{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tutorial (custom model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can define custom models to be tested in the state learner library. In this notebook, this is demonstrated with a DeepSurv neural network model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from pytmle import PyTMLE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_times = [5.0, 10.0, 15.0, 20.0, 25.0]\n",
    "np.random.seed(42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the Hodgkin's Disease dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"hodgkins_disease.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define a custom model (exemplified by a model based on DeepSurv) to be cross-fitted in the .fit method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# imports needed for pycox\n",
    "import torchtuples as tt\n",
    "from pycox.models import CoxPH\n",
    "\n",
    "in_features = 7 # age, female, extranod, stage2, medwidsi_S, medwidsi_N, chemo\n",
    "num_nodes = [32, 32]\n",
    "out_features = 1\n",
    "batch_norm = True\n",
    "dropout = 0.1\n",
    "output_bias = False\n",
    "\n",
    "net = tt.practical.MLPVanilla(in_features, num_nodes, out_features, batch_norm,\n",
    "                              dropout, output_bias=output_bias)\n",
    "\n",
    "\n",
    "model = CoxPH(net, tt.optim.Adam)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Instantiate the PyTMLE class and fit it with both the DeepSurv model and a RandomSurvivalForest in the library of the state learner, and a support vector machine classifier + random forest in the library of the stacking classifier for propensity scores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Estimating propensity scores...\n",
      "Estimating hazards and event-free survival...\n",
      "Estimating censoring survival...\n",
      "(CoxPH | CoxPH): 3.039067494120523\n",
      "(CoxPH | RandomSurvivalForest): 3.0488997726124407\n",
      "(RandomSurvivalForest | CoxPH): 3.2878747147243326\n",
      "(RandomSurvivalForest | RandomSurvivalForest): 3.267161259846405\n",
      "Starting TMLE update loop...\n",
      "Step 1: Norm PnEIC improved to 0.02552972784702802.\n",
      "Step 2: Norm PnEIC improved to 0.01879889731053828.\n",
      "Step 3: Norm PnEIC improved to 0.013366134071044776.\n",
      "Step 4: Norm PnEIC improved to 0.012506568462887686.\n",
      "Step 5: Norm PnEIC improved to 0.011800513341399812.\n",
      "Step 6: Norm PnEIC improved to 0.011189614114653643.\n",
      "Step 7: Norm PnEIC improved to 0.01065846380412103.\n",
      "Step 8: Norm PnEIC improved to 0.010202372383367327.\n",
      "Step 9: Norm PnEIC improved to 0.009824646524712764.\n",
      "Step 10: Norm PnEIC improved to 0.00954416156014202.\n",
      "Step 11: Norm PnEIC improved to 0.009406644125445627.\n",
      "Step 12: Norm PnEIC improved to 0.008857826295995429.\n",
      "Step 13: Norm PnEIC improved to 0.008695787777146047.\n",
      "Step 14: Norm PnEIC improved to 0.008553978805939081.\n",
      "Step 15: Norm PnEIC improved to 0.008419328831486756.\n",
      "Step 16: Norm PnEIC improved to 0.0082904473347252.\n",
      "Step 17: Norm PnEIC improved to 0.008166783516545822.\n",
      "Step 18: Norm PnEIC improved to 0.008047894515103305.\n",
      "Step 19: Norm PnEIC improved to 0.007933384318140045.\n",
      "Step 20: Norm PnEIC improved to 0.007822895653610199.\n",
      "Step 21: Norm PnEIC improved to 0.007716106194492411.\n",
      "Step 22: Norm PnEIC improved to 0.007612725315303317.\n",
      "Step 23: Norm PnEIC improved to 0.007512490898060883.\n",
      "Step 24: Norm PnEIC improved to 0.0074151663180456314.\n",
      "Step 25: Norm PnEIC improved to 0.007320537615986524.\n",
      "Step 26: Norm PnEIC improved to 0.007228410915022249.\n",
      "Step 27: Norm PnEIC improved to 0.007138610087925415.\n",
      "Step 28: Norm PnEIC improved to 0.00705097468448443.\n",
      "Step 29: Norm PnEIC improved to 0.00696535810888335.\n",
      "Step 30: Norm PnEIC improved to 0.006881626035099635.\n",
      "Step 31: Norm PnEIC improved to 0.006799655040726316.\n",
      "Step 32: Norm PnEIC improved to 0.006719331438764763.\n",
      "Step 33: Norm PnEIC improved to 0.006640550284849012.\n",
