plots/plots_e2e.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import re\n",
    "import time\n",
    "import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# lhidde: Fix for my machine to machine times new roman not bold.\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# if 'roman' in matplotlib.font_manager.weight_dict:\n",
    "#     del matplotlib.font_manager.weight_dict['roman']\n",
    "#     matplotlib.font_manager._rebuild()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "import plots.e2e_parse_benchlog as parser\n",
    "import os\n",
    "\n",
    "# e2e_base_dir = \"./../rofl_polybox_data/\" # for nicolas\n",
    "\n",
    "# e2e_base_dir = \"/Users/hidde/PhD/fl/fl-ansible/experiment_results/e2e/\" # for hidde\n",
    "e2e_base_dir = \"./data/endtoend/\" # for hidde\n",
    "\n",
    "configs = {\n",
    "    os.path.join(e2e_base_dir, \"shakespeare_e2e_1679952478/0\"): {\n",
    "        \"dataset\": \"Shakespeare\",\n",
    "        \"type\": \"linf_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 53.46\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"shakespeare_e2e_1679953917/0\"): {\n",
    "        \"dataset\": \"Shakespeare\",\n",
    "        \"type\": \"plain\",\n",
    "        \"version\": \"None\",\n",
    "        \"bandwidth\": 6.3\n",
    "    },\n",
    "    # CIFAR Large\n",
    "    os.path.join(e2e_base_dir, \"cifar_large_1680595169/0\"): {\n",
    "        \"dataset\": \"CIFAR Large\",\n",
    "        \"type\": \"l2_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 70.71\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_large_1679957384/0\"): { # Done\n",
    "        \"dataset\": \"CIFAR Large\",\n",
    "        \"type\": \"linf_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 143.15\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_large_1679957384/1\"): {\n",
    "        \"dataset\": \"CIFAR Large\",\n",
    "        \"type\": \"plain\",\n",
    "        \"version\": \"None\",\n",
    "        \"bandwidth\": 16.89\n",
    "    },\n",
    "\n",
    "    os.path.join(e2e_base_dir, \"cifar_small_1679945008/0\"): {\n",
    "        \"dataset\": \"CIFAR Small\",\n",
    "        \"type\": \"l2_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 13.51\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_small_1679945008/3\"): {\n",
    "        \"dataset\": \"CIFAR Small\",\n",
    "        \"type\": \"l2\",\n",
    "        \"version\": \"unoptimized\",\n",
    "        \"bandwidth\": 68.28\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_small_1679945008/1\"): {\n",
    "        \"dataset\": \"CIFAR Small\",\n",
    "        \"type\": \"linf_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 20.58\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_small_1679945008/4\"): {\n",
    "        \"dataset\": \"CIFAR Small\",\n",
    "        \"type\": \"linf\",\n",
    "        \"version\": \"unoptimized\",\n",
    "        \"bandwidth\": 20.66\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"cifar_small_1679945008/2\"): {\n",
    "        \"dataset\": \"CIFAR Small\",\n",
    "        \"type\": \"plain\",\n",
    "        \"version\": None,\n",
    "        \"bandwidth\": 2.45\n",
    "    },\n",
    "\n",
    "    # CIFAR Small L2\n",
    "    # MNIST Linf, L2\n",
    "    os.path.join(e2e_base_dir, \"mnist_e2e_1679941149_l2rsl/0\"): {\n",
    "        \"dataset\": \"MNIST\",\n",
    "        \"type\": \"l2_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 9.33\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"mnist_e2e_1679945880/0\"): {\n",
    "        \"dataset\": \"MNIST\",\n",
    "        \"type\": \"l2\",\n",
    "        \"version\": \"unoptimized\",\n",
    "        \"bandwidth\": 34.2\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"mnist_e2e_1679945880/1\"): {\n",
    "        \"dataset\": \"MNIST\",\n",
    "        \"type\": \"linf_opt\",\n",
    "        \"version\": \"optimized\",\n",
    "        \"bandwidth\": 10.55\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"mnist_e2e_1679945880/2\"): {\n",
    "        \"dataset\": \"MNIST\",\n",
    "        \"type\": \"linf\",\n",
    "        \"version\": \"unoptimized\",\n",
    "        \"bandwidth\": 10.65\n",
    "    },\n",
    "    os.path.join(e2e_base_dir, \"mnist_e2e_1679945880/3\"): {\n",
    "        \"dataset\": \"MNIST\",\n",
    "        \"type\": \"plain\",\n",
    "        \"version\": None,\n",
    "        \"bandwidth\": 1.3\n",
    "    }\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "labels = [f\"{config['dataset']} {config['type']} {config['version']}\" for _, config in configs.items()]\n",
    "\n",
    "x = np.arange(len(labels))  # the label locations\n",
    "width = 0.35  # the width of the bars\n",
    "\n",
    "data_server, observer_data, client_data = pd.DataFrame(), pd.DataFrame(), pd.DataFrame()\n",
    "for i, (path, config) in enumerate(configs.items()):\n",
    "    s, o, clients = parser.load_benchlog(path)\n",
    "\n",
    "    for v in [s, o, clients]:\n",
    "        idf = f\"{config['dataset']} {config['type']} {config['version']}\"\n",
    "        v[\"idf\"] = idf\n",
    "        for key in ['dataset', 'type', 'version', 'bandwidth']:\n",
    "            v[key] = config[key]\n",
    "\n",
    "    o[\"rel_time\"] = o['time'].diff().dt.total_seconds()\n",
    "    o[\"rel_time\"] = o[\"rel_time\"][1:]\n",
    "\n",
    "    data_server = data_server.append(s)\n",
    "    observer_data = observer_data.append(o)\n",
    "    client_data = client_data.append(clients)\n",
    "\n",
    "# display(observer_data)\n",
    "pd.set_option('display.max_rows', 500)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=[10, 5])\n",
    "\n",
    "# x = observer_data.groupby([\"idf\"])[\"time\"].diff().dt.total_seconds()\n",
    "\n",
    "# observer_data[\"rel_time\"] = observer_data['time'].diff().dt.total_seconds()\n",
    "# observer_data[observer_data[\"round\"] == 0][\"rel_time\"] = None\n",
    "\n",
    "# display(observer_data)\n",
    "# display(observer_data.groupby([\"idf\"])[\"rel_time\"].mean())\n",
    "\n",
    "y1 = observer_data.groupby([\"idf\"], sort=False)[\"rel_time\"].mean()\n",
    "y2 = data_server.groupby([\"idf\"], sort=False)[\"verification_complete\"].mean() / 1000\n",
    "rects1 = ax.bar(x - width/2, y1, width, label='Observer time')\n",
    "rects2 = ax.bar(x + width/2, y2, width, label='Server time')\n",
    "\n",
    "# display(y1)\n",
    "# display(y2)\n",
    "\n",
    "ax.set_ylabel(\"Round time\")\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(labels)\n",
    "ax.legend()\n",
    "\n",
    "# ax.bar_label(rects1, padding=3)\n",
    "# ax.bar_label(rects2, padding=3)\n",
    "for a, (b, c) in zip(x, zip(y1, y2)):\n",
    "    ax.text(a - width / 1.2, b + 5, round(b), color='black')\n",
    "    ax.text(a + width / 4, c + 5, round(c), color='black')\n",
    "\n",
    "# plt.show()\n",
    "plt.savefig(\"plots_output/e2e_cifar_linf.png\")\n",
    "\n",
    "# data_server.to_csv(\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "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>(verification_complete, sum)</th>\n",
       "      <th>(verification_complete, mean)</th>\n",
       "      <th>(verification_complete, sem)</th>\n",
       "      <th>mean_factor</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset</th>\n",
       "      <th>type</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">CIFAR Large</th>\n",
       "      <th>l2_opt</th>\n",
       "      <td>29598747</td>\n",
       "      <td>1.861556e+05</td>\n",
       "      <td>370.534011</td>\n",
       "      <td>8.56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>16636654</td>\n",
       "      <td>1.046330e+05</td>\n",
       "      <td>477.933341</td>\n",
       "      <td>4.81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>3480557</td>\n",
       "      <td>2.175348e+04</td>\n",
       "      <td>21.081351</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">CIFAR Small</th>\n",
       "      <th>l2</th>\n",
       "      <td>11049661</td>\n",
       "      <td>1.104966e+05</td>\n",
       "      <td>354.313867</td>\n",
       "      <td>59.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>2605994</td>\n",
       "      <td>2.632317e+04</td>\n",
       "      <td>110.232785</td>\n",
       "      <td>14.21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>8619593</td>\n",
       "      <td>8.619593e+04</td>\n",
       "      <td>344.936155</td>\n",
       "      <td>46.54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>2862986</td>\n",
       "      <td>2.891905e+04</td>\n",
       "      <td>68.745964</td>\n",
       "      <td>15.61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>185203</td>\n",
       "      <td>1.852030e+03</td>\n",
       "      <td>7.957265</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">MNIST</th>\n",
       "      <th>l2</th>\n",
       "      <td>7550711</td>\n",
       "      <td>4.719194e+04</td>\n",
       "      <td>157.609389</td>\n",
       "      <td>20.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>2394780</td>\n",
       "      <td>1.506151e+04</td>\n",
       "      <td>39.467406</td>\n",
       "      <td>6.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>6461782</td>\n",
       "      <td>4.038614e+04</td>\n",
       "      <td>126.644688</td>\n",
       "      <td>17.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>1854981</td>\n",
       "      <td>1.166655e+04</td>\n",
       "      <td>18.222016</td>\n",
       "      <td>5.18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>360611</td>\n",
       "      <td>2.253819e+03</td>\n",
       "      <td>5.469483</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Shakespeare</th>\n",
       "      <th>linf_opt</th>\n",
       "      <td>19476679</td>\n",
       "      <td>1.025088e+06</td>\n",
       "      <td>774.362802</td>\n",
       "      <td>1.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>15469128</td>\n",
       "      <td>7.734564e+05</td>\n",
       "      <td>253.505231</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      (verification_complete, sum)  \\\n",
       "dataset     type                                     \n",
       "CIFAR Large l2_opt                        29598747   \n",
       "            linf_opt                      16636654   \n",
       "            plain                          3480557   \n",
       "CIFAR Small l2                            11049661   \n",
       "            l2_opt                         2605994   \n",
       "            linf                           8619593   \n",
       "            linf_opt                       2862986   \n",
       "            plain                           185203   \n",
       "MNIST       l2                             7550711   \n",
       "            l2_opt                         2394780   \n",
       "            linf                           6461782   \n",
       "            linf_opt                       1854981   \n",
       "            plain                           360611   \n",
       "Shakespeare linf_opt                      19476679   \n",
       "            plain                         15469128   \n",
       "\n",
       "                      (verification_complete, mean)  \\\n",
       "dataset     type                                      \n",
       "CIFAR Large l2_opt                     1.861556e+05   \n",
       "            linf_opt                   1.046330e+05   \n",
       "            plain                      2.175348e+04   \n",
       "CIFAR Small l2                         1.104966e+05   \n",
       "            l2_opt                     2.632317e+04   \n",
       "            linf                       8.619593e+04   \n",
       "            linf_opt                   2.891905e+04   \n",
       "            plain                      1.852030e+03   \n",
       "MNIST       l2                         4.719194e+04   \n",
       "            l2_opt                     1.506151e+04   \n",
       "            linf                       4.038614e+04   \n",
       "            linf_opt                   1.166655e+04   \n",
       "            plain                      2.253819e+03   \n",
       "Shakespeare linf_opt                   1.025088e+06   \n",
       "            plain                      7.734564e+05   \n",
       "\n",
       "                      (verification_complete, sem)  mean_factor  \n",
       "dataset     type                                                 \n",
       "CIFAR Large l2_opt                      370.534011         8.56  \n",
       "            linf_opt                    477.933341         4.81  \n",
       "            plain                        21.081351         1.00  \n",
       "CIFAR Small l2                          354.313867        59.66  \n",
       "            l2_opt                      110.232785        14.21  \n",
       "            linf                        344.936155        46.54  \n",
       "            linf_opt                     68.745964        15.61  \n",
       "            plain                         7.957265         1.00  \n",
       "MNIST       l2                          157.609389        20.94  \n",
       "            l2_opt                       39.467406         6.68  \n",
       "            linf                        126.644688        17.92  \n",
       "            linf_opt                     18.222016         5.18  \n",
       "            plain                         5.469483         1.00  \n",
       "Shakespeare linf_opt                    774.362802         1.33  \n",
       "            plain                       253.505231         1.00  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Server benchlor\n",
    "\n",
    "results = data_server.groupby([\"dataset\", \"type\"]).agg({'verification_complete': ['sum', 'mean', 'sem']})\n",
    "results.reset_index()\n",
    "results_per_dataset = results.groupby(\"dataset\")\n",
    "# print(results_per_dataset)\n",
    "\n",
    "in_group = results['verification_complete']['mean']\\\n",
    "    .groupby([\"dataset\"])\\\n",
    "    .transform(lambda x: x / x[x.index.get_level_values('type') == 'plain'][0])\n",
    "# results['fraction'] = results['verification_complete']['mean'] / results[results['type'] == 'plain']['verification_complete']['mean']\n",
    "\n",
    "in_group = in_group.rename('mean_factor')\n",
    "\n",
    "results = results.join(in_group, on=['dataset', 'type'])\n",
    "results['mean_factor'] = results['mean_factor'].round(2)\n",
    "display(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "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>(rel_time, sum)</th>\n",
       "      <th>(rel_time, mean)</th>\n",
       "      <th>(rel_time, sem)</th>\n",
       "      <th>(acc, &lt;lambda&gt;)</th>\n",
       "      <th>(loss, min)</th>\n",
       "      <th>(bandwidth, mean)</th>\n",
       "      <th>mean_factor</th>\n",
       "      <th>bw_factor</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset</th>\n",
       "      <th>type</th>\n",
       "      <th></th>\n",
       "      <th></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=\"3\" valign=\"top\">CIFAR Large</th>\n",
       "      <th>l2_opt</th>\n",
       "      <td>23733.677</td>\n",
       "      <td>149.268409</td>\n",
       "      <td>0.505256</td>\n",
       "      <td>0.816200</td>\n",
       "      <td>0.540763</td>\n",
       "      <td>70.71</td>\n",
       "      <td>6.86</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>10097.524</td>\n",
       "      <td>63.506440</td>\n",
       "      <td>0.906456</td>\n",
       "      <td>0.849900</td>\n",
       "      <td>0.601752</td>\n",
       "      <td>143.15</td>\n",
       "      <td>2.92</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>3458.081</td>\n",
       "      <td>21.748937</td>\n",
       "      <td>0.016318</td>\n",
       "      <td>0.857900</td>\n",
       "      <td>0.551553</td>\n",
       "      <td>16.89</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">CIFAR Small</th>\n",
       "      <th>l2</th>\n",
       "      <td>10936.882</td>\n",
       "      <td>110.473556</td>\n",
       "      <td>0.357023</td>\n",
       "      <td>0.602700</td>\n",
       "      <td>1.121349</td>\n",
       "      <td>68.28</td>\n",
       "      <td>59.71</td>\n",
       "      <td>28.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>1799.007</td>\n",
       "      <td>18.171788</td>\n",
       "      <td>0.101841</td>\n",
       "      <td>0.586100</td>\n",
       "      <td>1.175898</td>\n",
       "      <td>13.51</td>\n",
       "      <td>9.82</td>\n",
       "      <td>6.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>8533.177</td>\n",
       "      <td>86.193707</td>\n",
       "      <td>0.347924</td>\n",
       "      <td>0.603000</td>\n",
       "      <td>1.122020</td>\n",
       "      <td>20.66</td>\n",
       "      <td>46.58</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>1440.460</td>\n",
       "      <td>14.550101</td>\n",
       "      <td>0.104054</td>\n",
       "      <td>0.607200</td>\n",
       "      <td>1.123885</td>\n",
       "      <td>20.58</td>\n",
       "      <td>7.86</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>183.176</td>\n",
       "      <td>1.850263</td>\n",
       "      <td>0.006123</td>\n",
       "      <td>0.607300</td>\n",
       "      <td>1.119677</td>\n",
       "      <td>2.45</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">MNIST</th>\n",
       "      <th>l2</th>\n",
       "      <td>7503.997</td>\n",
       "      <td>47.194950</td>\n",
       "      <td>0.158336</td>\n",
       "      <td>0.986500</td>\n",
       "      <td>0.069298</td>\n",
       "      <td>34.20</td>\n",
       "      <td>20.96</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>1698.451</td>\n",
       "      <td>10.682082</td>\n",
       "      <td>0.057340</td>\n",
       "      <td>0.973400</td>\n",
       "      <td>0.104768</td>\n",
       "      <td>9.33</td>\n",
       "      <td>4.74</td>\n",
       "      <td>7.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>6422.447</td>\n",
       "      <td>40.392748</td>\n",
       "      <td>0.126836</td>\n",
       "      <td>0.986700</td>\n",
       "      <td>0.068483</td>\n",
       "      <td>10.65</td>\n",
       "      <td>17.94</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>1228.299</td>\n",
       "      <td>7.725151</td>\n",
       "      <td>0.067717</td>\n",
       "      <td>0.987100</td>\n",
       "      <td>0.068014</td>\n",
       "      <td>10.55</td>\n",
       "      <td>3.43</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>358.076</td>\n",
       "      <td>2.252050</td>\n",
       "      <td>0.003521</td>\n",
       "      <td>0.987600</td>\n",
       "      <td>0.069459</td>\n",
       "      <td>1.30</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Shakespeare</th>\n",
       "      <th>linf_opt</th>\n",
       "      <td>16532.907</td>\n",
       "      <td>918.494833</td>\n",
       "      <td>1.716796</td>\n",
       "      <td>0.573490</td>\n",
       "      <td>1.405828</td>\n",
       "      <td>53.46</td>\n",
       "      <td>1.19</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>13917.818</td>\n",
       "      <td>773.212111</td>\n",
       "      <td>0.153057</td>\n",
       "      <td>0.572081</td>\n",
       "      <td>1.403638</td>\n",
       "      <td>6.30</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      (rel_time, sum)  (rel_time, mean)  (rel_time, sem)  \\\n",
       "dataset     type                                                           \n",
       "CIFAR Large l2_opt          23733.677        149.268409         0.505256   \n",
       "            linf_opt        10097.524         63.506440         0.906456   \n",
       "            plain            3458.081         21.748937         0.016318   \n",
       "CIFAR Small l2              10936.882        110.473556         0.357023   \n",
       "            l2_opt           1799.007         18.171788         0.101841   \n",
       "            linf             8533.177         86.193707         0.347924   \n",
       "            linf_opt         1440.460         14.550101         0.104054   \n",
       "            plain             183.176          1.850263         0.006123   \n",
       "MNIST       l2               7503.997         47.194950         0.158336   \n",
       "            l2_opt           1698.451         10.682082         0.057340   \n",
       "            linf             6422.447         40.392748         0.126836   \n",
       "            linf_opt         1228.299          7.725151         0.067717   \n",
       "            plain             358.076          2.252050         0.003521   \n",
       "Shakespeare linf_opt        16532.907        918.494833         1.716796   \n",
       "            plain           13917.818        773.212111         0.153057   \n",
       "\n",
       "                      (acc, <lambda>)  (loss, min)  (bandwidth, mean)  \\\n",
       "dataset     type                                                        \n",
       "CIFAR Large l2_opt           0.816200     0.540763              70.71   \n",
       "            linf_opt         0.849900     0.601752             143.15   \n",
       "            plain            0.857900     0.551553              16.89   \n",
