source": [
        "model_elmo = model1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u52fxfv3yDr2"
      },
      "source": [
        "from sklearn.metrics.pairwise import cosine_similarity"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V9Ww3j5Sq2MA"
      },
      "source": [
        "def vecFuncELMO(headers, items):\n",
        "  return model_elmo.get_elmo_vector_average(np.append(headers, items))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ObLo9kGRUj4X"
      },
      "source": [
        "sentences = ['This framework generates embeddings for each input sentence',\n",
        "              'Sentences are passed as a list of string.', \n",
        "              'The quick brown fox jumps over the lazy dog.', \n",
        "              'Быстрая коричневая лиса перепрыгнула через ленивую собаку.',\n",
        "              'Предложения передаются в виде списка строк.', \n",
        "              'Random forest','Случайный лес']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4U3cVaXMaBlT",
        "outputId": "be2b1188-f3d8-4814-ba1c-0399729a5fd3"
      },
      "source": [
        "temp = pd.DataFrame(sentences)\n",
        "\n",
        "temp[0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0    This framework generates embeddings for each i...\n",
              "1            Sentences are passed as a list of string.\n",
              "2         The quick brown fox jumps over the lazy dog.\n",
              "3    Быстрая коричневая лиса перепрыгнула через лен...\n",
              "4          Предложения передаются в виде списка строк.\n",
              "5                                        Random forest\n",
              "6                                        Случайный лес\n",
              "Name: 0, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5NMF55cjX95u",
        "outputId": "0223b476-c89c-4fb7-add9-43ed491cb9c2"
      },
      "source": [
        "ru_stopw = stopwords.words('russian')\n",
        "morph = pymorphy2.MorphAnalyzer()\n",
        "nlp = spacy.load('en')\n",
        "  # англ: токенизация, лемматизация, убираем стоп-слова/пунктуцацию\n",
        "temp['prepro'] = temp[0].apply(lambda x:[i.lemma_ for i in nlp(x) \n",
        "                                                                  if not i.is_stop and not i.is_punct])\n",
        "  # рус: лемматизация, убираем стоп-слова, объединяем в строку\n",
        "temp['prepro'] =  temp['prepro'].apply(lambda x:' '.join([morph.parse(word)[0].normal_form for word in x \n",
        "                                                                  if not morph.parse(word)[0].normal_form in ru_stopw]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:00:13,775 : INFO : Loading dictionaries from /usr/local/lib/python3.7/dist-packages/pymorphy2_dicts_ru/data\n",
            "2021-05-26 21:00:13,832 : INFO : format: 2.4, revision: 417127, updated: 2020-10-11T15:05:51.070345\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "id": "nmupmpM_e5vw",
        "outputId": "4ef92d71-6be5-4dd5-a325-6692a714ea23"
      },
      "source": [
        "temp"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>0</th>\n",
              "      <th>prepro</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>This framework generates embeddings for each i...</td>\n",
              "      <td>framework generate embedding input sentence</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Sentences are passed as a list of string.</td>\n",
              "      <td>sentence pass list string</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>The quick brown fox jumps over the lazy dog.</td>\n",
              "      <td>quick brown fox jump lazy dog</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Быстрая коричневая лиса перепрыгнула через лен...</td>\n",
              "      <td>быстрый коричневый лиса перепрыгнуть ленивый с...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Предложения передаются в виде списка строк.</td>\n",
              "      <td>предложение передаваться вид список строка</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>Random forest</td>\n",
              "      <td>random forest</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>Случайный лес</td>\n",
              "      <td>случайный лес</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                   0                                             prepro\n",
              "0  This framework generates embeddings for each i...        framework generate embedding input sentence\n",
              "1          Sentences are passed as a list of string.                          sentence pass list string\n",
              "2       The quick brown fox jumps over the lazy dog.                      quick brown fox jump lazy dog\n",
              "3  Быстрая коричневая лиса перепрыгнула через лен...  быстрый коричневый лиса перепрыгнуть ленивый с...\n",
              "4        Предложения передаются в виде списка строк.         предложение передаваться вид список строка\n",
              "5                                      Random forest                                      random forest\n",
              "6                                      Случайный лес                                      случайный лес"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s4JtwDe9e_9L",
        "outputId": "3d8dcb73-6a5f-48c9-f2fb-be7f4ab7a870"
      },
      "source": [
        "sentence_embeddings = model1.get_elmo_vector_average(temp['prepro'].to_list())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:06:53,796 : INFO : Warming up ELMo on 7 sentences...\n",
            "2021-05-26 21:06:59,084 : INFO : Warming up finished.\n",
            "2021-05-26 21:06:59,089 : INFO : Texts in the current batch: 7\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "id": "NSEQQUdeqewy",
        "outputId": "0fb1022f-b24f-4ed5-f80e-28e4a0f5e83d"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "sns.set(font_scale=1)\n",
        "plt.figure(figsize=(8, 6))\n",
        "plot=sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\",yticklabels=sentences)\n",
        "loc, labels = plt.yticks()\n",
        "plot.set_yticklabels(labels, rotation=20);"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "DGoxZ_yLfCst",
        "outputId": "8db7e182-6340-403c-e483-323ca925e627"
      },
      "source": [
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\",yticklabels=sentences)\n",
        "#sns.set_xticklabels(labels, rotation=rotation)\n",
        "sns."
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fac35380810>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "TMgZdYVwUkLU",
        "outputId": "873c3d27-1c43-4c45-fc5d-a3a472baaf0a"
      },
      "source": [
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fce870797d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aflDbtVHd-ci"
      },
      "source": [
        "## -- (V) Universal Encoder (TH-Hub)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MYXl3xIfd-xi",
        "outputId": "3728823f-bfc8-46bc-a71a-cfcbded48259"
      },
      "source": [
        "!pip install tensorflow_text>=2.0.0rc0\n",
        "!pip install numpy==1.19.5"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: numpy==1.19.5 in /usr/local/lib/python3.7/dist-packages (1.19.5)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YPgmU91qYyzs"
      },
      "source": [
        "import tensorflow_hub as hub"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QS3o6hJKZLQW"
      },
      "source": [
        "### пробуем загрузить сохраненную на диске модель"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iArh7tNaYTA_"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "loaded = hub.load('/content/drive/My Drive/bachelor/!Диплом/usencoder/')\n",
        "\n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "0Bu3DdRxZb3p"
      },
      "source": [
        "embed('i love you') ==loaded('i love you')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GRD8hdkWZPq1"
      },
      "source": [
        "используем загрузку с хаба"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LwzSaUpTenZx"
      },
      "source": [
        "import tensorflow_hub as hub\n",
        "import numpy as np\n",
        "import tensorflow_text\n",
        "import seaborn as sns"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "60xXLpS2grj2"
      },
      "source": [
        "embed = hub.load(\"https://tfhub.dev/google/universal-sentence-encoder-multilingual/3\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OGjaHrAIP0pd",
        "outputId": "fbad5c80-a58b-4b56-e9c7-a42ebd6cf9b8"
      },
      "source": [
        "%%timeit\n",
        "# заранее делаем эмбединг для уч.сущностей\n",
        "items_vecs = embed(dp_items['name'].values)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1 loop, best of 5: 18.4 s per loop\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "yKlg0JBasC7k"
      },
      "source": [
        "items_vecs.numpy().shape"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "GYzY0E1UUhaK"
      },
      "source": [
        "np_vecs = items_vecs.numpy()\n",
        "dp_items.id.values.shape"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "lODxP_-8sR6c"
      },
      "source": [
        "df = pd.DataFrame({'id':dp_items.id.values,\n",
        "              'emb':list(np_vecs)})\n",
        "df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "nYvODjqFyBgx"
      },
      "source": [
        "pd.read_json(df.to_json())"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "aFSOU7nUYnhJ"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "\n",
        "with open('/content/drive/My Drive/!Диплом/embs.json', 'w') as f:\n",
        "  f.write(df.to_json())\n",
        "  f.close()\n",
        "\n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Psi9VyPbYnju",
        "outputId": "3653163f-a53e-4761-b72d-c55156ab9360"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "\n",
        "with open('/content/drive/My Drive/bachelor/!Диплом/embs.json') as json_file:\n",
        "    read_json = pd.read_json(json_file)\n",
        "    json_file.close()\n",
        "\n",
        "drive.flush_and_unmount()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "oolzBs-2aO0o"
      },
      "source": [
        "read_json"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "8s9_T86f2FUg"
      },
      "source": [
        "%%timeit\n",
        "np.inner(np.vstack(read_json['emb']),np.vstack(read_json['emb']))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "wZL9J9wp1ip_"
      },
      "source": [
        "%%timeit\n",
        "cosine_similarity(np.vstack(read_json['emb']))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "cDs39_Cd3MX2"
      },
      "source": [
        "np.inner(np.vstack(read_json['emb'][:3]),np.vstack(read_json['emb'][-2:]))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "yMeaKME0hrP1"
      },
      "source": [
        "from sklearn.metrics.pairwise import cosine_similarity"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "F9-4OCKnR3C6"
      },
      "source": [
        "%%timeit\n",
        "items_vecs = embed(dp_items['name'].values)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "tFkOCGrI7T_x"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "cIGDYi0j7UN1"
      },
      "source": [
        "corr = np.inner(np.vstack(read_json['emb'][:3]),np.vstack(read_json['emb'][3:100]))\n",
        "new_df = pd.DataFrame(read_json['id'][:3],columns=['name'])\n",
        "\n",
        "corr_items = np.where(np.any(corr > 0.5, axis=0))\n",
        "vals = np.max(corr[:,corr_items],axis=0)[0]\n",
        "df = dp_items.iloc[corr_items].copy()\n",
        "df['similarity'] = vals\n",
        "\n",
        "df.sort_values('similarity',ascending=False).nlargest(10, 'similarity')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D9PiMYO9aYpu"
      },
      "source": [
        "Выше заранее вычисленные эмбеддинги записывали в файл и читали.\n",
        "Ниже -- проверяем работу моделей"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "id": "E_HUpO-xuvG7",
        "outputId": "e9ba235a-e3c7-4c55-f467-eda444cdac50"
      },
      "source": [
        "sentences = ['HTML','CSS','JavaScript','Neural networks','Нейронные сети','Random forest','Случайный лес']\n",
        "\n",
        "sentence_embeddings = embed(sentences)\n",
        "\n",
        "sentences_lab = ['HTML','CSS','JavaScript','Neural networks','[RUS] Neural networks','Random forest','[RUS] Random forest']\n",
        "import matplotlib.pyplot as plt\n",
        "plt.figure(figsize=(8, 6))\n",
        "plot=sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\",yticklabels=sentences_lab, xticklabels=sentences_lab)\n",
        "loc, labels = plt.yticks()\n",
        "plot.set_yticklabels(labels, rotation=15);\n",
        "loc, labels = plt.xticks()\n",
        "plot.set_xticklabels(labels, rotation=15, ha='right');"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x432 with 2 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "1i9tq7FuuvOY",
        "outputId": "112a2e48-1c0b-434a-c00c-e1654e6336c6"
      },
      "source": [
        "sentences = ['This framework generates embeddings for each input sentence',\n",
        "    'Sentences are passed as a list of string.', \n",
        "    'The quick brown fox jumps over the lazy dog.', \n",
        "    'Быстрая коричневая лиса перепрыгнула через ленивую собаку.',\n",
        "    'Предложения передаются в виде списка строк.', 'Random forest','Случайный лес']\n",
        "\n",
        "sentence_embeddings = embed(sentences)\n",
        "\n",
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f50f8482790>"
            ]
          },
          "metadata": {},
          "execution_count": 44
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IJObEqvOsUP2"
      },
      "source": [
        "### Проверяем для всех тем"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6X7m7Binfoc0"
      },
      "source": [
        "def plot_similarity(topics, items, all_vectors):\n",
        "  '''\n",
        "    Матрица корелляции + ее визуализация. \n",
        "    Аргументы: \n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей. \n",
        "    Возвращает: матрицу корелляции и график.\n",
        "  '''\n",
        "  len = topics.shape[0]\n",
        "\n",
        "  #оставляем только часть сверху справа.\n",
        "  corr = np.inner(all_vectors[:len], all_vectors[len:])\n",
        "  sns.set(font_scale=0.6)\n",
        "  #PuBuGn\n",
        "  g = sns.heatmap(corr,vmin=0,vmax=1,cmap=\"YlGnBu\")\n",
        "\n",
        "  return corr,g"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 230
        },
        "id": "sSmtnhPuMUAy",
        "outputId": "3757d0e0-41da-4282-f4dd-2dbbb163f944"
      },
      "source": [
        "topics, items_val = wp_topic['description'].values[:50], dp_items['name'].values[:1000]\n",
        "all_labels = np.append(topics,items_val)\n",
        "all_vectors = embed(all_labels)\n",
        "corr, plot = plot_similarity(topics, items_val, all_vectors)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-33-0e31eba9374a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mtopics\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mitems_val\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mwp_topic\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'description'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdp_items\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'name'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m1000\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mall_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtopics\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mitems_val\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0mall_vectors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0membed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mall_labels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m \u001b[0mcorr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplot\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplot_similarity\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtopics\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mitems_val\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mall_vectors\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'embed' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "b30OGkkCj5Fs"
      },
      "source": [
        "def corrItems(labels,topics,items,all_vectors,threshold=0.8):\n",
        "  '''\n",
        "    Поиск уч.сущностей, которые похожи на названия тем \n",
        "    (выбирает только одну сущность с макс.значением для каждой темы).\n",
        "    Аргументы: \n",
        "        labels -- список названий, которые будут выведены в итоге;\n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей;\n",
        "        threshold -- порог значений. \n",
        "    Возвращает: таблицу с темами и соответствующими уч.сущностями \n",
        "                + визуализацию матрицы корелляции.\n",
        "  '''\n",
        "  corr_m, plot = plot_similarity(topics,items,all_vectors)\n",
        "  corr_m_new = corr_m.copy()\n",
        "\n",
        "  new_df = pd.DataFrame(labels[0],columns=['name'])\n",
        "  new_df['value']=corr_m_new.max(axis=1)\n",
        "  new_df['item'] = corr_m_new.argmax(axis=1)\n",
        "  new_df['item_name'] = new_df['item'].apply(lambda x: labels[1][x])\n",
        "\n",
        "  return new_df[new_df.value>threshold], plot"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "RpIP-iI6OhI0"
      },
      "source": [
        "labels = [wp_topic['description'].values[:50], dp_items['name'].values[:1000]]\n",
        "all_labels = np.append(labels[0],labels[1])\n",
        "all_vectors = embed(all_labels)\n",
        "res, plot = corrItems(labels, wp_topic['description'].values[:50], dp_items['name'].values[:1000], all_vectors,threshold=0.7)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "tFDxVfnwXFZu"
      },
      "source": [
        "res"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0XgB6VIGk7zK"
      },
      "source": [
        "---\n",
        "Возвращаемся -> "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "yi0BLWnDbfuE"
      },
      "source": [
        "def corrItemsList(item_labels,topics,items,all_vectors,threshold=0.8):\n",
        "  '''\n",
        "    Поиск уч.сущностей, которые похожи на названия тем. \n",
        "    Выбирает все сущности, значение которых выше порога.\n",
        "    Аргументы: \n",
        "        item_labels -- список названий уч.сущностей, которые будут выведены в итоге\n",
        "                        (при выводе id, а не названий, не используется);\n",
        "        topics -- список названий тем; \n",
        "        items -- список названий уч.сущностей;\n",
        "        all_vectors -- список с векторами и тем, и уч.сущностей;\n",
        "        threshold -- порог значений. \n",
        "    Возвращает: таблицу с уч.сущностями + визуализацию матрицы корелляции.\n",
        "  '''\n",
        "  corr_m, plot = plot_similarity(topics,items,all_vectors)\n",
        "  corr_m_new = corr_m.copy()\n",
        "\n",
        "  #corr_items = item_labels[np.where(np.any(corr_m_new > threshold, axis=0))]\n",
        "  corr_items = np.where(np.any(corr_m_new > threshold, axis=0))\n",
        "\n",
        "  return np.unique(corr_items), plot"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "iIkxkTCVeOIr"
      },
      "source": [
        "all_vectors = embed(np.append(wp_topic['description'].values,dp_items['name'].values))\n",
        "resList,plot_l = corrItemsList(dp_items['name'].values, wp_topic['description'].values, dp_items['name'].values, all_vectors, threshold=0.8)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "vcBKOoOsbT0G"
      },
      "source": [
        "dp_items.iloc[resList]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "jgv57rDemrVB"
      },
      "source": [
        "# есть повторения в названиях сущностей, но вдруг повторы относились бы к разным областям?\n",
        "dp_items.iloc[resList].drop_duplicates(['name','domain_id'])['name'].values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "fUhILz-YrySD"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D71GlrsArzHt"
      },
      "source": [
        "### То же, но с вводом id дисциплины"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RkJHq6KOJwZa"
      },
      "source": [
        "def vecFuncUSE(headers):\n",
        "  return embed(headers)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "txudezEknorq"
      },
      "source": [
        "res = itemDiscSim(12223,vecFuncUSE,items_vecs)\n",
        "res"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "f5nOFMC6iwZs"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cPzg40TEaheE"
      },
      "source": [
        "## -- BERT"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dRzORw25YkTV"
      },
      "source": [
        "Долго вычисляет + странные рез-ты?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t1uvLKn6al8P"
      },
      "source": [
        "%%capture\n",
        "!pip install transformers"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "H_s524cpanMv",
        "outputId": "c890f513-d879-4605-adea-ce46235fe3bf"
      },
      "source": [
        "import torch\n",
        "from sklearn.metrics.pairwise import cosine_similarity\n",
        "from sklearn import preprocessing\n",
        "from transformers import AutoModel, AutoTokenizer, BertTokenizer\n",
        "\n",
        "torch.set_grad_enabled(False)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<torch.autograd.grad_mode.set_grad_enabled at 0x7fac98fe47d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 445,
          "referenced_widgets": [
            "8660ccc5f5be408b9a9085fc52424dc1",
            "bdd0af729cd144f9b5bd9f434e2a2bb4",
            "b79de08fa68342d2a19e444c1928023d",
            "5d39e44dfb3845e5a29dced91cdff4bb"
          ]
        },
        "id": "1gceh369anPo",
        "outputId": "178c14d7-68ac-466e-a09f-c0fef9410586"
      },
      "source": [
        "model = AutoModel.from_pretrained('bert-base-multilingual-cased')\n",
        "tokenizer = AutoTokenizer.from_pretrained('bert-base-multilingual-cased')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:00,551 : INFO : Lock 140377619456464 acquired on /root/.cache/huggingface/transformers/0a3fd51713dcbb4def175c7f85bddc995d5976ce1dde327f99104e4d33069f17.aa7be4c79d76f4066d9b354496ea477c9ee39c5d889156dd1efb680643c2b052.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "8660ccc5f5be408b9a9085fc52424dc1",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=714314041.0, style=ProgressStyle(descri…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:21,348 : INFO : Lock 140377619456464 released on /root/.cache/huggingface/transformers/0a3fd51713dcbb4def175c7f85bddc995d5976ce1dde327f99104e4d33069f17.aa7be4c79d76f4066d9b354496ea477c9ee39c5d889156dd1efb680643c2b052.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "Some weights of the model checkpoint at bert-base-multilingual-cased were not used when initializing BertModel: ['cls.predictions.transform.LayerNorm.bias', 'cls.seq_relationship.weight', 'cls.predictions.transform.LayerNorm.weight', 'cls.predictions.transform.dense.weight', 'cls.predictions.decoder.weight', 'cls.seq_relationship.bias', 'cls.predictions.bias', 'cls.predictions.transform.dense.bias']\n",
            "- This IS expected if you are initializing BertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
            "- This IS NOT expected if you are initializing BertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
