notebooks/ranking/final_model/data-for-final-model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "db4e0453",
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:14.863052Z",
     "iopub.status.busy": "2021-12-09T10:08:14.854286Z",
     "iopub.status.idle": "2021-12-09T10:08:15.871402Z",
     "shell.execute_reply": "2021-12-09T10:08:15.870735Z",
     "shell.execute_reply.started": "2021-12-09T10:06:39.774209Z"
    },
    "papermill": {
     "duration": 1.055397,
     "end_time": "2021-12-09T10:08:15.871590",
     "exception": false,
     "start_time": "2021-12-09T10:08:14.816193",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using \u001b[34mXGBoost 1.5.0 \u001b[0m\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from tqdm import tqdm\n",
    "\n",
    "from collections import Counter\n",
    "\n",
    "import gc\n",
    "import cloudpickle\n",
    "\n",
    "import os, sys\n",
    "import random\n",
    "\n",
    "import warnings\n",
    "\n",
    "import xgboost as xgb\n",
    "\n",
    "print('Using \\033[34mXGBoost', xgb.__version__, '\\033[0m')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec9ae19e",
   "metadata": {
    "papermill": {
     "duration": 0.031223,
     "end_time": "2021-12-09T10:08:16.208248",
     "exception": false,
     "start_time": "2021-12-09T10:08:16.177025",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Main Part"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6976e5d3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:16.277905Z",
     "iopub.status.busy": "2021-12-09T10:08:16.277212Z",
     "iopub.status.idle": "2021-12-09T10:08:24.868548Z",
     "shell.execute_reply": "2021-12-09T10:08:24.867966Z",
     "shell.execute_reply.started": "2021-12-09T10:06:40.905935Z"
    },
    "papermill": {
     "duration": 8.629053,
     "end_time": "2021-12-09T10:08:24.868712",
     "exception": false,
     "start_time": "2021-12-09T10:08:16.239659",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "images = pd.read_csv('../input/final-image-traindataset-part0-4-count-5/final_images.csv')\n",
    "images['image_id'] = images.index\n",
    "matchings = pd.read_csv('../input/combiner-target-dataset-0-of-5/final_matchings.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2bb5a494",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:25.158174Z",
     "iopub.status.busy": "2021-12-09T10:08:25.157139Z",
     "iopub.status.idle": "2021-12-09T10:08:38.764267Z",
     "shell.execute_reply": "2021-12-09T10:08:38.763289Z",
     "shell.execute_reply.started": "2021-12-09T10:06:59.019942Z"
    },
    "papermill": {
     "duration": 13.647232,
     "end_time": "2021-12-09T10:08:38.764472",
     "exception": false,
     "start_time": "2021-12-09T10:08:25.117240",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:43: FutureWarning: Index.ravel returning ndarray is deprecated; in a future version this will return a view on self.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "all_results = []\n",
    "\n",
    "parts = ['../input/data-for-model-00/valid-70', '../input/data-for-model-00/valid-71', \\\n",
    "    '../input/data-for-model-05/train-09']\n",
    "\n",
    "for N, part in enumerate(parts):\n",
    "    \n",
    "    ranks = []\n",
    "\n",
    "    for k in range(7):\n",
    "        _ = pd.read_csv('../input/train-ranks-for-model-0{0}/short-ranks-{1}.csv'.format(k, N))\n",
    "        _.drop(columns=['image_url', 'hyperref', 'target', 'count', 'undigit_filename'], inplace=True)\n",
    "        _.rename(columns={'rank': 'rank' + str(k)}, inplace=True)\n",
    "    \n",
    "        ranks.append(_)\n",
    "        \n",
    "    topk = 100\n",
    "\n",
    "    ranks[0].sort_values(by=['image_id', 'rank0'], inplace=True, kind='mergesort', ascending=False)\n",
    "    result = ranks[0]#.groupby('image_id').head(topk)\n",
    "    \n",
    "    for k, _ranks in enumerate(ranks[1:]):\n",
    "        rank_field = 'rank{}'.format(k + 1)\n",
    "        index = k + 1\n",
    "        ranks[index].sort_values(by=['image_id', rank_field], inplace=True, kind='mergesort', ascending=False)\n",
    "#         r = ranks[index].groupby('image_id').head(topk)\n",
    "\n",
    "        r = ranks[index]\n",
    "\n",
    "        result = result.merge(r, how='outer', on=['image_id', 'target_id'])\n",
    "\n",
    "\n",
    "    rank_fields = ['rank' + str(k) for k in range(len(ranks))]\n",
    "\n",
    "    for rank_field in rank_fields:\n",
    "        result[rank_field] = result[rank_field].fillna(np.float32(result[rank_field].min() - 0.5))\n",
    "\n",
    "    rank_fields = ['rank' + str(k) for k in range(7)]\n",
    "\n",
    "    stat_df = result.groupby('image_id', as_index=False).agg({ c: ['mean', 'std'] for c in rank_fields})\n",
