16_capstone2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cc67eabb-3bcd-44a9-8625-7ee6aff6e5ec",
   "metadata": {},
   "source": [
    "# Capstone project: Classify robot failures\n",
    "### Optional challenge: read the raw data files and format them correctly to save as CSV files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e1015d49-29d6-4fb5-834d-8d9d4e41c8c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "# Initialize lists to store data\n",
    "labels = []\n",
    "values = []\n",
    "all_labels = []\n",
    "\n",
    "def is_numeric_line(line):\n",
    "    try:\n",
    "        [int(num) for num in line.split()]\n",
    "        return True\n",
    "    except ValueError:\n",
    "        return False\n",
    "\n",
    "# Open the file and read through the data line by line\n",
    "with open('data/lp1.data', 'r') as file:\n",
    "    current_label = None\n",
    "    for line in file:\n",
    "        line = line.strip()\n",
    "        if line:\n",
    "            if is_numeric_line(line):\n",
    "                values.append([int(num) for num in line.split()])\n",
    "                labels.append(current_label)\n",
    "            else:\n",
    "                current_label = line.strip()\n",
    "                all_labels.append(current_label)\n",
    "\n",
    "# Create DataFrame for the numeric values\n",
    "columns = ['F_x', 'F_y', 'F_z', 'T_x', 'T_y', 'T_z']\n",
    "df = pd.DataFrame(values, columns=columns)\n",
    "\n",
    "# Add 'timestep' column\n",
    "df['timestep'] = [i % 15 + 1 for i in range(len(df))]\n",
    "\n",
    "# Add 'id' column (constant value per 15 rows)\n",
    "df['id'] = [i // 15 for i in range(len(df))]\n",
    "df = df[['id', 'timestep', 'F_x', 'F_y', 'F_z', 'T_x', 'T_y', 'T_z']]\n",
    "df.to_csv('data/lp1_X.csv', header=True, index=False)\n",
    "\n",
    "# Create DataFrame for the labels\n",
    "labels_df = pd.DataFrame(all_labels, columns=['label'])\n",
    "labels_df.to_csv('data/lp1_y.csv', header=True, index=False)\n",
    "\n",
    "# Print the DataFrames\n",
    "print(\"DataFrame with numeric values:\")\n",
    "print(df)\n",
    "print(\"\\nDataFrame with labels:\")\n",
    "print(labels_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e319547d-40da-4814-8dee-06d587b43d49",
   "metadata": {},
   "source": [
    "## 1. Initial setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c18acab5-e9c9-46b5-9701-5369648579b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import classification_report, precision_recall_fscore_support \n",
    "\n",
    "from sktime.classification.dictionary_based import WEASEL\n",
    "from sktime.classification.ensemble import BaggingClassifier\n",
    "from sktime.classification.feature_based import SummaryClassifier, Catch22Classifier\n",
    "\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier\n",
    "\n",
    "from tsfresh import extract_features, select_features\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10bfc911-fa1e-4187-80e4-c054e72515ed",
   "metadata": {},
   "source": [
    "### Read data (lp1 data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2168cac6-3444-40f7-a53c-2e42bac1e2dd",
   "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>id</th>\n",
       "      <th>timestep</th>\n",
       "      <th>F_x</th>\n",
       "      <th>F_y</th>\n",
       "      <th>F_z</th>\n",
       "      <th>T_x</th>\n",
       "      <th>T_y</th>\n",
       "      <th>T_z</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>62</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>61</td>\n",
       "      <td>-3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-2</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  timestep  F_x  F_y  F_z  T_x  T_y  T_z\n",
       "0   0         1   -1   -1   63   -3   -1    0\n",
       "1   0         2    0    0   62   -3   -1    0\n",
       "2   0         3   -1   -1   61   -3    0    0\n",
       "3   0         4   -1   -1   63   -2   -1    0\n",
       "4   0         5   -1   -1   63   -3   -1    0"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = pd.read_csv('data/lp1_X.csv')\n",
    "y = pd.read_csv('data/lp1_y.csv').squeeze()\n",
    "\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fa70ff34-5481-4ec1-843c-284aeb994d41",
   "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></th>\n",
       "      <th>F_x</th>\n",
       "      <th>F_y</th>\n",
       "      <th>F_z</th>\n",
       "      <th>T_x</th>\n",
       "      <th>T_y</th>\n",
       "      <th>T_z</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th>timestep</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">0</th>\n",
       "      <th>1</th>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>62</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>61</td>\n",
       "      <td>-3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-2</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>63</td>\n",