            "2021-05-26 21:03:24,888 : INFO : Lock 140377619483856 acquired on /root/.cache/huggingface/transformers/eff018e45de5364a8368df1f2df3461d506e2a111e9dd50af1fae061cd460ead.6c5b6600e968f4b5e08c86d8891ea99e51537fc2bf251435fb46922e8f7a7b29.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "bdd0af729cd144f9b5bd9f434e2a2bb4",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=995526.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:25,114 : INFO : Lock 140377619483856 released on /root/.cache/huggingface/transformers/eff018e45de5364a8368df1f2df3461d506e2a111e9dd50af1fae061cd460ead.6c5b6600e968f4b5e08c86d8891ea99e51537fc2bf251435fb46922e8f7a7b29.lock\n",
            "2021-05-26 21:03:25,192 : INFO : Lock 140377619483856 acquired on /root/.cache/huggingface/transformers/46880f3b0081fda494a4e15b05787692aa4c1e21e0ff2428ba8b14d4eda0784d.b33e51591f94f17c238ee9b1fac75b96ff2678cbaed6e108feadb3449d18dc24.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "b79de08fa68342d2a19e444c1928023d",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=1961828.0, style=ProgressStyle(descript…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:25,437 : INFO : Lock 140377619483856 released on /root/.cache/huggingface/transformers/46880f3b0081fda494a4e15b05787692aa4c1e21e0ff2428ba8b14d4eda0784d.b33e51591f94f17c238ee9b1fac75b96ff2678cbaed6e108feadb3449d18dc24.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:25,657 : INFO : Lock 140377618734416 acquired on /root/.cache/huggingface/transformers/f55e7a2ad4f8d0fff2733b3f79777e1e99247f2e4583703e92ce74453af8c235.ec5c189f89475aac7d8cbd243960a0655cfadc3d0474da8ff2ed0bf1699c2a5f.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "5d39e44dfb3845e5a29dced91cdff4bb",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=29.0, style=ProgressStyle(description_w…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "2021-05-26 21:03:25,788 : INFO : Lock 140377618734416 released on /root/.cache/huggingface/transformers/f55e7a2ad4f8d0fff2733b3f79777e1e99247f2e4583703e92ce74453af8c235.ec5c189f89475aac7d8cbd243960a0655cfadc3d0474da8ff2ed0bf1699c2a5f.lock\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "20n8FeDwEDvt"
      },
      "source": [
        "from sklearn import preprocessing"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gg9UZkC3fv6Q",
        "outputId": "aeebce7b-c8a1-4216-fa86-b698d28fe2b1"
      },
      "source": [],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['framework generate embedding input sentence',\n",
              " 'sentence pass list string',\n",
              " 'quick brown fox jump lazy dog',\n",
              " 'быстрый коричневый лиса перепрыгнуть ленивый собака',\n",
              " 'предложение передаваться вид список строка',\n",
              " 'random forest',\n",
              " 'случайный лес']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "bDOXveWufpRU",
        "outputId": "95e9832b-2b55-4d06-d836-edbb1b81a558"
      },
      "source": [
        "sentences = ['This framework generates embeddings for each input sentence',\n",
        "    'Sentences are passed as a list of string.', \n",
        "    'The quick brown fox jumps over the lazy dog.', \n",
        "    'Быстрая коричневая лиса перепрыгнула через ленивую собаку.',\n",
        "    'Предложения передаются в виде списка строк.', 'Random forest','Случайный лес']\n",
        "\n",
        "#tokens_pt = tokenizer(sentences, return_tensors=\"pt\", padding=True)\n",
        "tokens_pt = tokenizer(temp['prepro'].to_list(), return_tensors=\"pt\", padding=True)\n",
        "\n",
        "outputs = model(**tokens_pt)\n",
        "pooler_output = outputs.pooler_output\n",
        "\n",
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(preprocessing.normalize(pooler_output),preprocessing.normalize(pooler_output)),vmin=0,vmax=1,cmap=\"YlGnBu\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fac361228d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "UNwhPyVJvoXb",
        "outputId": "67fe4676-a7b1-42bf-94ae-2fda20bb6968"
      },
      "source": [],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f025cd972d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hf1dMQ9eD-eH",
        "outputId": "ebe0fc21-b62b-4bf7-f8b1-49ce1444e934"
      },
      "source": [
        "np.inner(pooler_output,pooler_output)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[126.636925,  80.4872  , 113.7118  ,  89.03408 ,  29.241615,\n",
              "         68.68401 , 102.90591 ],\n",
              "       [ 80.4872  ,  60.198616,  75.23463 ,  60.9293  ,  27.185085,\n",
              "         48.540928,  68.02994 ],\n",
              "       [113.7118  ,  75.23463 , 114.5472  ,  86.57979 ,  28.321049,\n",
              "         64.43546 ,  96.296906],\n",
              "       [ 89.03408 ,  60.9293  ,  86.57979 ,  72.562515,  27.899416,\n",
              "         49.463776,  79.38815 ],\n",
              "       [ 29.241615,  27.185085,  28.321049,  27.899416,  26.448711,\n",
              "         22.184502,  29.187   ],\n",
              "       [ 68.68401 ,  48.540928,  64.43546 ,  49.463776,  22.184502,\n",
              "         55.18467 ,  58.466663],\n",
              "       [102.90591 ,  68.02994 ,  96.296906,  79.38815 ,  29.187   ,\n",
              "         58.466663,  94.17383 ]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IdpgLxwoxvZu"
      },
      "source": [
        "## -- SBERT? (2 модели)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3utWTH0NY0AU"
      },
      "source": [
        "Странные рез-ты"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XiHNW0_X6GrM"
      },
      "source": [
        "\n",
        "\n",
        "*   pickle -- чтобы вычислить и загрузить потом\n",
        "*   Paraphrase Mining -- считает быстрее, нежели просто для каждой пары находить кос.сходство\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oTcWTPAxx7Ar"
      },
      "source": [
        "%%capture\n",
        "!pip install -U sentence-transformers"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jB0p_KgqyLdT"
      },
      "source": [
        "from sentence_transformers import SentenceTransformer\n",
        "model = SentenceTransformer('distiluse-base-multilingual-cased-v1')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NRFMlWDAwxRp"
      },
      "source": [
        "model_sbert = model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "ptfbODtHxxA6",
        "outputId": "3367d648-f593-48a7-bdc3-27803adf3a09"
      },
      "source": [
        "sentences = ['This framework generates embeddings for each input sentence',\n",
        "    'Sentences are passed as a list of string.', \n",
        "    'The quick brown fox jumps over the lazy dog.', \n",
        "    'Быстрая коричневая лиса перепрыгнула через ленивую собаку.',\n",
        "    'Предложения передаются в виде списка строк.', 'Random forest','Случайный лес']\n",
        "\n",
        "sentence_embeddings = model.encode(sentences, convert_to_tensor=True)\n",
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f99f5f663d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVkAAAD5CAYAAACNmUaBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAR7klEQVR4nO3df2wU9brH8c9sCwuVroB3ty3qPRqh1N/KKWLujYjF+vOa2KQC/iFtLCZE2oqBEjUVipRSrtb+0WqOUa8siQrJJkSvCdEUogISjAZvLiAiPcVYzVJii6un0ALd+5d7qW2X3fZ8d2an71cyCTMsMw+EfPL0me/MWtFoNCoAgBEeuwsAADcjZAHAIEIWAAwiZAHAIEIWAAwiZAHAoEy7CwAApzp27Jhef/11XXfddaqqqpIkDQwMqK6uTpJ0++23q6SkJO456GQBjDuRSESdnZ1DtkgkMuhz+fn5Wr169aBjX3/9tW699Va99NJLOnDgwCWvZbST/cutDSZPPyo7P7vF7hJGNNOXa3cJw8q0JttdwrC+Pd1hdwkjunFavt0lpKGx/ZtN/tfHE/7sf9b8m1pbW4ccr6ysjHWsIzl58qTy8vIkSZZlXfJajAsAuIJlJf6DeVlZ2bA/5vt8vkv+2ZycHP34448JX4uQBeAKVhLTT5/Pl1CghsNhNTc3q6OjQ16vVx6PR08++aQ++ugjHTp0SPPmzbvkOQhZAK6QTCebqNzcXDU1NQ05vn79+oTPQcgCcAUTIfvPQMgCcAXLyrC7hGERsgBcgU4WAAwiZAHAoGRWF6QSIQvAFehkAcAgQhYADPKwugAAzKGTBQCDCFkAMCgtQ/bYsWN6//331d/fL6/Xq8WLF2v27Nmpqg0AkpCGIfv2229rw4YNmjhxovr7+7Vu3Tpt2rQpVbUBQMI8Hmf+YB43+qPRqE6dOqWBgQGdOnVK0Wg0VXUBQFIseRLeUilu9K9atUqhUEgnT55Ubm6unn322VTVBQBJScuZbE5OjlasWJGqWgBg1BL5Khg7OHOIAQBJSstOFgDSBS+IAQCDnLq6wJlVAUCS6GQBwCRmsgBgDje+AMAglnABgEHMZAHAIMvDS7sBwBxnNrJmQ3bnZ7eYPP2oLfz3/XaXMKzj/3Of3SUMz6H/eR/62yS7SxhRx3Pn7S5hWB7LxX0VM1lncGrAAhgjQhYADHLoT1yELABXiHroZAHAHEIWAAxiJgsABjkzYwlZAC7BuAAADGJcAAAGZRCyAGCOMzOWkAXgDlHGBQBgUBI3viKRiCKRyJDjPp9PPp8vth8Oh9XU1CRJqqioUEFBgSSpvr5eZ8+e1fnz59XQ0CCPZ+THzRz6IBoAJMlKfAsGg1q4cOGQLRgMDjplKBRSdXW1NmzYoG3btsWOnz9/XvX19Zo6dap6e3vjlkUnC8AdkhgXlJWVqaSkZMjxi7tYSerq6lJeXp4yMzPV19cXO56VlaVly5YpJydHU6ZMiXstQhaAOySxuuDPY4GRBAIBhcNh+f1+eb1eSVJ3d7f6+/v11ltvqaGhQb///nvcoE1qXPDxxx8n83EASB3LSnxLUGlpqVpaWrRu3TotWbJE9fX1mjZtmnp7e7Vu3Tr9+uuvysrKinuOuJ1sa2vroP1PP/1U999/f8IFAkDKGFhdkJubq82bN8f2a2trJUkNDQ0JnyNuyLa3t2v16tWx/RMnTiRZIgCkiENv48cN2aefflpXXnllbL+mpsZ4QQAwKum4TnbWrFmD9nNycowWAwCjFeWxWgAwKB07WQBIG87MWEIWgEvwPlkAMIhxAQAY5MyMJWQBuESmMxfKErIAXCFKJwsABnHjCwAM4sYXABg0HjvZmb5ck6cflR/+t0TnBv5hdxnDuuGv++wuYVie+2+xu4RhtTdebXcJI+q7cNruEoblsZzbV3kzxngCZ973Gn+drFMDFsAYZTgzZcddyAJwJ76tFgBMcmYjS8gCcInxeOMLAFKGcQEAGMRLuwHAnCjjAgAwiJAFAIOYyQKAQSzhAgCD6GQBwCBe2g0A5vBYLQCY5MxGlpAF4BIO7WTjZn97e7u2bNmiU6dOSZJ27dqVkqIAIGkeK/EtlWXF+82Wlhbdeeedam1t1ZEjR7Rnz55U1QUAyUnHkM3OzlZBQYHq6ur0wQcf6Ntvv01VXQCQlGiGlfCWSnFDtqioSJJkWZaef/55PfzwwykpCgCSZlmJbykU98bXPffcM2h/6dKlRosBgFHj3QUAYJAzM5aQBeAOHtbJAoA5hCwAGGQZuKEVDofV1NQkSaqoqFBBQYEkqa2tTfv371dWVpZWrVoV9xyELABXSCZjI5GIIpHIkOM+n08+ny+2HwqFVF1dLb/fr8bGRtXV1enChQvavn278vPz5ff7L3kthzbYAJCcZFZwBYNBLVy4cMgWDAYHnbOrq0t5eXmaNGmS+vr6JEnd3d0aGBhQTU2Nenp69MMPP8Sti04WgCtYSbSMZWVlKikpGXL84i5WkgKBgMLhsPx+v7xeryTp8ssvj3WwPp9Pvb29ca9FyAJwhWTGBX8eC4yktLRUzc3NsixL5eXlqq+vV21trQoKCrRp0yadO3dO5eXl8euKRqPRxEtLTv/AV6ZOPWrnBv5hdwkjuuGv++wuYVie+2+xu4Rhfd94td0ljMip/888lnP7Km/GHWP68zf81+cJf/bIk/PHdK1kOPdfHACS4NA3HRKyANzBxBKufwajIZtpTTZ5+tFx8HoKp/5YPmHPj3aXMCyPda3dJaSdC9FzdpdgTDI3vlKJThaAKzi0kSVkAbgDj9UCgEEOfdMhIQvAHRgXAIBBhCwAGGQ5dF5AyAJwBTpZADCI1QUAYJBDpwWELAB3YFwAAAbxWC0AGEQnCwAGjcu3cAFAqjh1dUHcso4cOaJQKKSOjg5t2LBB33zzTarqAoCkJPNFiqkUN2RbWlp04403as2aNVqxYoW2bduWqroAICkeK/EtleKOCwKBgK6//no99thjmj59uqZMmZKqugAgKWm5TvaRRx6RJC1atEiSNG/ePPMVAcAoeCxj3wk7JnFDtrCwcNB+cXGx0WIAYLQy07GTBYB0kZadLACki7ScyQJAunDoMllCFoA70MkCgEEWM1kAMIfVBQBgEKsLAMAgZrIAYBCrCwDAIDpZADBoXM5kvz3dYfL0o/LQ3ybZXcKI2huvtruEYXmsa+0uYVhTr2u2u4QRRf6+xu4Sxh1WFwCAQeOykwWAVGEmCwAGEbIAYBBLuADAoExP4jPZSCSiSCQy5LjP55PP54vth8NhNTU1SZIqKipUUFAQ+72NGzfK5/Opqqoqfl0JVwUADpZMJxsMBtXa2jrkeGVl5aDQDIVCqq6ult/vV2Njo+rq6iRJX3zxha655hp1d3df8lqELABXSGYmW1ZWppKSkiHHL+5iJamrq0t5eXnKzMxUX1+fJOnMmTPat2+fHn/8ce3YseOS1yJkAbhCMq86/PNYYCSBQEDhcFh+v19er1eSdPjwYf3yyy967bXX1N7erkWLFiknJ2fEcxCyAFzBxOqC0tJSNTc3y7IslZeXq76+XrW1tSosLFRnZ6d27NgRN2AlQhaAS5hYXZCbm6vNmzfH9mtra2O/vuqqqy5500siZAG4RDKrC1KJkAXgCjyMAAAGZdhdwAgIWQCukJYviPn5559jvw6FQiotLdWMGTOMFwUAyUrLcUF5ebmKi4s1efJk7du3Tx6PR5WVlamqDQAS5tSQjbvqYdu2bZowYYJuvvlm3XbbbQQsAMea4El8S6W4l5s+fbpWrlwpy7KUlZWVqpoAIGkeK5rwlkoJ3fiaP3++5s+fb7oWABg1p44LWF0AwBVYwgUABtHJAoBBE3isFgDMoZMFAIMIWQAwiJAFAIMy0vHdBQCQLvhKcAAwKNOhKUvIAnAFxgUAYNC4vPF147R8k6cflY7nzttdwoj6Lpy2u4S0Evn7GrtLGNHMwp12lzC8GVPtrmBExz8cW16My5AFgFQhZAHAIB6rBQCDHLq4gJAF4A6MCwDAoAxCFgDMScuvBAeAdMG4AAAMyiRkAcAci5AFAHMcmrGELAB3oJMFAIN4GAEADLJYwgUA5qTlEq729nZdccUVCgaDOnPmjBYvXqxrr702VbUBQMIcmrHxQ3br1q2KRqNaunSppk2bpo0bN+rVV19NVW0AkLC07GQ7OjqUlZWlmTNnSpKmTJmSkqIAIFkmMjYcDqupqUmSVFFRoYKCAv3222/auHGjLMvS7NmzVV5eHvcccUO2srJS1kXrIu6+++6xVw0ABiSzhCsSiSgSiQw57vP55PP5YvuhUEjV1dXy+/1qbGxUXV2dsrOz1djYKElas+bS384RN2TvuOOOQfsLFy5M6C8AAKmWzBKuYDCo1tbWIccrKytVVVUV2+/q6lJeXp4yMzPV19c36LNtbW2aO3fuJa/F6gIArpDMTLasrEwlJSVDjl/cxUpSIBBQOByW3++X1+uNHd+zZ4++++47rVix4pLXsqLRqMHFZcfMnXqUBqJ8kaJbTM78F7tLGBFfpJi84x+WjenPf//rRwl/dtbl/5HQ58LhsJqbm2VZlsrLyxUKhbR8+XI9+uijKioqUnZ2tmpqauKeg5B1EEI2OYTsKLg4ZI9H/jvhz870PTKmayWDcQEAV3DoCi5CFoA78IIYADCI7/gCAIMcmrGELAB3YFwAAAY5NGMJWQDukJYviAGAdOHQjB1/IeuxnPtXdmptF6Ln7C4h/Th00b/1e7/dJRjj4ZsRAMAcbnwBgEEOzVhCFoA78G21AGAQ4wIAMMhyaC9LyAJwBcsiZAHAIGfOCwhZAK5gEbIAYBIhCwDGMJMFAINYXQAABjGTBQCj6GQBwBjLoY98xQ3ZXbt2acaMGXrnnXfk8Xi0YMECPfDAA6mqDQCSkIYhu3//fmVkZGjz5s2yLEvr168nZAE4UlrOZC+77DL19PRo586dys7O1vnz51NVFwAkxVKG3SUMK+6k+JlnntFdd92lzs5O9fT06IUXXkhVXQCQFMuyEt5SKW4n6/F4VFxcnKpaAGAM0nBcAADpgocRAMAoOlkAMIZ3FwCAQYwLAMAoxgUAYExaPowAAOkiLd9dAADpg5ksABjDjS8AMIhxAQAYlXgnG4lEFIlEhhz3+Xzy+Xyx/XA4rKamJklSRUWFCgoKJEkvv/yyzp49q9zcXD311FNxr2U4ZPPNnt5lvM58iRBG4fiH/N9PNUuzE/5sMNii1tbWIccrKytVVVUV2w+FQqqurpbf71djY6Pq6ur0008/yev1qqamRmvXrlV/f78mTpw44rXoZAGMO2VlZSopKRly/OIuVpK6urqUl5enzMxM9fX1DTomSdOmTdPp06cVCARGvBYhC2Dc+fNYYCSBQEDhcFh+v19erzd2bO/evZKknp4eTZ06Ne45rGg0Gh17yQDgPuFwWM3NzbIsS+Xl5QqFQqqtrdUrr7yi/v5+BQIBLVu2LO45CFkAMMiZC8sAwCUIWQAwiJAFAIMIWQAwiJAFAIMIWQAwyPEhGw6HVVNTo5qaGh09etTucgY5duyYVq5cqZaWFrtLGWT37t2qra1VdXW1vv/+e7vLifnqq6+0du1aLV++XIcOHbK7nEG6u7tVXFyszs5Ou0sZ5MCBA3riiSe0du1aHT9+3O5yYi5cuKDm5mZt2LBBbW1tdpfjaI5/4mu4Z4edIj8/X6tXr9aOHTvsLmWQoqIiFRUV6ejRo9q7d69mzZpld0mSpMLCQhUWFurIkSM6fPiwbrrpJrtLigkGg7r33nvtLmMIy7KUlZWlgYEBTZ8+3e5yYnbt2qXu7m5NmjQp9ogphuf4TvaP54QnTZoUe3YYlzYwMKD33ntPDz30kN2lDLJ9+3a9+OKLmjNnjt2lxHz22WeaM2eOsrKy7C5liLlz5+qNN97Q0qVLtXXrVrvLiTlx4oTmzZun5557Tlu2bLG7HEdzfMj+8exwX19f7NlhxBeNRtXY2KhFixYpJyfH7nIGWbx4sd588029++67dpcSc/DgQe3du1d79uxRKBSyu5xB/nhH6vTp09Xb22tzNf8vEAgoOztbGRkZjn2Pq1M4flxQWlo66NlhJ/njueaOjg7NnDlTDz74oN0lSZK2bt2qgwcP6uzZswqHw475MfiTTz7Rl19+qUgkoiVLlthdTszKlSslSS0tLcO+mclObW1t+vzzzxWJRAa9gs9u9913nxoaGrR7924tWLDA7nIcjXcXAIBBjh8XAEA6I2QBwCBCFgAMImQBwCBCFgAMImQBwKD/A79gBuZwfEuGAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qrCIeEcx7Sw0"
