ML Project.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "79d7eae7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: category-encoders in d:\\anaconda\\lib\\site-packages (2.5.0)\n",
      "Requirement already satisfied: pandas>=1.0.5 in d:\\anaconda\\lib\\site-packages (from category-encoders) (1.3.4)\n",
      "Requirement already satisfied: statsmodels>=0.9.0 in d:\\anaconda\\lib\\site-packages (from category-encoders) (0.12.2)\n",
      "Requirement already satisfied: numpy>=1.14.0 in d:\\anaconda\\lib\\site-packages (from category-encoders) (1.20.3)\n",
      "Requirement already satisfied: scikit-learn>=0.20.0 in d:\\anaconda\\lib\\site-packages (from category-encoders) (0.24.2)\n",
      "Requirement already satisfied: patsy>=0.5.1 in d:\\anaconda\\lib\\site-packages (from category-encoders) (0.5.2)\n",
      "Requirement already satisfied: scipy>=1.0.0 in d:\\anaconda\\lib\\site-packages (from category-encoders) (1.7.1)\n",
      "Requirement already satisfied: pytz>=2017.3 in d:\\anaconda\\lib\\site-packages (from pandas>=1.0.5->category-encoders) (2021.3)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in d:\\anaconda\\lib\\site-packages (from pandas>=1.0.5->category-encoders) (2.8.2)\n",
      "Requirement already satisfied: six in d:\\anaconda\\lib\\site-packages (from patsy>=0.5.1->category-encoders) (1.16.0)\n",
      "Requirement already satisfied: joblib>=0.11 in d:\\anaconda\\lib\\site-packages (from scikit-learn>=0.20.0->category-encoders) (1.1.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in d:\\anaconda\\lib\\site-packages (from scikit-learn>=0.20.0->category-encoders) (2.2.0)\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install category-encoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1b5f4aee",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Import libraries & packages\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "import category_encoders as ce\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "abf6ceb7",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Import dataset\n",
    "df = pd.read_csv('weatherAUS.csv')\n",
    "#Drop NaN values on RainTomorrow variable\n",
    "df.dropna(subset = ['RainTomorrow'], inplace=True)\n",
    "#Feature Engineering\n",
    "df['Date'] = pd.to_datetime(df['Date'])\n",
    "df['Year'] = df['Date'].dt.year\n",
    "df['Month'] = df['Date'].dt.month\n",
    "df['Day'] = df['Date'].dt.day\n",
    "df.drop('Date',inplace= True,axis=1)\n",
    "\n",
    "#Drop very strong positive correlation\n",
    "df.drop(['Temp9am','Temp3pm','Pressure9am'],inplace= True,axis=1)\n",
    "#Drop categorical variables (have practically no impact to the model performance as seen in EDA)\n",
    "df.drop(['Location','WindGustDir','WindDir9am','WindDir3pm'],inplace= True,axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9b5a705a",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Extract independent and dependent variables\n",
    "X = df.drop(['RainTomorrow'], axis=1)\n",
    "y = df['RainTomorrow']\n",
    "#Split training and test dataset\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d488fe5f",
   "metadata": {},
   "outputs": [],
   "source": [
    "numerical = [col for col in X_train.columns if X_train[col].dtypes != 'O']\n",
    "#replace missing values in X_train and X_test with respective column median in X_train\n",
    "for df1 in [X_train, X_test]:\n",
    "    for col in numerical:\n",
    "        col_median=X_train[col].median()\n",
    "        df1[col].fillna(col_median, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "174e2634",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Engineering outliers in numerical variables\n",
    "def max_value(df2, variable, top):\n",
    "    return np.where(df2[variable]>top, top, df2[variable])\n",
    "\n",
    "for df2 in [X_train, X_test]:\n",
    "    df2['Rainfall'] = max_value(df2, 'Rainfall', 2.0)\n",
    "    df2['Evaporation'] = max_value(df2, 'Evaporation', 14.6)\n",
    "    df2['WindSpeed9am'] = max_value(df2, 'WindSpeed9am', 37.0)\n",
    "    df2['WindSpeed3pm'] = max_value(df2, 'WindSpeed3pm', 40.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2084639a",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_copy = X_train.copy()\n",
    "# encode RainToday variable using category encoders because RainToday values are binary\n",
    "encoder = ce.BinaryEncoder(cols=['RainToday'])\n",
    "X_train = encoder.fit_transform(X_train)\n",
    "X_test = encoder.transform(X_test)\n",
    "#Creating X_train dataset that has been encoded using pd.get_dummies\n",
    "X_train = pd.concat([X_train[numerical], X_train[['RainToday_0', 'RainToday_1']]], axis=1)\n",
    "#Creating X_test dataset that has been encoded using pd.get_dummies\n",
    "X_test = pd.concat([X_test[numerical], X_test[['RainToday_0', 'RainToday_1']]], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "55f88eea",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Feature Scaling\n",
    "scaler = MinMaxScaler()\n",
    "col = X_train.columns\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)\n",
    "X_train = pd.DataFrame(X_train, columns=[col])\n",
