prediction_plotter.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import configparser\n",
    "from customDatasetMakers import state_to_dic, ian_dataset\n",
    "from dataSettings import get_denormalized_dic, get_normalized_dic\n",
    "from customModels import IanRNN, HiroLRAN\n",
    "from train_helpers import get_state_mask, get_sample_time_state_mask, masked_loss\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from prediction_helpers import get_ml_truth, get_ml_profile_warmup, get_ml_actuator_trajectory, get_ml_predictions, get_considered_models, get_fast_profile_prediction\n",
    "import pickle\n",
    "import os\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "LOAD MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using /projects/EKOLEMEN/profile_predictor/joe_hiro_models/HiroLRAN_betan7.tar\n"
     ]
    }
   ],
   "source": [
    "#model_name = 'HiroLRAN_everything'\n",
    "model_name = 'HiroLRAN_betan7'\n",
    "config_filename = f'/projects/EKOLEMEN/profile_predictor/joe_hiro_models/{model_name}config'\n",
    "config=configparser.ConfigParser()\n",
    "config.read(config_filename)\n",
    "output_filename_base=config['model']['output_filename_base']\n",
    "output_dir=config['model']['output_dir']\n",
    "profiles=config['inputs']['profiles'].split()\n",
    "actuators=config['inputs']['actuators'].split()\n",
    "parameters=config['inputs'].get('parameters','').split()\n",
    "calculations=config['inputs'].get('calculations','').split()\n",
    "nwarmup = 3\n",
    "prediction_length = 300\n",
    "\n",
    "considered_model = get_considered_models(config_filename, ensemble=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x14a77bafb890>]"
      ]
     },
     "execution_count": 25,
     "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": [
    "model_file=os.path.join(output_dir, f'{output_filename_base}.tar')\n",
    "saved_state=torch.load(model_file, map_location=torch.device('cpu'))\n",
    "plt.plot(saved_state['train_losses'][5:],c='r',label='train')\n",
    "plt.plot(saved_state['val_losses'][5:],c='b',label='validation')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "LOAD DATABASE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_filename = config['preprocess']['preprocessed_data_filenamebase'] + 'val.pkl'\n",
    "x_test, y_test, shots, start_times = ian_dataset(data_filename,profiles,parameters,calculations,actuators,sort_by_size=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "SELECT SHOT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shot 175285\n",
      "Start time 320\n"
     ]
    }
   ],
   "source": [
    "shot_index = 10\n",
    "print(f'Shot {shots[shot_index]}')\n",
    "print(f'Start time {start_times[shot_index]}')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "VISUALIZE MODEL PREDICTONS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Bucket 1/1 took 0s\n",
      "Took 0.19 s\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "actuator_trajectory = get_ml_actuator_trajectory(x_test[shot_index:shot_index+1], profiles, parameters, calculations, actuators, nwarmup=nwarmup, prediction_length=prediction_length)\n",
    "'''for i, actuator in enumerate(actuators):\n",
    "    plt.plot(actuator_trajectory[shot_index, i, :])\n",
    "    plt.ylabel(actuator)\n",
    "    plt.show()'''\n",
    "\n",
    "warmup_profiles, warmup_parameters = get_ml_profile_warmup(x_test[shot_index:shot_index+1], profiles, parameters, calculations, actuators, recorded_profiles=profiles, recorded_parameters=parameters, nwarmup=nwarmup)\n",
