CG_tcn_iterative_spline.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a398f945",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:12.403973Z",
     "iopub.status.busy": "2023-11-15T02:19:12.403406Z",
     "iopub.status.idle": "2023-11-15T02:19:26.927393Z",
     "shell.execute_reply": "2023-11-15T02:19:26.926017Z"
    },
    "id": "dhW1edGiuwGS",
    "papermill": {
     "duration": 14.536628,
     "end_time": "2023-11-15T02:19:26.930216",
     "exception": false,
     "start_time": "2023-11-15T02:19:12.393588",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/kaggle/input/optiver-trading-at-the-close/public_timeseries_testing_util.py\n",
      "/kaggle/input/optiver-trading-at-the-close/train.csv\n",
      "/kaggle/input/optiver-trading-at-the-close/example_test_files/sample_submission.csv\n",
      "/kaggle/input/optiver-trading-at-the-close/example_test_files/revealed_targets.csv\n",
      "/kaggle/input/optiver-trading-at-the-close/example_test_files/test.csv\n",
      "/kaggle/input/optiver-trading-at-the-close/optiver2023/competition.cpython-310-x86_64-linux-gnu.so\n",
      "/kaggle/input/optiver-trading-at-the-close/optiver2023/__init__.py\n"
     ]
    }
   ],
   "source": [
    "import numpy as np \n",
    "import pandas as pd \n",
    "from sklearn.linear_model import LinearRegression\n",
    "import pandas as pd\n",
    "from sklearn.experimental import enable_iterative_imputer\n",
    "from sklearn.impute import IterativeImputer\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Layer, Conv1D, Input, Dropout, Dense\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.layers import BatchNormalization\n",
    "from itertools import groupby\n",
    "from sklearn.model_selection import train_test_split\n",
    "import lightgbm as lgb\n",
    "import gc\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from itertools import combinations\n",
    "import plotly.express as px\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "from warnings import simplefilter\n",
    "simplefilter(action=\"ignore\", category=pd.errors.PerformanceWarning)\n",
    "import os\n",
    "for dirname, _, filenames in os.walk('/kaggle/input'):\n",
    "    for filename in filenames:\n",
    "        print(os.path.join(dirname, filename))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "26c63611",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:26.949404Z",
     "iopub.status.busy": "2023-11-15T02:19:26.948919Z",
     "iopub.status.idle": "2023-11-15T02:19:26.954666Z",
     "shell.execute_reply": "2023-11-15T02:19:26.953480Z"
    },
    "papermill": {
     "duration": 0.018926,
     "end_time": "2023-11-15T02:19:26.957835",
     "exception": false,
     "start_time": "2023-11-15T02:19:26.938909",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "kaggle_environment = True # True if running on Kaggle, don't forget to add the dataset!\n",
    "\n",
    "if kaggle_environment:\n",
    "    data_path = '/kaggle/input/'\n",
    "else:\n",
    "    data_path = 'kaggle/input/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "613fd176",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:26.977704Z",
     "iopub.status.busy": "2023-11-15T02:19:26.977261Z",
     "iopub.status.idle": "2023-11-15T02:19:51.687282Z",
     "shell.execute_reply": "2023-11-15T02:19:51.685164Z"
    },
    "papermill": {
     "duration": 24.723964,
     "end_time": "2023-11-15T02:19:51.690823",
     "exception": false,
     "start_time": "2023-11-15T02:19:26.966859",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "train = pd.read_csv('/kaggle/input/optiver-trading-at-the-close/train.csv')\n",
    "revealed_targets = pd.read_csv('/kaggle/input/optiver-trading-at-the-close/example_test_files/revealed_targets.csv')\n",
    "test = pd.read_csv('/kaggle/input/optiver-trading-at-the-close/example_test_files/test.csv')\n",
    "sample_submission = pd.read_csv('/kaggle/input/optiver-trading-at-the-close/example_test_files/sample_submission.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0c1f9796",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:51.711161Z",
     "iopub.status.busy": "2023-11-15T02:19:51.710689Z",
     "iopub.status.idle": "2023-11-15T02:19:51.728438Z",
     "shell.execute_reply": "2023-11-15T02:19:51.726606Z"
    },
    "papermill": {
     "duration": 0.031867,
     "end_time": "2023-11-15T02:19:51.732482",
     "exception": false,
     "start_time": "2023-11-15T02:19:51.700615",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "columns_to_difference = ['bid_price', 'ask_price']\n",
    "columns_to_impute = ['far_price', 'near_price']\n",
    "columns_to_interpolate = ['bid_price', 'ask_price']\n",
    "lagged_columns = ['far_price', 'near_price']\n",
    "n_lags = 2\n",
    "start_time = pd.to_datetime('2023-01-01 08:00:00')\n",
    "end_time = pd.to_datetime('2023-01-01 0:00:00')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55f81659",
   "metadata": {
    "papermill": {
     "duration": 0.008498,
     "end_time": "2023-11-15T02:19:51.749426",
     "exception": false,
     "start_time": "2023-11-15T02:19:51.740928",
     "status": "completed"
    },
    "tags": []
   },
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1f13d88b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:51.769007Z",
     "iopub.status.busy": "2023-11-15T02:19:51.768457Z",
     "iopub.status.idle": "2023-11-15T02:19:55.943444Z",
     "shell.execute_reply": "2023-11-15T02:19:55.941345Z"
    },
    "papermill": {
     "duration": 4.188574,
     "end_time": "2023-11-15T02:19:55.946820",
     "exception": false,
     "start_time": "2023-11-15T02:19:51.758246",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.experimental import enable_iterative_imputer\n",
    "from sklearn.impute import IterativeImputer\n",
    "from sklearn.base import BaseEstimator, TransformerMixin\n",
    "from sklearn.pipeline import Pipeline\n",
    "\n",
