fixed_p_modified_chen.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/marcus/anaconda3/lib/python3.6/site-packages/statsmodels/compat/pandas.py:56: FutureWarning: The pandas.core.datetools module is deprecated and will be removed in a future version. Please use the pandas.tseries module instead.\n",
      "  from pandas.core import datetools\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from math import ceil\n",
    "from CIoTS import *\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "max_ps = list(range(4, 9, 2))\n",
    "test_ps = list(range(2, 11, 2))\n",
    "runs = 20\n",
    "dimensions = 4\n",
    "data_length = 10000\n",
    "alpha = 0.05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "f1_scores = pd.DataFrame(columns=['true_p', 'p', 'mean_f1', 'std_f1'])\n",
    "runtimes = pd.DataFrame(columns=['true_p', 'p', 'mean_time', 'std_time'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done: p=4 run=1 p_test=2 exec_time=2.0252604484558105\n",
      "done: p=4 run=1 p_test=4 exec_time=7.836875915527344\n",
      "done: p=4 run=1 p_test=6 exec_time=21.853820323944092\n",
      "done: p=4 run=1 p_test=8 exec_time=54.23538565635681\n",
      "done: p=4 run=1 p_test=10 exec_time=82.95904469490051\n",
      "done: p=4 run=2 p_test=2 exec_time=0.8542451858520508\n",
      "done: p=4 run=2 p_test=4 exec_time=5.028505086898804\n",
      "done: p=4 run=2 p_test=6 exec_time=13.176971197128296\n",
      "done: p=4 run=2 p_test=8 exec_time=30.117101192474365\n",
      "done: p=4 run=2 p_test=10 exec_time=39.15522050857544\n",
      "done: p=4 run=3 p_test=2 exec_time=1.0705671310424805\n",
      "done: p=4 run=3 p_test=4 exec_time=3.7217624187469482\n",
      "done: p=4 run=3 p_test=6 exec_time=10.627224445343018\n",
      "done: p=4 run=3 p_test=8 exec_time=21.681575298309326\n",
      "done: p=4 run=3 p_test=10 exec_time=37.53425097465515\n",
      "done: p=4 run=4 p_test=2 exec_time=0.7710905075073242\n",
      "done: p=4 run=4 p_test=4 exec_time=3.1293275356292725\n",
      "done: p=4 run=4 p_test=6 exec_time=9.422035455703735\n",
      "done: p=4 run=4 p_test=8 exec_time=21.265116930007935\n",
      "done: p=4 run=4 p_test=10 exec_time=31.67174196243286\n",
      "done: p=4 run=5 p_test=2 exec_time=1.1952762603759766\n",
      "done: p=4 run=5 p_test=4 exec_time=5.241429328918457\n",
      "done: p=4 run=5 p_test=6 exec_time=10.943949222564697\n",
      "done: p=4 run=5 p_test=8 exec_time=25.129889965057373\n",
      "done: p=4 run=5 p_test=10 exec_time=32.176199197769165\n",
      "done: p=4 run=6 p_test=2 exec_time=0.8621993064880371\n",
      "done: p=4 run=6 p_test=4 exec_time=4.675207614898682\n",
      "done: p=4 run=6 p_test=6 exec_time=14.195857763290405\n",
      "done: p=4 run=6 p_test=8 exec_time=32.04996371269226\n",
      "done: p=4 run=6 p_test=10 exec_time=58.81268930435181\n",
      "done: p=4 run=7 p_test=2 exec_time=1.1530723571777344\n",
      "done: p=4 run=7 p_test=4 exec_time=5.676813125610352\n",
      "done: p=4 run=7 p_test=6 exec_time=15.43423342704773\n",
      "done: p=4 run=7 p_test=8 exec_time=28.966148376464844\n",
      "done: p=4 run=7 p_test=10 exec_time=47.967679023742676\n",
      "done: p=4 run=8 p_test=2 exec_time=1.3333792686462402\n",
      "done: p=4 run=8 p_test=4 exec_time=5.9045569896698\n",
      "done: p=4 run=8 p_test=6 exec_time=14.752938985824585\n",
      "done: p=4 run=8 p_test=8 exec_time=35.135146141052246\n",
      "done: p=4 run=8 p_test=10 exec_time=47.11895942687988\n",
      "done: p=4 run=9 p_test=2 exec_time=0.6003403663635254\n",
      "done: p=4 run=9 p_test=4 exec_time=3.30189847946167\n",
      "done: p=4 run=9 p_test=6 exec_time=11.567784070968628\n",
      "done: p=4 run=9 p_test=8 exec_time=25.146071910858154\n",
      "done: p=4 run=9 p_test=10 exec_time=47.062018156051636\n",
      "done: p=4 run=10 p_test=2 exec_time=0.6543042659759521\n",
      "done: p=4 run=10 p_test=4 exec_time=2.4763612747192383\n",
      "done: p=4 run=10 p_test=6 exec_time=6.508742570877075\n",
      "done: p=4 run=10 p_test=8 exec_time=14.338969469070435\n",
      "done: p=4 run=10 p_test=10 exec_time=19.815636157989502\n",
      "done: p=4 run=11 p_test=2 exec_time=0.7086412906646729\n",
      "done: p=4 run=11 p_test=4 exec_time=3.7303519248962402\n",
      "done: p=4 run=11 p_test=6 exec_time=7.832421779632568\n",
      "done: p=4 run=11 p_test=8 exec_time=19.22883415222168\n",
      "done: p=4 run=11 p_test=10 exec_time=27.24103569984436\n",
      "done: p=4 run=12 p_test=2 exec_time=0.8127777576446533\n",
      "done: p=4 run=12 p_test=4 exec_time=4.451983213424683\n",
      "done: p=4 run=12 p_test=6 exec_time=12.425765991210938\n",
      "done: p=4 run=12 p_test=8 exec_time=36.84986400604248\n",
      "done: p=4 run=12 p_test=10 exec_time=54.478264570236206\n",
      "done: p=4 run=13 p_test=2 exec_time=0.7848637104034424\n",
      "done: p=4 run=13 p_test=4 exec_time=3.3650262355804443\n",
      "done: p=4 run=13 p_test=6 exec_time=7.839848518371582\n",
      "done: p=4 run=13 p_test=8 exec_time=24.042402982711792\n",
      "done: p=4 run=13 p_test=10 exec_time=30.82375168800354\n",
      "done: p=4 run=14 p_test=2 exec_time=0.8192596435546875\n",
      "done: p=4 run=14 p_test=4 exec_time=6.426586151123047\n",
      "done: p=4 run=14 p_test=6 exec_time=15.62636399269104\n",
      "done: p=4 run=14 p_test=8 exec_time=39.370126724243164\n",
      "done: p=4 run=14 p_test=10 exec_time=55.34554314613342\n",
      "done: p=4 run=15 p_test=2 exec_time=0.9841494560241699\n",
      "done: p=4 run=15 p_test=4 exec_time=5.287997722625732\n",
      "done: p=4 run=15 p_test=6 exec_time=13.674142837524414\n",
      "done: p=4 run=15 p_test=8 exec_time=27.005751132965088\n",
      "done: p=4 run=15 p_test=10 exec_time=50.44912648200989\n",
      "done: p=4 run=16 p_test=2 exec_time=1.0094435214996338\n",
      "done: p=4 run=16 p_test=4 exec_time=5.179925203323364\n",
      "done: p=4 run=16 p_test=6 exec_time=11.963346242904663\n",
      "done: p=4 run=16 p_test=8 exec_time=22.321982860565186\n",
      "done: p=4 run=16 p_test=10 exec_time=31.829598426818848\n",
      "done: p=4 run=17 p_test=2 exec_time=0.4883434772491455\n",
      "done: p=4 run=17 p_test=4 exec_time=3.7226638793945312\n",
      "done: p=4 run=17 p_test=6 exec_time=13.21059250831604\n",
      "done: p=4 run=17 p_test=8 exec_time=27.37944483757019\n",
      "done: p=4 run=17 p_test=10 exec_time=41.67905306816101\n",
      "done: p=4 run=18 p_test=2 exec_time=1.0038683414459229\n",
      "done: p=4 run=18 p_test=4 exec_time=5.35983943939209\n",
      "done: p=4 run=18 p_test=6 exec_time=12.389490604400635\n",
      "done: p=4 run=18 p_test=8 exec_time=29.97350263595581\n",
      "done: p=4 run=18 p_test=10 exec_time=35.95038604736328\n",
      "done: p=4 run=19 p_test=2 exec_time=1.840397596359253\n",
      "done: p=4 run=19 p_test=4 exec_time=8.620524883270264\n",
      "done: p=4 run=19 p_test=6 exec_time=21.46791124343872\n",
      "done: p=4 run=19 p_test=8 exec_time=41.874364614486694\n",
      "done: p=4 run=19 p_test=10 exec_time=73.67135167121887\n",
      "done: p=4 run=20 p_test=2 exec_time=1.4882338047027588\n",
      "done: p=4 run=20 p_test=4 exec_time=5.957768440246582\n",
      "done: p=4 run=20 p_test=6 exec_time=15.908726215362549\n",
      "done: p=4 run=20 p_test=8 exec_time=39.62301683425903\n",
      "done: p=4 run=20 p_test=10 exec_time=58.838849782943726\n",
      "done: p=6 run=1 p_test=2 exec_time=0.46350693702697754\n",
      "done: p=6 run=1 p_test=4 exec_time=3.9238967895507812\n",
      "done: p=6 run=1 p_test=6 exec_time=6.3293962478637695\n",
      "done: p=6 run=1 p_test=8 exec_time=12.459447622299194\n",
      "done: p=6 run=1 p_test=10 exec_time=17.922335386276245\n",
      "done: p=6 run=2 p_test=2 exec_time=0.40010762214660645\n",
      "done: p=6 run=2 p_test=4 exec_time=2.5457658767700195\n",
      "done: p=6 run=2 p_test=6 exec_time=5.837512254714966\n",
