terra_validation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from ESRNN import ESRNN\n",
    "plt.style.use('ggplot')\n",
    "pd.options.display.max_rows = 999\n",
    "np.set_printoptions(threshold=np.inf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot\n",
    "def plot_prediction(y, y_hat):\n",
    "    n_y = len(y)\n",
    "    n_yhat = len(y_hat)\n",
    "    ds_y = np.array(range(n_y))\n",
    "    ds_yhat = np.array(range(n_y-n_yhat, n_y))\n",
    "\n",
    "    plt.plot(ds_y, y, label = 'y')\n",
    "    plt.plot(ds_yhat, y_hat, label='y_hat')\n",
    "    plt.legend(loc='upper left')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ffill_missing_dates_particular_serie(serie, min_date, max_date, freq):\n",
    "    date_range = pd.date_range(start=min_date, end=max_date, freq=freq)\n",
    "    unique_id = serie['unique_id'].unique()\n",
    "    df_balanced = pd.DataFrame({'ds':date_range, 'key':[1]*len(date_range), 'unique_id': unique_id[0]})\n",
    "\n",
    "    # Check balance\n",
    "    check_balance = df_balanced.groupby(['unique_id']).size().reset_index(name='count')\n",
    "    assert len(set(check_balance['count'].values)) <= 1\n",
    "    df_balanced = df_balanced.merge(serie, how=\"left\", on=['unique_id', 'ds'])\n",
    "\n",
    "    df_balanced['y'] = df_balanced['y'].fillna(method='ffill')\n",
    "\n",
    "    return df_balanced\n",
    "\n",
    "def ffill_missing_dates_per_serie(df, freq, fixed_max_date=None):\n",
    "    \"\"\"Receives a DataFrame with a date column and forward fills the missing gaps in dates, not filling dates before\n",
    "    the first appearance of a unique key\n",
    "\n",
    "    Parameters\n",
    "    ----------\n",
    "    df: DataFrame\n",
    "        Input DataFrame\n",
    "    key: str or list\n",
    "        Name(s) of the column(s) which make a unique time series\n",
    "    date_col: str\n",
    "        Name of the column that contains the time column\n",
    "    freq: str\n",
    "        Pandas time frequency standard strings, like \"W-THU\" or \"D\" or \"M\"\n",
    "    numeric_to_fill: str or list\n",
    "        Name(s) of the columns with numeric values to fill \"fill_value\" with\n",
    "    \"\"\"\n",
    "    if fixed_max_date is None:\n",
    "        df_max_min_dates = df[['unique_id', 'ds']].groupby('unique_id').agg(['min', 'max']).reset_index()\n",
    "    else:\n",
    "        df_max_min_dates = df[['unique_id', 'ds']].groupby('unique_id').agg(['min']).reset_index()\n",
    "        df_max_min_dates['max'] = fixed_max_date\n",
    "\n",
    "    df_max_min_dates.columns = df_max_min_dates.columns.droplevel()\n",
    "    df_max_min_dates.columns = ['unique_id', 'min_date', 'max_date']\n",
    "\n",
    "    df_list = []\n",
    "    for index, row in df_max_min_dates.iterrows():\n",
    "        df_id = df[df['unique_id'] == row['unique_id']]\n",
    "        df_id = ffill_missing_dates_particular_serie(df_id, row['min_date'], row['max_date'], freq)\n",
    "        df_list.append(df_id)\n",
    "\n",
    "    df_dates = pd.concat(df_list).reset_index(drop=True).drop('key', axis=1)[['unique_id', 'ds', 'y']]\n",
    "\n",
    "    return df_dates"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv('data/terra/train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "      <td>1970-01-13</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "      <td>1970-01-14</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "      <td>1970-01-15</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "      <td>1970-01-16</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "      <td>1970-01-17</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           unique_id         ds     y\n",
       "0  PI145619MAC Field Scanner Season 4 Range 27 Co... 1970-01-13  10.0\n",
       "1  PI145619MAC Field Scanner Season 4 Range 27 Co... 1970-01-14  10.0\n",
       "2  PI145619MAC Field Scanner Season 4 Range 27 Co... 1970-01-15  10.0\n",
       "3  PI145619MAC Field Scanner Season 4 Range 27 Co... 1970-01-16  10.0\n",
       "4  PI145619MAC Field Scanner Season 4 Range 27 Co... 1970-01-17  10.0"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_csv('data/terra/train.csv')\n",
    "data['ds'] = pd.to_datetime(data['day'], unit='d')\n",
    "data['unique_id'] = data['cultivar'] + data['sitename']\n",
    "data = data.rename(columns={'canopy_height':'y'})\n",
    "\n",
    "#Series must be complete in the frequency\n",
    "data = ffill_missing_dates_per_serie(data,'D')\n",
    "data = data.drop_duplicates(['unique_id','ds'])\n",
    "\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data[['unique_id','ds']]\n",
    "X_train['x'] = '1'\n",
    "y_train = data[['unique_id','ds','y']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/cchallu/anaconda3/envs/esrnn_torch/lib/python3.6/site-packages/pandas/plotting/_matplotlib/converter.py:103: FutureWarning: Using an implicitly registered datetime converter for a matplotlib plotting method. The converter was registered by pandas on import. Future versions of pandas will require you to explicitly register matplotlib converters.\n",
      "\n",
      "To register the converters:\n",
      "\t>>> from pandas.plotting import register_matplotlib_converters\n",
      "\t>>> register_matplotlib_converters()\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_plot = y_train[y_train['unique_id']=='PI157030MAC Field Scanner Season 4 Range 17 Column 14']\n",
    "plt.plot(data_plot['ds'], data_plot['y'], label = 'y')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>d3mIndex</th>\n",