      },
      "source": [
        "def vecFuncSBert(headers):\n",
        "  return model_sbert.encode(headers, convert_to_tensor=True)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZT1WSycLw_8l"
      },
      "source": [
        "item_vecs_sb = model_sbert.encode(dp_items['name'].values, convert_to_tensor=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "TDxZwW0UwkZ5",
        "outputId": "f7fb8723-6b8b-486d-cbec-887f884f86f0"
      },
      "source": [
        "res = itemDiscSim(12223,vecFuncSBert,item_vecs_sb)\n",
        "res\n",
        "# игра 0.9, хотя есть 100%соответствие в теме"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>similarity</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>17159</th>\n",
              "      <td>17923</td>\n",
              "      <td>Игра</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.901437</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>627</th>\n",
              "      <td>867</td>\n",
              "      <td>Методы построения</td>\n",
              "      <td>64</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>0.704804</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7509</th>\n",
              "      <td>8058</td>\n",
              "      <td>Методы разработки архитектуры</td>\n",
              "      <td>7</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>0.694217</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6131</th>\n",
              "      <td>6558</td>\n",
              "      <td>Методы конструирования технологической оснастки</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>36.0</td>\n",
              "      <td>0.684116</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18306</th>\n",
              "      <td>18885</td>\n",
              "      <td>Прототипирование игровых механик</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>17.0</td>\n",
              "      <td>0.653086</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id  ... similarity\n",
              "17159  17923  ...   0.901437\n",
              "627      867  ...   0.704804\n",
              "7509    8058  ...   0.694217\n",
              "6131    6558  ...   0.684116\n",
              "18306  18885  ...   0.653086\n",
              "\n",
              "[5 rows x 7 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KeC7cTacTzwv"
      },
      "source": [
        "\n",
        "\n",
        "---\n",
        "Проверим на 2-й модели. В док-ии написано, что хоть языков больше, но точность на 15-ти изначальных (из 1-й модели) ниже\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68,
          "referenced_widgets": [
            "11d040b31917490da154996fb715e446"
          ]
        },
        "id": "6_GT620J9PwQ",
        "outputId": "57ca2da4-aef9-48b8-fb9a-024604a8dde7"
      },
      "source": [
        "model_v2 = SentenceTransformer('distiluse-base-multilingual-cased-v2')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "11d040b31917490da154996fb715e446",
              "version_major": 2,
              "version_minor": 0
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, max=503702349.0), HTML(value='')))"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nr-VRK31zu3W"
      },
      "source": [
        "def vecFuncSBert2(headers):\n",
        "  return model_v2.encode(headers, convert_to_tensor=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "bgYf0AVOSV-k",
        "outputId": "60a91c06-965b-4eae-8d09-6f96095b7e17"
      },
      "source": [
        "sentences = ['This framework generates embeddings for each input sentence',\n",
        "    'Sentences are passed as a list of string.', \n",
        "    'The quick brown fox jumps over the lazy dog.', \n",
        "    'Быстрая коричневая лиса перепрыгнула через ленивую собаку.',\n",
        "    'Предложения передаются в виде списка строк.', 'Random forest','Случайный лес']\n",
        "\n",
        "sentence_embeddings = model_v2.encode(sentences, convert_to_tensor=True)\n",
        "sns.set(font_scale=0.6)\n",
        "sns.heatmap(np.inner(sentence_embeddings,sentence_embeddings),vmin=0,vmax=1,cmap=\"YlGnBu\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f99e9aa3f10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9xT4ryL_zn8N"
      },
      "source": [
        "item_vecs_sb = model_v2.encode(dp_items['name'].values, convert_to_tensor=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        },
        "id": "Zth8J0g4zn8O",
        "outputId": "d6dfe4ed-45c5-4c7b-8a99-0b85cf2196f0"
      },
      "source": [
        "res = itemDiscSim(12223,vecFuncSBert2,item_vecs_sb,threshold=0.4)\n",
        "res"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>similarity</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>7509</th>\n",
              "      <td>8058</td>\n",
              "      <td>Методы разработки архитектуры</td>\n",
              "      <td>7</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>0.536467</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1996</th>\n",
              "      <td>2267</td>\n",
              "      <td>Методики описания системной архитектуры</td>\n",
              "      <td>11</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>0.502733</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9361</th>\n",
              "      <td>9956</td>\n",
              "      <td>Методики описания архитектуры предприятия</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>18.0</td>\n",
              "      <td>0.491816</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1303</th>\n",
              "      <td>1554</td>\n",
              "      <td>Требования к структуре хорошей истории</td>\n",
              "      <td>9</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>0.487548</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>627</th>\n",
              "      <td>867</td>\n",
              "      <td>Методы построения</td>\n",
              "      <td>64</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>0.463450</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17817</th>\n",
              "      <td>18492</td>\n",
              "      <td>Структурный подход к нарративу</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "      <td>0.459777</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17159</th>\n",
              "      <td>17923</td>\n",
              "      <td>Игра</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0.453946</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>490</th>\n",
              "      <td>633</td>\n",
              "      <td>История</td>\n",
              "      <td>70</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>16.0</td>\n",
              "      <td>0.451369</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18306</th>\n",
              "      <td>18885</td>\n",
              "      <td>Прототипирование игровых механик</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>17.0</td>\n",
              "      <td>0.443785</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1312</th>\n",
              "      <td>1560</td>\n",
              "      <td>Теория игр</td>\n",
              "      <td>29</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>11.0</td>\n",
              "      <td>0.442421</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id                                       name  ...  domain_id similarity\n",
              "7509    8058              Методы разработки архитектуры  ...       17.0   0.536467\n",
              "1996    2267    Методики описания системной архитектуры  ...       17.0   0.502733\n",
              "9361    9956  Методики описания архитектуры предприятия  ...       18.0   0.491816\n",
              "1303    1554     Требования к структуре хорошей истории  ...       30.0   0.487548\n",
              "627      867                          Методы построения  ...       14.0   0.463450\n",
              "17817  18492             Структурный подход к нарративу  ...       62.0   0.459777\n",
              "17159  17923                                       Игра  ...        NaN   0.453946\n",
              "490      633                                    История  ...       16.0   0.451369\n",
              "18306  18885           Прототипирование игровых механик  ...       17.0   0.443785\n",
              "1312    1560                                 Теория игр  ...       11.0   0.442421\n",
              "\n",
              "[10 rows x 7 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NuiGRd_8cyin"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_xsS6ZR1vka6"
      },
      "source": [
        "---\n",
        "# Пререквизиты"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rOcTv6EWYrEP"
      },
      "source": [
        "Взяли рез-ты всех РПД из всех Уч.Планов, находящихся в списке ранее текущего. И что дальше?\n",
        "\n",
        "Можно у текущей РПД найти рез-ты + узнать их самые частые предм.области. Затем из чужих рез-в выбрать из самой частой предм.области.\n",
        "\n",
        "Или в зависимости от модуля (например, универсальный или элективы) как-то по-разному?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NbgZUPgwxUU4",
        "outputId": "074c992e-377f-4c19-f218-0704748c4beb"
      },
      "source": [
        "#учебные планы\n",
        "wp_iaplan.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>year</th>\n",
              "      <th>academic_plan_id</th>\n",
              "      <th>period_of_study</th>\n",
              "      <th>op_isu_id</th>\n",
              "      <th>ns_id</th>\n",
              "      <th>ap_isu_id</th>\n",
              "      <th>educational_profile</th>\n",
              "      <th>language</th>\n",
              "      <th>qualification</th>\n",
              "      <th>title</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>531</td>\n",
              "      <td>2019</td>\n",
              "      <td>473</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6960</td>\n",
              "      <td>3893</td>\n",
              "      <td>11687</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Нейротехнологии и программирование</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>493</td>\n",
              "      <td>2018</td>\n",
              "      <td>435</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6215</td>\n",
              "      <td>3108</td>\n",
              "      <td>10555</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Информатика и программирование</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>528</td>\n",
              "      <td>2020</td>\n",
              "      <td>470</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6959</td>\n",
              "      <td>3892</td>\n",
              "      <td>13280</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Компьютерные технологии в дизайне</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>526</td>\n",
              "      <td>2018</td>\n",