    "X_test = pd.DataFrame(X_test, columns=[col])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "5e85e263",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Modelling Random Forest Classifier\n",
    "model_rf = RandomForestClassifier(random_state=0)\n",
    "model_rf.fit(X_train, y_train)\n",
    "y_pred = model_rf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "49c9c70e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def check_model_performance():\n",
    "  #Check confusion matrix\n",
    "  cm = confusion_matrix(y_test, y_pred)\n",
    "  print('Confusion matrix\\n\\n', cm)\n",
    "  print('\\nTrue Positives(TP) = ', cm[0,0])\n",
    "  print('\\nTrue Negatives(TN) = ', cm[1,1])\n",
    "  print('\\nFalse Positives(FP) = ', cm[0,1])\n",
    "  print('\\nFalse Negatives(FN) = ', cm[1,0])\n",
    "\n",
    "  #Check accuracy\n",
    "\n",
    "  print('Model accuracy score: {0:0.4f}'. format(accuracy_score(y_test, y_pred)))\n",
    "  #Check ROC AUC\n",
    "  ROC_AUC_rf = roc_auc_score(y_test, model_rf.predict_proba(X_test)[:,1])\n",
    "  print('ROC AUC RF: {:.4f}'.format(ROC_AUC_rf))\n",
    "  #Check classification report\n",
    "  print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6678696c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix\n",
      "\n",
      " [[21106   961]\n",
      " [ 3089  3283]]\n",
      "\n",
      "True Positives(TP) =  21106\n",
      "\n",
      "True Negatives(TN) =  3283\n",
      "\n",
      "False Positives(FP) =  961\n",
      "\n",
      "False Negatives(FN) =  3089\n",
      "Model accuracy score: 0.8576\n",
      "ROC AUC RF: 0.8867\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          No       0.87      0.96      0.91     22067\n",
      "         Yes       0.77      0.52      0.62      6372\n",
      "\n",
      "    accuracy                           0.86     28439\n",
      "   macro avg       0.82      0.74      0.77     28439\n",
      "weighted avg       0.85      0.86      0.85     28439\n",
      "\n"
     ]
    }
   ],
   "source": [
    "check_model_performance()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4f227c4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocess_input():\n",
    "  input = {\n",
    "    'MinTemp': min_t, \n",
    "    'MaxTemp': max_t, \n",
    "    'Rainfall': rainfall, \n",
    "    'Evaporation': evap, \n",
    "    'Sunshine': sun,\n",
    "    'WindGustSpeed': wind_gust_speed,\n",
    "    'WindSpeed9am': wind_speed_9, \n",
    "    'WindSpeed3pm': wind_speed_3, \n",
    "    'Humidity9am': humid_9, \n",
    "    'Humidity3pm': humid_3, \n",
    "    'Pressure3pm': pres_3, \n",
    "    'Cloud9am': cloud_9, \n",
    "    'Cloud3pm': cloud_3, \n",
    "    'RainToday': rain_today, \n",
    "    'Year': year, \n",
    "    'Month': month, \n",
    "    'Day': day\n",
    "    }\n",
    "  input_arr = []\n",
    "  input_arr.append(input)\n",
    "  X_input = pd.DataFrame(input_arr)\n",
    "  #Preprocessing\n",
    "  # encode RainToday variable using category encoders because RainToday values are binary\n",
    "  encoder = ce.BinaryEncoder(cols=['RainToday'])\n",
    "  func_X_train_copy = encoder.fit_transform(X_train_copy)\n",
    "  X_input = encoder.transform(X_input)\n",
    "  #Creating X_test dataset that has been encoded using pd.get_dummies\n",
    "  X_input = pd.concat([X_input[numerical], X_input[['RainToday_0', 'RainToday_1']]], axis=1)\n",
    "  return X_input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "31a081d8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAEQCAYAAACk818iAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAABBXklEQVR4nO3dd5xU1fnH8c93l6YCIqCoSBEroqJYsMdEjdFoRCH2HsUSTYz+YjQCEjW2RGOMBbGhsQtEY+8NG2IPCihIEwRFQXp9fn+cM3AZZmfv7s622efN67525pxbnsvePXPm3HPPkZnhnHOueJTUdgDOOecKywt255wrMl6wO+dckfGC3TnniowX7M45V2S8YHfOuSLjBXsBSNpXkkkaWMX9nBz3c3JhIqtdkjaUdI+kqZKWx3NrVc3HLMjvwhXf9diQ1MuCPV5sJmmFpM3yrPdKYt2TazDEGiWpRFIfScMkTZG0SNJ8SZ9LGixpz1oKbQhwAvAacAXwF2BRLcVSaxIfNibpK0k5/+4kNZf0Y2LdzlU87pBC7MfVP41qO4AqWEaI/zfAn7MzJW0B/CSxXlGStCEwFNgTmAu8AIwHBGwBHAOcLul3ZvavGoyrCXAA8KKZHVdTxwVGAl2B72rwmGktAzoD+wPP58g/GmhB3blm/wO8A0yv7UBcxdSFi6eyZhAuuFMkDTCzZVn5pxEKtyeBXjUcW42QtDbwLNAdeAg428x+yFqnJfB/QMsaDm9DwjfCaTV5UDNbAIypyWNWwIvAT4HTyV2wn064picDPWswrpzMbA4wp7bjcBVXL5tiEm4nFCCHJBMlNQZOAt4CRpe1saQtJN0r6WtJSyRNi++3KGP9dpLulDRD0kJJH0k6KV+AklpLuio2iyyUNEfSS5J+XuGzXdMfCIX6m8Bx2YU6gJn9aGYDgL9nxbVujGtsbLr5QdJzkvbPcQ4r260l7SDpKUmzJS2Q9JqkPbLWnwhMim9PSjQtDIn5A+P7fXMcq3Ny3UR6O0l/j/HOj8cfG5sbuuSKNce+U/++kzHGZq6R8Xy/l/SQpPbZ26QwCxgOHCZp/azjbQ/sCtxNqLGvQVIvSfdJGhf/D+ZJel/S77KbdyQZ4W8A4KvE72BiYp1XY1oTSQPi/+fixO9pjTZ2SefHtGE54ttf4V7Kp5LWqvh/jyuU+lxjB3gQuJ5QO38skf4roB1wEbB5rg0l7UKoQbUA/gt8BmwNHEf4w9vPzEYl1m9D+KDoAoyIy0bAIHLXvpDUCXiV8PX7DULteh3CB9Gzks4ws9srfNar9I0/LzezFflWNLPFibhaET4MtgHeA24A2gJHAs9LOsvMbsuxm52BC4G3gTuAjkBv4CVJO5jZ2LjeDYRz/j3wMat+Nx9V4NxWit9M3gQ2IzQ1PUH4NtYJOIzQFDWhnH1U6PedcDbhevov4V5BT+AooHs858U5tsnndkLz2Ems/mF7OmDAnYQmxFyuBlYA7wJfA+sCPwP+CexCuJ+R8RfCN9XuMX92TJ/NmobF7Z8h/K5mlhW8mV0fP5CPkHS2md0CK5sE7yPcQznKzBaWtQ9XA8ys3i2EP4Cp8fUdhBrOJon8ZwlfIdcm3LQz4OREvoDPY/pxWfs+KqaPAUoS6YNj+j+y1t8ZWBrzBmblvUr4Qzw6K70VoZBbCLRLpJ+cHWue/4MOcd2lQLMK/v/dFre9DVAifYv4/7YY6JxI3zeuv0ZswBkx/Zas9M4xfUiO4w+MefvmyFtjO+DQXP/3Ma8J0CJHrAMTaZX5fWdi/BHYLmubB2LekSn/vzMx3Rdj+QIYk8hfC/gBeCG+HxHX75y1n81y7LsEuCeu3zMrb0iu/WRdnwZ8ArTNkZ/zegTaAFPi9ds9xvBiXPeUqv59+1L1pb43xUCoAZUCp8LKWvIBwP0W2ltz2YNQW3vbzO5PZpjZw4Q/rK2AveI+GxNqdnMJf/DJ9UcBq+0jbtOdUPMaZmYPZW0zG7gUaEao8VbGRvHnLDNL3dMknsvxwDzgYot/qTGuL4AbCYXliTk2f9PMhmSl3UX4YN01feiVtkYt0MyWmNnccrar0O87y41m9mlWWuZbVoXPOf5/3wFsJWmfmPxrwod93m9vZjY+R9oKQo0c4MCKxhP1N7PUN5vNbBbhW0dj4GFC5Wk/wt/c3ZWMwRVQvS/Yzexd4FPg1NjOeBrhvPL9kfSIP18uIz+TvmP8uTWh9v+RhRtK2V7NkbZ7/LlubK9dbQEybexd88SZj+LPio67nDmXj83s+xz52eeetEZThZktJdzIXq+CcVTEa4Smh4skPRvblHeSVJpy+4r+vpNyNc9MiT8re85DCN+0To/vTyf04nks30aS2ki6WtInsX3dYlv6+3GVyrT7Q+hJVCFmNoJQOdkKuJjwLeTMSh7fFVh9b2PPuJ1Q0/wFcArwvpl9mGf9dePPsrpxZdJbZa0/o4z1v8mR1ib+PCAuZWmeJy+fTG+TtpKaVaDWXtFzT5pdxjbLCN+aqoWZ/ShpN0K78a9YVTP9TtItwBXxA6YshT7nzM3NSp2zmc2Q9ATQO8a/F3CdmS0pa5t4X+Q9YFNCQXwv8H2MpRXhfkbTysRD7us3jeHAZYSK1B1mNq+S+3EFVu9r7NG/CV/TbyPUWgaXs36m1r1hGfkbZa2X+dmujPVz7Sezze/NTHmWU8qJNSczm0LoFtcI2Kec1XPFlfbcq0PmRm+uikWrXBuY2VQz+w2wAbAt8DtCL5MBccmnLpxztsGEtvVH4vvybqKfRijU/2JmPc3sbDPrZ2YDCc0hlZZsjktLUjNC5wUI9wcGSNqqKnG4wimKgj22WQ8FNgHms+qCK0umNr9vGfmZ9A/izzHAAmAHSevmWT/pnfhz73JiqYrMB1i/7O5u2SRlanNjWXUuuZoSfhp/fpAjr1Ay3TI75MjbOd+GFoy28LBV5ptQr3KOV9Hfd014gdAldBPgdVvVo6gsmd5da3QzpOxeNMvjz+r4NnU94cbpVYQHq9YGHo4FvqtlRVGwR/2Aw4EDU9xMe5NQwO0lqU8yI77fBxhHuKmWaUe+n9BVbmDW+jsTbqyuJt5UfYPQLezUXEFI2k7SBuWeWdn+QehOuDdwr3KMw6LwmPoAwkNKxK/79xOagC7LWnczQk14KeFbUHXJtOmeImllrV1SB3LUviVtq9yPxWe+QZV1kzyjQr/vmhBveh5BuGb7lrM6wMT4c99koqQdCW3cucyKPztWPMKySeoNnEX4f73UzJ4HriUU9NcX8liucoqljR0zm0xomkizrik8WPQCoZbxOKFWvhWh9jcXONFW7xv+Z8Kd//NiYZ7px34U8DSh7TfbsYQbc3dK+h2h//FsQi1te0KTwu7k6TdcznkskPQLwreV44BDJb0AfEn40N48xtwSOCex6UWED4NzYv/uV1jVj70FcI6ZfVWZmFLG/a6k1wkF6khJLxMK6UOB51izJr8/cL2ktwi/p5mE/8PDCM06fyvneJX5fVc7M/uA9N8S7gX+CNwg6aeEm5VbEJ6JGE64DrO9FLe5XdJQQk+o2WZ2U2Vjjh+wdxC+dR1rZplvBf0Iv8+zJL1kZrm+WbiaUtv9LSuzkOjHnmLdNfqxJ/K2ItRMpxNqqdMJfY23KmNfGxK6931LaNP/iNDXd19y9GOP27QgfCi8T/jDWgh8BTxFqKmtk1j35LJiLeccSwhd5oYDUwkPiWQerb8D2CPHNq2AawgFxGLCB84LwM9zrFvm+cX8icDErLTOlNGPPXH82wmF9GLgf/H/Y43tCD2Hrif0UPk2rj+R8IG2R9pYK/L7poJ97cv5/WRiui/l+mX1Y9+G8KDUTEKT4/uEtvcy4wHOJ/ThXxzXmZjIe5XYxF5GHKtdj4Tuje/EtCNyrN+JUODPBjat6t+5L5VfFH8hzjnnikQxtbE755zDC3bnnCs6XrA751yR8YLdOeeKTI12d1yr4zF+p9atYeHkv9R2CK5O2lLlr5NfRcqchZMfrPLx6oqi6cfunHPZynkgu2h5we6cK1pqoK3NXrA754qW19idc67IeMHunHNFJv1cLMWl0gV7HAZ2gZk1zP8551yd5zX2HCTlG+6zGaumZ3POuTrHC/bcJhJGciurAPd+6c65Oquh9oop76ynE2Z4b5xjaVG9oTnnXNVIJamX8velppLulDRJ0lxJH0o6KJG/n6QxkhZIekVSp0SeJF0jaVZcrpWkRH7nuM2CuI/9s459bDzufEmPSWqdL9byzmYUsKOZLc9eCJPoelOMc67OKmTBTmjhmEKYinBdoD/wSCyU2xLmQ+gPtCaUncm5aPsSJnXpTphk5xDgjET+g4QpHNsAlwBDJa0fzkHdCPM5n0CYkGYBcEt5geZzBqsmHl6NmS3Gx5pxztVhJQXsFWNm81l9aswnJX0F7EQokEeb2aMAkgYC30na2szGACcB15nZ1Jh/HXA6MEjSlkAPwiQ3C4Fhks4DegODCLOjPWFmr8dt+wOfS2phZUwDmrdgNrNvzKxS07Y551xtq0iNXVJfSaMSS965aCW1A7YERgPdCPMPAys/BMbHdLLz4+tk3oSsQjo7P7nv8cCSeOycUnd3lLRJ5tMmX5pzztUVFekVY2aDgcHp9qvGhEnh7zGzMZKaE6ZtTJrDqnuRzeP7ZF7z2M6enZfJb1/Gttn7XkNF+rF/Qmg7SvqMMFGyc87VOdXR3VFhp/8m1Jozk8TPY82ysCVhovRc+S2BeWZmkiq6bXb+Gipy1t1zpHXLkeacc3VESQWW8sUa9p2Em5i9zWxpzBpNooyUtA6wWUxfIz++TuZ1kdQiT35y312ApsC4suJMXbCb2ZQ0ac45V1eUlDRKvaR0K9AVODTe6Mz4D7CtpN6SmgEDgE/ijVOAe4HzJbWXtDFwATAEwMzGAR8Bl0pqJulwQs+ZYXHb+4FDJe0dPzAuA4aXdeMUUhbssQ/m6ZJelvRJTNtH0pFptnfOudogSlIv5e4r9Es/A9gB+EbSvLgcZ2bfEnqx/BX4AegJHJ3Y/DbgCeBT4H/AUzEt42hg57jt1UCfuE/MbDRwJqGAn0loWz87X6xpP6YuAw4AbiB0vwGYCvwDeCTlPpxzrkYVso3dzCaR59kdM3sR2LqMPAMujEuu/InAvnn2/QDwQNpY0xbsJxMeVPpO0q0x7SugS9oDOedcTUs83NmgpC3YSwl3ZmHV+DDNE2nOOVfnNNRBwNKe9dPA9XGo3syd4csJbUbOOVcnFbKNvT5JW2M/n3BXdw5hALB5wPPAidUUl3POVVkFersUlVRnbWY/Ar3iI7QdgSlm9k21Ruacc1VUbDXxtFIV7JJuAO43s/eAGdUakXPOFUoDbWNP+z1FwOOS5hO63DxgZmOrLyznnKs6v3mah5n9HtiE0Cm+A/COpPclnV+dwTnnXFVISr0Uk4oMKbDCzF4ws1OBbYFZwN+qLTLnnKuihtorJvXZSGou6XhJTxEGn1lGGDzeOefqJJWUpl6KSdqbp48CBwEfEKZwOsnMvqvOwJxzrsqKqyKeWtqbp6OAC8xscnUG45xzBVVkbedppe3Hfk11B+KccwXnBfvqJH1uZl3j6ymsGiNmNWbWsZpic865qvGmmDWcnnh9fHUH4pxzhWYlXmNfjZmNSLzdwMwezV5HUp9qico55wqhgRbsab+o3FlGeqoZvZ1zrlZI6Zcikrdgl9QlTpxaImnTzPu47A8sqpkwnXOuElSBpbxdSedIGiVpsaQhifTjEtPkzZO0QJJJ2inmD5S0NGudLontO0t6JW43JpatyeMeK2mSpPmSHpPUurxYy6uxfwl8AawNjI/vM8u9wMDy/zucc66WlCj9Ur5pwBXAXclEM7vfzJpnFsLQKxMIz/1kPJxcx8wmJPIeBD4E2gCXAEMlrQ8gqRthbtQTgHbAAuCW8gLN293RzErizl8zs5+UtzPnnKtTCtjEYmbDwy61M2HsrLKcBNwb5znNS9KWQA/g52a2EBgm6TzCxNiDgOOAJ8zs9bh+f+BzSS3MbG5Z+007CJgX6s65+qdUqRdJfWNTS2bpW9HDSeoE7ENo0Ug6VNL3kkZLOiuR3g2YkFVIfxzTM/kfZzLMbDywBNgyXxz5+rE/a2a/iK/foOx+7PvkO4BzztWaClTYzWwwVe8QciLwhpl9lUh7JO53BtCTUCufbWYPEuaOnpO1jzlA+/i6rPwW+YLI1xST/MS5I99OnHOuLrKa7+1yInDlajGYfZZ4+5akfwJ9CG3r84CWWftoCWRq8OXl55SvH/sDidf35NuJc87VSTXYj13SnsDGwNByVjVWfZcYDXTJajPvTpjQKJPfPXGMLkBTwgi7ZUrVxi7pGEmZ4QW2kvSapJclbZ1me+ecqxWF7e7YSFIzoBQoldRMUrJyfBIwLPumpqTDJK2nYFfgd8DjAGY2DvgIuDTu73Bge2BY3Px+Qvv83pLWAS4Dhue7cQrpH1C6Avg+vv478B7wOim63TjnXK0p7ANK/YCFwEWEYVYWxjRigX8kkKt142hCF/G5hCbua7JaQY4GdgZ+AK4G+pjZtwBmNho4k1DAzyS0rZ9dXqBph+1d38xmxOD3IrQPLQV8THbnXN1VWtDujgMp49kdM1sEtCoj75hy9jsR2DdP/gOsappJJW3B/q2kzYHtgPfMbLGktanQPWfnnKthRTZUQFppC/bLgfeB5cBRMW0/Ev0rnXOuzvGCvWxmNkTSI/H1gpj8LqFtyDnn6iYfj71cTQl3Z9sDXwNPmtn35WzjnHO1p4HW2NN2d9ydMAjYmYSuOGcAX8b0BqVJk0bcem1fxr51IzM/u4u3n76Kn+8bupk2blzKA4POY8ybN7Jw8oPsvVvX1bbdZ/dtePahfnzzvzsZ8+aNa+y74yZtefahfswaO4SPXv47P91r29Xy27ZuwZAbz2H6p3cw7dPbufufv62+E3UFNX78FE488RJ22ukoDjigLy+88PbKvIULFzFw4C307HksO+10FMcdd9HKvHfe+YQTTvgzO+10FD/72W9qI/R6zUqVeikmab+o3ACcbWZ7mNkxZrYncBawZulU5BqVljJ1+iwOOPIy2nX7DZdd9wj33fJ7Om7SFoC33hvLqefdzPSZP6yx7YIFi7nnkVf585X359z3vf86l49HT6T99qcz8G+P8MCt59G29aonhx8afD4zvp3NVnucS8cdz+SGwU9Vz0m6glq2bDlnn30FP/3pLowc+QCXXfZb/vjH6/jqq68B6N//ZubMmcczz9zKyJEPcPHFp63cdu21m9G79/5ceOEptRV+/ebjsee1JWG8g6ShwOaFDafuW7BwMX/9xzAmT/0OM+OZlz5k4pRv6bFdF5YuXc5Ndz7DW++NZcXyFWtsO+rj8Tw4fARfTZ65Rt7mm27IDttuyuXXD2XR4qU89sxIRo+dQq+DewKw397bsclGbbj4r/fz49yFLFu2nI9HT6zu03UFMGHCVGbO/J6TTz6M0tJSdt+9Oz16dOXxx19hwoSpvPzyu1x++Tm0br0upaWlbLvtqj+r7bffkl69fkaHDhvW4hnUYwV8QKk+SVuwf8GaN0p/TWieadA2aLsuW2y6IZ+Nm1ql/Wyz5SZ8NXkm8+avmrvk088msc0WYXTQXXtswbgJ07jj+rOZ+vFgRjxxBXv17FrW7lwdkmv0VjP44otJfPLJONq334Abb7yfnj2P5dBDz+G5596shSiLVGHHY6830hbs5wE3SXpH0sOS3iU8dfq7aousHmjUqJS7b/wt9w17nXHjp1VpX+us04wf5y5YLW3O3AU0b94MgPYbteaAn3TntbdH03mns/jn4Kd49I4LaLNe3kHeXB3QpcsmtG69LnfcMZylS5cxYsQHvPfe/1i0aDHffPMd48ZNokWLdXjjjXvo3/8MLrroBsaPn1LbYRcHb4opm5m9BWwG3EToz/4vYPOYnldyjONl876sUrB1iSTuuuFslixZxh/6D6ny/ubPX0SLFmutltay+drMmxdq8IsWLWHi5Jnc8/CrLFu2nEefeJup02ex+855h2V2dUDjxo24+eZLeO2199hrrxO5++7H+MUv9qJdu7Y0a9aExo0bcdZZR9GkSWN23XU7evbcjhEjPqztsItDA22KydvdUVI74B/AtoRpni4ws1kVOUByjOO1Oh5T7owi9cWgv/Vlg7br0uuka1i2bHmV9/fZuKls2mEDmq/TbGVzzHbbdOThx8Nn56efT+bg/XtU+Tiudmy99abcd9/VK98fffQf6dXrZ3TqtHEtRtUANGqYHdnLO+tbCPPsDQI6EAr5Bu/GK3/D1pu3p/epf2PR4qWr5TVp0oimTRuH141XvYZQy2/atDGNG5UiEV43LgXgy6++4ZPPJnHJeb1p2rQxvzpwZ7bduiOPPf0uAP999j1arbsOx/XZh5IScfjBu7Lxhq15e1Te0TtdHTFmzFcsXryEhQsXceedw5k583uOOGJ/dt65GxtttD633fYoy5Yt5/33P+Pdd//HXnvtCMCKFStYvHgJS5cuw8xYvHgJS5YsLedoLsOUfikmyjctn6SZwJZmNltSW+BDM+tQ2YMVQ429Y/u2jH37XyxatIRliZ4v5158Bw899iZj3ryRTh3WX22brfY4l8lTv2Pv3bry/CMDVst7/e3POPCoy8O+N2nL7dedxS47bs6Ur7/jvP5388qI/61cd89dt+KGK06lc4cNGDt+Gn+67F7eHDm2Gs+2Ziyc/JfaDqHaXXPNXQwd+jzLli1np522oX//M1bW1r/4YhL9+v2LsWMnsvHGG/CHP5zAAQeER0TeffdTTjzxz6vta9ddt+Xf/76qxs+h5m1Z5eK2S9+hqcucCYP7FE3xXl7B/qOZtUy8/97MWlf2YMVQsLvCawgFu6uMAhTsZwxLX7Df1rtoCvbyhhRoLOkUVt1aaCrp1OQKZnZXtUTmnHNVVWTdGNMqr2B/lzCHX8ZI4ITEewO8YHfO1U0N895p/oLdzPatoTicc67wShtmyd4wz9o51yCYlHopj6Rz4jM5iyUNSaR3lmSS5iWW/ol8SbpG0qy4XCutOmDc/hVJCySNkbR/1nGPlTRJ0nxJj0kq9z6nF+zOueJVUoGlfNMI8z+X1fzcysyax+XyRHpfoBfQnTA67iGEEXIzHgQ+BNoAlwBDJa0PIKkbcBuhCbwdsIAUc017we6cK14FHCvGzIab2WNAhR7SBE4CrjOzqWb2NXAdcDKApC2BHsClZrbQzIYBnwK947bHAU+Y2etmNg/oDxwhKe9YIl6wO+eKVwXGikkOfxKXvhU82iRJUyXdHZ/7yejG6tOIfhzTMnkTzGxunvyV25rZeGAJYcTdMqWdaCPnTEnxASbnnKubSpV6MbPBZrZzYhmc8ijfAbsAnYCdgBZActKF5sCcxPs5QPPYzp6dl8lvUca22fk5pZ0ar3F2gqTGQGnK7Z1zrsZZDfRjj00ko+LbGZLOAaZLamlmPwLzgJaJTVoC88zMJGXnZfIzNfjy8nMqbxCwNwh91ZtJej0rexOg3NEdnXOu1tTOA0qZp10zBx9NuHE6Mr7vHtMyeV0ktUg0x3QHHsjaNuxQ6kKYfzrvIFHl1djviMHtAtyZFfgM4OVytnfOudpTwHHWJTUilJmlQKmkZsAyQvPLbMKEROsRpgx91cwyTSj3AudLeppQdl5AGPocMxsn6SPgUkn9gIMIPWcyN0/vB96WtDdhhN3LgOFZbfJrKO8BpXviCb1jZmPS/gc451ydUNjuIf2ASxPvjwf+AowFrgQ2AH4EXgCOSax3G9CF0NsFQoX5tkT+0cAQ4AdgMtDHzL4FMLPRks4kFPBtgBeBU8oLNO8gYCtXkm4EHkpOrCFpD+BIMzuv3B1EPgiYy8UHAXO5VX0QsM6XPpu6zJn4l18UzcAyaT/PjmHVzYGM94FjCxuOc84VUKOS9EsRSdsrxljzQ6A0R5pzztUZaYYKKEZpC+Y3gCsklQDEnwNjunPO1U2FHVKg3khbY/898CShb+YkoCMwHTi0ugJzzrkqa6A19lQFu5lNldQD6Enovz4FGGlmK/Jv6Zxztcgn2sgvFuJvV2MszjlXWF6wr07S52bWNb6ewqqnqVZjZh2rKTbnnKsSK/WCPdvpidfHV3cgzjlXcN7GvjozG5F4/VrNhOOccwXkTTGrk3RZmh2Y2YDCheOccwXUMMv1vE0xHRKvmxEGpXkPyHR33BUYVn2hOedc1ZQUWf/0tPI1xawcaEbSQ8AxcdqmTNoRwK+rNzznnKu8hlqwpz3tg4DHstIeBw4uaDTOOVdAClPepVqKSdqC/Uvgt1lpZwPjCxuOc84VTgWmPC0qaR9QOg34j6QLga+B9oQB5o+orsCcc66qiq3ATivtkAIfStoC2A3YmDBOzNtmtrQ6g3POuapQA21jTz2kQJKZvS5pHUlNzGx+oYNyzrlCaKg19lSfZ5K2I0yeejur5j79CXBXNcXlnHNVVlqSfimPpHMkjZK0WNKQRPpukl6Q9L2kbyU9KmmjRP5ASUslzUssXRL5nSW9ImmBpDGS9s867rGSJkmaL+kxSa3LizXtF5VbgQFmtjWQaX55Ddgr5fbOOVfjCnzzdBpwBWtWaNcDBgOdgU7AXODurHUeNrPmiWVCIu9B4EPCnKaXAEMlrR/iVzfC/KgnAO2ABcAt5QWatimmG3BffG0AZjZf0lopt3fOuRpXyG6MZjY87nNnwvDlmfRnso55E6Himya+LYEewM/NbCEwTNJ5hAdCBwHHAU+Y2etx/f7A55JamNncsvabtsY+EdgpK6BdCd0gnXOuTlJJBRapb2xqySx9K3nYfYDRWWmHxqaa0ZLOSqR3AyZkFdIfx/RM/seZDDMbDywBtswXQNoae3/gKUmDgCaSLgbOZPURIJ1zrk6pSIXdzAYTmlSqcDxtDwwADkskPxL3O4MwWdEwSbPN7EGgOTAnazdzCF3KyZPfIl8cqWrsZvYk4enT9QlfMToBR5jZ82m2d8652lBSkn6pKkmbA88AvzezlfNBm9lnZjbNzJab2VvAP4E+MXse0DJrVy0J7fRp8nMqt8YuqZTQI2YbMzu7vPWdc66uqKlReyV1Al4ELjezf5ezurFq3MnRQJesNvPuwAOJ/O6J43QBmhLK5DKV+zllZsuB5YQRHp1zrt4oZK8YSY0kNQNKgVJJzWJae+Bl4GYzG5Rju8MkradgV+B3hLG2MLNxwEfApXF/hwPbs2rk3PsJ7fN7S1oHuAwYnu/GKaRvY78BeETSlcBUEtPkZXXbcc65OqPADyj1Ay5NvD8e+AuhPOxCKJxX5ptZ8/jyaEIXyaaE8vMaM7snsZ+jgSHAD8BkoI+ZfRv3MVrSmYQCvg3hW8EplENmOacyXX0laUUZWWZmpeXuIFqr4zHlH8w1OAsn/6W2Q3B10pZVLpZ3eWRE6jLnvSP3KprnVNOOFdNAR1xwztVnDXVIgbwFu6S1CV8/tgU+AK4ys8U1EZhzzlVVQ51oo7wa+03ALoQuPH0IbTznVndQzjlXCA10Lutye8UcRHjU9cL4+pDqD8k55wrDJ9rIbR0zmw5gZlMkrVsDMTnnXEH4eOxl5Ev6Kas602e/x8xerq7gnHOuKoqtJp5WeQX7TFYfonJW1vtM/03nnKtzim2S6rTyFuxm1rmG4nDOuYLzXjHOOVdkGmiFvWYL9m++PKkmD+fqidlLfFh/t6ZWTfIOOZ5KQ+3u6DV251zR8oK9DJJ+DpxMmMmjBWEc4NHA3Wb2QrVG55xzVVCihjk8VXlDCvwBuBC4nTCM5BzCIO/dgXskXWNm/6z2KJ1zrhIaeY09pz8CPzWzMVnpwyU9CLxCmA3EOefqHK+x57YOMK2MvG+AtQsbjnPOFU5DbWMvr5fnMOAJSftJWl9SE0ltJe0H/AcYWv0hOudc5ZRUYCkm5Z3PmcBbwD2EGbYXxp/3AO8AZ1VrdM45VwUlSr8Uk7wFu5ktMbOLzWwToDXQCWhjZpvE9CU1EqVzzlWCZKmX8velcySNkrRY0pCsvP0kjZG0QNIrcXLrTJ4kXSNpVlyuVWKsA0md4zYL4j72z9r3sZImSZov6TFJrcuLNfU3EDObbWZTzWx22m2cc642NVL6JYVpwBWsPl4WktoCw4H+hArwKODhxCp9gV6E3oTbE4Y/PyOR/yDwIWG+i0uAoZLWj/vuBtwGnAC0AxYAt5QXaOqCXdJ/c6Q9lXZ755yraSWy1Et5zGy4mT1GGAwx6QhgtJk9amaLgIFAd0lbx/yTgOtixfhr4DrCs0FI2hLoAVxqZgvNbBjwKdA7bnsc8ISZvW5m8wgfHkdIapH3vMs9m1XezJE2ogLbO+dcjapIG7ukvrGpJbP0TXmYbsDHmTdmNh8YH9PXyI+vk3kTzGxunvzkvscDS4C84y2kHlLAzK7JkXZV2u2dc66mVaTmamaDgcGVOExz4NustDmEJ/Uz+XOy8prHdvbsvEx++zK2zd53TqkLdkltgIOBjczsWkkbAyVmNjXtPpxzribVUG+XeYQn8pNaEoZfyZXfEphnZiapottm5+eU6gNN0k+AsYT2nv4xeQvg1jTbO+dcbShkG3seowk3RgGQtA6wWUxfIz++TuZ1yWozz85P7rsL0BQYly+gtN9UbgCOMrNfAMti2rvArim3d865GlfIXjGSGklqBpQCpZKaSWpEeFhzW0m9Y/4A4JPEUCz3AudLah9bOi4AhgCY2TjgI+DSuL/DCT1nhsVt7wcOlbR3/MC4DBie1Sa/hrQFe2czeym+zny0LcGH/XXO1WEFrrH3IzykeRFwfHzdz8y+JfRi+SvwA9ATODqx3W3AE4TeLv8DnoppGUcDO8dtrwb6xH1iZqMJD4reT5iqtAVwdnmBpi2YP5N0oJk9l0jbPwbqnHN1UiHb2M1sIKErY668F4Gty8gzwii5F5aRPxHYN89xHwAeqEisaQv2C4AnY7/1tSTdBhwKHFaRgznnXE0qtqEC0krVFGNm77CqQf8u4CtgVzN7rxpjc865Kmmog4ClqrFL2sHMPgKurd5wnHOucBqVNMzx2NN+UL0g6TNJ/SRtWq0ROedcgTTUGnva89mQMJvS1sDHkt6WdK6kDaovNOecqxoftjcPM1tuZk+Z2fGEEcb+CfQBplRncM45VxWFHLa3PqlQP/TY+f4Q4ChCv8s3qiMo55wrhGKriaeVdkiBgyXdR+ggfwHwGrCZme2ff0vnnKs9DbWNPW2N/e+EweB3jMNGOudcnddQe8WkKtjNbJvqDsQ55wqtoTbFlFmwS7rEzP4aX19W1npmNqA6AnPOuaoqre0Aakm+GvsmidcdqjsQ55wrtCoOx1tvlVmwm9lZiden1Ew4zjlXOA21KSZtr5jvy0ifWdhwnHOucBrqA0ppe8U0zk6Q1JiG24TlnKsHGhdbP8aU8hbskt4gTKzRTNLrWdmbAG9VV2DOOVdV3sae2x2AgF2AOxPpBswAXq6muJxzrsqKrYklrbwFu5ndAyDpncT8fc45Vy8Usq1Y0ryspLWAW8zsXEmdCfNUzE/kX2Nml8dtRZj27rSYdyfwpzi7EnH7uwnT6k0GzomzMlVK2geUxkhqR5i8ui2hFp/Ju6uyB3fOuepU4Knxmmdex4mlZwCPZq3WysyW5di8L9CLMGGRAS8AE4BBMf9B4G3g4LgMlbRFZu7Tiko70UYv4D7gC6AbYSalbYERhBmVnHOuzmlcfUMK9CGMnZV2IMSTgOvMbCqApOuA04FBkrYEegA/N7OFwDBJ5xEmyB5Uxv7ySnvP+ArgFDPbEZgff/YF3q/MQZ1zriZUpLujpL6SRiWWvnl2fRJwb6YpJWGSpKmS7pbUNpHeDfg48f7jmJbJm2Bmc8vIr7C0BXtHM8v+ynEPcGJlD+ycc9WtIgW7mQ02s50Ty+Bc+5TUEfgJoQzM+I7QyaQTsBPQArg/kd8cmJN4PwdoHtves/My+S0qe95p+7HPlNTOzGYAEyXtTjgR78funKuzqqlXzInACDP7KpNgZvOAUfHtDEnnANMltTSzH4F5QMvEPloC88zM4k3ZZF4mfy6VlLbGfjuwV3z9D+AVwleFWyp7YOecq26lstRLBZzI6rX1XDI7zHy0jCbcOM3oHtMyeV0ktSgjv8LS9oq5JvH6XkmvAuuY2eeVPbBzzlW3Qj94KmkPoD1ZvWEk9QRmEzqYrAfcCLxqZpkmlnuB8yU9TSj0LwD+BWBm4yR9BFwqqR9wELA94eZppVRoarwMM5tc2QM651xNaVT4IQVOAoZn3egE6AJcCWwA/EjoznhMIv+2uM6n8f0dMS3jaGAI8AOhH3ufynZ1BNCaN3VzrCRNYdVXi6TFwFRgOHBrGf03V5qz5NmG+Xyvy8tYUdshuDqoVZODq9xC/p+Jz6Qucw7vfFDRPKeatsZ+I3B8/DkF6Aj8lvB15HvC14oOwIXVEKNzzlWKDymQ38nAAWY2LZMg6RngeTPrJukV4EW8YHfO1SFesOe3EaG7TtJ8YOP4ehzQqkAxOedcQXjBnt8TwOOS/kpoU98EuDimA+wOTCx4dM45VwXVOKRAnZb2nvEZwLuEu7gfxp/vAWfG/AnALwsenXPOVUFJBZZikrYf+yLgorjkyv+mkEHVNwMuupf33v2CRQsX07ptS044ZT969d4dgJHvjOVvfx3KN9/8wLbbdWLAFcex0catAViyZBnXXT2M1176hGXLVrD9jptyUf8j2aBdKwA++egrrr9mOBMnzGDj9q25sN+v2aHHZrV1mq6CLr3oPt57dxwLFy6hTduWnHDKzzis9258+vFEBt/0DGM+m0pJqeix8+ZccPHhtF1/XSBcF9dfPZxXX/qU5ctWsP2OnflT/1+vvC7OOvVmJnw5nSVLlrFx+zb0/e0v+MnPtqvFM627GmpTTKrujgCSDiD0tdzAzA6VtDPQ0sxST7ZRrN0dx385nQ4d16dJk0ZMnDCDM0/9F/+4+Qw22ng9Dj/4ci4ZeDR777stt930NB99MJ677j8fgHvveolnnxrFv247i+Yt1uLKgQ+xcOESrr3hN8yZM58+h/yVP/X7NT/dvzvPP/0+f79qGP95ZgAt1127ls+4sIq1u+OEL6ezSeK6OOvUm7n+5tP54ft5LFywmJ57bk2j0hL+duUwvvv2R/456AwA/n3XSzz71AfceNuZNG/RjCsHPsyihUu45oZTAfhi7DQ23awdjRqV8r9PJnHu6bfy6JMXr/xgKBaF6O742vSnU5c5P9mo6serK9JOZn0ucCvhqap9YvJCwqiPDd5mm29Ekybhy4/CKHFMnfIdr7z4CV0225D9D9yRpk0bc/pZv+CLsdOYOGEGANO+nsVue25Nm7Ytadq0MQcc1IMJ48OXn08++orWbVqw/4E7UlpawkGH7kKr1s155aWPy4zD1S1dVrsuhARfT/mOPfbuyn4H7kDz5s1otlYTfn3MXnzy4cphR5j29ffstudWtGnbIl4XO668LgC22GpjGjUqjfuFZcuWM+Ob2TV6bvVFiSz1UkzS3jw9D9jPzCZK+lNMGwNsVS1R1UPXXPEITz4+ksWLlrJV103Yc59tuPXGJ9liq/Yr11lr7aa079CGCeOn07lLO351xG5cf/Vwvp05hxYt1uLZJ0exx15dw8oGZH+bMmP8F9Nr7qRclV17xdDEddGePfbZZo11Pnx/AptuvuHK9786oifXX/2fldfFc09+wO6Z6yI6/7e3894741iyZBm77bk1Xbt1qPZzqY8aalNM2oK9BeHBJFj1BGpjYEl5G8YxjfsC3HDzuZx82sEVjbFe+FO/I/m/i/vw6cdf8f57X9KkcSMWLFjMeus1X2295s3XYsH8xQB07LQBG260Hr/cbwClpSVstsVG/PGSPgBst8OmfPvtjzz39Pvsd8AOPPv0+0ydMovFi5bW+Lm5yruwXx8uuPgIPv14Ih/E6yLpi7HTuGvQ81x746kr0zp02oB2G63HIfsNXHld/N8lR6y23fU3n86ypcsZ+c44Jn01g5KSYrv9VxiNGmjBnvZqeJ01b5z+jjDKY17JMY6LtVDPKC0tYYcemzFzxmyGPTKCtdduyvz5i1ZbZ/68Ray9TlMArrn8ERYvXsoLI67ktZF/46f7dee8s8LwEa1arcPf/3kaD9z7Cr/Ytx/vvPk5u+625cobaK7+CNdFl3hdvLkyfcrkb/nD2YP5w0W92HGnVTfFr738UZYsXsbzI67g1ZHXsO9+2/OHs9YcGrxR41L22Lsr77w1ltdf+V+NnEt9E5pG0y3FJG3Bfi5wuKSJQAtJY4FfA+dXV2D12fLlK5g65Tu6bLYRX4xd+bAuCxcsZurUkA7wxbivOeSwnqy77jo0adKII4/dm9GfTmL2D+FZsB67bM49D/0fL755FQOvPJ5JE2eyzXYda+WcXNVlrguA6dO+59zTb+XUMw7g4EN3WW29L8ZN45eH7ZJ1XUxeeV2ssd9ly/k67tetThVYikmqgt3MphNmBzkKOJYwwlnPht7NEeD7WXN5/pkPWLBgMcuXr+DtNz/n+Wc+YOddt2Tf/bZn/JfTefmFj1i8eCl3DHqOLbbcmM5d2gGwTbeOPP3ESObNXciypcsZ+vAI1t9gXVrF5puxn09l2dLlzJu3iBuve5wN2rVi9z275gvH1RHZ18U7b47h+Wc+ZOddt2DmjNn89je30PvovTjiyD3X2LZrt448/cSoldfFsMR1MXHCDN5643MWLVrCsqXLeeaJUXz4/gR23Nm7webSUGvsqbs7FkIxdnf84ft5XHT+XXwxbhq2YgUbbtSao47bh1599gBg5Ntj+duVQ/lm+g90264TA644lo3btwFg9uz5XHfVMEa+M5alS5ez2eYbct4fD6fbdp0A6HfhPbz5xmcA7L5nV/7v4t60blPp2bLqrGLs7vjD9/O4+PwhfDHua1asMDbaqDVHHrc3vfrszh23PsvttzzHWms1WW2bV0eGaQ/mzJ7PdVcNZ+Q741i6dBldNt+I8/54GN2268RXE2Zweb8H+Gr8DEpKRYeO63Py6fuz737b18ZpVqtCdHf84LunUpc5Pdr+smiK97wFu6SvyD1cb4aZWeqqQjEW7K7qirFgd1VXiIL9w1lPpi5zdmxzSNEU7OX1ijmtjPSdCCM55h1/3TnnapN3d8zBzF5KvpfUFbgc+Cnwd8L47M45Vyc10HI99ZOnXST9G3gL+BzoYmZXmdn8ao3OOeeqoETplzQkvSppkaR5cRmbyNtP0hhJCyS9IqlTIk+SrpE0Ky7XSqtu2UrqHLdZEPexf5XOu5yTaC9pEGFExxnAFmbWPzFBq3PO1VnV1N3xHDNrHpetACS1JUwR2h9oDYwCHk5s0xfoBXQnTFR9CGHU3IwHCeVsG+ASYKik9SsW1irltbF/SZhg4+/A18CvlNUvyMzuquzBnXOuOtVgN8YjgNFm9mg4rgYC30na2szGELqIX2dmU2P+dcDpwCBJWwI9gJ+b2UJgmKTzgN7AoMoEU17B/i6hV8zPysg3wAt251ydVJGBFpLDn0SDzWzNR37hKklXA2OBS8zsVaAbsHKEPjObL2l8TB+TnR9fd4uvuwETzGxuGfkVVt7N030ru2PnnKttFekVEwvxXAV50p+AzwjjZB0NPCFpB6A58G3WunMI42wR8+dk5TWP7ezZeZn89lRS2kHAnHOu3il0S4yZvZt4e4+kY4CDCU3WLbNWbwlkauHZ+S2BeWZmksrbtsJ8SDjnXNGSLPVSSUb4/BhNuDEaj6t1gM1iOtn58XUyr4ukFmXkV5gX7M65olXIXjGSWkk6UFIzSY0kHUeYeOg54D/AtpJ6S2oGDAA+iTdOAe4Fzo89DTcGLgCGAJjZOOAj4NK478MJPWeGVfa8vSnGOVe0CtwrpjFh1ritgeWEm6K9zGxsOJZ6AzcB9xE6nhyd2PY2oAvwaXx/R0zLOJpQ0P8ATAb6mFl2m31qqQYBk/S9mbXOkT7TzDZIezAfK8bl4mPFuFwKMVbMpHlPpC5zOjU/tGgeVE1bY2+cnSCpMVBa2HCcc65wiqakrqC8BbukNwg3B5pJej0rexPCEAPOOVcnFds462mVV2O/g/ChtwtwZyLdCEMMvFxNcTnnXJU10HK93AeU7gGQ9E7i7q5zztULDXXY3rTdHc+WtEcyQdIekm4ofEjOOVcYPudpfscQRitLep8w/6lzztVJJbLUSzFJ2yvGWPNDoDRHmnPO1RkN9eZp2oL5DeAKSSUA8efAmO6cc3VSQ22KSVtj/z3wJDBd0iSgIzAdOLS6AnPOuapqqE0KqQp2M5sqqQfQk9B/fQow0sz8kUHnXJ3VUJtiUo8VEwvxt6sxFuecKyg10Dp7mQW7pM/NrGt8PYVwA3UNZtaxmmJzzrkqibcFG5x8NfbTE6+Pr+5AnHOu8BpmW0yZBbuZjUi8fq1mwnHOucKRF+yrk3RZmh2Y2YDCheOcc4XkBXu2DonXzYDewHtAprvjrlRhhg/nnKtu3saexcxOybyW9BBwjJkNS6QdAfy6esNzzrnKa6i9YtKe9UHAY1lpjxNm53bOuTpJFfhX7r6kppLulDRJ0lxJH0o6KOZ1lmSS5iWW/oltJekaSbPicq20qpd93P4VSQskjZG0f1XOO20/9i+B3wI3JtLOBsZX5eDOOVe9Clpjb0R4OPMnhHlJDwYekbRdYp1WZrYsx7Z9gV5Ad0LX8ReACcCgmP8g4Tmhg+MyVNIWlZ33NO1Zn0aYYXuqpHclTSXMsn1aZQ7qnHM1QVLqpTxmNt/MBprZRDNbYWZPAl8BO6UI5STgOjObamZfA9cBJ8cYtwR6AJea2cLY5P0p4b5mpaQdUuBDSVsAuwEbE8aJedvMllb2wM45V/3S94qR1JdQs84YbGaD86zfDtgSGJ1IniQpUyP/o5l9F9O7AR8n1vs4pmXyJpjZ3DLyKyz1kAJJZva6pHUkNTGz+ZU9uHPOVaeK9GOPhXiZBflq+5UaA/cD95jZGEnNCVOIfgS0AW6O+QfGTZoDcxK7mAM0j+3s2XmZ/Papg8+SqmCPbUj/BRYTBgF7mNDOdBJwVGUP7pxz1UmUFn6foQ/lv4ElwDkAZjaPVZMRzZB0DmE03JZm9iMwD2iZ2E1LYJ6ZmaTsvEz+XCopbRv7rcAAM9sayDS/vAbsVdkDO+dcdStkG3vcn4A7gXZA7zzN0ZmxtTI7Hk24cZrRnVVNOKOBLpJalJFfYWkL9m7AffG1QbiRAKxV2QM751z1K/hUG7cCXYFDzWzhyqNIPSVtJalEUhtCD8JXzSzTxHIvoQNKe0kbEzqfDAEws3GEJpxLJTWTdDiwPVV4ADRtG/tEwp3flfOeStqV0A3SOefqpEI+oCSpE3AGoUn6m0Qt/wxgBXAlsAHwI+Hm6TGJzW8DuhB6uwDcEdMyjiYU9D8QulL2qWxXR0hfsPcHnpI0CGgi6WLgTFYfAdI55+qYwo0VY2aTytnhg3m2NeDCuOTKnwjsW4XwVpPq4yz21zwIWJ/Qtt4JOMLMni9UIM45V2hSSeqlmJRbY5dUCowDtjGzs6s/JOecKwwfK6YMZrYcWE4Y4dE55+qRgt88rRfStrHfQBgT4UpgKolp8sxsQjXE5ZxzVeYTbeR3U/x5QFa6QTU8AeCccwWQtn96sUk7VkzDbKhyztVzDbPoyluwS1ob6AdsC3wAXGVmi2siMOecq6qGevO0vBr7TYSBbZ4B+hAGtzm3uoNyzrlCaKhNMeV9nB0E/NzMLoyvD6n+kJxzrlBKKrAUj/Jq7OuY2XQAM5siad0aiMk55wqiofaKUXjStYxMaQHwS1Z18nwMOCzxHjN7uRrjK1qS+uYbxN81TH5duEIor2CfSKLPeg5mZl0KHVRDIGmUme1c23G4usWvC1cIeZtizKxzDcXhnHOuQIrrjoFzzjkv2GuRt6O6XPy6cFWWt43dOedc/eM1duecKzJesDvnXJHxgr0KJA2UdF/5axbseKMl7Zsn/xlJJ9VUPC6o6eugKiR1lDQvTqDjilS9LtglvSrpB0lNU65/sqQR1R1XPNa+klbEP6K5ksZKOqUq+zSzbmb2atz/GoWJmR1kZvdU5Ri5SDpH0ihJiyUNKfT+q6oeXAcm6eas9BGSTq6B40+UtH/mvZlNNrPmcQKdQh9rP0ljJC2Q9Eqc/NnVgnpbsEvqDOxNeIDqV7UbTZmmmVlzoCXwJ+B2SdvUckyVMQ24ArirtgPJVk+ug/nAiTHWoiSpLTCcMPF9a2AU8HCtBtWA1duCHTgReAcYAqzW/CCpg6Thkr6VNEvSTZK6AoOA3WMtenZc91VJpyW2Xa02J+mfkqZI+lHS+5L2rmigFjwG/ABsI6mppBskTYvLDZnapqS2kp6UNFvS95LeUJxpN1P7kvQL4M/AUfFcPk6eS9z/bEnbJs5jfUkLJW0Q3x8i6aO43luSts8T//AY/6yKnnsNqA/XwewY36VlrSDpVEmfx28ezyVru5J+Hr/xzZF0i6TXMrFK2kzSy/H8vpN0v6RWMe/fQEfgiXiuF0rqHL9BNJJ0tKRRWXH8QdJ/4+umkv4uabKkGZIGSVqrjFM4AhhtZo+a2SJgINBd0tYV+H9yBVLfC/b743KgpHawcvLtJ4FJQGegPfCQmX0OnAm8Hb+Ktkp5nPeAHQi1kAeARyVVaP5XSSWSDgdaAZ8ClwC7xf12B3YljHsPcAFh+sH1gXaEAny1Pqlm9ixwJfBwPJfuWfmLCbWnYxLJRwKvmdlMST0Ite8zCEMx3wb8N/HhcoukWypyjrWovlwHfwV6S9oqO0NSL8Lv+QjC7/0N4MGY1xYYClxM+F2NBfZIbg5cBWwMdAU6EApVzOwEYDJwaDzXa7MO/V9gK0lbJNKOjecHcA2wZTzvzQn/hwMScc+WtFd82w34OJNnZvOB8THd1bB6WbDHi6kT8IiZvU+4gI6N2bsSLvI/mtl8M1tkZpVuTzWz+8xslpktM7PrgKbAGn+cZdg41gi/I9TWTjCzscBxwGVmNtPMvgX+ApwQt1kKbAR0MrOlZvaGVe5hgwdYvWBP/sGeDtxmZu+a2fLYLr+Y8GGDmZ1tZmdX4pg1qh5dB5jZN4RvCpflyD6DMInN52a2jPChvUOstR9MqAkPj3k3At8k9vulmb1gZovjtXQ98JOUMS0AHideJ7GA35rwIS/CdfIHM/vezObGuI5ObN8q8X/aHJiTdYg5QIs0sbjCqpcFO+Er9/Nm9l18/wCrvoZ3ACbFP4Iqk3RB/Io8JxbS6wJtU24+LV78rc1sBzN7KKZvTKhJZkyKaQB/A74Enpc0QdJFlQz9ZWAtST1jAbED8J+Y1wm4INa4Zsfz6pCIob6oL9dBxjWEbxXds9I7Af9M/C6+J9TE2xN+J1MyK8YP+amJuDaQ9JCkryX9CNxXwbiSFYBjgcdigb8+sDbwfiKuZ2N6LvMI95KSWgJzKxCLK5C0k1nXGbGN70igVFKm5tIUaBX/YKYAHSU1yvFHnavmO59wAWdsmDjW3oSbnvsRak0rJP0AVR7keRrhj3l0fN8xphFrRhcQCt5uwCuS3jOzl1Kcy6rMEOsjhD/aGcCTcd8Q/o/+amZ/reJ51Jr6eB2Y2SxJNwCXZ2Vlfh/3Z28Ta9GbJN4r+Z7QDGPA9nH/vVg1+TyUc50AzwNtJe1AuFb+ENO/AxYC3czs63L2AeFaXnmPQ9I6wGasusZdDaqPNfZewHJgG0ItdAdC2+IbhPbWkcB04GpJ60hqJmnPuO0MYBNJTRL7+wg4QtLakjYHfpPIawEsA74FGkkawJq1ksp4EOincEOzLaHd8j5YeVNz8/gH/GM811xd02YAnRVvrJbhAeAoQtPPA4n024EzY21e8f/pl5Jyfm2ON9qaAaWEgrSZpNquFPSifl4H1xPayLsm0gYBF8cPciStK+nXMe8pYDtJveL/+W9JfOjE2OYBsyW1B/6YdbwZQJlDa8cPvaGEb4qtgRdi+grCdfIPrbrh3l7SgWXs6j/AtpJ6x2tlAPCJmY0p+7/CVZf6WLCfBNwd++N+k1kItZTjCLWoQwk3eyYTvrYeFbd9mVCD+EZS5uv7P4AlhD+Aewg34TKeI8z3Oo7QXLKIxNfiKriC0B3sE8LN1A9iGsAWwIuEP9a3gVsyfdezPBp/zpL0Qa6DmNm7hJroxvE8MumjCO2nNxF66nwJnJzJj70fBiV21Y9Qe7sIOD6+7kftqpfXgZn9CFxLKEQzaf8hNNM8FJtT/keYipLYzPTruM0swgfZKMI9EQj3Z3oQ2rOfItw0T7qKUImYLen/ygjrAWB/4NGsbzd/Ilwb78S4XiRxX0Ghp83eMc5vgd6Em8Q/AD1JtMe7muWDgDlXj8RvaFOB48zsldqOx9VN9bHG7lyDIulASa1id9Q/E76NvFPLYbk6zAt25+q+3QldOb8jNC/1MrOFtRuSq8u8KcY554qM19idc67IeMHunHNFxgt255wrMl6wO+dckfGC3Tnnisz/A9LVTyT5cl27AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pickle\n",
    "Model = pickle.dumps(model_rf)\n",
    "\n",
    "import tkinter as tk\n",
    "from tkinter import ttk\n",
    "from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg\n",
    "\n",
    "def create_plot(): \n",
    "\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    cm_matrix = pd.DataFrame(data=cm, columns=['Actual Positive:1', 'Actual Negative:0'], \n",
