{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a3925b02-d72f-4936-aa50-bfa50c2487ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from neuralforecast.core import NeuralForecast\n",
    "from neuralforecast.models import TSMixer, TSMixerx, NHITS, MLPMultivariate, NBEATSx\n",
    "from neuralforecast.losses.pytorch import MSE, MAE, MAPE\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "import numpy as np\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "b28bd72a-d97a-4511-8130-3f0bb4fdccd4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Functions\n",
    "\n",
    "def createLag(data, amt=10):\n",
    "    \"\"\"\n",
    "    Create a lag inside dataframe, in business days\n",
    "\n",
    "    Input:\n",
    "        data -> Pandas dataframe \n",
    "        amt -> int\n",
    "\n",
    "    Output:\n",
    "        Copy of pandas Dataframe\n",
    "    \"\"\"\n",
    "    if 'ds' in data:\n",
    "        copy = data.copy()\n",
    "        copy['ds'] = copy['ds'] + pd.tseries.offsets.BusinessDay(amt)\n",
    "        return copy\n",
    "    else:\n",
    "        print(f\"No 'ds' column found inside dataframe\")\n",
    "        return data\n",
    "\n",
    "def trainTestValSplit(data, test_size, val_size):\n",
    "    \"\"\"\n",
    "    Splits data into train-test-validation sets\n",
    "\n",
    "    Input:\n",
    "        data -> Pandas dataframe\n",
    "        test_size -> Proportion of data for test set\n",
    "        val_size -> Proportiion of data fro validation set\n",
    "\n",
    "    Output:\n",
    "        This is not needed yet, actually\n",
    "    \"\"\"\n",
    "    pass\n",
    "\n",
    "def scaleStandard(df_col):\n",
    "    \"\"\"\n",
    "    Fits and returns a standard scaled version of a dataframe column\n",
    "    \"\"\"\n",
    "    scaler = StandardScaler()\n",
    "    df_col = scaler.fit_transform(df_col)\n",
    "    df_col = pd.DataFrame(df_col)\n",
    "    return df_col, scaler\n",
    "\n",
    "def logReturn(data, df_col):\n",
    "    \"\"\"\n",
    "    Perform log return for a dataframe column\n",
    "    \"\"\"\n",
    "    new_col = np.log1p(data[df_col].pct_change())\n",
    "    return new_col\n",
    "\n",
    "def transformData(data, log_return = [], standard_scale = []):\n",
    "    \"\"\"\n",
    "    Perform essential transformations towards the dataframe\n",
    "    \"\"\"\n",
    "    y_log_ret = False\n",
    "    y_std_scale = False\n",
    "\n",
    "    data.sort_values(by='ds', inplace=True)\n",
    "\n",
    "    if len(log_return) > 0:\n",
    "        \n",
    "        for col1 in log_return:\n",
    "            try:\n",
    "                #print(data[col1])\n",
    "                data[col1] = logReturn(data, col1)\n",
    "            except Exception as e:\n",
    "                print(e)\n",
    "                pass\n",
    "            \n",
    "        if 'y' in log_return:\n",
    "            y_log_ret = True\n",
    "\n",
    "    if len(standard_scale) > 0:\n",
    "        \n",
    "        for col2 in standard_scale:\n",
    "            try:\n",
    "                data[col2], _ = scaleStandard(data[[col2]])\n",
    "            except Exception as e:\n",
    "                print(e)\n",
    "                pass\n",
    "            \n",
    "        if 'y' in standard_scale:\n",
    "            data['y'], yScaler = scaleStandard(data[['y']])\n",
    "            y_std_scale = True\n",
    "\n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "dc3678db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total length is 1962, with validation and test size of 49 for each\n"
     ]
    }
   ],
   "source": [
    "# Exogenous \n",
    "\n",
    "Y_df = pd.read_csv(os.path.join('dataset', 'DatedBrent', 'priceForecast_nosent_1.csv')\n",
    "                  ).rename({'date' : 'ds', 'BrDa' : 'y'}, axis=1\n",
    "                          ).drop(columns=['Unnamed: 0'])\n",
    "Y_df['unique_id'] = 'Dated'\n",
    "Y_df['ds'] = pd.to_datetime(Y_df['ds'])\n",
    "\n",
    "# We make validation and test splits\n",
    "n_time = len(Y_df.ds.unique())\n",
    "val_size = int(.025 * n_time)\n",
    "test_size = int(.025 * n_time)\n",
    "\n",
    "print(f'Total length is {n_time}, with validation and test size of {val_size} for each')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "41f4b728",
   "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>ds</th>\n",
       "      <th>BrFu</th>\n",
       "      <th>Gas</th>\n",
       "      <th>OVX</th>\n",
       "      <th>DXY</th>\n",
       "      <th>GPRD</th>\n",
       "      <th>y</th>\n",
       "      <th>unique_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2024-08-02</td>\n",
       "      <td>73.52</td>\n",
       "      <td>2.3176</td>\n",
       "      <td>33.419998</td>\n",
       "      <td>103.209999</td>\n",