    "class CustomIterativeImputer(BaseEstimator, TransformerMixin):\n",
    "    def __init__(self, columns, max_iter=10, random_state=0):\n",
    "        self.columns = columns\n",
    "        self.max_iter = max_iter\n",
    "        self.random_state = random_state\n",
    "        self.imputer = None\n",
    "\n",
    "    def fit(self, X, y=None):\n",
    "        # Initialize the IterativeImputer\n",
    "        self.imputer = IterativeImputer(max_iter=self.max_iter, random_state=self.random_state)\n",
    "        self.imputer.fit(X[self.columns])\n",
    "        return self\n",
    "\n",
    "    def transform(self, X):\n",
    "        # Check is fit had been called\n",
    "        if self.imputer is None:\n",
    "            raise RuntimeError(\"You must call fit before calling transform\")\n",
    "        \n",
    "        # Fit the imputer and transform the data\n",
    "        X[self.columns] = self.imputer.transform(X[self.columns])\n",
    "        return X\n",
    "\n",
    "# Define the columns you want to impute\n",
    "columns_to_impute = ['far_price', 'near_price']\n",
    "\n",
    "# Create the pipeline with the custom imputer\n",
    "pipeline = Pipeline([\n",
    "    ('iterative_imputer', CustomIterativeImputer(columns=['far_price', 'near_price'])),\n",
    "    # Add other steps here if necessary\n",
    "])\n",
    "\n",
    "# Use the pipeline on a DataFrame `df`\n",
    "df_processed = pipeline.fit_transform(train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "157faced",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:55.967023Z",
     "iopub.status.busy": "2023-11-15T02:19:55.966584Z",
     "iopub.status.idle": "2023-11-15T02:19:56.009854Z",
     "shell.execute_reply": "2023-11-15T02:19:56.007870Z"
    },
    "papermill": {
     "duration": 0.057915,
     "end_time": "2023-11-15T02:19:56.013603",
     "exception": false,
     "start_time": "2023-11-15T02:19:55.955688",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>stock_id</th>\n",
       "      <th>date_id</th>\n",
       "      <th>seconds_in_bucket</th>\n",
       "      <th>imbalance_size</th>\n",
       "      <th>imbalance_buy_sell_flag</th>\n",
       "      <th>reference_price</th>\n",
       "      <th>matched_size</th>\n",
       "      <th>far_price</th>\n",
       "      <th>near_price</th>\n",
       "      <th>bid_price</th>\n",
       "      <th>bid_size</th>\n",
       "      <th>ask_price</th>\n",
       "      <th>ask_size</th>\n",
       "      <th>wap</th>\n",
       "      <th>target</th>\n",
       "      <th>time_id</th>\n",
       "      <th>row_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3180602.69</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999812</td>\n",
       "      <td>13380276.64</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999812</td>\n",
       "      <td>60651.50</td>\n",
       "      <td>1.000026</td>\n",
       "      <td>8493.03</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-3.029704</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>166603.91</td>\n",
       "      <td>-1</td>\n",
       "      <td>0.999896</td>\n",
       "      <td>1642214.25</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999896</td>\n",
       "      <td>3233.04</td>\n",
       "      <td>1.000660</td>\n",
       "      <td>20605.09</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-5.519986</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>302879.87</td>\n",
       "      <td>-1</td>\n",
       "      <td>0.999561</td>\n",
       "      <td>1819368.03</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999403</td>\n",
       "      <td>37956.00</td>\n",
       "      <td>1.000298</td>\n",
       "      <td>18995.00</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-8.389950</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>11917682.27</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.000171</td>\n",
       "      <td>18389745.62</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>2324.90</td>\n",
       "      <td>1.000214</td>\n",
       "      <td>479032.40</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-4.010200</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>447549.96</td>\n",
       "      <td>-1</td>\n",
       "      <td>0.999532</td>\n",
       "      <td>17860614.95</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999394</td>\n",
       "      <td>16485.54</td>\n",
       "      <td>1.000016</td>\n",
       "      <td>434.10</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-7.349849</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5237975</th>\n",
       "      <td>195</td>\n",
       "      <td>480</td>\n",
       "      <td>540</td>\n",
       "      <td>2440722.89</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.000317</td>\n",
       "      <td>28280361.74</td>\n",
       "      <td>0.999734</td>\n",
       "      <td>0.999734</td>\n",
       "      <td>1.000317</td>\n",
       "      <td>32257.04</td>\n",
       "      <td>1.000434</td>\n",
       "      <td>319862.40</td>\n",
       "      <td>1.000328</td>\n",
       "      <td>2.310276</td>\n",
       "      <td>26454</td>\n",
       "      <td>480_540_195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5237976</th>\n",
       "      <td>196</td>\n",
       "      <td>480</td>\n",
       "      <td>540</td>\n",
       "      <td>349510.47</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.000643</td>\n",
       "      <td>9187699.11</td>\n",
       "      <td>1.000129</td>\n",
       "      <td>1.000386</td>\n",
       "      <td>1.000643</td>\n",
       "      <td>205108.40</td>\n",
       "      <td>1.000900</td>\n",
       "      <td>93393.07</td>\n",
       "      <td>1.000819</td>\n",
       "      <td>-8.220077</td>\n",
       "      <td>26454</td>\n",
       "      <td>480_540_196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5237977</th>\n",