      "done: p=6 run=2 p_test=8 exec_time=12.7652747631073\n",
      "done: p=6 run=2 p_test=10 exec_time=20.020562171936035\n",
      "done: p=6 run=3 p_test=2 exec_time=0.44570279121398926\n",
      "done: p=6 run=3 p_test=4 exec_time=2.627903699874878\n",
      "done: p=6 run=3 p_test=6 exec_time=8.418219327926636\n",
      "done: p=6 run=3 p_test=8 exec_time=27.69229483604431\n",
      "done: p=6 run=3 p_test=10 exec_time=38.62348747253418\n",
      "done: p=6 run=4 p_test=2 exec_time=1.3735880851745605\n",
      "done: p=6 run=4 p_test=4 exec_time=5.915826797485352\n",
      "done: p=6 run=4 p_test=6 exec_time=16.50376844406128\n",
      "done: p=6 run=4 p_test=8 exec_time=46.568458557128906\n",
      "done: p=6 run=4 p_test=10 exec_time=64.20468306541443\n",
      "done: p=6 run=5 p_test=2 exec_time=0.7993214130401611\n",
      "done: p=6 run=5 p_test=4 exec_time=3.930022716522217\n",
      "done: p=6 run=5 p_test=6 exec_time=7.727311372756958\n",
      "done: p=6 run=5 p_test=8 exec_time=14.623368263244629\n",
      "done: p=6 run=5 p_test=10 exec_time=25.884487628936768\n",
      "done: p=6 run=6 p_test=2 exec_time=0.9219202995300293\n",
      "done: p=6 run=6 p_test=4 exec_time=3.2003860473632812\n",
      "done: p=6 run=6 p_test=6 exec_time=9.283292531967163\n",
      "done: p=6 run=6 p_test=8 exec_time=13.864828109741211\n",
      "done: p=6 run=6 p_test=10 exec_time=28.9912211894989\n",
      "done: p=6 run=7 p_test=2 exec_time=0.546025276184082\n",
      "done: p=6 run=7 p_test=4 exec_time=2.346289873123169\n",
      "done: p=6 run=7 p_test=6 exec_time=4.821955442428589\n",
      "done: p=6 run=7 p_test=8 exec_time=14.817166566848755\n",
      "done: p=6 run=7 p_test=10 exec_time=23.677170753479004\n",
      "done: p=6 run=8 p_test=2 exec_time=1.0448179244995117\n",
      "done: p=6 run=8 p_test=4 exec_time=7.998412609100342\n",
      "done: p=6 run=8 p_test=6 exec_time=8.699263095855713\n",
      "done: p=6 run=8 p_test=8 exec_time=20.165058374404907\n",
      "done: p=6 run=8 p_test=10 exec_time=28.278943300247192\n",
      "done: p=6 run=9 p_test=2 exec_time=1.0983400344848633\n",
      "done: p=6 run=9 p_test=4 exec_time=4.014044761657715\n",
      "done: p=6 run=9 p_test=6 exec_time=10.179778099060059\n",
      "done: p=6 run=9 p_test=8 exec_time=28.145370483398438\n",
      "done: p=6 run=9 p_test=10 exec_time=48.70680499076843\n",
      "done: p=6 run=10 p_test=2 exec_time=1.1925835609436035\n",
      "done: p=6 run=10 p_test=4 exec_time=4.962731838226318\n",
      "done: p=6 run=10 p_test=6 exec_time=13.785194635391235\n",
      "done: p=6 run=10 p_test=8 exec_time=27.167850494384766\n",
      "done: p=6 run=10 p_test=10 exec_time=43.495362520217896\n",
      "done: p=6 run=11 p_test=2 exec_time=1.3447315692901611\n",
      "done: p=6 run=11 p_test=4 exec_time=5.688276290893555\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "done: p=6 run=11 p_test=6 exec_time=16.08575987815857\n",
      "done: p=6 run=11 p_test=8 exec_time=33.05758762359619\n",
      "done: p=6 run=11 p_test=10 exec_time=49.37740349769592\n",
      "done: p=6 run=12 p_test=2 exec_time=0.9162611961364746\n",
      "done: p=6 run=12 p_test=4 exec_time=4.31618070602417\n",
      "done: p=6 run=12 p_test=6 exec_time=11.5531587600708\n",
      "done: p=6 run=12 p_test=8 exec_time=23.67918562889099\n",
      "done: p=6 run=12 p_test=10 exec_time=43.49035382270813\n",
      "done: p=6 run=13 p_test=2 exec_time=0.657717227935791\n",
      "done: p=6 run=13 p_test=4 exec_time=3.727285385131836\n",
      "done: p=6 run=13 p_test=6 exec_time=7.345951795578003\n",
      "done: p=6 run=13 p_test=8 exec_time=17.11294174194336\n",
      "done: p=6 run=13 p_test=10 exec_time=23.91374182701111\n",
      "done: p=6 run=14 p_test=2 exec_time=0.9778492450714111\n",
      "done: p=6 run=14 p_test=4 exec_time=4.502882480621338\n",
      "done: p=6 run=14 p_test=6 exec_time=9.341235160827637\n",
      "done: p=6 run=14 p_test=8 exec_time=17.210639238357544\n",
      "done: p=6 run=14 p_test=10 exec_time=27.7166268825531\n",
      "done: p=6 run=15 p_test=2 exec_time=1.1119017601013184\n",
      "done: p=6 run=15 p_test=4 exec_time=5.711415767669678\n",
      "done: p=6 run=15 p_test=6 exec_time=11.028361558914185\n",
      "done: p=6 run=15 p_test=8 exec_time=29.622695922851562\n",
      "done: p=6 run=15 p_test=10 exec_time=52.9062397480011\n",
      "done: p=6 run=16 p_test=2 exec_time=0.8254072666168213\n",
      "done: p=6 run=16 p_test=4 exec_time=5.270332098007202\n",
      "done: p=6 run=16 p_test=6 exec_time=10.558504343032837\n",
      "done: p=6 run=16 p_test=8 exec_time=25.45148468017578\n",
      "done: p=6 run=16 p_test=10 exec_time=36.069499015808105\n",
      "done: p=6 run=17 p_test=2 exec_time=1.4837875366210938\n",
      "done: p=6 run=17 p_test=4 exec_time=7.209232330322266\n",
      "done: p=6 run=17 p_test=6 exec_time=14.778461933135986\n",
      "done: p=6 run=17 p_test=8 exec_time=35.83038258552551\n",
      "done: p=6 run=17 p_test=10 exec_time=69.49191617965698\n",
      "done: p=6 run=18 p_test=2 exec_time=0.9505112171173096\n",
      "done: p=6 run=18 p_test=4 exec_time=3.0535600185394287\n",
      "done: p=6 run=18 p_test=6 exec_time=8.210403203964233\n",
      "done: p=6 run=18 p_test=8 exec_time=23.301844596862793\n",
      "done: p=6 run=18 p_test=10 exec_time=35.23658895492554\n",
      "done: p=6 run=19 p_test=2 exec_time=1.466275691986084\n",
      "done: p=6 run=19 p_test=4 exec_time=5.288707733154297\n",
      "done: p=6 run=19 p_test=6 exec_time=15.465732097625732\n",
      "done: p=6 run=19 p_test=8 exec_time=29.463562488555908\n",
      "done: p=6 run=19 p_test=10 exec_time=44.26550650596619\n",
      "done: p=6 run=20 p_test=2 exec_time=0.6799108982086182\n",
      "done: p=6 run=20 p_test=4 exec_time=6.495421648025513\n",
      "done: p=6 run=20 p_test=6 exec_time=7.296763896942139\n",
      "done: p=6 run=20 p_test=8 exec_time=17.075754404067993\n",
      "done: p=6 run=20 p_test=10 exec_time=25.959674835205078\n",
      "done: p=8 run=1 p_test=2 exec_time=0.9095408916473389\n",
      "done: p=8 run=1 p_test=4 exec_time=5.800484895706177\n",
      "done: p=8 run=1 p_test=6 exec_time=14.339200973510742\n",
      "done: p=8 run=1 p_test=8 exec_time=31.961408853530884\n",
      "done: p=8 run=1 p_test=10 exec_time=48.74232363700867\n",
      "done: p=8 run=2 p_test=2 exec_time=0.5479936599731445\n",
      "done: p=8 run=2 p_test=4 exec_time=4.713562250137329\n",
      "done: p=8 run=2 p_test=6 exec_time=12.163206577301025\n",
      "done: p=8 run=2 p_test=8 exec_time=22.177058219909668\n",
      "done: p=8 run=2 p_test=10 exec_time=40.385847091674805\n",
      "done: p=8 run=3 p_test=2 exec_time=1.5832436084747314\n",
      "done: p=8 run=3 p_test=4 exec_time=5.495720386505127\n",
      "done: p=8 run=3 p_test=6 exec_time=17.14287519454956\n",
      "done: p=8 run=3 p_test=8 exec_time=29.027281045913696\n",
      "done: p=8 run=3 p_test=10 exec_time=51.16912794113159\n",
      "done: p=8 run=4 p_test=2 exec_time=1.197075366973877\n",
      "done: p=8 run=4 p_test=4 exec_time=7.278475761413574\n",
      "done: p=8 run=4 p_test=6 exec_time=11.54115605354309\n",
      "done: p=8 run=4 p_test=8 exec_time=25.814289808273315\n",
      "done: p=8 run=4 p_test=10 exec_time=44.50191402435303\n",
      "done: p=8 run=5 p_test=2 exec_time=0.39020657539367676\n",
      "done: p=8 run=5 p_test=4 exec_time=3.63852858543396\n",
      "done: p=8 run=5 p_test=6 exec_time=8.718227863311768\n",
      "done: p=8 run=5 p_test=8 exec_time=18.30421233177185\n",
      "done: p=8 run=5 p_test=10 exec_time=38.13652205467224\n",
      "done: p=8 run=6 p_test=2 exec_time=1.1763648986816406\n",
      "done: p=8 run=6 p_test=4 exec_time=4.174854516983032\n",
      "done: p=8 run=6 p_test=6 exec_time=11.959004402160645\n",
      "done: p=8 run=6 p_test=8 exec_time=31.619673490524292\n",
      "done: p=8 run=6 p_test=10 exec_time=43.20965886116028\n",