       "      <td>207.799438</td>\n",
       "      <td>78.265</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2024-08-01</td>\n",
       "      <td>76.93</td>\n",
       "      <td>2.3980</td>\n",
       "      <td>33.689999</td>\n",
       "      <td>104.419998</td>\n",
       "      <td>139.878098</td>\n",
       "      <td>81.825</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2024-07-31</td>\n",
       "      <td>77.91</td>\n",
       "      <td>2.4425</td>\n",
       "      <td>31.490000</td>\n",
       "      <td>104.099998</td>\n",
       "      <td>135.206848</td>\n",
       "      <td>81.450</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2024-07-30</td>\n",
       "      <td>74.73</td>\n",
       "      <td>2.3443</td>\n",
       "      <td>30.820000</td>\n",
       "      <td>104.550003</td>\n",
       "      <td>95.696396</td>\n",
       "      <td>79.140</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2024-07-26</td>\n",
       "      <td>77.16</td>\n",
       "      <td>2.4170</td>\n",
       "      <td>26.490000</td>\n",
       "      <td>104.320000</td>\n",
       "      <td>105.654129</td>\n",
       "      <td>81.245</td>\n",
       "      <td>Dated</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1957</th>\n",
       "      <td>2014-08-15</td>\n",
       "      <td>97.35</td>\n",
       "      <td>2.5750</td>\n",
       "      <td>17.860001</td>\n",
       "      <td>81.419998</td>\n",
       "      <td>131.822342</td>\n",
       "      <td>100.775</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1958</th>\n",
       "      <td>2014-08-14</td>\n",
       "      <td>95.58</td>\n",
       "      <td>2.5403</td>\n",
       "      <td>17.190001</td>\n",
       "      <td>81.589996</td>\n",
       "      <td>106.742241</td>\n",
       "      <td>100.445</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1959</th>\n",
       "      <td>2014-08-13</td>\n",
       "      <td>97.59</td>\n",
       "      <td>2.6266</td>\n",
       "      <td>17.549999</td>\n",
       "      <td>81.599998</td>\n",
       "      <td>165.846603</td>\n",
       "      <td>101.855</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1960</th>\n",
       "      <td>2014-08-12</td>\n",
       "      <td>97.37</td>\n",
       "      <td>2.6045</td>\n",
       "      <td>17.540001</td>\n",
       "      <td>81.500000</td>\n",
       "      <td>150.571259</td>\n",
       "      <td>101.725</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1961</th>\n",
       "      <td>2014-08-08</td>\n",
       "      <td>97.65</td>\n",
       "      <td>2.7537</td>\n",
       "      <td>18.850000</td>\n",
       "      <td>81.389999</td>\n",
       "      <td>212.396439</td>\n",
       "      <td>103.315</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1962 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             ds   BrFu     Gas        OVX         DXY        GPRD        y  \\\n",
       "0    2024-08-02  73.52  2.3176  33.419998  103.209999  207.799438   78.265   \n",
       "1    2024-08-01  76.93  2.3980  33.689999  104.419998  139.878098   81.825   \n",
       "2    2024-07-31  77.91  2.4425  31.490000  104.099998  135.206848   81.450   \n",
       "3    2024-07-30  74.73  2.3443  30.820000  104.550003   95.696396   79.140   \n",
       "4    2024-07-26  77.16  2.4170  26.490000  104.320000  105.654129   81.245   \n",
       "...         ...    ...     ...        ...         ...         ...      ...   \n",
       "1957 2014-08-15  97.35  2.5750  17.860001   81.419998  131.822342  100.775   \n",
       "1958 2014-08-14  95.58  2.5403  17.190001   81.589996  106.742241  100.445   \n",
       "1959 2014-08-13  97.59  2.6266  17.549999   81.599998  165.846603  101.855   \n",
       "1960 2014-08-12  97.37  2.6045  17.540001   81.500000  150.571259  101.725   \n",
       "1961 2014-08-08  97.65  2.7537  18.850000   81.389999  212.396439  103.315   \n",
       "\n",
       "     unique_id  \n",
       "0        Dated  \n",
       "1        Dated  \n",
       "2        Dated  \n",
       "3        Dated  \n",
       "4        Dated  \n",
       "...        ...  \n",
       "1957     Dated  \n",
       "1958     Dated  \n",
       "1959     Dated  \n",
       "1960     Dated  \n",
       "1961     Dated  \n",
       "\n",
       "[1962 rows x 8 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "11685932-4158-49d9-861c-0b917d9f45e9",
   "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>ds</th>\n",
       "      <th>y</th>\n",
       "      <th>BrFu</th>\n",
       "      <th>Gas</th>\n",
       "      <th>OVX</th>\n",
       "      <th>DXY</th>\n",
       "      <th>GPRD</th>\n",
       "      <th>BrDa</th>\n",
       "      <th>unique_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2024-08-02</td>\n",
       "      <td>78.265</td>\n",
       "      <td>80.73</td>\n",
