ten_city_presentation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "696175c9-278e-4ec3-96ae-d819f44cf808",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T09:25:04.597862Z",
     "iopub.status.busy": "2023-06-08T09:25:04.596862Z",
     "iopub.status.idle": "2023-06-08T09:25:27.064252Z",
     "shell.execute_reply": "2023-06-08T09:25:27.063337Z",
     "shell.execute_reply.started": "2023-06-08T09:25:04.597862Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ready: 2023-06-08 12:25:27\n"
     ]
    }
   ],
   "source": [
    "%run functions.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fe99b644-38be-45e0-ad1e-bf32b03abd81",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T09:26:10.768138Z",
     "iopub.status.busy": "2023-06-08T09:26:10.768138Z",
     "iopub.status.idle": "2023-06-08T09:26:10.785776Z",
     "shell.execute_reply": "2023-06-08T09:26:10.784649Z",
     "shell.execute_reply.started": "2023-06-08T09:26:10.768138Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "size_x = 30.7 / 2.54 * 4\n",
    "size_y = 16   / 2.54 * 4\n",
    "font_size = 38"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "3fb7c96e-8126-45a2-b9a7-966a96f87723",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T19:20:29.916272Z",
     "iopub.status.busy": "2023-05-29T19:20:29.915275Z",
     "iopub.status.idle": "2023-05-29T19:20:30.315346Z",
     "shell.execute_reply": "2023-05-29T19:20:30.314285Z",
     "shell.execute_reply.started": "2023-05-29T19:20:29.916272Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df = pd.read_excel('files/ten_city/ratings_df.xlsx').fillna(0).astype(\n",
    "    {'Рубрика':str, \n",
    "'Пользователи 2020':int,\n",
    "'Пользователи 2021':int,\n",
    "'Пользователи 2022':int,\n",
    "'Пользователи 2023':int,\n",
    "'Пользователи, итого':int,\n",
    "'Пользователи, %' : float,\n",
    "'Запросы 2020':int,\n",
    "'Запросы 2021':int,\n",
    "'Запросы 2022':int,\n",
    "'Запросы 2023':int,\n",
    "'Запросы, итого':int,\n",
    "'Запросы, %' : float,                                                                      \n",
    "})\n",
    "\n",
    "df['Запросы, %'] = df['Запросы, %'].round(1)\n",
    "df['Пользователи, %'] = df['Пользователи, %'].round(1)\n",
    "\n",
    "df_users = df[[\n",
    "    'Рубрика',\n",
    "    'Пользователи 2020',\n",
    "    'Пользователи 2021',\n",
    "    'Пользователи 2022',\n",
    "    'Пользователи 2023',\n",
    "    'Пользователи, итого',\n",
    "    'Пользователи, %'\n",
    "]].iloc[list(range(15))+[-1]].sort_values('Пользователи, итого', ascending=False)\n",
    "\n",
    "df_requests = df[[\n",
    "    'Рубрика',\n",
    "    'Запросы 2020',\n",
    "    'Запросы 2021',\n",
    "    'Запросы 2022',\n",
    "    'Запросы 2023',\n",
    "    'Запросы, итого',\n",
    "    'Запросы, %'\n",
    "]].iloc[list(range(15))+[-1]].sort_values('Запросы, итого', ascending=False)\n",
    "\n",
    "#249FD8\n",
    "#7FA2D3\n",
    "#D2DDF1\n",
    "\n",
    "def make_hse_image(df, output_file_name=None, cols_without_color=[0], dedup_cols=None, col_width=1.0, row_height=0.625, font_size=28,\n",
    "                     header_color='#249FD8', row_colors=['#D2DDF1', 'w'], edge_color='w',\n",
    "                     bbox=[0, 0, 1, 1], header_columns=0,\n",
    "                     show=False, **kwargs):\n",
    "    rating_col = df.columns[-1]\n",
    "    data = df.sort_values(rating_col, ascending=False)\n",
    "    data_separeted_numbers = data.copy()\n",
    "    for col in data.columns[1:]:\n",
    "        data_separeted_numbers[col] = data[col].apply(make_sep)\n",
    "        if col in ['Пользователи, %', 'Запросы, %', 'Запросов на пользователя,\\nитого']:\n",
    "            data_separeted_numbers[col] = data[col].apply(lambda x: str(x).replace('.',','))\n",
    "        \n",
    "    size = (np.array([size_x, size_y]))\n",
    "    fig, ax = plt.subplots(figsize=size, dpi=2**8)\n",
    "    ax.axis('off')\n",
    "\n",
    "    mpl_table = ax.table(cellText=data_separeted_numbers.values, bbox=bbox, colLabels=data.columns, colLoc='center', loc='bottom')\n",
    "    mpl_table.auto_set_font_size(False)\n",
    "    cells_dict = dict(six.iteritems(mpl_table._cells))\n",
    "        \n",
    "    fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "    colors_dict = colorize_column(df, cols_not_to_color=cols_without_color)\n",
    "    colors_dict.update(colorize_column(data, cols_not_to_color=list(range(data.shape[1]-1))), filter_zero=True)\n",
    "    colors = ([color_list_full[i] for i in range(0, len(color_list_full), len(color_list_full) % 15)] + [color_list_full[-1]]) \n",
    "    \n",
    "    for k in sorted(cells_dict.keys()): #k[0] - строка, k[1] - столбец\n",
    "        cell = cells_dict[k]\n",
    "        cell.set_edgecolor(edge_color)\n",
    "        row_index = k[0]\n",
    "        column_index = k[1]\n",
    "        \n",
    "        # Высота и цвет заголовков\n",
    "        if row_index == 0:\n",
    "            rotation='horizontal' if k == (0,0) else 'vertical' \n",
    "            cell.set_text_props(color='w', rotation=rotation, ha='center')\n",
    "            cell.set_facecolor(header_color)\n",
    "        # Высота и цвет остальных строк\n",
    "        else:\n",
    "            ha = 'left' if column_index ==0 else 'right'\n",
    "            cell.set_facecolor(row_colors[k[0]%len(row_colors)])\n",
    "            cell.set_text_props(wrap=True, ha=ha)\n",
    "            \n",
    "        if row_index == 0:\n",
    "            cell.set_height(24)\n",
    "        elif row_index > 0 and row_index < 13:\n",
    "            cell.set_height(5)\n",
    "        else:\n",
    "            cell.set_height(4)\n",
    "        \n",
    "        # Ширина столбцов\n",
    "        if column_index == 0:\n",
    "            cell.set_width(35)\n",
    "        elif column_index == 6:\n",
    "            cell.set_width(10)\n",
    "        else:\n",
    "            cell.set_width(11)\n",
    "                        \n",
    "        # Цветвоая шкала\n",
    "        if row_index > 0 and column_index not in cols_without_color and row_index<16:\n",
    "            cell.set_facecolor(colors.pop(0).upper())       \n",
    "            \n",
    "        if row_index == 16:\n",
    "            cell.set_text_props(color='w')\n",
    "            cell.set_facecolor(header_color) \n",
    "            if column_index == 6:\n",
    "                cell.set_text_props(color=header_color)\n",
    "            \n",
    "        # Шрифт\n",
    "        cell.set_text_props(font=fpath, fontsize=font_size)\n",
    "    if output_file_name:\n",
    "        fig.savefig(f'{output_file_name}', bbox_inches='tight')\n",
    "    if show:\n",
    "        plt.show()\n",
    "    plt.close() \n",
    "\n",
    "def make_hse_image2(df, output_file_name=None, cols_without_color=[0], dedup_cols=None, col_width=1.0, row_height=0.625, font_size=28,\n",
    "                     header_color='#249FD8', row_colors=['#D2DDF1', 'w'], edge_color='w',\n",
    "                     bbox=[0, 0, 1, 1], header_columns=0,\n",
    "                     show=False, **kwargs):\n",
    "    rating_col = df.columns[-1]\n",
    "    data = df\n",
    "    data['Население'] = data['Население'].apply(make_sep)\n",
    "    size = (np.array([size_x, size_y]))\n",
    "    fig, ax = plt.subplots(figsize=size, dpi=2**8)\n",
    "    ax.axis('off')\n",
    "\n",
    "    mpl_table = ax.table(cellText=data.values, bbox=bbox, colLabels=data.columns, colLoc='center', loc='bottom')\n",
    "    mpl_table.auto_set_font_size(False)\n",
    "    cells_dict = dict(six.iteritems(mpl_table._cells))\n",
    "    \n",
    "    fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "    colors = ([color_list_full[i] for i in range(0, len(color_list_full), len(color_list_full) % 20)] + [color_list_full[-1]]) \n",
    "    \n",
    "    for k in sorted(cells_dict.keys()): #k[0] - строка, k[1] - столбец\n",
    "        cell = cells_dict[k]\n",
    "        cell.set_edgecolor(edge_color)\n",
    "        row_index = k[0]\n",
    "        column_index = k[1]\n",
    "        \n",
    "        # Высота и цвет заголовков\n",
    "        if row_index == 0:\n",
    "            rotation='horizontal' if k == (0,0) else 'vertical' \n",
    "            cell.set_text_props(color='w', ha='center')\n",
    "            cell.set_facecolor(header_color)\n",
    "            \n",
    "        # Высота и цвет остальных строк\n",
    "        else:\n",
    "            ha = 'left' if column_index in (0,1,2) else 'right'\n",
    "            cell.set_facecolor(row_colors[k[0]%len(row_colors)])\n",
    "            cell.set_text_props(wrap=True, ha=ha)\n",
    "        \n",
    "        if row_index == 0:\n",
    "            cell.set_height(24)\n",
    "        elif row_index > 0 and row_index < 5:\n",
    "            cell.set_height(5)\n",
    "        else:\n",
    "            cell.set_height(4) \n",
    "        \n",
    "        \n",
    "        # Ширина столбцов\n",
    "        if column_index == 0:\n",
    "            cell.set_width(5)\n",
    "            cell.set_text_props(ha='right')\n",
    "        elif column_index == 1:\n",
    "            cell.set_width(30)\n",
    "        elif column_index == 2:\n",
    "            cell.set_width(33)\n",
    "        else:\n",
    "            cell.set_width(32)\n",
    "            \n",
    "        if row_index == 0 and column_index == 0:\n",
    "            cell.set_text_props(ha='center')\n",
    "\n",
    "        # Цветвоая шкала\n",
    "        if row_index > 0 and column_index == 3:\n",
    "            cell.set_facecolor(colors.pop(0).upper())       \n",
    "            \n",
    "        if row_index == 18:\n",
    "            cell.set_text_props(color='w')\n",
    "            cell.set_facecolor(header_color) \n",
    "            if column_index == 1 or column_index == 2:\n",
    "                cell.set_text_props(color=header_color)\n",
    "            \n",
    "        # Шрифт\n",
    "        cell.set_text_props(font=fpath, fontsize=font_size)\n",
    "    if output_file_name:\n",
    "        fig.savefig(f'{output_file_name}', bbox_inches='tight')\n",
    "    if show:\n",
    "        plt.show()\n",
    "    plt.close()    \n",
    "\n",
    "\n",
    "def plt_requsts(data, file_name, index_col_to_colorize=5, dups=None, col_width=1.0, row_height=0.625, font_size=font_size, size_x=size_x, size_y=size_y,\n",
    "                     header_color='#C0504D', row_colors=['#F2DCDB', 'w'], edge_color='w',\n",
    "                     bbox=[0, 0, 1, 1], header_columns=0,\n",
    "                     ax=None, **kwargs):\n",
    "    size = (np.array([size_x, size_y]))\n",
    "    fig, ax = plt.subplots(figsize=size, dpi=2**8)\n",
    "    ax.axis('off')\n",
    "\n",
    "    mpl_table = ax.table(cellText=data.values, bbox=bbox, colLabels=data.columns, colLoc='center', loc='bottom')\n",
    "\n",
    "    mpl_table.auto_set_font_size(False)\n",
    "    cells_dict = dict(six.iteritems(mpl_table._cells))\n",
    "\n",
    "    fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "    color_list_full = ['#F96A6C', '#F96B6C', '#F96C6C', '#F96D6C', '#F96E6C', '#F96F6D', '#F9706D', '#F9716D', '#F9726D', '#F9736D', '#F9746E', '#F9756E', '#F9766E', '#F9776E', '#F9786E', '#F9796E', '#F9796F', '#F97A6F'\n",
    "                       , '#F97B6F', '#F97C6F', '#FA7C6F', '#FA7D6F', '#FA7E6F', '#FA7E70', '#FA7F70', '#FA8070', '#FA8170', '#FA8270', '#FA8370', '#FA8471', '#FA8571', '#FA8671', '#FA8771', '#FA8871', '#FA8972', '#FA8A72'\n",
    "                       , '#FA8B72', '#FA8C72', '#FA8D72', '#FA8E73', '#FB8F73', '#FB9073', '#FB9173', '#FB9273', '#FB9373', '#FB9374', '#FB9474', '#FB9574', '#FB9674', '#FB9774', '#FB9874', '#FB9875', '#FB9975', '#FB9A75'\n",
    "                       , '#FB9B75', '#FB9C75', '#FB9D75', '#FB9E76', '#FB9F76', '#FBA076', '#FBA176', '#FCA176', '#FCA276', '#FCA377', '#FCA477', '#FCA577', '#FCA677', '#FCA777', '#FCA877', '#FCA878', '#FCA978', '#FCAA78'\n",
    "                       , '#FCAB78', '#FCAC78', '#FCAD78', '#FCAD79', '#FCAE79', '#FCAF79', '#FCB079', '#FCB179', '#FCB279', '#FCB27A', '#FCB37A', '#FDB47A', '#FDB57A', '#FDB67A', '#FDB77A', '#FDB87B', '#FDB97B', '#FDBA7B'\n",
    "                       , '#FDBB7B', '#FDBC7B', '#FDBD7C', '#FDBE7C', '#FDBF7C', '#FDC07C', '#FDC17C', '#FDC27C', '#FDC27D', '#FDC37D', '#FDC47D', '#FDC57D', '#FDC67D', '#FEC77D', '#FEC77E', '#FEC87E', '#FEC97E', '#FECA7E'\n",
    "                       , '#FECB7E', '#FECC7E', '#FECC7F', '#FECD7F', '#FECE7F', '#FECF7F', '#FED07F', '#FED17F', '#FED280', '#FED380', '#FED480', '#FED580', '#FED680', '#FED781', '#FED881', '#FFD981', '#FFDA81', '#FFDB81'\n",
    "                       , '#FFDC81', '#FFDC82', '#FFDD82', '#FFDE82', '#FFDF82', '#FFE082', '#FFE182', '#FFE183', '#FFE283', '#FFE383', '#FFE483', '#FFE583', '#FFE683', '#FFE784', '#FFE884', '#FFE984', '#FFEA84', '#FFEB84'\n",
    "                       , '#FEEA83', '#FDEA83', '#FCEA83', '#FBEA83', '#FAE983', '#F9E983', '#F8E983', '#F7E883', '#F6E883', '#F5E883', '#F4E883', '#F4E783', '#F3E783', '#F2E783', '#F1E783', '#F0E683', '#EFE683', '#EEE683'\n",
    "                       ,'#EDE683', '#EDE582', '#ECE582', '#EBE582', '#EAE582', '#EAE482', '#E9E482', '#E8E482', '#E7E482', '#E6E482', '#E6E382', '#E5E382', '#E4E382', '#E3E382', '#E3E282', '#E2E282', '#E1E282', '#E0E282'\n",
    "                       , '#DFE182', '#DEE182', '#DDE182', '#DCE182', '#DCE081', '#DBE081', '#DAE081', '#D9E081', '#D8DF81', '#D7DF81', '#D6DF81', '#D5DF81', '#D4DE81', '#D3DE81', '#D2DE81', '#D1DD81', '#D0DD81', '#CFDD81'\n",
    "                       , '#CEDD81', '#CDDC81', '#CCDC81', '#CBDC81', '#CADB80', '#C9DB80', '#C8DB80', '#C7DB80', '#C7DA80', '#C6DA80', '#C5DA80', '#C4DA80', '#C3D980', '#C2D980', '#C1D980', '#C0D980', '#C0D880', '#BFD880'\n",
    "                       , '#BED880', '#BDD880', '#BDD780', '#BCD780', '#BBD780', '#BAD780', '#B9D780', '#B9D67F', '#B8D67F', '#B7D67F', '#B6D67F', '#B6D57F', '#B5D57F', '#B4D57F', '#B3D57F', '#B2D47F', '#B1D47F', '#B0D47F'\n",
    "                       , '#AFD47F', '#AED37F', '#ADD37F', '#ACD37F', '#ABD27F', '#AAD27F', '#A9D27F', '#A8D27F', '#A7D17E', '#A6D17E', '#A5D17E', '#A4D07E', '#A3D07E', '#A2D07E', '#A1D07E', '#A0CF7E', '#9FCF7E', '#9ECF7E'\n",
    "                       , '#9DCE7E', '#9CCE7E', '#9BCE7E', '#9ACE7E', '#9ACD7E', '#99CD7E', '#98CD7E', '#97CD7E', '#96CC7D', '#95CC7D', '#94CC7D', '#93CC7D', '#93CB7D', '#92CB7D', '#91CB7D', '#90CB7D', '#90CA7D', '#8FCA7D'\n",
    "                       , '#8ECA7D', '#8DCA7D', '#8CCA7D', '#8CC97D', '#8BC97D', '#8AC97D', '#89C97D', '#88C87D', '#87C87D', '#86C87D', '#85C77C', '#84C77C', '#83C77C', '#82C77C', '#81C67C', '#80C67C', '#7FC67C', '#7EC57C'\n",
    "                       , '#7DC57C', '#7CC57C', '#7BC57C', '#7AC47C', '#79C47C', '#78C47C', '#77C37C', '#76C37C', '#75C37C', '#74C37C', '#73C27B'\n",
    "                       , '#72C27B', '#71C27B', '#70C17B', '#6FC17B', '#6EC17B', '#6DC17B', '#6DC07B', '#6CC07B', '#6BC07B', '#6AC07B', '#69BF7B', '#68BF7B', '#67BF7B', '#66BF7B', '#66BE7B', '#65BE7B', '#64BE7B', '#63BE7B']\n",
    "    colors = ([color_list_full[i] for i in range(0, len(color_list_full), len(color_list_full) % 20)] + [color_list_full[-1]])\n",
    "    \n",
    "    for k in sorted(cells_dict.keys()): #k[0] - строка, k[1] - столбец\n",
    "        cell = cells_dict[k]\n",
    "        cell.set_edgecolor(edge_color)\n",
    "        row_index = k[0]\n",
    "        column_index = k[1]\n",
    "        \n",
    "        # Высота и цвет заголовков\n",
    "        if row_index == 0:\n",
    "            rotation='horizontal' if k == (0,0) else 'vertical' \n",
    "            cell.set_text_props(color='w', rotation=rotation, ha='center')\n",
    "            cell.set_facecolor(header_color)\n",
    "            \n",
    "        # Высота и цвет остальных строк\n",
    "        else:\n",
    "            ha = 'right' if k[1] !=0 else 'left'\n",
    "            cell.set_facecolor(row_colors[k[0]%len(row_colors)])\n",
    "            cell.set_text_props(wrap=True, ha=ha)\n",
    "        \n",
    "        if row_index == 0:\n",
    "            cell.set_height(24)\n",
    "        elif row_index > 0 and row_index < 17:\n",
    "            cell.set_height(4)\n",
    "        else:\n",
    "            cell.set_height(3)\n",
    "        \n",
    "        # Ширина столбцов\n",
    "        if column_index == 0:\n",
    "            cell.set_width(35)\n",
    "        elif column_index == 5:\n",
    "            cell.set_width(13)\n",
    "        else:\n",
    "            cell.set_width(13)\n",
    "        \n",
    "        # Раскраска дублей\n",
    "        if dups is not None:\n",
    "            if row_index > 0 and column_index == 0 and cell.get_text().get_text() in dups:\n",
    "                cell.set_facecolor('#FFC7CE')\n",
    "                cell.set_text_props(color='#9C0006')\n",
    "                \n",
    "        # Цветвоая шкала\n",
    "        if row_index > 0 and column_index == index_col_to_colorize:\n",
    "            cell.set_facecolor(colors.pop(0).upper()) \n",
    "        \n",
    "        # Шрифт\n",
    "        cell.set_text_props(font=fpath, fontsize=font_size)\n",
    "    fig.savefig(f'{file_name}', bbox_inches='tight')\n",
    "    plt.show()\n",
    "    plt.close()\n",
    "\n",
    "def make_plt_requsts(file_name, city, header_color, row_color, cnt=9999, size_x=size_x, size_y=size_y, font_size=font_size, img_w=17.5, img_h=10.8, blank=False, word_sample='files/report.docx'):\n",
    "    dt_now = str(datetime.datetime.now())[2:-9].replace('-','').replace(':','').replace(' ','_')\n",
    "    df = pd.read_excel(file_name).drop(['tx_1','tx_2','tx_3','tx_4','tx_5'], axis=1)\n",
    "    word_file = f'{file_name[:-5]}_report_{dt_now}.docx'\n",
    "    #print(word_file)\n",
    "    \n",
    "    df.columns = ['Рубрика', 'Поисковый запрос', 'Количество\\nзапросов 2020',\n",
    "           'Количество\\nзапросов 2021', 'Количество\\nзапросов 2022',\n",
    "           'Количество\\nзапросов 2023', 'Количество\\nзапросов, итого',\n",
    "           'Рейтинг\\nпо запросам, итого', 'Рейтинг рубрики']\n",
    "\n",
    "    for i in df.columns[2:-2]:\n",
    "        df[i] = df[i].apply(make_sep)\n",
    "        \n",
    "    dups = df['Поисковый запрос'][(df['Поисковый запрос'].duplicated())&(df['Рейтинг рубрики'] > 0)].values \n",
    "    df_text = pd.read_excel(file_name, usecols=['Рубрика', 'Рейтинг рубрики', 'Поисковый запрос', 'tx_1'])\n",
    "    df_text = df_text[df_text['tx_1'].notna()]\n",
    "    tx_1 = df_text[df_text['Поисковый запрос'].apply(lambda x: x not in dups)][['tx_1', 'Рейтинг рубрики']]\n",
    "    imgs = []\n",
    "    for i in (df['Рейтинг рубрики'].unique()[1:cnt+1]):  \n",
    "        dfx = df[df['Рейтинг рубрики'] == i].drop('Рубрика', axis=1).iloc[:20]\n",
    "        rubric = df[df['Рейтинг рубрики'] == i]['Рубрика'].iloc[0]\n",
    "        col_name = f'Поисковые запросы по рубрике\\n{rubric}'\n",
    "        img_file_name = file_name.replace('files/','')[:-5]\n",
    "        img_file_name_final = f'pictures/{img_file_name}_{i:02}.png'\n",
    "        dfx = dfx.rename({'Поисковый запрос':col_name}, axis=1).drop(['Рейтинг рубрики', 'Рейтинг\\nпо запросам, итого'], axis=1)\n",
    "        plt_requsts(dfx, img_file_name_final, size_x=size_x, size_y=size_y, font_size=font_size, header_color=header_color, row_colors=[row_color, 'w'], dups=dups)\n",
    "        imgs.append(img_file_name_final)\n",
    "    return imgs\n",
    "\n",
    "\n",
    "\n",
    "def make_ratings_word(file_name, header_color, row_color, cnt=9999, size_x=size_x, size_y=size_y, font_size=18, img_w=17.5, img_h=10.8, blank=False, word_sample='files/report.docx'):\n",
