tm_model_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hf8585/.conda/envs/torch/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import h5py as h5\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from tqdm import tqdm\n",
    "import pickle\n",
    "import sys\n",
    "sys.path.append(os.path.expanduser(\"~/TMPredictor/survival_tm/auton-survival\"))\n",
    "from auton_survival.preprocessing import Scaler\n",
    "import optuna\n",
    "from sklearn.model_selection import ParameterGrid\n",
    "#sys.path.append('/projects/EKOLEMEN/survival_tm/train_models/auton-survival')\n",
    "sys.path.append(os.path.expanduser(\"~/TMPredictor/survival_tm/auton-survival\"))\n",
    "from sklearn.model_selection import ParameterGrid\n",
    "from auton_survival.estimators import SurvivalModel\n",
    "from auton_survival.metrics import survival_regression_metric\n",
    "from auton_survival.models.dsm import DeepSurvivalMachines\n",
    "from sksurv.metrics import concordance_index_ipcw, brier_score, cumulative_dynamic_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"with open('data/x_train_memory_normed.pkl', 'rb') as f:\\n    x_train_df = pickle.load(f)\\nwith open('data/x_valid_memory_normed.pkl', 'rb') as f:\\n    x_valid_df = pickle.load(f)\\nwith open('data/outcomes_train_memory.pkl', 'rb') as f:\\n    outcomes_train_df = pickle.load(f)\\nwith open('data/outcomes_valid_memory.pkl', 'rb') as f:\\n    outcomes_valid_df = pickle.load(f)\""
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open('data/x_train_df_normed.pkl', 'rb') as f:\n",
    "    x_train_df = pickle.load(f)\n",
    "with open('data/x_valid_df_normed.pkl', 'rb') as f:\n",
    "    x_valid_df = pickle.load(f)\n",
    "with open('data/outcomes_train_df.pkl', 'rb') as f:\n",
    "    outcomes_train_df = pickle.load(f)\n",
    "with open('data/outcomes_valid_df.pkl', 'rb') as f:\n",
    "    outcomes_valid_df = pickle.load(f)\n",
    "\n",
    "'''with open('data/x_train_memory_normed.pkl', 'rb') as f:\n",
    "    x_train_df = pickle.load(f)\n",
    "with open('data/x_valid_memory_normed.pkl', 'rb') as f:\n",
    "    x_valid_df = pickle.load(f)\n",
    "with open('data/outcomes_train_memory.pkl', 'rb') as f:\n",
    "    outcomes_train_df = pickle.load(f)\n",
    "with open('data/outcomes_valid_memory.pkl', 'rb') as f:\n",
    "    outcomes_valid_df = pickle.load(f)'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('models/simple_models.pkl', 'rb') as file:\n",
    "    loaded_models = pickle.load(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'batch_size': 1000, 'distribution': 'LogNormal', 'iters': 200, 'k': 3, 'layers': [1000, 1000], 'learning_rate': 1e-05}\n"
     ]
    }
   ],
   "source": [
    "print(loaded_models[1][3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_peaks_in_data(data):\n",
    "    peaks = []\n",
    "    for i in range(1, len(data) - 1):\n",
    "        if data[i-1] < data[i] > data[i+1]:\n",
    "            peaks.append(i)\n",
    "    return peaks\n",
    "\n",
    "def fnr_fpr_calculator(model, normed_x, normed_t, normed_e, prediction_times, threshold=0.7):\n",
    "    out_survival = model.predict_survival(normed_x, prediction_times)\n",
    "    fnrs = []\n",
    "    fprs = []\n",
    "    shot_indices = find_peaks_in_data(normed_t)\n",
    "    for i, time in enumerate(prediction_times):\n",
    "        tm_prediction_per_shot = []\n",
    "        # 1 means correct TM prediction (TP), 0 means unpredicted TM (FN), -1 means no TM in shot and no TM predicted (TN), -2 means no TM in shot and yes TM predicted (FP)\n",
    "        # a TM is predicted when the survival prediction is 0 at any point in the shot. Check if better results when TM is consecutive 0s\n",
    "        survival_prediction = out_survival[:,i]\n",
    "        survival_prediction = (survival_prediction > threshold).astype(int)\n",
    "        for j, shot_index in enumerate(shot_indices[:-1]):\n",
    "            tm = (0 in survival_prediction[shot_indices[j]:shot_indices[j+1]])\n",
    "            if normed_e[shot_index] == 1 and tm:\n",
    "                tm_prediction_per_shot.append(1)\n",
    "            elif normed_e[shot_index] == 1 and not tm:\n",
    "                tm_prediction_per_shot.append(0)\n",
