demo.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "source": [
    "# Predicting Tritium Thermo-Desorption Spectra using twinLab"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*** \n",
    "### What is tritium desorption?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The interior of the reactor wall is bombarded by tritium isotopes from the tritium breeding loop and plasma bombardment, resulting in adsorption and diffusion through the material. \n",
    "\n",
    "Additionally to the mobile tritium which diffuses through the material, there are regions of the atomic lattice called trapping sites, whereby tritium can become trapped in a potential well of a given energy. Understanding both the diffusion of the mobile tritium and the energy and density of these traps is essential in predicting the amount of tritium which is retained within the fusion reactor structures. This has implications in the required tritium breeding ratio, tritium inventory, future waste classification and material degradation. Characteristing these trapping sites is therefore essential in making engineering predictions of future fusion powerplants.\n",
    "\n",
    "If the material is heated up above a certain termperature, the tritium can become excited and released from these trapping site. This can be measured through an experiment known as thermal-desorption spectrometry. A piece of material which has been implanted with tritium is heated at incremental steps, with the amount of tritium released being measured as a function of temperature. the physical properties of these traps can then be determined from the characteristics of the spectra."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*** \n",
    "### ML Challenge \n",
    "\n",
    "Computer simulations are required to model the behaviour of these materials over a range of parameters, but simulations are expensive in terms of computational power, and cannot be run at every point in parameter space under consideration. `twinLab` can be used to train simulation surrogate models using data from a sparse array of simulations. This allows for meaningful interpolation and extrapolation to unexplored regions of parameter space, together with a calibrated uncertainty estimate on the accuracy of the simulation surrogate.\n",
    "\n",
    "In this example, we look at the ability of `twinLab` Gaussian Processes to model of tritium in the wall of a fusion reactor using tritium desorption spectra. \n",
    "\n",
    "Taking a training set of known correlations between tritium desorption spectra (TDS) of form $f(T)$ and physical trapping properties ($E_1$, $E_2$, $E_3$, $n_1$, $n_2$), a Gaussian Process can then be traing to predict the shape of a TDS given a any new set of physical parameters along with the uncertainty in these predictions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "### Importing libraries"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, import the required libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "          ====== TwinLab Client Initialisation ======\n",
      "          Version     : 2.4.0\n",
      "          Server      : https://twinlab.digilab.co.uk\n",
      "          Environment : /Users/will/Projects/FusionEnergy/.env\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Third-party imports\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "# Project imports\n",
    "import twinlab as tl"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to provide `twinLab` with locations of data files (in `.csv` format) and directories, together with a chosen surrogate emulator name (here `tritium_desorption`) and a list of the input parameters that we are going to use as inputs to our model. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "### Importing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "editable": true,
    "slideshow": {
     "slide_type": ""
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataframe downloaded successfully\n"
     ]
    }
   ],
   "source": [
    "data = tl.load_example_dataset('tritium-desorption')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_train = data.iloc[:900,:]\n",
    "df_test = data.iloc[-100:,:]\n",
    "\n",
    "df_grid = pd.DataFrame(np.linspace(300, 800, 624))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "### View the training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>E1</th>\n",
       "      <th>E2</th>\n",
       "      <th>E3</th>\n",
       "      <th>n1</th>\n",
       "      <th>n2</th>\n",
       "      <th>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.726163</td>\n",
       "      <td>1.248766</td>\n",
       "      <td>1.128671</td>\n",
       "      <td>0.002004</td>\n",
       "      <td>0.000307</td>\n",
       "      <td>1.511613e+17</td>\n",
       "      <td>1.626357e+17</td>\n",
       "      <td>1.749599e+17</td>\n",
       "      <td>1.881609e+17</td>\n",
       "      <td>2.022474e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.318606e+17</td>\n",
       "      <td>2.304607e+17</td>\n",
       "      <td>2.290609e+17</td>\n",
       "      <td>2.276611e+17</td>\n",
       "      <td>2.262613e+17</td>\n",
       "      <td>2.248615e+17</td>\n",
       "      <td>2.234616e+17</td>\n",
       "      <td>2.220618e+17</td>\n",
       "      <td>2.206620e+17</td>\n",
       "      <td>2.192622e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.787837</td>\n",
       "      <td>0.925174</td>\n",
       "      <td>1.226117</td>\n",
       "      <td>0.001564</td>\n",
       "      <td>0.000768</td>\n",
       "      <td>7.621430e+16</td>\n",
       "      <td>8.255297e+16</td>\n",
       "      <td>8.942421e+16</td>\n",
       "      <td>9.684174e+16</td>\n",
       "      <td>1.048160e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.728998e+16</td>\n",
       "      <td>2.706896e+16</td>\n",
       "      <td>2.684793e+16</td>\n",
       "      <td>2.662690e+16</td>\n",
       "      <td>2.640587e+16</td>\n",
       "      <td>2.618484e+16</td>\n",
       "      <td>2.596381e+16</td>\n",
       "      <td>2.574279e+16</td>\n",
       "      <td>2.552176e+16</td>\n",
       "      <td>2.530073e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.710103</td>\n",
       "      <td>0.902457</td>\n",
       "      <td>1.558539</td>\n",
       "      <td>0.004860</td>\n",
       "      <td>0.000613</td>\n",
       "      <td>2.376907e+17</td>\n",
       "      <td>2.553641e+17</td>\n",
       "      <td>2.743015e+17</td>\n",
       "      <td>2.945655e+17</td>\n",
       "      <td>3.161688e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.079730e+16</td>\n",
       "      <td>2.078486e+16</td>\n",
       "      <td>2.077242e+16</td>\n",
       "      <td>2.075998e+16</td>\n",
       "      <td>2.074754e+16</td>\n",
       "      <td>2.073510e+16</td>\n",
       "      <td>2.072265e+16</td>\n",
       "      <td>2.071021e+16</td>\n",
       "      <td>2.069777e+16</td>\n",
       "      <td>2.068533e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.757797</td>\n",
       "      <td>1.187059</td>\n",
       "      <td>1.479970</td>\n",
       "      <td>0.002868</td>\n",
       "      <td>0.000841</td>\n",
       "      <td>1.491269e+17</td>\n",
       "      <td>1.609964e+17</td>\n",
       "      <td>1.737868e+17</td>\n",
       "      <td>1.875378e+17</td>\n",
       "      <td>2.022633e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.405455e+17</td>\n",
       "      <td>2.382445e+17</td>\n",