       "CIFAR Small l2               0.602700     1.121349              68.28   \n",
       "            l2_opt           0.586100     1.175898              13.51   \n",
       "            linf             0.603000     1.122020              20.66   \n",
       "            linf_opt         0.607200     1.123885              20.58   \n",
       "            plain            0.607300     1.119677               2.45   \n",
       "MNIST       l2               0.986500     0.069298              34.20   \n",
       "            l2_opt           0.973400     0.104768               9.33   \n",
       "            linf             0.986700     0.068483              10.65   \n",
       "            linf_opt         0.987100     0.068014              10.55   \n",
       "            plain            0.987600     0.069459               1.30   \n",
       "Shakespeare linf_opt         0.573490     1.405828              53.46   \n",
       "            plain            0.572081     1.403638               6.30   \n",
       "\n",
       "                      mean_factor  bw_factor  \n",
       "dataset     type                              \n",
       "CIFAR Large l2_opt           6.86        4.0  \n",
       "            linf_opt         2.92        8.0  \n",
       "            plain            1.00        1.0  \n",
       "CIFAR Small l2              59.71       28.0  \n",
       "            l2_opt           9.82        6.0  \n",
       "            linf            46.58        8.0  \n",
       "            linf_opt         7.86        8.0  \n",
       "            plain            1.00        1.0  \n",
       "MNIST       l2              20.96       26.0  \n",
       "            l2_opt           4.74        7.0  \n",
       "            linf            17.94        8.0  \n",
       "            linf_opt         3.43        8.0  \n",
       "            plain            1.00        1.0  \n",
       "Shakespeare linf_opt         1.19        8.0  \n",
       "            plain            1.00        1.0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "observer_results = observer_data.groupby([\"dataset\", \"type\"]).agg({'rel_time': ['sum', 'mean', 'sem'], 'acc': lambda x: x.iloc[-1], 'loss': ['min'], 'bandwidth': 'mean'})\n",
    "in_group = observer_results['rel_time']['mean']\\\n",
    "    .groupby([\"dataset\"])\\\n",
    "    .transform(lambda x: x / x[x.index.get_level_values('type') == 'plain'][0])\n",
    "in_group = in_group.rename('mean_factor')\n",
    "observer_results = observer_results.join(in_group, on=['dataset', 'type'])\n",
    "observer_results['mean_factor'] = observer_results[\"mean_factor\"].round(2)\n",
    "\n",
    "in_group = observer_results[('bandwidth', 'mean')]\\\n",
    "    .groupby([\"dataset\"])\\\n",
    "    .transform(lambda x: x / x[x.index.get_level_values('type') == 'plain'][0])\n",
    "in_group = in_group.rename('bw_factor')\n",
    "observer_results = observer_results.join(in_group, on=['dataset', 'type'])\n",
    "observer_results['bw_factor'] = observer_results[\"bw_factor\"].round(0)\n",
    "\n",
    "display(observer_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Series([], Name: rel_time, dtype: float64)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(observer_data[(observer_data[\"dataset\"] == \"Shakespeare\") & (observer_data[\"type\"] == \"l2\")][\"rel_time\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "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>acc</th>\n",
       "      <th>mean_s</th>\n",
       "      <th>sum_s</th>\n",
       "      <th>bandwidth_GB</th>\n",
       "      <th>mean_factor</th>\n",
       "      <th>bw_factor</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dataset</th>\n",
       "      <th>type</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=\"3\" valign=\"top\">CIFAR Large</th>\n",
       "      <th>l2_opt</th>\n",
       "      <td>0.816200</td>\n",
       "      <td>149.268409</td>\n",
       "      <td>23882.945409</td>\n",
       "      <td>70.71</td>\n",
       "      <td>6.860000</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>0.849900</td>\n",
       "      <td>63.506440</td>\n",
       "      <td>10161.030440</td>\n",
       "      <td>143.15</td>\n",
       "      <td>2.920000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>0.857900</td>\n",
       "      <td>21.748937</td>\n",
       "      <td>3479.829937</td>\n",
       "      <td>16.89</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">CIFAR Small</th>\n",
       "      <th>l2</th>\n",
       "      <td>0.602700</td>\n",
       "      <td>110.473556</td>\n",
       "      <td>11047.355556</td>\n",
       "      <td>68.28</td>\n",
       "      <td>59.709999</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>0.586100</td>\n",
       "      <td>18.171788</td>\n",
       "      <td>1817.178788</td>\n",
       "      <td>13.51</td>\n",
       "      <td>9.820000</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>0.603000</td>\n",
       "      <td>86.193707</td>\n",
       "      <td>8619.370707</td>\n",
       "      <td>20.66</td>\n",
       "      <td>46.580002</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>0.607200</td>\n",
       "      <td>14.550101</td>\n",
       "      <td>1455.010101</td>\n",
       "      <td>20.58</td>\n",
       "      <td>7.860000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>0.607300</td>\n",
       "      <td>1.850263</td>\n",
       "      <td>185.026263</td>\n",
       "      <td>2.45</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">MNIST</th>\n",
       "      <th>l2</th>\n",
       "      <td>0.986500</td>\n",
       "      <td>47.194950</td>\n",
       "      <td>7551.191950</td>\n",
       "      <td>34.20</td>\n",
       "      <td>20.959999</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>l2_opt</th>\n",
       "      <td>0.973400</td>\n",
       "      <td>10.682082</td>\n",
       "      <td>1709.133082</td>\n",
       "      <td>9.33</td>\n",
       "      <td>4.740000</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf</th>\n",
       "      <td>0.986700</td>\n",
       "      <td>40.392748</td>\n",
       "      <td>6462.839748</td>\n",
       "      <td>10.65</td>\n",
       "      <td>17.940001</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>linf_opt</th>\n",
       "      <td>0.987100</td>\n",
       "      <td>7.725151</td>\n",
       "      <td>1236.024151</td>\n",
       "      <td>10.55</td>\n",
       "      <td>3.430000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>0.987600</td>\n",
       "      <td>2.252050</td>\n",
       "      <td>360.328050</td>\n",
       "      <td>1.30</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Shakespeare</th>\n",
       "      <th>linf_opt</th>\n",
       "      <td>0.573490</td>\n",
       "      <td>918.494833</td>\n",
       "      <td>17451.401833</td>\n",
       "      <td>53.46</td>\n",
       "      <td>1.190000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>0.572081</td>\n",
       "      <td>773.212111</td>\n",
       "      <td>14691.030111</td>\n",
       "      <td>6.30</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           acc      mean_s         sum_s  bandwidth_GB  \\\n",
       "dataset     type                                                         \n",
       "CIFAR Large l2_opt    0.816200  149.268409  23882.945409         70.71   \n",
       "            linf_opt  0.849900   63.506440  10161.030440        143.15   \n",
       "            plain     0.857900   21.748937   3479.829937         16.89   \n",
       "CIFAR Small l2        0.602700  110.473556  11047.355556         68.28   \n",
       "            l2_opt    0.586100   18.171788   1817.178788         13.51   \n",
       "            linf      0.603000   86.193707   8619.370707         20.66   \n",
       "            linf_opt  0.607200   14.550101   1455.010101         20.58   \n",
       "            plain     0.607300    1.850263    185.026263          2.45   \n",
       "MNIST       l2        0.986500   47.194950   7551.191950         34.20   \n",
       "            l2_opt    0.973400   10.682082   1709.133082          9.33   \n",
       "            linf      0.986700   40.392748   6462.839748         10.65   \n",
       "            linf_opt  0.987100    7.725151   1236.024151         10.55   \n",
       "            plain     0.987600    2.252050    360.328050          1.30   \n",
       "Shakespeare linf_opt  0.573490  918.494833  17451.401833         53.46   \n",
       "            plain     0.572081  773.212111  14691.030111          6.30   \n",
       "\n",
       "                      mean_factor  bw_factor  \n",
       "dataset     type                              \n",
       "CIFAR Large l2_opt       6.860000          4  \n",
       "            linf_opt     2.920000          8  \n",
       "            plain        1.000000          1  \n",
       "CIFAR Small l2          59.709999         28  \n",
       "            l2_opt       9.820000          6  \n",
       "            linf        46.580002          8  \n",
       "            linf_opt     7.860000          8  \n",
       "            plain        1.000000          1  \n",
       "MNIST       l2          20.959999         26  \n",
       "            l2_opt       4.740000          7  \n",
       "            linf        17.940001          8  \n",
       "            linf_opt     3.430000          8  \n",
       "            plain        1.000000          1  \n",
       "Shakespeare linf_opt     1.190000          8  \n",
       "            plain        1.000000          1  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# output\n",
    "cols = [('acc', '<lambda>'), ('rel_time', 'mean'), ('rel_time', 'sum'), ('bandwidth', 'mean'), 'mean_factor', 'bw_factor']\n",
    "\n",
    "output_frame = pd.DataFrame()\n",
    "output_frame[['acc', 'mean_s', 'sum_s', 'bandwidth_GB', 'mean_factor', 'bw_factor']] = observer_results[cols]\n",
    "output_frame['mean_factor'] = output_frame['mean_factor'].round(2).astype(np.float32)\n",
    "output_frame['bw_factor'] = output_frame['bw_factor'].round(0).astype(np.int32)\n",
    "\n",
    "# Server terminates at round id 39.\n",
    "output_frame['sum_s'] = output_frame['sum_s'] + output_frame['mean_s']\n",
    "\n",
    "#str(round(x, 2)))\n",
    "display(output_frame)\n",
    "\n",
    "output_frame.to_csv(\"data/e2e_table.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "### Accuracy plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "def _preprocess(df, suffix):\n",
    "    # set start timestamp to 0 (1 because of x-axis log scale)\n",
    "    df[\"rel_time\"].iloc[0] = 1\n",
    "    df[\"cumtime\"] = df[\"rel_time\"].cumsum()\n",
    "    df = df.rename(columns={\"accuracy\": f\"accuracy_{suffix}\", \"rel_time\": f\"rel_time_{suffix}\", \"cumtime\": f\"cumtime_{suffix}\" })\n",
    "    return df\n",
    "\n",
    "def _select_df(df, idf):\n",
    "    plot_data = pd.DataFrame()\n",
    "    plot_data[['round', 'accuracy', 'rel_time']] = df[df['idf'] == idf] \\\n",
    "        [['round', 'acc', 'rel_time']]\n",
    "    return plot_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "CIFAR Large"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/hidde/opt/anaconda3/envs/tf37_standard/lib/python3.7/site-packages/pandas/core/indexing.py:190: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n",
      "  self._setitem_with_indexer(indexer, value)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'why not also show l2?'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plots.e2e_time import build_fig_e2e_cifar_time\n",
    "from functools import reduce\n",
    "\n",
    "df_u = _preprocess(_select_df(observer_data, \"CIFAR Large plain None\"), \"u\")\n",
    "df_l8p = _preprocess(_select_df(observer_data, \"CIFAR Large linf_opt optimized\"), \"l8p\")\n",
    "\n",
    "data_frames = [df_u, df_l8p]\n",
    "df = reduce(lambda  left, right: pd.merge(left, right, on=['round'], how='outer'), data_frames)\n",
    "fig_e2e_cifar_time = build_fig_e2e_cifar_time(df)\n",
    "display(fig_e2e_cifar_time)\n",
    "\n",
    "display(\"why not also show l2?\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "Shakespeare"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plots.e2e_time import build_fig_e2e_shakespeare_time\n",
    "\n",
    "df_u = _preprocess(_select_df(observer_data, \"Shakespeare plain None\"), \"u\")\n",
    "df_l8p = _preprocess(_select_df(observer_data, \"Shakespeare linf_opt optimized\"), \"l8p\")\n",
    "\n",
    "data_frames = [df_u, df_l8p]\n",
    "df = reduce(lambda  left, right: pd.merge(left, right, on=['round'], how='outer'), data_frames)\n",
    "fig_e2e_cifar_time = build_fig_e2e_shakespeare_time(df)\n",
    "display(fig_e2e_cifar_time)\n",
    "# display(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaYAAADoCAYAAACkVECrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOydeXhURda437rdnc4eIAuB7EBCCISEHUEREVFQBvd9n1FnHL/R+URHRQf3EUfHb9SfyziKIMFd9lVAZA9hCVsSSEIIWcm+dSfd6e76/dHpJoGskAXkvs+Tp++te6rq3Ns3dbqqTp0SUkpUVFRUVFQuFJSeVkBFRUVFRaUxqmFSUVFRUbmgUA2TioqKisoFhWqYVFRUVFQuKFTDpKKioqJyQaEaJhUVFRWVCwptTysghJgMPA8sllIuaIf8w8B1QBlgA56RUho6KqOioqKicmHSo4ZJCDEduBOYBnzdDvk/AY8DI6SUFiHEa8C3wA0dkWkNPz8/GR4e3tFbUVFRUVE5g71795ZIKf07mq9HDZOUco0QIhu4vy1ZIYQv8BYwR0ppaUj+DMgWQtwopVzaHpm26gkPD2fPnj3ndD8qKioqKqdpaN87zIUwx1TbTrnfAd7AbkeClPIkcIrThq09MioqKioqFzA9PscEtDcm0sSGz+Iz0guAMR2QUVHpNDZu3EhoaCgnT5486/P48eN4e3sjhCAvLw9PT0/neWZmJq6urnh6eiKlxGCwT4F6enpSUlLiLN8hk52djU6nw8XFxXnNYDCg0+mcxx4eHtTX16PT6airq8NiseDp6ek8ztWF4WmtotalD3pTCTWKDyaNB771+ZRoAzEIL/yshVRofEHaMCkeeFvLAEGVpheKtKCTZuoUdwA0Z5w3h0W4IKQVDVbnsRYb9cKutw0NAhsCiVS0uLm6oFMUauutCKDeKlEEaBSByWJDEQKdRlBvldikRHC6ATkzuprj1D1gL/XGAHTuRc5PRV+BqTwKfe9jCAlSgM61BJvFzZ6pIVHR1lJXEYlrr3SExoy0umCzuKH3qLKL2fRYzK5oXCpBAZtNoiBAW4vNpgGbDpQ6FLRYrVrQ1KERYBP1SIsrQmMGYUVa3RCK/RjnXQnsU+Si0XGDbmfReaHlpASBQBECjSIaSreX38+zHx9P/U+n1dUSF4Jhai+BDZ+lZ6RXAzEdkGkVKSV1dXW4urpisVgwm824u7tjs9mora3F3d0dIQRGoxG9Xo9Go6G2thatVotOp8NkMgGg1+upr6/HYrHg5uaG1WrFZDLh7u6OlBKj0YibmxuKomA0GnFxcUGr1VJXV4eiKLi4uGA2m7HZbG3qYjAYcHV1bVMXq9WKq6trE10cZXVUl/r6etzc3NrURafTodVqMZlMCCFwcXFpoktz99WSLhqNBp1Oh9lsRkqJXq936rJjxw6Cg4MJDg7Gw8ODjRs3EhAQwNChQ8nMzGTDhg1cddVVREdHs3jxYoQQ3HXXXaxfvx4hBIqiYLPZkFKi0WiQUmKz2cjKymLKlCkkJSWhKAr+/v7YbPYGory8nPLycjZs2EBERAQbNmwgLCyMX375hbCwMH7OqMWHYnrLCmw2Gyc1ofhQjI+1jDJ8MGi1yMJaPK2VKIpCjcYHW2EtnlYLPtYyKjV9yNf70z8vixpNAMJsQxrtgxye1kpAT74tAh9LCZXagfSvznKm9a/PokQXCHUNL7YW/MyFpLiPwMtSRrVrPF6WMqo0vpzS9ANpAyGo0fjQu/4U5bq+ANQpbvYGXyjYG8bGgyxnnreAgPpGx7q+exGGvtTXBAPg3jeZekNfLIYQTNoTaAL2YS0aSX1NCJ4hGxESqnOvxj1gL1IKhJC4AFIKXLxyMBaNtJcTsA9j0cgmxkZIqKuIwidiFfU1/XHvm0R9TX90Xrm49UlFSkejLwGJVrF/t6bKcPQ+J0AqeHqcQtoEQpEgFRC2RsZQQXE5fd7YUIpGj0YCivb0sQCES83pR6ScPu5phACQICTWM67F9BqK1Wptd7t3rlwIQ3kdxXjGuQYwn4OMEyHEo0KIPUKIPfn5+WzevBmAnJwcli61T0tVV1eTkJBAfb39X+zrr7+mqKgIgHXr1nH06FEAEhMTSUxMBODo0aOsW7cOgKKiIr7+2u7fUV9fT0JCAtXV1QAsXbqUnJwcADZv3szBgwcBSE5OZsuWLQBkZ2ezfPlyACorK0lISMBqtb82ixcvdv7KXrNmDRkZGQDs3LmTpKQkAFJTU1m/fj0AhYWFfPvttwCYzWYSEhKcv9h/+ukn8vLyANi0aROHDx8GYN++fWzbtg2ArKwsVqxYAdgb54SEBBzBgBMSEigrKwNg1apVZGZmArB9+3bn3N2RI0fYsGEDAAUFBXz44Yekp6dTW1tLQkICtbX20d0PP/yQFStWsHHjRjZs2MCRI0cAWLRoEQsXLmTjxo1kZmayatUqQkNDWbBgAe+//z4bN24kJCSEBQsWsGbNGhYsWIDJZGLBggVs3ryZ/fv3s3//ftLT0ykqKmL16tVoNBqqq6tZs2YNGo0Gi8XCmjVr8PHx4YsvvmDZwRJ+OXCCsrIyVq9eTWZm5ukyrB4sOelBsdWTjIwMwsPDSUlJoY+mllSXWHI0weRoQ/GW1fZzJZhMz5F4WCrwtFZy1H0EKa5xeFjt58fc4snXhXHMLR6f+hKOucVjEq5ku0YjpM2ZJ81tBHqbkWzXaPQ2ozPNx1LCUfcRFGv7U6LrT7G2P6W6fhgUL6SEKq0vNinsnwgkwt4aSYnARrmuL+5996DzPIlEAaHQu/4UjubCM2QTXiGbAAX3vnvxDNmEzjMX9757cfPfj3vfvQDoPHPpM3Q+3gNW4Bmyid5R39llPfLxGbiM3tGL8B3+KfreqfgMWoLvsM/wGbgEnWcePgOX4Bv7Ca59jqD3PYJv7Ce49U3CM2gb7v12OI91Xtn4DFzSkC8Hn4FLnHkcnz4Dl4BiRud94vSnsICwIYS14dgKoqFHIkHvc6LBINvThCKbnDvkmpz/hhCNTxos7fTw6bgf8OxQu3fO9fd0dHEhRDiQBTwkpfyyFbmPgT8CAVLK4kbphwB3KeXA9si0pc+oUaPk9u3b1R5TJ/WYtm3bRnh4OKGhoWf1mHJycjhx4gT9+vXjm2++4f777ycoKMipy/r161mzZg2zZs1i/Pjxzh7Txo0bWblyJbNmzeLyyy936rJp0yZWrFjBrFmzuPLKK/n555+d+ceNG8eOHTtYuXIlM2bMIDg4mIULFwKgi7qKovT99JYVXHXVVfzyyy9IKbn88stZsf0wlXjiTTXpbvFEVO1hXGQAaWlpREVFse9EOan6WILNx8l1GcAoTSZKRTZBQUHk5eVh7jeKvTV+hJuO0s98gnxdGNmu0YTVpdG/PptKTR/S3EcggMHG/fhYy86ScZz3NZ+kVNePfqbj5OkHOo1JP3M2Bfpw+69yJMF1GZx0jcLx+9297z60Ve7Ue9eirXbH4mVAcanCVD4Ife+MJgNHEtDoDAhtLVq3Eiy1fpjKI3HxzgEJEoGLt30+u65kKEJI9H2OIoQNa707GpcabBY3pMUVxaUKkAjF1tAha9TcNTscdfEiRNPzjjSrjryOPA1fa4tl9wTXhU/n8fgnOtzuCSH2SilHd7S+i8kwPQL8B4iTUh5slF4GrJBSPtAembb0GT16tFS98jqP9PR0FixYwAMPPEBkZGSL6S2dT506lQ0bNnDrrbcSHR1NTk4OCxYsYMKECWzdupXJkyfj5+dHbm4uO3bsIDIyko3HTVwxfAD5h7czdepUVq9eja+vL8XFxdh6hZJXIxjiWkpVVRUajYaIUVP4OsXCkPrDeNQVOXWs1PThmFs8UbXJzqG1TM9RDKzZS5CbmdxaF9Ld4hlQdxgPazWV2j6c1A8mpv4wHqYi6gJHccDQG29LKVVaP9ytVRg13nhayqnR9qKf+QTF+jDCjEc4qY/CpLghsGGfpbDh1nc/9YYAbDX9CK7LoDS0Cgy9qDEORueZ4xy6qq8JcZ7r8kMpqY2nT9Bq6hVXqnOvtvckGoaydB6F1FUMwrVPCqDY53bOmNMQis05lOUYwnIObQHSptgbS2HDXG0vUyjN9xzs1/Ptw1qOKtrBhdAYq9jxdfVl/nULzynvpWCYvLA7MTwnpfywIS0COA5cJ6Vc1x6ZtvRRDVPbOCb8GxsaB+np6Zw8eZKrr766SVpjo5OamsqiRYu47rrrCAkJwbFuLCUlhS+++AJvb28qKytxc3PDbDY7h0+HDRtGVlYWDzzwAPn5+c5h1jOp0vlyVB/H3UO1PH7ndF588UUMBsNZhgYgJCSEnJwcIsdN4/tjEFa1FzebkQpdACf1gwi0FhBIKS7GUwQHB7P7lIZsN/uQmn2Yq2kLGmtLJZWBxHgZOFTjgxRNR8sdE/FKZR9MGg+CTBkE1edweIC3c/7EWDQKwG5QwtdgLhhGH2sRptp+1EcewFA4Go/APRgLR+HRP5Hasmhce2VgKohD3+8AloIYtP0PIYSkrmxwwxzLSkBQVxaFm/9hzBXhuPQ6AVLjnOyWWBGKxFLbB61bGVazJxqXs+c+usNonNljaK430dxxa2mNP9uq05G/Pfo1lv0tbG/X3H1PD5/On+KfOIeyzs0wXQhzTLqGzyaOGEKIu4QQyUKIwQBSymrgNeBOcXpM4I/AUofBaY+MyvnjmM9JT09vku4wQKGhoU3Sq6qqCAwM5JNPPmH27Nn85z//wWg08tNPP7FkyRKnXHR0NFarlfLycmw2GwaDgfr6ehRFQafTcfjwYSZOnEhkZCS+vr5ER0cTEBAAQP/+/Zk4cSJxcXH4UsXdQ7V8nWLl8yWbsFqt+EaOIt09nplhJrx1NoKHTcDFxYWioiKmTZvG6iNl1FsspHmMYb/XlWS5DsEqdORpQykSfowePZrc3Fyig3wAgRQaEAKNkOikGb3NSC+tGZ3hFOHhSaSKGoJcTfQ1n8TPUogirfhYS8Em6TVgOUqvUoJMGZzShVItPPCtL8IzaBte9eVobWZCao8iq/uiyQtHH7KHIk0wRXUjIT8Sz+BtkB9JXXEctaXRuPkdxlQ2AGNZLNaiCHQh+7Aa/JFIXH3T8BmwguqcyZjKI3HzP4y06hqMkgBhRdFYERoLikYiBOjcyxACtPoaRMP0U+M/sDfAjr+26IisgzPra+9xa2mNP9uqsy3j25Jse/Je6DT3Pa05sYaPkz/sNh16OvLDZcDDDaf3CyGKpJTLG879gAjAyyEvpZzXYG++FEJUY/c5urNxme2RUTk/IiMjeeCBB1iwYAH33HMPNpuNlJQUdu/eTUBAAF999RUPP/ywsyeUm5vrdIIA+1yDj48PXl5e9OvXz5nucKEeNmwYR44c4dZbb2Xo0KFkZ2ezcOFCrrrqKrZv386gQYMYNmwYer2eBQsWMG3aNLZv346fnx/Jyck88sgjREZGol+yiS/2GwnoM5niYoHeRbLkpA5cxpOUC1coCgJwc3PDhAsWcdoV20WaGOjnRkVRLp6WCg4fzmTChAmsSjyK14A8PCoklZbRRBqT8TAV2ee/LDo0I4LIqMvCd8BqSjKm0Q8zuZpAhpqS0bhnkx94iur8y/AKW0WAMRCfqjIy/QPwCkrG+0Q4xqCjBNa4kSMmEOS5CUvwUZTyPngEJ+Lqn4rQV4LZFRF6GDcXV1x7Z1BXPBTXgCO49U1x6q/zyT/9hUnwifj59PPXNvjItTDPYzW7o3ExYrNqUTSWs643Tpc24SxHShq81uwebmcaAynBZtGhaOt/Ez2L3zLN9SzXnFgDcE49p47S05EfdgI7gUeaufYB8EEz6fPaUW6bMirnR2RkJLfddhv//e9/ne7TAPn59gaxqKjIaZhGjBiBRqNhx44djBs3jqSkJO6+++5m55wefvjhJnNMVVVVbNiwwTkMOGjQoCZzT450Nzc3li1bxqxZs5z5t+87jFUZTEENgMTSMAHvIutwtdRx5z0PUFFSyLJly/jD1JnkV0t+PGJkUoQbifl6xMld3D0umn37jmG1Wtm4/zjZ3qMIKt2IadhxwtOMpLhMIIbDjBnox46S4xS7/cIUz5nkHKmD6PXkZs1gvEc9bqKWwwHFGE/MINK3jOzKgeQPO4aLtQJv7W5s9W4YQ3PxMwykcPARwstqMPgdR2gk9ClDAFq3cvvD0tt9wD36JWEoGIupOgxXv5Q2HQqsZjcUrQksLqCrA5MH6A0IKeyOV0KiVPZG8SnHVBmMi3cu2Owdq/rq/ui88kEKhGJB2uxu0wjZxDiZKkNx8c5BKNazjY8NhOa3YZQat9mNp87OPJb1nvZe6UW4jkk0s47pYMnB1rJ2GhfTOiaVCwibzcb69eudRikkJITRo0cTEBBAQEAAvXr1csqaTCZ27drFQw89RGRkJDExMa06PoDd8E2dOrWJsWkt3Wq1MmvWLDZs2EBQUBAnT57k2imTSP31FJF+erJKavn9ZX25++p4lv70A1JC/LBoNm7Mo/+1gRw3nODnjCAeGulBuLckwkfhm9Sx7Co7SNTvwqhO8ycpx527h2qJ9L2BE6YsVkWvYHJ1EIm5o9F5JlPim0J/wtAEn8LFQ4umWkfvyCWkWVzx0OuY6PEYk0eO5u0jj6G31CIBk1KNADQuBmyAf2Ak0V5D2KJbxsg+k8gorMHLXSHPfMDZ6NeWDMHT/ygh+nhOBibh0c8e6EQ0uDJorN5odCZczFFUiTS0ikIfYihxScZcFot7n0zifaewp3QTClosNtDrNCBt1PcqZ1zfsewWu5FWARqJrWoALt5ZjO07lt2n7HVJbCAVu9ODkOgUHfU2K/peJ9AIDRpFR73NuXrJjgIIUFCwNawbcujcJs21/i05U3TAyaIzEK0ca/W1vDLhNYb7x3WfQr8BVMOkck4oikJcXByFhYVcccUV7Nmzh379+p3lENGS0XEMBT7wwAOcPHnyLK89OG1sHOu1Wku/+uqr2ZlezKlghcysbPwjR/DP7w7wf/eOYtwgP/ZmlTLnuwPERpVz++23N8lXfsiFBUf/zeOznuTW2Cvs6UDvQ1tZcDSREa6PYAhXuCu2EqEvY3N1LrnmXO4edi8/pn/PtZEubC3cjhAWCsmkMK9h2FJj/xDaOsJcJvP0lGsBmG6cQbGxhKTCROIDRnCw+CCPxD7CuH7jyaw4zj/3vMUdg+9kZeYqNPWXkVu7hcoTN9DXJYoa0048ArdgqfXnpLLL2RIO9B5IUV0RA3wGcqA4mRG9p7Bzvy+e4RmADaNylOnB17GGtZjKYjmk7MRV44rZakWvU+w2Q9GADXaf2k2wZwi5NTm410dh9E5nbN8x7D61u8EAWVAESGGjjyacSlsOCgpRvQdwrPwoNqxoUHDVuKLX6PFy8Sa3JsdplIQQjA4YxSlDITk1OYR4hmBrZJyqzdW4aOxTzzXmGjxdPDFb63HR6DDWG6m31eOj93Eea4QGq7SiERrqbfY5SccxgE3anD0Am2xYl9RgEB1Gsd0GsgUUFASnexeywegqQiG9Il01TB2kx73yLjRUr7zWMZlM6PV6p8G5//77iYqKatEtvKMefOdCfrmR/1t7lC1pdlfvW8aEsPFIIW/cHseoCF+nnMM4nZn+1bYs3HxyWZrzEQ8OfYipYdM4WHyAeUn/oM5ST72t7qw6AR4d/kcqTRV8e/QbJgdfRW/X3vTS92747EVxbTFfHv6C6REzWHtiDc+OeY7h/nEcLD7A20lvnXV+a+Rt/JD+PY/HzubX/a78nLMcj/7bsBVP5tqw37EmYyNuweuxSHuDK20CnUbLpOBJ/JKzievCr2N7/nYm9p/ImhNr0Cl65l42l625W9iat4UrgiYR5BnEV6mL8NUM4omx97A1d4szQoLjs7SuFIkkzi+OAyUHnJ8CQR83X4SE41XHuaL/FVixEdkrki15WyitLUEgKDAU4K5z54GYBxnuH8eP6T+gQSHXkOd8dv08+nFL5K0cLD5AekU6t0Teel7vwJlUVVVRVFTk9OhU6Ty0Wi2urq74+/vj6uraquxF6y5+oaEappZJS0tj0aJFXHPNNfz8889trk3qaurMVr7alsVX27MwW2y4u2j4/eSB1FslsSE+TYyPg71ZpaTkVXHf5RFN0qvN1by+61WOlqVx2+DbWZO1mmfHPMfbSW9htVkJ8gomxCuEYM9ggryCCfYMobi2iHf3/JPpETOc8o5fxm0Zn8ayAEvTf2L+kS+I97iJ7XvCqamz4NF3L3EhvuTKjRgq+qP1zkRiQxEKIwNGUlJTQ8bxvnj338ed0bdzoOQANw26meH+cby682UOFCcz97JXnPU7DEBXGYMLhaqqKk6dOuVcrC0udje5CwgpJRaLhZqaGoqLi+nbty8+Pj4tyquGqZNQDVPz7Nu3j8WLF2O1WgkPD2fGjBld2gtqi53pxcxbmUJhhb03c93wfvz5mij8vVv/BdccKaVHeGfPPympLcZFccFsM3PH4Du5Z8h9VJur8dR5ntW4tWR4nh3zHECTa43zvLbrFe6JvpcbI29uUt57OxewNa2UCmMdxqJRjB/kx6NTg9lXvo4fjv2IDbsX3BVBk7h7yL0EeQYB8H7iQuoNfZ3DhGfq+Fs2QC2RkZFB//79cXdvObisyvljNBopLCxkwIABLcqohqmTUA3T2Wzbto2ffvoJKSWTJ09m5syZKErXLYH7alsWMUHerfZ4+vd2Y853B4gM9OLpGUOID+vd4Xqs0soPx77n69QEbNgI9gyhwlTO9QNuOKsH1JgzjdKZ6Zf1m8AVwZNazNvYWJwsNfDvtUfZfsweQSvE153Hp0VQInbwU8YPGOrtMQzH9B3LvTH3EeHTciOgYic1NZXo6Gi1p9TFSClJS0tjyJAhLcqcq2FSnR9UWkRKyfr161m7di0AN9xwA1OmTOnyf/iYIO+z5oIMJgvfJ2bzzc5s3rg9jpHhfXjzjjgmDQ5Aq+m4kSytLeVfe9/hUIP76xVBV3KgeD/PjX2B4f5xxPoNb9b4AKRXpDebPtw/jmfHPNfqZPdw/ziG+8dhqLPwxZZMvt2VjcUqcddreHBSKD5905if8QkVpgoAhvnGcl/M/QzxbVdwfJUGVKPU9XTlM1Z7TGeg9phOs2LFCjZt2oQQgttuu43LLrus2+p2OCq8flscJTUm/rU6jarael68cSg3jAg+7/KXZSzh88P/xUffi5sG3cRP6T+22ANqqed0LthsklXJeXy0IZ1ygxkh4Pr4QGKH5LPixHcU1dodOCJ7RXFvzP3E+8erjWwHSU1NbfVXvErn0dazVntMKp1OZGQk27dv5+6772b48OHdWveoCF/mzBrGU1/txWKz/3gK9/MgJqhXGznbx8yBs6g213D9gOvZlLPpnHtAZ9LcMOSP6T8Q2SuStMxefLUti+JqeyTmYSE+xA4/zJ6SBexOrQQg1CuMe4fcx7h+41WDpNImhYWF/Pvf/2bt2rXs37+/XXlGjBjB/fffz1//+tcu1u7cUQ2TShOklM4GMTo6mpdeegkPD48e0WXjkUKnUbp8sD9v3zkCRTm3xjq/Jp9PD37ME/F/wd/dH0Uo3BtzH0CrzgGOobf20twwpL8ujL9vfZ3SzGuprwnB38uFGy43s6/6czY2RMoIdA/kriH3MCn4SjRCc073qHJpUltbS3l5ebvlp02bxtChQ7tQo/NHNUwqToxGI/Pnz2fq1KkMHjwYoMeM0i8pp1h7sAA4vS5pf3ZZsw4RbfFrzmY+OvAhtZZaFqZ8ydOjn+lsdZ2MivDljdvjmPPdAebeHEtafhVf/FoBbtfiE7GWoe7XYfM4yqpCe1w7L50X98bcz9Swa9ApujZKV1FpSmBgILGxsS1G2m+OefMu/IhtqmG6BGlu0WtlZSWffvopBQUFFBYW8sorr3Sp511rWG2Sd1fbG+5bx4Yw+/oYpgzt2+zi2Naos9Tx6cFP2HjSHsB0Yv/LeWz4n7pMbwejInx5bMognk7YR0OHj7g+sUSGGNhSsATM9kgD14Zfx+9jH0Gv0Xe5TiodpzsWh3cGGs1vr4d9IWx7odLNnLltRXFxMe+//z4FBQUIIbjpppt6zCgBJGeXYbZIrhkWyP9Ot0+sNu6J7M0qbbOMrMrj/O/mJ9l48mdcFBcej3+CZ8c8h6eLZ1erz96sUj5Yf8xplK4d3pexYw+xpWC9U+amQTfzePwTqlG6gOno9i5dgZSSlStXcvXVVzN//nxmz56Nl5cXI0eOJDk5udk8RUVF3HXXXbzwwgtMnTqV2bNnO6+tWrWKyZMns2DBAgA2bdrEzJkzefXVV5k/fz79+vUjKiqqyW4APYFqmC5BGseq27lzJx988AFlZWUoisKDDz7IyJEje1S/lLwq/nFHHK/dFtdkTslhnFLyqlrNX2Wq5G9bniG3JpdQrzDenfwe14VP7xZngi1pRfx10T6MZis6jeC+y0PYY/iKZZlLUFBw07pxx+A72XDyZw4WH+hyfVTOncb/Jw7j1N3RTQCCg4NJTEzk+++/Z8qUKaxatYri4mJuvfXWZkMuPfPMMyiKwptvvsmnn37Ku+++S0qKfQQiMDCQxMREHN7YUVFRpKSksGnTJvr06UNycjJubm68++673XJvLaEO5V2iOLat+PLLLwH7cMBDDz3Uo5OiRVV1bEkt4p4J4S06OYyK8OWE5RcOFle1uoD11qjbKa4t4g/DHkGv7Xg0iHNhVXIeby47gtUmcdEqvHHHENYXf4TGOw1p06DTujBn3EttrpNS6Rpa80K77bbbmDBhAgA7duzg+++/b3L9o48+ch4//vjjTqP07rvvkpub22yZ48eP54477gAgJyeHkJCQDusshCA+Ph5fX19mzZrFjBkzAHjttdd46KGH2Lx581l5Jk2ahL+/PwC+vvZh75KSEgBGjRqFn5+fUzY4OJigoCAmTZrErFmzALj22mtb7I11F6phuoTRaDQIIZBSctVVV/WoUZJS8uayI+zKKKHMYOLRKS3/Go3sFXlWo55SeoS0slTneqRYv+Hd6m79zc4T/N/aowDodQov3zaAH/LmkVV5HA+dB1abpDrreupjg8H/tCu6apwuPrqrp9QYIQR6/fwJ6mwAACAASURBVOlhX8fcVkZGBm5ubk1kf//735OXl8fcuXOdYZka75l25pyUoihNhu7d3d0xm82dfg8dQTVMlzB6vR43NzdGjhzJzp07iYqK6pF/OoBle3PZlVGCt5uWm8e0PnbfuFGfPfpZjpYfZXHqIgCeGf23bm3kpZT8Z1MG87ccB2BilB/TRupJyH6DIuMp+nv0Z3z/CYwMGEn9sGBS8qqczhvnsk5K5dx577332iU3YcIEZ+8JTg/fTZw4ke3bt5Oenu78P3n66afbVea59JZaIyAgAKCJsXKwdu1a3nrrLX744Qf8/Px47rnnOrXu7kA1TJcojn+2Bx98kMjISIYPH97tY+cO8suNvL/O3tt45voY/LzadggY7h/H/4x4kld2zsUq7fsyXRE0ifH9uy86hc0meXd1Kj8m5aAIeGHWMCLDqnl11ytUm6uI6h3FS+NfxkffEH3Zn7M8Cju6TkqlezlzTsmxg3JP/J803n/MsVP0hAkT2LVrVxO5P/3pTzzzzDNNhuwuNlTnh0uQ9PR05s+fz3333dfs5n1neiF1JTab5LWlhzGarVw9tC9ThwW2K59VWlmTtdpplCYFXckzY/6GVume31oWq42XfzrIj0k5uGgV/nFHPH375TFn+wtUm6sY3Xc0r0/8x2mjpHLR0dYml935fwJw8ODpbc0TEhK4+eabiY6Oxmq1NjFa1dXVLF++nIyMDP71r38B9iE/R6i1M+Xr6+ubDPU5ZHoS1TBdgmRlZWGxWEhISMBgMDjTHf90J0+e7DZdvkvMZv+Jcnp7uPDM9THtnhf6Nu0b9hXtRSC4IWImycX7u83Lrc5s5dmv97P+UCHuLhreu3ckJrdk3kh8DbPVxNTQa5gz7u+4dpPThUrX0NLOyj3xfwJw9OhRXnjhBZ566imOHz/OggULyMzM5Ouvv6agoICPP/4Ym83GG2+8wa5du7jnnnuYNm0ao0aN4vvvv2fIkCF8/vnn5Ofns2jRIpKTk1m2bBkHDx5kxYoVbN++ncTERFauXMmBAwc6tGi3s1GDuJ7BpRDEde/evSxatIjQ0NAejZdls0n+/GUS+7PLefuuEUyKDmhXvmJjEY+s/z02bDw49GFujrylSwKuNkd1bT2zF+/jwMkKfNx1vHfPSA7WrGFxmn2O6/aoO7lnyL1qnLse5LcYxDU8PJyXX36ZBx98sKdVaUJXBXFVe0yXILt37wZg7NixPaqHogg+eGB0h4wSQIGhAFetGzdE3MDNkbcATR0iuqrnVFpj4vEvkzhwsoIAb1c+emg0m0sSWJy2CAWFP8X9mXtj7lONkorKeaI6P1xilJWVkZ6ejlarZcSIET2tDlqN0iGjBPb9kF4YN6dTooG3l/zyWv6ycA+5ZUZCfd15555YFmW8z66CnegUHbNHP8tl/Se0XZCKyjlgtVqbXUz7W0U1TJcYSUlJSCmJjY3tsa2n0wurmP/rcZ6+fgi+nu0LySOl5IvD/yXOP75To4G3h+NFNTy5cA/F1Sai+nnx+h1R/L9Db5BaloKHzoOXxs8lxvfCjtascnFiNBr56KOPyM/PZ8GCBcTFxfX4SEd3oA7lXULYbLYeH8art9h45adDbEo5RcL2E+3Otz57Hcsyl/J20ltUmiq7TsEzOJxbwR+/2E1xtYkRYb159Y5w5u17kdSyFPzc/Jl3xTuqUVLpMtzd3Zk9ezZWq5Vt27ZdEkYJ1B7TJUVVVRWKotCrVy+ioqJ6RIfPf80k41QNwX3c+MPkge3Kk15+jE8PfgzAn+L+3G0u2LszS/nbN/upNVuZGOXPY9f58HLic5TVlRLmHcbcy17Fz+3iXSuionKhohqmS4hevXrxwgsvUFFR0SPRww/nVrBw63GEgBdvjMVd3/brV2Wq5K3db2KxWZgRcT1XhU7pBk1hU0ohc384SL1Vct3wftx0uWTuzucxWAwM9R3GnHEvdUukchWVS5EeH8oTQgwRQvwghHi/4bPV0NZCiFQhhGzm74cz5Eaccd0khAjq2ru58BFC0Lt3726vt67eymtLDmOTcPdl4cSHta2DVVp5d+87FNcWE9V7ML8f9kg3aArL9+by4ncHqLdKbhsXytXjKnkl8e8YLAYm9J/IKxNeU42SikoX0qM9JiFECLAZuFFKuVMIMRjYIYQYJ6XMaEb+auAU8AlgaHRpNvDjGeJPA88CjoVaJ6WUeZ18CxcNeXl5eHh40KtXrx6p/5MN6WSXGAj39+DRKYPaleebtMXsL9qHt4s3fxvzPDpN1+/wumhbFh/+fAyAP0weSN/QFN7Z8x8kkusjbuAPwx9Vtz5XUelienoo7y0gR0q5E0BKeVQIcQB4B7ixGXk/YIqU0hk/QwjhCcwDVjRKiwfypZT/7ErlLyZ++OEHsrOzefTRR4mOju7WuqWUaDQCnUYw96ZY9Lr2NewjAkay8eQGnhz5V/zd/btcx483pLNwWxYAf71uMHXev/DZIXtH/P6YB7gl8jZ1jZKKSjfQY4ZJCOEO3AzMP+PSbmC2EKK3lLK88QUp5bfNFPU7YJOUsqZR2lxgjBDCG1gopdzRiapfdJw6dYoTJ06g1+uJiIjo9vqFEPzPtMHcMT6MAO/2h+mJ8R3KJ1M/w0Xj0oXa2bdy/+fKFJbuzUWjCJ6fFU269Xt+Sd+ERmj4nxFPMiW057fQVlG5VOjJOaaRgCtQfEZ6AaBpuN4ebgecBqvBGJmAPOAPwDYhxDtCtG/8RUpJXV0dABaLBaPRCNhdrQ0Gg3PnR6PR6Ax0WFtb61z8ZjKZMJlMgD04Ym1tLWBfIOcoS0qJwWBwBk40Go1YLBYA6urqnHuhmM3mduliMBha1cXhIh4bG4ter2+ii6OsjuriuK+2dDHUmZy69HJVnM+lpfsqqypj/6l9Tl0UqTh1cdyX2Wx2PuPGulitVmfsvzOfcW1tbbP3ZaitY863+1m6NxcXrcKcWREkGT/jl5xN6DV65ox7iSmhVzf5vhvr0tr33ZIu5/t9X0zv3pnfd3e9ezabzZlus9mcdTQ+llI65R3HjjyNj1vL356yukOXzr6v9uZ3HLf07p0rPWmYHGGkS89Ir274bDMcgBDCC5gIrHKkSSmrpJR3SinHARHAcuzzTX9vpZxHhRB7hBB78vPznbtC5uTkOAMZVldXk5CQ4PzH+/rrrykqKgJg3bp1HD1q37YhMTGRxMREwB50cd26dQAUFRXx9ddfA/YvLiEhgepq+60uXbqUnJwcADZv3uyMIpycnMyWLVsAyM7OZvny5QBUVlaSkJDgfHkWL17s3KFyzZo1ZGTYp+d27txJYmJik6jCAIWFhXz7rd2Wm83mJsFcf/rpJ/Ly7FNxmzZt4vDhwwDs27ePbdu2AfYgsCtW2EdOy8vLSUhIcL6gCQkJlJWVAfDBN2u57d9bScwoYfv27U49jhw5woYNGwAoKChw7hZqNBp5cdULzN35Eiszl/P9999TUFAAwIYNGzhy5AgAe/bsYfv27QBkZmayapX96y8rKyMhIQGw/0MlJCRQXm7vdK9YsYKsLPsw3bZt29i3bx+1ZgtPfL6DzWnFeOi1PHm1D99mvmqf19J5M6Z4HMN6xQLw7bffUlhYCMD69etJTU0F7AuWd+7cCdgjOK9Zswaw7xi6ePFi53NPSEigstK+/mr58uVkZ2cDsGXLFuduoQcPHvxNvXtJSUmAPZ7a+vXrge579yoqKpwNZU1NjdMY1tXVUVVV5XwWjvfDZrNRVlbmbJzLy8udz7uqqsppGI1GIzU19sEZk8lERUUFYDeYjrrB/i46nlFLutTW1jq/B7PZ7CzLoYvjvlrSxWAwOJ+dyWRyvl+t6VJZWenUpbq62mlAWtLFarU20aWsrMypS2VlpfO4pXfvnJFStvsP+KAj8m2UdSt2x4Q/nJH+SEP6Le0o415gcRsyGmA1UAlo2ypz5MiRsra2VkopZX19vTQYDFJKKa1Wq6ypqZE2m01KKaXBYJAWi0VKKaXRaJRms1lKKWVdXZ2sq6uTUkppNpul0WiUUkppsVicZdlsNllTUyOtVquzrPr6eimllLW1tdJkMkkppTSZTO3SpaampkVdkpOT5VNPPSVff/31ZnVxlNVRXRxltaRLhcEkZ7y9SY77+1q5cOtxWVdX5yzLbDY3e1+rMlfKmUtmyFuX3ySzKrLO0sVxXyaTyfmMG+tisVhkTU1Ns8/YaDQ2Kau4okb+4bNdctzf18pr39oot2YelY+se1jOXDJDPrLuYZlbmXPWc3E848a6tPZ9t6TL+X7fF8u711gXR1nd8e6lpKRIq9XqTLdarc46Gh/bbDan7o5jR57Gx63lb09ZnaXLxx9/LAG5d+/eLr+v9uY/cuRIs9+34zsC9shzsA8dnWP6sxDCA5gnpTx6fiaRnIbPM/2GvRs+i9pRxu3A560JSCmtQogXgb2AP/ahwhYRQuDqap8H0Wq1aLX2R6QoCh4eHk65xuF8Gm9t3HhHSZ1Oh05n9yTTaDTOPEKIFsty1A3g4nJ6bqU1XRofn6nL/v37ARg3bpzzWmNdWruv89HlX2uOUFpjJjakF3dPCEejnHYaaPxcHGUdLUvjv4f+A8AT8X8h3CecxrRHF41G46z/zGfc+LnU1AueXJRMZlENgT6uPH1Tbz5NfZlKcyWDeg3ipfEv09u16WvZ0nNp7ftuSZfzfcYXy7vXli5d9e45rrV1LIRwbjPe+NihZ3vyO5xhWiurs3RJTk5Gp9MxbNiwLr+v9uZ33H9L3/e50tGhvHwgEfhACLFGCDHjPOpOBWqBfmekBwN12A1JiwghfIBxwNp21HW0oa4zhw1/00gpURQFrVbL6NEdjjx/zmxKKWTdwQL0OoWXbhrWxCg1R6WpknlJ/8AiLVwfcQNXhkzuMt1yy4w8+nkimUU1hPl58D836vjg8CtUmisZETCSNy5/6yyjpHJp8tW2LPZmNd9k7M0q5asGD87uIjk5mZiYmCaG+rdKRw3TdVLKT6WU04D/Ba4XQuwTQjzZMN/TbqSUVcAPwKQzLo0ElkopjW0UMQtYK6U0taO6ccB/pJTmjuh4sSOE4P777+e1117rtvVLZTUm3l6RAsAT10QR6uvRqrxVWnlnzzxKakuI7hPNw7F/6DLdMk5V89jnieSX1zKkvzf3XlfDh4feos5ax1UhV/PS+Lm4ad3aLkjlkiAmyJs53x04yzjtzSplzncHiAnybiFn52Oz2Th06BDx8fHdVmdP0iHDJKU83Og4VUr5Z+A64H4gVwjxgRCiI0HYXgUGCSGiAYQQQ4EY4IWG89lCiEQhRHMByZp44zkQQtwphFgohIhoOI8DHgae74BevykaD4t0NW+vTKHCWM/oAX24ZUxom/LFxmJyq3Px0feyL6JVumYR7cGT5fzpi92U1pgZGdGba67I5rMjH2CVVm6JvI2nRv6127ZlV7k4GBXhyxu3xzUxTg6j9MbtcYyK8O02XY4dO4bRaLxkDFOH/hOFEF5SyuqGYy3we+xGJAT70NtB4B0hhAT+JqVMa608KWWGEOJa4A0hRCYQBFwlpXT0kfti96xr8jO2YRhvFPBzM8WWAlcAR4QQ+7B77D0kpbx0NjMBiouLKS0tJSoqqlvj4v1uZDBZxQZenDUMpY0hPIBAj0Deu+p9io1F+HZRQNRdGSU8900ydfVWroj2IyI6kW+OrUIgeCT2UW4Y+LsuqVflwmH83PPzEvvzl3taPW8vu1659pzyOTw3VcPUPPuEEGOA27AbpDBgH/BnKeXKBpnPhBAjgGVCiPuklLtbK1BKmQjc0sK1Z4Bnmkmv5Oy5Kce1n7Ebs0uarVu3snXrVqZMmcLMmTO7rd4JUf6MH+TXplEyWU3oNfYJ0176XvTSd81Q44bDhbz800EsVsn0eH9c+q9lbfZ2tIqW/x01m8uDruiSelVUOhPVMLXOQOw9EgHsB56UUi4/U0hKuV8IkQp8gH1+R6UbsVgs7NtnX6TaHS+ylJLMohoG9bVPM7ZllOosdTy7ZTaj+o7i3iH3o1G6Jvbc0j05zFuZgpRwy/gAyjy/YW/BYTy0Hrww7kVi/Yd3Sb0qFx7n2lOB08N3N48J4aeknG4fxgO7YQoLC+uxWJfdzbmM8ZwAbpJSjmrOKDViNNC9QdlUADh8+DAGg4H+/fsTHBzc5fUt3ZPL/R/v4Mstx9uUlVLyUfKHnKjKYlfBTkzW9viudAwpJQu3HuetFXajdP9kX066fMaR0sP0cfXlH1e8rRollXbReE7p0SmRZ805dRfJycmt/sgsKirizjvv5LHHHnPuejtjxgznwuaLjY4apmJgpJRyWTtk3wQe7bhKKudL411quzroaF6ZkffXH8UmIahP2x5ta7JWsTn3F/QaPc+PnYO7rnO3d5dS8uH6Y3y0IR0h4JFpvdhj/oDs6mxCvEL456R3zlojpaLSHM05OjTnENHVFBYWcurUqVYN07x583j11Ve56qqruOGGG9i9ezdff/01CQkJztBOFxMdHcqLbHDzbhMp5UfnoI/KeVJZWUlaWhqKojBq1Kgurctmk7y+9DC1ZitXDw3kmmHNTvs5SStL47+HPgPgiRF/IdQ7rFP1sVhtzFuRwor9eWgUwaPTXVlX/C9q6muI7jOEl8bPxculQ6saVC5hUvKqmh22cxinlLyqbhnSc8wv1dTUOMNUOdDr9UyfPh2bzUZUVBRRUVH83//9H8899xw+Pj6MGTOGwsJCQkPb9pC9kOioYTIKId4D3KSUfwQQQvTH7i7+q2zYvkKl50hKSkJKSWxsLJ6eXbuZ3XeJ2ezPLqePpwvPXD+kVdkKUwXzdr+JRVqYOeB3XBk8uVN1MVts/P2HA2xOLUKvU3hwupUVee9htpkZ3+8ynh79jNPZQkWlPdx3ecs+VKMifLttnslhmN59913efffdpnqMGsX06dMRQpCZmUlOTg56vZ6XX36Z+fPns2fPHh599OIbuOqoYXoTeBL7PBMAUsp84C0hxFdCCCml3NWJ+ql0EHd3d3r37s3YsWO7tJ4TxTV8vCEdgOd/N5ReHq2vRp9/+HNK60oZ0ieGB4c93Km6GEwW/vbNfvYcL8PTVcsd08pYlrMQGzauC5/BY3F/VDf3U7loee6553juuedalXn22Wd56qmn0Gg0LFu2jO+++46ZM2fyl7/85aKMFCGklG1LOYSFOAb8DdjgWM/U6NrlwKtSyimdq2L3Mnr0aOmIgn2x4oiQ3JXrl576ai+7Mkq4YUQQL944rE35KlMlnx36Dw8OfRhft877pVlpNPPXRftIyaukj6eO6yZnsiHPPtxxz5D7uD3qDnVzv0uM1NRUhgxpvQev0jm09ayFEHullB2Oh9bRHlOBlHJJC9ckdk88lR6mOxbUvnjjMD7dlM6T1w5ul7y33oenR5+1JO28KKqq48mFe8gqNhDYy4Vx4/eyIe8XFKHweNwTTAs/dxdhFRWVnqOjLZhFCHHWPkkNm/A9D5SdnUWlOzCbzfzyyy/OvWa6Gj8vPXNmDcPTteUQQnk1eSxMWYDVZu30+k+WGnj080Syig1E9HVh6OhN7Dz1Cy4Nm/upRklF5eKlo4bpQ2CHEOIBIUSsEGKkEOIPQDIwHVA98XqIAwcOsHz5cr788ssuq8NssbEkKQeL1dambJ2ljn8kvsEPx77j+2NnhTQ8L44VVPHY57sprKhjSIiWvtHLOFS6Dy8Xb16f+CZjArt2fk1FRaVr6dBQnpRyiRAiHPgUcPxUFtiH8f6flPLtTtVOpd041i515fYWn2/OZMHW4+zPLufVW1teoCql5MPk9zlZnU2wZzC/G3hjp+mQnF3O7MX7qKmzED9QwRb4Dcer8glw78vLl71KsFfXLyhWUVHpWjocTllK+Z4Q4lvsPaS+2EMUbZRSZnS2cirto7S0lIyMDHQ6HSNGjOiSOg7nVPDVtuMIAbeMDWlVdtXxFWzJ/RVXjSvPj+u8RbQ7jhXz/HfJmOptjB9qpcxrIRXGCiJ8BjD3slfo49qnU+pRUVHpWc4pzn+Di3iTnWOFEAMAHynl/s5QTKX9OHpLw4cPb7KLaGdRZ7by6pJD2CTcMzGcuNCWN9JLLU3h88P/BeAvI58ixKtzFvatP1TAKz8dwmqTTBph4IRmMXWmWuL843h+7IudHkFCRUWl5+iwYWpwfogAXLAP4zlwBx4E7uwUzVTahc1mIykpCaDL1i59vPEYJ0uNRPh78OhVg1qUK68rZ17SW1illVkDb+y0yN0/7D7Ju6tTkRKuHldKSv03WK1WJgVfyZMj/9plezipqKj0DB3dj+lOYEEr+S6prcsvBDIyMigvL6d3794MGtSy0ThX9maV8u2uk2gUwdybY9HrWl6oqlW0hHuHE+gRyANDHzrvuqWUzN9ynP9sso8ST514koNG+xqlGwfdxINDH0YR3bfXlIqKSvfQ0R7THGB9w9+VwE6gpOHaVGBpC/lUugi9Xk9MTAxhYWFdsn5p5f58AB6aNIDo/j6tynq5ePH3y16m1lJ73rvB2myS99cd5Ztd2QghmXx5CgdrNgLw8LA/cOOgm86rfBUVlQuXjrYeQko5E0AIsQ2YLKVc0HD+A/As8GPnqqjSGmFhYTzyyCN0JIJHR3jpxmGMHtCHa2NbDtCaWZFJmHcYWkWLIhQ8dB7nVafFauPNZUdYfSAfrdbK+MuSOFKzG63Q8uSov3Z6nD0VFZULi47+xC5yHDQ4OVwlhHBvODcAl8b2ihcgXRV2R1EE18cHodU0/6rkVufwwra/MWfb8xjrjeddn6neyvPfJrP6QD6uegsjxm/iWM1u3LRuzJ3wimqUVFQuATpqmNKFEL8IId4RQngD3wPLhRAzhBCvAJM6X0WVllizZg3Hjh1zxsbrLKpq65n740FOVda2KldrqeUfu9+g1lKLr5svbtrz8wg01Fn466K9bD1ajLdnHVGjVpJtSKW3vjf/uHwecf7q7x4VlUuBjhqm54Ba7NtchEkpv2o4Xwm8CCzsXPVUWqKwsJD169czf/58LBZLp5b97upU1h0s4K0VKS3KSCn5cP/75FTnEOIVwv+MePK8em3lBjN/XpDEvhPl+PUx0G/YTxTWniTIM4h5k95hQK+B51y2isqlxF133cWvv/7aqoyUkokTJ3LixInuUaqDdMgwSSnLpZQzpJQBUspDDck3AbOAKVLKJztdQ5Vmcaxdio+P79Sw9ptSCll3sABXnYb/nR7dotyK48vZmrcFN60bz4+dc169pVOVtfzxi92k5VfRL7AMr4HfUm4qIar3YOZd8U8CPQLPuWwVlUuJffv2kZ+fz5VXXtmqnBCCp59+mrlz53aTZh2jQ4ZJCLG4YaNAJ1JKi5RyhZSydROt0mlYrVYcW3OMGzeu08otqzHxdkMv6Ylrogjxbd6JIaX0CPMP29dX/2XEUwR7tR4JojWySww8+vlusksMhIUWoAn6DoOlhjF9x/L6xDfx1rfuCaiicqnwySefIIRwbhzYHJ9++il333238/zNN9/kmmuu4YorriAmJoYJEyZQWFgIwMyZM1m9ejWVlZVdrntH6ehQ3gxOx8hT6SFSU1Oprq4mICCAsLDO2Z5cSsm8FSlUGOsZM8CXm8e0bGw2ndyIVVq5cdBNTAy6/JzrTMuv4rHPEzlVWcfAQcep8/0Rs83MNWHTeGHci7hqXc+5bBWV8+WrbVnszWp+aeberFK+2pbVrfokJyej0+mIiYlpUWbjxo1MmDDBeZ6UlITBYGD58uWkpKQQGhrKRx/ZY23rdDpiY2PZunVrl+veUTpqmOYBa1u6KITYcn7qqLQHxzDe2LFjO80bb+3BAn5NK8JDr2XOrKEoSsvlPh7/BE+OeIoHYs59Ee2+E2U8/uVuKoxmBsccotprJRIbdw6+iyfi/4JGUXecVelZYoK8mfPdgbOM096sUuZ8d4CYIO9u1Sc5OZmYmJhWh+5zcnIIDDw99J2UlMS///1veve2hxGLj4+npKTEeT0wMJDc3NyuU/oc6eg6pnzgMSFELJDXKF0AscDEzlJMpXlqamo4cuQIiqJ0aiTxshozGkXw1+nRBPZqfr7IJm0oQkERCleHXXPOdW1JK+LF7w9gtliIHp5ImSYJBYU/xj3OdRHTz7lcFZXOZFSEL2/cHsec7w7wxu1xjIrwdRolx3l3YbPZOHToELfddlurcu7u7tTW2r1pCwoKKCwsZOTIkc7ru3fv5oYbbnCe19XVdUl8zfOlo4bpGSAGuL6F612zylOlCVOmTMFgMODj03nzL/dMDOeKwf6E+DYfDHVb3laWZSzlb2Ofx8/N75zrWXMgn9eXHsYqzUSP+JUyUnBRXJg9+lnG97/snMtVUWmL8XPXtXjtuZkx3DjaPny9dE/OWR6pf/5yj/P4/z042mmUHvhkJ0cLmt+cc9aoYJ7/3VAA0vIr24yc0hrHjh3DaDQSH9/6konhw4eTlpZGaGgoSUlJWK1Wjh49SkxMDCtXriQ1NZXFixc75VNTU/n73/9+znp1FR01TF8DrsAvwJnbkvYDPuuoAkKIIcBr2Htj/YE3pZT72sgzAmgsYwYGSCnzGsk8DwwFTNgXBr8opez8rVS7GU9PT66/vqXfBR3HYrU5F8+G+jXv7JBTfZIP9v+bWkstuwsTmRFxbvV/uzOb99amITS1RI34mTLbCTx1nrw4/v+3d95hUpVXA/+drXQWqQIrSJGm9CY2jCaK0YiiifGLifrZYhSxISoiEAufii2G2GKKGRFRJBAkICIiiMCCS1n6siC9bZ+Z3dmZOd8fd2aYHWZ3Z7bNLry/55nn3vu2e95z796zbzvvs/RuWXa/ucFQl6jNlpIf/4SHigzTzTffzMKFC/nZz37G2rVrueuuu7jrrrvIy8ujU6dOLFq0iAYNrLHbrKwsvF4v/fr1q3H5oyVaw/QRkKSq28NFikjZu8eFT58KLANGpt+X/AAAIABJREFUq+oqEemBtUPusAr2d3oUy/2Rv4X2Y4hRmgb0V9WrfdcfAm8Cf4hGvtOd/dkOxv4zjbFX9WBkr7Zh0zhKHLy42lpEe2nHyxjV+Zqo76OqvP91Jn/9JpO4xHy69FtIrucIrRu2ZvKIqdW2NYbBUB7fT7kqonSjB6cGWk9wckzpxiGpzFm7j3VZJwLG6R/3RdbKr0prCSI3TLfffjvDhw+noKCAtWvXMnbsWN57L3x7YcaMGYwfP77GvMZUhWh3sK1oGkq5LZ0wTAP2qeoqX/nbRWQD8AoQdttTEekPHFTVl8uI7wk8FpL/PeAbEXm/Pu8XtXTpUuLi4hgyZAiNG1fNH53Hqzw3dzMHc5ws2Xw4rGFSVf70wxvsL9xPatNz+EP/B6N+ib1e5dWF2/h0zY8kNjxGhz5fkO/Jo1Ozzky+cAotq9AtaDDUNKFjSoPOPSsmY0zp6el06tSJlJSUctM1adKEP/3pT2RmZpKWllbmOLSqkpqayp133lkT4laZaLe9+KCsKKAlVjffpxGW1Qi4EfhbSNQa4DERaaGqOWGyPgsM8blE+qeqfhcS/2sg3leOnzTAC9wG1EvDVFJSwpIlS3A6nXTv3r3KhmnW93tJ35tDyyZJPP7zXmHTzMucy8qDKyq9iNbt8fLHuZtZtPEQDZvtp2X3hTg8Ts5veQFPDZtIk6QmVaqDwVCThJvoEG5CRG2Qnp7O8OHDI0p7+eWXA5SafReKiDB27Nhqka0miLYr7/YK4qPx4jkQy5AdCwk/hGVYBgJfBUf4jFEx1ozAu4B7RORV4Img8aOLsLr4AnM8VdUhIvnAkCjkq1Ns3rwZp9NJhw4d6NChQ5XKyjpWyNtf7QTgyV/0oXmjU6ef7snbw98yrP9DHhr4MB2bdozqHkUlHp7+ZAMrdxyjWatdND5nMS6vm4vaX8zDgx4lKb76vFUYDDXBlgP5YY2P3zhtOZBfK4bp8OHDHDlypMJuvNOJyngXb6+qcaE/4EGgSxRl+Sfbh65gK/Ad24RmUNV8Vb1FVYdh7aI7D2u8KXhaSTsgN8xEh4JwZYa5B0VFRQC43W4cDsvWer1e7HZ7YHsJh8OBx2Pdwul0UlJSAkBxcTHFxcWA1crxT930eDyBslQVu90ecL7qcDgC/u6KiopwuVwAuFyugCzff/89YK1dCpXFbrdHJEtRURFuj5epczbhcnu5bkAHRnRvFVaWTs068evu/8MNXW5kRPuLSsnidrsD9QonS569mHEfrmPljmOc1X4TDVMX4lE313T+OY8PeQLxSrk6Lksv/nq5XK5AvYJl8Xg82O32sDp2Op0V6rgiWfw6DpalvOddliyRPO+69O6VJ0s0716oLBU977JkKe/d88vi9XoD4V6vN3CP4HNVDaT3n/vzeDwefnNRZwad2zJs/kHntuQ3F3WOqKyqyrJu3TrAWioyZ84c5syZw9y5c5kzZw4LFiyIul5V1Uu4epX17lWWaA3TOFU9XEbcO1hjO9ES2sryr6x0lZdJVfcBY4CFwDgRCW79hWu5xZdVpojcIyJpIpJ28OBBli1bBliL1ebOtfY+LCgowGazBf7wZs6cydGj1i4gixYtYvt2az7I6tWrWb16NQDbt29n0SJriurRo0eZOXMmYD04m81GQYFlg+fOncu+ffsAWLZsGRs3bgSs5vvy5cvJyclhx44diAiDBg0iLy8Pm80WeHk++uijQLN94cKF7NplzRtZtWpVYNv1rVu3snjxYj5ckcXWg/k0iXcz7uqeuFwubDZb4CM6Z84cDhw4gIjQdE8z+rkHAJYPrhUrVgDWbJ758+cDkJOTg81mC7yg7/9zJr//YDXpe7Np0fFbEtp+jaJcnHwJF9j7ESdxZGRksGTJEsBaazF79mzA+rDZbLbASz179mwOHToEwJIlS8jIyAAgLS2NlStXApCZmRn448zOzsZmswHWH5TNZiMnx+oNnj9/PllZ1hDpihUrWL/eGg7dvHkzS5cuBeDAgQPMmTMHsD5yNpst8HGcNWtWwJXL4sWL2bp1K2AtYFy1ahVg7Sa8cOFCwOpG8U/L9Xg82Gy2gOuXefPmsXfvXgCWL18eGNjeuHFjnXv3APbu3cu8efMAqvTugfXf/6xZswDKfPfAGk/dvHkzEPm7Z7PZyM7OBiA3NzfwoSwsLAwYw6KiIvLz8wO68L8fXq+X7OzswMc5JycnoO/8/PyAYXQ4HBQWFgLWxzg3NxewDKb/3mC9i34dlSWL0+kMPAeXyxUoyy+Lv17+f0qnT5/OmDFjGDNmDDfccANjxozhmWeeAaz35aabbuLee+8lPz+fV155hWuuuYbFixeXkiUvLy8gS0FBQeBvrSxZPB5PKVmys7MDesnLywucl/XuVRpVrbYf8EUUaYdhdbk9HhL+qC/8kgjLGehLf7bveiFgD5OuAPiqovIGDhyoTqdTVVVLSkrUbrerqqrH49HCwkL1er2qqmq329XtdquqqsPhUJfLpaqqRUVFWlRUpKqqLpdLHQ6Hqqq63e5AWV6vVwsLC9Xj8QTKKikpUVVVp9OpxcXFqqpaXFysTqdTFy9erOPGjdP3338/rCyFhYURyeJ0OnVd1gm94bVv9NuM/aXK8svy2dZP9WD+wTJl8evFX69gWQ7mOPTG15bpsGcX6PX/elyv+/waHT33Ol269ystKioKlOWXpSwdl6UXf72Ki4sD9QqWxe12a2FhYVgdOxyOcnUciSx+HQfLUt7zLkuWip53WbLE4t2rSJZo3r1QWSp63tG8e8GybNmyRT0eTyDc4/EE7hF87vV6A7L7z/15gs/Lyx9JWbUhy8MPP6xbt27VmTNn6uWXX66//e1vNTc3V3/729+qw+GIul6RypKRkRH2efufEZCmlbAl0U5+KGu/pYZYezENjaK4rVhbZoRujdoRKALWRVjOdl85/i7B9cDVItJMVfN9cjcHmgArKipMRALz/BMSEkhIsFQUFxdXasJBo0YnF6IGr5xOTk4OnCcmJpKYaLkWjI+PD+QRkTLL8t8bICkpCVUNuCDy+8AKlSX4vCJZBnZuwKwHLiYxIe6Uspbv/4a/b/uAeXvm8vaV750ii59wesk6WsjYD9M4XlhA+16LcCdn0SC+AROGPsXAtoOCVVxKL5HqOBJZ4uPjA/lDdRysl2jqVZEs5T3vsmSJtl6xevcikSWady9Ulqo+7/JkiYuLq/BcRIiPjz/l3C9nJPn9s1TLK6s2ZFFVevbsSc+ePXn99dd58sknad68OUOGDOHYsWOcc845UdUrUln89S/reVeWaLvylmEtrg39LcTaq+mFSAvyGY1POXVzwYHAXFWNdCLFMOBdVfV3032A1YIKLncg1qy8WZHKV1dQVa655hoGDhxIjx49Kl3OwZyT6vQbpWB+zN/LWz+8CcDN5/0qKgeqWw7kcd/f1nDckUP7Pv/GnZxF8+QUnr942ilGyWAwVD8iQmZmJsuWLSM5OZnJkyfjdDpJS0sr5TuvvhDtrDw31kSDQ0FhijWmk67lL4oNx1QgTUR6quo2EemD5fLodgAReQy4Gfi5qh4XkVuwPJw/q6pZItIPuBO4OyCMaqaIvAPcirWBIcC9wJuqWvbOd3WUuLg4BgwYwIABAypdxqZ9udz3wRrGDEnl4VE9T1mL5Chx8OKa5ynyFHFZx5FReXZI232C8TN/oFhO0L7PfEricmjX+GymjPgjZzcObQwbDIaaYPz48YwbN474+Hj+/e9/88knn3DdddcxduzYat2vrbaI1jAtVNVp1XVzVd0lIlcBz4tIJtABuFxPLuRtizX7zt9HcAK4BMgQkfXAAuAOVS0JKfpBYJrPQAmwEXixuuSuTxS5PEz9fBMer5KcGH+KUVJV3lj/GgcKD9CpWaeoFtF+s/UIE2dvQJMP0ea8BZSInW4p3Zl04WRSkstfCGgwGKqPdu3a8fHHHweu77nnHu65554YSlQ1ovX8cL2IJAGNVDUXQEQ6AA1UNbMyAqjqaqzZdeHiHsdyHOu//hLLUFVUppvKzRCsU6xfv57t27dzySWX0LFjdOuI/MxYsoN9Jxx0adOEuy/vdkr83F1zWHXoOxolNOLJoU9H3IX3nx8O8MK/N5PQdA8tuy7EQwkD2wziiaFPVmk3W4PBYIh2B9tewB7Av606avmoGyAiL0hddLpUj1m5ciVr1qyp9H4p67JO8MnqH4mPEybdcAFJYcaW4iSeeIln3KBHaN8ksoW7H323h+fmbiapxRZadP0PXkq4PPUKJg6fZIySwWCoMtF25b2KZcw+CQ5U1U9F5BzgYV8aQxU5duwYu3fvJikpqVIrvu1Fbv4411oHcudlXejZPvymZtd3G83wsy+kbePwTlyDUVXeWbqLvy/PpFHbNJq0X4UCN533S27r9ds66QzSYDDUP6I1TKlAJ1UtDhP3NfAxxjBVC/4p4v369Ss1dTZS3lu2i8O5RfRs34zfXVLaIUeJt4TcohxaN7IcYURilDxe5ZUFW/k8bS/NUpfToNVGBOHuvvdybZfropbPYDAYyiJaw3S8DKME0ANrDZKhini93sCq+aFDo1kadpI7Lu1Crt3Fby/pEthvyc/fN3/A1/uWMn7IBPq3qXi2X4nby5TPN7EkYz8pXRaT1HwXiXGJPDLoMS7qcHGl5DMYDIayiNYw7ReRm1V1dnCgiFwITKeeeu6ua+zYsYO8vDxatmxJ165dK1VG80ZJTB5z6vZY3+xfxvzd80iQhIjGg4pcHibMSmd11n5adl9AfOMDNE5ozNPDJ3F+q/MrJZvBYDCUR7SG6WlgrYj8AcszQyLW4tULsfzQ3VS94p2Z+Lvxhg4dGvW4zZebDzGyZ9uwi2j35u8JLKK9q+899DirZ7ll5TtLeNS2nozD+2jZYx5xySdo2aAlz144lc7NO0cll8FgMERKtNPF94rICOB1rLVCCVgeFVYC41X1++oX8czjqquuokWLFgwZEt0uHUszDvPM7I1ckJrCO3cOJS7upFGzl9h5cfXzFHuKuTz1JxXuRHuioJiHPkwjK28vrXr+GxIKSW16DpMvnErrRq0rVS+DwWCIhGhbTPi8O1wrIslYmwPmqGrVfJwbStG2bVuuuy66CQUnCov5v/9Yji2u6nt2KaPkX0R70H6Qzs3O5f5+fyi3JXYwx9py/UjxLlr2WABxRfRu2Yenhz1D06SmlauUwWAwREi0vvIQkfY+F0LFqnoQaCkiI6tfNEOkqCrT5mWQ5yhhSJeW3Dg4tVR8Zt4u1h5eQ+OExkwY+hTJ5SyizTxSwD1/XcMxzyZadJ8LcUUMP/tCpoz4ozFKBkMMefvttxGRwDYppzPRehe/EFiM5c27DYCq7vcZq/eAB8qZtWeogIMHDzJz5kwuvvhihg0bFnG+hRsO8u32YzROTmDi6D6lWksA3VK68/zFL+J0O2nfpH2Z5Wzel8sjtnWUNF5H847fgMA15/6cu/veS7zEl5nPYDDUPOnp6SQmJtK7d+9Yi1LjRNtiegnYBZTyl6eqa4DVwMRqkuuMxO/l4ccff4w4z5E8J9O/2AbAI9f0pG3z8DPterfsw6C2g8ssZ3XmcR74x1o8KctpmmoZpd/0uo17+/7eGCWDoQ6Qnp5O796966VT1miJ1jA1VdUBqhpuEe1a4LZqkOmMxOPxBLZQjqa1NG/dAezFbi7p0Zpr+p1sDZV4Spj83STWHFpdYRlLMw7zqG0tiWcvonG7NOIkjgcHPMQve9xivDkYzlg+2/kpG49tCBu38dgGPtv5aa3J4vV62bRpU6W8wNRHojVMeeXEDQNaVUGWM5otW7ZQWFhIu3btSE1NrTiDj7su78rT1/dhwnV9ShmRv25+j/VH1/HuxrdxecrepX5u2j4mfpZG407zadhyK8nxyTw97Bl+2ulnVaqPwVDf6Z7SnZfWTjvFOG08toGX1k6je0r3WpNlx44dOBwOY5jKYIuIjA921ioWv8bq5vu2WqU7g1i92mrZRLt2SUS4bmBHWjY9uYPk1/uW8kXWAhLiEnhi6JMkxYdv+v/z293838I0mnedQ3LzvTRLasbzF7/IkHaV8zZhMJxO9G3dj/FDJpQyTn6jNH7IBPq27ldrsvgnPJwphina6eITge+Be0UkHWuBbT8sV0R5WE5cDVGSn5/P1q1biYuLY/DgsseB/Hi8ypuLtnHT0HNIbdm4VFxWXhZ/Tn8LgHv63kf3Fuedkl9V+fOXO/ho7QZanDeXhOQ82jZqy+QRf6RDhB7GDYb6xC/mRr75ZTgmrnyq3OtImTd6QaXylWeYcnJyuO2229ixYwcNGzakbdu2zJgxg27dTt3mpr4QVYtJVU8Aw7G2Uh8CXA0kATZgMJb3B0OUrF+/Hq/XS+/evWnatOIp2bO+38us73/k4X+tw+PVQHihq5AX1zyPy1PMFedcyVWdrj4lr8ervDgvg4/Xr+Gs8z4hITmPLs278tKlrxijZDDUUdLT0+nUqRMpKaduwCkijBs3jh07drBhwwauvfZa7rrrrhhIWX1UZoHtCeAB368UIrILqL9mOkaMGDGCpk2b0qJFiwrTZh0t5O2vdgIw7uqexAdNDX8r/U0O2w9xbvMu3Nfv/lO6BF1uL5M/28i3+9bSovsCJL6Efq378+TQp2mU2Kh6K2Uw1CEq21KBk913o869hoVZX9R6Nx5Yhmn48OFh41JSUrjyyisD1yNGjODVV+v3Jg9RG6ZwiEgL4BUi2F3WcCpJSUkMGjSownRuj5epn2/C5fZy3YAOXNyjTan40d1u4EDhfp4c+hTJ8cml4hzFbibMSmdD9kpSui5BxMtlHUcyduA4EuMSq7U+BsPpQuiY0gWt+tb6GNPhw4c5cuRIxONLr7/+Otdff30NS1WzVMkwiUhH4CHgbqAZoOXnMITi9XqJi4usR/Wf32ax9WA+7Zo3YNzVpzpg7XlWL964/C3ipHR5eQ4XD9vWscf1Nc07rwTghm438rs+d5yS1mAwWISb6BA8IaK2jJN/fKmwsJC5c+eWiktOTmbUqFGB6ylTprB7927efffdGperJqmUYRKRQcCjWN7E44ESrBl5F1WfaKc/LpeLadOm0aNHD8aMGUNCQtmPY/uhfP76TSYAE0efT+MGVtpjjmNk5u1i+NkXApxiaI7lFzH2wzSOJS6kSQfrBf/f8+/m+m6ja6JKBsNpw87cnWGNj9847czdWauGafr06UyfPr1U3KBBgwKG6bnnnuOLL75g8eLFNGpUv7vmo3VJ9AvgEeASQAA38Cbwf6p6WETWVb+Ipy+bNm0iJyeHgwcPlmuUAHYfLSQ+TrhxSCqDu7QErEW0/7f2BXbk7OChgQ9zxTlXlsqz74SdB/+5GkfKPBq12Em8JPDwoEe4tONlNVYng+F0YUz3snfx6du6X6115U2YMIEJEyaUm2bKlCkBo9S8efNakasmqdAwiUgD4HasqeDdsAzSTuBPwK9UNXiKePjROUNYgvddqohR/drTs30z2jU/6YD1vU3vsiNnB20atjll7dHOwwU8ZFuJt80cGjQ9QIP4hjw9/Bn61fKgrcFgqFkyMjKYPHkyXbt25bLLrH86ExISSEtLi7FkladcwyQizwH3AmdhGaQlwOuq+oUv/sbg9KpaUkNynnZkZ2ezc+dOEhISGDCg7O3NVTUwu+7c1k0C4Ut//Ir/7vnCt4j2KZolNQvEbfgxh8dmLSeh4xySGh4nJbkFU0ZM5dzmXWquQgaDISb06dMH1dNreL+ike/lwAYsozQVuMpvlAxVY+3atagqF1xwQZn9wU6Xm/99bzULNxws9eLtzs1khm8R7b19f0/3Fiddo6zaeYxxsxaT1GkmiQ2P075xB16+dLoxSgaDod5QrmFS1cWqeiUwCDgP2CEiD4tI4/LyGcrH6/VG1I03Y8lOthzI48MVWbg9lmEqdBUwbc0LuLwurjznp/ys01WB9F9uOsSEuQtofO4nxCcVcF6LHrx02Su0bdy2ZitkMBgM1UhEkx9U9QfgVhHpjDXWtENEPgZK7bEgIgNVdX11C3m6sW/fPrKzs0lJSeG88051GQSQtvsEs1f/SHycMOmGC0hMsP6HyC7Kxq0eujTvyr39fh/o5puzdh9vfPsfmnVdiMR5GNJuKOMHP1HupoAGg8FQF4lqVp6q7gEeEpHJwP1AJxH5CJjk23L9b1i+8wzl0KlTJx5//HHy8vLCrmGyF7l5bu5mAO68rAs9258cPzqnWSdeH/kGxZ5ikuOTUVX+8W0Wf0//nGbnfo2I8tNOP+P+fg8QH2f2UTIYDPWPSq2uVNUcVX0ey9PDN8ACEdkAnB9tWSLSS0Q+FZE3fceBFaS/WERWiUihiGwTkfsl1PeOlW6AiGjQr1hE6owzuPbt29OrV6+wca8v2sbhvCJ6tW/G7y6xxobyi0/uONIsuTmtG7VBVXnjv9v4MONDmp2zFBHllh638kD/scYoGQyGekuVPD/4tlF/R0T+CownSsMkIqnAMmC0qq4SkR7AdyIyzNcCC01/HtY09fcAO1ar7c9AAyDUOdSjPpn8swZ+VNUD0chXExQXF5OcnFxm/Modx5i//gBJCXFMuuECEuLjOOo4ysPLHuKKc67gd73vID4uHrfHy4vzN7H8uI3GZ2cgCL/v/weu7jyqzLINBoOhPlAtvvJU1Q28ICKXR5l1GrBPVVf5ytnua3m9AoRzTTAauNLnSBYR+QTYBIwlyDCJSH/goKq+HHVlapi3334bVeXWW2+lTZs2p8Sf1TiJTq0a84uBHTm3TRNcHhfT1rxAgSuffQX7EBGKSzxM/DSNja4PadgqiwRJ5ImhExh2tllGZjAY6j/VYpj8qOpPI00rIo2AG7HGpYJZAzwmIi1UNSckzuY3Sr77OUXkP8D/hqR7FhgiIs2Af6rqdxFXogY5cuQIe/bsITk5uczV2b06NOcf911IYrzVy/repnfYlbuTNo3a8sigx3C6vDz28Qr2JHxIcvPDNIxvwuQRz9KrZe/arIrBYDDUGLH04DkQqwvuWEj4ISz/e6eMNZXRFZeItd4KAJ8xKgYOAHcBK0TkFRGJ+aCLf4p4//79T+nOO1FYHDhvkBhPfJzw1d4vWbTnvyTGJfLk0KfwlCRz/4dfsjf5fRIbH6ZFUiteGfmyMUoGg+G0IpaGqZ3veCIkvMB3PLWfKwQRiQOuwlr8C4Cq5qvqLao6DGtyxjys8aZJkQilqhQVFQHgdrtxOByAtfbIbrcHFro6HA48Hg8ATqeTkhLL6UVxcTHFxZaRKSkpwel0ApbD1rVr1wIwZMgQ7HY7Xq8XgP1Hc7n1rZVM/XwTeYUOXC4XmbmZ/GXDDADu63c/Db3tuPuf8zje/AMSGuRwdsNUpo+cTmrTc7Db7RHJ4q+Xx+M5pV5+WRwOB263G4CioiJcLldA/mC9+OsVqpdQWfxlFRcXB8oKliWcjsuSxV8vl8sVqFewLB6PB7vdHniOwWUFy1JevcqTxV+vYFnKet7lyRKpjqvr3Qt+3tUtS11997xebyDc6/UG7hF8rqqB9P5zf57g8/LyR1JWbchS3fWKNL//vKx3r7LUhT0PHCHX/pZNJLvh3g4sVNXV4SJVdR8wBmvH3XEiErbrUkTuEZE0EUk7ePAgy5YtA6z1Rn438wUFBdhstsAf3syZMzl69CgAixYtYvv27QCsXr2a1astcbZv386iRYsAq7VUUFBA69at6dixIzabjYKCAlSVx//xLXnOEo4XFPPdt8vZuHEjf9v8V1xeF73j+9Cj8UXcbZuNs42N+EQH3Zr2pO+PF5CSZO1m+dFHH3H8+HEAFi5cyK5d1ryRVatWBYzh1q1bWbx4MWDt7zJr1ixLyS4XNpst8BGdM2cOBw5YDdOlS5eyebM1bX39+vWsWLECgKysLObPnw9Y2zrbbLbAC2qz2cjOzgZgwYIFZGZaHtFXrlwZ8N2VkZHBkiVLADh06BCzZ88GrA+bzWYLvNSzZ8/m0KFDACxZsoSMjAwA0tLSWLnS2r4jMzOTBQusTeCys7Ox2Wz+Z4/NZiMnx+oNnj9/PllZWQCsWLGC9eut5XabN29m6dKlABw4cIA5c+YA1kfOZrMFPo6zZs3i8OHDACxevJitW7cClgePVatWAbBr1y4WLlwIwPHjx/noo48A64/ZZrORl2fNrJw3bx579+4FYPny5QHv0Rs3bqyRd+/o0aPMnDkTsD4a/ncPYO7cuezbtw+AZcuWsXHjRsDyaL18udURsXfvXubNmwdAXl4eNpst8OGqq+9ebm5u4ENZWFgYMIZFRUXk5+cHdOF/P7xeL9nZ2YGPc05OTkDf+fn5AcPocDgoLCwErI9xbm4uYBlM/73Behf9OipLFqfTGXgOLpcrUJZfFn+9ypLFbrcHdFdcXBx4v8qTJS8vLyBLQUFB4G+tLFk8Hk8pWbKzswOy5OXlBc7LevcqjarG5AcMw5ox93hI+KO+8EsqyN8VmAskRnCvgb4yz64o7cCBA9XpdKqqaklJidrtdlVV9Xg8WlhYqF6vV1VV7Xa7ut1uVVV1OBzqcrlUVbWoqEiLiopUVdXlcqnD4VBV1Q8++EDHjRunX375pXq9Xi0sLFSPx6P/+WG/Dpv0X73i+SV6ONehTqdTi4uLNb84X99Nf1vXZe7XUTNm6LWfXavXfX6NPrfqeS1yFZWSpbCwMCJZ/PVyu92n1Mvj8QTqVVJSoqoakEVVtbi4uJRe/PUK1UuoLP6yioqKAmUFyxJOx2XJ4q9XcXFxoF7Bsrjdbi0sLFRVLaXjUFnKq1d5svjrFSxLWc+7PFki1XF1vXvBz7u6ZamL796WLVvU4/EEwj0eT+Aewederzcgu//cnyf4vLz8kZRVG7JUd70izZ+RkRH2efufEZCmlbAPojFy/ucbCzoMvK2qjwSFvwbcB7RU1dDWlD/NWcAM4Pd66gSJcOkbY43eH7lDAAAYlUlEQVRlpahquS2xwYMHa3V75S0pKWHy5Mk4nU4mTZpESorV0jmS5+TWP3+HvdjNpBvO55r+pZdZpe/N4YmF75Hc7hsARnW+jnv73WM29zMYymHr1q1lrhE8E/j1r3/NfffdF/A0HgkFBQV06NCBnJwcLr30Umw2G507d64wX0W6FpF1qjo4YkF8xOwLp6r5wKfApSFRA4G55Ril5lh7QI0LNkoicnY5txsGvFuRUaopEhMTmTRpEnfffXfAKKkqz83NwF7s5tKebRjVrz3rjqTx/qZ3cXvdrNh+hCcWvxowSrf1uoP7+t1rjJLBYCiT9evXc/DgwbBGSYPGxEJZt24d/fr1Iz4+nkcffZRnn322pkUtl2qdLl4JpgJpItJTVbeJSB+gN9bYESLyGHAz8HNVPS4iLbC67/4F9Pc5fEgABgB5wJsicgtwDfCsqmaJSD/gTqzt32NGcnJyqf8s/vPDAdbuPkHzRok8cV1vjjqOMj3tZQpLCsnLa8binWtIbr0dIY6xAx7mik4/iaH0BoOhPvDOO+9w6623Bq4nTZrE7t27ycvLY+fOnSxdupTt27czceJEnE4nhYWFPPjggzidToYMGQLAddddx7333kteXl7MNh2MqWFS1V0ichXwvIhkAh2Ay1U1y5ekLdbMuoa+7rjlWN4lQltZbiDVd34Ca4fdDBFZDywA7tAY7RVlt9tJSkoiMTGxVPhPerdj475chndrRdOGwhPLn6ewpJD2Sb35cvfXJLfYTzzJPHPh0wxsOygWohsMhnrGV199xQMPPBC4XrduHQUFBcyfPz9gZAYOHEhaWhodO3ZEVcnNzeWee+5hzJgxgNXDc8EFF/Dtt99y7bXXxqQeMe8XUtXVqjpGVcer6v+o6uaguMdVtY2q7lNVu6peoKoS5peoqod9eb5U1XNVtZGqXqyqL8bKKIE1a+rZZ5/lhx9+KBXeuEECT19/Plf0acc7G/9CZl4mDaUFe3OPkNRsP8nSlFdGvmSMksFgACyvMSISmMUZjn379tGuXbvA9bp163jjjTdKtXw6duzIAw88wMcff4zdbqdFixasWbMm0GICaNeuHfv376+ZikRAzA3T6Yzb7Wb9+vU4nU5atWoFWNtZFLk8gTSL9yziy72LERIoLHaR2OgEzRLa8KcrX6NrSrdYiW4wGOoY6enpJCYm0rt32QvqGzVqFJgCvn//ftxu9yk7ZH///feMHTuWb7/9lu7du7N3717sdjtdu3YNpCkqKqJhw1K7GtUqxjDVIJs3b8Zut9O+fXs6duzI7qOFPPyvdfzunVXYi9xk5mbyzsa/AOBxC/FJdtomn8tbP32Ndo3Lm8thMBjONNLT0+nduzdJSUllpunbty/btm0DrNZScCsIYNu2bcTFxfGTn/yEp556ipKSEtLT0xk8uPTEua1bt9KvX+x2MDKGqQYJ3qXW41Wmfr6JEo/S95wUGjdIoHlSK+JK2qBeIS6hhK5N+vLmT18mJTklxpIbDIbPdn7KxmMbwsZtPLaBz3Z+WmuyeL1eNm3aRP/+/ctNd/PNNwcWeYczTK+99ho9e/ZkwIAB3HTTTfzrX/9i/fr1pXbSzsrKwuv1xtQwxXpW3mlLXl4e27ZtIz4+nkGDBvGPb3ez7WA+7VIaMO6qnhQWlfDInI8panwAERjY8lImXvQoCXHmkRgMdYHuKd15ae00xg+ZQN/WJz/SG49tCITXFjt27MDhcFRomG6//XaGDx9OQUEBU6dOPSX+nXfeOSXs6quvLnU9Y8YMxo8fH9gdOxaYFlMNsXbtWlSVPn36sD/fywff7AbgmdHns/rQau6YPZ3cJvMQgZ92vIFnLx5vjJLBUIfo27of44dM4KW10wItp2CjFGysahr/hIeKDFOTJk3405/+FHAFFi2qSmpqKnfeeWel8lcX5ktYQ/j9jw0YNISpn2/C41V+Oewcjnu28OfN06ApoHBL97u49fwbYiuswXAG8Iu5Py8z7v7+DwQ22fzvnoXMSH+rVPzElU8Fzp+76IWAUXr467Fk5oU3Aj/rdBUPDBgLwK7cnXRL6V5p2SM1TACXXx7ttngnERHGjh1b6fzVhTFMNcTtt9/O/v372Xhc2H30AKktGzG4h/LKxpdAABX+0O9xruoSudsQg8EQe2qzpeQnPT2dTp06BTzHnO4Yw1RDiAipqal07KgkJsST7TzOyxueQuK8oHFMHDqFoR1O2XLKYDDUEPNGL4go3dWdRwVaT3Cy+27UudewMOsLNh7bEDBOr13+ZkRlVqW1BJZhGj48/A7VOTk53HbbbezYsYOGDRvStm1bZsyYQbdu9Xe5iRljqmZcLldgSwKwDFR8Yj7/+nESElcCKjw3YpoxSgZDPSB4TOl/et12yphTbXD48GGOHDlSZjeeiDBu3Dh27NjBhg0buPbaa7nrrrtqTb6awLSYqpkNGzbw0UcfkdxjJHff8BPSD+/kbzufIy7BBQoThz1L37Z9Yi2mwWCogHATHYInRNTWBAj/+FJhYWFgjy4/ycnJjBo1iiuvvDIQNmLECF599dUal6smMYapmlmzZg158Wex5VADVts+pXGnfxOXYHlEurfv/QxtP6SCEgwGQ11gZ+7OsMbHb5x25u6sVcM0ffp0pk+fXipu0KBBjBo1qlTY66+/zvXXX1/jctUkxjBVI8ePH2fbrj1kNrmI5JSdNOm0CInz0jb+Au4aNJph7cP3ERsMhrrHmO43lRnXt3W/WpsEMWHCBCZMiGzN1JQpU9i9ezfvvvtuDUtVsxjDVAV++cEL9D6rG5NH/xKw1i7tadCTuDY7aNLxG0Sge8NLeelnjxEv8RWUZjAYDJXnueee44svvmDx4sU0atQo1uJUCTP5oQr0PqsbaSWzmDz3E7xeL/NX78DR6UeaplpGKc7bkCcuvt0YJYPBUKNMmTKF+fPns3jx4pjtoVSdmBZTFZg8+pdMngtpJbMY92Euu1J/pHHLbaiCCBBXzCHHYdo0bhtrUQ0Gw2lKRkYGkydPpmvXroGdaxMSEkhLS4uxZJXHGKYq8NnOT7nxoh64V97EhqYf0SDOiyrEEY/ioX+bAfSLwWI8g8Fw5tCnTx9UNdZiVCumK68K+J083jyiF96i1lagJxkVDwlxCdzQbUxsBTQYDIZ6iDFMVcA/bXTKd1ORpGwoboEkFAPChCFP0a+NaS0ZDAZDtBjDVEXmrNyOy+0lIQFIzgHA647ni9V7YyuYwWAw1FOMYaoCk+d+QlrJLAYn3MpN51nddsPPvpDBCf8TmK1nMBgMhugwhqkKbMnexeDEX3HjRT34754v+FWPW9hyIoMbL+rB4MRfsSV7V6xFNBgMhnqHmZVXBT6586lT/Gld0KqvdX3RBCa3/mWsRTQYzkhUNaY7sJ4J1ORMQNNiqgIVOXmsTQ/EBoPBIjExEafTGWsxTnucTifJyck1UrYxTFUgEiePBoOhdmnTpg0HDhzA4XCcdut7Yo2qUlJSQnZ2Nvv376dly5Y1ch8xD640gwcP1vq8YtpgMEB+fj5Hjx6lpKQk1qKcdiQkJNCgQQNat25NgwYNyk0rIutUdXDU96i0dAaDwVBHadasGc2aNYu1GIZKYrryDAaDwVCniHmLSUR6AX8EDgLtgRdUdX0FeZ4E+gDFwFFgoqp6ok1jMBgMhrpHTA2TiKQCy4DRqrpKRHoA34nIMFUNuwhIRKYB/VX1at/1h8CbwB+iSWMwGAyGukmsu/KmAftUdRWAqm4HNgCvhEssIj2Bx4C3goLfA+4XkQGRpjEYDAZD3SVmhklEGgE3AmtCotYA14pIizDZfg3Eh+RJA7zAbVGkMRgMBkMdJZYtpoFAA+BYSPghLMMyMEyeiwAFTvgDVNUB5ANDokhTJqpKUVERAG63G4fDAYDX68VutwfWRTgcDjwea8jK6XQGpqUWFxdTXFwMQElJSWChn8fjCZSlqtjtdrxeb6Ast9sNQFFRES6XCwCXyxWRLHa7PSJZ/GUFy+IvK1pZ/PWqSBZ/WcXFxYGygmUJV6+yZPHXy+VyBeoVLIvH48Fut4fVcbAs0eg4WBZ/vYJlKe95lyVLVZ+3effMu1fR864r715lieUYUzvf8URIeIHv2KaMPLlhJjEUBKWPJE0pROQe4B7fZWHDhg23+86bA3khyUPDgq9bAcfD3aOaCCdPdearKF1Z8ZGGV3Rt9Bdd2Jmkv2jjzN9uZPE1/bfbqRyZykZVY/IDbsJq2dwVEn63L3xMmDybgf1hwg8AmyJNE4WM71YUFnwNpNWwzk6RpzrzVZSurPhIwyO4NvqLIuxM0l+0ceZvt/LvXkW6KuO6WvUXy668fb5j6FiSf1Xc0TLyhBt7ahaUPpI0kTI/grBwaWqKyt4r0nwVpSsrPtLwWOquKveLhf7q2rtXlftVh/6ijatr+qtP71648FrVXcxcEolIM+Aw8LaqPhIU/hpwH9BSrbGh4DzPA08BzVU13xfWHMgFpqrqs5GkqcE6pWkl3G8YLIz+qobRX+Uxuqsa1a2/mLWYfEbjU+DSkKiBwNxQo+TjA6xuvktD0nuBWVGkqSnereHyT3eM/qqG0V/lMbqrGtWqv5g6cRWRblhTuYer6jYR6YO14HaoqmaJyGPAzcDPVfW4L89fsFpDt/quPwYOqerDQeVWmMZgMBgMdZOYexcXkWHAeCAT6AC8qKqbfXEvA78DBqnqPl9YAtbC3KaAAHt8eTSozArTGAwGg6FuEnPDZDAYDAZDMLF2SXTGICKdReQ1EZkoIvfGWp76hog0EJFHRGR2rGWpj4jIAyJyTET2iMgNsZanviEid4jIFhHZKyI/jbU89RERuVRElkaS1himWkBEBPgYeFlVnwNG+ryqGyKnKZb3jtaxFqS+4Ru7jQdSgWcAm4i0iq1U9QcR6YDl07M38CrwUoxFqneISDIwmQhtjjFMtcNgoImqHvRdfwOMjaE89Q5VPYY1DmmIniJVfUNVi1T1Q2AX0CXWQtUXVPWAqi7xXS7DvIeVYSzwt0gTG8NUOwzB8gHo5yDWXlGG6DADopVAVUM/pF4gIxaynAaMAczs3igQkd5Y7ov2VZTWjzFMtUMKkB107QLaxkgWwxmMiAwC5qiqPday1CfE4nGsPd0+iLU89QXfMMYdwPvR5DOGqXY4geVJ3U9jKufQ0WCoNCISB/wKeDHWstQ31OJl4AJgsIj0jbVM9YQ7gL+rqjeaTMYw1Q7rsQae/aRiLSw2GGqT+4HXVbUk1oLUV3zjxMsBd6xlqSfcDnwlIoeBOcAIEVlbUSZjmCJEREaKyCIR+V0Z8b1E5FMRedN3DOwnpaprAa+I+LvvhhNl07a+UxX9BSerYTHrLFXVny/fN6p6UESSRORntSJ4HaEq+vPpq6nvPB6wA9tqR/LYU8Vv36Wq2k5V22FtDPudqla4L16NuXk/nX7AKOAfWIPvt4eJTwWOABf6rntgdd91C0rTG3gba9v3U8o4nX/VpL9WWNN0DwDDYl2n+qQ/rK1k3ECR7+cBfhPretUj/V2LNWHpLWAi0CnWdaovugtJOxJYFsl9jeeHCPHNLMkA7lDVv4fE2YAeGuRd17eQLF9VR9eqoHUUo7+qYfRXNYz+Kk8sdGe68iIn7F7BItIIq4m6JiRqDXCtiITbG+pMxOivahj9VQ2jv8pT67ozhilyympaDsSacXcsJPwQ1mr7cGMlZyJGf1XD6K9qGP1VnlrXnTFMVaed73giJLzAd2xTi7LUR4z+qobRX9Uw+qs8NaY7Y5iqj9CNDeN9R1dtC1JPMfqrGkZ/VcPor/JUu+6MYao6fjcbof2pzXzHo7UoS33E6K9qGP1VDaO/ylNjujOGqepsxRocPDskvCPW1Nx1tS5R/cLor2oY/VUNo7/KU2O6M4apiqhqPvApcGlI1EBgrqqGNnMNQRj9VQ2jv6ph9Fd5alJ3xjBFTqLvmBAmbirQTUR6QmD/m97AU7UkW33A6K9qGP1VDaO/ylPrugt3I0MIInIhcKfv8rciclRV5/njVXWXiFwFPC8imUAH4HJVzYqBuHUOo7+qYfRXNYz+Kk+sdGc8PxgMBoOhTmG68gwGg8FQpzCGyWAwGAx1CmOYDAaDwVCnMIbJYDAYDHUKY5gMBoPBUKcwhslgMBgMdQpjmAwGg8FQpzCGyWAwGAx1CmOYDIYgROR7Efku1nLEChFJEZHhIhInIleJyGci8mUN3u8yETEeaAylMIbJcFoiIreLiIrINhH5RkSWiUiRiLh958tEJMOX5vWgrNlAbqzk9iMiI32ypQfJ2zyCfA8Fpc8VkWVR3PMyYDywActjdEtgNCd9pdUEO4G3RaRTDd7DUM8wLokMpyUicjvQWlVfDgrbA6SoakpQ2PXASFV9uNaFLAcRGQl8jeV3bFkly1gGoKojI0h7GfAH4BZV9QaFHwR2RFJGZRGRlsB/gBtV9VBN3cdQfzBNaMPpih34qKJEqvrvSFoipzMi0hqwAZcFGyUfNb6Dq6qeEJHpwGcicnEYGQxnGKYrz3BaoqqzVTWij6qq/rOm5anjTAG2qGpmDGX4HOgC3BpDGQx1BGOYDAZAROJF5FoR+dw/2C8izUXkQd94zTYROVtEpojIdyJyTEQmiEiCiDwsIp+IyFERmSsizULK7isiH4vIUhE5LiKzRKRNFWS9wifTShEp8I1FnV/JsloA/wssjCKPiMgDvvp8LyJZIvJqaL19af/Xp68VIpInIodEZJOIfC0iF/vTqaoH+AZ4sjL1MJxmqKr5md8Z8QP2ALllxLUGRgElwLKQuEysSRG/BOJ8YZ8BCkwD2vrCegBe4I9BefsDGUAH3/UgoBBYiW+Mtwx5RvrKHxkSngIcxBo/A2uCwhbg/DBlLAutS5g0v/Pd56fl6CxUH38BvgQa+K6HYE0YWQMkB6V7BKtLtYvvejDgBnaVca9JPlm6xvpdMb/Y/kyLyWAAVPWYqi4EjoaJPgjkq+onenL8wz+F+r+qesRXxnbgCJbx8fMKMENVD/jSrAP+C4wArqiEqN2BVlgGClU9AUzGMoiVYZjvGNHGbr6N4+4DpqpqkU+GtcAfsQzUQ0HJHwIyVHW3L10aVquoq4i0ClO8vyvxsmgrYTi9MIbJYChNSZgwT5iwojLyFwMNAUSkAdZH9p6gKdzLgJ7AXqBjJeTb7Mub5utWbOUzmFsqURbAOb5jfoTpf+M77ggJ/8R3HB0U1gxoF5IuG6sVlROm7DzfsX2EshhOU4xhMhiqH/EdW2LNfH1ZVUcG/c5X1c6q+vdoC1ZVJ3AhMBN4CvhRRJ6rwiJV/xqlSFtc5/qOTULk2oc1g69FUPDfgVQR+QOAiJwLXA48r9aYUihO3/GUsSrDmYUxTAZDzZGHNWYyJjTC51mhUhMWVPW4qt6HNaa1GHgaeLn8XGVyzHdsHGH6/b7jeWHi3FgLZv2Mx5qGfqWIrARmAONU9cUyym7oO8Z8gbMhthjDZDiTiKMW33lVLQRWA6NF5HX/rDURaQi8BjSKtkwRGexbFIyq7lbV0cD3VG68Ck52ybWNMP1nvuPvQuRqj1WffwUFXwacUNUbVPUiVR2lqsHxofhbW1sjlMVwmmIMk+GMQERSgDZAUxEJ+xEWkWTgLKCdiMT5wuKwxjxa+OL9+MdBzg7K3xCrG6qNiPi78x7D6uJ6CMgRkb3AceAsVV1Tyeq8LiLn+e6ZgNVduLSSZf3Xd+wWGuGrT0ugvYjEA6jqIizjdLPfQPp0NBmYp6qfBBXxR+ABEdntm26fISLrRWS2iAwNI0sPrFbXt5Wsi+E0wRgmw2mPiLwBbAf8hmWziPwjJE03rIkFTbE+kFtEZIgvrDuWwdkiIheLyN+wPsQA74vIWyLyE6z/9FsAvXxpz1LVlVgth2+wDFRD4B3gripUqTOQISLrsaadf43VbVYZ1gLbsGbUBQjSRxOs+m8RkR6+6FuxjM7rIrIaqzsxi1O7LKdgdWc2xxqb6g0MAG4CvhSRs0LSDwb+o6rHK1kXw2mC8ZVnMNRBatNXnoj8EmtCQvfK3KeMMhOAT4FnVXVDUHgSVivz78AT/lajr5vzMHCRqv5QXXIY6iemxWQwnOH4ut/W+wxUdTERaBVslHz3cqnqXqyW2sagqHHAa8YoGcAYJoPBYHEH8CsRqa5WU0ugn4iM8Y9PgdWSEpFbgLn+Bbq+8aZU4JlqurehnmO68gyGOkhQV94GTk6fvl5V88rMZOV7CLjBd9kfSK+oKy8obxLwIPCNz0tDpfF15d2G5YevFdaC2h+xxrPe9617QkSuwlrk+76aj5HBhzFMBoOhFCKSrKrFtXSvBv6Wk8Hgxxgmg8FgMNQpzBiTwWAwGOoUxjAZDAaDoU5hDJPBYDAY6hTGMBkMBoOhTmEMk8FgMBjqFP8PmjFXiqoRcIYAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plots.e2e_time import build_fig_e2e_mnist_time\n",
    "from functools import reduce\n",
    "\n",
    "df_u = _preprocess(_select_df(observer_data, \"MNIST plain None\"), \"u\")\n",
    "# display(df_u)\n",
    "# display(observer_data)\n",
    "df_l8 = _preprocess(_select_df(observer_data, \"MNIST linf unoptimized\"), \"l8\")\n",
    "df_l8p = _preprocess(_select_df(observer_data, \"MNIST linf_opt optimized\"), \"l8p\")\n",
    "df_l2 = _preprocess(_select_df(observer_data, \"MNIST l2 unoptimized\"), \"l2\")\n",
    "df_l2rst = _preprocess(_select_df(observer_data, \"MNIST l2_opt optimized\"), \"l2rst\")\n",
    "#\n",
    "data_frames = [df_u, df_l8, df_l8p, df_l2, df_l2rst]\n",
    "df = reduce(lambda  left, right: pd.merge(left, right, on=['round'], how='outer'), data_frames)\n",
    "fig_e2e_mnist_time = build_fig_e2e_mnist_time(df)\n",
    "display(fig_e2e_mnist_time)\n",
    "\n",
    "# display(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "### Timing plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "from matplotlib.backends.backend_pdf import PdfPages\n",
    "\n",
    "# display(data_server)\n",
    "\n",
    "from matplotlib import cm\n",
    "\n",
    "records = [\n",
    "    \"MNIST linf_opt optimized\",\n",
    "    \"CIFAR Small linf_opt optimized\",\n",
    "    \"CIFAR Large linf_opt optimized\",\n",
    "    \"Shakespeare linf_opt optimized\"\n",
    "]\n",
    "metadata = {\n",
    "           \"CIFAR Large linf_opt optimized\": { \"label\": \"CIFAR-10 L\" },\n",
    "           \"CIFAR Small linf_opt optimized\": { \"label\": \"CIFAR-10 S\" },\n",
    "           \"MNIST linf_opt optimized\": { \"label\": \"MNIST\" },\n",
    "           \"Shakespeare linf_opt optimized\": { \"label\": \"Shakespeare\" },\n",
    "    }\n",
    "\n",
    "columns_server = [\n",
    "    # 'start',\n",
    "                  \"agg_done\", \"param_extraction_done\"] #, \"verification_complete\"]\n",
    "columns_client = [\"model_download\", \"model_training\",\n",
    "                       \"model_crypto\", \"model_upload\"]\n",
    "metadata_columns = {\n",
    "    'start': \"Training time\",\n",
    "    'agg_done': \"Reconstruction\",\n",
    "    'param_extraction_done': 'Verification time',\n",
    "    'model_download': 'Model download',\n",
    "    'model_training': 'Training time',\n",
    "    'model_crypto': 'Proof creation',\n",
    "    'model_upload': 'Model upload'\n",
    "}\n",
    "\n",
    "client_colors = [cm.get_cmap(\"Blues\")(col) for col in np.linspace(0.4, 0.8, len(columns_client))]\n",
    "server_colors = [cm.get_cmap(\"Greens\")(col) for col in np.linspace(0.4, 0.7, len(columns_server))]\n",
    "bar_colors = client_colors + server_colors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "client_colors = [cm.get_cmap(\"Blues\")(col) for col in np.linspace(0.0, 0.4, len(columns_client))]\n",
    "server_colors = [cm.get_cmap(\"Greens\")(col) for col in np.linspace(0.1, 0.4, len(columns_server))]\n",
    "bar_colors_non_current = client_colors + server_colors\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "df1: \n"
     ]
    },
    {
     "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>training</th>\n",
       "      <th>createproof</th>\n",
       "      <th>verifyproof</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>idf</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MNIST linf_opt optimized</th>\n",
       "      <td>5670.0</td>\n",
       "      <td>5134.981132</td>\n",
       "      <td>3960.496855</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Small linf_opt optimized</th>\n",
       "      <td>5529.0</td>\n",
       "      <td>8449.595960</td>\n",
       "      <td>15210.828283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Large linf_opt optimized</th>\n",
       "      <td>34576.0</td>\n",
       "      <td>26266.603774</td>\n",
       "      <td>41358.389937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shakespeare linf_opt optimized</th>\n",
       "      <td>864659.0</td>\n",
       "      <td>68236.578947</td>\n",
       "      <td>106771.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                training   createproof    verifyproof\n",
       "idf                                                                  \n",
       "MNIST linf_opt optimized          5670.0   5134.981132    3960.496855\n",
       "CIFAR Small linf_opt optimized    5529.0   8449.595960   15210.828283\n",
       "CIFAR Large linf_opt optimized   34576.0  26266.603774   41358.389937\n",
       "Shakespeare linf_opt optimized  864659.0  68236.578947  106771.000000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "idf\n",
       "MNIST linf_opt optimized         -0.731773\n",
       "CIFAR Small linf_opt optimized   -0.478892\n",
       "CIFAR Large linf_opt optimized   -0.595323\n",
       "Shakespeare linf_opt optimized   -0.897303\n",
       "dtype: float64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'The above is not valid for shakespeare, because verifyproof only is not dictating the new time!'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "df1_scaled: \n"
     ]
    },
    {
     "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>training</th>\n",
       "      <th>createproof</th>\n",
       "      <th>verifyproof</th>\n",
       "      <th>scale</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>idf</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MNIST linf_opt optimized</th>\n",
       "      <td>22.174025</td>\n",
       "      <td>20.081693</td>\n",
       "      <td>15.488564</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Small linf_opt optimized</th>\n",
       "      <td>12.116368</td>\n",
       "      <td>18.516625</td>\n",
       "      <td>33.333333</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Large linf_opt optimized</th>\n",
       "      <td>21.206598</td>\n",
       "      <td>16.110172</td>\n",
       "      <td>25.366461</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shakespeare linf_opt optimized</th>\n",
       "      <td>43.834309</td>\n",
       "      <td>3.459287</td>\n",
       "      <td>5.412808</td>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 training  createproof  verifyproof  scale\n",
       "idf                                                                       \n",
       "MNIST linf_opt optimized        22.174025    20.081693    15.488564  100.0\n",
       "CIFAR Small linf_opt optimized  12.116368    18.516625    33.333333  100.0\n",
       "CIFAR Large linf_opt optimized  21.206598    16.110172    25.366461  100.0\n",
       "Shakespeare linf_opt optimized  43.834309     3.459287     5.412808  100.0"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# transform data into df1 (which contains the required format\n",
    "\n",
    "data_df_server = data_server.groupby(\"idf\")[columns_server].mean()\n",
    "data_df_client = client_data.groupby(\"idf\")[columns_client].max()\n",
    "df1 = data_df_client.loc[records]\n",
    "df1_server = data_df_server.loc[records]\n",
    "\n",
    "# combine model download and training into training time (i.e., what it would be on plaintext)\n",
    "df1[\"training\"] = df1[\"model_download\"] + df1[\"model_training\"]\n",
    "\n",
    "# sum creating proofs (well formedness + range proof), model upload time + discrete log solve time\n",
    "df1[\"createproof\"] = df1[\"model_crypto\"] + df1[\"model_upload\"] + df1_server[\"agg_done\"]\n",
    "\n",
    "# overhead for verifying proofs\n",
    "df1[\"verifyproof\"] = df1_server[\"param_extraction_done\"]\n",
    "\n",
    "df1 = df1[[\"training\", \"createproof\", \"verifyproof\"]]\n",
    "\n",
    "print(f\"df1: \")\n",
    "display(df1)\n",
    "\n",
    "total_sum = (df1[\"verifyproof\"] - df1.sum(axis=1)) / df1.sum(axis=1)\n",
    "display(total_sum)\n",
    "display(\"The above is not valid for shakespeare, because verifyproof only is not dictating the new time!\")\n",
    "\n",
    "# scale the numbers such that 3* verify == 100%\n",
    "df1_scaled = pd.DataFrame()\n",
    "\n",
    "# scale either by verifyproof or by sum of training and createproof\n",
    "max_series = pd.concat([df1[\"verifyproof\"], df1[\"training\"] + df1[\"createproof\"]], axis=1).max(axis=1)\n",
    "\n",
    "def determine_scale(row):\n",
    "    if row[\"training\"] + row[\"createproof\"] > row[\"verifyproof\"]:\n",
    "        base = 2 * (row[\"training\"] + row[\"createproof\"])\n",
    "        return (base + row[\"verifyproof\"])\n",
    "    else:\n",
    "        return row[\"verifyproof\"] * 3\n",
    "\n",
    "# scale each column such that 3*max_series column is 100%\n",
    "df1[\"scale\"] = df1.apply(determine_scale, axis=1)\n",
    "\n",
    "for col in df1.columns.values: \n",
    "    df1_scaled[col] = df1[col] * 100 / df1[\"scale\"]\n",
    "print(f\"\\ndf1_scaled: \")\n",
    "display(df1_scaled)\n",
    "\n",
    "\n",
    "# mark for which phase (create, train, verify) we have data from a particular round\n",
    "verifyproof_rounds = [\"t-1\", \"t\", \"t+1\"]\n",
    "createproof_rounds = [\"t\", \"t+1\", \"t+2\"]\n",
    "training_rounds = [\"t\", \"t+1\", \"t+2\"]\n",
    "\n",
    "per_round_offset = 100.0 / 3.0\n",
    "\n",
    "with PdfPages(f\"plots_output/e2e_operation_times_optimistic_cont_nku.pdf\") as pdf:\n",
    "    fig, ax = plt.subplots(figsize=(7,4))\n",
    "    y_pos = np.arange(len(records))\n",
    "\n",
    "    bar_height = 0.3\n",
    "\n",
    "    from common import get_colorful_styles\n",
    "\n",
    "    colors, linestyles = get_colorful_styles()\n",
    "\n",
    "    color_d = {\n",
    "        \"t-1\": (0.66, 0.66, 0.66, 1),\n",
    "        \"t\": colors[2],\n",
    "        \"t+1\": colors[1],\n",
    "        \"t+2\": (0.44, 0.44, 0.44, 1),\n",
    "    }\n",
    "\n",
    "\n",
    "    bars = []\n",
    "\n",
    "    for i, round_lbl in enumerate([\"t-1\", \"t\", \"t+1\", \"t+2\"]):\n",
    "        \n",
    "        per_round_offset = pd.concat([df1_scaled[\"verifyproof\"], df1_scaled[\"training\"] + df1_scaled[\"createproof\"]], axis=1).max(axis=1)\n",
    "        \n",
    "#         if i is 0:\n",
    "#             current_round_offset = 0\n",
    "#         else:\n",
    "#             max_series = pd.concat([df1[\"verifyproof\"], df1[\"training\"] + df1[\"createproof\"]], axis=1).max(axis=1)\n",
    "\n",
    "#             current_round_offset = df1_scaled[\"verifyproof\"]\n",
    "    \n",
    "        if round_lbl in verifyproof_rounds:\n",
    "            # add the horizontal bars for verify (3rd lines)\n",