              "      <td>468</td>\n",
              "      <td>NaN</td>\n",
              "      <td>6959</td>\n",
              "      <td>3892</td>\n",
              "      <td>10559</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Компьютерные технологии в дизайне</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>523</td>\n",
              "      <td>2019</td>\n",
              "      <td>465</td>\n",
              "      <td>NaN</td>\n",
              "      <td>7140</td>\n",
              "      <td>4071</td>\n",
              "      <td>11678</td>\n",
              "      <td>NaN</td>\n",
              "      <td>ru</td>\n",
              "      <td>bachelor</td>\n",
              "      <td>Мобильные и сетевые технологии</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    id  year  ...  qualification                               title\n",
              "0  531  2019  ...       bachelor  Нейротехнологии и программирование\n",
              "1  493  2018  ...       bachelor      Информатика и программирование\n",
              "2  528  2020  ...       bachelor   Компьютерные технологии в дизайне\n",
              "3  526  2018  ...       bachelor   Компьютерные технологии в дизайне\n",
              "4  523  2019  ...       bachelor      Мобильные и сетевые технологии\n",
              "\n",
              "[5 rows x 11 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "egvcmtum4WeB"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EswL6Payun9E"
      },
      "source": [
        "объедияем таблицы"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4yZBMPI7w6z6",
        "outputId": "5745581a-c048-44fe-fc5d-a728607650bb"
      },
      "source": [
        "# распределение РПД по модулям в уч.плане\n",
        "wp_in_bmodules = wp_wp_disc_bm.merge(wp_wp_fos, left_on='id', right_on='work_program_change_in_discipline_block_module_id').iloc[:,:-2]\n",
        "wp_in_bmodules.head(4)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id_x</th>\n",
              "      <th>discipline_block_module_id</th>\n",
              "      <th>change_type</th>\n",
              "      <th>semester_hour</th>\n",
              "      <th>credit_units</th>\n",
              "      <th>code</th>\n",
              "      <th>subject_code</th>\n",
              "      <th>id_y</th>\n",
              "      <th>work_program_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>32043</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 3.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1</td>\n",
              "      <td>41411</td>\n",
              "      <td>2623</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>32044</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 2.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2</td>\n",
              "      <td>41412</td>\n",
              "      <td>2624</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>32045</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.0, 1.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3</td>\n",
              "      <td>41413</td>\n",
              "      <td>2625</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>32046</td>\n",
              "      <td>8462</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 0, 0, 0, 3.0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4</td>\n",
              "      <td>41414</td>\n",
              "      <td>2626</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    id_x  discipline_block_module_id  ...   id_y  work_program_id\n",
              "0  32043                        8461  ...  41411             2623\n",
              "1  32044                        8461  ...  41412             2624\n",
              "2  32045                        8461  ...  41413             2625\n",
              "3  32046                        8462  ...  41414             2626\n",
              "\n",
              "[4 rows x 9 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vvUHPVdu1A_Z",
        "outputId": "812cd4a3-3751-4985-f218-e093cfa1cc9c"
      },
      "source": [
        "# распределение РПД по модулям и блокам в уч.плане\n",
        "wp_in_ap = wp_in_bmodules.merge(wp_disc_bm.iloc[:,:-2], left_on='discipline_block_module_id',right_on='id')\n",
        "wp_in_ap.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id_x</th>\n",
              "      <th>discipline_block_module_id</th>\n",
              "      <th>change_type</th>\n",
              "      <th>semester_hour</th>\n",
              "      <th>credit_units</th>\n",
              "      <th>code</th>\n",
              "      <th>subject_code</th>\n",
              "      <th>id_y</th>\n",
              "      <th>work_program_id</th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>descipline_block_id</th>\n",
              "      <th>order</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>32043</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 3.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1</td>\n",
              "      <td>41411</td>\n",
              "      <td>2623</td>\n",
              "      <td>8461</td>\n",
              "      <td>Универсальный модуль</td>\n",
              "      <td>1695</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>32044</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0, 0, 2.0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>2</td>\n",
              "      <td>41412</td>\n",
              "      <td>2624</td>\n",
              "      <td>8461</td>\n",
              "      <td>Универсальный модуль</td>\n",
              "      <td>1695</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>32045</td>\n",
              "      <td>8461</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>1.0, 1.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>3</td>\n",
              "      <td>41413</td>\n",
              "      <td>2625</td>\n",
              "      <td>8461</td>\n",
              "      <td>Универсальный модуль</td>\n",
              "      <td>1695</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "    id_x  discipline_block_module_id  ...  descipline_block_id  order\n",
              "0  32043                        8461  ...                 1695      2\n",
              "1  32044                        8461  ...                 1695      2\n",
              "2  32045                        8461  ...                 1695      2\n",
              "\n",
              "[3 rows x 13 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wkzwe6Ld-tVD"
      },
      "source": [
        "outcomes_dom = wp_outcomes.merge(dp_items,left_on='item_id',right_on='id')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sJ36w7ZDcjNp"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Wxokabjr7iXD"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UXriwy9YGUmO"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YTNwvXqYwkYi"
      },
      "source": [
        "# какие РПД в учебном плане\n",
        "def wpInAplan(aplan_id = 435):\n",
        "  \n",
        "  # нужные блоки\n",
        "  blocks = wp_disc_b[wp_disc_b['academic_plan_id']  == aplan_id].id\n",
        "  # нужные модули в блоках + РПД\n",
        "  bmodules_and_wp = wp_in_ap[wp_in_ap['descipline_block_id'].isin(blocks)]\n",
        "\n",
        "  bmodules_and_wp = bmodules_and_wp.loc[:,['id_x','discipline_block_module_id',\n",
        "                                           'subject_code','work_program_id','id',\n",
        "                                           'name','descipline_block_id','order']]\n",
        "  #сортируем по порядку модулей и по порядку РПД\n",
        "  return bmodules_and_wp.sort_values(by=['order','subject_code'])  \n",
        "\n",
        "\n",
        "# список РПД, которые стоят в уч. плане раньше текущей\n",
        "def ealierWpInAplan(wp_id = 2627, aplan_id=435):\n",
        "  needed_aplan = wpInAplan(aplan_id)\n",
        "  # ищем место РПД в списке\n",
        "  wp_index = needed_aplan[needed_aplan.work_program_id==wp_id].index[0]\n",
        "\n",
        "  return needed_aplan.iloc[:wp_index,:].work_program_id  "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d4aKkP1qZOpr"
      },
      "source": [
        "def earlierWpRes(wp_id = 2623, to_return=10):\n",
        "  '''\n",
        "    Поиск учебных сущностей, которые часто встречаются в РПД,\n",
        "    стоящих ранее в связанных учебных планах.\n",
        "    Аргументы:\n",
        "        wp_id -- id дисциплины;\n",
        "        to_return -- кол-во рекомендаций.\n",
        "    Возвращает: таблицу с уч.сущностями\n",
        "  '''\n",
        "  # в каких уч.планах есть РПД:\n",
        "    # в каких блоках встречается РПД\n",
        "  blocks = wp_in_ap[wp_in_ap.work_program_id==wp_id].descipline_block_id.unique()\n",
        "    # в каких уч.планах находятся блоки\n",
        "  ap_ids = wp_disc_b[wp_disc_b.id.isin(blocks)].academic_plan_id.unique()\n",
        "\n",
        "  # составляем список РПД, которые находятся раньше, из всех уч.планов\n",
        "  earlier_wps = np.array([])\n",
        "  for ap in ap_ids:\n",
        "    earlier_wps = np.append(earlier_wps, ealierWpInAplan(wp_id, ap).values)\n",
        "  \n",
        "  # составляем список всех результатов для всех РДП \n",
        "  earlier_res, res_count = np.unique(wp_outcomes[wp_outcomes.workprogram_id.isin(\n",
        "                                      earlier_wps)].item_id.values, return_counts=True)\n",
        "  \n",
        "  # самые частые результаты\n",
        "  freq_res = pd.DataFrame([earlier_res,res_count]).T[0].values\n",
        "  res = dp_items[dp_items.id.isin(freq_res)].copy()\n",
        "\n",
        "  # записываем абсолютные и относительные величины частоты\n",
        "  len = res.shape[0]\n",
        "  res['val'] = 0\n",
        "  res['perc'] = 0\n",
        "  for item,value in zip(earlier_res,res_count):\n",
        "    res.loc[res['id']==item,'val'] = value\n",
        "    res.loc[res['id']==item,'perc'] = value/len\n",
        "  \n",
        "  return res.nlargest(to_return,'perc')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FiL-Fx-vZOra",
        "outputId": "6ce0f33a-1654-44b3-9475-ffa220be89e8"
      },
      "source": [
        "earlierWpRes(12186,to_return=15)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>val</th>\n",
              "      <th>perc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>6733</th>\n",
              "      <td>7203</td>\n",
              "      <td>Правовая охрана интеллектуальной собственности</td>\n",
              "      <td>18</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>7</td>\n",
              "      <td>0.017241</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17738</th>\n",
              "      <td>856</td>\n",
              "      <td>Нейронные сети</td>\n",
              "      <td>232</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>6</td>\n",
              "      <td>0.014778</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6763</th>\n",
              "      <td>7201</td>\n",
              "      <td>Научное исследование</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>5</td>\n",
              "      <td>0.012315</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1232</th>\n",
              "      <td>13255</td>\n",
              "      <td>Бизнес-модель</td>\n",
              "      <td>0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2179</th>\n",
              "      <td>2405</td>\n",
              "      <td>Организация научных исследований</td>\n",
              "      <td>16</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6720</th>\n",
              "      <td>7199</td>\n",
              "      <td>Международные наукометрические базы данных</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6721</th>\n",
              "      <td>7200</td>\n",
              "      <td>Международные научные парнерства</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6725</th>\n",