       "      <td>2.4966</td>\n",
       "      <td>24.230000</td>\n",
       "      <td>105.800003</td>\n",
       "      <td>128.757050</td>\n",
       "      <td>87.015</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2024-07-30</td>\n",
       "      <td>79.140</td>\n",
       "      <td>81.57</td>\n",
       "      <td>2.4652</td>\n",
       "      <td>24.170000</td>\n",
       "      <td>105.260002</td>\n",
       "      <td>144.024689</td>\n",
       "      <td>85.280</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2024-07-26</td>\n",
       "      <td>81.245</td>\n",
       "      <td>78.45</td>\n",
       "      <td>2.3877</td>\n",
       "      <td>24.450001</td>\n",
       "      <td>105.550003</td>\n",
       "      <td>90.600143</td>\n",
       "      <td>81.730</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2024-07-25</td>\n",
       "      <td>82.655</td>\n",
       "      <td>78.62</td>\n",
       "      <td>2.4020</td>\n",
       "      <td>24.799999</td>\n",
       "      <td>105.199997</td>\n",
       "      <td>134.210846</td>\n",
       "      <td>81.640</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2024-07-24</td>\n",
       "      <td>82.995</td>\n",
       "      <td>78.50</td>\n",
       "      <td>2.3944</td>\n",
       "      <td>25.540001</td>\n",
       "      <td>104.650002</td>\n",
       "      <td>118.305992</td>\n",
       "      <td>81.245</td>\n",
       "      <td>Dated</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1812</th>\n",
       "      <td>2014-09-26</td>\n",
       "      <td>95.250</td>\n",
       "      <td>97.35</td>\n",
       "      <td>2.5750</td>\n",
       "      <td>17.860001</td>\n",
       "      <td>81.419998</td>\n",
       "      <td>131.822342</td>\n",
       "      <td>100.775</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1813</th>\n",
       "      <td>2014-09-25</td>\n",
       "      <td>95.465</td>\n",
       "      <td>95.58</td>\n",
       "      <td>2.5403</td>\n",
       "      <td>17.190001</td>\n",
       "      <td>81.589996</td>\n",
       "      <td>106.742241</td>\n",
       "      <td>100.445</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1814</th>\n",
       "      <td>2014-09-24</td>\n",
       "      <td>94.585</td>\n",
       "      <td>97.59</td>\n",
       "      <td>2.6266</td>\n",
       "      <td>17.549999</td>\n",
       "      <td>81.599998</td>\n",
       "      <td>165.846603</td>\n",
       "      <td>101.855</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1815</th>\n",
       "      <td>2014-09-23</td>\n",
       "      <td>95.065</td>\n",
       "      <td>97.37</td>\n",
       "      <td>2.6045</td>\n",
       "      <td>17.540001</td>\n",
       "      <td>81.500000</td>\n",
       "      <td>150.571259</td>\n",
       "      <td>101.725</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1816</th>\n",
       "      <td>2014-09-19</td>\n",
       "      <td>96.645</td>\n",
       "      <td>97.65</td>\n",
       "      <td>2.7537</td>\n",
       "      <td>18.850000</td>\n",
       "      <td>81.389999</td>\n",
       "      <td>212.396439</td>\n",
       "      <td>103.315</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1817 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             ds       y   BrFu     Gas        OVX         DXY        GPRD  \\\n",
       "0    2024-08-02  78.265  80.73  2.4966  24.230000  105.800003  128.757050   \n",
       "1    2024-07-30  79.140  81.57  2.4652  24.170000  105.260002  144.024689   \n",
       "2    2024-07-26  81.245  78.45  2.3877  24.450001  105.550003   90.600143   \n",
       "3    2024-07-25  82.655  78.62  2.4020  24.799999  105.199997  134.210846   \n",
       "4    2024-07-24  82.995  78.50  2.3944  25.540001  104.650002  118.305992   \n",
       "...         ...     ...    ...     ...        ...         ...         ...   \n",
       "1812 2014-09-26  95.250  97.35  2.5750  17.860001   81.419998  131.822342   \n",
       "1813 2014-09-25  95.465  95.58  2.5403  17.190001   81.589996  106.742241   \n",
       "1814 2014-09-24  94.585  97.59  2.6266  17.549999   81.599998  165.846603   \n",
       "1815 2014-09-23  95.065  97.37  2.6045  17.540001   81.500000  150.571259   \n",
       "1816 2014-09-19  96.645  97.65  2.7537  18.850000   81.389999  212.396439   \n",
       "\n",
       "         BrDa unique_id  \n",
       "0      87.015     Dated  \n",
       "1      85.280     Dated  \n",
       "2      81.730     Dated  \n",
       "3      81.640     Dated  \n",
       "4      81.245     Dated  \n",
       "...       ...       ...  \n",
       "1812  100.775     Dated  \n",
       "1813  100.445     Dated  \n",
       "1814  101.855     Dated  \n",
       "1815  101.725     Dated  \n",
       "1816  103.315     Dated  \n",
       "\n",