    "            rects = ax.barh(y_pos - bar_height, left=i * per_round_offset, width=df1_scaled[\"verifyproof\"], height=bar_height, color=color_d[round_lbl], label=f\"{round_lbl}\", edgecolor=\"black\")\n",
    "    \n",
    "            if round_lbl == \"t\":\n",
    "                bars += [(bar, \"verify\") for bar in rects]\n",
    "            \n",
    "    \n",
    "        if round_lbl in training_rounds:\n",
    "            # add the horizontal bars for train model (1st lines)\n",
    "            label = f\"{round_lbl}\" if round_lbl not in verifyproof_rounds else None\n",
    "            left = (i-1) * per_round_offset\n",
    "            rects = ax.barh(y_pos + bar_height, left=left, width=df1_scaled[\"training\"], height=bar_height, color=color_d[round_lbl], label=label, edgecolor=\"black\")\n",
    "        \n",
    "            if round_lbl == \"t\":\n",
    "                bars += [(bar, \"train\") for bar in rects]\n",
    "    \n",
    "        if round_lbl in createproof_rounds:\n",
    "            # add the horizontal bars for create proofs (2nd lines)\n",
    "            left = (i-1) * per_round_offset + df1_scaled[\"training\"]\n",
    "            rects = ax.barh(y_pos, left=left, width=df1_scaled[\"createproof\"], height=bar_height, color=color_d[round_lbl], label=None, edgecolor=\"black\")\n",
    "        \n",
    "            if round_lbl == \"t\":\n",
    "                bars += [(bar, \"create\") for bar in rects]\n",
    "\n",
    "            \n",
    "    # set minor tick labels: which phase it is (train, create, verify)\n",
    "    ax.set_yticks([bar.get_y() + bar.get_height() / 2 + 0.0001 for bar, _ in bars], minor=True)        \n",
    "    ax.set_yticklabels([label for _, label in bars], minor=True, va=\"center\")\n",
    "    for tick in ax.yaxis.get_minor_ticks():\n",
    "        tick.label.set_fontsize(16)\n",
    "\n",
    "    # set major tick labels: which model is used\n",
    "    major_label_d = {\n",
    "        \"CIFAR Large linf_opt optimized\": \"C10-L\",\n",
    "        \"CIFAR Small linf_opt optimized\": \"C10-S\",\n",
    "        \"MNIST linf_opt optimized\": \"FMN\",\n",
    "        \"Shakespeare linf_opt optimized\": \"SHK\",\n",
    "    }\n",
    "\n",
    "    # the pad controls the distance between modeles (e.g., C10-L) and phase names (e.g., train)\n",
    "    ax.tick_params(axis='y', which='major', pad=45, size=0)\n",
    "    ax.set_yticks(y_pos)\n",
    "    ax.set_yticklabels([major_label_d[key]for key in records], rotation = 90, va=\"center\")\n",
    "    for tick in ax.yaxis.get_major_ticks():\n",
    "        tick.label.set_fontsize(16) \n",
    "\n",
    "    ax.set_xlabel(\"Round t and t+1 Progress [%]\")\n",
    "    ax.set_xlim(0, 100)\n",
    "\n",
    "    #ax.legend(ncol=4, bbox_to_anchor=(-0.48, 1), loc='lower left', fontsize='small')\n",
    "    \n",
    "    ax.legend(ncol=1, bbox_to_anchor=(1, 0), loc='lower left', fontsize='small',  handlelength=0.7)\n",
    "    \n",
    "    pdf.savefig(bbox_inches='tight', pad_inches=0)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "idf\n",
       "MNIST linf_opt optimized          42.255718\n",
       "CIFAR Small linf_opt optimized    33.333333\n",
       "CIFAR Large linf_opt optimized    37.316770\n",
       "Shakespeare linf_opt optimized    47.293596\n",
       "Name: new_round_time, dtype: float64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>new_round_time</th>\n",
       "      <th>old_round_time</th>\n",
       "      <th>percent_reduction</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>idf</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MNIST linf_opt optimized</th>\n",
       "      <td>42.255718</td>\n",
       "      <td>57.744282</td>\n",
       "      <td>26.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Small linf_opt optimized</th>\n",
       "      <td>33.333333</td>\n",
       "      <td>63.966326</td>\n",
       "      <td>47.89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CIFAR Large linf_opt optimized</th>\n",
       "      <td>37.316770</td>\n",
       "      <td>62.683230</td>\n",
       "      <td>40.47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Shakespeare linf_opt optimized</th>\n",
       "      <td>47.293596</td>\n",
       "      <td>52.706404</td>\n",
       "      <td>10.27</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                new_round_time  old_round_time  \\\n",
       "idf                                                              \n",
       "MNIST linf_opt optimized             42.255718       57.744282   \n",
       "CIFAR Small linf_opt optimized       33.333333       63.966326   \n",
       "CIFAR Large linf_opt optimized       37.316770       62.683230   \n",
       "Shakespeare linf_opt optimized       47.293596       52.706404   \n",
       "\n",
       "                                percent_reduction  \n",
       "idf                                                \n",
       "MNIST linf_opt optimized                    26.82  \n",
       "CIFAR Small linf_opt optimized              47.89  \n",
       "CIFAR Large linf_opt optimized              40.47  \n",
       "Shakespeare linf_opt optimized              10.27  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "savings = pd.DataFrame()\n",
    "savings[\"new_round_time\"] = pd.concat([df1_scaled[\"verifyproof\"], df1_scaled[\"training\"] + df1_scaled[\"createproof\"]], axis=1).max(axis=1)\n",
    "display(savings[\"new_round_time\"])\n",
    "savings[\"old_round_time\"] = df1_scaled[\"training\"] + df1_scaled[\"createproof\"] + df1_scaled[\"verifyproof\"]\n",
    "savings[\"percent_reduction\"] = round(-100 * (savings[\"new_round_time\"] - savings[\"old_round_time\"]) / savings[\"old_round_time\"], 2)\n",
    "display(savings)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Probabilistic Checking"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e753716b5db94307b36d924a301772c6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(HTML(value=''), FloatProgress(value=1.0, bar_style='info', layout=Layout(width='20px'), max=1.0…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0a3a1bd49b344e749782587d77b51b2d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(HTML(value=''), FloatProgress(value=1.0, bar_style='info', layout=Layout(width='20px'), max=1.0…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8a20a06ebdc5461388ae629cba8e714c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(HTML(value=''), FloatProgress(value=0.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4ce12e20d4e94ac386d6d0ef8cef3aba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(HTML(value=''), FloatProgress(value=0.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a9922450d552452e817225ccea74c70d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(HTML(value=''), FloatProgress(value=0.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "Probabilistic Checking:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "fail_prob_bound=1e-08 \t prob_weights_violating_bound=0.1\n",
      "(n_weights, n_checks): [(1000, 160), (6290, 173), (11580, 174), (16870, 174), (22160, 175), (27450, 175), (32740, 175), (38030, 175), (43320, 175), (48610, 175), (53900, 175), (59190, 175), (64480, 175), (69770, 175), (75060, 175), (80350, 175), (85640, 175), (90930, 175), (96220, 175), (101510, 175), (106800, 175), (112090, 175), (117380, 175), (122670, 175), (127960, 175), (133250, 175), (138540, 175), (143830, 175), (149120, 175), (154410, 175), (159700, 175), (164990, 175), (170280, 175), (175570, 175), (180860, 175), (186150, 175), (191440, 175), (196730, 175), (202020, 175), (207310, 175), (212600, 175), (217890, 175), (223180, 175), (228470, 175), (233760, 175), (239050, 175), (244340, 175), (249630, 175), (254920, 175), (260210, 175), (265500, 175), (270790, 175), (276080, 175), (281370, 175), (286660, 175), (291950, 175), (297240, 175), (302530, 175), (307820, 175), (313110, 175), (318400, 175), (323690, 175), (328980, 175), (334270, 175), (339560, 175), (344850, 175), (350140, 175), (355430, 175), (360720, 175), (366010, 175), (371300, 175), (376590, 175), (381880, 175), (387170, 175), (392460, 175), (397750, 175), (403040, 175), (408330, 175), (413620, 175), (418910, 175), (424200, 175), (429490, 175), (434780, 175), (440070, 175), (445360, 175), (450650, 175), (455940, 175), (461230, 175), (466520, 175), (471810, 175), (477100, 175), (482390, 175), (487680, 175), (492970, 175), (498260, 175), (503550, 175), (508840, 175), (514130, 175), (519420, 175), (524710, 175)]\n",
      "====================\n",
      "\n",
      "fail_prob_bound=1e-08 \t prob_weights_violating_bound=0.01\n",
      "(n_weights, n_checks): [(1000, 838), (6290, 1609), (11580, 1706), (16870, 1744), (22160, 1764), (27450, 1776), (32740, 1785), (38030, 1791), (43320, 1796), (48610, 1799), (53900, 1802), (59190, 1808), (64480, 1810), (69770, 1811), (75060, 1812), (80350, 1814), (85640, 1815), (90930, 1816), (96220, 1816), (101510, 1817), (106800, 1818), (112090, 1820), (117380, 1820), (122670, 1821), (127960, 1821), (133250, 1821), (138540, 1822), (143830, 1822), (149120, 1822), (154410, 1823), (159700, 1823), (164990, 1824), (170280, 1824), (175570, 1824), (180860, 1825), (186150, 1825), (191440, 1825), (196730, 1825), (202020, 1825), (207310, 1825), (212600, 1825), (217890, 1826), (223180, 1826), (228470, 1827), (233760, 1827), (239050, 1827), (244340, 1827), (249630, 1827), (254920, 1827), (260210, 1827), (265500, 1827), (270790, 1828), (276080, 1828), (281370, 1828), (286660, 1828), (291950, 1828), (297240, 1828), (302530, 1828), (307820, 1828), (313110, 1828), (318400, 1828), (323690, 1829), (328980, 1829), (334270, 1829), (339560, 1829), (344850, 1829), (350140, 1829), (355430, 1829), (360720, 1829), (366010, 1829), (371300, 1829), (376590, 1829), (381880, 1829), (387170, 1829), (392460, 1829), (397750, 1829), (403040, 1829), (408330, 1829), (413620, 1829), (418910, 1829), (424200, 1829), (429490, 1830), (434780, 1830), (440070, 1830), (445360, 1830), (450650, 1830), (455940, 1830), (461230, 1830), (466520, 1830), (471810, 1830), (477100, 1830), (482390, 1830), (487680, 1830), (492970, 1830), (498260, 1830), (503550, 1830), (508840, 1830), (514130, 1830), (519420, 1830), (524710, 1830)]\n",
      "====================\n",
      "\n",
      "fail_prob_bound=1e-08 \t prob_weights_violating_bound=0.001\n",
      "(n_weights, n_checks): [(1000, 1000), (6290, 5996), (11580, 9407), (16870, 11531), (22160, 12562), (27450, 13569), (32740, 14323), (38030, 14603), (43320, 15088), (48610, 15485), (53900, 15817), (59190, 15866), (64480, 16119), (69770, 16340), (75060, 16338), (80350, 16518), (85640, 16678), (90930, 16822), (96220, 16792), (101510, 16916), (106800, 17029), (112090, 16991), (117380, 17091), (122670, 17183), (127960, 17269), (133250, 17226), (138540, 17304), (143830, 17376), (149120, 17333), (154410, 17400), (159700, 17462), (164990, 17521), (170280, 17478), (175570, 17533), (180860, 17585), (186150, 17544), (191440, 17593), (196730, 17639), (202020, 17599), (207310, 17643), (212600, 17685), (217890, 17725), (223180, 17686), (228470, 17724), (233760, 17761), (239050, 17724), (244340, 17759), (249630, 17792), (254920, 17825), (260210, 17789), (265500, 17820), (270790, 17850), (276080, 17816), (281370, 17845), (286660, 17873), (291950, 17900), (297240, 17867), (302530, 17893), (307820, 17918), (313110, 17887), (318400, 17911), (323690, 17935), (328980, 17958), (334270, 17928), (339560, 17950), (344850, 17972), (350140, 17943), (355430, 17964), (360720, 17985), (366010, 17957), (371300, 17977), (376590, 17996), (381880, 18016), (387170, 17989), (392460, 18007), (397750, 18025), (403040, 17999), (408330, 18017), (413620, 18034), (418910, 18051), (424200, 18026), (429490, 18043), (434780, 18059), (440070, 18035), (445360, 18051), (450650, 18066), (455940, 18082), (461230, 18058), (466520, 18073), (471810, 18088), (477100, 18065), (482390, 18079), (487680, 18093), (492970, 18107), (498260, 18085), (503550, 18099), (508840, 18112), (514130, 18091), (519420, 18104), (524710, 18117)]\n",
      "====================\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Probabilistic Checking\n",
    "from plots.probabilistic_checking import build_prob_checking_data, build_fig_pcheck_num_required_checks\n",
    "lines = build_prob_checking_data()\n",
    "\n",
    "fig_pcheck_num_required_checks = build_fig_pcheck_num_required_checks(lines)\n",
    "\n",
    "print(\"Probabilistic Checking:\")\n",
    "display(fig_pcheck_num_required_checks)\n",
    "\n",
    "# print data\n",
    "for line in lines:\n",
    "    print(f\"fail_prob_bound={line['fail_prob_bound']} \\t prob_weights_violating_bound={line['prob_weights_violating_bound']}\")\n",
    "\n",
    "    lst = [(label, value) for label, value in zip(line[\"labels (n_weights)\"], line[\"values (n_checks)\"])]\n",
    "    \n",
    "    print(f\"(n_weights, n_checks): {lst}\")\n",
    "    print(\"====================\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 99/99 [00:41<00:00,  2.37it/s]\n",
      "100%|██████████| 99/99 [04:48<00:00,  2.92s/it]\n"
     ]
    },
    {
     "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>p_c</th>\n",
       "      <th>detection_rate</th>\n",
       "      <th>computed</th>\n",
       "      <th>std</th>\n",
       "      <th>task</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00001</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.00002</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.00003</td>\n",
       "      <td>0.826</td>\n",
       "      <td>0.817539</td>\n",
       "      <td>0.379109</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.00004</td>\n",
       "      <td>0.800</td>\n",
       "      <td>0.817539</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.00005</td>\n",
       "      <td>0.960</td>\n",
       "      <td>0.966711</td>\n",
       "      <td>0.195959</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.00006</td>\n",
       "      <td>0.960</td>\n",
       "      <td>0.966711</td>\n",
       "      <td>0.195959</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.00007</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.993927</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.00008</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.993927</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.00009</td>\n",
       "      <td>0.988</td>\n",
       "      <td>0.993927</td>\n",
       "      <td>0.108885</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.00010</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.998892</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.00011</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.998892</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.00012</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999798</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.00013</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999798</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.00014</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.00015</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.00016</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.00017</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.00018</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.00019</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.00020</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.00021</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.00022</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.00023</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.00024</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.00025</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.00026</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.00027</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.00028</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.00029</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.00030</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.00031</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.00032</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.00033</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.00034</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.00035</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.00036</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>0.00037</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>0.00038</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>0.00039</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>0.00040</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>0.00041</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>0.00042</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>0.00043</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>0.00044</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>0.00045</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>0.00046</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>0.00047</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>0.00048</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>0.00049</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>0.00050</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>0.00051</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>0.00052</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>0.00053</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>0.00054</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>0.00055</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>0.00056</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>0.00057</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>0.00058</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>0.00059</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>0.00060</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>0.00061</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>0.00062</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>0.00063</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>0.00064</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>0.00065</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>0.00066</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>0.00067</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>0.00068</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>0.00069</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>0.00070</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>0.00071</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>0.00072</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>0.00073</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>0.00074</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>0.00075</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>0.00076</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>0.00077</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>0.00078</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>0.00079</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>0.00080</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>0.00081</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>0.00082</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>0.00083</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>0.00084</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>0.00085</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>0.00086</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>0.00087</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>0.00088</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>0.00089</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>0.00090</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>0.00091</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>0.00092</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>0.00093</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>0.00094</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>0.00095</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>0.00096</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>0.00097</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>0.00098</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>0.00099</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>0.00001</td>\n",
       "      <td>0.970</td>\n",
       "      <td>0.967027</td>\n",
       "      <td>0.170587</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.00002</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999803</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>0.00003</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>0.00004</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>0.00005</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>0.00006</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>0.00007</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>0.00008</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>0.00009</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>0.00010</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>0.00011</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110</th>\n",