       "      <td>-3</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             F_x  F_y  F_z  T_x  T_y  T_z\n",
       "id timestep                              \n",
       "0  1          -1   -1   63   -3   -1    0\n",
       "   2           0    0   62   -3   -1    0\n",
       "   3          -1   -1   61   -3    0    0\n",
       "   4          -1   -1   63   -2   -1    0\n",
       "   5          -1   -1   63   -3   -1    0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = X.set_index(['id', 'timestep'])\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "16b5af67-c01c-4793-bd19-1d79732fa19b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "label\n",
       "obstruction     34\n",
       "normal          21\n",
       "collision       17\n",
       "fr_collision    16\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49017be7-25d7-4346-a471-27d2912f2c41",
   "metadata": {},
   "source": [
    "### Train/test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9ba9f310-b205-49b0-aea8-8c6ab84bdf22",
   "metadata": {},
   "outputs": [],
   "source": [
    "obstruction_indices = y.loc[y == 'obstruction'].index\n",
    "normal_indices = y.loc[y == 'normal'].index\n",
    "collision_indices = y.loc[y == 'collision'].index\n",
    "fr_collision_indices = y.loc[y == 'fr_collision'].index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "7464da97-377f-4577-b514-f91b943460b1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index([35, 36, 37, 38, 39, 41, 43, 44, 45, 46, 47, 48, 49, 50, 51, 53, 57,  0,\n",
      "        1,  2,  3,  4,  5,  6,  7,  8,  9, 18, 19, 20, 21, 22, 23, 24, 25, 52,\n",
      "       54, 55, 56, 58, 62, 68, 69],\n",
      "      dtype='int64')\n",
      "Index([59, 60, 61, 63, 64, 65, 66, 70, 72, 73, 75, 77, 78, 80, 83, 85, 86, 10,\n",
      "       11, 12, 13, 14, 15, 16, 17, 40, 42, 67, 26, 27, 28, 29, 30, 31, 32, 33,\n",
      "       34, 71, 74, 76, 79, 81, 82, 84, 87],\n",
      "      dtype='int64')\n"
     ]
    }
   ],
   "source": [
    "split_idx = math.floor(len(obstruction_indices)/2)\n",
    "train_obst_idx = pd.Index(obstruction_indices.tolist()[:split_idx])\n",
    "test_obst_idx = pd.Index(obstruction_indices.tolist()[split_idx:])\n",
    "\n",
    "split_idx = math.floor(len(normal_indices)/2)\n",
    "train_normal_idx = pd.Index(normal_indices.tolist()[:split_idx])\n",
    "test_normal_idx = pd.Index(normal_indices.tolist()[split_idx:])\n",
    "\n",
    "split_idx = math.floor(len(collision_indices)/2)\n",
    "train_col_idx = pd.Index(collision_indices.tolist()[:split_idx])\n",
    "test_col_idx = pd.Index(collision_indices.tolist()[split_idx:])\n",
    "\n",
    "split_idx = math.floor(len(fr_collision_indices)/2)\n",
    "train_frcol_idx = pd.Index(fr_collision_indices.tolist()[:split_idx])\n",
    "test_frcol_idx = pd.Index(fr_collision_indices.tolist()[split_idx:])\n",
    "\n",
    "train_index = pd.Index(pd.concat([pd.Series(idx) for idx in [train_obst_idx, train_normal_idx, train_col_idx, train_frcol_idx]]))\n",
    "test_index = pd.Index(pd.concat([pd.Series(idx) for idx in [test_obst_idx, test_normal_idx, test_col_idx, test_frcol_idx]]))\n",
    "\n",
    "print(train_index)\n",
    "print(test_index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "caa70a31-0787-4698-b8af-3db26b0c5ab9",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train = X.loc[train_index], y.iloc[train_index]\n",
    "X_test, y_test = X.loc[test_index], y.iloc[test_index]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3e3d880-31e4-45d7-bffd-2b42d8f9fb69",
   "metadata": {},
   "source": [
    "## 2. Modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8906a657-1494-4838-b032-0b1b2da4935f",
   "metadata": {},
   "outputs": [],
   "source": [
    "label_mapping = {\n",
    "    'obstruction': 0,\n",
    "    'normal': 1,\n",
    "    'collision': 2,\n",
    "    'fr_collision': 3\n",
    "}\n",
    "\n",
    "y_train = y_train.map(label_mapping)\n",
    "y_test = y_test.map(label_mapping)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd2515b3-1651-440f-a261-b21e810064ab",
   "metadata": {},
   "source": [
    "### Bagging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9a3d1955-f7cd-47ea-807e-4004102b8da6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.46      0.71      0.56        17\n",
      "           1       0.82      0.82      0.82        11\n",
      "           2       1.00      0.11      0.20         9\n",
      "           3       0.29      0.25      0.27         8\n",
      "\n",
      "    accuracy                           0.53        45\n",
      "   macro avg       0.64      0.47      0.46        45\n",
      "weighted avg       0.63      0.53      0.50        45\n",
      "\n"
     ]
    }
   ],
   "source": [
    "base_clf = WEASEL(alphabet_size=3,support_probabilities=True, random_state=42)\n",
    "\n",
    "clf = BaggingClassifier(base_clf, n_estimators=6, n_features=1, random_state=42)\n",