       "      <td>197</td>\n",
       "      <td>480</td>\n",
       "      <td>540</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0</td>\n",
       "      <td>0.995789</td>\n",
       "      <td>12725436.10</td>\n",
       "      <td>0.995789</td>\n",
       "      <td>0.995789</td>\n",
       "      <td>0.995789</td>\n",
       "      <td>16790.66</td>\n",
       "      <td>0.995883</td>\n",
       "      <td>180038.32</td>\n",
       "      <td>0.995797</td>\n",
       "      <td>1.169443</td>\n",
       "      <td>26454</td>\n",
       "      <td>480_540_197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5237978</th>\n",
       "      <td>198</td>\n",
       "      <td>480</td>\n",
       "      <td>540</td>\n",
       "      <td>1000898.84</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999210</td>\n",
       "      <td>94773271.05</td>\n",
       "      <td>0.999210</td>\n",
       "      <td>0.999210</td>\n",
       "      <td>0.998970</td>\n",
       "      <td>125631.72</td>\n",
       "      <td>0.999210</td>\n",
       "      <td>669893.00</td>\n",
       "      <td>0.999008</td>\n",
       "      <td>-1.540184</td>\n",
       "      <td>26454</td>\n",
       "      <td>480_540_198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5237979</th>\n",
       "      <td>199</td>\n",
       "      <td>480</td>\n",
       "      <td>540</td>\n",
       "      <td>1884285.71</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.002129</td>\n",
       "      <td>24073677.32</td>\n",
       "      <td>1.000859</td>\n",
       "      <td>1.001494</td>\n",
       "      <td>1.002129</td>\n",
       "      <td>250081.44</td>\n",
       "      <td>1.002447</td>\n",
       "      <td>300167.56</td>\n",
       "      <td>1.002274</td>\n",
       "      <td>-6.530285</td>\n",
       "      <td>26454</td>\n",
       "      <td>480_540_199</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5237980 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         stock_id  date_id  seconds_in_bucket  imbalance_size  \\\n",
       "0               0        0                  0      3180602.69   \n",
       "1               1        0                  0       166603.91   \n",
       "2               2        0                  0       302879.87   \n",
       "3               3        0                  0     11917682.27   \n",
       "4               4        0                  0       447549.96   \n",
       "...           ...      ...                ...             ...   \n",
       "5237975       195      480                540      2440722.89   \n",
       "5237976       196      480                540       349510.47   \n",
       "5237977       197      480                540            0.00   \n",
       "5237978       198      480                540      1000898.84   \n",
       "5237979       199      480                540      1884285.71   \n",
       "\n",
       "         imbalance_buy_sell_flag  reference_price  matched_size  far_price  \\\n",
       "0                              1         0.999812   13380276.64   1.001772   \n",
       "1                             -1         0.999896    1642214.25   1.001772   \n",
       "2                             -1         0.999561    1819368.03   1.001772   \n",
       "3                             -1         1.000171   18389745.62   1.001772   \n",
       "4                             -1         0.999532   17860614.95   1.001772   \n",
       "...                          ...              ...           ...        ...   \n",
       "5237975                       -1         1.000317   28280361.74   0.999734   \n",
       "5237976                       -1         1.000643    9187699.11   1.000129   \n",
       "5237977                        0         0.995789   12725436.10   0.995789   \n",
       "5237978                        1         0.999210   94773271.05   0.999210   \n",
       "5237979                       -1         1.002129   24073677.32   1.000859   \n",
       "\n",
       "         near_price  bid_price   bid_size  ask_price   ask_size       wap  \\\n",
       "0          0.999660   0.999812   60651.50   1.000026    8493.03  1.000000   \n",
       "1          0.999660   0.999896    3233.04   1.000660   20605.09  1.000000   \n",
       "2          0.999660   0.999403   37956.00   1.000298   18995.00  1.000000   \n",
       "3          0.999660   0.999999    2324.90   1.000214  479032.40  1.000000   \n",
       "4          0.999660   0.999394   16485.54   1.000016     434.10  1.000000   \n",
       "...             ...        ...        ...        ...        ...       ...   \n",
       "5237975    0.999734   1.000317   32257.04   1.000434  319862.40  1.000328   \n",
       "5237976    1.000386   1.000643  205108.40   1.000900   93393.07  1.000819   \n",
       "5237977    0.995789   0.995789   16790.66   0.995883  180038.32  0.995797   \n",
       "5237978    0.999210   0.998970  125631.72   0.999210  669893.00  0.999008   \n",
       "5237979    1.001494   1.002129  250081.44   1.002447  300167.56  1.002274   \n",
       "\n",
       "           target  time_id       row_id  \n",
       "0       -3.029704        0        0_0_0  \n",
       "1       -5.519986        0        0_0_1  \n",
       "2       -8.389950        0        0_0_2  \n",
       "3       -4.010200        0        0_0_3  \n",
       "4       -7.349849        0        0_0_4  \n",
       "...           ...      ...          ...  \n",
       "5237975  2.310276    26454  480_540_195  \n",
       "5237976 -8.220077    26454  480_540_196  \n",
       "5237977  1.169443    26454  480_540_197  \n",
       "5237978 -1.540184    26454  480_540_198  \n",
       "5237979 -6.530285    26454  480_540_199  \n",
       "\n",
       "[5237980 rows x 17 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_processed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c6f5a971",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:56.036772Z",