    "    stat_df.columns = [\"_\".join(x) for x in stat_df.columns.ravel()]\n",
    "    stat_df.rename(columns={ 'image_id_': 'image_id' }, inplace=True)\n",
    "\n",
    "    result = pd.merge(result, stat_df, on='image_id')\n",
    "\n",
    "    for rank_field in rank_fields:\n",
    "        result[rank_field + '_std'] = np.float32(result[rank_field + '_std'])\n",
    "        result[rank_field + '_var'] = np.float32(result[rank_field + '_mean'] / (result[rank_field + '_std'] + 1e-2))\n",
    "        result[rank_field + '_normed'] = np.float32((result[rank_field] - result[rank_field + '_mean']) / (result[rank_field + '_std'] + 1e-2))\n",
    "        \n",
    "    \n",
    "    y = pd.read_parquet(os.path.join(part, 'targets.parquet'))\n",
    "    image_ids = pd.read_parquet(os.path.join(part, 'image_ids.parquet'))\n",
    "    target_ids= pd.read_parquet(os.path.join(part, 'target_ids.parquet'))\n",
    "    \n",
    "    X = pd.concat([y, image_ids, target_ids], axis=1)\n",
    "    \n",
    "    \n",
    "    result = pd.merge(result, X, on=['image_id', 'target_id'])\n",
    "    \n",
    "        \n",
    "    all_results.append(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "01bf4ab2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:38.834919Z",
     "iopub.status.busy": "2021-12-09T10:08:38.834201Z",
     "iopub.status.idle": "2021-12-09T10:08:39.682207Z",
     "shell.execute_reply": "2021-12-09T10:08:39.682760Z",
     "shell.execute_reply.started": "2021-12-09T10:07:13.324073Z"
    },
    "papermill": {
     "duration": 0.886282,
     "end_time": "2021-12-09T10:08:39.682938",
     "exception": false,
     "start_time": "2021-12-09T10:08:38.796656",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "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>image_id</th>\n",
       "      <th>target_id</th>\n",
       "      <th>rank0</th>\n",
       "      <th>rank1</th>\n",
       "      <th>rank2</th>\n",
       "      <th>rank3</th>\n",
       "      <th>rank4</th>\n",
       "      <th>rank5</th>\n",
       "      <th>rank6</th>\n",
       "      <th>rank0_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>rank2_normed</th>\n",
       "      <th>rank3_var</th>\n",
       "      <th>rank3_normed</th>\n",
       "      <th>rank4_var</th>\n",
       "      <th>rank4_normed</th>\n",
       "      <th>rank5_var</th>\n",
       "      <th>rank5_normed</th>\n",
       "      <th>rank6_var</th>\n",
       "      <th>rank6_normed</th>\n",
       "      <th>MATCH</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>86262</th>\n",
       "      <td>68</td>\n",
       "      <td>137717</td>\n",
       "      <td>6.767020</td>\n",
       "      <td>7.589909</td>\n",
       "      <td>6.981566</td>\n",
       "      <td>8.064589</td>\n",
       "      <td>7.789616</td>\n",
       "      <td>7.245657</td>\n",
       "      <td>7.790066</td>\n",
       "      <td>-1.826752</td>\n",
       "      <td>...</td>\n",
       "      <td>2.304940</td>\n",
       "      <td>-0.217325</td>\n",
       "      <td>2.160819</td>\n",
       "      <td>-0.231773</td>\n",
       "      <td>2.001936</td>\n",
       "      <td>-0.216139</td>\n",
       "      <td>1.879396</td>\n",
       "      <td>-0.288423</td>\n",
       "      <td>2.098608</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86263</th>\n",
       "      <td>68</td>\n",
       "      <td>137719</td>\n",
       "      <td>5.563097</td>\n",
       "      <td>6.411661</td>\n",
       "      <td>5.403645</td>\n",
       "      <td>6.262428</td>\n",
       "      <td>7.076739</td>\n",
       "      <td>5.920494</td>\n",
       "      <td>6.193605</td>\n",
       "      <td>-1.826752</td>\n",
       "      <td>...</td>\n",
       "      <td>1.893191</td>\n",
       "      <td>-0.217325</td>\n",
       "      <td>1.726514</td>\n",
       "      <td>-0.231773</td>\n",
       "      <td>1.839937</td>\n",
       "      <td>-0.216139</td>\n",
       "      <td>1.575202</td>\n",
       "      <td>-0.288423</td>\n",
       "      <td>1.727637</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86264</th>\n",
       "      <td>68</td>\n",
       "      <td>137718</td>\n",
       "      <td>5.254476</td>\n",
       "      <td>6.209639</td>\n",
       "      <td>5.109680</td>\n",
       "      <td>6.214959</td>\n",
       "      <td>7.070467</td>\n",
       "      <td>5.503346</td>\n",
       "      <td>6.032679</td>\n",
       "      <td>-1.826752</td>\n",
       "      <td>...</td>\n",
       "      <td>1.816483</td>\n",
       "      <td>-0.217325</td>\n",
       "      <td>1.715074</td>\n",
       "      <td>-0.231773</td>\n",
       "      <td>1.838512</td>\n",