    "\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "clf_report = classification_report(y_test, y_pred)\n",
    "\n",
    "precision_bagging, recall_bagging, fscore_bagging, _ = precision_recall_fscore_support(y_test, y_pred, average='weighted')\n",
    "\n",
    "print(clf_report)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16909df1-de85-44e5-bdd4-503023eab0fd",
   "metadata": {},
   "source": [
    "### Summary classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1e5f2065-8450-4bcd-b375-8d61447763b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision 0.7805982905982906 \n",
      "Recall: 0.6222222222222222 \n",
      "F1-Score: 0.6196712487035068 \n",
      "===================\n",
      "Precision 0.6914529914529914 \n",
      "Recall: 0.6222222222222222 \n",
      "F1-Score: 0.5890793650793651 \n",
      "===================\n",
      "Precision 0.8459259259259259 \n",
      "Recall: 0.7333333333333333 \n",
      "F1-Score: 0.7297517297517297 \n",
      "===================\n",
      "Precision 0.45045045045045046 \n",
      "Recall: 0.4222222222222222 \n",
      "F1-Score: 0.3538148148148148 \n",
      "===================\n"
     ]
    }
   ],
   "source": [
    "knn_clf = KNeighborsClassifier(n_neighbors=1)\n",
    "ada_clf = AdaBoostClassifier(n_estimators=10, random_state=42)\n",
    "rf_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "svc_clf = SVC()\n",
    "\n",
    "classifiers = [knn_clf, ada_clf, rf_clf, svc_clf]\n",
    "\n",
    "for classifier in classifiers:\n",
    "    clf = SummaryClassifier(\n",
    "        summary_functions=(\"mean\", \"min\", \"max\", \"median\", \"sum\", \"skew\", \"kurt\", \"var\", \"std\", \"mad\", \"sem\", \"nunique\", \"count\"),\n",
    "        estimator=classifier,\n",
    "        n_jobs=-1,\n",
    "        random_state=42\n",
    "    )\n",
    "    \n",
    "    clf.fit(X_train, y_train)\n",
    "    \n",
    "    y_pred = clf.predict(X_test)\n",
    "    \n",
    "    precision, recall, fscore, _ = precision_recall_fscore_support(y_test, y_pred, average='weighted')\n",
    "    \n",
    "    print(f\"Precision {precision} \\nRecall: {recall} \\nF1-Score: {fscore} \\n===================\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a41a54a-030d-4436-8a08-52f0932900cc",
   "metadata": {},
   "source": [
    "Here, random forest achieves the best result, with a precision of 0.85, recall of 0.73 and F1-Score of 0.73."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ee59ab7-4d62-4f1a-8b2d-0b35f0b0cfb2",
   "metadata": {},
   "source": [
    "### TSFresh"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "2e14406c-bf14-4c37-b94a-b8a15064fea1",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.read_csv('data/lp1_X.csv')\n",
    "y = pd.read_csv('data/lp1_y.csv').squeeze()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "42e0b801-2b9b-49d4-840b-0dbb22f5bb36",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Feature Extraction: 100%|██████████████████████████████████████████████████████████████| 30/30 [00:05<00:00,  5.60it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision 0.8751262626262627 \n",
      "Recall: 0.8222222222222222 \n",
      "F1-Score: 0.8131801181414461 \n",
      "===================\n",
      "Precision 0.8438095238095239 \n",
      "Recall: 0.7333333333333333 \n",
      "F1-Score: 0.7091986113725244 \n",
      "===================\n",
      "Precision 0.8161236424394319 \n",
      "Recall: 0.6666666666666666 \n",
      "F1-Score: 0.642178077066966 \n",
      "===================\n",
      "Precision 0.14271604938271606 \n",
      "Recall: 0.37777777777777777 \n",
      "F1-Score: 0.207168458781362 \n",
      "===================\n"
     ]
    }
   ],
   "source": [
    "X_feat = extract_features(X, column_id='id', column_sort='timestep')\n",
    "X_feat = X_feat.dropna(axis=1)\n",
    "\n",
    "X_train, y_train = X_feat.iloc[train_index], y.iloc[train_index]\n",
    "X_test, y_test = X_feat.iloc[test_index], y.iloc[test_index]\n",
    "\n",
    "X_train_filtered = select_features(X_train, y_train, multiclass=True, n_significant=1)\n",
    "X_test_filtered = X_test[X_train_filtered.columns]\n",
    "\n",
    "classifiers = [knn_clf, ada_clf, rf_clf, svc_clf]\n",
    "\n",
    "for classifier in classifiers:\n",
    "       \n",
    "    classifier.fit(X_train_filtered, y_train)\n",
    "    \n",
    "    y_pred = classifier.predict(X_test_filtered)\n",
    "    \n",
    "    precision, recall, fscore, _ = precision_recall_fscore_support(y_test, y_pred, average='weighted')\n",
    "    \n",
    "    print(f\"Precision {precision} \\nRecall: {recall} \\nF1-Score: {fscore} \\n===================\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d047ec00-ce0e-4167-8bea-4bbfde506288",
   "metadata": {},
   "source": [