       "      <td>2.359435e+17</td>\n",
       "      <td>2.336425e+17</td>\n",
       "      <td>2.313415e+17</td>\n",
       "      <td>2.290405e+17</td>\n",
       "      <td>2.267395e+17</td>\n",
       "      <td>2.244385e+17</td>\n",
       "      <td>2.221375e+17</td>\n",
       "      <td>2.198365e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.880510</td>\n",
       "      <td>0.987709</td>\n",
       "      <td>1.305003</td>\n",
       "      <td>0.001307</td>\n",
       "      <td>0.000302</td>\n",
       "      <td>8.396414e+15</td>\n",
       "      <td>9.173453e+15</td>\n",
       "      <td>1.003567e+16</td>\n",
       "      <td>1.097871e+16</td>\n",
       "      <td>1.200347e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>3.434468e+16</td>\n",
       "      <td>3.408240e+16</td>\n",
       "      <td>3.382012e+16</td>\n",
       "      <td>3.355785e+16</td>\n",
       "      <td>3.329557e+16</td>\n",
       "      <td>3.303329e+16</td>\n",
       "      <td>3.277101e+16</td>\n",
       "      <td>3.250873e+16</td>\n",
       "      <td>3.224646e+16</td>\n",
       "      <td>3.198418e+16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 629 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         E1        E2        E3        n1        n2            y0  \\\n",
       "0  0.726163  1.248766  1.128671  0.002004  0.000307  1.511613e+17   \n",
       "1  0.787837  0.925174  1.226117  0.001564  0.000768  7.621430e+16   \n",
       "2  0.710103  0.902457  1.558539  0.004860  0.000613  2.376907e+17   \n",
       "3  0.757797  1.187059  1.479970  0.002868  0.000841  1.491269e+17   \n",
       "4  0.880510  0.987709  1.305003  0.001307  0.000302  8.396414e+15   \n",
       "\n",
       "             y1            y2            y3            y4  ...          y614  \\\n",
       "0  1.626357e+17  1.749599e+17  1.881609e+17  2.022474e+17  ...  2.318606e+17   \n",
       "1  8.255297e+16  8.942421e+16  9.684174e+16  1.048160e+17  ...  2.728998e+16   \n",
       "2  2.553641e+17  2.743015e+17  2.945655e+17  3.161688e+17  ...  2.079730e+16   \n",
       "3  1.609964e+17  1.737868e+17  1.875378e+17  2.022633e+17  ...  2.405455e+17   \n",
       "4  9.173453e+15  1.003567e+16  1.097871e+16  1.200347e+16  ...  3.434468e+16   \n",
       "\n",
       "           y615          y616          y617          y618          y619  \\\n",
       "0  2.304607e+17  2.290609e+17  2.276611e+17  2.262613e+17  2.248615e+17   \n",
       "1  2.706896e+16  2.684793e+16  2.662690e+16  2.640587e+16  2.618484e+16   \n",
       "2  2.078486e+16  2.077242e+16  2.075998e+16  2.074754e+16  2.073510e+16   \n",
       "3  2.382445e+17  2.359435e+17  2.336425e+17  2.313415e+17  2.290405e+17   \n",
       "4  3.408240e+16  3.382012e+16  3.355785e+16  3.329557e+16  3.303329e+16   \n",
       "\n",
       "           y620          y621          y622          y623  \n",
       "0  2.234616e+17  2.220618e+17  2.206620e+17  2.192622e+17  \n",
       "1  2.596381e+16  2.574279e+16  2.552176e+16  2.530073e+16  \n",
       "2  2.072265e+16  2.071021e+16  2.069777e+16  2.068533e+16  \n",
       "3  2.267395e+17  2.244385e+17  2.221375e+17  2.198365e+17  \n",
       "4  3.277101e+16  3.250873e+16  3.224646e+16  3.198418e+16  \n",
       "\n",
       "[5 rows x 629 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inputs"
   ]
  },
  {
   "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>E1</th>\n",
       "      <th>E2</th>\n",
       "      <th>E3</th>\n",
       "      <th>n1</th>\n",
       "      <th>n2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.726163</td>\n",
       "      <td>1.248766</td>\n",
       "      <td>1.128671</td>\n",
       "      <td>0.002004</td>\n",
       "      <td>0.000307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.787837</td>\n",
       "      <td>0.925174</td>\n",
       "      <td>1.226117</td>\n",
       "      <td>0.001564</td>\n",
       "      <td>0.000768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.710103</td>\n",
       "      <td>0.902457</td>\n",
       "      <td>1.558539</td>\n",
       "      <td>0.004860</td>\n",
       "      <td>0.000613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.757797</td>\n",
       "      <td>1.187059</td>\n",
       "      <td>1.479970</td>\n",
       "      <td>0.002868</td>\n",
       "      <td>0.000841</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.880510</td>\n",
       "      <td>0.987709</td>\n",
       "      <td>1.305003</td>\n",
       "      <td>0.001307</td>\n",
       "      <td>0.000302</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         E1        E2        E3        n1        n2\n",
       "0  0.726163  1.248766  1.128671  0.002004  0.000307\n",
       "1  0.787837  0.925174  1.226117  0.001564  0.000768\n",
       "2  0.710103  0.902457  1.558539  0.004860  0.000613\n",
       "3  0.757797  1.187059  1.479970  0.002868  0.000841\n",
       "4  0.880510  0.987709  1.305003  0.001307  0.000302"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.iloc[:,:5].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Outputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>y5</th>\n",
       "      <th>y6</th>\n",
       "      <th>y7</th>\n",
       "      <th>y8</th>\n",
       "      <th>y9</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.511613e+17</td>\n",
       "      <td>1.626357e+17</td>\n",
       "      <td>1.749599e+17</td>\n",
       "      <td>1.881609e+17</td>\n",
       "      <td>2.022474e+17</td>\n",
       "      <td>2.171753e+17</td>\n",
       "      <td>2.329431e+17</td>\n",
       "      <td>2.501679e+17</td>\n",
       "      <td>2.681171e+17</td>\n",
       "      <td>2.873808e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.318606e+17</td>\n",
       "      <td>2.304607e+17</td>\n",
       "      <td>2.290609e+17</td>\n",
       "      <td>2.276611e+17</td>\n",
       "      <td>2.262613e+17</td>\n",
       "      <td>2.248615e+17</td>\n",
       "      <td>2.234616e+17</td>\n",
       "      <td>2.220618e+17</td>\n",
       "      <td>2.206620e+17</td>\n",
       "      <td>2.192622e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7.621430e+16</td>\n",
       "      <td>8.255297e+16</td>\n",
       "      <td>8.942421e+16</td>\n",
       "      <td>9.684174e+16</td>\n",
       "      <td>1.048160e+17</td>\n",
       "      <td>1.133235e+17</td>\n",
       "      <td>1.223659e+17</td>\n",
       "      <td>1.323483e+17</td>\n",
       "      <td>1.428036e+17</td>\n",
       "      <td>1.541191e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.728998e+16</td>\n",
       "      <td>2.706896e+16</td>\n",
       "      <td>2.684793e+16</td>\n",
       "      <td>2.662690e+16</td>\n",
       "      <td>2.640587e+16</td>\n",
       "      <td>2.618484e+16</td>\n",
       "      <td>2.596381e+16</td>\n",
       "      <td>2.574279e+16</td>\n",
       "      <td>2.552176e+16</td>\n",
       "      <td>2.530073e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.376907e+17</td>\n",
       "      <td>2.553641e+17</td>\n",
       "      <td>2.743015e+17</td>\n",
       "      <td>2.945655e+17</td>\n",
       "      <td>3.161688e+17</td>\n",
       "      <td>3.390432e+17</td>\n",
       "      <td>3.631858e+17</td>\n",
       "      <td>3.895231e+17</td>\n",
       "      <td>4.169486e+17</td>\n",
       "      <td>4.463476e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.079730e+16</td>\n",
       "      <td>2.078486e+16</td>\n",