    "                                     index=['Predict Positive:1', 'Predict Negative:0'])\n",
    "    g = sns.heatmap(cm_matrix, annot=True, fmt='d', cmap='YlGnBu')\n",
    "    plt.title('Model Confusion Matrix', fontsize=20)\n",
    "\n",
    "    return g.figure\n",
    "\n",
    "def create_plot2():\n",
    "    #Display ROC Curve\n",
    "    from sklearn.metrics import roc_curve\n",
    "    y_score = model_rf.predict_proba(X_test)[:,1]\n",
    "    fpr, tpr, thresholds = roc_curve(y_test, y_score, pos_label = 'Yes')\n",
    "    x = plt.figure(figsize=(6,4))\n",
    "    plt.plot(fpr, tpr, linewidth=2)\n",
    "    plt.plot([0,1], [0,1], 'k--' )\n",
    "    plt.rcParams['font.size'] = 12\n",
    "    plt.title('ROC curve for RainTomorrow classifier')\n",
    "    plt.xlabel('False Positive Rate (1 - Specificity)')\n",
    "    plt.ylabel('True Positive Rate (Sensitivity)')\n",
    "    plt.show()\n",
    "    \n",
    "    return x.figure\n",
    "\n",
    "def New_Window():\n",
    "    Window = tk.Toplevel()\n",
    "    Window.title('Model Details')\n",
    "    Window.geometry(\"500x350\")\n",
    "    Window['bg']='#ffffff'\n",
    "    #Visualize confusion matrix\n",
    "    figure = create_plot()\n",
    "    canvas = FigureCanvasTkAgg(figure, master=Window)\n",
    "    canvas.draw()\n",
    "    canvas.get_tk_widget().pack()\n",
    "    \n",
    "    button2 = tk.Button(Window, text=\"Close\", bg='White', fg='Black',\n",
    "                              command=lambda: Window.destroy(),font='Arial 11 bold')\n",
    "    button2.pack()\n",
    "    \n",
    "def New_Window2():\n",
    "    Window2 = tk.Toplevel()\n",
    "    Window2.title('Model Details')\n",
    "    Window2.geometry(\"500x350\")\n",
    "    Window2['bg']='#ffffff'\n",
    "    \n",
    "    figure2 = create_plot2()\n",
    "    canvas = FigureCanvasTkAgg(figure2, master=Window2)\n",
    "    canvas.draw()\n",
    "    canvas.get_tk_widget().pack()\n",
    "    \n",
    "    button4 = tk.Button(Window2, text=\"Close\", bg='White', fg='Black',\n",
    "                              command=lambda: Window2.destroy(),font='Arial 11 bold')\n",
    "    button4.pack()\n",
    "    \n",
    "win = tk.Tk()\n",
    "\n",
    "win.title('Tomorrow Rain Predictions')\n",
    "\n",
    "win.geometry(\"750x250\")\n",
    "win['bg']='#070362'\n",
    "\n",
    "#Column 1\n",
    "MinTemp=ttk.Label(win,text=\"MinTemp\")\n",
    "MinTemp.grid(row=0,column=0,sticky='W')\n",
    "MinTemp_var=tk.DoubleVar()\n",
    "MinTemp_entrybox=ttk.Entry(win,width=16,textvariable=MinTemp_var)\n",
    "MinTemp_entrybox.grid(row=0,column=1)\n",
    "#Column 2\n",
    "MaxTemp=ttk.Label(win,text=\"MaxTemp\")\n",
    "MaxTemp.grid(row=1,column=0,sticky='W')\n",
    "MaxTemp_var=tk.DoubleVar()\n",
    "MaxTemp_entrybox=ttk.Entry(win,width=16,textvariable=MaxTemp_var)\n",
    "MaxTemp_entrybox.grid(row=1,column=1)\n",
    "#Column 3\n",
    "Rainfall=ttk.Label(win,text=\"Rainfall\")\n",
    "Rainfall.grid(row=2,column=0,sticky='W')\n",
    "Rainfall_var=tk.DoubleVar()\n",
    "Rainfall_entrybox=ttk.Entry(win,width=16,textvariable=Rainfall_var)\n",
    "Rainfall_entrybox.grid(row=2,column=1)\n",
    "#Column 4\n",
    "Evaporation=ttk.Label(win,text=\"Evaporation\")\n",
    "Evaporation.grid(row=3,column=0,sticky='W')\n",
    "Evaporation_var=tk.DoubleVar()\n",
    "Evaporation_entrybox=ttk.Entry(win,width=16,textvariable=Evaporation_var)\n",
    "Evaporation_entrybox.grid(row=3,column=1)\n",
    "#Column 5\n",
    "Sunshine=ttk.Label(win,text=\"Sunshine\")\n",
    "Sunshine.grid(row=4,column=0,sticky='W')\n",
    "Sunshine_var=tk.DoubleVar()\n",
    "Sunshine_entrybox=ttk.Entry(win,width=16,textvariable=Sunshine_var)\n",
    "Sunshine_entrybox.grid(row=4,column=1)\n",
    "#Column 6\n",
    "WindGustSpeed=ttk.Label(win,text=\"WindGustSpeed\")\n",
    "WindGustSpeed.grid(row=5,column=0,sticky='W')\n",
    "WindGustSpeed_var=tk.DoubleVar()\n",
    "WindGustSpeed_entrybox=ttk.Entry(win,width=16,textvariable=WindGustSpeed_var)\n",
    "WindGustSpeed_entrybox.grid(row=5,column=1)\n",
    "#Column 7\n",
    "WindSpeed9am=ttk.Label(win,text=\"WindSpeed9am\")\n",
    "WindSpeed9am.grid(row=0,column=2,sticky='W')\n",
    "WindSpeed9am_var=tk.DoubleVar()\n",
    "WindSpeed9am_entrybox=ttk.Entry(win,width=16,textvariable=WindSpeed9am_var)\n",
    "WindSpeed9am_entrybox.grid(row=0,column=3)\n",
    "#Column 8\n",
    "WindSpeed3pm=ttk.Label(win,text=\"WindSpeed3pm\")\n",
    "WindSpeed3pm.grid(row=1,column=2,sticky='W')\n",
    "WindSpeed3pm_var=tk.DoubleVar()\n",
    "WindSpeed3pm_entrybox=ttk.Entry(win,width=16,textvariable=WindSpeed3pm_var)\n",
    "WindSpeed3pm_entrybox.grid(row=1,column=3)\n",
    "#Column 9\n",
    "Humidity9am=ttk.Label(win,text=\"Humidity9am\")\n",
    "Humidity9am.grid(row=2,column=2,sticky='W')\n",
    "Humidity9am_var=tk.DoubleVar()\n",
    "Humidity9am_entrybox=ttk.Entry(win,width=16,textvariable=Humidity9am_var)\n",
    "Humidity9am_entrybox.grid(row=2,column=3)\n",
    "#Column 10\n",
    "Humidity3pm=ttk.Label(win,text=\"Humidity3pm\")\n",
    "Humidity3pm.grid(row=3,column=2,sticky='W')\n",
    "Humidity3pm_var=tk.DoubleVar()\n",