      "done: p=8 run=7 p_test=2 exec_time=0.7691638469696045\n",
      "done: p=8 run=7 p_test=4 exec_time=3.3725533485412598\n",
      "done: p=8 run=7 p_test=6 exec_time=7.205317258834839\n",
      "done: p=8 run=7 p_test=8 exec_time=15.445576190948486\n",
      "done: p=8 run=7 p_test=10 exec_time=23.185030937194824\n",
      "done: p=8 run=8 p_test=2 exec_time=1.156238079071045\n",
      "done: p=8 run=8 p_test=4 exec_time=4.94893217086792\n",
      "done: p=8 run=8 p_test=6 exec_time=12.809539318084717\n",
      "done: p=8 run=8 p_test=8 exec_time=28.94004774093628\n",
      "done: p=8 run=8 p_test=10 exec_time=42.41828203201294\n",
      "done: p=8 run=9 p_test=2 exec_time=0.9255785942077637\n",
      "done: p=8 run=9 p_test=4 exec_time=7.578350067138672\n",
      "done: p=8 run=9 p_test=6 exec_time=24.53645086288452\n",
      "done: p=8 run=9 p_test=8 exec_time=49.450920820236206\n",
      "done: p=8 run=9 p_test=10 exec_time=63.66039800643921\n",
      "done: p=8 run=10 p_test=2 exec_time=0.47385168075561523\n",
      "done: p=8 run=10 p_test=4 exec_time=4.432581186294556\n",
      "done: p=8 run=10 p_test=6 exec_time=7.474052906036377\n",
      "done: p=8 run=10 p_test=8 exec_time=13.944685220718384\n",
      "done: p=8 run=10 p_test=10 exec_time=30.333306550979614\n",
      "done: p=8 run=11 p_test=2 exec_time=0.7784173488616943\n",
      "done: p=8 run=11 p_test=4 exec_time=1.9783313274383545\n",
      "done: p=8 run=11 p_test=6 exec_time=6.980299472808838\n",
      "done: p=8 run=11 p_test=8 exec_time=9.786311388015747\n",
      "done: p=8 run=11 p_test=10 exec_time=12.061108827590942\n",
      "done: p=8 run=12 p_test=2 exec_time=0.9111225605010986\n",
      "done: p=8 run=12 p_test=4 exec_time=5.307244777679443\n",
      "done: p=8 run=12 p_test=6 exec_time=16.91193389892578\n",
      "done: p=8 run=12 p_test=8 exec_time=36.877289056777954\n",
      "done: p=8 run=12 p_test=10 exec_time=59.4916615486145\n",
      "done: p=8 run=13 p_test=2 exec_time=0.6285772323608398\n",
      "done: p=8 run=13 p_test=4 exec_time=3.8851263523101807\n",
      "done: p=8 run=13 p_test=6 exec_time=13.557451009750366\n",
      "done: p=8 run=13 p_test=8 exec_time=22.65037751197815\n",
      "done: p=8 run=13 p_test=10 exec_time=42.09564828872681\n",
      "done: p=8 run=14 p_test=2 exec_time=0.8634679317474365\n",
      "done: p=8 run=14 p_test=4 exec_time=2.90946364402771\n",
      "done: p=8 run=14 p_test=6 exec_time=10.182981491088867\n",
      "done: p=8 run=14 p_test=8 exec_time=18.18177056312561\n",
      "done: p=8 run=14 p_test=10 exec_time=25.36373805999756\n",
      "done: p=8 run=15 p_test=2 exec_time=0.9656147956848145\n",
      "done: p=8 run=15 p_test=4 exec_time=4.850516080856323\n",
      "done: p=8 run=15 p_test=6 exec_time=15.443260192871094\n",
      "done: p=8 run=15 p_test=8 exec_time=34.65247178077698\n",
      "done: p=8 run=15 p_test=10 exec_time=62.184354066848755\n",
      "done: p=8 run=16 p_test=2 exec_time=0.7349436283111572\n",
      "done: p=8 run=16 p_test=4 exec_time=3.1831884384155273\n",
      "done: p=8 run=16 p_test=6 exec_time=11.928561449050903\n",
      "done: p=8 run=16 p_test=8 exec_time=29.56243586540222\n",
      "done: p=8 run=16 p_test=10 exec_time=41.98190426826477\n",
      "done: p=8 run=17 p_test=2 exec_time=0.9662313461303711\n",
      "done: p=8 run=17 p_test=4 exec_time=3.5074124336242676\n",
      "done: p=8 run=17 p_test=6 exec_time=11.779945373535156\n",
      "done: p=8 run=17 p_test=8 exec_time=28.310731410980225\n",
      "done: p=8 run=17 p_test=10 exec_time=38.02334475517273\n",
      "done: p=8 run=18 p_test=2 exec_time=0.5209841728210449\n",
      "done: p=8 run=18 p_test=4 exec_time=2.279738187789917\n",
      "done: p=8 run=18 p_test=6 exec_time=5.422073602676392\n",
      "done: p=8 run=18 p_test=8 exec_time=11.767040729522705\n",
      "done: p=8 run=18 p_test=10 exec_time=20.213844299316406\n",
      "done: p=8 run=19 p_test=2 exec_time=0.9672756195068359\n",
      "done: p=8 run=19 p_test=4 exec_time=7.967986822128296\n",
      "done: p=8 run=19 p_test=6 exec_time=21.14949131011963\n",
      "done: p=8 run=19 p_test=8 exec_time=39.2610559463501\n",
      "done: p=8 run=19 p_test=10 exec_time=54.60033655166626\n",
      "done: p=8 run=20 p_test=2 exec_time=0.7498040199279785\n",
      "done: p=8 run=20 p_test=4 exec_time=5.116159200668335\n",
      "done: p=8 run=20 p_test=6 exec_time=11.845930337905884\n",
      "done: p=8 run=20 p_test=8 exec_time=24.391905784606934\n",
      "done: p=8 run=20 p_test=10 exec_time=38.539933919906616\n"
     ]
    }
   ],
   "source": [
    "for p in max_ps:\n",
    "    incoming_edges = 3 #max(ceil(dimensions*p/3), 1)\n",
    "    \n",
    "    f1 = {p_test: [] for p_test in test_ps}\n",
    "    runtime = {p_test: [] for p_test in test_ps}\n",
    "    \n",
    "    for run in range(runs):\n",
    "        generator = CausalTSGenerator(dimensions=dimensions, max_p=p, data_length=data_length, incoming_edges=incoming_edges)\n",
    "        ts = generator.generate()\n",
    "            \n",
    "        for p_test in test_ps:\n",
    "            start_time = time.time()\n",
    "            predicted_graph = pc_chen_modified(partial_corr_test, ts, p_test, alpha)\n",
    "            exec_time = time.time()-start_time\n",
    "            \n",
    "            f1[p_test].append(evaluate_edges(generator.graph, predicted_graph)['f1-score'])\n",
    "            runtime[p_test].append(exec_time)\n",
    "            print('done: p='+ str(p) + ' run='+str(run+1) + ' p_test='+str(p_test) + ' exec_time='+str(exec_time))\n",
    "    \n",
    "    for p_test in test_ps:\n",
    "        f1_scores = f1_scores.append({'true_p': p, 'p': p_test, 'mean_f1': np.mean(f1[p_test]),\n",
    "                                      'std_f1': np.std(f1[p_test])}, ignore_index=True)\n",
    "        runtimes = runtimes.append({'true_p': p, 'p': p_test, 'mean_time': np.mean(runtime[p_test]),\n",
    "                                    'std_time': np.std(runtime[p_test])}, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>true_p</th>\n",
       "      <th>p</th>\n",
       "      <th>mean_f1</th>\n",
       "      <th>std_f1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.313380</td>\n",
       "      <td>0.097676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.560156</td>\n",
       "      <td>0.116840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.452907</td>\n",
       "      <td>0.128258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.394392</td>\n",
       "      <td>0.142734</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.385415</td>\n",
       "      <td>0.133961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.240020</td>\n",
       "      <td>0.099248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.391057</td>\n",
       "      <td>0.119675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>6.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.661547</td>\n",
       "      <td>0.125512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>6.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.596114</td>\n",
       "      <td>0.122370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>6.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.563571</td>\n",
       "      <td>0.126436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>8.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.197006</td>\n",
       "      <td>0.085686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>8.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.284110</td>\n",
       "      <td>0.111221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>8.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.398530</td>\n",
       "      <td>0.095938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0.596487</td>\n",
       "      <td>0.126649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>8.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0.579478</td>\n",