       "      <td>2.077242e+16</td>\n",
       "      <td>2.075998e+16</td>\n",
       "      <td>2.074754e+16</td>\n",
       "      <td>2.073510e+16</td>\n",
       "      <td>2.072265e+16</td>\n",
       "      <td>2.071021e+16</td>\n",
       "      <td>2.069777e+16</td>\n",
       "      <td>2.068533e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.491269e+17</td>\n",
       "      <td>1.609964e+17</td>\n",
       "      <td>1.737868e+17</td>\n",
       "      <td>1.875378e+17</td>\n",
       "      <td>2.022633e+17</td>\n",
       "      <td>2.179179e+17</td>\n",
       "      <td>2.345022e+17</td>\n",
       "      <td>2.527081e+17</td>\n",
       "      <td>2.717244e+17</td>\n",
       "      <td>2.922129e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.405455e+17</td>\n",
       "      <td>2.382445e+17</td>\n",
       "      <td>2.359435e+17</td>\n",
       "      <td>2.336425e+17</td>\n",
       "      <td>2.313415e+17</td>\n",
       "      <td>2.290405e+17</td>\n",
       "      <td>2.267395e+17</td>\n",
       "      <td>2.244385e+17</td>\n",
       "      <td>2.221375e+17</td>\n",
       "      <td>2.198365e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>8.396414e+15</td>\n",
       "      <td>9.173453e+15</td>\n",
       "      <td>1.003567e+16</td>\n",
       "      <td>1.097871e+16</td>\n",
       "      <td>1.200347e+16</td>\n",
       "      <td>1.310722e+16</td>\n",
       "      <td>1.429071e+16</td>\n",
       "      <td>1.561696e+16</td>\n",
       "      <td>1.701622e+16</td>\n",
       "      <td>1.854882e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>3.434468e+16</td>\n",
       "      <td>3.408240e+16</td>\n",
       "      <td>3.382012e+16</td>\n",
       "      <td>3.355785e+16</td>\n",
       "      <td>3.329557e+16</td>\n",
       "      <td>3.303329e+16</td>\n",
       "      <td>3.277101e+16</td>\n",
       "      <td>3.250873e+16</td>\n",
       "      <td>3.224646e+16</td>\n",
       "      <td>3.198418e+16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 624 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             y0            y1            y2            y3            y4  \\\n",
       "0  1.511613e+17  1.626357e+17  1.749599e+17  1.881609e+17  2.022474e+17   \n",
       "1  7.621430e+16  8.255297e+16  8.942421e+16  9.684174e+16  1.048160e+17   \n",
       "2  2.376907e+17  2.553641e+17  2.743015e+17  2.945655e+17  3.161688e+17   \n",
       "3  1.491269e+17  1.609964e+17  1.737868e+17  1.875378e+17  2.022633e+17   \n",
       "4  8.396414e+15  9.173453e+15  1.003567e+16  1.097871e+16  1.200347e+16   \n",
       "\n",
       "             y5            y6            y7            y8            y9  ...  \\\n",
       "0  2.171753e+17  2.329431e+17  2.501679e+17  2.681171e+17  2.873808e+17  ...   \n",
       "1  1.133235e+17  1.223659e+17  1.323483e+17  1.428036e+17  1.541191e+17  ...   \n",
       "2  3.390432e+17  3.631858e+17  3.895231e+17  4.169486e+17  4.463476e+17  ...   \n",
       "3  2.179179e+17  2.345022e+17  2.527081e+17  2.717244e+17  2.922129e+17  ...   \n",
       "4  1.310722e+16  1.429071e+16  1.561696e+16  1.701622e+16  1.854882e+16  ...   \n",
       "\n",
       "           y614          y615          y616          y617          y618  \\\n",
       "0  2.318606e+17  2.304607e+17  2.290609e+17  2.276611e+17  2.262613e+17   \n",
       "1  2.728998e+16  2.706896e+16  2.684793e+16  2.662690e+16  2.640587e+16   \n",
       "2  2.079730e+16  2.078486e+16  2.077242e+16  2.075998e+16  2.074754e+16   \n",
       "3  2.405455e+17  2.382445e+17  2.359435e+17  2.336425e+17  2.313415e+17   \n",
       "4  3.434468e+16  3.408240e+16  3.382012e+16  3.355785e+16  3.329557e+16   \n",
       "\n",
       "           y619          y620          y621          y622          y623  \n",
       "0  2.248615e+17  2.234616e+17  2.220618e+17  2.206620e+17  2.192622e+17  \n",
       "1  2.618484e+16  2.596381e+16  2.574279e+16  2.552176e+16  2.530073e+16  \n",
       "2  2.073510e+16  2.072265e+16  2.071021e+16  2.069777e+16  2.068533e+16  \n",
       "3  2.290405e+17  2.267395e+17  2.244385e+17  2.221375e+17  2.198365e+17  \n",
       "4  3.303329e+16  3.277101e+16  3.250873e+16  3.224646e+16  3.198418e+16  \n",
       "\n",
       "[5 rows x 624 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_train.iloc[:,5:].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Temperature grid:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>300.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>300.802568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>301.605136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>302.407705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>303.210273</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            0\n",
       "0  300.000000\n",
       "1  300.802568\n",
       "2  301.605136\n",
       "3  302.407705\n",
       "4  303.210273"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_grid.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The 5 parameters $E_1$, $E_2$, $E_3$, $n_1$, and $n_2$ are physical properties of the material characterising the trapping of tritium isotopes.\n",
    "\n",
    "- $E_i$ refers to the detrapping energies of tritium trap sites in $eV$. \n",
    "- $n_i$ are the densities of the trapping site.\n",
    "\n",
    "The incrementing number $i$ relates to discrete trapping sites.\n",
    "\n",
    "It might be noticed that the trapping density $n_3$ which should compliment $E_3$ is not present. This trap is an implantation trap created through irradiation damage and is not considered in this model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### View the test data"
   ]
  },
  {
   "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>E1</th>\n",
       "      <th>E2</th>\n",
       "      <th>E3</th>\n",
       "      <th>n1</th>\n",
       "      <th>n2</th>\n",
       "      <th>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>300</th>\n",
       "      <td>0.700338</td>\n",
       "      <td>1.150021</td>\n",
       "      <td>1.710087</td>\n",
       "      <td>0.004111</td>\n",
       "      <td>0.000930</td>\n",
       "      <td>2.021919e+17</td>\n",
       "      <td>2.169475e+17</td>\n",
       "      <td>2.327300e+17</td>\n",
       "      <td>2.495828e+17</td>\n",
       "      <td>2.675127e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.974306e+17</td>\n",
       "      <td>2.932256e+17</td>\n",
       "      <td>2.890206e+17</td>\n",
       "      <td>2.848157e+17</td>\n",
       "      <td>2.806107e+17</td>\n",
       "      <td>2.764057e+17</td>\n",
       "      <td>2.722007e+17</td>\n",
       "      <td>2.679958e+17</td>\n",
       "      <td>2.637908e+17</td>\n",
       "      <td>2.595858e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>0.910138</td>\n",
       "      <td>1.132486</td>\n",
       "      <td>1.682557</td>\n",
       "      <td>0.004323</td>\n",
       "      <td>0.000657</td>\n",
       "      <td>9.677587e+15</td>\n",
       "      <td>1.061440e+16</td>\n",
       "      <td>1.165152e+16</td>\n",
       "      <td>1.278887e+16</td>\n",
       "      <td>1.402926e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>2.185340e+17</td>\n",
       "      <td>2.147866e+17</td>\n",
       "      <td>2.110392e+17</td>\n",
       "      <td>2.072918e+17</td>\n",
       "      <td>2.035444e+17</td>\n",
       "      <td>1.997969e+17</td>\n",
       "      <td>1.960495e+17</td>\n",