    "    dt_now = str(datetime.datetime.now())[2:-9].replace('-','').replace(':','').replace(' ','_')\n",
    "    df = pd.read_excel(file_name)\n",
    "    word_file = f'{file_name[:-17]}_report_{dt_now}.docx'\n",
    "    #print(word_file)\n",
    "    \n",
    "    df = df[(df['№']!='Итог')]\n",
    "    cols = [\n",
    "        'Рубрика' \n",
    "        ,'Пользователи, итого'\n",
    "        ,'Пользователи, %'\n",
    "        ,'Запросы, итого'\n",
    "        ,'Запросы, %'\n",
    "        ,'Запросов на пользователя,\\nитого'\n",
    "    ]\n",
    "\n",
    "    ratings_cols = [i for i in df.columns if 'ейтинг' in i and i not in  ('В рейтингах, раз','Рейтинг суммарный','Рейтинг средний','Вне рейтингов')]\n",
    "\n",
    "    for col in ratings_cols:\n",
    "        df[col] = df[col].astype('int').round()\n",
    "\n",
    "    for col in ['Пользователи, %', 'Запросы, %', 'Запросов на пользователя,\\nитого']:\n",
    "        df[col] = df[col].astype('float64').round(1)\n",
    "    \n",
    "    i=0\n",
    "    imgs = []\n",
    "    for col in ([ratings_cols[3], ratings_cols[28]]):  \n",
    "        i+=1\n",
    "        cols_filter = cols+[col]    \n",
    "        dfx = df[df[col]>0].sort_values(col)['Рубрика'].values\n",
    "        img_file_name = file_name.replace('files/','')[:-5]\n",
    "        img_file_name_final = f'pictures/{img_file_name}_{i:02}.png'\n",
    "        df_cols_filter = df[cols_filter]\n",
    "        new_cols = [\n",
    "        'Рубрика' \n",
    "        ,'Пользователи,\\nитого'\n",
    "        ,'Пользователи, %'\n",
    "        ,'Запросы, итого'\n",
    "        ,'Запросы, %'\n",
    "        ,'Запросов\\nна пользователя,\\nитого'\n",
    "        ] +[col.replace('Рейтинг ', 'Рейтинг\\n').replace('пользователям ', 'пользователям,\\n')]  \n",
    "        df_cols_filter.columns = new_cols\n",
    "            \n",
    "        make_ratings_image(df_cols_filter, output_file_name=img_file_name_final, cols_without_color=[0], header_color=header_color, row_colors=[row_color, 'w'], size_x=size_x, size_y=size_y, font_size=font_size)\n",
    "        txt =  col + '\\n' + ', '.join([word[::-1].replace('/', 'и', 1)[::-1].replace('/', ',').replace(' ,', ',').lower() for word in df[df[col]>0].sort_values(col)['Рубрика'].values]) + \":\"\n",
    "        #add_image(word_file, img_file_name_final, word=txt, img_w=img_w, img_h=img_h, blank=blank, word_sample=word_sample)\n",
    "        imgs.append(img_file_name_final)\n",
    "    return imgs\n",
    "\n",
    "def make_ratings_image(df, output_file_name=None, cols_without_color=[0], dedup_cols=None, col_width=1.0, row_height=0.625, font_size=18,\n",
    "                     header_color='#C0504D', row_colors=['#F2DCDB', 'w'], edge_color='w',\n",
    "                     bbox=[0, 0, 1, 1], header_columns=0,\n",
    "                     ax=None, size_x=size_x, size_y=size_y, **kwargs):\n",
    "    rating_col = df.iloc[:,-1].name\n",
    "    data = df[df[rating_col]>0].sort_values(rating_col)\n",
    "    data_separeted_numbers = data.copy()\n",
    "    for col in data.columns[1:]:\n",
    "        data_separeted_numbers[col] = data[col].apply(make_sep)\n",
    "        if col in ['Пользователи, %', 'Запросы, %', 'Запросов на пользователя,\\nитого']:\n",
    "            data_separeted_numbers[col] = data[col].apply(lambda x: str(x).replace('.',','))\n",
    "        \n",
    "    if ax is None:\n",
    "        size = (np.array([size_x, size_y]))\n",
    "        fig, ax = plt.subplots(figsize=size, dpi=2**8)\n",
    "        ax.axis('off')\n",
    "\n",
    "    mpl_table = ax.table(cellText=data_separeted_numbers.values, bbox=bbox, colLabels=data.columns, colLoc='center', loc='bottom')\n",
    "    mpl_table.auto_set_font_size(False)\n",
    "    cells_dict = dict(six.iteritems(mpl_table._cells))\n",
    "    \n",
    "    fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "    colors_dict = colorize_column(df, cols_not_to_color=cols_without_color)\n",
    "    colors_dict.update(colorize_column(data, cols_not_to_color=list(range(data.shape[1]-1))), filter_zero=True)\n",
    "    \n",
    "    for k in sorted(cells_dict.keys()): #k[0] - строка, k[1] - столбец\n",
    "        cell = cells_dict[k]\n",
    "        cell.set_edgecolor(edge_color)\n",
    "        row_index = k[0]\n",
    "        column_index = k[1]\n",
    "        \n",
    "        # Высота и цвет заголовков\n",
    "        if row_index == 0:\n",
    "            rotation='horizontal' if k == (0,0) else 'vertical' \n",
    "            cell.set_text_props(color='w', rotation=rotation, ha='center')\n",
    "            cell.set_facecolor(header_color)\n",
    "\n",
    "        # Высота и цвет остальных строк\n",
    "        else:\n",
    "            ha = 'left' if column_index==0 else 'right'\n",
    "            cell.set_facecolor(row_colors[k[0]%len(row_colors)])\n",
    "            cell.set_text_props(wrap=True, ha=ha)\n",
    "        \n",
    "        if row_index == 0:\n",
    "            cell.set_height(24)\n",
    "        elif row_index > 0 and row_index < 15:\n",
    "            cell.set_height(5)\n",
    "        else:\n",
    "            cell.set_height(6)\n",
    "    \n",
    "        # Ширина столбцов\n",
    "        if column_index == 0:\n",
    "            cell.set_width(35)\n",
    "        elif column_index == 6:\n",
    "            cell.set_width(10)\n",
    "        else:\n",
    "            cell.set_width(11)\n",
    "                                       \n",
    "        # Цветвоая шкала\n",
    "        if row_index > 0 and column_index not in cols_without_color:\n",
    "            row_index_origin = data[data.iloc[:,-1] == row_index].index[0]\n",
    "            color_value = colors_dict[column_index][row_index_origin]\n",
    "            cell.set_facecolor(color_value) \n",
    "              \n",
    "        # Шрифт\n",
    "        cell.set_text_props(font=fpath, fontsize=font_size)\n",
    "    if output_file_name:\n",
    "        fig.savefig(f'{output_file_name}', bbox_inches='tight')\n",
    "        #print(output_file_name)\n",
    "    plt.show()\n",
    "    plt.close()  \n",
    "    \n",
    "    \n",
    "def make_ratings_12(file_name,  header_color='#249FD8', row_color='#D2DDF1', cnt=9999, size_x=size_x, size_y=size_y, font_size=font_size, filter_str=''):\n",
    "    dt_now = str(datetime.datetime.now())[2:-9].replace('-','').replace(':','').replace(' ','_')\n",
    "    df = pd.read_excel(file_name)\n",
    "    word_file = f'{file_name[:-17]}_report_{dt_now}.docx'\n",
    "    #print(word_file)\n",
    "    \n",
    "    df = df[(df['№']!='Итог')]\n",
    "    cols = [\n",
    "        'Рубрика' \n",
    "        ,'Пользователи, итого'\n",
    "        ,'Пользователи, %'\n",
    "        ,'Запросы, итого'\n",
    "        ,'Запросы, %'\n",
    "        ,'Запросов на пользователя,\\nитого'\n",
    "    ]\n",
    "\n",
    "    ratings_cols = [i for i in df.columns if 'ейтинг' in i and filter_str in i.lower() and i not in  ('В рейтингах, раз','Рейтинг суммарный','Рейтинг средний','Вне рейтингов')]\n",
    "\n",
    "    for col in ratings_cols:\n",
    "        df[col] = df[col].astype('int').round()\n",
    "\n",
    "    for col in ['Пользователи, %', 'Запросы, %', 'Запросов на пользователя,\\nитого']:\n",
    "        df[col] = df[col].astype('float64').round(1)\n",
    "    \n",
    "    i=0\n",
    "    imgs = []\n",
    "    for col in ratings_cols[:cnt]:  \n",
    "        i+=1\n",
    "        cols_filter = cols+[col]    \n",
    "        dfx = df[df[col]>0].sort_values(col)['Рубрика'].values\n",
    "        img_file_name = 'pic_rating' if filter_str == '' else 'pic_antirating'\n",
    "        img_file_name_final = f'pics/{img_file_name}_{i:02}.png'.replace('ten_city/','')\n",
    "        df_cols_filter = df[cols_filter]\n",
    "        new_cols = [\n",
    "        'Рубрика' \n",
    "        ,'Пользователи,\\nитого'\n",
    "        ,'Пользователи, %'\n",
    "        ,'Запросы, итого'\n",
    "        ,'Запросы, %'\n",
    "        ,'Запросов\\nна пользователя,\\nитого'\n",
    "        ] +[col.replace('Рейтинг ', 'Рейтинг\\n').replace('Антирейтинг ', 'Антирейтинг\\n').replace('пользователям ', 'пользователям,\\n')]  \n",
    "        df_cols_filter.columns = new_cols\n",
    "            \n",
    "        make_ratings_image(df_cols_filter, output_file_name=img_file_name_final, cols_without_color=[0], header_color=header_color, row_colors=[row_color, 'w'], size_x=size_x, size_y=size_y, font_size=font_size)\n",
    "        txt =  col + '\\n' + ', '.join([word[::-1].replace('/', 'и', 1)[::-1].replace('/', ',').replace(' ,', ',').lower() for word in df[df[col]>0].sort_values(col)['Рубрика'].values]) + \":\"\n",
    "        #add_image(word_file, img_file_name_final, word=txt, img_w=img_w, img_h=img_h, blank=blank, word_sample=word_sample)\n",
    "        imgs.append(img_file_name_final)\n",
    "    return imgs\n",
    "\n",
    "\n",
    "def delete_recta():\n",
    "    prs = pptx.Presentation('pics/Презентация_Мамедов_Рамиль_230527_2145.pptx')\n",
    "    slides = prs.slides\n",
    "    slides_elements = {j:{i.name[:5]:i for i in slides[j].shapes} for j in range(len(slides))}\n",
    "\n",
    "    shapes = [[i for i in s.shapes if i.name[:5] == 'Recta'] for s in slides]\n",
    "    shapes = [i for i in shapes if i]\n",
    "\n",
    "    for shape in shapes:\n",
    "        shape_element = shape[0].element\n",
    "        shape_element.getparent().remove(shape_element)\n",
    "\n",
    "    prs.save('pics/Презентация_Мамедов_Рамиль_230527_2145.pptx')\n",
    "    \n",
    "    \n",
    "def save_pptx():\n",
    "    while True:\n",
    "        try:\n",
    "            prs.save('pics/test_new.pptx')\n",
    "            break\n",
    "        except:\n",
    "            pass\n",
    "\n",
    "    _ = subprocess.Popen([r\"C:\\Program Files\\Microsoft Office\\root\\Office16\\POWERPNT.EXE\",\n",
    "                         r\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\pics\\test_new.pptx\"])\n",
    "    print(str(datetime.datetime.now())[:-7], 'test_new.pptx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "86f0d822-2e8e-4b70-a209-dd4f55d992b8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T21:46:38.199258Z",
     "iopub.status.busy": "2023-05-28T21:46:38.198248Z",
     "iopub.status.idle": "2023-05-28T21:47:47.190376Z",
     "shell.execute_reply": "2023-05-28T21:47:47.189464Z",
     "shell.execute_reply.started": "2023-05-28T21:46:38.199258Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "now = str(datetime.datetime.now())[:-9].replace('-','').replace(' ','_').replace(':','')\n",
    "dfn = pd.read_excel(r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\ВШЭ\\_Backups\\Таблица_Население_230425_0030.xlsx\")\n",
    "pic1 = f'pics/table_{now}.png'\n",
    "make_hse_image2(dfn, pic1, show=True, font_size=38)\n",
    "\n",
    "now = str(datetime.datetime.now())[:-9].replace('-','').replace(' ','_').replace(':','')\n",
    "pic2= f'pics/users_{now}.png'\n",
    "make_hse_image(df_users, pic2, show=True, font_size=38)\n",
    "\n",
    "now = str(datetime.datetime.now())[:-9].replace('-','').replace(' ','_').replace(':','')\n",
    "pic3 = f'pics/requests_{now}.png'\n",
    "make_hse_image(df_requests, pic3, show=True, font_size=38)\n",
    "\n",
    "pic4, pic5 = make_ratings_word(\"files/ten_city/Мастер_таблица_2ГИС_10_городов_230428_1845.xlsx\", header_color='#249FD8', row_color='#D2DDF1', size_x=size_x, size_y=size_y, cnt=1, font_size=38, img_w=16.5, img_h=10, blank=False, word_sample='files/report_ten_city.docx')\n",
    "\n",
    "pic6, = make_plt_requsts('files/ten_city/requests_text_ten_city_230504_0256.xlsx', \"\", '#249FD8', '#D2DDF1', \n",
    "                   size_x=size_x, size_y=size_y, cnt=1, font_size=38, img_w=16.5, img_h=10, blank=True, word_sample='files/report_ten_city.docx')\n",
    "\n",
    "ratings_pics = make_ratings_12(\"files/ten_city/Мастер_таблица_2ГИС_10_городов_230428_1845.xlsx\", cnt=12,\n",
    "                              font_size=font_size/4, size_x = size_x/4, size_y=size_y/3)\n",
    "\n",
    "antiratings_pics = make_ratings_12(\"files/ten_city/Мастер_таблица_2ГИС_10_городов_230428_1845.xlsx\", cnt=12,\n",
    "                              font_size=font_size/4, size_x = size_x/4, size_y=size_y/3, filter_str='анти')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "908e1421-b4e3-4283-aa47-fff8f3ffb239",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T22:00:36.492282Z",
     "iopub.status.busy": "2023-05-28T22:00:36.492282Z",
     "iopub.status.idle": "2023-05-28T22:00:50.719942Z",
     "shell.execute_reply": "2023-05-28T22:00:50.719942Z",
     "shell.execute_reply.started": "2023-05-28T22:00:36.492282Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "requsts_pics = make_plt_requsts('files/ten_city/requests_text_ten_city_230504_0256.xlsx', \"\", '#249FD8', '#D2DDF1', \n",
    "                   size_x=size_x/4, size_y=size_y/3, cnt=12, font_size=font_size/4, img_w=16.5, img_h=10, blank=True, word_sample='files/report_ten_city.docx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2d76d5b6-17f5-43c0-99db-90d44a26a146",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T21:47:47.191377Z",
     "iopub.status.busy": "2023-05-28T21:47:47.191377Z",
     "iopub.status.idle": "2023-05-28T21:47:47.460535Z",
     "shell.execute_reply": "2023-05-28T21:47:47.459619Z",
     "shell.execute_reply.started": "2023-05-28T21:47:47.191377Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pptx\n",
    "from pptx.util import Cm, Pt\n",
    "prs = pptx.Presentation('pics/test.pptx')\n",
    "slides = prs.slides\n",
    "slides_elements = {j:{i.name[:5]:i for i in slides[j].shapes} for j in range(len(slides))}\n",
    "pics_to_add = [pic1, pic2, pic3, pic4, pic5, pic6]\n",
    "now = str(datetime.datetime.now())[:-9].replace('-','').replace(' ','_').replace(':','')[2:]\n",
    "\n",
    "for n, i in enumerate([2, 3, 4, 5, 7, 9]):\n",
    "    slides[i].shapes.add_picture(pics_to_add[n], Cm(0.7), Cm(2.4), Cm(30.7), Cm(16))\n",
    "    shape_i = slides_elements[i]['Капля']\n",
    "    p = shape_i.text_frame.add_paragraph()\n",
    "    p.text = now\n",
    "    p.font.size = Pt(12)\n",
    "\n",
    "#prs.save('pics/test_new.pptx')\n",
    "#prs.save(r\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\pics\\Презентация_Мамедов Рамиль.pptx\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "4638f928-224f-4a54-b76f-345c7b6d07fe",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T22:04:33.662364Z",
     "iopub.status.busy": "2023-05-28T22:04:33.662364Z",
     "iopub.status.idle": "2023-05-28T22:04:33.737083Z",
     "shell.execute_reply": "2023-05-28T22:04:33.737083Z",
     "shell.execute_reply.started": "2023-05-28T22:04:33.662364Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def add_pictures(slide, pics_list, x=7.7, y=5.4):\n",
    "    slide.shapes.add_picture(pics_list[0], Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[1], Cm(0.7+x), Cm(2.4), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[2], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[3], Cm(0.7+x*3), Cm(2.4), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[4], Cm(0.7), Cm(2.4+y), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[5], Cm(0.7+x), Cm(2.4+y), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[6], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[7], Cm(0.7+x*3), Cm(2.4+y), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[8], Cm(0.7), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[9], Cm(0.7+x), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[10], Cm(0.7+x*2), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "    slide.shapes.add_picture(pics_list[11], Cm(0.7+x*3), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "\n",
    "    \n",
    "antiratings_pics.insert(6, antiratings_pics[10])\n",
    "add_pictures(slides[6], ratings_pics)\n",
    "add_pictures(slides[8], antiratings_pics)\n",
    "add_pictures(slides[10], requsts_pics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "c73ad361-769d-49ad-94e5-7fa79f600410",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T19:24:57.158666Z",
     "iopub.status.busy": "2023-05-29T19:24:57.157667Z",
     "iopub.status.idle": "2023-05-29T19:24:57.177672Z",
     "shell.execute_reply": "2023-05-29T19:24:57.176675Z",
     "shell.execute_reply.started": "2023-05-29T19:24:57.158666Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def make_linear_plots(input_filename, rubrics_df, cnt=9999, size_x=size_x, size_y=size_y, font_size=font_size):\n",
    "    def zero(x,y):\n",
    "        if x < 0:\n",
    "            return str(0) +'%'\n",
    "        return str(round(x,1)) +'%'\n",
    "    \n",
    "    dt_now = str(datetime.datetime.now())[2:-9].replace('-','').replace(':','').replace(' ','_')\n",
    "\n",
    "    cols1 = ['Рубрика','Час','Количество запросов Астана','Количество запросов Бишкек','Количество запросов Москва','Количество запросов ОАЭ','Количество запросов Ташкент']\n",
    "    cols2 = ['Рубрика','Дата','Количество запросов Астана','Количество запросов Бишкек','Количество запросов Москва','Количество запросов ОАЭ','Количество запросов Ташкент']\n",
    "    cols3 = ['Рубрика','Неделя','Количество запросов Астана','Количество запросов Бишкек','Количество запросов Москва','Количество запросов ОАЭ','Количество запросов Ташкент']\n",
    "    cols4 = ['Рубрика','День недели','Количество запросов Астана','Количество запросов Бишкек','Количество запросов Москва','Количество запросов ОАЭ','Количество запросов Ташкент']\n",
    "    dff1 = pd.read_excel(input_filename, sheet_name='Часы')[cols1]\n",
    "    dff2 = pd.read_excel(input_filename, sheet_name='Даты')[cols2]\n",
    "    dff3 = pd.read_excel(input_filename, sheet_name='Недели')[cols3]\n",
    "    dff4 = pd.read_excel(input_filename, sheet_name='Дни недели')[cols4]\n",
    "    \n",
    "    i = 0\n",
    "    imgs = []\n",
    "    for rubric in tqdm(rubrics_df[:].iloc[:,0].values[0:cnt]):\n",
    "        dff1_test = dff1[dff1['Рубрика'] == rubric].set_index(dff1.columns[1])\n",
    "        dff2_test = dff2[dff2['Рубрика'] == rubric].set_index(dff2.columns[1])\n",
    "        dff3_test = dff3[dff3['Рубрика'] == rubric].set_index(dff3.columns[1])\n",
    "        dff4_test = dff4[dff4['Рубрика'] == rubric].set_index(dff4.columns[1])\n",
    "        dff4_test.index = (['пн','вт','ср','чт','пт','сб','вс'])\n",
    "        dff1_test.index = dff1_test.index.astype(str)\n",
    "        i+=1\n",
    "        j=0\n",
    "        for d in [dff1_test, dff2_test, dff3_test, dff4_test]: \n",
    "            j+=1\n",
    "            fig, ax = plt.subplots()\n",
    "            ax = sns.lineplot(data = d, linewidth=3, palette=['#4F81BD', '#F79646', '#C0504D', '#9BBB59', '#264478'])\n",
    "            plt.rcParams['figure.figsize']=size_x,size_y\n",
    "            plt.rcParams['font.size'] = font_size\n",
    "            plt.rcParams['figure.dpi'] = 2**8\n",
    "            plt.rcParams['figure.facecolor'] = 'white'\n",
    "            fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "            \n",
    "            \n",
    "            plt.legend(['Астана', 'Бишкек', 'Москва', 'ОАЭ', 'Ташкент'], prop=fpath)\n",
    "            plt.xticks(font=fpath)\n",
    "            plt.yticks(font=fpath)\n",
    "            plt.grid(visible=True, linewidth=0.3)\n",
    "            ax.set_xlabel('')\n",
    "            ax.yaxis.set_major_formatter(FuncFormatter(zero))\n",
    "            ax.xaxis.set_major_locator(ticker.MultipleLocator(1)) \n",
    "            ax.yaxis.set_major_locator(ticker.LinearLocator(numticks=6))  \n",
