Data_Preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "G-bec3cCXsVx"
   },
   "source": [
    "# Data Exploration and Preprocessing\n",
    "TCGA Reannotated Ovarian Cancer Clinical Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "id": "mTx2g1keOEjc"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy import stats\n",
    "from openpyxl import load_workbook\n",
    "import sidetable\n",
    "from functools import reduce\n",
    "import itertools\n",
    "import math\n",
    "import random\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "import torch.nn.functional as F\n",
    "from collections import namedtuple\n",
    "from sklearn.linear_model import LinearRegression, LogisticRegression\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "import time\n",
    "import statsmodels.api as sm\n",
    "from scipy import stats\n",
    "from lifelines import CoxPHFitter\n",
    "\n",
    "device = torch.device('cpu')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "68TNefuaXmnW"
   },
   "source": [
    "Import Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "id": "kEYLFUOhRVru"
   },
   "outputs": [],
   "source": [
    "# Villalobos 2018 reannotated TCGA data (https://ascopubs.org/doi/suppl/10.1200/CCI.17.00096)\n",
    "tcga_ov_1 = pd.read_excel('./Data/Villalobos_TCGA/ds_CCI.17.00096-1.xlsx',\n",
    "                          sheet_name='Master clinical dataset'\n",
    "                          )\n",
    "\n",
    "tcga_ov_2 = pd.read_excel('./Data/Villalobos_TCGA/ds_CCI.17.00096-2.xlsx',\n",
    "                          sheet_name='Months'\n",
    "                          )\n",
    "\n",
    "tcga_ov_3 = pd.read_excel('./Data/Villalobos_TCGA/ds_CCI.17.00096-3.xlsx',\n",
    "                          sheet_name='clinical_drug_all_OV.txt'\n",
    "                          )\n",
    "\n",
    "# TCGA Drug standardization (https://gdisc.bme.gatech.edu/cgi-bin/gdisc/tap5.cgi#)\n",
    "drugs = pd.read_csv('./Data/DrugCorrection1.csv')\n",
    "drugs['Correction'] = drugs['Correction'].str.strip()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "M2uBW25GhHv_"
   },
   "source": [
    "## Clean and Reorganize"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TgbCL7brzOid"
   },
   "source": [
    "TCGA 3: Clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "C8ZXyxh_Yo7J",
    "outputId": "2760169e-47f0-4789-c306-c3402cde0ba9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2463, 19)\n",
      "bcr_patient_barcode                 0\n",
      "bcr_drug_barcode                    0\n",
      "days_to_drug_therapy_end          339\n",
      "days_to_drug_therapy_start        145\n",
      "days_to_drug_treatment_end       2463\n",
      "days_to_drug_treatment_start     2463\n",
      "dosage_units                     2463\n",
      "drug_category                       0\n",
      "drug_dosage                      2463\n",
      "drug_name                          11\n",
      "initial_course                   2463\n",
      "number_cycles                     391\n",
      "regimen_indication                  2\n",
      "regimen_indication_notes         2328\n",
      "route_of_administration           243\n",
      "route_of_administration_notes    2126\n",
      "therapy_ongoing                   167\n",
      "total_dose                        746\n",
      "total_dose_units                  715\n",
      "dtype: int64 \n",
      "\n",
      "Replacement Rules:\n",
      "            drug_name_old            drug_name_new\n",
      "0         5F4 Leucovorin  Fluorouracil+Leucovorin\n",
      "1            90Y-HU3S193                  Hu3S193\n",
      "2                AMG 706                Motesanib\n",
      "3  Abagovomab or Placebo               Abagovomab\n",
      "4  Abagovomag vs Placebo               Abagovomab \n",
      "\n",
      "Unique Drugs: 56 \n",
      " ['Abagovomab', 'Aldesleukin', 'Altretamine', 'Amifostine', 'Aminocamptothecin', 'Anastrozole', 'Bevacizumab', 'CBP501', 'CEP -11981', 'Capecitabine', 'Carboplatin', 'Catumaxomab', 'Cediranib', 'Cetuximab', 'Chlorambucil', 'Cisplatin', 'Cyclophosphamide', 'Dexamethasone', 'Docetaxel', 'Dortezomib', 'Doxorubicin', 'Etoposide', 'Fluorouracil+Leucovorin', 'Gefitinib', 'Gemcitabine', 'Hu3S193', 'Ifosfamide', 'Ilex', 'Irinotecan', 'Irofulven', 'Letrozole', 'Leuprolide', 'Lily', 'Megestrol', 'Mitoxantrone', 'Motesanib', 'NOS', 'Oregovomab', 'Oxaliplatin', 'Paclitaxel', 'Paclitaxel+Carboplatin', 'Patupilone', 'Pemetrexed', 'SCH 58500', 'Sargramostim', 'Sorafenib', 'Tamoxifen', 'Topotecan', 'Topotecan+Carboplatin', 'Trabectedin', 'Trastuzumab', 'Vamydex', 'Vincristine', 'Vinorelbine', 'Vosaroxin', 'rBBX-01'] \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Drop columns with all missing values\n",
    "print(tcga_ov_3.shape)\n",
    "print(tcga_ov_3.isnull().sum(), '\\n')\n",
    "tcga_ov_3_drop = tcga_ov_3.dropna(axis=1, how='all')\n",
    "# Drop rows where drug is missing\n",
    "tcga_ov_3_drop = tcga_ov_3_drop[tcga_ov_3_drop['drug_name'].notna()]\n",
    "# Drop rows where start AND end are missing (right now it's actually OR, see below)\n",
    "tcga_ov_3_drop = tcga_ov_3_drop.dropna(how='any', subset=['days_to_drug_therapy_end', 'days_to_drug_therapy_start']) #change to how='all' when figure out what to do with start OR end missing\n",
    "\n",
    "# Standardize drug names\n",
    "    # Merge using values from TCGA drug standardization (https://gdisc.bme.gatech.edu/cgi-bin/gdisc/tap5.cgi)\n",
    "tcga_ov_3_clean = tcga_ov_3_drop.merge(drugs, how='left', left_on='drug_name', right_on='OldName').drop(['OldName'], axis=1)\n",
    "    # Additional replacement rules\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'Doxoribicin', 'Correction'] = 'Doxorubicin'\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'gemcitabin', 'Correction'] = 'Gemcitabine'\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'Hexlalen', 'Correction'] = 'Altretamine'\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'Cisplatin/Gemzar', 'Correction'] = 'Cisplatin' # This applies to only one line, which has another sample just for Gemzar (Gemcitabine)\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'Ilex', 'Correction'] = 'Ilex'\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'ILIZ', 'Correction'] = 'ILIZ'\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['drug_name'] == 'Lily', 'Correction'] = 'Lily'\n",
    "\n",
    "    # Print replacement rules and replace 'drug_name'\n",
    "drug_name_old = tcga_ov_3_clean['drug_name']\n",
    "drug_name_new = tcga_ov_3_clean['Correction']\n",
    "rules = pd.DataFrame({'drug_name_old': drug_name_old, 'drug_name_new': drug_name_new}).drop_duplicates().sort_values(by=['drug_name_old']).reset_index().drop('index', axis=1)\n",
    "\n",
    "pd.set_option('display.max_rows', None)\n",
    "print('Replacement Rules:\\n', rules.head(), '\\n')\n",
    "pd.reset_option('max_rows')\n",
    "\n",
    "tcga_ov_3_clean['drug_name'] = tcga_ov_3_clean['Correction']\n",
    "tcga_ov_3_clean = tcga_ov_3_clean.drop('Correction', axis=1)\n",
    "\n",
    "    # List of drugs in dataset\n",
    "drug_list = [x for x in list(tcga_ov_3_clean['drug_name'].drop_duplicates()) if str(x) != 'nan']\n",
    "drug_list.sort()\n",
    "print('Unique Drugs:', len(drug_list), '\\n', drug_list, '\\n')\n",
    "\n",
    "# Drop where therapy start = therapy end\n",
    "tcga_ov_3_clean = tcga_ov_3_clean[tcga_ov_3_clean['days_to_drug_therapy_end'] != tcga_ov_3_clean['days_to_drug_therapy_start']] "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "RlpTIsomL8nT"
   },
   "source": [
    "TCGA 3: Fix/standardize time variables and fix order of therapy lines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "id": "ycWmppEk6d5o"
   },
   "outputs": [],
   "source": [
    "# Fix values where start and end are switched\n",
    "tcga_ov_3_clean.loc[tcga_ov_3_clean['days_to_drug_therapy_start'] > tcga_ov_3_clean['days_to_drug_therapy_end'], ['days_to_drug_therapy_start', 'days_to_drug_therapy_end']] = tcga_ov_3_clean.loc[tcga_ov_3_clean['days_to_drug_therapy_start'] > tcga_ov_3_clean['days_to_drug_therapy_end'], ['days_to_drug_therapy_end', 'days_to_drug_therapy_start']].values\n",
    "\n",
    "# Set earliest drug therapy start to zero and subtract everything else by min days\n",
    "ther_start = tcga_ov_3_clean.groupby('bcr_patient_barcode')['days_to_drug_therapy_start']\n",
    "tcga_timefix = tcga_ov_3_clean.assign(start_day=ther_start.transform(min))\n",
    "      # keep start days for later use with tcga_ov_1\n",
    "tcga_start_days = tcga_timefix[['bcr_patient_barcode', 'start_day']].drop_duplicates()\n",
    "\n",
    "tcga_timefix['therapy_start'] = tcga_timefix['days_to_drug_therapy_start'] - tcga_timefix['start_day']\n",
    "tcga_timefix['therapy_end'] = tcga_timefix['days_to_drug_therapy_end'] - tcga_timefix['start_day']\n",
    "tcga_timefix = tcga_timefix.drop(['days_to_drug_therapy_end', 'days_to_drug_therapy_start', 'start_day'], axis=1)\n",
    "tcga_timefix = tcga_timefix.sort_values(by=['bcr_patient_barcode', 'therapy_start', 'therapy_end'])\n",
    "\n",
    "# Set up state list for each patient: barcode, timing, drug combo\n",
    "tcga_drug_lines = []\n",
    "for barcode in tcga_timefix['bcr_patient_barcode'].unique():\n",
    "    tcga_time = tcga_timefix[tcga_timefix['bcr_patient_barcode'] == barcode]\n",
    "    tcga_time = tcga_time[['therapy_start', 'therapy_end', 'drug_name']].drop_duplicates(keep='first').values.tolist() # Drop duplicate drugs that have different dosages or administration but same timing\n",
    "\n",
    "    points = [] # list of (offset, plus/minus, drug) tuples\n",
    "    for start,stop,drug in tcga_time:\n",
    "        points.append((start,'+',drug))\n",
    "        points.append((stop,'-',drug))\n",
    "    points.sort()\n",
    "\n",
    "    ranges = [] # output list of (start, stop, drug_set) tuples\n",
    "    current_set = []\n",
    "    last_start = None\n",
    "    for offset,pm,drug in points:\n",
    "        if pm == '+':\n",
    "            if last_start is not None:\n",
    "                ranges.append([last_start,offset,list(set(current_set.copy()))])\n",
    "            current_set.append(drug)\n",
    "            last_start = offset\n",
    "        elif pm == '-':\n",
    "            ranges.append([last_start,offset,list(set(current_set.copy()))])\n",
    "            current_set.remove(drug)\n",
    "            last_start = offset\n",
    "\n",
    "    # Finish off\n",
    "    if last_start is not None:\n",
    "        ranges.append([last_start,offset,list(set(current_set.copy()))])\n",
    "\n",
    "    # Remove the ranges where start = stop\n",
    "    range_drug = []\n",
    "    for i in range(len(ranges)):\n",
    "        if ranges[i][0] != ranges[i][1]: # add condition:  <& (ranges[i][2] != [])> to drop no-drug periods\n",
    "            range_drug.append(ranges[i])\n",
    "\n",
    "    # Sort drugs in each drug combo\n",
    "    for i in range(len(range_drug)):\n",
    "        range_drug[i][2].sort()\n",
    "\n",
    "    # Remove overlapping/back-to-back duplicate lines. Drop this section if decide to do something with dosages\n",
    "    ranges_final = []\n",
    "    for line in range(0, len(range_drug)-1):\n",
    "        if (range_drug[line+1][2] == range_drug[line][2]) & (range_drug[line+1][0] <= range_drug[line][1]):\n",
    "            range_drug[line][1] = range_drug[line+1][1]\n",
    "            range_drug[line+1][0] = range_drug[line][0]\n",
