{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8f9c5803",
   "metadata": {},
   "source": [
    "# 调优自动特征工程\n",
    "在`自动特征工程`中,我们的df字段类型都是由woodwork自动推断的，几乎是完全自动化的过程。 \n",
    "\n",
    "我们需要对字段类型引入更多的人为设置\n",
    "- 时间序列\n",
    "- 自定义原语\n",
    "- 类型纠正"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1603941",
   "metadata": {},
   "source": [
    "## 导入"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "53cfceb5",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\63517\\miniconda3\\envs\\data-analysis\\lib\\site-packages\\woodwork\\__init__.py:2: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81.\n",
      "  import pkg_resources\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ft: 1.31.0,  ww: 0.31.0\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import featuretools as ft\n",
    "import woodwork as ww\n",
    "from woodwork.column_schema import ColumnSchema\n",
    "from woodwork.logical_types import NaturalLanguage, Datetime,Boolean\n",
    "from featuretools.primitives import AggregationPrimitive, TransformPrimitive\n",
    "from featuretools.tests.testing_utils import make_ecommerce_entityset\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "import gc\n",
    "gc.enable()\n",
    "\n",
    "print(f'ft: {ft.__version__},  ww: {ww.__version__}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8d5ae2c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "application_train = pd.read_csv('data/application_train.csv')\n",
    "application_test = pd.read_csv('data/application_test.csv')\n",
    "bureau = pd.read_csv('data/bureau.csv')\n",
    "bureau_balance = pd.read_csv('data/bureau_balance.csv')\n",
    "credit_card_balance = pd.read_csv('data/credit_card_balance.csv')\n",
    "installments_payments = pd.read_csv('data/installments_payments.csv')\n",
    "previous_application = pd.read_csv('data/previous_application.csv')\n",
    "pos_cash_balance = pd.read_csv('data/POS_CASH_balance.csv')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa37c3c1",
   "metadata": {},
   "source": [
    "为了验证我们的处理过程，必须先抽样一些。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "89412ef7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# application_train = application_train.iloc[:1000, :]\n",
    "# application_test = application_test.iloc[:1000, :]\n",
    "# bureau = bureau.iloc[:1000, :]\n",
    "# bureau_balance = bureau_balance.iloc[:1000, :]\n",
    "# credit_card_balance = credit_card_balance.iloc[:1000, :]\n",
    "# installments_payments = installments_payments.iloc[:1000, :]\n",
    "# previous_application = previous_application.iloc[:1000, :]\n",
    "# pos_cash_balance = pos_cash_balance.iloc[:1000, :]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "753cff39",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(307511, 123) (48744, 123)\n"
     ]
    }
   ],
   "source": [
    "application_train['set'] = 'train'\n",
    "application_test['set'] = 'test'\n",
    "application_test['TARGET'] = np.nan\n",
    "print(application_train.shape, application_test.shape)\n",
    "app = pd.concat([application_train, application_test], ignore_index=True)\n",
    "\n",
    "app_target = app[['SK_ID_CURR', 'TARGET']]\n",
    "app_set = app[['SK_ID_CURR', 'set']]\n",
    "app = app.drop(columns=['set'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6f6df720",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([dtype('int64'), dtype('O'), dtype('float64')], dtype=object)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "application_train.dtypes.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5f862c3b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<module 'woodwork.logical_types' from 'c:\\\\Users\\\\63517\\\\miniconda3\\\\envs\\\\data-analysis\\\\lib\\\\site-packages\\\\woodwork\\\\logical_types.py'>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ww.logical_types"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da4e6346",
   "metadata": {},
   "source": [
    "## woodwork认识：\n",
    "- 物理类型\n",
    "- 逻辑类型：\n",
    "- 语义标签：额外数据含义\n",
    "\n",
    "逻辑类型是必须的， 语义标签是可选的"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0f67ff7",
   "metadata": {},
   "source": [
    "woodwork使用了Pandas 的 Accessor机制，这是扩展接口，在`import featuretools`时候，就把ww加上去了"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f58f07b0",
   "metadata": {},
   "source": [
    "woodwork初始时候会为其添加逻辑类型，语义标签"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "99ee10d5",
   "metadata": {},
   "source": [
    "### 语义标签"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7e483877",
   "metadata": {},
   "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>name</th>\n",
       "      <th>is_standard_tag</th>\n",
       "      <th>valid_logical_types</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>numeric</td>\n",
       "      <td>True</td>\n",
       "      <td>[Age, AgeFractional, AgeNullable, Double, Inte...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>category</td>\n",
       "      <td>True</td>\n",
       "      <td>[Categorical, CountryCode, CurrencyCode, Ordin...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>index</td>\n",
       "      <td>False</td>\n",
       "      <td>Any LogicalType</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>time_index</td>\n",
       "      <td>False</td>\n",
       "      <td>[Datetime, Age, AgeFractional, AgeNullable, Do...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>date_of_birth</td>\n",
       "      <td>False</td>\n",
       "      <td>[Datetime]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ignore</td>\n",
       "      <td>False</td>\n",
       "      <td>Any LogicalType</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>passthrough</td>\n",
       "      <td>False</td>\n",
       "      <td>Any LogicalType</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            name  is_standard_tag  \\\n",
       "0        numeric             True   \n",
       "1       category             True   \n",
       "2          index            False   \n",
       "3     time_index            False   \n",
       "4  date_of_birth            False   \n",
       "5         ignore            False   \n",
       "6    passthrough            False   \n",
       "\n",
       "                                 valid_logical_types  \n",
       "0  [Age, AgeFractional, AgeNullable, Double, Inte...  \n",
       "1  [Categorical, CountryCode, CurrencyCode, Ordin...  \n",
       "2                                    Any LogicalType  \n",
       "3  [Datetime, Age, AgeFractional, AgeNullable, Do...  \n",
       "4                                         [Datetime]  \n",
       "5                                    Any LogicalType  \n",
       "6                                    Any LogicalType  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ww.list_semantic_tags()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32346c1a",
   "metadata": {},
   "source": [
    "- `numeric`, `category` 标准语义标签和特定的逻辑类型关联\n",
    "- `index`,`time_index` woodwork为一些索引列添加标签，表明一些含义\n",
    "- `date_of_birth` 表明应该解释为出生日期\n",
    "- `ignore`,`passthrough` 应该被忽略，在ft过程中"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0dd57169",
   "metadata": {},
   "source": [
    "我们应该添加额外标签帮助解释"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fc026c70",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "63adac94",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "bc7408b2",
   "metadata": {},
   "source": [
    "### 逻辑类型"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5198407",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "043232ce",
   "metadata": {},
   "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>name</th>\n",
       "      <th>type_string</th>\n",
       "      <th>description</th>\n",
       "      <th>physical_type</th>\n",
       "      <th>standard_tags</th>\n",
       "      <th>is_default_type</th>\n",
       "      <th>is_registered</th>\n",
       "      <th>parent_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Address</td>\n",
       "      <td>address</td>\n",
       "      <td>Represents Logical Types that contain address ...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Age</td>\n",
       "      <td>age</td>\n",
       "      <td>Represents Logical Types that contain whole nu...</td>\n",
       "      <td>int64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Integer</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AgeFractional</td>\n",
       "      <td>age_fractional</td>\n",
       "      <td>Represents Logical Types that contain non-nega...</td>\n",
       "      <td>float64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Double</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AgeNullable</td>\n",
       "      <td>age_nullable</td>\n",
       "      <td>Represents Logical Types that contain whole nu...</td>\n",
       "      <td>Int64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>IntegerNullable</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Boolean</td>\n",
       "      <td>boolean</td>\n",
       "      <td>Represents Logical Types that contain binary v...</td>\n",
       "      <td>bool</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>BooleanNullable</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>BooleanNullable</td>\n",
       "      <td>boolean_nullable</td>\n",
       "      <td>Represents Logical Types that contain binary v...</td>\n",
       "      <td>boolean</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>categorical</td>\n",
       "      <td>Represents Logical Types that contain unordere...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>CountryCode</td>\n",
       "      <td>country_code</td>\n",
       "      <td>Represents Logical Types that use the ISO-3166...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Categorical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>CurrencyCode</td>\n",
       "      <td>currency_code</td>\n",
       "      <td>Represents Logical Types that use the ISO-4217...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Categorical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Datetime</td>\n",
       "      <td>datetime</td>\n",
       "      <td>Represents Logical Types that contain date and...</td>\n",
       "      <td>datetime64[ns]</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Double</td>\n",
       "      <td>double</td>\n",
       "      <td>Represents Logical Types that contain positive...</td>\n",
       "      <td>float64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>EmailAddress</td>\n",
       "      <td>email_address</td>\n",
       "      <td>Represents Logical Types that contain email ad...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Filepath</td>\n",
       "      <td>filepath</td>\n",
       "      <td>Represents Logical Types that specify location...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>IPAddress</td>\n",
       "      <td>ip_address</td>\n",
       "      <td>Represents Logical Types that contain IP addre...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Integer</td>\n",
       "      <td>integer</td>\n",
       "      <td>Represents Logical Types that contain positive...</td>\n",
       "      <td>int64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>IntegerNullable</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>integer_nullable</td>\n",
       "      <td>Represents Logical Types that contain positive...</td>\n",
       "      <td>Int64</td>\n",
       "      <td>{numeric}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>LatLong</td>\n",
       "      <td>lat_long</td>\n",
       "      <td>Represents Logical Types that contain latitude...</td>\n",
       "      <td>object</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>NaturalLanguage</td>\n",
       "      <td>natural_language</td>\n",
       "      <td>Represents Logical Types that contain text or ...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Ordinal</td>\n",
       "      <td>ordinal</td>\n",
       "      <td>Represents Logical Types that contain ordered ...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Categorical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>PersonFullName</td>\n",
       "      <td>person_full_name</td>\n",
       "      <td>Represents Logical Types that may contain firs...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>PhoneNumber</td>\n",
       "      <td>phone_number</td>\n",
       "      <td>Represents Logical Types that contain numeric ...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>PostalCode</td>\n",
       "      <td>postal_code</td>\n",
       "      <td>Represents Logical Types that contain a series...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Categorical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>SubRegionCode</td>\n",
       "      <td>sub_region_code</td>\n",
       "      <td>Represents Logical Types that use the ISO-3166...</td>\n",
       "      <td>category</td>\n",
       "      <td>{category}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Categorical</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>Timedelta</td>\n",
       "      <td>timedelta</td>\n",
       "      <td>Represents Logical Types that contain values s...</td>\n",
       "      <td>timedelta64[ns]</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>URL</td>\n",
       "      <td>url</td>\n",
       "      <td>Represents Logical Types that contain URLs, wh...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>Unknown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>Unknown</td>\n",
       "      <td>unknown</td>\n",
       "      <td>Represents Logical Types that cannot be inferr...</td>\n",
       "      <td>string</td>\n",
       "      <td>{}</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               name       type_string  \\\n",
       "0           Address           address   \n",
       "1               Age               age   \n",
       "2     AgeFractional    age_fractional   \n",
       "3       AgeNullable      age_nullable   \n",
       "4           Boolean           boolean   \n",
       "5   BooleanNullable  boolean_nullable   \n",
       "6       Categorical       categorical   \n",
       "7       CountryCode      country_code   \n",
       "8      CurrencyCode     currency_code   \n",
       "9          Datetime          datetime   \n",
       "10           Double            double   \n",
       "11     EmailAddress     email_address   \n",
       "12         Filepath          filepath   \n",
       "13        IPAddress        ip_address   \n",
       "14          Integer           integer   \n",