       "      <td>-0.216139</td>\n",
       "      <td>1.479445</td>\n",
       "      <td>-0.288423</td>\n",
       "      <td>1.690243</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86265</th>\n",
       "      <td>68</td>\n",
       "      <td>212117</td>\n",
       "      <td>5.243744</td>\n",
       "      <td>6.207840</td>\n",
       "      <td>5.616974</td>\n",
       "      <td>6.043822</td>\n",
       "      <td>6.644480</td>\n",
       "      <td>6.292782</td>\n",
       "      <td>6.555391</td>\n",
       "      <td>-1.826752</td>\n",
       "      <td>...</td>\n",
       "      <td>1.948858</td>\n",
       "      <td>-0.217325</td>\n",
       "      <td>1.673832</td>\n",
       "      <td>-0.231773</td>\n",
       "      <td>1.741708</td>\n",
       "      <td>-0.216139</td>\n",
       "      <td>1.660662</td>\n",
       "      <td>-0.288423</td>\n",
       "      <td>1.811706</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86266</th>\n",
       "      <td>68</td>\n",
       "      <td>137716</td>\n",
       "      <td>4.299229</td>\n",
       "      <td>5.391079</td>\n",
       "      <td>4.410022</td>\n",
       "      <td>5.383885</td>\n",
       "      <td>6.233411</td>\n",
       "      <td>5.200714</td>\n",
       "      <td>5.502605</td>\n",
       "      <td>-1.826752</td>\n",
       "      <td>...</td>\n",
       "      <td>1.633912</td>\n",
       "      <td>-0.217325</td>\n",
       "      <td>1.514793</td>\n",
       "      <td>-0.231773</td>\n",
       "      <td>1.648294</td>\n",
       "      <td>-0.216139</td>\n",
       "      <td>1.409975</td>\n",
       "      <td>-0.288423</td>\n",
       "      <td>1.567069</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>69464</td>\n",
       "      <td>244066</td>\n",
       "      <td>-6.730565</td>\n",
       "      <td>-6.215149</td>\n",
       "      <td>-6.592605</td>\n",
       "      <td>-5.714254</td>\n",
       "      <td>-5.980622</td>\n",
       "      <td>-5.797653</td>\n",
       "      <td>-2.003985</td>\n",
       "      <td>-3.807902</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.999150</td>\n",
       "      <td>-1.216273</td>\n",
       "      <td>-1.016111</td>\n",
       "      <td>-1.062039</td>\n",
       "      <td>-1.012258</td>\n",
       "      <td>-1.028829</td>\n",
       "      <td>-1.026801</td>\n",
       "      <td>-1.019766</td>\n",
       "      <td>0.341815</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>69464</td>\n",
       "      <td>292335</td>\n",
       "      <td>-6.730565</td>\n",
       "      <td>-6.215149</td>\n",
       "      <td>-6.592605</td>\n",
       "      <td>-5.714254</td>\n",
       "      <td>-5.980622</td>\n",
       "      <td>-5.797653</td>\n",
       "      <td>-2.078722</td>\n",
       "      <td>-3.807902</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.999150</td>\n",
       "      <td>-1.216273</td>\n",
       "      <td>-1.016111</td>\n",
       "      <td>-1.062039</td>\n",
       "      <td>-1.012258</td>\n",
       "      <td>-1.028829</td>\n",
       "      <td>-1.026801</td>\n",
       "      <td>-1.019766</td>\n",
       "      <td>0.316532</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>69464</td>\n",
       "      <td>5265</td>\n",
       "      <td>-6.730565</td>\n",
       "      <td>-6.215149</td>\n",
       "      <td>-6.592605</td>\n",
       "      <td>-5.714254</td>\n",
       "      <td>-5.980622</td>\n",
       "      <td>-5.797653</td>\n",
       "      <td>-2.084232</td>\n",
       "      <td>-3.807902</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.999150</td>\n",
       "      <td>-1.216273</td>\n",
       "      <td>-1.016111</td>\n",
       "      <td>-1.062039</td>\n",
       "      <td>-1.012258</td>\n",
       "      <td>-1.028829</td>\n",
       "      <td>-1.026801</td>\n",
       "      <td>-1.019766</td>\n",
       "      <td>0.314668</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>69464</td>\n",
       "      <td>186697</td>\n",
       "      <td>-6.730565</td>\n",
       "      <td>-6.215149</td>\n",
       "      <td>-6.592605</td>\n",
       "      <td>-5.714254</td>\n",
       "      <td>-5.980622</td>\n",
       "      <td>-5.797653</td>\n",
       "      <td>-2.191618</td>\n",
       "      <td>-3.807902</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.999150</td>\n",
       "      <td>-1.216273</td>\n",
       "      <td>-1.016111</td>\n",
       "      <td>-1.062039</td>\n",
       "      <td>-1.012258</td>\n",
       "      <td>-1.028829</td>\n",
       "      <td>-1.026801</td>\n",
       "      <td>-1.019766</td>\n",
       "      <td>0.278339</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>69464</td>\n",