              "      <td>7202</td>\n",
              "      <td>Методологические основы научного знания</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6727</th>\n",
              "      <td>7204</td>\n",
              "      <td>Управление в сфере науки в России</td>\n",
              "      <td>6</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15433</th>\n",
              "      <td>7308</td>\n",
              "      <td>Проектная деятельность</td>\n",
              "      <td>22</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>30.0</td>\n",
              "      <td>4</td>\n",
              "      <td>0.009852</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2204</th>\n",
              "      <td>2513</td>\n",
              "      <td>Автоматизация производств</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.007389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2233</th>\n",
              "      <td>2509</td>\n",
              "      <td>Технологии анализа данных</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.007389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2234</th>\n",
              "      <td>2510</td>\n",
              "      <td>Алгоритмы анализа данных</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.007389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2236</th>\n",
              "      <td>2516</td>\n",
              "      <td>Умный город</td>\n",
              "      <td>23</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.007389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2243</th>\n",
              "      <td>2512</td>\n",
              "      <td>Безопасность</td>\n",
              "      <td>10</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>14.0</td>\n",
              "      <td>3</td>\n",
              "      <td>0.007389</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id                                            name  ...  val      perc\n",
              "6733    7203  Правовая охрана интеллектуальной собственности  ...    7  0.017241\n",
              "17738    856                                  Нейронные сети  ...    6  0.014778\n",
              "6763    7201                            Научное исследование  ...    5  0.012315\n",
              "1232   13255                                   Бизнес-модель  ...    4  0.009852\n",
              "2179    2405                Организация научных исследований  ...    4  0.009852\n",
              "6720    7199      Международные наукометрические базы данных  ...    4  0.009852\n",
              "6721    7200                Международные научные парнерства  ...    4  0.009852\n",
              "6725    7202         Методологические основы научного знания  ...    4  0.009852\n",
              "6727    7204               Управление в сфере науки в России  ...    4  0.009852\n",
              "15433   7308                          Проектная деятельность  ...    4  0.009852\n",
              "2204    2513                       Автоматизация производств  ...    3  0.007389\n",
              "2233    2509                       Технологии анализа данных  ...    3  0.007389\n",
              "2234    2510                        Алгоритмы анализа данных  ...    3  0.007389\n",
              "2236    2516                                     Умный город  ...    3  0.007389\n",
              "2243    2512                                    Безопасность  ...    3  0.007389\n",
              "\n",
              "[15 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zXE62H3I9F99"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zKS83NUncT7I"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3FEueIyRcUIF"
      },
      "source": [
        "def earlierWpResDomain(wp_id = 2623, to_return=10):\n",
        "  '''\n",
        "    Поиск учебных сущностей, которые часто встречаются в РПД,\n",
        "    стоящих ранее в связанных учебных планах И в той же пр.области,\n",
        "    что и результаты заданной РПД.\n",
        "    Аргументы:\n",
        "        wp_id -- id дисциплины;\n",
        "        to_return -- кол-во рекомендаций.\n",
        "    Возвращает: таблицу с уч.сущностями\n",
        "  '''\n",
        "  # в каких уч.планах есть РПД:\n",
        "    # в каких блоках встречается РПД\n",
        "  blocks = wp_in_ap[wp_in_ap['work_program_id']==wp_id]['descipline_block_id'].unique()\n",
        "    # в каких уч.планах находятся блоки\n",
        "  ap_ids = wp_disc_b[wp_disc_b['id'].isin(blocks)]['academic_plan_id'].unique()\n",
        "\n",
        "  # составляем список РПД, которые находятся раньше, из всех уч.планов\n",
        "  earlier_wps = np.array([])\n",
        "  for ap in ap_ids:\n",
        "    earlier_wps = np.append(earlier_wps, ealierWpInAplan(wp_id, ap).values)\n",
        "\n",
        "  # результаты самой РПД\n",
        "  orig_res = wp_outcomes[wp_outcomes['workprogram_id'] == wp_id]['item_id'].values\n",
        "  # предм.области результатов\n",
        "  orig_domain = dp_items[dp_items['id'].isin(orig_res)]['domain_id'].unique()\n",
        "\n",
        "  # составляем список всех результатов для всех РДП \n",
        "  earlier_res, res_count = np.unique(wp_outcomes[wp_outcomes['workprogram_id'].isin(\n",
        "                                      earlier_wps)]['item_id'].values, return_counts=True)\n",
        "  \n",
        "  # возвращаем только рез-ты, которые из тех же предм.областей, что и оригинальная РПД\n",
        "  item_ids = outcomes_dom[(outcomes_dom['workprogram_id'].isin(earlier_wps) & \n",
        "                (outcomes_dom['domain_id'].isin(orig_domain)))\n",
        "              ].value_counts('item_id').index#[:to_return]\n",
        "  len = item_ids.shape[0]\n",
        "  res = dp_items[dp_items.id.isin(item_ids)].copy()\n",
        "\n",
        "  # записываем абсолютные и относительные величины частоты\n",
        "  len = res.shape[0]\n",
        "  res['val'] = 0\n",
        "  res['perc'] = 0\n",
        "  for item,value in zip(earlier_res,res_count):\n",
        "    res.loc[res['id']==item,'val'] = value\n",
        "    res.loc[res['id']==item,'perc'] = value/len\n",
        "  res = res.nlargest(to_return,'perc')\n",
        "\n",
        "  # если данных меньше заданного кол-ва рекомендаций, задействуем базовую модель\n",
        "  if res.shape[0]<to_return:\n",
        "    # cамые частые результаты\n",
        "    freq_res = pd.DataFrame([earlier_res,res_count]).T[0].values\n",
        "    res2 = dp_items[dp_items.id.isin(freq_res)].copy()\n",
        "\n",
        "    # записываем абсолютные и относительные величины частоты\n",
        "    len = res.shape[0]\n",
        "    res2['val'] = 0\n",
        "    res2['perc'] = 0\n",
        "    for item,value in zip(earlier_res,res_count):\n",
        "      res2.loc[res2['id']==item,'val'] = value\n",
        "      res2.loc[res2['id']==item,'perc'] = value/len\n",
        "    res2 = res2.nlargest(to_return - res.shape[0],'perc')\n",
        "    \n",
        "    res.append(res2)\n",
        "\n",
        "  return res.drop_duplicates()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hdDeU3vfcUIH",
        "outputId": "54ebc51d-88be-4afd-8274-cf6637a9aa97"
      },
      "source": [
        "earlierWpResDomain(12186,to_return=15)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "      <th>val</th>\n",
              "      <th>perc</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>7781</th>\n",
              "      <td>8355</td>\n",
              "      <td>Охрана интеллектуальной собственности авторски...</td>\n",
              "      <td>22</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "      <td>2</td>\n",
              "      <td>0.666667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7778</th>\n",
              "      <td>8364</td>\n",
              "      <td>Распоряжение правами на интеллектуальную собст...</td>\n",
              "      <td>19</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.333333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17920</th>\n",
              "      <td>8354</td>\n",
              "      <td>Правовые основы интеллектуальной собственности</td>\n",
              "      <td>20</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.333333</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         id                                               name  ...  val      perc\n",
              "7781   8355  Охрана интеллектуальной собственности авторски...  ...    2  0.666667\n",
              "7778   8364  Распоряжение правами на интеллектуальную собст...  ...    1  0.333333\n",
              "17920  8354     Правовые основы интеллектуальной собственности  ...    1  0.333333\n",
              "\n",
              "[3 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DLmTyzA_1wqZ"
      },
      "source": [
        "# checking"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ugR4qMgIzH3P"
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oTkqxwWMRagP",
        "outputId": "16745701-723d-4f76-ee7f-c6197db3434e"
      },
      "source": [
        "wp_outcomes.workprogram_id.unique().shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3848,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ULqmA3U3M4nJ",
        "outputId": "13017a95-8bc2-46a8-d31f-a58642148b74"
      },
      "source": [
        "wp_prereq.workprogram_id.unique().shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(294,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nXM-r-eP17If",
        "outputId": "1bef8791-0663-4855-ddd0-425379e630eb"
      },
      "source": [
        "%%timeit\n",
        "set(a).intersection(set(b))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The slowest run took 8.33 times longer than the fastest. This could mean that an intermediate result is being cached.\n",
            "100000 loops, best of 5: 7.24 µs per loop\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xqiHAd483OFP",
        "outputId": "9e1bac59-5eec-439d-8cad-8c82be2c9071"
      },
      "source": [
        "%%timeit\n",
        "np.intersect1d(a,b)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The slowest run took 23.41 times longer than the fastest. This could mean that an intermediate result is being cached.\n",
            "10000 loops, best of 5: 40.1 µs per loop\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ntI53Lpn53o7"
      },
      "source": [
        "def evalPrereq(to_return = 10, prereq_func = earlierWpResDomain):\n",
        "  all = wp_prereq.workprogram_id.unique().shape[0]\n",
        "  all_hits = np.array([])\n",
        "\n",
        "  for i in wp_prereq.workprogram_id.unique():\n",
        "    #orig_data = wp_prereq[wp_prereq.workprogram_id==i].item_id\n",
        "    #our_pred = prereq_func(i,to_return).id\n",
        "\n",
        "    orig_data_id = wp_prereq[wp_prereq.workprogram_id==i].item_id\n",
        "    orig_data = dp_items[dp_items.id.isin(orig_data_id)].name\n",
        "    our_pred = prereq_func(i,to_return).name\n",
        "\n",
        "    res = set(orig_data).intersection(set(our_pred))\n",
        "    all_hits = np.append(all_hits, np.array(len(res)/orig_data.shape[0]))\n",
        "\n",
        "\n",
        "  return all_hits.mean()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pJyeXWmvW3py",
        "outputId": "ee654141-2f30-416d-e947-75bd1d50abae"
      },
      "source": [
        "#for name\n",
        "nums = range(5,16)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalPrereq(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.12505938883489903\n",
            "0.1261931756829716\n",
            "0.12676006910700788\n",
            "0.1297969981643451\n",
            "0.13801695281287119\n",
            "0.14085141993305259\n",
            "0.15031854011445847\n",
            "0.1577205485368751\n",
            "0.16594050318540116\n",
            "0.16962531044163698\n",
            "0.17245977756181838\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.12505938883489903,\n",