    "true_profiles, true_params = get_ml_truth(y_test[shot_index:shot_index+1], profiles, parameters, calculations, nwarmup=nwarmup, prediction_length=-1)\n",
    "predicted_profiles, predicted_params = get_ml_predictions(x_test[shot_index:shot_index+1], y_test[shot_index:shot_index+1],\n",
    "                profiles, parameters, calculations, actuators,\n",
    "                considered_models=considered_model,\n",
    "                recorded_profiles=profiles,\n",
    "                recorded_actuators=actuators,\n",
    "                recorded_parameters=parameters,\n",
    "                prediction_length=-1,nwarmup=3,\n",
    "                use_fancy_normalization=False,\n",
    "                bucket_size=10000)\n",
    "\n",
    "combined_true_profiles = np.concatenate((warmup_profiles, true_profiles), axis=2)\n",
    "combined_predicted_profiles = np.concatenate((warmup_profiles, predicted_profiles), axis=2)\n",
    "combined_true_params = np.concatenate((warmup_parameters, true_params), axis=2)\n",
    "combined_predicted_params = np.concatenate((warmup_parameters, predicted_params), axis=2)\n",
    "for i, actuator in enumerate(actuators):\n",
    "    plt.plot(actuator_trajectory[0, i, :])\n",
    "    plt.plot(actuator_trajectory[0, i, :])\n",
    "    plt.ylabel(actuator)\n",
    "    plt.show()\n",
    "for i, parameter in enumerate(parameters):\n",
    "    plt.plot(combined_true_params[0, i, :], label='True')\n",
    "    plt.plot(combined_predicted_params[0, i, :], label='Predicted')\n",
    "    plt.ylabel(parameter)\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "for i, profile in enumerate(profiles):\n",
    "    plt.plot(combined_true_profiles[0, i, :, 0])\n",
    "    plt.plot(combined_predicted_profiles[0, i, :, 0])\n",
    "    plt.ylabel(profile)\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "PLOTS FROM ROLLOUT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('/projects/EKOLEMEN/profile_predictor/joe_hiro_models/test_rollout_alldiiid_ensemble.pkl', 'rb') as f:\n",
    "    data = pickle.load(f)\n",
    "with open('/projects/EKOLEMEN/profile_predictor/joe_hiro_models/rollout_alldiiid_ensemble.pkl', 'rb') as f:\n",
    "    data2 = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "directory = '/projects/EKOLEMEN/profile_predictor/joe_hiro_models/'\n",
    "\n",
    "with open(f'{directory}rollout_{model_name}.pkl', 'rb') as f:\n",
    "    all_data = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "ensemble_model_predictions_list = []\n",
    "for i in range(4):\n",
    "    with open(f'{directory}rollout_alldiiid_ensemble{i}_RESUMED.pkl', 'rb') as f:\n",
    "        all_data = pickle.load(f)\n",
    "        ensemble_model_predictions_list.append(all_data[list(all_data.keys())[index]])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import animation\n",
    "from IPython.display import HTML\n",
    "\n",
    "rhos = np.arange(33)/33\n",
    "profile_labels = ['Te / keV', 'Ti / keV', 'Rot / Hz', r'ne/$10^{19}m^{-3}$', 'q', r'$Z_{eff}$']\n",
    "actuator_labels = ['P_NBI / MW', 'torque / Nm', 'Ip / A','Bt / T','gas', 'P_ECH / MW', 'Triangularity bot', 'Triangularity top', r'$\\kappa$', 'a minor / m', r'Volume $m^3$', 'R major / m']\n",
    "timetrace = data['truth']['actuators'][actuators[0]]\n",
    "\n",
    "fig, axes = plt.subplots(nrows=6, ncols=3, figsize=(16, 12))\n",
    "#axes = axes.flatten()\n",
    "axes = np.concatenate([axes[:,0], axes[:,1], axes[:,2]], axis=0)\n",