    "            title_text = f'{i:02}/{j}. {rubric}. '\n",
    "            img_file_name = input_filename.replace('files/','')[:-5]\n",
    "            img_file_name_final = f'pictures/{img_file_name}_{i:02}_{j}.png'\n",
    "            imgs.append(img_file_name_final)\n",
    "            if 'Час' == d.index.name:\n",
    "                ax.set_title(f'{title_text}Запросы по часам, шт.', font=fpath)\n",
    "            elif 'Дата' == d.index.name:\n",
    "                ax.xaxis.set_major_locator(ticker.MultipleLocator(4)) \n",
    "                ax.set_title(f'{title_text}Запросы по датам, шт.', font=fpath)\n",
    "            elif 'Неделя' == d.index.name:\n",
    "                ax.set_title(f'{title_text}Запросы по неделям, шт.', font=fpath)\n",
    "            else:\n",
    "                ax.set_title(f'{title_text}Запросы по дням недели, шт.', font=fpath) \n",
    "            \n",
    "            ax.figure.savefig(img_file_name_final, bbox_inches='tight')\n",
    "            #plt.show()\n",
    "            plt.close()\n",
    "    return imgs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "1ba959ba-6f97-471c-b1e0-53c552dd8c45",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T22:30:17.743149Z",
     "iopub.status.busy": "2023-05-28T22:30:17.743149Z",
     "iopub.status.idle": "2023-05-28T22:30:50.607176Z",
     "shell.execute_reply": "2023-05-28T22:30:50.607176Z",
     "shell.execute_reply.started": "2023-05-28T22:30:17.743149Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "54eba4d33a8c4f2585745b0fd8c2b537",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rubrics = pd.DataFrame(pd.read_parquet(f\"files/ten_city/df_tx_grouped_filtered_ten_city\", columns=['rubricName'])['rubricName'].unique())\n",
    "rubrics['Рейтинг рубрики'] = rubrics.index + 1\n",
    "grapics_pics = make_linear_plots(f'files/ten_city/df_dates.xlsx', rubrics, 4, size_x=size_x/2, size_y=size_y/2, font_size=font_size/2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "6e134cac-184d-4b18-b152-bbf04d78b7bd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T19:51:39.600177Z",
     "iopub.status.busy": "2023-05-29T19:51:39.599177Z",
     "iopub.status.idle": "2023-05-29T19:51:40.071633Z",
     "shell.execute_reply": "2023-05-29T19:51:40.070568Z",
     "shell.execute_reply.started": "2023-05-29T19:51:39.600177Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "slides_elements = {j:{i.name[:5]:i for i in slides[j].shapes} for j in range(len(slides))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "fdaf104d-019b-4c6d-80c9-1002294480c1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T22:34:21.890453Z",
     "iopub.status.busy": "2023-05-28T22:34:21.890453Z",
     "iopub.status.idle": "2023-05-28T22:34:21.951010Z",
     "shell.execute_reply": "2023-05-28T22:34:21.950056Z",
     "shell.execute_reply.started": "2023-05-28T22:34:21.890453Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b68e20ffd0>"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = 15.3 \n",
    "y = 8\n",
    "slides[11].shapes.add_picture(grapics_pics[0], Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[11].shapes.add_picture(grapics_pics[1], Cm(0.7+x), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[11].shapes.add_picture(grapics_pics[2], Cm(0.7), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[11].shapes.add_picture(grapics_pics[3], Cm(0.7+x), Cm(2.4+y), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "b6fb082f-c53d-4cbe-bdaf-0692aad4d5cd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-28T22:34:23.125946Z",
     "iopub.status.busy": "2023-05-28T22:34:23.125946Z",
     "iopub.status.idle": "2023-05-28T22:34:24.057882Z",
     "shell.execute_reply": "2023-05-28T22:34:24.057882Z",
     "shell.execute_reply.started": "2023-05-28T22:34:23.125946Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "901d7862-6721-4c70-a615-59c3c89a1ee3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T19:42:13.182775Z",
     "iopub.status.busy": "2023-05-29T19:42:13.182775Z",
     "iopub.status.idle": "2023-05-29T19:43:06.629509Z",
     "shell.execute_reply": "2023-05-29T19:43:06.628634Z",
     "shell.execute_reply.started": "2023-05-29T19:42:13.182775Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f07cd43fcd6f43d5b2222ca6ab0ee743",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/24 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rubrics = pd.DataFrame(pd.read_parquet(f\"files/ten_city/df_tx_grouped_filtered_ten_city\", columns=['rubricName'])['rubricName'].unique())\n",
    "rubrics['Рейтинг рубрики'] = rubrics.index + 1\n",
    "rubrics_sample_index = [56,  9, 98,  7, 13,  5, 74, 60, 79, 73, 90, 34, 54, 58, 47, 89,  6, 46, 29, 65, 62, 39, 18, 26, 41, 81, 87, 83, 25, 16]\n",
    "rubrics_sample = rubrics.iloc[rubrics_sample_index]\n",
    "grapics_pics_24 = make_linear_plots(f'files/ten_city/df_dates.xlsx', rubrics_sample, 24, size_x=size_x/6, size_y=size_y/4, font_size=font_size/4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "1271df02-5894-465f-968e-ac1f4b77131e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T19:43:21.829735Z",
     "iopub.status.busy": "2023-05-29T19:43:21.829735Z",
     "iopub.status.idle": "2023-05-29T19:43:22.067650Z",
     "shell.execute_reply": "2023-05-29T19:43:22.066655Z",
     "shell.execute_reply.started": "2023-05-29T19:43:21.829735Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b76ce57160>"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = 5\n",
    "y = 4\n",
    "slides[12].shapes.add_picture(grapics_pics_24[0], Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[1], Cm(0.7+x), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[2], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[3], Cm(0.7+x*3), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[4], Cm(0.7+x*4), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[5], Cm(0.7+x*5), Cm(2.4), Cm(x), Cm(y))\n",
    "\n",
    "slides[12].shapes.add_picture(grapics_pics_24[6], Cm(0.7)    , Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[7], Cm(0.7+x)  , Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[8], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[9], Cm(0.7+x*3), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[10], Cm(0.7+x*4), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[11], Cm(0.7+x*5), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[12].shapes.add_picture(grapics_pics_24[12], Cm(0.7)    , Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[13], Cm(0.7+x)  , Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[14], Cm(0.7+x*2), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[15], Cm(0.7+x*3), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[16], Cm(0.7+x*4), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[17], Cm(0.7+x*5), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "\n",
    "slides[12].shapes.add_picture(grapics_pics_24[18], Cm(0.7)    , Cm(2.4+y*3), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[19], Cm(0.7+x)  , Cm(2.4+y*3), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[20], Cm(0.7+x*2), Cm(2.4+y*3), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[21], Cm(0.7+x*3), Cm(2.4+y*3), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[22], Cm(0.7+x*4), Cm(2.4+y*3), Cm(x), Cm(y))\n",
    "slides[12].shapes.add_picture(grapics_pics_24[23], Cm(0.7+x*5), Cm(2.4+y*3), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "id": "41f5cb2c-22c7-4971-b6cc-e77f25ee89c5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T21:07:01.060140Z",
     "iopub.status.busy": "2023-05-29T21:07:01.060140Z",
     "iopub.status.idle": "2023-05-29T21:07:02.399167Z",
     "shell.execute_reply": "2023-05-29T21:07:02.398241Z",
     "shell.execute_reply.started": "2023-05-29T21:07:01.060140Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 00:07:02 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "id": "2e2e1401-a223-4e8f-9367-c078e29f54b0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T21:42:47.276365Z",
     "iopub.status.busy": "2023-05-29T21:42:47.275363Z",
     "iopub.status.idle": "2023-05-29T21:42:54.160182Z",
     "shell.execute_reply": "2023-05-29T21:42:54.159298Z",
     "shell.execute_reply.started": "2023-05-29T21:42:47.276365Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from matplotlib.font_manager import fontManager, FontProperties\n",
    "def rounder(x,y):\n",
    "    return str(round(x))\n",
    "\n",
    "def prcnter(x,y):\n",
    "    if x == 0 or x > 1:\n",
    "        return ''\n",
    "    return str(round(x*100)) + '%'\n",
    "\n",
    "i = 0\n",
    "dates = [r\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\files\\ten_city\\df_dates_city.xlsx\",\n",
    "r\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\files\\ten_city\\df_dates_rubrics.xlsx\",\n",
    "r\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\files\\ten_city\\df_dates_rubrics_city.xlsx\"]\n",
    "\n",
    "pics = []\n",
    "for file in dates:\n",
    "    for col_r in ['hour_r', 'date_r', 'week_r', 'week_day_r']:\n",
    "        xr = pd.read_excel(file, sheet_name=col_r)\n",
    "        xr['diff'] = xr['diff'].apply(lambda x: 1 if x > 1 else x)\n",
    "        file_name = file.split('\\\\')[-1].replace('.xlsx','').replace('df_dates_','')\n",
    "        sns_plot = sns.displot(data=xr, x=\"diff\", kde=True)\n",
    "        sns.set(rc={'figure.figsize':(size_y/2, size_x/6), \"figure.dpi\": 2**8})\n",
    "        sns.set(rc={'axes.facecolor':'white', 'figure.facecolor':'white'})\n",
    "        fpath = r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\"\n",
    "        fontManager.addfont(fpath)\n",
    "        prop = FontProperties(fname=fpath)\n",
    "        fig = sns_plot.figure\n",
    "        ax = sns_plot.ax\n",
    "        ax.yaxis.set_major_formatter(FuncFormatter(rounder))\n",
    "        ax.xaxis.set_major_formatter(FuncFormatter(prcnter))\n",
    "        ax.yaxis.set_major_locator(ticker.LinearLocator(numticks=5))\n",
    "        ax.set_xlabel('')\n",
    "        ax.set_ylabel('')\n",
    "        one_dict = {'city':'городов','rubrics': 'бизнесов','rubrics_city':'бизнесов в городах'}\n",
    "        file_name = one_dict[file_name]\n",
    "        two_dict = {'hour_r': 'часам', 'date_r':'датам', 'week_r':'неделям', 'week_day_r':'дням недели'}\n",
    "        col_r = two_dict[col_r]\n",
    "        ax.set_title(f'Схожесть {file_name} по {col_r}'.replace('_r',''), y=0.98, font=prop.get_name())\n",
    "        i+=1\n",
    "        fig.savefig(f\"pictures/ten_city/displot_{i:02}.png\")\n",
    "        pics.append(f\"pictures/ten_city/displot_{i:02}.png\")\n",
    "        plt.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "8869b530-ddd1-41bb-9d63-bb7d1d06052c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T21:50:23.508378Z",
     "iopub.status.busy": "2023-05-29T21:50:23.507390Z",
     "iopub.status.idle": "2023-05-29T21:50:23.666506Z",
     "shell.execute_reply": "2023-05-29T21:50:23.665626Z",
     "shell.execute_reply.started": "2023-05-29T21:50:23.508378Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "slides_elements = {j:{i.name[:5]:i for i in slides[j].shapes} for j in range(len(slides))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "id": "dafa18be-09e5-4c13-b91b-8381725e368f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T21:50:24.138701Z",
     "iopub.status.busy": "2023-05-29T21:50:24.137696Z",
     "iopub.status.idle": "2023-05-29T21:50:24.283227Z",
     "shell.execute_reply": "2023-05-29T21:50:24.282362Z",
     "shell.execute_reply.started": "2023-05-29T21:50:24.138701Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b68324b790>"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = 16.1 / 2\n",
    "x = 10.2 / 2\n",
    "slides[13].shapes.add_picture(pics[0], Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[1], Cm(0.7+x), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[2], Cm(0.7), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[3], Cm(0.7+x), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[13].shapes.add_picture(pics[4], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[5], Cm(0.7+x+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[6], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[7], Cm(0.7+x+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[13].shapes.add_picture(pics[8], Cm(0.7+x*4), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[9], Cm(0.7+x+x*4), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[10], Cm(0.7+x*4), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[13].shapes.add_picture(pics[11], Cm(0.7+x+x*4), Cm(2.4+y), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "id": "89f3d044-bc2c-4ca8-9df1-9688e2c44e82",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T21:50:25.668324Z",
     "iopub.status.busy": "2023-05-29T21:50:25.667324Z",
     "iopub.status.idle": "2023-05-29T21:50:26.740128Z",
     "shell.execute_reply": "2023-05-29T21:50:26.739128Z",
     "shell.execute_reply.started": "2023-05-29T21:50:25.668324Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 00:50:26 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "id": "746886f0-4ce8-410a-8c1f-5954b27ce670",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:33:53.379205Z",
     "iopub.status.busy": "2023-05-29T22:33:53.378206Z",
     "iopub.status.idle": "2023-05-29T22:33:53.402205Z",
     "shell.execute_reply": "2023-05-29T22:33:53.401311Z",
     "shell.execute_reply.started": "2023-05-29T22:33:53.379205Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_matrix(file, city, cnt_graphics=50, cnt_shars=12, step=1, size_x=size_x, size_y=size_y, font_size=font_size, size_coef=1):\n",
    "    shar_nums = list(range(1,16))*500\n",
    "    start_range = 1\n",
    "    stop_range = cnt_shars*step+1\n",
    "    time = str(datetime.datetime.now())[2:16].replace('-','').replace(' ','_').replace(':','')\n",
    "    print(f'files/{city}/{city}_shars_{time}.docx')\n",
    "    df_list = []\n",
    "    pics=[]\n",
    "    for graphic_num in tqdm(range(cnt_graphics)):\n",
    "        shar_num = shar_nums.pop()\n",
    "\n",
    "        df = pd.read_excel(file)[['Рубрика','Запросы 2022', 'Запросы 2023', 'Запросы, итого']].iloc[range(start_range, stop_range, step)].set_index('Рубрика').fillna(0)\n",
    "        start_range = stop_range\n",
    "        stop_range = (graphic_num+2)*cnt_shars*step+1\n",
    "\n",
    "        df['Запросы, итого'] = (df['Запросы, итого'] - min(df['Запросы, итого'])) / (max(df['Запросы, итого']) - min(df['Запросы, итого']))*300\n",
    "        requests_total = pd.read_excel(file)['Запросы 2023'].sum()\n",
    "        df['Рост/Падение, %'] = df['Запросы 2023'] / df['Запросы 2022']\n",
    "        df['Доля рынка, %'] = df['Запросы 2023'] / requests_total\n",
    "        x = (df['Доля рынка, %'])\n",
    "        y = (df['Рост/Падение, %'])\n",
    "        sizes = (df['Запросы, итого']/1000)\n",
    "        x_avg = (x.min()+x.max())/2\n",
    "        y_avg = (y.min()+y.max())/2\n",
    "        xi, yi = df['Доля рынка, %'], df['Рост/Падение, %']\n",
    "        df.loc[(xi > x_avg) & (yi > y_avg), 'group'] = 'Звёзды:'\n",
    "        df.loc[(xi > x_avg)  & (yi <= y_avg), 'group'] = 'Коровы:'\n",
    "        df.loc[(xi <= x_avg)  & (yi > y_avg), 'group'] = 'Дети:'\n",
    "        df.loc[(xi <= x_avg)  & (yi <= y_avg), 'group'] = 'Собаки:'\n",
    "        df = df.sort_values(['Доля рынка, %', 'Рост/Падение, %'], ascending=[False, False])\n",
    "\n",
    "        plt.rcParams['figure.figsize'] = size_x, size_y\n",
    "        plt.rcParams['figure.dpi'] = 2**8\n",
    "        plt.rcParams['font.size'] = font_size\n",
    "        fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "        fig, ax = plt.subplots()\n",
    "        ax.scatter(x, y, sizes, c='#C0504D', alpha=0.7)\n",
    "\n",
    "        n = 0.02\n",
    "        m = 0.02\n",
    "        ax.set_ylim(ax.get_ylim()[0] * (1 - n), ax.get_ylim()[1] * (1 + n))\n",
    "        ax.set_xlim(ax.get_xlim()[0] * (1 - m), ax.get_xlim()[1] * (1 + m))\n",
    "        ax.spines['top'].set_visible(False)\n",
    "        ax.spines['bottom'].set_visible(False)\n",
    "        ax.spines['left'].set_visible(False)\n",
    "        ax.spines['right'].set_visible(False)\n",
    "        ax.set_xticklabels('')\n",
    "        ax.set_yticklabels('')\n",
    "        plt.tick_params(bottom = False, left=False)\n",
    "\n",
    "        for xi, yi, rubric, size in zip(x, y, x.index, sizes):\n",
    "            rs = rubric.split(' ')\n",
    "            if len(rs) > 1:\n",
    "                rsl = len(rs) // 2\n",
    "                rubric = ' '.join(rs[:rsl] + ['\\n'] + rs[rsl:])\n",
    "\n",
    "            shar_path = [r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\Шар_{:02}.png\".format(shar_num)]\n",
    "            shar_img = OffsetImage(plt.imread(shar_path[0], format=\"png\"), zoom=size/size_coef)\n",
    "            ab = AnnotationBbox(shar_img, (xi, yi), frameon=False)\n",
    "            ax.add_artist(ab)\n",
    "            plt.annotate(rubric, (xi, yi), ha='center', va='center', font=fpath, color='#404040')\n",
    "\n",
    "        ax.hlines((y.min()+y.max())/2, x.min(), x.max(), colors='gray', linewidth=1)\n",
    "        ax.vlines((x.min()+x.max())/2, y.min(), y.max(), colors='gray', linewidth=1)\n",
    "        \n",
    "        fig.savefig(f'pictures/{city}_shar_{shar_num}.png', bbox_inches='tight', transparent=False)\n",
    "        pics.append(f'pictures/{city}_shar_{shar_num}.png')\n",
    "        txt = ''\n",
    "        for cat in ['Звёзды:','Коровы:','Дети:','Собаки:']:\n",
    "            txt +=  cat + ' ' + ', '.join([word[::-1].replace('/', 'и', 1)[::-1].replace('/', ',').replace(' ,', ',').lower() for word in df[df['group'] == cat].index]) + \"\\n\"\n",
    "        txt = txt[::-1].replace('\\n',':', 1)[::-1]\n",
    "        #plt.show()\n",
    "        plt.close()\n",
    "        df_list.append(df)\n",
    "    return pics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "1966003f-f992-4bb4-8b1e-f8388db2af05",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:21:47.240139Z",
     "iopub.status.busy": "2023-05-29T22:21:47.239230Z",
     "iopub.status.idle": "2023-05-29T22:21:47.254136Z",
     "shell.execute_reply": "2023-05-29T22:21:47.253135Z",
     "shell.execute_reply.started": "2023-05-29T22:21:47.240139Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "file = r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\10 городов\\Мастер_таблица_2ГИС_10_городов_230419_1700_ext.xlsx\"\n",
    "city = 'ten_city'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "006cf8d9-c178-40e4-90e9-07f43e5e3c46",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:21:48.748785Z",
     "iopub.status.busy": "2023-05-29T22:21:48.748785Z",
     "iopub.status.idle": "2023-05-29T22:21:57.032795Z",
     "shell.execute_reply": "2023-05-29T22:21:57.031794Z",
     "shell.execute_reply.started": "2023-05-29T22:21:48.748785Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "matrix_pics = plot_matrix(file, city, cnt_graphics=1, cnt_shars=12, step=1, size_x=size_x, size_y=size_y, font_size=font_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "id": "68d5bd86-bb87-48be-aab2-01ac5b7b9163",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:39:16.637986Z",
     "iopub.status.busy": "2023-05-29T22:39:16.637986Z",
     "iopub.status.idle": "2023-05-29T22:39:16.767704Z",
     "shell.execute_reply": "2023-05-29T22:39:16.766823Z",
     "shell.execute_reply.started": "2023-05-29T22:39:16.637986Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "#slides[14].shapes.add_picture(matrix_pics[0], Cm(0.7), Cm(2.4), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7e4f0ebf-c091-4ba4-a89e-0cee89f083d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:33:55.337650Z",