       "      <td>0.138812</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    true_p     p   mean_f1    std_f1\n",
       "0      4.0   2.0  0.313380  0.097676\n",
       "1      4.0   4.0  0.560156  0.116840\n",
       "2      4.0   6.0  0.452907  0.128258\n",
       "3      4.0   8.0  0.394392  0.142734\n",
       "4      4.0  10.0  0.385415  0.133961\n",
       "5      6.0   2.0  0.240020  0.099248\n",
       "6      6.0   4.0  0.391057  0.119675\n",
       "7      6.0   6.0  0.661547  0.125512\n",
       "8      6.0   8.0  0.596114  0.122370\n",
       "9      6.0  10.0  0.563571  0.126436\n",
       "10     8.0   2.0  0.197006  0.085686\n",
       "11     8.0   4.0  0.284110  0.111221\n",
       "12     8.0   6.0  0.398530  0.095938\n",
       "13     8.0   8.0  0.596487  0.126649\n",
       "14     8.0  10.0  0.579478  0.138812"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f1_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHwCAYAAAC/hfaiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3Xd8jef/x/HXlT0kEkkQO0iMSITae7a26qLr+6t+O760VKlWq1VanYpWi7a6N6pqq1GrRo2KDEVjh4QkiBEh4/r9cZ9oRBCck/vk5PN8PO5Hc865z32/T4rPua/7GkprjRBCCCFKPiezAwghhBDCOqSoCyGEEA5CiroQQgjhIKSoCyGEEA5CiroQQgjhIKSoCyGEEA5CiroQBSil6iiltiulziilhpqdx5aUUuOVUqlKqWQbHX+1Uuoxy88PKqWW5XuttVLqH6XUWaXUnUqpCkqptZbf+0Rb5BHC0bmYHUAIO/Q8sFpr3QhAKdURGAM0Bk5qrWuYmM1qlFJVgRFAda31cVufT2v9PfB9vqdeAz7SWn9gyfMKkAr46mKeQEMpVQPYD7hqrbOL89xCWJNcqQtxpepAfL7H54AvgJHmxPmXUsqaX8SrA2k3U9CtlKPg77k6sPNmCrqVfy9ClFhS1IXIRyn1O9AR+MjSLBymtd6stf4W2FeE93sopb5TSqUppU4ppbYopSpYXiunlPpSKXVUKXVSKfVrvvc9rpRKUEqdUErNV0pVyveaVko9pZT6B/jH8lxdpdRyy/67lVL35du/h1Jqp6UZ+4hS6rlCcnYBlgOVLJ/zK8vzfZRS8Zbsq5VS9fK954BS6gWlVAxwrrBCqpTqqpTapZRKV0p9BKh8rz2ilPrD8vNeoCawwHL+H4H/A563PO6ilHJSSo1SSu21/D5nKaXKWd5fw/J7+a9S6hDwu+X5FkqpDZb8O5RSHfKdf7VS6nWl1HrL72aZUirQ8vJay39PWc7fspDPNlYp9bNSaqbl/X8ppRpe+adACBNprWWTTbZ8G7AaeKyQ57sAB67z3ieBBYAX4AzchtGcDLAImAn4A65Ae8vznTCanRsD7sCHwNp8x9QYBbgc4Al4A4eBgRi30Bpb3h9u2T8JaGv52R9ofJWsHYDEfI/DMFolulryPQ8kAG6W1w8A0UBVwLOQ4wUCp4F7LO9/FsjO+10CjwB/5Nv/ANAl3+OvgPH5Hg8DNgFVLL+XT4AfLa/VsPxevrH8PjyBykAa0APjgqWr5XFQvv+vey2f09Py+O0Cx3O5xv/bsUBWvs/3HP822Zv+51Y22bTWcqUuhJVlAQFAba11jtZ6m9b6tFIqGOgO/E9rfVJrnaW1XmN5z4PAF1rrv7TWF4AXgZaW+7x53tJan9Banwd6YXy5+FJrna21/guYg1Fs8jLUV0r5Ws71VxGz9wcWaa2Xa62zgPcwil+rfPtM0VoftuQoqAdG8/nPlve/D9xKB7wngdFa60TL72UscE+BFoKxWutzljwPAYu11ou11rla6+XAVkuuPF9qrfdY9p8FRN1gpm35Pt8kwANocXMfTwjrk6IuhHV9C/wG/GRpZn9XKeWKcXV7Qmt9spD3VAIO5j3QWp/FuMKsnG+fw/l+rg40tzQxn1JKncL4YlDR8vrdGIXsoFJqTWFNyVdRMEeu5bxXy1HY+y+9rrXW19n/eqoDc/N9xr+BHKDCVfJUB+4t8HtpAwTn2yf/l4wMoMwNZsr/+XKBRIzPLYRdkM4lQliR5QpuHDDOcqW9GNht+W85pZSf1vpUgbcdxShIACilvDGu9o/kP3S+nw8Da7TWXa+SYQvQ1/Jl4mmMK9KqRYh/FIjIl0NZ3ne1HAUl5T9PvvffrMPAo1rr9QVfyNeKUfD38q3W+vGbOFdRO+fl/3xOGLcGjt7E+YSwCblSF+I6LB22PDDuoypLZzi3q+zbUSkVoZRyxri/nAXkaK2TgCXANKWUv1LKVSnVzvK2H4CBSqkopZQ78Cbwp9b6wFUiLQTClFIPW47jqpRqqpSqp5RyU8Z48LKWLxinMa5ui2IW0FMp1dnyhWAEcAHYUMT3LwLClVJ3WZrIh/Jv68HN+Bh4QylVHUApFaSU6nuN/b8Deiul7lBKOVv+P3VQSlUpwrlSgFyMznvXclu+zzcM4/ezqQjHF6JYSFEX4vraAecxrrarWX5edpV9KwI/YxTTv4E1GMUG4GGMIr8LOI5RFNBarwRewbgvngTUAgZcLYzW+gxwu2WfoxhNyu9gdCbLO88BpdRp4H8Y95qvS2u927Lvhxgd73oDvbXWF4v4/lTgXuBtjNsHocAVV9k34ANgPrBMKXUGo3g2v8b5DwN9gZcwivRhjGGI1/13TmudAbwBrLc03V/tPvk8jL4HJzF+z3dZvjwJYReUcdtLCCHEtSilxmJ0gCzSlyQhzCBX6kIIIYSDkKIuhBBCOAhpfhdCCCEchFypCyGEEA5CiroQQgjhIErc5DOBgYG6Ro0aZscQQgghisW2bdtStdZBRdm3xBX1GjVqsHXrVrNjCCGEEMVCKXXw+nsZpPldCCGEcBBS1IUQQggHIUVdCCGEcBAl7p66EEII+5WVlUViYiKZmZlmRylxPDw8qFKlCq6urjd9DCnqQgghrCYxMREfHx9q1KiBsfquKAqtNWlpaSQmJhISEnLTx5HmdyGEEFaTmZlJQECAFPQbpJQiICDglls4bFrUlVLdlFK7lVIJSqlRhbxeTSm1Sim1XSkVo5TqYcs8QgghbO9GC3r/TzbS/5ONNkpTcljji5DNirpSyhmYCnQH6gP3K6XqF9jtZWCW1roRxtrQ02yVRwghhOM7deoU06bZdykZMmQIZcqUscmxbXml3gxI0Frv01pfBH4C+hbYRwO+lp/LAkdtmEcIIYSDu1ZRz8nJKeY0V9q6dSunTp2y2fFtWdQrA4fzPU60PJffWOAhpVQisBgYYsM8QgghHNyoUaPYu3cvUVFRjBw5ktWrV9OxY0ceeOABIiIiOHDgAA0aNLi0/3vvvcfYsWMB2Lt3L926deO2226jbdu27Nq164rjjx07locffphOnToRGhrKjBkzipwtJyeHkSNH8u67797y57waW/Z+L+zmQMF1Xu8HvtJaT1RKtQS+VUo10FrnXnYgpZ4AngCoVq2aTcIKIYSwrnEL4tl59PR199uZZOxTlPvq9Sv58mrv8Ku+/vbbbxMXF0d0dDQAq1evZvPmzcTFxRESEsKBAweu+t4nnniCjz/+mNDQUP78808GDx7M77//fsV+MTExbNq0iXPnztGoUSN69uyJj48Pbdu2LfS4P/zwA/Xr1+ejjz6iT58+BAcHX/dz3ixbFvVEoGq+x1W4snn9v0A3AK31RqWUBxAIHM+/k9b6U+BTgCZNmsgC8EIIIYqsWbNm1x0mdvbsWTZs2MC999576bkLFy4Uum/fvn3x9PTE09OTjh07snnzZu68885LXyQKc/ToUWbPns3q1atv6jMUlS2L+hYgVCkVAhzB6Aj3QIF9DgGdga+UUvUADyDFhpmEEEIUk2tdUeeXd4U+88mWNsnh7e196WcXFxdyc/9tDM4bQpabm4ufn981C3Oegr3UlVKcOXPmmlfq+/fvJyEhgdq1awOQkZFB7dq1SUhIuOHPcy02u6eutc4GngZ+A/7G6OUer5R6TSnVx7LbCOBxpdQO4EfgEa21XIkLIYS4KT4+Ppw5c+aqr1eoUIHjx4+TlpbGhQsXWLhwIQC+vr6EhIQwe/ZswJgMZseOHYUeY968eWRmZpKWlsbq1atp2rQpPj4+REdHF7rVr1+fnj17kpyczIEDBzhw4ABeXl5WL+hg4xnltNaLMTrA5X9uTL6fdwKtbZlBCCFE6REQEEDr1q1p0KAB3bt3p2fPnpe97urqypgxY2jevDkhISHUrVv30mvff/89gwYNYvz48WRlZTFgwAAaNmx4xTmaNWtGz549OXToEK+88gqVKlWy+ecqKpkmVgghhEP54YcfLnvcoUOHyx4PHTqUoUOHXvG+kJAQli5det3jh4WF8emnn95SxrNnz97S+69GiroQQghT2epeemkkRV0IIYQoorwx7fZKFnQRQlzflz2NTQhh16SoCyGEEA5CiroQQgjhIKSoCyGEMJfc3rEaKepCCCEchj0vvaq1ZvTo0YSFhVGvXj2mTJli9XNIURdCCOEw7Hnp1a+++orDhw+za9cu/v77bwYMGGD1c0hRF0II4TDseenV6dOnM2bMGJycjNJbvnz5W/uwhZBx6kIIIWxjyShIjr3+fskxxn+Lcl+9YgR0f/uqL9vz0qt79+5l5syZzJ07l6CgIKZMmUJoaOj1P/MNkKIuhBDCodnD0qt5x/Pw8GDr1q388ssvPProo6xbt+7GP9A1SFEXQghhG9e4or5M3hX6wEU2iWEPS6/Wr1+fKlWqcPfddwPQr18/Bg4ceMOf5XrknroQQgiHYa9LrwLceeedl5rz16xZQ1hYmDU/OiBFXQghhAPJv/TqyJEjr3g9/9KrvXr1umLp1c8//5yGDRsSHh7OvHnzCj1H3tKrLVq0uKGlV0eNGsWcOXOIiIjgxRdf5LPPPru5D3kN0vwuhBDCodjr0qt+fn4sWmSbWwx5pKgLIYQwl43upZdGUtSFEEKIIpKlV4UQQghRLKSoCyGEEA5CiroQQgjhIKSoCyGEMNXApQMZuNT6E7GURlLUhRBCOAx7Xnp15cqVNG7cmKioKNq0aUNCQoLVzyFFXQghhMOw56VXBw0axPfff090dDQPPPAA48ePt/o5pKgLIYRwGPa89KpSitOnTwOQnp5e5JnoboSMUxdCCGET72x+h10nriyMBeXtU5T76nXL1eWFZi9c9XV7Xnr1s88+o0ePHnh6euLr68umTZuu+3lvlBR1IYQQDs1ell6dPHkyixcvpnnz5kyYMIHhw4dbff53KepCCCFs4lpX1PnlXaF/2e1Lm+Swh6VXg4KC2LFjB82bNwegf//+dOvW7YY/y/XIPXUhhBAOw16XXvX39yc9PZ09e/YAsHz5curVq2flTy9X6kIIIRxI/qVXu3fvTs+ePS97Pf/SqyEhIVcsvTpo0CDGjx9PVlYWAwYMoGHDhlecI2/p1UOHDhV56VUXFxdmzJjB3XffjZOTE/7+/nzxxRe3/oELnsfqRxRCCCFMZK9Lr/br149+/frd8PtuhBR1IYQQprLVvfTSSIq6EEIIUUSy9KoQQgghioUUdSGEEFaltTY7Qolkjd+bFHUhhBBW4+HhQVpamhT2G6S1Ji0tDQ8Pj1s6jtxTF0IIYTVVqlQhMTGRlJQUs6OUOB4eHlSpUuWWjiFFXQghhNW4urped0pWYTvS/C6EEEI4CCnqQgghhIOQoi6EECXUwKUDi7RcqSg9pKgLIYQQDkKKuhBCCOEgpPe7EEKUVMmxZicQBdh6bfjrkSt1IYQQwkFIURdCCCEchBR1IYQQwkFIURdCCCEchBR1IYQQwkFIURdCCCEchE2LulKqm1Jqt1IqQSk1qpDXJyuloi3bHqXUKVvmEUIIIRyZzcapK6WcgalAVyAR2KKUmq+13pm3j9b62Xz7DwEa2SqPEEII4ehseaXeDEjQWu/TWl8EfgL6XmP/+4EfbZhHCCGEcGi2LOqVgcP5HidanruCUqo6EAL8bsM8QgghhEOzZVFXhTynr7LvAOBnrXVOoQdS6gml1Fal1NaUlBSrBRRCCCEciS2LeiJQNd/jKsDRq+w7gGs0vWutP9VaN9FaNwkKCrJiRCGEEMJx2LKobwFClVIhSik3jMI9v+BOSqk6gD+w0YZZhBBCCIdns6Kutc4GngZ+A/4GZmmt45VSryml+uTb9X7gJ6311ZrmhRBCCFEENl16VWu9GFhc4LkxBR6PtWUGIYQQorSQ9dSFEEIIazF5jXuZJlYIIYRwEFLUhRBCCAchRV0IIYRwEFLUhRBCCAchRV0IIYRwEFLUhRBCCAchRV0IIYRwEFLUhd3p/8lG+n8iswYLIcSNkqIuhBBCOAgp6kIIIYSDkKIuhBBCOAgp6kIIIYSDkKIuhBBCOAgp6kIIIYSDkKIuhBBCOAgp6kKI67t4Dk7uh8ObQWuz0wghrsLF7ABCCDuXfQFSd0HWefi8KwSEQtQD0PB+8A02O50QIh+5UhdCXNu6iUZBD6wLfT4C70BYOQ4m14fv7oG4XyAr0+yUQgjkSl0IcS3HdsK6SeAdZBTzxg8bW9peiP4BdvwIPw8EDz+IuAeiHoRKjUAps5MLUSpJURdCFC43B+YPAQ9f8K95+WsBtaDzK9DxJdi/BrZ/D9u/gy2fQfn6RvN8ZH8oU96c7EKUUtL8LoQo3JbP4MhW6PY2OLsWvo+TM9TqBPd8DiN2Q6/J4OYNy16GiXXhhwHw9wLIvli82YUopeRKXQhxpVOHYcU4qN0FIu6FbV9f/z2eftDkUWNL2Q3R38OOmbBnCXgFQMR90OhBqBhh+/xClFJypS6EuJzWsGi48XOvyTd3fzyoDnR9DZ6NhwdmQ422sPVz+LiNsW36GM6lWTe3EEKu1IUQBcTNgX+WGc3uftVu7VjOLhB2u7FlnIDYn40r+KUvGE30dbpB1ENGi4Cz/HMkxK2Sv0VCiH+dS4Mlz0PlJtDsCese26scNH/C2I7FW3rP/2Tcc/cuDw37G73ny9ez7nmFKEWk+V0I8a/fXoLMdOjzodEJzlYqhMMdb8CIXTDgB6jaDDZNh2kt4NOORie98ydtd34hHJRcqQshDAkrIOYnaPc8VKhfPOd0doW6PY3tbArEzjaa5xeNgKUvGc9HPQi1Otr2S4YQDkKKuhACLpyFBc9CYBi0e86cDGWCoOVgaDEIknYYzfOxsyD+F/CpBA0HGAU+sLY5+YQoAaSoCyFg1ZuQfggGLgUXd3OzKAWVoozt9tdhz1Jjcpv178Mfk6Bqc2Nym/C7jIlxhBCXSFEXorRL3AZ/Tocm/4XqLc1OczkXd6jf19jOJEPMTKPAL3gGloyCer2Nse812oGTdBESQoq6EKVZTpYxFWyZitDlVbPTXJtPRWj9DLQaCkf+gujvIHaO0URftqqxalzUA1AuxOykQphGiroQpdn6D+B4PAz4ETzKmp2maJSCKrcZ2x1vwq5FRue6tRNg7btQvbVx771+X3AvY3ZaIYqVFHUhSqvUf2DNu1D/Tqjbw+w0N8fV01gdLuIeSD9irBoX/QPMGwyLR0L4nUaBr95KVo4TpYIUdSFKo9xcmD/UKIo9JpidxjrKVjZ67rcdAYf/NK7e4+Ya//WvYRT3hveDX1WzkwphM9KzRIjS6K+v4NAGYwIYR1seVSmo1sKYQOe53dDvE2O621VvwPsR8HUfiJkFFzPMTiqE1cmVuhClzemjsPxVCGlnXL06MjdvY3x7wwFw8qAxLW309/DL4+DuC+H9oNFDUKWpNM8LhyBFXYjSRGtY9BzkXITeH5SuQuZfHTq8AO1GwsH1lsltZsNfX0NAqNFzvuEA8K1kdlIhbpo0vwtRmvw9H3Yvgo4vQbmaZqcxh5MThLSFftPhuT3Qdyp4B8HKcTA5HL67G+J+gaxMs5MKccPkSl2I0uL8SaNHeMVIaPGU2Wnsg7uP0fze6CFI22vpPf8j/DwQPPyMXvVRD0ClxqWrVUOUWFLUhSgtlo+Bc6nwwCxZu7wwAbWg08vQ4UXYv9a49779O2PFuKB6xsx1EfeBTwWzkwpxVfI3W4jSYP9a+OsbY0a2SlFmp7FvTs7GqnC1OhrL0Mb9YhT4ZS8bHQxDbzeu3sO6gYub2WmFuIwUdSEcXdZ5Y650/xDjKlQUnUdZaDLQ2FL2GMV9x0+wZwl4BRhX7lEPQHCk2UmFAKSjnBCOb/XbcGIf9JliTDYjbk5QGHQdB8/Gw4M/Q422sPVz+KQtfNwGNk2Hc2lmpxSlnFypC+HIknbAhg+h0cPGuHRx65xdILSrsWWcgLg5xr33paNg2StQp5sx/r92F3B2NTutKGWkqAvhqHKyjRXYvAKMdcmF9XmVg2aPG9uxeGPse8xM+HsBeJeHyPuMnvXl65mdVJQS0vwuhKPaNM24Uu8xATz9zU7j+CqEG9PuDv/bWPWuajP482OY1gI+7QibZxjDCoWwIblSF8IRndgHq96EOj2NJUhF8XF2NVa9q9vDGEIYM8voYLf4OfjtJajbE6IeMnrXOzmbnVY4GCnqQjgarWHBMKO49HxPJk0xk3cgtBxsbEk7YPv3EDsL4ueCT7AxLW3UgxAYanZS4SBs2vyulOqmlNqtlEpQSo26yj73KaV2KqXilVI/2DKPEKVC9A+wfw10GSvzmNuT4IbQ410YsRvu+8Z4vH4KfNQEPusK274yxsULcQtsdqWulHIGpgJdgURgi1JqvtZ6Z759QoEXgdZa65NKKQdbA1KIYnb2uNHEW60V3DbQ7DSiMC7uxi2R+n3hTLLRsW7798ZcAktGQb3exux1NdoZ89QLcQNs2fzeDEjQWu8DUEr9BPQFdubb53Fgqtb6JIDW+rgN8wjh+JY8D1kZxph0KQj2z6eiMctfq6Fw5C/j3nvcz0YTfdmq0PB+iLq/9C6+I26YLf/WVwYO53ucaHkuvzAgTCm1Xim1SSnVzYZ5hHBsuxYb92rbPy/3aEsapaDKbdBrEozYA/d8AYFhsHYCTGkEX/YwxsJfOGt2UmHnbHmlXljvHF3I+UOBDkAVYJ1SqoHW+tRlB1LqCeAJgGrVqlk/qRAlXeZpWDQCyodDq2fMTiNuhasHNLjb2NKPQMxPRvP8vKdg8fMQfqcxNW311mYnFXbIlkU9Eaia73EV4Ggh+2zSWmcB+5VSuzGK/Jb8O2mtPwU+BWjSpEnBLwZCiJXj4EwS9P9OFhlxJGUrQ9sR0GY4HN4M0d9B3FyI/p5c/+qUd8vkoKsrs3d+T5BPZYI8gwj0DCTAMwAXJxncVBrZ8v/6FiBUKRUCHAEGAA8U2OdX4H7gK6VUIEZz/D4bZhLC8RzaZCwP2mKw0YQrHI9SUK25sXV7G/5eyMToKSy2NH7Gb3n78t1R+LuXJcirAoFegZT3LE+gZyBBXkGXCn/ez27O8iXQkdisqGuts5VSTwO/Ac7AF1rreKXUa8BWrfV8y2u3K6V2AjnASK21rIggRFFlXzCmgi1bDTqONjuNKA5u3nzvls03nKZqjqJr5kXuD7uP1OOxpJxMICXzBKnOzqQ4nybVPY0Udy/+cVKk5V4g54o7oODr5msU+rzi7xVIkOeVxd/L1cuEDytulE3bZ7TWi4HFBZ4bk+9nDQy3bEKIG7VuIqTugQfngHsZs9OIYrDy0Ere2fwOnap2Iv3QH8R4ufFsl9eomLfD+VNwLA6SYy1bDBzfRU5uFiednEh19yIlIIRUv0qkePmT4uZBisol5cIJth3bRsr5FLJys644r5eL16UCn/cloGDhD/QMxNfNFyUTHplGbroIUVId2wnrJkFkfwjtYnYaUQx2pOzghbUvEBEYwdvt3mbwd22v3MnTD2q0MbY82RdxTt1NYHIsgcmx1E2OhYRN/052o5wgoDZUjEDXuIPTgbVIKRtMCjmknk8l5XwKKRkpl/4bnxZPSmIK57PPX3F6d2d3o9B7BhHkFXTp54LF39/DHyclwy6tTYq6ECVRbo7R7O7hC3e8ZXYaUQwOnT7EkJVDKO9Vng87f4ini2fR3+ziBhUjjC2P1pB+ON8VfSwkbkHFzaEsUBaoXabCv++rGAE17zLGzFvmrD+Xde5SsU89n8rxjOOXvgSkZqSy99ReNiVt4szFM1dGUi4EeAZcdtV/xc/S6e+GyW9KiJJo8ww4shXu+gy8A8xOI2zsZOZJBq0YhEYzvct0ynmUu/WDKgV+1Yytbs9/nz9/0lhGNn/z/YY1kNck7+plrEhXMQLvihF4V4ykRvlwcLv6PffM7MwrrvhTz6eSkmH89+jZo8SkxHAi88SVMVGU8yh3zav+8l5GR0Dp9CdFXYiS59QhWPka1O4KEfeYnUbYWGZ2JkN+H8KxjGN8dvtnVPetbtsTevoX2nxPyq7Lr+pj58DWL4zX8zXf/7tFQhlj5m8PFw+q+FShik+Va546KzeLtPNplxd+yxeBvJ/3nNhDWmYaOTrniveXdS/7b9EvcNWf/0uAI3f6k6IuREmiNSy09CvtNUlWYHNwObk5vLjuRWJSYpjYYSJR5aPMCeLiBsGRxpZHa+MLZv5Cf3gLxM35d5+CzfcVIy9rvi/I1cmVit4VqehdsdDX8+Tk5nDywskrrvrzfxHYemwrqedTC+305+3qXXjxL9AR0MfVp8R1+pOiLkRJEvszJCyHbu8YzabCob239T1WHFrB802fp2v1rmbHuZxS4F/d2Or1+vf58ychOe7yYr9vNeRmG69far6P/LfQl693zeb7gpydnAn0DCTQM5B61Lvqflpr0i+kG1f7BZr881oA4tLiSE1MLXKnv7ym/vy3Afzc/eym058UdSFKinNpsPQFqNwEmj1erKeOTzJ6SYcX61lLt293fst3f3/HQ/Ue4uH6D5sdp+g8/SGkrbHlyb4AKbsLNN//DFs/N15XThAQWkjzfdAtRVFK4efhh5+HH6H+V18PQWttdPor5Ko/r/NfwqkENh3dxJmsa3f6SyQLV8sxzbjKl6IuREnx20vGEKQ+H161+VI4hhUHVzBhywS6VOvCc02eMzvOrXNxv0rz/cECzfd/GqvU5SlT8SrN99a9KlZKUcatDGXcyhBSNuSa+57PPk/q+dTLin/+LwEX0FywHNMMUtSFKAn+WWEs7NHueahQ3+w0woaij0czat0oIoIieKvtWzg76hc4pcC/hrHV6/3v8xknCkyeEwv7VuVrvve+1Pv+Zpvvb4WniydVfapS1adqoa8P/KpJseS4GinqQti7C2dh4bPGUpztHOCqTVzVwdMHGfL7ECp4VeDDTh/i4eJhdqTi51UOQtoZW57sC4X0vp9t8+b7kkiKuhD2btUbkH4IHv3NaMYUDulE5gkGrRiEQllvLLqjcHGH4IbGlqcozfc+wVcWev8Qqzff2xMp6kLYs8StsGk6NH0MqrUwO42wkfPZ5xmycgjHM47z+R2fU81XRjZc17Wa7/MofEroAAAgAElEQVQX+uRYSFgJeePaXb2hYoPLi335+uB6AzP02TEp6kLYq+yLMH+ocbXR+VWz0wgbycnNYdTaUcSmxjK5w2QaBjW8/pvE1XmVg5rtjS1PVuaVzfc7ZhpLFoPRfB8YduVVvXegOZ/hFkhRF8JebfgAjsfDgB+NOd6Fw9Fa8+6Wd/n98O+MajaKztU7mx3JMbl6QKUoY8uTm3tl8/3Bjca9+jwlsPleiroQ9ihlD6x5F8L7Qd0eZqcRNvLtzm/5YdcPPFz/YR6s96DZcUoXJycoF2Js9fv8+/z1mu/dyhTofW9fzfdS1IWwN7m5sOAZY+at7u+anUbYyLIDy3hv63t0rd7VMcaiO4qrNt//fY3me+dLzffd0k9yyM3d+HtswhW9FHUh7M22L+HQBug79dKCGMKxbD++nRfXvUjDoIa82eZNu5liVFyFqwdUamRseXJz4dSBAs336+l/OpVzTk6mrcsgRV3YnfMXc3BxLlmLKFjN6aOw/FVjjG6UNMc6ogPpBxjy+xCCywQzpdOU0jkW3RE4ORmz25WrCfX7Xnp6yOeNCMjOYqxJRV2+Hgq7siv5NLFH09mRmM53mw6Sm6vNjlR8tIZFI4yZs3p/ICuwOaC082kMWjEIZ+XM9M7T8ffwNzuSsLKzzs4cdDfvi5oUdWE3MrNyGPZTNM5K4e3mzMu/xnHX9A3sPHra7GjFY+c82L0YOr5kfPsXDuV89nmG/D6E1POpfNjpQ6r6Fj7NqBC3Qoq6sBsTl+1mV/IZagZ5U7eiD5P7N+TwiQx6f/QH4xfu5NyFbLMj2s75k7B4pDFjVovBZqcRVpaTm8MLa18gLjWOd9q9Q2RQ5PXfJMRNkKIu7ML6hFRmrNvPQy2q4e/lhlKKfo2qsHJEe+5rUpXP/thPl0lr+C0+Ga0dsEl+2SuQkWaswOYsXV0cidaatze/zarDqxjVbBSdqnUyO5JwYFLUhenSM7IYMWsHNYO8Gd3j8hXI/LzceOuuCOYMaklZT1ee/HYbj3+zlcSTGSaltYF9a2D7t9BqyOVzWwuH8M3Ob/hp90/8X/3/44F6D5gdRzg4KerCVFprRv8aS+rZC7zfPwpPt8KXmbytejkWDGnDSz3qsj4hja6T1vLxmr1k5eQWc2IryzpvjEn3D4EOo8xOI6zstwO/8d7W97i9+u0MbzLc7DiiFJCiLkw1L/ooC2OSeLZrGJFV/K65r6uzE0+0q8WKEe1pXTuQt5fsoteUP9h64EQxpbWB1W/Byf3QZ4rdzEglrOOvY3/x0rqXaFS+EW+2lbHoonjInzJhmsSTGbzyaxxNqvvzv/a1ivy+yn6efPZ/Tfj04ds4k5nFPR9vZNScGE5lXLRhWhs4Gg0bPoJGD1++drQo8fan72fI70OoVKYSUzpOwd1ZlswVxUOKujBFTq5m+KwdaGBy/yicnW58TPbt4RVZPrw9T7SryextiXSauIY52xJLRke6nGyYPwS8AuD2181OI6wo9Xwqg1YMwsXJhWldpuHnce0WKCGsSYq6MMWna/exef8JxvUJp2o5r5s+jre7Cy/1qMeCp9tQPcCLEbN3cP+MTSQcP2vFtDawaSokx0CPCeApE5A4ioysDIasHELa+TSmdp5KVR8Ziy6KlxR1UezijqQzafluekYEc1fjylY5Zv1Kvsz5Xyve7BfBzqOn6f7BWiYu201mVo5Vjm9VaXth1ZtQt9dl00uKki07N5sX1r7AzhM7ebfduzQIbGB2JFEKSVEXxer8xRye+Wk75bzdeKNfA5QVp0J1clI80LwaK0d0oFdkJT78PYE73l/L2j0pVjvHLdMaFg4DZzfjKl2mgnUIeWPRVyeuZlSzUXSs1tHsSKKUkqIuitXbS/5mb8o5Jt4bhZ+Xm03OEeTjzuT+UfzwWHOcleI/X2zm6R/+4vjpTJuc74ZEfw/710LXceBbyew0wkq+iv+KmbtnMjB8IPfXvd/sOKIUk6Iuis2q3cf5euNBHm0dQpvQQJufr1XtQJYMa8uzXcJYtvMYnSeu4ZuNB8gxa5GYM8fgt9FQrRU0fsScDMLqlu5fyqRtk+hWoxvDbhtmdhxRyklRF8XixLmLPP9zDGEVyvB8tzrFdl53F2ee6RLKb8Pa0bCqH2PmxXPXtPXEHUkvtgyXLHkesjKMMelO8lfPEWxN3spLf7xE4/KNGd9mfLGPRT93MZtzFx14TQRxw+RfFmFzWmtGzYkhPSOL9/s3wsO18FnjbCkk0Jtv/9uMDwZEceRUJn0++oNxC+I5k5lVPAF2LYKdv0L75yEwtHjOaUWvBUzgtYAJZsewK/tO7WPoqqFULlOZKZ1kLLqwD1LUhc3N3prIsp3HGHlHHepX8jUth1KKvlGVWTmiPQ80r8ZXGw7QZdIaFscm2XZse2a6sU56+XBo9YztziOKTer5VAavHIybkxvTu0ynrHtZsyMJAUhRFzZ2MO0cYxfE07JmAP9tE2J2HADKeroy/s4IfhnUigBvdwZ//xePfrWFwydstEjMinFw9pixApuLbToHiuKTkZXBUyuf4kTmCaZ2nkoVnypmRxLiEinqwmayc3IZNjMaFyfFxPsa4nQTs8bZUqNq/sx/ujUv96zHn/tP0HXyGqauSuBithUXiTm4EbZ+Ds0HQZXbrHdcYYrs3GxGrh3JrhO7mNBuAuGB4WZHEuIyUtSFzUxdtZfth07xRr8IKvnZ52IlLs5OPNa2JiuGt6dDWHkm/LabnlPWsXm/FRaJycqEBUPBrxp0Gn3rxxOm0lrz1p9vsTZxLaObj6Z91fZmRxLiClLUhU1sP3SSKb//Q79Glend0P7HY1fy8+Tjh2/j8/9rQsbFHO77ZCMjZ+/gxLlbWCRm3URI3QO9JoObt/XCClN8EfcFs/bM4tEGj3JfnfvMjiNEoaSoC6s7dyGbZ2dGU9HXg3F9S1bzZOd6FVg+vB3/a1+LuduP0HniamZtPXzjHemOxcMfkyByANTuYpuwotgs3reY9/96n+4h3XmmsXR2FPZLirqwuvGLdnLwRAaT7muIr4er2XFumJebC6O612Xh0DbUCirD8z/H0P+TTew5dqZoB8jNgflDwaMs3PGmbcMKm9uSvIWX179MkwpNGN+6+MeiC3Ej5E+nsKpl8cn8uPkwT7arRfOaAWbHuSV1K/oy68mWvHN3BHuOn6HHB+t4d+kuzl+8ziIxmz+FI1uh2zvgXbJ/B6Xd3lN7eWbVM1T1qcr7Hd/HzVlGLwj7JkVdWM3xM5mM+iWW+sG+DO8aZnYcq3ByUvRvWo2Vw9vTN6oy01bvpevkNazadbzwN5w8CCtfh9pdIeKe4g0rrColI4XBKwbj7uzOtC7TZCy6KBKzZ/mToi6sQmvNCz/HcO5CNh8MiMLNxbH+aAWUcWfifQ358fEWuLs4MfCrLQz+fhvJ6fkWidEaFj5r/NxrkqzAVoLljUU/eeEkH3X+iMplrLNEsBC25lj/8grTfPfnIVbtTuGlHvUIreBjdhybaVkrgMXPtOW528NY+fdxOk9czRd/7DcWiYmdDXtXQucxxjA2USJl52bz3Jrn2H1yN++1f4/wgJLV2VOUbi5mBxAlX8Lxs7yxaCftw4L4T8vqZsexOXcXZ57uFErvhpV4+dc4Xlu4kxXb4vkm43lcqjSFZo+bHVHcJK014zeNZ92RdYxpOYZ2VdqZHUmIGyJFXdySi9m5DJu5HU9XZybcE4kqRU3O1QO8+ebRZiyKTcJp7pPk5p7mozJD+c/FXHw9in/RGnHrPo/7nDn/zOGxiMe4N+xes+MIccOkqItb8sHKPcQdOc3HD91GeV8PqxxzTNpIy09/WOV4tqSUopfnTtBrWR08kIk7nPl67xrG9KpPr8jgUvUlp6RbuG8hH/z1AT1CejC00VCz4whxU+Seurhpm/efYNrqvfRvUpVuDSqaHcccF87CwmEQGEaHx95h3lOtqeDrzpAft/N/X27hYNo5sxOKItictJlX1r9C04pNeb316/JlTJRYUtTFTTmdmcWzM6OpVs6LMb3rmx3HPL+Ph/TDlhXY3Ims4se8p9rwau/6/HXwJLdPXsuHK//hQvZ1xrYL0yScTGDYqmFU96nO5A6TZSy6KNFsWtSVUt2UUruVUglKqVGFvP6IUipFKRVt2R6zZR5hPWPnx5OUfp5J90Xh7V5K7+IkboU/P4amj0G1FpeednZSDGwdworh7elSrwITl++hxwfr2Lg3zcSwojDHM44zaOUg3F1kLLpwDDYr6kopZ2Aq0B2oD9yvlCrskm6m1jrKsn1mqzzCehbFJPHLX0d4ulMot1X3NzuOObIvwvwh4BMMnV8tdJeKZT2Y+mBjvnykKRdzcrl/xiaGz4om7eyFYg4rCnMu6xxPr3ya9AvpTOs8jUpl7H/hISGux5ZX6s2ABK31Pq31ReAnoK8NzyeKQXJ6Ji/NjaVhVT+GdKptdhzzrP8Aju80Jpnx8L3mrh3rlmfZsPY81bEWC3YcpdPENfy0+RC5uTe4SIywmuzcbEasGcGek3uY2H4i9QLqmR1JCKuwZVGvDBzO9zjR8lxBdyulYpRSPyulqtowj7hFubma52bv4GJ2Lu/3j8LVuZR2yUjZA2vfhfC7oE73Ir3F082ZkXfUZfHQttSp4MOoX2K595ON7Eo+beOwoqC8sejrj6zn5RYv07ZKW7MjCWE1tvxXubDuowUvTRYANbTWkcAK4OtCD6TUE0qprUqprSkpKVaOKYrqyw0H+CMhlTG96xMSWErXB8/NhQVDwdULur9zw28PreDDzCdbMOGeSPalnKXXlD94a8nfZJg4V3RpMyN2BnP+mcPjEY9zT5jMzy8ciy2LeiKQ/8q7CnA0/w5a6zStdd4NxhnAbYUdSGv9qda6ida6SVBQkE3CimvblXyad5buoku9CgxoWoobVLZ9CYc2Gkuqlil/U4dQSnFvk6qsHNGBuxpX5pM1++g6aS0rdh6zclhR0IK9C/hw+4f0rtmbIY2GmB1HCKuzZVHfAoQqpUKUUm7AAGB+/h2UUsH5HvYB/rZhHnGTMrNyGPZTNL4errxzd0TpHcObfgSWvwoh7SHqgVs+XDlvN969pyEzn2iBl5szj32zlSe/3crRU+etEFYU9GfSn4zZMIZmFZsxrtW40vvnWDg0mxV1rXU28DTwG0axnqW1jldKvaaU6mPZbahSKl4ptQMYCjxiqzzi5k1ctptdyWeYcE8kAWXczY5jDq1h0QjIzYbe71t1BbbmNQNYNLQtz3erw5o9KXSZtIbP1u0jOyfXaue4VQfc3uOA23tmx7hp/5z8h2GrhlHDtwaTO07G1dnV7EhC2IRNBxhrrRcDiws8Nybfzy8CL9oyg7g16xNSmbFuPw+3qE7HujfX3OwQdv4Ke5ZA19ehXE2rH97NxYnBHWrTO7ISr8yLY/yiv5nz1xHe7NeARtVK6bBBKzl27hiDVgzCy8WLaZ2n4et27dEKQpRkpbT7siiK9IwsRszaQc0gb17qUYqH/GScgMUjIbghtBhs01NVLefFl480ZfqDjTlx7gJ3Td/A6LmxpGdk2fS8jupc1jmeWvkUZy6eYWqXqQSXCb7+m4QowaSoi0JprRn9ayypZy/wfv8oPN1K8apjy18xCnufD8HZ9rPnKaXoHhHMiuHteaRVDX7cfIjOk1YzL/oIWsvY9qLKys1ixOoRJJxKYFKHSdQtV9fsSELYnBR1Uah50UdZGJPEs13DiKziZ3Yc8+xbDdu/g1ZDjCv1YuTj4cqrvcOZ/3QbKvl58sxP0Tz0+Z/sSzlbrDlKIq01r298nfVH1zOm5RhaV25tdiQhioUUdXGFxJMZvPJrHE2q+/O/9rXMjmOeixmwYJhxD73DFUsXFJsGlcsyd3BrXusbTszhdLq9v473V+whM0sWibmaT2I+YW7CXJ6MfJK7Qu8yO44QxUaKurhMTq5m+KwdaGBy/yicnUrxsJ81b8PJ/dD7A3D1NDWKs5PiPy1rsHJEe+5oUJH3V/xDjw/WsT4h1dRc9mhewjymRk+lT60+PBX1lNlxhChWN1XUlVIDrR1E2IdP1+5j8/4TjOsTTtVyXmbHMc/RaNjwETT+D4S0MzvNJeV9Pfjw/kZ8/WgzcrTmwc/+ZNhP20k5I4vEAGxK2sTYDWNpHtycsS3Hylh0Uerc7JX6OKumEHYh7kg6k5bvpmdEMHc1Lmya/lIiJwvmPw3egdD1NbPTFKp9WBC/DWvH0E61WRSbROeJq/n+z4OlepGYPSf38OyqZ6lRtgaTO8hYdFE6XbUrr1Iq5movARVsE0eY5fzFHJ75aTvlvN14o1+D0n2Fs3EqJMfCfd+Ap/2OEfdwdWb47XXoE1WZl3+NZfTcOH7elsgbd0ZQv1LpGoudfC7ZGIvu6sX0LtPxcfMxO5IQprjW+JwKwB3AyQLPK2CDzRIJU7y95G/2ppzju/82x8/Lzew45knbC6vfgrq9oF6f6+9vB2qXL8OPj7dg7vYjjF/0N70/+oNHW9dgWJcwvN1tPwTPbGcvnuWplU9xLuscX3f7moreFc2OJIRprvU3fiFQRmsdXfAFpdRqmyUSxW7V7uN8vfEg/20TQpvQQLPjmEdrWPAMOLtBjwlWnQrW1pRS3NW4Cp3qluedpbuYsW4/C2OSGNsnnDvCHbfIZeVmMXz1cPad2sfUzlOpU66O2ZGEMNW17qmP11r/UdgLWutbX81C2IW0sxd4/ucY6lTwYeQdpfwfxO3fwYF10HUc+FYyO81N8fNy4627Ivn5fy3x9XDlyW+38djXW0k8mWF2NKvTWjNuwzg2Jm1kTMsxtKrcyuxIxe6gay0OupbiYafiCtcq6j8DKKVWFlMWUcy01rz4izEF6eT+UXi4luJZ484cg2WjoXpraPyI2WluWZMa5Vg4tA0vdq/L+oRUuk5ayydr9pJlR4vE3KqPd3zMvL3zGNRwEP1C+5kdRwi7cK3mdyel1KtAmFJqeMEXtdaTbBdLFIfZWxNZtvMYo3vUK3Udq66w5HnIyjTGpDs5xvQNrs5OPNm+Fj0jgxk7P563luxi7vYjvNGvAbdVL2d2vFvya8KvTNsxjb61+jKo4SCz45imxsXnzI4g7My1/vUaAGRiFH6fQjZRgh1MO8fYBfG0rBnAf9uEmB3HXLsWGauwtX8eAkPNTmN1Vfy9mPGfJnzy8G2kn8/i7ukbefGXGE5lXDQ72k3ZcHQD4zaMo2VwS15t9WrpHqkhRAFXvVLXWu8G3lFKxWitlxRjJmFj2Tm5DJsZjYuTYuJ9DXEqzbPGZaYb66SXD4fWz5idxmaUUtwRXpE2tQOZvHwPX244wLL4Y4zuWY9+jSqXmMK4+8Ruhq8eTk2/mkzqMAlXJxmLLkR+121nlILueKau2sv2Q6d4o18ElfzMnf7UdCvGwtljlhXYHL9AeLu78HKv+sx/ujVVy3kxfNYO7p+xiYTj9r9ITPK5ZAavHIy3qzdTO0+ljFsZsyMJYXcc4+ahKLLth04y5fd/6NeoMr0blswe3lZzcANs/QKaD4Iqt5mdpliFVyrLL4NaMf7OBsQfPU33D9Yycdluu10k5szFMwxeOZiMrAymd5kuY9GFuAop6qXIuQvZPDszmoq+HozrG252HHNlZcL8oeBXDTqNNjuNKZycFA+1qM7vIzrQMyKYD39P4I7317J2T4rZ0S6TlWOMRd9/aj+TOkwizD/M7EhC2K0iFXWlVCul1ANKqf/kbbYOJqxv/KKdHDyRwaT7GuLr4fhNzde07j1I+wd6TQY3b7PTmCrIx533BzTi+8ea46QU//liM0//8BfHT2eaHQ2tNWM3jjUWamk1lpaVWpodSQi7dt05JJVS3wK1gGggr21OA9/YMJewsmXxyfy4+TCDOtSiec0As+OY61g8/DEZIgdA7S5mp7EbrWsHsuSZtny8Zi/TVu1lze4URnarw4PNq5uWadqOaczfO5/BUYPpW7uvaTmEKCmKMjF0E6C+1rr0Lv9Uwh0/k8moX2IJr+TLs11KedNlbg7MHwIeftDtLbPT2B0PV2eGdQmjT8NKvDIvjjHz4pmzLZEs5wBcPdKKNcvcf+by8Y6P6Ve7H/+L/F+xnluIkqooze9xgPRKKaG01rzwcwznLmTzfv8o3FxKeTeKPz+BI9ug+zvgVbInYLGlmkFl+O6/zXm/fxRHTp3nxMHenD7WktjEdIrj+/36I+sZt3EcrSq14pWWr5SYIXdCmK0oV+qBwE6l1GbgQt6TWuuSsYRVKffdn4dYtTuFcX3CCa1QyucMOnkQfn8dQm+HBnebncbuKaW4s1FlOtYpT4sPPuD8qTr0/ugPqgd40TMimF6RlagX7GP1grvrxC6Grx5Obb/aTGw/UcaiC3EDilLUx9o6hLCNhONneWPRTtqHBfGflubdF7ULWsPCZwEFPSeVqBXYzFbWyxXfihspE/QXzzV8m4UxSXyydh/TVu+lZqA3PSONAh9WocwtF/iks0kMXjEYHzcfGYsuxE24blHXWq8pjiDCui5m5zJs5nY8XZ2ZcE+kNF/GzIK9K6H7u+BX1ew0JZKT8wX6N61G/6bVSDt7gaXxySyKSWLqqgQ+/D2B2uXL0CsymF6RwdQuf+OtQqcvnmbwysGczz7PN92/oYJ3BRt8CiEcW1F6v7cAPgTqAW6AM3BOa13KVwCxbx+s3EPckdN8/NBtlPf1MDuOuc6lwtJRUKUpNH3M7DQOIaCMOw82r86DzauTcuYCS+OSWBiTxAcr/+H9Ff9Qp4IPvSKD6RkZTM2g619tZ+VkMXzVcA6cPsDHXT4m1N/x5uAXojgUpfn9I4zFXWZj9IT/DyB/4+zY5v0nmLZ6L/2bVKVbA+njyNIX4cIZYypYp1K8vKyNBPm483DLGjzcsgbHTmeyJDaJRbFJTFy+h4nL91A/2NfSRB9M9YAr5wTQWvPqhlf5M/lP3mzzJs2Dm5vwKYRwDEUp6mitE5RSzlrrHOBLpdQGG+cSN+l0ZhbPzoymWjkvxvSub3Yc8/2zHGJnQfsXoHw9s9M4vAq+HjzSOoRHWoeQlH6exbHJLIo5yoTfdjPht91EVC5Lz8hgekYEU7WcFwAfRX/Egn0LeDrqaXrX6m3yJxCiZCtKUc9QSrkB0Uqpd4EkoHRPwWXHxs6PJ/l0JrP/1xJv9yJ9Z7M74cFlrXOgC2eNznGBdaDtCOscUxRZcFlP/tsmhP+2CSHxZAZLYpNZGHOUt5fs4u0lu2hY1Y8aNWL5PfVT7g69mycinzA7shAlXlH+1X8YYzz708CzQFVAxgPZoUUxSfzy1xGGdg6lcTV/s+OY7/fXIT0RHv0NXNzNTlOqVfH34vF2NXm8XU0On8hgUWwSs+OXszJlOjnnwtgR3Ykvcg/QMyKYimVLeR8QUaK9mmruiI2i9H4/qJTyBIK11uOKIZO4Ccnpmbw0N5aGVf0Y0qm22XHMd3iLMdFM08egmtyjtSdVy3nRvsFFvkz8klpetelQdSzL4k7x+sKdvL5wJ01r+NMrshLdG1SUTp5C3KCi9H7vDbyH0fM9RCkVBbwmk8/Yj9xczXOzd3AxO5f3+0fh6lzKZ43LvmhMBetbCTqPMTuNKODo2aMMXjmYsu5lmXHHdMp7lefZzrA35SyLYpJYFJPEq/PjGbsgnmY1ytGrYSW6hVckyEdaW4T9ey1gAgAzTTp/USefaQasBtBaRyulatgskbhhX244wB8Jqbx1VwQhgdLdgfXvQ8rfcP9M8JCRl/bk9MXTDF4xmAvZF5jRfQblvcpfeq1WUBmGdg5laOdQ/jl2hoUxSSyMOcorv8bx6rw4WtQMoFdkJbo1qEg5bzcTP4UQ9qsoRT1ba51e6icvsVO7kk/zztJddKlXgQFNZVIVUnbD2gkQfhfU6WZ2GpHPxZyLDFs1jINnDvJJl0+o7X/120ShFXx4tqsPw7qEsvvYGRbFGOPgX5obyyvz4mhVK4BekcHcEV4RPy8p8ELkKUpRj1NKPQA4K6VCgaGADGmzA5lZOQz7KRpfD1feuTtCZo3LzYX5Q8HVy1iwRdgNrTVjNoxhS/IW3mr7Fs2CmxXpfUop6lb0pW5FX4Z3DWNn0ulLBf6FObGMnhtHm9BAekYEc3t4Rcp6yjzxonQrSlEfAozGWMzlR+A34HVbhhJFM3HZbnYln+HLR5oSUEbuN7LtCzi8CfpOgzLlr7+/KDYfbv+QRfsWMbTRUHrV7HVTx1BKEV6pLOGVyjLyjjrEHTnNwtijLIpJYuTPMbw0N5Z2oUH0jAyma/0K+HhIgRelT1F6v2dgFPXRto8jimp9Qioz1u3n4RbV6VhXChjpR2D5WAhpD1EPmJ1G5DN7z2xmxM7gnrB7eCzCOtP0KqWIqFKWiCplGdWtLjsS01kUYxT4lbuO4+biRPuwIHpFBtO5XgXKlNA5G4S4UUXp/d4EeAmokX9/rXWk7WKJa0nPyGLErB3UDPLmpR4ySxpaw6IRkJsNvT+QFdjsyNrEtbyx6Q3aVm7L6OajbXKLSClFVFU/oqr68WL3emw/fIqFMUdZHJvE8p3HcHdxomOd8vRqGEynuuXxcpMCLxxXUf50fw+MBGKBXNvGEdejtWb0r7Gknr3A3P+0xtNN5jInfi7sWQK3j4dyIWanERbxafE8t+Y5wvzDeK/9e7g42b6YOjkpbqvuz23V/XmlZ322HTrJwh1HWRyXzNL4ZDxcnehctwK9IoPpUKe8/P0RDqcof8tStNbzbZ5EFMm86KMsjEli5B11iKhipelUS7KME7DkeQiOguaDzE4jLI6cPcLTK5/G392fqZ2n4uXqVewZnJwUTWuUo2mNcozpHc6WAydYGHOUJbHJLIpNwsvNmS71KtAzMpj2YUF4uEqBFyVfUYr6q0qpz4CVGJ3lANBa/2KzVKJQiSczeOXXOJpU9+d/7WuZHcc+LHvFKOwP/QLO0qxqDxV13GEAACAASURBVNIvpBtj0XMu8NntnxHkFWR2JJydFC1qBtCiZgBje4fz5/4TLIxJYmlcEvN3HKWMuwtd61egZ0QwbcMCcXeRAi9KpqL8KzgQqAu48m/zuwakqBejnFzN8Fk70MDk/lE4O8l9Y/augujvoM2zECxdPOxB3lj0w2cO80nXT6jlZ39fPl2cnWhdO5DWtQN5rW84G/emsSgmiaXxyczdfgQfDxdur1+RXpHBtK4diJtLKZ+hUZQoRSnqDbXWETZPIq7p07X72Lz/BBPvbXhpycpS7WIGLBwG5Woay6oK0+XqXF5e/zJbj23l7bZv07RiU7MjXZersxPtwoJoFxbE63c2YP3eVBbFJPFbfDJz/kqkrKcrd4RXoFdkJVrWCpApmIXdK0pR36SUqq+13mnzNKJQcUfSmbR8Nz0jgrmrcWWz49iH1W/ByQPwfwvA1dPsNAKY8tcUluxfwjON/7+9+w6PqkrcOP496SEk9N6btBCKgKKIsogLiiCu0kRAkLKWRV3xZwMBddfepUSURQVpNpQiirgqIKAQElqkQ+g1QCD9/P6YkY2ICiGTm7nzfp4nT2Ymw8x7E5L31nOGc0PtG5yOc8HCvGfJt69fnqe7x/L9pkN8nriXeUn7mPljCqWKhdIptiJd4ipzWa3ShKjgpQg6n1JvC/Q3xmzDc0zdAFaXtBWO05k5DJ++mtJRYTzdPVajxgHsSYBlb0CLflCrndNpBJiZPJO3175Nj0t6MCh2kNNxLlp4SDAdGlagQ8MKpGfl8O3PB/k8cS+fJuzhgxW7KFs8jE6xFbmhSWVa1yqtw2FSZJxPqWsAbQc9M38DWw6m8f6gyzTGNUBOFsy5B6LKQcexTqcR4L+7/svTy5+mXdV2PHLZI65b8YwIDea6xhW5rnFF0rNy+Cb5AJ8l7uXDn3bz/g87KRcdzvWxFbkhrjIta5QiSAUvDjqv+dQLI4