      "Step 34: Norm PnEIC improved to 0.00656321453807256.\n",
      "Step 35: Norm PnEIC improved to 0.00648723435407574.\n",
      "Step 36: Norm PnEIC improved to 0.006412526490721732.\n",
      "Step 37: Norm PnEIC improved to 0.006339013808254429.\n",
      "Step 38: Norm PnEIC improved to 0.006266624847842752.\n",
      "Step 39: Norm PnEIC improved to 0.006195293474267316.\n",
      "Step 40: Norm PnEIC improved to 0.006124958570450705.\n",
      "Step 41: Norm PnEIC improved to 0.006055563773274096.\n",
      "Step 42: Norm PnEIC improved to 0.005987057241792404.\n",
      "Step 43: Norm PnEIC improved to 0.00591939145041587.\n",
      "Step 44: Norm PnEIC improved to 0.005852523000953846.\n",
      "Step 45: Norm PnEIC improved to 0.005786412448541076.\n",
      "Step 46: Norm PnEIC improved to 0.00572102413745785.\n",
      "Step 47: Norm PnEIC improved to 0.005656326043677415.\n",
      "Step 48: Norm PnEIC improved to 0.005592289621675556.\n",
      "Step 49: Norm PnEIC improved to 0.005528889653607046.\n",
      "Step 50: Norm PnEIC improved to 0.005466104099432896.\n",
      "Step 51: Norm PnEIC improved to 0.005403913946957157.\n",
      "Step 52: Norm PnEIC improved to 0.005342303061049861.\n",
      "Step 53: Norm PnEIC improved to 0.005281258031571782.\n",
      "Step 54: Norm PnEIC improved to 0.0052207680197234735.\n",
      "TMLE converged at step 54.\n"
     ]
    }
   ],
   "source": [
    "from sksurv.ensemble import RandomSurvivalForest\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "tmle = PyTMLE(df, \n",
    "              col_event_times=\"time\", \n",
    "              col_event_indicator=\"status\", \n",
    "              col_group=\"chemo\", \n",
    "              target_times=target_times, \n",
    "              g_comp=True,\n",
    "              verbose=3,)\n",
    "\n",
    "tmle.fit(cv_folds=5, \n",
    "         max_updates=100, \n",
    "         save_models=True,\n",
    "         models=[model, RandomSurvivalForest()],\n",
    "         propensity_score_models=[SVC(), RandomForestClassifier()],\n",
    "         labtrans=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<Figure size 1400x700 with 2 Axes>,\n",
       " array([<Axes: title={'center': 'Event 1'}, xlabel='Time', ylabel='Predicted Risk'>,\n",
       "        <Axes: title={'center': 'Event 2'}, xlabel='Time', ylabel='Predicted Risk'>],\n",
       "       dtype=object))"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1400x700 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmle.plot(g_comp=True, use_bootstrap=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since `save_models` was set to `True` in the `fit()` call, you can extract the models used for initial estimates and analyze them further. Note that all cross-fitted models are of type `PycoxWrapper` and only return the name of the wrapped class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'propensity_model': StackingClassifier(cv=5,\n",
      "                   estimators=[('0', SVC()), ('1', RandomForestClassifier())],\n",
      "                   final_estimator=LogisticRegression(max_iter=1000)), 'risks_model_fold_0': CauseSpecificCoxPH, 'censoring_model_fold_0': CoxPH, 'risks_model_fold_1': CauseSpecificCoxPH, 'censoring_model_fold_1': CoxPH, 'risks_model_fold_2': CauseSpecificCoxPH, 'censoring_model_fold_2': CoxPH, 'risks_model_fold_3': CauseSpecificCoxPH, 'censoring_model_fold_3': CoxPH, 'risks_model_fold_4': CauseSpecificCoxPH, 'censoring_model_fold_4': CoxPH}\n",
      "            risks_model       censoring_model      loss\n",
      "0                 CoxPH                 CoxPH  3.039067\n",
      "1                 CoxPH  RandomSurvivalForest  3.048900\n",
      "3  RandomSurvivalForest  RandomSurvivalForest  3.267161\n",
      "2  RandomSurvivalForest                 CoxPH  3.287875\n"
     ]
    }
   ],
   "source": [
    "print(tmle.models)\n",
    "print(tmle.state_learner_cv_fit)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can see that the DeepSurv model was chosen for both risks and censoring."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "PyTMLE",
   "language": "python",
   "name": "pytmle"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.19"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}