       "      <td>37884</td>\n",
       "      <td>-6.730565</td>\n",
       "      <td>-6.215149</td>\n",
       "      <td>-6.592605</td>\n",
       "      <td>-5.714254</td>\n",
       "      <td>-5.980622</td>\n",
       "      <td>-5.797653</td>\n",
       "      <td>-2.194801</td>\n",
       "      <td>-3.807902</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.999150</td>\n",
       "      <td>-1.216273</td>\n",
       "      <td>-1.016111</td>\n",
       "      <td>-1.062039</td>\n",
       "      <td>-1.012258</td>\n",
       "      <td>-1.028829</td>\n",
       "      <td>-1.026801</td>\n",
       "      <td>-1.019766</td>\n",
       "      <td>0.277262</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>258991 rows × 38 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       image_id  target_id     rank0     rank1     rank2     rank3     rank4  \\\n",
       "86262        68     137717  6.767020  7.589909  6.981566  8.064589  7.789616   \n",
       "86263        68     137719  5.563097  6.411661  5.403645  6.262428  7.076739   \n",
       "86264        68     137718  5.254476  6.209639  5.109680  6.214959  7.070467   \n",
       "86265        68     212117  5.243744  6.207840  5.616974  6.043822  6.644480   \n",
       "86266        68     137716  4.299229  5.391079  4.410022  5.383885  6.233411   \n",
       "...         ...        ...       ...       ...       ...       ...       ...   \n",
       "84        69464     244066 -6.730565 -6.215149 -6.592605 -5.714254 -5.980622   \n",
       "85        69464     292335 -6.730565 -6.215149 -6.592605 -5.714254 -5.980622   \n",
       "86        69464       5265 -6.730565 -6.215149 -6.592605 -5.714254 -5.980622   \n",
       "87        69464     186697 -6.730565 -6.215149 -6.592605 -5.714254 -5.980622   \n",
       "88        69464      37884 -6.730565 -6.215149 -6.592605 -5.714254 -5.980622   \n",
       "\n",
       "          rank5     rank6  rank0_mean  ...  rank2_normed  rank3_var  \\\n",
       "86262  7.245657  7.790066   -1.826752  ...      2.304940  -0.217325   \n",
       "86263  5.920494  6.193605   -1.826752  ...      1.893191  -0.217325   \n",
       "86264  5.503346  6.032679   -1.826752  ...      1.816483  -0.217325   \n",
       "86265  6.292782  6.555391   -1.826752  ...      1.948858  -0.217325   \n",
       "86266  5.200714  5.502605   -1.826752  ...      1.633912  -0.217325   \n",
       "...         ...       ...         ...  ...           ...        ...   \n",
       "84    -5.797653 -2.003985   -3.807902  ...     -0.999150  -1.216273   \n",
       "85    -5.797653 -2.078722   -3.807902  ...     -0.999150  -1.216273   \n",
       "86    -5.797653 -2.084232   -3.807902  ...     -0.999150  -1.216273   \n",
       "87    -5.797653 -2.191618   -3.807902  ...     -0.999150  -1.216273   \n",
       "88    -5.797653 -2.194801   -3.807902  ...     -0.999150  -1.216273   \n",
       "\n",
       "       rank3_normed  rank4_var  rank4_normed  rank5_var  rank5_normed  \\\n",
       "86262      2.160819  -0.231773      2.001936  -0.216139      1.879396   \n",
       "86263      1.726514  -0.231773      1.839937  -0.216139      1.575202   \n",
       "86264      1.715074  -0.231773      1.838512  -0.216139      1.479445   \n",
       "86265      1.673832  -0.231773      1.741708  -0.216139      1.660662   \n",
       "86266      1.514793  -0.231773      1.648294  -0.216139      1.409975   \n",
       "...             ...        ...           ...        ...           ...   \n",
       "84        -1.016111  -1.062039     -1.012258  -1.028829     -1.026801   \n",
       "85        -1.016111  -1.062039     -1.012258  -1.028829     -1.026801   \n",
       "86        -1.016111  -1.062039     -1.012258  -1.028829     -1.026801   \n",
       "87        -1.016111  -1.062039     -1.012258  -1.028829     -1.026801   \n",
       "88        -1.016111  -1.062039     -1.012258  -1.028829     -1.026801   \n",
       "\n",
       "       rank6_var  rank6_normed  MATCH  \n",
       "86262  -0.288423      2.098608      1  \n",
       "86263  -0.288423      1.727637      1  \n",
       "86264  -0.288423      1.690243      1  \n",
       "86265  -0.288423      1.811706      0  \n",
       "86266  -0.288423      1.567069      1  \n",
       "...          ...           ...    ...  \n",
       "84     -1.019766      0.341815      1  \n",
       "85     -1.019766      0.316532      0  \n",
       "86     -1.019766      0.314668      0  \n",
       "87     -1.019766      0.278339      0  \n",
       "88     -1.019766      0.277262      0  \n",
       "\n",
       "[258991 rows x 38 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_result = pd.concat(all_results)\n",
    "\n",