    "        if (range_drug[line][2] != range_drug[line+1][2]) | (range_drug[line][0] != range_drug[line+1][0]):\n",
    "            ranges_final.append(range_drug[line])\n",
    "    ranges_final.append(range_drug[len(range_drug)-1])\n",
    "\n",
    "    # Add the number of previous lines of therapy\n",
    "    for line in range(len(ranges_final)):\n",
    "        if line == 0:\n",
    "            ranges_final[line].append(0) \n",
    "        elif ranges_final[line-1][2] == []:\n",
    "            ranges_final[line].append(ranges_final[line-1][3])\n",
    "        else:\n",
    "            ranges_final[line].append(ranges_final[line-1][3] + 1)\n",
    "    \n",
    "    # Add treat transition var\n",
    "    for line in range(len(ranges_final)):\n",
    "        try:\n",
    "            if (ranges_final[line][2] != []) & (ranges_final[line+1][2] != []):\n",
    "                ranges_final[line].append(math.ceil((ranges_final[line][1] - ranges_final[line][0]) / 30))\n",
    "            elif (ranges_final[line][2] != []) & (ranges_final[line+1][2] == []):\n",
    "                ranges_final[line].append(math.ceil((ranges_final[line][1] - ranges_final[line][0]) / 30) - 1)\n",
    "            elif (ranges_final[line][2] == []) & (ranges_final[line+1][2] != []):\n",
    "                ranges_final[line].append(1)\n",
    "            elif (ranges_final[line][2] == []) & (ranges_final[line+1][2] == []):\n",
    "                ranges_final[line].append(0)\n",
    "        except IndexError:\n",
    "            ranges_final[line].append(0)\n",
    "    \n",
    "    # Add no treat transition var\n",
    "    for line in range(len(ranges_final)):\n",
    "        try:\n",
    "            if (ranges_final[line][2] != []) & (ranges_final[line+1][2] != []):\n",
    "                ranges_final[line].append(0)                \n",
    "            elif (ranges_final[line][2] != []) & (ranges_final[line+1][2] == []):\n",
    "                ranges_final[line].append(1)\n",
    "            elif (ranges_final[line][2] == []) & (ranges_final[line+1][2] != []):\n",
    "                ranges_final[line].append(math.ceil((ranges_final[line][1] - ranges_final[line][0]) / 30) - 1)\n",
    "            elif (ranges_final[line][2] == []) & (ranges_final[line+1][2] == []):\n",
    "                ranges_final[line].append(math.ceil((ranges_final[line][1] - ranges_final[line][0]) / 30))\n",
    "        except IndexError:\n",
    "            ranges_final[line].append(math.ceil((ranges_final[line][1] - ranges_final[line][0]) / 30)-1)\n",
    "\n",
    "    # Add patient barcodes\n",
    "    for line in range(len(ranges_final)):\n",
    "        ranges_final[line].insert(0, barcode)\n",
    "    \n",
    "\n",
    "    tcga_drug_lines.extend(ranges_final)\n",
    "\n",
    "tcga_drug_lines[3] = ['TCGA-04-1332', 0.0, 151.0, ['Carboplatin', 'Paclitaxel', 'Topotecan'], 0, 5, 1] # Special case to fix. Make sure to check this if make changes above\n",
    "\n",
    "# Back to df\n",
    "lines_df = pd.DataFrame(tcga_drug_lines, columns=['bcr_patient_barcode', 'start', 'end', 'therapy', 'previous_lines', 'treat', 'no_treat'])\n",
    "\n",
    "# List of patient barcodes\n",
    "tcga_barcodes = list(lines_df['bcr_patient_barcode'].unique())\n",
    "\n",
    "# Notes\n",
    "    # One thing to be aware of is that this code drops all values where therapy start and therapy end are equal\n",
    "    # Fix: drops all back to back duplicate treatment, should keep first (see TCGA-04-1365)\n",
    "        #####################################I think this was fixed, but need to double check"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IX1joLnWxyzL"
   },
   "source": [
    "TCGA 1: Clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 603
    },
    "id": "eH71y_18xy4L",
    "outputId": "4f8952d2-ddea-497a-9769-0881141d9053"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NaNs:  bcr_patient_barcode                                 4\n",
      "total_days_overall_survival                        28\n",
      "outcome_overall_survival_censoring                 16\n",
      "vital_status                                       34\n",
      "days_to_tumor_progression                         570\n",
      "                                                 ... \n",
      "5th_chemo_regimen_days_outcome                    524\n",
      "6th_chemo_regimen_days_outcome                    552\n",
      "Days off platinum prior to recurrence 1st line     99\n",
      "Last day of platinum 1st line                     108\n",
      "Chemotherapy number of lines of therapy            27\n",
      "Length: 72, dtype: int64\n"
     ]
    },
    {
     "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>bcr_patient_barcode</th>\n",
       "      <th>total_days_overall_survival</th>\n",
       "      <th>outcome_overall_survival_censoring</th>\n",
       "      <th>age_at_initial_pathologic_diagnosis</th>\n",
       "      <th>race</th>\n",
       "      <th>tumor_grade</th>\n",
       "      <th>tumor_stage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>1300.0</td>\n",
       "      <td>1</td>\n",
       "      <td>79.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TCGA-04-1332</td>\n",
       "      <td>1217.0</td>\n",
       "      <td>1</td>\n",
       "      <td>70.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TCGA-04-1336</td>\n",
       "      <td>1445.0</td>\n",
       "      <td>0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TCGA-04-1338</td>\n",
       "      <td>1418.0</td>\n",
       "      <td>0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TCGA-04-1342</td>\n",
       "      <td>531.0</td>\n",
       "      <td>1</td>\n",
       "      <td>80.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G2</td>\n",
       "      <td>IV</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>455</th>\n",
       "      <td>TCGA-61-2113</td>\n",
       "      <td>627.0</td>\n",
       "      <td>1</td>\n",
       "      <td>54.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>456</th>\n",
       "      <td>TCGA-61-2610</td>\n",
       "      <td>1550.0</td>\n",
       "      <td>1</td>\n",
       "      <td>61.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>457</th>\n",
       "      <td>TCGA-61-2611</td>\n",
       "      <td>417.0</td>\n",
       "      <td>1</td>\n",
       "      <td>40.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>458</th>\n",
       "      <td>TCGA-61-2612</td>\n",
       "      <td>173.0</td>\n",
       "      <td>1</td>\n",
       "      <td>63.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>459</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>561.0</td>\n",
       "      <td>1</td>\n",
       "      <td>73.0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>G3</td>\n",
       "      <td>IIIC</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>460 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    bcr_patient_barcode  total_days_overall_survival  \\\n",
       "0          TCGA-04-1331                       1300.0   \n",
       "1          TCGA-04-1332                       1217.0   \n",
       "2          TCGA-04-1336                       1445.0   \n",
       "3          TCGA-04-1338                       1418.0   \n",
       "4          TCGA-04-1342                        531.0   \n",
       "..                  ...                          ...   \n",
       "455        TCGA-61-2113                        627.0   \n",
       "456        TCGA-61-2610                       1550.0   \n",
       "457        TCGA-61-2611                        417.0   \n",
       "458        TCGA-61-2612                        173.0   \n",
       "459        TCGA-61-2613                        561.0   \n",
       "\n",
       "     outcome_overall_survival_censoring  age_at_initial_pathologic_diagnosis  \\\n",
       "0                                     1                                 79.0   \n",
       "1                                     1                                 70.0   \n",
       "2                                     0                                 55.0   \n",
       "3                                     0                                 78.0   \n",
       "4                                     1                                 80.0   \n",
       "..                                  ...                                  ...   \n",
       "455                                   1                                 54.0   \n",
       "456                                   1                                 61.0   \n",
       "457                                   1                                 40.0   \n",
       "458                                   1                                 63.0   \n",
       "459                                   1                                 73.0   \n",
       "\n",
       "      race tumor_grade tumor_stage  \n",
       "0    WHITE          G3        IIIC  \n",
       "1    WHITE          G3        IIIC  \n",
       "2    WHITE          G3        IIIB  \n",
       "3    WHITE          G3        IIIC  \n",
       "4    WHITE          G2          IV  \n",
       "..     ...         ...         ...  \n",
       "455  WHITE          G3         IIC  \n",
       "456  WHITE          G3        IIIC  \n",
       "457  WHITE          G3        IIIC  \n",
       "458  WHITE          G3        IIIC  \n",
       "459  WHITE          G3        IIIC  \n",
       "\n",
       "[460 rows x 7 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Keep subset of variables \n",
    "print('NaNs: ', tcga_ov_1.isnull().sum())\n",
    "tcga_ov_1_keep = tcga_ov_1[['bcr_patient_barcode',\n",
    "                            'total_days_overall_survival',\n",
    "                            'outcome_overall_survival_censoring',\n",
    "                            # 'vital_status',\n",
    "                            # 'days_to_tumor_progression',\n",
    "                            # 'days_to_death',\n",
    "                            # 'days_to_last_followup',\n",
    "                            # 'days_to_tumor_recurrence',\n",
    "                            # 'time_to_failure',  \n",
    "                            # 'Cycles_of_adjuvant_therapy',\n",
    "                            # 'Adjuvant_chemotherapy_dose_intensity',\n",
    "                            'age_at_initial_pathologic_diagnosis',\n",
    "                            # 'anatomic_organ_subdivision',\n",
    "                            # 'days_to_birth',\n",
    "                            # 'initial_pathologic_diagnosis_method',\n",
    "                            # 'person_neoplasm_cancer_status',\n",
    "                            # 'pretreatment_history',\n",
    "                            # 'primary_therapy_outcome_success', # ask EKO about this\n",
    "                            'race',\n",
    "                            # 'residual_tumor',\n",
    "                            # 'site_of_tumor_first_recurrence',\n",
    "                            # 'tissue_source_site',\n",
    "                            'tumor_grade',\n",
    "                            # 'tumor_residual_disease',\n",
    "                            'tumor_stage',\n",
    "                            # 'tumor_tissue_site'\n",
    "                            # 'year_of_initial_pathologic_diagnosis',\n",
    "                            # 'Days off platinum prior to recurrence 1st line',\n",