    "# Initialize line objects for each subplot\n",
    "predictions = [ax.plot([], [], lw=2, label='Predictions')[0] for ax in axes]\n",
    "targets = [ax.plot([], [], lw=2, label='Truth')[0] for ax in axes]\n",
    "ensemble_1 = [ax.plot([], [], lw=2, color='C0', alpha=0.3)[0] for ax in axes]\n",
    "ensemble_2 = [ax.plot([], [], lw=2, color='C0', alpha=0.3)[0] for ax in axes]\n",
    "ensemble_3 = [ax.plot([], [], lw=2, color='C0', alpha=0.3)[0] for ax in axes]\n",
    "ensemble_4 = [ax.plot([], [], lw=2, color='C0', alpha=0.3)[0] for ax in axes]\n",
    "# Set subplot limits and labels\n",
    "for i, ax in enumerate(axes):\n",
    "    if i < len(actuators):\n",
    "        if actuator_labels[i] == 'P_NBI / MW':\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2/1000)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]]/1000, alpha=0.3)\n",
    "        elif actuator_labels[i] == 'P_ECH / MW':\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2/1000000)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]]/1000000, alpha=0.3)\n",
    "        else:\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]], alpha=0.3)\n",
    "    else:\n",
    "        ax.set_xlim(0, 1)\n",
    "        if i -len(actuators) < len(profiles):\n",
    "            ax.set_ylim(0, np.max(data['truth']['profiles'][profiles[i-len(actuators)]][:])*1.2)\n",
    "            ax.set_ylabel(profile_labels[i-len(actuators)])\n",
    "        ax.legend()\n",
    "axes[5].set_xlabel('Time / ms')\n",
    "axes[11].set_xlabel('Time / ms')\n",
    "axes[-1].set_xlabel('rho')\n",
    "\n",
    "def animate(time):\n",
    "    for i in range(len(axes)):\n",
    "        if i < len(actuators):\n",
    "            if actuator_labels[i] == 'P_NBI / MW':\n",
    "                predictions[i].set_data(np.arange(len(timetrace[:time]))*20, data['truth']['actuators'][actuators[i]][:time]/1000)\n",
    "            else:\n",
    "                predictions[i].set_data(np.arange(len(timetrace[:time]))*20, data['truth']['actuators'][actuators[i]][:time])\n",
    "            #targets[i].set_data(np.arange(len(timetrace[:time]))*20, control_model['controlled'][controller_actuators[i]][:time,0][:])\n",
    "        else:\n",
    "            \n",
    "            predictions[i].set_data(rhos, np.mean(np.array([ensemble_model_predictions_list[j]['predictions']['profiles'][profiles[i-len(actuators)]][0][time] for j in range(4)]), axis=0))\n",
    "            \n",
    "            targets[i].set_data(rhos, data['truth']['profiles'][profiles[i-len(actuators)]][time])\n",
    "            ensemble_1[i].set_data(rhos, ensemble_model_predictions_list[0]['predictions']['profiles'][profiles[i-len(actuators)]][0][time])\n",
    "            ensemble_2[i].set_data(rhos, ensemble_model_predictions_list[1]['predictions']['profiles'][profiles[i-len(actuators)]][0][time])\n",
    "            ensemble_3[i].set_data(rhos, ensemble_model_predictions_list[2]['predictions']['profiles'][profiles[i-len(actuators)]][0][time])\n",
    "            ensemble_4[i].set_data(rhos, ensemble_model_predictions_list[3]['predictions']['profiles'][profiles[i-len(actuators)]][0][time])\n",
    "    return predictions + targets\n",
    "fig.suptitle('Simulation of shot 185885', fontsize=16)\n",
    "ani = animation.FuncAnimation(fig, animate, frames=len(timetrace), interval=50, blit=True)\n",
    "ani.save(f'{model_name}{index}.mp4', writer='ffmpeg')\n",