       "[1817 rows x 9 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_df_test = Y_df.copy().rename({'y' : 'BrDa'}, axis=1)\n",
    "Y_df_test = createLag(Y_df_test, amt=30)\n",
    "df = Y_df[['ds', 'y']].merge(Y_df_test, on = 'ds')\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "8dacad81-f2ed-48ee-a72b-8a3efdb4f112",
   "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>ds</th>\n",
       "      <th>y</th>\n",
       "      <th>BrFu</th>\n",
       "      <th>Gas</th>\n",
       "      <th>OVX</th>\n",
       "      <th>DXY</th>\n",
       "      <th>GPRD</th>\n",
       "      <th>BrDa</th>\n",
       "      <th>unique_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1815</th>\n",
       "      <td>2014-09-23</td>\n",
       "      <td>1.448573</td>\n",
       "      <td>-0.002872</td>\n",
       "      <td>-0.055705</td>\n",
       "      <td>17.540001</td>\n",
       "      <td>0.001351</td>\n",
       "      <td>150.571259</td>\n",
       "      <td>-0.015509</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1814</th>\n",
       "      <td>2014-09-24</td>\n",
       "      <td>1.424494</td>\n",
       "      <td>0.002257</td>\n",
       "      <td>0.008450</td>\n",
       "      <td>17.549999</td>\n",
       "      <td>0.001226</td>\n",
       "      <td>165.846603</td>\n",
       "      <td>0.001277</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1813</th>\n",
       "      <td>2014-09-25</td>\n",
       "      <td>1.468639</td>\n",
       "      <td>-0.020811</td>\n",
       "      <td>-0.033408</td>\n",
       "      <td>17.190001</td>\n",
       "      <td>-0.000123</td>\n",
       "      <td>106.742241</td>\n",
       "      <td>-0.013940</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1812</th>\n",
       "      <td>2014-09-26</td>\n",
       "      <td>1.457854</td>\n",
       "      <td>0.018349</td>\n",
       "      <td>0.013567</td>\n",
       "      <td>17.860001</td>\n",
       "      <td>-0.002086</td>\n",
       "      <td>131.822342</td>\n",
       "      <td>0.003280</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1811</th>\n",
       "      <td>2014-09-30</td>\n",
       "      <td>1.435279</td>\n",
       "      <td>-0.029925</td>\n",
       "      <td>-0.004437</td>\n",
       "      <td>19.520000</td>\n",
       "      <td>0.005634</td>\n",
       "      <td>216.154495</td>\n",
       "      <td>-0.010825</td>\n",
       "      <td>Dated</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2024-07-24</td>\n",
       "      <td>0.843081</td>\n",
       "      <td>0.007673</td>\n",
       "      <td>-0.006038</td>\n",
       "      <td>25.540001</td>\n",
       "      <td>-0.005527</td>\n",
       "      <td>118.305992</td>\n",
       "      <td>0.006235</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2024-07-25</td>\n",
       "      <td>0.826024</td>\n",
       "      <td>0.001527</td>\n",
       "      <td>0.003169</td>\n",
       "      <td>24.799999</td>\n",
       "      <td>0.005242</td>\n",
       "      <td>134.210846</td>\n",
       "      <td>0.004850</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2024-07-26</td>\n",
       "      <td>0.755292</td>\n",
       "      <td>-0.002165</td>\n",
       "      <td>-0.005971</td>\n",
       "      <td>24.450001</td>\n",
       "      <td>0.003322</td>\n",
       "      <td>90.600143</td>\n",
       "      <td>0.001102</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2024-07-30</td>\n",
       "      <td>0.649694</td>\n",
       "      <td>0.039000</td>\n",
       "      <td>0.031942</td>\n",
       "      <td>24.170000</td>\n",
       "      <td>-0.002751</td>\n",
       "      <td>144.024689</td>\n",
       "      <td>0.042519</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2024-08-02</td>\n",
       "      <td>0.605800</td>\n",
       "      <td>-0.010351</td>\n",
       "      <td>0.012657</td>\n",
       "      <td>24.230000</td>\n",
       "      <td>0.005117</td>\n",
       "      <td>128.757050</td>\n",
       "      <td>0.020141</td>\n",
       "      <td>Dated</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1816 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             ds         y      BrFu       Gas        OVX       DXY  \\\n",
       "1815 2014-09-23  1.448573 -0.002872 -0.055705  17.540001  0.001351   \n",
       "1814 2014-09-24  1.424494  0.002257  0.008450  17.549999  0.001226   \n",
       "1813 2014-09-25  1.468639 -0.020811 -0.033408  17.190001 -0.000123   \n",
       "1812 2014-09-26  1.457854  0.018349  0.013567  17.860001 -0.002086   \n",
       "1811 2014-09-30  1.435279 -0.029925 -0.004437  19.520000  0.005634   \n",
       "...         ...       ...       ...       ...        ...       ...   \n",
       "4    2024-07-24  0.843081  0.007673 -0.006038  25.540001 -0.005527   \n",
       "3    2024-07-25  0.826024  0.001527  0.003169  24.799999  0.005242   \n",