       "      <td>1.923021e+17</td>\n",
       "      <td>1.885547e+17</td>\n",
       "      <td>1.848073e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>0.864146</td>\n",
       "      <td>1.011063</td>\n",
       "      <td>1.535927</td>\n",
       "      <td>0.003682</td>\n",
       "      <td>0.000942</td>\n",
       "      <td>3.051326e+16</td>\n",
       "      <td>3.331790e+16</td>\n",
       "      <td>3.638907e+16</td>\n",
       "      <td>3.973493e+16</td>\n",
       "      <td>4.336344e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>6.721837e+16</td>\n",
       "      <td>6.657205e+16</td>\n",
       "      <td>6.592572e+16</td>\n",
       "      <td>6.527939e+16</td>\n",
       "      <td>6.463306e+16</td>\n",
       "      <td>6.398674e+16</td>\n",
       "      <td>6.334041e+16</td>\n",
       "      <td>6.269408e+16</td>\n",
       "      <td>6.204776e+16</td>\n",
       "      <td>6.140143e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>303</th>\n",
       "      <td>0.785364</td>\n",
       "      <td>0.914528</td>\n",
       "      <td>1.494170</td>\n",
       "      <td>0.004685</td>\n",
       "      <td>0.000446</td>\n",
       "      <td>1.469481e+17</td>\n",
       "      <td>1.591239e+17</td>\n",
       "      <td>1.722814e+17</td>\n",
       "      <td>1.864735e+17</td>\n",
       "      <td>2.017193e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.596674e+16</td>\n",
       "      <td>2.577841e+16</td>\n",
       "      <td>2.559007e+16</td>\n",
       "      <td>2.540174e+16</td>\n",
       "      <td>2.521341e+16</td>\n",
       "      <td>2.502508e+16</td>\n",
       "      <td>2.483675e+16</td>\n",
       "      <td>2.464842e+16</td>\n",
       "      <td>2.446009e+16</td>\n",
       "      <td>2.427175e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>304</th>\n",
       "      <td>0.850848</td>\n",
       "      <td>0.943984</td>\n",
       "      <td>1.476743</td>\n",
       "      <td>0.001971</td>\n",
       "      <td>0.000751</td>\n",
       "      <td>2.501309e+16</td>\n",
       "      <td>2.727755e+16</td>\n",
       "      <td>2.975959e+16</td>\n",
       "      <td>3.246143e+16</td>\n",
       "      <td>3.538895e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>3.578703e+16</td>\n",
       "      <td>3.551876e+16</td>\n",
       "      <td>3.525050e+16</td>\n",
       "      <td>3.498224e+16</td>\n",
       "      <td>3.471397e+16</td>\n",
       "      <td>3.444571e+16</td>\n",
       "      <td>3.417744e+16</td>\n",
       "      <td>3.390918e+16</td>\n",
       "      <td>3.364091e+16</td>\n",
       "      <td>3.337265e+16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 629 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           E1        E2        E3        n1        n2            y0  \\\n",
       "300  0.700338  1.150021  1.710087  0.004111  0.000930  2.021919e+17   \n",
       "301  0.910138  1.132486  1.682557  0.004323  0.000657  9.677587e+15   \n",
       "302  0.864146  1.011063  1.535927  0.003682  0.000942  3.051326e+16   \n",
       "303  0.785364  0.914528  1.494170  0.004685  0.000446  1.469481e+17   \n",
       "304  0.850848  0.943984  1.476743  0.001971  0.000751  2.501309e+16   \n",
       "\n",
       "               y1            y2            y3            y4  ...  \\\n",
       "300  2.169475e+17  2.327300e+17  2.495828e+17  2.675127e+17  ...   \n",
       "301  1.061440e+16  1.165152e+16  1.278887e+16  1.402926e+16  ...   \n",
       "302  3.331790e+16  3.638907e+16  3.973493e+16  4.336344e+16  ...   \n",
       "303  1.591239e+17  1.722814e+17  1.864735e+17  2.017193e+17  ...   \n",
       "304  2.727755e+16  2.975959e+16  3.246143e+16  3.538895e+16  ...   \n",
       "\n",
       "             y614          y615          y616          y617          y618  \\\n",
       "300  2.974306e+17  2.932256e+17  2.890206e+17  2.848157e+17  2.806107e+17   \n",
       "301  2.185340e+17  2.147866e+17  2.110392e+17  2.072918e+17  2.035444e+17   \n",
       "302  6.721837e+16  6.657205e+16  6.592572e+16  6.527939e+16  6.463306e+16   \n",
       "303  2.596674e+16  2.577841e+16  2.559007e+16  2.540174e+16  2.521341e+16   \n",
       "304  3.578703e+16  3.551876e+16  3.525050e+16  3.498224e+16  3.471397e+16   \n",
       "\n",
       "             y619          y620          y621          y622          y623  \n",
       "300  2.764057e+17  2.722007e+17  2.679958e+17  2.637908e+17  2.595858e+17  \n",
       "301  1.997969e+17  1.960495e+17  1.923021e+17  1.885547e+17  1.848073e+17  \n",
       "302  6.398674e+16  6.334041e+16  6.269408e+16  6.204776e+16  6.140143e+16  \n",
       "303  2.502508e+16  2.483675e+16  2.464842e+16  2.446009e+16  2.427175e+16  \n",
       "304  3.444571e+16  3.417744e+16  3.390918e+16  3.364091e+16  3.337265e+16  \n",
       "\n",
       "[5 rows x 629 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***\n",
    "### Uploading data to the `twinLab` cloud"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataframe is uploading.\n",
      "Processing dataset\n",
      "Dataset tritium_desorption was processed.\n"
     ]
    }
   ],
   "source": [
    "tl.set_api_key(\"SKsPqjaZiKO-eDKec3p8BQ\")\n",
    "dataset = tl.Dataset(\"tritium_desorption\")\n",
    "\n",
    "dataset.upload(df_train, verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define the columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "inputs = [\"E1\", \"E2\", \"E3\", \"n1\", \"n2\"]\n",
    "outputs = [f\"y{i}\" for i in range(len(df_grid))]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check dataset has uploaded"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>E1</th>\n",
       "      <th>E2</th>\n",
       "      <th>E3</th>\n",
       "      <th>n1</th>\n",
       "      <th>n2</th>\n",
       "      <th>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>400.000000</td>\n",
       "      <td>400.000000</td>\n",
       "      <td>400.000000</td>\n",
       "      <td>400.000000</td>\n",
       "      <td>400.000000</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>...</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "      <td>4.000000e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.847811</td>\n",
       "      <td>1.096564</td>\n",
       "      <td>1.439218</td>\n",
       "      <td>0.002780</td>\n",
       "      <td>0.000550</td>\n",
       "      <td>6.361675e+16</td>\n",
       "      <td>6.873802e+16</td>\n",
       "      <td>7.426884e+16</td>\n",
       "      <td>8.022546e+16</td>\n",
       "      <td>8.661425e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>1.576280e+17</td>\n",
       "      <td>1.556983e+17</td>\n",
       "      <td>1.537915e+17</td>\n",
       "      <td>1.519250e+17</td>\n",
       "      <td>1.501097e+17</td>\n",
       "      <td>1.483684e+17</td>\n",
       "      <td>1.466999e+17</td>\n",
       "      <td>1.450814e+17</td>\n",
       "      <td>1.435212e+17</td>\n",
       "      <td>1.420506e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.086283</td>\n",
       "      <td>0.113226</td>\n",
       "      <td>0.183739</td>\n",
       "      <td>0.001264</td>\n",
       "      <td>0.000254</td>\n",
       "      <td>6.795742e+16</td>\n",
       "      <td>7.317686e+16</td>\n",
       "      <td>7.878150e+16</td>\n",
       "      <td>8.479061e+16</td>\n",
       "      <td>9.120958e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>1.292721e+17</td>\n",
       "      <td>1.276754e+17</td>\n",
       "      <td>1.260886e+17</td>\n",