    "total_result.sort_values(by=['image_id'], inplace=True, kind='mergesort', ascending=True)\n",
    "\n",
    "total_result.to_parquet('train.parquet')\n",
    "\n",
    "total_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f557a9af",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:39.766156Z",
     "iopub.status.busy": "2021-12-09T10:08:39.765114Z",
     "iopub.status.idle": "2021-12-09T10:08:44.269586Z",
     "shell.execute_reply": "2021-12-09T10:08:44.268540Z",
     "shell.execute_reply.started": "2021-12-09T10:07:14.179986Z"
    },
    "papermill": {
     "duration": 4.552734,
     "end_time": "2021-12-09T10:08:44.269742",
     "exception": false,
     "start_time": "2021-12-09T10:08:39.717008",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:44: FutureWarning: Index.ravel returning ndarray is deprecated; in a future version this will return a view on self.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "all_results = []\n",
    "\n",
    "parts = ['../input/data-for-model-06/train-09']\n",
    "\n",
    "for N, part in enumerate(parts):\n",
    "    \n",
    "    N = 3\n",
    "    \n",
    "    ranks = []\n",
    "\n",
    "    for k in range(7):\n",
    "        _ = pd.read_csv('../input/train-ranks-for-model-0{0}/short-ranks-{1}.csv'.format(k, N))\n",
    "        _.drop(columns=['image_url', 'hyperref', 'target', 'count', 'undigit_filename'], inplace=True)\n",
    "        _.rename(columns={'rank': 'rank' + str(k)}, inplace=True)\n",
    "    \n",
    "        ranks.append(_)\n",
    "        \n",
    "    topk = 100\n",
    "\n",
    "    ranks[0].sort_values(by=['image_id', 'rank0'], inplace=True, kind='mergesort', ascending=False)\n",
    "    result = ranks[0]#.groupby('image_id').head(topk)\n",
    "    \n",
    "    for k, _ranks in enumerate(ranks[1:]):\n",
    "        rank_field = 'rank{}'.format(k + 1)\n",
    "        index = k + 1\n",
    "        ranks[index].sort_values(by=['image_id', rank_field], inplace=True, kind='mergesort', ascending=False)\n",
    "#         r = ranks[index].groupby('image_id').head(topk)\n",
    "\n",
    "        r = ranks[index]\n",
    "\n",
    "        result = result.merge(r, how='outer', on=['image_id', 'target_id'])\n",
    "\n",
    "\n",
    "    rank_fields = ['rank' + str(k) for k in range(len(ranks))]\n",
    "\n",
    "    for rank_field in rank_fields:\n",
    "        result[rank_field] = result[rank_field].fillna(np.float32(result[rank_field].min() - 0.5))\n",
    "\n",
    "    rank_fields = ['rank' + str(k) for k in range(7)]\n",
    "\n",
    "    stat_df = result.groupby('image_id', as_index=False).agg({ c: ['mean', 'std'] for c in rank_fields})\n",
    "    stat_df.columns = [\"_\".join(x) for x in stat_df.columns.ravel()]\n",
    "    stat_df.rename(columns={ 'image_id_': 'image_id' }, inplace=True)\n",
    "\n",
    "    result = pd.merge(result, stat_df, on='image_id')\n",
    "\n",
    "    for rank_field in rank_fields:\n",
    "        result[rank_field + '_std'] = np.float32(result[rank_field + '_std'])\n",
    "        result[rank_field + '_var'] = np.float32(result[rank_field + '_mean'] / (result[rank_field + '_std'] + 1e-2))\n",
    "        result[rank_field + '_normed'] = np.float32((result[rank_field] - result[rank_field + '_mean']) / (result[rank_field + '_std'] + 1e-2))\n",
    "        \n",
    "    \n",
    "    y = pd.read_parquet(os.path.join(part, 'targets.parquet'))\n",
    "    image_ids = pd.read_parquet(os.path.join(part, 'image_ids.parquet'))\n",
    "    target_ids= pd.read_parquet(os.path.join(part, 'target_ids.parquet'))\n",
    "    \n",
    "    X = pd.concat([y, image_ids, target_ids], axis=1)\n",
    "    \n",
    "    \n",
    "    result = pd.merge(result, X, on=['image_id', 'target_id'])\n",
    "    \n",
    "        \n",
    "    all_results.append(result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "9bf0bf04",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:44.345298Z",
     "iopub.status.busy": "2021-12-09T10:08:44.344651Z",
     "iopub.status.idle": "2021-12-09T10:08:44.569775Z",
     "shell.execute_reply": "2021-12-09T10:08:44.570266Z",
     "shell.execute_reply.started": "2021-12-09T10:07:19.193757Z"
    },
    "papermill": {
     "duration": 0.2653,
     "end_time": "2021-12-09T10:08:44.570448",
     "exception": false,
     "start_time": "2021-12-09T10:08:44.305148",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "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>image_id</th>\n",
       "      <th>target_id</th>\n",
       "      <th>rank0</th>\n",
       "      <th>rank1</th>\n",
       "      <th>rank2</th>\n",
       "      <th>rank3</th>\n",
       "      <th>rank4</th>\n",
       "      <th>rank5</th>\n",