     "iopub.status.busy": "2023-11-15T02:19:56.035333Z",
     "iopub.status.idle": "2023-11-15T02:19:56.052368Z",
     "shell.execute_reply": "2023-11-15T02:19:56.050511Z"
    },
    "papermill": {
     "duration": 0.031956,
     "end_time": "2023-11-15T02:19:56.055374",
     "exception": false,
     "start_time": "2023-11-15T02:19:56.023418",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "def preprocess_and_interpolate(df, columns_to_interpolate, freq='T', order=3):\n",
    "    \"\"\"\n",
    "    Preprocess and interpolate specified columns of a DataFrame minute-wise grouped by stock_id.\n",
    "\n",
    "    Parameters:\n",
    "    df (pandas.DataFrame): DataFrame to process.\n",
    "    columns_to_interpolate (list of str): Columns to apply interpolation.\n",
    "    freq (str): Frequency for resampling. Default is 'T' for minute-wise.\n",
    "    order (int): Order of spline interpolation. Default is 3.\n",
    "\n",
    "    Returns:\n",
    "    pandas.DataFrame: Processed and interpolated DataFrame.\n",
    "    \"\"\"\n",
    "    # Create a time range for the 8-hour day\n",
    "    start_time = pd.Timestamp('4:00:00')  # Market open at 8 AM\n",
    "    end_time = pd.Timestamp('16:00:00')  # Market close at 4 PM\n",
    "    # Assuming 540 seconds is the time span for the closing auction.\n",
    "    trading_seconds = (end_time - start_time).seconds - 540\n",
    "    seconds_in_day = (end_time - start_time).seconds\n",
    "\n",
    "    # Map 'seconds_in_bucket' to the actual time within the 8-hour window\n",
    "    df['actual_time'] = pd.to_timedelta(df['seconds_in_bucket'], unit='s') + start_time\n",
    "\n",
    "    # Ensure the actual time does not go beyond the market closing time\n",
    "    df['actual_time'] = df['actual_time'].apply(lambda x: x if x <= end_time else end_time)\n",
    "\n",
    "    # Set the timestamp as the index\n",
    "    df.set_index('actual_time', inplace=True)\n",
    "\n",
    "    # Initialize an empty DataFrame to store the results\n",
    "    resampled_dfs = []\n",
    "\n",
    "    # Group by 'stock_id' and process each group\n",
    "    for stock_id, group in df.groupby('stock_id'):\n",
    "        # Handle duplicates\n",
    "        if group.index.duplicated().any():\n",
    "            group = group[~group.index.duplicated(keep='first')]\n",
    "\n",
    "        # Resample the data\n",
    "        group_resampled = group.resample(freq).asfreq()\n",
    "\n",
    "        # Interpolate for each column\n",
    "        for column in columns_to_interpolate:\n",
    "            group_resampled[column] = group_resampled[column].interpolate(method='spline', order=order)\n",
    "        \n",
    "        # Add the stock_id back to the DataFrame\n",
    "        group_resampled['stock_id'] = stock_id\n",
    "\n",
    "        # Append the processed group to the list\n",
    "        resampled_dfs.append(group_resampled)\n",
    "\n",
    "    # Concatenate all the processed groups\n",
    "    df_resampled = pd.concat(resampled_dfs)\n",
    "    df_resampled.reset_index(inplace=True)\n",
    "    df_new = df_resampled.drop(['actual_time'], axis=1)\n",
    "\n",
    "    return df_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8e9470d7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:19:56.077830Z",
     "iopub.status.busy": "2023-11-15T02:19:56.077074Z",
     "iopub.status.idle": "2023-11-15T02:20:30.782480Z",
     "shell.execute_reply": "2023-11-15T02:20:30.781445Z"
    },
    "papermill": {
     "duration": 34.719808,
     "end_time": "2023-11-15T02:20:30.785575",
     "exception": false,
     "start_time": "2023-11-15T02:19:56.065767",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "#trial = df_processed.drop(['time_id'], axis=1)\n",
    "imputed_df = preprocess_and_interpolate(df_processed, ['bid_price', 'ask_price'])\n",
    "y = imputed_df['target'].values\n",
    "no_time = imputed_df.drop(['target'], axis=1)\n",
    "X = no_time.to_numpy(dtype='float32')\n",
    "scaler = StandardScaler()\n",
    "X_scaled = scaler.fit_transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "2c5ac76b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:20:30.806393Z",
     "iopub.status.busy": "2023-11-15T02:20:30.805721Z",
     "iopub.status.idle": "2023-11-15T02:20:30.838685Z",
     "shell.execute_reply": "2023-11-15T02:20:30.837203Z"
    },
    "papermill": {
     "duration": 0.046805,
     "end_time": "2023-11-15T02:20:30.841486",
     "exception": false,
     "start_time": "2023-11-15T02:20:30.794681",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>stock_id</th>\n",
       "      <th>date_id</th>\n",
       "      <th>seconds_in_bucket</th>\n",
       "      <th>imbalance_size</th>\n",
       "      <th>imbalance_buy_sell_flag</th>\n",
       "      <th>reference_price</th>\n",
       "      <th>matched_size</th>\n",
       "      <th>far_price</th>\n",
       "      <th>near_price</th>\n",
       "      <th>bid_price</th>\n",
       "      <th>bid_size</th>\n",
       "      <th>ask_price</th>\n",
       "      <th>ask_size</th>\n",
       "      <th>wap</th>\n",
       "      <th>target</th>\n",
       "      <th>time_id</th>\n",
       "      <th>row_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3180602.69</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999812</td>\n",
       "      <td>13380276.64</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999812</td>\n",
       "      <td>60651.50</td>\n",
       "      <td>1.000026</td>\n",
       "      <td>8493.03</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-3.029704</td>\n",
       "      <td>0</td>\n",
       "      <td>0_0_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>60</td>\n",
       "      <td>1218204.43</td>\n",
       "      <td>1</td>\n",
       "      <td>1.000562</td>\n",
