test_mlflow_sklearn.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MLFLOW для экспериментов с Sklearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import mlflow\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "\n",
    "# Загрузка данных Титаника из csv-файла\n",
    "data = pd.read_csv('titanic.csv')\n",
    "data = data [['Survived','Pclass','Age','Fare']]\n",
    "data = data.dropna(subset=['Age'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 714 entries, 0 to 890\n",
      "Data columns (total 4 columns):\n",
      " #   Column    Non-Null Count  Dtype  \n",
      "---  ------    --------------  -----  \n",
      " 0   Survived  714 non-null    int64  \n",
      " 1   Pclass    714 non-null    int64  \n",
      " 2   Age       714 non-null    float64\n",
      " 3   Fare      714 non-null    float64\n",
      "dtypes: float64(2), int64(2)\n",
      "memory usage: 27.9 KB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Укажем юзера который делает эксперименты\n",
    "os.environ['USER'] = 'Dima Kolesnikov'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Experiment: artifact_location='mlflow-artifacts:/1', creation_time=1732645400278, experiment_id='1', last_update_time=1732645400278, lifecycle_stage='active', name='TreeClassifier_test', tags={}>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Разделение данных на обучающую и тестовую выборки\n",
    "train, test = train_test_split(data, test_size=0.2)\n",
    "\n",
    "# Инициализация MLflow\n",
    "mlflow.set_tracking_uri(\"http://localhost:5050\")\n",
    "mlflow.set_experiment(\"TreeClassifier_test\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import itertools\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    Функция для построения матрицы ошибок.\n",
    "    cm - матрица ошибок\n",
    "    classes - список классов\n",
    "    normalize - если True, то значения матрицы ошибок нормализуются к 1\n",
    "    title - заголовок графика\n",
    "    cmap - цветовая схема для отображения графика\n",
    "    \"\"\"\n",
    "\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "        \n",
    "    plt.figure()\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Запуск контекста MLflow и сохранение гиперпараметров\n",
    "def experiment(run_name, max_depth, min_samples_split):\n",
    "    with mlflow.start_run(run_name=run_name):\n",
    "\n",
    "        mlflow.log_param(\"max_depth\", max_depth)\n",
    "        mlflow.log_param(\"min_samples_split\", min_samples_split)\n",
    "\n",
    "        # Создание и обучение модели решающего дерева\n",
    "        model = DecisionTreeClassifier(max_depth=max_depth, min_samples_split=min_samples_split)\n",
    "        model.fit(train.drop('Survived', axis=1), train['Survived'])\n",
    "\n",
    "        # Вычисление метрик и сохранение их в MLflow\n",
    "        preds = model.predict(test.drop('Survived', axis=1))\n",
    "        acc = accuracy_score(test['Survived'], preds)\n",
    "        cm = confusion_matrix(test['Survived'], preds)\n",
    "\n",
    "        mlflow.log_metric(\"accuracy\", acc)\n",
    "        mlflow.log_metric(\"tn\", cm[0][0])\n",
    "        mlflow.log_metric(\"fp\", cm[0][1])\n",
    "        mlflow.log_metric(\"fn\", cm[1][0])\n",
    "        mlflow.log_metric(\"tp\", cm[1][1])\n",
    "\n",
    "        # Визуализация матрицы ошибок и сохранение ее в MLflow\n",
    "        plot_confusion_matrix(cm, classes=['Not Survived', 'Survived'])\n",
    "        cm_fig = plt.gcf()\n",
    "        cm_fig.savefig('matrix.png')\n",
    "\n",
    "        mlflow.log_artifact(\"matrix.png\")\n",
    "\n",
    "        # Сохранение модели в MLflow\n",
    "        mlflow.sklearn.log_model(model, \"model\")   \n",
    "    mlflow.end_run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024/11/26 21:26:22 WARNING mlflow.models.model: Model logged without a signature and input example. Please set `input_example` parameter when logging the model to auto infer the model signature.\n",
      "2024/11/26 21:26:23 INFO mlflow.tracking._tracking_service.client: 🏃 View run DecisionTreeClassifier_run at: http://localhost:5050/#/experiments/1/runs/5350787e9e5e4501bc56d2bcd4799d0a.\n",
      "2024/11/26 21:26:23 INFO mlflow.tracking._tracking_service.client: 🧪 View experiment at: http://localhost:5050/#/experiments/1.\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Определение гиперпараметров модели\n",
    "max_depth = 10\n",
    "min_samples_split = 25\n",
    "run_name = 'DecisionTreeClassifier_run'\n",
    "\n",
    "experiment(run_name, max_depth, min_samples_split)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "patched_yolo_infer",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.8"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}