     "iopub.status.busy": "2023-05-29T22:33:55.337650Z",
     "iopub.status.idle": "2023-05-29T22:34:41.702707Z",
     "shell.execute_reply": "2023-05-29T22:34:41.701713Z",
     "shell.execute_reply.started": "2023-05-29T22:33:55.337650Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "matrix_pics_12 = plot_matrix(file, city, cnt_graphics=12, cnt_shars=12, step=1, size_x=size_x/4, size_y=size_y/3, font_size=font_size/3, size_coef=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "id": "9508b44b-6f66-4932-a88e-c724f3aefded",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:39:21.522039Z",
     "iopub.status.busy": "2023-05-29T22:39:21.522039Z",
     "iopub.status.idle": "2023-05-29T22:39:21.623580Z",
     "shell.execute_reply": "2023-05-29T22:39:21.623580Z",
     "shell.execute_reply.started": "2023-05-29T22:39:21.522039Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b68bdc0af0>"
      ]
     },
     "execution_count": 260,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = 7.7\n",
    "y = 5.3\n",
    "slides[15].shapes.add_picture(matrix_pics_12[0], Cm(0.7),     Cm(2.4), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[1], Cm(0.7+x),   Cm(2.4), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[2], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[3], Cm(0.7+x*3), Cm(2.4), Cm(x), Cm(y))\n",
    "\n",
    "slides[15].shapes.add_picture(matrix_pics_12[4], Cm(0.7),     Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[5], Cm(0.7+x),   Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[6], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[7], Cm(0.7+x*3), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[15].shapes.add_picture(matrix_pics_12[8],  Cm(0.7),     Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[9],  Cm(0.7+x),   Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[10], Cm(0.7+x*2), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[15].shapes.add_picture(matrix_pics_12[11], Cm(0.7+x*3), Cm(2.4+y*2), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "id": "7103aa45-0301-4057-9b59-d8c481c89648",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:39:23.275304Z",
     "iopub.status.busy": "2023-05-29T22:39:23.275304Z",
     "iopub.status.idle": "2023-05-29T22:39:24.564039Z",
     "shell.execute_reply": "2023-05-29T22:39:24.564039Z",
     "shell.execute_reply.started": "2023-05-29T22:39:23.275304Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 01:39:24 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f5c2e1f-976f-4aa7-9b0e-1c009ca0b85f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f143510e-100e-4ffa-b21b-55f47a8d2619",
   "metadata": {
    "tags": []
   },
   "source": [
    "### 1.10.\tАнализ Матриц BCG по городам"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "a9c45e29-b482-4ff3-a4d1-868f59a73f23",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T10:51:33.337494Z",
     "iopub.status.busy": "2023-06-08T10:51:33.336512Z",
     "iopub.status.idle": "2023-06-08T10:51:33.369079Z",
     "shell.execute_reply": "2023-06-08T10:51:33.368094Z",
     "shell.execute_reply.started": "2023-06-08T10:51:33.337494Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_matrix_cities(cnt_graphics=100, show=False, size_x=size_x, size_y=size_y, font_size=font_size, size_coef=1):\n",
    "    folder='ten_city'\n",
    "    shar_nums = list(range(1,16))*500\n",
    "    time = str(datetime.datetime.now())[2:16].replace('-','').replace(' ','_').replace(':','')\n",
    "    df_src = pd.read_parquet(f'files/ten_city/requests_for_shars')\n",
    "    df_new = df_src[['rubricName','city']].drop_duplicates().set_index(['rubricName','city'])\n",
    "    df_new['Запросы 2020'] = df_src[df_src['year_r'] == 2020].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2021'] = df_src[df_src['year_r'] == 2021].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2022'] = df_src[df_src['year_r'] == 2022].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2023'] = df_src[df_src['year_r'] == 2023].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new = df_new.fillna(0).astype(int)\n",
    "    df_new['Запросы, итого'] =  df_new['Запросы 2020'] + df_new['Запросы 2021'] + df_new['Запросы 2022'] + df_new['Запросы 2023']\n",
    "    df_new = df_new.reset_index().set_index('city')\n",
    "\n",
    "    rubrics = dd.read_parquet(f\"files/ten_city/df_tx_grouped_filtered_ten_city\")['rubricName'].unique().compute()\n",
    "    rubrics = [i for i in rubrics.to_dict().values()]\n",
    "    df_list = []\n",
    "    pics = []\n",
    "    for rubric in tqdm(rubrics[:cnt_graphics]):\n",
    "        shar_num = shar_nums.pop()\n",
    "        df = df_new.loc[df_new['rubricName']==rubric]\n",
    "        df['Запросы, итого'] = (df['Запросы, итого'] - min(df['Запросы, итого'])) / (max(df['Запросы, итого']) - min(df['Запросы, итого']))*300\n",
    "        requests_total = df_new['Запросы 2023'].sum()\n",
    "        df['Рост/Падение, %'] = df['Запросы 2023'] / df['Запросы 2022']\n",
    "        df['Доля рынка, %'] = df['Запросы 2023'] / requests_total\n",
    "        x = (df['Доля рынка, %'])\n",
    "        y = (df['Рост/Падение, %'])\n",
    "        sizes = (df['Запросы, итого']/1000)\n",
    "        x_avg = (x.min()+x.max())/2\n",
    "        y_avg = (y.min()+y.max())/2\n",
    "        xi, yi = df['Доля рынка, %'], df['Рост/Падение, %']\n",
    "        df.loc[(xi > x_avg) & (yi > y_avg), 'group'] = 'Звёзды:'\n",
    "        df.loc[(xi > x_avg)  & (yi <= y_avg), 'group'] = 'Коровы:'\n",
    "        df.loc[(xi <= x_avg)  & (yi > y_avg), 'group'] = 'Дети:'\n",
    "        df.loc[(xi <= x_avg)  & (yi <= y_avg), 'group'] = 'Собаки:'\n",
    "        df = df.sort_values(['Доля рынка, %', 'Рост/Падение, %'], ascending=[False, False])\n",
    "\n",
    "        plt.rcParams['figure.figsize'] = size_x, size_y\n",
    "        plt.rcParams['figure.dpi'] = 2**8\n",
    "        plt.rcParams['font.size'] = font_size\n",
    "        fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "        fig, ax = plt.subplots()\n",
    "        ax.scatter(x, y, sizes, c='#C0504D', alpha=0.7)\n",
    "\n",
    "        n = 0.02\n",
    "        m = 0.02\n",
    "        ax.set_ylim(ax.get_ylim()[0] * (1 - n), ax.get_ylim()[1] * (1 + n))\n",
    "        ax.set_xlim(ax.get_xlim()[0] * (1 - m), ax.get_xlim()[1] * (1 + m))\n",
    "        ax.spines['top'].set_visible(False)\n",
    "        ax.spines['bottom'].set_visible(False)\n",
    "        ax.spines['left'].set_visible(False)\n",
    "        ax.spines['right'].set_visible(False)\n",
    "        ax.set_xticklabels('')\n",
    "        ax.set_yticklabels('')\n",
    "        plt.tick_params(bottom = False, left=False)\n",
    "        plt.title(rubric, font=fpath, fontsize=40/size_coef, fontweight=1000)\n",
    "\n",
    "        for xi, yi, city, size in zip(x, y, x.index, sizes):\n",
    "            rs = city.split(' ')\n",
    "            if len(rs) > 1:\n",
    "                rsl = len(rs) // 2\n",
    "                city = ' '.join(rs[:rsl] + ['\\n'] + rs[rsl:])\n",
    "\n",
    "            shar_path = [r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\Шар_{:02}.png\".format(shar_num)]\n",
    "            shar_img = OffsetImage(plt.imread(shar_path[0], format=\"png\"), zoom=size/size_coef)\n",
    "            ab = AnnotationBbox(shar_img, (xi, yi), frameon=False)\n",
    "            ax.add_artist(ab)\n",
    "            plt.annotate(city, (xi, yi), ha='center', va='center', font=fpath, color='#404040')\n",
    "\n",
    "        ax.hlines((y.min()+y.max())/2, x.min(), x.max(), colors='gray', linewidth=1)\n",
    "        ax.vlines((x.min()+x.max())/2, y.min(), y.max(), colors='gray', linewidth=1)\n",
    "        \n",
    "        fig.savefig(f'pictures/{folder}_shar_{shar_num}.png', bbox_inches='tight', transparent=False)\n",
    "        pics.append(f'pictures/{folder}_shar_{shar_num}.png')\n",
    "        #plt.show()\n",
    "        plt.close()\n",
    "    return pics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "id": "29ef79db-b7b6-4440-9cc4-848d79bdab2c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:52:09.494750Z",
     "iopub.status.busy": "2023-05-29T22:52:09.494750Z",
     "iopub.status.idle": "2023-05-29T22:56:15.889605Z",
     "shell.execute_reply": "2023-05-29T22:56:15.888641Z",
     "shell.execute_reply.started": "2023-05-29T22:52:09.494750Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[########################################] | 100% Completed | 2.20 sms\n",
      "[########################################] | 100% Completed | 2.31 s\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f032b7119bc14025bda386b8cffe121c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/50 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "matrix_cities_pics = plot_matrix_cities(cnt_graphics=50, size_x=size_x, size_y=size_y, font_size=font_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "id": "910de91d-def1-4d41-b9f6-765cc2b78ac8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:59:56.328782Z",
     "iopub.status.busy": "2023-05-29T22:59:56.328782Z",
     "iopub.status.idle": "2023-05-29T22:59:56.599345Z",
     "shell.execute_reply": "2023-05-29T22:59:56.598345Z",
     "shell.execute_reply.started": "2023-05-29T22:59:56.328782Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b71d63a2e0>"
      ]
     },
     "execution_count": 278,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "slides[16].shapes.add_picture('pictures/ten_city_shar_2 — копия.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "id": "70aada6c-d282-4477-aa9b-9c9367b07341",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T22:59:57.144040Z",
     "iopub.status.busy": "2023-05-29T22:59:57.144040Z",
     "iopub.status.idle": "2023-05-29T23:00:14.270758Z",
     "shell.execute_reply": "2023-05-29T23:00:14.269757Z",
     "shell.execute_reply.started": "2023-05-29T22:59:57.144040Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 02:00:14 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b1c44a94-e4dc-4522-9c04-f68bd538bd51",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T10:25:14.764077Z",
     "iopub.status.busy": "2023-06-08T10:25:14.764077Z",
     "iopub.status.idle": "2023-06-08T10:25:14.792077Z",
     "shell.execute_reply": "2023-06-08T10:25:14.791073Z",
     "shell.execute_reply.started": "2023-06-08T10:25:14.764077Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "plot_matrix_cities() got an unexpected keyword argument 'size_coef'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "Input \u001b[1;32mIn [12]\u001b[0m, in \u001b[0;36m<cell line: 1>\u001b[1;34m()\u001b[0m\n\u001b[1;32m----> 1\u001b[0m matrix_cities_pics_12 \u001b[38;5;241m=\u001b[39m \u001b[43mplot_matrix_cities\u001b[49m\u001b[43m(\u001b[49m\u001b[43mcnt_graphics\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m12\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msize_x\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43msize_x\u001b[49m\u001b[38;5;241;43m/\u001b[39;49m\u001b[38;5;241;43m4\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msize_y\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43msize_y\u001b[49m\u001b[38;5;241;43m/\u001b[39;49m\u001b[38;5;241;43m3\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mfont_size\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mfont_size\u001b[49m\u001b[38;5;241;43m/\u001b[39;49m\u001b[38;5;241;43m3\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msize_coef\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m3\u001b[39;49m\u001b[43m)\u001b[49m\n",
      "\u001b[1;31mTypeError\u001b[0m: plot_matrix_cities() got an unexpected keyword argument 'size_coef'"
     ]
    }
   ],
   "source": [
    "matrix_cities_pics_12 = plot_matrix_cities(cnt_graphics=12, size_x=size_x/4, size_y=size_y/3, font_size=font_size/3, size_coef=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2d03201b-1195-4255-853e-8b68e18bca1a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "id": "2a4cf768-c075-43cf-ace5-2c755c686590",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T23:07:47.580447Z",
     "iopub.status.busy": "2023-05-29T23:07:47.580447Z",
     "iopub.status.idle": "2023-05-29T23:07:47.813468Z",
     "shell.execute_reply": "2023-05-29T23:07:47.812462Z",
     "shell.execute_reply.started": "2023-05-29T23:07:47.580447Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b68bd5aa90>"
      ]
     },
     "execution_count": 289,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x = 7.7\n",
    "y = 5.3\n",
    "slides[17].shapes.add_picture(matrix_pics_12[0], Cm(0.7),     Cm(2.4), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[1], Cm(0.7+x),   Cm(2.4), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[2], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[3], Cm(0.7+x*3), Cm(2.4), Cm(x), Cm(y))\n",
    "\n",
    "slides[17].shapes.add_picture(matrix_pics_12[4], Cm(0.7),     Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[5], Cm(0.7+x),   Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[6], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[7], Cm(0.7+x*3), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[17].shapes.add_picture(matrix_pics_12[8],  Cm(0.7),     Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[9],  Cm(0.7+x),   Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[10], Cm(0.7+x*2), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[17].shapes.add_picture(matrix_pics_12[11], Cm(0.7+x*3), Cm(2.4+y*2), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "id": "da4baa91-e7e7-4f5e-b8d0-15c6f539c71e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T23:07:48.613896Z",
     "iopub.status.busy": "2023-05-29T23:07:48.612895Z",
     "iopub.status.idle": "2023-05-29T23:07:50.092796Z",
     "shell.execute_reply": "2023-05-29T23:07:50.091798Z",
     "shell.execute_reply.started": "2023-05-29T23:07:48.613896Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 02:07:50 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "087b716d-e7c8-4c2a-904d-0cfda8f11622",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b5403c3d-8e9d-4323-855a-a3640ca03a81",
   "metadata": {
    "tags": []
   },
   "source": [
    "### 1.11.\tАнализ Матриц BCG по городам и бизнесам"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "e8d2f687-ca58-45d2-9fb7-9c3459fabd1f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T11:00:47.012612Z",
     "iopub.status.busy": "2023-06-08T11:00:47.011611Z",
     "iopub.status.idle": "2023-06-08T11:00:47.043548Z",
     "shell.execute_reply": "2023-06-08T11:00:47.041226Z",
     "shell.execute_reply.started": "2023-06-08T11:00:47.012612Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_matrix_rubric_cities(cnt_graphics=100, show=False, size_x=size_x, size_y=size_y, font_size=font_size, size_coef=1):\n",
    "    folder='ten_city'\n",
    "    shar_nums = list(range(1,16))*500\n",
    "    time = str(datetime.datetime.now())[2:16].replace('-','').replace(' ','_').replace(':','')\n",
    "    df_src = pd.read_parquet(f'files/ten_city/requests_for_shars')\n",
    "    df_new = df_src[['rubricName','city']].drop_duplicates().set_index(['rubricName','city'])\n",
    "    df_new['Запросы 2020'] = df_src[df_src['year_r'] == 2020].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2021'] = df_src[df_src['year_r'] == 2021].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2022'] = df_src[df_src['year_r'] == 2022].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new['Запросы 2023'] = df_src[df_src['year_r'] == 2023].groupby(['rubricName', 'city']).sum('cnt_requests')[['cnt_requests']]\n",
    "    df_new = df_new.fillna(0).astype(int)\n",
    "    df_new['Запросы, итого'] =  df_new['Запросы 2020'] + df_new['Запросы 2021'] + df_new['Запросы 2022'] + df_new['Запросы 2023']\n",
    "    df_new = df_new.reset_index()\n",
    "    df_new['city_rubric'] = df_new['city'] + '.' + df_new['rubricName']\n",
    "    df_new = df_new.set_index('city_rubric').sort_values('Запросы, итого', ascending=False)\n",
    "    cnt_shars = 12\n",
    "    step = 1\n",
    "    rubrics = dd.read_parquet(f\"files/ten_city/df_tx_grouped_filtered_ten_city\")['rubricName'].unique().compute()\n",
    "    rubrics = [i for i in rubrics.to_dict().values()]\n",
    "    df_list = []\n",
    "    start_range = 1\n",
    "    stop_range = cnt_shars*step+1\n",
    "    pics=[]\n",
    "    for graphic_num in tqdm(range(cnt_graphics)):\n",
    "        shar_num = shar_nums.pop()\n",
    "        \n",
    "        df = df_new.iloc[range(start_range, stop_range, step)]\n",
    "        start_range = stop_range\n",
    "        stop_range = (graphic_num+2)*cnt_shars*step+1\n",
    "        \n",
    "        df['Запросы, итого'] = (df['Запросы, итого'] - min(df['Запросы, итого'])) / (max(df['Запросы, итого']) - min(df['Запросы, итого']))*300\n",
    "        requests_total = df_new['Запросы 2023'].sum()\n",
    "        df['Рост/Падение, %'] = df['Запросы 2023'] / df['Запросы 2022']\n",
    "        df['Доля рынка, %'] = df['Запросы 2023'] / requests_total\n",
    "        x = (df['Доля рынка, %'])\n",
    "        y = (df['Рост/Падение, %'])\n",
    "        sizes = (df['Запросы, итого']/1000)\n",
    "        x_avg = (x.min()+x.max())/2\n",
    "        y_avg = (y.min()+y.max())/2\n",
    "        xi, yi = df['Доля рынка, %'], df['Рост/Падение, %']\n",
    "        df.loc[(xi > x_avg) & (yi > y_avg), 'group'] = 'Звёзды:'\n",
    "        df.loc[(xi > x_avg)  & (yi <= y_avg), 'group'] = 'Коровы:'\n",
    "        df.loc[(xi <= x_avg)  & (yi > y_avg), 'group'] = 'Дети:'\n",
    "        df.loc[(xi <= x_avg)  & (yi <= y_avg), 'group'] = 'Собаки:'\n",
    "        df = df.sort_values(['Доля рынка, %', 'Рост/Падение, %'], ascending=[False, False])\n",
    "\n",
    "        plt.rcParams['figure.figsize'] = size_x, size_y\n",
    "        plt.rcParams['figure.dpi'] = 2**8\n",
    "        plt.rcParams['font.size'] = font_size\n",
    "        fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "        fig, ax = plt.subplots()\n",
    "        ax.scatter(x, y, sizes, c='#C0504D', alpha=0.7)\n",
    "\n",
    "        n = 0.02\n",
    "        m = 0.02\n",
    "        ax.set_ylim(ax.get_ylim()[0] * (1 - n), ax.get_ylim()[1] * (1 + n))\n",
    "        ax.set_xlim(ax.get_xlim()[0] * (1 - m), ax.get_xlim()[1] * (1 + m))\n",
    "        ax.spines['top'].set_visible(False)\n",
    "        ax.spines['bottom'].set_visible(False)\n",
    "        ax.spines['left'].set_visible(False)\n",
    "        ax.spines['right'].set_visible(False)\n",
    "        ax.set_xticklabels('')\n",
    "        ax.set_yticklabels('')\n",
    "        plt.tick_params(bottom = False, left=False)\n",
    "        #plt.title(rubric, font=fpath, fontsize=40, fontweight=1000)\n",
    "\n",
    "        for xi, yi, city, size in zip(x, y, x.index, sizes):\n",
    "            rs = city.split(' ')\n",
    "            if len(rs) > 1:\n",
    "                rsl = len(rs) // 2\n",
    "                city = ' '.join(rs[:rsl] + ['\\n'] + rs[rsl:])\n",
    "\n",
    "            shar_path = [r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\Шар_{:02}.png\".format(shar_num)]\n",
    "            shar_img = OffsetImage(plt.imread(shar_path[0], format=\"png\"), zoom=size/size_coef)\n",
    "            ab = AnnotationBbox(shar_img, (xi, yi), frameon=False)\n",
    "            ax.add_artist(ab)\n",
    "            plt.annotate(city, (xi, yi), ha='center', va='center', font=fpath, color='#404040')\n",
    "\n",
    "        ax.hlines((y.min()+y.max())/2, x.min(), x.max(), colors='gray', linewidth=1)\n",
    "        ax.vlines((x.min()+x.max())/2, y.min(), y.max(), colors='gray', linewidth=1)\n",
    "        \n",