       "15  IntegerNullable  integer_nullable   \n",
       "16          LatLong          lat_long   \n",
       "17  NaturalLanguage  natural_language   \n",
       "18          Ordinal           ordinal   \n",
       "19   PersonFullName  person_full_name   \n",
       "20      PhoneNumber      phone_number   \n",
       "21       PostalCode       postal_code   \n",
       "22    SubRegionCode   sub_region_code   \n",
       "23        Timedelta         timedelta   \n",
       "24              URL               url   \n",
       "25          Unknown           unknown   \n",
       "\n",
       "                                          description    physical_type  \\\n",
       "0   Represents Logical Types that contain address ...           string   \n",
       "1   Represents Logical Types that contain whole nu...            int64   \n",
       "2   Represents Logical Types that contain non-nega...          float64   \n",
       "3   Represents Logical Types that contain whole nu...            Int64   \n",
       "4   Represents Logical Types that contain binary v...             bool   \n",
       "5   Represents Logical Types that contain binary v...          boolean   \n",
       "6   Represents Logical Types that contain unordere...         category   \n",
       "7   Represents Logical Types that use the ISO-3166...         category   \n",
       "8   Represents Logical Types that use the ISO-4217...         category   \n",
       "9   Represents Logical Types that contain date and...   datetime64[ns]   \n",
       "10  Represents Logical Types that contain positive...          float64   \n",
       "11  Represents Logical Types that contain email ad...           string   \n",
       "12  Represents Logical Types that specify location...           string   \n",
       "13  Represents Logical Types that contain IP addre...           string   \n",
       "14  Represents Logical Types that contain positive...            int64   \n",
       "15  Represents Logical Types that contain positive...            Int64   \n",
       "16  Represents Logical Types that contain latitude...           object   \n",
       "17  Represents Logical Types that contain text or ...           string   \n",
       "18  Represents Logical Types that contain ordered ...         category   \n",
       "19  Represents Logical Types that may contain firs...           string   \n",
       "20  Represents Logical Types that contain numeric ...           string   \n",
       "21  Represents Logical Types that contain a series...         category   \n",
       "22  Represents Logical Types that use the ISO-3166...         category   \n",
       "23  Represents Logical Types that contain values s...  timedelta64[ns]   \n",
       "24  Represents Logical Types that contain URLs, wh...           string   \n",
       "25  Represents Logical Types that cannot be inferr...           string   \n",
       "\n",
       "   standard_tags  is_default_type  is_registered      parent_type  \n",
       "0             {}             True           True             None  \n",
       "1      {numeric}             True           True          Integer  \n",
       "2      {numeric}             True           True           Double  \n",
       "3      {numeric}             True           True  IntegerNullable  \n",
       "4             {}             True           True  BooleanNullable  \n",
       "5             {}             True           True             None  \n",
       "6     {category}             True           True             None  \n",
       "7     {category}             True           True      Categorical  \n",
       "8     {category}             True           True      Categorical  \n",
       "9             {}             True           True             None  \n",
       "10     {numeric}             True           True             None  \n",
       "11            {}             True           True          Unknown  \n",
       "12            {}             True           True             None  \n",
       "13            {}             True           True          Unknown  \n",
       "14     {numeric}             True           True  IntegerNullable  \n",
       "15     {numeric}             True           True             None  \n",
       "16            {}             True           True             None  \n",
       "17            {}             True           True             None  \n",
       "18    {category}             True           True      Categorical  \n",
       "19            {}             True           True             None  \n",
       "20            {}             True           True          Unknown  \n",
       "21    {category}             True           True      Categorical  \n",
       "22    {category}             True           True      Categorical  \n",
       "23            {}             True           True          Unknown  \n",
       "24            {}             True           True          Unknown  \n",
       "25            {}             True           True             None  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ww.list_logical_types()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db42ecaf",
   "metadata": {},
   "source": [
    "#### unknown类型\n",
    "当woodwork类型推导没能成功，就设置`unknown`.  我们可以手动设置他"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d45248f",
   "metadata": {},
   "source": [
    "比如， 下面例子，国家代码没有推导类型成功，就设置了Unknown. 我们可以手动设置`CountryCode`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "aab23290",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Series: None (Physical Type = string) (Logical Type = Unknown) (Semantic Tags = set())>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s = pd.Series(['AU', 'US', 'UA'])\n",
    "unkown_series = ww.init_series(s)\n",
    "unkown_series.ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ff9fe108",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Series: None (Physical Type = category) (Logical Type = CountryCode) (Semantic Tags = {'category'})>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "countrycode_series = ww.init_series(unkown_series, 'CountryCode')\n",
    "countrycode_series.ww"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b10bd73",
   "metadata": {},
   "source": [
    "key"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa5be036",
   "metadata": {},
   "source": [
    "#### IntegerNullable \n",
    "表明是整数，但是可能会有空值，要小心"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "932e1f2b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Series: None (Physical Type = Int64) (Logical Type = IntegerNullable) (Semantic Tags = {'numeric'})>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "series = pd.Series([1, 2, None, 4], dtype=\"Int64\")\n",
    "intn_series = ww.init_series(series)\n",
    "intn_series.ww"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d95ca2b",
   "metadata": {},
   "source": [
    "#### ordinal 有序类型\n",
    "评分，排名等"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b5f1fcc",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "6f50eb2e",
   "metadata": {},
   "source": [
    "## ww处理"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abaa5b97",
   "metadata": {},
   "source": [
    "我们根据自动推导的，在进行微调"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "537c9d9a",
   "metadata": {},
   "source": [
    "### application"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "050ce50a",
   "metadata": {},
   "source": [
    "对于set和target这样的标签，我们手动处理，即不让进入ft过程"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "7703110b",
   "metadata": {},
   "outputs": [],
   "source": [
    "app.ww.init(name= 'app', index='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "afe0ca90",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'app'"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "app.ww.name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "dfdfecb4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TARGET</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_TYPE_SUITE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_INCOME_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_EDUCATION_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_FAMILY_STATUS</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_HOUSING_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_POPULATION_RELATIVE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_BIRTH</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_EMPLOYED</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_REGISTRATION</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ID_PUBLISH</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OWN_CAR_AGE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_MOBIL</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_EMP_PHONE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_WORK_PHONE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_CONT_MOBILE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_PHONE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_EMAIL</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OCCUPATION_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_FAM_MEMBERS</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_RATING_CLIENT</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_RATING_CLIENT_W_CITY</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WEEKDAY_APPR_PROCESS_START</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUR_APPR_PROCESS_START</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_REGION_NOT_LIVE_REGION</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_REGION_NOT_WORK_REGION</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVE_REGION_NOT_WORK_REGION</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_CITY_NOT_LIVE_CITY</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_CITY_NOT_WORK_CITY</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVE_CITY_NOT_WORK_CITY</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ORGANIZATION_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_1</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_2</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_3</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_AVG</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_MEDI</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FONDKAPREMONT_MODE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUSETYPE_MODE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TOTALAREA_MODE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WALLSMATERIAL_MODE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EMERGENCYSTATE_MODE</th>\n",
       "      <td>BooleanNullable</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OBS_30_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DEF_30_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OBS_60_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DEF_60_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_PHONE_CHANGE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_2</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_3</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_4</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_5</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_6</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_7</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_8</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_9</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_10</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_11</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_12</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_13</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_14</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_15</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_16</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_17</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_18</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_19</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_20</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_21</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_HOUR</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_DAY</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_WEEK</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_MON</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_QRT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_YEAR</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                               Logical Type Semantic Tag(s)\n",
       "Column                                                     \n",
       "SK_ID_CURR                          Integer       ['index']\n",
       "TARGET                      IntegerNullable     ['numeric']\n",
       "NAME_CONTRACT_TYPE              Categorical    ['category']\n",
       "CODE_GENDER                     Categorical    ['category']\n",
       "FLAG_OWN_CAR                        Boolean              []\n",
       "...                                     ...             ...\n",
       "AMT_REQ_CREDIT_BUREAU_DAY   IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_WEEK  IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_MON   IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_QRT   IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_YEAR  IntegerNullable     ['numeric']\n",
       "\n",
       "[122 rows x 2 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "app.ww.schema"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "01a91083",
   "metadata": {},
   "source": [
    "2. flag, is_not 设置为bool"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "efcd6141",
   "metadata": {},
   "outputs": [],
   "source": [
    "FLAG_DOCUMENTS = { f'FLAG_DOCUMENT_{i}':'Boolean' for i in range(2, 22)}\n",
    "app.ww.set_types(\n",
    "    logical_types = {\n",
    "        'FLAG_MOBIL': 'Boolean',\n",
    "        'FLAG_EMP_PHONE': 'Boolean',\n",
    "        'FLAG_WORK_PHONE': 'Boolean',\n",
    "        'FLAG_CONT_MOBILE': 'Boolean',\n",
    "        'FLAG_EMAIL': 'Boolean',\n",
    "        'FLAG_PHONE': 'Boolean',\n",
    "        'REG_CITY_NOT_LIVE_CITY': 'Boolean',\n",
    "        'REG_CITY_NOT_WORK_CITY': 'Boolean',\n",
    "        'LIVE_CITY_NOT_WORK_CITY': 'Boolean',\n",
    "        **FLAG_DOCUMENTS\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9dda836",
   "metadata": {},
   "source": [
    "3. 评级\n",
    "\n",