    "                            # 'Last day of platinum 1st line',\n",
    "                            # 'Chemotherapy number of lines of therapy'\n",
    "                            ]]\n",
    "\n",
    "# Drop cases that don't have a survival metric\n",
    "tcga_ov_1_keep = tcga_ov_1_keep.dropna(subset=['total_days_overall_survival'])\n",
    "tcga_ov_1_keep.drop(tcga_ov_1_keep[tcga_ov_1_keep['total_days_overall_survival'] == 'cannot assess'].index, inplace=True)\n",
    "\n",
    "# Only keep samples that are in the cleaned 'lines' data\n",
    "tcga_ov_1_keep = tcga_ov_1_keep[tcga_ov_1_keep['bcr_patient_barcode'].isin(tcga_barcodes)].sort_values(by=['bcr_patient_barcode']).reset_index().drop('index', axis=1)\n",
    "\n",
    "# Adjust final survival by start of therapy day\n",
    "tcga_start_days = tcga_start_days[tcga_start_days['bcr_patient_barcode'].isin(tcga_barcodes)].sort_values(by=['bcr_patient_barcode']).reset_index().drop('index', axis=1)\n",
    "tcga_ov_1_keep['total_days_overall_survival'] = tcga_ov_1_keep['total_days_overall_survival'] - tcga_start_days['start_day'] # this needs work, getting some negative values\n",
    "\n",
    "# Fill blanks\n",
    "tcga_ov_1_keep = tcga_ov_1_keep.replace(r'^\\s*$', np.nan, regex=True)\n",
    "tcga_ov_1_keep = tcga_ov_1_keep.replace(np.nan, 'NOT SPECIFIED', regex=True)\n",
    "\n",
    "tcga_ov_1_keep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "tcga_ov_1_keep.to_excel('tcga_1.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data summary table(s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "book = load_workbook('Table_1.xlsx')\n",
    "book.template = False\n",
    "\n",
    "writer = pd.ExcelWriter('Table_1.xlsx', engine='openpyxl')\n",
    "writer.book = book\n",
    "\n",
    "# Race\n",
    "try:\n",
    "    del book['race']\n",
    "except:\n",
    "    pass\n",
    "tcga_ov_1_keep.stb.freq(['race']).to_excel(writer, sheet_name='race')\n",
    "# Tumor grade\n",
    "try:\n",
    "    del book['tumor grade']\n",
    "except:\n",
    "    pass\n",
    "tcga_ov_1_keep.stb.freq(['tumor_grade']).to_excel(writer, sheet_name='tumor grade')\n",
    "# Tumor stage\n",
    "try:\n",
    "    del book['tumor stage']\n",
    "except:\n",
    "    pass\n",
    "tcga_ov_1_keep.stb.freq(['tumor_stage']).to_excel(writer, sheet_name='tumor stage')\n",
    "# Death\n",
    "try:\n",
    "    del book['death']\n",
    "except:\n",
    "    pass\n",
    "tcga_ov_1_keep.stb.freq(['outcome_overall_survival_censoring']).to_excel(writer, sheet_name='death')\n",
    "# Age\n",
    "try:\n",
    "    del book['age']\n",
    "except:\n",
    "    pass\n",
    "pd.DataFrame(tcga_ov_1_keep['age_at_initial_pathologic_diagnosis'].describe()).to_excel(writer, sheet_name='age')\n",
    "# Overall survival\n",
    "try:\n",
    "    del book['days survival']\n",
    "except:\n",
    "    pass\n",
    "pd.DataFrame(tcga_ov_1_keep['total_days_overall_survival'].describe()).to_excel(writer, sheet_name='days survival')\n",
    "\n",
    "writer.save()\n",
    "writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vRX2hrEOPvMu"
   },
   "source": [
    "Add death dummy to drug lines data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "id": "X95jfaknP5oS"
   },
   "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>patient</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>therapy</th>\n",
       "      <th>previous_lines</th>\n",
       "      <th>treat</th>\n",
       "      <th>no_treat</th>\n",
       "      <th>death</th>\n",
       "      <th>race</th>\n",
       "      <th>age</th>\n",
       "      <th>tumor_stage</th>\n",
       "      <th>tumor_grade</th>\n",
       "      <th>total_ev</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>0.0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>133.0</td>\n",
       "      <td>445.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>445.0</td>\n",
       "      <td>462.0</td>\n",
       "      <td>[Dortezomib]</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TCGA-04-1332</td>\n",
       "      <td>0.0</td>\n",
       "      <td>151.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel, Topotecan]</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>70.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TCGA-04-1332</td>\n",
       "      <td>151.0</td>\n",
       "      <td>396.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>70.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1912</th>\n",
       "      <td>TCGA-61-2610</td>\n",
       "      <td>500.0</td>\n",
       "      <td>561.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>61.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1913</th>\n",
       "      <td>TCGA-61-2610</td>\n",
       "      <td>561.0</td>\n",
       "      <td>1101.0</td>\n",
       "      <td>[Doxorubicin]</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>61.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1914</th>\n",
       "      <td>TCGA-61-2611</td>\n",
       "      <td>0.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>40.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1915</th>\n",
       "      <td>TCGA-61-2612</td>\n",
       "      <td>0.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>63.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1916</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>0.0</td>\n",
       "      <td>105.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>73.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1917 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           patient  start     end                               therapy  \\\n",
       "0     TCGA-04-1331    0.0   133.0             [Carboplatin, Paclitaxel]   \n",
       "1     TCGA-04-1331  133.0   445.0                                    []   \n",
       "2     TCGA-04-1331  445.0   462.0                          [Dortezomib]   \n",
       "3     TCGA-04-1332    0.0   151.0  [Carboplatin, Paclitaxel, Topotecan]   \n",
       "4     TCGA-04-1332  151.0   396.0                                    []   \n",
       "...            ...    ...     ...                                   ...   \n",
       "1912  TCGA-61-2610  500.0   561.0                                    []   \n",
       "1913  TCGA-61-2610  561.0  1101.0                         [Doxorubicin]   \n",
       "1914  TCGA-61-2611    0.0   238.0             [Carboplatin, Paclitaxel]   \n",
       "1915  TCGA-61-2612    0.0   131.0             [Carboplatin, Paclitaxel]   \n",
       "1916  TCGA-61-2613    0.0   105.0             [Carboplatin, Paclitaxel]   \n",
       "\n",
       "      previous_lines  treat  no_treat  death   race   age tumor_stage  \\\n",
       "0                  0      4         1      0  WHITE  79.0        IIIC   \n",
       "1                  1      1        10      0  WHITE  79.0        IIIC   \n",
       "2                  1      0         0      1  WHITE  79.0        IIIC   \n",
       "3                  0      5         1      0  WHITE  70.0        IIIC   \n",
       "4                  1      1         8      0  WHITE  70.0        IIIC   \n",
       "...              ...    ...       ...    ...    ...   ...         ...   \n",
       "1912               2      1         2      0  WHITE  61.0        IIIC   \n",
       "1913               2      0        17      1  WHITE  61.0        IIIC   \n",
       "1914               0      0         7      1  WHITE  40.0        IIIC   \n",
       "1915               0      0         4      1  WHITE  63.0        IIIC   \n",
       "1916               0      0         3      1  WHITE  73.0        IIIC   \n",
       "\n",
       "     tumor_grade  total_ev  \n",
       "0             G3         5  \n",
       "1             G3        11  \n",
       "2             G3         1  \n",
       "3             G3         6  \n",
       "4             G3         9  \n",
       "...          ...       ...  \n",
       "1912          G3         3  \n",
       "1913          G3        18  \n",
       "1914          G3         8  \n",
       "1915          G3         5  \n",
       "1916          G3         4  \n",
       "\n",
       "[1917 rows x 13 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge in final death event for each patient\n",
    "lines_df_2 = lines_df.merge(tcga_ov_1_keep[['bcr_patient_barcode', 'outcome_overall_survival_censoring', 'race', 'age_at_initial_pathologic_diagnosis', 'tumor_stage', 'tumor_grade']],\n",
    "                            on='bcr_patient_barcode').rename(columns={'bcr_patient_barcode': 'patient', 'outcome_overall_survival_censoring': 'death', 'age_at_initial_pathologic_diagnosis': 'age'})\n",
    "\n",
    "# Only keep death event on last line\n",
    "lines_df_2.loc[lines_df_2['patient'] == lines_df_2['patient'].shift(-1), 'death'] = 0\n",
    "\n",
    "# Fix no treat on last line where death = 0\n",
    "def fix_notreat(row):\n",
    "    if (row['treat'] == 0) & (row['no_treat'] == 0) & (row['death'] == 0):\n",
    "        return 1\n",
    "    else:\n",
    "        return row['no_treat']\n",
    "lines_df_2['no_treat'] = lines_df_2.apply(lambda row: fix_notreat(row), axis=1)\n",
    "\n",
    "# Total events\n",
    "lines_df_2['total_ev'] = lines_df_2.apply(lambda row: max(1, row['treat'] + row['no_treat'] + row['death']), axis=1)\n",
    "\n",
    "# Function to get indices of a therapy\n",
    "def get_index_pos(my_list, val):\n",
    "    return [i for i, x in enumerate(my_list) if x == val]\n",
    "\n",
    "lines_df_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 402
    },
    "id": "kId5QzCmZie1",
    "outputId": "35f4770e-50e1-4f61-f167-82d64a05b284"
   },
   "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>patient</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>therapy</th>\n",
       "      <th>previous_lines</th>\n",
       "      <th>treat</th>\n",
       "      <th>no_treat</th>\n",
       "      <th>death</th>\n",
       "      <th>race</th>\n",
       "      <th>age</th>\n",
       "      <th>tumor_stage</th>\n",
       "      <th>tumor_grade</th>\n",
       "      <th>total_ev</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>0.0</td>\n",
       "      <td>133.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>133.0</td>\n",
       "      <td>445.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>445.0</td>\n",
       "      <td>462.0</td>\n",
       "      <td>[Dortezomib]</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>79.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TCGA-04-1332</td>\n",
       "      <td>0.0</td>\n",
       "      <td>151.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel, Topotecan]</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>70.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TCGA-04-1332</td>\n",
       "      <td>151.0</td>\n",
       "      <td>396.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>70.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1259</th>\n",
       "      <td>TCGA-61-2610</td>\n",
       "      <td>500.0</td>\n",
       "      <td>561.0</td>\n",
       "      <td>[]</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>61.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1260</th>\n",
       "      <td>TCGA-61-2610</td>\n",
       "      <td>561.0</td>\n",
       "      <td>1101.0</td>\n",
       "      <td>[Doxorubicin]</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>17</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>61.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1261</th>\n",