    "Humidity3pm_entrybox=ttk.Entry(win,width=16,textvariable=Humidity3pm_var)\n",
    "Humidity3pm_entrybox.grid(row=3,column=3)\n",
    "#Column 11\n",
    "Pressure3pm=ttk.Label(win,text=\"Pressure3pm\")\n",
    "Pressure3pm.grid(row=4,column=2,sticky='W')\n",
    "Pressure3pm_var=tk.DoubleVar()\n",
    "Pressure3pm_entrybox=ttk.Entry(win,width=16,textvariable=Pressure3pm_var)\n",
    "Pressure3pm_entrybox.grid(row=4,column=3)\n",
    "#Column 12\n",
    "Cloud9am=ttk.Label(win,text=\"Cloud9am\")\n",
    "Cloud9am.grid(row=5,column=2,sticky='W')\n",
    "Cloud9am_var=tk.DoubleVar()\n",
    "Cloud9am_entrybox=ttk.Entry(win,width=16,textvariable=Cloud9am_var)\n",
    "Cloud9am_entrybox.grid(row=5,column=3)\n",
    "#Column 13\n",
    "Cloud3pm=ttk.Label(win,text=\"Cloud3pm\")\n",
    "Cloud3pm.grid(row=0,column=4,sticky='W')\n",
    "Cloud3pm_var=tk.DoubleVar()\n",
    "Cloud3pm_entrybox=ttk.Entry(win,width=16,textvariable=Cloud3pm_var)\n",
    "Cloud3pm_entrybox.grid(row=0,column=5)\n",
    "#Column 14\n",
    "Year=ttk.Label(win,text=\"Year\")\n",
    "Year.grid(row=1,column=4,sticky='W')\n",
    "Year_var=tk.IntVar()\n",
    "Year_entrybox=ttk.Entry(win,width=16,textvariable=Year_var)\n",
    "Year_entrybox.grid(row=1,column=5)\n",
    "#Column 15\n",
    "Month=ttk.Label(win,text=\"Month\")\n",
    "Month.grid(row=2,column=4,sticky='W')\n",
    "Month_var=tk.IntVar()\n",
    "Month_entrybox=ttk.Entry(win,width=16,textvariable=Month_var)\n",
    "Month_entrybox.grid(row=2,column=5)\n",
    "#Column 16\n",
    "Day=ttk.Label(win,text=\"Day\")\n",
    "Day.grid(row=3,column=4,sticky='W')\n",
    "Day_var=tk.IntVar()\n",
    "Day_entrybox=ttk.Entry(win,width=16,textvariable=Day_var)\n",
    "Day_entrybox.grid(row=3,column=5)\n",
    "#Column 17\n",
    "RainToday=ttk.Label(win,text=\"RainToday\")\n",
    "RainToday.grid(row=4,column=4,sticky='W')\n",
    "RainToday_var=tk.StringVar()\n",
    "RainToday_entrybox=ttk.Entry(win,width=16,textvariable=RainToday_var)\n",
    "RainToday_entrybox.grid(row=4,column=5)\n",
    "\n",
    "\n",
    "global DB\n",
    "import pandas as pd\n",
    "def Output():\n",
    "    global MinTemp_entrybox\n",
    "    global MaxTemp_entrybox\n",
    "    global Rainfall_entrybox\n",
    "    global Evaporation_entrybox\n",
    "    global Sunshine_entrybox\n",
    "    global WindGustSpeed_entrybox\n",
    "    global WindSpeed9am_entrybox\n",
    "    global WindSpeed3pm_entrybox\n",
    "    global Humidity9am_entrybox\n",
    "    global Humidity3pm_entrybox\n",
    "    global Pressure3pm_entrybox\n",
    "    global Cloud9am_entrybox\n",
    "    global Cloud3pm_entrybox\n",
    "    global Year_entrybox\n",
    "    global Month_entrybox\n",
    "    global Day_entrybox\n",
    "    global RainToday_entrybox\n",
    "    global min_t\n",
    "    global max_t\n",
    "    global rainfall\n",
    "    global evap\n",
    "    global sun\n",
    "    global wind_gust_speed\n",
    "    global wind_speed_9\n",
    "    global wind_speed_3\n",
    "    global humid_9\n",
    "    global humid_3\n",
    "    global pres_3\n",
    "    global cloud_9\n",
    "    global cloud_3\n",
    "    global year\n",
    "    global month\n",
    "    global day\n",
    "    global rain_today\n",
    "    min_t=MinTemp_entrybox.get()\n",
    "    max_t=MaxTemp_entrybox.get()\n",
    "    rainfall=Rainfall_entrybox.get()\n",
    "    evap=Evaporation_entrybox.get()\n",
    "    sun=Sunshine_entrybox.get()\n",
    "    wind_gust_speed=WindGustSpeed_entrybox.get()\n",
    "    wind_speed_9=WindSpeed9am_entrybox.get()\n",
    "    wind_speed_3=WindSpeed3pm_entrybox.get()\n",
    "    humid_9=Humidity9am_entrybox.get()\n",
    "    humid_3=Humidity3pm_entrybox.get()\n",
    "    pres_3=Pressure3pm_entrybox.get()\n",
    "    cloud_9=Cloud9am_entrybox.get()\n",
    "    cloud_3=Cloud3pm_entrybox.get()\n",
    "    year=Year_entrybox.get()\n",
    "    month=Month_entrybox.get()\n",
    "    day=Day_entrybox.get()\n",
    "    rain_today=RainToday_entrybox.get()\n",
    "    \n",
    "    DB=preprocess_input()\n",
    "    output=model_rf.predict(DB)\n",
    "    result=output\n",
    "    Predict_entrybox=ttk.Entry(win,width=16)\n",
    "    Predict_entrybox.grid(row=10,column=1)\n",
    "    Predict_entrybox.insert(1,str(result))\n",
    "    \n",
    "Predict_button=ttk.Button(win,text=\"Predict\",command=Output)\n",
    "Predict_button.grid(row=10,column=0,sticky='W')\n",
    "\n",
    "button = tk.Button(win, text=\"Check Model Visualization >>\", bg='White', fg='Black',\n",
    "                              command=lambda: New_Window())\n",
    "button.grid(row=12,column=0,sticky='W')\n",
    "button3 = tk.Button(win, text=\"Check Model ROC Curve >>\", bg='White', fg='Black',\n",
    "                              command=lambda: New_Window2())\n",
    "button3.grid(row=13,column=0,sticky='W')\n",
    "button5 = tk.Button(win, text=\"Exit\", bg='White', fg='Black',\n",
    "                              command=lambda: win.destroy())\n",
    "button5.grid(row=14,column=0,sticky='W')\n",
    "win.mainloop()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "11495505",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}