       "      <th>rank6</th>\n",
       "      <th>rank0_mean</th>\n",
       "      <th>...</th>\n",
       "      <th>rank2_normed</th>\n",
       "      <th>rank3_var</th>\n",
       "      <th>rank3_normed</th>\n",
       "      <th>rank4_var</th>\n",
       "      <th>rank4_normed</th>\n",
       "      <th>rank5_var</th>\n",
       "      <th>rank5_normed</th>\n",
       "      <th>rank6_var</th>\n",
       "      <th>rank6_normed</th>\n",
       "      <th>MATCH</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>86469</th>\n",
       "      <td>52</td>\n",
       "      <td>233063</td>\n",
       "      <td>7.178525</td>\n",
       "      <td>6.538880</td>\n",
       "      <td>6.433822</td>\n",
       "      <td>6.433228</td>\n",
       "      <td>7.546990</td>\n",
       "      <td>7.375566</td>\n",
       "      <td>6.843082</td>\n",
       "      <td>-2.993165</td>\n",
       "      <td>...</td>\n",
       "      <td>2.904336</td>\n",
       "      <td>-0.535776</td>\n",
       "      <td>2.401313</td>\n",
       "      <td>-0.563619</td>\n",
       "      <td>2.646683</td>\n",
       "      <td>-0.583808</td>\n",
       "      <td>2.562964</td>\n",
       "      <td>-0.594538</td>\n",
       "      <td>2.352816</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86470</th>\n",
       "      <td>52</td>\n",
       "      <td>233064</td>\n",
       "      <td>5.607776</td>\n",
       "      <td>5.350560</td>\n",
       "      <td>4.994696</td>\n",
       "      <td>4.913579</td>\n",
       "      <td>6.568535</td>\n",
       "      <td>5.669404</td>\n",
       "      <td>5.264292</td>\n",
       "      <td>-2.993165</td>\n",
       "      <td>...</td>\n",
       "      <td>2.459852</td>\n",
       "      <td>-0.535776</td>\n",
       "      <td>1.960638</td>\n",
       "      <td>-0.563619</td>\n",
       "      <td>2.376617</td>\n",
       "      <td>-0.583808</td>\n",
       "      <td>2.105134</td>\n",
       "      <td>-0.594538</td>\n",
       "      <td>1.947158</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86471</th>\n",
       "      <td>52</td>\n",
       "      <td>233065</td>\n",
       "      <td>4.395752</td>\n",
       "      <td>5.107599</td>\n",
       "      <td>4.264780</td>\n",
       "      <td>5.157209</td>\n",
       "      <td>5.308811</td>\n",
       "      <td>5.173628</td>\n",
       "      <td>5.063816</td>\n",
       "      <td>-2.993165</td>\n",
       "      <td>...</td>\n",
       "      <td>2.234413</td>\n",
       "      <td>-0.535776</td>\n",
       "      <td>2.031287</td>\n",
       "      <td>-0.563619</td>\n",
       "      <td>2.028918</td>\n",
       "      <td>-0.583808</td>\n",
       "      <td>1.972097</td>\n",
       "      <td>-0.594538</td>\n",
       "      <td>1.895647</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86472</th>\n",
       "      <td>52</td>\n",
       "      <td>127163</td>\n",
       "      <td>2.353567</td>\n",
       "      <td>3.259148</td>\n",
       "      <td>2.255012</td>\n",
       "      <td>3.231598</td>\n",
       "      <td>3.326448</td>\n",
       "      <td>3.246206</td>\n",
       "      <td>3.551573</td>\n",
       "      <td>-2.993165</td>\n",
       "      <td>...</td>\n",
       "      <td>1.613683</td>\n",
       "      <td>-0.535776</td>\n",
       "      <td>1.472890</td>\n",
       "      <td>-0.563619</td>\n",
       "      <td>1.481761</td>\n",
       "      <td>-0.583808</td>\n",
       "      <td>1.454894</td>\n",
       "      <td>-0.594538</td>\n",
       "      <td>1.507087</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86473</th>\n",
       "      <td>52</td>\n",
       "      <td>152990</td>\n",
       "      <td>1.777707</td>\n",
       "      <td>2.624292</td>\n",
       "      <td>0.927192</td>\n",
       "      <td>2.583427</td>\n",
       "      <td>3.171491</td>\n",
       "      <td>2.743405</td>\n",
       "      <td>2.506178</td>\n",
       "      <td>-2.993165</td>\n",
       "      <td>...</td>\n",
       "      <td>1.203577</td>\n",
       "      <td>-0.535776</td>\n",
       "      <td>1.284930</td>\n",
       "      <td>-0.563619</td>\n",
       "      <td>1.438991</td>\n",
       "      <td>-0.583808</td>\n",
       "      <td>1.319972</td>\n",
       "      <td>-0.594538</td>\n",
       "      <td>1.238481</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>69458</td>\n",
       "      <td>269097</td>\n",
       "      <td>-6.605800</td>\n",
       "      <td>-6.353398</td>\n",
       "      <td>-6.282269</td>\n",
       "      <td>-5.450472</td>\n",
       "      <td>-5.780550</td>\n",
       "      <td>-6.079898</td>\n",
       "      <td>0.146490</td>\n",