       "      <td>TCGA-61-2611</td>\n",
       "      <td>0.0</td>\n",
       "      <td>238.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>40.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1262</th>\n",
       "      <td>TCGA-61-2612</td>\n",
       "      <td>0.0</td>\n",
       "      <td>131.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>63.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1263</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>0.0</td>\n",
       "      <td>105.0</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>73.0</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1264 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           patient  start     end                               therapy  \\\n",
       "0     TCGA-04-1331    0.0   133.0             [Carboplatin, Paclitaxel]   \n",
       "1     TCGA-04-1331  133.0   445.0                                    []   \n",
       "2     TCGA-04-1331  445.0   462.0                          [Dortezomib]   \n",
       "3     TCGA-04-1332    0.0   151.0  [Carboplatin, Paclitaxel, Topotecan]   \n",
       "4     TCGA-04-1332  151.0   396.0                                    []   \n",
       "...            ...    ...     ...                                   ...   \n",
       "1259  TCGA-61-2610  500.0   561.0                                    []   \n",
       "1260  TCGA-61-2610  561.0  1101.0                         [Doxorubicin]   \n",
       "1261  TCGA-61-2611    0.0   238.0             [Carboplatin, Paclitaxel]   \n",
       "1262  TCGA-61-2612    0.0   131.0             [Carboplatin, Paclitaxel]   \n",
       "1263  TCGA-61-2613    0.0   105.0             [Carboplatin, Paclitaxel]   \n",
       "\n",
       "      previous_lines  treat  no_treat  death   race   age tumor_stage  \\\n",
       "0                  0      4         1      0  WHITE  79.0        IIIC   \n",
       "1                  1      1        10      0  WHITE  79.0        IIIC   \n",
       "2                  1      0         0      1  WHITE  79.0        IIIC   \n",
       "3                  0      5         1      0  WHITE  70.0        IIIC   \n",
       "4                  1      1         8      0  WHITE  70.0        IIIC   \n",
       "...              ...    ...       ...    ...    ...   ...         ...   \n",
       "1259               2      1         2      0  WHITE  61.0        IIIC   \n",
       "1260               2      0        17      1  WHITE  61.0        IIIC   \n",
       "1261               0      0         7      1  WHITE  40.0        IIIC   \n",
       "1262               0      0         4      1  WHITE  63.0        IIIC   \n",
       "1263               0      0         3      1  WHITE  73.0        IIIC   \n",
       "\n",
       "     tumor_grade  total_ev  \n",
       "0             G3         5  \n",
       "1             G3        11  \n",
       "2             G3         1  \n",
       "3             G3         6  \n",
       "4             G3         9  \n",
       "...          ...       ...  \n",
       "1259          G3         3  \n",
       "1260          G3        18  \n",
       "1261          G3         8  \n",
       "1262          G3         5  \n",
       "1263          G3         4  \n",
       "\n",
       "[1264 rows x 13 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create version of final dataset that only includes patients who died\n",
    "d_pats = list(tcga_ov_1_keep['bcr_patient_barcode'].loc[tcga_ov_1_keep['outcome_overall_survival_censoring'] == 1])\n",
    "\n",
    "lines_df_d = lines_df_2[lines_df_2['patient'].isin(d_pats)].reset_index().drop(['index'], axis=1)\n",
    "lines_df_d"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create version of data for regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>patient</th>\n",
       "      <th>therapy</th>\n",
       "      <th>previous_lines</th>\n",
       "      <th>treat</th>\n",
       "      <th>no_treat</th>\n",
       "      <th>death</th>\n",
       "      <th>race</th>\n",
       "      <th>tumor_stage</th>\n",
       "      <th>tumor_grade</th>\n",
       "      <th>ther_str</th>\n",
       "      <th>...</th>\n",
       "      <th>IIIB</th>\n",
       "      <th>IIIC</th>\n",
       "      <th>IV</th>\n",
       "      <th>NOT SPECIFIED_tumor_stage</th>\n",
       "      <th>G1</th>\n",
       "      <th>G2</th>\n",
       "      <th>G3</th>\n",
       "      <th>GB</th>\n",
       "      <th>GX</th>\n",
       "      <th>NOT SPECIFIED_tumor_grade</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TCGA-04-1331</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9291</th>\n",
       "      <td>TCGA-61-2612</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9292</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9293</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9294</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9295</th>\n",
       "      <td>TCGA-61-2613</td>\n",
       "      <td>[Carboplatin, Paclitaxel]</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>WHITE</td>\n",
       "      <td>IIIC</td>\n",
       "      <td>G3</td>\n",
       "      <td>Carboplatin,Paclitaxel</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9296 rows × 164 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           patient                    therapy  previous_lines  treat  \\\n",
       "0     TCGA-04-1331  [Carboplatin, Paclitaxel]               0      1   \n",
       "1     TCGA-04-1331  [Carboplatin, Paclitaxel]               0      1   \n",
       "2     TCGA-04-1331  [Carboplatin, Paclitaxel]               0      1   \n",
       "3     TCGA-04-1331  [Carboplatin, Paclitaxel]               0      1   \n",
       "4     TCGA-04-1331  [Carboplatin, Paclitaxel]               0      0   \n",
       "...            ...                        ...             ...    ...   \n",
       "9291  TCGA-61-2612  [Carboplatin, Paclitaxel]               0      0   \n",
       "9292  TCGA-61-2613  [Carboplatin, Paclitaxel]               0      1   \n",
       "9293  TCGA-61-2613  [Carboplatin, Paclitaxel]               0      1   \n",
       "9294  TCGA-61-2613  [Carboplatin, Paclitaxel]               0      1   \n",
       "9295  TCGA-61-2613  [Carboplatin, Paclitaxel]               0      0   \n",
       "\n",
       "      no_treat  death   race tumor_stage tumor_grade                ther_str  \\\n",
       "0            0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "1            0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "2            0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "3            0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "4            1      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "...        ...    ...    ...         ...         ...                     ...   \n",
       "9291         0      1  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "9292         0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "9293         0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "9294         0      0  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "9295         0      1  WHITE        IIIC          G3  Carboplatin,Paclitaxel   \n",
       "\n",
       "      ...  IIIB  IIIC  IV  NOT SPECIFIED_tumor_stage  G1  G2  G3  GB  GX  \\\n",
       "0     ...     0     1   0                          0   0   0   1   0   0   \n",
       "1     ...     0     1   0                          0   0   0   1   0   0   \n",
       "2     ...     0     1   0                          0   0   0   1   0   0   \n",
       "3     ...     0     1   0                          0   0   0   1   0   0   \n",
       "4     ...     0     1   0                          0   0   0   1   0   0   \n",
       "...   ...   ...   ...  ..                        ...  ..  ..  ..  ..  ..   \n",
       "9291  ...     0     1   0                          0   0   0   1   0   0   \n",
       "9292  ...     0     1   0                          0   0   0   1   0   0   \n",
       "9293  ...     0     1   0                          0   0   0   1   0   0   \n",
       "9294  ...     0     1   0                          0   0   0   1   0   0   \n",
       "9295  ...     0     1   0                          0   0   0   1   0   0   \n",
       "\n",
       "      NOT SPECIFIED_tumor_grade  \n",
       "0                             0  \n",
       "1                             0  \n",
       "2                             0  \n",
       "3                             0  \n",
       "4                             0  \n",
       "...                         ...  \n",
       "9291                          0  \n",
       "9292                          0  \n",
       "9293                          0  \n",
       "9294                          0  \n",
       "9295                          0  \n",
       "\n",
       "[9296 rows x 164 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def fix_treat(row):\n",
    "    if (row['patient'] == row['pat_lag']) & (row['ther_lag'] != ''):\n",
    "        return 1\n",
    "    else:\n",
    "        return 0\n",
    "\n",
    "def fix_death(row):\n",
    "    if row['treat'] == 1:\n",
    "        return 0\n",
    "    else:\n",
    "        return row['death']\n",
    "\n",
    "def fix_notreat(row):\n",
    "    return (1 - row['treat'] - row['death'])\n",
    "\n",
    "def treat_state(row):\n",
    "    if row['ther_str'] == '':\n",
    "        return 0\n",
    "    else:\n",
    "        return 1\n",
    "    \n",
    "def make_reg_df(df): \n",
    "    df_reg = df.loc[df.index.repeat(df.total_ev)].reset_index(drop=True)\n",
    "    df_reg['ther_str'] = [','.join(map(str, l)) for l in df_reg['therapy']]\n",
    "    df_reg['ther_lag'] = df_reg['ther_str'].shift(-1)\n",
    "    df_reg['pat_lag'] = df_reg['patient'].shift(-1)\n",
    "\n",
    "    df_reg['treat'] = df_reg.apply(lambda row: fix_treat(row), axis=1) # prob that next state is treatment\n",
    "    df_reg['death'] = df_reg.apply(lambda row: fix_death(row), axis=1) # prob that next state is death\n",
    "    df_reg['no_treat'] = df_reg.apply(lambda row: fix_notreat(row), axis=1) # prob that next state no treatment\n",
    "    \n",
    "    df_reg = df_reg.drop(columns=['total_ev', 'ther_lag', 'pat_lag', 'start', 'end'])\n",
    "    \n",
    "    df_reg['months'] =  df_reg.groupby('patient').cumcount() # time trend\n",
    "    \n",
    "    df_reg['treat_state'] = df_reg.apply(lambda row: treat_state(row), axis=1) # current state\n",
    "    \n",
    "    df_reg = df_reg.merge(pd.get_dummies(df_reg['ther_str']), left_index=True, right_index=True)\n",
    "    df_reg = df_reg.merge(pd.get_dummies(df_reg['race']), left_index=True, right_index=True)\n",
    "    df_reg = df_reg.merge(pd.get_dummies(df_reg['tumor_stage']), left_index=True, right_index=True)\n",
    "    df_reg = df_reg.merge(pd.get_dummies(df_reg['tumor_grade']), left_index=True, right_index=True)\n",
    "    df_reg = df_reg.rename(columns={'NOT SPECIFIED_x':'NOT SPECIFIED_race', 'NOT SPECIFIED_y':'NOT SPECIFIED_tumor_stage', 'NOT SPECIFIED':'NOT SPECIFIED_tumor_grade'})\n",
    "    \n",
    "    age = df_reg['age']\n",
    "    df_reg = df_reg.drop(['age'], axis=1)\n",
    "    df_reg.insert(11, 'age', age)\n",
    "    \n",
    "    return df_reg\n",
    "\n",
    "df_d_reg = make_reg_df(lines_df_d)\n",
    "\n",
    "df_reg = make_reg_df(lines_df_2)\n",
    "\n",