    "HTML(ani.to_html5_video())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4, 271)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "full_arrs = np.array([ensemble_model_predictions_list[j]['predictions']['profiles'][profiles[15-len(actuators)]][0] for j in range(4)])\n",
    "print(np.mean(full_arrs, axis=-1).shape)\n",
    "#print(np.mean(np.array([ensemble_model_predictions_list[j]['predictions']['profiles'][profiles[15-len(actuators)]][0][:, :] for j in range(4)]), axis=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, axes = plt.subplots(nrows=len(data['truth']['actuators'].keys())+len(data['truth']['parameters'].keys())+len(data['truth']['profiles'].keys()), ncols=1, figsize=(8, 20), sharex=True)\n",
    "actuator_labels = ['P_NBI / W', 'torque', 'Ip / A','Bt / T','gas', 'P_ECH / W']\n",
    "#profile_labels = ['Te / keV', 'Ti / keV', 'Rot / Hz', r'ne/$10^{19}m^{-3}$', 'q']\n",
    "#parameter_labels = ['betaN']\n",
    "actuator_labels = list(data['truth']['actuators'].keys())\n",
    "profile_labels = list(data['truth']['profiles'].keys())\n",
    "parameter_labels = list(data['truth']['parameters'].keys())\n",
    "time = np.arange(len(data['truth']['actuators'][actuators[0]]))*20\n",
    "\n",
    "for i, ax in enumerate(axes):\n",
    "    if i < len(actuators):\n",
    "        if actuator_labels[i] == 'P_NBI / MW':\n",
    "            ax.plot(time, data['truth']['actuators'][actuators[i]]/1000)\n",
    "            ax.set_ylabel(f'{actuator_labels[i]}')\n",
    "            ax.tick_params(axis='y', which='both', labelleft=True, labelright=False)\n",
    "            ax.yaxis.set_label_position('left')\n",
    "        else:\n",
    "            ax.plot(time, data['truth']['actuators'][actuators[i]])\n",
    "            ax.set_ylabel(f'{actuator_labels[i]}')\n",
    "            ax.tick_params(axis='y', which='both', labelleft=True, labelright=False)\n",
    "            ax.yaxis.set_label_position('left')\n",
    "    elif i < len(actuators) + len(parameters):\n",
    "        ax.plot(time[:], data['predictions']['parameters'][parameters[i-len(actuators)]][0], label='predicted')\n",
    "        ax.plot(time, data['truth']['parameters'][parameters[i-len(actuators)]], label='truth')\n",
    "        ax.set_ylabel(f'{parameter_labels[i-len(actuators)]}')\n",
    "        ax.tick_params(axis='y', which='both', labelleft=True, labelright=False)\n",
    "        ax.yaxis.set_label_position('left')\n",
    "    else:\n",
    "\n",
    "        #predicted_mean = np.mean(data['predictions']['profiles'][profiles[i - len(actuators) - len(parameters)]][0], axis=1)\n",
    "        predicted_mean = np.mean(np.mean(np.array([ensemble_model_predictions_list[j]['predictions']['profiles'][profiles[i-len(actuators)]][0] for j in range(4)]), axis=-1), axis=0)\n",
    "        predicted_std = np.std(np.mean(np.array([ensemble_model_predictions_list[j]['predictions']['profiles'][profiles[i-len(actuators)]][0] for j in range(4)]), axis=-1), axis=0)\n",
    "        truth_mean = np.mean(data['truth']['profiles'][profiles[i - len(actuators) - len(parameters)]], axis=1)\n",
    "        ax.plot(time, predicted_mean, label='predicted')\n",
    "        ax.plot(time, truth_mean, label='truth')\n",
    "        ax.set_ylabel(f'{profile_labels[i-len(actuators)-len(parameters)]}')\n",
    "        ax.tick_params(axis='y', which='both', labelleft=True, labelright=False)\n",
    "        ax.yaxis.set_label_position('left')\n",
    "        #ax.set_ylim(bottom=0)\n",
    "\n",
    "# Set the x-axis label for the bottom subplot\n",
    "axes[-1].set_xlabel('Time / ms')\n",
    "axes[len(actuators)].legend()\n",