              " 0.1261931756829716,\n",
              " 0.12676006910700788,\n",
              " 0.1297969981643451,\n",
              " 0.13801695281287119,\n",
              " 0.14085141993305259,\n",
              " 0.15031854011445847,\n",
              " 0.1577205485368751,\n",
              " 0.16594050318540116,\n",
              " 0.16962531044163698,\n",
              " 0.17245977756181838]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0eOsMrsn7jm5",
        "outputId": "cfcae776-eed2-47fe-cb18-4669c4f9d57d"
      },
      "source": [
        "nums = range(5,11)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalPrereq(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.12041086275780154\n",
            "0.12154464960587408\n",
            "0.12211154302991035\n",
            "0.12514847208724758\n",
            "0.13336842673577365\n",
            "0.13620289385595508\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.12041086275780154,\n",
              " 0.12154464960587408,\n",
              " 0.12211154302991035,\n",
              " 0.12514847208724758,\n",
              " 0.13336842673577365,\n",
              " 0.13620289385595508]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FD-r2Xhm1ejm"
      },
      "source": [
        "nums = range(11,21)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalPrereq(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wamm9nt4PXc0",
        "outputId": "763d2e80-982a-42cb-fdf5-8c4804c03c58"
      },
      "source": [
        "results = [0.12041086275780154,0.12154464960587408,0.12211154302991035,\n",
        "  0.12514847208724758,0.13336842673577365,0.13620289385595508,\n",
        "  0.14567001403736096,0.15307202245977755,0.16129197710830365,\n",
        "  0.16497678436453944,0.16781125148472087,0.1703055825504805,\n",
        "  0.1759745167908433,0.18099557283230752]\n",
        "len(results)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "14"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 507
        },
        "id": "XpZqRjQx1enC",
        "outputId": "07f3698c-7900-487e-f65b-f49c1ee0dc2c"
      },
      "source": [
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "sns.lineplot(x = range(5,16), y = results[:11], marker='o')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f42a1837910>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F7nG_q-GP_Js",
        "outputId": "2f35a96e-baf7-4622-c3d2-fd17438a637d"
      },
      "source": [
        "nums = range(5,21)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalPrereq(i, earlierWpRes)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.03767006802721088\n",
            "0.038236961451247165\n",
            "0.03993764172335601\n",
            "0.03993764172335601\n",
            "0.042205215419501126\n",
            "0.04486961451247166\n",
            "0.04486961451247166\n",
            "0.045436507936507935\n",
            "0.045436507936507935\n",
            "0.045436507936507935\n",
            "0.04628684807256236\n",
            "0.04628684807256236\n",
            "0.04628684807256236\n",
            "0.04628684807256236\n",
            "0.04628684807256236\n",
            "0.04968820861678004\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.03767006802721088,\n",
              " 0.038236961451247165,\n",
              " 0.03993764172335601,\n",
              " 0.03993764172335601,\n",
              " 0.042205215419501126,\n",
              " 0.04486961451247166,\n",
              " 0.04486961451247166,\n",
              " 0.045436507936507935,\n",
              " 0.045436507936507935,\n",
              " 0.045436507936507935,\n",
              " 0.04628684807256236,\n",
              " 0.04628684807256236,\n",
              " 0.04628684807256236,\n",
              " 0.04628684807256236,\n",
              " 0.04628684807256236,\n",
              " 0.04968820861678004]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tKgQvyQuP_N7"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N_kZNjKFeeo0",
        "outputId": "b636045d-3db1-47bf-cfb4-5b2581a61405"
      },
      "source": [
        "results = [0.03767006802721088,\n",
        " 0.038236961451247165,\n",
        " 0.03993764172335601,\n",
        " 0.03993764172335601,\n",
        " 0.042205215419501126,\n",
        " 0.04486961451247166,\n",
        " 0.04486961451247166,\n",
        " 0.045436507936507935,\n",
        " 0.045436507936507935,\n",
        " 0.045436507936507935,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04968820861678004]\n",
        "len(results)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "16"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "id": "_5jMNp0Ueeo5",
        "outputId": "cbafaec3-8d70-4967-fe25-fa6cef6bc200"
      },
      "source": [
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "sns.lineplot(x = range(5,16), y = results[:11], marker='o')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f42a184c610>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U3iJWwsaRs1x"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "21jK0ONVTVJt"
      },
      "source": [
        "т.к. не везде заполнены темы"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JnEHIE88Smtx",
        "outputId": "4fb5b0e7-a5a7-4afc-8ae8-722d05377533"
      },
      "source": [
        "wp_outcomes[wp_outcomes.workprogram_id.isin(wp_disc_s.work_program_id.unique())].workprogram_id.unique().shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(392,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 101
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aK3jOMc7SHbb"
      },
      "source": [
        "short_outcomes = wp_outcomes[wp_outcomes.workprogram_id.isin(wp_disc_s.work_program_id.unique())]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "39ZPiiShfDw7"
      },
      "source": [
        "def evalResUSE(to_return = 10):\n",
        "  all = short_outcomes.workprogram_id.unique()\n",
        "  all_hits = np.array([])\n",
        "\n",
        "  for i in all:\n",
        "    orig_data = short_outcomes[short_outcomes.workprogram_id==i].item_id\n",
        "    our_pred = itemDiscSim(i,vecFuncUSE,items_vecs, to_return=to_return).id\n",
        "\n",
        "    res = set(orig_data).intersection(set(our_pred))\n",
        "    all_hits = np.append(all_hits, np.array(len(res)/orig_data.shape[0]))\n",
        "\n",
        "  return all_hits.mean()\n",
        "  "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "njK1zR6OfDyp",
        "outputId": "2fc3c11c-c4fc-409d-8de2-91ac544130cf"
      },
      "source": [
        "nums = range(5,11)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalResUSE(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.16693985085455804\n",
            "0.18447026775969771\n",
            "0.19013128757888073\n",
            "0.21174506566587312\n",
            "0.21841554229808446\n",
            "0.22796277298689677\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.16693985085455804,\n",
              " 0.18447026775969771,\n",
              " 0.19013128757888073,\n",
              " 0.21174506566587312,\n",
              " 0.21841554229808446,\n",
              " 0.22796277298689677]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5j3VAXC4fD1G",
        "outputId": "17d9e73a-43cd-48d8-96b6-ed49f955e9fb"
      },
      "source": [
        "nums = range(11,16)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalResUSE(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.23235300435416895\n",
            "0.23689508217430802\n",
            "0.24545314825788428\n",
            "0.24989212623002552\n",
            "0.25497410490894296\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.23235300435416895,\n",
              " 0.23689508217430802,\n",
              " 0.24545314825788428,\n",
              " 0.24989212623002552,\n",
              " 0.25497410490894296]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gx0dh10efD2z"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ISlxqmPmiNlU",
        "outputId": "72393ac4-affe-4445-99b9-fb3664334e49"
      },
      "source": [
        "results = [0.16693985085455804,\n",
        " 0.18447026775969771,\n",
        " 0.19013128757888073,\n",
        " 0.21174506566587312,\n",
        " 0.21841554229808446,\n",
        " 0.22796277298689677,0.23235300435416895,\n",
        " 0.23689508217430802,\n",
        " 0.24545314825788428,\n",
        " 0.24989212623002552,\n",
        " 0.25497410490894296]\n",
        "len(results)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "11"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 89
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "id": "E15TQ4IkiNlW",
        "outputId": "dda4bf0a-d87e-4290-9695-8aa84a51d2c9"
      },
      "source": [
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "sns.lineplot(x = range(5,16), y = results[:11], marker='o')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f42567571d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 90
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uiv_zQT3fD6j"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S00N-ftFfD9C"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5o7U9K5kNtaA"
      },
      "source": [
        "def vecFuncUSE(headers):\n",
        "  return embed(headers)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "55l_rM1PLzc5"
      },
      "source": [
        "wp_topic_prepro.iloc[3535,-1]='not to do'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tmXoujteLb30",
        "outputId": "9dc7f3d3-922e-4fa3-b4ab-5c5a7bbd9e5b"
      },
      "source": [
        "wp_topic_prepro.iloc[3535]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Unnamed: 0                    3535\n",
              "id                            4449\n",
              "number                           5\n",
              "description              Not to do\n",
              "discipline_section_id         1544\n",
              "url_online_course_id           NaN\n",
              "prepro                   not to do\n",
              "Name: 3535, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ficnwJL-ilJ-"
      },
      "source": [
        "def evalRes1(to_return = 10):\n",
        "  all = short_outcomes.workprogram_id.unique()\n",
        "  all_hits = np.array([])\n",
        "\n",
        "  for i in all:\n",
        "    if i != 13299:\n",
        "      orig_data = short_outcomes[short_outcomes.workprogram_id==i].item_id\n",
        "     \n",
        "      our_pred = itemDiscMatches(discipline_id = i, to_return=to_return).id\n",
        "\n",
        "      res = set(orig_data).intersection(set(our_pred))\n",
        "      all_hits = np.append(all_hits, np.array(len(res)/orig_data.shape[0]))\n",
        "\n",
        "  return all_hits.mean()\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3VqUZpYhilJ_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "07c492b5-7ab8-45ce-8de7-f0b503a84b31"
      },
      "source": [
        "nums = range(5,16)\n",
        "results = []\n",
        "for i in nums:\n",
        "  a = evalRes1(i)\n",
        "  results.append(a)\n",
        "  print(a)\n",
        "results"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.07908526879671102\n",
            "0.09144118156822484\n",
            "0.10217175179367996\n",
            "0.10994057078372658\n",
            "0.1244617877026417\n",
            "0.13324485956116114\n",
            "0.14778526729210595\n",
            "0.160698543984155\n",
            "0.17047831256799345\n",
            "0.18082623990413382\n",
            "0.19312330629327182\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.07908526879671102,\n",
              " 0.09144118156822484,\n",
              " 0.10217175179367996,\n",
              " 0.10994057078372658,\n",
              " 0.1244617877026417,\n",
              " 0.13324485956116114,\n",
              " 0.14778526729210595,\n",
              " 0.160698543984155,\n",