    "df_reg"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Treatment summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\bmurp\\miniconda3\\envs\\d14med\\lib\\site-packages\\scipy\\stats\\stats.py:2500: RuntimeWarning: invalid value encountered in true_divide\n",
      "  return (a - mns) / sstd\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(array([  0.5,   2.5,   4.5,   6.5,   8.5,  10.5,  12.5,  14.5,  16.5,\n",
       "         18.5,  20.5,  22.5,  24.5,  26.5,  28.5,  30.5,  32.5,  34.5,\n",
       "         36.5,  38.5,  40.5,  42.5,  44.5,  46.5,  48.5,  50.5,  52.5,\n",
       "         54.5,  56.5,  58.5,  60.5,  62.5,  64.5,  66.5,  68.5,  70.5,\n",
       "         72.5,  74.5,  76.5,  78.5,  80.5,  82.5,  84.5,  86.5,  88.5,\n",
       "         90.5,  92.5,  94.5,  96.5,  98.5, 100.5, 102.5, 104.5, 106.5,\n",
       "        108.5, 110.5, 112.5, 114.5, 116.5, 118.5]),\n",
       " [Text(0, 0.5, '0'),\n",
       "  Text(0, 2.5, '2'),\n",
       "  Text(0, 4.5, '4'),\n",
       "  Text(0, 6.5, '6'),\n",
       "  Text(0, 8.5, '8'),\n",
       "  Text(0, 10.5, '10'),\n",
       "  Text(0, 12.5, '12'),\n",
       "  Text(0, 14.5, '14'),\n",
       "  Text(0, 16.5, '16'),\n",
       "  Text(0, 18.5, '18'),\n",
       "  Text(0, 20.5, '20'),\n",
       "  Text(0, 22.5, '22'),\n",
       "  Text(0, 24.5, '24'),\n",
       "  Text(0, 26.5, '26'),\n",
       "  Text(0, 28.5, '28'),\n",
       "  Text(0, 30.5, '30'),\n",
       "  Text(0, 32.5, '32'),\n",
       "  Text(0, 34.5, '34'),\n",
       "  Text(0, 36.5, '36'),\n",
       "  Text(0, 38.5, '38'),\n",
       "  Text(0, 40.5, '40'),\n",
       "  Text(0, 42.5, '42'),\n",
       "  Text(0, 44.5, '44'),\n",
       "  Text(0, 46.5, '46'),\n",
       "  Text(0, 48.5, '48'),\n",
       "  Text(0, 50.5, '50'),\n",
       "  Text(0, 52.5, '52'),\n",
       "  Text(0, 54.5, '54'),\n",
       "  Text(0, 56.5, '56'),\n",
       "  Text(0, 58.5, '58'),\n",
       "  Text(0, 60.5, '60'),\n",
       "  Text(0, 62.5, '62'),\n",
       "  Text(0, 64.5, '64'),\n",
       "  Text(0, 66.5, '66'),\n",
       "  Text(0, 68.5, '68'),\n",
       "  Text(0, 70.5, '70'),\n",
       "  Text(0, 72.5, '72'),\n",
       "  Text(0, 74.5, '74'),\n",
       "  Text(0, 76.5, '76'),\n",
       "  Text(0, 78.5, '78'),\n",
       "  Text(0, 80.5, '80'),\n",
       "  Text(0, 82.5, '82'),\n",
       "  Text(0, 84.5, '84'),\n",
       "  Text(0, 86.5, '86'),\n",
       "  Text(0, 88.5, '88'),\n",
       "  Text(0, 90.5, '90'),\n",
       "  Text(0, 92.5, '92'),\n",
       "  Text(0, 94.5, '94'),\n",
       "  Text(0, 96.5, '96'),\n",
       "  Text(0, 98.5, '98'),\n",
       "  Text(0, 100.5, '100'),\n",
       "  Text(0, 102.5, '102'),\n",
       "  Text(0, 104.5, '104'),\n",
       "  Text(0, 106.5, '106'),\n",
       "  Text(0, 108.5, '108'),\n",
       "  Text(0, 110.5, '110'),\n",
       "  Text(0, 112.5, '112'),\n",
       "  Text(0, 114.5, '114'),\n",
       "  Text(0, 116.5, '116'),\n",
       "  Text(0, 118.5, '118')])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1152 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Heatmap data\n",
    "df_heat = df_reg.stb.freq(['ther_str', 'months'])[['ther_str', 'months', 'count']].sort_values(by=['months', 'ther_str'])\n",
    "df_heat = df_heat.pivot_table(index=['months'], columns='ther_str', values='count').reset_index().rename_axis(None, axis=1).set_index('months')\n",
    "df_heat = df_heat.fillna(value=0).drop([''], axis=1)\n",
    "s = df_heat.sum()\n",
    "df_heat = df_heat[s.sort_values(ascending=False).index]\n",
    "# convert to z-scores\n",
    "for row in df_heat.index:\n",
    "    stats.zscore(df_heat.iloc[row])\n",
    "    df_heat.iloc[row] = stats.zscore(df_heat.iloc[row])\n",
    "df_heat = df_heat.fillna(value=0)\n",
    "\n",
    "# Make heatmap\n",
    "sns.set(font_scale=1)\n",
    "\n",
    "cmap = sns.color_palette(\"Blues\", as_cmap=True)\n",
    "plt.figure(figsize=(16,16))\n",
    "plt.rcParams[\"font.family\"] = \"Arial\"\n",
    "ax = sns.heatmap(data=df_heat,\n",
    "                 cmap=cmap, \n",
    "                 square=1,\n",
    "                 vmin=0, vmax=3,\n",
    "                 cbar_kws={\"shrink\": .5},\n",
    "                 )\n",
    "plt.title('Relative Patients Receiving Each Treatment by Month/State', size=18, fontweight='bold')\n",
    "plt.ylabel('Months from start of treatment', size=14)\n",
    "plt.xticks(size=12, rotation='vertical')\n",
    "plt.yticks(size=12)\n",
    "# cbar = df_heat.collections[0].colorbar\n",
    "# cbar.ax.tick_params(labelsize=6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "WnwEXGPRAq7A"
   },
   "source": [
    "## Create MDP objects\n",
    "\n",
    "* This should probably be the start of the next notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ZvulJ_dUAv2h"
   },
   "source": [
    "State and Action set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Ciq1VT6qIHUH",
    "outputId": "321fd508-7692-4145-8519-23ddd96ea4a9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "128\n",
      "[[], ['Abagovomab'], ['Aldesleukin'], ['Altretamine'], ['Amifostine', 'Carboplatin', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Carboplatin', 'Paclitaxel'], ['Amifostine', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Paclitaxel'], ['Aminocamptothecin'], ['Anastrozole']]\n",
      "[['Abagovomab'], ['Aldesleukin'], ['Altretamine'], ['Amifostine', 'Carboplatin', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Carboplatin', 'Paclitaxel'], ['Amifostine', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Paclitaxel'], ['Aminocamptothecin'], ['Anastrozole'], ['Anastrozole', 'Doxorubicin']]\n",
      "108\n",
      "[[], ['Aldesleukin'], ['Altretamine'], ['Amifostine', 'Carboplatin', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Carboplatin', 'Paclitaxel'], ['Amifostine', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Paclitaxel'], ['Aminocamptothecin'], ['Anastrozole'], ['Anastrozole', 'Doxorubicin']]\n",
      "[['Aldesleukin'], ['Altretamine'], ['Amifostine', 'Carboplatin', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Carboplatin', 'Paclitaxel'], ['Amifostine', 'Cisplatin', 'Paclitaxel'], ['Amifostine', 'Paclitaxel'], ['Aminocamptothecin'], ['Anastrozole'], ['Anastrozole', 'Doxorubicin'], ['Anastrozole', 'Tamoxifen']]\n",
      "['AMERICAN INDIAN OR ALASKA NATIVE', 'ASIAN', 'BLACK OR AFRICAN AMERICAN', 'NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER', 'NOT SPECIFIED', 'WHITE']\n",
      "['AMERICAN INDIAN OR ALASKA NATIVE', 'ASIAN', 'BLACK OR AFRICAN AMERICAN', 'NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER', 'WHITE']\n",
      "['AMERICAN INDIAN OR ALASKA NATIVE', 'ASIAN', 'BLACK OR AFRICAN AMERICAN', 'NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER', 'NOT SPECIFIED', 'WHITE']\n",
      "['AMERICAN INDIAN OR ALASKA NATIVE', 'ASIAN', 'BLACK OR AFRICAN AMERICAN', 'NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER', 'WHITE']\n",
      "['IA', 'IB', 'IC', 'IIA', 'IIB', 'IIC', 'IIIA', 'IIIB', 'IIIC', 'IV', 'NOT SPECIFIED']\n",
      "['IA', 'IB', 'IC', 'IIA', 'IIB', 'IIC', 'IIIA', 'IIIB', 'IIIC', 'IV']\n",
      "['IIB', 'IIC', 'IIIA', 'IIIB', 'IIIC', 'IV', 'NOT SPECIFIED']\n",
      "['IIB', 'IIC', 'IIIA', 'IIIB', 'IIIC', 'IV']\n",
      "['G1', 'G2', 'G3', 'GB', 'GX', 'NOT SPECIFIED']\n",
      "['G1', 'G2', 'G3', 'GB', 'GX']\n",
      "['G1', 'G2', 'G3', 'GB', 'GX', 'NOT SPECIFIED']\n",
      "['G1', 'G2', 'G3', 'GB', 'GX']\n"
     ]
    }
   ],
   "source": [
    "# States\n",
    "state_set = ['N', 'T', 'D']\n",
    "\n",
    "# Combos/actions\n",
    "combos_all = list(lines_df_2['therapy'])\n",
    "combos_all.sort()\n",
    "combos_all = list(combos_all for combos_all,_ in itertools.groupby(combos_all))\n",
    "print(len(combos_all))\n",
    "print(combos_all[0:10])\n",
    "\n",
    "combos_all_drugs = combos_all.copy()\n",
    "combos_all_drugs.remove([])\n",
    "print(combos_all_drugs[0:10]) \n",
    "\n",
    "# Combos for deceased patients\n",
    "combos_d = list(lines_df_d['therapy'])\n",
    "\n",
    "combos_d.sort()\n",
    "combos_d = list(combos_d for combos_d,_ in itertools.groupby(combos_d))\n",
    "print(len(combos_d))\n",
    "print(combos_d[0:10])\n",
    "\n",
    "combos_drugs_d = combos_d.copy()\n",
    "combos_drugs_d.remove([])\n",
    "print(combos_drugs_d[0:10]) \n",
    "\n",
    "###### \"2\" versions are just dropping the 'NOT SPECIFIED' value\n",
    "# Race\n",
    "races_all = list(lines_df_2['race'].unique())\n",
    "races_all.sort()\n",
    "print(races_all)\n",
    "races_all2 = races_all.copy()\n",
    "races_all2.remove('NOT SPECIFIED')\n",
    "print(races_all2)\n",
    "    # Death only dataset\n",
    "races_d = list(lines_df_d['race'].unique())\n",
    "races_d.sort()\n",
    "print(races_d)\n",
    "races_d2 = races_d.copy()\n",
    "races_d2.remove('NOT SPECIFIED')\n",
    "print(races_d2)\n",
    "\n",
    "# Tumor stage\n",
    "tumor_stages_all = list(lines_df_2['tumor_stage'].unique())\n",
    "tumor_stages_all.sort()\n",
    "print(tumor_stages_all)\n",
    "tumor_stages_all2 = tumor_stages_all.copy()\n",
    "tumor_stages_all2.remove('NOT SPECIFIED')\n",
    "print(tumor_stages_all2)\n",
    "    # Death only dataset\n",
    "tumor_stages_d = list(lines_df_d['tumor_stage'].unique())\n",
    "tumor_stages_d.sort()\n",
    "print(tumor_stages_d)\n",
    "tumor_stages_d2 = tumor_stages_d.copy()\n",
    "tumor_stages_d2.remove('NOT SPECIFIED')\n",
    "print(tumor_stages_d2)\n",
    "\n",
    "# Tumor grade\n",
    "tumor_grades_all = list(lines_df_2['tumor_grade'].unique())\n",
    "tumor_grades_all.sort()\n",
    "print(tumor_grades_all)\n",
    "tumor_grades_all2 = tumor_grades_all.copy()\n",
    "tumor_grades_all2.remove('NOT SPECIFIED')\n",
    "print(tumor_grades_all2)\n",
    "    # Death only dataset\n",
    "tumor_grades_d = list(lines_df_d['tumor_grade'].unique())\n",
    "tumor_grades_d.sort()\n",
    "print(tumor_grades_d)\n",
    "tumor_grades_d2 = tumor_grades_d.copy()\n",
    "tumor_grades_d2.remove('NOT SPECIFIED')\n",
    "print(tumor_grades_d2)\n",
    "\n",
    "# Function to create random array for descriptive variables\n",
    "def rand_bin_array(K, N):\n",
    "    arr = np.zeros(N)\n",
    "    arr[:K]  = 1\n",
    "    np.random.shuffle(arr)\n",
    "    return arr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Static transition matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "DEeUFm3ZM8RU",
    "outputId": "0bcf7d3f-2c6a-410c-c70d-b3173b3f1b72"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "T_a shape:\n",
      "    N  T  D\n",
      " N [0. 0. 0.] \n",
      " T [0. 0. 0.]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[[0.84903439, 0.15096561, 0.        ],\n",
       "        [0.        , 0.        , 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.25      , 0.75      , 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.26315789, 0.68421053, 0.05263158]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.11111111, 0.88888889, 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.33333333, 0.66666667, 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.33333333, 0.66666667, 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.16666667, 0.83333333, 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.5       , 0.5       , 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.2       , 0.8       , 0.        ]],\n",
       "\n",
       "       [[0.        , 0.        , 0.        ],\n",
       "        [0.        , 1.        , 0.        ]]])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Treatment States = {No Treatment, Treatment, Death}\n",