    "## 3. Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "7159e5cd-6ce3-4cda-ab14-7b8f7d1ae6a0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnYAAAHWCAYAAAD6oMSKAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAAA9hAAAPYQGoP6dpAABKuklEQVR4nO3df3zN9f//8fvZ2IYxvzfTWPlNzAwzEmo15aOUfsivURRFZZUfhVVi0htTvPMOQ+8S/fCjd7z9aFlvTAhLMuS3ettYYgwb2/P7h+/Ou2PDDtsOL7fr5XIul17P1/P1ej1ep3Pm/nq+fhybMcYIAAAANz03VxcAAACAwkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiXBrv//Oc/6ty5s/z9/WWz2bR48eKrLpOQkKBmzZrJ09NTtWvX1pw5c4q8TgAAgJuBS4NdRkaGgoKCNG3atAL1379/vzp16qQOHTooKSlJL7/8svr166cVK1YUcaUAAAA3Ppsxxri6CEmy2WxatGiRunTpctk+w4YN09KlS7V9+3Z7W7du3XTixAktX768GKoEAAC4cZVwdQHOWL9+vcLDwx3aIiIi9PLLL192mczMTGVmZtqnc3JydPz4cVWqVEk2m62oSgUAACgUxhidOnVK/v7+cnO78snWmyrYpaSkyNfX16HN19dX6enpOnv2rEqVKpVnmZiYGL311lvFVSIAAECROHz4sG677bYr9rmpgt21GDFihKKiouzTJ0+eVI0aNXT48GGVK1fOhZUBAABcXXp6ugICAlS2bNmr9r2pgp2fn59SU1Md2lJTU1WuXLl8R+skydPTU56ennnay5UrR7ADAAA3jYJcQnZTPccuLCxM8fHxDm2rVq1SWFiYiyoCAAC4cbg02J0+fVpJSUlKSkqSdPFxJklJSTp06JCki6dRe/fube8/YMAA7du3T0OHDtXOnTv197//XZ9//rmGDBniivIBAABuKC4Ndj/++KOCg4MVHBwsSYqKilJwcLBGjx4tSTpy5Ig95EnS7bffrqVLl2rVqlUKCgrSxIkTNXPmTEVERLikfgAAgBvJDfMcu+KSnp4uHx8fnTx5kmvsAADADc+Z7HJTXWMHAACAyyPYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AABuEdOmTVNgYKC8vLwUGhqqjRs3XrF/bGys6tWrp1KlSikgIEBDhgzRuXPn7PP/85//qHPnzvL395fNZtPixYuLeA9wNQQ7AABuAQsWLFBUVJSio6O1ZcsWBQUFKSIiQkePHs23/7x58zR8+HBFR0crOTlZs2bN0oIFC/T666/b+2RkZCgoKEjTpk0rrt3AVdiMMcbVRRSn9PR0+fj46OTJkypXrpyrywEAoFiEhoaqRYsWmjp1qiQpJydHAQEBGjx4sIYPH56n/6BBg5ScnKz4+Hh72yuvvKINGzZo7dq1efrbbDYtWrRIXbp0KbJ9uFU5k10YsQMAwOKysrK0efNmhYeH29vc3NwUHh6u9evX57tM69attXnzZvvp2n379mnZsmV68MEHi6VmXJsSri4AAAAUrbS0NGVnZ8vX19eh3dfXVzt37sx3me7duystLU133XWXjDG6cOGCBgwY4HAqFjceRuwAAEAeCQkJGjdunP7+979ry5YtWrhwoZYuXaoxY8a4ujRcASN2AABYXOXKleXu7q7U1FSH9tTUVPn5+eW7zKhRo9SrVy/169dPktS4cWNlZGTo2Wef1RtvvCE3N8aGbkT8XwEAwOI8PDwUEhLicCNETk6O4uPjFRYWlu8yZ86cyRPe3N3dJUm32H2XNxVG7AAAuAVERUUpMjJSzZs3V8uWLRUbG6uMjAz17dtXktS7d29Vr15dMTExkqTOnTtr0qRJCg4OVmhoqPbs2aNRo0apc+fO9oB3+vRp7dmzx76N/fv3KykpSRUrVlSNGjWKfydBsAMA4Fbw5JNP6tixYxo9erRSUlLUtGlTLV++3H5DxaFDhxxG6EaOHCmbzaaRI0fq999/V5UqVdS5c2eNHTvW3ufHH39Uhw4d7NNRUVGSpMjISM2ZM6d4dgwOeI4dAADADYzn2AEAANyCCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAW4fJgN23aNAUGBsrLy0uhoaHauHHjFfvHxsaqXr16KlWqlAICAjRkyBCdO3eumKoFAAC4cZVw5cYXLFigqKgoTZ8+XaGhoYqNjVVERIR27dqlqlWr5uk/b948DR8+XHFxcWrdurV2796tPn36yGazadKkSS7YAwDAzSxw+FJXlwALODC+k6tLsHPpiN2kSZPUv39/9e3bVw0bNtT06dNVunRpxcXF5ds/MTFRbdq0Uffu3RUYGKj7779fTz311FVH+QAAAG4FLgt2WVlZ2rx5s8LDw/9XjJubwsPDtX79+nyXad26tTZv3mwPcvv27dOyZcv04IMPFkvNAG4tzlwq0r59e9lstjyvTp3+dyT/5ptvqn79+ipTpowqVKig8PBwbdiwoTh2BcAtwmWnYtPS0pSdnS1fX1+Hdl9fX+3cuTPfZbp37660tDTdddddMsbowoULGjBggF5//fXLbiczM1OZmZn26fT09MLZAQCW5uylIgsXLlRWVpZ9+o8//lBQUJAef/xxe1vdunU1depU3XHHHTp79qwmT56s+++/X3v27FGVKlWKZb8AWJvLb55wRkJCgsaNG6e///3v2rJlixYuXKilS5dqzJgxl10mJiZGPj4+9ldAQEAxVgzgZuXspSIVK1aUn5+f/bVq1SqVLl3aIdh1795d4eHhuuOOO9SoUSNNmjRJ6enp2rZtW3HtFgCLc1mwq1y5stzd3ZWamurQnpqaKj8/v3yXGTVqlHr16qV+/fqpcePGeuSRRzRu3DjFxMQoJycn32VGjBihkydP2l+HDx8u9H0BYC3XcqnIpWbNmqVu3bqpTJkyl93GRx99JB8fHwUFBRVK3QDgsmDn4eGhkJAQxcfH29tycnIUHx+vsLCwfJc5c+aM3NwcS3Z3d5ckGWPyXcbT01PlypVzeAHAlVzpUpGUlJSrLr9x40Zt375d/fr1yzPvm2++kbe3t7y8vDR58mStWrVKlStXLrTaAdzaXHoqNioqSjNmzNDcuXOVnJysgQMHKiMjQ3379pUk9e7dWyNGjLD379y5sz788EPNnz9f+/fv16pVqzRq1Ch17tzZHvAAwNVmzZqlxo0bq2XLlnnmdejQQUlJSUpMTFTHjh31xBNP6OjRoy6oEoAVufQ5dk8++aSOHTum0aNHKyUlRU2bNtXy5cvtR8mHDh1yGKEbOXKkbDabRo4cqd9//11VqlRR586dNXbsWFftAgALupZLRXJlZGRo/vz5evvtt/OdX6ZMGdWuXVu1a9dWq1atVKdOHc2aNcvhIBYArpVLg50kDRo0SIMGDcp3XkJCgsN0iRIlFB0drejo6GKoDMCt6q+XinTp0kXS/y4Vudzfq1xffPGFMjMz1bNnzwJtKycnx+HOfQC4Hi4PdgBwI4qKilJkZKSaN2+uli1bKjY2Ns+lItWrV1dMTIzDcrNmzVKXLl1UqVIlh/aMjAyNHTtWDz30kKpVq6a0tDRNmzZNv//+u8OdswBwPQh2AJAPZy8VkaRdu3Zp7dq1WrlyZZ71ubu7a+fOnZo7d67S0tJUqVIltWjRQmvWrFGjRo2KZZ8AWJ/NXO52UotKT0+Xj4+PTp48yR2yAHCL47diURiK+rdinckuN9UDigEAAHB5BDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYRAlXFwDg5hE4fKmrS8BN7sD4Tq4uAbA0RuwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALAIgh0AAIBFEOwAAAAsgmAHAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAMAALCIEs50Tk5O1vz587VmzRodPHhQZ86cUZUqVRQcHKyIiAh17dpVnp6eRVUrAAAArqBAI3ZbtmxReHi4goODtXbtWoWGhurll1/WmDFj1LNnTxlj9MYbb8jf31/vvvuuMjMzi7puAAAAXKJAI3Zdu3bVa6+9pi+//FLly5e/bL/169drypQpmjhxol5//fXCqhEAAAAFUKBgt3v3bpUsWfKq/cLCwhQWFqbz589fd2EAAABwToFOxRYk1F1PfwAAAFy/Ao3Yvf/++wVe4YsvvnjNxQAAAODaFSjYTZ482WH62LFjOnPmjP16uxMnTqh06dKqWrUqwQ4AAMBFCnQqdv/+/fbX2LFj1bRpUyUnJ+v48eM6fvy4kpOT1axZM40ZM6ao6wUAAMBlOP2A4lGjRumDDz5QvXr17G316tXT5MmTNXLkyEItDgAAAAXndLA7cuSILly4kKc9OztbqamphVIUAAAAnOd0sLv33nv13HPPacuWLfa2zZs3a+DAgQoPDy/U4gAAAFBwTge7uLg4+fn5qXnz5vL09JSnp6datmwpX19fzZw5syhqBAAAQAE49VuxklSlShUtW7ZMu3fv1s6dOyVJ9evXV926dQu9OAAAABSc08EuV2BgoIwxqlWrlkqUuObVAAAAoJA4fSr2zJkzeuaZZ1S6dGk1atRIhw4dkiQNHjxY48ePL/QCAQAAUDBOB7sRI0bop59+UkJCgry8vOzt4eHhWrBgQaEWBwAAgIJz+hzq4sWLtWDBArVq1Uo2m83e3qhRI+3du7dQiwMAAEDBOT1id+zYMVWtWjVPe0ZGhkPQAwAAQPFyOtg1b95cS5cutU/nhrmZM2cqLCys8CoDAACAU5w+FTtu3Dg98MAD2rFjhy5cuKApU6Zox44dSkxM1Pfff18UNQIAAKAAnB6xu+uuu5SUlKQLFy6ocePGWrlypapWrar169crJCSkKGoEAABAAVzTA+hq1aqlGTNmFHYtAAAAuA5Oj9i5u7vr6NGjedr/+OMPubu7F0pRAAAAcJ7Twc4Yk297ZmamPDw8nC5g2rRpCgwMlJeXl0JDQ7Vx48Yr9j9x4oReeOEFVatWTZ6enqpbt66WLVvm9HYBAACspsCnYt9//31JF++CnTlzpry9ve3zsrOz9Z///Ef169d3auMLFixQVFSUpk+frtDQUMXGxioiIkK7du3K95EqWVlZuu+++1S1alV9+eWXql69ug4ePKjy5cs7tV0AAAArKnCwmzx5sqSLI3bTp093OO3q4eGhwMBATZ8+3amNT5o0Sf3791ffvn0lSdOnT9fSpUsVFxen4cOH5+kfFxen48ePKzExUSVLlpR08TdrAQAA4ESw279/vySpQ4cOWrhwoSpUqHBdG87KytLmzZs1YsQIe5ubm5vCw8O1fv36fJf5+uuvFRYWphdeeEFLlixRlSpV1L17dw0bNozr+wAAwC3P6btiV69eXSgbTktLU3Z2tnx9fR3afX19tXPnznyX2bdvn7777jv16NFDy5Yt0549e/T888/r/Pnzio6OzneZzMxMZWZm2qfT09MLpX4AAIAbjdPBLjs7W3PmzFF8fLyOHj2qnJwch/nfffddoRV3qZycHFWtWlUfffSR3N3dFRISot9//13vvffeZYNdTEyM3nrrrSKrCQAA4EbhdLB76aWXNGfOHHXq1El33nnnNf8+bOXKleXu7q7U1FSH9tTUVPn5+eW7TLVq1VSyZEmH064NGjRQSkqKsrKy8r0rd8SIEYqKirJPp6enKyAg4JpqBgAAuJE5Hezmz5+vzz//XA8++OB1bdjDw0MhISGKj49Xly5dJF0ckYuPj9egQYPyXaZNmzaaN2+ecnJy5OZ28Uktu3fvVrVq1S77qBVPT095enpeV60AAAA3A6efY+fh4aHatWsXysajoqI0Y8YMzZ07V8nJyRo4cKAyMjLsd8n27t3b4eaKgQMH6vjx43rppZe0e/duLV26VOPGjdMLL7xQKPUAAADczJwesXvllVc0ZcoUTZ069ZpPw+Z68skndezYMY0ePVopKSlq2rSpli9fbr+h4tChQ/aROUkKCAjQihUrNGTIEDVp0kTVq1fXSy+9pGHDhl1XHQAAAFZgM5f7KYm/ePTRRx2mv/vuO1WsWFGNGjWyP08u18KFCwu3wkKWnp4uHx8fnTx5UuXKlXN1OcBNJXD4UleXgJvcgfGdXF2CAz7TKAxF/bl2JrsUaMTOx8fHYfqRRx659uoAAABQJAoU7GbPnl3UdQAAAOA6OX3zBAAAAG5MTt88ERwcnO9NEzabTV5eXqpdu7b69OmjDh06FEqBAAAAKBinR+w6duyoffv2qUyZMurQoYM6dOggb29v7d27Vy1atNCRI0cUHh6uJUuWFEW9uIlMmzZNgYGB8vLyUmhoqDZu3HjZvnPmzJHNZnN4eXl5FWO1AADc/JwesUtLS9Mrr7yiUaNGObS/8847OnjwoFauXKno6GiNGTNGDz/8cKEVipvLggULFBUVpenTpys0NFSxsbGKiIjQrl27VLVq1XyXKVeunHbt2mWfvt7H6QAAcKtxesTu888/11NPPZWnvVu3bvr8888lSU899ZTDP9C49UyaNEn9+/dX37591bBhQ02fPl2lS5dWXFzcZZex2Wzy8/Ozv3KfZwgAAArG6WDn5eWlxMTEPO2JiYn2U2c5OTmcRruFZWVlafPmzQoPD7e3ubm5KTw8XOvXr7/scqdPn1bNmjUVEBCghx9+WL/88ktxlAsAgGU4fSp28ODBGjBggDZv3qwWLVpIkjZt2qSZM2fq9ddflyStWLFCTZs2LdRCcfNIS0tTdnZ2nhE3X19f7dy5M99l6tWrp7i4ODVp0kQnT57U3/72N7Vu3Vq//PKLbrvttuIoGwCAm57TwW7kyJG6/fbbNXXqVP3zn/+UdPEf5RkzZqh79+6SpAEDBmjgwIGFWyksLSwsTGFhYfbp1q1bq0GDBvrHP/6hMWPGuLAyAABuHk4HO0nq0aOHevTocdn5pUqVuuaCcPOrXLmy3N3dlZqa6tCempoqPz+/Aq2jZMmSCg4O1p49e4qiRAAALIkHFKPQeXh4KCQkRPHx8fa2nJwcxcfHO4zKXUl2drZ+/vlnVatWrajKBADAcgo0YlexYkXt3r1blStXVoUKFa74GIrjx48XWnG4eUVFRSkyMlLNmzdXy5YtFRsbq4yMDPXt21eS1Lt3b1WvXl0xMTGSpLffflutWrVS7dq1deLECb333ns6ePCg+vXr58rdAADgplKgYDd58mSVLVtWkhQbG1uU9cAinnzySR07dkyjR49WSkqKmjZtquXLl9tvqDh06JDc3P43YPznn3+qf//+SklJUYUKFRQSEqLExEQ1bNjQVbsAAMBNx2aMMa4uojilp6fLx8dHJ0+eVLly5VxdDnBTCRy+1NUl4CZ3YHwnV5fggM80CkNRf66dyS7XdI3d3r17NXLkSD311FM6evSoJOnf//43zx0DAABwIaeD3ffff6/GjRtrw4YNWrhwoU6fPi1J+umnnxQdHV3oBQIAAKBgnA52w4cP1zvvvKNVq1bJw8PD3n7PPffohx9+KNTiAAAAUHBOB7uff/5ZjzzySJ72qlWrKi0trVCKAgAAgPOcDnbly5fXkSNH8rRv3bpV1atXL5SiAAAA4Dyng123bt00bNgwpaSkyGazKScnR+vWrdOrr76q3r17F0WNAAAAKACng924ceNUv359BQQE6PTp02rYsKHuvvtutW7dWiNHjiyKGgEAAFAATv9WrIeHh2bMmKFRo0Zp+/btOn36tIKDg1WnTp2iqA8AAAAF5HSw27dvn+644w7VqFFDNWrUKIqaAAAAcA2cDna1a9fWbbfdpnbt2ql9+/Zq166dateuXRS1AQAAwAlOX2N3+PBhxcTEqFSpUpowYYLq1q2r2267TT169NDMmTOLokYAAAAUgNPBrnr16urRo4c++ugj7dq1S7t27VJ4eLg+//xzPffcc0VRIwAAAArA6VOxZ86c0dq1a5WQkKCEhARt3bpV9evX16BBg9S+ffsiKBEAAAAF4XSwK1++vCpUqKAePXpo+PDhatu2rSpUqFAUtQEAAMAJTge7Bx98UGvXrtX8+fOVkpKilJQUtW/fXnXr1i2K+gAAAFBATl9jt3jxYqWlpWn58uUKCwvTypUr1bZtW/u1dwAAAHANp0fscjVu3FgXLlxQVlaWzp07pxUrVmjBggX69NNPC7M+AAAAFJDTI3aTJk3SQw89pEqVKik0NFSfffaZ6tatq6+++krHjh0rihoBAABQAE6P2H322Wdq166dnn32WbVt21Y+Pj5FURcAAACc5HSw27RpU1HUYUmBw5e6ugTc5A6M7+TqEgAAN5ECnYo9dOiQUyv9/fffr6kYAAAAXLsCBbsWLVroueeeu+Jo3cmTJzVjxgzdeeed+uqrrwqtQAAAABRMgU7F7tixQ2PHjtV9990nLy8vhYSEyN/fX15eXvrzzz+1Y8cO/fLLL2rWrJkmTJigBx98sKjrBgAAwCUKNGJXqVIlTZo0SUeOHNHUqVNVp04dpaWl6ddff5Uk9ejRQ5s3b9b69esJdQAAAC7i1M0TpUqV0mOPPabHHnusqOoBAADANXL6OXYAAAC4MRHsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFjENQW7f/7zn2rTpo38/f118OBBSVJsbKyWLFlSqMUBAACg4JwOdh9++KGioqL04IMP6sSJE8rOzpYklS9fXrGxsYVdHwAAAArI6WD3wQcfaMaMGXrjjTfk7u5ub2/evLl+/vnnQi0OAAAABed0sNu/f7+Cg4PztHt6eiojI6NQigIAAIDznA52t99+u5KSkvK0L1++XA0aNCiMmgAAAHANnPpJMUmKiorSCy+8oHPnzskYo40bN+qzzz5TTEyMZs6cWRQ1AgAAoACcDnb9+vVTqVKlNHLkSJ05c0bdu3eXv7+/pkyZom7duhVFjQAAACgAp4LdhQsXNG/ePEVERKhHjx46c+aMTp8+rapVqxZVfQAAACggp66xK1GihAYMGKBz585JkkqXLk2oAwAAuEE4ffNEy5YttXXr1qKoBQAAANfB6Wvsnn/+eb3yyiv67bffFBISojJlyjjMb9KkSaEVBwAAgIJzOtjl3iDx4osv2ttsNpuMMbLZbPZfogAAAEDxcjrY7d+/vyjqAAAAwHVyOtjVrFmzKOoAAADAdXI62EnS3r17FRsbq+TkZElSw4YN9dJLL6lWrVqFWhwAAAAKzum7YlesWKGGDRtq48aNatKkiZo0aaINGzaoUaNGWrVqVVHUCAAAgAJwesRu+PDhGjJkiMaPH5+nfdiwYbrvvvsKrTgAAAAUnNMjdsnJyXrmmWfytD/99NPasWNHoRQFAAAA5zkd7KpUqaKkpKQ87UlJSfwKBQAAgAs5fSq2f//+evbZZ7Vv3z61bt1akrRu3Tq9++67ioqKKvQCAQAAUDBOB7tRo0apbNmymjhxokaMGCFJ8vf315tvvunw0GIAAAAUL6eDnc1m05AhQzRkyBCdOnVKklS2bNlCLwwAAADOuaZfnrhw4YLq1KnjEOh+/fVXlSxZUoGBgYVZHwAAAArI6Zsn+vTpo8TExDztGzZsUJ8+fQqjJgAAAFwDp4Pd1q1b1aZNmzztrVq1yvduWQAAABQPp4OdzWazX1v3VydPnlR2dvY1FTFt2jQFBgbKy8tLoaGh2rhxY4GWmz9/vmw2m7p06XJN2wUAALASp4Pd3XffrZiYGIcQl52drZiYGN11111OF7BgwQJFRUUpOjpaW7ZsUVBQkCIiInT06NErLnfgwAG9+uqratu2rdPbBAAAsCKnb5549913dffdd6tevXr2ULVmzRqlp6fru+++c7qASZMmqX///urbt68kafr06Vq6dKni4uI0fPjwfJfJzs5Wjx499NZbb2nNmjU6ceKE09sFAACwGqdH7Bo2bKht27bpiSee0NGjR3Xq1Cn17t1bO3fu1J133unUurKysrR582aFh4f/ryA3N4WHh2v9+vWXXe7tt99W1apV8/1ps0tlZmYqPT3d4QUAAGBFTo/YSRcfSDxu3Ljr3nhaWpqys7Pl6+vr0O7r66udO3fmu8zatWs1a9asAt+oERMTo7feeut6SwUAALjhFXjELi0tTQcPHnRo++WXX9S3b1898cQTmjdvXqEXd6lTp06pV69emjFjhipXrlygZUaMGKGTJ0/aX4cPHy7iKgEAAFyjwCN2gwcPlr+/vyZOnChJOnr0qNq2bSt/f3/VqlVLffr0UXZ2tnr16lXgjVeuXFnu7u5KTU11aE9NTZWfn1+e/nv37tWBAwfUuXNne1tOTs7FHSlRQrt27VKtWrUclvH09JSnp2eBawIAALhZFXjE7ocfftBDDz1kn/74449VsWJFJSUlacmSJRo3bpymTZvm1MY9PDwUEhKi+Ph4e1tOTo7i4+MVFhaWp3/9+vX1888/Kykpyf566KGH1KFDByUlJSkgIMCp7QMAAFhJgUfsUlJSHH4u7LvvvtOjjz6qEiUuruKhhx5STEyM0wVERUUpMjJSzZs3V8uWLRUbG6uMjAz7XbK9e/dW9erVFRMTIy8vrzw3aJQvX16SnL5xAwAAwGoKHOzKlSunEydOqGbNmpKkjRs3OtyVarPZlJmZ6XQBTz75pI4dO6bRo0crJSVFTZs21fLly+03VBw6dEhubk7fvAsAAHDLKXCwa9Wqld5//33NmDFDCxcu1KlTp3TPPffY5+/evfuaT4UOGjRIgwYNyndeQkLCFZedM2fONW0TAADAagoc7MaMGaN7771Xn3zyiS5cuKDXX39dFSpUsM+fP3++2rVrVyRFAgAA4OoKHOyaNGmi5ORkrVu3Tn5+fgoNDXWY361bNzVs2LDQCwQAAEDBOPWA4sqVK+vhhx+2T//222/y9/eXm5ubOnXqVOjFAQAAoOCu666Ehg0b6sCBA4VUCgAAAK7HdQU7Y0xh1QEAAIDrxHNEAAAALOK6gt3rr7+uihUrFlYtAAAAuA5O3TxxqREjRhRWHQAAALhOhXYq9vDhw3r66acLa3UAAABwUqEFu+PHj2vu3LmFtToAAAA4qcCnYr/++usrzt+3b991FwMAAIBrV+Bg16VLF9lstis+4sRmsxVKUQAAAHBegU/FVqtWTQsXLlROTk6+ry1bthRlnQAAALiKAge7kJAQbd68+bLzrzaaBwAAgKJV4FOxr732mjIyMi47v3bt2lq9enWhFAUAAADnFTjYtW3b9orzy5Qpo3bt2l13QQAAALg2BT4Vu2/fPk61AgAA3MAKHOzq1KmjY8eO2aeffPJJpaamFklRAAAAcF6Bg92lo3XLli274jV3AAAAKF6F9ssTAAAAcK0CBzubzZbnAcQ8kBgAAODGUeC7Yo0x6tOnjzw9PSVJ586d04ABA1SmTBmHfgsXLizcCgEAAFAgBQ52kZGRDtM9e/Ys9GIAAABw7Qoc7GbPnl2UdQAAAOA6cfMEAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFjEDRHspk2bpsDAQHl5eSk0NFQbN268bN8ZM2aobdu2qlChgipUqKDw8PAr9gcAALhVuDzYLViwQFFRUYqOjtaWLVsUFBSkiIgIHT16NN/+CQkJeuqpp7R69WqtX79eAQEBuv/++/X7778Xc+UAAAA3FpcHu0mTJql///7q27evGjZsqOnTp6t06dKKi4vLt/+nn36q559/Xk2bNlX9+vU1c+ZM5eTkKD4+vpgrBwAAuLG4NNhlZWVp8+bNCg8Pt7e5ubkpPDxc69evL9A6zpw5o/Pnz6tixYr5zs/MzFR6errDCwAAwIpcGuzS0tKUnZ0tX19fh3ZfX1+lpKQUaB3Dhg2Tv7+/Qzj8q5iYGPn4+NhfAQEB1103AADAjcjlp2Kvx/jx4zV//nwtWrRIXl5e+fYZMWKETp48aX8dPny4mKsEAAAoHiVcufHKlSvL3d1dqampDu2pqany8/O74rJ/+9vfNH78eH377bdq0qTJZft5enrK09OzUOoFAAC4kbl0xM7Dw0MhISEONz7k3ggRFhZ22eUmTJigMWPGaPny5WrevHlxlAoAAHDDc+mInSRFRUUpMjJSzZs3V8uWLRUbG6uMjAz17dtXktS7d29Vr15dMTExkqR3331Xo0eP1rx58xQYGGi/Fs/b21ve3t4u2w8AAABXc3mwe/LJJ3Xs2DGNHj1aKSkpatq0qZYvX26/oeLQoUNyc/vfwOKHH36orKwsPfbYYw7riY6O1ptvvlmcpQMAANxQXB7sJGnQoEEaNGhQvvMSEhIcpg8cOFD0BQEAANyEbuq7YgEAAPA/BDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFkGwAwAAsAiCHQAAgEUQ7AAAACyCYAcAAGARBDsAAACLINgBAABYBMEOAADAIgh2AAAAFnFDBLtp06YpMDBQXl5eCg0N1caNG6/Y/4svvlD9+vXl5eWlxo0ba9myZcVUKQAAwI3L5cFuwYIFioqKUnR0tLZs2aKgoCBFRETo6NGj+fZPTEzUU089pWeeeUZbt25Vly5d1KVLF23fvr2YKwcAALixuDzYTZo0Sf3791ffvn3VsGFDTZ8+XaVLl1ZcXFy+/adMmaKOHTvqtddeU4MGDTRmzBg1a9ZMU6dOLebKAQAAbiwuDXZZWVnavHmzwsPD7W1ubm4KDw/X+vXr811m/fr1Dv0lKSIi4rL9AQAAbhUlXLnxtLQ0ZWdny9fX16Hd19dXO3fuzHeZlJSUfPunpKTk2z8zM1OZmZn26ZMnT0qS0tPTr6f0AsnJPFPk24C1Fcfn1Bl8pnG9+EzDior6c527fmPMVfu6NNgVh5iYGL311lt52gMCAlxQDeAcn1hXVwAULj7TsKLi+lyfOnVKPj4+V+zj0mBXuXJlubu7KzU11aE9NTVVfn5++S7j5+fnVP8RI0YoKirKPp2Tk6Pjx4+rUqVKstls17kHuFbp6ekKCAjQ4cOHVa5cOVeXAxQKPtewGj7TNwZjjE6dOiV/f/+r9nVpsPPw8FBISIji4+PVpUsXSReDV3x8vAYNGpTvMmFhYYqPj9fLL79sb1u1apXCwsLy7e/p6SlPT0+HtvLlyxdG+SgE5cqV448FLIfPNayGz7TrXW2kLpfLT8VGRUUpMjJSzZs3V8uWLRUbG6uMjAz17dtXktS7d29Vr15dMTExkqSXXnpJ7dq108SJE9WpUyfNnz9fP/74oz766CNX7gYAAIDLuTzYPfnkkzp27JhGjx6tlJQUNW3aVMuXL7ffIHHo0CG5uf3v5t3WrVtr3rx5GjlypF5//XXVqVNHixcv1p133umqXQAAALgh2ExBbrEACllmZqZiYmI0YsSIPKfKgZsVn2tYDZ/pmw/BDgAAwCJc/ssTAAAAKBwEOwAAAIsg2MGuffv2Do+RKUq7du2Sn5+fTp06VSzbuxGkpaWpatWq+u2331xdCi4xatQoPfvss64uw85ms2nx4sXFvl1jjJ599llVrFhRNptNSUlJxV7DlQwfPlyDBw92dRnADY1g50J9+vSRzWazvypVqqSOHTtq27ZtLqln4cKFGjNmTLFsa8SIERo8eLDKli2r06dPq2TJkpo/f75Dn27duslms+nAgQMO7YGBgRo1apQk6c0333R4D3Nf9evXz7PNzz77TO7u7nrhhRfyrWnGjBkKCgqSt7e3ypcvr+DgYPtjdgq6rSuF48qVK6t3796Kjo4uyFtUaI4dO6aBAweqRo0a8vT0lJ+fnyIiIrRu3bpireNGlZKSoilTpuiNN96wt/31u1myZEndfvvtGjp0qM6dO+fCSove8uXLNWfOHH3zzTc6cuSIy542cODAgXyD5auvvqq5c+dq3759LqnrVjFnzpxrft7rjXCQlJWVpcDAQP34448urcNVCHYu1rFjRx05ckRHjhxRfHy8SpQoof/7v/9zSS0VK1ZU2bJli3w7hw4d0jfffKM+ffpIkry9vdW8eXMlJCQ49EtISFBAQIBD+/79+3Xw4EHdc8899rZGjRrZ38Pc19q1a/Nsd9asWRo6dKg+++yzPP9Ax8XF6eWXX9aLL76opKQkrVu3TkOHDtXp06cd+hV0W5fTt29fffrppzp+/HiBl7leXbt21datWzV37lzt3r1bX3/9tdq3b68//vij2GooallZWde87MyZM9W6dWvVrFnToT33u7lv3z5NnjxZ//jHP4o9lBe3vXv3qlq1amrdurX8/PxUooTzT8QyxujChQtFUN3Fg6OIiAh9+OGHRbL+wpbfgeBfX2+++aYkadGiRWrVqpV8fHxUtmxZNWrUyOEAcc6cOfkuP3PmTNfs2GVc7iAp9wcIcn355Zfy8vLSxIkT7X1sNpvGjx/v0G/x4sUOvxCVkJAgm82mRo0aKTs726Fv+fLlNWfOHEkXf/zg1Vdf1bBhwwpx724iBi4TGRlpHn74YYe2NWvWGEnm6NGj9rahQ4eaOnXqmFKlSpnbb7/djBw50mRlZTksN2bMGFOlShXj7e1tnnnmGTNs2DATFBRkn3/+/HkzePBg4+PjYypWrGiGDh1qevfu7bD9du3amZdeesk+XbNmTTN27FjTt29f4+3tbQICAsw//vEPh+2uW7fOBAUFGU9PTxMSEmIWLVpkJJmtW7dedr/fe+8907x5c4e2ESNGmHr16tmnd+zYYXx8fMy4ceNMZGSkvT0uLs54enqas2fPGmOMiY6OdtjPy9m3b58pVaqUOXHihAkNDTWffvqpw/yHH37Y9OnT54rrKMi2Ln0P83P77bebmTNnXrXmwvDnn38aSSYhIeGyffbv35/n/1nucqtXrzbGGLN69WojySxfvtw0bdrUeHl5mQ4dOpjU1FSzbNkyU79+fVO2bFnz1FNPmYyMDPt62rVrZwYNGmReeuklU758eVO1alXz0UcfmdOnT5s+ffoYb29vU6tWLbNs2TL7MhcuXDBPP/20CQwMNF5eXqZu3bomNjbWoebc784777xjqlWrZgIDA81bb71lGjVqlGf/goKCzMiRIy+7/40aNTJTp07Nd/1/9eijj5rg4GD7dFpamunWrZvx9/c3pUqVMnfeeaeZN2+ewzLt2rUzgwcPNq+99pqpUKGC8fX1NdHR0Q59du/ebdq2bWs8PT1NgwYNzMqVK40ks2jRInufbdu2mQ4dOhgvLy9TsWJF079/f3Pq1Kk89Y4dO9ZUrVrV+Pj4mLfeesucP3/evPrqq6ZChQqmevXqJi4u7rLvQ2RkpJFkf9WsWdMYY8y5c+fM4MGDTZUqVYynp6dp06aN2bhxo3253M/GsmXLTLNmzUzJkiXN6tWrTXZ2thk3bpz9/2OTJk3MF198YV/u+PHjpnv37qZy5crGy8vL1K5d217fX+uQZNq1a2dfbu7cuea222677H7cSI4cOWJ/xcbGmnLlyjm0nTp1ynz77bemZMmSZsKECWbnzp1m165dZtGiReb555+3r2f27Nl5lj1y5Ig5c+ZMnm1mZmZed92zZ882Pj4+Ti83ZswYExER4dB26XdpxowZxsPDw+GzGBkZaby8vEz58uXN8ePH7e25/57kyv2seXl55fks+/j4mNmzZ9unjx8/bjw8PMz27dud3o+bHSN2N5DTp0/rk08+Ue3atVWpUiV7e9myZTVnzhzt2LFDU6ZM0YwZMzR58mT7/E8//VRjx47Vu+++q82bN6tGjRp5jmjfffddffrpp5o9e7bWrVun9PT0Al3DM3HiRDVv3lxbt27V888/r4EDB2rXrl2SLv6GYOfOndW4cWNt2bJFY8aMKdAR0po1a9S8eXOHtg4dOmjXrl06cuSIJGn16tW66667dM899ziM2K1evVphYWHy8vK66nb+avbs2erUqZN8fHzUs2dPzZo1y2G+n5+ffvjhBx08eNCp9V6Lli1bas2aNUW+HeniaKi3t7cWL16szMzM617fm2++qalTpyoxMVGHDx/WE088odjYWM2bN09Lly7VypUr9cEHHzgsM3fuXFWuXFkbN27U4MGDNXDgQD3++ONq3bq1tmzZovvvv1+9evXSmTNnJF38WcHbbrtNX3zxhXbs2KHRo0fr9ddf1+eff+6w3vj4eO3atUurVq3SN998o6efflrJycnatGmTvc/WrVu1bds2+y/ZXOr48ePasWNHns/jpbZv367ExER5eHjY286dO6eQkBAtXbpU27dv17PPPqtevXpp48aNefa/TJky2rBhgyZMmKC3335bq1atsu/ro48+Kg8PD23YsEHTp0/P8x3KyMhQRESEKlSooE2bNumLL77Qt99+m+dnF7/77jv997//1X/+8x9NmjRJ0dHR+r//+z9VqFBBGzZs0IABA/Tcc89d9hrPKVOm6O2339Ztt92mI0eO2N/HoUOH6quvvtLcuXO1ZcsW1a5dWxEREXlGnYcPH67x48crOTlZTZo0UUxMjD7++GNNnz5dv/zyi4YMGaKePXvq+++/l3TxlN2OHTv073//W8nJyfrwww9VuXJlSbK/h99++62OHDmihQsX2rfTsmVL/fbbb3ku0bgR+fn52V8+Pj6y2WwObd7e3vrXv/6lNm3a6LXXXlO9evVUt25ddenSRdOmTXNY16XL+vn5qVSpUnrzzTfVtGlTzZw5U7fffrv9b+OJEyfUr18/ValSReXKldM999yjn376yb6+n376SR06dFDZsmVVrlw5hYSE5Dl1uWLFCjVo0EDe3t72EewrmT9/vjp37nzZ+RMmTNDgwYM1f/78PN/J8PBw+fn5OVz+cjmDBw9WdHT0Ff+mVahQQW3atMlzic8twdXJ8lYWGRlp3N3dTZkyZUyZMmWMJFOtWjWzefPmKy733nvvmZCQEPt0aGioeeGFFxz6tGnTxmF0ydfX17z33nv26QsXLpgaNWpcdcSuZ8+e9umcnBxTtWpV8+GHHxpjjPnwww9NpUqV7KNnxlw8GtNVRuyCgoLM22+/7dCWkZFhPDw87CMejz/+uJkwYYI5f/68KVOmjNm3b58xxpgaNWqYt956y75cdHS0cXNzs7+Hua/nnnvO3ic7O9sEBASYxYsXG2OMOXbsmPHw8LCv0xhj/vvf/5pWrVoZSaZu3bomMjLSLFiwwGRnZzu1rYKM2A0ZMsS0b9/+in0K05dffmkqVKhgvLy8TOvWrc2IESPMTz/9ZJ/vzIjdt99+a+8TExNjJJm9e/fa25577jmHI/Z27dqZu+66yz594cIFU6ZMGdOrVy9725EjR4wks379+svuwwsvvGC6du1qn46MjDS+vr55RiceeOABM3DgQPv04MGDr/heb9261Ugyhw4dcmj/63fT09PTSDJubm7myy+/vOy6jDGmU6dO5pVXXrFPX7r/xhjTokULM2zYMGOMMStWrDAlSpQwv//+u33+v//9b4cRu48++shUqFDBnD592t5n6dKlxs3NzaSkpNjrrVmzpsPntV69eqZt27b26dz3/rPPPrts/ZMnT7aP1BljzOnTp03JkiUdRrizsrKMv7+/mTBhgjHmf5+N3O+XMRdH+UqXLm0SExMd1v/MM8+Yp556yhhjTOfOnU3fvn3zrSO/z2SukydPXnUU+kZ0uVGwmJgYU6VKFfPzzz87vawxF/8ulSlTxnTs2NFs2bLF/t0ODw83nTt3Nps2bTK7d+82r7zyiqlUqZL5448/jDEXR6p79uxpkpOTze7du83nn39ukpKS7NsrWbKkCQ8PN5s2bTKbN282DRo0MN27d79sjX/88Yex2Wzmhx9+cGjPHbEbOnSo8fb2dvgbcmmfhQsXGi8vL3P48GFjzOVH7H7//XdTrVo1h3/TLh2xM8aYYcOGOYz23ioYsXOxDh06KCkpSUlJSdq4caMiIiL0wAMPOIwcLViwQG3atLEf4Y0cOVKHDh2yz9+1a5datmzpsN6/Tp88eVKpqakObe7u7goJCblqfU2aNLH/d+4R49GjR+3bbdKkicPo2aV15Ofs2bN5RtxKly6tFi1a2Efnvv/+e7Vv314lSpRQ69atlZCQoH379unQoUPq0KGDw7L16tWzv4e5r7fffts+f9WqVcrIyNCDDz4o6eJ1Ovfdd5/i4uLsfapVq6b169fr559/1ksvvaQLFy4oMjJSHTt2VE5OToG3VRClSpWyj04Vh65du+q///2vvv76a3Xs2FEJCQlq1qyZ/XoUZ/z18+Dr66vSpUvrjjvucGjL/Xzkt4y7u7sqVaqkxo0bOywjyWG5adOmKSQkRFWqVJG3t7c++ugjh8+8JDVu3NhhBE2S+vfvb7+GMisrS/PmzdPTTz992f05e/asJOU7Apz73dywYYMiIyPVt29fde3a1T4/OztbY8aMUePGjVWxYkV5e3trxYoVeer86/5LFz9rufuanJysgIAA+fv72+eHhYU59E9OTlZQUJDKlCljb2vTpo1ycnLso+fSxes///rzi76+vg7vc+57f+n/nyvZu3evzp8/rzZt2tjbSpYsqZYtWyo5Odmh719HPffs2aMzZ87ovvvus48ae3t76+OPP9bevXslSQMHDtT8+fPVtGlTDR06VImJiQWqqVSpUpJUrN+hojR48GC1aNFCjRs3VmBgoLp166a4uLg8o1EnT550eC/9/Pzs87KysvTxxx8rODhYTZo00dq1a7Vx40Z98cUXat68uerUqaO//e1vKl++vL788ktJF691Dg8PV/369VWnTh09/vjjCgoKsq/z/Pnzmj59upo3b65mzZpp0KBBio+Pv+x+HDp0SMYYh89yrn//+9+aMGGClixZonvvvfey63jkkUfUtGnTq17LWrp0aUVHRysmJkYnT568bD9/f/9iOQtzoyHYuViZMmVUu3Zt1a5dWy1atNDMmTOVkZGhGTNmSJLWr1+vHj166MEHH9Q333yjrVu36o033riui8WdUbJkSYdpm83mEHSuReXKlfXnn3/mae/QoYNWr16tX375RWfPnlWzZs0kSe3atdPq1au1evVqlS5dWqGhoQ7LeXh42N/D3FfVqlXt82fNmqXjx4+rVKlSKlGihEqUKKFly5Zp7ty5efblzjvv1PPPP69PPvlEq1at0qpVq+ynjgqyrYI4fvy4qlSp4tQy18vLy0v33XefRo0apcTERPXp08f+xzM3DJi//AjN+fPn813PXz8PuXeM/lV+n4/8+ly6Hkn25ebPn69XX31VzzzzjFauXKmkpCT17ds3z2f+r0EnV+fOneXp6alFixbpX//6l86fP6/HHnss332RZD/1l9/nMfe7GRQUpLi4OG3YsMHhFP57772nKVOmaNiwYVq9erWSkpIUERGRp86i+A7l52rvc1FuW3L8/5F709HSpUsdDoJ27NhhDxa5B7BDhgzRf//7X91777169dVXr7qd3FPAxf0dKiplypTR0qVLtWfPHo0cOVLe3t565ZVX1LJlS4fwWrZsWYf38q9BuGbNmg7vx08//aTTp0+rUqVKDmFw//799mAdFRWlfv36KTw8XOPHj7e35ypdurRq1apln/7rAUl+rnSQ1KRJEwUGBio6OjrPDWmXevfddzV37tw8Bw6XeuaZZ1SpUiW9++67l+1T3AfRNwqC3Q3GZrPJzc3N/iVJTExUzZo19cYbb9iPvC49AqlXr57DdUWSHKZ9fHzk6+vr0Jadna0tW7ZcV6316tXTzz//7HBkeWkd+QkODtaOHTvytHfo0EG//vqr5s2bp7vuukvu7u6SpLvvvlvff/+9EhIS1KZNmzyjNFfyxx9/aMmSJZo/f77DH8WtW7fqzz//1MqVKy+7bMOGDSVdvMapMG3fvl3BwcGFuk5nNWzY0L5fuf8g/PX6GVc+v2zdunVq3bq1nn/+eQUHB6t27dp5/tG5nBIlSigyMlKzZ8/W7Nmz1a1bN/sIT35q1aqlcuXK5ft5/Cs3Nze9/vrrGjlypP27uW7dOj388MPq2bOngoKCdMcdd2j37t0F31FJDRo00OHDhx3e+x9++CFPn59++snhc7hu3Tq5ubmpXr16Tm3PWbVq1ZKHh4fDo3HOnz+vTZs22b8f+WnYsKE8PT116NChPAdCAQEB9n5VqlRRZGSkPvnkE8XGxuqjjz6SJPt3/NI7H6WL35+SJUuqUaNGhbWbN4RatWqpX79+mjlzprZs2aIdO3ZowYIF9vlubm4O7+NfR8ovPcg5ffq0qlWrlufswq5du/Taa69Juni97C+//KJOnTrpu+++U8OGDbVo0SL7OvI7KDBX+AXSKx0kVa9eXQkJCfr999/VsWPHKz6/9O6771ZERIRGjBhx2T7Sxe/62LFjNWXKFP33v//Nt48rDqJvBAQ7F8vMzFRKSopSUlKUnJyswYMH6/Tp0/YLUOvUqaNDhw5p/vz52rt3r95//32HL590cSh/1qxZmjt3rn799Ve988472rZtm8Nt4oMHD1ZMTIyWLFmiXbt26aWXXtKff/7p0MdZ3bt3V05Ojp599lklJydrxYoV+tvf/iZJV1xvRESE1q9fn+ePduvWreXp6akPPvhA7dq1s7e3bNlSR48e1ZIlS/KchpWkCxcu2N/D3Fdqaqok6Z///KcqVaqkJ554Qnfeeaf9FRQUpAcffNA+AjNw4ECNGTNG69at08GDB/XDDz+od+/eqlKlisOpsSttK9exY8fy/EHN7XPmzBlt3rxZ999/vzNv9TX7448/dM899+iTTz7Rtm3btH//fn3xxReaMGGCHn74YUkXj2pbtWplv/D9+++/18iRI4ulvvzUqVNHP/74o1asWKHdu3dr1KhRBTpgyNWvXz999913Wr58+RVPw0oX/7EMDw8v0CNrHn/8cbm7u9svaq9Tp45WrVqlxMREJScn67nnnsvzWbia8PBw1a1bV5GRkfrpp5+0Zs0ah0dFSFKPHj3k5eWlyMhIbd++XatXr9bgwYPVq1cv+2nsolKmTBkNHDhQr732mpYvX64dO3aof//+OnPmjJ555pnLLle2bFm9+uqrGjJkiObOnau9e/dqy5Yt+uCDDzR37lxJ0ujRo7VkyRLt2bNHv/zyi7755hs1aNBAklS1alWVKlVKy5cvV2pqqsPptjVr1qht27ZXDOw3u8DAQJUuXfqaDyqbNWumlJQUlShRIk+wzg1gklS3bl0NGTJEK1eu1KOPPqrZs2dfc81XO0iqWbOmvv/+e6WkpFw13I0fP17/+te/tH79+itu8/HHH1ejRo301ltv5Tv/RjiIdgWCnYstX75c1apVU7Vq1RQaGmq/6619+/aSpIceekhDhgzRoEGD1LRpUyUmJtofzpurR48eGjFihF599VU1a9ZM+/fvV58+fRyGxIcNG6annnpKvXv3VlhYmLy9vRUREeH03aV/Va5cOf3rX/9SUlKSmjZtqjfeeEOjR4+WlP9wfK4HHnhAJUqU0LfffuvQ7uXlpVatWunUqVP2/ZckT09Pe3t+we6XX36xv4e5r9xnksXFxemRRx7JN2h27dpVX3/9tdLS0hQeHq4ffvhBjz/+uOrWrauuXbvKy8tL8fHxDncoX2lbuebNm6fg4GCHV+6p9SVLlqhGjRpq27btVd7dwuHt7a3Q0FBNnjxZd999t+68806NGjVK/fv319SpU+394uLidOHCBYWEhOjll1/WO++8Uyz15ee5557To48+qieffFKhoaH6448/9Pzzzxd4+Tp16qh169aqX79+ntP2+enXr5/mz59/1VOUJUqU0KBBgzRhwgRlZGRo5MiRatasmSIiItS+fXv5+fnleV7X1bi5uWnRokU6e/asWrZsqX79+mns2LEOfUqXLq0VK1bo+PHjatGihR577DHde++9Dv//itL48ePVtWtX9erVS82aNdOePXu0YsUKVahQ4YrLjRkzRqNGjVJMTIwaNGigjh07aunSpbr99tslXRyVGzFihJo0aaK7775b7u7u9jsYS5Qooffff1//+Mc/5O/vbz8IkS6equ/fv3/R7XAxe/PNNzV06FAlJCRo//792rp1q55++mmdP39e99133zWtMzw8XGFhYerSpYtWrlypAwcOKDExUW+88YZ+/PFHnT17VoMGDVJCQoIOHjyodevWadOmTfZgfS0KcpCU+1zSo0ePKiIiQunp6fn2a9y4sXr06KH333//qtsdP3684uLi8g3Ba9asKbaD6BuKi2/eQBEJDw93uKP1UtnZ2aZu3bpXfL7Xtfjkk09MyZIl832+0l9NnTrV3H///YW67ZtBfs/QQ+HKyckxtWrVMhMnTixw/xYtWuR5Bh1uPMuWLTMNGjQw58+fd3UpTrvcna3fffed6dq1qwkICDAeHh7G19fXdOzY0axZs+aqyxpz+edrpqenm8GDBxt/f39TsmRJExAQYHr06GEOHTpkMjMzTbdu3ezb9Pf3N4MGDbI/4SC/7V16h2p+li1bZqpXr+5wd3Z+z4T87bffTJ06dUyrVq3MyZMn8+2zf/9+4+Hhke9dsX/++adD3/vvv99IcrgrNjEx0ZQvX/6q/xZZEcHOAjIyMszEiRPN9u3bTXJyshk9erSRZFatWmXvc+DAAfPRRx+ZXbt2mW3btplnn33WlCxZ0uzYseO6tj137lyzZs0as2/fPrNo0SJTvXp106NHj6sud/78efPOO++Y9PT069r+zeTYsWPm3XffNTk5Oa4uxbKOHj1q3n//fVOmTBmHB51ezdatW83HH39chJWhMHzxxRd5HqeBG8eNdJD0xBNPmLFjx7q6DJewGXOFqyFxUzh79qw6d+6srVu36ty5c6pXr55GjhypRx991N7n8OHD6tatm7Zv3y5jjO68806NHz9ed99993Vte8KECfr73/+ulJQUVatWTV26dNHYsWNVunTp690twGk2m02VK1fWlClT1L17d1eXA9xykpKS9PPPP6tXr14uqyErK0sTJkzQK6+8YulrMS+HYAcAAGAR3DwBAABgEQQ7AAAAiyDYAQAAWATBDgAAwCIIdgAAABZBsAOAIpSQkCCbzaYTJ04UeJnAwEDFxsYWWU0ArItgB+CW1qdPH9lsNg0YMCDPvBdeeEE2m019+vQp/sIA4BoQ7ADc8gICAjR//nydPXvW3nbu3DnNmzdPNWrUcGFlAOAcgh2AW16zZs0UEBCghQsX2tsWLlyoGjVqKDg42N6WmZmpF198UVWrVpWXl5fuuusubdq0yWFdy5YtU926dVWqVCl16NBBBw4cyLO9tWvXqm3btipVqpQCAgL04osv5vsj5gDgLIIdAEh6+umnNXv2bPt0XFyc+vbt69Bn6NCh+uqrrzR37lxt2bJFtWvXVkREhI4fPy7p4k/3Pfroo+rcubOSkpLUr18/DR8+3GEde/fuVceOHdW1a1dt27ZNCxYs0Nq1azVo0KCi30kAlkewAwBJPXv21Nq1a3Xw4EEdPHhQ69atU8+ePe3zMzIy9OGHH+q9997TAw88oIYNG2rGjBkqVaqUZs2aJUn68MMPVatWLU2cOFH16tVTjx498lyfFxMTox49eujll19WnTp11Lp1a73//vv6+OOPde7cueLcZQAWVMLVBQDAjaBKlSrq1KmT5syZI2OMOnXqpMqVK9vn7927V+fPn1ebNm3sbSVLllTLli2VnJwsSUpOTlZoaKjDesPCwhymf/rpJ23btk2ffvqpvc0Yo5ycHO3fv18NGjQoit0DcIsg2AHA//f000/bT4lOmzatSLZx+vRpPffcc3rxxRfzzONGDQDXi2AHAP9fx44dlZWVJZvNpoiICId5tWrVkoeHh9atW6eaNWtKks6fP69Nmzbp5ZdfliQ1aNBAX3/9tcNyP/zwg8N0s2bNtGPHDtWuXbvodgTALYtr7ADg/3N3d1dycrJ27Nghd3d3h3llypTRwIED9dprr2n58uXasWOH+vfvrzNnzuiZZ56RJA0YMEC//vqrXnvtNe3atUvz5s3TnDlzHNYzbNgwJSYmatCgQUpKStKvv/6qJUuWcPMEgEJBsAOAvyhXrpzKlSuX77zx48era9eu6tWrl5o1a6Y9e/ZoxYoVqlChgqSLp1K/+uorLV68WEFBQZo+fbrGjRvnsI4mTZro+++/1+7du9W2bVsFBwdr9OjR8vf3L/J9A2B9NmOMcXURAAAAuH6M2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwCIIdAACARRDsAAAALIJgBwAAYBEEOwAAAIsg2AEAAFgEwQ4AAMAiCHYAAAAWQbADAACwiP8HyA7lgUzyZ8IAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = ['Bagging (WEASEL)', 'Summary (Random forest)', 'TSFresh (KNN)']\n",
    "y = [0.50, 0.73, 0.81]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "ax.bar(x, y)\n",
    "ax.set_xlabel('Model')\n",
    "ax.set_ylabel('F1-Score (weighted)')\n",
    "ax.set_ylim(0, 1)\n",
    "\n",
    "for i, v in enumerate(y):\n",
    "    ax.text(x=i, y=v+0.05, s=str(v), ha='center')\n",
    "\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "95c1fe75-1389-4f87-980c-a4cb85658f79",
   "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": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}