    "对于一些异常的，我们可以代替为np.nan， ft是不会处理的"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "dc54daee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 3], dtype=int64)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "app['REGION_RATING_CLIENT'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "2e18938c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 2,  1,  3, -1], dtype=int64)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "app['REGION_RATING_CLIENT_W_CITY'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "26982b30",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "224393   -1\n",
       "Name: REGION_RATING_CLIENT_W_CITY, dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "app['REGION_RATING_CLIENT_W_CITY'][app['REGION_RATING_CLIENT_W_CITY'] == -1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "a2555083",
   "metadata": {},
   "outputs": [],
   "source": [
    "app['REGION_RATING_CLIENT_W_CITY'] = app['REGION_RATING_CLIENT_W_CITY'].replace(-1, np.nan)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "bfe0b59d",
   "metadata": {},
   "outputs": [],
   "source": [
    "app.ww.set_types(\n",
    "    logical_types = {\n",
    "        'REGION_RATING_CLIENT': ww.logical_types.Ordinal(order=[1,2,3]),\n",
    "        'REGION_RATING_CLIENT_W_CITY': ww.logical_types.Ordinal(order=[1,2,3]),\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c86746b3",
   "metadata": {},
   "source": [
    "4. 时间段，应该为分类"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "eb3cb96b",
   "metadata": {},
   "outputs": [],
   "source": [
    "app.ww.set_types(\n",
    "    logical_types = {\n",
    "        'HOUR_APPR_PROCESS_START': 'Categorical',\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c3b771f9",
   "metadata": {},
   "source": [
    "### bureau"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "7ca4a0c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "bureau.ww.init(name= 'bureau', index='SK_ID_BUREAU')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "89e39f41",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_ACTIVE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_CURRENCY</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_DAY_OVERDUE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT_ENDDATE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ENDDATE_FACT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_MAX_OVERDUE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_CREDIT_PROLONG</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_DEBT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT_UPDATE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                           Logical Type Semantic Tag(s)\n",
       "Column                                                 \n",
       "SK_ID_CURR                      Integer     ['numeric']\n",
       "SK_ID_BUREAU                    Integer       ['index']\n",
       "CREDIT_ACTIVE               Categorical    ['category']\n",
       "CREDIT_CURRENCY             Categorical    ['category']\n",
       "DAYS_CREDIT                     Integer     ['numeric']\n",
       "CREDIT_DAY_OVERDUE              Integer     ['numeric']\n",
       "DAYS_CREDIT_ENDDATE     IntegerNullable     ['numeric']\n",
       "DAYS_ENDDATE_FACT       IntegerNullable     ['numeric']\n",
       "AMT_CREDIT_MAX_OVERDUE           Double     ['numeric']\n",
       "CNT_CREDIT_PROLONG              Integer     ['numeric']\n",
       "AMT_CREDIT_SUM                   Double     ['numeric']\n",
       "AMT_CREDIT_SUM_DEBT              Double     ['numeric']\n",
       "AMT_CREDIT_SUM_LIMIT             Double     ['numeric']\n",
       "AMT_CREDIT_SUM_OVERDUE           Double     ['numeric']\n",
       "CREDIT_TYPE                 Categorical    ['category']\n",
       "DAYS_CREDIT_UPDATE              Integer     ['numeric']\n",
       "AMT_ANNUITY                      Double     ['numeric']"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau.ww.schema"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de36aef4",
   "metadata": {},
   "source": [
    "1. id不参与"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "7a9b5874",
   "metadata": {},
   "outputs": [],
   "source": [
    "bureau.ww.set_types(semantic_tags={'SK_ID_CURR':'ignore'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "1076d49e",
   "metadata": {},
   "outputs": [],
   "source": [
    "bureau_balance = bureau_balance.reset_index().rename(columns = {'index':'bureaubalance_index'})\n",
    "bureau_balance.ww.init(name='bureau_balance', index='bureaubalance_index')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "01680ee2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>bureaubalance_index</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>STATUS</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                    Logical Type Semantic Tag(s)\n",
       "Column                                          \n",
       "bureaubalance_index      Integer       ['index']\n",
       "SK_ID_BUREAU             Integer     ['numeric']\n",
       "MONTHS_BALANCE           Integer     ['numeric']\n",
       "STATUS               Categorical    ['category']"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_balance.ww.schema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "d310e71c",
   "metadata": {},
   "outputs": [],
   "source": [
    "bureau_balance.ww.set_types(semantic_tags={'SK_ID_BUREAU':'ignore'})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88214411",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "51861296",
   "metadata": {},
   "source": [
    "### previous"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "06fd6728",
   "metadata": {},
   "outputs": [],
   "source": [
    "previous_application.ww.init(name='previous', index='SK_ID_PREV')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "adbf87cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_APPLICATION</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DOWN_PAYMENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WEEKDAY_APPR_PROCESS_START</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUR_APPR_PROCESS_START</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_LAST_APPL_PER_CONTRACT</th>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NFLAG_LAST_APPL_IN_DAY</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_DOWN_PAYMENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_INTEREST_PRIMARY</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_INTEREST_PRIVILEGED</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CASH_LOAN_PURPOSE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_DECISION</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PAYMENT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE_REJECT_REASON</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_TYPE_SUITE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CLIENT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_GOODS_CATEGORY</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PORTFOLIO</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PRODUCT_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CHANNEL_TYPE</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SELLERPLACE_AREA</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_SELLER_INDUSTRY</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_PAYMENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_YIELD_GROUP</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PRODUCT_COMBINATION</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_FIRST_DRAWING</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_FIRST_DUE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_DUE_1ST_VERSION</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_DUE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_TERMINATION</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NFLAG_INSURED_ON_APPROVAL</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                                Logical Type Semantic Tag(s)\n",
       "Column                                                      \n",
       "SK_ID_PREV                           Integer       ['index']\n",
       "SK_ID_CURR                           Integer     ['numeric']\n",
       "NAME_CONTRACT_TYPE               Categorical    ['category']\n",
       "AMT_ANNUITY                           Double     ['numeric']\n",
       "AMT_APPLICATION                       Double     ['numeric']\n",
       "AMT_CREDIT                            Double     ['numeric']\n",
       "AMT_DOWN_PAYMENT                      Double     ['numeric']\n",
       "AMT_GOODS_PRICE                       Double     ['numeric']\n",
       "WEEKDAY_APPR_PROCESS_START       Categorical    ['category']\n",
       "HOUR_APPR_PROCESS_START              Integer     ['numeric']\n",
       "FLAG_LAST_APPL_PER_CONTRACT          Boolean              []\n",
       "NFLAG_LAST_APPL_IN_DAY               Integer     ['numeric']\n",
       "RATE_DOWN_PAYMENT                     Double     ['numeric']\n",
       "RATE_INTEREST_PRIMARY                 Double     ['numeric']\n",
       "RATE_INTEREST_PRIVILEGED              Double     ['numeric']\n",
       "NAME_CASH_LOAN_PURPOSE           Categorical    ['category']\n",
       "NAME_CONTRACT_STATUS             Categorical    ['category']\n",
       "DAYS_DECISION                        Integer     ['numeric']\n",
       "NAME_PAYMENT_TYPE                Categorical    ['category']\n",
       "CODE_REJECT_REASON               Categorical    ['category']\n",
       "NAME_TYPE_SUITE                  Categorical    ['category']\n",
       "NAME_CLIENT_TYPE                 Categorical    ['category']\n",
       "NAME_GOODS_CATEGORY              Categorical    ['category']\n",
       "NAME_PORTFOLIO                   Categorical    ['category']\n",
       "NAME_PRODUCT_TYPE                Categorical    ['category']\n",
       "CHANNEL_TYPE                     Categorical    ['category']\n",
       "SELLERPLACE_AREA                     Integer     ['numeric']\n",
       "NAME_SELLER_INDUSTRY             Categorical    ['category']\n",
       "CNT_PAYMENT                  IntegerNullable     ['numeric']\n",
       "NAME_YIELD_GROUP                 Categorical    ['category']\n",
       "PRODUCT_COMBINATION              Categorical    ['category']\n",
       "DAYS_FIRST_DRAWING           IntegerNullable     ['numeric']\n",
       "DAYS_FIRST_DUE               IntegerNullable     ['numeric']\n",
       "DAYS_LAST_DUE_1ST_VERSION    IntegerNullable     ['numeric']\n",
       "DAYS_LAST_DUE                IntegerNullable     ['numeric']\n",
       "DAYS_TERMINATION             IntegerNullable     ['numeric']\n",
       "NFLAG_INSURED_ON_APPROVAL    IntegerNullable     ['numeric']"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application.ww.schema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "c0551cc8",
   "metadata": {},
   "outputs": [],
   "source": [
    "previous_application.ww.set_types(semantic_tags={'SK_ID_CURR':'ignore'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "fbb533d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "previous_application.ww.set_types(\n",
    "    logical_types = {\n",
    "        'HOUR_APPR_PROCESS_START': 'Categorical',\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "1b970b69",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0], dtype=int64)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application['NFLAG_LAST_APPL_IN_DAY'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "b7589ff0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<IntegerArray>\n",
       "[0, 1, <NA>]\n",
       "Length: 3, dtype: Int64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application['NFLAG_INSURED_ON_APPROVAL'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "814a704a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "673065"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application['NFLAG_INSURED_ON_APPROVAL'].isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47447a1c",
   "metadata": {},
   "source": [
    "nan比例过大，我们不应该设置`NFLAG_INSURED_ON_APPROVAL` 为布尔, 而是Categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "e4c2d639",
   "metadata": {},
   "outputs": [],
   "source": [
    "previous_application.ww.set_types(\n",
    "    logical_types = {\n",
    "        'NFLAG_LAST_APPL_IN_DAY': 'Boolean',\n",
    "        'NFLAG_INSURED_ON_APPROVAL': 'Categorical'\n",
    "        }\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "d8874a0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "credit_card_balance = credit_card_balance.reset_index().rename(columns = {'index':'credit_index'})\n",
    "credit_card_balance.ww.init(name='credit', index = 'credit_index')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "1023c60f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>credit_index</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_BALANCE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_LIMIT_ACTUAL</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_POS_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INST_MIN_REGULARITY</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT_TOTAL_CURRENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_RECEIVABLE_PRINCIPAL</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_RECIVABLE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_TOTAL_RECEIVABLE</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_CURRENT</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_POS_CURRENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT_MATURE_CUM</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                               Logical Type Semantic Tag(s)\n",
       "Column                                                     \n",
       "credit_index                        Integer       ['index']\n",
       "SK_ID_PREV                          Integer     ['numeric']\n",
       "SK_ID_CURR                          Integer     ['numeric']\n",
       "MONTHS_BALANCE                      Integer     ['numeric']\n",
       "AMT_BALANCE                          Double     ['numeric']\n",
       "AMT_CREDIT_LIMIT_ACTUAL             Integer     ['numeric']\n",
       "AMT_DRAWINGS_ATM_CURRENT             Double     ['numeric']\n",
       "AMT_DRAWINGS_CURRENT                 Double     ['numeric']\n",
       "AMT_DRAWINGS_OTHER_CURRENT           Double     ['numeric']\n",
       "AMT_DRAWINGS_POS_CURRENT             Double     ['numeric']\n",
       "AMT_INST_MIN_REGULARITY              Double     ['numeric']\n",
       "AMT_PAYMENT_CURRENT                  Double     ['numeric']\n",
       "AMT_PAYMENT_TOTAL_CURRENT            Double     ['numeric']\n",
       "AMT_RECEIVABLE_PRINCIPAL             Double     ['numeric']\n",
       "AMT_RECIVABLE                        Double     ['numeric']\n",