    "# All or death-only action sets\n",
    "combos = combos_d # Switch to _all when use full dataset\n",
    "combos_drugs = combos_drugs_d\n",
    "\n",
    "# Descriptive States (all or death-only)\n",
    "races = races_d\n",
    "races2 = races_d2\n",
    "tumor_stages = tumor_stages_d\n",
    "tumor_stages2 = tumor_stages_d2\n",
    "tumor_grades = tumor_grades_d\n",
    "tumor_grades2 = tumor_grades_d2\n",
    "\n",
    "# Dataset to use\n",
    "trans_data = df_d_reg # change to df_reg if using full dataset\n",
    "\n",
    "########################### Currently not using any of the below in this cell\n",
    "nactions = len(combos) # action set is list of unique drug combos, not including choosing no drugs\n",
    "nstates = len(state_set) # three state probabilities\n",
    "\n",
    "# Initialize transition matrix\n",
    "transitions = np.zeros((nactions,nstates-1,nstates))\n",
    "print('T_a shape:\\n',\n",
    "      '   N  T  D\\n',\n",
    "      'N', transitions[0][0],'\\n',\n",
    "      'T', transitions[0][1])\n",
    "\n",
    "# Calculate probabilities\n",
    "for i in range(nactions):\n",
    "    inds = get_index_pos(list(lines_df_2['therapy']), combos[i])\n",
    "    \n",
    "    n_prob = lines_df_2[['no_treat']].iloc[inds, :].sum().values[0] / lines_df_2[['total_ev']].iloc[inds, :].sum().values[0]\n",
    "    t_prob = lines_df_2[['treat']].iloc[inds, :].sum().values[0] / lines_df_2[['total_ev']].iloc[inds, :].sum().values[0]\n",
    "    d_prob = lines_df_2[['death']].iloc[inds, :].sum().values[0] / lines_df_2[['total_ev']].iloc[inds, :].sum().values[0]\n",
    "    # placeholder so that every action has a chance of death. Will update with regression to calc probs\n",
    "    # if d_prob == 0:\n",
    "    #   d_prob = .1\n",
    "    #   n_prob = n_prob - (d_prob / 2)\n",
    "    #   t_prob = t_prob - (d_prob / 2)\n",
    "    ###################################################################################################\n",
    "\n",
    "    if combos[i] == []:\n",
    "        # N --> \n",
    "        transitions[i][0][0] = n_prob\n",
    "        transitions[i][0][1] = t_prob\n",
    "        transitions[i][0][2] = d_prob\n",
    "        # T -->\n",
    "        transitions[i][1][0] = 0\n",
    "        transitions[i][1][1] = 0\n",
    "        transitions[i][1][2] = 0\n",
    "    else:\n",
    "        # N --> \n",
    "        transitions[i][0][0] = 0\n",
    "        transitions[i][0][1] = 0\n",
    "        transitions[i][0][2] = 0\n",
    "        # T -->\n",
    "        transitions[i][1][0] = n_prob\n",
    "        transitions[i][1][1] = t_prob\n",
    "        transitions[i][1][2] = d_prob\n",
    "\n",
    "# Get tranistion probabilities function\n",
    "def get_probs(act_choice, curr_state, trans_set=transitions, act_set=combos, states_set=state_set):\n",
    "    return trans_set[act_set.index(act_choice)][states_set.index(curr_state)]\n",
    "\n",
    "transitions[0:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "fuspJEgVFPKy"
   },
   "source": [
    "Cox Proportional Hazard Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.04884743176118048, 0.014315723284023594)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fit regressions\n",
    "    # Survival\n",
    "X_cox = trans_data.drop(['previous_lines', '', 'no_treat', 'treat', 'patient', 'therapy', 'race', 'tumor_stage', \n",
    "                         'tumor_grade', 'ther_str', 'NOT SPECIFIED_tumor_stage', 'NOT SPECIFIED_tumor_grade', \n",
    "                         'NOT SPECIFIED_race'], axis=1)\n",
    "cph_surv = CoxPHFitter(penalizer=0.1)\n",
    "cph_surv.fit(X_cox, duration_col='months', event_col='death', robust=True) # add this to do specific cols: , formula=\"fin + wexp + age * prio\"\n",
    "\n",
    "    # Treatment\n",
    "X_cox = trans_data.drop(['previous_lines', '', 'treat', 'death', 'patient', 'therapy', 'race', 'tumor_stage',\n",
    "                         'tumor_grade', 'ther_str', 'NOT SPECIFIED_tumor_stage', 'NOT SPECIFIED_tumor_grade',\n",
    "                         'NOT SPECIFIED_race'], axis=1)\n",
    "cph_treat = CoxPHFitter(penalizer=0.1)\n",
    "cph_treat.fit(X_cox, duration_col='months', event_col='no_treat', robust=True)\n",
    "\n",
    "def cph_probs(data, state, months, action, age, race, tumor_stage, tumor_grade):\n",
    "    month = min(months, max(X_cox['months']))\n",
    "    if action == []:\n",
    "        pass\n",
    "    else:\n",
    "        act_ind = combos_drugs.index(action) + 2\n",
    "        \n",
    "    act_arr = np.zeros(len(combos_drugs)) # sim patient vectors = 0 if 'NOT SPECIFIED'\n",
    "    if action == []:\n",
    "        pass\n",
    "    else:\n",
    "        act_arr[combos_drugs.index(action)] = 1\n",
    "\n",
    "    x_test = np.concatenate((np.array([age]), np.array([state_set.index(state)]), act_arr, race, tumor_stage, tumor_grade))\n",
    "    x_test = x_test.reshape(1,-1)\n",
    "#     haz = (base_haz * part_haz.values[0]).iloc[months].values[0]\n",
    "    surv = cph_surv.predict_survival_function(x_test).iloc[month].values[0]\n",
    "    treat_prob = cph_treat.predict_survival_function(x_test).iloc[month].values[0]\n",
    "    return surv, treat_prob\n",
    "\n",
    "race_arr = np.array([0, 0, 0, 0, 1])\n",
    "ts_arr = np.array([0, 0, 1, 0, 0, 0])\n",
    "tg_arr = np.array([0, 0, 1, 0, 0])\n",
    "\n",
    "# check that it works\n",
    "cph_probs(X_cox, 'T', 119, ['Carboplatin', 'Paclitaxel'], 80, race_arr, ts_arr, tg_arr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# CPH figs\n",
    "# cph.plot_partial_effects_on_outcome(covariates='age', values=[40, 50, 60, 70, 80], cmap='coolwarm')\n",
    "\n",
    "# cph.print_summary()\n",
    "# cph_treat.print_summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.07607657694501867, 0.9146841888244316, 0.009239234230549809]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.array(trans_data.loc[:, 'months':])\n",
    "# OLS Linear, event on time and treatment, only patients who died\n",
    "        # Treatment prob\n",
    "y = np.array(trans_data['treat'])\n",
    "t_reg = sm.OLS(y, X)\n",
    "t_reg = t_reg.fit()\n",
    "\n",
    "\n",
    "        # No treatment prob\n",
    "y = np.array(trans_data['no_treat'])\n",
    "n_reg = sm.OLS(y, X)\n",
    "n_reg = n_reg.fit()\n",
    "\n",
    "        # Death prob\n",
    "y = np.array(trans_data['death'])\n",
    "d_reg = sm.OLS(y, X)\n",
    "d_reg = d_reg.fit()\n",
    "\n",
    "# MinMax and sum to 1 scaling for negative probs\n",
    "scaler = MinMaxScaler(feature_range=(0.01, 0.99)) # maybe change min to the overall proportion of deaths to total transitions\n",
    "\n",
    "def scaleMM(inprobs):\n",
    "    scaler.fit(inprobs.reshape(-1, 1))\n",
    "    outprobs = scaler.transform(inprobs)\n",
    "    outprobs = outprobs / outprobs.sum()\n",
    "    return outprobs\n",
    "\n",
    "# Function to calculate transition probabilities based on regressions\n",
    "def reg_probs(data, state, months, action, age, race, tumor_stage, tumor_grade):\n",
    "    act_ind = combos.index(action) + 3\n",
    "    race_ind = races.index(race) + len(combos) + 3\n",
    "    ts_ind = tumor_stages.index(tumor_stage) + len(combos) + len(races) + 3\n",
    "    tg_ind = tumor_grades.index(tumor_grade) + len(combos) + len(races) + len(tumor_stages) + 3\n",
    "    x_test = np.zeros(data.shape[1]-10)\n",
    "    x_test[0] = months\n",
    "    x_test[1] = age\n",
    "    x_test[2] = state_set.index(state)\n",
    "    x_test[act_ind] = 1\n",
    "    x_test[race_ind] = 1\n",
    "    x_test[ts_ind] = 1\n",
    "    x_test[tg_ind] = 1\n",
    "    x_test = x_test.reshape(1,-1)\n",
    "    probs = np.array([[n_reg.predict(x_test)[0], t_reg.predict(x_test)[0], d_reg.predict(x_test)[0]]])\n",
    "    probs = scaleMM(probs)[0].tolist()\n",
    "    return probs\n",
    "\n",
    "# check that it works\n",
    "reg_probs(trans_data, 'T', 50, ['Carboplatin', 'Paclitaxel'], 58, 'WHITE', 'IIIA', 'G3')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note: visualize probability of death over time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# # If want formatted regression output\n",
    "# print(t_reg.summary())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7Xbe4sqhFYNq"
   },
   "source": [
    "Q-Network: MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "id": "NX_SB9BTFcA_"
   },
   "outputs": [],
   "source": [
    "# Define DQN, one hidden layer MLP for now\n",
    "class DQN(nn.Module):\n",
    "    def __init__(self, in_features, hidden_size, out_features):\n",
    "        super(DQN, self).__init__()\n",
    "        self.fc = nn.Linear(in_features, hidden_size)\n",
    "        self.relu = nn.ReLU()\n",
    "        self.fc_out = nn.Linear(hidden_size, out_features)\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = self.fc(x)\n",
    "        out = self.relu(out)\n",
    "        out = self.fc_out(out)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a bigger NN\n",
    "class DQN_3(nn.Module):\n",
    "    def __init__(self, in_features, out_features):\n",
    "        super(DQN_3, self).__init__()\n",
    "        self.fc1 = nn.Linear(in_features, 128)\n",
    "        self.fc2 = nn.Linear(128, 256)\n",
    "        self.fc3 = nn.Linear(256, 256)\n",
    "        self.relu = nn.ReLU()\n",
    "        self.fc_out = nn.Linear(256, out_features)\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = self.fc1(x) # in\n",
    "        out = self.relu(out)\n",
    "        out = self.fc2(out) # hidden 1\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out) # hidden 2\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out) # hidden 3\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out) # hidden 4\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out) # hidden 5\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out) # hidden 6\n",
    "        out = self.relu(out)\n",
    "        out = self.fc_out(out) # out\n",
    "        return out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "qwi3cx2gsa4M"
   },
   "source": [
    "Replay Memory\n",
    "\n",
    "https://pytorch.org/tutorials/intermediate/reinforcement_q_learning.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "id": "rZ9jxVEmsefd"
   },
   "outputs": [],
   "source": [
    "Transition = namedtuple('Transition',\n",
    "                        ('state', 'action', 'next_state', 'reward'))\n",
    "\n",
    "class ReplayMemory(object):\n",
    "    def __init__(self, capacity):\n",
    "        self.capacity = capacity\n",
    "        self.memory = []\n",
    "        self.position = 0\n",
    "\n",
    "    def push(self, *args):\n",
    "        \"\"\"Saves a transition.\"\"\"\n",
    "        if len(self.memory) < self.capacity:\n",
    "            self.memory.append(None)\n",
    "        self.memory[self.position] = Transition(*args)\n",
    "        self.position = (self.position + 1) % self.capacity\n",
    "\n",
    "    def sample(self, batch_size):\n",
    "        return random.sample(self.memory, batch_size)\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.memory)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EB4ZbWbXCFzI"
   },
   "source": [
    "Model training loop\n",
    "\n",
    "https://pytorch.org/tutorials/intermediate/reinforcement_q_learning.html\n",
    "\n",
    "Use Q update formua: $Q^{\\pi}(s,a) \\leftarrow Q^{\\pi}(s,a) + \\alpha \\cdot [r + \\gamma \\cdot Q^{\\pi}(s', a') - Q^{\\pi}(s,a)]$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "id": "G0b-nblgCDVR"