       "      <td>15342674.90</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>1.000455</td>\n",
       "      <td>14005.50</td>\n",
       "      <td>1.000562</td>\n",
       "      <td>10085.04</td>\n",
       "      <td>1.000517</td>\n",
       "      <td>-0.200272</td>\n",
       "      <td>6</td>\n",
       "      <td>0_60_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>120</td>\n",
       "      <td>1216057.90</td>\n",
       "      <td>1</td>\n",
       "      <td>1.000133</td>\n",
       "      <td>15460827.57</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999812</td>\n",
       "      <td>43482.46</td>\n",
       "      <td>1.000026</td>\n",
       "      <td>68224.23</td>\n",
       "      <td>0.999895</td>\n",
       "      <td>2.290011</td>\n",
       "      <td>12</td>\n",
       "      <td>0_120_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>180</td>\n",
       "      <td>1445736.98</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999705</td>\n",
       "      <td>15642349.64</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999598</td>\n",
       "      <td>18658.00</td>\n",
       "      <td>0.999705</td>\n",
       "      <td>9330.00</td>\n",
       "      <td>0.999669</td>\n",
       "      <td>-3.029704</td>\n",
       "      <td>18</td>\n",
       "      <td>0_180_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>240</td>\n",
       "      <td>1783022.73</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999705</td>\n",
       "      <td>15687146.86</td>\n",
       "      <td>1.001772</td>\n",
       "      <td>0.999660</td>\n",
       "      <td>0.999491</td>\n",
       "      <td>37778.40</td>\n",
       "      <td>0.999705</td>\n",
       "      <td>37320.00</td>\n",
       "      <td>0.999598</td>\n",
       "      <td>2.130270</td>\n",
       "      <td>24</td>\n",
       "      <td>0_240_0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1995</th>\n",
       "      <td>199</td>\n",
       "      <td>88</td>\n",
       "      <td>300</td>\n",
       "      <td>16037831.49</td>\n",
       "      <td>1</td>\n",
       "      <td>1.000818</td>\n",
       "      <td>16836941.87</td>\n",
       "      <td>1.229196</td>\n",
       "      <td>1.100820</td>\n",
       "      <td>1.000818</td>\n",
       "      <td>197658.00</td>\n",
       "      <td>1.001218</td>\n",
       "      <td>4305.16</td>\n",
       "      <td>1.001210</td>\n",
       "      <td>37.330390</td>\n",
       "      <td>4870</td>\n",
       "      <td>88_300_199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1996</th>\n",
       "      <td>199</td>\n",
       "      <td>88</td>\n",
       "      <td>360</td>\n",
       "      <td>11933481.61</td>\n",
       "      <td>1</td>\n",
       "      <td>1.005618</td>\n",
       "      <td>20929716.97</td>\n",
       "      <td>1.240886</td>\n",
       "      <td>1.106020</td>\n",
       "      <td>1.005618</td>\n",
       "      <td>32883.12</td>\n",
       "      <td>1.006018</td>\n",
       "      <td>17605.00</td>\n",
       "      <td>1.005879</td>\n",
       "      <td>1.660585</td>\n",
       "      <td>4876</td>\n",
       "      <td>88_360_199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1997</th>\n",
       "      <td>199</td>\n",
       "      <td>88</td>\n",
       "      <td>420</td>\n",
       "      <td>10558331.77</td>\n",
       "      <td>1</td>\n",
       "      <td>1.006018</td>\n",
       "      <td>22304866.80</td>\n",
       "      <td>1.241785</td>\n",
       "      <td>1.106420</td>\n",
       "      <td>1.005618</td>\n",
       "      <td>70366.86</td>\n",
       "      <td>1.006018</td>\n",
       "      <td>31839.90</td>\n",
       "      <td>1.005894</td>\n",
       "      <td>-7.939935</td>\n",
       "      <td>4882</td>\n",
       "      <td>88_420_199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1998</th>\n",
       "      <td>199</td>\n",
       "      <td>88</td>\n",
       "      <td>480</td>\n",
       "      <td>2195659.82</td>\n",
       "      <td>1</td>\n",
       "      <td>1.005218</td>\n",
       "      <td>30667538.75</td>\n",
       "      <td>1.006818</td>\n",
       "      <td>1.006418</td>\n",
       "      <td>1.004818</td>\n",
       "      <td>49787.84</td>\n",
       "      <td>1.005218</td>\n",
       "      <td>11308.50</td>\n",
       "      <td>1.005144</td>\n",
       "      <td>-4.569888</td>\n",
       "      <td>4888</td>\n",
       "      <td>88_480_199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1999</th>\n",
       "      <td>199</td>\n",
       "      <td>88</td>\n",
       "      <td>540</td>\n",
       "      <td>1347450.34</td>\n",
       "      <td>1</td>\n",
       "      <td>1.005218</td>\n",
       "      <td>31515748.23</td>\n",
       "      <td>1.006418</td>\n",
       "      <td>1.006018</td>\n",
       "      <td>1.004818</td>\n",
       "      <td>234319.36</td>\n",
       "      <td>1.005218</td>\n",
       "      <td>82853.61</td>\n",
       "      <td>1.005114</td>\n",
       "      <td>-7.860065</td>\n",
       "      <td>4894</td>\n",
       "      <td>88_540_199</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2000 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      stock_id  date_id  seconds_in_bucket  imbalance_size  \\\n",
       "0            0        0                  0      3180602.69   \n",
       "1            0        0                 60      1218204.43   \n",
       "2            0        0                120      1216057.90   \n",
       "3            0        0                180      1445736.98   \n",
       "4            0        0                240      1783022.73   \n",
       "...        ...      ...                ...             ...   \n",
       "1995       199       88                300     16037831.49   \n",
       "1996       199       88                360     11933481.61   \n",
       "1997       199       88                420     10558331.77   \n",