    "        fig.savefig(f'pictures/{folder}_shar_{shar_num}.png', bbox_inches='tight', transparent=False, facecolor='white')\n",
    "        pics.append(f'pictures/{folder}_shar_{shar_num}.png')\n",
    "        #plt.show()\n",
    "        plt.close()\n",
    "    return pics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96eeae32-bae8-4f7c-a7b7-0183f49e78dc",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T23:21:04.649916Z",
     "iopub.status.busy": "2023-05-29T23:21:04.649916Z",
     "iopub.status.idle": "2023-05-29T23:21:17.504245Z",
     "shell.execute_reply": "2023-05-29T23:21:17.503331Z",
     "shell.execute_reply.started": "2023-05-29T23:21:04.649916Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "cnt_matrix_rubric_cities = plot_matrix_rubric_cities(cnt_graphics=1, size_x=size_x, size_y=size_y, font_size=font_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "id": "e63b7604-79de-4eb9-9f82-7b5c0ca4ee10",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T23:29:41.863414Z",
     "iopub.status.busy": "2023-05-29T23:29:41.863414Z",
     "iopub.status.idle": "2023-05-29T23:29:42.021944Z",
     "shell.execute_reply": "2023-05-29T23:29:42.020960Z",
     "shell.execute_reply.started": "2023-05-29T23:29:41.863414Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x2b6e655ff10>"
      ]
     },
     "execution_count": 308,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "slides[18].shapes.add_picture('pictures/ten_city_shar_15.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "id": "e5ec6121-631e-449e-8e41-1b53992fec92",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-29T23:29:44.986215Z",
     "iopub.status.busy": "2023-05-29T23:29:44.986215Z",
     "iopub.status.idle": "2023-05-29T23:29:46.409226Z",
     "shell.execute_reply": "2023-05-29T23:29:46.408228Z",
     "shell.execute_reply.started": "2023-05-29T23:29:44.986215Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-05-30 02:29:46 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "8ca8dcdc-cd61-4a96-8f18-a0995817452e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T11:00:53.919044Z",
     "iopub.status.busy": "2023-06-08T11:00:53.919044Z",
     "iopub.status.idle": "2023-06-08T11:01:52.935302Z",
     "shell.execute_reply": "2023-06-08T11:01:52.933937Z",
     "shell.execute_reply.started": "2023-06-08T11:00:53.919044Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[########################################] | 100% Completed | 5.56 sms\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "376aa1184eef484a8a82129f52ef99e2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/30 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "matrix_cities_pics_12 = plot_matrix_rubric_cities(cnt_graphics=30, size_x=size_x/4, size_y=size_y/3, font_size=font_size/3, size_coef=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "8613ceac-fe4c-4da9-837d-a35d66977e8f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T11:01:52.941432Z",
     "iopub.status.busy": "2023-06-08T11:01:52.940347Z",
     "iopub.status.idle": "2023-06-08T11:01:54.644431Z",
     "shell.execute_reply": "2023-06-08T11:01:54.643060Z",
     "shell.execute_reply.started": "2023-06-08T11:01:52.941432Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<pptx.shapes.picture.Picture at 0x24c12bbee50>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x = 7.7\n",
    "y = 5.3\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[0], Cm(0.7),     Cm(2.4), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[1], Cm(0.7+x),   Cm(2.4), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[2], Cm(0.7+x*2), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[3], Cm(0.7+x*3), Cm(2.4), Cm(x), Cm(y))\n",
    "\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[4], Cm(0.7),     Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[5], Cm(0.7+x),   Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[6], Cm(0.7+x*2), Cm(2.4+y), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[7], Cm(0.7+x*3), Cm(2.4+y), Cm(x), Cm(y))\n",
    "\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[8],  Cm(0.7),     Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[9],  Cm(0.7+x),   Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[10], Cm(0.7+x*2), Cm(2.4+y*2), Cm(x), Cm(y))\n",
    "slides[19].shapes.add_picture(matrix_cities_pics_12[11], Cm(0.7+x*3), Cm(2.4+y*2), Cm(x), Cm(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "5cb5e7da-ba5a-489a-bd57-fdaf4dcf18e5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T11:01:54.648447Z",
     "iopub.status.busy": "2023-06-08T11:01:54.647447Z",
     "iopub.status.idle": "2023-06-08T11:02:16.712685Z",
     "shell.execute_reply": "2023-06-08T11:02:16.711391Z",
     "shell.execute_reply.started": "2023-06-08T11:01:54.648447Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-06-08 14:02:16 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "save_pptx()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5a468066-ab57-436d-a090-456204e39fe8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "6c350ef1-0060-4638-9b47-5f67e33c22e3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-08T10:49:11.741075Z",
     "iopub.status.busy": "2023-06-08T10:49:11.741075Z",
     "iopub.status.idle": "2023-06-08T10:49:11.748077Z",
     "shell.execute_reply": "2023-06-08T10:49:11.747077Z",
     "shell.execute_reply.started": "2023-06-08T10:49:11.741075Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "size_x = 30.7 / 2.54 * 4\n",
    "size_y = 16   / 2.54 * 4\n",
    "fontsize = 38"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "275fb5cf-989b-4781-9c70-f918894661c0",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Москва матрицы"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9340ccb0-92c2-4ae3-a1db-3457f8ed8c05",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Бизнес"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9024f78c-a456-48fc-93ef-fc6e8f7eb623",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-31T22:00:37.556851Z",
     "iopub.status.busy": "2023-05-31T22:00:37.555846Z",
     "iopub.status.idle": "2023-05-31T22:00:39.357191Z",
     "shell.execute_reply": "2023-05-31T22:00:39.357191Z",
     "shell.execute_reply.started": "2023-05-31T22:00:37.556851Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "city = 'Москва'\n",
    "df1 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df1 = df1[df1['city'] == city]\n",
    "df1['city_rubric_param'] = df1['city'] + '. ' + df1['rubric'] + '. ' + 'Пользователи'\n",
    "df1 = df1[['date_r', 'cnt_users', 'city_rubric_param']].rename({'cnt_users':'cnt'}, axis=1)\n",
    "\n",
    "df2 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df2 = df2[df2['city'] == city]\n",
    "df2['city_rubric_param'] = df2['city'] + '. ' + df2['rubric'] + '. ' + 'Запросы'\n",
    "df2 = df2[['date_r', 'cnt_requests', 'city_rubric_param']].rename({'cnt_requests':'cnt'}, axis=1)\n",
    "df_dates_users = pd.concat([df1, df2])\n",
    "df_dates_users['cnt_all'] = df_dates_users.groupby('city_rubric_param')['cnt'].transform(np.sum)\n",
    "df_dates_users['city_rubric_param_cnt'] = df_dates_users['cnt_all'].astype(str) + '_' + df_dates_users['city_rubric_param']\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_users, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param_cnt'])\n",
    "t = dfx.corr().fillna(0).astype(np.float16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "53ca35ed-290b-4179-99be-9a44d74d10d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-31T22:00:40.724034Z",
     "iopub.status.busy": "2023-05-31T22:00:40.724034Z",
     "iopub.status.idle": "2023-05-31T22:00:45.166570Z",
     "shell.execute_reply": "2023-05-31T22:00:45.166570Z",
     "shell.execute_reply.started": "2023-05-31T22:00:40.724034Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_requests'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Запросы'\n",
    "df_requests = df_all.copy()\n",
    "\n",
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_users'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Пользователи'\n",
    "df_users = df_all.copy()\n",
    "\n",
    "df_100 = pd.concat([df_requests, df_users]).iloc[:30]\n",
    "\n",
    "df1 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df1 = df1[df1['city'] == city]\n",
    "df1['city_rubric_param'] = df1['city'] + '. ' + df1['rubric'] + '. ' + 'Пользователи'\n",
    "df1 = df1[['date_r', 'cnt_users', 'city_rubric_param']].rename({'cnt_users':'cnt'}, axis=1)\n",
    "\n",
    "df2 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "df2 = df2[df2['city'] == city]\n",
    "df2['city_rubric_param'] = df2['city'] + '. ' + df2['rubric'] + '. ' + 'Запросы'\n",
    "df2 = df2[['date_r', 'cnt_requests', 'city_rubric_param']].rename({'cnt_requests':'cnt'}, axis=1)\n",
    "df_dates_users = pd.concat([df1, df2])\n",
    "\n",
    "df_dates_users['cnt_all'] = df_dates_users.groupby('city_rubric_param')['cnt'].transform(np.sum)\n",
    "df_dates_users['city_rubric_param_cnt'] = df_dates_users['cnt_all'].astype(str) + '_' + df_dates_users['city_rubric_param']\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_users, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param_cnt'])\n",
    "t_h = dfx.corr().fillna(0).astype(np.float16)\n",
    "\n",
    "t_h = t_h.rename_axis([None]).rename_axis([None], axis=1)\n",
    "t1 = t_h.stack()\n",
    "t1 = t1.reset_index()\n",
    "t1.columns = ['r1', 'r2', 'corr']\n",
    "t1 = t1[t1.r1 != t1.r2]\n",
    "\n",
    "t1_100 = df_100.merge(t1, left_on='cnt_city_rubric', right_on='r1')\n",
    "\n",
    "t1_100['r1_cnt'] = t1_100.r1.apply(lambda x: x.split('_')[0])\n",
    "t1_100['r2_cnt'] = t1_100.r2.apply(lambda x: x.split('_')[0])\n",
    "\n",
    "t1_100.r1 = t1_100.r1.apply(lambda x: x.split('_')[1])\n",
    "t1_100.r2 = t1_100.r2.apply(lambda x: x.split('_')[1])\n",
    "t1_100['r1_city'] = t1_100.r1.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r2_city'] = t1_100.r2.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r1_rubric'] = t1_100.r1.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r2_rubric'] = t1_100.r2.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r1_param'] = t1_100.r1.apply(lambda x: x.split('. ')[2])\n",
    "t1_100['r2_param'] = t1_100.r2.apply(lambda x: x.split('. ')[2])\n",
    "\n",
    "t1_100.r1_cnt = t1_100.r1_cnt.astype(int)\n",
    "t1_100.r2_cnt = t1_100.r2_cnt.astype(int)\n",
    "\n",
    "t1_100 = t1_100[t1_100.r2_cnt > 1000]\n",
    "\n",
    "t1_100 = t1_100[['r1', 'r2', 'r1_city', 'r1_rubric', 'r1_param', 'r2_city', 'r2_rubric', 'r2_param', 'corr', 'r1_cnt', 'r2_cnt']]\n",
    "\n",
    "df_100.shape, t1_100.shape\n",
    "\n",
    "t2 = t1_100[(t1_100.r1_cnt > 10000) & (t1_100.r2_cnt > 10000)]\n",
    "t2['cnt_diff'] = abs(t2.r1_cnt - t2.r2_cnt)\n",
    "t1_m = t2[t2['corr'] < 0].sort_values(['cnt_diff', 'r1_cnt'], ascending=[True, False])\n",
    "fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e43527c9-48b6-4ad0-a4ec-ecd2dd535380",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:26:22.787369Z",
     "iopub.status.busy": "2023-06-01T04:26:22.786373Z",
     "iopub.status.idle": "2023-06-01T04:26:36.344940Z",
     "shell.execute_reply": "2023-06-01T04:26:36.344940Z",
     "shell.execute_reply.started": "2023-06-01T04:26:22.786373Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "negative_corr = t1_m.iloc[:1000]\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "negative_corr_rubrics = [i.split('. ')[1] for i in negative_corr_rubrics]\n",
    "#negative_corr_rubrics = [i.replace('Москва. ', '') for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_users[df_dates_users['city_rubric_param'].apply(lambda x: ((x.split('. ')[1] in negative_corr_rubrics) and (len(x) < 35)))]\n",
    "df_dates_negative['rubric'] = df_dates_negative['city_rubric_param'].apply(lambda x: x.split('.')[1])\n",
    "dfx = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['rubric']).iloc[:,:80]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=False, cbar=False)\n",
    "plt.xticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.yticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cd2e4571-0f34-4066-8dc6-4b976caa8198",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:19:04.480877Z",
     "iopub.status.busy": "2023-06-01T04:19:04.480877Z",
     "iopub.status.idle": "2023-06-01T04:19:32.927754Z",
     "shell.execute_reply": "2023-06-01T04:19:32.927754Z",
     "shell.execute_reply.started": "2023-06-01T04:19:04.480877Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_dates_all = df_dates_users\n",
    "df_dates_all['city_rubric'] = df_dates_all['city_rubric_param']\n",
    "df_dates_all['rubric'] = df_dates_all['city_rubric_param'].apply(lambda x: x.split('. ')[1])\n",
    "\n",
    "positive_corr = t2[t2['corr'] >= 0.99]\n",
    "positive_corr_rubrics = pd.concat([positive_corr.r1, positive_corr.r2]).drop_duplicates().values\n",
    "positive_corr_rubrics = [i.split('. ')[1] for i in positive_corr_rubrics]\n",
    "\n",
    "df_dates_positive = df_dates_all[df_dates_all['rubric'].apply(lambda x: ((x in positive_corr_rubrics[:500]) and (len(x) < 25)))]\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_positive, values='cnt', aggfunc='sum', index=['date_r'], columns=['rubric']).iloc[:,:30]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_1.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3c211ba2-df29-45de-b653-0a2daef5486c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:19:58.802367Z",
     "iopub.status.busy": "2023-06-01T04:19:58.802367Z",
     "iopub.status.idle": "2023-06-01T04:20:17.916926Z",
     "shell.execute_reply": "2023-06-01T04:20:17.916018Z",
     "shell.execute_reply.started": "2023-06-01T04:19:58.802367Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "negative_corr = t1_m.iloc[:500]\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "negative_corr_rubrics = [i.split('. ')[1] for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_all[df_dates_all['rubric'].apply(lambda x: (x in negative_corr_rubrics[:500]) and (len(x) < 25))]\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['rubric']).iloc[:,:30]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_2.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "88a0520f-6fa6-4472-a0c0-187e8e880815",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:20:17.918924Z",
     "iopub.status.busy": "2023-06-01T04:20:17.918924Z",
     "iopub.status.idle": "2023-06-01T04:22:14.928423Z",
     "shell.execute_reply": "2023-06-01T04:22:14.928423Z",
     "shell.execute_reply.started": "2023-06-01T04:20:17.918924Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "zero_corr = t1[t1['corr'] < abs(0.41)]\n",
    "zero_corr_rubrics = pd.concat([zero_corr.r1, zero_corr.r2]).drop_duplicates().values\n",
    "zero_corr_rubrics = [i.split('. ')[1] for i in zero_corr_rubrics]\n",
    "\n",
    "df_dates_zero = df_dates_all[df_dates_all['rubric'].apply(lambda x: (x in zero_corr_rubrics[:500]) and (len(x) < 25) )]\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_zero, values='cnt', aggfunc='sum', index=['date_r'], columns=['rubric']).iloc[:,:30]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_3.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "f05fae31-d772-42e3-97f1-e3d9c365ad11",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:27:12.721632Z",
     "iopub.status.busy": "2023-06-01T04:27:12.721632Z",
     "iopub.status.idle": "2023-06-01T04:27:17.792113Z",
     "shell.execute_reply": "2023-06-01T04:27:17.791113Z",
     "shell.execute_reply.started": "2023-06-01T04:27:12.721632Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-06-01 07:27:17 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "slides[35].shapes.add_picture('pics/corr_matrix.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[36].shapes.add_picture('pics/corr_matrix_1.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[37].shapes.add_picture('pics/corr_matrix_2.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[38].shapes.add_picture('pics/corr_matrix_3.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "save_pptx()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7559abb-da72-4a8f-8421-33bb44084571",
   "metadata": {},
   "source": [
    "## Бизнес - город"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "abb7ba5f-47c3-450a-a31b-23940b9c3a37",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T03:49:12.313021Z",
     "iopub.status.busy": "2023-06-01T03:49:12.313021Z",
     "iopub.status.idle": "2023-06-01T03:49:13.724747Z",
     "shell.execute_reply": "2023-06-01T03:49:13.724747Z",
     "shell.execute_reply.started": "2023-06-01T03:49:12.313021Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_all = pd.concat([pd.read_parquet(f'files/ten_city/df_dates_{col_r}').rename({col_r: 'date_r'}, axis=1)[['rubricName', 'date_r', 'cnt', 'city', 'year_r']] for col_r in ['hour_r', 'date_r', 'week_r', 'week_day_r']])\n",
    "rs = \"('\" + \"','\".join(list(df_all.rubricName.dropna().unique())) + \"')\"\n",
    "\n",
    "df_schema = StructType([StructField('user', StringType(), True), StructField('proj', StringType(), True), StructField('rubricName', StringType(), True), StructField('rubricID', IntegerType(), True), StructField('time', StringType(), True), StructField('date', StringType(), True), StructField('org', StringType(), True), StructField('branch', StringType(), True), StructField('prod', StringType(), True), StructField('tx', StringType(), True), StructField('fl', StringType(), True), StructField('devmod', StringType(), True), StructField('lat', StringType(), True), StructField('lon', StringType(), True)])\n",
    "df_list = [i for i in files_list_chain]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7f48b679-bff8-4ede-8476-74606406a66a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-25T03:54:07.328136Z",
     "iopub.status.busy": "2023-05-25T03:54:07.328136Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "912ff55004a9413b96cde913c08472b6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for file in tqdm(df_list[:4]):\n",
    "    df = spark.read.option(\"header\",\"true\").csv(file, schema=df_schema)\n",
    "    df.createOrReplaceTempView('df')\n",
    "    city = file.split('\\\\')[2]\n",
    "    df_new = spark.sql(f'''\n",
    "    select \n",
    "        rubricName as rubric\n",
    "        , \"{city}\" city\n",
    "        , user\n",
    "        , SUBSTR(time, 6, 5) date_r\n",
    "        from df\n",
    "    ''').where(f'rubric in {rs}')\n",
    "\n",
    "    df_new.write.partitionBy('rubric').mode('append').parquet(f'files/ten_city/df_dates_users_city2')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "273c6555-8c37-4d98-8ac3-10697ffacdc4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-05-25T18:10:52.485602Z",
     "iopub.status.busy": "2023-05-25T18:10:52.485602Z",
     "iopub.status.idle": "2023-05-25T20:34:51.377901Z",
     "shell.execute_reply": "2023-05-25T20:34:51.372842Z",
     "shell.execute_reply.started": "2023-05-25T18:10:52.485602Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8c91dd3886ed4c8785fdc73e196b1e2d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/731 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_all = pd.concat([pd.read_parquet(f'files/ten_city/df_dates_{col_r}').rename({col_r: 'date_r'}, axis=1)[['rubricName', 'date_r', 'cnt', 'city', 'year_r']] for col_r in ['hour_r', 'date_r', 'week_r', 'week_day_r']])\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubricName']\n",
    "df_all = df_all[~df_all.rubricName.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "rubrics = df_all.rubricName.unique()\n",
    "\n",
    "for rubric in tqdm(rubrics[910:]):\n",
    "    try:\n",
    "        df = spark.read.parquet(fr\"C:\\Users\\Mamed\\_Notebooks\\2GIS_2023\\files\\ten_city\\df_dates_users_city2\\rubric={rubric.replace(' ', '%20').replace('/','%2F')}\")\n",
