{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook walks through a basic example of using the GPU-accelerated estimators from [RAPIDS](https://rapids.ai/) cuML and [DMLC/XGBoost](https://github.com/dmlc/xgboost) with TPOT for classification tasks. You must have access to an NVIDIA GPU and have cuML installed in your environment. Running this notebook without cuML will cause TPOT to raise a `ValueError`, indicating you should install cuML."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tpot import TPOTClassifier\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "NSAMPLES = 50000\n",
    "NFEATURES = 20\n",
    "SEED = 12\n",
    "\n",
    "# For cuML with TPOT, you must use CPU data (such as NumPy arrays)\n",
    "X, y = make_classification(\n",
    "    n_samples=NSAMPLES,\n",
    "    n_features=NFEATURES,\n",
    "    n_informative=NFEATURES,\n",
    "    n_redundant=0,\n",
    "    class_sep=0.55,\n",
    "    n_classes=2,\n",
    "    random_state=SEED,\n",
    "    \n",
    ")\n",
    "\n",
    "X = X.astype(\"float32\")\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y, random_state=SEED)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that for cuML to work correctly, you must set `n_jobs=1` (the default setting)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Optimization Progress', max=30.0, style=ProgressStyle(des…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Generation 1 - Current best internal CV score: 0.9695733333333334\n",
      "Generation 2 - Current best internal CV score: 0.9695733333333334\n",
      "Generation 3 - Current best internal CV score: 0.9695733333333334\n",
      "Generation 4 - Current best internal CV score: 0.9705333333333334\n",
      "Generation 5 - Current best internal CV score: 0.9705333333333334\n",
      "Best pipeline: KNeighborsClassifier(input_matrix, n_neighbors=20, weights=uniform)\n",
      "0.97704\n"
     ]
    }
   ],
   "source": [
    "# TPOT setup\n",
    "GENERATIONS = 5\n",
    "POP_SIZE = 100\n",
    "CV = 5\n",
    "\n",
    "tpot = TPOTClassifier(\n",
    "    generations=GENERATIONS,\n",
    "    population_size=POP_SIZE,\n",
    "    random_state=SEED,\n",
    "    config_dict=\"TPOT cuML\",\n",
    "    n_jobs=1, # cuML requires n_jobs=1, the default\n",
    "    cv=CV,\n",
    "    verbosity=2,\n",
    ")\n",
    "\n",
    "tpot.fit(X_train, y_train)\n",
    "\n",
    "preds = tpot.predict(X_test)\n",
    "print(accuracy_score(y_test, preds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "import numpy as np\n",
      "import pandas as pd\n",
      "from cuml.neighbors import KNeighborsClassifier\n",
      "from sklearn.model_selection import train_test_split\n",
      "\n",
      "# NOTE: Make sure that the outcome column is labeled 'target' in the data file\n",
      "tpot_data = pd.read_csv('PATH/TO/DATA/FILE', sep='COLUMN_SEPARATOR', dtype=np.float64)\n",
      "features = tpot_data.drop('target', axis=1)\n",
      "training_features, testing_features, training_target, testing_target = \\\n",
      "            train_test_split(features, tpot_data['target'], random_state=12)\n",
      "\n",
      "# Average CV score on the training set was: 0.9705333333333334\n",
      "exported_pipeline = KNeighborsClassifier(n_neighbors=20, weights=\"uniform\")\n",
      "# Fix random state in exported estimator\n",
      "if hasattr(exported_pipeline, 'random_state'):\n",
      "    setattr(exported_pipeline, 'random_state', 12)\n",
      "\n",
      "exported_pipeline.fit(training_features, training_target)\n",
      "results = exported_pipeline.predict(testing_features)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "tpot.export('tpot_classification_cuml_pipeline.py')\n",
    "print(tpot.export())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}