       "AMT_TOTAL_RECEIVABLE                 Double     ['numeric']\n",
       "CNT_DRAWINGS_ATM_CURRENT    IntegerNullable     ['numeric']\n",
       "CNT_DRAWINGS_CURRENT                Integer     ['numeric']\n",
       "CNT_DRAWINGS_OTHER_CURRENT  IntegerNullable     ['numeric']\n",
       "CNT_DRAWINGS_POS_CURRENT    IntegerNullable     ['numeric']\n",
       "CNT_INSTALMENT_MATURE_CUM   IntegerNullable     ['numeric']\n",
       "NAME_CONTRACT_STATUS            Categorical    ['category']\n",
       "SK_DPD                              Integer     ['numeric']\n",
       "SK_DPD_DEF                          Integer     ['numeric']"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "credit_card_balance.ww.schema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "cfc5a247",
   "metadata": {},
   "outputs": [],
   "source": [
    "credit_card_balance.ww.set_types(semantic_tags={'SK_ID_CURR':'ignore', 'SK_ID_PREV':'ignore'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "b6c9849f",
   "metadata": {},
   "outputs": [],
   "source": [
    "installments_payments = installments_payments.reset_index().rename(columns = {'index':'installments_index'})\n",
    "\n",
    "installments_payments.ww.init(name = 'installments', index='installments_index')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "24f38556",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>installments_index</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NUM_INSTALMENT_VERSION</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NUM_INSTALMENT_NUMBER</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_INSTALMENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ENTRY_PAYMENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INSTALMENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT</th>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                           Logical Type Semantic Tag(s)\n",
       "Column                                                 \n",
       "installments_index              Integer       ['index']\n",
       "SK_ID_PREV                      Integer     ['numeric']\n",
       "SK_ID_CURR                      Integer     ['numeric']\n",
       "NUM_INSTALMENT_VERSION           Double     ['numeric']\n",
       "NUM_INSTALMENT_NUMBER           Integer     ['numeric']\n",
       "DAYS_INSTALMENT                  Double     ['numeric']\n",
       "DAYS_ENTRY_PAYMENT      IntegerNullable     ['numeric']\n",
       "AMT_INSTALMENT                   Double     ['numeric']\n",
       "AMT_PAYMENT                      Double     ['numeric']"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments.ww.schema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "1f56c28f",
   "metadata": {},
   "outputs": [],
   "source": [
    "installments_payments.ww.set_types(semantic_tags={'SK_ID_CURR':'ignore', 'SK_ID_PREV':'ignore'})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7a974f79",
   "metadata": {},
   "source": [
    "`NUM_INSTALMENT_VERSION` 更换类型为整数，这也会影响到后面得特征矩阵"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "2e5c63be",
   "metadata": {},
   "outputs": [],
   "source": [
    "installments_payments.ww.set_types(\n",
    "    logical_types = {\n",
    "        'NUM_INSTALMENT_VERSION': 'Integer',\n",
    "        'DAYS_INSTALMENT': 'Integer'\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "09276350",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments['DAYS_INSTALMENT'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "213edce0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments['NUM_INSTALMENT_VERSION'].isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "67d7711d",
   "metadata": {},
   "outputs": [],
   "source": [
    "pos_cash_balance = pos_cash_balance.reset_index().rename(columns = {'index':'cash_index'})\n",
    "pos_cash_balance.ww.init(name='cash', index='cash_index')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e570fce7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>cash_index</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                          Logical Type Semantic Tag(s)\n",
       "Column                                                \n",
       "cash_index                     Integer       ['index']\n",
       "SK_ID_PREV                     Integer     ['numeric']\n",
       "SK_ID_CURR                     Integer     ['numeric']\n",
       "MONTHS_BALANCE                 Integer     ['numeric']\n",
       "CNT_INSTALMENT         IntegerNullable     ['numeric']\n",
       "CNT_INSTALMENT_FUTURE  IntegerNullable     ['numeric']\n",
       "NAME_CONTRACT_STATUS       Categorical    ['category']\n",
       "SK_DPD                         Integer     ['numeric']\n",
       "SK_DPD_DEF                     Integer     ['numeric']"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_cash_balance.ww.schema"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "e090fd93",
   "metadata": {},
   "outputs": [],
   "source": [
    "pos_cash_balance.ww.set_types(semantic_tags={'SK_ID_CURR':'ignore', 'SK_ID_PREV':'ignore'})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b920d11c",
   "metadata": {},
   "source": [
    "## 构建es"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e534c5c",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "13ef33da",
   "metadata": {},
   "source": [
    "featuretools对于已经初始化的ww，有些要求：\n",
    "- 具备index\n",
    "- name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "55a96aab",
   "metadata": {},
   "outputs": [],
   "source": [
    "es = ft.EntitySet(id='clients')\n",
    "\n",
    "# 有主键唯一列\n",
    "es = es.add_dataframe( dataframe=app,)\n",
    "es = es.add_dataframe( dataframe=bureau)\n",
    "es = es.add_dataframe( dataframe=previous_application)\n",
    "\n",
    "# 没有主键唯一的列，需要make_index, 创建一列主键\n",
    "es = es.add_dataframe(dataframe=bureau_balance)\n",
    "es = es.add_dataframe( dataframe=credit_card_balance)\n",
    "es = es.add_dataframe( dataframe=installments_payments)\n",
    "es = es.add_dataframe(dataframe=pos_cash_balance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "56ed6406",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 父亲dfname, 父亲列名； 字dfname, 子列名\n",
    "es = es.add_relationship(\"app\", \"SK_ID_CURR\", \"bureau\", \"SK_ID_CURR\")\n",
    "es = es.add_relationship(\"bureau\", \"SK_ID_BUREAU\", \"bureau_balance\", \"SK_ID_BUREAU\")\n",
    "\n",
    "es = es.add_relationship(\"app\", \"SK_ID_CURR\", \"previous\", \"SK_ID_CURR\")\n",
    "es = es.add_relationship(\"previous\", \"SK_ID_PREV\", \"cash\", \"SK_ID_PREV\")\n",
    "es = es.add_relationship(\"previous\", \"SK_ID_PREV\", \"installments\", \"SK_ID_PREV\")\n",
    "es = es.add_relationship(\"previous\", \"SK_ID_PREV\", \"credit\", \"SK_ID_PREV\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5406e34",
   "metadata": {},
   "source": [
    "在构建完关系后，语义标签上会携带外键"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "f9eaa11c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TARGET</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_TYPE_SUITE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_INCOME_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_EDUCATION_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_FAMILY_STATUS</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_HOUSING_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_POPULATION_RELATIVE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_BIRTH</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_EMPLOYED</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_REGISTRATION</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ID_PUBLISH</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OWN_CAR_AGE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_MOBIL</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_EMP_PHONE</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_WORK_PHONE</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_CONT_MOBILE</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_PHONE</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_EMAIL</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OCCUPATION_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_FAM_MEMBERS</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_RATING_CLIENT</th>\n",
       "      <td>category</td>\n",
       "      <td>Ordinal: [1, 2, 3]</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REGION_RATING_CLIENT_W_CITY</th>\n",
       "      <td>category</td>\n",
       "      <td>Ordinal: [1, 2, 3]</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WEEKDAY_APPR_PROCESS_START</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUR_APPR_PROCESS_START</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_REGION_NOT_LIVE_REGION</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_REGION_NOT_WORK_REGION</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVE_REGION_NOT_WORK_REGION</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_CITY_NOT_LIVE_CITY</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>REG_CITY_NOT_WORK_CITY</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVE_CITY_NOT_WORK_CITY</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ORGANIZATION_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_1</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_2</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EXT_SOURCE_3</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_AVG</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>APARTMENTS_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASEMENTAREA_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BEGINEXPLUATATION_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YEARS_BUILD_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COMMONAREA_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ELEVATORS_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENTRANCES_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMAX_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLOORSMIN_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LANDAREA_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LIVINGAREA_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAPARTMENTS_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NONLIVINGAREA_MEDI</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FONDKAPREMONT_MODE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUSETYPE_MODE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TOTALAREA_MODE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WALLSMATERIAL_MODE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>EMERGENCYSTATE_MODE</th>\n",
       "      <td>boolean</td>\n",
       "      <td>BooleanNullable</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OBS_30_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DEF_30_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OBS_60_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DEF_60_CNT_SOCIAL_CIRCLE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_PHONE_CHANGE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_2</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_3</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_4</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_5</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_6</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_7</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_8</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_9</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_10</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_11</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_12</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_13</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_14</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_15</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_16</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_17</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_18</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_19</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_20</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_DOCUMENT_21</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_HOUR</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_DAY</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_WEEK</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_MON</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_QRT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_YEAR</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                           Physical Type     Logical Type Semantic Tag(s)\n",
       "Column                                                                   \n",
       "SK_ID_CURR                         int64          Integer       ['index']\n",
       "TARGET                             Int64  IntegerNullable     ['numeric']\n",
       "NAME_CONTRACT_TYPE              category      Categorical    ['category']\n",
       "CODE_GENDER                     category      Categorical    ['category']\n",
       "FLAG_OWN_CAR                        bool          Boolean              []\n",
       "...                                  ...              ...             ...\n",
       "AMT_REQ_CREDIT_BUREAU_DAY          Int64  IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_WEEK         Int64  IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_MON          Int64  IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_QRT          Int64  IntegerNullable     ['numeric']\n",
       "AMT_REQ_CREDIT_BUREAU_YEAR         Int64  IntegerNullable     ['numeric']\n",
       "\n",
       "[122 rows x 3 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['app'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "b1d27ddf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_ACTIVE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_CURRENCY</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_DAY_OVERDUE</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT_ENDDATE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ENDDATE_FACT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_MAX_OVERDUE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_CREDIT_PROLONG</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_DEBT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CREDIT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_CREDIT_UPDATE</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                       Physical Type     Logical Type  \\\n",
       "Column                                                  \n",
       "SK_ID_CURR                     int64          Integer   \n",