    "        df.createOrReplaceTempView('df')\n",
    "        df_new = spark.sql(f'''\n",
    "        select \n",
    "            \"{rubric}\" rubric\n",
    "            ,city\n",
    "            , date_r\n",
    "            ,count(distinct user) cnt_users\n",
    "            ,count(*) cnt_requests\n",
    "        from df\n",
    "            group by \n",
    "                    \"{rubric}\"\n",
    "                    ,city\n",
    "                    ,date_r\n",
    "\n",
    "        ''')\n",
    "        df_new.write.mode('append').parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "    except:\n",
    "        pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "442d06ac-8194-4458-b3eb-c832f84ff669",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T03:49:31.497717Z",
     "iopub.status.busy": "2023-06-01T03:49:31.497717Z",
     "iopub.status.idle": "2023-06-01T03:49:37.215024Z",
     "shell.execute_reply": "2023-06-01T03:49:37.215024Z",
     "shell.execute_reply.started": "2023-06-01T03:49:31.497717Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "city = 'Москва'\n",
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_requests'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Запросы'\n",
    "df_requests = df_all.copy()\n",
    "\n",
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_users'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Пользователи'\n",
    "df_users = df_all.copy()\n",
    "\n",
    "df_100 = pd.concat([df_requests, df_users]).iloc[:30]\n",
    "\n",
    "df1 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df1 = df1[df1['city'] == city]\n",
    "\n",
    "df1['city_rubric_param'] = df1['city'] + '. ' + df1['rubric'] + '. ' + 'Пользователи'\n",
    "df1 = df1[['date_r', 'cnt_users', 'city_rubric_param']].rename({'cnt_users':'cnt'}, axis=1)\n",
    "\n",
    "df2 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df2 = df2[df2['city'] == city]\n",
    "\n",
    "df2['city_rubric_param'] = df2['city'] + '. ' + df2['rubric'] + '. ' + 'Запросы'\n",
    "df2 = df2[['date_r', 'cnt_requests', 'city_rubric_param']].rename({'cnt_requests':'cnt'}, axis=1)\n",
    "df_dates_users = pd.concat([df1, df2])\n",
    "\n",
    "df_dates_users['cnt_all'] = df_dates_users.groupby('city_rubric_param')['cnt'].transform(np.sum)\n",
    "df_dates_users['city_rubric_param_cnt'] = df_dates_users['cnt_all'].astype(str) + '_' + df_dates_users['city_rubric_param']\n",
    "df_dates_users['city_rubric'] = df_dates_users['city_rubric_param'].apply(lambda x: x.split('. ')[0] + '. ' + x.split('. ')[1])\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_users, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param_cnt'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2a79e9ad-1b5e-4479-a005-f22e84d093ae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T03:49:37.215024Z",
     "iopub.status.busy": "2023-06-01T03:49:37.215024Z",
     "iopub.status.idle": "2023-06-01T03:53:24.584817Z",
     "shell.execute_reply": "2023-06-01T03:53:24.583820Z",
     "shell.execute_reply.started": "2023-06-01T03:49:37.215024Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 2min 55s\n",
      "Wall time: 3min 47s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "t_h = dfx.corr().fillna(0).astype(np.float16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "96d09336-b04c-4d27-9aa6-1f9f4deb3cae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T03:59:13.097695Z",
     "iopub.status.busy": "2023-06-01T03:59:13.097695Z",
     "iopub.status.idle": "2023-06-01T04:01:53.506746Z",
     "shell.execute_reply": "2023-06-01T04:01:53.505713Z",
     "shell.execute_reply.started": "2023-06-01T03:59:13.097695Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 2min 17s\n",
      "Wall time: 2min 40s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "t_h = t_h.rename_axis([None]).rename_axis([None], axis=1)\n",
    "t1 = t_h.stack()\n",
    "t1 = t1.reset_index()\n",
    "t1.columns = ['r1', 'r2', 'corr']\n",
    "t1 = t1[t1.r1 != t1.r2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "bfe57465-27de-4475-a253-77e4c6eb4a6f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:01:53.507714Z",
     "iopub.status.busy": "2023-06-01T04:01:53.507714Z",
     "iopub.status.idle": "2023-06-01T04:02:37.864695Z",
     "shell.execute_reply": "2023-06-01T04:02:37.863778Z",
     "shell.execute_reply.started": "2023-06-01T04:01:53.507714Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 37.5 s\n",
      "Wall time: 44.3 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "t1_100 = df_100.merge(t1, left_on='cnt_city_rubric', right_on='r1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "560a7cca-91ce-4844-9f11-d16a4fd7e25d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:02:37.866786Z",
     "iopub.status.busy": "2023-06-01T04:02:37.865697Z",
     "iopub.status.idle": "2023-06-01T04:02:42.330925Z",
     "shell.execute_reply": "2023-06-01T04:02:42.330006Z",
     "shell.execute_reply.started": "2023-06-01T04:02:37.866786Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((30, 5), (379020, 11))"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t1_100['r1_cnt'] = t1_100.r1.apply(lambda x: x.split('_')[0])\n",
    "t1_100['r2_cnt'] = t1_100.r2.apply(lambda x: x.split('_')[0])\n",
    "\n",
    "t1_100.r1 = t1_100.r1.apply(lambda x: x.split('_')[1])\n",
    "t1_100.r2 = t1_100.r2.apply(lambda x: x.split('_')[1])\n",
    "t1_100['r1_city'] = t1_100.r1.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r2_city'] = t1_100.r2.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r1_rubric'] = t1_100.r1.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r2_rubric'] = t1_100.r2.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r1_param'] = t1_100.r1.apply(lambda x: x.split('. ')[2])\n",
    "t1_100['r2_param'] = t1_100.r2.apply(lambda x: x.split('. ')[2])\n",
    "\n",
    "t1_100.r1_cnt = t1_100.r1_cnt.astype(int)\n",
    "t1_100.r2_cnt = t1_100.r2_cnt.astype(int)\n",
    "\n",
    "t1_100 = t1_100[t1_100.r2_cnt > 1000]\n",
    "\n",
    "t1_100 = t1_100[['r1', 'r2', 'r1_city', 'r1_rubric', 'r1_param', 'r2_city', 'r2_rubric', 'r2_param', 'corr', 'r1_cnt', 'r2_cnt']]\n",
    "\n",
    "df_100.shape, t1_100.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "f7da5ac9-5f7f-4793-9fd6-dcb591e610ed",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:02:42.332931Z",
     "iopub.status.busy": "2023-06-01T04:02:42.332931Z",
     "iopub.status.idle": "2023-06-01T04:02:43.074856Z",
     "shell.execute_reply": "2023-06-01T04:02:43.074856Z",
     "shell.execute_reply.started": "2023-06-01T04:02:42.332931Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_dates_users['city'] = df_dates_users.city_rubric.apply(lambda x: x.split('. ')[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d08c32dc-0a17-4723-9443-d601cf685b56",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:32:55.334091Z",
     "iopub.status.busy": "2023-06-01T04:32:55.334091Z",
     "iopub.status.idle": "2023-06-01T04:33:23.642525Z",
     "shell.execute_reply": "2023-06-01T04:33:23.642525Z",
     "shell.execute_reply.started": "2023-06-01T04:32:55.334091Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "t2 = t1_100[(t1_100.r1_cnt > 10000) & (t1_100.r2_cnt > 10000)]\n",
    "t2['cnt_diff'] = abs(t2.r1_cnt - t2.r2_cnt)\n",
    "t1_m = t2[t2['corr'] < 0].sort_values(['cnt_diff', 'r1_cnt'], ascending=[True, False])\n",
    "fpath = Path(mpl.get_data_path(), r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\arialnarrow.ttf\")\n",
    "\n",
    "negative_corr = t1_m.iloc[:1000]\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "#negative_corr_rubrics = [i.replace('Москва. ', '') for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_users[df_dates_users['city_rubric_param'].apply(lambda x: ((x in negative_corr_rubrics) and (len(x.replace('. Запросы','')) < 35)))]\n",
    "df_dates_negative['city_rubric_param'] = df_dates_negative['city_rubric_param']\n",
    "dfx_pred = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Алматы' in i]].iloc[:,:15]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Астана' in i]].iloc[:,:15]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Москва' in i]].iloc[:,:15]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Калининград' in i]].iloc[:,:15]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'СПБ' in i]].iloc[:,:15]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Ташкент' in i]].iloc[:,:15]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Новосибирск' in i]].iloc[:,:15]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'ОАЭ' in i]].iloc[:,:15]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Бишкек' in i]].iloc[:,:15]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if 'Запросы' in i and 'Владивосток' in i]].iloc[:,:15]\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "dfx.columns = [i.replace('. Запросы','') for i in dfx.columns]\n",
    "\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=False, cbar=False)\n",
    "plt.xticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.yticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_other.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d58aabaf-04d6-456c-a917-17e928bd776b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:33:23.642525Z",
     "iopub.status.busy": "2023-06-01T04:33:23.642525Z",
     "iopub.status.idle": "2023-06-01T04:34:03.470037Z",
     "shell.execute_reply": "2023-06-01T04:34:03.470037Z",
     "shell.execute_reply.started": "2023-06-01T04:33:23.642525Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_dates_all = df_dates_users\n",
    "df_dates_all['city_rubric'] = df_dates_all['city_rubric_param']\n",
    "\n",
    "positive_corr = t2[t2['corr'] >= 0.95]\n",
    "positive_corr_rubrics = pd.concat([positive_corr.r1, positive_corr.r2]).drop_duplicates().values\n",
    "#positive_corr_rubrics = [i.split('_')[1] for i in positive_corr_rubrics]\n",
    "\n",
    "df_dates_positive = df_dates_all[df_dates_all['city_rubric'].apply(lambda x: ((x in positive_corr_rubrics[:500]) and (len(x.replace('. Запросы','')) < 35)))]\n",
    "\n",
    "dfx_pred = pd.pivot_table(df_dates_positive, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if 'Запросы' in i and 'Владивосток' in i]].iloc[:,:3]\n",
    "\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "dfx.columns = [i.replace('. Запросы','') for i in dfx.columns]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_1_other.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2006876a-cb61-4912-985b-bc3ffa03aaaa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:34:03.470037Z",
     "iopub.status.busy": "2023-06-01T04:34:03.470037Z",
     "iopub.status.idle": "2023-06-01T04:34:31.340321Z",
     "shell.execute_reply": "2023-06-01T04:34:31.340321Z",
     "shell.execute_reply.started": "2023-06-01T04:34:03.470037Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "negative_corr = t1_m.iloc[:500]\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "#negative_corr_rubrics = [i.split('_')[1] for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_all[df_dates_all['city_rubric'].apply(lambda x: (x in negative_corr_rubrics[:500]) and (len(x.replace('. Запросы','')) < 35))]\n",
    "\n",
    "dfx_pred = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if 'Запросы' in i and 'Владивосток' in i]].iloc[:,:3]\n",
    "\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "dfx.columns = [i.replace('. Запросы','') for i in dfx.columns]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_2_other.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5a107c14-6ae1-4104-8508-b4b24da7329f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:34:31.340321Z",
     "iopub.status.busy": "2023-06-01T04:34:31.340321Z",
     "iopub.status.idle": "2023-06-01T04:36:16.966437Z",
     "shell.execute_reply": "2023-06-01T04:36:16.966437Z",
     "shell.execute_reply.started": "2023-06-01T04:34:31.340321Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "zero_corr = t1[t1['corr'] < abs(0.41)]\n",
    "zero_corr_rubrics = pd.concat([zero_corr.r1, zero_corr.r2]).drop_duplicates().values\n",
    "zero_corr_rubrics = [i.split('_')[1] for i in zero_corr_rubrics]\n",
    "\n",
    "df_dates_zero = df_dates_all[df_dates_all['city_rubric'].apply(lambda x: (x in zero_corr_rubrics[:500]) and (len(x.replace('. Запросы','')) < 35) )]\n",
    "\n",
    "dfx_pred = pd.pivot_table(df_dates_zero, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if 'Запросы' in i and 'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if 'Запросы' in i and 'Владивосток' in i]].iloc[:,:3]\n",
    "\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "dfx.columns = [i.replace('. Запросы','') for i in dfx.columns]\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_3_other.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "b217ea86-c635-4479-9796-81f4ba57683d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:38:01.266718Z",
     "iopub.status.busy": "2023-06-01T04:38:01.266718Z",
     "iopub.status.idle": "2023-06-01T04:38:03.781800Z",
     "shell.execute_reply": "2023-06-01T04:38:03.781203Z",
     "shell.execute_reply.started": "2023-06-01T04:38:01.266718Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-06-01 07:38:03 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "slides[39].shapes.add_picture('pics/corr_matrix_other.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[40].shapes.add_picture('pics/corr_matrix_1_other.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[41].shapes.add_picture('pics/corr_matrix_2_other.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[42].shapes.add_picture('pics/corr_matrix_3_other.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "save_pptx()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93db5a51-e380-4aa7-a927-e65642f09868",
   "metadata": {},
   "source": [
    "## Бизнес город показатель"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "40df01eb-23e6-411b-a24f-a9fec3d0a284",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:40:49.498155Z",
     "iopub.status.busy": "2023-06-01T04:40:49.498155Z",
     "iopub.status.idle": "2023-06-01T04:41:33.257740Z",
     "shell.execute_reply": "2023-06-01T04:41:33.256709Z",
     "shell.execute_reply.started": "2023-06-01T04:40:49.498155Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "negative_corr = t1_100.sample(1000)\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "#negative_corr_rubrics = [i.replace('Москва. ', '') for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_users[df_dates_users['city_rubric_param'].apply(lambda x: ((x in negative_corr_rubrics) and (len(x) < 35)))]\n",
    "df_dates_negative['city_rubric_param'] = df_dates_negative['city_rubric_param']\n",
    "dfx_pred = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if  'Алматы' in i]].iloc[:,:8]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if  'Астана' in i]].iloc[:,:8]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if  'Москва' in i]].iloc[:,:8]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if  'Калининград' in i]].iloc[:,:8]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if  'СПБ' in i]].iloc[:,:8]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if  'Ташкент' in i]].iloc[:,:8]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if  'Новосибирск' in i]].iloc[:,:8]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if  'ОАЭ' in i]].iloc[:,:8]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if  'Бишкек' in i]].iloc[:,:8]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if  'Владивосток' in i]].iloc[:,:8]\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=False, cbar=False)\n",
    "plt.xticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.yticks(font=fpath, fontsize=fontsize*0.5)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_other2.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97816615-0339-4007-aad0-939dc707683c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:45:13.113828Z",
     "iopub.status.busy": "2023-06-01T04:45:13.113828Z",
     "iopub.status.idle": "2023-06-01T04:45:44.764205Z",
     "shell.execute_reply": "2023-06-01T04:45:44.764205Z",
     "shell.execute_reply.started": "2023-06-01T04:45:13.113828Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "df_dates_all = df_dates_users\n",
    "df_dates_all['city_rubric'] = df_dates_all['city_rubric_param']\n",
    "\n",
    "positive_corr = t2[t2['corr'] >= 0.95]\n",
    "positive_corr_rubrics = pd.concat([positive_corr.r1, positive_corr.r2]).drop_duplicates().values\n",
    "#positive_corr_rubrics = [i.split('_')[1] for i in positive_corr_rubrics]\n",
    "\n",
    "df_dates_positive = df_dates_all[df_dates_all['city_rubric'].apply(lambda x: ((x in positive_corr_rubrics[:500]) and (len(x) < 35)))]\n",
    "\n",
    "dfx_pred = pd.pivot_table(df_dates_positive, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if  'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if  'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if  'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if  'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if  'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if  'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if  'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if  'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if  'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if  'Владивосток' in i]].iloc[:,:3]\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_1_other2.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "180951c6-b82e-454b-9e71-e6f04c26a291",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:46:01.304800Z",
     "iopub.status.busy": "2023-06-01T04:46:01.304800Z",
     "iopub.status.idle": "2023-06-01T04:46:47.127562Z",
     "shell.execute_reply": "2023-06-01T04:46:47.127562Z",
     "shell.execute_reply.started": "2023-06-01T04:46:01.304800Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "negative_corr = t1_100.sample(1000)\n",
    "negative_corr_rubrics = pd.concat([negative_corr.r1, negative_corr.r2]).drop_duplicates().values\n",
    "#negative_corr_rubrics = [i.replace('Москва. ', '') for i in negative_corr_rubrics]\n",
    "\n",
    "df_dates_negative = df_dates_users[df_dates_users['city_rubric_param'].apply(lambda x: ((x in negative_corr_rubrics) and (len(x) < 35)))]\n",
    "df_dates_negative['city_rubric_param'] = df_dates_negative['city_rubric_param']\n",
    "dfx_pred = pd.pivot_table(df_dates_negative, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if  'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if  'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if  'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if  'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if  'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if  'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if  'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if  'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if  'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if  'Владивосток' in i]].iloc[:,:3]\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_2_other2.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d960535e-88ca-4174-8f3d-200184dea968",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:46:47.127562Z",
     "iopub.status.busy": "2023-06-01T04:46:47.127562Z",
     "iopub.status.idle": "2023-06-01T04:50:29.022572Z",
     "shell.execute_reply": "2023-06-01T04:50:29.022572Z",
     "shell.execute_reply.started": "2023-06-01T04:46:47.127562Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "zero_corr = t1[t1['corr'] < abs(0.41)]\n",
    "zero_corr_rubrics = pd.concat([zero_corr.r1, zero_corr.r2]).drop_duplicates().values\n",
    "zero_corr_rubrics = [i.split('_')[1] for i in zero_corr_rubrics]\n",
    "\n",