       "      <td>0.00012</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>0.00013</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>0.00014</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>0.00015</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>0.00016</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>0.00017</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>0.00018</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>0.00019</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>0.00020</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>0.00021</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>0.00022</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121</th>\n",
       "      <td>0.00023</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>0.00024</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>0.00025</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>0.00026</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>0.00027</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126</th>\n",
       "      <td>0.00028</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>0.00029</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>0.00030</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>0.00031</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130</th>\n",
       "      <td>0.00032</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131</th>\n",
       "      <td>0.00033</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>0.00034</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133</th>\n",
       "      <td>0.00035</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134</th>\n",
       "      <td>0.00036</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135</th>\n",
       "      <td>0.00037</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>0.00038</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137</th>\n",
       "      <td>0.00039</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>0.00040</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>0.00041</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>0.00042</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>0.00043</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>0.00044</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>0.00045</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>0.00046</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145</th>\n",
       "      <td>0.00047</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146</th>\n",
       "      <td>0.00048</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>0.00049</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148</th>\n",
       "      <td>0.00050</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149</th>\n",
       "      <td>0.00051</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150</th>\n",
       "      <td>0.00052</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>151</th>\n",
       "      <td>0.00053</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>0.00054</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>153</th>\n",
       "      <td>0.00055</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>154</th>\n",
       "      <td>0.00056</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155</th>\n",
       "      <td>0.00057</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>156</th>\n",
       "      <td>0.00058</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>0.00059</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158</th>\n",
       "      <td>0.00060</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>159</th>\n",
       "      <td>0.00061</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160</th>\n",
       "      <td>0.00062</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161</th>\n",
       "      <td>0.00063</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>162</th>\n",
       "      <td>0.00064</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163</th>\n",
       "      <td>0.00065</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164</th>\n",
       "      <td>0.00066</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>0.00067</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>166</th>\n",
       "      <td>0.00068</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167</th>\n",
       "      <td>0.00069</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>168</th>\n",
       "      <td>0.00070</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>169</th>\n",
       "      <td>0.00071</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170</th>\n",
       "      <td>0.00072</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>171</th>\n",
       "      <td>0.00073</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172</th>\n",
       "      <td>0.00074</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>0.00075</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174</th>\n",
       "      <td>0.00076</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>175</th>\n",
       "      <td>0.00077</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>176</th>\n",
       "      <td>0.00078</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>177</th>\n",
       "      <td>0.00079</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>178</th>\n",
       "      <td>0.00080</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179</th>\n",
       "      <td>0.00081</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>180</th>\n",
       "      <td>0.00082</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181</th>\n",
       "      <td>0.00083</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>182</th>\n",
       "      <td>0.00084</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183</th>\n",
       "      <td>0.00085</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>184</th>\n",
       "      <td>0.00086</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>185</th>\n",
       "      <td>0.00087</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186</th>\n",
       "      <td>0.00088</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>0.00089</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188</th>\n",
       "      <td>0.00090</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>0.00091</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>0.00092</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191</th>\n",
       "      <td>0.00093</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>192</th>\n",
       "      <td>0.00094</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>0.00095</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>0.00096</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>195</th>\n",
       "      <td>0.00097</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196</th>\n",
       "      <td>0.00098</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197</th>\n",
       "      <td>0.00099</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         p_c  detection_rate  computed       std task\n",
       "0    0.00001           0.000  0.000000  0.000000  FMN\n",
       "1    0.00002           0.000  0.000000  0.000000  FMN\n",
       "2    0.00003           0.826  0.817539  0.379109  FMN\n",
       "3    0.00004           0.800  0.817539  0.400000  FMN\n",
       "4    0.00005           0.960  0.966711  0.195959  FMN\n",
       "5    0.00006           0.960  0.966711  0.195959  FMN\n",
       "6    0.00007           0.996  0.993927  0.063119  FMN\n",
       "7    0.00008           0.996  0.993927  0.063119  FMN\n",
       "8    0.00009           0.988  0.993927  0.108885  FMN\n",
       "9    0.00010           0.996  0.998892  0.063119  FMN\n",
       "10   0.00011           1.000  0.998892  0.000000  FMN\n",
       "11   0.00012           1.000  0.999798  0.000000  FMN\n",
       "12   0.00013           1.000  0.999798  0.000000  FMN\n",
       "13   0.00014           1.000  0.999963  0.000000  FMN\n",
       "14   0.00015           1.000  0.999963  0.000000  FMN\n",
       "15   0.00016           1.000  0.999993  0.000000  FMN\n",
       "16   0.00017           1.000  0.999993  0.000000  FMN\n",
       "17   0.00018           1.000  0.999993  0.000000  FMN\n",
       "18   0.00019           1.000  0.999999  0.000000  FMN\n",
       "19   0.00020           1.000  0.999999  0.000000  FMN\n",
       "20   0.00021           1.000  1.000000  0.000000  FMN\n",
       "21   0.00022           1.000  1.000000  0.000000  FMN\n",
       "22   0.00023           1.000  1.000000  0.000000  FMN\n",
       "23   0.00024           1.000  1.000000  0.000000  FMN\n",
       "24   0.00025           1.000  1.000000  0.000000  FMN\n",
       "25   0.00026           1.000  1.000000  0.000000  FMN\n",
       "26   0.00027           1.000  1.000000  0.000000  FMN\n",
       "27   0.00028           1.000  1.000000  0.000000  FMN\n",
       "28   0.00029           1.000  1.000000  0.000000  FMN\n",
       "29   0.00030           1.000  1.000000  0.000000  FMN\n",
       "30   0.00031           1.000  1.000000  0.000000  FMN\n",
       "31   0.00032           1.000  1.000000  0.000000  FMN\n",
       "32   0.00033           1.000  1.000000  0.000000  FMN\n",
       "33   0.00034           1.000  1.000000  0.000000  FMN\n",
       "34   0.00035           1.000  1.000000  0.000000  FMN\n",
       "35   0.00036           1.000  1.000000  0.000000  FMN\n",
       "36   0.00037           1.000  1.000000  0.000000  FMN\n",
       "37   0.00038           1.000  1.000000  0.000000  FMN\n",
       "38   0.00039           1.000  1.000000  0.000000  FMN\n",
       "39   0.00040           1.000  1.000000  0.000000  FMN\n",
       "40   0.00041           1.000  1.000000  0.000000  FMN\n",
       "41   0.00042           1.000  1.000000  0.000000  FMN\n",
       "42   0.00043           1.000  1.000000  0.000000  FMN\n",
       "43   0.00044           1.000  1.000000  0.000000  FMN\n",
       "44   0.00045           1.000  1.000000  0.000000  FMN\n",
       "45   0.00046           1.000  1.000000  0.000000  FMN\n",
       "46   0.00047           1.000  1.000000  0.000000  FMN\n",
       "47   0.00048           1.000  1.000000  0.000000  FMN\n",
       "48   0.00049           1.000  1.000000  0.000000  FMN\n",
       "49   0.00050           1.000  1.000000  0.000000  FMN\n",
       "50   0.00051           1.000  1.000000  0.000000  FMN\n",
       "51   0.00052           1.000  1.000000  0.000000  FMN\n",
       "52   0.00053           1.000  1.000000  0.000000  FMN\n",
       "53   0.00054           1.000  1.000000  0.000000  FMN\n",
       "54   0.00055           1.000  1.000000  0.000000  FMN\n",
       "55   0.00056           1.000  1.000000  0.000000  FMN\n",
       "56   0.00057           1.000  1.000000  0.000000  FMN\n",
       "57   0.00058           1.000  1.000000  0.000000  FMN\n",
       "58   0.00059           1.000  1.000000  0.000000  FMN\n",
       "59   0.00060           1.000  1.000000  0.000000  FMN\n",
       "60   0.00061           1.000  1.000000  0.000000  FMN\n",
       "61   0.00062           1.000  1.000000  0.000000  FMN\n",
       "62   0.00063           1.000  1.000000  0.000000  FMN\n",
       "63   0.00064           1.000  1.000000  0.000000  FMN\n",
       "64   0.00065           1.000  1.000000  0.000000  FMN\n",
       "65   0.00066           1.000  1.000000  0.000000  FMN\n",
       "66   0.00067           1.000  1.000000  0.000000  FMN\n",
       "67   0.00068           1.000  1.000000  0.000000  FMN\n",
       "68   0.00069           1.000  1.000000  0.000000  FMN\n",
       "69   0.00070           1.000  1.000000  0.000000  FMN\n",
       "70   0.00071           1.000  1.000000  0.000000  FMN\n",
       "71   0.00072           1.000  1.000000  0.000000  FMN\n",
       "72   0.00073           1.000  1.000000  0.000000  FMN\n",
       "73   0.00074           1.000  1.000000  0.000000  FMN\n",
       "74   0.00075           1.000  1.000000  0.000000  FMN\n",
       "75   0.00076           1.000  1.000000  0.000000  FMN\n",
       "76   0.00077           1.000  1.000000  0.000000  FMN\n",
       "77   0.00078           1.000  1.000000  0.000000  FMN\n",
       "78   0.00079           1.000  1.000000  0.000000  FMN\n",
       "79   0.00080           1.000  1.000000  0.000000  FMN\n",
       "80   0.00081           1.000  1.000000  0.000000  FMN\n",
       "81   0.00082           1.000  1.000000  0.000000  FMN\n",
       "82   0.00083           1.000  1.000000  0.000000  FMN\n",
       "83   0.00084           1.000  1.000000  0.000000  FMN\n",
       "84   0.00085           1.000  1.000000  0.000000  FMN\n",
       "85   0.00086           1.000  1.000000  0.000000  FMN\n",
       "86   0.00087           1.000  1.000000  0.000000  FMN\n",
       "87   0.00088           1.000  1.000000  0.000000  FMN\n",
       "88   0.00089           1.000  1.000000  0.000000  FMN\n",
       "89   0.00090           1.000  1.000000  0.000000  FMN\n",
       "90   0.00091           1.000  1.000000  0.000000  FMN\n",
       "91   0.00092           1.000  1.000000  0.000000  FMN\n",
       "92   0.00093           1.000  1.000000  0.000000  FMN\n",
       "93   0.00094           1.000  1.000000  0.000000  FMN\n",
       "94   0.00095           1.000  1.000000  0.000000  FMN\n",
       "95   0.00096           1.000  1.000000  0.000000  FMN\n",
       "96   0.00097           1.000  1.000000  0.000000  FMN\n",
       "97   0.00098           1.000  1.000000  0.000000  FMN\n",
       "98   0.00099           1.000  1.000000  0.000000  FMN\n",
       "99   0.00001           0.970  0.967027  0.170587  C10\n",
       "100  0.00002           1.000  0.999803  0.000000  C10\n",
       "101  0.00003           1.000  0.999999  0.000000  C10\n",
       "102  0.00004           1.000  1.000000  0.000000  C10\n",
       "103  0.00005           1.000  1.000000  0.000000  C10\n",
       "104  0.00006           1.000  1.000000  0.000000  C10\n",
       "105  0.00007           1.000  1.000000  0.000000  C10\n",
       "106  0.00008           1.000  1.000000  0.000000  C10\n",
       "107  0.00009           1.000  1.000000  0.000000  C10\n",
       "108  0.00010           1.000  1.000000  0.000000  C10\n",
       "109  0.00011           1.000  1.000000  0.000000  C10\n",
       "110  0.00012           1.000  1.000000  0.000000  C10\n",
       "111  0.00013           1.000  1.000000  0.000000  C10\n",
       "112  0.00014           1.000  1.000000  0.000000  C10\n",
       "113  0.00015           1.000  1.000000  0.000000  C10\n",
       "114  0.00016           1.000  1.000000  0.000000  C10\n",
       "115  0.00017           1.000  1.000000  0.000000  C10\n",
       "116  0.00018           1.000  1.000000  0.000000  C10\n",
       "117  0.00019           1.000  1.000000  0.000000  C10\n",
       "118  0.00020           1.000  1.000000  0.000000  C10\n",
       "119  0.00021           1.000  1.000000  0.000000  C10\n",
       "120  0.00022           1.000  1.000000  0.000000  C10\n",
       "121  0.00023           1.000  1.000000  0.000000  C10\n",
       "122  0.00024           1.000  1.000000  0.000000  C10\n",
       "123  0.00025           1.000  1.000000  0.000000  C10\n",
       "124  0.00026           1.000  1.000000  0.000000  C10\n",
       "125  0.00027           1.000  1.000000  0.000000  C10\n",
       "126  0.00028           1.000  1.000000  0.000000  C10\n",
       "127  0.00029           1.000  1.000000  0.000000  C10\n",
       "128  0.00030           1.000  1.000000  0.000000  C10\n",
       "129  0.00031           1.000  1.000000  0.000000  C10\n",
       "130  0.00032           1.000  1.000000  0.000000  C10\n",
       "131  0.00033           1.000  1.000000  0.000000  C10\n",
       "132  0.00034           1.000  1.000000  0.000000  C10\n",
       "133  0.00035           1.000  1.000000  0.000000  C10\n",
       "134  0.00036           1.000  1.000000  0.000000  C10\n",
       "135  0.00037           1.000  1.000000  0.000000  C10\n",
       "136  0.00038           1.000  1.000000  0.000000  C10\n",
       "137  0.00039           1.000  1.000000  0.000000  C10\n",
       "138  0.00040           1.000  1.000000  0.000000  C10\n",
       "139  0.00041           1.000  1.000000  0.000000  C10\n",
       "140  0.00042           1.000  1.000000  0.000000  C10\n",
       "141  0.00043           1.000  1.000000  0.000000  C10\n",
       "142  0.00044           1.000  1.000000  0.000000  C10\n",
       "143  0.00045           1.000  1.000000  0.000000  C10\n",
       "144  0.00046           1.000  1.000000  0.000000  C10\n",
       "145  0.00047           1.000  1.000000  0.000000  C10\n",
       "146  0.00048           1.000  1.000000  0.000000  C10\n",
       "147  0.00049           1.000  1.000000  0.000000  C10\n",
       "148  0.00050           1.000  1.000000  0.000000  C10\n",
       "149  0.00051           1.000  1.000000  0.000000  C10\n",
       "150  0.00052           1.000  1.000000  0.000000  C10\n",
       "151  0.00053           1.000  1.000000  0.000000  C10\n",
       "152  0.00054           1.000  1.000000  0.000000  C10\n",
       "153  0.00055           1.000  1.000000  0.000000  C10\n",
       "154  0.00056           1.000  1.000000  0.000000  C10\n",
       "155  0.00057           1.000  1.000000  0.000000  C10\n",
       "156  0.00058           1.000  1.000000  0.000000  C10\n",
       "157  0.00059           1.000  1.000000  0.000000  C10\n",
       "158  0.00060           1.000  1.000000  0.000000  C10\n",
       "159  0.00061           1.000  1.000000  0.000000  C10\n",
       "160  0.00062           1.000  1.000000  0.000000  C10\n",
       "161  0.00063           1.000  1.000000  0.000000  C10\n",
       "162  0.00064           1.000  1.000000  0.000000  C10\n",
       "163  0.00065           1.000  1.000000  0.000000  C10\n",
       "164  0.00066           1.000  1.000000  0.000000  C10\n",
       "165  0.00067           1.000  1.000000  0.000000  C10\n",
       "166  0.00068           1.000  1.000000  0.000000  C10\n",
       "167  0.00069           1.000  1.000000  0.000000  C10\n",
       "168  0.00070           1.000  1.000000  0.000000  C10\n",
       "169  0.00071           1.000  1.000000  0.000000  C10\n",
       "170  0.00072           1.000  1.000000  0.000000  C10\n",
       "171  0.00073           1.000  1.000000  0.000000  C10\n",
       "172  0.00074           1.000  1.000000  0.000000  C10\n",
       "173  0.00075           1.000  1.000000  0.000000  C10\n",
       "174  0.00076           1.000  1.000000  0.000000  C10\n",
       "175  0.00077           1.000  1.000000  0.000000  C10\n",
       "176  0.00078           1.000  1.000000  0.000000  C10\n",
       "177  0.00079           1.000  1.000000  0.000000  C10\n",
       "178  0.00080           1.000  1.000000  0.000000  C10\n",
       "179  0.00081           1.000  1.000000  0.000000  C10\n",
       "180  0.00082           1.000  1.000000  0.000000  C10\n",
       "181  0.00083           1.000  1.000000  0.000000  C10\n",
       "182  0.00084           1.000  1.000000  0.000000  C10\n",
       "183  0.00085           1.000  1.000000  0.000000  C10\n",
       "184  0.00086           1.000  1.000000  0.000000  C10\n",
       "185  0.00087           1.000  1.000000  0.000000  C10\n",
       "186  0.00088           1.000  1.000000  0.000000  C10\n",
       "187  0.00089           1.000  1.000000  0.000000  C10\n",
       "188  0.00090           1.000  1.000000  0.000000  C10\n",
       "189  0.00091           1.000  1.000000  0.000000  C10\n",
       "190  0.00092           1.000  1.000000  0.000000  C10\n",
       "191  0.00093           1.000  1.000000  0.000000  C10\n",
       "192  0.00094           1.000  1.000000  0.000000  C10\n",
       "193  0.00095           1.000  1.000000  0.000000  C10\n",
       "194  0.00096           1.000  1.000000  0.000000  C10\n",
       "195  0.00097           1.000  1.000000  0.000000  C10\n",
       "196  0.00098           1.000  1.000000  0.000000  C10\n",
       "197  0.00099           1.000  1.000000  0.000000  C10"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tqdm import tqdm\n",
    "\n",
    "from plots.probabilistic_checking import detection_prop_hypergeometric\n",
    "\n",
    "df = pd.DataFrame(columns=['p_c', 'detection_rate'])\n",
    "num_samples = 500\n",
    "\n",
    "for task, mal_update_size, mal_update_outside_bound in zip([\"FMN\", \"C10\"], [44426, 274442], [36320, 224607]):\n",
    "\n",
    "    mal_update = np.zeros([mal_update_size], dtype=np.bool)\n",
    "    mal_update[:mal_update_outside_bound] = True\n",
    "    mal_update = np.random.permutation(mal_update)\n",
    "\n",
    "    p_cs = np.arange(0.00001, 0.001, 0.00001)\n",
    "    # print(p_cs)\n",
    "\n",
    "    for p_c in tqdm(p_cs,):\n",
    "        # we select p_c at random\n",
    "        num_select = int(mal_update_size * p_c)\n",
    "    #     print(f\"Selecting {num_select}\")\n",
    "        hits = 0\n",
    "        computed = detection_prop_hypergeometric(mal_update_size, mal_update_outside_bound, num_select)\n",
    "    #     print(computed)\n",
    "        for i in range(num_samples):\n",
    "            selection_indices = np.random.choice(mal_update_size, num_select, replace=False)\n",
    "            if np.any(mal_update[selection_indices]) == True:\n",
    "                hits += 1\n",
    "\n",
    "        std_array = np.zeros(num_samples)\n",
    "        std_array[:hits] = 1\n",
    "        std = np.std(std_array)\n",
    "        df = df.append({\n",
    "            'p_c': p_c,\n",
    "            'detection_rate': float(hits) / float(num_samples),\n",
    "            'std': std,\n",
    "            'computed': computed,\n",