   },
   "outputs": [],
   "source": [
    "def optimize_model():\n",
    "    if len(memory) < BATCH_SIZE:\n",
    "        return\n",
    "    transitions = memory.sample(BATCH_SIZE)\n",
    "    # Transpose the batch (see https://stackoverflow.com/a/19343/3343043 for\n",
    "    # detailed explanation). This converts batch-array of Transitions\n",
    "    # to Transition of batch-arrays.\n",
    "    batch = Transition(*zip(*transitions))\n",
    "\n",
    "    # Compute a mask of non-final states and concatenate the batch elements\n",
    "    # (a final state would've been the one after which simulation ended)\n",
    "    non_final_mask = torch.tensor(tuple(map(lambda s: s is not None,\n",
    "                                          batch.next_state)), device=device, dtype=torch.bool)\n",
    "    non_final_next_states = torch.stack([s for s in batch.next_state\n",
    "                                                if s is not None])\n",
    "\n",
    "    state_batch = torch.stack(batch.state)\n",
    "    action_batch = torch.stack(batch.action)\n",
    "    reward_batch = torch.cat(batch.reward)\n",
    "\n",
    "    # Compute Q(s_t, a) - the model computes Q(s_t), then we select the\n",
    "    # columns of actions taken. These are the actions which would've been taken\n",
    "    # for each batch state according to policy_net\n",
    "    state_action_values = policy(state_batch).gather(1, action_batch).squeeze()\n",
    "\n",
    "    # Compute V(s_{t+1}) for all next states.\n",
    "    # Expected values of actions for non_final_next_states are computed based\n",
    "    # on the \"older\" target_net; selecting their best reward with max(1)[0].\n",
    "    # This is merged based on the mask, such that we'll have either the expected\n",
    "    # state value or 0 in case the state was final.\n",
    "    next_state_values = torch.zeros(BATCH_SIZE, device=device)\n",
    "    next_state_values[non_final_mask] = target_net(non_final_next_states).max(1)[0].detach()\n",
    "\n",
    "    # Compute the expected Q values\n",
    "    expected_state_action_values = state_action_values + ALPHA * (reward_batch + (GAMMA * next_state_values) - state_action_values)\n",
    "    expected_state_action_values = expected_state_action_values\n",
    "#     expected_state_action_values = (next_state_values * GAMMA) + reward_batch # old version\n",
    "\n",
    "    # Compute Huber loss\n",
    "    loss = F.smooth_l1_loss(state_action_values, expected_state_action_values)\n",
    "\n",
    "    # Optimize the model\n",
    "    optimizer.zero_grad()\n",
    "    loss.backward()\n",
    "    for param in policy.parameters():\n",
    "        param.grad.data.clamp_(-1, 1)\n",
    "    optimizer.step()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QZLSsYJeheqC"
   },
   "source": [
    "## Environment\n",
    "\n",
    "Designed based on: https://github.com/MJeremy2017/reinforcement-learning-implementation/blob/master/GridWorld/gridWorld.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OZX71-HbqIf9"
   },
   "source": [
    "State Class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "id": "vh4gKHOnAydB"
   },
   "outputs": [],
   "source": [
    "# Globals\n",
    "START_STATE = 'T'\n",
    "GAMMA = 0.99\n",
    "ALPHA = 0.01\n",
    "EPS_START = 0.9\n",
    "EPS_END = 0.05\n",
    "EPS_DECAY = 100 # Increase to slow decay\n",
    "BATCH_SIZE = 100\n",
    "TARGET_UPDATE = 10\n",
    "IN_FEATURES = 3 + len(races2)+len(tumor_stages2)+len(tumor_grades2)\n",
    "\n",
    "# policy = DQN(2, 100, len(combos_drugs)).to(device) # input features are the state and the number of previous states\n",
    "# target_net = DQN(2, 100, len(combos_drugs)).to(device)\n",
    "policy = DQN_3(IN_FEATURES, len(combos_drugs)).to(device)\n",
    "target_net = DQN_3(IN_FEATURES, len(combos_drugs)).to(device)\n",
    "target_net.load_state_dict(policy.state_dict())\n",
    "target_net.eval()\n",
    "\n",
    "optimizer = optim.RMSprop(policy.parameters())\n",
    "memory = ReplayMemory(10000)\n",
    "\n",
    "# State class initializes the simulated patient vector\n",
    "class State:\n",
    "    def __init__(self, state=START_STATE):\n",
    "        self.state = state\n",
    "        self.age = round(np.random.normal(trans_data['age'].mean(), trans_data['age'].std(), 1)[0])\n",
    "        self.race = random.choice(races) # choose from all including 'NOT SPECIFIED'\n",
    "        self.r_arr = np.zeros(len(races2)) # sim patient vectors = 0 if 'NOT SPECIFIED'\n",
    "        if self.race == 'NOT SPECIFIED':\n",
    "            pass\n",
    "        else:\n",
    "            self.r_arr[races2.index(self.race)] = 1\n",
    "        self.tumor_stage = random.choice(tumor_stages) \n",
    "        self.ts_arr = np.zeros(len(tumor_stages2))\n",
    "        if self.tumor_stage == 'NOT SPECIFIED':\n",
    "            pass\n",
    "        else:\n",
    "            self.ts_arr[tumor_stages2.index(self.tumor_stage)] = 1\n",
    "        self.tumor_grade = random.choice(tumor_grades)\n",
    "        self.tg_arr = np.zeros(len(tumor_grades2))\n",
    "        if self.tumor_grade == 'NOT SPECIFIED':\n",
    "            pass\n",
    "        else:\n",
    "            self.tg_arr[tumor_grades2.index(self.tumor_grade)] = 1\n",
    "        \n",
    "        self.isEnd = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cdgwb3ALqMS5"
   },
   "source": [
    "Environment Class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "id": "9Rub4E2hlARd"
   },
   "outputs": [],
   "source": [
    "class CancerEnv:\n",
    "    def __init__(self):\n",
    "        self.State = State()\n",
    "        self.s_a = []\n",
    "        self.round = 0\n",
    "        self.actions = combos_drugs\n",
    "        # initial state and action reward\n",
    "        self.state_values = {'N':1, 'T':1, 'D':-1}       # not sure if need these, need action values instead. Can it be like: Combo: Value - xt?\n",
    "        self.statevals_list = [1, 1, -1]      # for [N, T, D]\n",
    "        self.s_a_values = {} \n",
    "        for i in state_set:\n",
    "            for j in combos:\n",
    "                self.s_a_values[repr([i, j])] = 0\n",
    "    \n",
    "    # Choose Action\n",
    "    def chooseAction(self):\n",
    "        # choose action with most expected value\n",
    "        mx_exp_reward = 0\n",
    "        exp_rate = EPS_END + (EPS_START - EPS_END) * math.exp(-1. * self.round / EPS_DECAY)\n",
    "        if self.State.state == 'N':\n",
    "            action = []\n",
    "        else:\n",
    "            if np.random.uniform(0, 1) <= exp_rate:\n",
    "                action = self.actions[np.random.choice(len(self.actions))]\n",
    "            else:\n",
    "                # greedy action\n",
    "                with torch.no_grad():\n",
    "                    state_t = torch.cat((torch.FloatTensor([state_set.index(self.State.state), len(self.s_a), self.State.age]),\n",
    "                                         torch.FloatTensor(self.State.r_arr),\n",
    "                                         torch.FloatTensor(self.State.ts_arr),\n",
    "                                         torch.FloatTensor(self.State.tg_arr))) #torch.FloatTensor([state_set.index(self.State.state), len(self.s_a)])\n",
    "\n",
    "                    act = policy(state_t).max(0)[1].item()\n",
    "                    action = self.actions[act]\n",
    "        return action\n",
    "\n",
    "        # Update step\n",
    "    def nxtPosition(self, action, months):\n",
    "        # Choose probabilities from transition matrix\n",
    "#         probs = get_probs(action, self.state) # static probs\n",
    "        # Get regression probs\n",
    "            # OLS\n",
    "#         probs = reg_probs(trans_data, self.State.state, months, action, self.State.age, self.State.race, \n",
    "#                           self.State.tumor_stage, self.State.tumor_grade)\n",
    "\n",
    "#         nxtState = random.choices(state_set, weights=probs, k=1)[0]\n",
    "\n",
    "            # CPH\n",
    "        surv, treat_prob = cph_probs(X_cox, self.State.state, months, action, self.State.age, self.State.r_arr, \n",
    "                                     self.State.ts_arr, self.State.tg_arr)\n",
    "        # Return next state\n",
    "        if np.random.uniform(0,1) > surv:\n",
    "            nxtState = 'D'\n",
    "        else:\n",
    "            nxtState = random.choices(['T', 'N'], weights=[treat_prob, 1-treat_prob], k=1)[0]\n",
    "        return nxtState\n",
    "\n",
    "    def takeAction(self, action):\n",
    "        position = self.nxtPosition(action, len(self.s_a))\n",
    "        return position # State(state=position)\n",
    "      \n",
    "        # Give reward: for now, simple rewards\n",
    "    def giveReward(self):\n",
    "        if self.State.state == 'N':\n",
    "            return 1 \n",
    "        elif self.State.state == 'T':\n",
    "            return 1\n",
    "        elif self.State.state == 'D':\n",
    "            return -1 \n",
    "    \n",
    "    def isEndFunc(self):\n",
    "        if self.State.state == 'D':\n",
    "            self.State.isEnd = True\n",
    "\n",
    "    def reset(self):\n",
    "        self.s_a = []\n",
    "        self.State = State()\n",
    "\n",
    "    def play(self, rounds=1000, verbose=True):\n",
    "        total_reward_list = [] \n",
    "        since = time.time()\n",
    "        while self.round < rounds:\n",
    "            # to the end of game back propagate reward\n",
    "            if self.State.isEnd:\n",
    "                # back propagate\n",
    "                reward = self.giveReward()\n",
    "                # explicitly assign end state to reward values\n",
    "                # self.s_a_values[repr([self.State.state, action])] = reward  # this is optional\n",
    "                # for s in reversed(self.s_a):\n",
    "                #     reward = self.s_a_values[repr([s[0], s[1]])] + self.lr * (reward - self.s_a_values[repr([s[0], s[1]])]) # I think i need to update this formula because using (s,a) now\n",
    "                #     self.s_a_values[repr([s[0], s[1]])] = reward\n",
    "                # Calculate total reward\n",
    "                total_reward = 0\n",
    "                for s in self.s_a:\n",
    "                    total_reward += self.state_values[s[0]]\n",
    "                if verbose:\n",
    "                    print(self.s_a)\n",
    "                    print('Total reward: ', total_reward)\n",
    "                    print('----------------------------------------------------')\n",
    "                total_reward_list.append(total_reward)\n",
    "                \n",
    "                if self.round % 1000 == 0:\n",
    "                    time_elapsed = time.time() - since\n",
    "                    print('Rounds complete: ', self.round,'/', rounds)\n",
    "                    print('Time elapsed: {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))\n",
    "\n",
    "                self.reset()\n",
    "                self.round += 1\n",
    "            else:\n",
    "                action = self.chooseAction()\n",
    "                try:\n",
    "                    act_ind = torch.tensor([combos_drugs.index(action)])\n",
    "                except ValueError:\n",
    "                    pass\n",