       "      <td>1.245359e+17</td>\n",
       "      <td>1.230176e+17</td>\n",
       "      <td>1.216074e+17</td>\n",
       "      <td>1.202565e+17</td>\n",
       "      <td>1.189713e+17</td>\n",
       "      <td>1.177443e+17</td>\n",
       "      <td>1.165804e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.700338</td>\n",
       "      <td>0.902457</td>\n",
       "      <td>1.103538</td>\n",
       "      <td>0.000522</td>\n",
       "      <td>0.000100</td>\n",
       "      <td>1.963621e+14</td>\n",
       "      <td>2.115386e+14</td>\n",
       "      <td>2.298551e+14</td>\n",
       "      <td>2.515298e+14</td>\n",
       "      <td>2.763921e+14</td>\n",
       "      <td>...</td>\n",
       "      <td>1.460660e+16</td>\n",
       "      <td>1.447977e+16</td>\n",
       "      <td>1.435294e+16</td>\n",
       "      <td>1.422611e+16</td>\n",
       "      <td>1.409927e+16</td>\n",
       "      <td>1.397244e+16</td>\n",
       "      <td>1.385901e+16</td>\n",
       "      <td>1.379510e+16</td>\n",
       "      <td>1.373119e+16</td>\n",
       "      <td>1.366728e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.776996</td>\n",
       "      <td>1.004006</td>\n",
       "      <td>1.280775</td>\n",
       "      <td>0.001742</td>\n",
       "      <td>0.000322</td>\n",
       "      <td>3.537697e+15</td>\n",
       "      <td>3.885903e+15</td>\n",
       "      <td>4.276821e+15</td>\n",
       "      <td>4.708195e+15</td>\n",
       "      <td>5.180239e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>5.655948e+16</td>\n",
       "      <td>5.534897e+16</td>\n",
       "      <td>5.479281e+16</td>\n",
       "      <td>5.426104e+16</td>\n",
       "      <td>5.364998e+16</td>\n",
       "      <td>5.317452e+16</td>\n",
       "      <td>5.268290e+16</td>\n",
       "      <td>5.221749e+16</td>\n",
       "      <td>5.175208e+16</td>\n",
       "      <td>5.128668e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>0.846752</td>\n",
       "      <td>1.089826</td>\n",
       "      <td>1.435188</td>\n",
       "      <td>0.002727</td>\n",
       "      <td>0.000586</td>\n",
       "      <td>3.677477e+16</td>\n",
       "      <td>3.990621e+16</td>\n",
       "      <td>4.332506e+16</td>\n",
       "      <td>4.702728e+16</td>\n",
       "      <td>5.106477e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>1.074164e+17</td>\n",
       "      <td>1.056007e+17</td>\n",
       "      <td>1.036920e+17</td>\n",
       "      <td>1.023300e+17</td>\n",
       "      <td>1.009598e+17</td>\n",
       "      <td>9.926862e+16</td>\n",
       "      <td>9.720451e+16</td>\n",
       "      <td>9.593318e+16</td>\n",
       "      <td>9.500944e+16</td>\n",
       "      <td>9.441496e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>0.923479</td>\n",
       "      <td>1.195031</td>\n",
       "      <td>1.588289</td>\n",
       "      <td>0.003888</td>\n",
       "      <td>0.000756</td>\n",
       "      <td>1.154598e+17</td>\n",
       "      <td>1.245779e+17</td>\n",
       "      <td>1.343340e+17</td>\n",
       "      <td>1.446751e+17</td>\n",
       "      <td>1.557012e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>2.294064e+17</td>\n",
       "      <td>2.254695e+17</td>\n",
       "      <td>2.217749e+17</td>\n",
       "      <td>2.188035e+17</td>\n",
       "      <td>2.164698e+17</td>\n",
       "      <td>2.153460e+17</td>\n",
       "      <td>2.142763e+17</td>\n",
       "      <td>2.121939e+17</td>\n",
       "      <td>2.109158e+17</td>\n",
       "      <td>2.095455e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>0.997323</td>\n",
       "      <td>1.299867</td>\n",
       "      <td>1.748338</td>\n",
       "      <td>0.004999</td>\n",
       "      <td>0.000995</td>\n",
       "      <td>2.376907e+17</td>\n",
       "      <td>2.553641e+17</td>\n",
       "      <td>2.743015e+17</td>\n",
       "      <td>2.945655e+17</td>\n",
       "      <td>3.161688e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>5.888639e+17</td>\n",
       "      <td>5.804446e+17</td>\n",
       "      <td>5.720253e+17</td>\n",
       "      <td>5.663434e+17</td>\n",
       "      <td>5.610157e+17</td>\n",
       "      <td>5.556881e+17</td>\n",
       "      <td>5.518733e+17</td>\n",
       "      <td>5.513362e+17</td>\n",
       "      <td>5.507992e+17</td>\n",
       "      <td>5.502622e+17</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8 rows × 629 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               E1          E2          E3          n1          n2  \\\n",
       "count  400.000000  400.000000  400.000000  400.000000  400.000000   \n",
       "mean     0.847811    1.096564    1.439218    0.002780    0.000550   \n",
       "std      0.086283    0.113226    0.183739    0.001264    0.000254   \n",
       "min      0.700338    0.902457    1.103538    0.000522    0.000100   \n",
       "25%      0.776996    1.004006    1.280775    0.001742    0.000322   \n",
       "50%      0.846752    1.089826    1.435188    0.002727    0.000586   \n",
       "75%      0.923479    1.195031    1.588289    0.003888    0.000756   \n",
       "max      0.997323    1.299867    1.748338    0.004999    0.000995   \n",
       "\n",
       "                 y0            y1            y2            y3            y4  \\\n",
       "count  4.000000e+02  4.000000e+02  4.000000e+02  4.000000e+02  4.000000e+02   \n",
       "mean   6.361675e+16  6.873802e+16  7.426884e+16  8.022546e+16  8.661425e+16   \n",
       "std    6.795742e+16  7.317686e+16  7.878150e+16  8.479061e+16  9.120958e+16   \n",
       "min    1.963621e+14  2.115386e+14  2.298551e+14  2.515298e+14  2.763921e+14   \n",
       "25%    3.537697e+15  3.885903e+15  4.276821e+15  4.708195e+15  5.180239e+15   \n",
       "50%    3.677477e+16  3.990621e+16  4.332506e+16  4.702728e+16  5.106477e+16   \n",
       "75%    1.154598e+17  1.245779e+17  1.343340e+17  1.446751e+17  1.557012e+17   \n",
       "max    2.376907e+17  2.553641e+17  2.743015e+17  2.945655e+17  3.161688e+17   \n",
       "\n",
       "       ...          y614          y615          y616          y617  \\\n",
       "count  ...  4.000000e+02  4.000000e+02  4.000000e+02  4.000000e+02   \n",
       "mean   ...  1.576280e+17  1.556983e+17  1.537915e+17  1.519250e+17   \n",
       "std    ...  1.292721e+17  1.276754e+17  1.260886e+17  1.245359e+17   \n",
       "min    ...  1.460660e+16  1.447977e+16  1.435294e+16  1.422611e+16   \n",
       "25%    ...  5.655948e+16  5.534897e+16  5.479281e+16  5.426104e+16   \n",
       "50%    ...  1.074164e+17  1.056007e+17  1.036920e+17  1.023300e+17   \n",
       "75%    ...  2.294064e+17  2.254695e+17  2.217749e+17  2.188035e+17   \n",
       "max    ...  5.888639e+17  5.804446e+17  5.720253e+17  5.663434e+17   \n",
       "\n",
       "               y618          y619          y620          y621          y622  \\\n",
       "count  4.000000e+02  4.000000e+02  4.000000e+02  4.000000e+02  4.000000e+02   \n",
       "mean   1.501097e+17  1.483684e+17  1.466999e+17  1.450814e+17  1.435212e+17   \n",
       "std    1.230176e+17  1.216074e+17  1.202565e+17  1.189713e+17  1.177443e+17   \n",
       "min    1.409927e+16  1.397244e+16  1.385901e+16  1.379510e+16  1.373119e+16   \n",
       "25%    5.364998e+16  5.317452e+16  5.268290e+16  5.221749e+16  5.175208e+16   \n",
       "50%    1.009598e+17  9.926862e+16  9.720451e+16  9.593318e+16  9.500944e+16   \n",
       "75%    2.164698e+17  2.153460e+17  2.142763e+17  2.121939e+17  2.109158e+17   \n",