       "SK_ID_BUREAU                   int64          Integer   \n",
       "CREDIT_ACTIVE               category      Categorical   \n",
       "CREDIT_CURRENCY             category      Categorical   \n",
       "DAYS_CREDIT                    int64          Integer   \n",
       "CREDIT_DAY_OVERDUE             int64          Integer   \n",
       "DAYS_CREDIT_ENDDATE            Int64  IntegerNullable   \n",
       "DAYS_ENDDATE_FACT              Int64  IntegerNullable   \n",
       "AMT_CREDIT_MAX_OVERDUE       float64           Double   \n",
       "CNT_CREDIT_PROLONG             int64          Integer   \n",
       "AMT_CREDIT_SUM               float64           Double   \n",
       "AMT_CREDIT_SUM_DEBT          float64           Double   \n",
       "AMT_CREDIT_SUM_LIMIT         float64           Double   \n",
       "AMT_CREDIT_SUM_OVERDUE       float64           Double   \n",
       "CREDIT_TYPE                 category      Categorical   \n",
       "DAYS_CREDIT_UPDATE             int64          Integer   \n",
       "AMT_ANNUITY                  float64           Double   \n",
       "\n",
       "                                             Semantic Tag(s)  \n",
       "Column                                                        \n",
       "SK_ID_CURR              ['foreign_key', 'numeric', 'ignore']  \n",
       "SK_ID_BUREAU                                       ['index']  \n",
       "CREDIT_ACTIVE                                   ['category']  \n",
       "CREDIT_CURRENCY                                 ['category']  \n",
       "DAYS_CREDIT                                      ['numeric']  \n",
       "CREDIT_DAY_OVERDUE                               ['numeric']  \n",
       "DAYS_CREDIT_ENDDATE                              ['numeric']  \n",
       "DAYS_ENDDATE_FACT                                ['numeric']  \n",
       "AMT_CREDIT_MAX_OVERDUE                           ['numeric']  \n",
       "CNT_CREDIT_PROLONG                               ['numeric']  \n",
       "AMT_CREDIT_SUM                                   ['numeric']  \n",
       "AMT_CREDIT_SUM_DEBT                              ['numeric']  \n",
       "AMT_CREDIT_SUM_LIMIT                             ['numeric']  \n",
       "AMT_CREDIT_SUM_OVERDUE                           ['numeric']  \n",
       "CREDIT_TYPE                                     ['category']  \n",
       "DAYS_CREDIT_UPDATE                               ['numeric']  \n",
       "AMT_ANNUITY                                      ['numeric']  "
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['bureau'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "6bc3914e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>bureaubalance_index</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>STATUS</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                    Physical Type Logical Type  \\\n",
       "Column                                           \n",
       "bureaubalance_index         int64      Integer   \n",
       "SK_ID_BUREAU                int64      Integer   \n",
       "MONTHS_BALANCE              int64      Integer   \n",
       "STATUS                   category  Categorical   \n",
       "\n",
       "                                          Semantic Tag(s)  \n",
       "Column                                                     \n",
       "bureaubalance_index                             ['index']  \n",
       "SK_ID_BUREAU         ['foreign_key', 'numeric', 'ignore']  \n",
       "MONTHS_BALANCE                                ['numeric']  \n",
       "STATUS                                       ['category']  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['bureau_balance'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "f90363d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_APPLICATION</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DOWN_PAYMENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WEEKDAY_APPR_PROCESS_START</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HOUR_APPR_PROCESS_START</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FLAG_LAST_APPL_PER_CONTRACT</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NFLAG_LAST_APPL_IN_DAY</th>\n",
       "      <td>bool</td>\n",
       "      <td>Boolean</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_DOWN_PAYMENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_INTEREST_PRIMARY</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RATE_INTEREST_PRIVILEGED</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CASH_LOAN_PURPOSE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_DECISION</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PAYMENT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CODE_REJECT_REASON</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_TYPE_SUITE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CLIENT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_GOODS_CATEGORY</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PORTFOLIO</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_PRODUCT_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CHANNEL_TYPE</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SELLERPLACE_AREA</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_SELLER_INDUSTRY</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_PAYMENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_YIELD_GROUP</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PRODUCT_COMBINATION</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_FIRST_DRAWING</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_FIRST_DUE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_DUE_1ST_VERSION</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_LAST_DUE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_TERMINATION</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NFLAG_INSURED_ON_APPROVAL</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                            Physical Type     Logical Type  \\\n",
       "Column                                                       \n",
       "SK_ID_PREV                          int64          Integer   \n",
       "SK_ID_CURR                          int64          Integer   \n",
       "NAME_CONTRACT_TYPE               category      Categorical   \n",
       "AMT_ANNUITY                       float64           Double   \n",
       "AMT_APPLICATION                   float64           Double   \n",
       "AMT_CREDIT                        float64           Double   \n",
       "AMT_DOWN_PAYMENT                  float64           Double   \n",
       "AMT_GOODS_PRICE                   float64           Double   \n",
       "WEEKDAY_APPR_PROCESS_START       category      Categorical   \n",
       "HOUR_APPR_PROCESS_START          category      Categorical   \n",
       "FLAG_LAST_APPL_PER_CONTRACT          bool          Boolean   \n",
       "NFLAG_LAST_APPL_IN_DAY               bool          Boolean   \n",
       "RATE_DOWN_PAYMENT                 float64           Double   \n",
       "RATE_INTEREST_PRIMARY             float64           Double   \n",
       "RATE_INTEREST_PRIVILEGED          float64           Double   \n",
       "NAME_CASH_LOAN_PURPOSE           category      Categorical   \n",
       "NAME_CONTRACT_STATUS             category      Categorical   \n",
       "DAYS_DECISION                       int64          Integer   \n",
       "NAME_PAYMENT_TYPE                category      Categorical   \n",
       "CODE_REJECT_REASON               category      Categorical   \n",
       "NAME_TYPE_SUITE                  category      Categorical   \n",
       "NAME_CLIENT_TYPE                 category      Categorical   \n",
       "NAME_GOODS_CATEGORY              category      Categorical   \n",
       "NAME_PORTFOLIO                   category      Categorical   \n",
       "NAME_PRODUCT_TYPE                category      Categorical   \n",
       "CHANNEL_TYPE                     category      Categorical   \n",
       "SELLERPLACE_AREA                    int64          Integer   \n",
       "NAME_SELLER_INDUSTRY             category      Categorical   \n",
       "CNT_PAYMENT                         Int64  IntegerNullable   \n",
       "NAME_YIELD_GROUP                 category      Categorical   \n",
       "PRODUCT_COMBINATION              category      Categorical   \n",
       "DAYS_FIRST_DRAWING                  Int64  IntegerNullable   \n",
       "DAYS_FIRST_DUE                      Int64  IntegerNullable   \n",
       "DAYS_LAST_DUE_1ST_VERSION           Int64  IntegerNullable   \n",
       "DAYS_LAST_DUE                       Int64  IntegerNullable   \n",
       "DAYS_TERMINATION                    Int64  IntegerNullable   \n",
       "NFLAG_INSURED_ON_APPROVAL        category      Categorical   \n",
       "\n",
       "                                                  Semantic Tag(s)  \n",
       "Column                                                             \n",
       "SK_ID_PREV                                              ['index']  \n",
       "SK_ID_CURR                   ['foreign_key', 'numeric', 'ignore']  \n",
       "NAME_CONTRACT_TYPE                                   ['category']  \n",
       "AMT_ANNUITY                                           ['numeric']  \n",
       "AMT_APPLICATION                                       ['numeric']  \n",
       "AMT_CREDIT                                            ['numeric']  \n",
       "AMT_DOWN_PAYMENT                                      ['numeric']  \n",
       "AMT_GOODS_PRICE                                       ['numeric']  \n",
       "WEEKDAY_APPR_PROCESS_START                           ['category']  \n",
       "HOUR_APPR_PROCESS_START                              ['category']  \n",
       "FLAG_LAST_APPL_PER_CONTRACT                                    []  \n",
       "NFLAG_LAST_APPL_IN_DAY                                         []  \n",
       "RATE_DOWN_PAYMENT                                     ['numeric']  \n",
       "RATE_INTEREST_PRIMARY                                 ['numeric']  \n",
       "RATE_INTEREST_PRIVILEGED                              ['numeric']  \n",
       "NAME_CASH_LOAN_PURPOSE                               ['category']  \n",
       "NAME_CONTRACT_STATUS                                 ['category']  \n",
       "DAYS_DECISION                                         ['numeric']  \n",
       "NAME_PAYMENT_TYPE                                    ['category']  \n",
       "CODE_REJECT_REASON                                   ['category']  \n",
       "NAME_TYPE_SUITE                                      ['category']  \n",
       "NAME_CLIENT_TYPE                                     ['category']  \n",
       "NAME_GOODS_CATEGORY                                  ['category']  \n",
       "NAME_PORTFOLIO                                       ['category']  \n",
       "NAME_PRODUCT_TYPE                                    ['category']  \n",
       "CHANNEL_TYPE                                         ['category']  \n",
       "SELLERPLACE_AREA                                      ['numeric']  \n",
       "NAME_SELLER_INDUSTRY                                 ['category']  \n",
       "CNT_PAYMENT                                           ['numeric']  \n",
       "NAME_YIELD_GROUP                                     ['category']  \n",
       "PRODUCT_COMBINATION                                  ['category']  \n",
       "DAYS_FIRST_DRAWING                                    ['numeric']  \n",
       "DAYS_FIRST_DUE                                        ['numeric']  \n",
       "DAYS_LAST_DUE_1ST_VERSION                             ['numeric']  \n",
       "DAYS_LAST_DUE                                         ['numeric']  \n",
       "DAYS_TERMINATION                                      ['numeric']  \n",
       "NFLAG_INSURED_ON_APPROVAL                            ['category']  "
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['previous'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "c01741f9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>cash_index</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                      Physical Type     Logical Type  \\\n",
       "Column                                                 \n",
       "cash_index                    int64          Integer   \n",
       "SK_ID_PREV                    int64          Integer   \n",
       "SK_ID_CURR                    int64          Integer   \n",
       "MONTHS_BALANCE                int64          Integer   \n",
       "CNT_INSTALMENT                Int64  IntegerNullable   \n",
       "CNT_INSTALMENT_FUTURE         Int64  IntegerNullable   \n",
       "NAME_CONTRACT_STATUS       category      Categorical   \n",
       "SK_DPD                        int64          Integer   \n",
       "SK_DPD_DEF                    int64          Integer   \n",
       "\n",
       "                                            Semantic Tag(s)  \n",
       "Column                                                       \n",
       "cash_index                                        ['index']  \n",
       "SK_ID_PREV             ['foreign_key', 'numeric', 'ignore']  \n",
       "SK_ID_CURR                            ['numeric', 'ignore']  \n",
       "MONTHS_BALANCE                                  ['numeric']  \n",
       "CNT_INSTALMENT                                  ['numeric']  \n",
       "CNT_INSTALMENT_FUTURE                           ['numeric']  \n",
       "NAME_CONTRACT_STATUS                           ['category']  \n",
       "SK_DPD                                          ['numeric']  \n",
       "SK_DPD_DEF                                      ['numeric']  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['cash'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "d1bdec13",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>credit_index</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_BALANCE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_CREDIT_LIMIT_ACTUAL</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_DRAWINGS_POS_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INST_MIN_REGULARITY</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT_TOTAL_CURRENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_RECEIVABLE_PRINCIPAL</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_RECIVABLE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_TOTAL_RECEIVABLE</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_CURRENT</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_DRAWINGS_POS_CURRENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT_MATURE_CUM</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>category</td>\n",