       "2    2024-07-26  0.755292 -0.002165 -0.005971  24.450001  0.003322   \n",
       "1    2024-07-30  0.649694  0.039000  0.031942  24.170000 -0.002751   \n",
       "0    2024-08-02  0.605800 -0.010351  0.012657  24.230000  0.005117   \n",
       "\n",
       "            GPRD      BrDa unique_id  \n",
       "1815  150.571259 -0.015509     Dated  \n",
       "1814  165.846603  0.001277     Dated  \n",
       "1813  106.742241 -0.013940     Dated  \n",
       "1812  131.822342  0.003280     Dated  \n",
       "1811  216.154495 -0.010825     Dated  \n",
       "...          ...       ...       ...  \n",
       "4     118.305992  0.006235     Dated  \n",
       "3     134.210846  0.004850     Dated  \n",
       "2      90.600143  0.001102     Dated  \n",
       "1     144.024689  0.042519     Dated  \n",
       "0     128.757050  0.020141     Dated  \n",
       "\n",
       "[1816 rows x 9 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data = transformData(df,\n",
    "                         log_return = ['BrFu', 'Gas', 'DXY', 'BrDa'],\n",
    "                         standard_scale = ['y'])\n",
    "test_data.dropna(inplace=True)\n",
    "test_data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b862eb0-afca-489b-9289-0fa2efaaf1d4",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "0d42f7f7-8c6f-4708-88e8-f4e4fdd081db",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Seed set to 12345678\n",
      "Seed set to 12345678\n",
      "Seed set to 12345678\n"
     ]
    }
   ],
   "source": [
    "horizon = 30\n",
    "input_size = 30*2\n",
    "models = [\n",
    "    TSMixer(h=horizon,\n",
    "                input_size=input_size,\n",
    "                n_series=1,\n",
    "                max_steps=1000,\n",
    "                val_check_steps=100,\n",
    "                early_stop_patience_steps=5,\n",
    "                scaler_type='identity',\n",
    "                loss=MAPE(),\n",
    "                valid_loss=MAPE(),\n",
    "                random_seed=12345678,\n",
    "                ),\n",
    "    TSMixerx(h=horizon,\n",
    "                input_size=input_size,\n",
    "                n_series=1,\n",
    "                max_steps=1000,\n",
    "                val_check_steps=100,\n",
    "                early_stop_patience_steps=5,\n",
    "                scaler_type='identity',\n",
    "                dropout=0.7,\n",
    "                loss=MAPE(),\n",
    "                valid_loss=MAPE(),\n",
    "                random_seed=12345678,\n",
    "                futr_exog_list=['Gas', 'DXY', 'BrFu', 'BrDa'],\n",
    "                ),\n",
    "    NBEATSx(h=horizon,\n",
    "                input_size=horizon,\n",
    "                max_steps=1000,\n",
    "                val_check_steps=100,\n",
    "                early_stop_patience_steps=5,\n",
    "                scaler_type='identity',\n",
    "                loss=MAPE(),\n",
    "                valid_loss=MAPE(),\n",
    "                random_seed=12345678,\n",
    "                futr_exog_list=['Gas', 'DXY', 'BrFu', 'BrDa']\n",
    "                ),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "1ef4261e-2256-4173-b6fa-a75a12062b11",
   "metadata": {},
   "outputs": [],
   "source": [
    "nf = NeuralForecast(\n",
    "    models=models,\n",
    "    freq='D')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "1157bb88-22fb-45c7-a36b-73149b873dd6",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
      "\n",
      "  | Name          | Type                     | Params | Mode \n",
      "-------------------------------------------------------------------\n",
      "0 | loss          | MAPE                     | 0      | train\n",
      "1 | valid_loss    | MAPE                     | 0      | train\n",
      "2 | padder        | ConstantPad1d            | 0      | train\n",
      "3 | scaler        | TemporalNorm             | 0      | train\n",
      "4 | norm          | ReversibleInstanceNorm1d | 2      | train\n",
      "5 | mixing_layers | Sequential               | 8.2 K  | train\n",
      "6 | out           | Linear                   | 1.8 K  | train\n",
      "-------------------------------------------------------------------\n",
      "10.0 K    Trainable params\n",
      "0         Non-trainable params\n",
      "10.0 K    Total params\n",
      "0.040     Total estimated model params size (MB)\n",
      "29        Modules in train mode\n",