       "1998       199       88                480      2195659.82   \n",
       "1999       199       88                540      1347450.34   \n",
       "\n",
       "      imbalance_buy_sell_flag  reference_price  matched_size  far_price  \\\n",
       "0                           1         0.999812   13380276.64   1.001772   \n",
       "1                           1         1.000562   15342674.90   1.001772   \n",
       "2                           1         1.000133   15460827.57   1.001772   \n",
       "3                           1         0.999705   15642349.64   1.001772   \n",
       "4                           1         0.999705   15687146.86   1.001772   \n",
       "...                       ...              ...           ...        ...   \n",
       "1995                        1         1.000818   16836941.87   1.229196   \n",
       "1996                        1         1.005618   20929716.97   1.240886   \n",
       "1997                        1         1.006018   22304866.80   1.241785   \n",
       "1998                        1         1.005218   30667538.75   1.006818   \n",
       "1999                        1         1.005218   31515748.23   1.006418   \n",
       "\n",
       "      near_price  bid_price   bid_size  ask_price  ask_size       wap  \\\n",
       "0       0.999660   0.999812   60651.50   1.000026   8493.03  1.000000   \n",
       "1       0.999660   1.000455   14005.50   1.000562  10085.04  1.000517   \n",
       "2       0.999660   0.999812   43482.46   1.000026  68224.23  0.999895   \n",
       "3       0.999660   0.999598   18658.00   0.999705   9330.00  0.999669   \n",
       "4       0.999660   0.999491   37778.40   0.999705  37320.00  0.999598   \n",
       "...          ...        ...        ...        ...       ...       ...   \n",
       "1995    1.100820   1.000818  197658.00   1.001218   4305.16  1.001210   \n",
       "1996    1.106020   1.005618   32883.12   1.006018  17605.00  1.005879   \n",
       "1997    1.106420   1.005618   70366.86   1.006018  31839.90  1.005894   \n",
       "1998    1.006418   1.004818   49787.84   1.005218  11308.50  1.005144   \n",
       "1999    1.006018   1.004818  234319.36   1.005218  82853.61  1.005114   \n",
       "\n",
       "         target  time_id      row_id  \n",
       "0     -3.029704        0       0_0_0  \n",
       "1     -0.200272        6      0_60_0  \n",
       "2      2.290011       12     0_120_0  \n",
       "3     -3.029704       18     0_180_0  \n",
       "4      2.130270       24     0_240_0  \n",
       "...         ...      ...         ...  \n",
       "1995  37.330390     4870  88_300_199  \n",
       "1996   1.660585     4876  88_360_199  \n",
       "1997  -7.939935     4882  88_420_199  \n",
       "1998  -4.569888     4888  88_480_199  \n",
       "1999  -7.860065     4894  88_540_199  \n",
       "\n",
       "[2000 rows x 17 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imputed_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "4752ce0b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:20:30.861290Z",
     "iopub.status.busy": "2023-11-15T02:20:30.860194Z",
     "iopub.status.idle": "2023-11-15T02:20:30.868225Z",
     "shell.execute_reply": "2023-11-15T02:20:30.866762Z"
    },
    "papermill": {
     "duration": 0.020859,
     "end_time": "2023-11-15T02:20:30.870998",
     "exception": false,
     "start_time": "2023-11-15T02:20:30.850139",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "def generate_features(cumulative_test_df, current_test, pipeline):\n",
    "    transformed_df = pipeline.transform(cumulative_test_df)\n",
    "    # Only return rows corresponding to the current test dataframe\n",
    "    transformed_df = preprocess_and_interpolate(transformed_df, ['bid_price', 'ask_price'])\n",
    "    test_df = transformed_df.iloc[-len(current_test):]\n",
    "    #print(test_df)\n",
    "    test_df.to_numpy(dtype='float32')\n",
    "    output = scaler.fit_transform(test_df)\n",
    "    return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d20440aa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:20:30.890726Z",
     "iopub.status.busy": "2023-11-15T02:20:30.890302Z",
     "iopub.status.idle": "2023-11-15T02:21:17.751297Z",
     "shell.execute_reply": "2023-11-15T02:21:17.750148Z"
    },
    "papermill": {
     "duration": 46.874087,
     "end_time": "2023-11-15T02:21:17.753809",
     "exception": false,
     "start_time": "2023-11-15T02:20:30.879722",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, None, 1)]         0         \n",
      "                                                                 \n",
      " tcn (TCN)                   (None, None, 64)          146240    \n",
      "                                                                 \n",
      " dense (Dense)               (None, None, 1)           65        \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 146,305\n",
      "Trainable params: 145,281\n",
      "Non-trainable params: 1,024\n",
      "_________________________________________________________________\n",
      "Epoch 1/17\n",
      "63/63 [==============================] - 9s 40ms/step - loss: 6.7373\n",
      "Epoch 2/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.6286\n",
      "Epoch 3/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.6054\n",
      "Epoch 4/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.5884\n",
      "Epoch 5/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.5777\n",
      "Epoch 6/17\n",
      "63/63 [==============================] - 2s 35ms/step - loss: 6.5673\n",
      "Epoch 7/17\n",
      "63/63 [==============================] - 2s 38ms/step - loss: 6.5674\n",
      "Epoch 8/17\n",
      "63/63 [==============================] - 2s 38ms/step - loss: 6.5631\n",
      "Epoch 9/17\n",
      "63/63 [==============================] - 2s 37ms/step - loss: 6.5585\n",