       "      <td>Categorical</td>\n",
       "      <td>['category']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                           Physical Type     Logical Type  \\\n",
       "Column                                                      \n",
       "credit_index                       int64          Integer   \n",
       "SK_ID_PREV                         int64          Integer   \n",
       "SK_ID_CURR                         int64          Integer   \n",
       "MONTHS_BALANCE                     int64          Integer   \n",
       "AMT_BALANCE                      float64           Double   \n",
       "AMT_CREDIT_LIMIT_ACTUAL            int64          Integer   \n",
       "AMT_DRAWINGS_ATM_CURRENT         float64           Double   \n",
       "AMT_DRAWINGS_CURRENT             float64           Double   \n",
       "AMT_DRAWINGS_OTHER_CURRENT       float64           Double   \n",
       "AMT_DRAWINGS_POS_CURRENT         float64           Double   \n",
       "AMT_INST_MIN_REGULARITY          float64           Double   \n",
       "AMT_PAYMENT_CURRENT              float64           Double   \n",
       "AMT_PAYMENT_TOTAL_CURRENT        float64           Double   \n",
       "AMT_RECEIVABLE_PRINCIPAL         float64           Double   \n",
       "AMT_RECIVABLE                    float64           Double   \n",
       "AMT_TOTAL_RECEIVABLE             float64           Double   \n",
       "CNT_DRAWINGS_ATM_CURRENT           Int64  IntegerNullable   \n",
       "CNT_DRAWINGS_CURRENT               int64          Integer   \n",
       "CNT_DRAWINGS_OTHER_CURRENT         Int64  IntegerNullable   \n",
       "CNT_DRAWINGS_POS_CURRENT           Int64  IntegerNullable   \n",
       "CNT_INSTALMENT_MATURE_CUM          Int64  IntegerNullable   \n",
       "NAME_CONTRACT_STATUS            category      Categorical   \n",
       "SK_DPD                             int64          Integer   \n",
       "SK_DPD_DEF                         int64          Integer   \n",
       "\n",
       "                                                 Semantic Tag(s)  \n",
       "Column                                                            \n",
       "credit_index                                           ['index']  \n",
       "SK_ID_PREV                  ['foreign_key', 'numeric', 'ignore']  \n",
       "SK_ID_CURR                                 ['numeric', 'ignore']  \n",
       "MONTHS_BALANCE                                       ['numeric']  \n",
       "AMT_BALANCE                                          ['numeric']  \n",
       "AMT_CREDIT_LIMIT_ACTUAL                              ['numeric']  \n",
       "AMT_DRAWINGS_ATM_CURRENT                             ['numeric']  \n",
       "AMT_DRAWINGS_CURRENT                                 ['numeric']  \n",
       "AMT_DRAWINGS_OTHER_CURRENT                           ['numeric']  \n",
       "AMT_DRAWINGS_POS_CURRENT                             ['numeric']  \n",
       "AMT_INST_MIN_REGULARITY                              ['numeric']  \n",
       "AMT_PAYMENT_CURRENT                                  ['numeric']  \n",
       "AMT_PAYMENT_TOTAL_CURRENT                            ['numeric']  \n",
       "AMT_RECEIVABLE_PRINCIPAL                             ['numeric']  \n",
       "AMT_RECIVABLE                                        ['numeric']  \n",
       "AMT_TOTAL_RECEIVABLE                                 ['numeric']  \n",
       "CNT_DRAWINGS_ATM_CURRENT                             ['numeric']  \n",
       "CNT_DRAWINGS_CURRENT                                 ['numeric']  \n",
       "CNT_DRAWINGS_OTHER_CURRENT                           ['numeric']  \n",
       "CNT_DRAWINGS_POS_CURRENT                             ['numeric']  \n",
       "CNT_INSTALMENT_MATURE_CUM                            ['numeric']  \n",
       "NAME_CONTRACT_STATUS                                ['category']  \n",
       "SK_DPD                                               ['numeric']  \n",
       "SK_DPD_DEF                                           ['numeric']  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['credit'].ww"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "351f406f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Physical Type</th>\n",
       "      <th>Logical Type</th>\n",
       "      <th>Semantic Tag(s)</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>installments_index</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['index']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['foreign_key', 'numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric', 'ignore']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NUM_INSTALMENT_VERSION</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NUM_INSTALMENT_NUMBER</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_INSTALMENT</th>\n",
       "      <td>int64</td>\n",
       "      <td>Integer</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DAYS_ENTRY_PAYMENT</th>\n",
       "      <td>Int64</td>\n",
       "      <td>IntegerNullable</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_INSTALMENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AMT_PAYMENT</th>\n",
       "      <td>float64</td>\n",
       "      <td>Double</td>\n",
       "      <td>['numeric']</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "                       Physical Type     Logical Type  \\\n",
       "Column                                                  \n",
       "installments_index             int64          Integer   \n",
       "SK_ID_PREV                     int64          Integer   \n",
       "SK_ID_CURR                     int64          Integer   \n",
       "NUM_INSTALMENT_VERSION         int64          Integer   \n",
       "NUM_INSTALMENT_NUMBER          int64          Integer   \n",
       "DAYS_INSTALMENT                int64          Integer   \n",
       "DAYS_ENTRY_PAYMENT             Int64  IntegerNullable   \n",
       "AMT_INSTALMENT               float64           Double   \n",
       "AMT_PAYMENT                  float64           Double   \n",
       "\n",
       "                                             Semantic Tag(s)  \n",
       "Column                                                        \n",
       "installments_index                                 ['index']  \n",
       "SK_ID_PREV              ['foreign_key', 'numeric', 'ignore']  \n",
       "SK_ID_CURR                             ['numeric', 'ignore']  \n",
       "NUM_INSTALMENT_VERSION                           ['numeric']  \n",
       "NUM_INSTALMENT_NUMBER                            ['numeric']  \n",
       "DAYS_INSTALMENT                                  ['numeric']  \n",
       "DAYS_ENTRY_PAYMENT                               ['numeric']  \n",
       "AMT_INSTALMENT                                   ['numeric']  \n",
       "AMT_PAYMENT                                      ['numeric']  "
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['installments'].ww"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb41357c",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b7d266b6",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "3604641e",
   "metadata": {},
   "source": [
    "## 添加interesting values\n",
    "就是where = 条件聚合。\n",
    "\n",
    "比如设置agg原语`mean`,  产生`MEAN(prev.AMT_CREDIT)`\n",
    "\n",
    "如果另外设置where原语`count` 和 兴趣 `{\"NAME_CONTRACT_STATUS\": [\"Approved\", \"Refused\"]}` \n",
    "就会多两个特征\n",
    "- `COUNT(prev.AMT_CREDIT where NAME_CONTRACT_STATUS==Approved) `\n",
    "- `COUNT(prev.AMT_CREDIT where NAME_CONTRACT_STATUS==Refused) `\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "a8298367",
   "metadata": {},
   "outputs": [],
   "source": [
    "es.add_interesting_values(dataframe_name='previous', values= {\n",
    "    \"NAME_CONTRACT_STATUS\": [\"Approved\", \"Refused\"]\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "e8ab2918",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'dataframe_name': 'previous',\n",
       " 'entityset_id': 'clients',\n",
       " 'interesting_values': ['Approved', 'Refused']}"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['previous'].ww.columns['NAME_CONTRACT_STATUS'].metadata"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fbe005fe",
   "metadata": {},
   "source": [
    "我们确实为这个列添加了where值"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3039e7b7",
   "metadata": {},
   "source": [
    "## seed feature\n",
    "没什么特别，就是构造了一列"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "d1dbe1db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "196114.0212179794"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application['AMT_CREDIT'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "3fc6ea37",
   "metadata": {},
   "outputs": [],
   "source": [
    "FLAG_LATED = ft.Feature(es['installments'].ww['DAYS_ENTRY_PAYMENT']) > ft.Feature(es['installments'].ww['DAYS_INSTALMENT'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "20c014e4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Feature: DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "FLAG_LATED"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84350b20",
   "metadata": {},
   "source": [
    "FLAG_LATED标识了一种条件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "e81429c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# \n",
    "FLAG_DUE = ft.Feature(es['bureau_balance'].ww['STATUS']).isin(['1', '2', '3', '4', '5'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Feature: STATUS.isin(['1', '2', '3', '4', '5'])>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "FLAG_DUE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "a50977dc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Boolean\n"
     ]
    }
   ],
   "source": [
    "print(FLAG_DUE.column_schema.logical_type)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7257fdc5",
   "metadata": {},
   "source": [
    "## 自定义特征原语\n",
    "对于自定义原语，我们一定要小心其性能"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "8c5daf19",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1670214"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "es['previous'].ww['NAME_CONTRACT_STATUS'].value_counts().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "b96231f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "class NormalizedModeCount(AggregationPrimitive):\n",
    "    \"\"\" 计算出现最多的次数占比总数的比例。\n",
    "    \"\"\"\n",
    "    name = 'normalized_mode_count'\n",
    "    input_types = [ColumnSchema(semantic_tags={'category'})]\n",
    "    return_type = ColumnSchema(semantic_tags={\"numeric\"})\n",
    "\n",
    "    def get_function(self):\n",
    "        def normalized_mode_count(column):\n",
    "            if len(column) == 0:\n",
    "                return 0\n",
    "            counts = column.value_counts()\n",
    "            if len(counts) == 0:\n",
    "                return 0\n",
    "            return counts.max()/counts.sum()\n",
    "        return normalized_mode_count"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5a78e5c",
   "metadata": {},
   "source": [
    "比如对于NAME_CONTRACT_STATUS， 表明以往 申请 通过或者拒绝的比例"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "e6313578",
   "metadata": {},
   "outputs": [],
   "source": [
    "class MaxConsecutive(AggregationPrimitive):\n",
    "    \"\"\" 最大连续次数，一般针对bool\n",
    "    \"\"\"\n",
    "    name = 'max_consecutive'\n",
    "    input_types = [ColumnSchema(logical_type=Boolean)]\n",
    "    return_type = ColumnSchema(semantic_tags={\"numeric\"})\n",
    "    def get_function(self):\n",
    "        def max_consecutive(column):\n",
    "            v = column.values\n",
    "            if len(v) == 0: return 0\n",
    "            \n",
    "            # 在首尾补 0 方便计算切换点\n",
    "            calls = np.concatenate(([0], v, [0]))\n",
    "            # 寻找从 0 变 1 和从 1 变 0 的位置\n",
    "            diffs = np.diff(calls.astype(int))\n",
    "            starts = np.where(diffs == 1)[0]\n",
    "            ends = np.where(diffs == -1)[0]\n",
    "            \n",
    "            if len(starts) == 0: return 0\n",
    "            # 长度即为 结束索引 - 开始索引\n",
    "            return np.max(ends - starts)\n",
    "        return max_consecutive"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6ea42ef9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "15e8303d",
   "metadata": {},
   "source": [
    "```{warning} \n",
    "我们必须清晰，哪些原语作用哪些列！"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "af7a6aab",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c5ab6260",
   "metadata": {},
   "source": [
    "## dfs"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "097d9aac",
   "metadata": {},
   "source": [
    "特征数估计：\n",
    "主表100列，从表50列，兴趣特征1*3个分类值， 原语5个。 \n",
    "50* 5 * 3 + 100"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "65bb4416",
   "metadata": {},
   "source": [
    "\n",
    "外键和索引不需要管"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "36f8f51d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 4h 41min 18s\n",
      "Wall time: 5h 46min 14s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "default_agg_primitives = [\n",
    "    \"count\",  # index\n",
    "    \"mean\", \"max\", \"sum\", \"std\",  # numeric\n",
    "    \"mode\", \"num_unique\", # categorical\n",
    "    'percent_true' # boolean\n",
    "    ]\n",
    "default_trans_primitives =  [\"month\", \"weekday\"]\n",
    "\n",
    "# 返回特征矩阵； 特征\n",
    "feature_matrix, features = ft.dfs(\n",
    "    entityset = es,\n",
    "    target_dataframe_name = 'app', # 最后要关联到这个表，以这个为主\n",
    "    agg_primitives= default_agg_primitives + [NormalizedModeCount, MaxConsecutive],\n",
    "    trans_primitives=default_trans_primitives,\n",
    "    max_depth=2,\n",
    "    seed_features=[FLAG_LATED, FLAG_DUE],\n",
    "    # n_jobs=2,        # 使用2个核\n",
    "    where_primitives=['count', 'mean', 'percent_true'],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "a5195ba6",
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_matrix.to_parquet(\"ft_tuning_feature_matrix.parquet\")\n",
    "ft.save_features(features, \"ft_tuning_feature_definitions.json\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f2dedd8",
   "metadata": {},
   "source": [
    "耗时1h30min"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "76bd0b8e",
   "metadata": {},
   "source": [