      "0         Modules in eval mode\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6fda2e4a4cc8470d8c1b9682bf7080f7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sanity Checking: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b5e626d565df4752a645cceddfbc5885",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Training: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "de306cabc38c4c1299e829d48a0963fd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1fba904242824a168780b53a6d62ef03",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eafa8930b2d54bd8a9719069d6b6c319",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "16081f443ef4422e9d3596fd6b9178af",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3665f4b2e5a141f59c8d98385c15aa99",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "157a7800b1d64fdeae430ee6f3103382",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6e8683f1054f48f2a175ee2debf81e27",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dced14b5390948fa8bf678fe81c4cdc2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f0899b3ef2e842c0a24d076d8944fade",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0a9a10ea906c4e53886e9fa62b5efeb3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "`Trainer.fit` stopped: `max_steps=1000` reached.\n",
      "Trainer already configured with model summary callbacks: [<class 'pytorch_lightning.callbacks.model_summary.ModelSummary'>]. Skipping setting a default `ModelSummary` callback.\n",
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7c15d29b08454c02ba907c6acf2810d8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Predicting: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
      "\n",
      "   | Name                | Type                     | Params | Mode \n",
      "--------------------------------------------------------------------------\n",
      "0  | loss                | MAPE                     | 0      | train\n",
      "1  | valid_loss          | MAPE                     | 0      | train\n",
      "2  | padder              | ConstantPad1d            | 0      | train\n",
      "3  | scaler              | TemporalNorm             | 0      | train\n",
      "4  | norm                | ReversibleInstanceNorm1d | 2      | train\n",
      "5  | temporal_projection | Linear                   | 1.8 K  | train\n",
      "6  | feature_mixer_hist  | FeatureMixing            | 8.8 K  | train\n",
      "7  | feature_mixer_futr  | FeatureMixing            | 8.6 K  | train\n",
      "8  | first_mixing        | MixingLayer              | 33.1 K | train\n",
      "9  | mixing_block        | Sequential               | 33.9 K | train\n",
      "10 | out                 | Linear                   | 65     | train\n",
      "--------------------------------------------------------------------------\n",
      "86.3 K    Trainable params\n",
      "0         Non-trainable params\n",
      "86.3 K    Total params\n",
      "0.345     Total estimated model params size (MB)\n",
      "56        Modules in train mode\n",
      "0         Modules in eval mode\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e603809b378041c4944303ff79fdd7cb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sanity Checking: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "968c9417bdf8403a916460717b06297a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Training: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8e5c8280f3e640e6b32c3e8f6cb7f51c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "59f628f61c7e4d48a71f2b14d0b8e8ed",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e435f7c4034a416b81016c8b265472a6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bfd2a72365ba4e4fa0981ca4d18e2582",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "82cc2fd8da924e3d8b56b647301743a0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f9cf10b217f341218a09c58e2680de64",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7c4162caae4041669188600835ee1277",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4c0a64db0d2946598c73c0565de94d64",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2b4095836fac44478a5b7d29923f6492",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1bfb3a5ca61848abac884dda424ba8f3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "`Trainer.fit` stopped: `max_steps=1000` reached.\n",
      "Trainer already configured with model summary callbacks: [<class 'pytorch_lightning.callbacks.model_summary.ModelSummary'>]. Skipping setting a default `ModelSummary` callback.\n",
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ccd2baca479b4ce8801170f7fd9979cf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Predicting: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
      "\n",
      "  | Name         | Type          | Params | Mode \n",
      "-------------------------------------------------------\n",
      "0 | loss         | MAPE          | 0      | train\n",
      "1 | valid_loss   | MAPE          | 0      | train\n",
      "2 | padder_train | ConstantPad1d | 0      | train\n",
      "3 | scaler       | TemporalNorm  | 0      | train\n",
      "4 | blocks       | ModuleList    | 2.9 M  | train\n",
      "-------------------------------------------------------\n",