      "Epoch 10/17\n",
      "63/63 [==============================] - 2s 37ms/step - loss: 6.5577\n",
      "Epoch 11/17\n",
      "63/63 [==============================] - 2s 37ms/step - loss: 6.5595\n",
      "Epoch 12/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.5608\n",
      "Epoch 13/17\n",
      "63/63 [==============================] - 2s 37ms/step - loss: 6.5564\n",
      "Epoch 14/17\n",
      "63/63 [==============================] - 2s 36ms/step - loss: 6.5510\n",
      "Epoch 15/17\n",
      "63/63 [==============================] - 3s 40ms/step - loss: 6.5500\n",
      "Epoch 16/17\n",
      "63/63 [==============================] - 2s 37ms/step - loss: 6.5493\n",
      "Epoch 17/17\n",
      "63/63 [==============================] - 2s 35ms/step - loss: 6.5590\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x79d0b6535480>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Layer, Conv1D, Input, Dropout, Dense\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.layers import BatchNormalization\n",
    "\n",
    "\n",
    "class TemporalBlock(Layer):\n",
    "    def __init__(self, n_filters, kernel_size, dilation_rate):\n",
    "        super(TemporalBlock, self).__init__()\n",
    "        self.conv1 = Conv1D(\n",
    "            filters=n_filters,\n",
    "            kernel_size=kernel_size,\n",
    "            padding='causal',\n",
    "            dilation_rate=dilation_rate\n",
    "        )\n",
    "        self.batch_norm1 = BatchNormalization()\n",
    "        self.conv2 = Conv1D(\n",
    "            filters=n_filters,\n",
    "            kernel_size=kernel_size,\n",
    "            padding='causal',\n",
    "            dilation_rate=dilation_rate\n",
    "        )\n",
    "        self.batch_norm2 = BatchNormalization()\n",
    "        self.dropout = Dropout(0.5)\n",
    "    \n",
    "    def call(self, inputs, training=False):\n",
    "        x = self.conv1(inputs)\n",
    "        x = self.batch_norm1(x, training=training)\n",
    "        x = tf.nn.relu(x)\n",
    "        x = self.dropout(x, training=training)\n",
    "        x = self.conv2(x)\n",
    "        x = self.batch_norm2(x, training=training)\n",
    "        x = tf.nn.relu(x)\n",
    "        x = self.dropout(x, training=training)\n",
    "        return x + inputs  # Residual connection\n",
    "\n",
    "class TCN(Layer):\n",
    "    def __init__(self, n_filters, kernel_size, dilation_rates):\n",
    "        super(TCN, self).__init__()\n",
    "        self.blocks = [TemporalBlock(n_filters, kernel_size, dr) for dr in dilation_rates]\n",
    "\n",
    "    def call(self, inputs, training=False):\n",
    "        x = inputs\n",
    "        for block in self.blocks:\n",
    "            x = block(x, training=training)\n",
    "        return x\n",
    "\n",
    "# Define the TCN model\n",
    "n_filters = 64\n",
    "kernel_size = 5\n",
    "dilation_rates = [1, 2, 4, 8]  # exadilation rates\n",
    "\n",
    "input_layer = Input(shape=(None, 1))  # univariate time series; change 1 to n_features if multivariate\n",
    "tcn_layer = TCN(n_filters, kernel_size, dilation_rates)(input_layer)\n",
    "output_layer = Dense(1)(tcn_layer)  # regression task\n",
    "\n",
    "m = Model(inputs=input_layer, outputs=output_layer)\n",
    "optimizer = tf.keras.optimizers.Adam(learning_rate=1e-3, clipvalue=1.0)\n",
    "m.compile(optimizer=optimizer, loss='mae')\n",
    "m.summary()\n",
    "early_stopping = EarlyStopping(monitor='loss', patience=10, restore_best_weights=True)\n",
    "m.fit(X_scaled, y, epochs=17, callbacks=[early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4f35f7ae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:17.885846Z",
     "iopub.status.busy": "2023-11-15T02:21:17.885316Z",
     "iopub.status.idle": "2023-11-15T02:21:19.104458Z",
     "shell.execute_reply": "2023-11-15T02:21:19.102986Z"
    },
    "papermill": {
     "duration": 1.289351,
     "end_time": "2023-11-15T02:21:19.107296",
     "exception": false,
     "start_time": "2023-11-15T02:21:17.817945",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 1s 9ms/step\n"
     ]
    }
   ],
   "source": [
    "ok = m.predict(X_scaled)\n",
    "y_pred = ok[:, 0, 0] \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "bedcdca4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:19.236493Z",
     "iopub.status.busy": "2023-11-15T02:21:19.235323Z",
     "iopub.status.idle": "2023-11-15T02:21:19.242083Z",
     "shell.execute_reply": "2023-11-15T02:21:19.241011Z"
    },
    "papermill": {
     "duration": 0.073529,
     "end_time": "2023-11-15T02:21:19.244353",
     "exception": false,
     "start_time": "2023-11-15T02:21:19.170824",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2000\n"
     ]
    }
   ],
   "source": [
    "print(len(y))\n",
    "\n",
    "y_test = y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "ed641440",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:19.438712Z",
     "iopub.status.busy": "2023-11-15T02:21:19.437177Z",
     "iopub.status.idle": "2023-11-15T02:21:19.864697Z",
     "shell.execute_reply": "2023-11-15T02:21:19.863426Z"
    },
    "papermill": {
     "duration": 0.495717,
     "end_time": "2023-11-15T02:21:19.867702",
     "exception": false,
     "start_time": "2023-11-15T02:21:19.371985",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.figure(figsize=(10, 6))\n",
    "plt.plot(y_test, label='Actual')\n",
    "plt.plot(y_pred, label='Predicted')\n",
    "plt.title('Actual vs Predicted Values')\n",
    "plt.xlabel('Samples')\n",
    "plt.ylabel('Values')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b3d9a5bd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:20.012424Z",
     "iopub.status.busy": "2023-11-15T02:21:20.011985Z",
     "iopub.status.idle": "2023-11-15T02:21:20.017651Z",
     "shell.execute_reply": "2023-11-15T02:21:20.016774Z"
    },
    "papermill": {
     "duration": 0.078829,