    "                # append trace\n",
    "                self.s_a.append([self.State.state, action])\n",
    "                if verbose:\n",
    "                    print('current state {} action {}'.format(self.State.state, action))\n",
    "                \n",
    "                s_t = torch.cat((torch.FloatTensor([state_set.index(self.State.state), len(self.s_a), self.State.age]),\n",
    "                                 torch.FloatTensor(self.State.r_arr),\n",
    "                                 torch.FloatTensor(self.State.ts_arr),\n",
    "                                 torch.FloatTensor(self.State.tg_arr))) \n",
    "                # by taking the action, it reaches the next state\n",
    "                self.State.state = self.takeAction(action)\n",
    "                # save transition vals for memory\n",
    "                s_t1 = torch.cat((torch.FloatTensor([state_set.index(self.State.state), len(self.s_a)+1, self.State.age]),\n",
    "                                  torch.FloatTensor(self.State.r_arr),\n",
    "                                  torch.FloatTensor(self.State.ts_arr),\n",
    "                                  torch.FloatTensor(self.State.tg_arr)))\n",
    "                r_t = torch.FloatTensor([self.giveReward()])\n",
    "                # store transition in memory\n",
    "                if action != []:\n",
    "                    memory.push(s_t, act_ind, s_t1, r_t)\n",
    "                else:\n",
    "                    pass\n",
    "                # mark is end\n",
    "                self.isEndFunc()\n",
    "                if verbose:\n",
    "                    print('nxt state', self.State.state)\n",
    "                    print('---------------------')\n",
    "            # perform an optimizaiton step\n",
    "            optimize_model()\n",
    "            # update target model\n",
    "            if self.round % TARGET_UPDATE == 0:\n",
    "                target_net.load_state_dict(policy.state_dict())\n",
    "        time_elapsed = time.time() - since\n",
    "        print('Rounds complete: ', rounds,'/', rounds)\n",
    "        print('Time elapsed: {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))\n",
    "        return total_reward_list"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jpBhfZoBSw10"
   },
   "source": [
    "Treat patients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 757
    },
    "id": "7o5Wl8r3Sv-4",
    "outputId": "a75a8109-3699-47dd-d7f2-4cd3a2d3831c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rounds complete:  0 / 50000\n",
      "Time elapsed: 0m 0s\n",
      "Rounds complete:  1000 / 50000\n",
      "Time elapsed: 2m 28s\n",
      "Rounds complete:  2000 / 50000\n",
      "Time elapsed: 4m 45s\n",
      "Rounds complete:  3000 / 50000\n",
      "Time elapsed: 7m 6s\n",
      "Rounds complete:  4000 / 50000\n",
      "Time elapsed: 9m 13s\n",
      "Rounds complete:  5000 / 50000\n",
      "Time elapsed: 11m 16s\n",
      "Rounds complete:  6000 / 50000\n",
      "Time elapsed: 13m 9s\n",
      "Rounds complete:  7000 / 50000\n",
      "Time elapsed: 14m 59s\n",
      "Rounds complete:  8000 / 50000\n",
      "Time elapsed: 17m 2s\n",
      "Rounds complete:  9000 / 50000\n",
      "Time elapsed: 18m 45s\n",
      "Rounds complete:  10000 / 50000\n",
      "Time elapsed: 20m 35s\n",
      "Rounds complete:  11000 / 50000\n",
      "Time elapsed: 22m 35s\n",
      "Rounds complete:  12000 / 50000\n",
      "Time elapsed: 24m 20s\n",
      "Rounds complete:  13000 / 50000\n",
      "Time elapsed: 26m 4s\n",
      "Rounds complete:  14000 / 50000\n",
      "Time elapsed: 27m 34s\n",
      "Rounds complete:  15000 / 50000\n",
      "Time elapsed: 29m 23s\n",
      "Rounds complete:  16000 / 50000\n",
      "Time elapsed: 31m 4s\n",
      "Rounds complete:  17000 / 50000\n",
      "Time elapsed: 32m 50s\n",
      "Rounds complete:  18000 / 50000\n",
      "Time elapsed: 34m 43s\n",
      "Rounds complete:  19000 / 50000\n",
      "Time elapsed: 36m 46s\n",
      "Rounds complete:  20000 / 50000\n",
      "Time elapsed: 39m 1s\n",
      "Rounds complete:  21000 / 50000\n",
      "Time elapsed: 41m 31s\n",
      "Rounds complete:  22000 / 50000\n",
      "Time elapsed: 44m 32s\n",
      "Rounds complete:  23000 / 50000\n",
      "Time elapsed: 47m 28s\n",
      "Rounds complete:  24000 / 50000\n",
      "Time elapsed: 50m 18s\n",
      "Rounds complete:  25000 / 50000\n",
      "Time elapsed: 51m 42s\n",
      "Rounds complete:  26000 / 50000\n",
      "Time elapsed: 53m 5s\n",
      "Rounds complete:  27000 / 50000\n",
      "Time elapsed: 54m 22s\n",
      "Rounds complete:  28000 / 50000\n",
      "Time elapsed: 55m 42s\n",
      "Rounds complete:  29000 / 50000\n",
      "Time elapsed: 56m 57s\n",
      "Rounds complete:  30000 / 50000\n",
      "Time elapsed: 58m 9s\n",
      "Rounds complete:  31000 / 50000\n",
      "Time elapsed: 59m 20s\n",
      "Rounds complete:  32000 / 50000\n",
      "Time elapsed: 60m 33s\n",
      "Rounds complete:  33000 / 50000\n",
      "Time elapsed: 61m 46s\n",
      "Rounds complete:  34000 / 50000\n",
      "Time elapsed: 63m 2s\n",
      "Rounds complete:  35000 / 50000\n",
      "Time elapsed: 64m 16s\n",
      "Rounds complete:  36000 / 50000\n",
      "Time elapsed: 65m 31s\n",
      "Rounds complete:  37000 / 50000\n",
      "Time elapsed: 66m 44s\n",
      "Rounds complete:  38000 / 50000\n",
      "Time elapsed: 67m 55s\n",
      "Rounds complete:  39000 / 50000\n",
      "Time elapsed: 69m 7s\n",
      "Rounds complete:  40000 / 50000\n",
      "Time elapsed: 70m 18s\n",
      "Rounds complete:  41000 / 50000\n",
      "Time elapsed: 71m 31s\n",
      "Rounds complete:  42000 / 50000\n",
      "Time elapsed: 72m 44s\n",
      "Rounds complete:  43000 / 50000\n",
      "Time elapsed: 73m 56s\n",
      "Rounds complete:  44000 / 50000\n",
      "Time elapsed: 75m 8s\n",
      "Rounds complete:  45000 / 50000\n",
      "Time elapsed: 76m 23s\n",
      "Rounds complete:  46000 / 50000\n",
      "Time elapsed: 77m 34s\n",
      "Rounds complete:  47000 / 50000\n",
      "Time elapsed: 81m 40s\n",
      "Rounds complete:  48000 / 50000\n",
      "Time elapsed: 87m 10s\n",
      "Rounds complete:  49000 / 50000\n",
      "Time elapsed: 92m 33s\n",
      "Rounds complete:  50000 / 50000\n",
      "Time elapsed: 98m 4s\n"
     ]
    }
   ],
   "source": [
    "if __name__ == '__main__':\n",
    "    env = CancerEnv()\n",
    "    reward_hist = env.play(rounds=50000, verbose=False)\n",
    "    # print(ag.s_a_values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot total rewards"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "plt.rcParams[\"font.family\"] = \"Arial\"\n",
    "plt.plot(reward_hist)\n",
    "plt.title('Total Reward Per Round', fontsize=18, fontweight='bold')\n",
    "plt.xlabel('Epoch', fontsize=14)\n",
    "plt.ylabel('Months Survival', fontsize=14)\n",
    "plt.xticks(fontsize=12)\n",
    "plt.yticks(fontsize=12)\n",
    "plt.savefig('reward.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "t4G7VixwubZe"
   },
   "source": [
    "## Scratch/notes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lhzLwojuFxnC"
   },
   "source": [
    "Notes\n",
    "* Add in \"remission\" as a state (i.e. no treatment)\n",
    "  - Calculate similar to probability of death: probability that a no-treatment period follows a treatment\n",
    "  - Then need to calculate probability following \"remission\" for each:\n",
    "    - Death\n",
    "    - remission, i.e. no treatment\n",
    "    - More treatment i.e. \"progression/recurrence\" \n",
    "  - No-treatment (\"[]\") will not be in the action set\n",
    "    - Unless already in the no-treatment state\n",
    "    - Importantly, this means that the agent cannot decide to take a patient off of treatement. Not being on chemo is not a choice made by the AI \"doctor\", but rather, is a result of the previous treatment. Agent can, however, choose not to add drugs if already achieved remission (i.e. in the no-treatment state) \n",
    "\n",
    "* \"Grid world\" verion of action space\n",
    "  - 56 dimensional object with up to 5 drugs activated at once\n",
    "  - Action set is [Start drug {n}, Stop drug {n}, Do nothing]\n",
    "    - Cannot [Stop drug {n}] if resulting state is no treatment\n",
    "    - Cannot [Start drug {n}] if resulting len(combo) > 5\n",
    "    - Cannot start a drug that is already in the combo\n",
    "  - \\> 4mm possible combinations, only 127 drug combos in data - is this necessary? \n",
    "\n",
    "* Rewards\n",
    "  - I think in the final version this should be a probability distribution returning a number of days of survival/until treatment failure after each action\n",
    "  - For starters, just try to get the average days survived when therapy doesn't result in death.\n",
    "  - Also need to think of something for when patients don't die\n",
    "    - Maybe just probability of death and a final non-death stop state?\n",
    "    - Death vs. no death in the end doesn't matter unless the no deaths get a final reward\n",
    "\n",
    "* Bellman action value equation: Q(s,a) = r + γ(max(Q(s’,a’))\n",
    "  - \"This says that the Q-value for a given state (s) and action (a) should represent the current reward (r) plus the maximum discounted (γ) future reward expected according to our own table for the next state (s’) we would end up in\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "i1T5dOpFDHsd"
   },
   "source": [
    "Intermittant reward matrix for each patient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "id": "1kse5iCiC84F"
   },
   "outputs": [],
   "source": [
    "# fix because including no drug periods and using df\n",
    "\n",
    "# tcga_rewards = {}\n",
    "# for barcode in tcga_lines_keys:\n",
    "#     rewards = []\n",
    "#     for line in range(len(tcga_drug_lines[barcode])):\n",
    "#         try:\n",
    "#             rewards.append(tcga_drug_lines[barcode][line+1][0] - tcga_drug_lines[barcode][line][0])\n",
    "#         except:\n",
    "#             rewards.append(tcga_ov_1_keep.loc[tcga_ov_1_keep['bcr_patient_barcode'] == barcode, 'total_days_overall_survival'].iloc[0] - tcga_drug_lines[barcode][line][0])\n",
    "\n",
    "#     tcga_rewards[barcode] = rewards\n",
    "\n",
    "# tcga_rewards"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "id": "se4yPE1y3pna"
   },
   "outputs": [],
   "source": [
    "# negs = []\n",
    "# for barcode in tcga_rewards:\n",
    "#     if tcga_rewards[barcode][-1] < 0:\n",
    "#         negs.append(barcode)\n",
    "\n",
    "# negs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "id": "3Rk6ADD7bzKc"
   },
   "outputs": [],
   "source": [
    "# Add in final state for each patient: [time, 'death']\n",
    "# Figure out what to do with patients where overall survival is < the end of the last therapy line"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "id": "xaT1Me5DZm5I"
   },
   "outputs": [],
   "source": [
    "# define transition matrix\n",
    "# def transition(state, action): \n",
    "#     if state,action = (living, drug A): \n",
    "#         then return 1 if np.random() < 0.8 …. \n",
    "        \n",
    "#     if state, action = (living, drug B), return 1 \n",
    "    \n",
    "#     if ..\n"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyOSHFAB03nAwKMJtubJJWPU",
   "collapsed_sections": [],
   "name": "Data_Preprocessing.ipynb",
   "provenance": []
  },
  "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}