       "      <td>-3.142374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.996508</td>\n",
       "      <td>-0.702036</td>\n",
       "      <td>-0.983249</td>\n",
       "      <td>-0.675824</td>\n",
       "      <td>-1.010808</td>\n",
       "      <td>-0.659865</td>\n",
       "      <td>-1.029812</td>\n",
       "      <td>-0.697645</td>\n",
       "      <td>0.736987</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>69458</td>\n",
       "      <td>269098</td>\n",
       "      <td>-6.605800</td>\n",
       "      <td>-6.353398</td>\n",
       "      <td>-6.282269</td>\n",
       "      <td>-5.450472</td>\n",
       "      <td>-5.780550</td>\n",
       "      <td>-6.079898</td>\n",
       "      <td>0.146490</td>\n",
       "      <td>-3.142374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.996508</td>\n",
       "      <td>-0.702036</td>\n",
       "      <td>-0.983249</td>\n",
       "      <td>-0.675824</td>\n",
       "      <td>-1.010808</td>\n",
       "      <td>-0.659865</td>\n",
       "      <td>-1.029812</td>\n",
       "      <td>-0.697645</td>\n",
       "      <td>0.736987</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>69458</td>\n",
       "      <td>269100</td>\n",
       "      <td>-6.605800</td>\n",
       "      <td>-6.353398</td>\n",
       "      <td>-6.282269</td>\n",
       "      <td>-5.450472</td>\n",
       "      <td>-5.780550</td>\n",
       "      <td>-6.079898</td>\n",
       "      <td>-0.787915</td>\n",
       "      <td>-3.142374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.996508</td>\n",
       "      <td>-0.702036</td>\n",
       "      <td>-0.983249</td>\n",
       "      <td>-0.675824</td>\n",
       "      <td>-1.010808</td>\n",
       "      <td>-0.659865</td>\n",
       "      <td>-1.029812</td>\n",
       "      <td>-0.697645</td>\n",
       "      <td>0.486037</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>69458</td>\n",
       "      <td>247179</td>\n",
       "      <td>-6.605800</td>\n",
       "      <td>-6.353398</td>\n",
       "      <td>-6.282269</td>\n",
       "      <td>-5.450472</td>\n",
       "      <td>-5.780550</td>\n",
       "      <td>-6.079898</td>\n",
       "      <td>-0.823042</td>\n",
       "      <td>-3.142374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.996508</td>\n",
       "      <td>-0.702036</td>\n",
       "      <td>-0.983249</td>\n",
       "      <td>-0.675824</td>\n",
       "      <td>-1.010808</td>\n",
       "      <td>-0.659865</td>\n",
       "      <td>-1.029812</td>\n",
       "      <td>-0.697645</td>\n",
       "      <td>0.476603</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>69458</td>\n",
       "      <td>23299</td>\n",
       "      <td>-6.605800</td>\n",
       "      <td>-6.353398</td>\n",
       "      <td>-6.282269</td>\n",
       "      <td>-5.450472</td>\n",
       "      <td>-5.780550</td>\n",
       "      <td>-6.079898</td>\n",
       "      <td>-0.828115</td>\n",
       "      <td>-3.142374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.996508</td>\n",
       "      <td>-0.702036</td>\n",
       "      <td>-0.983249</td>\n",
       "      <td>-0.675824</td>\n",
       "      <td>-1.010808</td>\n",
       "      <td>-0.659865</td>\n",
       "      <td>-1.029812</td>\n",
       "      <td>-0.697645</td>\n",
       "      <td>0.475241</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>86559 rows × 38 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       image_id  target_id     rank0     rank1     rank2     rank3     rank4  \\\n",
       "86469        52     233063  7.178525  6.538880  6.433822  6.433228  7.546990   \n",
       "86470        52     233064  5.607776  5.350560  4.994696  4.913579  6.568535   \n",
       "86471        52     233065  4.395752  5.107599  4.264780  5.157209  5.308811   \n",
       "86472        52     127163  2.353567  3.259148  2.255012  3.231598  3.326448   \n",
       "86473        52     152990  1.777707  2.624292  0.927192  2.583427  3.171491   \n",
       "...         ...        ...       ...       ...       ...       ...       ...   \n",
       "85        69458     269097 -6.605800 -6.353398 -6.282269 -5.450472 -5.780550   \n",
       "86        69458     269098 -6.605800 -6.353398 -6.282269 -5.450472 -5.780550   \n",
       "87        69458     269100 -6.605800 -6.353398 -6.282269 -5.450472 -5.780550   \n",
       "88        69458     247179 -6.605800 -6.353398 -6.282269 -5.450472 -5.780550   \n",
       "89        69458      23299 -6.605800 -6.353398 -6.282269 -5.450472 -5.780550   \n",
       "\n",
       "          rank5     rank6  rank0_mean  ...  rank2_normed  rank3_var  \\\n",
       "86469  7.375566  6.843082   -2.993165  ...      2.904336  -0.535776   \n",