    "df_dates_zero = df_dates_all[df_dates_all['city_rubric'].apply(lambda x: (x in zero_corr_rubrics[:500]) and (len(x) < 35) )]\n",
    "\n",
    "dfx_pred = pd.pivot_table(df_dates_zero, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric'])\n",
    "dfx_1 = dfx_pred[[ i for i in dfx_pred.columns if  'Алматы' in i]].iloc[:,:3]\n",
    "dfx_2 = dfx_pred[[ i for i in dfx_pred.columns if  'Астана' in i]].iloc[:,:3]\n",
    "dfx_3 = dfx_pred[[ i for i in dfx_pred.columns if  'Москва' in i]].iloc[:,:3]\n",
    "dfx_4 = dfx_pred[[ i for i in dfx_pred.columns if  'Калининград' in i]].iloc[:,:3]\n",
    "dfx_5 = dfx_pred[[ i for i in dfx_pred.columns if  'СПБ' in i]].iloc[:,:3]\n",
    "dfx_6 = dfx_pred[[ i for i in dfx_pred.columns if  'Ташкент' in i]].iloc[:,:3]\n",
    "dfx_7 = dfx_pred[[ i for i in dfx_pred.columns if  'Новосибирск' in i]].iloc[:,:3]\n",
    "dfx_8 = dfx_pred[[ i for i in dfx_pred.columns if  'ОАЭ' in i]].iloc[:,:3]\n",
    "dfx_9 = dfx_pred[[ i for i in dfx_pred.columns if  'Бишкек' in i]].iloc[:,:3]\n",
    "dfx_10 = dfx_pred[[i for i in dfx_pred.columns if  'Владивосток' in i]].iloc[:,:3]\n",
    "dfx = (dfx_1\n",
    ".merge(dfx_2, left_index=True, right_index=True)\n",
    ".merge(dfx_3, left_index=True, right_index=True)\n",
    ".merge(dfx_4, left_index=True, right_index=True)\n",
    ".merge(dfx_5, left_index=True, right_index=True)\n",
    ".merge(dfx_6, left_index=True, right_index=True)\n",
    ".merge(dfx_7, left_index=True, right_index=True)\n",
    ".merge(dfx_8, left_index=True, right_index=True)\n",
    ".merge(dfx_9, left_index=True, right_index=True)\n",
    ".merge(dfx_10, left_index=True, right_index=True)\n",
    ")\n",
    "\n",
    "t_h = dfx.corr().fillna(0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax = sns.heatmap(t_h, annot=True, cbar=False, annot_kws={'size':round(fontsize * 0.6)})\n",
    "plt.xticks(font=fpath, fontsize=fontsize)\n",
    "plt.yticks(font=fpath, fontsize=fontsize)\n",
    "plt.rcParams['figure.dpi'] = 2**8\n",
    "plt.rcParams['figure.figsize']=size_x,size_y\n",
    "ax.set_xlabel('')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "fig.savefig(f'pics/corr_matrix_3_other2.png', bbox_inches='tight', transparent=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "ff72a473-8f9b-4973-ae0a-d17b4cda13db",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-01T04:50:29.022572Z",
     "iopub.status.busy": "2023-06-01T04:50:29.022572Z",
     "iopub.status.idle": "2023-06-01T04:50:35.250464Z",
     "shell.execute_reply": "2023-06-01T04:50:35.242851Z",
     "shell.execute_reply.started": "2023-06-01T04:50:29.022572Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-06-01 07:50:35 test_new.pptx\n"
     ]
    }
   ],
   "source": [
    "prs = pptx.Presentation('pics/test_new.pptx')\n",
    "slides = prs.slides\n",
    "x, y = 30.7, 16\n",
    "slides[43].shapes.add_picture('pics/corr_matrix_other2.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[44].shapes.add_picture('pics/corr_matrix_1_other2.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[45].shapes.add_picture('pics/corr_matrix_2_other2.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "slides[46].shapes.add_picture('pics/corr_matrix_3_other2.png', Cm(0.7), Cm(2.4), Cm(x), Cm(y))\n",
    "save_pptx()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2512cb92-a803-4551-a52f-d3ded9a6123b",
   "metadata": {},
   "source": [
    "# Формулы корелляций"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1972e5cb-4bd3-4faa-acfd-7aabebbb0a7d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T19:31:01.228029Z",
     "iopub.status.busy": "2023-06-07T19:31:01.227030Z",
     "iopub.status.idle": "2023-06-07T19:31:06.999389Z",
     "shell.execute_reply": "2023-06-07T19:31:06.998389Z",
     "shell.execute_reply.started": "2023-06-07T19:31:01.228029Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ready: 2023-06-07 22:31:06\n"
     ]
    }
   ],
   "source": [
    "%run functions.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "beb31f0c-4341-4522-b122-c3ab5552256e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T19:36:51.329450Z",
     "iopub.status.busy": "2023-06-07T19:36:51.329450Z",
     "iopub.status.idle": "2023-06-07T19:37:06.964487Z",
     "shell.execute_reply": "2023-06-07T19:37:06.963490Z",
     "shell.execute_reply.started": "2023-06-07T19:36:51.329450Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 11.2 s\n",
      "Wall time: 15.6 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "exclude_rubrics = ['Общая рубрика ЗМК','Заказ такси','Спектакли','Детские дома и Приюты','Кладбища','Крематории','Памятники / Ритуальные принадлежности','Помощь в организации похорон','Прощальные залы','Воскресные школы','Дацаны','Духовные учебные заведения','Иконописные мастерские','Медресе','Мечети','Молитвенные комнаты','Монастыри','Паломнические поездки','Приходы','Религиозные организации','Религиозные товары','Синагога','Храмы / Соборы / Церкви','Часовни','Радиостанции','Телеканалы','Телекомпании / Радиокомпании','Герои с нашего двора','Отлов бездомных животных','Приюты для животных','Ритуальные услуги для животных','Выгодные покупки с 2ГИС','Ликвидация нефтеразливов / газонефтеводопроявлений','Библиотеки','Библиотечные коллекторы','Городские оркестры','Кинопрокатные компании','Киностудии','Мавзолей','Музеи','Планетарий','Реставрация памятников истории и культуры','Театры','Филармония','Хоровые капеллы','Художественные выставки / Галереи','Аварийные службы','Пожарная охрана','Скорая медицинская помощь','Службы аварийных комиссаров','Службы спасения / экстренного вызова','Справочно-информационные услуги','Телефоны доверия','Управление ГО и ЧС','Штрафстоянки','Эвакуация автомобилей','Администрации поселений','Администрации районов / округов городской власти','Администрации районов / округов региональной власти','Администрация города / городского округа','Администрация области / края','Правительство','Пенсионные фонды','Фонд обязательного медицинского страхования','Фонды социального страхования','Законодательная власть','Представительная власть','Военная комендатура','Военные комиссариаты','Благотворительные фонды','Венчурные фонды','Гарантийные фонды','ГОРПО / РАЙПО','Гранты','Залоговые фонды','Избирательные комиссии','Избирательные участки','Нотариальные палаты','Общественные группы','Общественные организации','Политические партии','Приёмные депутатов','Приёмные Президента','Приёмные Уполномоченных по правам человека','Территориальные общественные самоуправления','Гидрометеослужба','Организации природоохраны','Инспекции','Налоговые инспекции','Оператор системы взимания платы','Органы государственного надзора','Отделы ветеринарно-санитарного контроля','Службы мелиорации земель','Федеральные казначейства','Федеральные службы','Посольства / Консульства','Представительства субъектов РФ','ГИБДД','Главное управление по вопросам миграции МВД РФ','Изоляторы временного содержания / Спецприёмники','Исправительные учреждения','Медвытрезвители','Народные дружины','Общественные пункты охраны порядка','ОМВД / УМВД / ГУМВД / МВД','Отделения полиции','Прокуратура','Следственный комитет','Службы судебных приставов','Судебная / внесудебная экспертиза','Суды','Таможня','Участковые пункты полиции','Государственные архивы','ЗАГСы','Бэби-боксы','Детские дома / Приюты','Дома престарелых','Дома ребёнка','Молочные кухни','МФЦ','Ночлежные дома','Отделы доставки пенсий и пособий','Службы занятости населения','Социальные службы','Центры квотирования рабочих мест','Антиквариат','Багетные мастерские','Подвеска / Подсветка для картин','Филателия / Нумизматика','Художественные мастерские','Художественные товары','Общая рубрика ЗМК','Заказ такси','Спектакли','Детские дома и Приюты','Автовокзалы и Автостанции','Аэропорты','Контрольные браслеты','Онлайн-заправки 2ГИС','Спички','Лицеи-интернаты','Экспертиза авиационных происшествий','Саморегулируемые организации (СРО)','Аренда жид тупика','Ритуальные принадлежности','Железнодорожные вокзалы','Ликвидация нефтеразливов и газонефтеводопроявлений','Больницы','Дома и дворцы культуры','Санатории и Профилактории','Телекомпании и Радиокомпании','Организация театральных и концертных мероприятий','Акиматы районов города','Управление ГО и ЧС', 'ЦОН','ОМВД / УМВД / ГУМВД и МВД ','CD / DVD и BD','Лучшие достопримечательности города','Изоляторы временного содержания и Спецприёмники','Администрации районов и округов городской власти','Администрация города и городского округа','Полицейские академии','Акимат города','Почтовые отделения', 'Взрослые поликлиники', 'Кладбища','Крематории','Памятники / Ритуальные принадлежности','Помощь в организации похорон','Прощальные залы','Воскресные школы','Дацаны','Духовные учебные заведения','Иконописные мастерские','Медресе','Мечети','Молитвенные комнаты','Монастыри','Паломнические поездки','Приходы','Религиозные организации','Религиозные товары','Синагога','Храмы / Соборы / Церкви','Часовни','Радиостанции','Телеканалы','Телекомпании / Радиокомпании','Герои с нашего двора','Отлов бездомных животных','Приюты для животных','Ритуальные услуги для животных','Выгодные покупки с 2ГИС','Ликвидация нефтеразливов / газонефтеводопроявлений','Библиотеки','Библиотечные коллекторы','Городские оркестры','Кинопрокатные компании','Киностудии','Мавзолей','Музеи','Планетарий','Реставрация памятников истории и культуры','Театры','Филармония','Хоровые капеллы','Художественные выставки / Галереи','Аварийные службы','Пожарная охрана','Скорая медицинская помощь','Службы аварийных комиссаров','Службы спасения / экстренного вызова','Справочно-информационные услуги','Телефоны доверия','Управление ГО и ЧС','Штрафстоянки','Эвакуация автомобилей','Администрации поселений','Администрации районов / округов городской власти','Администрации районов / округов региональной власти','Администрация города / городского округа','Администрация области / края','Правительство','Пенсионные фонды','Фонд обязательного медицинского страхования','Фонды социального страхования','Законодательная власть','Представительная власть','Военная комендатура','Военные комиссариаты','Благотворительные фонды','Венчурные фонды','Гарантийные фонды','ГОРПО / РАЙПО','Гранты','Залоговые фонды','Избирательные комиссии','Избирательные участки','Нотариальные палаты','Общественные группы','Общественные организации','Политические партии','Приёмные депутатов','Приёмные Президента','Приёмные Уполномоченных по правам человека','Территориальные общественные самоуправления','Гидрометеослужба','Организации природоохраны','Инспекции','Налоговые инспекции','Оператор системы взимания платы','Органы государственного надзора','Отделы ветеринарно-санитарного контроля','Службы мелиорации земель','Федеральные казначейства','Федеральные службы','Посольства / Консульства','Представительства субъектов РФ','ГИБДД','Главное управление по вопросам миграции МВД РФ','Изоляторы временного содержания / Спецприёмники','Исправительные учреждения','Медвытрезвители','Народные дружины','Общественные пункты охраны порядка','ОМВД / УМВД / ГУМВД / МВД','Отделения полиции','Прокуратура','Следственный комитет','Службы судебных приставов','Судебная / внесудебная экспертиза','Суды','Таможня','Участковые пункты полиции','Государственные архивы','ЗАГСы','Бэби-боксы','Детские дома / Приюты','Дома престарелых','Дома ребёнка','Молочные кухни','МФЦ','Ночлежные дома','Отделы доставки пенсий и пособий','Службы занятости населения','Социальные службы','Центры квотирования рабочих мест','Воскресные школы','Дацаны','Духовные учебные заведения','Иконописные мастерские','Медресе','Мечети','Молитвенные комнаты','Монастыри','Паломнические поездки','Приходы','Религиозные организации','Религиозные товары','Синагога','Храмы / Соборы / Церкви','Часовни','Антиквариат','Багетные мастерские','Подвеска / Подсветка для картин','Филателия / Нумизматика','Художественные мастерские','Художественные товары','Банкоматы', 'Железнодорожные вокзалы', 'Музеи','Центры паровых коктейлей', 'Школы', 'Университеты', 'Больницы','Табачные изделия / Товары для курения','Художественные выставки / Галереи','Научно-исследовательские институты', 'Храмы / Соборы / Церкви', 'Студенческие общежития'\n",
    "]\n",
    "city = 'Москва'\n",
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_requests'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Запросы'\n",
    "df_requests = df_all.copy()\n",
    "\n",
    "df_all = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df_all = df_all[df_all['city'] == city]\n",
    "df_all['city_rubric'] = df_all['city'] + '. ' + df_all['rubric']\n",
    "df_all = df_all[~df_all.rubric.apply(lambda x: (x is None or x.isnumeric()))]\n",
    "df_all['cnt_all'] = df_all.groupby('city_rubric')['cnt_users'].transform(np.sum)\n",
    "df_all = df_all.drop(['date_r', 'cnt_requests', 'cnt_users'], axis=1).drop_duplicates()\n",
    "df_all = df_all.sort_values(['cnt_all'], ascending=False).reset_index().reset_index().set_index('level_0').drop('index', axis=1)\n",
    "df_all['cnt_city_rubric'] = df_all['cnt_all'].astype(str) + '_' + df_all['city_rubric'] + '. ' + 'Пользователи'\n",
    "df_users = df_all.copy()\n",
    "\n",
    "df_1000 = pd.concat([df_requests, df_users])\n",
    "df_1000 = df_1000[df_1000.rubric.apply(lambda x: x not in exclude_rubrics)]\n",
    "df_1000['param'] = df_1000.cnt_city_rubric.apply(lambda x: x.split('. ')[2])\n",
    "\n",
    "df_1000['r_cnt'] = df_1000.groupby('rubric')['cnt_all'].transform(np.sum)\n",
    "df_1000_1 = df_1000.sort_values(['r_cnt', 'city'], ascending=[False, True]).rename({'cnt_city_rubric':'cnt_city_rubric_param'}, axis=1)\n",
    "df_1000_1['city_rubric_param'] = df_1000_1['city_rubric'] + '. ' + df_1000_1['param']\n",
    "\n",
    "rubrics = [*df_1000_1.rubric.drop_duplicates().values]\n",
    "cities = [*df_1000_1.city.unique()] * 139\n",
    "params = [*df_1000_1.param.values]\n",
    "\n",
    "iter_set = list(zip(rubrics, cities, params))\n",
    "iter_df = pd.DataFrame(iter_set, columns=['rubric', 'city', 'param'])\n",
    "iter_df['city_rubric_param'] = iter_df['city'] + '. ' + iter_df['rubric'] + '. ' + iter_df['param']\n",
    "iter_df = df_1000_1[['city_rubric_param', 'cnt_city_rubric_param']].merge(iter_df, left_on='city_rubric_param', right_on='city_rubric_param')\n",
    "\n",
    "## Переменные формулы\n",
    "df1 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df1 = df1[df1['city'] == city]\n",
    "df1['city_rubric_param'] = df1['city'] + '. ' + df1['rubric'] + '. ' + 'Пользователи'\n",
    "df1 = df1[['date_r', 'cnt_users', 'city_rubric_param']].rename({'cnt_users':'cnt'}, axis=1)\n",
    "\n",
    "df2 = pd.read_parquet(f'files/ten_city/df_dates_users_city_agg2')\n",
    "#df2 = df2[df2['city'] == city]\n",
    "df2['city_rubric_param'] = df2['city'] + '. ' + df2['rubric'] + '. ' + 'Запросы'\n",
    "df2 = df2[['date_r', 'cnt_requests', 'city_rubric_param']].rename({'cnt_requests':'cnt'}, axis=1)\n",
    "df_dates_users = pd.concat([df1, df2])\n",
    "df_dates_users['cnt_all'] = df_dates_users.groupby('city_rubric_param')['cnt'].transform(np.sum)\n",
    "df_dates_users['city_rubric_param_cnt'] = df_dates_users['cnt_all'].astype(str) + '_' + df_dates_users['city_rubric_param']\n",
    "\n",
    "dfx = pd.pivot_table(df_dates_users, values='cnt', aggfunc='sum', index=['date_r'], columns=['city_rubric_param_cnt'])\n",
    "dfx = dfx.loc[:,iter_df.cnt_city_rubric_param]\n",
    "t = dfx.corr().fillna(0).astype(np.float16)\n",
    "\n",
    "t = t.rename_axis([None]).rename_axis([None], axis=1)\n",
    "t1 = t.stack()\n",
    "t1 = t1.reset_index()\n",
    "t1.columns = ['r1', 'r2', 'corr']\n",
    "t1 = t1[t1.r1 != t1.r2]\n",
    "\n",
    "t1_100 = t1 #df_100.merge(t1, left_on='cnt_city_rubric', right_on='r1')\n",
    "\n",
    "t1_100['r1_cnt'] = t1_100.r1.apply(lambda x: x.split('_')[0])\n",
    "t1_100['r2_cnt'] = t1_100.r2.apply(lambda x: x.split('_')[0])\n",
    "\n",
    "t1_100.r1 = t1_100.r1.apply(lambda x: x.split('_')[1])\n",
    "t1_100.r2 = t1_100.r2.apply(lambda x: x.split('_')[1])\n",
    "t1_100['r1_city'] = t1_100.r1.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r2_city'] = t1_100.r2.apply(lambda x: x.split('. ')[0])\n",
    "t1_100['r1_rubric'] = t1_100.r1.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r2_rubric'] = t1_100.r2.apply(lambda x: x.split('. ')[1])\n",
    "t1_100['r1_param'] = t1_100.r1.apply(lambda x: x.split('. ')[2])\n",
    "t1_100['r2_param'] = t1_100.r2.apply(lambda x: x.split('. ')[2])\n",
    "\n",
    "t1_100.r1_cnt = t1_100.r1_cnt.astype(int)\n",
    "t1_100.r2_cnt = t1_100.r2_cnt.astype(int)\n",
    "\n",
    "#t1_100 = t1_100[t1_100.r2_cnt > 1000]\n",
    "\n",
    "t1_100 = t1_100[['r1_city', 'r1_rubric', 'r1_param', 'r2_city', 'r2_rubric', 'r2_param', 'corr', 'r1_cnt', 'r2_cnt']]\n",
    "\n",
    "iter_set = list(zip(rubrics, cities, params))\n",
    "\n",
    "new_set = list(set(zip(list(t1_100['r1_rubric'].values), list(t1_100['r1_city'].values))))\n",
    "\n",
    "iter_set = [i for i in iter_set if i[:2] in new_set]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "3dc64d69-3717-4067-9602-4b88525b8e02",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T19:38:04.998504Z",
     "iopub.status.busy": "2023-06-07T19:38:04.998504Z",
     "iopub.status.idle": "2023-06-07T19:38:37.622751Z",
     "shell.execute_reply": "2023-06-07T19:38:37.621755Z",
     "shell.execute_reply.started": "2023-06-07T19:38:04.998504Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Алкогольные напитки Новосибирск\n",
      "Бытовая техника СПБ\n",
      "Услуги косметолога Ташкент\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Зоотовары Алматы\n",
      "Продажа легковых автомобилей Астана\n",
      "Канцелярские товары / Учебные принадлежности Бишкек\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Хостелы Новосибирск\n",
      "Колледжи ОАЭ\n",
      "Автозапчасти для иномарок СПБ\n"
     ]
    }
   ],
   "source": [
    "dfo = pd.read_excel(r\"C:\\Users\\Mamed\\Downloads\\Рубрикатор_230605_0030_отметки красным.xlsx\")[['Отрасль', 'Рубрика']].dropna(subset='Отрасль')\n",
    "t1_100_all = t1_100\n",
    "t1_100_city = t1_100#[(t1_100.r1_city == result_city) & (t1_100.r2_city == variables_city) & (t1_100.r1_param == result_param) & (t1_100.r2_param == variables_param)]\n",
    "r_list = [iter_df.iloc[:55]]\n",
    "\n",
    "### f13\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[46:49]):\n",
    "    print(rubric, city)\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r1_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r1_otrasl'}, axis=1)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r2_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r2_otrasl'}, axis=1)\n",
    "    f1_i = f1_i[f1_i['r1_otrasl'] == f1_i['r2_otrasl']]\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f13 = pd.concat(df_list)\n",
    "f13['group'] = 'f13'\n",
    "\n",
    "\n",
    "### f14\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[49:52]):\n",
    "    print(rubric, city)\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r1_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r1_otrasl'}, axis=1)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r2_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r2_otrasl'}, axis=1)\n",
    "    f1_i = f1_i[f1_i['r1_otrasl'] != f1_i['r2_otrasl']]\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f14 = pd.concat(df_list)\n",
    "f14['group'] = 'f14'\n",
    "\n",
    "\n",
    "### f15\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[55:58]):\n",
    "    print(rubric, city)\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r1_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r1_otrasl'}, axis=1)\n",
    "    f1_i = f1_i.merge(dfo, left_on='r2_rubric', right_on='Рубрика').drop('Рубрика',axis=1).rename({'Отрасль':'r2_otrasl'}, axis=1)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i_1 = f1_i[f1_i['r1_otrasl'] != f1_i['r2_otrasl']].iloc[:10]\n",