    "            'task': task\n",
    "        }, ignore_index=True)\n",
    "\n",
    "display(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plots.e2e_time import setup_plt\n",
    "from matplotlib.backends.backend_pdf import PdfPages\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from common import output_dir, get_colorful_styles\n",
    "setup_plt()\n",
    "\n",
    "name=\"pcheck_real_vector_detection\"\n",
    "colors, _ = get_colorful_styles()\n",
    "\n",
    "# compute where filter probability is at least 1e-08\n",
    "task_colors = {\n",
    "    \"FMN\": colors[1],\n",
    "    \"C10\": colors[2]\n",
    "}\n",
    "\n",
    "\n",
    "# display(df)\n",
    "# display(error_rate1e8)\n",
    "with PdfPages(f\"{output_dir}/{name}.pdf\") as pdf:\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "\n",
    "    ##########################\n",
    "    # Draw all the lines\n",
    "    ##########################\n",
    "\n",
    "    # lineplot\n",
    "    for task in [\"FMN\", \"C10\"]:\n",
    "        df_t = df[df[\"task\"] == task]\n",
    "#         plt.errorbar(df_t[\"p_c\"], df_t[\"detection_rate\"], yerr=df_t[\"std\"]/10.0, label=f\"{task}-P\", color=task_colors[task])\n",
    "        plt.plot(df_t[\"p_c\"], df_t[\"computed\"], label=f\"{task}-P\", color=colors[3])\n",
    "\n",
    "    \n",
    "        target = 1-1e-08\n",
    "        error_rate1e8 = df_t[df_t[\"computed\"] >= target].nsmallest(1, \"computed\")\n",
    "\n",
    "        target = 1-1e-06\n",
    "        error_rate1e6 = df_t[df_t[\"computed\"] >= target].nsmallest(1, \"computed\")\n",
    "        \n",
    "        label = None\n",
    "        if task == \"FMN\":\n",
    "            plt.scatter(error_rate1e8[\"p_c\"], error_rate1e8[\"computed\"], label=\"$\\\\delta=1e-08$\", color=colors[3])\n",
    "            plt.scatter(error_rate1e6[\"p_c\"], error_rate1e6[\"computed\"], label=\"$\\\\delta=1e-06$\", color=colors[4])\n",
    "        else:\n",
    "            plt.scatter(error_rate1e8[\"p_c\"], error_rate1e8[\"computed\"], label=None, color=colors[3])\n",
    "            plt.scatter(error_rate1e6[\"p_c\"], error_rate1e6[\"computed\"], label=None, color=colors[4])\n",
    "                \n",
    "\n",
    "    ##########################\n",
    "    # General Format\n",
    "    ##########################\n",
    "    #ax.set_title(\"Hello World\")\n",
    "    legend_properties = {'weight':'bold'}\n",
    "    ax.legend(loc=\"lower right\", prop=legend_properties)   # 'best', 'upper right', 'upper left', 'lower left',\n",
    "    # 'lower right', 'right', 'center left',  'center right',\n",
    "    # 'lower center', 'upper center', 'center'\n",
    "    ax.grid(True, axis=\"y\", linestyle=':', color='0.6', zorder=0, linewidth=1.2)\n",
    "\n",
    "    ##########################\n",
    "    # Y - Axis Format\n",
    "    ##########################\n",
    "    ax.set_ylim(ymin=0, ymax=1.05)\n",
    "    ax.set_ylabel(\"Detection rate\")\n",
    "    #ax.set_yticks(yticks)\n",
    "#     ylabels = [f\"{round(y)}k\" for y in ax.get_yticks()]\n",
    "#     ax.set_yticklabels(ylabels)\n",
    "\n",
    "    ##########################\n",
    "    # X - Axis Format\n",
    "    ##########################\n",
    "    ax.set_xlim(xmin=0, xmax=0.0005)\n",
    "    ax.set_xlabel(\"Percentage of checks $p_c$\")\n",
    "    #ax.set_xticks(xticks)\n",
    "    xlabels = [f\"{round(x, 2)}%\" for x in ax.get_xticks()*100]\n",
    "    ax.set_xticklabels(xlabels)\n",
    "\n",
    "    pdf.savefig(bbox_inches='tight', pad_inches=0)\n",
    "    plt.close()\n",
    "\n",
    "    display(fig)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [00:44<00:00,  2.22it/s]\n",
      "100%|██████████| 100/100 [05:20<00:00,  3.21s/it]\n"
     ]
    },
    {
     "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>p_c</th>\n",
       "      <th>detection_rate</th>\n",
       "      <th>computed</th>\n",
       "      <th>std</th>\n",
       "      <th>task</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0001</td>\n",
       "      <td>0.078</td>\n",
       "      <td>0.077675</td>\n",
       "      <td>0.268172</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.226</td>\n",
       "      <td>0.231115</td>\n",
       "      <td>0.418239</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0005</td>\n",
       "      <td>0.378</td>\n",
       "      <td>0.359053</td>\n",
       "      <td>0.484888</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0007</td>\n",
       "      <td>0.452</td>\n",
       "      <td>0.465722</td>\n",
       "      <td>0.497691</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0009</td>\n",
       "      <td>0.556</td>\n",
       "      <td>0.554656</td>\n",
       "      <td>0.496854</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0011</td>\n",
       "      <td>0.620</td>\n",
       "      <td>0.628800</td>\n",
       "      <td>0.485386</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0013</td>\n",
       "      <td>0.714</td>\n",
       "      <td>0.690611</td>\n",
       "      <td>0.451889</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.0015</td>\n",
       "      <td>0.766</td>\n",
       "      <td>0.742140</td>\n",
       "      <td>0.423372</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.0017</td>\n",
       "      <td>0.786</td>\n",
       "      <td>0.785094</td>\n",
       "      <td>0.410127</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.0019</td>\n",
       "      <td>0.780</td>\n",
       "      <td>0.817236</td>\n",
       "      <td>0.414246</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.0021</td>\n",
       "      <td>0.846</td>\n",
       "      <td>0.847691</td>\n",
       "      <td>0.360949</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.0023</td>\n",
       "      <td>0.860</td>\n",
       "      <td>0.873077</td>\n",
       "      <td>0.346987</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.0025</td>\n",
       "      <td>0.890</td>\n",
       "      <td>0.894235</td>\n",
       "      <td>0.312890</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.0027</td>\n",
       "      <td>0.900</td>\n",
       "      <td>0.911869</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.0029</td>\n",
       "      <td>0.936</td>\n",
       "      <td>0.926566</td>\n",
       "      <td>0.244753</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.0031</td>\n",
       "      <td>0.928</td>\n",
       "      <td>0.938815</td>\n",
       "      <td>0.258488</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.0033</td>\n",
       "      <td>0.946</td>\n",
       "      <td>0.949022</td>\n",
       "      <td>0.226018</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.0035</td>\n",
       "      <td>0.946</td>\n",
       "      <td>0.956658</td>\n",
       "      <td>0.226018</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.0037</td>\n",
       "      <td>0.970</td>\n",
       "      <td>0.963891</td>\n",
       "      <td>0.170587</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.0039</td>\n",
       "      <td>0.968</td>\n",
       "      <td>0.969918</td>\n",
       "      <td>0.176000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.0041</td>\n",
       "      <td>0.970</td>\n",
       "      <td>0.974940</td>\n",
       "      <td>0.170587</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.0043</td>\n",
       "      <td>0.976</td>\n",
       "      <td>0.979125</td>\n",
       "      <td>0.153049</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>0.0045</td>\n",
       "      <td>0.982</td>\n",
       "      <td>0.982611</td>\n",
       "      <td>0.132951</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>0.0047</td>\n",
       "      <td>0.982</td>\n",
       "      <td>0.985516</td>\n",
       "      <td>0.132951</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>0.0049</td>\n",
       "      <td>0.992</td>\n",
       "      <td>0.987936</td>\n",
       "      <td>0.089084</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>0.0051</td>\n",
       "      <td>0.990</td>\n",
       "      <td>0.989952</td>\n",
       "      <td>0.099499</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>0.0053</td>\n",
       "      <td>0.986</td>\n",
       "      <td>0.991459</td>\n",
       "      <td>0.117490</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>0.0055</td>\n",
       "      <td>0.994</td>\n",
       "      <td>0.992887</td>\n",
       "      <td>0.077227</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.0057</td>\n",
       "      <td>0.990</td>\n",
       "      <td>0.994076</td>\n",
       "      <td>0.099499</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.0059</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.995067</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.0061</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.995892</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>0.0063</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.996579</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>0.0065</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.997152</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>0.0067</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.997628</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.0069</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.998025</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.0071</td>\n",
       "      <td>0.994</td>\n",
       "      <td>0.998322</td>\n",
       "      <td>0.077227</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>0.0073</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.998603</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>0.0075</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.998837</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>0.0077</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.999032</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>0.0079</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999194</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>0.0081</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999329</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>0.0083</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999441</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>0.0085</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999535</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>0.0087</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999613</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>0.0089</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999671</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>0.0091</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999726</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>0.0093</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999772</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>0.0095</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999811</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>0.0097</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999842</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>0.0099</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999869</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>0.0100</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999879</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>0.0104</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999916</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>0.0108</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999942</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>0.0112</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999960</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>0.0116</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999972</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>0.0120</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999980</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>0.0124</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999986</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>0.0128</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>0.0132</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>0.0136</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999995</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>0.0140</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999997</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>0.0144</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>0.0148</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>0.0152</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>0.0156</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>0.0160</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>0.0164</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>0.0168</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>0.0172</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>0.0176</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>0.0180</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>0.0184</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>0.0188</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>0.0192</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>0.0196</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>0.0200</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>0.0204</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>0.0208</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>0.0212</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>0.0216</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>0.0220</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>0.0224</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>0.0228</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>0.0232</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>0.0236</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>0.0240</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>0.0244</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>0.0248</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>0.0252</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>0.0256</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>0.0260</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>0.0264</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>0.0268</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>0.0272</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>0.0276</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>0.0280</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>0.0284</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>0.0288</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>0.0292</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>0.0296</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>FMN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>0.0001</td>\n",
       "      <td>0.812</td>\n",
       "      <td>0.811897</td>\n",
       "      <td>0.390712</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>0.0003</td>\n",
       "      <td>0.996</td>\n",
       "      <td>0.993747</td>\n",
       "      <td>0.063119</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>0.0005</td>\n",
       "      <td>0.998</td>\n",
       "      <td>0.999792</td>\n",
       "      <td>0.044677</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>0.0007</td>\n",
       "      <td>1.000</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>0.0009</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>0.0011</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>0.0013</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>0.0015</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>0.0017</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>0.0019</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110</th>\n",
       "      <td>0.0021</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>0.0023</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>0.0025</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>0.0027</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>0.0029</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>0.0031</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>0.0033</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>0.0035</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>0.0037</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>0.0039</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>0.0041</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121</th>\n",
       "      <td>0.0043</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>0.0045</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>0.0047</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>0.0049</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>0.0051</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126</th>\n",
       "      <td>0.0053</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>0.0055</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>0.0057</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>0.0059</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130</th>\n",
       "      <td>0.0061</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131</th>\n",
       "      <td>0.0063</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>0.0065</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133</th>\n",
       "      <td>0.0067</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134</th>\n",
       "      <td>0.0069</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135</th>\n",
       "      <td>0.0071</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>0.0073</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137</th>\n",
       "      <td>0.0075</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>0.0077</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>0.0079</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>0.0081</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>0.0083</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>0.0085</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>0.0087</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>0.0089</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145</th>\n",
       "      <td>0.0091</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146</th>\n",
       "      <td>0.0093</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>0.0095</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148</th>\n",
       "      <td>0.0097</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149</th>\n",
       "      <td>0.0099</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150</th>\n",
       "      <td>0.0100</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>151</th>\n",
       "      <td>0.0104</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>0.0108</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>153</th>\n",
       "      <td>0.0112</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>154</th>\n",
       "      <td>0.0116</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155</th>\n",
       "      <td>0.0120</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>156</th>\n",
       "      <td>0.0124</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>0.0128</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158</th>\n",
       "      <td>0.0132</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>159</th>\n",
       "      <td>0.0136</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160</th>\n",
       "      <td>0.0140</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161</th>\n",
       "      <td>0.0144</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>162</th>\n",
       "      <td>0.0148</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163</th>\n",
       "      <td>0.0152</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164</th>\n",
       "      <td>0.0156</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>0.0160</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>166</th>\n",
       "      <td>0.0164</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167</th>\n",
       "      <td>0.0168</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>168</th>\n",
       "      <td>0.0172</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>169</th>\n",
       "      <td>0.0176</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>170</th>\n",
       "      <td>0.0180</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>171</th>\n",
       "      <td>0.0184</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>172</th>\n",
       "      <td>0.0188</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>173</th>\n",
       "      <td>0.0192</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174</th>\n",
       "      <td>0.0196</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>175</th>\n",
       "      <td>0.0200</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>176</th>\n",
       "      <td>0.0204</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>177</th>\n",
       "      <td>0.0208</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>178</th>\n",
       "      <td>0.0212</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>179</th>\n",