       "max    5.610157e+17  5.556881e+17  5.518733e+17  5.513362e+17  5.507992e+17   \n",
       "\n",
       "               y623  \n",
       "count  4.000000e+02  \n",
       "mean   1.420506e+17  \n",
       "std    1.165804e+17  \n",
       "min    1.366728e+16  \n",
       "25%    5.128668e+16  \n",
       "50%    9.441496e+16  \n",
       "75%    2.095455e+17  \n",
       "max    5.502622e+17  \n",
       "\n",
       "[8 rows x 629 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_summary = dataset.summarise()\n",
    "dataset_summary"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dimensional reduction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "variance = dataset.analyse_output_variance(outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>Number of Dimensions</th>\n",
       "      <th>Cumulative Variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.265579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0.734992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0.895238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0.969070</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>0.989724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>0.996053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>0.998313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>0.999210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>0.999563</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Number of Dimensions  Cumulative Variance\n",
       "0                     0             0.000000\n",
       "1                     1             0.265579\n",
       "2                     2             0.734992\n",
       "3                     3             0.895238\n",
       "4                     4             0.969070\n",
       "5                     5             0.989724\n",
       "6                     6             0.996053\n",
       "7                     7             0.998313\n",
       "8                     8             0.999210\n",
       "9                     9             0.999563"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "variance.iloc[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9, 1.01)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(variance['Number of Dimensions'], variance['Cumulative Variance'], 'x--')\n",
    "plt.xscale('log')\n",
    "plt.ylim(0.9,1.01)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*** \n",
    "### Setting up an emulator"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case we are training a functional model, which means that we want to return a *function* at every point in parameter space ($E_1$, $E_2$, $E_3$, $n_1$, $n_2$). In this case, our function describes the tritium desorption rate, $D$, of the material of the reactor wall as a function of temperature (rate of emitted nuclei per reactor wall area). The training of the surrogate is agnostic to the values of reactor temperature, $T$, so we must provide this by hand (`df_grid` above, from `file_grid`). The output of our model will therefore be the function $D(T; E_1, E_2, E_3, n_1, n_2)$.\n",
    "\n",
    "The `twinLab` model achieves this by predicting the value of $D$ at $\\sim 500$ points in $T$ in a regularly-spaced grid between $300\\mathrm{K}$ and $800\\mathrm{K}$. The correlations between points adjacent in $T$ are incorporated naturally by the model, and `twinLab` provides a model uncertainty. Here we call the outputs `y`, rather than `D`, as per the typical data-science convention."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Instantiate emulator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "emulator = tl.Emulator('tritium_desorption') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Set emulator parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimator_params = tl.EstimatorParams(detrend=False, covar_module='M32', estimator_type='single_task_gp')\n",
    "train_params = tl.TrainParams(output_retained_dimensions=5,\n",
    "                              estimator_params=estimator_params, \n",
    "                              train_test_ratio=0.8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train the emulator "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "emulator.train(dataset, inputs, outputs, train_params)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Check emulator score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE  = 3.112807054807116e+34\n"
     ]
    }
   ],
   "source": [
    "print(f\"MSE  = {emulator.score(tl.ScoreParams(metric='MSE', combined_score=True))}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>E1</th>\n",
       "      <th>E2</th>\n",
       "      <th>E3</th>\n",
       "      <th>n1</th>\n",
       "      <th>n2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>300</th>\n",
       "      <td>0.700338</td>\n",
       "      <td>1.150021</td>\n",
       "      <td>1.710087</td>\n",
       "      <td>0.004111</td>\n",
       "      <td>0.000930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>301</th>\n",
       "      <td>0.910138</td>\n",
       "      <td>1.132486</td>\n",
       "      <td>1.682557</td>\n",
       "      <td>0.004323</td>\n",
       "      <td>0.000657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>302</th>\n",
       "      <td>0.864146</td>\n",
       "      <td>1.011063</td>\n",
       "      <td>1.535927</td>\n",
       "      <td>0.003682</td>\n",
       "      <td>0.000942</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>303</th>\n",
       "      <td>0.785364</td>\n",
       "      <td>0.914528</td>\n",
       "      <td>1.494170</td>\n",
       "      <td>0.004685</td>\n",
       "      <td>0.000446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>304</th>\n",
       "      <td>0.850848</td>\n",
       "      <td>0.943984</td>\n",
       "      <td>1.476743</td>\n",
       "      <td>0.001971</td>\n",
       "      <td>0.000751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           E1        E2        E3        n1        n2\n",
       "300  0.700338  1.150021  1.710087  0.004111  0.000930\n",
       "301  0.910138  1.132486  1.682557  0.004323  0.000657\n",
       "302  0.864146  1.011063  1.535927  0.003682  0.000942\n",
       "303  0.785364  0.914528  1.494170  0.004685  0.000446\n",
       "304  0.850848  0.943984  1.476743  0.001971  0.000751"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test[inputs].head()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the trained emulator on `X` ($E_1, E_2, E_3, n_1, n_2$) from the evaluation file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>y5</th>\n",
       "      <th>y6</th>\n",
       "      <th>y7</th>\n",
       "      <th>y8</th>\n",
       "      <th>y9</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.604086e+17</td>\n",
       "      <td>1.724532e+17</td>\n",
       "      <td>1.853707e+17</td>\n",
       "      <td>1.991937e+17</td>\n",
       "      <td>2.139295e+17</td>\n",
       "      <td>2.295314e+17</td>\n",
       "      <td>2.459972e+17</td>\n",
       "      <td>2.639582e+17</td>\n",
       "      <td>2.826609e+17</td>\n",
       "      <td>3.027089e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>3.125125e+17</td>\n",
       "      <td>3.087092e+17</td>\n",
       "      <td>3.049670e+17</td>\n",
       "      <td>3.013496e+17</td>\n",
       "      <td>2.977851e+17</td>\n",
       "      <td>2.943683e+17</td>\n",
       "      <td>2.910358e+17</td>\n",