j81uLkA0xZtoNBbWvRtl5Zp+MUDcvegH1J0OM9iCzldJqAt+7QOkZ8O4IGpRvwfLvnC+VadCdFhAbTKbYSnWIrcSozm683HmBu4l6mr9zFlGU7qBATzvVNKtElrhLNq6ngpfC5+zfwT/ScuAyAGUPbOJzktw6fzOCh2YnUrxDNiL/WdzpO0XB4C3zzDDToAo26Op0m4KWcSOGuRXdROqK0Y9eiO6lYWAhd4irTJa4yaRnZfLVhP3MT9zJ1+U4mL9lO5RIRXN+kEjfEVaJZtZKu24MhRVNAl3pRZa3lkY+SSD2VxbsDW2tQDPAMBfvZcAgOg+tfcDpNwEvNSOWuRXeRnZvNuA7jKBsZ2HuSosJD6NasCt2aVeFEetaZgp+ybDuTvt9GlZKRdImrRJe4ysRWiVHBi8/4tNSNMZ2AV4FgYJK19pnfed4twCyglbX2R19m8gezfkxh4fr9PHZ9QxpWinE6TtGw+j3Y/h10eQViKjmdJqBl5mQyfPFwUk6kEN8xntolazsdqUiJjgile/OqdG9eldTTWXy5fj9zE/fw9vfbmPjtVqqXLsYNcZ5d9I0qqeClYPms1I0xwcCbQEcgBVhpjJlz9qVxxphoPHO0L/dVFn+y43Aaoz9bR5vaZRjUVuOYA3BiHyx8HGpcCS36O50moOXaXB77/jF+2v8Tz7V7jpYVWzodqUgrERnKLZdW5ZZLq3LsVCYL1+3ns8Q9xH+7lfHfbKFW2Si6xHl20devEK2Cl4vmyy311sBma+1WAGPMdKAbcPb17k8CzwEP+jCLX8jOyeW+GQmEBBle7NFUJ9n8Yv5DkJXumYEtSNcGO+mVVa+wYPsC7r/0fjrX6ux0HL9SslgYPVpVo0erahxJy+SLdfv4PHEPby7ezOtfb6Zu+eLc4D3Jrl6FaKfjip/yZalXAXbluZ8CXJb3CcaY5kA1a+3nxpiAL/U3F29h9c5jvN67OZVLakAVADZ8Dus/hb+MhLL1nE4T0LI4xuS1k+lZvyd3NL7D6Th+rXRUGL1bV6d36+ocOpnB/LX7mJu4h9e+3sSrizZRv0I0N3i34OuUK+50XPEjviz1c21m2jNfNCYIeBkY8KcvZMwQYAhA9erVCyhe0bJ651Fe+3oT3ZtX4camlZ2OUzSkp8K8B6F8Y7hyuNNpApa1liyOkWUOcE3Va3i49cPaTVyAyhYP5/bLa3D75TU4cCLdM4tc4l5e/upnXvryZxpWivHsom9SiZplo5yOK0WcL0s9BaiW535VYE+e+9FALPCN9w9ERWCOMabr2SfLWWvjgXiAli1bWlwmLSOb+2ckUDEmgjHdGjsdp+j48gk4uR96TYXgUKfTBKTdJ3czeulosoIOEGSL8Wy7Z11/LbqTykdH0P+KmvS/oib7UtOZl7SXuUl7ef6LZJ7/IpnYKjHc0KQyXeIqaQ4IOSdf/nauBOoZY2oBu4FeQJ9fvmitTQXOXAdjjPkGeDAQz35/au56dhw5xfTBlxMTofICPFvpP02GNvdAlUudThNwcm0uM5Nn8vJPLwMQmlueEEoE3LXoTqpYIoKBbWsxsG0tdh87zfykvXyeuJdnF2zk2QUbaVq1BEfSMokMC2b51sNEhYdQPDyEqPAQoiNCCA8J0h6VAOSzUrfWZhtj7gG+wHNJ2zvW2nXGmLHAj9baOb56b3+ycN0+Plixi79fU4fLapdxOk6RMNjuZezhnVQqWR3aP+p0nICz6/gunlj2BCv3raRNpTaMvmI03Wfd63SsgFalZCR3XlWbO6+qza4jp85swe86ehqAnvE//ObfBAcZosKCKR4eQvGIkDOl/0vxez4HUzw8lOLhwUT9skLg/Xzm+REhFAsN1om7fsKn+9GstfOAeWc9Nup3nnuNL7MURQdOpPPwR0k0rhzD/dde4nScIqNr6hEqZWd5rkkP0zHEwpJrc5m2YRqvrX6NYBPMmCvG0L1ud23tFTHVShdj6NV1GHp1HW56cwlZObk8en1DTmZkk5aRzUnvR1pGNmkZOZxI997O9Dy+LzX9V8/LPc8DmlFhnuIvHuFdIQj7316BKO9KQfGwX69A/HpFIvjMY5rC1nd0cMwh1lr+b3YiaRnZvNqrGWEh+k+OtfDdC9yYepTvo6JpW7eD04kCxvbU7YxaOorVB1ZzVZWrGNVmFBWjKjodS/5EeEgQ4SFBXFk3fyP6WWtJz8o95wrBr2/neD6nZ3My85cVhmxSjp4iLdOz8nAyI5vM7Nzzzn2u0j+zkhD26z0FntvBRIXlWanwfl2HGX5Npe6Q95fvZHHyQcZ0bUzd8romlZwsmPsArHqXZVHRTClbnrZOZwoAObk5vLf+Pd5IeIOw4DCebvs0N9a+UX8kA4QxhsiwYCLDgikXHX7Rr5eZnXtmhSAt07sS4N1jkJaRzYmM/60QnL0CcehkJtsPnzrz2KnMnPN6z5Agc9bhhN/bU+BZMSjuXWk481ielQQ3HGZQqTtg84GTPD13PVdfUo5+bWo4Hcd56cdh1gDYsgjajSB+x8egUvG5Lce2MHLJSJIOJdG+WntGXj6ScsXKOR1L/FhYSBBhIWGUirr4aaJzcq13L4Dnw3MYIed3DjPkXWHwHHLYm+cwQ9p5HmYwBu9eguB8n4OQkZ1DsIN/v1TqhSwzO5f7ZqwmMjSY52+J0xZR6m6Y1gMObICur0OLfvCfT5xO5WrZudlMXjuZ8WvGExUaxXPtnqNTzU76vyhFSnCQISYitECuCLLWcjor59d7DfKca3DmtvdQw8mMrF+tQBxJO/WrlYmsnD9eQ3ByY1+lXsheXfQza3cfZ0LfSykfE+F0HGftWwtTb4WME3DbLNAxdJ9LPpLMyCUj2XBkA9fVuI5HL3uUMpG66kLczRhDsbAQioWFeEZIuUgZ2TlnVg5OnrXX4MWFP3vOD3KISr0Qrdh2hHHfbKFny2p0ig3wk5A2L4KZ/SE8GgbOh4pNnE7kalk5WUxKmkR8UjwxYTG8dM1LdKzR0elYIn4pPCSY8JBgSp/jMMN7y3Y4kOh/VOqF5Hh6FvfPSKB66WKMurGR03Gcteo9z9zo5RtCn5lQoorTiVxt/eH1jFwykp+P/sz1ta7n4dYPUyqilNOxRMQHVOqFZPScdew7ns6sYW2ICg/Qb7u18PVT8N0LULs99HgXIjRfvK9k5mQyYc0E3ln7DqUjSvNa+9doX72907FExIcCtF0K19zEvXy0ajfDO9SjRfUA3ULKzoQ590DiDGje1zOwjMZz95mkg0mMWjqKzcc207VOVx5q9RAlwks4HUtEfEyl7mP7UtN59OMkmlUryT1/qet0HGecPgozboft30H7x6Hdg7pkzUfSs9MZlzCOKeunUC6yHOM6jOOqqlc5HUtEColK3Ydycy0PzlpDZnYuL/dsFphDIx7d4TnD/chW6B4PTXs6nci1Eg4kMHLJSLYf387f6v2Nf7b8J9FhGthIJJCo1H1o8tLtfL/5EP++uQm1AnEe5N2rYFpPyMmA2z+GWtpi9IXT2ad5bdVrTN0wlUpRlYjvGE+bym2cjiUiDlCp+8jGfcd5dsFGrm1YgV6tqv35P3Cb5AUw+w4oVhb6fwblGzidyJVW7lvJE0ufYNeJXfSs35P7L72fqNAAXIEUEUCl7hPpWTncNz2BmIhQnv1bk8AbqWvFWzD/IagY57lkLbqC04lc51TWKV7+6WWmJ0+navGqvPPXd2hVsZXTsUTEYSp1H3hxYTIb951g8oBWlCl+8ZMk+I3cXPhqFCx9HS7pDLe8ralTfeCHvT8weulo9pzcQ9+Gfbm3+b0UCy3mdCwRKQJU6gVsyeZDvPXdNm6/vAbtG5R3Ok7hyToNHw+F9Z9Cq8HQ+VkICnY6lauczDzJiz+9yOyfZ1MzpiZTOk+hefnmTscSkSJEpV6AUk9l8c+Za6hdLopHr2/odJzCk3YYpveGXcvhuqegzT26ZK2Afb/7e0YvHc3B0wcZ0HgAdze7m4iQAJ87QER+Q6VeQKy1PPpJEodOZvBxvyuJDAuQrdTDW2DqLZ7Z1m6dAo1vcjqRq6RmpPL8yuf5dMun1ClRh5eueYm4cnFOxxKRIkqlXkA+SdjN3MS9jPhrfZpUDZCRu3Yuhw96eW73/wyqX+ZsHpf5Ztc3jF02liPpRxjcZDDDmg4jLPji56kWEfdSqReAlKOnGPXJOlrVLMWwq+s4HadwrP8UPhzsmYzlttlQJkCWuxAcSz/GMyufYe7WudQrVY/XO7xO4zKNnY4lIn5ApX6RcnItD8xcgwVe6tGM4CCXH0u2Fpa9CQsfh6qtoPcHEFXW6VSu8dWOr3jqh6dIzUjl703/zuAmgwnVGPkicp5U6hcp/tutrNh2hBdvbUq10i6/rCg3BxY8DCvioWFXuDkeQiOdTuUKR9KP8K/l/+KL7V/QsHRDJnacSP3S9Z2OJSJ+RqV+EdbuTuWlL5O5oUklbm7h8jnBM9PgwzsheZ7n7PaOT0JQAI5lX8CstXyx/Qv+tfxfnMw6yb3N7+WO2DsIDdLWuYhcOJV6Pp3OzGH49NWUjgrj6e6x7h417sR++KAn7F0D178ArQc7ncgVDp0+xFM/PMWinYuILRPLk1c+Sd1SATqTn4gUCJV6Pj0zfwNbDqbx/qDLKFnMxWckH0z2XLKWdgh6TYP6nZ1O5PestXy+9XOeWfEM6dnp3H/p/fRr1I+QIP06isjF0V+RfFicfIApy3YwqG0t2tZz8Uli27+H6X0gOBwGzIUqLZxO5Pf2p+3nyR+e5L8p/6VpuaaMvXIstUvUdjqWiLiESv0CHT6ZwUOzE6lfIZoRf3XxiUyJM+GTu6B0bbhtFpSq4XQiv2at5ZPNn/D8yufJys1iRMsR3NbwNoI1lK6IFCCV+gWw1vLIR0mknsri3YGtiQh14R9ka+G7F+Drp6BGW+j1PkSWcjqVX9t7ci9jlo1hyZ4lXFrhUsZeMZbqMdWdjiUiLqRSvwCzfkxh4fr9PH5DQxpWinE6TsHLyYK5D8Cqd6FJD+j2BoQE0CxzBcxay6yfZ/HSTy+Ra3N5pPUj9GrQiyCjqwZExDdU6udpx+E0Rn+2jivqlGHglbWcjlPw0o/DrAGwZRG0GwHtH9OkLBch5UQKo5eOZvm+5VxW8TJGXzGaqtFVnY4lIi6nUj8P2Tm53DcjgZAgw4s9mhLktlHjju+BqbfCgQ3Q9XVo0c/pRH4r1+YyfeN0Xln1CkEmiFFtRnFLvVvcfcmjiBQZAV3qow6P8N76/g+f9+biLazeeYzXezenUgmXjaC2by1M6+HZUr9tFtTt4HQiv7Xj+A5GLRnFqgOruLLylTzR5gkqFa/kdCwRCSABXernY/XOo7z29Sa6N6/CjU0rOx2nYG1eBDP7Q3g0DJwPFZs4ncgv5eTm8P6G93lj9RuEBoUy9oqx3FT3Jm2di0ihU6n/gbSMbO6fkUDFmAjGdHPZLFmr3oPP74NyDaDPTM9sa3LBtqZuZdSSUaw5uIarq17NyMtHUiGqgtOxRCRAqdT/wFNz17PjyCmmD76cmAiXjMVtLSx+Gr59Hmq3hx7vQoQLz+T3sezcbKasm8K4hHFEhETwr7b/okvtLto6FxFHqdR/x8J1+/hgxS7+fk0dLqtdxuk4BSM7E+bcA4kzoHlf6PIKaFrPC7bp6CZGLRnF2sNr6VC9A49f/jhlI108sqCI+A2V+jkcOJHOwx8l0bhyDPdfe4nTcQrG6aMw43bY/h20fxzaPahL1i5QVm4W7yS9w4TECUSHRvP81c/z1xp/1da5iBQZKvWzWGv5v9mJpGVk82qvZoSFuGCgkKM7PJesHdkK3eOhaU+nE/mdjUc2MnLJSDYe2Uinmp145LJHKB1R2ulYIiK/olI/y/vLd7I4+SBjuzWmbvlop+NcvN2rYFpPyMmA2z+GWlc5ncivZOVkEZ8Uz6TESZQIL8Er17xChxq67E9EiiaVeh6bD5zk6bnrufqSctx+uQsmMEleALPvgGJlof9nUL6B04n8yrrD6xi5ZCSbjm6iS+0u/F+r/6NkREmnYzmikRuHRRZxIZW6V2Z2LvfNWE1kaDDP3xLn/8dJV7wF8x+CinGeS9aidZnV+crIyWDCmglMXjuZMhFleOMvb3B1taudjiUi8qdU6l6vLvqZtbuPM6HvpZSPiXA6Tv7l5sJXo2Dp63BJZ7jlbQiLcjqV30g8mMjIJSPZmrqVm+rexIhWI4gJ01aqiPgHlTqwYtsRxn2zhZ4tq9EptqLTcfIvKx0+HgrrP4FWg6Hzs6D5us9LenY6b6x+g/c2vEf5YuUZf+142lZp63QsEZELEvClnmbDuX9GAtVLF2PUjY2cjpN/aYdhem/YtRyuewra3KNL1s7Tqv2rGLV0FDuO7+DWS27lgUsfoHhYcadjiYhcsIAv9YnpHdmXk86sYW2ICvfTb8fhLZ5L1lJT4NYp0PgmpxP5hVNZp3ht9WtM2zCNysUr89Z1b3F5pcudjiUikm9+2mIF47usBizKasLwDnVpUb2U03HyZ9cK+KCXZ/jX/p9B9cucTuQXVu5byaglo0g5mULvBr25r8V9FAst5nQsEZGLEtClXiPoINeFJnDPXzo7HSV/1n8KHw2BmMpw22woU8fpREVeWlYaL//0MjOSZ1AtuhqT/zqZlhVbOh1LRKRABHSpVw8+zPDI+YQGP+Z0lAtjLSx7ExY+DlVbQe/pEOWS8el9aOmepYxZOoa9aXu5vdHt3Nv8XiJDIp2OJSJSYAK61P1Sbg4seBhWxEPDrnBzPISqmP7IicwTvPjji3y46UNqxtTk3c7v0qx8M6djiYgUOJW6P8lMgw/vhOR5nrPbOz4JQS4Ym96Hvk35ljHLxnDo9CHuiL2Du5reRUSIH49DICLyB3xa6saYTsCrQDAwyVr7zFlfHwbcDeQAJ4Eh1tr1vszkt07shw96wt41cP0L0Hqw04mKtNSMVJ5b+Rxztsyhbsm6vNr+VWLLxjodS0TEp3xW6saYYOBNoCOQAqw0xsw5q7SnWWsneJ/fFXgJ6OSrTH7rYDJMvQXSDkGvaVDfT0/sKyRf7/yaJ394kqPpRxkSN4ShcUMJCw5zOpaIiM/5cku9NbDZWrsVwBgzHegGnCl1a+3xPM+PAqwP8/in7d/D9D4QHA4D5kKVFk4nKrKOph/l3yv+zfxt86lfqj7jOoyjYZmGTsdyhcmdJjsdQUTOgy9LvQqwK8/9FOA3F1EbY+4GHgDCgL/4MI//SZwFn94FpWrBbbOglAtmjvORhdsX8vTypzmeeZy7mt3FnbF3Ehoc6nQsEZFC5ctSP9cYpb/ZErfWvgm8aYzpAzwO9P/NCxkzBBgCUL169QKOWQRZC9+9CF8/CTXaQq/3IdJPB8fxscOnD/P08qf5cseXNCrTiLeue4tLSl3idCwREUf4stRTgGp57lcF9vzB86cD48/1BWttPBAP0LJlS3fvos/JgrkPwKp3oUkP6PYGhIQ7narIsdYyf9t8/r3i36RlpTG8xXAGNB5ASJAu6BCRwOXLv4ArgXrGmFrAbqAX0CfvE4wx9ay1m7x3bwA2EcgyTsDM/rBlEbQbAe0f06Qs53Dw1EGe/OFJFu9aTFzZOMZeOZY6JTWanoiIz0rdWpttjLkH+ALPJW3vWGvXGWPGAj9aa+cA9xhjrgWygKOcY9d7wDi+B6b2gAProevr0KKf04mKHGstc7bM4dmVz5KZk8mDLR+kb8O+BGt6WRERwMfXqVtr5wHzznpsVJ7bw335/n5j31qY1gPSj3tOiKvbwelERc6+tH2MXTaW73Z/R4vyLRhzxRhqlqjpdCwRkSJFByCdtnmRZ5d7eDQMnA8VmzidqEix1vLRpo944ccXyLE5PNz6YXo36E2Q0Uh6IiJnU6k7adV78Pl9UK4B9JkJJao4nahI2XNyD6OXjmbZ3mW0qtiKMW3GUC2m2p//QxGRAKVSd4K1sPhp+PZ5qN0eerwLETFOpyoyLJaD5NL90+4APH7Z49xa/1ZtnYuI/AmVemHLzoQ590DiDGjeF7q8Ahok5Yyf9v9EMlmcNJY25Vox+orRVC5e2elYIiJ+QaVemE4fgxl9Yft30P5xaPegLlnzWrV/FeMSxrF833JCgRo2hIkdJ2L0/REROW8BXepjyp4EYGZhvNmxnTD1Vji8BbrHQ9OehfGuRd7qA6sZlzCOH/b+QJmIMoxoOYJFK18lCKNCFxG5QAFd6oVmz2qY1hOy0+H2j6HWVU4nclzCgQTGJYxj2d5llI4ozYMtH6RH/R5EhkSyeOVrTscTEfFLKnVfS14As++AYmWh3xwo38DpRI5KOJDA+DXjWbpn6W/KXERELo5K3ZdWToJ5I6BinOeStegKTidyzJqDaxifMJ4le5ZQOqI0/7z0n/So34NiocWcjiYi4hoqdV/IzYWvRsHS1+GSznDL2xAW5XQqRyQeTGTcmnEs2b2EUuGleODSB+hZv6fKXETEB1TqBS0rHT4eCus/gVaDofOzEIBjkycdTGLcmnF8v/t7SoaX5L4W99G7QW+VuYiID6nUC1LaYZjeG3Yth+uegjb3BNwla2sPrWX8mvF8m/ItJcNLMrzFcPo06HNhZa6hckVE8kWlXlAOb/FcspaaArdOgcY3OZ2oUK07tI7xa8bz35T/UiK8BMNbDKd3g95EhQbmYQcRESeo1AvCrhXwQS/P8K/9P4PqlzmdqNCsO7yOCQkT+CblG2LCYvhH83/Qp2EflbmIiANU6hdr/afw0RCIqQy3zYYydZxOVCg2HN7AuDXj+GaXp8zvbX4vfRr0oXhYcaejiYgELJV6flkLP4yDLx6Dqq2g93SIKuN0Kp/beGQj4xLGsXjXYqLDorm72d3c1vA2osOinY4mIhLwVOr5kZsDCx6BFROhYVe4OR5C3T14SvKRZMYljOPrXV8THRrNXc3uom/DvipzEZEiRKV+oTLT4MM7IXme5+z2jk9CkHunBE0+ksyENRP4audXnjJvehe3NbqNmDBNFSsiUtSo1C/EyQOeMdz3JsD1L0DrwU4n8pnkI8lMTJzIlzu+pHhocf7e9O/0bdRXZS4iUoSp1M/XwWSYegukHYJe06B+Z6cT+cTPR39mwpoJZ8p8WNNh9G3YlxLhJZyOJiIif0Klfj62fw/T+0BwOAyYC1VaOJ2owG0+upnxa8azcMdCokKjGBo3lNsb3a4yFxHxIyr1P5M4Cz69C0rVgttmQakaTicqUJuPbmZC4gQWbl9IZEgkg5sMpn/j/ipzERE/pFL/PdbCdy/C109CjbbQ632ILOV0qgKz5dgWJqyZwBfbvyAyJJI7m9xJv0b9KBlR0uloIiKSTyr1c8nJgrkPwKp3oUkP6PYGhIQ7napAbD22lQmJE1iwbQERIREMajKI/o36q8xF/NCMoW2cjiBncfpnolI/W8YJmNkftiyCdiOg/WOumJRla+pWJq6ZyPxt84kIiWBg7ED6N+5PqQj37H0QEQl0KvW8ju+BqT3gwHro+jq06Od0oou2LXUbExM9ZR4eHM4dsXfQv3F/SkeUdjqaiIgUMJX6L/athWk9IP2454S4uh2cTnRRtqduZ2LiROZtm0d4cDj9G/VnQOwAlbmIiIup1AG2fA0z+kF4NAyc79fzee84voOJayYyd9tcwoLC6NeoHwMaD6BMpPvHpRcRCXQBX+rXnDrhmQe9XAPoMxNKVHE6Ur7sPL6TiYkT+Xzr54QFhXF7w9sZEDuAspFlnY4mIiKFJKBLvUPacYamHoba7aHHuxDhf0Og7jq+60yZhwSF0LdhX+6IvUNlLiISgAK61H+MKEb5nGy63zYLgkOdjnNBdh3fRXxSPJ9t+YyQoBD6NOzDwNiBKnMRkQAW0KWeGhzCBzGl6e5Hhb7rxC7eSnyLOVvmEGyC6d2gNwNjB1KuWDmno4mIiMMCutT9ScqJFN5Keos5m+cQZILo1aAXA2MHUr5YeaejiYhIEaFSL+J2jwLAVQAADeNJREFUn9zNW4lv8enmTwkyQfSo34NBTQapzEVE5DdU6kXUnpN7iE+M59PNn2KM4db6tzIodhAVoio4HU1ERIoolXoRs/fkXt5KeouPN3+MwXDLJbcwqMkgKkZVdDqaiIgUcSr1ImLvyb1MSprER5s/AuBv9f7GnU3uVJmLiMh5U6k7bF/aPiYlTeLDTR8CcHPdm7mzyZ1UKl7J4WQiIuJvVOoO+aXMP9r0ERZL97rdGdxksMpcRETyTaVeyPan7T+zZW6t5aZ6NzG4yWAqF6/sdDQREfFzKvVCcuDUASYlTWL2z7Ox1tKtbjcGxw2mSnH/HGteRESKHpW6jx04dYB31r7DrORZ5NpcutXtxp1N7qRqdFWno4mIiMuo1H3k4KmDnjL/eRbZudl0rdOVwXGDqRZdzeloIiLiUir1Anbo9CHeTnr7TJnfWOdGhsQNUZmLiIjPBXSpR4UV3OIfOn2Id9a+w8zkmWTnZtOldheGxA2hekz1AnsPERGRPxLQpV4QDp0+xOS1k5mZPJPM3Ey61O7C0LihKnMRESl0KvV8Onz6MJPXTmZG8owzZT4kbgg1Ymo4HU1ERAKUSv0CHT59mCnrpjA9eToZORlcX+t6hsYNpWaJmk5HExGRAKdSP09H0o/wn3X/YfpGT5l3rtWZoXFDqVWiltPRREREAB+XujGmE/AqEAxMstY+c9bXHwDuBLKBg8BAa+0OX2a6UEfTj/Kfdf/hg40fkJ6d7inzpkOpXaK209FERER+xWelbowJBt4EOgIpwEpjzBxr7fo8T1sNtLTWnjLG/B14Dujpq0wX4mj6Uaasm8K0jdNIz06nU61ODIsbRu2SKnMRESmafLml3hrYbK3dCmCMmQ50A86UurV2cZ7n/wD09WGe83Is/RhT1k9h2oZpnM4+TaeanRjadCh1StZxOpqIiMgf8mWpVwF25bmfAlz2B88fBMz3YZ4/lJqRypR1U5i6YSqns09zXc3rGBY3jLql6joVSURE5IL4stTNOR6z53yiMX2BlsDVv/P1IcAQgOrVC/b672wsr69+nakbppKWlcZ1Na5jWNNh1CtVr0DfR0RExNd8WeopQN6xUasCe85+kjHmWuAx4Gprbca5XshaGw/EA7Rs2fKcKwb5cYQcdpBNQmI8HWt0ZFjTYVxS6pKCenkREZFC5ctSXwnUM8bUAnYDvYA+eZ9gjGkOTAQ6WWsP+DDLOUVgiCaISTfOpH7p+oX99iIiIgUqyFcvbK3NBu4BvgA2ADOtteuMMWONMV29T3seKA7MMsYkGGPm+CrPuRQjiLqEqtBFRMQVfHqdurV2HjDvrMdG5bl9rS/fX0REJJD4bEtdRERECpdKXURExCVU6iIiIi6hUhcREXEJlbqIiIhLqNRFRERcQqUuIiLiEip1ERERl1Cpi4iIuIRKXURExCVU6iIiIi7h07HfRfJjcqfJTkcQEfFL2lIXERFxCZW6iIiIS6jURUREXEKlLiIi4hIqdREREZdQqYuIiLiESl1ERMQlVOoiIiIuoVIXERFxCZW6iIiIS6jURUREXEKlLiIi4hIqdREREZdQqYuIiLiESl1ERMQlVOoiIiIuoVIXERFxCZW6iIiIS4Q4HcBRFZs4nUBERKTAaEtdRETEJVTqIiIiLqFSFxERcQmVuoiIiEuo1EVERFxCpS4iIuISKnURERGXUKmLiIi4hEpdRETEJVTqIiIiLqFSFxERcQmVuoiIiEuo1EVERFxCpS4iIuISKnURERGXUKmLiIi4RIjTAZw0udNkpyOIiIgUGG2pi4iIuIRKXURExCV8WurGmE7GmGRjzGZjzMPn+Ho7Y8wqY0y2MeYWX2YRERFxO5+VujEmGHgT6Aw0AnobYxqd9bSdwABgmq9yiIiIBApfnijXGthsrd0KYIyZDnQD1v/yBGvtdu/Xcn2YQ0REJCD4cvd7FWBXnvsp3sdERETEB3xZ6uYcj9l8vZAxQ4wxPxpjfjx48OBFxhIREXEnX5Z6ClAtz/2qwJ78vJC1Nt5a29Ja27JcuXIFEk5ERMRtfFnqK4F6xphaxpgwoBcwx4fvJyIiEtB8VurW2mzgHuALYAMw01q7zhgz1hjTFcAY08oYkwLcCkw0xqzzVR4RERG38+kwsdbaecC8sx4blef2Sjy75UVEROQiaUQ5ERERl1Cpi4iIuIRKXURExCVU6iIiIi6hUhcREXEJlbqIiIhLqNRFRERcQqUuIiLiEsbafM2x4hhjzEFgRwG+ZFngUAG+npO0LEWTW5bFLcsBWpaiyC3LAQW/LDWstec18YnflXpBM8b8aK1t6XSOgqBlKZrcsixuWQ7QshRFblkOcHZZtPtdRETEJVTqIiIiLqFSh3inAxQgLUvR5JZlcctygJalKHLLcoCDyxLwx9RFRETcQlvqIiIiLhGwpW6MqWaMWWyM2WCMWWeMGe50pvwyxkQYY1YYY9Z4l2WM05kuhjEm2Biz2hjzudNZLoYxZrsxJskYk2CM+dHpPBfDGFPSGDPbGLPR+zvTxulM+WGMqe/9efzycdwYc5/TufLDGHO/9/d9rTHmA2NMhNOZ8ssYM9y7HOv87edhjHnHGHPAGLM2z2OljTFfGmM2eT+XKqw8AVvqQDbwT2ttQ+By4G5jTCOHM+VXBvAXa21ToBnQyRhzucOZLsZwYIPTIQpIe2ttMxdcqvMqsMBa2wBoip/+fKy1yd6fRzPgUuAU8LHDsS6YMaYK8A+gpbU2FggGejmbKn+MMbHAYKA1nv9bXYwx9ZxNdUH+A3Q667GHgUXW2nrAIu/9QhGwpW6t3WutXeW9fQLPH6kqzqbKH+tx0ns31PvhlydLGGOqAjcAk5zOIh7GmBigHfA2gLU201p7zNlUBaIDsMVaW5CDWRWmECDSGBMCFAP2OJwnvxoCP1hrT1lrs4H/At0dznTerLXfAkfOergbMMV7ewpwU2HlCdhSz8sYUxNoDix3Nkn+eXdZJwAHgC+ttf66LK8ADwG5TgcpABZYaIz5yRgzxOkwF6E2cBCY7D0sMskYE+V0qALQC/jA6RD5Ya3dDbwA7AT2AqnW2oXOpsq3tUA7Y0wZY0wx4HqgmsOZLlYFa+1e8GxAAuUL640DvtSNMcWBD4H7rLXHnc6TX9baHO8uxapAa+8uLb9ijOkCHLDW/uR0lgJypbW2BdAZz+Gddk4HyqcQoAUw3lrbHEijEHcn+oIxJgzoCsxyOkt+eI/RdgNqAZWBKGNMX2dT5Y+1dgPwLPAlsABYg+fwqORDQJe6MSYUT6FPtdZ+5HSeguDdLfoNvz3G4w+uBLoaY7YD04G/GGPedzZS/llr93g/H8Bz3La1s4nyLQVIybP3ZzaekvdnnYFV1tr9TgfJp2uBbdbag9baLOAj4AqHM+WbtfZta20La207PLuyNzmd6SLtN8ZUAvB+PlBYbxywpW6MMXiOEW6w1r7kdJ6LYYwpZ4wp6b0diecXfqOzqS6ctfYRa21Va21NPLtGv7bW+uXWhzEmyhgT/ctt4Do8uxn9jrV2H7DLGFPf+1AHYL2DkQpCb/x017vXTuByY0wx79+yDvjpyYsAxpjy3s/VgZvx758NwBygv/d2f+DTwnrjkMJ6oyLoSuB2IMl7LBrgUWvtPAcz5VclYIoxJhjPitpMa61fXw7mAhWAjz1/bwkBpllrFzgb6aLcC0z17rbeCtzhcJ588x637QgMdTpLfllrlxtjZgOr8OyqXo1/j8j2oTGmDJAF3G2tPep0oPNljPkAuAYoa4xJAZ4AngFmGmMG4VkBu7XQ8mhEOREREXcI2N3vIiIibqNSFxERcQmVuoiIiEuo1EVERFxCpS4iIuISKnURKRTGmG+MMf4+sY1IkaZSFxERcQmVuoifM8Z84p00Zt0vE8d4J/j5j3eO6iRjzP3ex/9hjFlvjEk0xkz3PjbaGPNgntdba4yp6f3Y6J3AZa0xZqox5lpjzBLvPNGtvc+P8s4pvdI74Us37+ORxpjp3veaAUT+Tv7txphnjTErvB91ffwtE3GtQB5RTsQtBlprj3iHCF5pjPkQqAlU8c61zS/DCOOZiKWWtTYjz2N/pC6e0bCGACuBPkBbPJOhPIpnSsnH8AzpO9D7miuMMV/hGbHtlLU2zhgTh2f0s99z3Frb2hjTD89MfV0u5BsgIh7aUhfxf/8wxqwBfsAzZWU9PEO51jbGvG6M6QT8MgNhIp7hXvtyfjNhbbPWJllrc4F1wCLrGYYyCc+KA3jGtX/YO9zyN0AEUB3PHOzvA1hrE73v/Xs+yPO5zXnkEpFzUKmL+DFjzDV4JvBpY61timcM8Ajv2NlN8ZTs3cAk7z+5AXgTuBT4yRgTgqfc8/4tiMhzOyPP7dw893P5354+A/zNWtvM+1HdO50meOaUPx/2d26LyAVQqYv4txLAUWvtKWNMA+ByAGNMWSDIWvshMBJoYYwJAqpZaxcDDwElgeLAdrxTqRpjWuCZo/tCfAHc650tDGNMc+/j3wK3eR+LBeL+4DV65vm87ALfX0S8dExdxL8tAIYZYxKBZDy74AGqAJO9RQ7wCBAMvG+MKYFn6/pla+0x7zH4ft7d5yuBny8ww5N4joMneot9O55j4uO9GRKBBGDFH7xGuDFmOZ4Njd4X+P4i4qVZ2kTEUcaY7UBLa+0hp7OI+DvtfhcREXEJbamLiIi4hLbURUREXEKlLiIi4hIqdREREZdQqYuIiLiESl1ERMQlVOoiIiIu8f+jAkVZ2jXMqAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f13c0a50048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,8))\n",
    "plt.title('f1 scores for different p')\n",
    "plt.xlabel('assumed p')\n",
    "plt.ylabel('mean f1')\n",
    "handles = []\n",
    "labels = []\n",
    "for p in max_ps:\n",
    "    plt.errorbar(x=f1_scores.loc[f1_scores['true_p']==p, 'p'],\n",
    "                 y=f1_scores.loc[f1_scores['true_p']==p, 'mean_f1'],\n",
    "                 yerr=f1_scores.loc[f1_scores['true_p']==p, 'std_f1'],\n",
    "                 label='true p='+str(p))\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f14138f1e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,8))\n",
    "plt.title('runtime for different p')\n",
    "plt.xlabel('assumed p')\n",
    "plt.ylabel('runtime in s')\n",
    "handles = []\n",
    "labels = []\n",
    "for p in max_ps:\n",
    "    plt.errorbar(x=runtimes.loc[runtimes['true_p']==p, 'p'],\n",
    "                 y=runtimes.loc[runtimes['true_p']==p, 'mean_time'],\n",
    "                 yerr=runtimes.loc[runtimes['true_p']==p, 'std_time'],\n",
    "                 label='true p='+str(p))\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}