      "2.9 M     Trainable params\n",
      "3.7 K     Non-trainable params\n",
      "2.9 M     Total params\n",
      "11.509    Total estimated model params size (MB)\n",
      "32        Modules in train mode\n",
      "0         Modules in eval mode\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a484e33c9b2d41978c077020a0346df2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Sanity Checking: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "13436754d0074356a592d89e3e969752",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Training: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "188d56c7c1b841d2ae1274146f59c3e8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1033525eb07945dd886f826a47ec9656",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c6bf7e30cbef4688b618252a6836be43",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a3731edcb4204340b8ebaf597f178eda",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e5931192f0394792b16aa3a8b5a3617a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3858bc3c32a94e1ab79065d6335e1ea4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c5470a5426df4ef687b0f2846e45f11d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Validation: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Trainer already configured with model summary callbacks: [<class 'pytorch_lightning.callbacks.model_summary.ModelSummary'>]. Skipping setting a default `ModelSummary` callback.\n",
      "GPU available: True (cuda), used: True\n",
      "TPU available: False, using: 0 TPU cores\n",
      "HPU available: False, using: 0 HPUs\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dedfdcf9fdc3416fa0deeb8cba3d844e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Predicting: |          | 0/? [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\VM\\miniconda3\\envs\\pytorch\\lib\\site-packages\\neuralforecast\\core.py:209: FutureWarning: In a future version the predictions will have the id as a column. You can set the `NIXTLA_ID_AS_COL` environment variable to adopt the new behavior and to suppress this warning.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "Y_hat_df = nf.cross_validation(df=df,\n",
    "                               val_size=val_size,\n",
    "                               test_size=test_size,\n",
    "                               n_windows=None\n",
    "                               )\n",
    "Y_hat_df = Y_hat_df.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "e74d3f1a-88a9-4430-9041-34aa770c1514",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Y_hat_df['y'] = yScaler.inverse_transform(Y_hat_df[['y']])\n",
    "# Y_hat_df['TSMixer'] = yScaler.inverse_transform(Y_hat_df[['TSMixer']])\n",
    "# Y_hat_df['TSMixerx'] = yScaler.inverse_transform(Y_hat_df[['TSMixerx']])\n",
    "# Y_hat_df['NBEATSx'] = yScaler.inverse_transform(Y_hat_df[['NBEATSx']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "e1aa49a5-63e8-4b3d-adeb-154de6dd7f46",
   "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>unique_id</th>\n",
       "      <th>ds</th>\n",
       "      <th>cutoff</th>\n",
       "      <th>TSMixer</th>\n",
       "      <th>TSMixerx</th>\n",
       "      <th>NBEATSx</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-04-10</td>\n",
       "      <td>2024-04-09</td>\n",
       "      <td>1.177287</td>\n",
       "      <td>1.390748</td>\n",
       "      <td>1.221220</td>\n",
       "      <td>1.254183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-04-11</td>\n",
       "      <td>2024-04-09</td>\n",
       "      <td>1.155755</td>\n",
       "      <td>1.350257</td>\n",
       "      <td>1.241936</td>\n",
       "      <td>1.285286</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-04-12</td>\n",
       "      <td>2024-04-09</td>\n",
       "      <td>1.223631</td>\n",
       "      <td>1.333797</td>\n",
       "      <td>1.155751</td>\n",
       "      <td>1.362289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-04-16</td>\n",
       "      <td>2024-04-09</td>\n",
       "      <td>1.191017</td>\n",
       "      <td>1.371189</td>\n",
       "      <td>1.099887</td>\n",
       "      <td>1.259200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-04-17</td>\n",
       "      <td>2024-04-09</td>\n",
       "      <td>1.183228</td>\n",
       "      <td>1.362661</td>\n",
       "      <td>1.161999</td>\n",
       "      <td>1.177431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>595</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-07-24</td>\n",
       "      <td>2024-05-21</td>\n",
       "      <td>1.090267</td>\n",
       "      <td>0.998137</td>\n",
       "      <td>0.934902</td>\n",
       "      <td>0.843081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>596</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-07-25</td>\n",
       "      <td>2024-05-21</td>\n",