       "      <td>2.878774e+17</td>\n",
       "      <td>2.848975e+17</td>\n",
       "      <td>2.820278e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6.497008e+15</td>\n",
       "      <td>7.281823e+15</td>\n",
       "      <td>8.166451e+15</td>\n",
       "      <td>9.150083e+15</td>\n",
       "      <td>1.023586e+16</td>\n",
       "      <td>1.142136e+16</td>\n",
       "      <td>1.270815e+16</td>\n",
       "      <td>1.417875e+16</td>\n",
       "      <td>1.574461e+16</td>\n",
       "      <td>1.748520e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>2.444026e+17</td>\n",
       "      <td>2.413479e+17</td>\n",
       "      <td>2.383238e+17</td>\n",
       "      <td>2.353722e+17</td>\n",
       "      <td>2.325562e+17</td>\n",
       "      <td>2.298956e+17</td>\n",
       "      <td>2.273681e+17</td>\n",
       "      <td>2.249290e+17</td>\n",
       "      <td>2.225930e+17</td>\n",
       "      <td>2.203410e+17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.723385e+16</td>\n",
       "      <td>2.986779e+16</td>\n",
       "      <td>3.275569e+16</td>\n",
       "      <td>3.591093e+16</td>\n",
       "      <td>3.934059e+16</td>\n",
       "      <td>4.303555e+16</td>\n",
       "      <td>4.699747e+16</td>\n",
       "      <td>5.143731e+16</td>\n",
       "      <td>5.612075e+16</td>\n",
       "      <td>6.124772e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>7.907983e+16</td>\n",
       "      <td>7.813263e+16</td>\n",
       "      <td>7.719632e+16</td>\n",
       "      <td>7.626738e+16</td>\n",
       "      <td>7.541711e+16</td>\n",
       "      <td>7.458494e+16</td>\n",
       "      <td>7.381477e+16</td>\n",
       "      <td>7.302499e+16</td>\n",
       "      <td>7.222869e+16</td>\n",
       "      <td>7.147155e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.530478e+17</td>\n",
       "      <td>1.654752e+17</td>\n",
       "      <td>1.788941e+17</td>\n",
       "      <td>1.933508e+17</td>\n",
       "      <td>2.088636e+17</td>\n",
       "      <td>2.253865e+17</td>\n",
       "      <td>2.429211e+17</td>\n",
       "      <td>2.622293e+17</td>\n",
       "      <td>2.824274e+17</td>\n",
       "      <td>3.042431e+17</td>\n",
       "      <td>...</td>\n",
       "      <td>5.700515e+16</td>\n",
       "      <td>5.612460e+16</td>\n",
       "      <td>5.524457e+16</td>\n",
       "      <td>5.433518e+16</td>\n",
       "      <td>5.344123e+16</td>\n",
       "      <td>5.250270e+16</td>\n",
       "      <td>5.154834e+16</td>\n",
       "      <td>5.066236e+16</td>\n",
       "      <td>4.981514e+16</td>\n",
       "      <td>4.896772e+16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3.021194e+16</td>\n",
       "      <td>3.290791e+16</td>\n",
       "      <td>3.584748e+16</td>\n",
       "      <td>3.904069e+16</td>\n",
       "      <td>4.249362e+16</td>\n",
       "      <td>4.619667e+16</td>\n",
       "      <td>5.015138e+16</td>\n",
       "      <td>5.455264e+16</td>\n",
       "      <td>5.918040e+16</td>\n",
       "      <td>6.422099e+16</td>\n",
       "      <td>...</td>\n",
       "      <td>5.264406e+16</td>\n",
       "      <td>5.194033e+16</td>\n",
       "      <td>5.124237e+16</td>\n",
       "      <td>5.053531e+16</td>\n",
       "      <td>4.987387e+16</td>\n",
       "      <td>4.922421e+16</td>\n",
       "      <td>4.859895e+16</td>\n",
       "      <td>4.798611e+16</td>\n",
       "      <td>4.738708e+16</td>\n",
       "      <td>4.680578e+16</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 624 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             y0            y1            y2            y3            y4  \\\n",
       "0  1.604086e+17  1.724532e+17  1.853707e+17  1.991937e+17  2.139295e+17   \n",
       "1  6.497008e+15  7.281823e+15  8.166451e+15  9.150083e+15  1.023586e+16   \n",
       "2  2.723385e+16  2.986779e+16  3.275569e+16  3.591093e+16  3.934059e+16   \n",
       "3  1.530478e+17  1.654752e+17  1.788941e+17  1.933508e+17  2.088636e+17   \n",
       "4  3.021194e+16  3.290791e+16  3.584748e+16  3.904069e+16  4.249362e+16   \n",
       "\n",
       "             y5            y6            y7            y8            y9  ...  \\\n",
       "0  2.295314e+17  2.459972e+17  2.639582e+17  2.826609e+17  3.027089e+17  ...   \n",
       "1  1.142136e+16  1.270815e+16  1.417875e+16  1.574461e+16  1.748520e+16  ...   \n",
       "2  4.303555e+16  4.699747e+16  5.143731e+16  5.612075e+16  6.124772e+16  ...   \n",
       "3  2.253865e+17  2.429211e+17  2.622293e+17  2.824274e+17  3.042431e+17  ...   \n",
       "4  4.619667e+16  5.015138e+16  5.455264e+16  5.918040e+16  6.422099e+16  ...   \n",
       "\n",
       "           y614          y615          y616          y617          y618  \\\n",
       "0  3.125125e+17  3.087092e+17  3.049670e+17  3.013496e+17  2.977851e+17   \n",
       "1  2.444026e+17  2.413479e+17  2.383238e+17  2.353722e+17  2.325562e+17   \n",
       "2  7.907983e+16  7.813263e+16  7.719632e+16  7.626738e+16  7.541711e+16   \n",
       "3  5.700515e+16  5.612460e+16  5.524457e+16  5.433518e+16  5.344123e+16   \n",
       "4  5.264406e+16  5.194033e+16  5.124237e+16  5.053531e+16  4.987387e+16   \n",
       "\n",
       "           y619          y620          y621          y622          y623  \n",
       "0  2.943683e+17  2.910358e+17  2.878774e+17  2.848975e+17  2.820278e+17  \n",
       "1  2.298956e+17  2.273681e+17  2.249290e+17  2.225930e+17  2.203410e+17  \n",
       "2  7.458494e+16  7.381477e+16  7.302499e+16  7.222869e+16  7.147155e+16  \n",
       "3  5.250270e+16  5.154834e+16  5.066236e+16  4.981514e+16  4.896772e+16  \n",
       "4  4.922421e+16  4.859895e+16  4.798611e+16  4.738708e+16  4.680578e+16  \n",
       "\n",
       "[5 rows x 624 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>y0</th>\n",
       "      <th>y1</th>\n",
       "      <th>y2</th>\n",
       "      <th>y3</th>\n",
       "      <th>y4</th>\n",
       "      <th>y5</th>\n",
       "      <th>y6</th>\n",
       "      <th>y7</th>\n",
       "      <th>y8</th>\n",
       "      <th>y9</th>\n",
       "      <th>...</th>\n",
       "      <th>y614</th>\n",
       "      <th>y615</th>\n",
       "      <th>y616</th>\n",
       "      <th>y617</th>\n",
       "      <th>y618</th>\n",
       "      <th>y619</th>\n",
       "      <th>y620</th>\n",
       "      <th>y621</th>\n",
       "      <th>y622</th>\n",
       "      <th>y623</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.333374e+15</td>\n",
       "      <td>4.635510e+15</td>\n",
       "      <td>4.957953e+15</td>\n",
       "      <td>5.300706e+15</td>\n",
       "      <td>5.663830e+15</td>\n",
       "      <td>6.046135e+15</td>\n",
       "      <td>6.447603e+15</td>\n",
       "      <td>6.881517e+15</td>\n",
       "      <td>7.331366e+15</td>\n",
       "      <td>7.810235e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>9.774266e+15</td>\n",
       "      <td>9.635869e+15</td>\n",
       "      <td>9.498771e+15</td>\n",
       "      <td>9.363464e+15</td>\n",
       "      <td>9.227142e+15</td>\n",
       "      <td>9.098218e+15</td>\n",
       "      <td>8.967923e+15</td>\n",
       "      <td>8.851280e+15</td>\n",
       "      <td>8.746417e+15</td>\n",
       "      <td>8.642351e+15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.955286e+15</td>\n",
       "      <td>4.231417e+15</td>\n",
       "      <td>4.526136e+15</td>\n",
       "      <td>4.839459e+15</td>\n",
       "      <td>5.171448e+15</td>\n",