    "我们需要检查我们的特征确实生效了"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "4332fa95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<Feature: NORMALIZED_MODE_COUNT(bureau.CREDIT_ACTIVE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau.CREDIT_CURRENCY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau.CREDIT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau_balance.STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.CHANNEL_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.CODE_REJECT_REASON)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.HOUR_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_CASH_LOAN_PURPOSE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_CLIENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_CONTRACT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_GOODS_CATEGORY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_PAYMENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_PORTFOLIO)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_PRODUCT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_SELLER_INDUSTRY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_TYPE_SUITE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NAME_YIELD_GROUP)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.NFLAG_INSURED_ON_APPROVAL)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.PRODUCT_COMBINATION)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.WEEKDAY_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau.MODE(bureau_balance.STATUS))>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau_balance.bureau.CREDIT_ACTIVE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau_balance.bureau.CREDIT_CURRENCY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(bureau_balance.bureau.CREDIT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.MODE(cash.NAME_CONTRACT_STATUS))>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(previous.MODE(credit.NAME_CONTRACT_STATUS))>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.CHANNEL_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.CODE_REJECT_REASON)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.HOUR_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_CASH_LOAN_PURPOSE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_CLIENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_CONTRACT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_GOODS_CATEGORY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_PAYMENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_PORTFOLIO)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_PRODUCT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_SELLER_INDUSTRY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_TYPE_SUITE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NAME_YIELD_GROUP)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.NFLAG_INSURED_ON_APPROVAL)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.PRODUCT_COMBINATION)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(cash.previous.WEEKDAY_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.CHANNEL_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.CODE_REJECT_REASON)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.HOUR_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_CASH_LOAN_PURPOSE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_CLIENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_CONTRACT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_GOODS_CATEGORY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_PAYMENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_PORTFOLIO)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_PRODUCT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_SELLER_INDUSTRY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_TYPE_SUITE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NAME_YIELD_GROUP)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.NFLAG_INSURED_ON_APPROVAL)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.PRODUCT_COMBINATION)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(installments.previous.WEEKDAY_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.CHANNEL_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.CODE_REJECT_REASON)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.HOUR_APPR_PROCESS_START)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_CASH_LOAN_PURPOSE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_CLIENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_CONTRACT_STATUS)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_CONTRACT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_GOODS_CATEGORY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_PAYMENT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_PORTFOLIO)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_PRODUCT_TYPE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_SELLER_INDUSTRY)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_TYPE_SUITE)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NAME_YIELD_GROUP)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.NFLAG_INSURED_ON_APPROVAL)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.PRODUCT_COMBINATION)>,\n",
       " <Feature: NORMALIZED_MODE_COUNT(credit.previous.WEEKDAY_APPR_PROCESS_START)>]"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[f for f in features if f.primitive.name == 'normalized_mode_count']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "b7ddcf88",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<Feature: MAX_CONSECUTIVE(previous.FLAG_LAST_APPL_PER_CONTRACT)>,\n",
       " <Feature: MAX_CONSECUTIVE(previous.NFLAG_LAST_APPL_IN_DAY)>,\n",
       " <Feature: MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5']))>,\n",
       " <Feature: MAX_CONSECUTIVE(cash.previous.FLAG_LAST_APPL_PER_CONTRACT)>,\n",
       " <Feature: MAX_CONSECUTIVE(cash.previous.NFLAG_LAST_APPL_IN_DAY)>,\n",
       " <Feature: MAX_CONSECUTIVE(installments.previous.FLAG_LAST_APPL_PER_CONTRACT)>,\n",
       " <Feature: MAX_CONSECUTIVE(installments.previous.NFLAG_LAST_APPL_IN_DAY)>,\n",
       " <Feature: MAX_CONSECUTIVE(credit.previous.FLAG_LAST_APPL_PER_CONTRACT)>,\n",
       " <Feature: MAX_CONSECUTIVE(credit.previous.NFLAG_LAST_APPL_IN_DAY)>]"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[f for f in features if f.primitive.name == 'max_consecutive']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "0b2e966f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<Feature: PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT)>,\n",
       " <Feature: MAX(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT))>,\n",
       " <Feature: MEAN(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT))>,\n",
       " <Feature: MEAN(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT) WHERE NAME_CONTRACT_STATUS = Approved)>,\n",
       " <Feature: MEAN(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT) WHERE NAME_CONTRACT_STATUS = Refused)>,\n",
       " <Feature: STD(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT))>,\n",
       " <Feature: SUM(previous.PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT))>,\n",
       " <Feature: PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT WHERE previous.NAME_CONTRACT_STATUS = Refused)>,\n",
       " <Feature: PERCENT_TRUE(installments.DAYS_ENTRY_PAYMENT > DAYS_INSTALMENT WHERE previous.NAME_CONTRACT_STATUS = Approved)>]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[f for f in features if '>' in f.get_name()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "252c5ae1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<Feature: MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5']))>,\n",
       " <Feature: PERCENT_TRUE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5']))>,\n",
       " <Feature: MAX(bureau.MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: MAX(bureau.PERCENT_TRUE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: MEAN(bureau.MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: MEAN(bureau.PERCENT_TRUE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: STD(bureau.MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: STD(bureau.PERCENT_TRUE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: SUM(bureau.MAX_CONSECUTIVE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>,\n",
       " <Feature: SUM(bureau.PERCENT_TRUE(bureau_balance.STATUS.isin(['1', '2', '3', '4', '5'])))>]"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[f for f in features if 'isin' in f.get_name()]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c0b8e1e",
   "metadata": {},
   "source": [
    "特征按照我们预想的添加了。"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1e8070b6",
   "metadata": {},
   "source": [
    "## modeling\n",
    "- lgbm不需要one-hot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "89e34909",
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_matrix = pd.read_parquet(\"ft_tuning_feature_matrix.parquet\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "44c6005c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(356255, 1891)"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_matrix.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "bdec30fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_fm = feature_matrix.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "f9a7612b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0            1\n",
       "1            0\n",
       "2            0\n",
       "3            0\n",
       "4            0\n",
       "          ... \n",
       "356250    <NA>\n",
       "356251    <NA>\n",
       "356252    <NA>\n",
       "356253    <NA>\n",
       "356254    <NA>\n",
       "Name: TARGET, Length: 356255, dtype: Int64"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_fm['TARGET']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "b06ac067",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(307511, 1892) (48744, 1891)\n"
     ]
    }
   ],
   "source": [
    "final_fm = pd.merge(final_fm, app_set, on='SK_ID_CURR', how='left')\n",
    "\n",
    "train = final_fm[final_fm['set'] == 'train']\n",
    "test = final_fm[final_fm['set'] == 'test']\n",
    "\n",
    "train, test = train.align(test, join = 'inner', axis = 1)\n",
    "train = train.drop(columns=['set'])\n",
    "test = test.drop(columns = ['TARGET', 'set'])\n",
    "print(train.shape, test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "44141c3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_labels = train['TARGET']\n",
    "train_ids = train['SK_ID_CURR']\n",
    "test_ids = test['SK_ID_CURR']\n",
    "train_features = train.drop(columns=['TARGET', 'SK_ID_CURR'])\n",
    "test_features = test.drop(columns=['SK_ID_CURR'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "9b1cd6d8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "# 1. 定义清理函数\n",
    "def clean_names(df):\n",
    "    # 替换所有非字母、数字的字符为下划线\n",
    "    # 这里的正则 [^A-Za-z0-9_] 会匹配空格、斜杠、括号等所有特殊字符\n",
    "    df.columns = [re.sub(r'[^A-Za-z0-9_]+', '_', col) for col in df.columns]\n",
    "    # 顺便处理一下可能出现的重复下划线，比如 __\n",
    "    df.columns = [re.sub(r'_+', '_', col).strip('_') for col in df.columns]\n",
    "    return df\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "408c3601",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Info] Number of positive: 24825, number of negative: 282686\n",
      "[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 2.594608 seconds.\n",
      "You can set `force_col_wise=true` to remove the overhead.\n",
      "[LightGBM] [Info] Total Bins 293128\n",
      "[LightGBM] [Info] Number of data points in the train set: 307511, number of used features: 1666\n",
      "[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.080729 -> initscore=-2.432486\n",
      "[LightGBM] [Info] Start training from score -2.432486\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {\n",
       "  /* Definition of color scheme common for light and dark mode */\n",
       "  --sklearn-color-text: #000;\n",
       "  --sklearn-color-text-muted: #666;\n",
       "  --sklearn-color-line: gray;\n",
       "  /* Definition of color scheme for unfitted estimators */\n",
       "  --sklearn-color-unfitted-level-0: #fff5e6;\n",
       "  --sklearn-color-unfitted-level-1: #f6e4d2;\n",
       "  --sklearn-color-unfitted-level-2: #ffe0b3;\n",
       "  --sklearn-color-unfitted-level-3: chocolate;\n",
       "  /* Definition of color scheme for fitted estimators */\n",
       "  --sklearn-color-fitted-level-0: #f0f8ff;\n",
       "  --sklearn-color-fitted-level-1: #d4ebff;\n",
       "  --sklearn-color-fitted-level-2: #b3dbfd;\n",
       "  --sklearn-color-fitted-level-3: cornflowerblue;\n",
       "\n",
       "  /* Specific color for light theme */\n",
       "  --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, white)));\n",
       "  --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, black)));\n",
       "  --sklearn-color-icon: #696969;\n",
       "\n",
       "  @media (prefers-color-scheme: dark) {\n",
       "    /* Redefinition of color scheme for dark theme */\n",
       "    --sklearn-color-text-on-default-background: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-background: var(--sg-background-color, var(--theme-background, var(--jp-layout-color0, #111)));\n",
       "    --sklearn-color-border-box: var(--sg-text-color, var(--theme-code-foreground, var(--jp-content-font-color1, white)));\n",
       "    --sklearn-color-icon: #878787;\n",
       "  }\n",
       "}\n",
       "\n",
       "#sk-container-id-1 {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 pre {\n",
       "  padding: 0;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-hidden--visually {\n",
       "  border: 0;\n",
       "  clip: rect(1px 1px 1px 1px);\n",
       "  clip: rect(1px, 1px, 1px, 1px);\n",
       "  height: 1px;\n",
       "  margin: -1px;\n",
       "  overflow: hidden;\n",
       "  padding: 0;\n",
       "  position: absolute;\n",
       "  width: 1px;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-dashed-wrapped {\n",
       "  border: 1px dashed var(--sklearn-color-line);\n",
       "  margin: 0 0.4em 0.5em 0.4em;\n",
       "  box-sizing: border-box;\n",
       "  padding-bottom: 0.4em;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-container {\n",
       "  /* jupyter's `normalize.less` sets `[hidden] { display: none; }`\n",
       "     but bootstrap.min.css set `[hidden] { display: none !important; }`\n",
       "     so we also need the `!important` here to be able to override the\n",
       "     default hidden behavior on the sphinx rendered scikit-learn.org.\n",
       "     See: https://github.com/scikit-learn/scikit-learn/issues/21755 */\n",
       "  display: inline-block !important;\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-text-repr-fallback {\n",
       "  display: none;\n",
       "}\n",
       "\n",
       "div.sk-parallel-item,\n",
       "div.sk-serial,\n",