       "      <td>1.089488</td>\n",
       "      <td>0.950186</td>\n",
       "      <td>0.956440</td>\n",
       "      <td>0.826024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>597</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-07-26</td>\n",
       "      <td>2024-05-21</td>\n",
       "      <td>0.982749</td>\n",
       "      <td>1.021258</td>\n",
       "      <td>0.885282</td>\n",
       "      <td>0.755292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>598</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-07-30</td>\n",
       "      <td>2024-05-21</td>\n",
       "      <td>1.083079</td>\n",
       "      <td>0.930227</td>\n",
       "      <td>0.883445</td>\n",
       "      <td>0.649694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>599</th>\n",
       "      <td>Dated</td>\n",
       "      <td>2024-08-02</td>\n",
       "      <td>2024-05-21</td>\n",
       "      <td>1.033108</td>\n",
       "      <td>0.968054</td>\n",
       "      <td>0.857022</td>\n",
       "      <td>0.605800</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>600 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    unique_id         ds     cutoff   TSMixer  TSMixerx   NBEATSx         y\n",
       "0       Dated 2024-04-10 2024-04-09  1.177287  1.390748  1.221220  1.254183\n",
       "1       Dated 2024-04-11 2024-04-09  1.155755  1.350257  1.241936  1.285286\n",
       "2       Dated 2024-04-12 2024-04-09  1.223631  1.333797  1.155751  1.362289\n",
       "3       Dated 2024-04-16 2024-04-09  1.191017  1.371189  1.099887  1.259200\n",
       "4       Dated 2024-04-17 2024-04-09  1.183228  1.362661  1.161999  1.177431\n",
       "..        ...        ...        ...       ...       ...       ...       ...\n",
       "595     Dated 2024-07-24 2024-05-21  1.090267  0.998137  0.934902  0.843081\n",
       "596     Dated 2024-07-25 2024-05-21  1.089488  0.950186  0.956440  0.826024\n",
       "597     Dated 2024-07-26 2024-05-21  0.982749  1.021258  0.885282  0.755292\n",
       "598     Dated 2024-07-30 2024-05-21  1.083079  0.930227  0.883445  0.649694\n",
       "599     Dated 2024-08-02 2024-05-21  1.033108  0.968054  0.857022  0.605800\n",
       "\n",
       "[600 rows x 7 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_hat_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "10d24df3-901f-449a-ad2e-cb6a7a32ff74",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19831c741f0>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "Y_plot = Y_hat_df[Y_hat_df['unique_id']=='Dated']\n",
    "cutoffs = Y_hat_df['cutoff'].unique()[::horizon]\n",
    "Y_plot = Y_plot[Y_hat_df['cutoff'].isin(cutoffs)]\n",
    "\n",
    "plt.figure(figsize=(20,5))\n",
    "plt.plot(Y_plot['ds'], Y_plot['y'], label='True')\n",
    "for model in models:\n",
    "    plt.plot(Y_plot['ds'], Y_plot[f'{model}'], label=f'{model}')\n",
    "plt.xlabel('Datestamp')\n",
    "plt.ylabel('OT')\n",
    "plt.grid()\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "55f84c08-6ffd-49d7-8798-4cbc5209e550",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TSMixer horizon 30 - MAE: 0.259\n",
      "TSMixer horizon 30 - MSE: 0.092\n",
      "TSMixer horizon 30 - MAPE: 0.363\n",
      "TSMixerx horizon 30 - MAE: 0.274\n",
      "TSMixerx horizon 30 - MSE: 0.113\n",
      "TSMixerx horizon 30 - MAPE: 0.367\n",
      "NBEATSx horizon 30 - MAE: 0.195\n",
      "NBEATSx horizon 30 - MSE: 0.067\n",
      "NBEATSx horizon 30 - MAPE: 0.234\n"
     ]
    }
   ],
   "source": [
    "from neuralforecast.losses.numpy import mse, mae, mape\n",
    "\n",
    "for model in models:\n",
    "    mae_model = mae(Y_hat_df['y'], Y_hat_df[f'{model}']) \n",
    "    mse_model = mse(Y_hat_df['y'], Y_hat_df[f'{model}'])\n",
    "    mape_model = mape(Y_hat_df['y'], Y_hat_df[f'{model}'])\n",
    "    print(f'{model} horizon {horizon} - MAE: {mae_model:.3f}')\n",
    "    print(f'{model} horizon {horizon} - MSE: {mse_model:.3f}')\n",
    "    print(f'{model} horizon {horizon} - MAPE: {mape_model:.3f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "28baeafb-d96e-405d-b460-bd2001e7f160",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt\n",
    "# Y_plot = Y_hat_df[Y_hat_df['unique_id']=='Future']\n",
    "# cutoffs = Y_hat_df['cutoff'].unique()[::horizon]\n",
    "# Y_plot = Y_plot[Y_hat_df['cutoff'].isin(cutoffs)]\n",
    "\n",
    "# plt.figure(figsize=(20,5))\n",
    "# plt.plot(Y_plot['ds'], Y_plot['y'], label='True')\n",
    "# for model in models:\n",
    "#     plt.plot(Y_plot['ds'], Y_plot[f'{model}'], label=f'{model}')\n",
    "# plt.xlabel('Datestamp')\n",
    "# plt.ylabel('OT')\n",
    "# plt.grid()\n",
    "# plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5be8e1f9-a92d-4edb-a992-8021b605333c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pytorch",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}