       "      <td>5.521015e+15</td>\n",
       "      <td>5.888144e+15</td>\n",
       "      <td>6.285020e+15</td>\n",
       "      <td>6.696512e+15</td>\n",
       "      <td>7.134619e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>8.903862e+15</td>\n",
       "      <td>8.777929e+15</td>\n",
       "      <td>8.653179e+15</td>\n",
       "      <td>8.530079e+15</td>\n",
       "      <td>8.406089e+15</td>\n",
       "      <td>8.288827e+15</td>\n",
       "      <td>8.170363e+15</td>\n",
       "      <td>8.064267e+15</td>\n",
       "      <td>7.968849e+15</td>\n",
       "      <td>7.874176e+15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.991605e+15</td>\n",
       "      <td>4.270230e+15</td>\n",
       "      <td>4.567609e+15</td>\n",
       "      <td>4.883755e+15</td>\n",
       "      <td>5.218729e+15</td>\n",
       "      <td>5.571435e+15</td>\n",
       "      <td>5.941856e+15</td>\n",
       "      <td>6.342282e+15</td>\n",
       "      <td>6.757451e+15</td>\n",
       "      <td>7.199463e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>8.987640e+15</td>\n",
       "      <td>8.860506e+15</td>\n",
       "      <td>8.734565e+15</td>\n",
       "      <td>8.610289e+15</td>\n",
       "      <td>8.485111e+15</td>\n",
       "      <td>8.366724e+15</td>\n",
       "      <td>8.247119e+15</td>\n",
       "      <td>8.140007e+15</td>\n",
       "      <td>8.043679e+15</td>\n",
       "      <td>7.948100e+15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.173933e+15</td>\n",
       "      <td>4.465102e+15</td>\n",
       "      <td>4.775854e+15</td>\n",
       "      <td>5.106195e+15</td>\n",
       "      <td>5.456188e+15</td>\n",
       "      <td>5.824687e+15</td>\n",
       "      <td>6.211673e+15</td>\n",
       "      <td>6.629966e+15</td>\n",
       "      <td>7.063639e+15</td>\n",
       "      <td>7.525317e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>9.407247e+15</td>\n",
       "      <td>9.274105e+15</td>\n",
       "      <td>9.142214e+15</td>\n",
       "      <td>9.012053e+15</td>\n",
       "      <td>8.880931e+15</td>\n",
       "      <td>8.756924e+15</td>\n",
       "      <td>8.631617e+15</td>\n",
       "      <td>8.519422e+15</td>\n",
       "      <td>8.418541e+15</td>\n",
       "      <td>8.318435e+15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.008248e+15</td>\n",
       "      <td>4.288025e+15</td>\n",
       "      <td>4.586632e+15</td>\n",
       "      <td>4.904082e+15</td>\n",
       "      <td>5.240438e+15</td>\n",
       "      <td>5.594596e+15</td>\n",
       "      <td>5.966542e+15</td>\n",
       "      <td>6.368614e+15</td>\n",
       "      <td>6.785488e+15</td>\n",
       "      <td>7.229314e+15</td>\n",
       "      <td>...</td>\n",
       "      <td>9.025623e+15</td>\n",
       "      <td>8.897947e+15</td>\n",
       "      <td>8.771471e+15</td>\n",
       "      <td>8.646665e+15</td>\n",
       "      <td>8.520952e+15</td>\n",
       "      <td>8.402060e+15</td>\n",
       "      <td>8.281943e+15</td>\n",
       "      <td>8.174373e+15</td>\n",
       "      <td>8.077635e+15</td>\n",
       "      <td>7.981649e+15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 624 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             y0            y1            y2            y3            y4  \\\n",
       "0  4.333374e+15  4.635510e+15  4.957953e+15  5.300706e+15  5.663830e+15   \n",
       "1  3.955286e+15  4.231417e+15  4.526136e+15  4.839459e+15  5.171448e+15   \n",
       "2  3.991605e+15  4.270230e+15  4.567609e+15  4.883755e+15  5.218729e+15   \n",
       "3  4.173933e+15  4.465102e+15  4.775854e+15  5.106195e+15  5.456188e+15   \n",
       "4  4.008248e+15  4.288025e+15  4.586632e+15  4.904082e+15  5.240438e+15   \n",
       "\n",
       "             y5            y6            y7            y8            y9  ...  \\\n",
       "0  6.046135e+15  6.447603e+15  6.881517e+15  7.331366e+15  7.810235e+15  ...   \n",
       "1  5.521015e+15  5.888144e+15  6.285020e+15  6.696512e+15  7.134619e+15  ...   \n",
       "2  5.571435e+15  5.941856e+15  6.342282e+15  6.757451e+15  7.199463e+15  ...   \n",
       "3  5.824687e+15  6.211673e+15  6.629966e+15  7.063639e+15  7.525317e+15  ...   \n",
       "4  5.594596e+15  5.966542e+15  6.368614e+15  6.785488e+15  7.229314e+15  ...   \n",
       "\n",
       "           y614          y615          y616          y617          y618  \\\n",
       "0  9.774266e+15  9.635869e+15  9.498771e+15  9.363464e+15  9.227142e+15   \n",
       "1  8.903862e+15  8.777929e+15  8.653179e+15  8.530079e+15  8.406089e+15   \n",
       "2  8.987640e+15  8.860506e+15  8.734565e+15  8.610289e+15  8.485111e+15   \n",
       "3  9.407247e+15  9.274105e+15  9.142214e+15  9.012053e+15  8.880931e+15   \n",
       "4  9.025623e+15  8.897947e+15  8.771471e+15  8.646665e+15  8.520952e+15   \n",
       "\n",
       "           y619          y620          y621          y622          y623  \n",
       "0  9.098218e+15  8.967923e+15  8.851280e+15  8.746417e+15  8.642351e+15  \n",
       "1  8.288827e+15  8.170363e+15  8.064267e+15  7.968849e+15  7.874176e+15  \n",
       "2  8.366724e+15  8.247119e+15  8.140007e+15  8.043679e+15  7.948100e+15  \n",
       "3  8.756924e+15  8.631617e+15  8.519422e+15  8.418541e+15  8.318435e+15  \n",
       "4  8.402060e+15  8.281943e+15  8.174373e+15  8.077635e+15  7.981649e+15  \n",
       "\n",
       "[5 rows x 624 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_mean, df_std = emulator.predict(df_test[inputs], verbose=False)\n",
    "\n",
    "display(df_mean.head())\n",
    "display(df_std.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "*** \n",
    "### Test the emulator "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x120ba9ad0>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "i = 50\n",
    "\n",
    "if i > len(df_test): \n",
    "    raise ValueError('The index given is out of the range of the test dataframe')\n",
    "\n",
    "df_mean_i = df_mean.iloc[i].values\n",
    "df_std_i = df_std.iloc[i].values\n",
    "df_test_i = df_test.iloc[i,5::].values\n",
    "params = df_test.iloc[i,:5].values\n",
    "\n",
    "fig = plt.figure(figsize=(8,6))\n",
    "plt.plot(df_grid[::5], df_test_i[::5],'xr', label='Test data')\n",
    "plt.plot(df_grid, df_mean_i, label='GP Prediction')\n",
    "plt.fill_between(df_grid[0], df_mean_i-2*df_std_i, df_mean_i+2*df_std_i, alpha=0.4)\n",
    "plt.title(f'D(E1={round(params[0],2)},E2={round(params[1],2)},E3={round(params[2],2)}, n1={params[3]:.2e}, n2={params[4]:.2e})')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load the streamlit App"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[0m\n",
      "\u001b[34m\u001b[1m  You can now view your Streamlit app in your browser.\u001b[0m\n",
      "\u001b[0m\n",
      "\u001b[34m  Local URL: \u001b[0m\u001b[1mhttp://localhost:8501\u001b[0m\n",
      "\u001b[34m  Network URL: \u001b[0m\u001b[1mhttp://192.168.2.25:8501\u001b[0m\n",
      "\u001b[0m\n",
      "\u001b[34m\u001b[1m  For better performance, install the Watchdog module:\u001b[0m\n",
      "\n",
      "  $ xcode-select --install\n",
      "  $ pip install watchdog\n",
      "            \u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!streamlit run app.py"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.7"
  },
  "vscode": {
   "interpreter": {
    "hash": "8b4841770c9e34caa6ae18daa4abc2a1f1538c85b37a4e54ebb9908de806c0dc"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}