    "    f1_i_2 = f1_i[f1_i['r1_otrasl'] == f1_i['r2_otrasl']].iloc[:10]\n",
    "    f1_i = pd.concat([f1_i_1, f1_i_2])\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f15 = pd.concat(df_list)\n",
    "f15['group'] = 'f15'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "31b91476-6903-4f19-9257-f6e303679e5c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T19:38:49.072009Z",
     "iopub.status.busy": "2023-06-07T19:38:49.072009Z",
     "iopub.status.idle": "2023-06-07T20:06:12.992887Z",
     "shell.execute_reply": "2023-06-07T20:06:12.991880Z",
     "shell.execute_reply.started": "2023-06-07T19:38:49.072009Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/3 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 11min 27s\n",
      "Wall time: 27min 23s\n"
     ]
    },
    {
     "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>r1_city</th>\n",
       "      <th>r1_rubric</th>\n",
       "      <th>r1_param</th>\n",
       "      <th>r2_city</th>\n",
       "      <th>r2_rubric</th>\n",
       "      <th>r2_param</th>\n",
       "      <th>corr</th>\n",
       "      <th>r1_cnt</th>\n",
       "      <th>r2_cnt</th>\n",
       "      <th>avg_corr</th>\n",
       "      <th>r1_rubric_city</th>\n",
       "      <th>r2_rubric_city</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>group</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>f1</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f10</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f11</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f12</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f2</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f3</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f4</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f5</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f6</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f7</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f8</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>f9</th>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "      <td>60</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       r1_city  r1_rubric  r1_param  r2_city  r2_rubric  r2_param  corr  \\\n",
       "group                                                                     \n",
       "f1          60         60        60       60         60        60    60   \n",
       "f10         60         60        60       60         60        60    60   \n",
       "f11         60         60        60       60         60        60    60   \n",
       "f12         60         60        60       60         60        60    60   \n",
       "f2          60         60        60       60         60        60    60   \n",
       "f3          60         60        60       60         60        60    60   \n",
       "f4          60         60        60       60         60        60    60   \n",
       "f5          60         60        60       60         60        60    60   \n",
       "f6          60         60        60       60         60        60    60   \n",
       "f7          60         60        60       60         60        60    60   \n",
       "f8          60         60        60       60         60        60    60   \n",
       "f9          60         60        60       60         60        60    60   \n",
       "\n",
       "       r1_cnt  r2_cnt  avg_corr  r1_rubric_city  r2_rubric_city  \n",
       "group                                                            \n",
       "f1         60      60        60              60              60  \n",
       "f10        60      60        60              60              60  \n",
       "f11        60      60        60              60              60  \n",
       "f12        60      60        60              60              60  \n",
       "f2         60      60        60              60              60  \n",
       "f3         60      60        60              60              60  \n",
       "f4         60      60        60              60              60  \n",
       "f5         60      60        60              60              60  \n",
       "f6         60      60        60              60              60  \n",
       "f7         60      60        60              60              60  \n",
       "f8         60      60        60              60              60  \n",
       "f9         60      60        60              60              60  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "### f1\n",
    "\n",
    "\n",
    "\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[:3]):\n",
    "    f1_i = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric)].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f1 = pd.concat(df_list)\n",
    "f1['group'] = 'f1'\n",
    "\n",
    "### f2\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[3:6]):\n",
    "    f1_i = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric)].sort_values(['corr', 'r2_cnt'], ascending=[True,False])\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f2 = pd.concat(df_list)\n",
    "f2['group'] = 'f2'\n",
    "\n",
    "### f3\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[6:9]):\n",
    "    t1_c1 = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric)]\n",
    "    t1_c1['corr_abs'] = t1_c1['corr'].abs()\n",
    "    t1_c1.sort_values(['corr_abs', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = t1_c1.iloc[:]\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20].drop('corr_abs', axis=1)\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f3 = pd.concat(df_list)\n",
    "f3['group'] = 'f3'\n",
    "\n",
    "### f4\n",
    "#### органическая – для построения тренда используются смысловые ряды переменных: запросы базируются на запросах, пользователи на пользователях, динамика на динамике, схожесть на схожести\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[12:15]):\n",
    "    f1_i = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric) & ((t1_100_city.r1_param == t1_100_city.r2_param) & (t1_100_city.r2_param == param))].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f4 = pd.concat(df_list).sort_values(['r1_city','r1_rubric','r1_param'])\n",
    "f4['group'] = 'f4'\n",
    "\n",
    "### f5\n",
    "\n",
    "#### синтетическая – в формулу замешиваются показатели наибольшей предсказательной силы, без учёта их физической природы\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[15:18]):\n",
    "    t1_c1 = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric)].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_i = t1_c1.iloc[:]\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f5 = pd.concat(df_list).sort_values(['r1_city','r1_rubric','r1_param'])\n",
    "f5['group'] = 'f5'\n",
    "\n",
    "### f6\n",
    "\n",
    "#### адаптационная – задействуется оптимизационный алгоритм вариативности, чередующий переменные, не допуская преобладания единственного типа.\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[18:21]):\n",
    "    f1_1 = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric) & (t1_100_city.r2_param == 'Пользователи')].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_1 = f1_1[f1_1.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_1 = f1_1.iloc[:10]\n",
    "    r_list.append(f1_1.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_1.r2_rubric.drop_duplicates())\n",
    "    \n",
    "    f1_2 = t1_100_city[(t1_100_city.r1_city == city) & (t1_100_city.r1_rubric == rubric) & (t1_100_city.r2_param == 'Запросы')].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_2 = f1_2[f1_2.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_2 = f1_2.iloc[:10]\n",
    "    \n",
    "    f1_i = pd.concat([f1_1, f1_2])\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f6 = pd.concat(df_list)\n",
    "f6['group'] = 'f6'\n",
    "\n",
    "#### f7\n",
    "df_list = []\n",
    "t1_100_all = t1_100\n",
    "for rubric, city, param  in tqdm(iter_set[21:24]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric) & (t1_100_all.r2_city == city) & (t1_100_all.r2_rubric != rubric)].sort_values(['corr', 'r2_cnt'], ascending=[False,False])\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f7 = pd.concat(df_list)\n",
    "f7['group'] = 'f7'\n",
    "\n",
    "### f8\n",
    "\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[24:27]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric) & (t1_100_all.r2_city != city) & (t1_100_all.r2_rubric != rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f8 = pd.concat(df_list)\n",
    "f8['group'] = 'f8'\n",
    "\n",
    "\n",
    "### f9\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[9:12]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f9 = pd.concat(df_list)\n",
    "f9['group'] = 'f9'\n",
    "\n",
    "### f10\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[27:30]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f10 = pd.concat(df_list)\n",
    "f10['group'] = 'f10'\n",
    "\n",
    "### f11\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[30:33]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f11 = pd.concat(df_list)\n",
    "f11['group'] = 'f11'\n",
    "\n",
    "### f12\n",
    "df_list = []\n",
    "for rubric, city, param  in tqdm(iter_set[33:36]):\n",
    "    f1_i = t1_100_all[(t1_100_all.r1_city == city) & (t1_100_all.r1_rubric == rubric)]\n",
    "    f1_i.sort_values(['corr', 'r2_cnt'], ascending=[False,False], inplace=True)\n",
    "    f1_i = f1_i[f1_i.r2_rubric.apply(lambda x: (x not in pd.concat(r_list).values) & (len(x) <= 50) & (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "    f1_i = f1_i.iloc[:20]\n",
    "    f1_i['avg_corr'] = f1_i['corr'].mean()\n",
    "    r_list.append(f1_i.r1_rubric.drop_duplicates())\n",
    "    r_list.append(f1_i.r2_rubric.drop_duplicates())\n",
    "    df_list.append(f1_i)\n",
    "f12 = pd.concat(df_list)\n",
    "f12['group'] = 'f12'\n",
    "\n",
    "f_all = pd.concat([f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12])\n",
    "f_all['r1_rubric_city'] = f_all['r1_city'] + '. ' + f_all['r1_rubric']\n",
    "f_all['r2_rubric_city'] = f_all['r2_city'] + '. ' + f_all['r2_rubric']\n",
    "\n",
    "f_all = f_all[f_all.r2_rubric.apply(lambda x: (x not in f1_i.r1_rubric.values))].drop_duplicates(subset='r2_rubric', keep=\"first\")\n",
    "f_all.groupby('group').count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "b7f20ca7-cb09-4404-b8b6-de2ed9d371a7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T20:09:15.531247Z",
     "iopub.status.busy": "2023-06-07T20:09:15.531247Z",
     "iopub.status.idle": "2023-06-07T20:09:15.903879Z",
     "shell.execute_reply": "2023-06-07T20:09:15.902896Z",
     "shell.execute_reply.started": "2023-06-07T20:09:15.531247Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "f_new = pd.concat([f13, f14, f15])\n",
    "\n",
    "f_new['r1_rubric_city'] = f_new['r1_city'] + '. ' + f_new['r1_rubric']\n",
    "f_new['r2_rubric_city'] = f_new['r2_city'] + '. ' + f_new['r2_rubric']\n",
    "f_new['r1_rubric_city_param'] = f_new['r1_rubric_city'] + '. ' + f_new['r1_param']\n",
    "f_new['r2_rubric_city_param'] = f_new['r2_rubric_city'] + '. ' + f_new['r2_param']\n",
    "\n",
    "f_all['r1_otrasl'] = ''\n",
    "f_all['r2_otrasl'] = ''\n",
    "f_all['r1_rubric_city_param'] = f_all['r1_rubric_city'] + '. ' + f_all['r1_param']\n",
    "f_all['r2_rubric_city_param'] = f_all['r2_rubric_city'] + '. ' + f_all['r2_param']\n",
    "\n",
    "f_new = f_new[f_all.columns]\n",
    "f_all_new = pd.concat([f_all, f_new])\n",
    "f_all_new.to_excel('f_all_new2.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "d02d3778-36c6-4ca4-9f60-e2a7b6ddb801",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-07T20:29:19.462161Z",
     "iopub.status.busy": "2023-06-07T20:29:19.462161Z",
     "iopub.status.idle": "2023-06-07T20:29:19.811568Z",
     "shell.execute_reply": "2023-06-07T20:29:19.810589Z",
     "shell.execute_reply.started": "2023-06-07T20:29:19.462161Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "f_cr = f_all_new[['r1_city','r1_rubric']].drop_duplicates()\n",
    "cr = list(zip(f_cr['r1_city'].values, f_cr['r1_rubric'].values))\n",
    "\n",
    "def replace_slash(s):\n",
    "    return s[::-1].replace('/', 'и', 1)[::-1].replace('СПБ', 'Санкт-Петербург')\n",
    "\n",
    "docx_name = 'Формулы_корреляции_Москва_230607_2000.docx'\n",
    "doc = docx.Document('files/report.docx')\n",
    "doc.save(docx_name)\n",
    "p = doc.add_paragraph()\n",
    "n1, n2, n3 = 0,0,0\n",
    "for city, rubric in cr:\n",
    "    f_i = f_all_new[(f_all_new.r1_city == city) & (f_all_new.r1_rubric == rubric)]\n",
    "    \n",
    "    if f_i['group'].iloc[0] in ('f1', 'f2', 'f3'): # 1-3 Бизнес\n",
    "        col1 = 'r1_rubric'\n",
    "        col2 = 'r2_rubric'   \n",
    "        col3 = col2\n",
    "        col4 = col2\n",
    "    elif f_i['group'].iloc[0] in ('f4', 'f5', 'f6'): # 4-6 Город-бизнес-показатель\n",
    "        col1 = 'r1_rubric_city_param'\n",
    "        col2 = 'r2_rubric_city_param'\n",
    "        col3 = col2\n",
    "        col4 = col2\n",
    "    elif f_i['group'].iloc[0] in ('f7', 'f8', 'f9'): # 7-9 Город-бизнес\n",
    "        col1 = 'r1_rubric_city'\n",
    "        col2 = 'r2_rubric_city'\n",
    "        col3 = col2\n",
    "        col4 = col2\n",
    "    elif f_i['group'].iloc[0] =='f10': # 10-12 Комбо\n",
    "        n1 += 1\n",
    "        if n1 == 1:\n",
    "            col1 = 'r1_rubric'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city'\n",
    "            col4 = col3\n",
    "        if n1 == 2:\n",
    "            col1 = 'r1_rubric_city'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city'\n",
    "            col4 = col3\n",
    "        if n1 == 3:\n",
    "            col1 = 'r1_rubric_city_param'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city'\n",
    "            col4 = col3\n",
    "    elif f_i['group'].iloc[0] =='f11': # 10-12 Комбо\n",
    "        n2 += 1\n",
    "        if n2 == 1:\n",
    "            col1 = 'r1_rubric'\n",
    "            col2 = 'r2_rubric_city'\n",
    "            col3 = 'r2_rubric_city_param'\n",
    "            col4 = col3\n",
    "        if n2 == 2:\n",
    "            col1 = 'r1_rubric_city'\n",
    "            col2 = 'r2_rubric_city'\n",
    "            col3 = 'r2_rubric_city_param'\n",
    "            col4 = col3\n",
    "        if n2 == 3:\n",
    "            col1 = 'r1_rubric_city_param'\n",
    "            col2 = 'r2_rubric_city'\n",
    "            col3 = 'r2_rubric_city_param'\n",
    "            col4 = col3\n",
    "    elif f_i['group'].iloc[0] =='f12': # 10-12 Комбо\n",
    "        n3 += 1\n",
    "        if n3 == 1:\n",
    "            col1 = 'r1_rubric'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city'\n",
    "            col4 = 'r2_rubric_city_param' \n",
    "        if n3 == 2:\n",
    "            col1 = 'r1_rubric_city'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city'\n",
    "            col4 = 'r2_rubric_city_param' \n",
    "        if n3 == 3:\n",
    "            col1 = 'r1_rubric_city_param'\n",
    "            col2 = 'r2_rubric'\n",
    "            col3 = 'r2_rubric_city' \n",
    "            col4 = 'r2_rubric_city_param'       \n",
    "            \n",
    "    head = f_i['group'].iloc[0] +': '+ f_i[col1].iloc[0] + ' = '\n",
    "    left_sq = 'f{'\n",
    "    body = '\\n\\t,'.join([f'{i:.4f} x {replace_slash(j)}' for i, j in list(zip(f_i['corr'].values, [*f_i[col2].values[:6], *f_i[col3].values[6:13], *f_i[col4].values[13:]]))])\n",
    "    right_sq = '}\\n'\n",
    "    \n",
    "    if f_i['group'].iloc[0] in ('f13', 'f14', 'f15'):\n",
    "        head = f_i['group'].iloc[0] +': '+ f_i['r1_otrasl'].iloc[0] + '. ' + f_i['r1_rubric'].iloc[0] + ' = '\n",
    "        left_sq = 'f{'\n",
    "        body = '\\n\\t,'.join([f'{i:.4f} x {replace_slash(j)}' for i, j in list(zip(f_i['corr'].values, [*f_i['r2_rubric'].values[:6], *f_i['r2_rubric'].values[6:13], *f_i['r2_rubric'].values[13:]]))])\n",
    "        right_sq = '}\\n'        \n",
    "        \n",
    "    #print(head, left_sq, body, right_sq)\n",
    "    runner = p.add_run(head)\n",
    "    runner.font.highlight_color = WD_COLOR_INDEX.YELLOW\n",
    "    runner2 = p.add_run(left_sq)\n",
    "    runner2.italic = True\n",
    "    p.add_run(body)\n",
    "    runner3 = p.add_run(right_sq)\n",
    "    runner3.italic = True\n",
    "    doc.save(docx_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "801415b4-0da2-461d-8bed-e5781ddf650f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "attachments": {
    "e6fc99ed-6622-4695-93de-c8ac1a25e3d8.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "9a78ab22-7f94-4d90-86a2-d3aafef70ab1",
   "metadata": {},
   "source": [
    "![image.png](attachment:e6fc99ed-6622-4695-93de-c8ac1a25e3d8.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e43621af-fb33-4df8-8e5b-ab96f7425578",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "1f41a371-d3e5-4939-8195-c7a75639a0a5",
   "metadata": {},
   "source": [
    "# Вставка ракеты в презентацию"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c36aa513-3f0e-48eb-b535-6c2a4f77144a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-14T23:24:38.334203Z",
     "iopub.status.busy": "2023-06-14T23:24:38.334203Z",
     "iopub.status.idle": "2023-06-14T23:24:51.080082Z",
     "shell.execute_reply": "2023-06-14T23:24:51.079087Z",
     "shell.execute_reply.started": "2023-06-14T23:24:38.334203Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ready: 2023-06-15 02:24:51\n"
     ]
    }
   ],
   "source": [
    "%run functions.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f0107e9d-a319-4545-baf9-7fd1d1201203",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-06-14T23:25:39.000237Z",
     "iopub.status.busy": "2023-06-14T23:25:39.000237Z",
     "iopub.status.idle": "2023-06-14T23:25:39.015012Z",
     "shell.execute_reply": "2023-06-14T23:25:39.015012Z",
     "shell.execute_reply.started": "2023-06-14T23:25:39.000237Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "19c18e2f-9fd5-404c-9a3b-5b3ed9d59c1a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-07-09T06:01:56.248412Z",
     "iopub.status.busy": "2023-07-09T06:01:56.232935Z",
     "iopub.status.idle": "2023-07-09T06:01:56.826385Z",
     "shell.execute_reply": "2023-07-09T06:01:56.826385Z",
     "shell.execute_reply.started": "2023-07-09T06:01:56.248412Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pptx\n",
    "import numpy as np\n",
    "from pptx.util import Cm\n",
    "\n",
    "# задайте положение ракеты по горизонтали в см\n",
    "left  = 23.6 \n",
    "# начальное и конечное положение верхушки ракеты в см\n",
    "first_top, last_top = 16.66, 1.3\n",
    "# номер слайда на котором с\n",
    "first_slide_num, last_slide_num = 3, 136\n",
    "presentation_path = r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\Python ШТ\\Презентация_ШТ_162_Эффективный Python_230709_0730.pptx\"\n",
    "\n",
    "prs = pptx.Presentation(presentation_path)\n",
    "slides = prs.slides\n",
    "top_coordinates = np.linspace(first_top, last_top, last_slide_num-first_slide_num)\n",
    "p1 = r\"C:\\Users\\Mamed\\YandexDisk\\_Documents\\Проекты\\2ГИС 2023\\Лого\\Ракета1.png\"\n",
    "\n",
    "n = 0\n",
    "for i, top_i in zip(range(first_slide_num-1, last_slide_num-1), top_coordinates):\n",
    "    if n < 4:\n",
    "        n += 1\n",
    "    p_i = p1.replace('Ракета1', f'Ракета{n}')\n",
    "    slides[i].shapes.add_picture(p_i, Cm(left), Cm(top_i))\n",
    "\n",
    "prs.save(presentation_path.replace('.pptx', '_new.pptx'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e510b091-cdc7-4af8-b5db-53db4cd05340",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}