       "      <th>cultivar</th>\n",
       "      <th>sitename</th>\n",
       "      <th>day</th>\n",
       "      <th>canopy_height</th>\n",
       "      <th>ds</th>\n",
       "      <th>unique_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>34</td>\n",
       "      <td>PI145619</td>\n",
       "      <td>MAC Field Scanner Season 4 Range 27 Column 11</td>\n",
       "      <td>133</td>\n",
       "      <td>341</td>\n",
       "      <td>1970-05-14</td>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>69</td>\n",
       "      <td>PI145619</td>\n",
       "      <td>MAC Field Scanner Season 4 Range 27 Column 12</td>\n",
       "      <td>133</td>\n",
       "      <td>336</td>\n",
       "      <td>1970-05-14</td>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 27 Co...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>91</td>\n",
       "      <td>PI145619</td>\n",
       "      <td>MAC Field Scanner Season 4 Range 33 Column 15</td>\n",
       "      <td>133</td>\n",
       "      <td>348</td>\n",
       "      <td>1970-05-14</td>\n",
       "      <td>PI145619MAC Field Scanner Season 4 Range 33 Co...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>118</td>\n",
       "      <td>PI145626</td>\n",
       "      <td>MAC Field Scanner Season 4 Range 3 Column 7</td>\n",
       "      <td>133</td>\n",
       "      <td>399</td>\n",
       "      <td>1970-05-14</td>\n",
       "      <td>PI145626MAC Field Scanner Season 4 Range 3 Col...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>145</td>\n",
       "      <td>PI145626</td>\n",
       "      <td>MAC Field Scanner Season 4 Range 3 Column 8</td>\n",
       "      <td>133</td>\n",
       "      <td>364</td>\n",
       "      <td>1970-05-14</td>\n",
       "      <td>PI145626MAC Field Scanner Season 4 Range 3 Col...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   d3mIndex  cultivar                                       sitename  day  \\\n",
       "0        34  PI145619  MAC Field Scanner Season 4 Range 27 Column 11  133   \n",
       "1        69  PI145619  MAC Field Scanner Season 4 Range 27 Column 12  133   \n",
       "2        91  PI145619  MAC Field Scanner Season 4 Range 33 Column 15  133   \n",
       "3       118  PI145626    MAC Field Scanner Season 4 Range 3 Column 7  133   \n",
       "4       145  PI145626    MAC Field Scanner Season 4 Range 3 Column 8  133   \n",
       "\n",
       "   canopy_height         ds                                          unique_id  \n",
       "0            341 1970-05-14  PI145619MAC Field Scanner Season 4 Range 27 Co...  \n",
       "1            336 1970-05-14  PI145619MAC Field Scanner Season 4 Range 27 Co...  \n",
       "2            348 1970-05-14  PI145619MAC Field Scanner Season 4 Range 33 Co...  \n",
       "3            399 1970-05-14  PI145626MAC Field Scanner Season 4 Range 3 Col...  \n",
       "4            364 1970-05-14  PI145626MAC Field Scanner Season 4 Range 3 Col...  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_test = pd.read_csv('data/terra/test_real.csv')\n",
    "data_test['ds'] = pd.to_datetime(data_test['day'], unit='d')\n",
    "data_test['unique_id'] = data_test['cultivar'] + data_test['sitename']\n",
    "data_test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = data_test[['unique_id','ds','canopy_height']]\n",
    "X_test.columns = ['unique_id', 'ds', 'y']\n",
    "uniques = X_test['unique_id'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=============== Training ESRNN  ===============\n",
      "\n",
      "========= Epoch 0 finished =========\n",
      "Training time: 1.69194\n",
      "Training loss (60 prc): 4.06712\n",
      "Train finished! \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Train model\n",
    "esrnn = ESRNN(max_epochs=1, batch_size=16, learning_rate=3e-4, gradient_clipping_threshold=20,\n",
    "              dilations=[[1, 7], [28]], add_nl_layer=True, per_series_lr_multip=1.0, \n",
    "              seasonality=[], input_size=7, output_size=50, max_periods=20, level_variability_penalty=80,\n",
    "              training_percentile=60, rnn_weight_decay=0)\n",
    "esrnn.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "========= Predicting ESRNN =========\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/cchallu/anaconda3/envs/esrnn_torch/lib/python3.6/site-packages/pandas/core/frame.py:7138: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version\n",
      "of pandas will change to not sort by default.\n",
      "\n",
      "To accept the future behavior, pass 'sort=False'.\n",
      "\n",
      "To retain the current behavior and silence the warning, pass 'sort=True'.\n",
      "\n",
      "  sort=sort,\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_hat = esrnn.predict(X_test[['unique_id']])\n",
    "X_plot = y_train.append(y_hat)\n",
    "plot_id = 0\n",
    "y_test_plot = X_plot.loc[X_plot['unique_id']==uniques[plot_id]]\n",
    "plot_prediction(y_test_plot['y'], y_test_plot['y_hat'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "========= Predicting ESRNN =========\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_hat = esrnn.predict(X_test[['unique_id','ds','y']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "32.685025169735866"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.abs(y_hat['y_hat']-y_hat['y']).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "esrnn_torch",
   "language": "python",
   "name": "esrnn_torch"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}