       "      <td>0.0216</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>180</th>\n",
       "      <td>0.0220</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181</th>\n",
       "      <td>0.0224</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>182</th>\n",
       "      <td>0.0228</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183</th>\n",
       "      <td>0.0232</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>184</th>\n",
       "      <td>0.0236</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>185</th>\n",
       "      <td>0.0240</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186</th>\n",
       "      <td>0.0244</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>0.0248</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188</th>\n",
       "      <td>0.0252</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>0.0256</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>0.0260</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191</th>\n",
       "      <td>0.0264</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>192</th>\n",
       "      <td>0.0268</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>0.0272</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>0.0276</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>195</th>\n",
       "      <td>0.0280</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>196</th>\n",
       "      <td>0.0284</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197</th>\n",
       "      <td>0.0288</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198</th>\n",
       "      <td>0.0292</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199</th>\n",
       "      <td>0.0296</td>\n",
       "      <td>1.000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>C10</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        p_c  detection_rate  computed       std task\n",
       "0    0.0001           0.078  0.077675  0.268172  FMN\n",
       "1    0.0003           0.226  0.231115  0.418239  FMN\n",
       "2    0.0005           0.378  0.359053  0.484888  FMN\n",
       "3    0.0007           0.452  0.465722  0.497691  FMN\n",
       "4    0.0009           0.556  0.554656  0.496854  FMN\n",
       "5    0.0011           0.620  0.628800  0.485386  FMN\n",
       "6    0.0013           0.714  0.690611  0.451889  FMN\n",
       "7    0.0015           0.766  0.742140  0.423372  FMN\n",
       "8    0.0017           0.786  0.785094  0.410127  FMN\n",
       "9    0.0019           0.780  0.817236  0.414246  FMN\n",
       "10   0.0021           0.846  0.847691  0.360949  FMN\n",
       "11   0.0023           0.860  0.873077  0.346987  FMN\n",
       "12   0.0025           0.890  0.894235  0.312890  FMN\n",
       "13   0.0027           0.900  0.911869  0.300000  FMN\n",
       "14   0.0029           0.936  0.926566  0.244753  FMN\n",
       "15   0.0031           0.928  0.938815  0.258488  FMN\n",
       "16   0.0033           0.946  0.949022  0.226018  FMN\n",
       "17   0.0035           0.946  0.956658  0.226018  FMN\n",
       "18   0.0037           0.970  0.963891  0.170587  FMN\n",
       "19   0.0039           0.968  0.969918  0.176000  FMN\n",
       "20   0.0041           0.970  0.974940  0.170587  FMN\n",
       "21   0.0043           0.976  0.979125  0.153049  FMN\n",
       "22   0.0045           0.982  0.982611  0.132951  FMN\n",
       "23   0.0047           0.982  0.985516  0.132951  FMN\n",
       "24   0.0049           0.992  0.987936  0.089084  FMN\n",
       "25   0.0051           0.990  0.989952  0.099499  FMN\n",
       "26   0.0053           0.986  0.991459  0.117490  FMN\n",
       "27   0.0055           0.994  0.992887  0.077227  FMN\n",
       "28   0.0057           0.990  0.994076  0.099499  FMN\n",
       "29   0.0059           0.998  0.995067  0.044677  FMN\n",
       "30   0.0061           0.996  0.995892  0.063119  FMN\n",
       "31   0.0063           0.996  0.996579  0.063119  FMN\n",
       "32   0.0065           0.998  0.997152  0.044677  FMN\n",
       "33   0.0067           0.998  0.997628  0.044677  FMN\n",
       "34   0.0069           0.998  0.998025  0.044677  FMN\n",
       "35   0.0071           0.994  0.998322  0.077227  FMN\n",
       "36   0.0073           0.998  0.998603  0.044677  FMN\n",
       "37   0.0075           1.000  0.998837  0.000000  FMN\n",
       "38   0.0077           0.998  0.999032  0.044677  FMN\n",
       "39   0.0079           1.000  0.999194  0.000000  FMN\n",
       "40   0.0081           1.000  0.999329  0.000000  FMN\n",
       "41   0.0083           1.000  0.999441  0.000000  FMN\n",
       "42   0.0085           1.000  0.999535  0.000000  FMN\n",
       "43   0.0087           1.000  0.999613  0.000000  FMN\n",
       "44   0.0089           1.000  0.999671  0.000000  FMN\n",
       "45   0.0091           1.000  0.999726  0.000000  FMN\n",
       "46   0.0093           1.000  0.999772  0.000000  FMN\n",
       "47   0.0095           1.000  0.999811  0.000000  FMN\n",
       "48   0.0097           1.000  0.999842  0.000000  FMN\n",
       "49   0.0099           1.000  0.999869  0.000000  FMN\n",
       "50   0.0100           1.000  0.999879  0.000000  FMN\n",
       "51   0.0104           1.000  0.999916  0.000000  FMN\n",
       "52   0.0108           1.000  0.999942  0.000000  FMN\n",
       "53   0.0112           1.000  0.999960  0.000000  FMN\n",
       "54   0.0116           1.000  0.999972  0.000000  FMN\n",
       "55   0.0120           1.000  0.999980  0.000000  FMN\n",
       "56   0.0124           1.000  0.999986  0.000000  FMN\n",
       "57   0.0128           1.000  0.999991  0.000000  FMN\n",
       "58   0.0132           1.000  0.999993  0.000000  FMN\n",
       "59   0.0136           1.000  0.999995  0.000000  FMN\n",
       "60   0.0140           1.000  0.999997  0.000000  FMN\n",
       "61   0.0144           1.000  0.999998  0.000000  FMN\n",
       "62   0.0148           1.000  0.999998  0.000000  FMN\n",
       "63   0.0152           1.000  0.999999  0.000000  FMN\n",
       "64   0.0156           1.000  0.999999  0.000000  FMN\n",
       "65   0.0160           1.000  0.999999  0.000000  FMN\n",
       "66   0.0164           1.000  1.000000  0.000000  FMN\n",
       "67   0.0168           1.000  1.000000  0.000000  FMN\n",
       "68   0.0172           1.000  1.000000  0.000000  FMN\n",
       "69   0.0176           1.000  1.000000  0.000000  FMN\n",
       "70   0.0180           1.000  1.000000  0.000000  FMN\n",
       "71   0.0184           1.000  1.000000  0.000000  FMN\n",
       "72   0.0188           1.000  1.000000  0.000000  FMN\n",
       "73   0.0192           1.000  1.000000  0.000000  FMN\n",
       "74   0.0196           1.000  1.000000  0.000000  FMN\n",
       "75   0.0200           1.000  1.000000  0.000000  FMN\n",
       "76   0.0204           1.000  1.000000  0.000000  FMN\n",
       "77   0.0208           1.000  1.000000  0.000000  FMN\n",
       "78   0.0212           1.000  1.000000  0.000000  FMN\n",
       "79   0.0216           1.000  1.000000  0.000000  FMN\n",
       "80   0.0220           1.000  1.000000  0.000000  FMN\n",
       "81   0.0224           1.000  1.000000  0.000000  FMN\n",
       "82   0.0228           1.000  1.000000  0.000000  FMN\n",
       "83   0.0232           1.000  1.000000  0.000000  FMN\n",
       "84   0.0236           1.000  1.000000  0.000000  FMN\n",
       "85   0.0240           1.000  1.000000  0.000000  FMN\n",
       "86   0.0244           1.000  1.000000  0.000000  FMN\n",
       "87   0.0248           1.000  1.000000  0.000000  FMN\n",
       "88   0.0252           1.000  1.000000  0.000000  FMN\n",
       "89   0.0256           1.000  1.000000  0.000000  FMN\n",
       "90   0.0260           1.000  1.000000  0.000000  FMN\n",
       "91   0.0264           1.000  1.000000  0.000000  FMN\n",
       "92   0.0268           1.000  1.000000  0.000000  FMN\n",
       "93   0.0272           1.000  1.000000  0.000000  FMN\n",
       "94   0.0276           1.000  1.000000  0.000000  FMN\n",
       "95   0.0280           1.000  1.000000  0.000000  FMN\n",
       "96   0.0284           1.000  1.000000  0.000000  FMN\n",
       "97   0.0288           1.000  1.000000  0.000000  FMN\n",
       "98   0.0292           1.000  1.000000  0.000000  FMN\n",
       "99   0.0296           1.000  1.000000  0.000000  FMN\n",
       "100  0.0001           0.812  0.811897  0.390712  C10\n",
       "101  0.0003           0.996  0.993747  0.063119  C10\n",
       "102  0.0005           0.998  0.999792  0.044677  C10\n",
       "103  0.0007           1.000  0.999993  0.000000  C10\n",
       "104  0.0009           1.000  1.000000  0.000000  C10\n",
       "105  0.0011           1.000  1.000000  0.000000  C10\n",
       "106  0.0013           1.000  1.000000  0.000000  C10\n",
       "107  0.0015           1.000  1.000000  0.000000  C10\n",
       "108  0.0017           1.000  1.000000  0.000000  C10\n",
       "109  0.0019           1.000  1.000000  0.000000  C10\n",
       "110  0.0021           1.000  1.000000  0.000000  C10\n",
       "111  0.0023           1.000  1.000000  0.000000  C10\n",
       "112  0.0025           1.000  1.000000  0.000000  C10\n",
       "113  0.0027           1.000  1.000000  0.000000  C10\n",
       "114  0.0029           1.000  1.000000  0.000000  C10\n",
       "115  0.0031           1.000  1.000000  0.000000  C10\n",
       "116  0.0033           1.000  1.000000  0.000000  C10\n",
       "117  0.0035           1.000  1.000000  0.000000  C10\n",
       "118  0.0037           1.000  1.000000  0.000000  C10\n",
       "119  0.0039           1.000  1.000000  0.000000  C10\n",
       "120  0.0041           1.000  1.000000  0.000000  C10\n",
       "121  0.0043           1.000  1.000000  0.000000  C10\n",
       "122  0.0045           1.000  1.000000  0.000000  C10\n",
       "123  0.0047           1.000  1.000000  0.000000  C10\n",
       "124  0.0049           1.000  1.000000  0.000000  C10\n",
       "125  0.0051           1.000  1.000000  0.000000  C10\n",
       "126  0.0053           1.000  1.000000  0.000000  C10\n",
       "127  0.0055           1.000  1.000000  0.000000  C10\n",
       "128  0.0057           1.000  1.000000  0.000000  C10\n",
       "129  0.0059           1.000  1.000000  0.000000  C10\n",
       "130  0.0061           1.000  1.000000  0.000000  C10\n",
       "131  0.0063           1.000  1.000000  0.000000  C10\n",
       "132  0.0065           1.000  1.000000  0.000000  C10\n",
       "133  0.0067           1.000  1.000000  0.000000  C10\n",
       "134  0.0069           1.000  1.000000  0.000000  C10\n",
       "135  0.0071           1.000  1.000000  0.000000  C10\n",
       "136  0.0073           1.000  1.000000  0.000000  C10\n",
       "137  0.0075           1.000  1.000000  0.000000  C10\n",
       "138  0.0077           1.000  1.000000  0.000000  C10\n",
       "139  0.0079           1.000  1.000000  0.000000  C10\n",
       "140  0.0081           1.000  1.000000  0.000000  C10\n",
       "141  0.0083           1.000  1.000000  0.000000  C10\n",
       "142  0.0085           1.000  1.000000  0.000000  C10\n",
       "143  0.0087           1.000  1.000000  0.000000  C10\n",
       "144  0.0089           1.000  1.000000  0.000000  C10\n",
       "145  0.0091           1.000  1.000000  0.000000  C10\n",
       "146  0.0093           1.000  1.000000  0.000000  C10\n",
       "147  0.0095           1.000  1.000000  0.000000  C10\n",
       "148  0.0097           1.000  1.000000  0.000000  C10\n",
       "149  0.0099           1.000  1.000000  0.000000  C10\n",
       "150  0.0100           1.000  1.000000  0.000000  C10\n",
       "151  0.0104           1.000  1.000000  0.000000  C10\n",
       "152  0.0108           1.000  1.000000  0.000000  C10\n",
       "153  0.0112           1.000  1.000000  0.000000  C10\n",
       "154  0.0116           1.000  1.000000  0.000000  C10\n",
       "155  0.0120           1.000  1.000000  0.000000  C10\n",
       "156  0.0124           1.000  1.000000  0.000000  C10\n",
       "157  0.0128           1.000  1.000000  0.000000  C10\n",
       "158  0.0132           1.000  1.000000  0.000000  C10\n",
       "159  0.0136           1.000  1.000000  0.000000  C10\n",
       "160  0.0140           1.000  1.000000  0.000000  C10\n",
       "161  0.0144           1.000  1.000000  0.000000  C10\n",
       "162  0.0148           1.000  1.000000  0.000000  C10\n",
       "163  0.0152           1.000  1.000000  0.000000  C10\n",
       "164  0.0156           1.000  1.000000  0.000000  C10\n",
       "165  0.0160           1.000  1.000000  0.000000  C10\n",
       "166  0.0164           1.000  1.000000  0.000000  C10\n",
       "167  0.0168           1.000  1.000000  0.000000  C10\n",
       "168  0.0172           1.000  1.000000  0.000000  C10\n",
       "169  0.0176           1.000  1.000000  0.000000  C10\n",
       "170  0.0180           1.000  1.000000  0.000000  C10\n",
       "171  0.0184           1.000  1.000000  0.000000  C10\n",
       "172  0.0188           1.000  1.000000  0.000000  C10\n",
       "173  0.0192           1.000  1.000000  0.000000  C10\n",
       "174  0.0196           1.000  1.000000  0.000000  C10\n",
       "175  0.0200           1.000  1.000000  0.000000  C10\n",
       "176  0.0204           1.000  1.000000  0.000000  C10\n",
       "177  0.0208           1.000  1.000000  0.000000  C10\n",
       "178  0.0212           1.000  1.000000  0.000000  C10\n",
       "179  0.0216           1.000  1.000000  0.000000  C10\n",
       "180  0.0220           1.000  1.000000  0.000000  C10\n",
       "181  0.0224           1.000  1.000000  0.000000  C10\n",
       "182  0.0228           1.000  1.000000  0.000000  C10\n",
       "183  0.0232           1.000  1.000000  0.000000  C10\n",
       "184  0.0236           1.000  1.000000  0.000000  C10\n",
       "185  0.0240           1.000  1.000000  0.000000  C10\n",
       "186  0.0244           1.000  1.000000  0.000000  C10\n",
       "187  0.0248           1.000  1.000000  0.000000  C10\n",
       "188  0.0252           1.000  1.000000  0.000000  C10\n",
       "189  0.0256           1.000  1.000000  0.000000  C10\n",
       "190  0.0260           1.000  1.000000  0.000000  C10\n",
       "191  0.0264           1.000  1.000000  0.000000  C10\n",
       "192  0.0268           1.000  1.000000  0.000000  C10\n",
       "193  0.0272           1.000  1.000000  0.000000  C10\n",
       "194  0.0276           1.000  1.000000  0.000000  C10\n",
       "195  0.0280           1.000  1.000000  0.000000  C10\n",
       "196  0.0284           1.000  1.000000  0.000000  C10\n",
       "197  0.0288           1.000  1.000000  0.000000  C10\n",
       "198  0.0292           1.000  1.000000  0.000000  C10\n",
       "199  0.0296           1.000  1.000000  0.000000  C10"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tqdm import tqdm\n",
    "\n",
    "from plots.probabilistic_checking import detection_prop_hypergeometric\n",
    "\n",
    "df = pd.DataFrame(columns=['p_c', 'detection_rate'])\n",
    "num_samples = 500\n",
    "\n",
    "for task, mal_update_size, mal_update_outside_bound in zip([\"FMN\", \"C10\"], [44426, 274442], [889, 16467]):\n",
    "\n",
    "    mal_update = np.zeros([mal_update_size], dtype=np.bool)\n",
    "    mal_update[:mal_update_outside_bound] = True\n",
    "    mal_update = np.random.permutation(mal_update)\n",
    "\n",
    "    p_1 = np.arange(0.0001, 0.01, 0.0002)\n",
    "    p_cs = np.arange(0.01, 0.03, 0.0004)\n",
    "    p_cs = np.concatenate([p_1, p_cs])\n",
    "#     print(p_cs)\n",
    "    # print(p_cs)\n",
    "\n",
    "    for p_c in tqdm(p_cs,):\n",
    "        # we select p_c at random\n",
    "        num_select = int(round(mal_update_size * p_c))\n",
    "    #     print(f\"Selecting {num_select}\")\n",
    "        hits = 0\n",
    "        computed = detection_prop_hypergeometric(mal_update_size, mal_update_outside_bound, num_select)\n",
    "    #     print(computed)\n",
    "        for i in range(num_samples):\n",
    "            selection_indices = np.random.choice(mal_update_size, num_select, replace=False)\n",
    "            if np.any(mal_update[selection_indices]) == True:\n",
    "                hits += 1\n",
    "\n",
    "        std_array = np.zeros(num_samples)\n",
    "        std_array[:hits] = 1\n",
    "        std = np.std(std_array)\n",
    "        df = df.append({\n",
    "            'p_c': p_c,\n",
    "            'detection_rate': float(hits) / float(num_samples),\n",
    "            'std': std,\n",
    "            'computed': computed,\n",
    "            'task': task\n",
    "        }, ignore_index=True)\n",
    "\n",
    "display(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/hidde/opt/anaconda3/envs/tf37_standard/lib/python3.7/site-packages/ipykernel_launcher.py:49: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 438.356x216.735 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plots.e2e_time import setup_plt\n",
    "from matplotlib.backends.backend_pdf import PdfPages\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from common import output_dir, get_colorful_styles\n",
    "setup_plt()\n",
    "\n",
    "name=\"pcheck_real_vector_detection\"\n",
    "colors, _ = get_colorful_styles()\n",
    "\n",
    "# compute where filter probability is at least 1e-08\n",
    "task_colors = {\n",
    "    \"FMN\": colors[1],\n",
    "    \"C10\": colors[2]\n",
    "}\n",
    "\n",
    "def binomial_ci(x, n, alpha=0.05):\n",
    "    #x is number of successes, n is number of trials\n",
    "    from scipy import stats\n",
    "    if x==0:\n",
    "        c1 = 0\n",
    "    else:\n",
    "        c1 = stats.beta.interval(1-alpha, x,n-x+1)[0]\n",
    "    if x==n:\n",
    "        c2=1\n",
    "    else:\n",
    "        c2 = stats.beta.interval(1-alpha, x+1,n-x)[1]\n",
    "    return c1, c2\n",
    "\n",
    "def apply_bin(r):\n",
    "    c1, c2 = binomial_ci(r * num_samples, num_samples)\n",
    "    return abs(c1 - r), c2 - r\n",
    "\n",
    "# display(df)\n",
    "# display(error_rate1e8)\n",
    "with PdfPages(f\"{output_dir}/{name}.pdf\") as pdf:\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "\n",
    "    ##########################\n",
    "    # Draw all the lines\n",
    "    ##########################\n",
    "\n",
    "    # lineplot\n",
    "    for task in [\"FMN\", \"C10\"]:\n",
    "        \n",
    "        df_t = df[df[\"task\"] == task]\n",
    "        \n",
    "        df_t[\"yerr\"] = df_t[\"detection_rate\"].apply(apply_bin)\n",
    "        yerr = df_t[\"yerr\"].to_numpy()\n",
    "        yerr = np.array([[x, y] for x, y in yerr]).T\n",
    "\n",
    "        plt.errorbar(df_t[\"p_c\"], df_t[\"detection_rate\"], \n",
    "                     yerr=yerr,\n",
    "#                      yerr=df_t[\"std\"]/10.0, \n",
    "                     label=f\"{task}-P\", color=task_colors[task])\n",
    "#         plt.plot(df_t[\"p_c\"], df_t[\"computed\"], label=f\"{task}-P\", color=colors[3])\n",
    "\n",
    "    \n",
    "        target = 1-1e-08\n",
    "        error_rate1e8 = df_t[df_t[\"computed\"] >= target].nsmallest(1, \"computed\")\n",
    "\n",
    "        target = 1-1e-06\n",
    "        error_rate1e6 = df_t[df_t[\"computed\"] >= target].nsmallest(1, \"computed\")\n",
    "        \n",
    "#         label = None\n",
    "#         if task == \"FMN\":\n",
    "#             plt.scatter(error_rate1e8[\"p_c\"], error_rate1e8[\"computed\"], label=\"$\\\\delta=1e-08$\", color=colors[3])\n",
    "#             plt.scatter(error_rate1e6[\"p_c\"], error_rate1e6[\"computed\"], label=\"$\\\\delta=1e-06$\", color=colors[4])\n",
    "#         else:\n",
    "#             plt.scatter(error_rate1e8[\"p_c\"], error_rate1e8[\"computed\"], label=None, color=colors[3])\n",
    "#             plt.scatter(error_rate1e6[\"p_c\"], error_rate1e6[\"computed\"], label=None, color=colors[4])\n",
    "                \n",
    "\n",
    "    ##########################\n",
    "    # General Format\n",
    "    ##########################\n",
    "    #ax.set_title(\"Hello World\")\n",
    "    legend_properties = {'weight':'bold'}\n",
    "    handles, labels = ax.get_legend_handles_labels()\n",
    "    ax.legend(handles[::-1], labels[::-1], loc=\"lower right\", prop=legend_properties)   # 'best', 'upper right', 'upper left', 'lower left',\n",
    "    # 'lower right', 'right', 'center left',  'center right',\n",
    "    # 'lower center', 'upper center', 'center'\n",
    "\n",
    "\n",
    "    ax.grid(True, axis=\"y\", linestyle=':', color='0.6', zorder=0, linewidth=1.2)\n",
    "\n",
    "    ##########################\n",
    "    # Y - Axis Format\n",
    "    ##########################\n",
    "    ax.set_ylim(ymin=0, ymax=1.05)\n",
    "    ax.set_ylabel(\"Detection rate\")\n",
    "    #ax.set_yticks(yticks)\n",
    "#     ylabels = [f\"{round(y)}k\" for y in ax.get_yticks()]\n",
    "#     ax.set_yticklabels(ylabels)\n",
    "\n",
    "    ##########################\n",
    "    # X - Axis Format\n",
    "    ##########################\n",
    "    ax.set_xlim(xmin=0, xmax=0.01)\n",
    "    ax.set_xlabel(\"Percentage of checks $p_c$\")\n",
    "    #ax.set_xticks(xticks)\n",
    "    xlabels = [f\"{round(x, 2)}%\" for x in ax.get_xticks()*100]\n",
    "    ax.set_xticklabels(xlabels)\n",
    "\n",
    "    pdf.savefig(bbox_inches='tight', pad_inches=0)\n",
    "    plt.close()\n",
    "\n",
    "    display(fig)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}