       "86470  5.669404  5.264292   -2.993165  ...      2.459852  -0.535776   \n",
       "86471  5.173628  5.063816   -2.993165  ...      2.234413  -0.535776   \n",
       "86472  3.246206  3.551573   -2.993165  ...      1.613683  -0.535776   \n",
       "86473  2.743405  2.506178   -2.993165  ...      1.203577  -0.535776   \n",
       "...         ...       ...         ...  ...           ...        ...   \n",
       "85    -6.079898  0.146490   -3.142374  ...     -0.996508  -0.702036   \n",
       "86    -6.079898  0.146490   -3.142374  ...     -0.996508  -0.702036   \n",
       "87    -6.079898 -0.787915   -3.142374  ...     -0.996508  -0.702036   \n",
       "88    -6.079898 -0.823042   -3.142374  ...     -0.996508  -0.702036   \n",
       "89    -6.079898 -0.828115   -3.142374  ...     -0.996508  -0.702036   \n",
       "\n",
       "       rank3_normed  rank4_var  rank4_normed  rank5_var  rank5_normed  \\\n",
       "86469      2.401313  -0.563619      2.646683  -0.583808      2.562964   \n",
       "86470      1.960638  -0.563619      2.376617  -0.583808      2.105134   \n",
       "86471      2.031287  -0.563619      2.028918  -0.583808      1.972097   \n",
       "86472      1.472890  -0.563619      1.481761  -0.583808      1.454894   \n",
       "86473      1.284930  -0.563619      1.438991  -0.583808      1.319972   \n",
       "...             ...        ...           ...        ...           ...   \n",
       "85        -0.983249  -0.675824     -1.010808  -0.659865     -1.029812   \n",
       "86        -0.983249  -0.675824     -1.010808  -0.659865     -1.029812   \n",
       "87        -0.983249  -0.675824     -1.010808  -0.659865     -1.029812   \n",
       "88        -0.983249  -0.675824     -1.010808  -0.659865     -1.029812   \n",
       "89        -0.983249  -0.675824     -1.010808  -0.659865     -1.029812   \n",
       "\n",
       "       rank6_var  rank6_normed  MATCH  \n",
       "86469  -0.594538      2.352816      1  \n",
       "86470  -0.594538      1.947158      1  \n",
       "86471  -0.594538      1.895647      1  \n",
       "86472  -0.594538      1.507087      0  \n",
       "86473  -0.594538      1.238481      0  \n",
       "...          ...           ...    ...  \n",
       "85     -0.697645      0.736987      0  \n",
       "86     -0.697645      0.736987      0  \n",
       "87     -0.697645      0.486037      0  \n",
       "88     -0.697645      0.476603      0  \n",
       "89     -0.697645      0.475241      0  \n",
       "\n",
       "[86559 rows x 38 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_result = pd.concat(all_results)\n",
    "\n",
    "total_result.sort_values(by=['image_id'], inplace=True, kind='mergesort', ascending=True)\n",
    "\n",
    "total_result.to_parquet('valid.parquet')\n",
    "\n",
    "total_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "72839db4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-12-09T10:08:44.647092Z",
     "iopub.status.busy": "2021-12-09T10:08:44.646235Z",
     "iopub.status.idle": "2021-12-09T10:08:44.649571Z",
     "shell.execute_reply": "2021-12-09T10:08:44.650016Z",
     "shell.execute_reply.started": "2021-12-09T10:07:19.446473Z"
    },
    "papermill": {
     "duration": 0.046182,
     "end_time": "2021-12-09T10:08:44.650179",
     "exception": false,
     "start_time": "2021-12-09T10:08:44.603997",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['image_id',\n",
       " 'target_id',\n",
       " 'rank0',\n",
       " 'rank1',\n",
       " 'rank2',\n",
       " 'rank3',\n",
       " 'rank4',\n",
       " 'rank5',\n",
       " 'rank6',\n",
       " 'rank0_mean',\n",
       " 'rank0_std',\n",
       " 'rank1_mean',\n",
       " 'rank1_std',\n",
       " 'rank2_mean',\n",
       " 'rank2_std',\n",
       " 'rank3_mean',\n",
       " 'rank3_std',\n",
       " 'rank4_mean',\n",
       " 'rank4_std',\n",
       " 'rank5_mean',\n",
       " 'rank5_std',\n",
       " 'rank6_mean',\n",
       " 'rank6_std',\n",
       " 'rank0_var',\n",
       " 'rank0_normed',\n",
       " 'rank1_var',\n",
       " 'rank1_normed',\n",
       " 'rank2_var',\n",
       " 'rank2_normed',\n",
       " 'rank3_var',\n",
       " 'rank3_normed',\n",
       " 'rank4_var',\n",
       " 'rank4_normed',\n",
       " 'rank5_var',\n",
       " 'rank5_normed',\n",
       " 'rank6_var',\n",
       " 'rank6_normed',\n",
       " 'MATCH']"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_result.columns.tolist()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 42.177932,
   "end_time": "2021-12-09T10:08:47.532947",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2021-12-09T10:08:05.355015",
   "version": "2.3.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}