    "            elif normed_e[shot_index] == 0 and not tm:\n",
    "                tm_prediction_per_shot.append(-1)\n",
    "            elif normed_e[shot_index] == 0 and tm:\n",
    "                tm_prediction_per_shot.append(-2)\n",
    "        fnr = tm_prediction_per_shot.count(0) / (tm_prediction_per_shot.count(1) + tm_prediction_per_shot.count(0))\n",
    "        fpr = tm_prediction_per_shot.count(-2) / (tm_prediction_per_shot.count(-1) + tm_prediction_per_shot.count(-2))\n",
    "        #print(f'fnr: {fnr} from {tm_prediction_per_shot.count(1) + tm_prediction_per_shot.count(0)} shots, fpr: {fpr} from {tm_prediction_per_shot.count(-1) + tm_prediction_per_shot.count(-2)} shots')\n",
    "        fnrs.append(fnr)\n",
    "        fprs.append(fpr)\n",
    "    auc_fnr = np.trapz(fnrs, prediction_times)\n",
    "    auc_fpr = np.trapz(fprs, prediction_times)\n",
    "    return auc_fpr, auc_fnr, fprs, fnrs, prediction_times"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('models/simple_models.pkl', 'rb') as file:\n",
    "    loaded_models = pickle.load(file)\n",
    "for model1 in loaded_models[0:2]:\n",
    "    print(loaded_models[0][3])\n",
    "    auc_sums = []\n",
    "    for threshold in [0.7, 0.75, 0.8, 0.85, 0.9]:\n",
    "        model = model1[0]\n",
    "        t_valid = np.array(outcomes_valid_df['time'])\n",
    "        e_valid = np.array(outcomes_valid_df['event'])\n",
    "        auc_fpr, auc_fnr, fprs, fnrs, prediction_times = fnr_fpr_calculator(model, x_valid_df, t_valid, e_valid, [200], threshold=threshold)\n",
    "        auc_sums.append(fprs[0]+fnrs[0])\n",
    "    print(np.min(auc_sums))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n",
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x14ab3cb6cc50>]"
      ]
     },
     "execution_count": 42,
     "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": [
    "thresholds = np.linspace(0.5, 1, 20)\n",
    "fprs_list = []\n",
    "fnrs_list = []\n",
    "for threshold in thresholds:\n",
    "    auc_fpr, auc_fnr, fprs, fnrs, prediction_times = fnr_fpr_calculator(loaded_models[1][0], x_valid_df, t_valid, e_valid, [200], threshold=threshold)\n",
    "    fprs_list.append(fprs[0])\n",
    "    fnrs_list.append(fnrs[0])\n",
    "plt.plot(fprs_list, fnrs_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/hf8585/TMPredictor/survival_tm/auton-survival/auton_survival/estimators.py:206: FutureWarning: DataFrame.interpolate with method=bfill is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.\n",
      "  return survival_predictions.sort_index(axis=0).interpolate().interpolate(method='bfill').T[times].values\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x14ab1de7f410>]"
      ]
     },
     "execution_count": 22,
     "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": [
    "out_survival = loaded_models[1][0].predict_survival(x_valid_df, [200])\n",
    "t_valid = np.array(outcomes_valid_df['time'])\n",
    "e_valid = np.array(outcomes_valid_df['event'])\n",
    "start = 100000\n",
    "plt.plot(out_survival[start:start+2000, 0])\n",
    "plt.plot(e_valid[start:start+2000])\n",
    "plt.plot(1-t_valid[start:start+2000]/t_valid[start:start+2000].max())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction_times = [20, 50, 100, 200]\n",
    "out_survival = model.predict_survival(x_valid_df, [200])\n",
    "peaks = find_peaks_in_data(np.array(outcomes_valid_df['time']))\n",
    "print(peaks[789])\n",
    "for i in range(0, 20):\n",
    "    \n",
    "    peak_number = 786 + i\n",
    "    start_index = peaks[peak_number]\n",
    "    end_index = peaks[peak_number+1]\n",
    "    times = np.arange(0, (end_index - start_index)*20, 20)\n",
    "    #plt.plot(times, out_survival[start_index:end_index,2], label='Survival in 100ms')\n",
    "    plt.plot(times, out_survival[start_index:end_index,0], label='Survival in 200ms')\n",
    "\n",
    "    if (np.array(outcomes_valid_df['event'])[start_index]==1):\n",
    "        plt.title('YES TM')\n",
    "    else:\n",
    "        plt.title('NO TM')\n",
    "    plt.xlabel('Time / ms')\n",
    "    plt.ylabel('Survival probability')\n",
    "    plt.ylim(0.5, 1)\n",
    "    plt.legend()\n",
    "    plt.show()"
   ]
  }
 ],
 "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
}