    "axes[0].set_title('Actuators')\n",
    "axes[len(actuators)].set_title('Predicted Profiles')\n",
    "plt.subplots_adjust(hspace=0.1)\n",
    "plt.savefig(f'{model_name}{index}.png')\n",
    "# Display the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import animation\n",
    "from IPython.display import HTML\n",
    "\n",
    "rhos = np.arange(33)/33\n",
    "profile_labels = ['Te / keV', 'Ti / keV', 'Rot / Hz', r'ne/$10^{19}m^{-3}$', 'q']\n",
    "actuator_labels = ['P_NBI / MW', 'torque / Nm', 'Ip / A','Bt / T','gas', 'P_ECH / MW', 'Triangularity bot', 'Triangularity top', r'$\\kappa$', 'a minor / m', r'Volume $m^3$', 'R major / m']\n",
    "timetrace = data['truth']['actuators'][actuators[0]]\n",
    "\n",
    "fig, axes = plt.subplots(nrows=6, ncols=3, figsize=(16, 12))\n",
    "#axes = axes.flatten()\n",
    "axes = np.concatenate([axes[:,0], axes[:,1], axes[:,2]], axis=0)[:-1]\n",
    "# Initialize line objects for each subplot\n",
    "predictions = [ax.plot([], [], lw=2, label='Predictions')[0] for ax in axes]\n",
    "targets = [ax.plot([], [], lw=2, label='Truth')[0] for ax in axes]\n",
    "\n",
    "# Set subplot limits and labels\n",
    "for i, ax in enumerate(axes):\n",
    "    if i < len(actuators):\n",
    "        if actuator_labels[i] == 'P_NBI / MW':\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2/1000)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]]/1000, alpha=0.2)\n",
    "        elif actuator_labels[i] == 'P_ECH / MW':\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2/1000000)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]]/1000000, alpha=0.2)\n",
    "        else:\n",
    "            ax.set_xlim(0, 20*len(timetrace))\n",
    "            ax.set_ylim(0, np.max(data['truth']['actuators'][actuators[i]])*1.2)\n",
    "            ax.set_ylabel(actuator_labels[i])\n",
    "            ax.plot(np.arange(len(timetrace[:]))*20, data['truth']['actuators'][actuators[i]], alpha=0.2)\n",
    "    else:\n",
    "        ax.set_xlim(0, 1)\n",
    "        if i -len(actuators) < len(profiles):\n",
    "            ax.set_ylim(0, np.max(data['truth']['profiles'][profiles[i-len(actuators)]][:])*1.2)\n",
    "            ax.set_ylabel(profile_labels[i-len(actuators)])\n",
    "        ax.legend()\n",
    "axes[len(actuators)].set_xlabel('Time / ms')\n",
    "axes[-1].set_xlabel('rho')\n",
    "\n",
    "def animate(time):\n",
    "    for i in range(len(axes)):\n",
    "        if i < len(actuators):\n",
    "            if actuator_labels[i] == 'P_NBI / MW':\n",
    "                predictions[i].set_data(np.arange(len(timetrace[:time]))*20, data['truth']['actuators'][actuators[i]][:time]/1000)\n",
    "            else:\n",
    "                predictions[i].set_data(np.arange(len(timetrace[:time]))*20, data['truth']['actuators'][actuators[i]][:time])\n",
    "            #targets[i].set_data(np.arange(len(timetrace[:time]))*20, control_model['controlled'][controller_actuators[i]][:time,0][:])\n",
    "        else:\n",
    "            predictions[i].set_data(rhos, data['predictions']['profiles'][profiles[i-len(actuators)]][0][time])\n",
    "            targets[i].set_data(rhos, data['truth']['profiles'][profiles[i-len(actuators)]][time])\n",
    "    return predictions + targets\n",
    "\n",
    "ani = animation.FuncAnimation(fig, animate, frames=6, interval=150, blit=True)\n",
    "ani.save(f'{model_name}{index}.mp4', writer='ffmpeg')\n",
    "HTML(ani.to_html5_video())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}