              " 0.17047831256799345,\n",
              " 0.18082623990413382,\n",
              " 0.19312330629327182]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "id": "MmBArHgfpimm",
        "outputId": "a6766032-71fb-41ae-e384-6c27faa4e2af"
      },
      "source": [
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "sns.lineplot(x = nums, y = results, marker='o')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f4246ae2c90>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 135
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HY0cdOudpipn"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QMy548tQOIBF"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xgkKL98XOXRI"
      },
      "source": [
        "def vecFuncSBert(headers):\n",
        "  return model_sbert.encode(headers, convert_to_tensor=True)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mFzm-TvLOXRL"
      },
      "source": [
        "item_vecs_sb = model_sbert.encode(dp_items['name'].values, convert_to_tensor=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CY8wEVcYOIO3"
      },
      "source": [
        "def evalResSB(to_return = 10):\n",
        "  all = short_outcomes.workprogram_id.unique()\n",
        "  hits = 0\n",
        "\n",
        "  for i in all:\n",
        "    orig_data = short_outcomes[short_outcomes.workprogram_id==i].item_id\n",
        "    our_pred = itemDiscSim(i,vecFuncSBert,item_vecs_sb, to_return=to_return).id\n",
        "\n",
        "    if set(orig_data).intersection(set(our_pred)):\n",
        "      hits += 1\n",
        "\n",
        "  return hits/all.shape[0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TgY3an8MsSBw",
        "outputId": "f9e8fd83-4464-44f9-efb5-d9b7afb30c5d"
      },
      "source": [
        "wp_prereq.workprogram_id.unique()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 2811,  6986,    67,     5, 12186,  6990,  6991,  6992,  4861,\n",
              "        3858,  4931,  4171,  2728,  2865,  6994,  2986,  6995,  3148,\n",
              "        2777,  3410, 12799,  5982,  4911, 14522,  2639,  5983,  6993,\n",
              "        3388,  6997, 12838,  6449,  3842,  7050, 11289, 11290,  3433,\n",
              "        4714, 12525,  4713, 13707,  7898,  8156,  6194, 10907, 12801,\n",
              "        6441, 12811, 12891,  7464, 12812,  6442,  6443,  6626, 13966,\n",
              "        6634, 13972, 12270,  7496,  3205,  2746, 12813, 13977, 12894,\n",
              "        2747,  4173,  3063,  6149,  5020, 12814,  5027,  3197, 12524,\n",
              "       12436,  7273,  3840, 14525, 11272, 11554,  3115,  5146, 11273,\n",
              "       11275, 11276,  5224, 11277, 11278, 11279, 12805,  4947,  4707,\n",
              "        4706, 14527,  2641, 14528, 14026, 14027, 14031, 11472, 13300,\n",
              "       13352,  3351, 14041,  3352, 12440,  3118, 14054, 12933, 12533,\n",
              "       14064, 13832, 13372,  6635,  3465, 12534, 11394,  6632,  6633,\n",
              "        6643,  6644,  6645, 13375,  4888, 13701,  3831, 13638, 12515,\n",
              "       13127,  3442, 14510,  4906, 12535,  4905,  8824, 10513,  6510,\n",
              "        3110,  2640, 13287, 13632, 12221, 12517, 11470, 12518, 12519,\n",
              "       12520, 12223,  3839, 14325, 12521, 12522, 12849, 11064, 12536,\n",
              "        3856,  4555,  8064, 12800, 14396,  6199, 14408, 11523, 14415,\n",
              "        3460, 11555, 14523, 12844,  4884, 13124,  6210, 12804, 14471,\n",
              "       12937, 12851, 12852, 12853, 12854, 14002, 11406,  4907, 14499,\n",
              "       11558, 14007, 13698, 12523, 11282, 12750, 12751, 12863, 12864,\n",
              "       12865, 11285, 13978, 13897, 11286, 12938, 13590, 13888, 13289,\n",
              "       11471, 11635, 12079, 12866, 13893, 12867, 12868, 12752,  4262,\n",
              "        5125,  5129,  8356,  5131,  7593,  5140,  6152, 12423,  8057,\n",
              "       14515, 12526,  6532, 13121,  4671, 12870, 12871,  5088, 11863,\n",
              "       11648, 14526,  5048,  3434,  5154, 10492, 11638, 14518, 14520,\n",
              "       11812,  3131, 12495, 13154,  4715, 14530,  9657,  2662, 13288,\n",
              "       12222, 13699,  3463, 13642,  4703, 12845, 13290, 14533, 14534,\n",
              "       14532, 12839, 11811, 12494, 12496, 12491, 11782,  4904, 12607,\n",
              "       11283, 12463,  3439, 12438, 12464,  7808,  2635, 10243, 12453,\n",
              "       14536, 13118, 12081, 12730, 13710, 11473, 11484, 13353, 11061,\n",
              "        5980, 11033,  5998, 13799,  4908, 13887,  6202,  6185,  3849,\n",
              "       13157, 13855,  5786,  6638, 13129, 14538])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 146
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "s7i1y7gnrCi8",
        "outputId": "8940addd-f620-46e2-df3c-49912eae870c"
      },
      "source": [
        "dp_items[dp_items.id.isin(wp_prereq[wp_prereq.workprogram_id == 12186].item_id)]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>7778</th>\n",
              "      <td>8364</td>\n",
              "      <td>Распоряжение правами на интеллектуальную собст...</td>\n",
              "      <td>19</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7781</th>\n",
              "      <td>8355</td>\n",
              "      <td>Охрана интеллектуальной собственности авторски...</td>\n",
              "      <td>22</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17920</th>\n",
              "      <td>8354</td>\n",
              "      <td>Правовые основы интеллектуальной собственности</td>\n",
              "      <td>20</td>\n",
              "      <td>uploaded</td>\n",
              "      <td>1.0</td>\n",
              "      <td>27.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         id  ... domain_id\n",
              "7778   8364  ...      27.0\n",
              "7781   8355  ...      27.0\n",
              "17920  8354  ...      27.0\n",
              "\n",
              "[3 rows x 6 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 151
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "yPUfteKMqsTD",
        "outputId": "71438856-374d-43a4-8d3c-b88918f95315"
      },
      "source": [
        "dp_items[dp_items.id.isin(wp_outcomes[wp_outcomes.workprogram_id == 12223].item_id)]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>name</th>\n",
              "      <th>value</th>\n",
              "      <th>source</th>\n",
              "      <th>author_id</th>\n",
              "      <th>domain_id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>17814</th>\n",
              "      <td>18491</td>\n",
              "      <td>Разработка игрового нарратива</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17815</th>\n",
              "      <td>18493</td>\n",
              "      <td>Разработка персонажа игры</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17816</th>\n",
              "      <td>18495</td>\n",
              "      <td>Использования игровых механик как средств выра...</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17817</th>\n",
              "      <td>18492</td>\n",
              "      <td>Структурный подход к нарративу</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17818</th>\n",
              "      <td>18496</td>\n",
              "      <td>Конструирование опыта игрока через взаимодейст...</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17819</th>\n",
              "      <td>18494</td>\n",
              "      <td>Создание интерактивных элементов игрового пове...</td>\n",
              "      <td>1</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>62.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "          id  ... domain_id\n",
              "17814  18491  ...      62.0\n",
              "17815  18493  ...      62.0\n",
              "17816  18495  ...      62.0\n",
              "17817  18492  ...      62.0\n",
              "17818  18496  ...      62.0\n",
              "17819  18494  ...      62.0\n",
              "\n",
              "[6 rows x 6 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 142
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PVldiHWZZi7E"
      },
      "source": [
        "# Оценка рез-в работы. Два графика в одном"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1WfXewXQbXNp"
      },
      "source": [
        "resultsbest = [0.16693985085455804,\n",
        " 0.18447026775969771,\n",
        " 0.19013128757888073,\n",
        " 0.21174506566587312,\n",
        " 0.21841554229808446,\n",
        " 0.22796277298689677,0.23235300435416895,\n",
        " 0.23689508217430802,\n",
        " 0.24545314825788428,\n",
        " 0.24989212623002552,\n",
        " 0.25497410490894296]\n",
        "resultsbase = [0.07908526879671102,\n",
        " 0.09144118156822484,\n",
        " 0.10217175179367996,\n",
        " 0.10994057078372658,\n",
        " 0.1244617877026417,\n",
        " 0.13324485956116114,\n",
        " 0.14778526729210595,\n",
        " 0.160698543984155,\n",
        " 0.17047831256799345,\n",
        " 0.18082623990413382,\n",
        " 0.19312330629327182]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 556
        },
        "id": "in6amkVDZrEI",
        "outputId": "5e402888-b8a1-41be-ba44-33329a6fc20e"
      },
      "source": [
        "\n",
        "\n",
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "\n",
        "sns.set(font_scale=1.6)\n",
        "\n",
        "sns.set_style(\"whitegrid\")\n",
        "#sns.set(style=\"whitegrid\", rc={\"lines.linewidth\": 3})\n",
        "ax = sns.lineplot(x = range(5,16), y = resultsbase[:11], marker='o', color='blue')\n",
        "ax.set(xticks=range(5,16))\n",
        "\n",
        "ax.set_title('Outcomes')\n",
        "\n",
        "ax.set(xlabel='Recommendations', ylabel='Hit rate')\n",
        "sns.set(rc={\"lines.linewidth\": 3})\n",
        "sns.lineplot(x = range(5,16), y = resultsbest[:11], marker='P', color='red', linestyle=\"--\")\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f1b3485eb10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O6AxEBw0gU8S"
      },
      "source": [
        "resultsbest = [0.12041086275780154,0.12154464960587408,0.12211154302991035,\n",
        "  0.12514847208724758,0.13336842673577365,0.13620289385595508,\n",
        "  0.14567001403736096,0.15307202245977755,0.16129197710830365,\n",
        "  0.16497678436453944,0.16781125148472087,0.1703055825504805,\n",
        "  0.1759745167908433,0.18099557283230752]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "42e2JEU2gelA"
      },
      "source": [
        "resultsbase = [0.03767006802721088,\n",
        " 0.038236961451247165,\n",
        " 0.03993764172335601,\n",
        " 0.03993764172335601,\n",
        " 0.042205215419501126,\n",
        " 0.04486961451247166,\n",
        " 0.04486961451247166,\n",
        " 0.045436507936507935,\n",
        " 0.045436507936507935,\n",
        " 0.045436507936507935,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236,\n",
        " 0.04628684807256236]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 556
        },
        "id": "rbGiTcNDgVCA",
        "outputId": "3603ce48-78af-4c3e-bc83-a1fecfc35d6c"
      },
      "source": [
        "sns.set(rc={'figure.figsize':(10,8)})\n",
        "sns.set(font_scale=1.6)\n",
        "sns.set_style(\"whitegrid\")\n",
        "\n",
        "ax = sns.lineplot(x = range(5,16), y = resultsbase[:11], marker='o', color='blue')\n",
        "ax.set(xlabel='Recommendations', ylabel='Hit rate')\n",
        "ax.set_title('Prerequisites')\n",
        "ax.set(xticks=range(5,16))\n",
        "sns.set(rc={\"lines.linewidth\": 3})\n",
        "sns.lineplot(x = range(5,16), y = resultsbest[:11], marker='P', color='red', linestyle=\"--\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f1b3475d750>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7hWP_Rv6ap9w"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    }
  ]
}