       "div.sk-item {\n",
       "  /* draw centered vertical line to link estimators */\n",
       "  background-image: linear-gradient(var(--sklearn-color-text-on-default-background), var(--sklearn-color-text-on-default-background));\n",
       "  background-size: 2px 100%;\n",
       "  background-repeat: no-repeat;\n",
       "  background-position: center center;\n",
       "}\n",
       "\n",
       "/* Parallel-specific style estimator block */\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item::after {\n",
       "  content: \"\";\n",
       "  width: 100%;\n",
       "  border-bottom: 2px solid var(--sklearn-color-text-on-default-background);\n",
       "  flex-grow: 1;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel {\n",
       "  display: flex;\n",
       "  align-items: stretch;\n",
       "  justify-content: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  position: relative;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:first-child::after {\n",
       "  align-self: flex-end;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:last-child::after {\n",
       "  align-self: flex-start;\n",
       "  width: 50%;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-parallel-item:only-child::after {\n",
       "  width: 0;\n",
       "}\n",
       "\n",
       "/* Serial-specific style estimator block */\n",
       "\n",
       "#sk-container-id-1 div.sk-serial {\n",
       "  display: flex;\n",
       "  flex-direction: column;\n",
       "  align-items: center;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  padding-right: 1em;\n",
       "  padding-left: 1em;\n",
       "}\n",
       "\n",
       "\n",
       "/* Toggleable style: style used for estimator/Pipeline/ColumnTransformer box that is\n",
       "clickable and can be expanded/collapsed.\n",
       "- Pipeline and ColumnTransformer use this feature and define the default style\n",
       "- Estimators will overwrite some part of the style using the `sk-estimator` class\n",
       "*/\n",
       "\n",
       "/* Pipeline and ColumnTransformer style (default) */\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable {\n",
       "  /* Default theme specific background. It is overwritten whether we have a\n",
       "  specific estimator or a Pipeline/ColumnTransformer */\n",
       "  background-color: var(--sklearn-color-background);\n",
       "}\n",
       "\n",
       "/* Toggleable label */\n",
       "#sk-container-id-1 label.sk-toggleable__label {\n",
       "  cursor: pointer;\n",
       "  display: flex;\n",
       "  width: 100%;\n",
       "  margin-bottom: 0;\n",
       "  padding: 0.5em;\n",
       "  box-sizing: border-box;\n",
       "  text-align: center;\n",
       "  align-items: start;\n",
       "  justify-content: space-between;\n",
       "  gap: 0.5em;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label .caption {\n",
       "  font-size: 0.6rem;\n",
       "  font-weight: lighter;\n",
       "  color: var(--sklearn-color-text-muted);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label-arrow:before {\n",
       "  /* Arrow on the left of the label */\n",
       "  content: \"▸\";\n",
       "  float: left;\n",
       "  margin-right: 0.25em;\n",
       "  color: var(--sklearn-color-icon);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {\n",
       "  color: var(--sklearn-color-text);\n",
       "}\n",
       "\n",
       "/* Toggleable content - dropdown */\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content {\n",
       "  max-height: 0;\n",
       "  max-width: 0;\n",
       "  overflow: hidden;\n",
       "  text-align: left;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content pre {\n",
       "  margin: 0.2em;\n",
       "  border-radius: 0.25em;\n",
       "  color: var(--sklearn-color-text);\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-toggleable__content.fitted pre {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {\n",
       "  /* Expand drop-down */\n",
       "  max-height: 200px;\n",
       "  max-width: 100%;\n",
       "  overflow: auto;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {\n",
       "  content: \"▾\";\n",
       "}\n",
       "\n",
       "/* Pipeline/ColumnTransformer-specific style */\n",
       "\n",
       "#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator-specific style */\n",
       "\n",
       "/* Colorize estimator box */\n",
       "#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted input.sk-toggleable__control:checked~label.sk-toggleable__label {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label label.sk-toggleable__label,\n",
       "#sk-container-id-1 div.sk-label label {\n",
       "  /* The background is the default theme color */\n",
       "  color: var(--sklearn-color-text-on-default-background);\n",
       "}\n",
       "\n",
       "/* On hover, darken the color of the background */\n",
       "#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "/* Label box, darken color on hover, fitted */\n",
       "#sk-container-id-1 div.sk-label.fitted:hover label.sk-toggleable__label.fitted {\n",
       "  color: var(--sklearn-color-text);\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Estimator label */\n",
       "\n",
       "#sk-container-id-1 div.sk-label label {\n",
       "  font-family: monospace;\n",
       "  font-weight: bold;\n",
       "  display: inline-block;\n",
       "  line-height: 1.2em;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-label-container {\n",
       "  text-align: center;\n",
       "}\n",
       "\n",
       "/* Estimator-specific */\n",
       "#sk-container-id-1 div.sk-estimator {\n",
       "  font-family: monospace;\n",
       "  border: 1px dotted var(--sklearn-color-border-box);\n",
       "  border-radius: 0.25em;\n",
       "  box-sizing: border-box;\n",
       "  margin-bottom: 0.5em;\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-0);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-0);\n",
       "}\n",
       "\n",
       "/* on hover */\n",
       "#sk-container-id-1 div.sk-estimator:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-2);\n",
       "}\n",
       "\n",
       "#sk-container-id-1 div.sk-estimator.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-2);\n",
       "}\n",
       "\n",
       "/* Specification for estimator info (e.g. \"i\" and \"?\") */\n",
       "\n",
       "/* Common style for \"i\" and \"?\" */\n",
       "\n",
       ".sk-estimator-doc-link,\n",
       "a:link.sk-estimator-doc-link,\n",
       "a:visited.sk-estimator-doc-link {\n",
       "  float: right;\n",
       "  font-size: smaller;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1em;\n",
       "  height: 1em;\n",
       "  width: 1em;\n",
       "  text-decoration: none !important;\n",
       "  margin-left: 0.5em;\n",
       "  text-align: center;\n",
       "  /* unfitted */\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted,\n",
       "a:link.sk-estimator-doc-link.fitted,\n",
       "a:visited.sk-estimator-doc-link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "div.sk-estimator:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link:hover,\n",
       ".sk-estimator-doc-link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "div.sk-estimator.fitted:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover,\n",
       "div.sk-label-container:hover .sk-estimator-doc-link.fitted:hover,\n",
       ".sk-estimator-doc-link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "/* Span, style for the box shown on hovering the info icon */\n",
       ".sk-estimator-doc-link span {\n",
       "  display: none;\n",
       "  z-index: 9999;\n",
       "  position: relative;\n",
       "  font-weight: normal;\n",
       "  right: .2ex;\n",
       "  padding: .5ex;\n",
       "  margin: .5ex;\n",
       "  width: min-content;\n",
       "  min-width: 20ex;\n",
       "  max-width: 50ex;\n",
       "  color: var(--sklearn-color-text);\n",
       "  box-shadow: 2pt 2pt 4pt #999;\n",
       "  /* unfitted */\n",
       "  background: var(--sklearn-color-unfitted-level-0);\n",
       "  border: .5pt solid var(--sklearn-color-unfitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link.fitted span {\n",
       "  /* fitted */\n",
       "  background: var(--sklearn-color-fitted-level-0);\n",
       "  border: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "\n",
       ".sk-estimator-doc-link:hover span {\n",
       "  display: block;\n",
       "}\n",
       "\n",
       "/* \"?\"-specific style due to the `<a>` HTML tag */\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link {\n",
       "  float: right;\n",
       "  font-size: 1rem;\n",
       "  line-height: 1em;\n",
       "  font-family: monospace;\n",
       "  background-color: var(--sklearn-color-background);\n",
       "  border-radius: 1rem;\n",
       "  height: 1rem;\n",
       "  width: 1rem;\n",
       "  text-decoration: none;\n",
       "  /* unfitted */\n",
       "  color: var(--sklearn-color-unfitted-level-1);\n",
       "  border: var(--sklearn-color-unfitted-level-1) 1pt solid;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link.fitted {\n",
       "  /* fitted */\n",
       "  border: var(--sklearn-color-fitted-level-1) 1pt solid;\n",
       "  color: var(--sklearn-color-fitted-level-1);\n",
       "}\n",
       "\n",
       "/* On hover */\n",
       "#sk-container-id-1 a.estimator_doc_link:hover {\n",
       "  /* unfitted */\n",
       "  background-color: var(--sklearn-color-unfitted-level-3);\n",
       "  color: var(--sklearn-color-background);\n",
       "  text-decoration: none;\n",
       "}\n",
       "\n",
       "#sk-container-id-1 a.estimator_doc_link.fitted:hover {\n",
       "  /* fitted */\n",
       "  background-color: var(--sklearn-color-fitted-level-3);\n",
       "}\n",
       "</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>LGBMClassifier(max_depth=3, n_jobs=-1, random_state=42)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator fitted sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label fitted sk-toggleable__label-arrow\"><div><div>LGBMClassifier</div></div><div><span class=\"sk-estimator-doc-link fitted\">i<span>Fitted</span></span></div></label><div class=\"sk-toggleable__content fitted\"><pre>LGBMClassifier(max_depth=3, n_jobs=-1, random_state=42)</pre></div> </div></div></div></div>"
      ],
      "text/plain": [
       "LGBMClassifier(max_depth=3, n_jobs=-1, random_state=42)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features = clean_names(train_features)\n",
    "test_features = clean_names(test_features)\n",
    "\n",
    "from lightgbm import LGBMClassifier\n",
    "lgbm_model = LGBMClassifier(\n",
    "    n_estimators=100,      # 对应 max_iter，树的个数\n",
    "    learning_rate=0.1,     # 学习率\n",
    "    max_depth=3,           # 树的最大深度\n",
    "    random_state=42,       # 保证结果可复现\n",
    ")\n",
    "lgbm_model.fit(train_features, train_labels)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "2beb8ba2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "features_importance = pd.DataFrame(\n",
    "    {\n",
    "        'importance': lgbm_model.feature_importances_,\n",
    "        'feature': lgbm_model.feature_name_\n",
    "    }\n",
    ")\n",
    "features_importance_plot = features_importance.sort_values(by='importance', ascending=False).head(20)\n",
    "\n",
    "plt.figure(figsize=(8, 6), dpi=100) \n",
    "sns.barplot(data=features_importance_plot, x='importance', y='feature')\n",
    "\n",
    "plt.yticks(fontsize=7) # 进一步微调\n",
    "plt.title('Feature Importance', fontsize=14)\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "a2b54eca",
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "import os\n",
    "\n",
    "def submit(ids, pred, name, feature_count=None):\n",
    "    \"\"\"\n",
    "    ids: 测试集的 SK_ID_CURR\n",
    "    pred: 模型预测概率\n",
    "    name: 你的实验备注 (如 'lgb_v1', 'baseline')\n",
    "    feature_count: 可选，记录模型使用了多少个特征\n",
    "    \"\"\"\n",
    "    # 1. 创建提交 DataFrame\n",
    "    submit_df = pd.DataFrame({\n",
    "        'SK_ID_CURR': ids,\n",
    "        'TARGET': pred\n",
    "    })\n",
    "\n",
    "    # 2. 生成时间戳 (格式: 0213_1530)\n",
    "    timestamp = time.strftime(\"%m%d_%H%M\")\n",
    "    \n",
    "    # 3. 构造文件名\n",
    "    # 格式: 0213_1530_lgb_v1_f542.csv\n",
    "    f_str = f\"_f{feature_count}\" if feature_count else \"\"\n",
    "    filename = f\"{timestamp}_{name}{f_str}.csv\"\n",
    "    \n",
    "    # 4. 确保保存目录存在 (可选)\n",
    "    if not os.path.exists('submissions'):\n",
    "        os.makedirs('submissions')\n",
    "    \n",
    "    save_path = os.path.join('submissions', filename)\n",
    "    \n",
    "    # 5. 保存并打印提示\n",
    "    submit_df.to_csv(save_path, index=False)\n",
    "    \n",
    "    return submit_df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "7fbe2ec1",
   "metadata": {},
   "outputs": [],
   "source": [
    "lgbm_model_pred = lgbm_model.predict_proba(test_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "0a9fe02c",
   "metadata": {},
   "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>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>307511</th>\n",
       "      <td>100001</td>\n",
       "      <td>0.072112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>307512</th>\n",
       "      <td>100005</td>\n",
       "      <td>0.162662</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>307513</th>\n",
       "      <td>100013</td>\n",
       "      <td>0.029653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>307514</th>\n",
       "      <td>100028</td>\n",
       "      <td>0.034053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>307515</th>\n",
       "      <td>100038</td>\n",
       "      <td>0.139559</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        SK_ID_CURR    TARGET\n",
       "307511      100001  0.072112\n",
       "307512      100005  0.162662\n",
       "307513      100013  0.029653\n",
       "307514      100028  0.034053\n",
       "307515      100038  0.139559"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submit_df = submit(test['SK_ID_CURR'], lgbm_model_pred[:, 1], \n",
    "    name='lgbm_baseline',\n",
    "    feature_count=train_features.shape[1]\n",
    "    )\n",
    "submit_df.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07b952f4",
   "metadata": {},
   "source": [
    "得分76，差不多"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "54eb6587",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "data-analysis",
   "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": "3.9.25"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}