     "end_time": "2023-11-15T02:21:20.020234",
     "exception": false,
     "start_time": "2023-11-15T02:21:19.941405",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2000,)\n",
      "(2000,)\n"
     ]
    }
   ],
   "source": [
    "print(y_test.shape)\n",
    "print(y_pred.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bb4b667a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:20.162101Z",
     "iopub.status.busy": "2023-11-15T02:21:20.161378Z",
     "iopub.status.idle": "2023-11-15T02:21:20.170543Z",
     "shell.execute_reply": "2023-11-15T02:21:20.168764Z"
    },
    "papermill": {
     "duration": 0.08674,
     "end_time": "2023-11-15T02:21:20.173733",
     "exception": false,
     "start_time": "2023-11-15T02:21:20.086993",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE: 6.538532251716811\n",
      "MSE: 113.3199041906764\n",
      "R2 Score: -0.00031303672672100546\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score\n",
    "\n",
    "mae = mean_absolute_error(y_test, y_pred)\n",
    "mse = mean_squared_error(y_test, y_pred)\n",
    "r2 = r2_score(y_test, y_pred)\n",
    "\n",
    "print(f\"MAE: {mae}\")\n",
    "print(f\"MSE: {mse}\")\n",
    "print(f\"R2 Score: {r2}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "a96f6167",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:20.314276Z",
     "iopub.status.busy": "2023-11-15T02:21:20.313495Z",
     "iopub.status.idle": "2023-11-15T02:21:20.335823Z",
     "shell.execute_reply": "2023-11-15T02:21:20.334928Z"
    },
    "papermill": {
     "duration": 0.096145,
     "end_time": "2023-11-15T02:21:20.338802",
     "exception": false,
     "start_time": "2023-11-15T02:21:20.242657",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "if kaggle_environment:\n",
    "    import optiver2023\n",
    "    env = optiver2023.make_env()\n",
    "    iter_test = env.iter_test()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "a91a545e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:20.479884Z",
     "iopub.status.busy": "2023-11-15T02:21:20.479035Z",
     "iopub.status.idle": "2023-11-15T02:21:20.483778Z",
     "shell.execute_reply": "2023-11-15T02:21:20.482958Z"
    },
    "papermill": {
     "duration": 0.07909,
     "end_time": "2023-11-15T02:21:20.486264",
     "exception": false,
     "start_time": "2023-11-15T02:21:20.407174",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# test_in= generate_features(cumulative_test_df, test_in,pipeline)\n",
    "# ok = m.predict(test_in)\n",
    "# print(ok[:, 0, 0]  )\n",
    "# sample_prediction[\"target\"] =  ok[:, 0, 0] \n",
    "\n",
    "# print(sample_prediction)\n",
    "# #env.predict(sample_prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "92da4dcd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-11-15T02:21:20.626698Z",
     "iopub.status.busy": "2023-11-15T02:21:20.625946Z",
     "iopub.status.idle": "2023-11-15T02:25:03.312955Z",
     "shell.execute_reply": "2023-11-15T02:25:03.311763Z"
    },
    "papermill": {
     "duration": 222.760446,
     "end_time": "2023-11-15T02:25:03.315795",
     "exception": false,
     "start_time": "2023-11-15T02:21:20.555349",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This version of the API is not optimized and should not be used to estimate the runtime of your code on the hidden test set.\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 14ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 12ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 11ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 11ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 10ms/step\n",
      "7/7 [==============================] - 0s 8ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 9ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 7ms/step\n",
      "7/7 [==============================] - 0s 6ms/step\n",
      "\n",
      "==================================================\n",
      "counter: 165\n"
     ]
    }
   ],
   "source": [
    "if kaggle_environment:\n",
    "    # To count how many time the \"for loop\" runs.\n",
    "    counter = 0\n",
    "\n",
    "    # init 3 empty lists\n",
    "    test_ls, revealed_targets_ls, sample_prediction_ls = [], [], []\n",
    "    cumulative_test_df = pd.DataFrame()\n",
    "\n",
    "    for (test_in, revealed_targets, sample_prediction) in iter_test:\n",
    "        # Append the dataframe that API return into the list.\n",
    "        test_ls.append(test_in.copy())\n",
    "        revealed_targets_ls.append(revealed_targets.copy())\n",
    "        sample_prediction_ls.append(sample_prediction.copy())\n",
    "        cumulative_test_df = pd.concat([cumulative_test_df, test_in], axis=0, ignore_index=True)\n",
    "\n",
    "        # Generate features\n",
    "        test_transformed = generate_features(cumulative_test_df, test_in, pipeline)\n",
    "        predictions =  m.predict(test_transformed)\n",
    "\n",
    "\n",
    "        # Writes our predictions \n",
    "        sample_prediction[\"target\"] = predictions[:, 0, 0] \n",
    "        \n",
    "        # This line submit our predictions.\n",
    "        env.predict(sample_prediction)\n",
    "        counter += 1\n",
    "\n",
    "    print('\\n', '=' * 50, sep=\"\")\n",
    "    print(f\"counter: {counter}\")"
   ]
  }
 ],
 "metadata": {
  "kaggle": {
   "accelerator": "none",
   "dataSources": [
    {
     "databundleVersionId": 6640818,
     "sourceId": 57891,
     "sourceType": "competition"
    }
   ],
   "dockerImageVersionId": 30558,
   "isGpuEnabled": false,
   "isInternetEnabled": false,
   "language": "python",
   "sourceType": "notebook"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 358.225093,
   "end_time": "2023-11-15T02:25:06.299788",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2023-11-15T02:19:08.074695",
   "version": "2.4.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}